Nelson - NIS, A Comparative Analysis

National Innovation Systems
This page intentionally left blank

National
Innovation
Systems
A Comparative Analysis
Edited by
RICHARD R. NELSON
New York Oxford

OXFORD UNIVERSITY PRESS
1993
Oxford University Press
Oxford New York Toronto
Delhi Bombay Calcutta Madras Karachi
Kuala Lumpur Singapore Hong Kong Tokyo
Nairobi Dar es Salaam Cape Town
Melbourne Auckland Madrid
and associated companies in
Berlin Ibadan
Copyright 1993 by Oxford University Press, Inc.

Published by Oxford University Press, Inc.,
198 Madison Avenue, New York, New York 10016-4314
Oxford is a registered trademark of Oxford University Press
All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system, or transmitted, in any form or by any means,
electronic, mechanical, photocopying, recording, or otherwise,
without the prior permission of Oxford University Press.
Library of Congress Cataloging-in-Publication Data

National innovation systems: a comparative analysis /
edited by Richard R. Nelson.
p. cm. Includes bibliographical references and index.
ISBN 0-19-507616-8 ISBN 0-19-507617-6 (pbk.)
1. Technological innovations. 2. Technology and state.
I. Nelson, Richard R.
T173.8.N36 1993 338.9'26dc20 92-342
23456789
Printed in the United States of America
on acid-free paper
Acknowledgments
We, the authors of the chapters in this book and the members of the steering com-
mittee, are indebted to many parties for making this project possible. We would
like to thank the American Enterprise Institute, Columbia University's Center on
Japanese Economy and Business, and the German Marshall Fund of the United
States for providing the basic funding that enabled us to start the project and carry
it through. Funds provided by the Sloan Foundation to the Consortium on Com-
petition and Cooperation proved invaluable when, at the end, our expenses out-
ran our basic funding.
Our basic strategy for writing a book together was to meet together several
times so that all of us could discuss chapter drafts and help each other toward
something that was coherent, as contrasted with being a collection of freestanding
country essays. These meetings were hosted by Columbia University's School of
International and Public Affairs, by the Maastrict Economic Research Institute
on Innovation and Technology, by the Science Policy Research Unit of the Uni-
versity of Sussex, and by Stanford University's Center for Economic Policy
Research. Many thanks to our hosts and conference organizers.
Finally we, the participants in this project, would like to thank each other. It
was a fascinating and rewarding venture for us. Each of us learned much from the
others. We together hope that you, the readers, will learn as much as we did about
a fascinating set of issues.
Contents
Contributors, ix
1. Technical Innovation and National Systems, 3
Richard R. Nelson and Nathan Rosenberg
Part I Large High-Income Countries, 23

2. The U.S. National Innovation System, 29
David C. Mowery and Nathan Rosenberg
3. The Japanese System of Innovation: Past, Present, and Future, 76
Hiroyuki Odagiri and Akira Goto
4. The National System for Technical Innovation in Germany, 115
Otto Keck
5. National Innovation Systems: Britain, 158
William Walker
6. The French National System of Innovation, 192
Francois Chesnais
7. The National System of Innovation: Italy, 230
Franco Malerba
Part II Smaller High-Income Countries, 261

8. Comparing the Danish and Swedish Systems of Innovation, 265
Charles Edquist and Bengt-Ake Lundvall
9. The Canadian System of Industrial Innovation, 299
Donald G. McFetridge
10. The Australian Innovation System, 324
Robert G. Gregory
Part III Lower Income Countries, 353

11. National System of Industrial Innovation: Dynamics of Capability Building
in Korea, 357
Linsu Kim
viii CONTENTS
12. National Systems Supporting Technical Advance in Industry:

The Case of Taiwan, 384
Chi-Ming Hou and San Gee
13. National Systems Supporting Technical Advance in Industry:
The Brazilian Experience, 414
Carl J. Dalhman and Claudio R. Frischtak
14. National Systems of Innovation Supporting Technical Advance in Industry:
The Case of Argentina, 451
Jorge M. Katz and Nestor A. Bercovich
15. The Innovation System of Israel: Description, Performance, and
Outstanding Issues, 476
Morris Teubal
Part IV National Innovation Systems, 503

16. A Retrospective, 505
Richard R. Nelson
Index, 525
Contributors
Nestor A. Bercovich Professor Jorge Katz

National Council for Scientific and Economic Commission
Technological Research Latin America and the Caribbean
Francois Chesnais Professor Otto Keck

Directorate for Science Technology Department of Political Science
and Industry Free University of Berlin
OECD
Professor Linsu Kim
Carl Dahlman College of Business Administration
Industrial Development Division Korea University
World Bank
Bengt-Ake Lundvall
Professor Charles Edquist Institut for Produktion
Department of Technology and Social Aalburg Universitet Center
Change
Linkoping University Professor Franco Malerba
Universita Commerciale
L. Bocconi
Claudio Frischtak
Industrial Development Division
Professor Donald McFetridge
World Bank
Department of Economics
Carleton University
Professor Akira Goto
Department of Economics Professor David Mowery
Hitosubashi University Center for Research Management
University of California, Berkeley
Professor Robert Gregory
Department of Economics Professor Richard Nelson
Australia National University School of International and Public
Affairs
Chi-Ming Hou Columbia University
Director, Institute of International
Economy Professor Hiroyuki Odagiri
Chung-hun Institute for Economic Institute of Socio-Economic Planning
Research University of Tsukuba
X CONTRIBUTORS
Professor Nathan Rosenberg Professor Morris Teubal

Department of Economics Department of Economics
Stanford University Hebrew University
Dr. Gee San Professor William Walker

Research Fellow and Deputy Director Science Policy Research Unit
Institute of International Economy University of Sussex
Chung-Hua Institute for Economic
Research
Steering Committee
for National Innovation Project
Claude Barfield Professor Nathan Rosenberg

American Enterprise Institute Department of Economics
Stanford University
Professor Giovanni Dosi
University of Rome Professor Jon Sigurdson
Institute for Policy Science
Professor Christopher Freeman Saitama University
Science Policy Research Unit
University of Sussex Professor Luc Soete
Professor Hugh Patrick University of Limburg
Graduate School of Business
Columbia University
Professor David Teece
Professor Keith Pavitt Haas School of Business
Science Policy Research Unit University of California, Berkeley
University of Sussex
National Innovation Systems
1
Technical Innovation and National Systems
RICHARD R. NELSON
NATHAN ROSENBERG
WHAT IS THIS STUDY ABOUT
This book is about national systems of technical innovation. The heart of the work
consists of studies of 15 countries, including the large market-oriented industrialized
ones, several smaller high-income countries, and a number of newly industrializing
states. The studies have been carefully designed, developed, and written to illuminate
the institutions and mechanisms supporting technical innovation in the various coun-
tries, the similarities and differences across countries and how these came to be, and
to permit at least preliminary discussion of how the differences matter.
The book has been written more despite than because of the recent great interest
in the topic considered. The slowdown of growth since the early 1970s in all of the
advanced industrial nations, the rise of Japan as a major economic and technological
power, the relative decline of the United States, and widespread concerns in Europe
about being behind both have led to a rash of writing and policy concerned with sup-
porting the technical innovative prowess of national firms. At the same time the
enhanced technical sophistication of Korea, Taiwan, and other NICs has broadened
the range of nations whose firms are competitive players in fields that used to be the
preserve of only a few, and has led other nations who today have a weak manufacturing
sector to wonder how they might emulate the performance of the successful NICs.
There clearly is a new spirit of what might be called "technonationalism" in the air,
combining a strong belief that the technological capabilities of a nation's firms are a
key source of their competitive prowess, with a belief that these capabilities are in a
sense national, and can be built by national action.
It is this climate that has given rise to the current strong interest in national inno-
vation systems, and their similarities and differences, and in the extent and manner
that these differences explain variation in national economic performance. There may
now be more awareness and research on such national differences than on any other
area where comparative institutional analysis would seem interesting and illu-
minating.
3
4 NATIONAL INNOVATION SYSTEMS
The project that led to this book was born of the current strong interest in national
innovation systems,' and came out of a belief on the part of the participants that much
of the writing and argument on this subject has been somewhat hyped and rather hap-
hazard. In addition, many of the allegedly comparative studies concentrated on one
countryin recent times Japanwith comparisons with other countries mainly
implied. The actual comparative studies tended to involve only two or a very small
group of countries. This fact is serious in view of the absence of a well-articulated and
verified analytic framework linking institutional arrangements to technological and
economic performance. In the absence of such a framework, there were (and are) only
weak constraints on the inclinations of analysts to draw possibly spurious causal links
between differences in institutional structures that clearly are there, and differences in
performance that clearly are there also. Different authors have focused on different
things, and made different arguments about why some feature was an important factor
behind strong or weak performance. The broadening of the set of countries considered
simultaneously seemed an important way to tighten these constraints by enlarging the
number of "points" that a causal theory had to "fit."
The way we have been putting the matter clearly signals that the orientation of
this project has been to carefully describe and compare, and try to understand, rather
than to theorize first and then attempt to prove or calibrate the theory. However, a
comparative study such as this requires, at least, some agreement on basic terms and
concepts.
There is, first, the concept of a national innovation system itself. Each of the terms
can be interpreted in a variety of ways, and there is the question of whether, in a world
in which technology and business are increasingly transnational, the concept as a
whole makes much sense. Consider the term "innovation." In this study we interpret
the term rather broadly, to encompass the processes by which firms master and get into
practice product designs and manufacturing processes that are new to them, if not to
the universe or even to the nation. We do so for several reasons. First, the activities
and investments associated with becoming the leader in the introduction of a new
product or process, and those associated with staying near the head of the pack, or
catching up, are much less sharply distinguishable than is commonly presumed. More-
over, the strictly Schumpeterian innovator, the first firm to bring a new product to
market, is frequently not the firm that ultimately captures most of the economic rents
associated with the innovation. Second, much of the interest in innovative capability
is tied to concern about economic performance, and here it is certainly the broader
concept rather than the narrower one (the determinants of being first) that matters.
This means that our orientation is not limited to the behavior of firms at the forefront
of world's technology, or to institutions doing the most advanced scientific research,
although in some countries the focus is here, but is more broadly on the factors influ-
encing national technological capabilities.
Then there is the term "system." Although to some the word connotes something
that is consciously designed and built, this is far from the orientation here. Rather the
concept is of a set of institutions whose interactions determine the innovative perfor-
mance, in the sense above, of national firms. There is no presumption that the system
was, in some sense, consciously designed, or even that the set of institutions involved
works together smoothly and coherently. Rather, the "systems" concept is that of a set
of institutional actors that, together, plays the major role in influencing innovative per-
TECHNICAL INNOVATION AND NATIONAL SYSTEMS 5
formance. The broad concept of innovation that we have adopted has forced us to
consider much more than simply the actors doing research and development. Indeed,
a problem with the broader definition of innovation is that it provides no sharp guide
to just what should be included in the innovation system, and what can be left out.
More on this later.
Finally, there is the concept of "national" system. On the one hand, the concept
may be too broad. The system of institutions supporting technical innovation in one
field, say Pharmaceuticals, may have very little overlap with the system of institutions
supporting innovations in another field, say aircraft. On the other hand, in many fields
of technology, including both Pharmaceuticals and aircraft, a number of the institu-
tions are or act transnational. Indeed, for many of the participants in this study, one
of the key interests was in exploring whether, and if so in what ways, the concept of a
"national" system made any sense today. National governments act as if it did. How-
ever, that presumption, and the reality, may not be aligned.
The studies in this project are unified by at least broad agreement on the defini-
tional and conceptual issues discussed above. They were also guided by certain com-
mon understandings of the way technical advance proceeds, and the key processes and
institutional actors involved, that are now widely shared among scholars of technical
advance. In a way these understandings do provide a common analytic framework,
not wide enough to encompass all of the variables and relationships that likely are
important, not sharp enough to tightly guide empirical work, but broad enough and
pointed enough to provide a common structure in which one can have some confi-
dence. This basic common structure is discussed in the following section.
Technical Advance: An Overview of the Processes and Institutions Involved
To understand national innovation systems, it is essential to understand how technical

advance occurs in the modern world, and the key processes and institutions involved.
This section aims to provide such an overview.
We begin by describing two essential, and to some extent conflicting, aspects of
technical advance. In the modern era most technologies are associated with various
fields of science that illuminate them, and lend invaluable understanding and tech-
nique to efforts to advance technology. At the same time efforts at innovation almost
always involve a large element of trial and error and try again learning. Next we turn
to the institutions involved in industrial innovation, starting with firms, and then con-
sidering supporting institutions such as universities and government agencies and pol-
icies. We also point to some important intersectoral and intercountry differences.
Finally we consider the concept of "national systems" and flag some issues about the
relevance of national borders in a world in which business and technology are increas-
ingly transnational.
THE INTERTWINING OF SCIENCE AND TECHNOLOGY
Today, R&D facilities, staffed by university trained scientists and engineers attached
to business firms, universities, or government agencies, are the principal vehicles
through which technological advance proceeds, in fields such as electrical equipment
and systems, chemical products and processes, and aviation. Most contemporary
examinations of national capabilities in technology focus on these kinds of institutions
and industries. One of the principal messages of this chapter is that this view of tech-
nical advance and its sources is somewhat too narrow. However, it certainly is a central
part of the picture.
An important characteristic of technical advance in all of the above fields, and in
many other areas, is that it is strongly supported by various fields of science. However,
the connections between science and technology are complex and vary in certain
respects from field to field in certain essential respects.
Science as Leader and Follower
It is widely believed that new science gives rise to new technology. Although we shall
argue that this is at best an oversimplification, the statement is quite true in regard to
the rise of the electrical equipment industries. The very existence of these industries is
inseparable from the history of theoretical and experimental physics in the nineteenth
century. The emergence of electricity as a new source of power, and the wide range of
new products that came to be built on itincandescent light, telephone, gramo-
phonewere the legitimate offspring of a scientific research enterprise that began with
Faraday's demonstration of electromagnetic induction in 1831. Several decades later
Maxwell's research opened up vast new vistas that led to Hertz's confirmation, in
1887, of the existence of radio waves and the possibility of detecting them at a distance,
and then to modern radio and television (Cohen, 1948 and Dunsheath, 1962).
Thus, the discovery of radio waves, which gave birth to radio and television,
occurred not because scientists were searching for useful applications of their research.
Rather, Hertz was pursuing a logic internal to the discipline of science itself, exploring
the implications of an earlier theory by searching for empirical evidence that might
confirm the theory. Hertz searched forand foundradio waves because an earlier
scientific theory had predicted their existence. Establishing their existence led to the
work of Marconi and others in wireless communication.
In contrast with the electrical equipment industries, the industries producing
chemical products, or using chemical reactions in the manufacture of other products,
long antedated the rise of modern science. Some, such as tanning and dyeing and brew-
ing, are almost as old as civilization itself.
However, in the last four or five decades of the nineteenth century, a systematic
body of scientific knowledge about chemistry grew up that laid a new basis for chem-
ical-based innovation. Chemistry became a laboratory discipline in which research
could be carried out by trained professionals making use of well-understood methods
and experimental procedures. In the 1860s Kekule managed to disentangle the molec-
ular structure of benzene, a breakthrough of major significance for thousands ofaro-
matics, including dyes and drugs, but ultimately, for all of organic chemistry. These
new understandings were invaluable in enabling scientists in industry to search con-
structively for new chemical compounds and to devise better production processes
(Beer, 1959).
These stories about advances in physics and chemistry as scientific disciplines
appear to show scientific development as autonomous, evolving according to an inter-
nal logic of its own, with technology being illuminated as a by-product. But appear-
ances are deceiving. Faraday followed the tradition of his mentor, Humphrey Davy,
in having a strong interest in practical devices and a belief in the value of science in
inventing. Modern chemistry grew out of the ancient discipline of alchemy, which was
concerned with finding ways to transform base materials into valuable ones.
The advent of new technologies often leads to scientific work aimed at under-
standing these technologies, so as to enable them to be improved. Sometimes new
technology leads to whole new scientific disciplines. Sadi Carnot's work in the early
part of the nineteenth century, which led to the new field of thermodynamics, was
largely motivated by theoretical interest in the workings of the steam engine (Cardwell,
1971). The science of polymer chemistry emerged in the twentieth century, in large
part resulting from research, performed inside industrial laboratories, to develop
materials that could better fulfill the changing requirements of industry.
In addition, the rise of the modern chemical industry led to the rise of a new dis-
cipline expressly aimed to service its needschemical engineering. Chemical engi-
neering involved not simply the practical application of the science of chemistry, but
the merger of chemistry and mechanical engineering. Somewhat more precisely,
chemical engineering involved the application of mechanical engineering skills and
methods to the specialized task of producing chemical products on a large scale.
A modern chemical process plant is not a scaled-up version of the laboratory glass
tubes and reactors in which scientific discoveries were originally made. Such scaling-
up is neither technically nor economically feasible. Rather, entirely different processes
have to be invented. The transition from the glass equipment with which W. H.
Carothers produced the first polymers in the DuPont laboratories, or the transition
from the laboratory synthesis of polyethylene, or terephthalic acid, to the large-scale
commercial manufacture of such products, is a transition that involved years of seri-
ous development effort and significant further inventive activity. Indeed, the com-
plexities of the transition are so great that chemical engineers have devised a unique
transitional technology for it, the pilot plant.
The rise of scientific understanding supporting aircraft design reflects a similar
story. Again the technology, or a primitive version of it, came first, and the "science"
or engineering discipline developed to support it. Thus the frail apparatus that the
Wright brothers managed to get airborne for a few seconds in 1903 had very little well-
understood "science" behind its design. However, the promise of those early flying
machines gave rise to the modern disciplines of aerodynamics and aeronautical engi-
neering.
Thus saying that new technologies have given rise to new sciences is at least as
true as the other way around. And it is more on the mark to say that with the rise of
modern science-based technologies, much of science and much of technology have
become intertwined. This is the principal reason why, in the present era, technology is
advanced largely through the work of men and women who have university training
in science of engineering. This intertwining, rather than serendipity, is the principal
reason why, in many fields, university research is an important contributor to tech-
nical advance, and universities as well as corporate labs are essential parts of the inno-
vation system.
Thus, the problems, or observations, that originate in industry are explored not
only by industrial scientists. They feed into, and stimulate, the entire scientific com-
munity. Edison's attempts, in 1883, to improve the incandescent lamp led him to
observe the flow of electricity inside the lamp, across a gap separating a hot filament
from a metal wire. Edison had observed the existence of electrons before their exis-
tence had even been postulated. Although Edison did not appreciate the significance
of his observations at the time, the Edison Effect formed the basis for much twentieth-
century science, including atomic physics and the numerous electronics technologies.
The Limits of Science, Learning by Trying, and Cumulative Incremental

Technological Advance
It is insufficiently appreciated that successful innovation in high technology industries

often is not so much a matter of invention, as a patent examiner would define inven-
tion, as it is a matter of design, in the sense of trying to devise a product or process that
will achieve a desirable cluster of performance characteristics, subject to certain cost
constraints. This engineering design capability is a very sophisticated and costly busi-
ness. McDonnell Douglas recently estimated that the redesign of the wing for a new
wide-bodied jet that would be a successor to the DC-10 would likely cost a billion dol-
lars. Moreover, determining where "design" ends and "research" begins is a matter of
some real difficulty as soon as one deals with relationships that cannot be optimized
by referring to the codified data in the engineering handbooks.
Those aiming for a major design advance almost always are in a position of not
knowing whether a design will work or how well it will until they test it out. In the
chemical industries, a pilot plant may be thought of as an intermediate-scale technol-
ogy, incorporating design principles and mechanical expertise that have no counter-
part in laboratory research. Its ultimate purpose is to increase the confidence in the
technical feasibility and the underlying economics of a larger-scale plant that is both
newly designed and involves a very large financial commitment. Such designing activ-
ity is far from laboratory research in the sense that it cannot be deduced from the find-
ings of that research.
In some essential respects, a new aircraft prototype performs a role that is close
to that of the pilot plant in the chemical processing industries. In both cases there are
significant uncertainties attached to technical designs that incorporate significant ele-
ments of novelty. Since the technical uncertainties readily translate into huge financial
losses if new designs are prematurely introduced into practice, it is prudent to test on
a small scale, and to resolve the expensive uncertainties at the technological frontier
in a less costly rather than more costly fashion. Testing of aircraft prototypes and
chemical pilot plants are specialized modalities for the reduction of technological
uncertainties in innovation.
Through such vehicles as building and testing pilot plants and prototypes, and
testing experimental new drugs, the activities aimed to advance technology generate
new knowledge as well as new products and processes. As we have noted, in many cases
new scientific understanding follows rather than leads, as when the science of aero-
dynamics created theoretical understanding of the factors determining lift and drag,
after the first flying machines had been built and flown, or when William Shockley
developed a theory of holes and electrons in semiconductors to explain how and why
the transistor he and his colleagues at Bell Labs devised actually worked. The new
device or process that works, sort of, or surprisingly well, stimulates both efforts to
explain and understand, and efforts to refine, improve, and variegate.
The Wright Brothers' 1903 machine was scarcely more than a large, ungainly
bicycle with attached wings. (The resemblance to the bicycle was no coincidence, since
the Wright Brothers had previously been designers as well as manufacturers of bicy-
cles.) Their airplane's parts were secured by baling wire and glue, and its total flight
was only a few hundred yards. Not until the 1930s did aircraft shed their struts and
external bracing wire, the non-load-carrying skin involving the use of doped fabrics,
and assume their stressed-skin monocoque construction form. Only with the design
and development of the DC-3 did the airplane finally become a reliable means of trans-
portation on commercial routes (Miller and Sawers, 1970).
But the performance gap that separates the DC-3 of 50 years ago from today's
wide-bodied aircraft, equipped with powerful jet engines, swept-back wings, sophisti-
cated electronics, and capable of flying over most weather turbulence, is also immense.
It almost has to be said of the airplane that everything of economic significance is
attributable to the subsequent improvements, since 1903, that have been made within
the original, crude framework of the Wright Brothers' flying machine.
The point made here with respect to performance improvement of aircraft is, in
fact, a point of broad generality. Most industrial R&D expenditures are on products
that have long been in existencesuch as aircraft, automobiles, and cameras (which
have been in existence fully 150 years). It is these existing products that serve to define
the framework within which improvements can be identified and undertaken. Even
the transistor, which has so drastically transformed the world in the second half of the
twentieth century, has been around for more than 40 years. Its introduction in the late
1940s laid the groundwork for the continuing microelectronics revolution. Yet the
original transistor was a fragile, unreliable, and expensive piece of apparatus. It was
only the subsequent improvements in that original, primitive device that made the
later microelectronics revolution possible.
In this as in other cases, the advance of technology went hand in hand with the
advance of science. The invention of the transistor in 1948 rapidly transformed solid-
state physics from a small subspecialty into the largest subdiscipline of physics. This
was true within the university scientific community as well as within private industry.
Similarly, the advent of the laser in the 1960s, along with the feasibility of using optical
fibers for transmission purposes, led to a great expansion in the science of optics, where
advances in science now offered the prospect of sharply increased economic payoffs.
It is important not to confuse the highly valued autonomy of the individual scientist,
in shaping his or her own research agenda, with the determination of research-funding
agencies to commit resources to those areas of scientific research that appear to offer
the most attractive future returns. Public and private institutions may well be expected
to define future returns rather differently, but neither is likely to be indifferent to the
size of these returns.
THE MAJOR INSTITUTIONAL ACTORS
Because of many misconceptions, it is well to recapitulate just what the rise of science
based technology led to, and what it did not. The rise of science based technology did
lead to a dramatic change in the nature of the people and institutions involved in tech-
nical advance. Through much of the nineteenth century strong formal education in a
science provided an inventor with little or no advantage in problem solving, although
from time to time inventors would consult with scientists. By 1900 formal training in
chemistry was becoming virtually a requirement for successful inventive effort in the
chemical products industries. By 1910 or so the days when unschooled geniuses such
as Thomas Alva Edison could make major advances in the electrical technologies
where coming to an end, and the major electrical companies were busy staffing their
laboratories with university trained scientist and engineers.
Firms and Industrial Research Laboratories
By the beginning of World War I the industrial research laboratory, a facility dedicated
to research and the development of new or improved products and processes, and
staffed by university trained scientists and engineers, had become the principal locus
of technical advance in the chemical and electrical industries, and was beginning to
become important more and more widely. The industrial laboratories were teamed
with universities that trained their new R&D scientists and engineers, and that under-
took research in the new applied sciences and engineering disciplines as well as in the
more traditional basic sciences.
As we have stressed, the rise of science based technologies, and industrial research
laboratories dedicated to "invention," did not lead to routinization of innovation, as
some predicted. R&D continues to be an activity in which dead ends often are reached,
and a lot of trying, testing, and revising is required before a successful result is
achieved.
There are several reasons why the industrial research laboratory, rather than uni-
versity laboratories or government facilities, became the dominant locus of the R&D
part of innovation in most (but not all) fields. First, after a technology has been around
for a period of time, to orient R&D fruitfully, one needs detailed knowledge of its
strengths and weaknesses and areas where improvements would yield big payoffs, and
this knowledge tends to reside with those who use the technology, generally firms and
their customers and suppliers. In addition, over time firms in an industry tend to
develop capabilities for doing certain kinds of R&D that, although drawing on public
scientific knowledge, transcends it, being largely based on practice. Second, profiting
from innovation in many cases requires the integration of activity and planning of
R&D, production, and marketing, which tends to proceed much more effectively
within an organization that itself does all of these. For these reasons, although it is
common to see a significant university or other outsider role in inventing and inno-
vating when a technology is just coming into being (as biotechnology in the early
1980s), the process of cumulative improvement and variegation, which we have
pointed out accounts for the majority of R&D and innovation, tends to be the business
of incumbent firms.
It is important, however, to recognize the lack of distinctness surrounding the
concept of R&D. Partly the matter is one of accounting and nomenclature. Many
small firms engage in significant design and development work, yet do not have a for-
mally designated R&D department or facility; their design and development work
may or may not be accounted and reported as R&D. In many firms process engineer-
ing is located organizationally in production not in R&D; again, the work involved
may or may not be counted as R&D. But the matter is not simply, or basically, one of
nomenclature. The lines between R&D, and other activities, such as designing prod-
ucts for particular customers, problem solving on production processes, or monitoring
a competitor's new products, are inherently blurry. In a number of industries even
firms at the frontier invest significant resources in staying up with relevant develop-
ments elsewhere in science and technology, including prominently the work and
achievements of other firms.
In developing countries, what is an innovation for an indigenous firm may largely
involve learning to produce a product or employ technology that has been employed
for some time by firms in the highly industrialized economies. Learning to make or
use a product may require a considerable amount of study as well as the taking apart
of products and processes to find out how they work, that is, "reverse engineering."
Although generally not so counted, reverse engineering is very much like R&D. As
cases such as the economic development of Korea show, as a company and a country
catch on, such work increasingly reaches out to build something different and begins
to get counted as R&D.
Moreover, even if it is defined quite broadly, R&D usually is only a small part of
the resources and problem solving that go into innovation. The amounts that must be
invested in new equipment and plant to produce a new product, or embody a new
process, generally exceed the R&D costs many fold. New organizations may be called
for, or a different division of work, or new skills on the part of the work force, and new
approaches in marketing. It may take considerable time and effort to get these changes
made and the new system operating smoothly. And, as we have stated earlier, in many
cases innovation is a continuing business, with product and process engineers learning
from experience and making modifications on that basis, customers' feeding back
complaints and suggestions, management learning how to smooth out rough spots,
and so on.
Thus although the chapters that follow will discuss R&D spending, R&D must
be recognized as only part of the larger innovation picture. In many of the chapters
discussion of R&D will be augmented by descriptions and analyses of the different
kinds of firms active in innovation. Thus it has been argued that the style of manage-
ment and organization common in U. S. firms, although once a source of innovative
strength, is now a disadvantage, and that various aspects of Japanese firms help them
to be effective innovators. An important objective of this study is to illuminate and
explore propositions such as these.
Other Institutional Actors
The modern industrial laboratory and the modern research university grew up as com-
panions. The details of this companionship have been considerably different from one
country to another, as later chapters will indicate. In general, however, universities
play an extremely important role in technical advance, not only as places where indus-
trial scientists and engineers are trained, but as the source of research findings and tech-
niques of considerable relevance to technical advance in industry.
Universities in most countries are, first, the places where much of the basic
research in fundamental sciences such as physics is undertaken, although the reliance
on universities as a locus of basic research, as contrasted with national laboratories,
varies across countries. Research in basic sciences such as physics tends to be guided
by the internal logic of the discipline rather than by expectation of particular practical
application. Although occasionally research in basic science will provide understand-
ings or techniques that directly lead to product or process advances, as has been the
case recently in molecular biology, this is not usual. However, even if a nation has only
narrow economic motives in funding university research, it cannot afford to com-
pletely neglect the basic sciences, since training in these is an essential part of training
in the applied sciences and engineering disciplines.
Many fields of academic science are expressly applications orientated. The very
names "material science," "computer science," and "pathology" signal fields of
inquiry closely linked to particular practical problems. So too the engineering disci-
plines, which were expressly established not only to train people for work in industry,
but also to develop the scientific foundations of industrial technologies.
In certain cases, university based institutions have been directly oriented toward
helping a particular industry or other client advance its technologies. Thus the agri-
cultural experimentation stations, attached to the land grant universities of the United
States, are an important source of new technology for farmers, and for a wide range of
agriculture product processing industries. In countries that have a significant phar-
maceutical industry, university faculty in medical schools tend to have close relation-
ships with firms in the industry. In several countries, certain engineering schools have
a mandated responsibility to provide help to firms in the region. In many countries,
including the United States and Germany, universities are the home of institutes
designed to help particular industries.
In almost all nations universities now are funded to a substantial degree by gov-
ernments. However, the organization and means of funding, and orientation of uni-
versity research, differ significantly across nations. It is widely believed that the Amer-
ican university system provides more effective stimulus and help to technical advance
than university systems in most other countries. One purpose of this project is to inves-
tigate this and kindred propositions. There is a related question. To the extent that a
nation's universities support technical advance, how effectively is this support chan-
neled to help national firms? Many observers of the American system have argued that
in many cases foreign firms are benefiting as much as American. To what extent is this
true, or general, and what are the implications?
Recent American complaints, moreover, could easily be duplicated elsewhere.
Great Britain provided a remarkable degree of intellectual leadership in the field of
molecular biology, but most of the commercial exploitation of this research, so far, has
been in the United States. Earlier in this century, in fact until the Second World War,
Germany was undoubtedly the world leader in the sciences of aerodynamics, in large
measure due to Prandtl's great contributions at the University of Gottingen and, later,
at the Kaiser Wilhelm Institute for Fluid Mechanics. Nevertheless, the American com-
mercial aircraft designers and aeronautical engineers benefited immensely by being
able to draw on German aerodynamic research (Hallion, 1977).
Government laboratories also are an important part of many national innovation
systems. In the United States, government laboratories play important roles in, for
example, the fields of agriculture, health, and nuclear energy. However, with only a
few exceptions, agriculture being a major one, in the United States government labo-
ratories are tied to public sector missions in contrast to being established to help civil-
ian industry. And, again with a few exceptions, in the United States the universities,
rather than government laboratories, are seen as the appropriate sites for fundamental
research.
In other countries public laboratories play a significantly larger, or different, role.
Thus the Federal Republic of Germany has a network of Max Planck Institutes dedi-
cated to basic research, and Fraunhofer Laboratories dedicated to applied research.
Much of the latter's work is aimed at helping industry. In France, a substantial share
of government funded basic research is undertaken in public laboratories that are inde-
pendent of universities. This project aims to map out some of these salient differences.
Today public monies support not only research at universities and government
laboratories, but R&D in industry. For the most part government support of industrial
R&D in the United States has been limited to projects of direct governmental interest,
principally those involving military and space needs. But as the new Sematech venture
shows, even the United States is not adverse to using public monies to help the devel-
opment of industrial technologies that are principally of civilian use, if a strong "public
interest" argument can be articulated. In many other countries, simply helping an
important industry has been accepted as a suitable governmental mission. These dif-
ferences are an important subject of this comparative study.
As we have already stressed, innovation involves much more than R&D, and the
set of institutions that influences the technological capabilities of a nation and how
these are advanced extends far beyond those that directly impinge on innovation. The
character and effectiveness of a nation's system of schooling, training, and retraining
not only determine the supply of skills from engineer to machine tender, but also influ-
ence the attitudes of workers toward technical advance. So too do the patterns of
labormanagement bargaining and negotiation, dispute resolution, and the degree of
mutual commitment of firm and workers. Financial institutions, and the way firms
are governed and controlled, profoundly influence the technical activities that are fea-
sible and that managers choose to undertake.
More generally, it is somewhat artificial to try to describe and analyze a nation's
innovation system as something separable from its economic system more broadly
defined, or to depict the policies concerned with innovation as quite apart from those
concerned with the economy, education, or national security. The descriptions of
national innovation systems that will be presented in the following chapters, although
concentrating on the institutions and mechanisms discussed above, will inevitably in
some cases go well beyond these.
Interindustry Differences
Some characteristics of technical advance and the principal institutional actors are
quite general, holding across a wide range of industries. However, there are important
interindustry differences in the nature of technical change, the sources, and how the
involved actors are connected to each other, and it is useful to sketch some of these
here. Nations differ in the mix of industries and these differences alone strongly influ-
ence the shapes of national innovation systems.
A number of industries produce products that can be characterized as complex
systems. Much of electrical technology is of this sort. An aircraft is a complex system.
The size of the system, in terms of the number of critical components, can be very
large, as in jet aircraft or telecommunication systems, or relatively small, as in a tele-
vision set or an automatic loom. In either case technical advance in system-product

technologies tends to proceed through a combination of improvements in compo-
nents, and modifications in overall system design to take advantage of or drive these,
punctuated from time to time by the introduction of a significantly new system.
Technical advance in such fields generally stems from the work of component
and material producers, as well as systems designers. In general the larger the system,
the greater the role of the component producers. However, the effective incorporation
of better components into a system often requires significant R&D work by system
assemblers. The integration of component and systems R&D generally involves some
combination of independent initiatives mediated by the market, contracting, and
express cooperation, with the balance differing from industry to industry and from
country to country. In some cases government programs may facilitate coordination.
In some systems technologies users of the system play a major role in inducing
technical advance, and they may directly support it. Thus the major airlines will often
engage in extensive discussion with airframe manufacturers regarding the modifica-
tions they desire, and what they will pay for. In turn, this may lead to extended dis-
cussions between the airplane producer and the designers and producers of engines.
In the case of a drastic new aircraft design, such as the Boeing 747 or 767, the
manufacturer may not even proceed to the advanced design stage until it has substan-
tial purchase commitments from airline buyers. A firm such as Boeing, moreover,
functions as a designer and assembler of new aircrafts, but is dependent on thousands
of outside component suppliers. Similarly, IBM can also be described as a designer and
assembler of complex, system-like products, and is dependent on a large number of
outside component suppliers. In both the Boeing and IBM cases, components sup-
pliers, like ultimate product buyers, may be in many different countries. In both cases,
also, components suppliers may perform significant amounts of R&D. In the case of
IBM computers, where the number of ultimate buyers is very large, such buyers are
less intimately involved in shaping the details of product design than in the case of
aircraft, where the number of buyers is far smaller.
Technical advance in the industries producing fine chemical products, from syn-
thetic materials to pharmaceutical, is different in a number of respects. First, in these
industries innovation largely involves the introduction of discrete new products or
product classes, such as nylon or valium, that are not in general subject to the contin-
uing incremental improvement that marks systems technologies. However, they may
be subject to tailoring to fit the needs of different kinds of customers, or for new uses.
Indeed a striking aspect of technical advance in these fields is the discovery of new uses
by producers or users. Second, since the products do not involve complex systems,
input suppliers in general do not play a big role; however, process equipment suppliers
may. Indeed in some cases new products may require major process innovation. Pro-
cess equipment suppliers often do the key design work here. In many cases chemical
product companies and equipment suppliers may cooperate. And although product
innovation may be discrete in these industries, process innovation after the initial
design may be continuous and incremental as described earlier.
Still other industries produce bulk commodities, from steel to milk. Here product
innovation is minimal and technical change basically involves new or improved ways
to produce or process the product in question. Equipment and input suppliers and
processing firms often are the major sources of innovation. Where product producers
are large they may do a considerable amount of process R&D on their own or contract
for it. Where the product suppliers are small, they may bond together to get work done
on production processes or modes of processing products, or on standards for inputs
and outputs. Governmental agencies may organize and fund such work. Under such
arrangements close and durable modes of interaction may form between producer
cooperatives and the input suppliers, and processing firms.
TECHNOLOGICAL COMMUNITIES, BOUNDARIES, AND NATIONAL

INNOVATION SYSTEMS
Technological advance proceeds through the interaction of many actors. Above we

have considered some of the key interactions involved, between component and sys-
tems producers, upstream and downstream firms, universities and industry, and gov-
ernment agencies and universities and industries.
The important interactions, the networks, are not the same in all industries or
technologies. We have pointed to differences between systems technologies and chem-
ical product technologies in the nature of interindustry interactions. In some technol-
ogies universities play a key role, for example, Pharmaceuticals and computers. In oth-
ers they play a more modest one, for example, aircraft and steel. Government funding
is important in some industries, such as aircraft and agriculture, and unimportant in
others. But although its shape and character differ, in virtually all fields one must
understand technical advance as proceeding through the work of a community of
actors.
To what extent are there "national" communities? To what degree, and through
what mechanisms, do the individuals and institutions that advance technology divide
up into "national systems?"
We suspect that the answer to this question varies over time and from field to
field. The nineteenth-century history of the evolution of shoe making machinery in
the United States involved a collection of mechanic-inventors, familiar with various
aspects of the shoe making process, and with each other's inventions in the form of
artifacts and products (see Thomson, 1989). The community was national because the
relevant American industries used technology that was different than in Europe,
because of physical proximity, and because of shared language and culture. Studies of
the rise of Bessemer steel technology in the United States (Allen, 1983) and Great Brit-
ain (Morison, 1966) show two different national communities at work, with some
international exchange, but with the bulk of the interaction among nationals. The dye-
stuffs industry grew up as a largely German industry. In this case common training in
German universities, and links between the company laboratories and the universities,
clearly delineated the network. In the United States and Denmark, communities con-
cerned with farming and the processing of farm products grew up associated with gov-
ernment-supported research programs, and cooperatives of various sorts. Military
contracting, a reserved military market, and military R&D money built and protected
distinctly American technological communities in semiconductors, computers, and
jet aircraft in the early days of those technologies.
The discussion above has been of particular fields of technology, where for a vari-
ety of reasons distinctly national communities of actors formed. But what of the prop-
osition that there are national systems in a broader sense, encompassing very wide
ranges of technology, although perhaps specializing and stressing a narrower range?
What might define and delineate these broader national systems?
Certainly the policies and programs of national governments, the laws of a nation,
and the existence of a common language and a shared culture define an inside and
outside that can broadly affect how technical advance proceeds. Put another way,
national differences and boundaries tend to define national innovation systems, partly
intentionally, partly not. Further, general perceptions about national societies and cul-
tures tend to reify national systems.
Thus for the quarter century after World War II both Americans and citizens of
other countries recognized a distinctly American model that had a number of partic-
ular features. In the first place, as writers such as Servan Schreiber (1968) pointed out
in the late 1960s, in many of the key industries American firms were larger than their
European countries, spent more on R&D, and had a distinctive management style.
Furthermore the U. S. government spent much more on industrial R&D than did the
European governments, principally through defense contracts. The U. S. university
research system was stronger. To a very large extent firms situated in America were
American owned, and although overseas branches were becoming increasingly impor-
tant, by far most establishments owned by Americans operated in the United States.
Most of the goods produced by American firms were sold overwhelmingly to the
domestic market. Although foreigners were coming to the U. S. universities in increas-
ing numbers, most of the students being taught and the faculty of American univer-
sities were Americans, and the graduates went to work in American firms. The monies
of the U. S. government were almost exclusively spent in U. S. institutions.
As we shall see, the innovation systems of the major European industrial nations
differed in important aspects from the American. One can argue that the European
systems were much less strictly "national" systems than was the American. For one
thing, even the largest of the European nations was small compared with the United
States, and as a result there was much more importing and exporting as a percentage
of GNP. For another, increasingly over the period there was a sense of European com-
munity that in some ways eroded the significance of national borders and particular
citizenship. Third, and foreshadowing subsequent broader developments, American
firms in European countries were playing a significant role in many industries. How-
ever, as well shall see, in the early postwar era there was a strong sense that although
subject to strong influences from abroad, there was a distinctive and to a large extent
self-contained English, German, and French systems.
Until the 1970s there was no strong competitor to the American system as a broad
model of how an innovation system should be designed. This standing as a model sys-
tem was a natural reflection of the U. S. technological preeminence that marked the
postwar years. Earlier, the United States had been an imitator in many respects. The
American research university that arose early in the twentieth century was consciously
modeled on the German university system, and the R&D organization of American
chemical companies similarly was patterned on a German model. We note that this
proclivity of both private and public institutions to reform themselves toward what
are regarded as the leading models is a strong force toward similarities in national inno-
vation systems.
As European productivity and income levels have caught up with American lev-
els, and Japan has emerged as a leading economic and technological power, the attrac-
tion of the American model has waned, and Japanese institutions have waxed as tar-
gets for emulation. Many strongly held beliefs about the Japanese innovation system
are, at best, only partly correct. Nonetheless, it is widely thought that the Japanese
government targets certain key civilian technologies and orchestrates work on them,
and that this, as well as strong interfirm cooperation, leads to a more powerful and
efficient innovation system. These features currently are the models of fashion in
Europe, and, to an increasing degree, in the United States.
The rise of Japan as a model has enhanced the belief that an explicit national
technology policy can be effective; indeed it is now widely argued that a nation will fall
progressively behind if it does not have an explicit technology policy. This has led over
the past decade to a rash of national programs designed to enable national industries
to stay ahead or catch up technologically. These will be discussed in later chapters.
NATIONAL SYSTEMS AND TRANSNATIONAL TECHNOLOGY
Although the notion that there are distinctly national communities of technologists
fits the evidence in some cases, it does not fit so well in others. Thus the early history
of radio was one of transnational activity, involving inventors and companies in Great
Britain, the United States, and Western Europe, all building on each other's work. The
development of synthetic fibers involved similar transnational interactions. This is not
to argue that the networks of interaction were not more dense within countries or
regions than across borders; however, the latter interactions were sufficiently impor-
tant that the idea that there were in these fields nearly disjoined technological com-
munities does not ring right.
There is good reason to believe that in recent years, just as the idea of national
innovation systems has become widely accepted, technological communities have
become transnational as never before. There has been, first, a strong trend for manu-
facturing business firms to become multinational. This trend was set, initially, by
American based firms; somewhat later European and Japanese firms joined in. In the
1980s international joint ventures on particular product design and development proj-
ects, or on large scale research projects, began to crop up in a number of industries.
These developments partly result from and certainly reinforce other trends
toward internationalization of technology. Engineers and applied scientists now are
taught pretty much the same thing in schools in different countries. The dramatic low-
ering of national barriers to trade following the war, and the recent convergence of
living standards and factor prices in the major industrial nations, means that, increas-
ingly, firms face roughly the same market environment wherever their home base. For
both of these reasons, by the mid-1980s the technologies known and employed in the
major industrial nations were pretty much the same in most industries. Two decades
earlier there were major differences across the industrialized countries in the technol-
ogies employed and even in what the engineers knew how to do. But no longer.
More fundamentally, the internationalization of business and technology erodes
the extent to which national borders, and citizenship, define boundaries that are mean-
ingful in analyzing technological capabilities and technical advance. And these devel-
opments have both stimulated and been reinforced by the rise of transnational public
programs of R&D support, such as Eureka, and the increasing activity of organizations
such as the E.G. All this raises the following question: "to what extent does it make
sense any more to talk about "national innovation systems?".
As will be shown in the following chapters, although there are many areas of sim-
ilarity between the systems of countries in comparable economic settings, there still
are some striking differences as well. Japanese firms in the semiconductor business
tend to be different than American, German, or French firms. The university systems
are different and play different roles in the national R&D systems. The development
paths of Korea and Taiwan have been very different and so too are their present orga-
nization of industry and structure of R&D.
And the reasons for these differences reside, to a significant degree, in differences
in national histories and cultures including the timing of a country's entry into the
industrialization process. These have profoundly shaped national institutions, laws,
and policies (Landes, 1969). How they have will be one of the central topics of the
chapters that follow. Also, at the present time many national governments are com-
mitted to trying to define, and protect, or advance, what are regarded as specifically
national technological capabilities in key areas.
On the other hand, although there certainly are durable and important differences
in national characteristics that shape national innovation systems and constrain their
evolution, these systems have shown striking adaptability. U. S. public support of uni-
versity research across the board became policy only after World War II and, if one
reflects on it, it seems incompatible with the traditional American norm of small gov-
ernment. And countries clearly copy each other. The American copying of German
higher education was repaid when the Europeans later copied American large scale
public finance of university research. Europeans and Americans recently have been
attempting to copy what they see as successful cooperative research programs in Japan,
although adding important national wrinkles. And although important national dif-
ferences remain, it is not clear how much these matter to "national" firms who often
have the opportunity to set up shop in another country when it is advantageous
to do so.
There is a tension caused by the attempts of national governments to form and
implement national technology policies, in a world where business and technology are
increasingly transnational. We discuss some of the key issues later.
A GUIDE TO THIS VOLUME
The sequencing of the country studies that are the heart of this volume is an important
part of its analytic structure. Although each study can stand alone on its own merits,
the principal purpose of this project has been to map out what is similar and what is
different about national systems.
Above we have laid out some of the shared understandings about technological
change, and the processes and institutions involved, that have broadly guided the
work. We have tried to highlight the wide range of factors, organizations, and policies
influencing the capabilities of a nation's firms to innovate, in the broad sense in which
we are using that term.
This fact posed a problem regarding the overall design of the study. The desire for
comparability seemed to call for a relatively elaborate list of things all country chapters
would cover. Yet it was apparent that the most interesting features of a country's inno-
vation system varied significantly across countries, and we wanted our study to high-
light these. Limits on resources and space foreclosed doing both.
Our compromise involved two strategic decisions. First, we agreed on a limited
list of features all country studies would describe, for example, the allocation of R&D
activity and the sources of its funding, the characteristics of firms and the important
industries, the roles of universities, and government policies expressly aimed to spur
and mold industrial innovation. Beyond these the authors were encouraged to select
and highlight what they thought were the most important and interesting character-
istics of their country. But second, considerable effort was put into identifying the
kinds of comparisonssimilarities or differencesthat seemed most interesting and
important to make. In general these did not involve comparisons across all countries,
but rather among a small group where for various reasons comparison was apt.
Thus because they are the countries that currently are in the technological avant-
garde, various comparisons between the United States and Japan have been made by
others, and we felt we needed to explore these, and other interesting differences and
similarities that came to light in the course of discussion of the country chapters as
these developed. Since a central objective of the study was to broaden comparison, the
description and analysis of Germany, and to some extent those of Britain and France,
were designed, along with those of the United States and Japan, to enable comparisons
to be drawn in areas that seemed relevant. On reflection, it became clear that interest-
ing comparisons among Italy, France, and Britain could be drawn. Part I contains
chapters on these five large relatively affluent countries. The introduction aims to
point the reader's attention toward the similarities and differences that, in discussions
among the group, seemed most salient.
Part II contains studies of several high-income countries that are small, in the
sense that their population is small and hence their internal markets are limited. The
innovation systems of Denmark, Sweden, Canada, and Australia all reflect that they
are "small" high-income countries in the above sense. Comparisons across these coun-
tries are also interesting because all have a strong natural resources or agricultural base.
Part III contains studies of five lower income countries struggling with the indus-
trialization problem, some with striking success and others with less. The chapters on
Korea, Taiwan, Brazil, Argentina, and Israel provide a fascinating comparative pic-
ture of the evolving innovation systems of an important group of developing countries,
that have structured themselves in different ways and that have had quite different
experiences with industrialization.
As stated at the outset of this chapter, this study aimed to explore the usefulness
and the limitations of the concept of national innovation systems, not to reify the term.
In the concluding chapter we will look back, and try to provide an assessment of the
extent to which it has been useful to try to carve out something called an innovation
system from the complex and variegated institutional structures that make up national
economies, and the extent to which this is artificial and awkward. We also shall reflect
on the manner and extent to which national institutions matter in a world where busi-
ness, trade, and technology are increasingly transnational, and on the future of
national systems in such a world.
Although the reader must be the judge, we, the authors, come away from our
shared intellectual voyage with the belief that, yes, it does make sense to think of
national innovation systems, if one is careful to recognize the shadiness and, to some
extent, the arbitrariness of both the institutional and national borders. We also believe
we have shed significant new light on what is similar and what is different about
national systems, and the reasons behind the similarities and differences.
We are far less sure about another central issue. That is the extent to which the
particular features of a nation's technical innovation system matter centrally in affect-
ing a nation's overall economic performance in such dimensions as productivity and
income and their growth, export, and import performance.
There are certain matters we are sure about, because both general understanding
and the comparative case study evidence point to them strongly, and we will discuss
these in the concluding chapter. One is that in manufacturing at least, the efforts of
governments and universities may support, but cannot be a substitute for the tech-
nological efforts of firms. Another is the importance of a nation's education and train-
ing system. A third is that a nation's fiscal monetary and trade policies must spur, even
compel, national firms to compete on world markets. However, we, the authors, have
been impressed by the diversity of "national systems" that seem to be compatible with
relatively strong, and weak, economic performance in particular contexts. Partly this
may be because there are a variety of alternative arrangements for accomplishing basi-
cally the same thing; a number of our studies, when looked at together, suggest that
this is so. Partly it may be because the performance of the innovation system is a larger
factor behind economic performance in some contexts than in others.
But we are getting ahead of our story.
NOTE
1. We will discuss the concept of a "national innovation system" in more detail later. Three
of the participants in this project more or less independently began to use the term and the basic
conception in our work which fed into Dosi et al. (1988). See the chapters there by Freeman,
Lundvall, and Nelson in Part V, which was titled "National Innovation Systems."
REFERENCES
Allen, R. C. (1983). "Collective Invention." Journal vation." In G. Dosi, C. Freeman, R. Nelson, G.

of Economic Behavior and Organization 1(4): 1- Silverberg, and L. Soete (eds.), Technical Change
24. and Economic Theory. London: Pinter Publish-
Beer, J. (1959). The Emergence of the German Dye ers.
Industry. Urbana, IL: University of Illinois Press. Hallion, R. (1977). Legacy of Flight. Seattle: Uni-
Cardwell, D.S.L. (1971). From Watt to Clausius. Ith- versity of Washington Press.
aca: Cornell University Press. Landes, D. (1969). The Unbound Prometheus. Cam-
Cohen, I. B. (1948). Science, Servant of Man. Bos- bridge, England: Cambridge University Press.
ton: Little, Brown. Lundvall, B. A. (1988). "Innovation as an Interac-
Dosi, G., Freeman, C., Nelson, R., Silverberg, G., live Process: From User-Producer Interaction to
and Soete, L. (1988). Technical Change and Eco- the National System of Innovation." In G. Dosi,
nomic Theory. London: Pinter Publishers. C. Freeman, R. Nelson, G. Silverberg, and L.
Dunsheath, P. (1962). A History of Electrical Engi- Soete (eds.), Technical Change and Economic
neering. London: Faber & Faber. Theory. London: Pinter Publishers.
Freeman, C. (1988). "Japan, a New System of Inno- Miller, R., and Sawers, D. (1970). The Technical
Development of Modern Aviation. New York: L. Soete (eds.), Technical Change and Economic
Praeger. Theory. London: Pinter Publisher.
Morison, E. (1966). "Almost the Greatest Inven- Servan Schreiber, J. J. (1968). The American Chal-
tion." In E. Morison, (ed.), Man, Machines, and lenge. New York: Atheneum Press.
Modern Times. Cambridge, MA: MIT Press. Thomson, R. (1989). The Path to Mechanized Shoe
Nelson, R. R. (1988). "Institutions Supporting Production in the United Stales. Chapel Hill, NC:
Technical Change in the United States." In G. University of North Carolina Press.
Dosi, C. Freeman, R. Nelson, G. Silverberg, and
PARTI
LARGE HIGH-INCOME COUNTRIES

Part I is concerned with the innovation systems of six large affluent highly
industrialized nations (Table I.I). We note that the statistics for Germany
relate to the old Federal Republic, rather than the newly unified country. The
purpose of this introduction is to provide some basic statistics about these
countries.
First, the countries are populous. Each one had a population exceeding
50 million, with the United States having over 200 million people, and Japan
over 100 million. This differentiates this group of countries sharply from the
group of small affluent countries that are the subject of Part II.
And all are affluent. If one measures living standards by gross domestic
product (GDP) per capita using exchange rates to convert the nondollar cur-
rencies into dollars, as of 1989 Japan stood at the top of the pack, with the
United States and West Germany close, and with Italy and the United King-
dom bringing up the rear of this group. Measures of living standards are very
sensitive to how currencies are compared. If "purchasing power parity" as
contrasted with the official exchange rate is used to effect conversion, the
United States moves to the top by a considerable distance, and the other five
are seen as bunched at a level about 70% of that in the United States. By either
measure, the population of these countries is very well off compared with
those of the countries considered in Part III.
Although all of these countries are affluent now, the six differ significantly
in when they began their strong economic development. Great Britain was the
pioneer in the first industrial revolution, with the United States following
strongly by the mid-nineteenth century; Germany's strong development
begins later in the nineteenth century, with France and Italy lagging. At the
turn of the century Japan had hardly begun her modernization. By the begin-
ning of World War II the United States had a large per capita income lead.
Great Britain was a clear second, then Germany, France, and Italy, with Japan
by far at the bottom of the per capita league, although not necessarily last in
terms of command of technology.
Of the countries in this group Japan had the highest growth of gross
domestic product per capita over the period between 1965 and 1988, with the
United Kingdom and the United States slowest. It is not coincidental that of
these countries Japan had the highest ratio of gross domestic investment to
gross domestic product.
It may surprise some readers that of these countries, after the United
States, Japan had the smallest ratio of exports to GDP. There is an obvious
strong negative correlation between the size of the internal market of a country
and its exports as a fraction of gross national product (GNP). However, it is
West Germany that is off the regression line, not Japan, although it should be
noted that a large share of its exports went to neighboring European states. In
25
26 LARGE HIGH-INCOME COUNTRIES
all of these countries manufacturing exports accounted for the majority of

total exports. In the small high-income countries treated in Part II, exports
generally accounted for a significantly larger fraction of GDP than in the coun-
tries in Part I.
The literacy rate in all of these countries is very high, as is the enrollment
ratio in secondary education, although here the United Kingdom and Italy lag
somewhat behind the rest. The United States has a significantly larger fraction
of students going on to third-level education than do the other countries. How-
ever, there may be something of a statistical artifact here in that secondary
education is more intensive in some of the other countries than in the United
States.
A smaller fraction of university students in the United States major in
science and engineering, compared with the situation in some other countries,
particularly Japan and West Germany. Reflecting this Japan, not the United
States, stands at the top of the list in countries in terms of scientists and engi-
neers as a fraction of the total population. The United States follows with West
Germany a close third.
Table I.I. Comparison of the Six Large High-Income Countries

United West United
States Japan Germany France Kingdom Italy
GDP/capita, 1989 official 19,840 21,020 18,480 16,090 12,810 13,330
exchange rates
GDP/capita, 1988 19,558 14,228 14,161 13,603 13,428 12,985
purchasing power parity
Population, 1988 246,329 122,613 61,451 55,873 57,065 57,441
Average growth rate GDP/ 1.6 4.3 2.5 2.5 1.8 3.0
hour average 1965-1988
Gross domestic invest/GDP 16 29 19.5 23.5 21 23
average 1965-1988
Manufacturing output/GDP 22 29 44 27 27 27
Manufacturing exports/ 5 9 24 13 17 14
GDP
Total exports/GDP 7.4 16.5 32.4 17.0 20.7 15.5
Literacy rate >95 >95 >95 >95 >95 >95
Secondary level enrollment 98 96 94 92 83 75
rate
Third-level enrollment rate 60 28 30 31 22 24
Scientists and engineers/ 0.33 0.48 0.25 0.19 0.17 0.12
population
R&D/GNP 2.9 2.9 2.9 2.3 2.3 1.2
Private R&D/total R&D 48 78 64 42 49 42
Business R&D/total R&D 72.5 66.0 72.2 58.9 67.0 57.2
Private business R&D/total 66.4 98.0 86.5 69.0 68.2 71.7
business R&D
LARGE HIGH-INCOME COUNTRIES 27
The United States, Japan, and West Germany stand at the top of the
group in R&D as a fraction of GDP. France and Britain stand lower, with Italy
far behind. In all of these countries the majority of R&D is undertaken in busi-
ness enterprises. However, a significantly larger fraction of that work is
financed by government in the United States, the United Kingdom, and
France as contrasted with Japan and Germany, largely reflecting differences
in the military R&D budget.
2
The U. S. National Innovation System
DAVID C.MOWERY
NATHAN ROSENBERG
A descriptive analysis of the U. S. national innovation "system" is a gargantuan task,

made all the more ambitious by our efforts to begin this discussion in the early twen-
tieth century. We then examine the pre-World War II and postwar U. S. innovation
"systems." In the introductory and the concluding sections of the paper, we summa-
rize some of the elements that appear to distinguish this national innovation system
from those of other industrial and industrializing economies. Because of the dearth of
reliable quantitative data for the pre-1953 era, most of our comparative assessments
cover the postwar period or are largely qualitative. Moreover, the lack of data on other
dimensions of the post-1953 U. S. innovation system (e.g., comprehensive statistics
on the adoption of new technologies) means that our quantitative discussion of this
period relies heavily on R&D investment data, with all of their limitations.
One of the most salient distinguishing traits of the U. S. innovation system is its
enormous scalefor a substantial portion of the postwar era, the national R&D
investment of the United States was larger than those of all other OECD nations com-
bined. The relative importance of three key sectors within the U. S. innovation sys-
temindustry, universities, and the federal governmentas performers and as fun-
ders of R&D also contrasts somewhat with the role of these institutions in other
national innovation systems. The roles of these three sectors have changed consider-
ably during the past 70 years.
Another structural contrast between the U. S. and other national innovation sys-
tems that is particularly noteworthy for the postwar period is the importance of new
firms in the commercialization of new technologies within the U. S. economy. Rela-
tively small startup firms have played a significant role in the development and diffu-
sion of microelectronics, computer hardware and software, biotechnology, and robot-
ics during the past four decades. Their role appears to have been more significant
within the U. S. economy than in the other economies included in this comparative
analysis, with the possible exceptions of Taiwan and Denmark. This role may decline
somewhat in importance in the future, an issue to which we return later.
Two public policies in particular contributed to contrasts between the structure
of the U. S. national innovation system and those of other nations. The antitrust stat-
utes of the United States have had complex effects on the structure and performance
of this innovation system. Another policy-related point of contrast concerns the
29
important role of military R&D and procurement within the U. S. innovation system.
Frequently cited as an important source of commercial strength in high-technology
industries during the early postwar years, this large military R&D investment may now
yield a far smaller commercial payoff.
Although it has invested large sums in R&D throughout the postwar period, the
U. S. federal government has not based this investment on any economic strategy. The
fragmented structure of R&D programs' finance and administration in both the Exec-
utive branch and Congress has supported a high degree of pluralism and diversity in
publicly funded R&D programs. Conversely, however, this structure has precluded
any comprehensive oversight of the structure or economic effects of publicly financed
R&D. Evaluations of the economic benefits flowing from the large public R&D invest-
ment in the postwar United States are rare.
Some indicators of the basic research capabilities of the U. S. innovation system
(e.g., the share of postwar Nobel Prizes won for research performed within U. S. lab-
oratories, or citation analyses of scientific papers) suggest that the research perfor-
mance of the system is very strong. The U. S. innovation system has not succeeded,
however, in maintaining pre-1973 rates of growth in real earnings; nor has it enabled
U. S. productivity growth to match that of other industrial economies or prevented a
significant deterioration in the U. S. current account. U. S. economic performance has
been impaired by many factors during the past 2 decades, and the national R&D sys-
tem is by no means the sole contributor to widening trade deficits or to slow growth in
earnings and productivity. Nor are economic indices necessarily the most appropriate
criteria for an evaluation of the performance of this system. Nevertheless, these per-
ceived deficiencies are driving much of the current debate over the structure of U. S.
innovation and trade policies.
The recent technological performance of U. S. firms appears to be relatively weak
in several areas. U. S. firms have been slower than their counterparts in a number of
other industrial economies to adopt new manufacturing technologies and, some
observers suggest, do not utilize these technologies (e.g., robotics, computer-integrated
manufacturing) as intensively or as effectively as foreign firms.' Detailed comparisons
of the performance of U. S. and Japanese automobile firms suggest that U. S. firms
have been hampered by much longer development cycles for new products. Still other
analyses have faulted the ability of U. S. firms in a wide array of industries to com-
mercialize new technologies rapidly and effectively. These weaknesses, especially the
first (in view of the importance of capital costs in investment decisions), may be only
loosely related to the structure and performance of U. S. institutions for R&D.
Nevertheless, this and other evidence (some of which, such as international trade per-
formance, is affected by other factors, such as macroeconomic policy), has provoked
a wide-ranging debate over the need for new ways of organizing and funding public
and private R&D within the U. S. economy. This debate is discussed briefly later.
Discussion of new approaches to the organization and finance of innovation in
the United States has been parallelled by signs of change in the structure of the inno-
vation systems of Western Europe and Japan. Institutional change in all three of these
economic regions is occurring simultaneously with growing international trade in
high-technology products and increasing international interdependence of "national"
innovation systems. National trade and technology policies appear to be more and
more tightly interconnected. Change in the international environment, combined
with the weak apparatus for the formulation and oversight of technology policy in the
THE U. S. NATIONAL INNOVATION SYSTEM 31
United States, has created serious challenges for U. S. policymakers in the coordina-
tion of trade and technology policies. We briefly discuss this issue in our concluding
section.
THE U. S. SYSTEM BEFORE 1945
The Origins of U. S. Industrial Research

The expansion of the American economy during the late nineteenth and early twen-
tieth centuries combined with innovations in transportation, communications, and
production technologies yielded manufacturing operations of unprecedented scale
(Chandler, 1977). These production operations built on a long-established pattern of
technological innovation and adaptation that relied largely on mechanical skills,
rather than on formal scientific research. As David (1975), Rosenberg (1972), and oth-
ers have noted, growth in manufacturing productivity and output in the nineteenth
century U. S. economy was achieved in part through the development of the "Amer-
ican system of manufactures" for the production of light machines and other mechan-
ical devices.
Innovation in this sector did not rely heavily on scientific research.2 The resource
endowment of the United States, which favored the development of machinery for
agricultural and transportation applications, its enormous, protected domestic mar-
ket, and the ability of the United States to exploit foreign sources of knowledge
(importing machinery, blueprints, and skilled tinkerers from Europe and elsewhere)
all supported these developments.3 But innovation during much of this period, which
supported growth in U. S. productivity and per capita income to levels exceeding those
of Great Britain by 1913 (Nelson, 1990), relied on few of the institutions associated
with R&D in the late twentieth century.
The enormous mass-production operations that typified much of U. S. manufac-
turing during this period also were associated with a system of work organization and
hierarchy that if not unique to the United States was more systematically pursued in
this economy. The twentieth-century "American system of manufactures" was
inspired by the theories and experiments of Frederick Taylor and by the practices of
Henry Ford (among others), and spurred by the challenges of managing an ethnically
diverse and heterogeneous workforce. This system emphasized the division of opera-
tions into very narrow, relatively unskilled tasks, each of which was performed repeat-
edly by a single worker who was closely supervised by lower-level managers. Special-
ized capital equipment also was utilized in the repeated performance of these
taskshigh levels of capital intensity and specialization made for high costs of design
changes and meant that long production runs of a single product design were central
elements of this manufacturing system. Workers had little responsibility for the pace
and structure of the work process or for product quality. Fluctuations in product
demand were often managed through layoffs.
Each element of this system reinforced othersunstable employment and nar-
rowly defined work classifications both supported low levels of firm investment in
worker training and skills. Elaborate job classification systems received additional sup-
port in unionized establishments, since unions derived considerable power over both
members and management by defending these systems. A number of observers have
suggested that this "Fordist/Taylorist" system of work organization contributed to an
adversarial atmosphere of labor-management relations and low levels of investment

in worker skills that have impeded U. S. firms' efforts to adopt new technologies and
improve product quality (Walton and McKersie, 1990; Lazonick, 1991).
The materials analysis and quality control laboratories that were established
within many of these new, large factories were among the first industrial employers of
scientists and research personnel. These plant-level laboratories gradually expanded
and were supplemented by the foundation of central laboratories devoted to longer
term research. Although the development of much of the original testing and materials
analysis research was a response to changes in the structure of production, the expan-
sion and elaboration of these activities reflected change in the organizational structure
of the firm. The development of these research facilities was associated with expansion
and diversification of the firm's activities and products and substitution of intrafirm
control of these activities for market control.
Structural change in many large U. S. manufacturing firms, including their
investment in industrial research, was influenced by U. S. antitrust policy. The increas-
ingly stringent judicial interpretation of the Sherman Antitrust Act in the late nine-
teenth century made agreements among firms for the control of prices and output
more frequent targets of civil prosecution. The 1895-1904 U.S. merger wave, partic-
ularly the surge in mergers after 1898, was in part a response to this new legal environ-
ment. Finding that the legality of informal or formal price-fixing and market-sharing
agreements was under attack, firms resorted to horizontal mergers to control prices
and markets.4 Effective use of mergers for this purpose frequently required strong cen-
tral control of the firm's subsidiaries.
The influence of antitrust policy on the growth of industrial research, however,
extended beyond its effects on corporate structure. The incentives created by the Sher-
man Act for horizontal mergers were reduced by the Northern Securities decision of
1904. Nonetheless, judicial interpretations of the Sherman Act and Justice Depart-
ment prosecution of a widening array of firms increased corporate reliance on indus-
trial research and innovation to forestall or offset the effects of antitrust prosecution.
Industrial research supported corporate diversification and the use of patents to attain
or retain market power without running afoul of antitrust law.5
These early research laboratories focused in part on developing inventions cre-
ated by in-house research, but also monitored the environment for technological
threats and opportunities for the acquisition of new technologies, in many cases
through the purchase of patents or firms. Many of Du Font's major product and pro-
cess innovations, for example, were obtained by the firm at an early point in their
development, often on the advice of the central research laboratory (Mueller, 1962;
Hounshell and Smith, 1988). For much of the pre-1940 period, Du Pont research
focused on developing inventions acquired from external sources; nylon and neoprene
were exceptions to this rule. The research facilities of AT&T, General Electric, and, to
a lesser extent, Eastman Kodak, performed similar monitoring roles during this
period.6
The Growth of Industrial Research
Although recent historiography on U. S. industrial research has focused primarily on

the electrical industry (an exception is Hounshell and Smith, 1988), the limited data
on the growth of industrial research activity during the early twentieth century suggest
that it was dominated by the chemicals industry and related industries. The chemicals,
glass, rubber, and petroleum industries accounted for nearly 40% of the number of
laboratories founded during 1899-1946. The chemicals sector also dominated
research employment during 1921-1946. In 1921, the chemicals, petroleum, and rub-
ber industries accounted for slightly more than 40% of total research scientists and
engineers in manufacturing. The dominance of chemicals-related industries as
research employers was supplemented during the period by industries whose product
and process technologies drew heavily on physics. Electrical machinery and instru-
ments accounted for less than 10% of total research employment in 1921. By 1946,
however, these two industries contained more than 20% of all scientists and engineers
employed in industrial research in U. S. manufacturing, and the chemicals-based
industries had increased their share to slightly more than 43% of total research employ-
ment.
Table 2.1 provides data on research laboratory employment for 1921, 1927,
1933, 1944, and 1946 in 19 two-digit manufacturing industries and in manufacturing
overall (excluding miscellaneous manufacturing industries). Employment of scientists
Table 2.1. Employment of Scientists and Engineers in Industrial Research

Laboratories in U.S. Manufacturing Firms, 1921-1946
1921 1927 1933 1940 1946
Food/beverages 116 354 651 1712 2510
(.19) (.53) (.973) (2.13) (2.26)
Paper 89 189 302 752 770
(.49) (.87) (1.54) (2.79) (1.96)
Chemicals 1102 1812 3255 7675 14066
(5.2) (6.52) (12.81) (27.81) (30.31)
Petroleum 159 465 994 2849 4750
(1.83) (4.65) (11.04) (26.38) (28.79)
Rubber products 207 361 564 1000 1069
(2.04) (2.56) (5.65) (8.35) (5.2)
Stone/clay/glass 96 410 569 1334 1508
(.38) (1.18) (3.25) (5.0) (3.72)
Primary metals 297 538 850 2113 2460
(.78) (.93) (2.0) (3.13) (2.39)
Fabricated metal products 103 334 500 1332 1489
(.27) (.63) (1.53) (2.95) (1.81)
Nonelectrical machinery 127 421 629 2122 2743
(.25) (.65) (1.68) (3.96) (2.2)
Electrical machinery 199 732 1322 3269 6993
(1.11) (2.86) (8.06) (13.18) (11.01)
Transportation equipment 83 256 394 1765 4491
(.204) (.52) (1.28) (3.24) (4.58)
Instruments 127 234 581 1318 2246
(.396) (.63) (2.69) (4.04) (3.81)
Total (including R&D 2775 6320 10927 27777 45941
employment n.e.c.).
Note: Figures in parentheses represent research intensity, denned as employment of scientists and engi-
neers per 1,000 production workers.
Source: Mowery (1981).
and engineers in industrial research within manufacturing grew from roughly 3,000 in
1921 to nearly 46,000 by 1946.7 The ordering of industries by research intensity is
remarkably stablechemicals, rubber, petroleum, and electrical machinery are
among the most research-intensive industries, accounting for 48-58% of total employ-
ment of scientists and engineers in industrial research within manufacturing, through-
out this period. Similar stability is revealed in the geographic concentration of indus-
trial research employment during this period. Five states (New York, New Jersey,
Pennsylvania, Ohio, and Illinois) contained more than 70% of the professionals
employed in industrial research in 1921 and 1927; their share declined modestly, to
slightly more than 60%, by 1940 and 1946. The major prewar research employers
remained among the most research-intensive industries well into the postwar period
despite the growth in federal funding for research in industry. Chemicals, rubber,
petroleum, and electrical machinery accounted for more than 53% of industrial
research employment in 1940 and represented 40.3% of research employment in
industry in 1984 (National Science Foundation, 1985).
An exception to the pattern of stability in research intensity is transportation
equipment, which increased in research intensity throughout the period, and by 1946
was among the five most research-intensive manufacturing industries. The upward
movement in the relative research intensity of this industry (which includes aircraft)
is attributable to federal support of research and federal procurement during 1940-
1946, and to the rapid growth of the automobile industry throughout 1921-1946.
Government funding of wartime research in industry also contributed to research
employment growth within electrical machinery and instruments after 1940.
Schumpeter argued (1954) that in-house industrial research had supplanted the
inventor-entrepreneur (a hypothesis supported by Schmookler, 1957) and would rein-
force, rather than erode, the position of dominant firms. The data on research employ-
ment and firm turnover among the 200 largest firms suggest that during 1921 -1946 at
least, the effects of industrial research were consistent with his predictions. Industrial
research significantly improved firms' prospects for remaining in the ranks of the 200
largest firms during this period (Mowery, 1983). The growth of industrial research dur-
ing 1921-1946 among the 200 largest firms is associated with a decline in turnover
within this group. Industrial research contributed to the stabilization of market struc-
ture in the unstable economic environment of the 1921-1946 period (Edwards, 1975;
Kaplan, 1964; Collins and Preston, 1961). To the extent that federal antitrust policy
contributed to industrial research investment by large firms during this period, the pol-
icy paradoxically may have aided the survival of these firms and the growth of a rela-
tively stable, oligopolistic market structure in some U. S. manufacturing industries.
Interestingly, and in contrast to the usual statement of one of the Schumpeterian
"hypotheses," these results suggest that firm conduct (R&D employment) was an
important influence on market structure (turnover).
Publicly Funded Research and the Universities

In spite of the permissive implications of the "general welfare" clause of the U. S. Con-
stitution, federal support for science prior to World War II was limited by a strict inter-
pretation of the role of the federal government. During World War I, the military oper-
ated the R&D and production facilities for the war effort; with the exception of the
munitions industry, where the federal government relied on Du Pont, the necessary
technical and scientific expertise simply was not available in the private sector. When
one of the armed services identified a scientific need, a person with the appropriate
qualifications was drafted into that branch. One legacy of wartime programs for tech-
nology development was the National Advisory Committee on Aeronautics (NACA),
founded in 1915 to "investigate the scientific problems involved in flight and to give
advice to the military air services and other aviation services of the government"
(Ames, 1925). NACA, which was absorbed by the National Aeronautics and Space
Administration in 1958, made important contributions to the development of new
aeronautics technologies for both civilian and military applications throughout its
existence, but was particularly important during the era before 1940.
For 1940, the last year that was not dominated by the vast expenditures associated
with wartime mobilization, total federal expenditures for research, development, and
R&D plant amounted to $74.1 million. Of that, Department of Agriculture expendi-
tures amounted to $29.1 million, or 39%. In 1940, the Department of Agriculture's
research budget exceeded that of the agencies that would eventually be combined in
the Department of Defense, whose total research budget amounted to $26.4 million.
Between them, these categories accounted for 75% of all federal R&D expenditures.
The claimants on the remaining 25%, in descending order of importance, were the
Department of the Interior ($7.9 million), the Department of Commerce ($3.3 mil-
lion), the Public Health Service ($2.8 million), and the National Advisory Committee
on Aeronautics ($2.2 million).
Federal expenditures for R&D throughout the 1930s constituted 12-20% of total
U. S. R&D expenditures. Industry accounted for about two-thirds of the total. The
remainder came from universities,8 state governments, private foundations, and
research institutes. One estimate suggests that state funds may have accounted for as
much as 14% of university research funding during 1935-193'6 (National Resources
Planning Board, 1942, p. 178).'Moreover, the contribution of state governments to
nonagricultural university research appears from these data to have exceeded the fed-
eral contribution.
To a greater extent than was true of Germany or Great Britain, industrial and
academic research developed in parallel in the United States. The pursuit of research
was recognized as an important professional activity within both U. S. industry and
higher education only in the late nineteenth century, and research in both venues was
influenced by the example (and in the case of U. S. industry, by the competitive pres-
sure) of German industry and academia.
Linkages between academic and industrial research were powerfully influenced
by the decentralized structure and funding of the U. S. higher education system, espe-
cially public universities. Public funding meant that the size of the U. S. higher edu-
cation system outstripped that of such European nations as Great Britain.9 Of equal
or greater importance, however, was the fact that public funding for many U. S. uni-
versities was provided by state governments, rather than by the federal government.
The politics of state funding meant that both the curriculum and research of U. S.
public universities were more closely geared to commercial opportunities than was
true in many European systems of higher education.10 Especially within emerging sub-
fields of engineering and, to a lesser extent, within mining and metallurgy, state uni-
versity systems often introduced new programs as soon as the requirements of the local
economy became clear.
The use of scientific knowledge and problem-solving techniques in industry was
accelerated by growth in the pool of technically trained personnelespecially engi-
neers. This expansion was the result in part of growth in the number of engineering
schools and programs in the second half of the nineteenth century. The training of
these engineers was, to be sure, often elementary in character and did not prepare them
for work at the scientific frontier. Indeed, before 1940 there are few if any areas of
scientific research in which U. S. universities or scholars could be described as oper-
ating at the scientific frontier.
Although the situation was improving in the decade before 1940, Cohen (1976)
noted that virtually all "serious" U. S. scientists completed their studies at European
universities, and Thackray et al. (1985) argue that American chemistry research dur-
ing this period attracted attention (in the form of citations in other scientific papers)
as much because of its quantity as its quality." Interestingly, recent citation analyses
suggest that American physics research had begun to acquire a world-class research
reputation by the 1930s, before the infusion of scientific brilliance resulting from the
emigration to the United States of European scientists.12 The rise of American physics
research to scientific eminence is reflected as well in the award of Nobel Prizes to Lang-
muir, Millikan, Compton, and Davisson in physics during this periodtwo of these
recipients made their pathbreaking discoveries as employees of major U. S. industrial
research laboratories. Nevertheless, the current eminence of U. S. scientific research
in a broad array of disciplines is largely a postwar phenomenon.
Regardless of the quality of the scientific research performed within the U. S.
research system before World War II, it was the larger body of scientific knowledge,
and not merely frontier science, that was relevant to the needs of an expanding indus-
trial establishment.13 Thus, engineers and other technically trained personnel served
as valuable carriers of scientific knowledge. As a result, the number of people bringing
the knowledge and methods of science to bear on industrial problems was vastly
greater than the limited number of individuals that society chose to label "scientists"
at any particular time. Moreover, as was noted above, the scale of the U. S. higher
educational system exceeded those of other industrial nations during this period. As
in the postwar Japanese research system, this broad-based system of training scientists
and engineers aided the diffusion and utilization of advanced scientific and engineer-
ing knowledge. Even where it did not advance the knowledge frontier, higher educa-
tion appears to have been an important instrument for scientific and engineering
"catch-up" in the United States during the early twentieth century.
Agricultural Research
Although the focus of this and other papers is technical advance in industry, the U. S.
agricultural sector deserves brief mention. Agricultural products have long been
important U. S. exports, and U. S. industrial development for much of the last 150
years relied heavily on the exploitation of linkages between agriculture and industry,
as in the development of advanced technologies for food processing and in the growth
of a U. S. technological advantage in farm machinery and equipment (Patel and Pavitt,
Table 2.2. Defense R&D as a Share of Federal R&D Spending, 1960-1990

Year Percent Year Percent
1960 80 1976 50
1961 77 1977 51
1962 70 1978 49
1963 62 1979 48
1964 55 1980 51
1965 50 1981 54
1966 49 1982 61
1967 52 1983 64
1968 52 1984 66
1969 54 1985 67
1970 52 1986 69
1971 52 1987 69
1972 54 1988 67
1973 54 1989(est.) 60
1974 52 1990(est.) 65
1975 51
Source: Budget of the U.S. Government (Washington, D.C.: US Government Printing Office, 1989).
1986). The data cited above note the prominent role of agriculture as a recipient of
federal and state research funds during the pre-1940 era.
Much of the foundation for the extensive system of publicly supported higher
education in the United States was in fact laid down during the nineteenth century as
a means of financing research and other services for the agricultural sector. The Morrill
Act of 1862 provided the wherewithal for the founding of state universities to pursue
research and education in the "agricultural and mechanical arts." Further support for
agricultural research was provided in the Hatch Act of 1887 and the Adams Act of
1906, which established state experimental stations to perform agricultural research.
Table 2.2 includes data on funding for the pre-1940 period and illustrates two points:
(1) the important role of state funds in financing this research system, a role that was
greater during the pre-1940 period than in the postwar period; and (2) the sizable por-
tion of the budget devoted to extension activities, including testing and support for the
dissemination of best-practice techniques in an "industry" in which local conditions
and problems required considerable modification of seeds, techniques, and equip-
ment.
Extension activity appears to have been especially important for the pre-1940
growth of U. S. agriculture, which was almost entirely extensive rather than intensive
in character (Parker, 1972). Figure 2.1 illustrates the nearly flat trend in output per
acre or per man-hour for the pre-1940 period. Much of the growth in agricultural out-
put during this period depended on the expansion of cultivated land and the dissem-
ination of seed strains that were suited to local growing conditions. These functions
relied as much on extension as on scientific research. Beginning in 1940, however, agri-
cultural productivity grew rapidly, as a result of the exploitation of advances in bio-
logical and chemistry research. Hybrid seed corn, for example, came into widespread
use only in the late 1930s. The scientific research component of this considerable state
and federal investment in agricultural research and extension thus began to reap sub-
stantial payoffs only after a number of decades.14
Figure 2.1. Output per acre, 1868-1984. Source: U.S.D.A., Agricultural Statistics, 1962, 1982,
1987. From W. N. Parker, "Agriculture." In L. E. David, R. A. Easterlin, and W. N. Parker
(eds.), American Economic Growth, p. 374. New York: Harper & Row, 1972.
Conclusion
Much of the structure of the private sector components of the U. S. national innova-
tion system took shape during the 1900-1940 period. Closely linked with the rise of
the giant multiproduct corporation that began at the turn of the century, industrial
research contributed to the stability and survival of these firms. Before 1940 federal
support for research that was not agricultural was very limited and may well have been
exceeded by state government support. Although university research budgets before
1940 were miniscule by later standards, the system was one in which the requirements
of industry, agriculture, and mining were recognized and accommodated. As research
within industrial establishments grew in importance, university research during this
period often involved various forms of collaboration with private industry.
THE POSTWAR SYSTEM
Introduction
World War II transformed the U. S. R&D system. Federal government support for
industrial and academic research expanded dramatically, although in contrast to other
nations, nongovernmental institutions retained primary responsibility for the perfor-
mance of much of this R&D. World War II also transformed the global technological
and competitive environment within which U. S. firms operated. The United States
emerged from wartime as an unchallenged leader in a much broader range of tech-
nologies than was true at any point before 1940, and federal funding built a strong
scientific research capability during the postwar years. Moreover, the demands of
reconstruction were to prolong U. S. technological and economic supremacy. Because
a central point of contrast between the prewar and postwar research systems is the
upsurge in federal government involvement in the national R&D system, this section
devotes considerable attention to the contours of federal R&D support in basic, com-
mercial, and military research.
World War II and Its Aftermath

With war preparations and the entry of the United States into World War II in Decem-
ber 1941, the bucolic picture of federal R&D expenditures discussed above was trans-
formed. Funding for the primary categories of prewar R&D, which were not war
related, grew only slightly during the war in dollar terms and declined substantially in
real terms. Total federal R&D expenditures (in 1930 dollars) rose from $83.2 million
in 1940 to a peak of $1,313.6 million in 1945. Over the same period, the research
expenditures of the Department of Defense rose from $29.6 to $423.6 million (in 19 30
dollars).
The success and the organizational structure of the massive federal wartime R&D
program yielded several important legacies. The successful completion of the Man-
hattan Project, whose research budget in the peak years 1944 and 1945 substantially
exceeded that of the Department of Defense, created a research and weapons produc-
tion complex that eventually would usher in the age of truly "big science." Paradoxi-
cally, the Manhattan Project's success in creating weapons of unprecedented destruc-
tive power contributed to rosy postwar perceptions of the constructive possibilities of
large-scale science for the advance of societal welfare.
Far smaller in financial terms, but highly significant as an institutional innova-
tion, was the Office of Scientific Research and Development (OSRD), a civilian agency
directed by Vannevar Bush. The OSRD was not under military control. Although it
employed federal funds on wartime scientific research projects, OSRD entered into
contracts with the private sector for the performance of that research and allowed full
reimbursement of research costs. OSRD also relied heavily on universities as research
performers. The largest single recipient of OSRD grants and contracts during wartime
(and the inventor of that device beloved of university research administrators, insti-
tutional overhead) was M.I.T., with 75 contracts for a total of more than $ 116 million.
The largest corporate recipient of OSRD funds, Western Electric, accounted for only
$17 million (Pursell, 1977, p. 364).
The contrast between the organization of wartime R&D in World War I and
World War II reflects the far more advanced state of development of university and
private sector research capabilities during the second global conflict. The contractual
arrangements developed by OSRD during World War II allowed the Office to tap the
broad array of private sector scientific capabilities that had developed during the inter-
war period. '5 Members of the scientific community were called on to recommend and
to guide as well as to participate in scientific research with military payoffs. OSRD was
not subordinated to the military and had direct access to the President and to the per-
tinent congressional appropriations committees.
The success of these wartime contractual arrangements with the private sector
contributed to a feature of postwar publicly funded American R&D that distinguishes
it from both the prewar period and other countries.16 In 1940 federal R&D went to
support research performed within the federal establishment itselfby government
civil servants, as in the National Bureau of Standards, the Department of Agriculture,
and the Public Health Service, or by state institutions financed by federal grants, as in
the agricultural experiment stations. In the postwar period, by contrast, most federal
R&D funds have supported the performance of research by nongovernmental orga-
nizations.
Postwar R&D Expenditures
Two salient features of postwar R&D spending are the magnitude of the overall
national R&D investment and the size of the federal R&D budget. Throughout this
period, federal R&D spending has been a large fraction of a very large national R&D
investment. The total volume of resources devoted to R&D since the end of World
War II is large not only by comparison with our earlier history, but also by comparison
with other Organization for Economic Cooperation and Development (OECD) mem-
ber countries. Indeed, as late as 1969, when the combined R&D expenditures of the
largest foreign industrial economies (West Germany, France, the United Kingdom,
and Japan) were $11.3 billion, those for the United States were $25.6 billion. Not until
the late 1970s did the combined total for those four countries exceed that of the United
States (Danhof, 1968, p. 192).
Of the two components of national R&D spending, private and federal spending,
the latter has been more volatile, reaching a peak of about two-thirds of total R&D in
the mid-1960s and declining substantially after this point. Over the same period, pri-
vate R&D has tracked GNP growth more closely and therefore has grown more stead-
ily (see Table 2.3). Total R&D spending was slightly more than 1% of GNP in the
immediate postwar years. The percentage grew rapidly in the second half of the 1950s
and peaked at almost 3% in the mid-1960s, after which it declined until the second
half of the 1970s.
Within the postwar R&D system, federal expenditures have financed somewhere
between one-half and two-thirds of total R&D, the great bulk of which is performed
by private industry. In 1985 73% of all federally funded R&D was performed in private
industry, and only 12% in federal intramural laboratories (although 47% of all R&D
Table 2.3. Sources of Funds for Research and Development by Sector: 1953-1989 (Dollars in Millions)
Current Dollars Real (1982) Dollars"
United Federal Universities Other United Federal Universities Other Total Federal
Year States Govt. Industry and Colleges Nonprofit States Govt. Industry and Colleges Nonprofit Private* (%)
1953 5,124 2,753 2,245 72 54 19,744 10,590 8.671 276 208 9,155 53.6
1955 6,172 3,502 2,520 88 62 22,760 12,923 9,282 326 229 9,837 56.8
1960 13,523 8,738 4,516 149 120 43,648 28,191 14,591 479 387 15,457 64.6
1965 20,044 13,012 6,548 267 217 59,351 38,532 19,384 791 643 20,818 64.9
1970 26,134 14,892 10,444 461 337 62,405 35,636 24,851 1,111 807 26,769 57.1
1975 35,213 18,109 15,820 749 535 59,883 30,986 26,679 1,302 916 28,897 51.7
1980 62,594 29,453 30,914 1,326 901 73,237 34,548 36,067 1,565 1,057 38,689 47.2
1985 107,757 51,668 52,358 2,377 1,354 96,999 46,463 47,188 2,131 1,217 50,536 47.9
1989(est.) 132,350 62,700 64,035 3,800 1,815 105,029 49,720 50,863 3,007 1,439 55,309 47.3
"Based on GNP implicit price deflator.
^Total for three columns including industry, universities and colleges, and other nonprofit.
Sources: National Science Foundation, SRS. National Patterns of R&D Resources (1989). NSF 89-308.
was financed by the federal government). The remaining 15% is a critical component
of federal R&D spending. Approximately 3% of federal R&D supports federally
funded research and development centers (FFRDCs) administered by universities and
colleges, 3% is allocated to other nonprofit institutions, and 9% supports university
research.17
Federal funds have been especially important in supporting basic research.
Although that share has been declining for the past several years and now is at its lowest
level in 20 years, federal funds still represent two-thirds of total basic research spend-
ing. Only 15% of federally funded basic research currently is performed within the fed-
eral research establishment. Universities have increased in importance as basic
research performers during this period. In 1953, less than one-third of all basic research
was performed in universities and FFRDCs at universities and colleges. In recent
years, however, these institutions have performed more than one-half of all basic
research.
Support for basic research is concentrated in a few agencies within the federal
budget. By far the largest federal obligations are in the Department of Health and
Human Services, where the basic research budget consists overwhelmingly of the
expenditures of the National Institutes of Health. The next largest obligations, in
descending order, are in NSF, DOD, DOE, and NASA.
Military R&D Funding
The military services have dominated the federal R&D budget for the past 30 years,
falling below 50% of federal R&D obligations in only 3 years (see Table 2.4). In 1960
defense research constituted no less than 80% of federal R&D funds. It declined
sharply from that level (a decline offset by the growth of the space program) and hov-
ered around the 50% level until the early 1980s, when it rose swiftly again.
The dominant role of the defense budget within the total federal R&D budget has
another important implication. The defense R&D budget is far more development
intensive than the rest of the federal R&D budget. This characteristic of the dominant
component of federal R&D spending imparts a strong bias to the overall federal R&D
budget in favor of development. If the 1982 federal budget is broken down into defense
Table 2.4. Trends in Federal R&D Expenditures

Obligations (in Billions of Dollars)
Year Defense" All Other Total Basic Research* Defense (%) All Other (%)
1960 6.1 1.5 7.6 0.6 80 20
1965 7.3 7.3 14.6 1.4 50 50
1970 8.0 7.3 15.3 1.9 52 48
1975 9.7 9.3 19.0 2.6 51 49
1980 15.1 14.7 29.8 4.7 51 49
1985 33.4 16.1 49.5 7.8 67 33
1990(est.) 44.0 23.3 67.3 11.2 65 35
Încludes military-related programs of the Departments of Defense and Energy.
''Included in totals for conduct of R&D.
Source: Budget of the U.S. Government, 1990. Executive Office of the President, Office of Management and Budget, "Special
Analysis J" (1989, 1990).
and nondefense components, the share of basic, applied, and development expendi-
tures within each total is as follows:
1983 Federal R&D Expenditures18

Defense (%) Nondefense (%)
Basic 3.2 33.7
Applied 11.0 35.3
Development 85.8 31.0
Total 100.0 100.0
Sources:
1958 data: National Science Foundation, Research and
Development in Industry: 1974, Tables B-3, B-6, and B-9.
1972 data: National Science Foundation, National Patterns
of Science and Technology Resources: 1980, Tables 37-39.
1986 data: National Science Foundation, National Patterns
of Science and Technology Resources: 1989, Tables B-25, B-
26, and B-27. Implicit GNP deflator used for conversion to
1987 dollars.
The largest items in the DOD R&D budget involve the development of advanced
weapons systems, construction and testing of prototypes, and so on. Conversely, DOD
devotes a smaller share of its R&D budget to basic and applied research than any other
major federal R&D funding agency.19
As a result of the development emphasis in defense R&D and the large size of the
defense R&D budget, the distribution of the federal R&D budget across industry sec-
tors is highly concentrated. Nearly 80% of all federal R&D in 1984 went to two indus-
try sectorsaircraft and missiles (over 50%) and electrical machinery (over 25%).20
Nonelectrical machines was a distant third, and motor vehicles and other transpor-
tation equipment fourth (see Table 2.5).
Have military expenditures strengthened the commercial innovative capabilities
of U. S. firms during the postwar period? Assessing the commercial impact of military
R&D spending is complicated by the fact that the influence of military R&D spending
can easily be confounded with that of military procurement. The benefits that are
sometimes perceived to flow from military R&D are in fact frequently the product of
military R&D plus massive military procurement. This overlap between the influence
of Pentagon R&D spending and Pentagon procurement is accentuated by the practice
of paying a percentage of military procurement contracts to defense suppliers as an
"independent R&D" allowance that is generally not included in either the formal
defense R&D budget or the reported R&D expenditures of recipient firms.
In the semiconductor industry, for example, the role of military procurement
may well have outweighed the direct influence of military R&D expenditures (Utter-
back and Murray, 1977). The large procurement needs of the military and NASA were
vital in the early years of new product development in electronics. From the mid-1950s
to the late 1960s, the federal government (mainly the military and NASA) accounted
for a large, although declining, share of the output of semiconductor devices. By the
end of the 1960s the computer industry displaced the military as the largest end user
market for integrated circuits. Profits and overhead from military procurement con-
tracts supported company-funded R&D and thereby may have generated more civil-
Table 2.5. R&D Funds by Industry
1958
R&D Funds by Industry Total Federal Company
Current dollars
Total 8389 4759 3630
Chemicals and allied products 792 126 666
Industrial chemicals 553 110 443
Drugs and medicines and other chemicals 239 16 223
Petroleum refining and extraction 246 12 234
Rubber products 89 21 68
Primary metals 131 14 117
Ferrous metals and products 80 2 78
Nonferrous metals and products 51 12 39
Fabricated metal products 162 57 105
Nonelectrical machinery 781 343 438
Electrical machinery 1969 1337 632
Communication equipment and electronic components 868 615 253
Motor vehicles and other transportation equipment 856 296 560
Aircraft and missiles 2609 2276 333
Professional and scientific instruments 294 137 157
Scientific and mechanical measuring instruments 156 93 63
Optical, surgical, photographic, and other instruments 138 44 94
All other manufacturing industries 343 78 265
Nonmanufacturing industries 117 62 55
1989 constant dollars
Total 35674 20238 15437
1972
Current dollars
Total 19552 8017 11535
Table 2.5. R&D Funds by Industry (Continued)
1972
Total 45730 18751 26979
1986
Current dollars
Total 80629 27782 52847
Petroleum refining and extraction NA NA 1867
Primary metals NA NA 809
Ferrous metals and products NA NA 388
Table 2.5. R&D Funds by Industry (Continued)
1986
Scientific and mechanical measuring instruments NA NA 1959
Optical, surgical, photographic, and other instruments NA NA 2617
All other manufacturing industries NA NA 3172
Total 89485 30834 58652
cals
Petroleum refining and extraction NA NA 2072
Primary metals NA NA 898
Ferrous metals and products NA NA 431
Communication equipment and ronic components 13412 4874 8537
nstruments
Scientific and mechanical measuring instruments NA NA 2174
Optical, surgical, photographic, i ther instruments NA NA 2904
All other manufacturing industries NA NA 3520
Sources:
1958 data: National Science Foundation, Research and Development in Industry: 1974, Tables B-3, B-6, and B-9.
1972 data: National Science Foundation, National Patterns of Science and Technology Resources: 1980, Tables 37-39.
1986 data: National Science Foundation, National Patterns of Science and Technology Resources: 1989, Tables B-25, B-26,
and B-27. Implicit GNP deflator used for conversion to 1987 dollars.
ian spillovers than R&D that was directly funded by the military. In addition, direct
financial support from the Pentagon was available for the construction of production
facilities by winners of contracts under the provisions of the Defense Production Act.
Much of this defense-related procurement demand also was covered by "Buy Ameri-
can" provisions, which further favored U. S. over foreign suppliers (Malerba, 1985).
Defense procurement lowered marketing-based barriers to entry. Lower entry
barriers allowed small firms, such as General Radio, Texas Instruments, and Transi-
tron, to direct their development efforts to meeting the performance and design
requirements of a single large customer in the 1950s. The relatively modest barriers to
entry were associated with the entry and rapid growth of numerous young, relatively
small firms in the industry.
Granted that military-civilian technological spillovers have, at certain times and
in certain industries, been a significant economic phenomenon, are they as large today
as they were 20 or 30 years ago, and are spillovers likely to be rising or falling in the
years ahead? The answers to these questions vary across different technologies. The
commercial spillovers from defense research and procurement also appear to fluctuate
over time within a specific technology. A number of factors influence the magnitude
of such spillovers, but among the most important is the generic similarity of civilian
and military requirements for a technology. Although generalizations on this issue are
hazardous, increasing divergence in these requirements in a broad array of technolo-
gies appears to have reduced the economic importance of military-civil spillovers in
recent years.
Frequently, commercial and military requirements for performance, cost, rug-
gedness, and so on more closely resemble one another early in the development of a
new technology. This broad similarity in requirements appears to have been associated
with significant spillovers in microelectronics in the early 1960s, when the demands
of the commercial and military markets for miniaturization, low heat in operation,
and ruggedness did not diverge too dramatically. During the 1950s and 1960s, the jet
engine was applied in military strategic bombers, transports, and tankers, all of which
had fuselage design and engine performance requirements that resembled some of
those for commercial air transports. The jet engine was a prime example of a military
spillover to the civilian economy. Over time, however, the size and even the direction
of spillovers in these technologies appear to have changed.21 The changing relationship
between military and commercial technologies in microelectronics influenced the
establishment of the Sematech (Semiconductor Manufacturing Technology) research
consortium, funded jointly by private industry and the U. S. military (see below for
further discussion).
University Research and Federal Funding

Another change in the structure of the postwar U. S. research system from its prewar
outlines is the expansion of research in U. S. institutions of higher learning. Much of
this growth in research of course reflects the expansion in federal support for university
research during and after World War IIindeed, industrial funding now may account
for a smaller share of university research than was true during the 1930s (the industry
share of university research funding in the 1970s was well below that of the early
1950s).
By any measure, the expansion of academic research was immense. From an esti-
mated level of nearly $420 million (1982 dollars) in 1935-1936, university research
(excluding FFRDCs) grew to more than $2 billion (1982 dollars) in 1960 and $8.5
billion in 1985, nearly doubling as a share of GNP during 1960-1985 (from 0.13 to
0.25). The increase in federal support of university research has transformed major
U. S. universities into centers for the performance of scientific research, an unprece-
dented role.
The huge increase in federal expenditures on university research has taken the
form of contracts and grants for specific research projects. Most of the "demand" for
scientific research has emanated from a centralized federal authority, although a num-
ber of federal departments and agencies with distinctly separate missions and goals
have contributed to this demand. On the supply side has been a heterogeneous range
of institutions, public and private, committed to both research and education, depen-
dent on the federal government for financial support but otherwise determined to
maintain their autonomy.
The federal government did not confine itself to expanding the demand for uni-
versity research. Federal actions on the supply side enlarged the pool of scientific per-
sonnel and supported the acquisition of the physical equipment and facilities essential
to the performance of high-quality research. After World War II, federal programs
increased financial aid for students in higher education. The best known was the G.I.
Bill, which provided substantial financial support to all veterans who enrolled in col-
lege-level educational programs; others include graduate fellowships supported by
NSF and AEC funds, training fellowships from the National Institutes of Health, and
the National Defense Education Act fellowships. Federal funds also made it possible
for universities to purchase increasingly expensive scientific equipment and advanced
instrumentation, central to the expansion of both research and teaching functions of
the university scientific community.
By simultaneously providing funds for university education and for the support
of research within the university community, the federal government strengthened the
university commitment to research (a commitment that, before World War II, ran a
very poor second to teaching) and reinforced the link between research and teaching.
The combination of research and teaching in higher education has been carried much
further in the United States than elsewhere. In Europe and Japan, for example, a larger
fraction of research is carried out in specialized research institutes not connected
directly with higher education and in government-operated laboratories.22
Research in Industry
As the above discussion makes clear, private industry retained its dominance as a per-
former of research amid shifts in the sources of the funding for this research. In 1985,
although it performed 73% of total U. S. research and development, industry
accounted for slightly more than 50% of total funding. Its continued primacy as a per-
former of R&D, however, meant continued growth in employment within industrial
researchfrom less than 50,000 in 1946 (Table 2.1) to roughly 300,000 scientists and
engineers in 1962, 376,000 in 1970, and almost 600,000 in 1985 (Birr, 1966; U. S.
Bureau of the Census, 1987, p. 570).
Although the R&D facilities of established firms expanded greatly as a result of
hostilities and the following Cold War, relatively young industrial firms have also
played a prominent role in the postwar U. S. industrial innovation system. The suc-
cessive waves of new product technologies that have swept through the postwar U. S.
economy, including semiconductors, computers, and biotechnology, have been com-
mercialized in large part through the efforts of new firms.23 The role of small firms in
commercializing new technologies in the United States during this period appears to
contrast with the pattern in both Japan and Western Europe, where established firms
in electronics, Pharmaceuticals, and other industries have played a more significant
role in new technology development.
Several factors have contributed to this prominent role of new, small firms in the
postwar U. S. innovation system. The large basic research establishments in univer-
sities, government, and a number of private firms served as important "incubators"

for the development of innovations that "walked out the door" with individuals who
established firms to commercialize them. This pattern has been particularly significant
in the biotechnology, microelectronics, and computer industries. Indeed, high levels
of labor mobility within regional agglomerations of high-technology firms have served
both as an important channel for technology diffusion and as a magnet for other firms
in similar or related industries. At least one scholar has argued that the far lower levels
of interfirm labor mobility in Japan would restrict technology transfer without the off-
setting influence of cooperative research projects (Saxonhouse, 1982, 1986).
The foundation and survival of vigorous new firms also depend on a sophisticated
private financial system that can support new firms during their infancy. The U. S.
venture capital market played an especially important role in the establishment of
many microelectronics firms during the 1950s and 1960s, and has contributed to the
growth of the biotechnology and computer industries. Throughout the 1970s, $100-
200 million of funds annually flowed into this industry from the venture capital com-
munity, and one informed observer has suggested that by the early 1980s, flows of ven-
ture capital for high-technology firms may have been as much as $2-4 billion
annually. This abundant supply of venture capital was gradually supplemented by
public equity offerings.24
Commercialization of microelectronics and biotechnology innovations by new
firms was aided by a relatively permissive intellectual property regime in these indus-
tries that aided technology diffusion and reduced the burden on young firms of liti-
gation over innovations that may have originated in part within established firms or
other research installations. In microelectronics, liberal licensing and cross-licensing
policies were one byproduct of the 1956 consent decree that settled the federal antitrust
suit against AT&T. In biotechnology, continuing uncertainty over the strength and
breadth of intellectual property protection may have discouraged litigation.
Postwar U. S. antitrust policy also contributed to the importance of startup firms.
The 1956 settlement of the AT&T case significantly improved the environment for
startup firms in microelectronics, because of the liberal patent licensing terms of the
consent decree and because the decree prohibited AT&T from commercial activities
outside of telecommunications. As a result, the firm with the greatest technological
capabilities in microelectronics was effectively forestalled from entry into commercial
production of microelectronic devices, creating substantial opportunities for entry by
startup firms. A 1956 consent decree settling another antitrust suit against IBM also
mandated liberal licensing by this pioneer computer firm of its punchcard and com-
puter patents at reasonable rates (Flamm, 1988). The major antitrust suits of this
period also may have indirectly affected the prospects for startup firms, since a number
of established firms that were involved in antitrust litigation during the late 1940s and
1950s were deterred from continuing their prewar policies of technology acquisition
through the acquisition of smaller firms.
During much of the postwar period, U. S. military procurement aided the growth
of new firms. As was noted earlier, the U. S. military market in the 1950s and 1960s
provided an important springboard for startup firms in microelectronics and com-
puters, who faced relatively low marketing and distribution barriers to entry into this
market.25 The benefits of the military market were enhanced further by the substantial
possibilities for technological spillovers from military to civilian applications. Some of
the effects of military procurement on startup firms' success, and on the spillovers
from military to commercial applications, were a result of policy. In contrast to Euro-
pean military procurement, the U. S. armed services were willing to award major pro-
curement contracts to firms with little track record in serving the military (or, in many
cases, any) market. In industries such as microelectronics, these contracts attracted
startup firms as well as enterprises that historically had mainly served civilian markets
and that remained concerned with extracting commercial applications from their mil-
itary technology development efforts.26
The industrial research facilities of many of the pioneers of research, such as Gen-
eral Electric, Du Pont, RCA, and Kodak underwent considerable change during the
postwar period. The wartime demonstration of the significant potential for commer-
cial and military applications of scientific research, combined with vast increases in
government funding for research in defense-related technologies, led a number of
these firms to expand their central research facilities and to shift applied research to
the product divisions.
Especially during the early postwar period, buoyant domestic and international
markets supported robust profits and rapid expansion of R&D in both the central lab-
oratories and the divisional laboratories. Central R&D facilities focused increasingly
on fundamental research in many of these large firms, leaving the development and
application of new technologies, as well as the improvement of established products
and processes, to the divisional laboratories. Federal research contracts were awarded
to the central research facility or to a dedicated divisional laboratoryfor reasons of
both policy (accounting regulations governing federal contracts) and security; research
for government contracts often was carried out in separate facilities. In some cases, as
in that of Du Pont, the use of the central laboratory and Development Department as
"scanning devices," searching out promising technologies or firms for acquisition, was
ruled out by senior management as a result of increasing antitrust restrictions on
expansion through acquisition. As a result, internal discovery and development of new
products became paramount.27 The data on basic research for 1953 and 1960 are less
reliable than those for later years, but suggest nonetheless that the share of total U. S.
basic research financed by industry during the postwar period may well have been at
its peak during the 1950s and early 1960s.
As the fundamental research activities of the central research laboratories
expanded (expansion that often relied on federal funds), and as manufacturing firms
diversified into new product lines through acquisition and internal development, the
ties between the central research laboratory and the increasingly diverse and in some
cases geographically distant product divisions of these firms were weakened.28 Internal
communications between the fundamental and applied research operations deterio-
rated, making it more difficult to commercialize the work of the central research facil-
ity and eroding the contributions of central research to the activities of the product
divisions.29
Severe competitive pressures from foreign firms, increases in the real cost of cap-
ital, and a slowdown in the growth rate of the domestic economy in the 1970s may
have caused the returns to R&D investment to decline during the mid-1970s,30 and the
rate of growth in real industry expenditures on R&D declined. Industry funding of
basic research shrank, and many of the central research facilities of the giant corpo-
rations entered a period of budgetary austerity or cutbacks. After a resurgence in the
early 1980s, the rate of growth in industry-funded R&D declined, and the National
Science Foundation in early 1990 reported that real industry-financed R&D had
declined during 1988-1989.31
Research in Agriculture
The rapid growth in postwar agricultural productivity (measured either in terms of

yield per acre or output per hour) suggests that the returns to the federal and state
investment in agricultural research (as opposed to extension) increased sharply during
the postwar period. This research program is almost alone among federal R&D pro-
grams in having been the subject of a series of economic evaluations. These studies
have consistently found high returns to the public investments in agricultural research
(Evenson, 1982, reviews a number of these studies), and the U. S. agricultural research
system has been cited in some reports (e.g., OTA 1990) as an exemplary program of
support for technology adoption and adaptation. Public budgetary support for the
agricultural research and extension network continued to expand during the postwar
period, although the state government share of this budget declined somewhat. Devel-
opments in postwar agricultural research also point out the important interaction
between research investment decisions and the appropriability of the returns, a vari-
able influenced by intellectual property rights.
The Plant Patent Act of 1930 was an important early development in the estab-
lishment of private property rights for plant varieties, and its passage directly influ-
enced the subsequent commercialization of hybrid strains of corn and other plant vari-
eties. The provisions of the Plant Patent Act were expanded and strengthened in the
Plant Variety Protection Act (PVPA) of 1970. The U. S. Supreme Court's 1980 deci-
sion in Diamond v. Chakrabarty, upholding the patentability of living organisms, fur-
ther expanded the coverage of intellectual property rights in agricultural research. In
response to these developments, privately financed agricultural R&D has grown rap-
idly and the balance of public and private funding has shifted: "private spending
accounted for roughly one-fourth of total R&D relevant to agriculture in the 1950s
but. .. this had risen to approximately 40 percent in the 1960s and 50 percent in the
1970s" (Evenson, 1982, p. 242). Evenson (1983) notes that the 1970 PVPA produced
a dramatic increase in the development of new strains in soybeans and other crops.32
Although the balance of public and private funding has shifted during the postwar
period, the public and private research investments appear to be complementary,
rather than duplicative, as one might expect in a research system that is sensitive to
local political and economic demands.33
Although the agricultural research system's economic returns and blend of
research and diffusion support have been widely praised, other analysts have criticized
the system's weaknesses in scientific research, notably the failure to develop stronger
expertise in molecular biology and related fields.34 The resource allocation mechanism
for the system also is hardly a model of scientific self-governance (i.e., peer review and
evaluation of research proposals). Historically, two-thirds of the federal agricultural
research funds provided to land-grant universities by the U. S. Department of Agri-
culture have been allocated among the states on a formula basis, rather than on the
basis of scientific merit or research excellence (U. S. Congressional Office of Technol-
ogy Assessment, 1986). The share of the federal agricultural research budget allocated
on a competitive basis has slowly increased in recent years, but remains low. In both
of these areas, the agricultural system appears to exhibit some tension between polit-
ical responsiveness and research excellence.
Partly in response to criticism of the research performance of the agricultural
research system, the balance of federal funding between extension activities and
research also has shifted during the past 25 years. Federal funding for extension grew
more rapidly than funding for research during 1966-1975, but these trends were
reversed after 1975. Combined with expanding private investment in agricultural
research, the Congressional Office of Technology Assessment has argued that the bal-
ance of the public/private system of agricultural research is tipping away from its his-
toric support for technology adoption toward an expanded role in technology gener-
ation (1986, see Chapter 12). Nevertheless, the ability of this system to advance the
agricultural research frontier remains uncertain.
POLICY CHALLENGES AND STRUCTURAL CHANGE IN THE 1980S
Introduction
Change in the international environment during the past 15 years appears to have
reduced the contributions of the U. S. national innovation system to growth in domes-
tic incomes and competitiveness. Through the early 1970s, the returns to publicly
funded research were more easily captured by U. S. firms because of their considerable
technological lead over foreign firms and because they performed much of the publicly
funded research. In addition, the commercial competitiveness of U. S. firms in some
industries may have received less support in recent years from defense spending than
was true of the 1950s and 1960s.
A postwar process of convergence in both the economic and technological
spheres has brought many foreign economies to levels of income per capita, produc-
tivity, and R&D investment that approach or exceed those of the United States (see
Cyert and Mowery, 1987; Mowery, 1988; or Nelson, 1990, for more detailed discus-
sion). Convergence at the aggregate level, however, is not matched by any uniform
pattern of decline in the relative strength of U. S. performance in all industrial sectors.
As Nelson (1990) points out, the United States has preserved its export and patenting
"market share" in a number of high-technology sectors, although in such industries as
scientific instruments, consumer electronics, or steel and automobiles, U. S. exports
have declined dramatically. Since 1973, moreover, a number of indicators suggest that
U. S. living standards (e.g., real earnings) have stagnated or declined in real terms, and
aggregate productivity growth has remained low. These developments, which have
both contributed to and have been exacerbated by inept U. S. macroeconomic policy,
have affected U. S. debates on trade and technology policies. Restrictive trade policies
and nationalistic (or xenophobic) responses to foreign investment in the United States
have gained considerable currency in recent years, although the extensive offshore
markets and investments of many U. S. multinational firms have tempered any moves
to completely cut off foreign access to U. S. markets. Science and technology policies
also are being affected by these economic and political trends.
Since foreign firms now are more technologically sophisticated and technology is
more internationally mobile, the competitive advantages that accrued in the past from
U. S. basic research and a strong knowledge base have been eroded. Faster interna-
tional transfer of new technologies has undercut a major source of America's postwar
superiority in high-technology markets.35 Both the public and private sectors have
responded to these changes in the environment. Private firms have pursued a number
of new organizational approaches to exploit R&D and innovation outside of the
firmthese include domestic and international consortia or alliances and domestic
university-industry research linkages. The federal government also has undertaken
new initiatives in research funding, trade policy, and intellectual property protection,
in order to increase the domestic economic returns to public and private R&D invest-
ments in the United States. These policy initiatives may reduce some of the structural
contrasts between the United States and other national innovation systems, if the
influence of antitrust policy on U. S. firms' technology strategies becomes less signifi-
cant and if the importance of new firms in technology commercialization declines.
The Growth (Rebirth?) of University-Industry Research Cooperation
During the past decade, financial support from industry has established a number of
research facilities on university campuses to conduct research with potential com-
mercial value. Important initiatives are coming from the federal government as well
as private industry. The National Science Foundation has embarked on a program to
establish a number of interdisciplinary research centers in engineering and other sci-
entific disciplines on university campuses. The financial structure of these centers also
is relatively novel, since it combines "seed-money" support from the federal govern-
ment (as well, in many cases, as state and local governments) with major contributions
from private corporations that are affiliated with the centers.
The phenomenon of university-industry research collaboration is not new, hav-
ing been well-established before 1940. Indeed, the share of university research expen-
ditures financed by industry appears to have declined through much of the postwar
period. In 1953, industry financed 11% of university research, a share that declined to
5.5% in 1960 and 2.7% in 1978. By 1985-1986, estimates suggest that industrial funds
accounted for no more than 5% of university research. The recent development of
closer research ties between universities and industry represents a restoration of a link-
age that was weakened during the 1950s and 1960s, rather than a fundamental depar-
ture.36
There is a vast array of forms of research collaboration between universities and
industry, making generalizations virtually impossible. The relationship between uni-
versity research and commercial technology varies considerably across industries. No
single model or description of the constraints, advantages, and disadvantages of such
collaboration is likely to be accurate for all university-industry collaborations.
A fundamental motive for closer ties between university and industrial research
is the fact that U. S. universities account for a growing share of total U. S. basic
research. In 1953, industry accounted for 58% of the combined basic research budget
(from all sources) of the universities and industry; in 1978, universities accounted for
76% of the combined basic research budget of universities and industry (National
Commission on Research, 1980, pp. 8-9). Nevertheless, private industry financial
support of university research remains modest. Industry provided about 20% of the
funds for all basic research in the mid-1980s, but it funds a much smaller percentage
of the basic research performed at the universitiesaround 6% of the total.
The growing role of U. S. universities in the performance of basic research has
been associated with a recognition by U. S. industry that more fields of research at the
universities now hold out significant promise of generating findings that may be of
great commercial significance. The connection between university research and com-
mercial technology appears to be particularly close in biotechnology, a factor that
influences the character of many university-industry research relationships in this
field, and may distinguish them from university-industry research collaborations in
other fields.37 Increased pressure to reduce R&D costs, to monitor a wider range of
emerging areas of scientific research, and to speed the commercialization of scientific
research has driven many U. S. firms to attempt to develop relationships with an array
of external institutions (see below for additional discussion), including research uni-
versities in the United States and abroad, to complement and enhance the pay off from
their in-house R&D activities.38
By virtue of their mission as educational, as well as research, institutions, U. S.
universities are critically important sources of scientific and engineering personnel.
Participant firms can employ collaborative ventures as "filters" for hiring research per-
sonnel, observing the performance of potential researchers before making employ-
ment commitments. Moreover, the importance of people as vehicles for the transfer
of scientific and technological knowledge means that the hiring by firms of the grad-
uates of these programs facilitates the transfer of knowledge and technology from uni-
versity to industry. The interdisciplinary character of modern technological and
research challenges makes this training and hiring benefit particularly important.
The growing perceived economic significance of U. S. university research, com-
bined with expanded funding of such research by non-U. S. firms, have sparked con-
cerns that foreign firms are "unfairly" gaining access to high-quality research within
the relatively open U. S. university system (the "unfairness" presumably stems from
the belief that foreign, especially Japanese, universities offer fewer opportunities for
U.S. firms to gain access to world-class scientific and engineering research) allows for-
eign firms to improve their competitiveness vis-a-vis U. S. firms (e.g., Business Week,
1989, pp. 73-74). Congressional and state legislative policymakers have raised ques-
tions about programs such as the Industrial Liaison Program of M.I.T., through which
foreign firms receive briefings on academic research in exchange for financial contri-
butions.
Although the empirical basis for this criticism appears to be rather weak and no
action has yet been taken by either state or the federal governments, the debate under-
lines the complexities introduced by the increased quantity and rate of international
flows of scientific and technological knowledge. When national innovation systems
vary in their fundamental structure, a "level playing field" in this sphere is appealing
in the abstract and very difficult to create in reality. Nevertheless, as policymakers and
managers alike increasingly view national innovation systems as important strategic
assets in global economic competition, demands for such "leveling" are likely to
expand, as we note below.
Industrial Research in an Era of Structural Change

The structure of the U. S. industrial research system also appears to be undergoing
significant change. Although the outcome of current experiments and reorganization
cannot be predicted with confidence, competitive and cost pressures appear to be lead-
ing a number oflarge U. S. firms to reorganize their corporate research activities and,
in particular, to exploit external sources of new technologies more aggressively. The
growth in university-industry research cooperation is one example of thisothers
include international collaborative ventures and domestic research consortia. Even as
the historic dominance of industrial research by the in-house research laboratory may
be declining, changes in the policy and competitive environment may also reduce the
role of new firms in the commercialization of new technologies within the U. S. econ-
omy. This section discusses both of these developments in what must be a speculative
and tentative treatment.
Restructuring U. S. Industrial Research?

One of the most widespread forms of institutional innovation in the U. S. national
innovation system reflects the effort by firms to develop external sources of research
and development expertise. These efforts have resulted in considerable expansion of
collaboration in R&D that involves U. S. and foreign firms, as well as U. S. uni-
versities.
U. S. firms have expanded both international and domestic collaborative research
efforts, and both types of collaboration are responses to the same factors: the rising
costs and risks of product development, the increased breadth of the scientific and
technological knowledge base needed to compete in high-technology industries
(reflected in instances of "technological convergence" such as the interpenetration of
telecommunications and computer technologies, biotechnology and Pharmaceuticals,
etc.), more rapid product cycles in some industries, and more severe competitive pres-
sure from foreign firms.39 In addition to these factors common to both types of collab-
oration, change in the prospects for new startup firms in some high-technology indus-
tries may be increasing their predisposition to pursue technology commercialization
through collaboration with larger domestic or foreign firms, rather than pursuing this
goal independently (see below for further discussion). In some instances, these collab-
orative ventures have resulted in the acquisition of the smaller startup by the
larger firm.
The consequences of these new organizational structures are still uncertain, since
in most cases collaboration is less than 10 years old. Nevertheless, there appear to be
important contrasts in structure and motives between the international and domestic
research and development ventures into which U. S. firms have entered. One must
distinguish among at least three broad categories of research collaboration: collabo-
rative ventures between U. S. and foreign firms, research collaboration among U. S.
firms, and domestic university-industry research collaboration. International collab-
orative ventures focus mainly on development, production, and marketing, rather
than precommercial research. Thus far, domestic collaborations among U. S. firms
have been concerned with research that is less applied in character and less closely
linked to a specific commercial product. In general, however, these domestic collab-
orative ventures do not focus primarily on basic research despite the intentions and
founding aspirations of several of them. Finally, university-industry research collab-
oration appears to incorporate scientific and engineering research of a more funda-
mental character. These different forms of collaboration are not substitutes but com-
plements. Large U. S. corporations in the computer and electronics industries, for
example, appear to utilize all three forms of collaboration simultaneously.
International collaborative ventures involving U. S. firms rarely focus on the
"precommercial" research activities that are the intended target of the efforts of
domestic research consortia among firms and university-industry collaborations.
These ventures are a response by U. S. firms to the factors noted above as causes of
domestic research collaboration, but in addition reflect the growing technological
strengths of many foreign firms and the increasingly prominent role of nontariff trade
barriers and government support for the development of advanced technologies. Inter-
national collaboration has grown dramatically in a diverse array of U. S. manufactur-
ing industries, although its long-term prospects appear brighter in some industries
than in others. The consequences for U. S. competitiveness of such collaboration thus
far appear to be fairly benignin most cases, international collaboration is a response
to, rather than a cause of, declining U. S. competitiveness. In other industries, collab-
oration has assisted U. S. firms in strengthening their technological and production
skills.
International collaborative ventures create some complex policy issues in the
trade and technology policy spheres. Along with numerous other factors, international
collaborative ventures will accelerate international technology transfer in the global
economy of the future. Nevertheless, the technology transfer operating within inter-
national collaborative ventures involving U. S. firms is likely to remain a two-way flow
for the foreseeable futurethere are few documented examples of U. S. firms "giving
away" critical technological assets to foreign enterprises within these ventures. Reflect-
ing the close and interdependent relationship between trade and technology policies,
trade policy clearly influences the development of international collaborative ven-
tures. Indeed, international private collaborative ventures appear to be an important
complement to the growing number of "closed" (to foreign firms) cooperative tech-
nology development programs sponsored by governments in the industrial world.40
Increasingly, however, trade policy formulation and implementation will be affected
by the operation and existence of these venturesnotably, the internationalization of
sources of R&D and components.
Change in the Role of the Startup Firm?

The revival of faith in the "magic of the market" within U. S. policymaking circles
during the 1980s paradoxically has been combined with more frequent expressions of
concern over the impact of the efficient U. S. capital market on the growth of new,
high-technology firms.4' The U. S. venture capital market and other institutions that
have spawned innovative firms and new industries now are criticized by some observ-
ers for breeding "managerial myopia" (an excessive focus on short-term results) and
providing insufficient support for technology development over the long run. Other
critics suggest that the reliance on startup firms for the development of new technol-
ogies within the U. S. economy has resulted in excessive transfer of technological
know-how to the foreign firms (often, Japanese firms) that recently have expanded
their investments in these enterprises. In this view (which is by no means universally
held), many of the widely remarked difficulties of the U. S. economy in commercial-
izing new technologies stem in part from an excessive reliance on startup firms for
commercialization.
A number of factors may change the future role of the startup firm in the com-
mercialization of technologies within the U. S. economy. Although the evidence on
both of these hypotheses is mixed, it is possible that the rate of formation of new high-
technology firms may decline or that these new firms will less frequently develop into
large commercial enterprises. Some recent accounts suggest that venture capital mar-
kets now are a less important source of support for startup firms, as a result of the
increasing costs of new product development (especially in biotechnology and micro-
electronics), and the attraction of alternative investment opportunities.42 Acquisition
of startups by established U. S. and foreign firms also has become more common. Par-
ticularly in biotechnology, new firms rarely have developed into mature enterprises,
instead being acquired.43
The changing public policy environment in the United States also may reduce
the role of startup firms in the future. To the extent that postwar U. S. antitrust policy
before the 1980s tended to discourage acquisitions by large industrial firms as a means
to develop and commercialize new technologies, relaxation of this policy under the
Reagan and Bush Administrations (see below) may increase the likelihood that startup
firms will be acquired by larger enterprises, rather than remaining independent. Sim-
ilarly, the effects of the efforts of the 1980s to strengthen domestic protection for intel-
lectual property may reduce the viability of startup firms in at least one industry in
which they have been very important. The passage of the Semiconductor Chip Pro-
tection Act of 1984 significantly strengthened protection for chip designs in microe-
lectronics. In response, established semiconductor firms have become far more willing
to sue startup and established firms over alleged infringement, and startup firms have
been especially hard hit (Gupta, 1988). Intellectual property protection in biotech-
nology and computer software remains far more uncertain, however, and litigation
over intellectual property may pose less of an obstacle to the establishment of new
firms in these industries.
Change in the structure of markets for high-technology goods also has diminished
the ability of new firms to grow considerably in size and scope. The U. S. military mar-
ket no longer plays a strategic role in the computer and semiconductor industries com-
parable to its position in the 1960s, and the possibilities for military-civilian technol-
ogy spillovers appear to have declined in many areas of these technologies.
Biotechnology firms in particular are attempting to enter industries (e.g., food pro-
cessing or Pharmaceuticals) in the U. S. and foreign economies that are far more heav-
ily regulated than was true of semiconductors. As a result, the costs of new product
introduction and the marketing-related entry barriers faced by these firms have risen,
even as the willingness of prospective purchasers to provide capital for production
facilities has declined. For this and other reasons, including the greater interest by for-
eign firms in the technological assets of U. S. startup firms, collaborative ventures
involving startup and established U. S. and foreign firms have grown considerably in
recent years (Pisano et al., 1988). These ventures frequently focus on technology
exchange (often combined with the acquisition by an established firm of a substantial
portion of the equity of the new firm) and/or marketing (including navigating domes-
tic and foreign product regulations), rather than joint development of new products.
Summarizing this speculative discussion, the strength and effects of these changes
in the economic environment of the startup firm remain uncertain. If the formerly
significant role of the startup firm in high-technology industries does diminish, a key
contrast between the U. S. national innovation and those of many other nations will
be reduced in importance.
Public Policy Initiatives
Along with the governments of other industrial and industrializing nations, the U. S.
government during the past decade has recognized the important role of innovation
in economic performance and has begun to take steps intended to increase the domes-
tic economic payoff from the large public and private U. S. investment in R&D. The
contrast between the position of the newly elected Reagan Administration in 1981,
denying any role for the federal government in the development and commercializa-
tion of new civilian technologies,44 and the Reagan Administration of 1987-1988, is
dramatic. By 1987-1988, the Administration had organized a symposium on com-
mercial applications of high-temperature superconductivity (HTS) that was restricted
to U. S. nationals, had proposed legislation for speeding the development of commer-
cial applications of HTS that included provisions to restrict access by foreign nationals
to the results of publicly funded basic research in the U. S., and had launched two
military-funded research programs in civilian technology development.
These initiatives, which had broad bipartisan support, represented a considerable
shift in the focus of Federal programs aimed at civil technology development. Previous
postwar federal programs to support civil technology development in such areas as
energy, nuclear energy, and even housing construction, typically were aimed at tech-
nologies for which policymakers felt that market incentives were insufficient, causing
the social returns to substantially exceed the private returns to the innovator. Federal
programs designed to aid in the development of commercial applications of scientific
discoveries such as HTS, however, are aimed at technologies for which the private and
social returns to innovation both may be very high. Instead, these programs attempt
to influence the distribution of the private returns between U. S. and foreign firms.
Although the Bush Administration has expressed opposition to aggressive public sup-
port for commercial technology development as "industrial policy," the Administra-
tion is not unified in this opposition, and bipartisan Congressional pressure for action
remains strong.
The Changing Relationship Between Military and Civilian Technologies
As was noted earlier, military technologies (and therefore military R&D and procure-
ment budgets) may now be less fertile sources of commercial technologies than was
true during the 1950s and 1960s. New technologies in some important areas now
appear first in civil applications and are subsequently incorporated into weapons sys-
tems. Accordingly, several Pentagon research programs, such as Sematech, have
focused on supporting the development of commercial technologies, in the belief that
these technologies will yield advances in military applications.
A second important influence on recent Pentagon research and technology devel-
opment initiatives is the changing market for the products of U. S. high-technology
firms. As the military share of total demand for many high-technology systems and
components (especially in computers and microelectronics) has declined, the eco-
nomic viability of major defense suppliers increasingly depends on their ability to
compete in civilian markets. Recent military-funded research programs thus are
designed to address the commercial competitive strengths of U. S. firms.
The Sematech consortium is dedicated to the development of advanced manu-
facturing processes for commercial memory chips, not military components. Similar
concerns led to the 1988 foundation of the National Center for Manufacturing Sci-
ences (NCMS), funded in part by the military to support research in manufacturing
technologies and to the $30 million Defense Advanced Research Projects Agency
(DARPA) research program in high-definition television (HDTV) technologies begun
in 1989.45 Many (although not all) of the applications of HDTV will be in civilian mar-
kets. Both Sematech and the NCMS exclude foreign firms, and DARPA programs in
HDTV also are open only to U. S.-owned firms (Wolf, 1989; Mowery and Rosenberg,
1989b).
These programs have considerable political support, but it is difficult to predict
whether they will spawn similar initiatives in other technologies. Several observers
have suggested that U. S. military research and procurement programs should be struc-
tured and managed to strengthen U. S. producers of civilian high-technology products
(Gansler, 1989). Civilian and military policymakers in the Pentagon, however, may
resist any diversion of funds in a flat or shrinking budget from military to "dual-pur-
pose" R&D and procurement programs. The emphasis within these new military
R&D programs on civilian technology development has some resemblance to earlier
European initiatives. Indeed, a central justification for the subsidies extended by West-
ern European governments to the Airbus Industrie consortium is the desire of these
governments to maintain a military aerospace industrial base by supporting the com-
mercial activities of their national firms. Needless to say, the presence of a similar pro-
gram in the U. S. microelectronics industry will impede the efforts of U. S. trade pol-
icymakers to reduce European government subsidies to Airbus.
In addition to their unhappy implications for trade policy, the recent popularity
of military research and procurement programs as instruments for improving the civil-
ian technological and commercial capabilities of U. S. firms overlooks the poor track
record of similar programs in Western Europe. Airbus is a technological and political
(not financial) success, but it is exceptional in a list of costly programs supporting
weapons development by "national champions," or European consortia of "cham-
pions," that yielded little by way of improved military security or enhanced commer-
cial competitiveness.46
Intellectual Property and Antitrust

Other recent U. S. initiatives in technology policy have improved enforcement of intel-
lectual property protection and reduced antitrust restrictions on collaboration in
research. The 1982 legislation that established the Court of Appeals for the Federal
Circuit has strengthened the protection granted to patentholders.47 The U. S. govern-
ment has also pursued stronger international protection for intellectual property rights
in both bilateral and multilateral international trade negotiations.
In antitrust policy, the Reagan Administration adopted a substantially more
lenient enforcement posture than its predecessors, arguing that international compe-
tition had significantly reduced the dangers of market power being acquired through
domestic merger and acquisition activity. Justice Department guidelines and review
procedures for mergers were relaxed somewhat, and major federal antitrust suits
against high-technology firms were dropped or settled in the early 1980s. The Reagan
Administration supported the 1984 National Cooperative Research Act, which
reduced the antitrust penalties for collaboration among firms in precommercial
research. The NCRA has been credited with easing the founding of the Microelectron-
ics and Computer Technology Corporation, an early reesarch consortium involving
U. S. computer and electronics firms.
The number of research consortia in U. S. industry has grown since the NCRA's
passage, with 111 cooperative ventures registered under the terms of the Act from 1984
through June 1988. Some critics of the NCRA argue that more consortia would have
been established, or existing consortia would be more effective, if the Act's protection
against treble-damage antitrust penalties were extended beyond precommercial
research. In this view, consortia that are restricted to research cannot move quickly to
profitably apply new technologies commercially.48 In response to this criticism, legis-
lation has been introduced in the Congress to reduce antitrust penalties against con-
sortia that engage in production, a proposal that has gained White House support.49
In both antitrust and intellectual property policy, the Reagan and Bush Admin-
istrations are strengthening the returns to innovators. The policy initiatives in antitrust
have also been influenced by the example of Japanese success in cooperative research
and technology development, although the recent antitrust legislative proposals
extend the scope of cooperation well beyond the precompetitive research stage in
which Japanese firms most often collaborate. This policy focus, however, fails to
address one of the most serious weaknesses of the U. S. national innovation system,
the slow pace of domestic adoption of new technologies in manufacturing. Policies
designed to increase the rewards to innovators in some instances will increase the costs
associated with the adoption of the technologies produced by the innovators, and
thereby may hamper diffusion (David, 1986). Moreover, the "lessons" drawn from the
Japanese experience by U. S. policymakers appear to overlook the emphasis within
Japanese industry and technology policies on support for domestic technology adop-
tion and on strong domestic competition in technology commercialization.
The Merger of Technology and Trade Policies

The growing political salience of national science and technology policies has blurred
the boundaries between U. S. technology and trade policies and has complicated pol-
icy formulation in each area. Technology-intensive industries now are at the center of
bilateral trade disputes and negotiations. The commercial aircraft, telecommunica-
tions equipment, computer, and microelectronics industries all were the subjects of
special provisions in the 1988 Omnibus Trade and Competitiveness Act.50
At least one recent international negotiation dealt directly with the structural dif-
ferences between the U. S. and foreign national innovation systems. Negotiations over
renewal in 1988 of the U. S.-Japan Agreement on Scientific Cooperation, formerly of
concern only within the scientific community, for the first time involved trade poli-
cymakers in both governments. The central trade-related issues in these talks con-
cerned intellectual property rights within Japan and the assurance by the Japanese
government of access by U. S. firms to publicly funded research in Japanese labora-
tories that was comparable to Japanese firms' access to publicly funded research in
U. S. research facilities.51
Still another illustration of the influence on U. S. trade policy of the evolving tech-
nology policy agenda is intellectual property rights, which are a high priority for U. S.
negotiators in the Uruguay Round of multilateral trade negotiations. Bringing intel-
lectual property issues into trade policy (as has been done through Section 301 of the
Trade Act and may be done in the future, if U. S. goals in the Uruguay Round are
achieved, through the General Agreement on Tariffs and Trade) provides a powerful
enforcement mechanism, restrictions on market access for the products of nations that
provide insufficient protection, that current multilateral agreements lack. U. S. nego-
tiators have also pursued this issue in bilateral negotiations with Thailand, Taiwan,
South Korea, and other nations. Faced with the threat of restrictions on their firms'
access to U. S. markets, these and other foreign governments have revised domestic
policies to achieve standards of protection and breadth of coverage comparable to
those in the United States. The intellectual property rights issue also illustrates the sig-
nificant extension of the reach of trade policy issues beyond the borders of trading
nations into domestic policy.
Concluding Remarks
At least some of the changes in the structure of the U. S. innovation system that are
foreshadowed by these organizational experiments and public policy initiatives may
revive several of the elements of this system that were characteristic of the pre-1940
period. The high-technology startup firm, for example, is normally associated with the
postwar era; as was noted earlier, before 1940 much of the in-house research activity
of large manufacturing firms focused on the acquisition of technologies from smaller
entrepreneurial firms (or from individual entrepreneurs), often including the acqui-
sition of the firm. Similarly, the expansion in university-industry collaboration and in
the role of state governments in supporting these and other research activities involves
the revival of another key component of the pre-1940 system. Still another element of
the pre-1940 system, and a far less desirable one, that has been strengthened somewhat
in recent years is the protectionist and isolationist tone of some contributions to the
U.S. debate over trade and technology policies.
CONCLUSION
As was noted in the first part of this paper, the importance of each of three key com-
ponents within the U. S. national innovation system has changed over the course of
this century. In the early twentieth century research facilities in both universities and
industry were established and informal linkages between at least some universities
(often, public universities) and industrial research establishments were developed. The
federal government played a modest role as a supporter of research in the nonagricul-
tural sector, and state governments funded both public higher education and the
"engineering extension" activities of many of these universities.
The research system that had developed within U. S. industry and academia by
the end of the 1930s exceeded the size of its British counterpart, the only non-U. S.
system for which reliable data exist during this period, and probably outstripped that
of any other industrial economy, with the possible exception of Germany. Nonethe-
less, the quality of U. S. fundamental research in academia and industry was only spo-
radically on a par with the best of German or British research. The rise of the U. S.
economy to a position of world leadership in manufacturing output and productivity
during the first 15 years of this century did not rely on world-class domestic scientific
research (Nelson, 1990).
This structure was transformed beyond recognition by World War II and the state
of armed peace that developed within 5 years after the end of hostilities. Federal
research funding expanded and displaced the role of state governments as actors in this
innovation system and contributed to some weakening in the informal ties that linked
many corporate and academic research institutions. The powerful role of the federal
government within the postwar U. S. innovation system was not linked to any eco-
nomic strategy, however, instead being motivated largely by national security con-
cerns. During both the pre-1940 and postwar eras, little if any strategic planning
underpinned public intervention in the U. S. innovation system, and policymakers
devoted minimal attention to its domestic economic payoffs.
Whether or not the policy was based on a comprehensive strategy, the interaction
between federal and private R&D expenditures significantly influenced the perfor-
mance of the U. S. national innovation system during the postwar period. A large,
well-financed federal defense R&D program increased the demand for a limited supply
of professional engineers and scientists. Although that demand also expanded the sup-
ply of such trained persons (expansion aided by federal research and fellowship funds
for higher education), it also raised their wages and salaries and increased the cost of
privately financed R&D.52
Comparisons with other countries provide some support for concern over the
commercial implications of defense R&D spending. Japan and West Germany have
had very small military budgets since the end of World War II, and in both countries
the ratios of civilian R&D to GNP have been substantially higher than in the United
States for many years (National Science Board, 1981, pp. 214-215). The opportunity
costs to the U. S. economy of high levels of defense R&D spending have been high.
The relative economic performance of the United States, West Germany, Japan, and
other advanced industrial economies in the past few decades does not support the pre-
sumption that large expenditures on military R&D have improved this nation's eco-
nomic strength. Moreover, recent efforts by the military services to broaden their sup-
port of commercial innovation may create as many problems as they solve.
The historical perspective of this paper suggests that U. S. antitrust policy has
exercised an important and (by comparison with other industrial nations) unique
influence on the U. S. innovation system. Antitrust policy was partly responsible for
the formation of the giant firms that were among the earliest investors in industrial
research, and influenced their decisions to invest in industrial research. For most of
the post-1945 period, antitrust policy remained an important influence on corporate
R&D investment by large firms. In the case of Du Pont, R&D investment soared as
the firm attempted to develop "new nylons" (largely ignoring the fact that with the
exception of nylon and neoprene, the bulk of the firm's prewar innovations had been
acquired, rather than originating in Du Pont fundamental research). AT&T pursued
liberal licensing policies in microelectronics as a direct result of antitrust policy, lead-
ing to rapid entry into the semiconductor industry by new firms.
Antitrust policy may have led to higher postwar levels of industry-financed
investment in R&D; this investment, however, did not necessarily improve the eco-
nomic performance of the U. S. innovation system. Nevertheless, viewed in terms of
its effects on the domestic diffusion of new technological knowledge and findings, post-
war antitrust policy may well have aided the performance of the U. S. innovation sys-
tem, in view of the support that this policy provided for the important role of small,
startup firms in technology commercialization. Whether this structure is equally well-
suited to a global environment in which technological knowledge travels almost as rap-
idly across national boundaries as it does within these boundaries remains uncertain
(see Nelson, 1990, for additional discussion).
Recent and prospective changes in U. S. antitrust policy may reduce the extent
of the contrast between U. S. and foreign nations' antitrust policies. Combined with
other policy initiatives in intellectual property, a more lenient U. S. antitrust policy
may reduce the importance of new firms as agents for the commercialization of new
technologies. The effects of these policy changes on the adoption of new technologies,
an area in which U. S. performance has been weak, are uncertain but may prove det-
rimental. More generally, however, the current U. S. debate over science and tech-
nology policy exhibits little awareness of the importance of technology adoption for
international competitiveness, nor is there much acknowledgment of any role for pub-
lic policy in supporting technology adoption. As the case of agricultural research
makes clear, public programs designed to support technology adoption may perform
less well in scientific research. In addition, of course, the contemporary U. S. debate's
focus on trade and technology policies as containing the entirety of a "solution" to
problems of competitiveness is almost certainly misguided.
In all three of the major industrial regions of the world economyJapan, the
United States, and Western Europefar-reaching structural change now appears to
be occurring in national innovation systems. In Western Europe, national innovation
systems may be supplemented by a regional one, under the sponsorship of BRITE,
EUREKA, RACE, JESSI, and other programs. Japan's transition from a position of
"catch-up" within the global economy to a position of technological leadership in
many areas also is likely to require new institutions for technology development and
commercialization. The movement of the U. S. economy from a position of global
technological and economic dominance to the position of first among equals is inten-
sifying political debate over the trade and technology policies of this government
through most of the postwar period. This debate paradoxically has been fueled by the
success of the multilateral political and economic strategies, to which U. S. trade and
technology policies contributed, in the demise of the Warsaw Pact.
Within the U. S. debate, the economic and other consequences of increased inter-
national technological interdependence in both the civilian and military spheres are
only dimly perceived. It seems inevitable nonetheless that technology policy issues will
figure ever more prominently on the trade and national security agenda. Even as the
U. S. and other governments attempt to intervene strategically in their "national inno-
vation systems" (treating the European Communities as a single such system), the
growing economic and technological interpenetration of the major industrial and
industrializing economies of the world appears to be making these "national" systems
increasingly permeable.
Unfortunately, although the increased concern of a number of recent U. S. tech-
nology policy initiatives with commercial development of the fruits of basic research
investments arguably is a positive development, the mercantilistic flavor of many of
them is not. Proposals to restrict scientific and technological cooperation at the water's
edge fly in the face of the growing interdependence of national R&D systems. To the
extent that U. S. policymakers design technology initiatives that ignore the increasing
interdependence of U. S. and foreign scientific and technological research, U. S. and
foreign technological and economic development will be hampered.
At the international level, the challenge to policymakers seeking to manage inter-
national interdependence is spilling across all areas of economic policy. The devel-
opment of macroeconomic policy coordination among the industrialized nations has
required nearly two decades, and still leaves much to be desired. The problems of inter-
national interdependence and convergence in microeconomic policies and institu-
tions are only now beginning to receive attention (Ostry, 1990). In the multilateral
trade talks that almost certainly will begin shortly after the conclusion of the Uruguay
Round, a multilateral antitrust or competition policy is likely to be a central topic,
which may provide further impetus to some convergence in the structure of industrial
nations' national innovation systems.
Although economic factors may be forcing some convergence in the structure of
national innovation systems within the industrial world (just as these forces have
forced some convergence in the financial systems of the industrial economies), the
speed and effectiveness with which they will operate are uncertain, especially in view
of the political sensitivity of many of the affected issues and interests. Moreover, our
understanding of the management and organization of the innovation process is so
imperfect that debates over "fair" and "unfair," "open" and "closed," or "efficient"
and "inefficient" innovation systems will remain poorly informed for the foreseeable
future.
NOTES
1. There is a growing literature on this topic, although it focuses on a small number of tech-
nologies and often does not present internationally comparable data. See, among others, Flamm
(1988), Edquist and Jacobsson (1988), Kelley and Brooks (1988), and Jaikumar (1987).
2. "[ T]he coupling between science and technological innovation remained very loose dur-
ing this period [the nineteenth century] because, in many industrial activities, innovations did
not require scientific knowledge. This was true of the broad range of metal-using industries in
the second half of the nineteenth century, in which the United States took a position of distinct
technological leadership. Indeed, following the American display at the Crystal Palace Exhibi-
tion in 1851, the British came to speak routinely of'the American system of manufactures.'. . .
In the second half of the nineteenth century, America provided the leadership in developing a
new production technology for manufacturing such products as reapers, threshers, cultivators,
repeating rifles, hardware, watches, sewing machines, typewriters, and bicycles" (Mowery and
Rosenberg, 1989b, p. 27).
3. Concerning the characteristics of the trajectory of technological advance within the
U. S. economy, Abramovitz (1986) notes that "The path of technological change which in those
years [1870-1945] offered the greatest opportunities for advance was at once heavily scale-
dependent and biased in a labor-saving but capital- and resource-using direction. In both
respects America enjoyed great advantages compared with Europe or Japan. Large-scale pro-
duction was favored by a large, rapidly growing, and increasingly prosperous population. It was
supported also by a striking homogeneity of tastes. This reflected the country's comparative
youth, its rapid settlement by migration from a common base on the Atlantic, and the weakness
and fluidity of its class divisions. Further, insofar as the population grew by immigration, the
new Americans and their children quickly accepted the consumption patterns of their adopted
country because the prevailing ethos favored assimilation to the dominant native white culture.
At the same time, American industry was encouraged to explore the rich possibilities of a labor-
saving but capital- and resource-using path of advance. The country's resources of land, forest,
and minerals were particularly rich and abundant, and supplies of capital grew rapidly in
response to high returns" (p. 397).
4. See Stigler (1968). The Supreme Court ruled in the Trans Missouri Association case in
1898 and the Addyslon Pipe case in 1899 that the Sherman Act outlawed all agreements among
firms on prices or market sharing. Data in Thorelli (1954) and Lamoreaux (1985) indicate a
sharp increase in merger activity between the 1895-1898 and 1899-1902 periods.
5. The Du Pont Company's research activities focused increasingly on diversification out
of the black and smokeless powder businesses even before the antitrust decision of 1913 that
forced the divestiture of a portion of the firm's black powder and dynamite businesses. Discuss-
ing Du Font's early industrial research, Hounshell and Smith (1989) argue that "Du Font's initial
diversification strategy was based on utilizing the company's plants, know-how, and R&D capa-
bilities in smokeless powder (i.e., nitrocellulose) technology. The goal was to find uses for Du
Font's smokeless powder plants because political developments in Washington after 1907 [Con-
gressional restrictions on procurement by the Navy of powder from "trusts" and the 1913 anti-
trust decision that forced divestiture of much of Du Font's black powder and dynamite opera-
tions] signaled a significant decline, if not end, to Du Font's government business" (p. 57). The
1911 consent decree settling the federal government's antitrust suit against General Electric left
GE's patent licensing scheme untouched, enabling the firm to maintain an effective cartel within
the U. S. electric lamp market for years to come without further acquisitions, which would have
violated the consent decree (Bright, 1949). During the interwar period, Du Pont and General
Electric both utilized patent licensing arrangements as a basis for international cartel agreements
(see Taylor and Sudnik, 1984; Reid, 1989).
6. The ability of firms to utilize their industrial research laboratories to monitor their tech-
nological environments and evaluate patents or firms for possible acquisition was aided by
stronger protection for intellectual property in the late nineteenth century. Stronger intellectual
property protection increased the appropriability of the returns from innovation and facilitated
the development of a market for the acquisition and sale of patents. Federal court decisions in
the 1890s upholding the validity of patents covering goods not in production increased the utility
of large patent portfolios for defensive purposes.
7. The data in Table 2.1 were drawn originally from the National Research Council surveys
of industrial research employment, as tabulated in Mowery (1981). The surveys' coverage of
research laboratories in 1921 is somewhat suspect, and the data for that year should be treated
with caution.
8. "Universities" throughout this paper refers to all institutions of higher education, and
therefore covers a broader and more diverse array of institutions than those defined as univer-
sities in other national innovation systems.
9. In the early 1920s, roughly 42,000 students were enrolled in British universities; the fig-
ure rose to 50,000-60,000 by the late 1930s. By contrast, American institutions of higher learn-
ing awarded over 48,000 degrees in 1913 alone, nearly 10 years earlier, and more than 216,000
degrees in 1940. With a total population 35% that of the United States, Britain had only about
6% as many students in higher education in the late 1930s (Briggs, 1981; U. S. Bureau of the
Census, 1975, p. 386). The size of the higher educational system was an important "supply-side"
influence on the growth of German industrial research; Beer (1959) cites the high rate of pro-
duction of chemistry Ph.D.s by German higher education in the late nineteenth century as an
important influence on the growth of industrial research in the German chemicals industry. As
the supply of professional chemists exceeded available academic employment opportunities,
emigration or industrial research were the only alternatives open to the German graduate
chemist.
10. Even in the private universities, however, the applications of scientific research, rather
than the intrinsic importance of science, were emphasized by academic administrators. Cohen
(1976) notes that "By the mid-1840's the stress on practicality produced schools of science at
both Harvard and Yale in which a dominant theme was the utility of the sciences" (p. 374).
11. "[FJrom comparative obscurity before World War I, American chemistry rose steadily
in esteem to a position of international dominance. Almost half the citations in the Annual
Reports [Annual Reports in Chemistry, described on the page as "a central British review jour-
nal"] in 1975 were to American publications. Similarly, almost half the citations to non-Ger-
man-language literature in Chemische Berichte [the "central German chemical journal"] in
1975 went to American work. It is striking that this hegemony is the culmination of a fifty-year
trend of increasing presence, and not merely the result of post-World War II developments. Sec-
ond, it is clear that the increasing attention received in the two decades before World War II
reflected the growing volume of American chemistry, rather than a changed assessment of its
worth. Since World War II, however, in both Chemische Berichte and the Annual Reports,
American chemistry has been cited proportionately more than is warranted by increasing quan-
tity alone. The prominence of American work within the international literature has been sus-
tained by quality" (Thackray et al., 1985, p. 157; emphasis in original).
12. "Around the 1920s, American physics came of age. From a state of distinct inferiority
before World War I it improved until, by the early 1930s, it was the equal or superior of physics
anywhere in the world . . . For example, an early citation study, covering all the reference cita-
tions found in papers published in a number of physics journals in 1934, turned up only 21
citations to papers published in the leading American journal, the Physical Review, in the inter-
val 1895-1914, compared with 169 citations to the German Annalen der Physik. But in citations
to papers published in 1930-33, the Physical Review beat the Annalen three to one, and as of
1933 it had become the most-cited of all physics journals" (Weart, 1979, p. 298).
13. Moreover, to a much greater extent in the United States than elsewhere, technically
trained engineers moved into positions of industrial leadership. See Chandler (1962, p. 317).
14. Evenson (1982) argues that interregional disparities in agricultural productivity and
incomes also were influenced by federal policies toward research investments: "the federal gov-
ernment through its investment decisions has been very influential in changing the research sys-
tem, even though state governments have provided the majority of the funds. In the 1930s and
1940s it located much of its investment in the 'lagging' regions, chiefly the South. In this way it
had a major impact on the regional nature of productivity" (pp. 251-253).
15. The differences in the arrangements during the two world wars between the federal gov-
ernment and the private sector for defense-related research and development had significant
effects on the diffusion of technological know-how during and after each conflict. Limited
involvement by private firms in military R&D during World War I meant that "spillovers" from
military to commercial innovation were limited. World War II appears to have had a different
effect. In the chemicals industry, for example, Hounshell and Smith (1988) argue that World
War II created significant new competitive threats to the Du Pont Company, because of the large-
scale involvement of private firms in the operation of complex chemical production processes:
"Because the wartime emergncy served as a great levelerexposing other companies to truly
large-scale projects and manufacturing operations while forcing Du Pont to yield much of its
proprietary knowledgeDu Pout's executives foresaw that firms such as Allied, Union Carbide,
Monsanto, and Dow would become far more competitive after the war. This competition would
be manifested not only in the marketplace but also in the laboratory" (p. 332).
16. For a careful, although now somewhat dated, treatment of these contractual issues, sec
Danhof(1968).
17. National Science Foundation (1985, p. 3). For a listing of FFRDCs by location and
sponsoring agency, along with federal obligations for 1981, see National Science Board (1983,
p. 310).
18. Congressional Budget Office (1984, p. 53). In addition to expenditures of the Depart-
ment of Defense, defense includes expenditures for military programs in the Department of
Energy.
19. Although no more than 3.2% of federal defense R&D in 1982 went to basic research,
the absolute size of this budget is still very large, and basic research supported by military agen-
cies has been a significant component of federally supported basic research. The Office of Naval
Research has been a supporter of basic research for 40 years, and the Defense Advanced Research
Projects Agency (DARPA) has played a crucial role in the early stages of several research pro-
grams that have yielded significant civilian applications, most notably in computer technology
(seeFlamm, 1988).
20. Ergas (1987) argues that this concentration is a common feature of "mission-oriented"
R&D programs: "the goals of mission-oriented R&D are centrally decided and clearly set out,
generally in terms of complex systems meeting the needs of a particular government agency. .. .
Concentration also extends to the range of technologies covered. Virtually by its nature,
mission-oriented research focuses on a small number of technologies of particular strategic
importanceprimarily in aerospace, electronics, and nuclear energy. As a result, government
R&D funding in these countries is heavily biased toward a few industries that are generally con-
sidered to be in the early stages of the technology life cycle" (p. 194).
21. Flamm and McNaugher (1989) suggest that changes in Defense Department R&D pol-
icy have contributed to declining military-civilian technology spillovers. They cite declines in
the share of basic research in DOD R&D spending, as well as increased Congressional demands
that these R&D programs yield near-term applications in weapons systems, as two factors that
have reduced such spillovers.
22. See National Research Council (1982) and Okimoto and Saxonhouse (1987). Sharp
(1989) argues that the less prominent role played in scientific research by European universities
has contributed to the slower growth of small biotechnology firms: "A researcher at a CNRS
laboratory in France, or at a Max Planck Institute laboratory in Germany, is the full time
employee of that institution. As such his/her prime responsibility is to public, not private science.
Moeover, as a full time employee, he/she will not find it easy to undertake the 'mix' of research
frequently undertaken by an American professor, who combines an academic post with consul-
tancy in the private sector. Indeed the tradition of funding U S academic posts for only nine
months of the year, expecting the academic who wishes to carry out research in the summer to
raise research funds to meet the remaining three months of salary, explicitly encourages the
entrepreneurial academic. In stark contrast, his/her German opposite number at a Max Planck
Institute will find all research costs, including staff and equipment, met as part of institutional
overheads. The opportunity cost of leaving such a research environment for the insecurity of the
small firm is all the greater since, once off the academic ladder in West Germany, it is more
difficult to climb back on again. The same goes for the opposite number in France, and with the
additional disincentive that French researchers are civil servants and dropping out of the system
means both losing security of tenure/accumulated benefits and difficulty in re-entry should the
need arise. In the circumstances, it is not perhaps so surprising that few spin-offs from public
sector research arise, nor, for that matter, that in Europe most such spin-offs are to be found in
the U K, where the organisation of academic science most closely matches that of the U S. In
the U K, it is notable thatwith the exception of Celltech and the Agricultural Genetics Com-
pany (AGC)most of the spin-offs from biotechnology have come from the universities" (pp.
12-13).
23. This is not to deny the major role played by such large firms as IBM in computers and
AT&T in microelectronics. In other instances, large firms have acquired smaller enterprises and
applied their production or marketing expertise to expand markets for a new product technol-
ogy. Nonetheless, it seems apparent that startup firms have been far more active in commer-
cializing new technologies in the United States than in other industrial economies. Malerba's
analysis of the evolution of the microelectronics industry in Western Europe and in the United
States (1985) emphasizes the greater importance of startup firms in the United States.
24. See Perry (1986) and Mowery and Steinmueller (1991). Sharp (1989) argues that "the
venture capital market in Europe is underdeveloped. The most active venture capital market is
in the U K where some half dozen funds specialising in investment in biotechnology are active
and an estimated total of over $ 1 billion invested since 1980 . .. The doyen of this market is the
Rothschild Fund Biotechnology Investments Ltd (BIL)now capitalised at $200 million and
the largest specialist fund in Europe. By contrast, the largest German venture capital fund,
Techno Venture Management, established in 1984, had an initial capitalisation of $10 million
and in 1989 is worth only $50 million. The availability of venture capital, however, is only one
part of the equation. BIL, for example, whose investments span biotechnology and medical tech-
nology, have not found in Europe the quality of investment they are seeking. 75 per cent of their
investments are in the U S, only 25 per cent in Europe, and these concentrated almost entirely
in the U K. This pattern of investment is mirrored by nearly all the investment funds, all of which
invest a large proportion of their investments in biotechnology in the small firm sector in the
U S, and only a very small proportion in small firms in Europe" (pp. 9-10).
25. Discussing the early years of the semiconductor industry, Tiltori (1971) noted that
"The defense market has been particularly important for new firms .. . these firms often have
started by introducing new products and concentrating in new semiconductor fields where the
military has usually provided the major or only market. Fortunately for them, the armed forces
have not hesitated to buy from new and untried firms. In early 1953, for example, before Tran-
sitron had made any significant sales, the military authorized the use of its gold-bonded diode.
This approval has been called the real turning point for the new firms. During 1959, new firms
accounted for 63 percent of all semiconductor sales and 69 percent of military sales" (p. 91).
Describing a similar situation in the early computer industry, Flamm (1988) argues that "the
many start-up computer firms entering the U. S. industry in the early and middle 1950s were
chasing after a reasonably large market, dominated by military demand. For almost of these
producers, the military was the first, and generally the best customer. About eighty different orga-
nizations, including numerous small start-ups that later merged with larger producers or disap-
peared, produced computers in the United States during the 1950s. The U. S. military, or defense
contractors, paid for or purchased the first machines made by most of these groups" (pp. 78-79).
26. "European governments provided only limited funds to support the development of
both electronic component and computer technology in the 1950s and were reluctant to pur-
chase new and untried technology for use in their military and other systems. European govern-
ments also concentrated their limited support on defense-oriented engineering and electronics
firms. The American practice was to support military technology projects undertaken by indus-
trial and business equipment firms that were mainly interested in commercial markets. These
firms viewed their military business as a development vehicle for technology that eventually
would be adapted and sold in the open marketplace" (Flamm, 1988, p. 134).
27. "The happy experience of neoprene and nylon in the 1930s suggested a way in which
Du Pont could stay ahead of the competition, continue to grow, and avoid antitrust litigation.
By expanding its fundamental research effort dramatically, not only in the Chemical Depart-
ment but also in the industrial departments, the Executive Committee expected to reap a harvest
of new nylons" (Hounshell and Smith, 1988, p. 327).
28. See Graham (1986a), Wise (1985), Hounshell and Smith (1985), Sturchio (1985), and
Rosenbloom(1985).
29. Graham (1986b) argued in her analysis of the RCA and Alcoa laboratories that "The
problem both corporate laboratories encountered was that, while they could influence actual
strategy formulation through informal but effective channels, they encountered heavy opposi-
tion at the execution stage. The operating tasks required to develop and commercialize major
innovations were quite different from the tasks involved in other forms of corporate R&D activ-
ity, and the methods and priorities that were effective for these other activities proved dysfunc-
tional for their radical innovations" (p. 189).
30. See Baily and Chakrabarti (1988, especially pp. 42-43), who argue that such a decline
did occur, but attribute it largely to exhaustion of technological opportunities. The empirical
research on the returns to R&D investment, to say nothing of fluctuations in these returns over
time, yields mixed results [see Scherer (1983) and Griliches (1980, 1986)].
31. Markoff (1989) reports that recent survey data from the National Science Foundation
indicate that inflation-adjusted R&D spending by industry shrank by 0.9% between 1988 and
1989, the first reduction in real industry-funded R&D spending since 1974-1975. For a less
gloomy assessment of this development, see the Economist (1990, pp. 65-72).
32. "Prior to the 1970 act roughly a dozen or so public SAES and USDA breeder programs
and a smaller number of private breeder programs were in place. After 1970 some 35 additional
private programs were added. Of the 244 soybean varieties granted certificates since 1970, only
37 were granted to public research programs. The remainder were granted to approximately 35
different private firms" (Evenson, 1983, p. 971).
33. "Overall, the experiment stations have generally moved their work into areas where
they have a comparative advantage vis-a-vis the private sector. In direct competition with mar-
ket-oriented private firms, the public sector does poorly and generally does not invest heavily in
research of that type. It tends to be pressed into a good deal of work of a testing a certifying nature,
designed to help farmers make choices among suppliers of inputs. In recent years it has played
a major role in facilitating adjustment to regulations both in the chemical inputs fields and in
food technology" (Evenson, 1982, p. 275).
34. Evenson (1983) notes that "Given the major developments in the 'mother' biological
sciences, particularly in molecular biology, it would seem reasonable to expect a significant pro-
portion of the [agricultural research] system scientists to have backgrounds in this field. These
data show that this is not the case. Less than one percent of the researchers in the system, even
today after more than two decades of scientific revolution in the field, are trained in the field"
(p. 973). Other critiques of the U. S. agricultural research system are in U. S. Congressional Office
of Technology Assessment (1981, 1986).
35. Abramovitz's work on "catch-up" in postwar economic growth yields two insights into
the process of relative U. S. decline. First, Abramovitz points out that the process of convergence
in levels of productivity that has been underway in his sample of 16 nations since 1870 acceler-
ated sharply after World War II, reflecting more rapid rates of international technology transfer,
investment, and trade. Another important factor in more rapid postwar convergence, according
to Abramovitz (1990), was the fact that change in the structure of foreign industry, consumer
demand, and financial markets facilitated the absorption by foreign economies of the scale-
dependent, capital-intensive, manufacturing technologies of U. S. firms. No longer was it the
case that foreign economies' "underdeveloped financial markets and their still low levels of
income restricted capital accumulation, while the combination of low incomes and small pop-
ulations limited their domestic markets and therefore, also the base on which large-scale com-
petitive exports could be built" (1990, p. 6).
36. Indeed, one might argue that the weakening of university-industry research linkages
during a significant portion of the postwar period was the real departure from historical trends.
Hounshell and Smith (1988) cite a 1945 memo from Elmer Bolton, director of what was to
become the Du Pont Company's central research laboratory, that made a case for greater self-
reliance by the firm in its basic research: "Three things were necessary: Du Pont had to
strengthen its research organizations and house them in modern research facilities; the com-
pany's existing processes had to be improved and new processes and products developed; and
'fundamental research, which will serve as a background for new advances in applied chemistry,
should be expanded not only in the Chemical Department but should [also] be increased in our
industrial research laboratories and the Engineering Department.' Bolton stressed that it was no
longer 'possible to rely to the same extent as in the past upon university research to supply this
background so that in future years it will be necessary for the Company to provide this knowledge
to a far greater extent through its own efforts.' To 'retain its leadership' Du Pont had 'to under-
take on a much broader scale fundamental research in order to provide more knowledge to serve
as a basis for applied research'" (p. 355). Swann (1988, pp. 170-181) also argues that research
links between U. S. universities and the Pharmaceuticals industry weakened significantly in the
immediate aftermath of World War II, in part as a result of vastly increased federal research
funding for academic research in the health sciences.
37. This close relationship is due in part to the nature of biotechnology. Recombinant
DNA and genetic engineering techniques in many ways represent radical scientific break-
throughs that are being transferred to industry and reduced to practice. In Gomory's terminology
(1988), biotechnology is a "ladder" technology, that is, a case in which
the new idea is dominant and the product forms itself around the new idea or new tech-
nology. Those who understand that idea or technology are often scientists, and they there-
fore play leading roles in its introduction, (p. 11)
Another example of a ladder technology cited by Gomory is the transistor. In contrast to bio-
technology, of course, the transistor was first developed within industry. The different origin of
these two major scientific discoveries may reflect the shifting role of industry and universities as
basic research performers. An interesting empirical study of university-industry research collab-
oration that tends to support the characterization of biotechnology as a unique area of interac-
tion is Blumenthal et al. (1986).
38. A recent study by the Organization for Economic Cooperation and Development
(OECD, 1984) quotes a Xerox Corporation research executive's description of the firm's invest-
ment in the Center for Integrated Systems (CIS) at Stanford University: "Xerox's contribution
to CIS is very small compared to what we are investing internally in the same kind of research.
For little additional investment we enlarge our perspective by participating in a broad program
of basic research. We envision opportunities for joint interaction with the university and with
other companies, as well as the ability to recruit students. On a per-dollar basis it should be a
good investment" (quoted in OECD, 1984, p. 47).
39. A consortium like Bellcore, established in the wake of the divestiture of the Bell oper-
ating companies by AT&T, is a response to a very different and unique set of circumstances. In
most respects, Bellcore, serving the needs of the noncompeting businesses of the Bell operating
companies, more closely resembles the Electric Power Research Institute or the Gas Research
Institute than ventures such as the Software Productivity Consortium or the Microelectronics
and Computer Technology Corporation (MCC).
40. Chesnais (1988) has noted that an interesting complementary relationship may be
developing between closed domestic research programs in the EC and the United States, such as
JESSI and Sematech, and international product development and technology exchange agree-
ments in microelectronics: "one finds a combination between domestic alliances in pre-compet-
itive R&D (with all of the provisos attached to this notion), and a wide range of technology
exchange and cross-licensing agreements among oligopolist rivals at the international level"
(1988, p. 95; emphasis in original).
41. See Ferguson (1983, 1988), Florida and Kenney (1988), Borrus (1988), White House
Science Council (1988), and Business Week, 6/24/89. The comments of Jorde and Teece (1989)
illustrate one important line of argument within this critique: "Companies like Sun Microsys-
tems, Genentech, Compaq, Advanced Micro Devices, and Apple Computer are archetypical
examples [of startups]. Whereas large integrated firms like IBM and Exxon have relied upon
integration and administrative processes to effectuate coordination, the 'Silicon Valley' startups
have in the main eschewed integration and relied excessively on outsourcing . . . Market pro-
cesses have in some instances replaced administrative ones. This is particularly true where 'hol-
low corporations'those without significant in-house research, manufacturing, and distribu-
tionhave come to replace economic activity that in an earlier period took place inside
vertically integrated enterprises. In some cases this has left industries with inadequate strategic
coordination, particularly when they are competing against firms located in industrial structures
that are less fragmented and which are supported by governments that engage in directed indus-
trial policies" (p. 29).
42. See Economist (6/24/89, pp. 73-74). Other experts, however, assert that startup activity
remains strong. According to T.J. Rodgers, CEO of Cypress Semiconductor of San Jose, Cali-
fornia, "More chip companies were started in the five years between 1980 and 1985 than in the
two decades between 1960 and 1980" (1990, p. 25).
43. See Gupta (1982). A recent article in The Economist (2/10/89) asserts that "The dream
is dead. Biotechnology will never produce an entrepreneurial success to rival the out-of-nowhere
rise of the electronics industry's Apple, Compaq, or Intel" (p. 67).
44. Glenn R. Schleede, executive associate director of the Office of Management and Bud-
get, commented in 1981 that "By far the most important change [made in science and technol-
ogy policies by the Reagan Administration] came from this administration's redefinition of the
federal role. In the R&D spectrum stretching from the most esoteric basic research out through
the actual commercialization of a technology, we have drawn the line for federal intervention
and support back much farther toward the basic research end. In the civilian or domestic sector,
we do not think the government should be funding demonstration, product development, and
commercialization efforts" (quoted in Barfield, 1982, p. 41).
45. See Davis (1989). In still another new initiative, the Defense Advanced Research Proj-
ects Agency (DARPA) announced on April 9, 1990 that it was investing $4 million in Gazelle
Microcircuits, a small firm in the Silicon Valley region engaged in developing gallium-arsenide
components for civilian and military applications. The DARPA investment was made, and
explicitly justified, as a means of denying the firm's technology to potential foreign purchasers:
"The investment, made under an experimental two-year program approved by Congress, not
only allows the Defense Advanced Research Projects Agency, or Darpa, to earn a return on its
investment, but could also prevent the new company from having to sell its advanced technology
to a foreign company."
'"They in fact had been contacted by the Japanese and they were strapped for cash,' said
Richard L. Dunn, Darpa's general counsel. 'This agreement may have saved them from that.'"
In the event that the firm considers such a sale of its technology, DARPA is empowered to
arrange for a domestic buyer first; if no domestic buyer can be found, DARPA is allowed to
recover its investment, raising the costs to a foreign firm of investing in the U. S. firm (see Pollack,
1990).
46. Nelson (1984) concludes that "one rather clear lesson of the post-World War II expe-
rience is that trying to blend commercial and military procurement objectives is a mistake. If a
program is aimed specifically at enhancing competitive strength, it should stand separate from
procurement programs" (p. 73). Lorell (1980) provides a good review of the history of trans-
European weapons development consortia and programs.
47. See Perry (1986), among other accounts. According to Katz and Ordover (1990), at
least 14 Congressional bills passed during the 1980s focused on strengthening domestic and
international protection for intellectual property rights, and the Court of Appeals for the Federal
Circuit created in 1982 has upheld patent rights in roughly 80% of the cases argued before it, a
considerable increase from the pre-1982 rate of 30% for the Federal bench.
48. "Although the Act does take useful steps to assist innovation, the limited shelter it pro-
vides from antitrust covers only research activity. This is one reason why only 111 ventures had
been registered under the Act from 1984 through June 1988" (Jorde and Teece, 1989, p. 32).
49. Harris and Mowery (1990) present a critical discussion of the Congressional legislation.
50. The Act calls for a series of reports on U. S. firms' access to foreign markets for tele-
communications equipment, creates a new provision (Sec. 1315) for dealing with subsidized
international consortia, widely viewed as a provision directed at Airbus; and creates new pro-
visions for "fast-track" antidumping investigations in industries (such as microelectronics) with
short product life cycles.
51. Reciprocal access is a concept that is more easily stated than implemented in national
R&D systems that differ as sharply as do those of the United States, where publicly funded
research accounts for nearly 50% of all national R&D and where relatively open institutions such
as universities play a very important role in basic research, and Japan, in which corporate fund-
ing of R&D is far more significant. U.S. firms almost certainly would reject a policy that required
assurances of equal access to the research facilities of U. S. and Japanese corporations.
52. Comparisons of the pre- and post-1940 pattern of research employment (Mowery,
1981) suggest that federal funding was associated with some displacement of research activity
away from sectors receiving little or no federal research funds, such as chemicals and petroleum,
and toward the sectors that did receive massive defense-related federal R&D support (instru-
ments, electrical machinery, and transport equipment).
REFERENCES Agreements Between Firms." 5*77 Review 4: 51-

119.
Cohen, I. B. (1976). "Science and the Growth of the
Abramovitz, M. (1986). "Catching Up, Forging American Republic." Review of Politics 38(3):
Ahead, and Falling Behind." Journal of Economic 359-98.
History 46: 385-406. Collins, N. R., and F.L.E. Preston (1961). "The Size
Abramovitz, M. (1990). "The Catch-Up Factor in Structure of the Largest Industrial Firms." Amer-
Postwar Economic Growth." Economic Inquiry ican Economic Review 51: 986-1011.
28(1): 1-18. Congressional Budget Office (1984). Federal Sup-
Ames, J. (1925). Statement of NACA Chairman to port for R&D and Innovation. Washington, D. C.:
the President's Aircraft Board. Washington, D. C: Congressional Budget Office.
U. S. Government Printing Office. Cyert, R. M., and Mowery, D. C., eds., Technology
Baily, M. N., and Chakrabarti, A. K.. (1988). Inno- and Employment: Innovation and Growth in the
vation and the Productivity Crisis. Washington, U. S. Economy. Washington, D. C.: National
D. C.: Brookings Institution. Academy Press, 1987.
Barfield, C. E. (1982). Science Policy from Ford to Danhof, C. (1968). Government Contracting and
Reagan. Washington, D. C.: American Enterprise Technological Change. Washington, D. C.: The
Institute. Brookings Institution.
Beer, J. H. (1959). The Emergence of the German David, P. A. (1975). Technical Choice, Innovation,
Dye Industry. Urbana, IL: University of Illinois and Economic Growth. New York: Cambridge
Press. University Press.
Birr, K. (1966). "Science in American Industry." In David, P. A. (1986). "Technology Diffusion, Public
D. Van Tassel and M. Hall (eds.), Science and Policy, and Industrial Competitiveness." In R.
Society in the U. S. Homewood, IL: Dorsey. Landau and N. Rosenberg (eds.), The Positive
Blumenthal, D., Gluck, M., Louis, K. S., and Wise, Sum Strategy: Harnessing Technology for Eco-
D. (1986). "Industrial Support of University nomic Growth. Washington, D. C.: National
Research in Biotechnology." Science 231: 242- Academy Press.
46. Davis, B. (1989). "Pentagon Seeks to Spur U. S.
Borrus, M. G. (1988). Competing for Control. Cam- Effort to Develop 'High-Definition' TV." Wall
bridge: Ballinger. Street Journal 1/4/89, 29.
Briggs,A.(1981). "Social History 1900-1945."InR. Davis, L. E., and North, D. C. (1971). Institutional
Floud and D. W. McCloskey (eds.), The Eco- Change and American Economic Growth. New
nomic History of Britain Since 1700, Vol. 2. Cam- York: Cambridge University Press.
bridge, England: Cambridge University Press. Economist (1989a). "Venture-Capital Drought." 6/
Bright, A. A. (1949). The Electric Lamp Industry. 24/89, 73-4.
New York: Macmillan. Economist (1989b). "Test-Tube Trauma." 2/10/89,
Business Week (1989a). "Is the U. S. Selling Its 67.
High-Tech Soul to Japan?" 6/26/89, 117-18. Economist (1990). "Out of the Ivory Tower." 2/3/
Business Week (1989b). "Advanced Bio Class? 90, 65-72.
That's over in Hitachi Hall." 8/7/89, 73-4. Edquist, C., and Jacobsson, S. (1988). Flexible Auto-
Chandler, A. D., Jr. (1962). Strategy and Structure: mation: The Global Diffusion of New Technology
Chapters in the History of Industrial Enterprise. in the Engineering Industry. New York: Basil
Cambridge, MA: MIT Press. Blackwell.
Chandler, A. D., Jr. (1977). The Visible Hand. Cam- Edwards, R. C. (1975). "Stages in Corporate Stabil-
bridge, MA: Harvard University Press. ity and Risks of Corporate Failure." Journal of
Chesnais, F. (1988). "Technical Co-Opcration Economic History 35: 418-57.
Ergas, H. (1987). "Does Technology Policy Mat- Hounshell, D. A., and Smith, J. K., Jr. (1988). Sci-
ter?" In H. Brooks and B. Guile (eds.), Technology ence and Corporate Strategy: Du Pont R&D,
and Global Industry. Washington, D. C.: National 1902-1980 Strategy. New York: Cambridge Uni-
Academy Press. versity Press.
Evenson, R. E. (1982). "Agriculture." In R. R. Nel- Jaikumar, R., (1989). "Postindustrial Manufactur-
son (ed.), Government and Technical Progress. ing." Harvard Business Review 64: 69-76.
New York: Pergamon. Jorde, T. M., andTeece, D. J. (1989). "Competition
Evenson, R. E. (1983). "Intellectual Property Rights and Cooperation: Striking the Right Balance."
and Agribusiness Research and Development: California Management Review 31, 25-37.
Implications for the Public Agricultural Research Kaplan, A.D.H. (1964). Big Business in a Competi-
System." American Journal of Agricultural Eco- tive System. Washington, D. C.: Brookings Insti-
nomics 65: 967-75. tution.
Ferguson, C. H. (1983). "The Microelectronics Katz, M. L., and Ordover, J. A. (1990). "R&D Com-
Industry in Distress." Technology Review petition and Cooperation." Brookings Papers on
86(6):24-37. Economic Activity: Microeconomics 1990, 137-
Ferguson, C. H. (1988). "Beyond Entrepreneurial- 92.
ism to U. S. Competitiveness: From the People Kelley, M. R., and Brooks, H. (1933). The State of
Who Brought You Voodoo Economics." Harvard Computerized Automation in U. S. Manufactur-
Business Review 66(3):55-62. ing. Cambridge, MA: Center for Business and
Flamm, K. (1988). Creating the Computer. Wash- Government, Kennedy School of Government,
ington, D. C.: Brookings Institution. Harvard University.
Flamm, K., and McNaugher, T. (1989). "Rational- Lamoreaux, N. The Great Merger Movement in
izing Technology Investments." In J. D. Steinbru- American Business, 1895-1904. New York: Cam-
ner(ed.), Restructuring American Foreign Policy. bridge University Press.
Washington, D. C.: Brookings Institution. Lazonick, W. (1990). Competitive Advantage on the
Florida, R. L., and Kenney, M. (1988). "Venture Shop Floor. Cambridge, MA: Harvard University
Capital-Financed Innovation and Technological Press.
Change in the USA." Research Policy 17: 119- Lorell, M. A. (1980). Multinational Development of
37. Large Aircraft: The European Experience. Santa
Gansler, J. (1989). Affording Defense. Cambridge, Monica, CA: RAND Corporation.
MA: MIT Press. Malerba, F. (1985). The Semiconductor Business.
Gomory, R. E. (1988). "Reduction to Practice: The Madison, WI: University of Wisconsin Press.
Development and Manufacturing Cycle." In Markoff, J. (1989). "A Corporate Lag in Research
Industrial R&D and U. S. Technological Leader- Funds is Causing Worry." New York Times 1/23/
ship. Washington, D. C.: National Academy 89, Al.
Press. Mowery, D. C. (1981). "The Emergence and Growth
Graham, M.B.W. (1986a). RCA and the Videodisc: of Industrial Research in American Manufactur-
The Business of Research. Cambridge, England: ing, 1899-1946." Ph.D. diss., Stanford Univer-
Cambridge University Press. sity.
Graham, M.B.W. (1986b). "Corporate Research Mowery, D. C. (1983). "Industrial Research, Firm
and Development: The Latest Transformation." Size, Growth, and Survival, 1921-1946." Journal
Technology in Society 7: 179-95. of Economic History 43:953-80.
Griliches, Z. (1980). "R&D and the Productivity Mowery, D. C. (1984). "Firm Structure, Govern-
Slowdown." American Economic Review 70:343- ment Policy, and the Organization of Industrial
48. Research: Great Britain and the United States,
Griliches, Z. (1986). "Productivity, R&D and Basic 1900-1950." Business History Review 58:504-
Research at the Firm Level in the 1970s." Amer- 31.
ican Economic Review 76:141-54. Mowery, D. C., ed. (1988). International Collabo-
Gupta, U. (1982). "Biotech Start-Ups are Increas- rative Ventures in V. S. Manufacturing. Cam-
ingly Bred Just to Be Sold." Wall Street Journal If bridge, MA: Ballinger Publishers.
19/82, B2. Mowery, D. C. and Rosenberg, N. (1989a). Tech-
Gupta, U. (1988). "Start-Ups Face Big-Time Legal nology and the Pursuit of Economic Growth. New
Artillery." Wall Street Journal \ 1/20/88, B2. York: Cambridge University Press.
Harris, R. G., and Mowery, D. C. (1990). "New Mowery, D. C., and Rosenberg, N. (1989b). "New
Plans for Joint Ventures: The Results May Be an Developments in U. S. Technology Policy: Impli-
Unwelcome Surprise." The American Enterprise. cations for Competitiveness and International
September/October, 52-55. Trade Policy." California Management Review
Hounshell, D. A., and Smith, J. K. (1985). "Du 27:107-24.
Pont: Better Things for Better Living Through Mowery, D. C., and Steinmueller, W. E. (1991).
Research." Presented at "The R&D Pioneers," "Government Policy and Industry Evolution in
Hagley Museum and Library, Wilmington, Dela- the U. S. Integrated Circuit Industry: What Les-
ware, October 7. sons for Newly Industrializing Economies?" CCC
Working Paper, Center for Research in Manage- national Collaborative Ventures in U. S. Manu-
ment, University of California, Berkeley. facturing. Cambridge, MA: Ballinger.
Mueller, W. F. (1962). "The Origins of the Basic Pollack, A. (1990). "Technology Company Gets $4
Inventions Underlying Du Font's Major Product Million U. S. Investment." New York Times 4/
and Process Innovation, 1920 to 1950." In The 10/90, C17.
Rate and Direction of Inventive Activity. Prince- Pursell, C. (1977). "Science Agencies in World War
ton: Princeton University Press. II: The OSRD and its Challengers." In N. Rein-
National Commission on Research, (1980). Industry gold (ed.), The Sciences in the American Context.
and the Universities. Washington, D. C. National Washington, D. C.: Smithsonian Institution.
Commission on Research. Reich, R. B., and Mankin, E. (1986). "Joint Ven-
National Research Council (1982). "Research in tures with Japan Give Away Our Future." Har-
Europe and the United States." In Outlook for Sci- vard Business Review 64:78-86.
ence and Technology: The Next Five Years. San Reid, P. P. (1989). "Private and Public Regimes:
Francisco: Freeman. International Cartelization of the Electrical
National Resources Planning Board (1942). Equipment Industry in an Era of Hegemonic
ResearchA National Resource, Vol. 1. Wash- Change, 1919-1939." Unpublished Ph.D. disser-
ington, D. C.: U. S. Government Printing Office. tation, Johns Hopkins School of Advanced Inter-
National Science Board (1981). Science Indicators, national Studies.
1980. Washington, D. C.: U. S. Government Rodgers, T. J. (1990). "Landmark Messages from
Printing Office. the Microcosm." Harvard Business Review Jan-
National Science Board (1983). Science Indicators, uary-February, 24-30.
1982. Washington, D. C.: U. S. Government Rosenberg, N. (1972). Technology and American
Printing Office. Economic Growth. New York: Harper & Row.
National Science Foundation (1985). Science and Rosenberg, N., and Steinmueller, W. E. (1988).
Technology Data Book. Washington, D. C.: "Why Can't Americans Learn to be Better Imita-
National Science Foundation. tors?" American Economic Review 78(2):229-34.
Nelson, R. R. (1984). High-Technology Policies: A Rosenbloom, R. S. (1985). "The R&D Pioneers,
Five Nation Comparison. Washington, D. C.: Then and Now." Presented at "The R&D Pio-
American Enterprise Institute. neers," Hagley Museum and Library, Wilming-
Nelson, R. R. (1990). "U. S. Industrial Competitive- ton, Delaware, October 7.
ness: Where Did It Come From and Where Did It Saxonhouse, G. R. (1982). "Japanese High Tech-
Go?" Research Policy 19:117-32. nology, Government Policy, and Evolving Com-
Okimoto, D. I., and Saxonhouse, G. R. (1987). parative Advantage in Goods and Services" pre-
"Technology and the Future of the Economy." In sented at the Japanese Political Economy
K. Yamamuraand Y. Yasuba (eds.), The Political Research Conference, Honolulu, Hawaii.
Economy of Japan, Vol. 1. The Domestic Trans- Saxonhouse, G. R. (1986). "Why Japan is Win-
formation. Stanford: Stanford University Press. ning." Issues in Science and Technology 3:50-62.
Organization for Economic Cooperation and Devel- Scherer, F. M. (1983). "R&D and Declining Produc-
opment (OECD) (1984). Industry and University: tivity Growth." American Economic Review 73:
New Forms of Co-operation and Communication. 215-18.
Paris: OECD. Schmookler, J. (1957). "Inventors Past and Pres-
Ostry, S. (1990). The Political Economy of Policy ent." Review of Economics and Statistics^1): 321-
Making: Trade and Innovation Policies in the 33.
Triad. New York: Council on Foreign Relations. Schumpcter, J. A. (1954). Capitalism, Socialism and
Parker, W. N. (1972). "Agriculture." In L. E. Davis Democracy, 3d ed. New York, Harper & Row.
et al. (eds.), An Economist's History of the United Sharp, M. (1989). "European Countries in Science-
States. New York: Harper & Row. Based Competition: The Case of Biotechnology."
Patel, P., and Pavitt, K. (1986). "Measuring DRC Discussion Paper#72, SPRU, University of
Europe's Technological Performance: Results Sussex.
and Prospects." In H. Ergas (ed.), A European Stigler, G. J. (1968). "Monopoly and Oligopoly by
Future in High Technology? Brussels: Center for Merger." In G. J. Stigler (ed.), The Organization
European Policy Studies. of Industry. Homewood, IL: Irwin.
Perry, N. J, (1986). "The Surprising Power of Pat- Sturchio, J. L. (1985). "Experimenting with
ents." Fortune 6/23/86, 57-63. Research: Kenneth Mees, Eastman Kodak, and
Perry, W. J. (1986). "Cultivating Technological the Challenges of Diversification." Presented at
Innovation." In R. Landau and N. Rosenberg "The R&D Pioneers," Hagley Museum and
(eds.), The Positive Sum Strategy. Washington, Library, Wilmington, Delaware, October?.
D. C.: National Academy Press. Swann, J. P. (1988). Academic Scientists and the
Pisano, G. P., Shan, W., and Teece, D. J. (1988). Pharmaceutical Industry. Baltimore, MD: Johns
"Joint Ventures and Collaboration in the Biotech- Hopkins University Press.
nology Industry." In D. C. Mowery (ed.), Inter- Taylor, G. D., and Sudnik, P. E. (1984). Du Pont and
the International Chemical Industry. Boston: ship of Research and Development on the Devel-
Twayne. opment of the Civilian Electronics Industry."
Thackray, A., Sturchio, J. L., Carroll, P. T., and Bud, Center for Policy Alternatives working paper #
R. (1985). Chemistry in America, 1876-1976: 77-5, M.I.T.
Historical Indicators. Dordrecht: Reidel. Walton, R. E., and McKersie, R. B. (1989). "Man-
Thorelli, H. B. (1954). Federal Antitrust Policy. Bal- aging New Technology and Labor Relations: An
timore, MD: Johns Hopkins University Press. Opportunity for Mutual Influence." In D. C.
Tilton, J. E. (1971). The International Diffusion of Mowery and B. E. Henderson, eds., The Chal-
Technology: The Case of Transistors. Washing- lenge of New Technology to Labor-Management
ton, D. C: Brookings Institution. Relations. Washington, D. C.: U. S. Department
U. S. Bureau oftheCensus(1975). Historical Statis- of Labor.
tics oj the United Slates: Colonial Times to 1970, Weart, S. (1979). "The Physics Business in America,
Vol. 1. Washington, D. C.: U. S. Government 1919-1940." In N. Reingold (ed.), The Sciences in
Printing Office. the American Perspective. Washington, D. C.:
U. S. Bureau of the Census (1987). 1987 Statistical Smithsonian Institution.
Abstract of the United States. Washington, D. C.: White House Science Council (1988). High-Tem-
U. S. Government Printing Office. perature Superconductivity: Perseverance and
U. S. Congressional Office of Technology Assess- Cooperation on the Road to Commercialization.
ment (1981). An Assessment of the United States Washington, D. C.: Office of Science and Tech-
Food and Agricultural Research System. Wash- nology Policy.
ington, D. C.: U. S. Government Printing Office. Wise, G. (1985). "R&D at General Electric, 1878-
U. S. Congressional Office of Technology Assess- 1985." Presented at "The R&D Pioneers," Hagley
ment (1986). Technology, Public Policy, and the Museum and Library, Wilmington, Delaware,
Changing Structure of American Agriculture. October?.
Washington, D. C.: U. S. Government Printing Wolf, J. (1989). "Europeans Fear Obstacles by U. S.
Office. on Advanced TV." Wall Street Journal 5/31/89,
Utterback, J. M, and Murray, A. E. (1977). "The A16.
Influence of Defense Procurement and Sponsor-
3
The Japanese System of Innovation: Past,
Present, and Future
HIROYUKIODAGIRI
AKIRAGOTO
It was commonly believed that the history of the Japanese economy was that of a coun-
try desperately trying to catch up with technologically advanced nations. When Japan
opened its country around the time of the Meiji Restoration of 1868 following more
than two centuries of seclusionism, the leaders had to realize how much Japan was
behind Western countries in many aspects of technology. Naturally, the government
made efforts to import superior technology, hire engineers from abroad, educate its
people, and encourage the entrepreneurs to assimilate foreign technologies and apply
them in Japanese factories. In addition, the determination to catch up in terms of mil-
itary capacity gave the government a strong incentive to support technological
advances and make domestic procurement possible. Such a catch-up process has been
observed in many late-developing countries, such as Prussia and Russia in the nine-
teenth century or South Korea and Taiwan in recent decades. Japan's experience was
by no means unusual and neither were the policies taken by the government.
However, one cannot attribute Japan's "success" solely to the latecomer advan-
tage or to government policies because neither of these can lead to successful economic
development unless the private sectorinvestors, managers, engineers, workersis
willing and able to respond to the opportunities open to them. In fact, at the time of
the Meiji Restoration, entrepreneurs appeared in various industries and most people
had been sufficiently educated to read, write, and count. The willingness to start an
unfamiliar business and ability to assimilate new technology were present, and these
were the main moving forces behind Japan's development. It is also noted that even
under the seclusionism during the Tokugawa (Edo) era, technological importation did
take place (though not on a large scale) and indigenous technologies and skills were
gradually developed. One aim of this chapter is to investigate the conditions necessary
for successful catch-up by studying the development of Japanese industries and tech-
nologies by means of both a general historical description and industrial cases.
Another aim is to study how Japanese businesses have been (and are) behaving and
organizing, because such a micro aspect provides an important key to understanding
the Japanese innovation system.
76
THE JAPANESE SYSTEM OF INNOVATION 77
The paper is separated into four parts. The first section gives a historical overview
of Japan's technological advance from past to present. The second section discusses
the technological development of three industriesiron and steel, electrical and com-
munication equipment, and automobiles. The next section discusses the present Jap-
anese management system, since understanding the managerial aspects of Japanese
firms is indispensable to understanding why the firms are motivated to innovate and
how they maintain efficiency in innovation. The final section concludes the chapter
by discussing future prospects.
FROM PAST TO PRESENT
The Tokugawa Era and Before: Up to 1868
Major technology importation before the Tokugawa era (1603-1868) took place
twice. The first is from the seventh to ninth century when the emperor's government
sent envoys to China and immigration was frequent from China and Korea. The sec-
ond is the sixteenth century, particularly when guns were first introduced by the Por-
tuguese. This was the period of civil war and the strong demand for guns caused a
number of blacksmiths to start producing guns either by copying or with the help of
Portuguese gunsmiths. The technological level in guns, ships, and other arms produc-
tion soon caught up with the West.
The Tokugawa government, fearing that Christians would disobey their rule,
secluded the country in 1639 and restricted foreign trade to the Chinese and the Dutch.
Thus the only Western contact for Japanese was through a number of Dutchmen
allowed to live in Nagasaki in Kyushu, an island in the southwest. The Tokugawa gov-
ernment monopolized trade, though evidence suggests that some of the powerful feu-
dal lords under the Tokugawa rule did trade with the Chinese and other foreigners.
With the request of the government the Dutch provided information on foreign affairs
and science regularly, some of which were diffused to other lords. The Dutch in Naga-
saki were also a source of information on many aspects of science and technology, such
as medical science, biology, and geography. Several Dutch books were translated into
Japanese, and a number of Japanese studied with Dutch doctors or other scientists
living in the Dutch quarters of Nagasaki. Thus the seclusionism by no means implied
a complete seclusion from foreign scientific and technological information. Actually,
the accumulation of Western knowledge through the Dutch can be assumed to have
contributed to a rapid absorption of the Western economic and technological system
when the Tokugawa government opened the country to non-Dutch Western countries
in 1854. For instance, the first Western-style iron furnace was made by a samurai who
had studied with a Dutch book.
In terms of indigenous technology too, the Tokugawa era was hardly a static
period. Because of the importance of agriculture and mining for the feudal lords'
finances, increases in productivity in these areas were encouraged and many improve-
ments were made and diffused. Because of the steep flow of rivers, climatic conditions
(rainy season in June and typhoons in summer-autumn), and the importance of water
in rice crops, much investment was directed toward improving rivers and making irri-
gation systems. Consequently, the technological level in civil engineering is believed

to have been very high.
Machine engineering was another area for indigenous technology at the time. The
innovators were called karakuri masters because karakuri, a moving mechanism, is
the most essential part of any machine. These mechanisms were applied to many sorts
of machines and tools, ranging from dolls and clocks to textile looms and rice-polish-
ing machines.1 Again, we will show later that there was a continuous development
from such indigenous technology to imported Western technology after the Meiji Res-
toration. H. Tanaka, probably the most important karakuri master toward the end of
the Tokugawa era, became a pioneer in the electrical equipment industry, and S. Toy-
oda (born a year before the Restoration) invented internationally copied looms.
Therefore, although the Japanese failed to invent steam engines, among other
things, their technological level was not too much behind the West. This fact should
be emphasized because it is in marked difference from the cases of many developing
countries today. In addition, the educational level was highprobably higher in ele-
mentary education than the United States, Britain, France, and Germany, though
lower in higher education because in these countries science education in universities
had started by the early nineteenth century.
Basically there were two school systems. The first was the schools owned by feudal
local governments, which were usually compulsory to the children of samurais (i.e.,
the employees of respective local governments). Many of them also admitted the chil-
dren of wealthier farmers and merchants. The second was private schools, called ter-
akoya, since many of them were run by Buddhist temples (tera). The length of edu-
cation was various but, most typically, went from the age of 6 to 12, similar to the
present elementary school system. They mainly taught reading, writing, and the use
ofsoroban (abacus) to calculate. In addition to these schools for children, there were
a number of private or public higher education systems. Some of these taught Japanese
or Chinese studies, for instance, Confucianism. However, there were private schools
teaching medical science using Dutch books. The best known of the latter case was
"Tekijuku," a private school in Osaka taught by K. Ogata, who had learned Dutch
medical science in Nagasaki. Among his students were Y. Fukuzawa, who later intro-
duced Western democratic thoughts to Japan and established the first private univer-
sity, Keio, and M. Ohmura, who later designed the Meiji military system (but was
assassinated before the system was completed).
There were not many terakoyas and other schools in the early Tokugawa era but
they became more and more popular, and, according to a very rough estimate, there
were almost ten thousand such schools in Japan toward the end of the era. Many of
them were small with 10 or fewer children, but there were schools with more than 500
in large cities such as Edo (Tokyo) and Kyoto. The ratio of school enrollment (or atten-
dance) varies among estimates and among regions, ranging from about 25 to almost
100% for boys and a lower percentage for girls (Umihara, 1988). These estimates may
be misleading because they may include those children enrolled as pupils but who sel-
dom came. Nevertheless, the ratio of, let us say, 50% under the noncompulsory system
is surprisingly high and shows the eagerness of parents to invest in child education.
This fact suggests that the literacy rate in Japan in the seventeenth and eighteenth cen-
tury was likely higher than in Europe and America. The introduction of the public
education system in the Meiji era would not have been so smooth were it not for this
wide educational background.
The Meiji Era: 1868-1911
The end of seclusionism in 1854 and the inauguration of a nonfeudal central govern-
ment following the Meiji Restoration of 1868 prompted the Japanese government and
the public at large to import advanced foreign technology and to catch up with the
Western countries economically and militarily. The government thus started an orga-
nized effort to "modernize" the country, including the provision of infrastructures for
transportation, communication, utilities, education, and finance.
Generally speaking, diffusion of science and technology from one country (or
countries) to another can be made by transferring written information (e.g., books,
papers, and drawings), people (e.g., hiring foreigners and sending Japanese to study
abroad), goods (e.g., importing machines and plants), and capital (i.e., foreign direct
investment). All these methods were used in Meiji Japan. Particularly in the case of
importing social systems, hire (yatoi) of foreign teachers and advisors was common.
For instance, when the Meiji government started to establish a national system of edu-
cation including compulsory elementary education, an American, D. Murray, was
hired as an advisor and gave American influences to the system, though these influ-
ences were diluted when nationalistic education was later emphasized.
It took about three decades of trial and error before the government could estab-
lish a countrywide elementary education system. In 1874 2 years after the government
started the effort, there were about twenty thousand schools, less than half of what the
government had planned. Some of these schools were converted from terakoyas and
taught by former terakoya teachers. The enrollment ratio was 46% for boys and 17%
for girls. By 1904 when the compulsory 6-year education system was finally estab-
lished, the ratio had increased to 99% for boys and 96% for girls. Thus illiteracy among
the youth was nearly absent by the beginning of this century. The secondary education
system also became commonly available and by 1920 more than half of the children
out of elementary schools proceeded to 2-year or 5-year secondary schools (Kaigo,
1971).
For the higher education system, particularly in the field of technology and engi-
neering education, a British influence was introduced. Kobusho (the Ministry of
Industries), itself started with the advice of a British railway engineer, hired a British,
H. Dyer, to make plans for an engineering college. In 1873, Kogakuryo (the College
of Engineering) was established with Dyer's plan and eight more British professors
were hired besides Dyer himself who became the head. With the eagerness and high
quality of both the instructors and the students, the college was quite successful. It
hired more foreign (mostly British) professors and then started to replace them with
Japanese graduates from the College. In 1886, the College merged with another col-
lege, established by the Tokugawa government to teach science and technology with
the help of the Dutch, French, and German, and became the Engineering Department
of Imperial University (later renamed the University of Tokyo). The College and the
University produced graduates who later founded many of the major Japanese man-
ufacturing companies as we will later show.
It is noteworthy that Dyer later exported his Japanese experience of engineering
education to his hometown, Glasgow. He had been recommended to the Japanese gov-
ernment by a professor at the University of Glasgow, and came to Japan at the age of
25 right after receiving a master's degree there. The program he made for Kogakuryo
emphasized the interaction between classroom studies and on-site training at the lab-
oratory works he made within Kogakuryo as well as Kobusho's works. This balance
between the two aspects of education was lacking in European schools at the time and
Dyer's originality should be noted. He left in 1882 to go back to Glasgow and made
efforts to introduce a similar engineering education program when a technical college
was founded there. Apparently, the Japanese government emphasized engineering
education at the time when more developed countries regarded pure science as supe-
rior to engineering. This background gave Dyer an opportunity to experiment with his
ideas on engineering education and, with his Japanese experience, he persuaded his
people of its importance.
During the early Meiji era, in particular the 1870s and early 1880s, the govern-
ment built and owned plants and factories in industries such as mining, railroad, ship-
building, machinery, and textile, because it was still difficult for the private sector to
finance the required investment and take risks. In addition, personnel with advanced
Western technological knowledge were scarce in the private sector. However, the gov-
ernment's investment program in industries was neither quantitatively spectacular
compared, say, to that of today's developing countries nor always successful. In fact,
most government-owned factories suffered losses and, with the tight government bud-
get, were gradually privatized and sold to emerging private entrepreneurs (mostly mer-
chants and ex-samurais).
The government retained plants in military-related industries, such as shipbuild-
ing, aircraft, munitions, and steel, and in public utilities including telecommunica-
tion. The military-related production occupied a significant portion of Japan's econ-
omy at the time, because the Meiji government was keen to build up its military
capacity to deal with the threat of Russia and other countries that were colonizing
China. In 1907, the largest operation in machinery industry (including shipbuilding,
vehicles, general machinery, tools, and parts) in terms of the number of blue-collar
workers was the navy shipyard in Kure with about 21,500 employees, followed by
another navy shipyard and two arsenals (Sawai, 1990). The largest private plant, Mit-
subishi's Nagasaki Shipyard, ranked fifth with less than 10,000 employees, whose
important customer was again the navy. The military-owned plants were also a center
of technological development. They hired a large percentage of scarce engineers and
imported advanced machinery from abroad. Their technology was subsequently
transferred to the private sector as the engineers and skilled workers moved from the
military plants to the private sector, especially during the disarmament period follow-
ing the Russo-Japanese War of 1904-1905.
The military not only produced goods within its own shipyards and arsenals, but
also procured them from the private sector. Since the military, for obvious defense
reasons, preferred to procure goods domestically, procurement gave domestic produc-
ers in shipbuilding, steel, machines, electrical equipment, and so on, who were under
competitive pressure from larger and technically advanced foreign firms, a chance to
increase their production and accumulate knowledge through experience. It should be
noted that Unequal Treaties the Tokugawa government was forced to sign in 1858
with the United States, the United Kingdom, the Netherlands, France, and Russia
deprived the government of the right to set tariffs to imports. In 1902, for instance, the
proportion of tariffs in government revenue was 5% in Japan versus 21 % in the United
Kingdom and 45% in the United States. Thus, until 1911 when the treaties were
revised, the government had little means to protect domestic producers from foreign
competitors except preferential procurement.
The economy started to grow after about two decades of the Restoration (see
Table 3.1). GNP more than doubled in the 30-year period from 1885 to 1914, the year
World War I started. In terms of industrial composition, food processing and textile
were the largest industries before the turn of the century. Then metal, machinery,
chemical, and other heavy industries started to grow fast. During the first four decades
of the twentieth century, these heavy industries grew at an annual rate of more than
10% whereas the manufacturing industries as a whole grew at about 6%. Their growth
was particularly rapid in the 1920s and the 1930s, and their share in the manufacturing
sector exceeded 50% in the 1930s.
Technological progress was an important source of this growth. According to
Minami (1981), nearly 70% of the growth of the private mining and manufacturing
sector was accounted for by the "residual" factors that include technological progress.
This technological progress came from both indigenous (traditional and domestic)
technology and the technology imported from advanced countries. Indigenous tech-
nology was important not only on its own, particularly in traditional industries, but
also in providing the ability to select among the technologies available in developed
countries, and in adapting and assimilating them to fit domestic conditions. This fact
was most notable in the textile industry, the second largest manufacturing industry at
the time (next to food processing) and the largest exporting industry before World War
II. The case study of iron and steel industry will provide another example.
This fact notwithstanding, the role of indigenous technology was limited in mod-
ern industries, such as metal and machinery, where imported technology played a far
greater role. As discussed earlier, technology transfer from abroad was made through
many channels. Many foreign engineers and specialists were hired, though they were
gradually replaced by the Japanese educated domestically at the above-mentioned
College of Engineering and other schools and those who had studied abroad and came
back. Other channels of technology transfer were importation of advanced machinery
(and reverse engineering), licensing agreements, and foreign direct investment into
Japan. The latter two increased since the turn of the century because the government
liberalized foreign direct investment and also joined the Paris Convention, though the
patent system itself was introduced as early as 1885. The significant role these various
means of technology transfer played in Japan's industrial and technological develop-
ment will be discussed in more detail in the case studies.
The number of patents granted during July 1885 to February 1902 was 4817 (Pat-
ent Office, 1955). In comparison, the number was 27,136 in 1902 in the United States,
13,714 in the United Kingdom, 12,026 in France, and 10,610 in Germany. Hence the
number of patents in Japan was hardly comparable with the Western countries.
Among these 4817 patents, 2175 (45%) were related to machinery, 728 (15%) to chem-
icals, 52(1 %) to electric equipment, and 1862 (39%) to miscellaneous. Therefore, inso-
far as we can infer from the number of patents, R&D in the machinery industry seems
to have been relatively active around the turn of the century.
Table 3.1. Gross National Expenditures and Production: 1875-1940
Levels (at 1 934- 1 936 prices) Annual Growth Rates (%) Composition (%)
1875- 1885- 1900- 1915-

1875 1885 1900 1915 1940 1885 1900 1915 1940 1875 1885 1900 1915 1940
Population (thousand) 35036 38176 44056 53110 71933 0.9 1 1.3 1.2
GNE (million yen) 3852 6238 8522 22848 3.3 2.1 4 100 100 100 100
Personal consumption 3284 5270 6806 13389 3.2 1.7 2.7 85.3 84.5 79.9 58.6
expenditure
Government consumption 283 538 769 3377 4.4 2.4 6.1 7.3 8.6 9 14.8
expenditure
Gross domestic fixed 346 703 1176 7070 4.8 3.5 7.4 9 11.3 13.8 30.9
capital formation
Surplus on current account -61 -279 -224 -988 10.7 -1.5 6.1 -1.6 -4.5 -2.6 -4.3
Exports and factor 68 275 1020 3973 9.8 9.1 5.6 1.8 4.4 12 17.4
income from abroad
Imports and factor 129 554 1244 4961 10.2 5.5 5.7 3.3 8.9 14.6 21.7
income paid abroad
GNE per Capita (yen) 101 142 160 318 2.3 0.8 2.8
Production (million yen)
Mining 7.1 21.3 90.3 290.6 763.7 11.6 10.1 8.1 3.9
Coal 4 9.2 53.2 145.5 399.8 8.7 12.4 6.9 4.1
Manufacturing 742.6 877.9 2101 4029.4 20210.2 1.7 6 4.4 6.7 100 100 100 100 100
Food 457.5 454 1018.5 1356.3 2634.7 -0.1 5.5 1.9 2.7 61.6 51.7 48.5 33.7 13
Textile 64.1 129.7 508.7 1133.6 3454.3 7.3 9.5 5.5 4.6 8.6 14.8 24.2 28.1 17.1
Chemicals 75.6 99.1 186.4 427.6 3342.3 2.7 4.3 5.7 8.6 10.2 11.3 8.9 10.6 16.5
Machines 4.9 12.2 68.9 362.2 5580.5 9.6 12.2 11.7 11.6 0.7 1.4 3.3 9 27.6
Iron and steel 2.3 3.2 5.8 89.1 2494.8 3.4 4 20 14.3 0.3 0.4 0.3 2.2 12.3
Nonferrous metals 6.1 13.8 22.7 145.4 556.2 8.5 3.4 13.2 5.5 0.8 1.6 1.1 3.6 2.8
Others 132.1 165.9 290 515.2 2147.4 2.3 3.8 3.9 5.9 17.8 18.9 13.8 12.8 10.6
Source: Estimates of Long-Term Economic Statistics of Japan since 1868. K.Ohkawa, M. Shinohara, and M. Umemura (eds.). Tokyo: Toyo Kelzal Shinposha, 1988.
Between the Two World Wars: 1914-1945
It is reasonable to say that the Japanese economy took off around the time of World
War I. As will be shown later, many companies in steel, machinery, chemical, and
other heavy industries had been established by the mid-1910s and started growing.
Also, the country's scientific and engineering base started to be formed. The education
system, which had been expanded by this time to include several universities and other
higher education institutions, started to supply many trained engineers. More and
more skilled workers capable of handling advanced equipment became available.
Some of them spun off with their acquired skills. Academic associations were formed
and professional journals started. Access to foreign technological information became
easier with the extended availability of foreign books and journals, and through trad-
ing companies.
To enhance the scientific and technological base further, various proposals were
made by industrialists, policy-makers, the military, and scientists; for instance, the cre-
ation of basic research institutions, the increase in national industrial laboratories, and
enhanced science and technology education. The industries felt a strong need to
upgrade its scientific and technological capability to accommodate the growth of tech-
nology-based industries. World War I accelerated this trend, by convincing the mili-
tary of the importance of high technology and the private sector's capability to con-
tribute to national defense. In addition, the import of equipment and intermediate
goods became difficult during the war, giving the private sector an incentive to produce
previously imported advanced equipment and materials domestically, such as
machine tools, various chemicals, aluminum, and steel. This fact further enhanced the
need for advanced technology.
More universities and vocational schools were established by the public as well as
the private sector. Several national research institutions were founded. Hiroshige
(1973) counts 38 national research laboratories founded during 1914-1930 including
those annexed to national universities and to the military. Some of them were founded
by reorganizing existing laboratories or testing institutions. For instance, the Electric
Research Institution was founded in 1918 by upgrading the laboratory within the Min-
istry of Communications, which dates back to 1873 when Kobusho started an exper-
imental electrical factory within it.2 In 1890 it had a staff of about 30 with 7 engineers,
the rest being assistants and shop workers. It studied both technologies related to elec-
tric power and those related to communication; however, the major part of its duty
was to test the equipment procured by the Ministry. Another example is the Industrial
Research Institution established in 1900 with a staff of 20. By 1920 the staff had
expanded to 220, including 48 researchers and engineers. It had five departments
inspection, chemistry [japan (a type of lacquer), matches, oil, and fat, etc.], ceramics,
dye, and (since 1909) electrochemistry. Other research institutions present or estab-
lished during this period include those studying measures, silk, geography, mining,
fuel, and airplanes (within the navy) besides several institutions related to agriculture.
One of the largest and most productive research institutions established during
this period was the Institute of Physical and Chemical Research (Rikagaku Kenkyu-
sho, known as Riken). The proposal for Riken was first made by a chemist, J. Taka-
mine, with the aim of fostering scientific progress and thereby contributing to indus-
tries. Thus the aim was not purely academic but also practical. After a long discussion
involving businesses, the government, and the academics, it was established in 1917
with roughly half of the funds from the government and half from the private sector,
and with a staff of 22, of whom 5 were full-time (including both researchers and sup-
portive staff). In 4 years the staff increased to 63, with 30 full-time members. By 1945
it produced about 800 patents in Japan and 200 overseas. Some of them were devel-
oped to products, which Riken's subsidiary companies sold. These include chemical
products, such as vitamins and photographic paper, and machinery, such as piston
rings and measuring equipment. In terms of the contributions to the national econ-
omy, these products were hardly as major as, say, automobiles and electrical products.
To promote more basic research, the Science Council (Gakujutsu Shinkokai,
known as Gakushiri) was established in 1933 with funds from both the public and pri-
vate sector. Its purpose was twofold: first, to increase research funds at universities and
other research institutions, and second, to promote efficient research management by,
for instance, encouraging interorganizational research collaboration. The research
funds were granted both on an individual basis and on a project basis. In the beginning,
most grants were given on an individual basis but, by 1942, more than 80% were given
on a project basis, with each project pursued by members from a number of institu-
tions. Among the grants given on a project basis in 1942, about two-thirds were given
to the projects in the field of engineering, and the rest to medical science, natural sci-
ence, social science, and humanities. The top three projects in terms of research grants
were jet fuel, wireless communication, and atomic nucleus. The emphasis on military-
related technology was apparent.3
Companies also started their own R&D laboratories. In 1923, there were 162 pri-
vate R&D laboratories affiliated with companies, cooperatives, and other private foun-
dations. Of these 71 were in chemistry (including Pharmaceuticals, dye, paint, rubber,
cement, ceramics, and paper), 27 in metals and machinery, and 24 in food. Among
the large companies having their own laboratories were Shibaura Seisakusho and
Tokyo Denki (later to be merged to become Toshiba), Mitsubishi Shipbuilding (later
Mitsubishi Heavy Industries), Nippon Kokan (now NKK), Oji Paper, and Takeda,
Tanabe, and Sankyo (all in Pharmaceuticals). However, they include testing or devel-
opment sections within factories and how many of them deserve to be called labora-
tories in the present sense is unknown. Most of them were small.4
After the breakout of the Chinese-Japanese War in 1937, the production of heavy
industries increased rapidly as the expansion of armaments speeded up. Compared to
the 1934-1936 average, the production of steel more than doubled and machinery
almost quadrupled by 1944. The share in total manufacturing production of heavy
industries (machinery, metal, and chemical) increased from 33.9% in 1931 to 55.8%
in 1937, and then to 70.2% in 1942.5
The R&D activity also increased. Although the reliability of the data may be ques-
tioned, one survey in 1930 shows the number of research organizations (including
departments and institutions within universities, government laboratories, Riken, and
laboratories affiliated with private companies) to be 349 in 1930, which spent 30 mil-
lion yen or 0.22% of the GNP.6 In 1942, the number of private research organizations
was 711, employing a staff of 33,400 and spending 590 million yen or approximately
1% of the GNP. In addition, there were 443 public research organizations (including
university departments, etc.) employing a staff of 16,160 and spending 296 million
yen.7
These increased production and R&D activities enabled some of the Japanese
manufacturing industries to start building world-class production facilities and devel-
oping advanced products; for instance, large-scale furnace and open hearth for steel
production, aircrafts, ships, alloys, and communication equipment. However, even
these industries depended on American and European technology in many aspects,
and imported technology actively until the late 1930s. The stoppage of technological
flow from abroad during World War II had a serious impact, and the increase in R&D
efforts at the time reflects Japan's desperate effort to fill the void created by this stop-
page. Consequently, despite the increased R&D effort, the technological gap from the
West widened in such key munitions industries as aircraft and shipbuilding, which had
almost caught up with the world technological frontier before the war. This gap partly
explains Japan's defeat in the war.8
The Post-World War II Era: 1945 to the early 1970s
The impact of World War II on the Japanese economy was, needless to say, devastat-
ing. The production index of the manufacturing industry in 1946 fell to 26% of the
prewar peak level in 1934-1936 and the supply of food to 51%. Many of the plants
and equipment had been destroyed by bombing during the war and others were
destroyed or taken away by the Allied Forces. Still, more than two-thirds of the pro-
duction capacity had been left intact in most of the heavy industries. The R&D
resources that had been increased before and during the war, and the many workers
who had worked in the munitions plants and gained production experience returned
to civilian production. Of course, not all of these resources were useful. Some of the
facilities had been hastily built or were hastily converted with little use after the war,
and some of the R&D knowledge or production skills were obsolete. Yet, as Japan's
military spending after the war was practically zero, most of these resources could be
transferred for the purpose of civilian production.
Owing to various drastic economic measures and the boom during the Korean
War, the manufacturing production regained the prewar peak level within 5 years.
From the mid-1950s to 1973 (the year of the first oil crisis), the economy grew at an
annual rate of approximately 10%. Various growth accounting studies show that this
fast growth, which lasted almost two decades, was the result of a high rate of capital
accumulation combined with technological progress.
As implied at the end of the previous subsection, after the end of the war Japan
had to realize that it had again been left behind. The process of catching up started
again. As in the Meiji era, it followed a dual track: encouraging the importation of
advanced technology and promoting a domestic technology base. To make an effec-
tive technology transfer, the government allocated its scarce foreign currency selec-
tively to those firms capable of adapting and improving imported technology.9 This
gave the government, notably the Ministry of International Trade and Industry
(MITI), a strong power over the industries. In fact, most of the firms with adequate
technological background were given the allocation of foreign currency. However, a
few years delay often occurred and, as Peck and Tamura (1976, p. 553) said, "in a fast-
growing economy it can be of importance whether government policy delays or speeds
up the use of a particular technology by only three or four years."
Nevertheless, despite the effect this control undoubtedly had on individual firms,
it seems unlikely to have affected the overall pattern and pace of technology impor-
tation in a major way. A more profound influence, in our view, was provided by the
restriction of import and direct investment because the only way for foreign firms to
exploit their technological superiority was to sell their technology, even though they
might have preferred to export the product or start their own production in Japan.
That is, Japan's restriction of trade and investment seems to have made it possible for
Japan to import technology only. Therefore, the trade and capital liberalization in the
1960s and 1970s inevitably caused MITI to lose its control over technology and indus-
tries in general.
Another important consequence was that the Japanese manufacturing industries
remained mostly in the hands of Japanese shareholders, in marked contrast with Euro-
pean countries where many U. S. firms started their own businesses or purchased exist-
ing European firms as their subsidiaries. There are exceptions such as IBM, NCR, and
a number of petroleum refiners, which established their subsidiaries in Japan using the
"grandfather rights" to resume their pre-World War II operations or using the "yen-
based stock purchase system" that allowed foreign direct investment into Japan with
strict conditions during 1958-1963. However, major firms in the automobile, electric
equipment, and steel industries remained Japanese owned and the stakes foreign firms
had in Japanese electric and communication equipment companies in fact decreased
during the postwar period. Even after the capital liberalization in the 1970s, foreign
direct investment into Japan remained low. Thus the management of most Japanese
firms has been, and in large part still is, not exposed to the pressure of the international
capital market. The consequence of this on corporate motivation will be discussed
later.
It is difficult to evaluate these policies. On the one hand, it is fair to assume that
they have substantially benefitted domestic firms. On the other hand, they may have
lessened market competition by restricting entry of foreign producers either through
import or investment. In addition, the restriction of foreign direct investment pre-
vented the inflow of capital when the Japanese economy needed it. Even though the
intense competition among domestic firms, and the high saving ratio after the 1950s
mitigated these negative effects, the overall effect has to be assessed in a broader
context.10
We also note that this government policy was feasible only because there were
many firms eager to import technology in expectation of a high return. Since these
technology-importing firms could start manufacturing those products that had been
imported previously, import substitution took place reducing imports and eventually
increasing exports, thereby enabling Japan to economize on precious foreign currency.
The amount of foreign currency thus saved by far exceeded the payments to technol-
ogy importation. Through the 1950s and 1960s the limited availability of foreign cur-
rency worked as a ceiling to economic growth, that is, economic booms could not be
sustained because of acute shortage of foreign currency in boom years when imports
increased. Technology importation had the long-run effect of pushing this ceiling
upward and making fast economic growth sustainable.
In evaluating the role imported technology played in Japan's development, one
should also note the following two facts. First, trade in technology became very active
after the war in every country. Although Japan's payment to technology importation
(64 billion yen) was the largest among major countries even as recently as 1988, its
percentage to GNP, 0.17%, was about the same as in France (0.18%), Germany
(0.17%), and the United Kingdom (0.16%), though larger than the United States
(0.04%) (STA, 1990). That is, under the GATT-IMF regime, trade in both goods and
technology increased worldwide and Japan is hardly peculiar in this regard, though,
no doubt, it was one of the major beneficiaries of this trend.
Second, technology transfer would not have succeeded without the prewar tech-
nology base and a rapid postwar increase in R&D expenditures. As discussed earlier,
heavy industries, such as machines and steel, had a large share in the 1930s and early
1940s, owing partly to the military build-up. When the war ended, a large number of
researchers, engineers, and skilled workers were released from such production and
could be utilized in civilian production. As for the contribution of domestic R&D
effort to technology transfer, empirical findings indicate that those who imported tech-
nology also invested in their own R&D.'' Domestic R&D was essential to enable firms
to evaluate, adapt, and improve imported technology.
Emphasis on Own Innovation: Since the 1960s
As Japan became a serious competitor in the international market, the terms of tech-
nology importation became less favorable. In addition, as Japanese industries
advanced technologically and became world-class, the backlog of technologies avail-
able for import decreased. Consequently, during the 1960s when the Japanese econ-
omy grew at a high rate and started to compete internationally, the need for increased
efforts to develop its own technology became more urgent and government policies to
promote domestic R&D began to be emphasized. However, the size of the incentives
provided through tax breaks, subsidies, and low-interest loans were modest (see Table
3.2). In total, they amount to a little more than a hundred billion yen in 1983, or 2.6%
of industrial R&D expenditures and 2.3% of industrial R&D expenditures plus pay-
ment to technology importation. Thus, financial incentives by the government to the
private sector has been modest and doubtlessly smaller than in other countries where
as much as a one-third of industrial R&D is financed by the government.
Table 3.2 also reveals a clear downward trend in the importance of subsidies: the
proportion to industrial R&D was almost 8% in 1960, still lower than in other coun-
tries, but gradually declined to 2.6% in 1983. In parallel with the generally decreasing
role of industrial policy in Japan's economic growth, government support to industrial
R&D has been decreasing rather dramatically in the past two to three decades.
Private firms, under increased competition from other Japanese firms and from
American and European firms in foreign and domestic markets (owing to liberated
trade and direct investment), felt an urgent need to enhance their technological capa-
bility. Their R&D expenditures more than tripled in the latter half of the 1960s. Rel-
ative to these increased private efforts, the government policy measures were modest
indeed and, in our opinion, most of the R&D projects the government supported
would have taken place anyway.
One policy measure that has attracted considerable interest in the West is joint or
cooperative research efforts based on the Act on the Mining and Manufacturing Indus-
try Technology Research Association enforced in 1961. During 1961-1987, 87 such
research associations (RAs) were formed. In 1983, there were 44 RAs of which 38
received subsidies of 33 billion yen, or about half of their total R&D expenditures
Table 3.2. Government Support to Industrial R&D in Japan0''' (in Billion Yen)
(a) Subsidies and (b) Preferential Tax (c) Low Interest
Year (A) Total Research Contracts Treatment Rate Loan
1960 9.8 0.7 9.1

1965 16.4 3.1 13.3
1970 31.0 11.0 19.1 0.9
1975 64.7 29.8 33.0 1.9
1980 101.0 60.8 38.0 2.2
1983 117.7 58.7 57.0 2.0
(B) R&D Expenditure (A)/(B) (Q Payment for (A)/(B + C)

Year by Industry (%) Technology Importation (%)
1960 124.4 7.88 34.2 6.18

1965 252.4 6.50 59.6 5.26
1970 823.3 3.77 155.1 3.17
1975 1684.8 3.84 211.3 3.41
1980 3142.3 3.21 326.2 2.91
1983 4560.1 2.58 493.8 2.33
"Adapted from Goto and Wakasugi (1988), Table 1.
b
a is the amount of subsidies and research contracts, b is the amount of tax forgone through the preferential tax treatments to
promote R&D. c is the interest payments savings due to the low interest rate loan to promote R&D. A = (a) + (b) + (c).
Sources: (a) Somu-cho, "Kagaku gijyutsu kenkyu chosa hokoku" (Report on the Survey of Research and Development), each
year.
(b) Documents submitted to Tax System Council.
(c) Estimated from Japan Development Bank documents.
(Wakasugi and Goto, 1985). Since the total government R&D subsidies were about
59 billion yen in the same year (see Table 3.2), we infer that more than one-half of
them were supplied through RAs. This reflects an important aspect of RAs: for MITI,
RAs have been a convenient way to distribute its subsidies to promote the technology
MITI (and participating firms) believed important, most notably semiconductors and
computers, and have been used to avoid favoring particular firms and to minimize the
cost of supervising the use of subsidies. From this viewpoint, it is not surprising that
only two of the 87 associations had joint research facilities; in all other cases, each
member firm simply took its share of research funds and carried out the research in its
own laboratory. Therefore, how coordinated the research really was among partici-
pating firms within each RA is doubtful except for a few cases.12
The effectiveness of these RAs in generating new technology is also doubtful.
According to Wakasugi and Goto (1985) and Fujishiro (1988), RA's productivity as
measured by the number of patents divided by its R&D expenditures was considerably
lower than that of industries, though neither of these authors implies that it is the best
measure of R&D efficiency.13 Perhaps RAs were the product of the time of transition:
the firms were then feeling a need to promote more long-range and large-scale, if not
basic, research but were still unable, technologically and financially, to conduct it indi-
vidually. Their role has been declining as more and more collaborative research activ-
ities are now carried out by research institutions funded jointly by companies or under
intercompany technology agreements.
The first oil crisis of 1973 ended the high-growth era. Around the same time,
Japan adopted a flexible exchange system, resulting in a major appreciation of yen.
The annual growth rate fell from roughly 10% of the high-growth era to less than 5%
in the 1970s and 1980s. There were two consequences of these drastic changes on tech-
nology. First, developing energy-saving production processes became one of the major
targets in R&D for many businesses. Such R&D, together with the accumulation of
incremental innovation at production sites, made it possible for Japanese manufac-
turing industries to increase energy efficiency. For instance, the steel industry devel-
oped the so-called oilless steel-making process and reduced energy consumption dras-
tically. This shift in emphasis toward energy R&D was also apparent in government
R&D: the proportion of energy R&D among total government R&D funding
increased from 7.5% in 1975 to 16.3% in 1985 (National Science Foundation, 1988,
p. 55). The Sunshine Project aimed at developing alternative energy-generating tech-
nology and the Moonlight Project to develop energy-conserving technology started in
1974 and 1978, respectively.
Second, the entire industrial structure shifted toward an energy-saving, technol-
ogy-intensive, and high value-added one. Energy-intensive industries scaled down and
high-technology industries expanded rapidly. The striking example of the former was
the aluminum smelting industry. Japan was the second largest aluminum-producing
country in the world in the early 1970s. After a decade and two oil crises, it had been
virtually wiped out while the import had soared (Goto, 1988). By contrast, high-tech-
nology industriessemiconductors, computers, fine chemicals, and suchgrew
rapidly.
Through the 1970s to 1980s, therefore, Japanese firms had to cope with the dras-
tically changing environment. Yet, after the two oil crises and two sharp appreciations
of yen (65% against the U. S. dollar in 1977-1978, and 92% in 1986-1988), Japanese
firms found their share to have considerably increased in the world export market.
This increase was most prominent in high-technology products where the Japanese
share rose from 7.2% in 1965 to 19.8% in 1986. We may attribute this success in
Japan's adaptation to changing environment to two factors. The first is the high rate
of investment in R&D and in plants and equipment. The R&D expenditures grew 4.4
times from 1973 to 1987 and, as a proportion to GNP, from 2.0 to 2.8%. The ratio of
investment in plants and equipment to GNP fell gradually after the first oil crisis from
about 19% in 1970, but remained at 14 to 15%, and then grew in the 1980s to reach
19% again in 1990. The amount of investment in plants and equipment in Japan even
surpassed that of the United States, whose economy is 1.8 times larger than Japan's.
These active investments in R&D and in plants and equipment reinforced each other,
increasing productivity and improving product quality. In addition, as the second fac-
tor, the management system of Japanese firms contributed to their adapting to the
rapidly changing environment, as we shall discuss later.
With increasing strength of the Japanese firms in the world market, the focus of
the science and technology policy is gradually shifting from the promotion of indus-
trial technological capability to emphasize such new issues as the promotion of basic
research and the globalization of innovation. These issues will be discussed in the final
section.
TECHNOLOGICAL DEVELOPMENT IN THREE INDUSTRIES
In this section, we will describe the development of three industries, iron and steel,
electrical and communication equipment, and automobiles, since the Meiji era; this
will complement the general historical description previously given. '4
Iron and Steel
In 1901, the government constructed Yawata Steel Works (and then ran it) wholly
dependent on German technology. This case, it has been argued, epitomizes the two
salient characteristics of Japan's industrial and technological developmentgovern-
ment support and technology importation. Is such an argument justified? Are these
two characteristics really evident? A close examination of the history of the iron and
steel industry actually gives a different picture.
The first attempt to produce iron and steel using Western technology was made
almost a half century before Yawata. This was during the 1850s, still in the Tokugawa
era but after the government opened the country to the United States, the United
Kingdom, France, and Russia under military threat from these countries. Before this
period, iron and steel had been produced from iron sand and charcoal using an indig-
enous small-scale production method. However, with the threat of Western countries,
the Tokugawa government and some of the powerful feudal lords considered it urgent
to construct strong cannons. Because this created a demand for quality steel, they con-
structed reverberating furnaces copying the technology described in a Dutch book (an
example of technology import through the Dutch during the Tokugawa era) but
entirely with Japanese hands. These furnaces produced wrought iron on a small scale
out of pig iron made with indigenous technology. One of the engineers who made these
furnaces, T. Ohshima, proposed building a blast furnace to produce pig iron. Ohshima
was a samurai of Nambu Clan, northeast of Honshu. Nambu had an iron mine and
thus a history of iron making. Adding to this background, Ohshima went to Nagasaki
to study Dutch and learned the steel- and cannon-making technology with the book
mentioned earlier. He thus knew both indigenous and Western technology, and, with
the financial support of private investors, constructed a small blast furnace in
Kamaishi, a city near an iron mine in Nambu. Although the production met with dif-
ficulties at first (in firebricks, ventilation, etc.), in about a year the production started
to go smoothly and several more furnaces were built.
In 1874, the Meiji government decided to build a modern furnace in Kamaishi
and hired Ohshima and a German engineer to plan it. The German engineer proposed
building a large-scale modern mill (with a railway system to carry the product and iron
ore) around two imported furnaces. Ohshima opposed this plan and instead proposed
building five relatively small furnaces and using more modest transportation method.
The government, which never questioned the superiority of Western technology,
rejected Ohshima's plan and imported furnaces, railway cars, and other equipment
from Britain and hired British engineers to construct a mill based on the German engi-
neer's plan. It started the operation in 1890, immediately met with difficulties, and was
closed down after 196 days of operation. The cause was obvious. At first, the mill used
charcoal but, after a short while, the supply of such a large quantity of charcoal was
found out to be impossible. It thus decided to switch to coke as fuel but the coke it
could obtain did not have a sufficiently high quality. Needless to say, technology can-
not be free from social and economic conditions. Apparently, neither the European
engineers nor the Meiji government was aware of this simple principle in the belief of
superior Western technology. By contrast, Ohshima, who knew the way iron had been
produced by that time, was aware of the conditions surrounding iron production in
Japan.
Kamaishi restarted its iron production several years after the government closed
the mill. C. Tanaka, a private entrepreneur who bought the mill from the government,
restarted the production not with the big furnace but with a new small furnace (built
after Ohshima's technology). Again several improvements had to be made before the
operation became commercially viable, but he could expand the production gradually
and in 1894, with the help of K. Noro, a professor at the Imperial University (now
University of Tokyo), repaired and improved the British-built big furnaces and suc-
ceeded in restarting them.
The lesson of Kamaishi appears to have been forgotten by the time the govern-
ment built an integrated steel mill in Yawata, Kyushu. The government had been
experimenting with steel production within the arsenals using the pig iron produced
in Kamaishi or with traditional technology, and imported iron. However, from the
military viewpoint it was considered necessary to produce steel using a modern inte-
grated technology and on a large scale. Thus the government decided in 1896 to build
a mill in Yawata and let M. Ohshima (T. Ohshima's son) make the plans. Ohshima
had studied in Germany and decided to build the mill with German technology, com-
missioning a German company to make a detailed plan, buying German equipment,
and hiring German engineers for the top engineering posts. Completed in 1901, it was
a spectacularly modern and large plant. The excitement, however, did not last long.
The operational problems mounted and the mill was forced to stop operations in
about a year. The major problem was the different quality of coke between Germany
and Japan. Yawata was chosen as the site basically because there was a big coal mine
nearby, and the mill used the coke produced with this coal. The coke production mill
was designed by Germans as well, who never understood the difference between Ger-
man and Japanese coal and were also under a budget constraint by that time. Conse-
quently, the coke mill was not only inadequate for Japanese coal but also inefficient
owing to the use of cheaper obsolete equipment. The coke thus produced was inevi-
tably poor in quality, and made steel production extremely difficult.
Again, like in Kamaishi, Noro was asked to improve the situation. Noro had ear-
lier proposed his own plan for the Yawata mill, which, not surprisingly, suggested a
more modest production scale and a gradual expansion. But the government rejected
it and he had resigned from government posts for an unrelated reason. Noro made a
thorough investigation of Yawata, improved the coke production process, and made
several changes to the furnaces to fit them to the attributes of Japanese coke. These
changes took a few years but in 1904 the mill restarted its operation successfully.
Yawata Mill was owned and run by the government, mostly because of its role as
the supplier of steel to weapon production. Though it was obviously a dominant steel
producer, private enterprises were not absent. In addition to Tanaka's Kamaishi mill,
Sumitomo, Kawasaki, and Mitsubishi (in occupied Korea) entered into the market
and so did Nippon Kokan (now NKK) with the initiative of K. Imaizumi, Noro's stu-
dent, who had resigned from Yawata when his proposal to privatize Yawata was not
accepted. Hence, private profit-seeking entry did take place. Because of the large initial
investment, many of the entrants were zaibatsu companies or financially supported by
zaibatsus (for instance, Asano, Okura, and other relatively small zaibatsus invested in
Nippon Kokan). Some of them built integrated mills, but some imported pig iron and
scrap iron from India, America, and other countries. In 1934, a partly government-
owned company, Nippon Steel, was established to achieve a stable supply of steel
(which the army considered indispensable). Yawata was absorbed into this company
and the government urged private steel companies to join. Kamaishi and a few other
companies complied with this request but Nippon Kokan, Sumitomo Metal, Kawa-
saki Steel, and other companies kept their independence.
Since the aim of this subsection is to show that Japan had a certain technological
background when it decided to import Western technology and that technology
importation alone cannot guarantee successful industrial development, we shall dis-
cuss the postwar development only briefly. In terms of innovation, two developments
had a significant impact. The first is the importation of the basic oxygen furnace
method from Austria and its improvement and adaptation to Japanese steel mills dur-
ing the 1950s. Japan was ahead of other countries not only in the speed of adoption
but also in the extent it improved the technology (see Lynn, 1982). For instance,
Yawata invented the oxygen converter gas recovery system as an effective device to
save energy use and reduce pollution, which has been used worldwide. The second is
the introduction of the continuous casting method and strip mills during the 1950s
and 1960s.
Two factors explain why these technologies were most effectively utilized in
Japan. One is the strong propensity of Japanese firms to invest in production pro-
cesses. Since both of these new technologies have to be embodied in capital equipment
and require a large-scale integrated production flow, their advantage can be exploited
only with new investment. Japanese iron and steel manufacturers in fact invested in a
number of new and integrated large plants, and productivity has increased tremen-
dously. For instance, between the two plants of Nippon Steel, a new one at Kimitsu
had a labor productivity (steel production per worker) 2.5 times higher than an old
one at Yawata in 1973, though Yawata's productivity itself had increased several times
in the preceding 30 years (lida, 1979). The other factor is the constant effort to improve
efficiency within the plants. Some of them were conspicuous, for instance, the intro-
duction of computers to control operation; however, some were made more at the
shop-floor level based on learning by doing and workers' proposals. As we shall discuss
later, Japanese firms have been relatively successful in attaining companywide
involvement in productivity increase.
As a result, Japan's productivity has outpaced the other countries and the com-
panies started to export technology, for instance, imparting know-hows in plant con-
struction and operation to many countries including both developing countries such
as Brazil and developed countries such as Italy and the United States. Since 1974 its
royalty receipts have exceeded its payments, the first among Japanese industries.
Let us conclude this subsection. In terms of industrial policy, the role played by
the government and the military in the development of the steel industry was obvi-
ously large (as in many other countries). Steel production is inseparable from the mil-
itary needs, as exemplified by the fact that the first steel production under Western
technology was made to make cannons. However, both the cases of Kamaishi and
Yawata imply that the government is less qualified than the private sector in selecting
appropriate technology. In either case, the government placed too much confidence
in the Western technology as it was and disregarded local conditions, such as infra-
structure and the accumulated indigenous technology. By contrast, T. Ohshima's and
Tanaka's attempts at Kamaishi suggest that private enterprises were better suited for
this purpose.
Electrical and Communication Equipment
The electrical and communication equipment industry gives another example of a

combination of Japan's indigenous technology or its own R&D efforts with imported
Western technology. Such a combination was feasible and effective because, though
Japan technologically lagged behind America and Europe (as in other modern indus-
tries) in the late nineteenth century when electrical and communication equipment
started to be introduced, the lag was not large. Take an example of the electric light.
T. W. Swan invented an electric incandescent bulb in 1878 in the United Kingdom
and T. A. Edison, in 1879 in the United States. The first incandescent light was exhib-
ited in Japan in 1884, only 5 years after Edison's invention. Needless to say, all the
equipment used in this exhibition was imported. However, in a similar exhibition in
the following year, a domestically produced generator was already used. This generator
was designed by I. Fujioka, a young professor at the College of Engineering (later the
Engineering Department of the University of Tokyo, as explained earlier) who had
studied with a British professor at the College. Fujioka then undertook to make a bulb.
He left the university to join Tokyo Electric Light Company (now Tokyo Electric
Power Company), was sent to America and Europe to learn the technology there,
bought bulb-manufacturing equipment in the United Kingdom, and came back in
1887. He started the development effort and then, to continue this effort indepen-
dently from Tokyo Electric Light Company, established a new company, Hakunet-
susha, in 1890. Even with the equipment brought from the United Kingdom, the
development met with continuous problems. For instance, following Swan's inven-
tion, his team first used cotton yarn for filament, but the result was unsatisfactory, and
it took some time before they learned about Edison's use of Japanese bamboo and did
the same. Hakunetsusha started selling the bulbs in 1890. Thus, although the United
States was unquestionably earlier in innovation and was dominant in production size,
the start of the industry in Japan did not lag by more than a generation.
As discussed earlier, Japan did not have a means to protect the domestic indus-
tries because of Unequal Treaties. Hakunetsusha met with fierce competition from
imports. Its bulb cost about 60% more than the imports. The company, however, man-
aged to survive with the booms after the Sino-Japanese war of 1894-1895 and the
Russo-Japanese war of 1904-1905, and also because of the international price increase
caused by an electric bulb cartel of European manufacturers in 1903. In the meantime,
it continued its effort to improve the product and made two innovations, a dual coil
bulb and a frosted bulb, that were copied worldwide.
Hakunetsusha consists of one-half of the present Toshiba. The other half came
from Tanaka Seisakusho, established by H. Tanaka. Born in 1799, Tanaka was one of
the most original and productive inventor-engineers (karakuri masters) in the Toku-
gawa era. At the age of 14, he already invented a loom. At 20 he made a "karakuri
doll" with a hydraulic mechanism, which became very popular. The list of his inven-
tions and developments is surprisingly comprehensivea clock, a torch lamp, a fur-
nace, a pump, a fire extinguisher, a tobacco cutter, an ice maker, a bicycle, an oil press,
a rice-polishing machine, a ship, a dredger, guns, and many more. He even made the
first steam engine in the country based on Western technology. In 1873, at the age of
74, he was invited by Kobusho (the Ministry of Industries) to come to Tokyo from his
hometown in Kyushu to make telegraphs at the ministry's small factory. After this
experience, he established his own company, Tanaka Seisakusho, in 1875 and made
telegraphs, switches, and miscellaneous electrical and other equipment. For Tanaka,
therefore, technological development was continuous, with little discontinuity
between indigenous and Western technologies. Tanaka's role in Japan's industrial
development was important also because among those under Tanaka's direction in the
Kobusho factory were several engineers who later became pioneers themselves. These
include S. Miyoshi, who later helped Fujioka in establishing Hakunetsusha, and K.
Oki, the founder of the present Oki (to be discussed soon).
Although H. Tanaka died in 1881, his son took over and expanded the company
with the production of torpedoes and mines at the request of the navy. In a few years
it became one of the largest manufacturing companies at the time, but as the navy
started to use competitive bids and then completed its own work, the demand
decreased substantially and the company started to lose money. The main creditor to
the company, Mitsui Bank, took over the virtually bankrupt company and renamed
it Shibaura Engineering Works. The company's main products were steam engines,
power generators, and other heavy electrical equipment.
Even though Shibaura had developed its own generators (after copying imported
ones), it found its technology still behind the American's. Thus in 1910 it formed an
affiliation with GE of America. GE acquired about a quarter of the share and allowed
Shibaura the use of GE's technology. For instance, engineers were sent to GE to learn
its technology and production methods.
Hakunetsusha, renamed Tokyo Electric in 1898, was also affiliated with GE, with
GE acquiring a 55% share. Since it started with the production of electric bulbs, the
main business was light electrical equipment and appliances, in contrast to Shibaura,
which was biased toward heavy equipment. Thus the two were considered comple-
mentary and, with GE's support, the two merged in 1939 to become Tokyo Shibaura
Electric (later Toshiba). GE owned 33% and Mitsui owned 15%. The relation with GE
continued until the beginning of the war and, after the war, resumed in 1953 with GE's
24% share, though this percentage has decreased substantially since then.
Toshiba's main rival, Hitachi, is rather unique among the Japanese producers in
that it relied on its own R&D efforts (together with reverse engineering) without form-
ing any affiliation with foreign companies. It was established by N. Odaira, a graduate
of the University of Tokyo's Engineering Department, in 1911 as a plant affiliated to
Hitachi Mine, and became independent 9 years later. It started with the production of
generators, blowers, transformers, and other heavy electrical equipment to be used in
mining, but soon began selling to other customers and expanding the business to
include industrial machines and lighter electrical equipment.
Hitachi never entered into any affiliation with foreign companies, in contrast to
Toshiba, which affiliated with GE, Mitsubishi Electric, which affiliated in 1923 with
Westinghouse (with 10% ownership), and Fuji Electric, which started in 1923 as a joint
venture between Siemens (30% ownership) and Furukawa Mine. That Hitachi became
the largest among these four main manufacturers must suggest either that Japan's
technological lag was not serious or that Japanese engineers were capable of develop-
ing the products by themselves with the help of reverse engineering. It must also be
noted that even companies with foreign affiliations made a significant amount of its
own R&D efforts. Thus the two predecessors of Toshiba, Tokyo Electric and Shibaura
Engineering, established their laboratories in 1918 and 1921, respectively, while affil-
iated with GE. Hitachi established its laboratory in 1934, though it started an inde-
pendent R&D section much earlier in 1918.
Let us turn to communication equipment. The Meiji government started tele-
graphs as a government business in 1868 with a British engineer using equipment
imported from Britain. Kobusho, the ministry, wished to supply them domestically
and, as mentioned above, asked H. Tanaka to undertake it. Tanaka succeeded in mak-
ing them in Kobusho's experimental factory and later in his own Tanaka Seisakusho.
In 1877, only a year after A. G. Bell's invention, Kobusho experimented with the first
telephone in Japan using American equipment. The engineers at the Kobusho factory
started its domestic production by copying and in the next year came up with its first
product, however unsatisfactory the quality was. One of the engineers, K. Oki, felt the
business had a bright future and left Kobusho to start his own company named Mei-
kosha (later Oki). The telephone business itself was not established until 1890, partly
because of scepticism about its merit (for instance, hiring a messenger was thought to
be a cheaper and more reliable way to convey information) and partly because of the
controversy within the government as to whether the business should be run by the
public or private sector. However, soon after the start (as a government business),
the public was easily convinced of its convenience, and the demand surged. Both
imported and domestic equipment (receivers, switchboards, wire, etc.) were used and,
as the sole domestic producer, Meikosha prospered.
Western Electric (WE) of the United States saw a growing market in Japan and
decided to enter through a joint venture. At first, it thought of teaming with Meikosha,
but the prospering Oki proposed conditions too severe for WE to accept. WE finally
decided to team with K. Iwadare, who had been WE's agent in Japan, and established
Nippon Electric Co. (NEC) in 1898 with WE owning 54% of the share and Iwadare
owning 33%. To start the business, NEC bought one of the largest manufacturing fac-
tories at the time, owned by S. Miyoshi but on the brink of bankruptcy as a conse-
quence of its too rapid expansion and the depression at the time. As mentioned earlier,
Miyoshi was another Tanaka disciple at the Kobusho factory and also a collaborator
with Fujioka in establishing Hakunetsusha. Hence, one can see a historical connection
from Tanaka, the master of pre-Meiji technology, not only to Oki but also to NEC, in
addition to Toshiba.
Iwadare's career illustrates another (and newer) type of engineer at the time, that
is, university educated. He studied electrical engineering at the College of Engineering
a year after Fujioka, worked at Kobusho for 4 years, went to the United States and got
a job at GE, came back to Japan after about a year to be an engineering manager of
Osaka Electric Lamp Company, and, after 8 years, quit this post to start his own busi-
ness as an agent for GE and WE. Thus, his career, like Fujioka's, is more "Western"
than Oki's or Miyoshi's in that he studied technology in a university designed and
taught by Westerners, and then was employed as an engineer in large companies at the
time.
In the beginning, NEC not only produced telephone and other communication
equipment with WE's technology, but also copied WE's management methods
(including the accounting system and work management) and bought WE's supplies
ranging from machines and materials to notepaper. However, it gradually replaced
them with their own or with products procured domestically. With the expansion of
the telephone system, the company expanded rapidly and competed against Oki and
new entrants including Toshiba, Hitachi, and Fuji (whose communication equipment
business was later separated as Fujitsu). NEC started the R&D department in 1926,
which was expanded into a separate laboratory in 1939. From just assimilating WE's
technology at first, it gradually increased its own R&D efforts and started producing
innovations that surpassed WE's.
The technological and industrial development of Japan's electrical and commu-
nication equipment industry shows those characteristics that can be commonly found
in many other industries. First, Japan was technologically behind America and Europe
but the lag was in a tolerable rangea marked difference from present developing
countries. Second, as a consequence, imported technologies were not entirely discon-
tinuous from indigenous technologies. Third, therefore, some of the entrepreneurs/
engineers could utilize the skills and knowledge accumulated in the pre-Meiji era.
Fourth, at the same time, those engineers educated by the colleges established by the
Meiji government and taught by foreign professors greatly contributed to assimilating
Western technology. Fifth, foreign technologies were actively imported through joint
ventures, technology contracts, reverse engineering, and such. Japanese had an ability
to absorb or adapt the imported technologies, and an ability to run the joint ventures
successfully. It should also be noted that a strong desire to become independent of
foreign dominance of technologyone may call it nationalismwas common among
the entrepreneurs/engineers as illustrated most clearly by Hitachi's Odaira. Finally, a
strong entrepreneurship and a strong will to enter into uncertain and unknown fields
were present. This conclusion does not imply that the government played no role. It
did play an important role in establishing science and engineering education systems,
introducing new technologies such as the telegraph, telephone, and weaponry, guar-
anteeing demand through military procurement, and so forth. In our view, however,
such policies would have never resulted in Japan's development were it not for the
willingness and capability in the private sector.
This conclusion is reinforced by the postwar experience. Although the govern-
ment continued to affect the industry through its nurturing policy during the 1950s
and 1960s, procurement by Nippon Telephone and Telegram, and research done
within MITI's Agency of Industrial Science and Technology and NTT laboratories, its
significance was more limited than previously because of the minimal defense research
and military procurement. As for the nurturing policy, the computer industry was the
main subject. Yet subsidies given to the industry, its sizable portion being granted
through research associations discussed earlier, were very small compared to the total
R&D expenditures of the industry. Apparently, it was the competitive threat from
domestic rivals and foreign giants such as IBM that fostered firms to R&D activities.
Their R&D expenditures increased more than 40-fold during the 20-year period from
1965 to 1985.
The importation of technology and its improvement within Japanese laboratories
and factories continued. After World War II, almost all the firms resumed their rela-
tion with American or European firms. Even firms such as Hitachi, which kept its
independence before the war, entered into contracts to import technology from com-
panies such as RCA, GE, Phillips, and Western Electric. Royalty payments thus
greatly exceeded receipts. However, with continuous innovation by Japanese firms,
the receipts have increased much faster than the payment: 13.4 times as opposed to
2.5 times from 1973 to 1988. Hence, though payment still dominates receipt, the dif-
ference has narrowed considerably: in 1988, payment was 1.65 times larger.15
Automobiles
The history of the development of the automobile industry also shows similar char-
acteristics: the dominance of foreign producers in the beginning, a number of (domes-
tic) entry attempts, the acquisition of technology first with reverse engineering, the
gradual accumulation of technologies with learning by doing, the procurement by the
military and the transportation authority, support by the government through finan-
cial incentives and standard setting, and the entrepreneurs' risk-taking behavior in the
form of physical investment and R&D. After several decades of struggle, Japanese car
manufacturers have now caught up with the European and American rivals techno-
logically, and are now surpassing them in certain areas. It is therefore worthwhile to
trace the Japanese experience in this industry in some detail.
The first car introduced into Japan was an European-made steam-driven one
imported in 1897. However, it was in 1903 when two cars (one steam-driven and the
other electric-driven) were used as buses in a large exposition that many Japanese peo-
ple saw cars for the first time. One of the visitors to this exposition saw them and started
an effort to make one domestically. This effort produced one steam-driven car, which
had so much trouble (particularly in terms of the tires) in the test run that he gave up
any commercial application. Four years later, the first domestic production of a
motor-driven car was made with the main parts (including, it seems, engines)
imported at first but gradually replaced by domestic production. This effort also col-
lapsed after selling 10 cars, 8 of which were bought by the police to add to its stock of
41 European cars and 11 American cars. In the meantime, import of foreign cars
increased and completely dominated the market: more than 200 cars had been
imported by 1910 and 6800 by 1920. The Great Earthquake in Tokyo of 1923
increased this trend further because all the train services had been disrupted and cars
were urgently needed to deliver food and other necessities. Thus by 1925 more than
16,000 cars had been imported, and in this year Ford established an assembly plant in
Yokohama. This was followed by GM, which established a similar plant in Osaka 2
years later. Thus the market became dominated by these two American producers.
More than 10 attempts had been made during the Meiji era to produce cars
domestically but all failed, and the first significant one was made by a company named
Kaishinsha established in 1911 by M. Hashimoto, who had studied engineering in the
United States for 3 years. He started the business by importing and repairing British
cars, and then with the technology learned from these British cars and his experience
with American cars, began to develop a passenger car domestically. His first car was a
failure because of lack of the casting technology needed to make a sufficiently durable
engine, but he finally succeeded in producing a viable product in 1913. He began sell-
ing this car with a brand name "Dat," but, in competition with Ford and GM, could
not sell in sufficiently large numbers to make the production commercially viable.
The first government initiative was made by the army who, after a survey of mil-
itary procurement policy in Britain, France, and Germany, concluded that it was
urgent to foster domestic production of trucks and buses for military purposes. In 1918
it started a policy of subsidizing domestic production and civilian purchase of trucks
and buses. Although the subsidy itself was modest, this policy gave an expectation of
increased procurement and three companies started production. One was Kaishinsha,
who, with the difficulty in the passenger car market, decided to enter into the produc-
tion of trucks; the second was Tokyo Gas Electric (now defunct), which inherited the
trial truck production in the arsenal; and the third was Tokyo Ishikawajima Zosensho
(now IHI), which, with the profits from shipbuilding during the World War I boom,
entered into the production of passenger cars and then trucks, first by disassembling a
Fiat car to learn the technology and then by buying technology from Wolseley, a Brit-
ish company. Despite the army's support, domestic production was small. During
1919-1931, these three companies together produced 2575 cars and the market
(including trucks and buses) was totally dominated by Ford, GM, and other imports.
The next governmental initiative occurred in 1930 when the Ministry of Railways
made a detailed open test of the three domestic producers, and then in 1932 when the
Ministry of Commerce and Industry (now MITI), in collaboration with the Ministry
of Railways, urged the three companies to collaborate in a joint venture to develop a
so-called "standard" car (truck and bus) with the brand name "Isuzu" (the ancestor
of current Isuzu). The joint venture lasted for only 2 years with the production of 750
cars, but the experience was utilized in these companies' efforts to produce their own
trucks and buses (both for military and civilian purposes).
Although the major part of Kaishinsha was absorbed in this joint venture and
later by Isuzu, the business of small passenger cars (called "Datsun" by this time) was
retained by G. Ayukawa, an energetic entrepreneur in the process of expanding its Nis-
san Zaibatsu (one of the so-called new zaibatsus), who had acquired Kaishinsha by that
time. Ayukawa established Nissan Motors for this purpose. Another origin of Nissan
Motors was a company called Jitsuyo Jidosha, which was started with the technolog-
ical guidance of American engineers using the engines of American Harley-Davidson
motorcycles. Hence, Nissan's technology originated both from a Japanese's effort to
copy British and American cars and the technology brought by hired Americans.
The number of entry attempts around this time was surprisingly large. To trucks,
Mitsubishi and Kawasaki made some efforts. To motorcycles and three-wheel cars,
more than 10 companies made efforts, some of whom, like Daihatsu and Mazda, sur-
vived until later, and some of whom eventually exited. A number of efforts were also
made in the passenger car market, including Mitsui Shipbuilding, but mostly by small-
scale nonzaibatsu companies. Most failed but an important exception was Toyota.
Two facts deserve emphasis. The first is the frequency of entry attempts, clearly indi-
cating the presence of risk-taking spirits and the will as well as capability to challenge
new technologies. In short, Schumpeterian entrepreneurship was abundant. Second,
though this entrepreneurship was present among established zaibatsus, their role was
limited. In fact, they gave up rather early. By contrast, independent entrepreneurs such
as Ayukawa, Toyoda (Toyota), and Matsuda (Mazda), not to mention Hashimoto and
others in the early period, were more persistent.
Sakichi Toyoda was a typical inventor and entrepreneur of pre-World War II
Japan, much like Thomas Edison, though in a different field. Born in 1867 in a poor
carpenter's family, he invented the first (wooden-made hand-driven) loom in 1890
after several years of struggle. He started selling the looms and making cloth using
them, but the business did not prosper. The struggle continued until he invented auto-
matic looms in 1897. This time the business was a success with a strong demand both
for the looms and the high-quality cloth made with them. He worked further to
improve the loom, and to make a wider cloth faster and more efficiently. The company
started exporting the products and the quality of his loom was proved when the world's
largest loom manufacturer, Pratt (U. K.), bought the technology for one hundred
thousand pounds. It is worth noting that Toyoda's development effort over his lifetime
was a purely private and independent enterprise. Though Mitsui Bussan (the largest
trading company at the time, now Mitsui Co.) at one time gave financial support to
Toyoda's enterprise, Sakichi parted the relationship fairly soon after a disagreement.
Sakichi's son, Kiichiro, plowed back the profits from loom sales and the fee from
Pratt into the development of cars. He organized a development team by hiring engi-
neers from fields such as steel, tools, and three-wheel cars from other companies. He
visited several professors to learn about technology, some of whom were his former
classmates when he studied at the University of Tokyo's Engineering Department.
And he sent his engineer to the United States to visit the Ford plant and get ideas on
how to make a factory, and purchase necessary machines. The team started by disas-
sembling American cars. Finding that the American technology was far advanced,
they decided to imitate Chevrolet's engine and use many of Chevrolet's and Ford's
parts. As in the previous experience of Kaishinsha, casting was a problem. Although
Toyoda had the casting technology to manufacture looms, they needed a more com-
plex and finer technology to produce an engine. After a seemingly endless process of
trial and learning in casting and other aspects of technology, the first viable product, a
truck, was introduced in 1935 to sell to the army. A passenger car followed in 1936.
An important policy initiative was taken in 1936 to foster domestic car produc-
tion. The government restricted the production to licensed companies who received
financial and other support. In contrast, the biggest producers at the time, Ford and
GM, were banned from increasing their production level in the future. The license was
given to Toyota (Toyoda's brand name for cars), Nissan, and (later) Isuzu. Toyota and
Nissan thus became the dominant producers and increased production. In 1938, the
production level of Toyota and Nissan combined was 6568 for nonmilitary sales ver-
sus 12,441 of Ford and GM. For military sales, they were 5930 and 5999, respectively,
showing the army's willingness to support domestic producers. Both Ford and GM
exited the Japanese market in 1939, and the production level of Toyota and Nissan
increased further, each exceeding 15,000 around 1941, with Isuzu following with
about half the production level. By this time, most of the demand was for trucks for
military purposes.
Without question, therefore, military procurement played an important role in
supporting domestic car production. However, the quality of domestically produced
cars never caught up with that of Ford and GM before the war. Complaints of break-
downs were frequent, including broken shafts, water dripping out of radiators, and
early wear of moving parts.16
Except for the short period of the Korean war, the postwar production shifted to
civilian uses. The big question at the time was whether Japan should protect and nur-
ture domestic producers following the infant industry theory, or import cheaper and
better-quality cars from abroad. The first view was taken mostly by MITI and the pro-
ducers, whereas the latter was taken by the Ministry of Transport, the Bank of Japan,
users such as taxi companies, and dealers of imported cars. The government settled on
the first position and restricted imports quantitatively and with high tariffs. This pro-
tection guaranteed a growing market to domestic producers, giving them a chance to
lower costs and improve quality through scale economies, learning by doing, and
imported technology. Compared to this effect, the financial incentives provided by the
government, such as the provision of low-interest loans through Japan Development
Bank and accelerated depreciation allowed to selected equipment, seem less impor-
tant. Government support specifically aimed at R&D was very limited.
To catch up technologically, Toyota relied on its own efforts but Nissan, Isuzu,
and Hino imported technology from, respectively, Austin (U. K.), Rootes (U. K.), and
Renault (France). In the latter case, at first they imported major parts including
engines, but gradually replaced them with domestically produced ones. All these com-
panies incorporated the technology they learned from the production of licensed mod-
els into the development of their own models such as Datsun, and terminated their
relationship with the foreign producers by 1965.
These producers, all established before the war, were not the only car manufac-
turers. Entry attempts were many, counting nearly 30 companies between 1945 and
1960. These included ex-zaibatsu companies, such as Mitsubishi Heavy Industries,
and reentrants, such as Mazda, but the majority were new independents including
Honda and Suzuki, both of which started as motorcycle producers. However, except
for these companies, few of them survived more than 5 years. These private efforts
were against MITI's wish to integrate domestic car production to two groups, Toyota
and Nissan, to attain scale economies large enough to compete against American
giants after trade and capital liberalization. MITI's guidance bore some fruits in the
form of a Nissan-Prince merger and an alliance between Toyota and Hino (and later
Daihatsu). As a consequence, market concentration to the two groups increased, but
the entry of several powerful independents kept the market highly competitive, which
together with the increasing importance of global competition contributed to main-
taining and even intensifying the innovation efforts of Japanese car makers.
This experience of the Japanese car industry throws lights on some key issues.
How could Japan succeed in establishing its own car industry under the presence (in
the prewar period) of technologically far advanced Ford and GM? Two factors appear
most important. One is the presence of engineers and entrepreneurs, such as Toyoda
and Ayukawa, who were willing to take risks and sustain efforts under adversity. The
other is the general ability of engineers to absorb foreign technology and the ability of
workers to absorb new production processes. Needless to say, the education system
discussed earlier contributed to this effect. Also, the emphasis on engineering educa-
tion in universities helped, as shown by the episode of Toyoda's visiting professors to
ask technical questions.
As in any late-developing countries, reverse engineering was the first source of
foreign technology. Visiting advanced factories of foreign producers, hiring foreign
engineers, and purchasing foreign technology were other important means of acquir-
ing technology.
The government's role was important in two aspects, military and other procure-
ment, and the protection of domestic producers. These policies, by guaranteeing
demand, encouraged investment in capital and R&D (mostly development rather
than research, particularly in the early period). Nevertheless, one should recall that
many entry attempts were made even before government policies or against MITI's
guidance. It is misleading indeed to attribute Japan's success solely to government pol-

icies.
Concluding Remarks
This section gave three case studies to show how Japanese industries acquired and
developed technology in its broad sense. Since the urgent economic target for Meiji
Japan was to acquire more advanced Western technology and build competitive
industries, our cases focused on "modern" industries and mostly the pre-World War
II era. In these industries, intensive efforts were made to transfer technologies from the
Western countries. The means were various. In steel, especially the case of Yawata
Work, the basic strategy was to import a whole setplan, technology, equipment, and
personnel. In automobiles and electrical/communication equipment, reverse engi-
neering and technology import through licence or joint ventures were common. The
balance between these two means varied between the two industries. Apparently, joint
ventures were more common in the electrical/communication equipment industry.
The balance also varied across companies within each industry; thus, Toshiba, NEC,
and Nissan relied more on imported technology than did Hitachi and Toyota. It is
difficult to assess the relative merit of the two strategies.
We also revealed a significant role played by indigenous technology. A right
choice of technology to import is easier to make with knowledge of indigenous tech-
nology, and imported technology has to be modified with the help of indigenous tech-
nology that is compatible with local conditions. The cases of Kamaishi and Yawata
Steel Works clearly demonstrate this fact. Reverse engineering needs understanding
of the basic mechanism behind products and the skills to imitate, which the masters
of indigenous technology, such as H. Tanaka, tended to have.
In addition to these technology-specific factors, there were, first, the persistent
efforts and inquisitive minds of industrial leaders and, second, educational back-
ground and skills of middle-to-top engineers and key workers. The government's hur-
ried effort to establish a compulsory education system and engineering schools was
clearly a contributing factor, though the effort would not have achieved its purpose
were it not for the parents' enthusiasm for education and the background of wide-
spread private education in the pre-Meiji era. Besides, education may be necessary but
not sufficient to produce entrepreneurial will, which, as Schumpeter has so forcefully
argued, is the key factor in economic development. A full inquiry into the factors sup-
porting such will is beyond the scope of this study. Some of them may be similar to
the Japanese business system today, to be discussed in the next section; however, some
are not. For instance, long-term employment was not a norm before World War I as
illustrated by NEC's Iwadare's career.17
Some observers stress the role zaibatsus played. Zaibatsus, such as Mitsui, Mit-
subishi, and Sumitomo, were big, diversified, and family-owned businesses often with
government or military connections, and were relatively strong in finance, mining,
and shipbuilding as they acquired these businesses when the government privatized
them. Among the three industries studied here, their presence was relatively large in
steel because the investment required large funds. Even in this industry, however, they
remained as followers to giant Nippon Steel. In the other two, their role was further
limited. In fact, it was independent entrepreneurs who made risky investment and per-
sistent efforts to catch up with the West technologically. Some of them had certain
relations with zaibatsus, including Ayukawa, whose business expanded into one of the
so-called new zaibatsus; Tanaka's company, which became partly owned by Mitsui
because of bankruptcy; NEC, which formed a tie with Sumitomo after it intermediated
between WE and Sumitomo in WE's entry into the wire business18; and Toyota's auto-
matic looms, which were sold by Mitsui's trading company from time to time. In nei-
ther of these cases, however, did zaibatsu (that is, established zaibatsus) make a first
move or become a dominant player. As Ayukawa's case suggests, zaibatsu is, after all,
an enterprise that has successfully expanded and diversified; hence, it hardly diners
from large diversified enterprises in other capitalist economiesfrom conglomerates
in the United States and European countries to chaebols in Korea (see Chapter 11 on
Korea) and other equivalents in developing countries. Thus, it is misleading to empha-
size the peculiarities of Japan's experience in reference to zaibatsu.
Zaibatsus were dissolved after the war and the present Mitsui Group, Mitsubishi
Group, or any other group is no more than a very loose federation of basically inde-
pendent companies, unlike pre-World War II zaibatsu where there was a holding com-
pany that served as headquarters. Thus it is now frequently observed that a Mitsui
company purchases from or sells to, say, a Mitsubishi company, or gets loans from
Mitsubishi Bank. On this difference between prewar zaibatsu and present business
groups, see Goto (1982) and Odagiri (1992, Chapter 7).
We have discussed the development of the three industries after World War II
only briefly, because the discussion of the prewar experience alone was long enough,
and the literature on the postwar experience is relatively plentiful. Komiya et al.
(1988), in our view, gives the best discussion both on the postwar industrial policy in
general and individual studies on the industries discussed in this section."
We have discussed the general shift among Japanese industries from technology
import to their own R&D efforts, and described the postwar R&D policies. More gen-
erally, the industrial policy (not particularly specific to R&D) may be summarized as
follows.
1. Even in the early years the amount of government funding to industries, including
subsidies, tax credits, and low-interest loans, was rather modest, though some
industries, such as iron and steel, coal mining, shipping, and electric power, bene-
fitted significantly.
2. There is a clear tendency that such government funding has decreased in amount
and importance over time; in fact, it appears minimal and by no means larger than
in other industrialized countries in recent years.20
3. The restriction on imports and foreign direct investment into Japan was probably
the most important policy until the early 1970s. Restricting the growing Japanese
market, already the second largest in the capitalist economy in the late 1960s, to
Japanese firms who were competing intensively among themselves gave a strong
incentive to invest in plants, equipment, and R&D. In addition, because postwar
Japan's Peace Constitution meant that the military was no longer a significant cus-
tomer to businesses, industries such as automobiles, which had been helped by mil-
itary procurement before the war but was still in its infancy relative to American
and European producers, might have been wiped out were the market made open
to foreign competition. However, such restriction on trade and investment has
been drastically reduced in the 1960s and the 1970s so that the rates of tariff and
the number of manufactured goods under import restriction are among the small-
est in the world.
4. Procurement by government corporations such as Nippon Telephone and Tele-
graph (NTT, privatized in 1985) and Japan National Railways (split into six
regional companies and privatized in 1987) helped the communication and elec-
tronic equipment industry, and the rolling stock industry, respectively. However,
they now buy from foreign suppliers as well.
5. MITI's role in the collection and diffusion of information may have been signifi-
cant, as it could obtain information on overseas markets through its Japan External
Trade Organization (JETRO) and technological information through its Agency of
Industry Science and Technology, which has a number of laboratories. However,
this role has also declined as firms themselves accumulated international experi-
ences and technological knowledge. In addition, Japanese firms have various other
"networks" through, for instance, trading companies (sogo shosha), trade associ-
ations, and buyer-supplier relationships.
Thus, government policies, especially the protection of domestic markets, played

a significant role in Japan's postwar industrial development at least until the early
1970s. However, as argued in the beginning of this chapter, government policies can
attain their purposes only if the private sector is willing and able to take the opportu-
nities given to them (or even to overcome the obstacles the policies create, as evidenced
by Honda's experience). To understand why Japanese firms were willing and able, and
to understand why they continue to be so despite the weakening role of government
policies, one has to know how Japanese businesses today make decisions and carry out
innovation efforts. This is why a detailed examination will be given in the next section
on innovation inside Japanese firms. It will be preceded by descriptions of the current
R&D efforts and performance in the aggregate.
INNOVATION IN TODAY'S JAPAN
Overview
Table 3.3 shows Japan's current R&D efforts in comparison to other major developed
countries. Clearly, the role of government in funding R&D is smaller in Japan than in
the other countries. One reason for this smaller share of government funds is the small
defense-related expenditures in Japan. However, even when defense-related R&D
expenditures are excluded, the proportion of government funds in total R&D expen-
ditures is the smallest.
The share of government funds in industrial R&D expenditures is a mere 1.2%
in Japan; that is, Japanese companies are financing almost all of their R&D out of their
own funds. This is in contrast to other countries where 11 to 34% of industrial R&D
expenditures are supported by government funds. Again, the gap narrows consider-
ably if defense-related industries are excluded because, without aircraft and missile
industries, the comparable percentage for the United States is reduced to 7.7 (Eads and
Nelson, 1986). Yet, in comparison with Japan's 1.2%, a substantial gap remains
between the two countries.
Table 3.3. R&D Expenditures and Personnel in Five Countries"

Germany United
Japan (Y) United States ($) (DM) France (F) Kingdom ()
Total R&D expenditures (in billions, 10909 142.0 64.82 141.9 10334*
local currency)
Total R&D expenditures (in billion 10909 19596 4757 3068 2359*
yen)
Total R&D expenditures/GNP (%) 2.69 2.73 2.89 2.33 2.19*
Nondefense R&D expenditures/ 2.66 1.95 2.75 1.79 1.77*
GNP (%)
Government funds/total R&D 17.1 46.4 33.2 49.3 36.7*
expenditures (%)
Government funds/nondefense 16.4 25.1 29.9 33.9 21.6*
R&D expenditures (%)
Industry expenditures/total R&D 75.5 72.0 73.0 56.8*** 66.4*
expenditures (%)
Industry funds/industry R&D 98.6 65.7 86.8 73.0*** 71.4*
expenditures (%)
Government funds/industry R&D 1.2 34.3 11.3 22.4*** 16.5*
expenditures (%)
R&D personnel (in thousands) 462 949* 166** 115* 102*
R&D personnel/ 1 0,000 population 37 39* 27** 21* 18*
a
Year 1989 except * 1988, **1987, and ***1983. R&D expenditures and personnel exclude humanities and social sciences in
Japan and the United Kingdom but include them in other countries. Germany refers to the Federal Republic (West) before
unification.
Source: Science and Technology Agency (1991).
Table 3.3 also shows the number of R&D personnel. The number of research per-
sonnel was about 462,000 in 1989, growing three times in the preceding two decades.
Of these 64% of them were working in industry. Although the number of research per-
sonnel is about half that of the United States, per capita Japan has about the same
number as the United States and more than the European countries. One notable fea-
ture of the Japanese research personnel is a relatively larger proportion of engineers
(42% of the total research personnel in 1989) than scientists (16%).21 The same ten-
dency exists for university degrees: the number of Ph.Ds granted was 860 in science
and 1404 in engineering in Japan in 1988. In contrast, these were 7438 and 3236 in
the United States and 2894 and 1020 in Germany, respectively. The same can be said
of the undergraduate students in Japan. In 1988, there were 368,000 students in engi-
neering departments in contrast to 62,000 in science departments. This emphasis on
engineering education started in the Meiji era as shown earlier.22
An active R&D activity results in patents and, more generally, technological
progress. The number of patent application in Japan (by Japanese or foreigners)
increased from 131,000 in 1970 to 341,000 in 1987. This is in marked contrast with
European countries where the number of application decreased in the 1970s to the
early 1980s, and with the United States, where the number increased only gradually
to 134,000 in 1987. Although this international comparison of patent numbers may
be biased owing to the Japanese system under which patents are granted separately
even when inventions behind them were closely related technologically, the fact that
the share of Japanese applications in the U. S. patents is also increasing rapidlyfrom
5.1% in 1970 to 20.3% in 1988suggests that Japan's increased R&D efforts are
indeed bearing fruits (STA, 1990).
Another popular measure of technological progress is the rate of change in total
factor productivity (TFP), which is affected not only by R&D but also by other mis-
cellaneous factors, such as demand shocks, and the "distance" from the best practice
in the world. Table 3.4 shows an impressive rate of TFP increase in Japan, especially
before 1973. The extremely high rate of economic growth in this period is explained
mainly by this TFP increase, which, in turn, appears to be explained by the combi-
nation of successful technology importation and extensive domestic R&D, among
other factors.
Innovation Inside Japanese Firms23
Motivations
Inherently, the choice of the level of R&D is a dynamic problem because the costs and
returns take place at different and over many periods. Thus the firm's willingness to
invest in R&D very much depends on how much the firm is inclined to growth (i.e.,
long-run objectives) as opposed to short-run objectives. It has often been argued that
Japanese firms are more growth oriented than American or European firms. Several
institutional factors are important.
It is first noted that the capital market constraints to the firm are presumably not
as strong in Japan as in the United States or the United Kingdom. Most of the large
Table 3.4. Annual Productivity and Output and Input Growth in Business Sector (in
Percentages)
OECD United United
Average States Japan Germany France Kingdom
1960s to 1973
Output 5.2 3.8 9.7 4.6 6.4 3.2
Factor input 2.4 2.3 3.5 1.8 2.1 1.2
TFP" 2.8 1.5 6.1 2.8 4.3 2.0
Labor productivity 4.1 2.2 8.6 4.9 5.9 3.3
Capital productivity -0.4 0.3 -2.4 -1.1 0.6 -0.7
1973-1979
Output 2.9 2.8 3.8 2.4 3.5 1.1
Factor input 2.2 2.9 2.0 0.6 1.4 0.9
TFP 0.7 -0.1 1.8 1.8 2.1 0.2
Capital productivity -1.4 -0.9 -3.0 -1.1 -1.2 -1.9
1979-1986
Output 2.3 2.2 3.8 1.6 1.5 1.4
Factor input 1.7 2.2 2.1 0.8 0.2 0.3
TFP 0.6 0.0 1.7 0.8 1.3 1.1
Capital productivity -1.3 -1.0 -2.0 -1.3 -1.4 -0.8
Memorandum
1985 capital share 32.2 34.3 22.6 34.8 30.8 32.5
"TFP, total factor productivity.
Source: Steven Englander, A., and Mittelstadt, A. (1988). "Total Factor Productivity: Macroeconomic and Structural Aspects
of the Slowdown." OECD Economic Studies, No. 10, Spring.
shareholders are banks and other firms that are basically friendly to the firm. Often
these shareholdings are reciprocal, that is, the firm also owns their shares.24 As a con-
sequence, it is unlikely that these shareholders interfere with the management of the
firm unless it is in trouble. These shareholders are themselves controlled by the man-
agers sympathetic to the management of this firm.25
Hence, Japanese managers tend to have a fair amount of discretionary power.
They are typically promoted from within among the pool of those having worked with
the company for, say, three decades and gradually climbed up the promotion ladder.
They consider themselves, first and foremost, as representing the employees rather
than the shareholders. Indeed, if they have worked with the employees all the time
since their graduation from the schools, how can they stop feeling closer to the
employees than to the shareholders with whom they have probably been acquainted
only briefly or they have never met?
That managers tend to pursue growth more than the value of the firm has been
put forth by many writers, most notably Marris (1964), for several reasons such as
pecuniary and nonpecuniary gains managers receive from growth more than from
shareholder wealth. Add to this the managers' strong identification with the employees
and the employees' long-term attachment to the firms common in Japan. Whether this
attachment should be called "lifetime" employment is doubtful, because it is not
explicitly written in any labor contract, and de facto dismissal has taken place more
frequently than is usually presumed. Yet it has been and is still regarded as a norm by
both employers and employees, and the employer-employee relationship has been
more stable in Japan than in any other country.
The result of these considerations is an even stronger motivation toward growth,
for the employees are most concerned with the long-term survival and performance
of the company and the prospects of promotion. Since the positions for promotion are
larger the faster the corporate organization expands, the employees are more growth-
oriented than the shareholders, and, with the weak capital market constraints and with
the managers identifying themselves with the employees, it appears reasonable to
assume that Japanese firms are more growth oriented and place a great emphasis on
the acquisition of technological knowledge.
The Managers' Background

There is another difference between Japanese managers and American or British man-
agerstheir background. In Japan, the largest proportion of directors (who, unlike in
the United States, are mostly full-time and insiders) came from production and tech-
nology departments, followed by marketing and export, which together accounted for
50 to 70%, far larger than those with financial and accounting origins, which are
merely 5 to 19% (Kono, 1984, p. 33). More recently, according to the report in Nihon
Keizai Shimbun (May 25,1987), among the 126 presidents who assumed the post dur-
ing the first 6 months of 1987 in the firms listed in the Tokyo Stock Exchange, 36%
had a marketing origin and 28% had a production/R&D origin. Those with a financial
origin again accounted for a mere 11 %. This lack of importance of financial back-
ground is in contrast to the United States or the United Kingdom where financial expe-
rience has been found most helpful in attaining managerial positions.26
The better knowledge and experience of Japanese managers in production/R&D
provide them with a better understanding of the potentials and limitations of R&D
projects, more accurate evaluation of the outcomes from R&D, and more favorable
general attitudes toward R&D. Similarly, their better knowledge and experience in
sales/marketing provide them with a keen understanding of what kinds of products
are in demand in the markets. These familiarities of Japanese managers with techno-
logical seeds and market needs are particularly valuable in technologically rapidly
changing markets.
R&D-Production-Sales Linkages
The Japanese internal labor market system is characterized not only by long-term
attachment but also by a carefully organized training and rotation scheme. It is not
unusual for the firm to provide several months of well-programmed training at the
time of entry of the workers, not only so that they can acquire technical knowledge
and skills but also to familiarize them with diverse activities of the firm, and then mid-
career training every 5 years or so. In addition, rotation of workers from, say R&D to
production in the case of engineers, or one shop to another in the case of production
workers, is more common than in other countries. These practices help workers
acquire a company-wide view and acquire flexibility to changing work environment.
Furthermore, long-term employment naturally leads the workers to develop per-
sonal linkages across departments. If you are working in R&D for many years, you
will have many opportunities to meet and talk with other people in the company, and
opportunities to visit other departments and other plants to discuss problems of
mutual concern. Thus a close human relation is created between the R&D department
and other departments within the firm. Consequently, the R&D staff will be more
familiar with the technological needs arising from production and marketing, and the
non-R&D departments will be more familiar with what is going on in the laboratories.
Consequently, the company R&D departments tend to be more production based and
market oriented. The advantage is that the research undertaken will be more com-
mercially relevant and the introduction of a new product into the production and mar-
keting stages will be faster. On the other hand, truly original and basic research may
not be emphasized.
Two other features of Japanese firms make this R&D-production-sales link even
more effective. One is the infrequent adoption in Japan of multidivisional forms with
substantial discretion within each division. The weaker divisional separation not only
fosters easy rotation across divisions (which probably is why a multidi visional form is
unpopular) but also facilitates a companywide use of technological knowledge. The
other is a close link with suppliers and subcontractors. The stable buyer-supplier rela-
tionship with a constant flow of information is usually observed in Japanese compa-
nies. As a result they tend to share the threat of market competition as well as the need
for innovation. Improvements in the product or production process by the supplier or
subcontractor will be noted and rewarded by the parent company and, if possible, will
be utilized in other firms in the group. In many occasions, cooperative R&D will be
carried out between the parent and the supplier(s). For instance, when an automobile
assembler develops a new car model, it is essential that suitable components are devel-
oped at the same time. Such development will be done by the supplier(s) with close
communication with the assembler's developing team and often with the latter's tech-
nical advice. See Asanuma (1985) and Clark and Fujirnoto (1991) for a detailed study
on the automobile industry, and Imai et al. (1985) for a general discussion.
Introduction of New Process/Product

The introduction of a new process or new product into production and marketing may
be smoother in Japanese firms for several reasons. Most important is the close com-
munication between R&D and production as previously discussed. In particular, in
the course of developing a product or process, the production staff tends to participate
from a rather early stage. This may not be efficient in terms of development itself but,
since the views of the production department are reflected in the development of the
final product/process, the transfer from development to production is quicker, with
little bottleneck or disruption occurring in the process.
It is also not unusual that the engineers in the development team are transferred
to the production department and actively involve themselves in applying the new
product/process into production (Sakakibara and Westney, 1985). Such a transfer is
common in the careers of Japanese engineers because the ability for research tends to
reach a peak in the late thirties (or even earlier) and, afterward, the management as
well as the engineers themselves tend to use their ability and expertise in more or less
administrative and managerial positions in the production and R&D departments.
The transfer, therefore, not only smooths the introduction of a particular new product/
process in which the engineer has been involved, but also increases the technological
knowledge in general in the production department. It also fosters an even tighter per-
sonal link between R&D and production.
Even if the production manager is familiar with the new product/process, its
introduction to production cannot be smoothly done unless the workers are skilled
and flexible. Here again the advantage of Japanese firms in terms of labor management
has to be emphasized. In Japan, under the expectation of lifetime employment, the
management has an ample incentive to provide training to both white-collar or blue-
collar workers. Hence, in-company education, on-the-job training, and rotation are
provided to the workers. Their skills are therefore not only high but also flexible and
wide, which makes adaptation of a new product/process into shop-floors much easier
and faster.
In this adaptation process, it is important for management to have a free hand in
reorganizing the work organization. The prevalence of single and companywide trade
unions in Japan is indispensable for this purpose. For instance, when automobile man-
ufacturers introduced industrial robots, most of the welding jobs were eliminated and
yet none of the Japanese firms discharged the welders for this reason. They were all
retrained and transferred to other shops, such as metalwork and assembling; hence,
no incidence of grievance or industrial action was reported. If, on the contrary, work-
ers in separate jobs are organized into different unions, such transfers would have been
impossible and the introduction of new processes would have caused disputes. The
relation between innovation and labor relations can be indeed quite important, how-
ever distant they may appear. That Japanese firms starting production in the United
Kingdom have all sought a single union representation has to be understood in this
light.
Synthesis
In this section, we have looked into the inside of the so-called black box and came up
with four major factors in Japanese management favorable in making large R&D
efforts, making them efficiently, and applying the outcome swiftly and smoothly into
manufacturing and marketing. They were a bias to growth maximization; familiarity

of management with research, production, and marketing; close R&D-production-
sales links; and the smooth transfer of new processes and products into production.
Behind all these factors, we stressed the importance of the human aspects of manage-
ment. The internal labor practices of Japanese management, particularly, a long-term
worker-company attachment with internal training and internal promotion, have
been conducive to the growth preference of the managers and employees, interde-
partmental personal contacts, and easy adaptability of production processes to new
processes and products. Innovation is predominantly a labor-intensive process. How-
ever powerful a computer may be, it needs a human brain to start with ideas and make
a final evaluation of alternative projects. However labor-saving a new process may be,
it needs experienced workers to install and operate it. Creating the human resources
and organizations that are most suitable for these needs is the key to successful inno-
vation.
These internal aspects, we note, must be accompanied by external factors. An
organization, however capable of being efficient and flexible, need not be so unless
external threats of competition and rapid changes in industrial structure exist. Com-
petition among rival firms has been very intensive in many industries in Japan, even
where the concentration ratio is not particularly low. Entry into growing industries has
been fast and frequent. And shifts in industrial structure have been drastic in the past
decades, with some of the top industries now fading into obscurity. These facts pro-
vided the firms with a strong sense of crisis and a motivation to be innovative. To
maintain the smooth internal labor system, growth and diversification have to be
sought internally and the firm has to acquire the necessary technology internally. Con-
sequently, under the threat of competition and of rapidly changing technology and
industrial structures, the firm's survival is threatened unless it keeps innovating. Com-
petition in industries, itself a product of the Japanese management system (through
the effort to grow internally and diversify, for instance), should never be disregarded
as a key to a nation's technical progress.
THE FUTURE!
What shall we expect of the Japanese innovation system in the future? We suggested
that the weight of government policies will further decline, particularly because the
government is losing most of its control tools through deregulation and liberalization.
The industries are less dependent on the government, and government-business per-
sonal relations are weakening.
There are a few signs suggesting that a change is inevitable in the Japanese man-
agement system. The first is the move by many firms toward increased offshore pro-
duction. How much this move will make the R&D-production linkage difficult is an
important question. The second is the change in financial markets, such as deregula-
tion and an increase in mergers and acquisitions. Again, it is important to know how
much impact this change will make on the way Japanese firms make decisions, and
particularly on their attitudes toward R&D. The third is the changing environment
surrounding labor relations. As the age composition of the Japanese working popu-
lation is rapidly changing toward that of more senior workers, and as technological
change accelerates, it will become more and more difficult for the firms to maintain
lifetime employment, and some of the advantages of the Japanese management sys-
tem may be lost. An increase in offshore production and immigrant workers may fur-
ther add to this tendency.
The fourth is an increasing need for global research. As international technolog-
ical race intensifies, and as globalization in production increases, the firm has to be
alert to technological development in any part of the world, to technological needs
from any production site worldwide, and to international market needs. The need for
global research is therefore increasingly felt, and many companies have started estab-
lishing research institutions overseas or increasing technological collaborations with
foreign firms. Whether the Japanese firm is suitable for this globalization is an inter-
esting question. Sakakibara (1988), for instance, suggests that American managers are
more experienced in employing people with different values, personalities, and edu-
cational backgrounds, and, therefore, have a higher flexibility and the adaptability
needed in international transplantation of technology management. However, the
heterogeneity of scientists in the United States and homogeneity in Japan (the differ-
ence being not absolute but only relative) may suggest that American scientists work-
ing in U. S.-based Japanese firms may adjust more easily than Japanese scientists
working in American firms in Japan. To this extent, American firms may be handi-
capped in obtaining technological knowledge originating from Japan, and the lan-
guage barrier may further aggravate this handicap. In addition, in view of the R&D-
production link in Japanese firms, foreign firms may not be able to carry out research
efficiently in Japan unless they have production facilities as well (Westney and Sak-
akibara, 1986).
Finally, more and more emphasis will be placed on basic research by both private
firms and policy-makers. As Japan has caught up with the best practices in the world
in many fields and become a leader in some of them, technological seeds that can be
imported have become scarce. Even if Japanese firms find technologies they wish to
import, it is now common that seller firms attach various restrictive conditions in their
offer. Consequently, Japanese firms are now keen to create technology of their own,
in particular in the field of basic research. In addition, recent development in high tech-
nology has demonstrated the importance of basic research for commercial success.
Many Japanese firms have thus set up basic research laboratories independently
of existing research facilities that have been biased toward applied research and devel-
opment. Although it is sometimes doubtful how basic their version of basic research
truly is, the shift of emphasis toward the basic end of technological spectrum is appar-
ent. The government is also emphasizing the need for more basic research, in the belief
that, as the status of the Japanese economy in the world rises, its contribution to "inter-
national public goods" has to be increased as well. Basic research that contributes to
the world stock of knowledge is viewed as such an international public good to which
Japan is expected to contribute.
In terms of expenditures on basic research, the increase has been attained far
more by industries than universities. During the 10-year period from 1977 to 1987,
industries increased the expenditures by 15.8% annually and universities by 6.5%. As
a consequence, whereas in 1977 universities undertook 64% of Japan's basic research
and companies undertook 20%, the percentages in 1987 were 50 and 34, respectively
(STA, 1990). Needless to say, universities have been traditionally assumed to be the
prime and most appropriate performers of basic research because industries are
unlikely to undertake research with externality and distant commercial applicability.
One wonders, therefore, whether the increasing dependence on the private effort can
lead to a healthy development of basic research. Insufficient funds in universities is also
a problem. In 1987, R&D expenditure per researcher was 25 million yen in companies
but only 10 million yen in universities (STA, 1990).
All the above discussions suggest that further changes are inevitable and necessary
to the Japanese innovation system. Which way to change can gravely influence future
technological performance, and innovation to create a leader's innovation system out
of a follower's system, however successful it may have been in the past, is very much
in need.
NOTES
We thank Chris Freeman, Richard Nelson, Hugh Patrick, and Nathan Rosenberg for help-
ful comments to earlier versions.
1. That the Japanese made rice-polishing machines while the West made mills reflects the
difference in eating habits. Some authors suggest that this difference caused the Japanese to be
good at the machines with reciprocating motion but behind the West at those with rotary
motion.
2. We will show later how this factory relates to indigenous karakuri technology and to
present Toshiba.
3. For national research institutions, Riken, and Gakushin discussed above, see Hiroshige
(1973) and Kamatani( 1988).
4. The data here are from Kogyo Chosa Shuho, 2(2), 1924, reprinted in Nihon Kagaku
Gijutsushi Taikei, Vol. 3, Tokyo: Daiichi Hoki Shuppan, 1967.
5. See Miyazaki and Ito (1989) for further details of the historical development of the Jap-
anese industry.
6. A survey conducted by Resources Bureau of the Government, quoted in Hiroshige
(1973, p. 115).
7. A survey conducted by the Technology Agency of the Cabinet, quoted in the Agency of
Industrial Science and Technology (1964, p. 125).
8. Take the case of Zero fighters known for their speed and maneuverability in the begin-
ning. Karasawa (1986) shows that during the 5-year war period, their horsepower could be
increased only by 20% whereas Americans and Germans more than doubled the horsepower of
their fighters, thereby surpassing Zero fighters. Also, guns on the fighters were difficult to aim
and suffered from short range and slowness. The weaker technology in communication and
radar was also evident. In the army, too, the trucks and other cars had great trouble, as will be
discussed later. Hence, the technological gap between Japan and America seems to have been
very large by the end of the war, not to mention America's development of atomic bombs.
9. For Japan's technology importation and the role of government policy up to the mid-
1970s, see Peck and Tamura (1976) and Peck and Goto (1981).
10. We will return to the discussion of the postwar industrial policy later.
11. Blumenthal (1976) and Odagiri (1983) found a significantly positive correlation
between the amount of royalty payments for imported technology by industries or firms and
their R&D expenditures.
12. Some authors, such as Saxonhouse (1985), emphasized the information-disseminating
role of cooperative research. It is difficult to estimate how important this role was in RAs. It has
been reported that in the much publicized case of the Very-Large-Scale Integrated Circuits
(VLSI) Research Association, it took considerable talent and effort (and drinks!) by the director
to have the scientists from different companies mingle (Sakakibara, 1981). Since this RA was
quite exceptional in having a common laboratory and a dedicated director from a third party
(namely, a government research institution and not companies), we imagine that communica-
tion among scientists from different companies must have been infrequent in most other RAs.
13. Two arguments may be made. First, research subjects of the RAs are in principle closer
to the basic end, need longer lead time, and have a smaller chance of success than the research
subjects of individual firms. Second, when most of the actual and potential rivals are within an
RA, there may be little need for protective patents, that is, to apply for patents only to preempt
rivals.
14. The discussion in this section is based on miscellaneous Japanese sources including
company histories published by respective companies. References will not be given individually.
15. Bureau of Statistics, Management and Coordination Agency, Report on the Survey of
Research and Development.
16. This fact supports the discussion in footnote 7.
17. See Hirschmeier and Yui (1975).
18. When WE decided to enter the wire business in Japan, NEC recommended Sumitomo,
one of the so-called four big zaibatsus before World War II, as the partner because Sumitomo
had a big copper business. A new company, Sumitomo Electric Wire (now Sumitomo Electric
Industries), was established, jointly owned by WE, Sumitomo, and NEC. In return, Sumitomo
acquired a share of NEC.
19. For other views, see Johnson (1982) and Okimoto (1989).
20. This finding agrees with the declining government support on R&D shown in Table
3.2.
21. The number of researchers (or scientists and engineers) is difficult to compare inter-
nationally because the definition is different and the necessary data to adjust the Japanese figures
to full-time equivalents are lacking. See National Science Foundation (1988).
22. The data on education are from Ministry of Education, Gakko Kihon Chosa Houko-
kusho.
23. Fora more detailed discussion of the topic in this section, see Odagiri (1992).
24. As suggested at the end of the previous section, it is incorrect to attribute this ownership
pattern to post-zaibatsu business groups. The discussion in the text applies to most firms whether
they belong to such groups or not.
25. This reciprocal shareholding is one reason why hostile takeovers are infrequent in
Japan. The labor practices in Japan are another reason. See Odagiri (1992, Chapter 5).
26. See Fidler (1981) and Kohn/Ferry (1981) for the United Kingdom and Browne and
Motamedi (1977) for the United States.
REFERENCES
Agency of Industrial Science and Technology, the Blumenthal, T. (1976). "Japan's Technological
Ministry of International Trade and Industry. Strategy." Journal of Development Economics 3:
(1964). Gijutsu Kakushin to Nihon no Kogyo: 245-55.
Kogyo Gijutsu-in 15 Nen no Ayumi. [Innovation Browne, W. G., and Motamedi, K. K. (1977).
and Industries in Japan: The 15 Year History of "Transition at the Top." California Management
the Agency of Industrial Science and Tech- Review 20(2): 67-73.
nology]. Tokyo: Nikkan Kogyo Shimbunsha. Clark K.. B., and Fujimoto, T. (1991). Product Devel-
Asanuma, B. (1985). "The Contractual Framework opment Performance:Strategy, Organization, and
for Parts Supply in the Japanese Automotive Management in the World Auto Industry. Boston:
Industry." Journal of Japanese Studies Summer, Harvard Business School Press.
54-78. Eads, G. C., and Nelson, R. R. (1986). "Japanese
High Technology Policy: What Lessons for the Lynn, L. H. (1982). How Japan Innovates. Boulder:
United States?" In Hugh Patrick (ed.), Japan's Westview Press.
High Technology Industries, 243-69. Seattle: Marris, R. L. (1964). An Economic Theory of'Man-
University of Washington Press. agerial' Capitalism. London: Macmillan.
Fidler, J. (1981). The British Business Elite: Its Atti- Minami, R. (1981). Nippon no Keizai Hatten [Eco-
tudes to Class, Status and Power. London: Rou- nomic Development in Japan]. Tokyo: Toyo Kei-
tledge and Kegan Paul. zai Shinposha.
Fujishiro, N. (1988). "Computer Sangyo ni Okeru Miyazaki, M., and Ito, O. (1989). "Senji Sengo no
Kyodo-Kenkyu no Yakuwari." [The Role of Sangyo to Kigyo" [Industries and Firms during
Joint R&D in the Computer Industry]. Unpub- and after the War]. In T. Nakamura (ed.), Nihon
lished master's thesis, University of Tsukuba. Keizaishi 7: Keikakuka to Minshuka, 165-235.
Goto, A. (1982). "Business Groups in a Market [The History of Japanese Economy, Vol. 7: Plan-
Economy." European Economic Review 19: 53- ning and Democratization], Tokyo: Iwanami
70. Shoten.
Goto, A. (1988) "Japan: A Sunset Industry." In Mer- National Science Foundation (1988). The Science
ton Peck (ed.), The World Aluminium Industry in and Technology Resources of Japan: A Compari-
a Changing Energy Era. Washington, D. C.: son with the United States. Washington, D. C.:
Resources for the Future. U. S. Government Printing Office.
Goto, A., and Wakasugi, R. (1988). "Technology Odagiri, H. (1983). "R&D Expenditures, Royalty
Policy." In R. Komiya, M. Okuno, and K. Suzu- Payments, and Sales Growth in Japanese Manu-
mura (eds.), Industrial Policy of Japan, 183-204. facturing Corporations." Journal of Industrial
New York: Academic Press. Economics 32( 1): 61 -71.
Hiroshige, T. (1973). Kagaku no Shakaishi [The Odagiri, H. (1992). Growth through Competition,
Social History of Science]. Tokyo: Chuo Koran Competition through Growth: Strategic Manage-
Sha. ment and the Economy in Japan. Oxford: Oxford
Hirschmeier, J., and Yui, T. (1975). The Develop- University Press.
ment of Japanese Business, 1600-1973. London: Okimoto, D. I. (1989). Between MITI and the Mar-
George Allen & Unwin. ket: Japanese Industrial Policy for High Technol-
lida, K. (1979). Nippon Tekkou Gijutsu Shi [The ogy. Stanford: Stanford LJniversity Press.
History of Iron and Steel Technology in Japan]. Patent Office (1955). TokkyoSeido 70NenShi[The
Tokyo: Toyo Keizai Shinposha. 70 Year History of the Patent System]. Tokyo:
Imai, K., Nonaka, I., and Takeuchi, H. (1985). Hatsumei Kyokai.
"Managing the New Product Development Pro- Peck, M. J., and Goto, A. (1981). "Technology and
cess: How Japanese Companies Learn and Economic Growth: The Case of Japan." Research
Unlearn." In K. B. Clark, R. H. Hayes, and C. Policy 10(3): 222-43.
Lorenz (eds.), The Uneasy Alliance, 337-75. Bos- Peck, M. J., and Tamura, S. (1976). "Technology."
ton: Harvard Business School Press. In H. Patrick and H. Rosovsky (eds.), Asia's New
Johnson, C. (1982). MITI and the Japanese Miracle. Giant, 525-85. Washington, D. C.: The Brook-
Stanford: Stanford University Press. ings Institution.
Kaigo, M. (ed.) (1971). Nihon Kindai Kyouiku Jiten Sakakibara, K. (1981). "Soshiki to Innovation: Jirei
[Encyclopedia on Modern Education in Japan]. Kenkyu, Cho LSI Gijutsu Kenkyu Kumiai"
Tokyo: Heibon Sha. [Organization and Innovation: A Case Study of
Kamatani, C. (1988) Gijutsu Taikoku 100 Nen no VLSI Technology Association]. The Hitotsubashi
Kei: Nihon no Kindaika to Kokuritsu Kenkyuu Review 86(2): 160-75.
Kikan [The 100 Year Strategy to an Innovative Sakakibara, K. (1988). "Managing Global Innova-
Country: Japan's Modernization and the tion: The Challenge of International Transplan-
National Research Institutions]. Tokyo: Heibon- tation of Technology Management." Working
sha. Paper No. 8803, Hitotsubashi University.
Karasawa, E. (1986). "Koku Gijutsu no Hizumi." Sakakibara, K., and Westney, D. E. (1985). "Com-
[Deficiencies in Aircraft Technology]. In K. Has- parative Study of the Training, Careers, and Orga-
egawa (ed.), Nihon Kindai to Senso, Vol. 6, 193- nization of Engineers in the Computer Industry in
224. [Modern Japan and Wars]. Kyoto: PHP. the United States and Japan." Hitotsubashi Jour-
Kohn/Ferry International in Conjunction with the nal of Commerce & Management 20( 1): 1-20.
London Business School (1981). British Corpo- Sawai, M. (1990). "Kikai Kogyo." [Machinery
rate LeadersA Profile. London: Kohn/Ferry Industry]. In S. Nishikawa and T. Abe (eds.),
International. Nihon Keizaishi 4: Sangyo-ka no Jidai, 213-53.
Komiya, R., Okuno, M., and Suzumura, K. (eds.) Tokyo: Iwanami Shoten.
(1988). Industrial Policy of Japan. New York: Saxonhouse, G. (1985). "Japanese Cooperative
Academic Press. R&D Ventures: A Market Evaluation." Discus-
Kono, T. (1984). Strategy and Structure of Japanese sion Paper No. 156, Department of Economics,
Enterprises. London: Macmillan. University of Michigan.
Science and Technology Agency (STA) (1990, Technological Innovation]. In Y. Okamoto and
1991). Kagaku Gijutsu Hakusho, [White Paper on T. Wakasugi (eds.), Gijutsu Kakushin to Kigyo
Science and Technology]. Tokyo: Printing Kodo, 193-217. Tokyo: Tokyo University Press.
Bureau of the Ministry of Finance. Westney, D. E., and Sakakibara, K. (1986). "The
Umihara, T. (1988). Kinsei no Gakko to Kyouiku Role of Japan-Based R&D in Global Technology
[Schools and Education in the Tokugawa Japan]. Strategy." In M. Horwitch (ed.), Technology in the
Kyoto: Shibunkaku. Modem Corporation: A Strategic Perspective,
Wakasugi, R., and Goto, A. (1985). "Kyodo Kenkyu 217-32. Oxford: Pergamon Press.
Kaihatsu to Gijutsu Kakushin." [Joint R&D and
4
The National System for Technical
Innovation in Germany
OTTO KECK1
Germany is a special case, for several reasons. One is its recent political history. After
World War II the country was divided into two states with opposed political-economic
systems, the Federal Republic of Germany in the West (with a capitalist economy and
a pluralist democracy) and the German Democratic Republic in the East (with a cen-
trally planned economy and an authoritarian socialist political system). The contrast
between a strong economy in the West and an acerbating economic crisis in the East
was one of the reasons for the revolution that in 1989 and early 1990 brought down
the socialist system in the East. By joining the Federal Republic in October 1990, the
Eastern part with a population of 17 million adopted the political and economic insti-
tutions of the Western part, including those relating to technology and science.
A second factor that makes Germany a special case is the export performance of
its economy. West Germany, with a population of 61 million, exported in the year
1988 goods of a total value of $323 billion, about the same as the United States ($320
billion), and more than Japan ($265 billion). On a per capita base, this is 4.0 times
more that the United States and 2.4 more than Japan.2
Institutional forms that today are taken for granted in most national systems for
technical innovation, such as the research-oriented university that combines its edu-
cational function with the advancement of scientific knowledge, and the science-based
firm with an in-house R&D laboratory separated from production, were pioneered as
social innovations in nineteenth-century Germany. Soon they were adopted by other
countries. Also in technical education Germany provided an impulse for emulation
in some other countries. Hence a historical description of the origins of the German
system for technical innovation may not only help in understanding the present sys-
tem in Germany, but also shed some light on the cross-fertilization among different
national systems of technical innovation that occurred by emulating institutional
forms across national boundaries while adapting them to specific national environ-
ments and sometimes improving them. This is the third reason why Germany may be
regarded as a special case.
The performance of the West Germany economy is sometimes explained by ref-
erence to a supposed national character, in particular to the Germans' reputation for
115
being hard-working people. The gap in performance between the Western and the
Eastern part of the country does not contradict such an explanation, since relative to
other socialist states the economy in the German Democratic Republic did have a high
standing. However, the explanation is refuted by data on annual working time. Indus-
trial workers in Germany have a contractual working time of 1615 hours per year,
compared to 2201 hours in Japan, 1904 hours in the United States, and 1775 hours
in France.3 The technical capability of German industry, built on the hard work of
former generations, now enables industrial workers to work fewer hours per year than
in other advanced countries.
This chapter therefore looks to institutional structures rather than national char-
acter to explain economic performance. Among the many institutions that contribute
to economic performance it focuses on those that relate to the technological capability
of industry. The first section describes the evolution of the innovation system up to
the beginning of the twentieth century, the second carries the historical account to the
present time, and the third analyzes the present system and discusses some of its chal-
lenges.
HISTORICAL ORIGINS IN THE NINETEENTH CENTURY
Among European states, Germany was a latecomer, in both political and economic
terms. Its development had experienced a setback in the seventeenth century through
the devastations of the Thirty Years' War, which reduced the population by about a
third. The slow pace of development was reflected in the fragmentation of the country:
as of 1789 there existed 314 independent territories and more than 1400 imperial
knighthoods. Many of the territories had their own laws, currency, weights and mea-
sures, taxes, and custom tolls. The way to unification was cumbersome. The Napo-
leonic wars and the Congress of Vienna reduced the number of territories to 39 by the
year 1815. In 1834 Prussia with some other German states formed a customs union to
which most other German states acceded until 1867. Political union was eventually
achieved in 1871. The German empire was a federal state, in which central govern-
ment was responsible only for some state functions, mainly foreign policy and the mil-
itary. The education system was under the jurisdiction of the federal states.4
Throughout the nineteenth century there was a conflict between the advocates of
political reform and industrialization (coming mainly from the German bourgeoisie)
and those striving to preserve the autocratic political order. One of the leading protag-
onists of industrial development was Friedrich List, from 1817 to 1820 professor of
political economy at the University of Tubingen. When his views made him clash with
the political authorities, he took refuge in exile in the United States. In opposition to
the classical political economy of Adam Smith and David Ricardo he advocated an
evolutionary perspective focusing on the development of the productive forces. He
proposed a customs union comprising central Europe that by means of protective tolls
would enable domestic industry to catch up through import substitution, using
imported machinery.5
In the first third of the nineteenth century Germany turned to foreign countries,
mainly to Britain, but also to Belgium, for new machinery and for skilled workers to
bring advanced technology to its industries. New types of machinery in the cotton,
TECHNICAL INNOVATION IN GERMANY 117
woollen, and linen industries, the first steam engines, and the first locomotives were
imported from Britain.
British and Belgian artisans were instrumental in transferring technical know-
how to Germany in the machine-building and iron and steel industries. To protect the
technological lead of its industries, Britain prohibited by law the emigration of skilled
workers until 1824 and, for many of its advanced industries, up to 1843 the export of
machinery, including models and drawings, as well as tools and utensils. Yet these laws
were difficult to enforce, and in a number of cases the British government granted
export licenses. German governments often provided financial support for the pur-
chase of foreign machines, which were sometimes used as demonstration objects.
Technical knowledge was also acquired through German visitors, often with encour-
agement and financial support by the government, and sometimes by industrial espi-
onage.6
Given the backward state of the polity and the economy, the government had a
key role in the country's development. The customs union, political unification, abol-
ishment of traditional restrictions on the freedom to engage in business (Gewerbef-
reiheit), construction of roads and canals, encouragement of railway construction, and
creation of a capable civil service were all important factors. And so was the mainly
government-financed system for education and research in technology, science, and
business.
The Education System
In the early nineteenth century, France was the center of science in the world.7 Many
German scientists, for example the chemist Justus Liebig, went to Paris to learn the
state of the art. French institutions of higher education, in particular the Ecole Poly-
technique, served as stimulus and model (though not always correctly represented).
But unlike in France the university became the institutional focus of scientific research
in Germany. This was not a preordained development. In the eighteenth century,
academies of science had been founded in several German states, and scientific
research was primarily their task. Many universities were in a poor state, and some
voices favored the idea of abolishing them altogether.
However, some of the states managed to break away from this condition by
reforming their universities or by establishing new ones with a reformed curriculum.8
Prominent among these were the universities of Halle and Gottingen. The latter was
founded in 1742 by the Kingdom of Hanover on the initiative of a noble civil servant,
Gerlach Adolf von Miinchhausen. Aiming to attract students from abroad (meaning
from other German states) he designed a new curriculum and established the practice
of selecting professors on the basis of their literary reputation derived from their pub-
lications. Gottingen soon became a center of scholarship.
Although the reforms introduced by Gottingen and other universities were
important for the further development of the German university system, the origins
of the orientation toward research were more widespread. In a study on the emerging
chemical community in Germany, Hufbauer counted 11 laboratories in the year 1780,
eight of which were located at medical departments and two at mining schools.
Although these laboratories were small and poorly equipped by later standards, they
are evidence that empirical scientific research had a hold at the universities.9 In late
eighteenth century some apothecaries expanded their pharmacies into private insti-
tutes that trained pharmacists, manufactured such drugs and chemicals as were tra-
ditionally custom-made by the pharmacies, and also engaged in laboratory research.
Some of these institutes reached such a high level that their courses were certified by
the government to be equivalent to university courses.10
Up to the present time the standard rhetoric in Germany dates the origin of the
modern university to the founding of the Berlin University in 1809 /1810. What made
Berlin special was the association of the new university with reform ideas coming from
German idealist philosophy. The new ideal of the university as proposed by Wilhelm
von Humboldt, Johann Gottlieb Fichte, Friedrich Schleiermacher, and others was
predicated on the assumption that the formation of an individual personality was
more important for human beings than the acquisition of useful knowledge and skills.
For idealist philosophy, becoming a personality was the highest goal in individual life,
the uppermost realization of human destiny that sets humans apart from animal life.
This ideal survived the fall of idealist philosophy and provided the personal orientation
(and often the personal ideology) for the more traditional groups of German intellec-
tuals throughout the nineteenth century.'' According to its idealist conception the uni-
versity was to be an autonomous community of teachers and students, where those by
devoting themselves to science would develop their individual personalities. Devotion
to science implied an orientation toward research, not only for the professors but also
for the students. Teaching at its best would introduce the student to doing creative
research.
However, the influence of the idealist conception of the university is generally
overestimated.12 First, the key proponents of the idealistic ideal had only a short reign
in the government of the university. Wilhelm von Humboldt left his office as director
of religious affairs and public education in the Prussian Ministry of the Interior 4
months before the university was formally opened, and Fichte stepped down as the
first rector of the university after about a year. Institutional details were worked out by
civil servants in the upper echelons of the Prussian bureaucracy, and they moderated
idealism by utilitarian considerations of what type of education the Prussian state
deemed desirable for its future civil servants, lawyers, doctors, and high-school
teachers.
Second, the idealist concept of science focused on such elevating fields as philos-
ophy, mathematics, and the humanities. Laboratory research was given low standing,
and empirical science had to fight for emancipation from the domination of idealistic
natural philosophy.13
Third, during the upheavals of the Napoleonic wars more than a dozen univer-
sities vanished, mainly small and poor ones. Governance of the universities by the
state bureaucracy improved as the territorial reordering increased the average size of
individual states and as general administrative reforms were introduced after the
Napoleonic wars.
By the middle of the nineteenth century the research orientation was firmly estab-
lished at German universities. It was supported by an institutional base, comprising
institutes with laboratories for the natural sciences, and specialized libraries (called
"Seminars") for the humanities.14 At Berlin, for example, funds for this institutional
base doubled every decade between 1830 and 1870. University research in Germany
rose to a high level, and in some fields, such as medicine, chemistry and physics
ascended to world leadership.'5
Student numbers did not grow between 1830 and 1870 (see Fig. 4.1), but there-
after they increased rapidly: from 14,000 in 1870 to nearly 60,000 in 1914. Govern-
ment funds for the universities increased even faster than the number of students: from
1860 to 1910 by a factor of about five in real terms.16 This expansion fostered special-
ization, and many universities then created separate departments for natural science.17
Contrary to much wishful rhetoric stressing their autonomy, the rise of the Ger-
man universities took place under close supervision by state officials.'8 Most promi-
nent among these was Friedrich Althoff, a senior official in the Prussian Ministry of
Culture from 1882 to 1907, who pursued, with great zeal and political skill, a policy
of expanding the Prussian universities and raising their standard still further. He filled
professorial chairs with the best people in their fields and provided them with adequate
facilities.19 As he relied on a network of private confidants to make his selections, often
cajoled departments into accepting his canditates, and sometimes appointed them
against their will; his style produced controversy. Nevertheless, even those objecting
to his ways acknowledged his merit.
Whereas in the area of science the German university system in the nineteenth
century accomplished very much, it did nothing for engineering. In the mind of pro-
fessors and administrators, engineering lacked the dignity of science, and for this rea-
son it was not admitted to the university. More accurately, it was driven out, for in the
late eighteenth and early nineteenth century, university education for civil servants
was rather broad and, in addition to law, political economy, and political science, corn-
Figure 4.1. Students in higher education, 1830 to 1940. Source: Hartmut Titze, Das Hochs-
chulstudium in Preussen und Deutschland 1820-1944, p. 26. Gottingen: Vandenhoek und
Ruprecht, 1987.
prised the knowledge of techniques for agriculture, mining, and manufacturing. The
German political economists of the eighteenth century, the so-called "cameralists,"
recognized the importance of these techniques for the economy and established "tech-
nology" (in German "Technologic"), in the precise meaning of this word in the sense
of knowledge of techniques, as a subject for university teaching.20
Some engineering schools had been founded in the eighteenth century to train
civil servants for government service: as administrators in the mining industry (which
at that time consisted maily of state enterprises), as civil engineers and architects, or
as military engineers and artillery officers. The graduates from these schools were with
some exceptions reluctant to go into private industry, since this conferred a much
lower social status than public service.21
In the early 1820s Prussia took the lead in establishing a system of schools to train
technicians for private industry. It soon comprised about 20 vocational schools in the
provinces providing 1-year full-time courses for craftsmen and factory shopmasters
and, above the provincial schools, the Technical Institute (Gewerbeinstitut) in Berlin
offering a 2-year course for technicians with the objective of enabling them to set up
and manage factories.22 Most major German states quickly followed by establishing
polytechnical schools.23
The vocational schools expanded by offering one or more years of preparatory
courses and most were gradually transformed into secondary schools for general edu-
cation that differed from the traditional secondary school in Germany, the "Gymna-
sium," only by not teaching Latin and Greek and by stressing mathematics and nat-
ural sciences. In the 1870s their students were admitted to the university.
The polytechnical schools gradually improved their teaching and raised entrance
requirements. They gained in social recognition, when their graduates were accepted
for public service. To fashion themselves as equals of the universities, they stressed the
use of scientific method and mathematics. The organizational politics of the poly-
technical schools became part of the effort of the German engineering profession, orga-
nized since 1856 in the Union of German Engineers (Verein Deutscher Ingenieure),
to obtain a social recognition equal to the university-based professions.24
In the 1870s the polytechnical schools were elevated to higher status. They were
now called Technische Hochschulen, required similar entrance qualifications as the
universities, and distinguished their graduates from lesser kinds of engineers by the
special designation of Diplom-Ingenieur. A further step toward equal status with
the universities was achieved in 1899, when after long political battles against the bitter
resistance of the universities the King of Prussia decided in person to give the Techn-
ische Hochschulen in Prussia the right to grant doctoral degrees.25 The other states
soon followed.
The fact that the Technische Hochschulen made their way by emulating the uni-
versities had some negative repercussions. One was an overemphasis on theory, which
led their education in a direction that may not quite have been in line with what indus-
try needed.26 Only in the last decade of the century, after sometimes bitter controver-
sies, did experimental laboratories become generally established in the departments
for machine construction. Here the German Technische Hochschulen in fact learned
from engineering schools in the United States. The World Exposition at Chicago in
1893 demonstrated the very high standard of the American machine tool industry, and
Alois Riedler, one of the leading German engineering professors, saw one of the rea-
sons for it in the engineering education at American universities, in particular their
use of laboratories.27
As polytechnical schools were upgraded to university level and vocational schools
were transformed into secondary schools for general education, a gap opened at the
middle level of technical education. There is a scattered history of specialized technical
and commercial schools in the nineteenth century, but these could not fill the gap.
Toward the end of the century the states created new technical schools offering courses
up to 2 years. The lower technical middle schools were open to everybody who had
gone through compulsory school training (normally age 6 to 14) and had 3 to 6 years
of practical experience, while the higher technical middle schools normally required 4
years of elementary and 6 years of secondary school. These new schools were first
established in mechanical engineering and textiles, and later also in other fields such
as electrical engineering. However, the middle-level schools financed by the govern-
ment were also tempted to direct their education more to careers in government than
in private industry, they were slow to take up new developments in electrical technol-
ogy and automobiles, and some of the smaller states could not affort them. For these
reasons there existed many private technical schools, which before World War I had
more students than the state schools.28
The basic level of technical training was provided by the traditional apprentice-
ship. The old craft guilds were abolished in the first half of the century, but the appren-
ticeship system, though beset with weaknesses and misuse, lived on with some
reforms.29 Although some of the old crafts vanished and the relative importance of the
craft sector decreased in this period, the craft sector as a whole still grew and provided
a pool of skilled manpower on which firms in the new industries could draw.30 In the
1860s some firms in the machine construction industry began their own apprentice-
ship programs; in the 1880s firms in the electrical and optical industries did the same.
Toward the end of the nineteenth century the apprenticeship system was reor-
ganized. Chambers of trade were charged with examination. Part-time schools were
established in 1897 for further education of all those who had finished the normal
school (Volksschule), supplementing the practical training in firms and craftshops. As
these schools initially continued the general education of the Volksschule, some firms
created their own schools, more oriented toward specialized vocational training, and
had them approved by the government.31
Around 1900 a number of business schools were founded at the higher education
level, mainly on initiatives of individuals from commerce and industry, chambers of
commerce, or city governments. They later developed into university-level institu-
tions, and some were expanded into full universities. These schools played an impor-
tant role in developing business economics into a specialized subject, later taught at
many universities and Technische Hochschulen.32 Education at these schools became
a routine entry into business management. Also for commercial education middle-
level schools were established, the Handelsschulen and Hohere Handelsschulen. At the
lowest level, the apprenticeship system trained people for commerce; so a system of
several levels emerged, similar to the one for technical education.
Today it is taken for granted in most industrialized countries that the basis of all
specialized training is a general education for everybody between the age of 6 and 13
or 14. Germany established general education in the eighteenth century, that is before
industrialization, whereas in Britain it came only toward the end of the nineteenth
century, long after industrialization. By many indicators of literacy, Germany was
among the top group of European nations in the first half of the nineteenth century.
In terms of primary school enrollment rate per capita, the United States and Germany
were until about 1880 far above France or Britain.13
By the beginning of the twentieth century Germany had established a sophisti-
cated system for education in scientific, technical, and commercial matters, reaching
from elementary school to the doctoral level.34 There were close connections between
the different levels in most areas of specialization, as the teachers for the schools at a
given level were normally educated at one of the higher levels. There was also a flow
of knowledge between universities and Technische Hochschulen, as many areas of sci-
ence such as chemistry were pursued in both, though usually with a greater emphasis
on applied science in the Technische Hochschulen. And there were links between the
education system and industrial firms, not only through the supply of trained per-
sonnel, but also through consultancy by professors in engineering and in areas of
applied science.
The relative importance of the various schools of technical and scientific educa-
tion is indicated by the data on student enrollments for Prussia given in Table 4.1.
They show that the universities dominated, and that student numbers at the middle
Table 4.1 Student Enrollments in Prussia in the Years 1891

and 1911
1891 1911
Technical schools at the middle level:
Lower-middle
Textiles 44 200
Machinery 574 1653
Upper-middle
Textiles 344 800
Machinery 54 1107
Total middle-level 1016 3760
Schools in higher education
Universities of 12826 27501
Theology
Law
Medicine
Mathematics and natural science
Economics, agriculture, forestry
Technische Hochschulen 1910 4064
Business schools 1277
Mining academies 245 216
Veterinary medicine 665 635
Agriculture 544 890
Forestry 120 139
Theology 370 626
Total higher education: 16680 35348
Sources: P. Lundgreen, "Educational Expansion and Economic Growth in Nine-
teenth-Century German."In L. Stone(ed.), Schooling and Society, pp. 20-66. Bal-
timore: Johns Hopkins University Press, 1976 (for lower and upper middle level);
Hartmut Titze, Das Hochschulstudium in Preuften und Deutschland 1820-1944,
pp. 37-38. Gottingen: Vandenhoek and Ruprecht 1987 (for higher education).
level of technical education were of the same order of magnitude as students at Techn-
ische Hochschulen.
What set this system apart from that of other countries was not only the relatively
high standard of research at universities and Technische Hochschulen, but also its
large size. From 1820 to 1850 Prussia trained, relative to the size of its population,
more technicians than France.35 Measured as percentage of the 20- to 24-year-old age
group, student enrollment in higher education was around the year 1880 about the
same as in Sweden, but considerably larger than in France or Italy. In mathematics
and natural science Germany then educated about two times more university students
than did Italy and eight times more than France. In institutes of higher education other
than universities, which comprise mainly technical schools, Germany trained twice as
many students as Italy or France.36 During the first decade of the twentieth century
about 30,000 engineers graduated from colleges and universities in Germany com-
pared to about 21,000 in the United States. Relative to the size of the population this
means two times more in Germany than in the United States.37 In 1913 there were
about 10 times more engineering students in Germany than in England and Wales.38
The German universities and, by the end of the nineteenth century, also the
Technische Hochschulen attracted many foreign students and served as stimulus for
reform or for emulation in other countries, among them Britain, France, and the
United States.39 Sometimes they were idealized, in particular by individuals trying to
promote reforms in their own countries.
Despite its strengths, the German educational system did have some weak
points.40 There were enormous tensions between universities and Technische Hochs-
chulen and, more generally, between the neohumanistic ideals of the German bour-
goisie and the worlds of industry and technology. Their intellectual orientation made
university professors often averse to exploiting new ideas for commercial purposes.41
Neither the industrialization process at large nor the system for technical and scientific
education in particular was the result of a unified ideology.42
Specialized Research Organizations

In addition to the universities, the Technische Hochschulen, and the academies of sci-
ence, central government and the federal states financed at the beginning of the twen-
tieth century some 40 to 50 research institutes for specialized research in applied areas
such as weather and atmosphere, geography and geology, health, shipbuilding,
hydroengineering, biology, agriculture, fishery, and forestry.43 Some of them had mil-
itary purposes; most were oriented toward public tasks such as public health or safety
regulation and some toward supporting technical innovation in the business sector.
Among the latter, a major new departure was the Imperial Institute of Physics
and Technology (Physikalisch-technische Reichsanstalt, now the Physikalisch-techn-
ische Bundesanstalt), founded in 1887 mainly with finance from central government
for work on standards and measures, for the development of precision instruments,
and for basic research in this area. The initiative came from scientists and industrial-
ists, among them Werner von Siemens, founder and director of the firm carrying this
name, who made a donation toward the cost of the institute. Its task was defined as
"physical investigations and measurements which primarily aim at solving scientific
problems of great impact in theoretical or technical respect and require larger means
in terms of instruments, materials and working time of observers and calculators than
can be provided by private people or educational institutions."44 By 1913 its total staff
numbered 139, of which 50 were academically trained scientists and engineers.45 The
Institute served as an example for similar institutions in other countries, for example,
the National Physical Laboratory in England (founded 1900), the National Bureau of
Standards in the United States (1901), and the Institute of Physical and Chemical
Research in Japan (1917). 46
When the chemical industry proposed a "Chemical-Technical Imperial Insti-
tute" similar to the Physical-Technical Imperial Institute, this project became part of
Friedrich AlthofFs plan to set up a number of special research institutes outside the
university system so as to complement the universities' research capabilities for basic
and applied research.47 American precedents such as the Rockefeller foundation had
fostered in Germany the idea to tap industry as a source for research funds. About 10
million Mark were brought together until 1911, when the Kaiser-Wilhelm-Society was
founded.48 A second source of income was the membership fees of industrial firms that
joined the Society. The state contributed real estate, the salaries of some of the institute
directors, and in some cases further financial support. To 1914 five institutes were
established: for chemistry, physical chemistry, coal research, biology, and experimen-
tal therapy. The first three, which did applied research, were nearly totally financed by
industry.
Some smaller research institutes financed jointly by government and industry
precede the Kaiser-Wilhelm-Society. The first may have been an experimental station
founded in 1874 by the Association of Spirit Manufactureres (Verein der Spiritusfa-
brikanten). When in the 1880s it was associated with the Agricultural College (Land-
wirtschaftliche Hochschule) at Berlin, government provided the ground and the build-
ing, while the industry association paid for the equipment and operating costs.49 At the
university of Gottingen the mathematician Felix Klein persuaded industrialists to sup-
port several new institutes for applied research. One of them was an experimental sta-
tion for motor aviation, which under Ludwig Prandtl's direction soon held a leading
position in aerodynamics.50 In biomedical research, private industry contributed funds
to some institutes; for example, that of Emil Behring at Marburg and that of Paul Ehr-
lich at Frankfurt am Main.51
At the turn of the century dozens of foundations for scientific research were
founded, some of which became the backbone of new research institutes, often affili-
ated with universities of Technische Hochschulen. Also some technical associations,
as for example the Verein Deutscher Ingenieure, provided funds for research proj-
ects.
Government Finance
In the five decades between 1860 and 1913 government funds for higher education and
scientific research increased in real terms by a factor of about nine (see Table 4.2).
Relative to gross national product, they more than doubled between 1860 and 1880,
and thereafter they grew at about the same rate as the gross national product. As can
be seen from Table 4.2, the federal states provided most of the funds for higher edu-
cation and science. However, their share fell from 100% in 1860 (when there was no
central government) to less than 80% in 1900.
Table 4.2. Public Expenditure for Science in Germany and the Federal Republic of Germany
Total Public Expenditure for Science" Contribution by
At Constant
At Current Prices As Percent of
Prices (Million Total As Percent of Central Federal
(Million Mark as of Government Net Social Government States
Year Mark) 1913} Expenditure Product (As Percent of Total)
German Empire
1860 6.0 10.7 1.0 0.06 100.0
1870 10.5 14.7 1.1 0.08 2.2 97.8
1880 27.3 33.6 1.5 0.16 10.1 89.9
1890 32.7 37.4 1.0 0.14 18.0 82.0
1900 53.2 59.9 1.1 0.16 21.2 78.8
1910 91.2 94.8 1.1 0.20 22.9 77.1
1913 101.9 101.9 1.2 0.19 21.9 78.1
1925 282.9 206.8 2.2 0.42 39.2 60.8
1930 359.6 241.3 1.9 0.55 29.5 70.5
1938 513.4 446.6 1.6 0.52 46.8 53.2
Federal Republic
1955 1,208 494 2.8 0.75 15.2 84.8
1960 2,352 844 4.0 0.89 26.7 73.3
1970 14,205 4,943 8.4 2.34 36.3 63.7
1980 30,249 5,038 7.1 2.31 38.2 61.8
1988 42,759 5,842 7.8 2.30 35.5 64.5
"In addition to research and development public expenditure for science includes expenditure for teaching and for scientific
and technical services.
Sources: Data on the German empire from F. Pfetsch, Datenhandbuch zur Wissenschaftsentwicklung. Cologne: Zentrum fur
historische Sozialforschung, 1982; F. Pfetsch, "Staatliche Wissenschaftsforderung 1870-1980." In R. vom Bruch and R. A.
Miiller (eds.), Formen aufferstaatlicher Wissenschaftsforderung, pp. 113-138. Stuttgart: Steiner, 1990. Data on the Federal
Republic from Bundesbericht Forschung 1984 (Bundestagsdrucksache 10/1543), 1984; Bundesminister fur Forschung und
Technologic, Faktenbericht 1990 zum Bundesbericht Forschung 1988, Bonn, 1990. Data on net social product and the price
index for state consumption from W. G. Hoffmann, F. Krumbach, and H. Hesse, Das Wachstum der deutschen Wirtschaft
seilder Mine des 19. Jahrhunderts, pp. 598-60 I.Berlin: Springer, 1965;StatistischesBundesamt, Statistisches Jahrbuch 1989.
Stuttgart: Metzler-Poeschel, 1990.
Industry
The first major science-based industry in Germany was the beet-sugar industry. The
close similarity of the content of the sugar beet and of cane sugar was recognized as
early as 1747 by the pharmacist/chemist Andreas Sigismund Markgraf at the Berlin
Academy of Sciences, who in 1753 became director of the Academy's new chemical
laboratory. His successor Franz Carl Achard developed a process for extracting and
refining sugar from beets. With financial support from the Prussian King he estab-
lished a commercial sugar factory, which, because of high operating cost, however,
became a commercial failure. When Napoleon cut off European trade with Britain,
the beet-sugar industry experienced a first, though artificial boom.53 In the late nine-
teenth century the beet-sugar industry became a major exporter for Germany: up to
1898 the value of sugar exports exceeded that of machinery, and in 1913 it still
exceeded that of synthetic dyestuffs.54 In addition to chemical research, this industry
had a base in agricultural research. The sugar content of beets was increased from ini-
tially about 2 to 15.5%.
In 1866 the Association of the Beet Sugar Industry (Verein fur die Riibenzuck-
erindustrie) founded a chemical laboratory, which may have been the first institute for
industrial cooperative research. In 1903 this laboratory was affiliated with the agricul-
tural college (Landwirtschaftliche Hochschule) at Berlin and then financed jointly by
government and industry.55
The origin of the pharmaceutical industry was in the private pharmaceutical insti-
tutes that existed in the early decades of the nineteenth century. Two of them later
turned into chemical-pharmaceutical firms. Many of their graduates founded their
own business. One of them was Heinrich Emanuel Merck, who in the 1820s began to
produce Pharmaceuticals on a larger scale in an apothecary shop at Darmstadt
founded in 1654. By 1900 the Merck company employed 800 workers, including 50
chemists, pharmacists, engineers, and doctors; by 1913 it was the largest German firm
specializing in Pharmaceuticals. In 1899 the American subsidiary Merck & Co. was
founded.
Also some other leaders in Pharmaceuticals including Schering AG and J. D. Rie-
del were started by pharmacists. Still others such as von Heyden, de Han, and C. F.
Boehringer started as druggists or as makers of fine chemicals. In the last third of the
century technical innovation in this industry drew heavily on the advanced state of
medical and biological research at the universities. By 1913 Germany was the largest
exporter of Pharmaceuticals with a 30.3% share of world exports, far ahead of Britain
(21.3%), the United States (13%), and France (11.9%).56
Important lines of business in the chemical industry supplied inputs to the textile
industry, in particular for bleaching and dyeing. Here Germany was far behind Britain
and Belgium up to the middle of the nineteenth century. Around that time many Ger-
man chemists worked in Britain, as the industry in Germany could not absorb all the
graduates with formal university training. German firms acquired a technological and
commercial lead first in synthetic dyes. The main producers were BASF, Hoechst, and
Bayer, all founded in the 1860s. Later they became Germany's largest chemical com-
panies.57 It was in the German synthetic dyestuffs industry that scientific research first
became a continuous company function separated from production.58 This union of
science and business was not always easily achieved. While BASF and Hoechst were
founded as partnerships of chemists and businessmen, Bayer was started by a dye mer-
chant and a dyer. It fell behind its competitors until, after several false starts, it hired
a capable chemist, Carl Duisberg, who placed the firm's business on a new footing by
synthesizing a new dye and later became the company's chief executive. By 1913 Ger-
many produced about three-quarters to four-fifths of total world output in synthetic
dyes and accounted for 90% of world exports. The main German firms had 15% of
their production located abroad.59
Toward the end of the century new or improved processes for some bulk chem-
icals provided an opportunity to compete effectively with established producers. Elec-
trolysis opened new fields of activity, and first steps were made in synthetic fibers. At
the beginning of the twentieth century photochemicals and new Pharmaceuticals
broadened the range of business. By 1913 the German chemical industry accounted
for 24 percent of world production, whereas the United States contributed 35 percent,
Britain 21 percent, and France 9 percent. About 35 percent of domestic German pro-
duction were exported, so Germany held a share of 27 percent of world exports, fol-
lowed by Great Britain with 16 percent, France and the United States, each with a
share of 10 percent.60
Although some parts of the industry were favored by geographic conditions (Ger-
many had a virtual monopoly of potash at that time), the dyestuffs, synthetic fertiliz-
ers, and pharmaceutical industries are evidence that technological innovation, based
on the country's educational and research systems, was the key factor that enabled the
industry to establish itself as leader on the world export market. According to one esti-
mate nearly 3000 chemists were employed in German industry by 1897.61
For the mining and metal processing industries the mining schools trained gen-
erations of administrators and managers in the eighteenth and nineteenth centuries.
Professors of these schools were instrumental in the transfer of technology from
abroad and their graduates pioneered some new processes, as, for example, the extrac-
tion of petrol and paraffin from lignite.62
In iron and steel, one of the central industries in the nineteenth century, appli-
cation of science was a slow process. The first known instance of a chemist to be
employed by an iron works dates back to 1820s, and for two decades this was the only
known instance.63 However, by the middle of the century, when railroad construction
brought a boom to iron and steel production, most of the larger works employed a
chemist for the analysis of ore inputs and of outputs, while smaller works commis-
sioned their analyses to apothecaries, independent chemical institutes, or institutes at
technical universities. In the 1860s the first plans for a cooperative "testing station"
were discussed, but they materialized only in 1917 with the foundation of the institute
for iron research within the Kaiser-Wilhelm-Society.
Although German steel output exceeded that of Britain by the end of the nine-
teenth century, major innovations such as the Bessemer and the Thomas processes
were made in Britain, and Britain retained a technological lead. Some authors take the
lower price of German steel on the world market at the beginning of the twentieth
century as evidence for a technological lead by German industry and explain this lead
by the existence of cartels that enabled German firms to produce more efficiently than
British competitors.64 Wengenroth has shown, however, that lower prices reflected
poorer quality.65
The German machine construction industry was able to free itself by the middle
of the century from dependence on British technology in some areas of machine con-
struction, including steam locomotives.66 In many areas British and later American
firms had a technological lead until the end of the nineteenth century and beyond,
although some new firms were established in Germany on the basis of new inventions,
for example, printing presses.67 Toward the end of the nineteenth century, when elec-
tric power opened new lines of machine construction and changed the design and the
manufacture of many traditional machines, German firms were able to move to the
technological front in additional areas.
By 1913 Germany accounted for about 27% of world production in machinery.
About 26% of domestic production was exported. According to industry estimates,
Germany held a 29% share of world exports, compared with 28% for Britain and 27%
for the United States.68
The electrotechnical industry began with the construction of telegraph lines. The
German leader was Siemens and Halske, founded in 1847 by the artillery officer Wer-
Table 4.3. Foreign Trade of Germany in the Year 1913 (in Million Marks)
Exports Growth of
Minus RCA* RCA* Exports
Product group Exports" Imports" Imports 1913 1907 1907 = 100
Agriculture, forestry, food 1,729 7,041 -5,312 -131 -139 161

Beet sugar4 264 Od 264 855 614 137
Minerals and fuels'' 841 1,102 -261 -17 -29 168
Minerals 98 627 -529 -177 -135 103
Fuels excluding petroleum 723 290 433 101 38 183
Petroleum 21 186 -166 -212 -84 182
Wax, paraffin, soap 52 27 25 74 76 156
Chemicals and Pharmaceuticals' 999 449 551 90 90 166
Dyes and dye products 298 21 277 275 274 126
Basic chemicals 377 281 96 39 28 192
Pharmaceuticals^ 117 98 19 27 24 178
Explosives 101 2 99 399 316 226
Fibers and textiles*' 1,581 873 708 69 74 108
From wool 444 241 203 71 73 101
From cotton 535 224 311 97 104 110
Leather and leather products 553 167 386 129 35 363
Rubber and rubber products 128 24 104 176 145 271
Products from wood and cork* 164 70 94 101 79 190
Paper, books, pictures, paintings 368 77 291 166 166 129
Products from stones and minerals' 34 33 1 13 -45 164
Products from clay 113 7 106 287 265 106
Glass and glass products 146 17 129 224 226 127
Products from gold and silver 74 29 45 102 222 75
Other metals and products thereof 1,903 673 1,230 113 102 164
Iron and iron products 1,336 105 1,231 264 212 169
Machinery1 680 80 600 223 177 176
Steam-powered vehicles'1 75 2 73 384 333 163
Combustion engines' 40 2 38 317 255 255
Electrotechnical products 290 13 277 321 317 173
Vehicles' 161 18 142 227 106 135
Motor vehicles'" 98 15 83 200 2 676
Rail vehicles'1 36 2 34 295 407 164
Ships 15 29 -14 -55 -52 101
Firearms 16 2 14 233 140 275
Clocks and watches 30 30 1 11 17 114
Musical instruments 84 4 80 306 290 130
Toys 103 1 102 459 487 129
Gold and silver 101 436 -335 -135 21 41
Not adequately declared 31 2 29 (14) (16) (16)
Total 10,199 11,206 -1,007 0 0 144
"The figures include only the official category of "Spezialhandcl" (i.e., imports from foreign countries into tax-free areas and
exports to foreign countries from these areas are excluded).
^Revealed comparative advantage (RCA) for industry /is denned as 100 X In [(m//j:/)/(E] mjfLxj)}, where mare imports and
j
x are exports.
^Comprising categories 176b-i and 1761.
Îmports were 56 thousand.
*CoaI tar and coal tar products have hcen moved from "Minerals and fuels" to "Chemicals."
^This item includes ether, alcohols, and cosmetics. It comprises product groups 4D and 4G of the German foreign trade clas-
sification, excluding for both years categories 371, 374, 375a, 377, 378, 379, and incompletely declared products, for year 1907
categories 390c, 390d, and for the year 1913 category 390b.
^'Including textures from plant fibers and plant material, also brooms, brushes, and sieves.
ner Siemens and the mechanic J. G. Halske. The invention of the dynamoelectric prin-
ciple by Werner Siemens in 1866 opened the use of electricity for power and changed
this industry's relationship to science. Siemens and Halske hired the first university-
trained physicists in 1872, and by 1882 the first professorial chair for electrical engi-
neering was established at a Technische Hochschule.69 Another major firm was AEG
(Allgemeine Elektricitats-Gesellschaft), started in 1883 on the basis of a license of the
Edison patents for Germany. A third large firm founded by Johann Siegmund Schuck-
ert (who had worked for some time with Edison in the United States) was merged in
1903 with Siemens, and from then on Siemens and AEG divided among them about
half of the domestic market for electrotechnical products.
By 1913 Germany accounted for 34.9% of world production, compared with
28.9% for the United States and 16.0% for Britain.70 About 25% of German production
was exported. Germany held a 46.4% share of world exports, followed by Britain (with
the strong presence of American and German subsidiaries) with 22.0% and the United
States with 15.7%.
The combustion engine and the motor vehicle provided another new line of busi-
ness in the machine construction industry. With the inventions by Otto, Diesel, Daim-
ler, Benz, and Bosch, German firms were among the early technical leaders. However,
they failed to turn this into a commercial lead. It was only after the turn of the century
that motor vehicles with combustion engines were produced in larger numbers.71 At
that time the leaders were France and the United States. In 1913 France accounted for
33.4% of world exports, followed by the United States with 23.7% and Germany with
17.2%. About 40% of German production was exported.72
Catching up and Taking the Lead

By the beginning of the twentieth century the results of rapid industrialization became
visible at home and on world markets. Although in 1870 German gross domestic prod-
uct ($21 billion at 1970 U. S. relative prices) was less than that of Britain (30 billion),
the United States (30 billion), and France (24 billion), by 1913 it was larger ($72 billion
at 1970 U. S. relative prices) than that of Britain (68 billion) and France (47 billion),
though the United States had surged ahead (176 billion). Per capita gross national
product was in 1913 still about 23% lower than that of Britain, then the wealthiest
country in Europe. Also Switzerland, Belgium, Denmark, the Netherlands, and Nor-
way had a higher per capita gross national product.73
In all manufacturing industries, Germany had an export surplus by 1913 (see
Table 4.3). The majority of exports, however, was still in the older industries, such as
agriculture, fibers and textiles, or metals and metal products. The newer industries
such as electrotechnical or motor vehicles contributed a small share (6.7 and 0.1%,
''Including carvings from plant and animal materials.

'Not including clay.
'As in the source, steam-powered vehicles are grouped with machinery rather than with rail vehicles.
^Comprising categories 892a-b and 893a-c for the year 1907, categories 892a-c and 893a-d for the year 3913.
^Comprising categories 894d-e, k, o.
m
Road vehicles only. Comprises categories 915a-c, c.
Source: Germany, Kaiserliches Statistisches Amt, Statistik des Deutschen Reichs, Vols. 189 (1909) and 270 (1914).
respectively), yet they were highly dynamic, as they had a high revealed comparative
advantage and rapid export growth. Highly dynamic were also machinery, coal (fuels
excluding petroleum), leather and leather products, rubber and rubber products, prod-
ucts from wood and cork, metals and products thereof, and firearms.
Although synthetic dyes were a showpiece of a science-based industry, in 1913
dyes and dye products contributed only a third of exports in chemicals. The export to
import ratio (indicated by the RCA index) in dyes and dye products is lower than in
most industries, and from 1907 exports in this product group grew less than total
exports.
Patenting activity in the United States can also serve as an overall indicator for
the technological capability of German industry. Whereas in 1883 German firms were
granted about half as many American patents as British firms, by 1913 they accounted
for 34% of all foreign patents in the United States, while Britain had fallen behind with
23%.74
By the beginning of the twentieth century, many German firms operated on a
worldwide scale.75 Their accumulated foreign direct investment was estimated in 1914
at about 1500 million dollars. This was still significantly less than the foreign direct
investment of Britain (6500 million dollars) or the United States (about 14,300 million
dollars).
Within a few generations the German economy had nearly caught up with the
British in terms of per capita social product, and in several industries it was now among
the world's technological leaders.76 This is not an exceptional case as other countries
such as Switzerland or Denmark achieved a similar development. In the literature, the
German case has often been described as an instance of the "advantages of backward-
ness," implying that a follower country adopting new technology from abroad can
move faster than a leader country, since the latter faces some retardation as a result of
old vintages of capital stock and the organizational resistance associated with old tech-
nologies.77 Certainly Germany could not have industrialized as quickly as it did with-
out the transfer of technology from countries such as Britain, Belgium, and later the
United States. But the historical account given here suggests that it could take the lead
in some industries not because of the "advantages of backwardness," but only because
it established new institutional forms that enabled German firms to move quickly as
new product areas or new processes were opened up by inventions and by advances in
scientific knowledge.
CONTINUITY AND CHANGE IN THE TWENTIETH CENTURY
At the beginning of the twentieth century Germany was on its way to join the wealth-
iest countries and had a dynamic industry that was moving rapidly into world markets.
But political development had not kept pace with economic development.78 The auto-
cratic political order blocked development toward a more democratic system and the
political elite was unable to handle the foreign policy challenges that came with the
country's rise to an industrial power. World War I began a series of political and eco-
nomic crises that ended only toward the late 1940s.
World War I cut German industry oft" from its main export markets, stimulated
efforts in the United States, France and Britain to substitute for German imports by
domestic production, and enabled firms in neutral countries to capture market shares.
The peace treaty of Versailles entailed the loss of some regions with a significant part
of Germany's mineral resources, heavy reparations, and the expropriation of German
patents and of direct investment in the United States, France, and Britain. A hyper-
inflation in 192 3 annihilated monetary assets. The short economic upswing that fol-
lowed was ended by the slump of 1929. The subsequent recession was one of factors
that brought the National Socialists to power in 1933.
Firms reacted to the economic crisis by increasing their effort toward different
forms of nonmarket cooperation. Cartels had been a normal part of the Germany
economy since the second half of the nineteenth century and could be enforced using
the legal system. In the 1920s the number of domestic cartels increased, and German
firms participated in many international cartels.79 Furthermore, there was a move-
ment towards mergers.80 In the chemical industry the three leaders and some other
firms joined in the IG Farben.81 Technical associations such as the Association of Ger-
man Engineers (Verein Deutscher Ingenieure) had for a long time provided a platform
for exchange of technical information and for standardization. Economic planning
during World War I led to the introduction of new standards in many industries, and
these efforts were continued in the Committee for German Industry Norms DIN. For
standards and norms a complex network of technical and business associations
evolved.82 The well-organized industry associations were then used by the national-
socialist government to impose elements of a command economy while leaving much
of formal industry structure as it was.
Their technological basis enabled German firms in some industries to return
quickly to the world market after World War I. In chemicals Germany recaptured its
prewar export position by 1929, and in machinery it held 35.0% of world exports by
1931, even surpassing its share as of 1913 (29.2%). The electrotechnical industry, how-
ever, recovered by 1932 a share of only 34.9%, compared with 46.4% in 1913. The
motor vehicles industry, which had a prewar share of 17.2%, virtually disappeared
from the world market. In the 1920s its share never exceeded 3%, and by 1928 Ger-
many was a net importer. The case of the automobile industry reflects the difficulty
German firms had in procuring the capital required for mass production. In this
period, the largest German automobile manufacturer, Opel, sold out to General
Motors. When in 1931 Ford completed a factory in Germany, these two companies
accounted for 71% of the production capacity for passenger cars.83
The economic situation also affected scientific research. During the war and up
to 1922 many new research institutes had been established by the state and by the
Kaiser-Wilhelm-Society.84 Some institutes of the Kaiser-Wilhelm-Society redirected
their activities toward the war effort. Between 1918 and 1923 the number of institutes
increased from 9 to 16, and Prandtl's institute for aerodynamics was then taken over
by the Kaiser-Wilhelm-Society.
In the years after World War I a surge in student enrollment in higher education
(see Fig. 4.1) resulted in unemployment for academics. Increases in government
expenditure for scientific institutions did not keep pace with inflation, so the real value
of salaries deteriorated and the purchase of books, equipment, and materials had to be
cut.85 As a reaction to these difficulties a new organization was founded in 1920, the
Notgemeinschaft derDeutschen Wissenschaft (Emergency Association of German Sci-
ence). It provided research money mainly to individual scientists on the basis of appli-
cations. In 1920 it received about 20 million Marks from the central government,
which is 4% of all expenditure for science and higher education by states and central
government together.86 For donations to be collected from industry a special organi-
zation was created, the Stifterverband der Notgemeinschaft der Deutschen Wlssen-
schaft (Donors' Union of the Emergency Association of German Science). To 1922
the Stifterverband collected a sum of about 100 million Marks, which was invested as
an endowmentand a little later was virtually annihilated by inflation. The Stifterv-
erband continued to depend on further donations, and to 1933 contributed about 1 to
3% of the income of the Notgemeinschaft. German firms provided additional funds
for research through scientific societies in specific disciplines.87
In these years the Notgemeinschaft received some donations from abroad, for
example, from the Rockefeller Foundation and the General Electric Company in the
United States, However, more than 80% of its income came from the state.88 In 1929
the Notgemeinschaft changed its name to Deutsche Forschungsgemeinschaft. Accord-
ingly, the Stifterverband had to change its name and became the Stifterverband fur die
Deutsche Wissemchqft.
The Kaiser-Wilhelm-Society ran into financial problems, as its endowment,
invested in government bonds, shrank as a result of inflation. From 1921 it received
additional funds on a regular basis from government. From 1924 to 1933 public funds
accounted for 65% of its budget; 27.8% came from domestic private sources, and 7%
from foreign sources, mainly from foundations. In the wake of the slump of 1929 both
government and industry reduced their funds.89
Nevertheless, the bleak picture of German science in the 1920s, as it is painted in
the literature, appears to be overdrawn. In 1925 expenditure for higher education and
science by states and central government was in real terms 2.0 times higher than in
1913 (see Table 4.2), and in 1930 it was 2.4 times higher. Obviously, the official
descriptions of the state of German science in these years contain a good deal of rhet-
oric, which historians have barely begun to separate from reality.90
The scarce data that are available indicate that the industry sector in the 1920s
generally increased its R&D efforts.91 In the depression of 1929 and the following years,
however, industry in Germany as in other countries laid off research personnel,
reduced its spending on R&D, and concentrated on traditional rather than radically
new innovations.92
In the period of National Socialism the number of students in higher education
was drastically reduced (see Fig. 4.1). Although a large majority of academics tolerated
the authoritarian rule of the National Socialists, many scientists and engineers were
removed from their posts. Researchers in all fields of scholarship were forced to emi-
grate, including leaders in their field such as the physicist Albert Einstein and the math-
ematician John von Neumann. As many of the emigrees were unwilling or unable to
return, the national-socialist period left a damaging effect on the quality of German
science for more than one or two decades.93
Although some clandestine military R&D had continued during the 1920s in
spite of the injunctions of the Versailles treaty, in 1935 rearmament was officially
resumed and a little later a plan was implemented to make the economy independent
in strategic materials. Since the planning machinery of the National Socialists was put
on top of existing industrial structures, it did not radically change the innovation
system.
During World War II, large parts of industrial plants were destroyed. After the
war, some of what remained was taken as reparations by the allied powers. Subsidiaries
of German firms abroad and all patents and trademarks were disowned. Some scien-
tists and engineers were moved to the allied countries to be employed in military, aero-
space, and nuclear technologies.94
After this, the two Germanys, in which the country was divided, made an
astounding economic recovery. The fact that technical knowledge and skills still
existed was a key factor in this. In the Eastern part of the country the Soviet Union
introduced a centrally planned economy, and the existing innovation system was
replaced step by step by something different.95 In the Western part the allies abolished
the planning structures of the national-socialist economy. They deconcentrated some
industries effectively, in particular the chemical and steel industries; however, in other
industries such as banking their deconcentration policy had only a temporary impact.
Furthermore, they introduced a trade union structure that virtually avoided conflicts
among specialized trade unions within firms. In all these changes the basic compo-
nents of the innovation system were reconstructed: the firms and their laboratories,
the schools, the universities and Technische Hochschulen, the Kaiser-Wilhelm-Soci-
ety (that in 1948 became the Max-Planck-Society), the Deutsche Forschungsgemein-
schaft (recreated in 1951), government research institutes, and business and technical
associations.96 Most organizations could take guidance in their reconstruction from
their history before the national-socialist period.
THE PRESENT SYSTEM
This section focuses on the innovation system in West Germany as it existed before
unification. Space does not allow us to describe the innovation system of the former
German Democratic Republic and the way it is being transformed by adopting the
institutional structures of West Germany. As a starting point I take the pattern of tech-
nological capabilities shown in the export performance of West German industry in
the years before unification.97
In describing the national system for technical innovation, one has to keep in
mind that in any country this is only one among several factors that account for eco-
nomic performance. Other factors that deserve to be mentioned in the case of Ger-
many but cannot be discussed in this chapter for lack of space are the system of indus-
trial relations that has limited trade union conflict within industries, social policies and
labor market policies that have eased the phasing out of noncompetitive production
facilities, and the German banking system that enables banks to support the restruc-
turing of industries.98
Export Performance
Table 4.4 presents the structure of West German foreign trade in the year 1988 for the
all one-digit and for selected two- or three-digit commodity groups. The major Ger-
man exports are machinery of various types (machines for special industries, metal-
working machines, general industry machinery, and electrical machinery together had
exports worth $65.8 billion U. S.), furthermore road vehicles ($54.7 billion U. S.), and
Table 4.4. West German Foreign Trade in 1988 (in Million U.S. Dollars)
Exports Growth of
SITC Classification" Exports Imports Imports 1988 1983 (1983 = 100)
0 Food and live animals 12A63 23,617 -10,154 -82 -82 182
061 Sugar and honey 355 366 -11 -29 10 91
1 Beverages and tobacco 1,847 2,498 -651 -56 -49 166
2 Crude materials excluding fuels; 5,886 16,081 -10,195 -126 -131 186
233 Synthetic rubber 575 553 22 -22 10 163
26 Textile fibers and their wastee 1,164 1,455 -291 -48 -39 149
28 Metalliferous ores and scrap) 1,436 4,163 -2,727 -132 -151 209
3 Mineral fuels _4 J 2K) JMZ1 -14,963 -177 -183 73
32 Coal, coke, and briquettes 1,467 506 961 81 80 97
33 Petroleum, petroleum productsucts 2,002 14,710 -12,708 -225 -461 80
tured
341 Gas, natural and manufactured 218 3,304 -3,086 -296 -154 84
4 Animal and vegetable oil and fatit __872 819 53 -19 -17 129
5 Chemical and Pharmaceuticals 42,810 _23j752 19,058 33 47 193
53 Dyes, tanning, and colors 4,443 1,161 3,282 109 131 206
541 Medicaments and 5,074 2,956 2,118 28 48 239
Pharmaceuticals
55 Perfume, cosmetics, soap 2,309 1,274 1,035 34 46 217
58 Plastic materials 11,397 6,720 4,677 27 42 209
6 Basic manufactures 58,961 46,460 12,501 -2 188
^3
61 Leather, leather manufacturesres 1,050 1,317 -267 -48 -30 170
62 Rubber manufactures 2,954 2,582 372 -12 6 207
63 Wood and cork manufactures es 1,239 1,531 -292 -47 -62 222
64 Paper, paperboard, articles thereof 6,877 5,947 930 -11 -3 236
65 Textile yarn, fabrics 10,738 8,833 1,905 -6 -3 197
662 Products from clay 989 740 249 3 6 174
663 Polishing stones, abrasives etc.
etc. 1,461 781 680 37 36 219
664, 665 Glass and glassware 2,373 1,432 941 25 23 256
666 Pottery 561 312 249 33 35 193
67 Iron and steel 13,825 9,494 4,331 12 37 173
68 Nonferrous metals 6,443 7,533 -1,090 -41 -20 173
69 Other manufactures of metals als 10,016 5,113 4,903 42 66 181
7 Machines and transport equipmentlent 155,263 72,972 82,291 50 73 202
71 Power generating equipmentit 9,220 4,855 4,365 39 74 188
72 Machines for special industries
ries 20,044 4,691 15,353 120 135 202
73 Metalworking machinery 5,577 1,985 3,592 78 121 203
74 General industry machineryf 21,730 7,538 14,192 80 98 214
751 Office machines 1,048 1,065 -17 -27 -24 200
752 Automatic data processing; 4,342 6,342 -2,000 -63 -21 234
equipment
759 Parts of office and ADP 2,447 3,46 1 -1,014 -60 -22 189
machines
761-763 Television and sound 2,732 3,6 1 5 -883 -54 -46 221
equipment
764 Telecommunications equipmentpment 3,273 3,378 -105 -29 44 173
nery
771-775, 778 Electrical machinery 18,469 9,972 8,497 36 62 230
776 Semiconductors, valves 3,435 3,727 -292 -34 -39 300
78 Road vehicles 54,678 17,831 36,847 86 115 205
791 Railway vehicles 564 83 481 166 160 127
792 Aircraft 1,085 2,626 -1,541 -114 -39 32
793 Ships 753 339 414 54 138 72
Table 4.4. West German Foreign Trade in 1988 (in Million U.S. Dollars) (Continued)
Exports Growth of
SITC Classification" Exports Imports Imports 1988 1983 (1983 - 100)
8 Miscellaneous manufactured goods 35,900 37,424 -1,524 -30 _nl! .226
812 Sanitary, plumbing, heating 1,210 896 314 4 3 289
fixtures
821 Furniture 3,989 2,797 1,192 10 19 239
84 Clothing and accessories 5,377 14,515 -9,138 -125 -108 212
87 Scientific instruments 7,952 4,004 3,948 43 56 228
881-883 Photographic goods 2,184 2,433 --249 -36 -5 205
884 Optical goods 732 670 62 -17 25 187
885 Watches and clocks 769 719 50 -19 7 192
892 Printed matter 2,926 1,123 1,803 70 95 231
893 Articles of plastics 3,976 2,097 1,879 38 48 261
894 Toys and sporting goods 1,118 1,377 --259 -46 24 208
895 Office and stationery supplies 665 276 389 62 79 228
896 Works of art 1,344 361 983 106 -20 78
897 Jewelry 698 659 39 -20 60 157
898 Musical instruments 1,653 1,588 65 -22 -4 249
9 Goods not classified elsewhere 13,985 7,497 -3,512 -89 _Z12 _84
95 1 War firearms, ammunition 684 321 363 50 104 1421
Total 323,196 250,293 72,903 0 0 191
"Data for one- and two-digit SITC groups in 1988 are taken from volume 1, data for three-digit groups from volume 2 of the
International Trade Statistics Yearbook. Because of the transition from SITC 2 to SITC 3 there may be small inconsistencies
between the first and the second set of data, and for some one-digit or two-digit groups the comparability to earlier years may
be only approximate.
*Revealed comparative advantage (RCA) for industryi is defined as 100 X In [(m,/x,)/ (E m// Xj)], where m are imports and
}
x are exports. '
chemicals and Pharmaceuticals ($42.8 billion U. S.). These industries together

accounted for 50.5% of exports and 27.5% of imports. Other areas with significant net
exports were scientific instruments, manufactures of metals (including hand tools, fit-
tings, and nails), power generating equipment, iron and steel manufactures, textile
yarn and fabrics, plastic articles, printed matter, and furniture.
Rather than being concentrated in a few product groups, German exports are
spread over many product groups. In some four-digit or five-digit industries German
industry accounts for a high share of world exports: 57.7% in rotary printing presses,
53.0% in reciprocating pumps, 46.3% in textile spinning machines, 42.4% in high-pres-
sure hydroelectric conduits of steel, 42.2% in coke of coal (one of the few areas where
Germany has a natural resource advantage), and 40.0% in combined harvesters-
threshers." However, the contribution of this top-league to total exports is small:
$4483 million U. S. or less than 2%.1(X)
In some product groups, the pattern of imports reflects the country's poor endow-
ment in natural resources, for example, in petroleum and natural gas. But there are
also some high-technology sectors, in which West Germany was a net importer: ADP
machines, aircraft, television and sound equipment, telecommunications, photo-
graphic goods, and semiconductors.
History goes a long way toward explaining the pattern of technological strength
and weakness as it is reflected in West Germany's export performance. In most indus-

tries that today are net exporters Germany performed well on the world market in
1913 (compare Tables 4.3 and 4.4). This applies not only to those industries that now
account for the bulk of German exports (machinery, electrotechnical products, motor
vehicles, and chemicals), but also to manufactures of metals, in particular iron and
steel, as well as textile yarns and fabrics, rubber products, paper and printed matter,
products from stones and minerals, products from clay and pottery, glass and glass-
ware, railway vehicles, jewelry, firearms, and musical instruments. Even a new prod-
uct group such as plastic articles is related to industries in which Germany was strong
at the beginning of the century: chemicals and machinery.
However, not all industries that were strong exporters in 1913 were still strong in
the 1980s, for example, leather and leather products, products of wood and cork, and
clothing. These became sensitive to the cost of unskilled or semiskilled labor, so firms
had to concentrate on high-quality products or specialized inputs, be they machinery
or materials.
What at first appears as historical continuity partly reflects the impact of political
events on some industries that newly emerged or that took large strides in technical
change. After World War II the allies prohibited R&D for military technology as well
as for some areas of civilian technology, including nuclear technology, aeronautics,
rocket propulsion, marine propulsion, radar, and remote and automatic control. Spe-
cial permission was required for work on such things as electronic valves, ball and
roller bearings, synthetic rubber, synthetic oil, and radioactivity other than for medical
purposes."" The key injunctions remained in force until the Federal Republic became
a sovereign state in 1955. They effectively wiped out the military and aeronautics
industries and in some product groups kept German firms for some time away from
the technological front. This is one of the reasons for the relatively poor export per-
formance of the German aircraft, electronics, and telecommunication industries.
Historical continuity appears also at the firm level. Table 4.5 lists the 25 largest
manufacturing companies in Germany. Of these 19 were founded before 1913, and
three of those for which a later founding date is given (VEBA, Ruhrkohle, and Preus-
sag) are mergers of enterprises many of which date back to the nineteenth century.
Although there have been recurrent alarms about the technical weakness of the
West German industry, from the technology gap discussion in the 1960s to the euro-
sklerosis cries in the early 1980s, its performance has been better than these alarms
made one expect.102 Although from the mid-1970s to the mid-1980s there has been a
slowly declining trend in Germany's world-market share in high-technology products
however defined and also in goods with more than average research intensity, the trend
was reversed in the late 1980s.103
Nevertheless, there are areas for concern.104 First, West German exports are con-
centrated in Europe, which takes about 70%. The dynamic markets of the newly indus-
trializing countries are less well represented than the size of their economies would
suggest.
Second, to some extent the good performance in European markets is the result
of trade protection against non-European imports. In the passenger car industry, for
example, the countries of the European Community accounted for about 64% of
exports in 1988 and were the main factor in export growth in the 1980s, whereas
exports to North America declined.105
Table 4.5. The 25 Largest Manufacturing Companies in Germany (Year 1989)
Founded Sales (Billion Employees R&D Expenditure
Company Main Activity (Year) DM)" (Thousand) (Million DM)
Daimler-Benz* Vehicles/electrical/ 1882 76.4 368.2 5494
aerospace
Volkswagen Vehicles 1938 65.4 250.6 2300
Siemens' Electrical/ electronics 1847 61.1 365.0 6875
VEBA Energy/chemicals 1929 49.2 94.5 370
BASF Chemicals 1865 47.6 137.0 1954
Hoechst Chemicals 1863 45.9 169.3 2621
Bayer Chemicals 1863 43.3 170.2 2695
RWE Energy 1898 38.9 78.2 n.a.
Thyssen Steel/machinery 1867 34.2 133.8 735
Bosch Electrical 1886 30.6 174.7 1803
BMW Vehicles 1916 26.5 66.3 ca. 3300
Ruhrkohle Coal mining 1968 23.4 124.8 273
Mannesmann Machinery 1890 22.3 125.8 518
Opel Vehicles 1862 20.8 54.6 706
Metallgesellschaft Metals/plant 1881 20.1 24.5 n.a.
engineering
Ford Vehicles 1925 19.8 48.2 n.a.
Krupp Steel/machinery 1811 17.7 63.6 275
MAN Machinery 1840 17.1 63.7 415
Deutsche Shell Petroleum 1902 16.9 3.3 47
Preussag"' Energy/oil 1923 16.4 65.7 200
Hoesch Steel 1871 15.9 52.0 n.a.
Degussa Chemicals/metals 1873 14.4 33.7 422
Bertelsmann Publishing 1835 12.5 43.7
IBM Electronics 1910 12.4 31.1 n.a.
Deutschland
Henkel Chemicals 1876 11.6 38.1 359
"One billion = 1000 million.
Not including Messerschmitt-Bolkow-Blohm (MBB), which was acquired by Daimler-Benz in 1989.
c
Not including Nixdorf, which was acquired by Siemens in 1990.
"Încluding Salzgitter, which was acquired by Preussag in 1989.
Sources: Die Zc'il, 17 August 1990, p. 24; Handbuch der deutschen Aktiengeselischaften 1989/90. Darmstadt: Hoppenstaedt,
1990/91; data on R&D from the firms' annual reports and personal correspondence.
Third, the dynamism of Japanese industry is about to affect a broad number of

industries. Although in some smaller industries such as television and sound equip-
ment, and photographic goods, West German firms have lost their markets to Japa-
nese firms in recent decades, in other sectors such as machine tools they have met the
Japanese challenge.106 Patent statistics suggest, however, that Japan is improving its
technological position in many industries. In the period 1975 to 1985 Japanese patents
in the United States increased from 8.9 to 17.9%, while German patents grew only
from 8.5 to 9.5%. Relative to population or gross domestic product, Germany has now
about the same share as Japan, but the trend is a matter for concern. Japan is catching
up with Germany in the chemical and mechanical industries, moving ahead in phar-
maceuticals and instruments, and increasing its lead it electronics, data processing,
and communications as well as in material science and transportation equipment.
Moreover, Germany has relatively fewer important patents (as measured by a high
frequency of citations).107
Research and Development in Industry
A key factor in the technological strength of any country is the innovation activities of
business enterprises. R&D is only a part of these innovation activities, but in many
industries it is an essential part.'08 In Germany about 63% of total national R&D is
financed by the business sector, a much higher percentage than in the United States,
France, Britain, or Italy, but a lower percentage than in Japan where it is 78% (see
Table 1.1 in the introduction to this part of the book). R&D expenditure financed by
the business sector grew from 0.60% of the gross national product in 1962 to 1.87% in
1989 (see Table 4.6). This is more than in France (0.96%), in Britain (1.15%), Italy
(0.50%), or the United States (1.43%), about the same as in Sweden (1.83%) and Swit-
zerland (1.76% in 1983), but less than in Japan (1.98%).109 At the aggregate level, gov-
ernment is not an important source of funds for R&D performed in the business sec-
tor. In 1987 it financed 4.9 billion DM, which is about 12% of all R&D performed by
the business sector (see Table 4.7).
About 31 % of domestic industrial R&D capability (as measured by R&D employ-
ees) is accounted for by the seven top spenders: Siemens, Daimler-Benz, Bayer,
Hoechst, Volkswagen, and BASF."0 Siemens alone stands for about 37% of R&D in
the electrotechnical industry; Bayer, Hoechst, and BASF for about 46% in the chem-
ical industry; and Volkswagen and Daimler-Benz for about 53% in the motor vehicle
Table 4.6. R&D Expenditure in West Germany 1962 to 1989
Financed by Business Financed by Government
Performed by Business Performed by

Total
Sector Business Sector
R&D Percent Percent Subtotal

(Billion Billion of Percent Billion of Billion Billion
Year DM) DM Total ofGNP DM Total Percent" DM DM Percent*
1962 4.5 2.2 47.9 0.60 2.1 47.0 98.1 2.3 0.3 13.9
1963 5.4 2.7 49.6 0.70 2.6 48.7 98.1 2.6 0.4 13.2
1965 7.9 4.1 51.3 0.89 4.0 50.2 97.8 3.7 0.6 12.3
1967 9.7 4.8 49.4 0.97 4.6 47.4 96.0 4.8 1.0 17.5
1969 12.3 6.4 52.2 1.07 6.2 50.8 97.2 5.7 1.0 14.2
1971 18.0 8.7 48.5 1.16 8.5 47.0 96.9 8.7 2.0 18.2
1973 20.5 9.6 47.0 1.05 9.4 46.2 98.2 10.4 2.3 19.5
1975 24.6 11.8 47.8 1.15 11.6 47.0 98.3 12.0 2.9 19.2
1977 27.7 14.1 50.9 1.18 13.8 49.9 98.2 12.6 2.9 16.6
1979 34.5 18.6 57.0 1.34 18.3 53.2 98.3 15.1 4.4 19.0
1981 39.4 22.1 56.1 1.43 21.4 54.4 96.9 16.7 4.8 18.2
1983 43.9 25.5 57.9 1.52 24.7 56.2 97.0 17.8 5.3 17.3
1985 52.3 31.1 59.5 1.69 30.1 57.6 96.8 20.5 6.0 16.3
1987 59.5 36.8 61.9 1.82 35.7 60.1 97.0 21.7 5.3 12.7
1989 66.7 42.4 63.5 1.87 41.1 61.6 97.0 23.0 5.4 11.3
0
Percent of billion DM financed by business.
Percent of billion DM performed by business sector.
Sources: Bundesminister fur Forschung und Technologic (ed.), Faktenbericht 1990 zum Bundesforschungsbericht 1988.
Bonn: Bundesminister fur Forschung und Technologic, April 1990, Table VI/3, p. 340sq. GNP from Statistisches Bundesamt
(ed.), Statistisches Jahrbuch 1983 fur die Bundesrepublik Deiitschland. Stuttgart: Kohlhammer 1983, Table 23.2, p. 528: Sta-
tistisches Bundesamt (ed.), Statistisches Jahrbuch 1989. Stuttgart: Metzler-Poeschel, 1989, Table 24.2, p. 542.
Table 4.7. West German R&D Expenditure in 1987 by Source or Funds and Sector of
Performance (Million DM)
Source of Funds
Sector of Private
Performance Business Government Nonprofit Abroad Total
Business 35,739 4,899 62 629 41,329
Government 86 6,990 112 100 7,288
Higher education 525 7,814 0 0 8,339
Private nonprofit 54 157 64 9 284
Abroad 427 1,306 0 _ 1,733
Total 36,831 21,166 238 738 58,973
Source: H. Echterhoff-Severitt, C. Grenzmann, R. Marquardt, E. Menner, A. Weisburger, J. Wudtke, Forschung undEntwick-
lung in der Wirlschaft 1987, mil erslen Dalen 1989. Essen: SV-Gemeinnutzige Gesellschaft fur Wissenschaftsstatistik, 1990.
industry. Patent statistics confirm the concentration of technological capability: five

German firms account for 29% of German patents in the United States.'"
As Table 4.7 shows, 97% (35.8 billion DM) of the funds provided by the business
sector are also spent in the business sector. Of this sum about 2.5 billion DM or 7% is
for outside R&D performed by other business enterprises, and 338 million DM or 1%
goes to institutes of cooperative industrial R&D. The latter had a total budget of 617
million DM in 1987, of which 55% was funded by the participating business enter-
prises, 17% by a program of the Federal Ministry for the Economy, and the rest mainly
through projects in other government programs."2 The business sector also finances
some R&D performed in the higher education sector (525 million DM in 1987), in
government laboratories (86 million DM), and in domestic private not-for-profit insti-
tutions (54 million DM).
According to the statistics summarized in Table 4.7 in 1987 the German business
sector received 629 million DM from abroad for R&D performed for foreign sources
and paid 427 million DM to foreign R&D performers. Data by the Deutsche Bundes-
bank put receipts at 3111 million DM and expenditure at 2549 million DM."3 The
latter figures appear to reflect more adequately the internationalization of R&D."4
According to one estimate 40,000 R&D employees (about 14% of R&D employees in
domestic industry) work in subsidiaries of foreign firms in Germany, and about the
same number work abroad in subsidiaries of German firms. In 1989 Hoechst spent
about 40% of its R&D funds in foreign subsidiaries, Bayer 33%, and Siemens 20%.l15
The Education System
As Table 1.1 in the introduction to Part I of this book and similar tables in other parts
show, most advanced Western countries do not differ in elementary schooling (for
which the literacy rate may serve as a rough indicator). However, there are significant
differences in the structure and duration of second-level education. In Germany, the
enrollment rate at age 17 is 98.7%, which is the highest among all OECD countries (see
Table 4.8). The system of vocational training has gone through some changes as some
skills disappeared, new ones emerged, and many skills changed, but it has the same
basic structure as established around the turn of the century, with part of the training
Table 4.8. International Comparison of Indicators for Education

Expenditure Per
Enrollment Rate as Percent of Age Students Qualifying as Student as
Group Percent of Age Group Percent of Per
Capita GDP
Age 17 Age 20
Level Level Level Second Third
Country Total Part-Time Total Part-Time 5 6 7 Level Level
Australia 50.3 10.4 n.a. n.a. 9.7 16.5 1.6 19' 55
Austria n.a. n.a. n.a. n.a. 3.6 6.8 0.5 37 50
6
Belgium 86.4 4.8 46.0 7.4 15.8 14.9 31 38
Canada 78.5 0.1 36.3 1.3 13.3 24.5 4.1 n.a. 41
Denmark 75.4 n.a. 36.4 n.a. 10.3 12.7 ft 31 39
Finland 83.3" 22.4" 17.0 11.8 0.9 29 40
France 79.7 10.3 28.1 0.7 14.7 15.3 6.2 26 30
Germany 99.7 48.4 36.6 19.5 7.7 12.4 1.5 27 37
Greece 58.7" 29.0" 4.9 10.9 0.4 15 36
Ireland 64.7" 17.6" 1.0 10.8 5.2 25 73
Italy n.a. n.a. n.a. n.a. 0.4 7.9 1.4 n.a. n.a.
Japan 90.5 1.6 n.a. n.a. 11.1 21.9 1.5 23 118
31.9 0.2 16.5 b
Netherlands 78.3 6.8 22 82
New Zealand 38.8 1.4 16.1 3.0 4.1 11.2 4.5 15 56
Norway 76.2 2.0 23.4 2.8 36.3 16.1 7.3 38J 36
Portugal n.a. n.a. n.a. n.a. n.a. n.a. n.a. 25 60
Spain 53.1" n.a. n.a. 0.1 14.7 0.5 n.a. n.a.
Sweden 83.0 9.5" n.a. n.a. n.a. 38d 40
Switzerland 83.1 56.3 30.2 15.1 0.9 6.1 0.3 28'' 52
United Kingdom 49.3 16.2 23.8 9.9 12.3 14.2 4.6 28 85
United States 89.0 0.4 35.7 3.8 12.7 24.1 9.7 29 39
a
Full-time students only,
'included in Level 6.
'First level and second level.
^Second level second stage only.
Source: Organisation for Economic Cooperation and Development, Education in OECD Countries, 1986-87. Paris: Organ-
isation for Economic Cooperation and Development, 1989.
provided by business firms and other organizations that are the future employers, and
part by vocational schools.'l6
The higher education sector, once a showpiece of the German education system,
no longer is so. In 1985, 7.7% of the corresponding age group received a qualification
of level five (as defined by the OECD), meaning completion of a course of education
usually outside the university and of shorter duration and of a quality lower than a full
university degree; 12.4% completed a first full university degree (level 6), and 1.5% a
postgraduate education (level 7). Many other advanced industrial countries did better
in all these categories: Australia, Belgium, Canada, France, Japan, Norway, the United
Kingdom, and the United States (see Table 4.8).
One may argue that the quality of education varies among countries. But as far
as expenditure per student is a measure, this argument does not help in the case of the
German higher education sector. With 37% of per capita GDP Germany spends about
the same per student as Belgium, Denmark, Greece, Norway, and the United States,
but other countries spend much more, such as the Netherlands (82%) or the United
Kingdom (85%). Japan is top with 118%.
Comparison with Japan shows a long-term trend in the educational level of the
labor force that is unfavorable for Germany. Per 10,000 labor force, Japan had in 1965
about 8% more scientists and engineers employed in nonacademic jobs than West
Germany. The difference increased by the mid-1980s to about 27%. "7
The higher education sector has been neglected in Germany since the mid-1970s.
In real terms neither expenditure nor the number of staff has grown since 1975. At the
same time the number of students has increased by 65%. "8 In particular for the uni-
versities, the neglect by government has not only been financial, but also one of gov-
ernance. Since the institutional reforms of the late 1960s and early 1970s ended in
widespread disappointment, little has been done to install governance structures that
would enable them to tackle deficiencies in teaching and to adapt with speed and flex-
ibility to new developments in science and technology, in particular such develop-
ments that open up new connections between areas that previously were distinct and
separate from each other."9 There is a broad consensus that new structures are nec-
essary that give the universities more responsibility and at the same time increase
incentives for them to be efficient. A government report in 1988 stated bluntly that a
continuation of trends may pose a risk for the whole research system.120
About 14% of total national R&D is performed in the higher education sector (see
Table 4.7), down from about 20% in 1975. Relative to the size of their population,
Japan spends about 40% and the United States about 30% more than Germany (mea-
sured in purchasing power) on R&D performed in the higher education sector. Nev-
ertheless, in some areas German research is first rate, as indicated by the Nobel prizes
won by German scientists in the 1980s. In other areas, for example, clinical research,
there is a consensus that the situation is unsatisfactory.12'
For specific research projects funds are provided through the Deutsche For-
schungsgemeinschaft. In 1989 its budget was 1188 million DM, of which 61% was
financed by the Federal Government and 38% by the federal states. In addition to the
funding of research the Deutsche Forschungsgemeinschaft has the function of advising
government in scientific matters. Except for the setting of standards for toxic sub-
stances in occupational health and safety, this function remained marginal.122
Closely linked to the universities are the research activities of the Max-Planck-
Society (the successor to the Kaiser-Wilhelm-Society). In 1989 it comprised about 60
institutes and had a budget of 1079 million DM, of which one-half is financed by the
Federal Government and the other half by the federal states. The Max-Planck-Society
focuses on the natural sciences, on which it spends about 80% of its funds. Leading
scientists are normally recruited from universities. The institutes usually participate
in postgraduate education. Whereas its predecessor, the Kaiser-Wilhelm-Society, had
major activities in applied research, and included institutes such as leather research or
textile research, the Max-Planck-Society after World War II moved toward basic
research, which had brought it international recognition in previous decades.123
The creation of better links between industry and the higher education sector was
recognized in the 1980s by federal and state governments as a task for technology pol-
icy. State governments prodded universities to be more sensitive to the needs of
regional industry. Most universities and, in some regions, technical and commercial
colleges (Fachhochschulen) now have a special office for technology transfer.124
Technology Policy by Federal and State Governments
An official technology policy in the sense of a set of government policies designed to

support technical change and to guide its direction has existed at the federal level only
since the late 1960s, and at the state level only since the 1980s. Military R&D, which
in other countries, especially in the United States, after World War II became a vehicle
to support R&D far beyond the narrow boundaries of armaments, was forbidden in
West Germany by allied law until 1955. Compared to countries such as the United
States, Britain, or France, West Germany spends little on defense R&D: in 1989 it was
3023 million DM, or 13% of total government-financed R&D.
The Federal Ministry for Research and Technology (Bundesministerium fur For-
schung und Technologic), which is the main R&D spender in the Federal Government
(see Table 4.9), was created in 1955 as Ministry for Atomic Questions. In the 1950s
and early 1960s it had a precarious position between the Ministry for Economy on one
Table 4.9. Government Expenditure for Research and Development in 1989

Million DM 1981 = 100
Federal government 13,900 134
By department
Ministry for Research and Technology 7,695 130
Ministry for Defense 3,052 196
Ministry for Education and Science 1,120 127
Ministry for the Economy 992 93
By type of support
Direct 7,094 131
Indirect 420 91
Institutional 5,414 143
National laboratories 2,354 146
Departmental laboratories 936 146
Higher education 380 139
Deutsche Forschungsgemeinschaft 610 129
Max-Planck-Society 457 129
Fraunhofer-Society 155 186
International 1,234 174
By technology (selected areas)
Space 1,291 195
Nuclear (including fusion) 893 65
Nonnuclear energy 450 62
Information technology (including prod ction technology) 804 214
Environment and climate 741 173
Health 542 159
Materials 549 185
Aircraft 777 163
Biotechnology 246 248
State governments 9,100 132
Higher education 8,030
Other 1,070
Government total 23,000 133
Tax reduction (not included in total) 674
Source: Bundesminister fur Forschung und Technologie, Faktenberichl 1990 zum BundesberichtForschunt 1988. Bonn: Bun-
desminister fur Forschung und Technologie, 1990.
side, which as an advocate of the market order was skeptical about government
involvement in industrial innovation, and the federal states on the other side, which
under the West German constitution have the primary responsibility for scientific
research. However, it created a secure organizational domain for itself by founding
nuclear research laboratories in collaboration with federal statesat a time when the
federal nuclear research laboratories in the United States were already looking for new
tasks.125
In the public discussion of the 1960s Neo-Schumpeterians like Kenneth Gal-
braith and Neomarxists agreed that direct government support for industrial R&D is
a necessity, and this opinion was reinforced by the public discussion on the technology
gap between the United States and Europe.126 In 1962 the responsibility of the Atom
Ministry was expanded into space research and technology and its name, which pre-
viously had changed slightly several times, now became the Federal Ministry for Sci-
entific Research. A program for electronic data processing was begun in 1966, but only
in 1969, with the start of a program for new technologies, the ministry's responsibility
was broadened to technology in general (and its name then became Ministry for Edu-
cation and Science). Government support for civilian aircraft, begun in 1962,
remained under the Ministry for the Economy, and military R&D under the Ministry
of Defense.127 In 1972 the Ministry for Scientific Research was split up in two parts:
federal responsibilities for education including higher education remained in the Min-
istry for Education and Science, while technology and R&D outside the higher edu-
cation sector became the province of the Ministry for Research and Technology.128
Until the 1970s the Federal Ministry for Research and Technology and its pred-
ecessors were widely regarded as successful. West German manufacturers of nuclear
power plants, Siemens and AEG, managed to catch up (using American licenses) in
short time, and this was claimed by the ministry and generally accepted as confirming
the effectiveness of the ministry's subsidies.129 In the mid-1970s the ministry won
acceptance for the idea that government support for industrial R&D is a key in mod-
ernizing the structure of industry.130
However, as research by this author revealed, contrary to this image AEG and
Siemens developed their nuclear reactors mainly with their own funds; most govern-
ment funds were spent on reactor types that later were not used for commercial elec-
tricity generation, and the government laboratories, apart from some safety research
in the 1970s and 1980s, contributed very little to the development of those reactor
types that were commercially used.131 A study by A. D. Little about the ministry's sup-
port for computers and electronic data processing cast doubt on its effectiveness as far
as large computer systems were concerned.'32
In the 1980s the failure of some large projects, for example, the fast breeder reac-
tor and the high-temperature reactor, became obvious even to the general public. By
that time the ministry had begun to learn from its experience and to strive for more
effective policies.133 As this author has argued elsewhere, the failures in government
support to technology can be explained by the simple fact that if nearly all costs are
financed by government, firms have an incentive to carry out a project even if they
expect it to have no commercial use. This problem can be avoided by cost sharing
between government.134 The ministry has more and more adopted a strategy to finance
only a fraction of project cost, usually between 30 and 70%. With some exceptions,
such as space, projects financed totally by government are being phased out.
As another major policy innovation the ministry devised new programs and tai-
lored existing ones so as to strengthen cooperation among firms or between firms and
public research organizations. In selected areas such as microelectronics, robotics,
computer-aided design and manufacturing, or biotechnology, criteria for subsidies are
explicitly defined so as to encourage firms to participate in cooperative projects. Such
projects are useful to the firms less for the funds they provide than for their informa-
tional effects: by alerting firms to technological frontiers, by pooling precompetitive
technical information, and by creating links between suppliers and users of specialized
inputs. 135 Since 1978 the ministry has supported research contracted by firms to other
firms or to government laboratories. Annual expenditure was 28 million DM in 1989.
Another program subsidizes the costs of personnel seconded by firms to research orga-
nizations. Expenditure for this program in 1989 was 22 million DM.136
A genuine institutional innovation was the Fraunhofer-Society, which from
inauspicious beginnings in 1949 grew to a large organization carrying out applied
research mainly on contract with clients in industry and government.137 In 1989 it had
a budget of about 560 million DM, of which about 155 million was provided by the
federal government as institutional support that is made contingent on its success in
securing contracts. Six of its institutes work for the Defense Ministry and are totally
financed with public funds. The Fraunhofer-Society has close links to universities and,
being dependent on contracts, a strong orientation toward serving clients. Hence it
provides a link between universities and industry, and thus helped to reduce the gap
that opened in the German innovation system as the Max-Planck-Society moved
toward basic research.
Indirect government support included such measures as tax credits (674 million
DM in 1989), subsidies, and special depreciation rates for investments in R&D up to
a certain limit (449 million DM and 225 million DM, respectively, in 1989), and two
programs of subsidies for R&D personnel in small and intermediate enterprises. All
these programs were discontinued. Since 1983 the government experimented with
support for newly created technology-oriented enterprises. Expenditure was 53.5 mil-
lion DM in 1989. Although this subsidy helped to strengthen the infrastructure for risk
capital, the rather small demand for it suggested that the availability of risk capital was
not an important barrier to industrial innovation.138
Institutional support by the federal government is heavily concentrated on
national laboratories (2354 million DM in 1989) and departmental laboratories (936
million DM). Both types of organizations have grown since 1981 a little faster than
total government expenditure for R&D (see Table 4.9). The role of the departmental
research laboratories has not been controversial, as most of them carry out R&D for
noncontroversial state functions. Of the laboratories more or less directly related to
industrial innovation the major ones are the Physikalisch-TechnischeBundesanstalt,
successor to the Physikalisch-technische Reichsanstalt mentioned in the first section,
with an annual budget of 372 million DM, working on standards and measures; the
Bundesansta.ltfur Materialforschung undprufung, with an annual budget of 125 mil-
lion DM, working on materials and on safety aspects of chemicals; and the Bundesan-
staltfur Geowissemchaften und Rohstoffe, with an annual budget of 69 million DM,
working on geology and raw materials.
Since the 1970s there were attempts to diversify the work of the national labora-
tories into technologies outside nuclear power and to improve their links with indus-
trial technology. Some laboratories have increasingly taken on work in environmental
R&D and climate research in recent years. Schemes were set up to help the national
laboratories to spin off new firms and to improve their links with industry.
In the priority given by Federal Government to different areas of technology there
have been some notable shifts in recent years (see Table 4.9). The largest share now
goes to space technology. Funds for this area have been strongly increased recently as
a result of a controversial political decision to participate in new ambitious Western
European collaborative projects.139 The second field of government support is nuclear
technology, although the level of support for this technology was reduced by 35% since
1981. A similar reduction has been effected on nonnuclear energy technology.
Support for aircraft technology has grown by 63% since 1981 to an annual expen-
diture of 777 million DM. One of the largest projects is the Airbus family of passenger
airplanes developed and manufactured by a European consortium of firms. Although
Airbus can be regarded as a technical success, it has not yet crossed the threshold to a
normal commercial business, as the participating firms were skilfull to motivate their
governments to spend more and more money to enlarge the Airbus family by devel-
oping more and more types of airplanes. In a controversial move the West German
government has brought the German aircraft firms that so far depended on its subsi-
dies into a system of mixed ownership with management leadership by Daimler-Benz,
hoping that after the next round of subsidized projects the Airbus program will finally
become a normal commercial venture.140
Since the 1980s, federal states and some cities supported science parks to attract
new high-technology firms to their region or to facilitate the spin-off of new firms from
existing research organizations. Innovation centers were established providing space
and infrastructure facilities for new science-based firms. A study in 1989 found 50
innovation centers in operation and 50 more under construction or in the planning
stage. Average investment was 5.5 million DM.141 In addition some federal states have
taken initiatives to create new R&D organizations. The state of North-Rhine-West-
falia, for example, created a Wissenschaftszentrum with several decentralized research
institutes in areas such as "technology and work" and climate research.142
A System Fallen into Oblivion and Partly Recovered
If the German innovation system of the 1980s is put into historical perspective, one
gets the impression that it reflects mainly the momentum of organizations that have
existed for a long time, have survived the period of wars and crises, and since then have
grown in size. In the reconstruction period after World War II each organization pri-
marily looked after itself, and the system as such fell into oblivion. The federal states
cared for their educational institutions, including higher education. The Max-Planck-
Society (successor to the Kaiser- Wilhelm-Society) shed many of its activities in applied
research and focused on basic research. The federal government was hesitant in assum-
ing responsibilities for science and technology, and where it did, as in nuclear power,
aerospace, and electronic data processing, its programs for supporting industrial tech-
nology were ineffective. Eager to build a secure organizational domain, the Atom Min-
istry and its successors build up national laboratories that contributed little to indus-
trial technology. Many of the foundations had vanished in two inflations, and for new
ones to be created (for example, the Volkswagen Foundation) it needed new stimuli
such as the technology gap discussion. The old innovation system had been driven by
the dynamism of the Prussian bureaucracy that because of Prussia's preponderance
could force other federal states to follow suit. After World War II none of the West
German federal states had the size and dynamism to take on the leadership previously
provided by Prussia, and the new institutions that were built up to coordinate the tech-
nical and scientific efforts of the Federal and state governments, as, for example, the
Wissenschaftsrat and the Bund-Lander-Kommission fur Bildungsplanung und For-
schungsforderung, were slow to assume their responsibilities and catered only to parts
of the innovation system.143
It was only in the late 1970s and 1980s that the Federal Ministry for Research and
Technology began hesitantly to assume a role as manager of a national innovation
system, designed programs so as to strengthen cooperation and the flow of personnel
and information between different organizations within the system, and fostered new
institutions such as the Fraunhofer-Society that provided new links among different
components of the system. By that time, however, the splitting up of federal respon-
sibilities between the Federal Ministry for Research and Technology and the Federal
Ministry for Education and Science had created a new barrier for policymakers to con-
sider the system as a whole.
The fact that the system as a whole fell into oblivion is one of the reasons why
German industry today shows a strong technical capability in those areas where it has
a long tradition of technological strength. Where radically new areas of technology
emerged in the decades after World War II, as, for example, computers and micro-
electronics, or where because of the post-World War II policies of the allied countries
German industry had to start anew as in aircraft, industry developed less technological
dynamism (with the exception of nuclear power).144 In technologies where government
organizations play a key role as customers, such as telecommunications, German
industry has not regained its earlier technological dynamism.145
While the system was partly recovered from oblivion in the late 1970s and in the
1980s and first steps were taken by government to strengthen the links among its com-
ponents, Japan has caught up with Germany or overtaken it on nearly all counts that
make up for a strong national technological capability: business financed R&D as per-
centage of gross domestic product, patents held in the United States, scientists and
engineers in nonacademic jobs per 10,000 labor force, qualifications as percent of age
group in all levels of higher education, and percent of per capita gross domestic prod-
uct spent per student in higher education.
Challenges
Although West German industry has performed well in export markets in the 1980s
there is no reason for complacency. The fact that other countries with even less favor-
able endowments of natural resources such as Japan or Switzerland achieve a higher
per capita gross national product shows that Germany could perform still better. The
major part of exports are directed toward Western European countries. Automobile
exports, which are a major factor in West Germany's export performance, have ben-
efitted from trade protection against Japanese imports in other European countries.
In the face of Japan's large scientific and technological potential German firms will
face a strong competitor on the world market not only in selected industries such as
photographic equipment, consumer electronics, or semiconductors, but also in those
industries that so far have been traditional areas of German technological strength.
In the next few years, a good deal of the attention and energy of decision makers
in industry and government as well as of the financial resources over which they decide
will be devoted to reconstructing the economy in the Eastern part of the country and
to establishing ties with the emerging market economies in Eastern Europe. Given the
speed with which unification had to proceed because of the precarious economic sit-
uation in the former German Democratic Republic and because of the geopolitical
situation, it was unavoidable that the present system of technical innovation was
extended roughly as it is to the Eastern part of the country. There was no time to have
a general discussion and a detailed analysis of the system's strengths and weaknesses.
However, in the face of a new challenge from Japan over broad areas of technology,
some reforms in the German system for technological innovation cannot wait for long
if German industry is to hold its place on world markets.
One major challenge for reform is the higher education sector. Although there are
areas where scientific research is first rate and some minor reforms of postgraduate
education are on the way, the higher education sector is one of the weak components
in the country's innovation system.
Given the close relationship between the national capability for technical inno-
vation and the education system, there is a need for closer coordination of government
policies for technology with policies for education at various levels. The separation of
federal responsibilities in two ministries, one for Education and Science and one for
Research and Technology, does not appear to be helpful in meeting this need.
A further challenge is the increasing internationalization of business. The trend
for firms to locate different parts of their activities in different countries may have
slowed down for German industry a little because of the reconstruction in the Eastern
part of the country, but it will continue and may even accelerate again as German firms
participate in the reconstruction of the Eastern European economies. For the Federal
Government it means that its policies adopt, willingly or unwillingly, more and more
the character of regional policies designed to provide the infrastructure and support
systems that keep the country attractive for high-wage business activities.
Finally, there is the challenge of European unification. In the past the impact of
the policies of the European Communities on the West German innovation system
has been mainly in terms of project funds, which were administered in such a way as
to foster links among firms and research organizations in other European countries.
Although the funds were substantial for single projects, as, for example, for the JESSI
project in memory chips, they make up only a few percent of the total national R&D
and so far have not effected significant institutional changes within the nation's inno-
vation system.146 At present the combined impact of Community policies for techno-
logical innovation and of Community policies for a unified European market on the
German system for technical innovation is difficult to assess, but questions concerning
what form the emerging European innovation system will have, what the German
innovation system may contribute to it, and how it will have to adapt to it still remain
to be answered.
NOTES
1. This paper was written while the author was at the Science Center Berlin and later at the
European University Institute at Florence. The author thanks Richard Nelson, Nathan Rosen-
berg, and Peter Walther for comments.
2. World Bank (1989, pp. 6-9); Statistisches Bundesamt (1990, p. 236).
3. Injormationsdienst des Institute der deutschen Wirtschaft 17, 17(25 April 1991)p.5.
4. Schnabel (1934) is a classic account of German history in the nineteenth century that
gives due attention to science and technology. For economic history see Stolper et al. (1967),
Henderson (1975), Treue (1975), Borchardt (1976), Aubin and Zorn (1973-1976, Vol. 2), Hen-
ning( 1979), and Fischer (1985).
5. On List see Henderson (1984) and Schefold (1990).
6. Henderson (1954/1972; pp. 1-9, 139-166), Ritter(1961), Mieck (1965), Weber (1975),
Weber (1983), Seeling (1983), and Radkau (1989).
7. Ben-David (197l,Ch. 6).
8. Turner (1974) and McClelland (1980, Chapters 2 and 3).
9. Hufbauer(1982).
10. D. Pohl(1972).
11. Turner (1980) and Ringer (1969). The university system implied a redistribution favor-
ing the upper classes (Borchardt, 1965).
12. McClelland (1980, Chapters 4-6).
13. Schnabel (1934, Vol. 3) and Schmauderer(1976).
14. Cahan (1985) and Turner [in Jeismann and Lundgreen (1987, pp. 221-249)].
15. Ben-David (1971, pp. 186-192), and Forman etal. (1975).
16. Titze(1987, pp. 27-29) (student numbers), Pfetsch( 1974, pp. 85-88, 186) (budget fig-
ures), and W. Hoffmann et al. (1965, pp. 598-601) (price index for public consumption).
17. On specialization see Lundgreen (in Jarausch, 1983, pp. 149-179) and on natural sci-
ence departments Riese (1977, pp. 80-93).
18. See Ben-David (1971) and McClelland (1980).
19. See Brocke and Backhaus (both in Backhaus, in press).
20. Troitzsch(1966).
21. Lundgreen (in Sodan, 1988).
22. On the Gewerbeinstitut and the careers of some of its graduates see Henderson (1958,
Chapter 6).
23. Schnabel (1925), K6nig( 1981), and Gispen (1989).
24. Ludwig and K6nig( 1981).
25. Manegold(1970).
26. Kocka(1978,p. 313) and Fischer (1978, p. 87).
27. Fischer (1978, p. 88) and Konig (in Sodan, 1988, pp. 186-189).
28. Konig (in Sodan, 1988, pp. 183-213), Harney, Lundgreen, Schmiel, Treese (all in Jeis-
mann and Lundgreen, 1987), Gruner (1967), and Gispen (1989, Chapter 7).
29. Straatmann (in Jeismann and Lundgreen, 1987, pp. 271-281).
30. Fischer (in Aubin and Zorn, 1973-1976, Vol. 2, pp. 557-562) and Engelhardt (1984).
31. E. Hoffmann (1962) and Adelmann (1979).
32. An enthusiastic account is given by Locke (1984, Chapters 4-6).
33. Cipolla (1969) and Easterlin (1981).
34. A good contemporary description is Lexis (1904).
35. Fischer and Lundgreen (1975, p. 557).
36. Flora etal. (1983, pp. 553-663).
37. Kocka(1980, p. 96).
38. Locke (1985, p. 187).
39. Spath (in Riirup, 1979, Vol. 1, pp. 189-208) and Lundgreen (1990). For Britain see
Haines (1969) and Alter (1982); for France see Paul (1972) and Fox and Weisz (1980). In the
United States some of those initiating the graduate school believed to follow the German model,
but, as Ben-David (1971, Chapter 8) showed, in fact built something different.
40. Ben-David (1971, pp. 129-133).
41. Some examples are given by Schmauderer (1976).
42. Landes (1969, p. 346f.) argues that ideological consensus favored German industriali-
zation.
43. Pfetsch (1974, pp. 91-99), Lundgreen (1986), Lundgreen et al. (1986), and Lundgreen
(in Vierhaus and Brocke, 1990).
44. Griewank(1927, p. 23).
45. Cahan(1989,p. 196f.).
46. Pfetsch (1970), Bortfeldt et al (1987), and Cahan (1989).
47. Johnson (1990).
48. Burchardt (1975) and Brocke and Burchardt (both in Vierhaus and Brocke, 1990).
49. Brocke (in Vierhaus and Brocke 1990, p. 90).
50. Manegold(1970).
51. Lenoir(1992).
52. Brocke (in Vierhaus and Brocke, 1990, pp. 109-119); Lundgreen (in Riirup, 1979).
53. Fischer (1978, p. 75f.) and Miiller (in Liirmer, 1979, pp. 215-243).
54. W. Hoffmann et al. (1965, p. 522); exports of synthetic dyes from Beer (1959, p. 134)
(exchange rate 4 Marks/dollar).
55. Brocke (in Vierhaus and Brocke, 1990, p. 90).
56. D. Pohl (1972), Vershofen (1949-58), Haber (1971, p. 133f.), and Hertner (1986, pp.
115-118).
57. Beer(1959)andHaber(1958,pp. 126-136).
58. For the laboratory of Bayer see Beer (1959).
59. Haber (1971, p. 121) and Plumpe (1990, p. 52).
60. Ausschuss(1930, p. 8; 1932, p. 85). These data differ slightly from another source cited
by Haber (1971, p. 108).
61. Haber (1971, p. 14); see also Jeffrey (in Cocks and Jarausch 1990, pp. 123-142).
62. Timm(1974).
63. Troitzsch (1977, pp. 35-42).
64. Most recently Allen (1979) and Webb (1980).
65. Wengenroth(1986).
66. Fremdling(1977).
67. For printing presses see Porter (1990, pp. 180-195).
68. Ausschuss(1932,pp. 33, 85, 174).
69. Hughes (1983, Chapters).
70. Based on an estimate by the German electrical manufacturers association (reported by
Czada, 1969, pp. 136-147). Another estimate gives figures of 31 % for Germany and 35% for the
United States (see Hertner 1986, p. 125).
71. W. Hoffmann et al. (1965, p. 358).
72. Ausschuss(1932, pp. 85, 238).
73. Maddison(1982).
74. Pavitt and Soete( 1982).
75. Kabisch (1982), Braun (1983), Dunning (1983), Hertner (1986), and Schroter (1990).
76. Kindleberger(1975).
77. Ames and Rosenberg (1963) critically review this literature.
78. This view goes back to Veblen (1915) and Weber (1917/1980).
79. On German and international cartels see Maschke (1969), Cornish (1979), H. Pohl
(1985, 1988), Wurm (1989), and Fischer (in Aubin and Zorn, 1973-1976, Vol. 2, p. 811).
80. Feldenkirchen (1988).
81. Plumpe (1990).
82. Brady (1933).
83. Ausschuss(1932, pp. 85, 174, 176, 200, 237-239) and Czada (1969, p. 144). Chandler
(1990, chapters 12 and 14) provides case histories of firms in the 1920s.
84. See the list in Schreiber( 1923, pp. 13-15).
85. Schreiber(1923).
86. Zierold (1968, pp. 38-39), Nipperdey/Schmugge (1970), Diiwell (1971), Schroeder-
Gudehus (1974), Jarausch (1985), and Feldman (1987); government expenditure for science
from Pfetsch( 1982, p. 65).
87. Richter (1972, 1979), Forman (1973), H. Pohl (1983), and Feldmann (in Bruch and
Mtiller 1990).
88. Zierold (1968, p. 234), Treue (in Aubin and Zorn, 1976, Vol. 2, p. 114), and H. Pohl
(1983, p. 59).
89. Witt and Diiwell (both in Vierhaus and Brocke, 1990).
90. For the Kaiser-Wilhelm-Society see Witt (in Vierhaus and Brocke, 1990).
91. Haber (1971, p. 354) (for the chemical industry) and Erker (1990, p. 86) (for Siemens).
92. Plumpe (1990, pp. 471 -477 (on IG Farben) and Erker (1990, p. 86) (on Siemens).
93. On science and technology under National Socialism see Beyerchen (1977), Ludwig
(1974), Mertcns and Richter (1979), Troger (1986), and Jarausch (1986). On the emigration of
scientists and engineers see Fleming and Bailyn (1969) and Mock (1986). On the philosophical
and ideological traditions in science and technology preceding National Socialism see Ringer
(1969) and Herf (1984).
94. Lasby (1971) and Gimbel (1990); for aerodynamics see Hanle (1982).
95. Bentley(1984).
96. Stamm(1981)andOsietzki(1984).
97. For formal institutional details see Geimer and Geimer (1981), Massow (1983), and
Meyer-Krahmer (1992).
98. For the German trade union system good entries are Berghahn and Karsten (1989) and
Markovits (1986); for labor market policies Schettkat and Wagner (1990), Matzner and Wagner
(1990), and Soskice (1990); for the banking system Zysman (1984, Chapter 5).
99. All data are for 1988; see United Nations (1990).
100. According to Porter (1990) the top 50 industries (in terms of shares of total world
export) in 1985 accounted in Germany for 10% of total exports, 53% in Korea, 49% in
Japan, 42% in Switzerland, 34% in the United States, 30% in Sweden, 27% in Italy, and 18% in
Britain.
101. Law No. 25, Official Gazette of the Control Council for Germany No. 6 (30 April
1946), pp. 138-143.
102. Servan-Schreiber (1968), OECD (1968, 1970), Majer (1973), and Nussbaum (1983).
For a historical account of the public discussion see Krieger (1987).
103. BMFT (1989). Earlier studies on German foreign trade in high technology are sur-
veyed by Schmietow (1988, Chapters 3 and 4). For selected areas see Grupp and Legler (1987).
104. BMFT/BMBW(1988).
105. Hild(1989).
106. Vieweg(1991).
107. Narin and Olivastro (1987).
108. According to data collected by Ifo-Institut (Penzkofer and Schmalholz, 1990), R&D
accounted for 26% of total innovation expenditure in manufacturing industry in 1988.
109. All figures except for Switzerland are for the year 1987. In that year the German figure
was 1.82%. See BMFT (1990, p. 378sq.)
110. Wortmann (1991, p. 39). The data on R&D expenditure in Table 4.5 also include
expenditure abroad, so they cannot be directly related to national R&D capability.
111. Patel and Pavitt( 1989).
112. Echterhoffet al. (1990, pp. 20, 51), BMFT (1990, pp. 50, 362).
113. Deutsche Bundesbank (1990).
114. Tulder and Junne (1988) and Wortmann (1990).
115. Wortmann(1991).
116. Sorge (1991) and Blossfeld (1992).
117. National Science Board (1987, p. 227).
118. Wissenschaftsrat(1988, pp. 121-122,200-207,233-259).
119. On universities see Katzenstein (1987, Chapter 7) and Oehler(1989).
120. BMFT/BMBW(1988, p. 7).
121. Wissenschaftsrat (1988, p. 64); an account of accomplishments is "Bin Wissenschafts-
wunder?," The Economist (11 November 1989), 145-152; for bibliometric research see Daniel
and Fisch (1990); for clinical research Freund (1991).
122. Stamm (1981, pp. 109-140), Zierold (1968), Nipperdey and Schmugge (1980), and
Hartmann and Neidhardt (in Daniel and Fisch, 1990).
123. Vierhaus and Brocke (1990), Stamm (1981, pp. 85-108), Hohn and Schimank (1990,
Chapter 4), and BMFT (1990, p. 266).
124. Schimank (1988).
125. For the Karlsruhe nuclear research center see Keck (1981, Chapter 3); for a general-
ization of this view see Hohn and Schimank (1990, Chapter 7); for the American discussion see
Weinberg(1967).
126. Galbraith (1967/1974) and Hirsch (1974).
127. For government support to aerospace see Schulte-Hillen (1975).
128. For federal policies up to the 1970s see Braunling and Harmsen (1975), Keck (1976),
and Schmitzetal. (1976).
129. An example of the ministry's claim is BMWF(1965, p. 18); an example of its uncrit-
ical acceptance is in a report by the Commission on Monopolies (Monopolkommission, 1977,
p. 367).
130. Hauff and Scharpf (1975/77).
131. Keck (1981).
132. Sommerlatte et al. (1982).
133. Lorenzen(1985).
134. Keck (1988).
135. An interesting case history is Hausler et al. (1991).
136. BMFT (1988, p. 186; 1990, p. 50).
137. Hohn and Schimank (1990, Chapter 6).
138. BMFT (1988, pp. 189; 1990, pp. 50-51), Bundesregierung(1989, pp. 99-101, 164,
166), Meyer-Krahmer et al. (1983), "Fue-Personalkostenzuschuss-Programm," Deutsches Insti-
tutftir Wirtschaftsforschung Wochenbericht (8 March 1990), 119-122, and Kulicke and Krupp
(1987).
139. Humphreys (1989, pp. 147-150)and Weyer(1990).
140. On the aircraft industry see Hornschild and Neckermann (1989).
141. Steinberg (1989).
142. Ellwein and Bruder (1982), Bruder (1983), Gibb (1985), Dose and Drexler (1987),
Hilpert (1990), Hucke and Wollmann (1989), Jurgens and Krumbein (1991), Staudt (1987/88),
and Sabel et al. (1989). For a survey of the activities of federal states see BMFT (1988, pp. 201-
229; 1990, pp. 187-230).
143. On the early work of the Wissenschaftsrat see Berger (1974); on its later work see Block
and Krull (in Daniel and Fisch, 1990).
144. For microelectronics see Friebe and Gerybadze (1984) and Malerba (1985); for
nuclear power see Keck (1980 and 1981).
145. For telecommunications see Grupp (1991) and the chapter on Germany in Grupp
and Schnoring (1990-1991).
146. For European community policies in science and technology see Klodt et al. (1988)
and Starbatty and Vetterlein (1990).
REFERENCES erbsfdhigkeit der deutschen Industrie. December

7.
Adelmann, G. (1979). "Beruflichc Ausbildung und BMFT (Bundesminister fur Forschung und Tech-
Wciterbildung in der deutschen Wirtschaft 1871- nologic) (1990). Faktenbericht 1990zum Bundes-
1918." In H. Pohl (ed.), Berufliche Aus- und Wei- bericht Forschung 1988. Bonn.
terbildung der deutschen Wirtschaft seit dem 19. BMwF (Bundesminister fur wissenschaftliche For-
Jahrhundert, 9-52. Wiesbaden: Steiner. schung) (1965). Bundesbericht Forschung II.
Allen, R. C. (1979). "International Competition in (Bundestagsdrucksache V/2054).
Iron and Steel, 1850-1913." Journal of Economic Borchardt, K. (1965). "Zum Problem der Erzic-
///.ory39:911-38. hungs- und Aus-bildungsinvcstitionen im 19.
Alter, P. (1982). Wissenschafi, Staat, Mdzene. Stutt- Jahrhundert." In H. Aubin, E. Ennen, H. Kellen-
gart: Klett-Cotta. bcnz, T. Mayer, F. Metz, M. Miller, J. Schmithu-
Ames, E., and Rosenberg, N. (1963). "Changing sen (eds.), Beitrdge zur Wirtschafts- undStadtges-
Technological Leadership and Industrial chichte, 380-92. Wiesbaden: Steiner.
Growth." Economic Journal 73(3): 13-31. Borchardt, K. (1976). "Germany-," In C. Cipolla
Aubin, H., and Zorn, W. (eds.) (1973-1976). Hand- (ed.), The Fontana Economic History of Europe:
buck der deutschen Wirtschafts- und Sozialges- Vol. 4, Part 1. Brighton: Harvester, London:
chichte, Vol. 1. Stuttgart: Union Verlag, 1973, Barnes & Noble.
Vol. 2. Stuttgart: Klett, 1976. Bortfeldt, J., Hauser, W. and Rechenberg, H. (1987).
Ausschuss zur Untersuchung der Erzeugungs- und Forschen, Messen, Prufen. Weinheim: Physik-
Absatzbedingungen der deutschen Wirtschaft. Verlag.
(1930). Die deutsche chemische Industrie. Berlin: Brady, R. A. (1933). The Rationalization Movement
Mittler. in German Industry. Berkeley, CA: University of
Ausschuss zur Untersuchung der Erzeugungs- und California Press.
Absatzbedingungen der deutschen Wirtschaft Braun, H.-J. (1983). "Technolgietransfer im Mas-
(1932). Der deutsche A ussenhandel unter der Ein- chinenbau von Deutschland in die USA, 1870-
wirkung der weltwirtschaftlichen Strukturwan- 1939." Technikgeschichte 50(3), 238-52.
dlungen. Berlin: Mittler. Braunling, G., and Harmsen, D. M. (1975). DieFor-
Backhaus, J. G. (in press). The Economics of Science derungsprinzipien und Instrumente der For-
Policy. Berlin: Springer. schungs- und Technologiepolitik. Gottingen: Otto
Beer, J. J. (1959). The Emergence of the German Dye Schwartz.
Industry. Urbana, IL: University of Illinois Press. Bruch, R. vom, and Muller, R. A. (eds.) (1990). Far-
Ben-David, J. (1971). The Scientist's Role in Society. men ausserstaatlicher Wissenschajisforderung.
Englewood Cliffs, NJ: Prentice-Hall. Stuttgart: Steiner.
Bentley, R. (1984). Technological Change in the Bruder, W. (1983). "Innovation Behavior of Small
German Democratic Republic. Boulder: West- and Medium-Scale Firms." Research Policy 12,
view. 213-25.
Berger, R. (1974). Zur Slellung des Wissenschafts- Bundesregierung. (1989). Bericht fiber die Entwick-
rals bei der wissenschaftspolitischen Beralung von lungder Finanzhilfen desBundes und der Steuerv-
Bund und Ldndern. Baden-Baden: Nomos. ergiinstigungenfurdieJahre 1987bis 1990. (Bun-
Berghahn, V. R., and Karsten, D. (1989). Industrial destagsdrucksache 11/5116).
Relations in West Germany. Oxford: Berg. Burchardt, L. (1975). Wissenschaftspolitik im Wil-
Beyerchen, A. D. (1977). Scientists under Hitler. helminischen Deutschland. Gottingen: Vanden-
New Haven: Yale University Press. hoek & Ruprecht.
Blossfeld, H. -P. (1992). "Unterschiedliche Systeme Cahan, D. (1985). "The Institutionalization of Ger-
der Berufsausbildung und Anpassung an Struk- man Physics, 1865-1914." Historical Studies in
turveranderungen im internationalen Vergleich." the Physical and Biological Sciences 15(2): 1-65.
In D. Sadowski and A. Timmesfeld (eds.), Okon- Cahan, D. (1989). An Institute for an Empire. Cam-
omie und Tolitik beruflicher Bildung: Euro- bridge, England: Cambridge University Press.
pdische Entwicklungen 45-60. Berlin: Duncker & Chandler, A.D., Jr. (1990). Scale and Scope. Cam-
Humblot. bridge, MA: Harvard University Press.
BMBW/BMFT (Bundesministerium fur Bildung Cipolla, C. M. (1969). Literacy and Development in
und Wissenschaft/Bundesministerium fur For- the West. Harmondsworth: Penguin.
schung und Technologic) (1988). Leistungsstand Cornish, W. R. (1979). "Legal Control over Cartels
und Perspektiven von Wissenschafi, Forschung and Monopolization, 1880-1914: A Compari-
und Technologic in der Bundesrepublik Deutsch- son." In N. Horn and J. Kocka (eds.), Law and the
land. Bonn. Formation of the Big Enterprises in the 19th and
BMFT (Bundesminister fur Forschung und Tech- Early 20th Centuries, 280-305. Gottingen: Van-
nologie) (1988). Bundesbericht Forschung 1988. denhoek & Ruprecht.
Bonn. Cocks, G., and Jarausch, K. H. (eds.). (1990). Ger-
BMFT (Bundesminister fur Forschung und Tech- man Professions, 1800-1950. New York: Oxford
nologie) (1989). Zur technologischen Wettbew- University Press.
Czada, P. (1969). Die Berliner Elektrotechnische Technical Personnel." In C. Tilly (ed.), The For-
Industrie in dcr Weimarer Zeit. Berlin: Collo- mation of National States in Western Europe,
quium Verlag. 456-561. Princeton: Princeton University
Daniel, H.-D., and Fisch, R. (eds.) (1990). Sciento- Press.
metrics in the Federal Republic of Germany. (Spe- Fleming, D., and Bailyn, B. (1969). The Intellectual
cial Issue of Scientometrics Vol. 19, Nos. 5-6). Migration. Cambridge, MA: Harvard University
Deutsche Bundesbank. (1990). "Patent und Lizenz- Press.
verkehr mil dem Ausland sowie sonstiger Aus- Flora, P., Alber, J., Eichenberg, R., Kohl, J., Kraus,
tausch von technischem Wissen durch Dienstleis- F., Pfenning, W., and Seebohm, K. (1983-1987).
tungen in den Jahren 1988 und 1989." State, Economy, and Society in Western Europe
Monatsberichte der Deutschen Bundesbank 42(5): 1815-1975, 2 Vols. Frankfurt: Campus; London:
28-43. Macmillan; Chicago: St. James.
Dose, N., and Drexler, A. (eds.) (1987). Technologie- Forman, P. (1974). "Financial Support and Political
parks. Opladen: Westdeutscher-Verlag. Alignments of the German Physicists after First
Dunning, J. H. (1983). "Changes in the Level and World War." Minerva 10(2): 39-66.
Structure of International Production." In M. Forman, P., Heilbron, J., and Weart, S. (1975).
Casson (ed.), The Growth of International Busi- "Physics Circa 1900." Historical Studies of the
ness, 84-139. Winchester, MA: George Allen & Physical and Biological Sciences 5: 1-185.
Unwin. Fox, R., and Weisz, G. (eds.) (1980). The Organi-
Diiwell, K. (1971). "Staat und Wisserischaft in der zation of Science and Technology in France 1808-
Weimarer Epoche." In T. Schieder (ed.), Beitrdge 1914. Cambridge, England: Cambridge Univer-
zur Geschichte der Weimarer Repub/ik, 31-74 sity Press.
(Historische '/.eitschriji Beiheft 1). Miinchen: Old- Fremdling, R. (1977). "Railroads and German Eco-
enbourg. nomic Growth." Journal of Economic History
Easterlin, R. A. (1981). "Why Isn't the Whole World 37(3): 583-604.
Developed?" Journal oj Economic History 41 (1): Freund, H.-J. (1991). "Problemfeld klinische For-
1-19. schung." Forschung: Mitteilungen der DFG. 3(3):
EchterhofT-Severitt, H., Grenzmann, C, Marquardt, 30.
R., Menner, E., Weissburger, A., and Wudtke, J. Friebe, K. P., and Gerybadze, A. (eds.) (1984).
(1990). Forschung und Entwicklung in der Wirt- Microelectronics in Western Europe. Berlin: Erich
schafi 1987. Essen: SV-Gemeinniitzige Gesells- Schmidt.
chaft fur Wissenschaftsstatistik. Galbraith, J. K. (1967/1974). The New Industrial
Ellwein, T., and Bruder, W. (1982). Innovationso- State. Harmondsworth: Penguin (first edition
rientierte Regionalpolitik. Opladen: West- 1967, a German translation was published in
deutscher Verlag. 1967).
Engelhardt, U. (ed.) (1984). Handwerker in der Geimer, H., andGeimer, R. (1981). Research Orga-
Industrialisierung. Stuttgart: Klett-Cotta. nization and Science Promotion in the Federal
Erker, P. (1990). "Die Verwissenschaftlichung der Republic of Germany. Miinchen: Saur.
Industrie." Zeitschrift fur Unternehmensges- Gibb, J. M. (ed.) (1985). Science Parks and Innova-
chichte 35(2): 73-94. tion Centres. Amsterdam: Elsevier.
Feldenkirchen, W. (1988). "Concentration in Ger- Gimbel, J. (1990). Science, Technology, and Repa-
man Industry 1870-1939." In H. Pohl (ed.), The rations. Stanford: Stanford University Press.
Concentration Process in the Entrepreneurial Gispen, K. (1989). New Profession, Old Order. Cam-
Economy Since the Late 19th Century, 113-146. bridge, England: Cambridge University Press.
Stuttgart: Steiner. Griewank, K. (1927). Staat und Wissenschaft im
Feldman, G. DD. (1987). "The Politics of Wissen- Deutschen Reich. Freiburg: Herder.
schaftspolitik in Weimar Germany." In C. Maier Griiner, G. (1967). Die Entwicklung der hoheren
(ed.), Changing Boundaries of the Political: technischen Fachschulen im deutschen Sprachge-
Essays on the Evolving Balance Between the State biet. Braunschweig: Westermann.
and Society, Public and Private in Europe. 255- Grupp, H. (1991). "A Quantitative Assessment of
85. Cambridge, U K: Cambridge University Press. Innovation Dynamics and R&D Management in
Fischer, W. (1978). "The Role of Science and Tech- Japanese and West German Telecommunica-
nology in the Economic Development of Modern tions." R&D Management 21 (4):271 -90.
Germany." In W. Beranek and G. Ranis (eds.), Grupp, H., and Legler, H. (1987). Spitzentechnik,
Science, Technology and Economic Development, Gebrauchslechnik, Innovationspotential und
71-113. New York: Praeger. Preise. Cologne: TUV Rheinland.
Fischer, W. (1985). "Deutschland 1850-1914." In Grupp, H., and Schnoring, Th. (eds.) (1990-91).
W. Fischer (ed.), Handbuch der Europdischen Forschung und Entwicklung fur die Telekommu-
Wirtschafts- und Sozialgeschichle, Vol. 5, 357- nikation, 2 Vols. Berlin: Springer.
442. Stuttgart: Klett-Cotta. Haber, L. F. (1958). The Chemical Industry during
Fischer, W., and Lundgreen, P. (1975). "The the Nineteenth Century. Oxford: Oxford Univer-
Recruitment and Training of Administrative and sity Press.
Haber, L. F. (1971). The Chemical Industry, 1900- Innovation and Industrial Change." In S. Bulmer
1930. Oxford: Oxford University Press. (ed.), The Changing Agenda of West German Pub-
Haines, G. (1969). Essays on German Influence lic Policy, 128-54. Aldershot, England; Dart-
upon English Education and Science, 1850-1919. mouth; Brokkfield, VT: Gower.
Hamden, CT: Archon Books. Jarausch, K. H. (ed.) (1983). The Transformation of
Hanle, P. A. (1982). Bringing Aerodynamics to Higher Learning 1860-1930. Stuttgart: Klett.
America. Cambridge, MA: Harvard University Jarausch, K. H. (1985). "The Crisis of German Pro-
Press. fessions 1918-1933." Journal of Contemporary
HaurF, V., and Scharpf, F. W. (1975/77). Moderni- History 20: 379-98.
sierung der Volkswirlschaft, 2nd ed. Cologne: Jarausch, K. H. (1986). "The Perils of Professional-
Europaische Verlagsanstalt (first edition 1975). ism." German Studies Review 9: 107-37.
Hausler, J., Hohn, H.-W., and Liitz, S. (1991). "The Jeismann, K.-E., and Lundgreen, P. (eds.) (1987).
Architecture of an R&D Collaboration." Handbuch der deutschen Kildungsgeschichte, Vol.
Cologne: Max-Planck-Institut fur Gesellschafts- 3. Munchen: Beck.
forschung. Johnson, J. A. (1990). The Kaiser's Chemists.
Henderson, W. O. (1954/1972). Britain and Indus- Chapel Hill: University of North Carolina Press.
trial Europe, 3rd ed., 1972, (first edition 1954). JUrgens, U., and Krumbein, W. (eds.) (1991). Indus-
Leicester: Leicester University Press. triepolitische Strategien. Berlin: Edition Sigma.
Henderson, W. O. (1958). The State and the Indus- Kabisch, T. R. (1982). Deutsches Kapital in den
trial Revolution in Prussia, 1740-1870, Liver- USA. Stuttgart: Klett-Cotta.
pool: Liverpool University Press. Katzenstein, P. J. (1987). Policy and Politics in West
Henderson, W. O. (1975). The Rise oj German Germany. Philadelphia, PA: Temple University
Industrial Power, 1834-1914. London: Temple Press.
Smith. Keck, O. (1976). "West German Science Policy
Henderson, W. O. (1984). Friedrich List. London: Since the Early 1960's." Research Policy 5: 116-
Cass. 57.
Henning, F.-W. (1979). Die Industrialisierung in Keck, O. (1980). "Government Policy and Techni-
Deutschland, 5th ed. Paderborn. cal Choice in the West German Reactor Pro-
Herf, J. (1984). Reactionary Modernism, Cam- gramme." Research Policy 9: 302-56.
bridge, England: Cambridge University Press. Keck, O. (1981). Policymaking in a Nuclear Pro-
Hertner, P. (1986). "German Multinational Enter- gram. Lexington, MA: Lexington Books.
prise Before 1914." In P. Hertner and G. Jones Keck, O. (1988). "A Theory of White Elephants."
(eds.), Multinationals, 113-34. Aldershot, Research Policy 17:187-201.
England: Gower. Kindleberger, C. P. (1975). "Germany's Over-Tak-
Hild, R. (1989). "Exportboom in der Automobilin- ing of England." WeltwirtschaftlichesArchiv 111:
dustrie." ifo-Schnelldienst 42(25): 19-32. 253-81,477-504.
Hilpert, U. (ed.). (1990). Regional Innovation and Klodt, H., Horfmeyer, M., Krieger-Boden, C., and
Decentralization. London: Routledge. Soltwedel, R. (1988). Forschungspolitik unter EG-
Hirsch, J. (1974). "Scientific-Technical Progress and Kontrolle. Tubingen: Mohr.
the Political System." German Political Studies, Kocka, J. (1978). "Entrepreneurs and Managers in
Vol. 1. 107-39. Beverly Hills: Sage. German Industrialization." In T. Mathias and M.
Hoffmann, E. (1962). Zur Geschichte der Berufser- M. Tostan (eds.), The Cambridge Economic His-
ziehung in Deutschland. Bielefeld: Bertelsmann. tory of Europe, Vol. 7: The Industrial Economies:
Hoffmann, W. G., Grumbach, F., and Hesse, H. Capital, Labour, and Enterprise, Part 1, 492-589
(1965). Das Wachstum der deutschen Wirtschafl Cambridge, Engl.: Cambridge University Press.
seit der Mitte des 19. Jahrhunderts. Berlin: Kocka, J. (1980). "The Rise of Modern Industrial
Springer. Enterprise in Germany." In A. D. Chandler and
Hohn, H.-W., and Schimank, U. (1990). Konflikte H. Daems (eds.), Managerial Hierarchies, 77-
undGleichgewichteim Forschungssystem, Frank- 116. Cambridge, MA: Harvard University Press.
furt am Main: Campus. Konig, W. (1981). "Stand und Aufgaben der For-
Hornschild, K., and Neckermann, G. (1989). Die schung zur Geschichte der deutschen Polytechn-
deutsche Luft- und Raumfahrlindustrie. Frank- ischen Schulen und Technischen Hochschulen im
furt am Main: Campus. 19. Jahrhundert." Technikgeschichte 48(1): 47-
Hucke, J., and Wollmann, H. (eds.) (1989). Dezen- 67.
trale Technologiepolitik?Base\, Switzerland: Birk- Krieger, W. (1987). "Zur Geschichte von Technol-
hauser. ogiepolitik und Forschungsforderung in der Bun-
Hufbauer, K. (1982). The Formation of the German desrepublik Deutschland." Vierteljahreshefte fur
Chemical Community, 1720-1785. Berkeley, CA: Zeitgeschichte 35: 247-71.
University of California Press. Kulicke, M., and Krupp, H. (1987). "The Forma-
Hughes, T. P. (1983). Networks of Power. Baltimore, tion, Relevance and Public Promotion of New
MD: Johns Hopkins University Press. Technology Based Firms." Technovation 6:47-
Humphreys, P. (1989). "Policies for Technological 56.
Landes, D. (1969). The Unbound Prometheus. Cam- Matzner, E., and Wagner, M. (1990). The Employ-
bridge, England: Cambridge University Press. ment Impact of New Technology. Aldershot,
Larmer, K. (ed.) (1979). Studien zur Geschichte der England: Avebury/Gower.
Produktivkrdfte. Berlin: Akademie Verlag. McClelland, C.E. (1980). Stale, Society and Univer-
Lasby, C. G. (1971). Project Paperclip. New York: sity in Germany 1700-1914. Cambridge,
Atheneum. England: Cambridge University Press.
Lenoir, T. (1992). Politik im Tempel der Wissen- Mehrtens, H., and Richter, S. (eds.) (1979). Natur-
schaft. Frankfurt am Main: Campus. wissenschaft, Technik und NS-fdeologie. Frank-
Lexis, W. (ed.) (1904). Das Unterrichtswesen im furt am Main: Suhrkamp.
Deutschen Reich, 4 Vols. Berlin: Asher (also avail- Meyer-Krahmer, F. (1990). Science and Technology
able in English). in the Federal Republic of Germany. London:
Locke, R. R. (1984). The End of the Practical Man. Longman.
Greenwich, CT: JAI Press. Meyer-Krahmer, F., Gielow, G., and Kuntze, U.
Locke, R. R. (1985). "Relationship between Higher (1983). "Impacts of Government Incentives
Educational and Managerial Cultures in Britain Towards Industrial Innovation." Research Policy
and West Germany." In P. Joynt and M. Warner 12: 153-69.
(eds.), Managing in Different Cultures, 166-214. Mieck, I. (1965). Preussische Gewerbepo/itik in Ber-
Oslo: Universitetsforlaget. lin 1806-1844. Berlin: de Gruyter.
Lorenzen, H.-P. (1985). Effektive Forschungs- und Mock, W. (1986). Technische Intelligenz im Exil.
Technologiepolitik. Frankfurt am Main: Campus. Diisseldorf: VDI-Verlag.
Ludwig, K.-H. (1974). Technik und Ingenieure im Monopolkommission. (1977). Hauplgulachlen
Dritten Reich. Diisseldorf: Droste. 1973/1975. Baden-Baden: Nomos.
Ludwig, K. -H., and Konig, W. (eds.) (1981). Tech- Narin, F., and Olivastro, D. (1987). Identifying
nik, Ingenieure und Gesellschafl. Diisseldorf: Areas of Strength and Excellence in F.R.G. Tech-
VDI-Verlag. nology. Report to the Bundesministerium fur For-
Lundgreen, P. (1976). "Educational Expansion and schung und Technologic.
Economic Growth in Nineteenth-Century Ger- National Science Board (United States). (1987). Sci-
many." In L. Stone (ed.), Schooling and Society, ence & Engineering Indicators 1987. Washington,
20-66. Baltimore, MD: Johns Hopkins Univer- D.C.: U.S. Government Printing Office.
sity Press. Nipperdey, T., and Schmugge, L. (1970). 50 Jahre
Lundgreen, P. (1986). Standardization, Testing, Forschungsforderung in Deutschland. Bonn:
Regulation. Universitat Bielefeld, Forschungs- Deutsche Forschungsgemeinschaft.
schwerpunkt Wissenschaftsforschung. Nussbaum, B. (1983). The World After Oil. New
Lundgreen, P. (1990). "Engineering Education in York: Simon and Schuster (a German translation
Europe and the USA, 1750-1930." Annals of 'Sci- was published in 1984).
ence 47: 33-75. Oehler, C. (1989). Die Hochschulentwicklung in der
Lundgreen, P., Horn, B., Krohn, W., Kiippers, G., Bundesrepublik Deutschland seit 1945. Frankfurt
and Paslack, R. (1986). Staatliche Forschung in am Main: Campus.
Deutschland 1870-1980. Frankfurt am Main: Organisation for Economic Cooperation and Devel-
Campus. opment. (1968). Gaps in Technology: General
Maddison, A. (1982). Phases of Capitalist Develop- Report. Paris: OECD.
ment. Oxford: Oxford University Press. Organisation for Economic Cooperation and Devel-
Majer, H. (1973). Die "Technologische Liicke" opment. (1970). Gaps in Technology: Analytical
zwischen der Bundesrepublik Deutschland und Report. Paris: OECD.
den Vereiniglen Staaten von Amerika. Tubingen: Osietzki, M. (1984). Wissenschafisorganisation und
Mohr. Restauration. Cologne: Bohlau.
Malerba, F. (1985). The Semiconductor Business. Patel, P., and Pavitt, K. (1989). "A Comparison of
London: Pinter. Technological Activities in West Germany and
Manegold, K.-H. (1970). Universitat, Technische the United Kingdom." National Westminister
Hochschule und Industrie. Berlin: Duncker & Bank Quarterly Review May, 27-42.
Humblot. Paul, H. W. (1972). The Sorcerer's Apprentice.
Markovits, A. S. (1986). The Politics of the West Ger- Gainsville: University of Florida Press.
man Trade Unions. Cambridge, England: Cam- Pavitt, K., and Soete, L. (1982). "International Dif-
bridge University Press. ferences in Economic Growth and the Interna-
Maschke, E. (1969). "Outline of the History of Ger- tional Location of Innovation." In H. Giersch
man Cartels from 1873 to 1914." In F. Crouzet, (ed.), Emerging Technologies, 105-33. Tubingen:
W. H. Challnor, and W. M. Stern (eds.), Essays in Mohr.
European Economic History, 226-58. London: Penzkofer, H., and Schmalholz, H. (1990). "Zwan-
Arnold. zig Jahre Innovationsforschung im Ifo-Institut
Massow, V. von (1983). Organization and Promo- und zehn Jahre Ifo-Innovationstest." IFO-
tion of Science in the Federal Republic of Ger- Schnelldienst 43(14): 14-22.
many. Bonn: Inter-Nationes. Pfetsch, F. (1970). "Scientific Organisation and Sci-
ence Policy in Imperial Germany." Minerva 8(4): bewerbsfahigkeit der Bundesrepublik. Bad Hom-
556-80. burg: DIE Verlag H. Schafer.
Pfetsch, F. (1974). Zur Entwick/ung der Wissen- Schmitz, K., Riehle, R., Narr, W.-D., Koch, C., and
schaftspolitik in Deutschland, 1750-1914. Berlin: Albrecht, U. (1976). Der Staat und die Steuerung
Duncker & Humblot. der Wissenschaft. Gbttingen: Otto Schwartz.
Pfetsch, F. (1982). Dalenhandbuch zur Wissenschaf- Schnabel, F. (1925). "Die Anfange des technischen
Isentwicklung. Cologne: Zentrum fur historische Hochschulwesens." In Festschrift anliisslich des
Sozialforschung. 100-jdhrigen Bestehens der Technischen Hochs-
Plumpe, G. (1990). Die IG-Farbenindustrie AG: chule Fridericiana zu Karlsruhe, 1 -44. Karlsruhe:
Wirtschaft, Technik und Politik 1904-1945. Ber- Muller.
lin: Duncker & Humblot. Schnabel, F. (1929-1934). Deutsche Geschichte im
Pohl, D. (1972). Zur Geschichte der pharmazeu- 19. Jahrhundert, 3 Vols. Freiburg: Herder.
tischen Privatinstitute von 1779 bis 1873. Doc- Schreiber, C. (1923). Die Not der deutschen Wissen-
toral Dissertation, Marburg. schaft und der geistigen Arbeiter. Leipzig: Quelle
Pohl, H. (1983). "Die Forderung schulischer Ausbil- & Meyer.
dung und wissenschaftlicher Forschung durch die Schroeder-Gudehus, B. (1974). "The Argument for
deutsche Wirtschaft von 1918 bis 1945." In H. the Self-Government and Public Support of Sci-
Pohl (ed.), Wirtschaft, Schule und Universitdt, ence in Weimer Germany." Minerva 10: 537-70.
42-77. Wiesbaden: Steiner. Schroter, H. G. (1990). "Die Auslandsinvestitionen
Pohl, H. (ed.) (1985). Kartette und Kartellgesetzge- der deutschen chemischen Industrie 1870 bis
bungin Praxis und Rechtsprechung vom 19. Jahr- 1930." Zeitschrift fur Vnternehmensgeschichte
hundert bis zur Gegenwart. Stuttgart: Steiner. 35(1): 1-22.
Pohl, H. (ed.) (1988). Wettbewerbsbeschrdnkungen Schulte-Hillen, J. (1975). Die Luft- und Raumfahrt-
auf internationalen Mdrkter. Stuttgart: Steiner. politik der Bundesrepublik Deutschland. Gottin-
Porter, M. E. (1990). The Competitive Advantage of gen: Otto Schwartz.
Nations. London: Macmillan. Seeling, H. (1983). Wallonische Industrie-Pioniere
Radkau, J. (1989). Technik in Deutschland. Frank- in Deutschland. Liittich: Wahle.
furt am Main: Suhrkamp. Servan-Schreiber, J.-J. (1968). Die amerikanische
Richter, S. (1972). Forschungsjbrderung in Deutsch- Herausforderung. (German translation of Le Deft
land, 1920-1936. Dusseldorf: VDI-Verlag. Americain), Hamburg: Hoffmann und Kampe.
Richter, S. (1979). "Wirtschaft und Forschung." Sodan, G. (ed.) (1988). Die Technische Fachhochs-
Technikgeschichte 46: 20-44. chule Berlin im Spektrum Berliner Bildungsges-
Riese, R.(1977). DieHochschuleaufdem Wegezum chichte. 1-44. Berlin: Technische Fachhochschule
wissenschaftlichen Grossbetrieb. Stuttgart: KJett. Berlin.
Ringer, F. (1969). The Decline of the German Man- Sommerlatte, T., Beckerhoff, D., Walsh, I., Stelter,
darins. Cambridge, MA: Harvard University K., Klekotto, I. (1982). Die Entwicklung der
Press. Dalenverarbeitung in der Bundesrepublik
Ritter, U. P. (1961). Die Rolle des Staates in den Deutschland. Report BMFT-FB-DV 82-004,
Friihstadien der Industrialisierung. Berlin: Bundesministerium fur Forschung und Technol-
Duncker & Humblot. ogic.
Riirup, R. (ed.) (1979). Wissenschaft und Gesells- Sorge, A. (1991). "Strategic Fit and the Societal
chaft, 2 Vols. Berlin: Springer. Effect: Interpreting Cross-National Comparisons
Sabel, C. F., Herrigel, G. B., Deeg, R., and Kazis, R. of Technology, Organization and Human
(1989). "Regional Prosperities Compared: Mas- Resources." Organization Studies 12(2): 161-90.
sachusetts and Baden-Wurttemberg in the Soskice, D. (1990). "Reinterpretating Corporatism
1980s." Economy and Society 18(4): 374-404. and Explaining Unemployment." In R. Brunetta
Schefold, B. (ed.) (1990). Studien zur Enlwicklung and C. Dell'Aringa (eds.), Labour Relations and
der okonomischen Theorie 10. Berlin: Duncker & Economic Performance. 170-211, London: Mac-
Humblot. millan.
Schettkat, R., and Wagner, M. (eds.) (1990). Tech- Stamm, T. (1981). Zwischen Staat und Selbstver-
nologischer Wandel und Beschaftigung. Berlin: de waltung. Cologne: Verlag Wissenschaft und Poli-
Gruyter. tik.
Schimank, U. (1988). "The Contribution of Univer- Starbatty, J., and Vetterlein, U. (1990). Die Tech-
sity Research to the Technological Innovation of nologiepolilik der Europaischer Gemeinschaft.
the German Economy." Research Policy 17: 329- Baden-Baden: Nomos.
40. Statistisches Bundesamt. (1990). Statistisches Jahr-
Schmauderer, E. (1976). "Die Stellung des Wissen- buch fur das Ausland 1990. Stuttgart: Metzler-
schaftlers zwischen chemischer Forschung und Poeschel.
chemischer Industrie im 19. Jahrhundert." In W. Staudt, E. (1987/88). "Technologic- und Regional-
TReue und K. Mauel (eds.), Naturwissenschaft, politik der Lander." List-Forum 14: 93-110.
Technik und Wirtschaft im 19. Jahrhundert, Vol. Steinberg, R. (1989). "Innovation Centres and Their
2, 614-53. Gottirgen: Vandenhoek & Ruprecht. Importance for the Growth of New Technology-
Schmietow, E. A. (1988). Die technologische Wett- Based Firms." Technovalion 9: 681-94.
Stolper, G., Hauser, K., and Borchardt, K. (1967). Vieweg, H.-G. (1991). "Der Werkzeugmaschinen-
The German Economy: 1870 to the Present. Lon- bau." IFO-Schnelldienst 44(21): 22-9.
don: Weidenfeld & Nicholson. Webb, S. B. (1980). "Tariffs, Cartels, Technology,
Timra, A. (1974). Bergbau und Wissenschafi. Dort- and Growth in the German Steel Industry, 1879-
mund: Gesellschaft fur Westfalische Wirtschafts- 1914." Journal of Economic History 40(2): 309-
geschichte. 29.
Titze, H. (1987). Das Hochschulstudium inPreussen Weber, M. (1917/1980). "Parliament and Democ-
und Deutschland 1820-1944. Gottingen: Van- racy in Germany." In J. Winckelmann (ed.). Max
denhoek & Ruprecht. Weber: Gesammelte politische Schriften. 4th ed.
Treue, W. (1975). Gesellschaft, Wirtschafi und Tech- Tubingen: Mohr.
nik Deulschlands im 19. Jahrhundert. Milnchen: Weber, W. (1975). "Industriespionage als technolo-
Deutscher Taschenbuch Verlag. gischer Transfer in der Friihindustrialisierung
Troger, J. (ed.)(1986). Hochschule und Wissenscha.fi Deutschlands." Technikgeschichte 42: 287-305.
im Dritten Reich. Frankfurt am Main: Campus. Weber, W. (1983). "Preussische Transferpolitik,
Troitzsch, U., (1966). Ansdtze technologischen Den- 1780-1820." Technikgeschichte SO: 181-96.
kens bei den Kameralisten des 17. und 18. Jahr- Weinberg, A. M. (1967). Reflections on Big Science.
hunderts. Berlin: Duncker & Humblot. Cambridge, MA: MIT Press.
Troitzsch, U. (1977). Innovation, Organisation und Wengenroth, U. (1986). Unternehmensstrategien
Wissenschafi beim Aufbau von Hiittenwerken im und technischer Fortschritt. Gottingen: Vanden-
Ruhrgebiet, 1850-1870. Dortmund: Gesellschaft hoek & Ruprecht.
fur Westfalische Wirtschaftsgeschichte. Weyer, J. (1990). "Strategies for the Social Construc-
Tulder, R. van, and Junne, G. (1988). European tion of Technology." In H. K.rupp(ed.), Technik-
Multinationals in Core Technologies. Chichester politik angesichts der Umweltkatastrophe, 250-
and New York: John Wiley. 61. Heidelberg: Physica.
Turner, S. (1974). "University Reformers and Pro- Wissenschaftsrat (1988). Empfehlungen des Wissen-
fessorial Scholarship in Germany, 1760-1806." schaftsrates zu den Perspektiven der Hochschulen
In L. Stone (ed.), The University in Society, Vol. in den 90er Jahren. Cologne: Wissenschaftsrat.
2, 495-531. Princeton: Princeton University World Bank. (1989). The World Bank Atlas 1989.
Press. Washington, D.C.: The World Bank.
Turner, S. (1980). "The Prussian Universities and Wortmann, M. (1990). "Multinationals and the
the Concept of Research." Internationales Archiv Internationalization of R&D." Research Policy
fur Sozialgeschichte der deutschen Literatur 5: 19: 175-83.
68-93. Wortmann, M. (1991). Country Study on the Fed-
United Nations. (1990). International Trade Statis- eral Republic of Germany. Commission of the
tical Yearbook. New York: United Nations. European Communities, Monitor-Fast Program,
Veblen, T. (1915). Imperial Germany and the Indus- Prospective Dossier No. 2: "Globalization of
trial Revolution. London and New York: Mac- Economy and Technology," Vol. 17. Brussels:
millan. Commission of the European Communities.
Vershofen, W. (1949-1958). Die Anfange der chem- Wurm, C. A. (ed.) (1989). Internationale Kartelle
isch-pharmazeutischen Industrie, 3 Vols. Berlin: und Aussenpolilik. Stuttgart: Steiner.
Deutscher Betriebswirte-Verlag; Aulendorf: Can- Zierold, K. (1968). Forschungsforderung in Drei
tor. Eochen. Wiesbaden: Steiner.
Vierhaus, R., and vom Brocke, B. (eds.) (1990). For- Zysman, J. (1983). Governments, Markets, and
schung im Spannungsfeld von Politik und Gesells- Growth. Ithaca, NY: Cornell University Press.
chaft. Stuttgart: Deutsche Verlagsanstalt.
5
National Innovation Systems: Britain
WILLIAM WALKER
There are two great puzzles in Britain's economic history. The first is why this com-
paratively small country on the northwest fringe of the European Continent became
the hub of the eighteenth- and nineteenth-century industrial revolutions, and domi-
nated the international economy over so long a period (the industrial supremacy of
the United States in this century seems short-lived in comparison). The second is why
Britain's industrial leadership begin to ebb away in the last decades of the nineteenth
century, and why the decline that followed was so prolonged, continuous, and seem-
ingly irreversible.
The period of apparent discontinuity in the late nineteenth and early twentieth
centuries, when the long retreat began, has gained particular attention from historians.
Explanations of the change in economic fortunes are of three main kinds. The first is
that the culture and institutions that sustained industrial development in the period
of expansion proved inappropriate to the new industries that emerged in the 1880s and
1890s, and which underpinned economic advance in much of the twentieth century
(Landes, 1969). As Nelson and Rosenberg described in Chapter 1, the chemical and
electrical industries required a greater and more systematic engagement in science and
education, and the automobile industry a more scientific approach to industrial man-
agement, than had hitherto been practiced. The diffusion of new techniques was also
inhibited by attachment to the oldby what historians have referred to as the "dis-
advantages of being first" to experience the industrial revolution.
The second explanation is that during the rise and fall of dominant nations, suc-
cess initially breeds success, but within a few generations success becomes a source of
failure. In the British case, resources became overextended as the Empire grew, middle
class culture turned against industrial enterprise, and a rentier mentality took hold
(Hobsbawm, 1987). In time, failure developed its own pathology. Industries became
obsessed with defending rather than expanding their territories, the power of organized
labor increased as managerial authority and competence weakened, and international
opportunities were narrowed by the country's loss of nerve. During the twentieth cen-
tury, Britain also escaped the traumas of invasion or defeat in war, so that its social
fabric was less disturbed than in other European countries or in Japan, allowing greater
institutional continuity and thus inertia.
The third explanation is that the spread of industrialization, especially in Europe
and North America, was bound to undermine Britain's economic hegemony, and that
158
NATIONAL INNOVATION SYSTEMS: BRITAIN 159
conditions unusually favorable to Britain in the eighteenth and nineteenth centuries

no longer applied in the twentieth century. For instance, the North Atlantic trading
system lost some of its former importance, and the railways that Britain pioneered
brought great improvements in transport, allowing land-locked regions of Europe and
the United States to be industrialized and united politically (Mackinder, 1904). The
advantages that came from being a maritime nation were thereby diminished. More-
over, by exporting its capital and technology, and by maintaining an open interna-
tional trading system, Britain helped other nations to challenge its supremacy, as did
the United States in the twentieth century (Stein, 1984).
For a long time, it was not generally accepted that Britain had suffered a serious
loss of economic vitality. In the first half of this century, decline was to some degree
masked by the military defeat of Britain's principal European rival, Germany, and by
the economic misery experienced by the United States and other industrial countries
during the Great Depression. Imperial markets also provided foreign income and a
relatively safe haven between the Wars (Svennilson, 1954). The weakness of Britain's
international position really became apparent in the 1950s and 1960s, politically
through the inability to hold the Empire together, and economically through the per-
sistent failure to match the growth rates of other industrial countries, or to stem the
loss of trade shares in domestic and foreign markets.
In the postwar era, the 1960s and 1980s stand out as the decades in which the
most determined and coherent efforts were made to halt the decline, with attention
focused on what had come to be regarded as systemic defects in the economy and its
management. However, the approaches taken could hardly have been more different.
During the 1960s, the guiding assumption, whether under governments of left or right,
was that the market economy could no longer be left to its own devices. The state had
to intervene financially and in other ways to increase investment and improve indus-
trial management, to ensure that economies of scale were realized and inventions
turned into successful innovations, and to redress the inequities of income and oppor-
tunity that were seen as inherent to the capitalist system.
In stark contrast, the guiding assumption in the 1980s and 1990s (so far) has been
that the market economy must be left to its own devices, and that Britain's economic
deficiencies have stemmed in large part from the state's creeping protection of indi-
viduals, firms, and sectional interests, and from delusions about its managerial powers.
The aim has therefore been to restore the spirit of "free enterprise" and "individual
responsibility," giving the market free rein wherever possible to decide the allocation
of resources. Hence the privatization of state monopolies, the attack on trade union-
ism, the shift away from direct taxation, and a whole range of measures and induce-
ments that came to be known under the rubric "Thatcherism."
In the 1960s and 1980s, the ways in which Britain's "system of innovation" were
conceived, and the roles ascribed to it in economic development, were thus very dif-
ferent. Broadly speaking, the first, implicitly Schumpeterian, of the above three expla-
nations of economic decline held sway in the 1960s. The managed restructuring of the
industrial economy, the allocation of increased resources to R&D and education, and
the adoption of more systematic approaches to the management of industrial and tech-
nological activity were seen as the route to recovery.
In the 1980s, the second explanation carried greater weight. The government's
deepest belief, romantic and behavioral rather than managerialist, was that the indi-
vidual energies released by the emancipation from past constraints could alone drive
economic modernization. Technical and other forms of advance would follow, but
could not precede, the revolution in social attitudes and behavior. Restoring faith in
free market capitalism, and rooting out the culture of failure, were seen as precondi-
tions for economic and thus technological recovery.
In some respects, Britain's economic performance in the 1980s relative to other
OECD countries showed an improvement over the 1960s and 1970s. Despite the mac-
roeconomic troubles with which the decade began and ended, productivity rose
strongly over much of the period, and in some areas, particularly in services, the old
dynamism seemed to be returning. There were few signs, however, of a reversal of the
historic decline in Britain's abilities to establish new technological capacities. Whereas
other countries' spending on research and development (R&D) increased, Britain's
stagnated; the share of world technological output, as measured by patents, continued
to decline; and the science and education systems were beset with problems. The
growth of high technology activity that did occur was substantially the result of the
expansion of defense expenditure, and of U. S. and Japanese inward investment in
electronics and other fields. Only in chemicals and Pharmaceuticals and in a few engi-
neering niches (aeroengines being a significant example) could it be claimed that Brit-
ain maintained its position among the leaders in the development of civil technology.
In the 1960s, industry and government were still acting in the shadow of the
Empire. It was taken for granted that Britain should remain at the forefront of tech-
nology. This is not the case today. The first thing to emphasize about Britain's contem-
porary innovation system is therefore that its development, whether by industry or the
state, has become a relatively low priority. This may be the natural economic behavior
of a country that now has relatively low income levels and needs to catch up with inter-
national best practice, and whose manufacturing companies no longer match the scale
or sophistication of their main foreign rivals. But it is also the consequence of the
greater dynamism in Britain of services and other activities, and of a prevailing eco-
nomic culture, even ideology, which has come to place quick gains before the patient,
long-term development of industrial capabilities.
BRITAIN'S ECONOMIC STRUCTURE: STRENGTH OUTSIDE MANUFACTURING
Britain has a population of 57 million people. Its economic output in 1988 was
360bn ($580bn at the exchange rate then current). In both population and GDP, it
is closest to France and Italy among the industrial nations. Britain's economic struc-
ture and praxis are, however, distinct from those of its European neighbors (Holland
provides perhaps the nearest European equivalent). In several respects, Britain has
most in common with the United States, despite the great disparity in size: the strength
of resource-based industries, the scale and functioning of its capital markets, the heavy
commitments to defense production, and the attachment to individualism and to lib-
eral economic ideals.
The structural differences between the British and Continental European econ-
omies are most noticable outside manufacturingin agriculture, energy, and trade-
able servicesand in areas of manufacturing that are linked to resource trading (e.g.,
food processing, petrochemicals). It is in these areas that Britain's position in inter-
national commerce is strongest. As a result, the priority given to advanced manufac-

turing and thus to technological development has often been less evident in Britain
than in other European countries.
Taking these other sectors in turn, agriculture forms a small, if generally efficient,
part of the British economy. In 1989, it accounted for 1.5% of GDP and 1.3% of
employment (see Table 5.1). Unlike Italy and France, Britain remains a net importer
of agricultural produce, and with the exception of whisky, is not a significant producer
of high-value items such as wine and cheese.
This nevertheless understates Britain's role in international commerce in food-
stuffs. It is home to some of the world's largest food, drink, and tobacco companies
whose origins can be traced back to the eighteenth and nineteenth centuries (British
American Tobacco, Unilever, Rank Hovis McDougall, Tate & Lyle, Cadbury
Schweppes). These firms are today highly diversified and multinational. Of the 29 Brit-
ish firms in the top 200 non-U. S. industrial companies in 1987, no fewer than 10
belonged in this category (Germany had none). Their combined sales were $71 bn (see
Table 5.2).
Where energy is concerned, Britain has the most extensive primary energy
resources in Europe. It is still a major if declining coal producer, and during the 1970s
and 1980s it became Europe's largest producer of oil, and its second largest producer
of natural gas (after the Netherlands). This accounts for the substantial rise evident in
Table 5.1 in the energy industries' share of national output in the late 1970s and early
1980s.
Outside the coal industry, the firms operating in this area are again highly mul-
tinational. BP and Shell are global actors and the largest firms based outside the United
States, with combined sales of $144bn in 1987. Like RTZ, which is now the world's
largest minerals producer and trader, they owe their positions partly to the territories,
capabilities, and linkages acquired before the Empire was dismantled.
Table 5.1. Structural Change in the British Economy, 1979-1989

GDP Per
GDP (%) Employment (%) Employee
1979-1989
1979 1983 1989 1979 1983 1989 (Constant Prices)
Agriculture 2.1 2.0 1.5 1.6 1.7 1.3 172
Energy, water 8.0 11.5 5.2 3.1 3.1 2.1 230
Manufacturing 28.2 23.5 22.2 31.3 26.2 23.0 245
Construction 6.1 6.0 6.9 5.3 5.0 4.7 251
Distribution, hotels 13.2 12.7 14.2 18.4 19.5 20.3 190
Transport, 7.6 7.0 6.9 6.4 6.4 6.0 180
communications
Banking, finance 11.5 13.6 19.7 7.1 8.9 11.8 131
insurance"
Other services 23.3 23.7 23.4 26.7 29.2 30.9 165
Total 100 100 100 100 100 100 206
"Excludes leasing.
Source: United Kingdom National Accounts, 1990 Edition. London: Central Statistical Office, 1990.
Table 5.2. Number (and Sales) of Top 500 Non-U.S. Industrial Companies, by Selected
Sectors, 1987
Federal Republic
United Kingdom of Germany France Japan
No. ($bn) No. ($bn) No. ($bn) No. ($bn)
Principal manufacturing sectors 32(123.6) 40(314.4) 22(160.4) 92(513.1)
Chemicals, etc. 5(29.8) 9(77.9) 5(25.0) 22(69.5)
Electricals, electronics 8(33.3) 6(40.4) 6(41.9) 28(188.9)
Metals, mechanical engineering 10(25.2) 16(79.0) 4(34.1) 23(101.3)
Motor vehicles, aerospace 9(35.3) 9(117.1) 7(59.4) 19(153.4)
Selected other production sectors 32(213.3) 7(33.8) 15(65.3) 28(98.6)
Mining, petroleum" 10(112.5) 7(33.8) 6(43.8) 8(43.7)
Food, drink, tobacco" 22(100.8) 0 9(21.5) 20(54.9)
Total, selected industries 64(336.9) 47(348.2) 37(225.7) 120(611.7)
"Fifty percent of the sales of the Anglo-Dutch firms, Shell and Unilever, have been attributed to the United Kingdom.
Source: Derived from "The International 500." Fortune, August 1988.
With regard to services, the formation of a dynamic international capital market,

based in the City of London, played an important part in Britain's rise to economic
dominance in the nineteenth century. Today, Britain's position in international
finance and insurance remains out of all proportion to the size of its productive econ-
omy. Following sterling's decline as an international currency in the 1950s and 1960s,
the City of London retained its role as a leading financial center partly by gaining com-
mand over the burgeoning Eurodollar market, and partly by becoming a major player
in the recycling of OPEC's windfall gains after the oil price rises in the 1970s. During
the 1980s, the City of London has expanded at a prodigious rate, due to its financing
of economic expansion in Europe and elsewhere, to its invasion by U. S. and Japanese
financial companies, and to its increasing technological linkage to the New York and
Tokyo capital markets. Many of the services offered by the City of London may be
classed as producer services, but its clientele is international. Unlike Tokyo or Frank-
furt, its recent expansion has not reflected the underlying dynamism of the domestic
production base.
This is not the only area of services in which Britain has a significant international
presence. In hotels (Grand Metropolitan, Trusthouses Forte), retail (Marks & Spencer,
Sainsbury), air transport (British Airways), and advertising, publishing, and property,
British firms have been expanding at home and abroad. They have, for instance, been
among the most aggressive purchasers of assets in the United States during the 1980s,
to the extent that Britain has recently surpassed Japan as the U. S.'s largest foreign
investor.
Before moving on to manufacturing, three implications are worth drawing here.
First, the above areas have increasingly come to form the heart of the British economy.
The shift toward services is, of course, an international trend, but in Britain it has been
especially pronounced. Manufacturing comprised less than one-quarter of national
output in 1985, against one-third in 1960. Only the United States among the advanced
countries has a lower share of output coming from manufacturing.
The increasing concentration of resources in the tradeable service sector should
be seen in the British case as a sociopolitical as well as an economic phenomenon.

Employment there carries greater prestige than in manufacturing, and is generally
more lucrative, so that it acts as a magnet to the social elite, whose contours it increas-
ingly defines. The leaning toward services is reflected in the composition of the gov-
ernment that has held power since 1979, with its domination by people whose careers
have been made in services such as retail, property, finance, and law. As a conse-
quence, there is a natural tendency among policy-makers, for instance with regard to
the interest and exchange rates, to place the demands of the tradeable service sector
above those of manufacturing. By the same token, the political administration often
betrays limited understanding of what it takes to be a successful manufacturer in the
contemporary international economy. Britain is in these respects at the opposite pole
to Germany and Japan.
Second, the norms of managerial behavior in British manufacturing have become
strongly influenced by practices in these other areas, and particularly by the nature and
power of British capital markets. To an exceptional degree among European countries,
short-run profit maximization and asset trading have become the primary objective of
the business manager (Ingham, 1984; Lazonick, 1990). This may be attributed, inter
alia, to the heavy reliance on stock market finance, and to the existence of a large sec-
ondary market in issued shares, which exerts pressure on companies to maximize the
income of stakeholders; in the 1980s, to the use of monetary instruments, and specif-
ically high real interest rates, as primary tools of macroeconomic policy; to the com-
paratively slight involvement of banks in industrial finance and decision making, and
to their unwillingness to shield companies from the vagaries of financial markets; and
to the ease of corporate takeover in Britain, as in the United States, which again
encourages high profits and dividends to maintain stock ratings and thus the selling
price of companies. The relatively open market in corporate assets has in turn fed the
well-documented tendency among large firms to give growth by acquisition and
merger a higher priority than organic growth, and has encouraged the "financial engi-
neering" practiced by conglomerates such as the Hanson Trust and BTR, and even by
some of Britain's largest high technology companies (GEC being a notable example).
The incentive structure thereby created helps explain why British manufacturing
companies repeatedly display higher profits than their European and Japanese coun-
terparts, while tending to invest less in fixed capital and R&D. This was well demon-
strated in the recent takeover of the electronics firm, Plessey, by GEC and Siemens.
International comparisons showed that GEC and Plessey had been consistently more
profitable than their international rivals, and certainly more so than Siemens, but had
equally consistently slipped down the league table of electronics companies measured
by size and market share (Morgan et al., 1989). A recent consultant's report proudly
announced that Britain's large firms were "the best in Europe," since they occupied 6
of the top 10, and 28 out of the top 50 positions when ranked according to a bundle
of profitability measures (Sunday Times, 22 April 1990). There were only two German
companies on the list.
Third, these nonmanufacturing sectors are predominantly technology users,
rather than technology producers. Moreover, with the advent of information technol-
ogy, they have become heavy and often highly sophisticated technology users. In 1987,
the service sector accounted for one-half of British investment in plant and machinery,
against one-third in 1977. In a number of contexts (e.g., bank cash dispensers, value-
added networks (VANs), stock control systems, mobile telephones), Britain leads
other European countries in the application of new technologies. This raises the ques-
tion of whether the growth of investment in new technology in these sectors has, by
pulling on the supply side, led to the establishment of new technological capabilities
and to a creative interplay between users and producers in Britain.
Unfortunately, there have been few studies of this issue. Some indigenous capa-
bilities have undoubtedly taken root (for instance, in relation to systems software and
mobile telephones), leading to new competitive advantages. However, the available
evidence suggests that this has been the exception rather than the rule. The main ben-
eficiaries appear to have been foreign multinationals. In the North Sea, for instance,
the more advanced technologies have been developed largely in the United States and
elsewhere (Surrey and Cook, 1983). In relation to the financial and retail sectors, IBM
and DEC have much the largest market shares in computing and networking, while
Japanese companies dominate the market for peripherals and office equipment.
Looking on the bright side, many of these products are now being supplied from
multinational facilities within the United Kingdomthe dynamism of the service sec-
tor is even leading some foreign firms to locate product development in Britain. DEC
is, for instance, pioneering VAN service products in the United Kingdom because of
the advanced nature of the financial service sector (Morgan and Davies, 1989).
Although the amount of value that is added to their products in the United Kingdom
is not known, the rapid uptake of new techniques in the British service sector seems
to be one of the factors attracting multinational investment to the United Kingdom.
For all the above reasons, the relationship between the tertiary sectors and man-
ufacturing has become an increasingly significant issue in Britain. Some argue that the
growing predominance of services and the decline of the manufacturing base do not
matter. According to this view, the growth of services is a sign of modernity, and
reflects Britain's true comparative advantage. However, the massive trade imbalance
that has developed in the 1980s suggests that macroeconomic stability may not be
achievable without a stronger manufacturing base. In 1989, Britain's trade imbalance
reached 22bn, or 4.5% of GDP (Table 5.3). Although this has been partly caused by
misjudgments in macroeconomic policy that caused consumer demand to run ahead
of supply capacities, Britain seems to be falling into a trap whereby the trade returns
from an expanding service sector are being outweighed by its propensity to suck in
Table 5.3. British Trading Performance, 1978-1988 (bn)

Invisibles Current
v isible Trade Account
X- I
Exports Imports
Year (X) (I) Total Oil Nonoil X- I X- I
1978 35.0 36.6 -1.6 -2.0 + 0.4 + 2.5 + 0.9
1980 47.1 45.8 + 1.3 + 0.3 + 1.0 + 1.7 + 0.3
1982 55.3 53.1 + 2.2 +4.4 --2.4 + 2.5 + 4.7
1984 70.3 74.8 -4.5 + 6.9 -11.5 + 6.7 + 2.2
1986 72.7 81.4 0 -7
o. / + 4.1 -12.8 + 8.9 + 0.2
1988 80.2 100.7 -20.5 + 2.3 -22.9 + 5.6 -14.9
Source: United Kingdom Balance of Payments. 1988 Edition. London: Central Statistical Office, 1988.
imports of manufactured goods. Equally, the income growth from the expansion of
tradeable services and from higher productivity is creating demand for consumer dur-
ables and other goods that the manufacturing sector cannot presently meet.
THE MANUFACTURING BASE
In 1987, manufacturing GDP in Britain amounted to 85bn, or 23% of national out-

put. Medium- and high-technology sectors, as denned by the OECD, were responsible
for 46% of manufacturing sales and 69% of exports in 1986 (Table 5.4). This placed
Britain midway between Italy, which has the lowest commitment to these sectors
among the major industrial countries (43% of sales and 51% of exports), and Japan,
which has the highest (53% of sales and 82% of exports).
Like other industrial countries, the sectoral composition of output and trade has
been moving in the direction of higher technology manufacturing, although in Britain
rather gradually (from 41 % of sales in 1971 to 46% in 1986). What has been most pro-
nounced in Britain has been the decline of metal-based manufacturing, extending
from iron and steel to mechanical engineering to motor vehicles, aerospace being the
only exception. Contributing close to one-half of manufactured exports in 1971, the
share had fallen to less than one-third in 1986. Compensating gains were made in
chemicals and Pharmaceuticals, electronics, and aerospace. The shares of low-tech-
nology manufacturing industries in both production and trade remained compara-
tively stable over the period, although the historic decline of textiles continued.
These trends suggest that a satisfactory "upward" shift in Britain's technology
base occurred in the 1970s and 1980s. However, as indicators of the vitality of British
Table 5.4. Production and Export Shares in U.K. Manufacturing, 1971 and
1986
Production Shares (%) Export Shares (
1971 1986 1971 1986
Medium and high technology

Chemicals, etc. 11.5 14.7 12.2 18.8
Electricals, electronics 9.7 11.7 12.3 19.3
Mechanical engineering 8.3 8.8 17.8 13.1
Motor vehicles 6.7 5.8 12.7 7.0
Aerospace 1.2 2.4 3.6 6.9
Other 3.6 2.8 5.0 3.6
Subtotal 41.1 46.2 63.6 68.7
Low technology
Food, drink, and tobacco 18.0 20.6 6.4 6.8
Textiles, etc. 7.9 6.0 8.0 5.9
Metals 16.4 7.1 13.0 8.9
Paper and printing 8.8 9.3 2.4 2.7
Other 7.8 10.8 6.6 7.0
Subtotal 58.9 53.8 36.4 31.3
Source: A. Buxton, "Technology and Structural Change." NEDO, London, 1990, Fables I and II, using
OECD figures and definitions.
industries, they have to be treated with caution due to the unusual weight of multi-
national investment in the British industrial economy. Table 5.5 shows that one-fifth
of manufacturing GDP in 1987 came from foreign firms locating production in Brit-
ain, the greatest concentration of multinational investment being in the medium- and
high-technology sectors (aerospace again excepted). Only Belgium has a higher pro-
portion among European countries. By the mid-1990s, the proportion of manufac-
turing in foreign hands could reach one-third or even higher. Britain is now experi-
encing a new wave of foreign direct investment, with Japan this time in the vanguard.
Many Japanese firms, particularly in the electronics and automobile sectors, have cho-
sen Britain as their favored point of entry to the extended European market.
Hence the broad statistics of industrial production and trade provide a partial,
and sometimes misleading, measure of the performance of British-based manufactur-
ing enterprises. The structural changes in British manufacturing in recent years have
been the result of both deindustrializationthe withdrawal from areas such as metal-
workingand positive restructuring involving the creation of new manufacturing
capacities. A substantial, even the major, part of the latter can be attributed to foreign
investment. Without it, the structure of the British economy today would appear
much less modern.
Further insights into the scale and sectoral distribution of British manufacturing
can be gained by comparing Britain's large firms with those of other countries. Table
5.2 shows that Britain's share of large firms in the Fortune list of the top 500 non-
U. S. industrial companies in 1987 was next only to that of Japan. However, one-half
of them were in the areas of traditional strength in mining and petroleum, and food,
drink, and tobacco. Indeed, the sales of British firms in these sectors dwarfed those
from other countries ($213bn against $34bn, $65bn, and $99bn for Germany, France,
and Japan, respectively). They also exceeded by a substantial margin the sales of Brit-
ish firms in the main manufacturing sectors ($213bn against $124bn), in contrast to
the other countries where the balance was weighted heavily in the other direction.
Table 5.5. Foreign Companies' Share of

U.K. Gross Value Added (1987)
%
Motor vehicles 45
Office machinery, data processing 37
Chemicals, Pharmaceuticals 32
Rubber, plastics 24
Instruments 22
Mechanical engineering 21
Elcctricals, electronics 17
Paper and Publishing 15
Food, drink, and tobacco 15
Metals 10
Textiles, leather, clothing 4
Timber, furniture 3
Aerospace, other transport equipment 2
Total manufacturing 19
Source: UK Census of Production (1987).
As regards the manufacturing sectors, one is struck in Table 5.2 by the relatively
high incidence of British large firms. But their output is comparatively small by inter-
national standards. One reason is that Britain has very few "giant" manufacturing
companies in the medium- and high-technology sectors. The average sales per firm in
these areas in the Fortune list was $3.8bn for Britain, against $5.6bn for Japan, $7.3bn
for France, and $7.9bn for Germany. In 1987, only one British manufacturing com-
pany, ICI, had sales ($ 18bn) that exceeded $ lObn. The next largest was GEC with sales
of $7.8bn. In comparison, France had five, Germany nine, and Japan twelve.
At the other end of the scale, a number of studies have shown that Britain is rel-
atively poorly endowed with small firms (Bolton Committee, 1971; Ganguly and Ban-
nock, 1985). Although the rate at which small firms are created may be high, so is the
casualty rate. Their share of total manufacturing employment in the late 1970s, mea-
sured by establishment, was 24% in Britain compared to 30% in West Germany, 34%
in the United States, 52% in Switzerland, and 54% in Japan. A recent comparison of
Dutch and British productivity found that the number of persons per manufacturing
enterprise in 1984 was 254 in the Netherlands and 413 in Britain (van Ark, 1990).
Britain appears to lack the diversity evident in the German Handwerk sector, in Italian
textile and shoe manufacturing, and in the French food and wine sectors.
There thus appears to be a lack of strength at both ends of the spectrum where
manufacturing is concerned: there is no Siemens, or Fiat, or Mannesmann in Britain;
but nor is there the abundance of sophisticated small firms that is found in some of
these other countries. Again, one has to look outside medium- and high-technology
manufacturing to find "giant" British firms (such as BP, Unilever, BAT), and to the
service sector to find a profusion of relatively dynamic small firms. The growth in the
number of enterprises recorded in Britain during the 1980s, including those involved
in computer software, has been substantially a service sector phenomenon.
In identifying where Britain's relative strengths and weaknesses lie in manufac-
turing, one has to distinguish at the outset between areas where the strength is indig-
enous and where it derives from the presence of foreign multinationals. Among the
former, three stand out:
1. Chemicals and Pharmaceuticals. Here Britain possesses some of the world's leading
companies (ICI, Glaxo, Beecham), although Table 5.2 has shown that British out-
put in chemicals is dwarfed by that of Germany. Although there is significant
inward investment by U. S. firms in particular, this is offset by substantial outward
investment. The sector gains additional strength from the chemical and petro-
chemical activities of the major British oil companies.
2. Aerospace. As will become apparent below, strength in this sector derives largely
from Britain's heavy postwar commitment to defense procurement. The leading
firms are British Aerospace (airframes and guided weapons), Rolls-Royce (aeroen-
gines), and Lucas Aerospace, Dowty and Smiths Industries (engine controls,
hydraulics, and other subassemblies). For the same reason, Britain has a strong
international presence in defense electronics (GECnow incorporating Ferranti
and Plesseyand British Aerospace Dynamics). None of these firms is strongly
multinational in the sense that they locate significant R&D and production abroad.
However, all are notable participants in collaborative aerospace projects within
Europe and across the Atlantic.
3. Food, drink, and tobacco. Britain's substantial competitive advantage in this area
has already been discussed. The firms have long been strongly multinational, hav-
ing large production bases outside the United Kingdom.
Britain still has indigenous capabilities in telecommunications, and in electrical
and mechanical engineering, but without any distinct competitive advantage outside
some niches. As filieres, they no longer have depth and cohesion, and thus a substantial
international standing. Moreover, the privatization of the telecommunication and
electric utilities appears to be causing their hitherto protected equipment suppliers to
fall under foreign domination, rather than reviving their fortunes. Siemens has gained
the upper hand in the telecommunications equipment industry following its acquisi-
tion of a half share in the main British supplier, GPT; and investment in the combined-
cycle power plants, which are expected to provide most additional generating capacity
in the 1990s, will be based on foreign technology supplied by General Electric, Asea-
Brown-Boveri, and Siemens in particular.
To the areas of indigenous capability should be added those in which production
in Britain relies heavily on foreign investment. Two stand out:
1. Motor vehicles. Ford and General Motors (Vauxhall) have had large production
facilities in Britain since the 1920s (their design and R&D centers have, however,
been largely moved to Germany). Ford has recently acquired Jaguar, and General
Motors has acquired Lotus, the most innovative of the Britain's small car produc-
ers. The remaining British volume-car manufacturer, Rover, has become increas-
ingly dependent on its links with Honda. The automobile industry in Britain is
about to experience a period of strong expansion because of the decisions by Nissan
and Toyota to locate their main European production bases in northern England.
2. Electronics. Companies with large manufacturing facilities in Britain include IBM,
DEC, and Fujitsu following its recent takeover of ICL (computers); Hitachi, Sony,
Matsushita, and Toshiba (consumer electronics); Motorola, NEC, and Intel (semi-
conductors); and Rank Xerox and Cannon (office electronics). Scotland now has
the largest concentration of semiconductor manufacturing in Europe ("Silicon
Glen"). In semiconductors there is no internationally significant capability remain-
ing in British hands.
These judgments about Britain's indigenous strengths and weaknesses are generally
supported by indicators of competitive advantage in technology and trade. Table 5.6
shows that Britain's revealed technological advantage lies in aerospace, pharmaceuti-
cals, food products, coal and petroleum, chemicals, and mechanical engineering (in
that order). A glance at Table 5.6 is sufficient to indicate that the distribution of advan-
tage in Britain quite closely matched that of France and the United States, but that
there was a strong negative correlation with that of Japan and to a lesser degree with
that of Germany.
Sectoral differences in performance emerge more clearly from the trade statistics
(Table 5.7). Among the medium- and high-technology industries, the trade surpluses
in chemicals and aerospace increased, whereas the balance in electronics and motor
vehicles changed from a slight surplus in 1978 to very large deficits in 1986 (other trade
measures display similar patterns). All European countries faced deteriorating trade
balances in electronics over the period in question, but the greatest decline was expe-
Table 5.6. Revealed Technological Advantage by Sector, 1978-1986"
United United
France Germany Japan States Kingdom
Aerospace 3.64 1,06 0.18 1.24 2.10
Pharmaceuticals 1.36 0.96 0.74 0.83 1.72
Food products 1.09 0.75 0.69 1.07 1.43
Coal and petroleum 1.75 0.71 0.81 1.35 1.13
Chemicals 0.95 1.27 0.87 0.99 1.03
Mechanical engineering 0.98 1.18 0.69 1.00 1.01
Electricals, electronics 1.24 0.79 1.21 1.03 0.95
Motor vehicles 0.60 1.06 1.54 0.69 0.68
Office equipment 0.74 0.53 1.80 0.95 0.64
"Revealed technical advantage (RTA) is a measure of technical specification; it is the ratio of a country's share of U.S. pat-
enting in a given sector to its share of U.S. patenting in all product groups. An RTA in excess of 1 indicates above average
specialization.
Source: Cantwell & Hodson (1990).
rienced by Britain. An important question in relation to Britain's recent industrial per-

formance is, therefore, why the chemical and aerospace industries have done compar-
atively well, whereas the electronics and motor vehicle industries have fared badly.
These figures also provide a warning for Britain: multinational investment, which
has been growing most strongly in these last two sectors, does not so far appear to be
correcting the trade imbalance. If it is hastening the decline of indigenous capabilities
and replacing them with assembly plants, as some are claiming, it may instead be wors-
ening the trade situation. One should note, however, that the chemical sectors have
also experienced strong inward investment without a deteriorating trade balance. It is
Table 5.7. U.K. Trade Balance by Sector, 1978-1986 (m)

Exports Minus Exports Minus Change in Balance
Imports (1978) Imports (1986) 1978-1986
Medium and high technology
Chemicals, drugs + 1206 + 2306 + 1100
Electricals, electronics +486 -2183 -2669
Mechanical engineering + 2298 + 1725 -573
Motor vehicles + 311 -4127 -4438
Aerospace + 237 + 1634 + 1397
Other + 376 -1332 -1708
Subtotal +4914 -1977 -6891
Low technology
Food, drink, tobacco -1644 -3252 -1608
Textiles, etc. -707 -3289 -2582
Metals -400 -652 -252
Paper and printing -773 -2088 -1315
Other -1290 -3071 -1781
Subtotal -4814 -12352 -7538
Total + 100 -14329 -14429
Sora?:Mayes(1987).
only where there is an underlying weakness that foreign firms can behave like cuckoos
in the nest.
Come what may, the development of the British industrial economy in the 1990s
will depend significantly, and to a greater extent than for other West European econ-
omies, on the behavior of foreign multinational companies: the scale of their invest-
ments, the degree of local content in their products, their willingness to locate R&D
and design capabilities in Britain and to use British personnel, and their attitudes
toward future investment in the United Kingdom as against other parts of Europe
(now including Eastern Europe). In particular, the outcome will depend on the deci-
sions of Japanese companies, and on how Japan generally conducts its trade relations
with the enlarged European market. As always, the price of decline is a loss of eco-
nomic sovereignty.
THE INNOVATION SYSTEM: PRODUCTIVITY AND R&D
So far, the picture has been one of continuing retrenchment in Britain's manufactur-
ing industries. In one respect, however, industrial performance in Britain was more
than satisfactory in the 1980sin the growth of industrial productivity.
The Productivity Conundrum

For Britain in the 1980s, the indicators of technical progress present us with a dichot-
omy. As we shall see, R&D and patenting statistics suggest that Britain's technological
standing continued to deteriorate. The growth of productivity was, however, the fastest
among OECD countries over much of the decade. Between 1980 and i 987, output per
person employed increased in real terms by 24% in the United Kingdom, against 22%
in Japan, 12-14% in France, Italy, and Germany, and 7% in the United States. On the
face of it, this suggests that notable advances occurred in Britain's application but not
in its development of technology.
The productivity increases have become the subject of much debate among econ-
omists. There appears to be consensus on two points (Layard and Nickell, 1989; Oul-
ton, 1990; Feinstein and Matthews, 1990). The first is that productivity growth in the
1980s was not, by and large, investment or output led. It was not until the late 1980s
that manufacturing investment and output recovered sufficiently to overtake the levels
reached in the late 1970s. Instead, it was particularly associated with reductions in
overmanning and the scrapping of the large "tail" of inefficient plants. Freeman has
referred to this as the Verdun rather than the Verdoorn effect (Freeman, 1989). The
second is that there was significant improvement in the management of the remaining
productive assets. This came especially from the reduced resistance to change among
the labor force, and the accompanying reassertion of managerial authority within
firms. Some have stressed the importance of the "shock" of 1980-1981, when the
combination of domestic recession and an overvalued exchange rate threatened the
survival of many firms and brought high unemployment (Metcalf, 1988). This, in
turn, weakened the influence of trades unions, as did the government's refusal to inter-
vene in wage bargaining and its legislation to curb their powers.
It remains to be seen whether the productivity improvements can be sustained in
the 1990s. As recession has taken hold in the early 1990s, productivity has fallen along
with industrial output. Nevertheless, it can justifiably be claimed that Britain came
closer, if not close enough, to operating according to international best practice during
the 1980s.
Why, then, were the gains in manufacturing productivity not accompanied by
equivalent gains in Britain's innovative capabilities? One interpretation is that Britain
was beginning to act as a low-wage, low-productivity economy. By the end of the
1970s, a productivity gap had opened up between Britain and other leading European
economies, let alone with the United States and Japan. It was, therefore, economically
rational to seek profits and strengthening competitiveness by raising productive effi-
ciency and by improving product quality and design, rather than by pursuing monop-
oly rents through innovation. Except in high-technology industries where the neglect
of innovation tends to be fatal, Britain could, according to this view, afford to lower
the priority given to the development of new technology, for the time being at least.
Another interpretation is that changes in social relations and incentives in the
1980s strongly supported productivity improvement, but not improvement in inno-
vative performance. The combination of a pliant labor force and strong incentives to
reduce costs and maximize profits brought large gains in productive efficiency, but
those same incentives discouraged investment in R&D and in new productive capac-
ity. The growth of profits has in fact outstripped the growth of R&D and capital invest-
ment. In general, the increased surplus generated by higher productivity has tended to
be absorbed by higher dividend and interest payments, and by higher taxes, or has been
put aside to raise money earnings, or to engage in company purchases (Table 5.8).
A charitable view of innovation strategies in the 1980s is that, as in relation to
production, priority was given to squeezing more out of less. We shall see that this
applied to the approaches adopted by the government as well as producers. The gov-
ernment constrained spending on R&D, and made the achievement of "value-for-
money" a guiding principle when defining its policies on science, R&D, education,
and training, as it did in relation to every other aspect of public expenditure. Within
firms, more emphasis may also have been placed on incremental innovation and
improvement in design, rather than on the radical innovations that seemed appropri-
ate when they still felt able to challenge the technology leaders.
Table 5.8. Allocation of Industrial and Commercial Surplus

1977 1987
Gross profits (% of GDP) 15.4 18.5

Dividend and interest payments (% of 20.0 25.5
gross income)
U.K. tax on income (% of gross income) 9.7 15.2
Percentage of income undistributed 62.5 51.7
Investment (% of gross income) 37.8 34.8
Industrial R&D" (% of gross income) 7.5 5.9
Financial surplus (% of income 5.9 10.3
undistributed)
"Intramural R&D funded from all sources.
Source: UK National Accounts; "R&D 1988: Annual Review of Government
Funded R&D," Cabinet Office, HMSO, 1989.
Although operating as a follower rather than a leader has its economic advan-
tages, it carries the risk of growing technological backwardness. The British manufac-
turing economy finds itself in a paradoxical situation. Low wages and productivity lev-
els encourage specialization in areas of relatively low technology, and lead to emphasis
being placed on catching up with international best practice in both design and pro-
duction. Yet Britain has, as indicated in Table 5.6, one of the most "modern" and
R&D-intensive industrial structures among the OECD countries. Although the para-
dox may in part be resolved by operating at the low end of the high technology spec-
trum, British manufacturing still remains vulnerable to underinvestment in R&D, and
to general inefficiency in the innovation system.
Patterns and Trends in R&D Expenditure
What, in outline, were the main trends in the scale and pattern of British R&D expen-
diture over the past two decades? Six aspects deserve attention.
Declining Investment in R&D Relative to Other Countries
Throughout the postwar period, Britain has been a relatively high spender on R&D.
The historic trend in R&D spending has been closer to that of the United States than
to that of France, Germany or Japan (see Table 5.9). From high levels in the 1960s,
U. S. and U. K. expenditures on R&D as a percentage of GNP dipped in the 1970s,
and then recovered somewhat in the 1980s. Second to the United States in the mid-
1960s, Britain had fallen to fifth place in the mid-1980s after Sweden, Japan, Ger-
many, and the United States. The OECD commented in 1989 that the "United King-
dom was the only country where growth in R&D expenditure [in the 1980s] was lower
than growth in GDP."
Britain is also distinguished by its relatively low expenditure on basic research
(Table 5.10). Per capita, it had fallen in 1987 well behind that recorded by France,
Germany, and the United States (but not Japan). Moreover, the increase in the gov-
ernment's science budget over the previous decade was the lowest among the industrial
countries.
Heavy Commitment to Defense Technology
Like the United States, Britain stands out for the unusually high proportion of funds
spent on defense R&D. The proportion increased in the late-1970s and early-1980s in
response to the expansion of NATO defense spending and to the launching of a num-
ber of large equipment projects. It has since declined slightly and is expected to decline
more steeply as defense needs diminish. In 1986, defense accounted for 51% of gov-
Table 5.9. R&D Expenditure as a Percentage of GNP

Federal Republic United United
France of Germany Japan Kingdom States
1964 1.8 1.6 1.5 2.3 2.9
1975 1.8 2.2 2.0 2.1 2.2
1986 2.3 2.7 2.8 2.4 2.7
Source: "International Science and Technology Data Update: 1988." NSF,
December 1988.
Table 5.10. Per Capita Government Expenditure on Basic Research (%)

Federal Republic United United
France ofGermany Japan Kingdom States
PPP exchange rates
1975 36.6 56.2 20.4 43.0 45.3
1987 57.7 66.0 30.6 49.1 61.1
Increase, 1975-1987 (%) 58 17 50 14 26
Official exchange rates
1975 45.2 77.1 30.1 40.9 45.3
1987 71.8 88.1 42.2 45.3 52.9
Increase, 1975-1987 (%) 59 14 40 11 28
Source: Irvine et al. (1990).
ernment, and 20% of total R&D expenditure (this included the science budget).
Removing this component of R&D expenditure thus presents Britain in a less flatter-
ing light, leaving it trailing some way behind Japan, Germany, Sweden, and Switzer-
land in its spending on industrial innovation.
In recent years, the manner in which Britain's defense R&D statistics are com-
piled has become the subject of some debate (House of Lords, 1990). Unlike its U. S.
counterpart, the British Ministry of Defence spends little on basic or applied research.
The great majority of funds go to development projects, and a significant proportion
to activities (e.g., design and production startup) that fall outside the Frascati defini-
tion of R&D. In addition, expenditure on nuclear warhead production in the United
Kingdom is counted in the government's R&D figures, so that its extent can be dis-
guised.
However measured, the scale of resources committed to defense purposes would
remain high by international standards. Moreover, those who oppose changes in def-
initions argue that the preproduction activities recorded as R&D still absorb a large
proportion of skilled manpower, with possible opportunity costs for the wider high
technology base.
The Shift from Public to Private Investment in R&D
During the 1950s and 1960s, there was a steady increase in the proportion of R&D
financed by government. It peaked in the early 1970s, at around one-half of total
expenditure. A significant shift has since occurred toward private financing of R&D
in Britain. By 1986, the government's share of expenditure had fallen to 39%. This was
due both to the higher rate of growth of industry-financed R&D and the government's
own deliberate reduction of spending on civil R&D from the mid-1980s onward. The
government has attempted to withdraw support from "near-market" research, and
from development work generally, out of the belief that industry alone should decide
which technologies to bring to market and should carry all the risks involved.
In the second half of the 1980s, a number of initiatives begun in the early 1980s
were therefore terminated, including the Alvey project, which had been Britain's
response to the Japanese Fifth Generation Computing Project. Only in relation to Air-
bus is substantial development aid still provided for civil purposes. Instead, attention
has shifted to supporting precompetitive R&D, albeit less generously than in most
other industrial countries. In 1987-1988, the Department of Trade and Industry
(DTI) spent 92.5m on this type of R&D. The role of government R&D laboratories
also declined over the period. Whereas 26% of government-funded R&D was con-
ducted intramurally in 1975, the proportion had fallen to 16% in 1986.
Increasing International Cooperation in R&D
In virtually every area, the autonomy of the British innovation system diminished dur-
ing the 1970s and 1980s. In aerospace, most technological development is now carried
out in cooperation with U. S. or European partners, whether in relation to civil (Air-
bus) or defense (Tornado, EH 101 Helicopter, European Fighter Aircraft, Harrier AV-
8B) projects. In electronics, much of the R&D that is supported with public funds is
now carried out under the aegis of European Community programs, with ESPRIT
alone accounting for one-half of government R&D expenditure in the area of infor-
mation technology. British electronics firms have also exhibited the general interna-
tional leaning toward bilateral and multilateral collaboration in R&D.
Britain is therefore becoming increasingly integrated into the European innova-
tion system. In defense, it is doing so enthusiastically and from a position of strength,
despite its resistance to the creation of a Europe-wide defense R&D program (see
below). In civilian fields, by contrast, it is very much a follower rather than a leader in
Europe. Britain was not one of the main instigators of the Single European Market,
and has not usually welcomed the Community's expanded role in technological devel-
opment. Its desire to receive its juste retourhas left it with no option but to participate.
The lead has come instead from industries and governments in France, Germany,
Italy, and (in electronics) the Netherlands, as well as from the European Commission
itself (Sharp, 1989). Thus Britain has ceded its leadership in civil aerospace to France
(although Rolls-Royce remains the preeminent aeroengine manufacturer); it is a reluc-
tant participant in the European space program; and in electronics generally, it has not
joined with Siemens, Thomson, Philips, Olivetti, and others in trying to mobilize
resources to mount a technological challenge to U. S. and Japanese firms.
Multinational Investment in R&D
Another notable trend has been the growing proportion of R&D carried out by foreign
multinational companiesincreasing from 4% of total funding of industrial R&D
(private and public) in 1967 to 13% in 1986 (Table 5.11). As a result, a growing part
of the British innovation system has become an appendage of foreign innovation sys-
tems. This has spawned a lively, if inconclusive, debate about its effect on indigenous
technological capabilities. Stoneman has argued that a potential "internal brain drain"
Table 5.11. Sources of Funds for Industrial R&D,

1967-1986 (%)
Government Overseas Own Funds Total
1967 29 4 67 100
1972 33 6 60 100
1978 29 8 63 100
1983 30 7 63 100
1986 23 13 64 100
Source. Cabinet Office (1988), Table 2.1.
is being created from domestic to foreign-owned firms, with the results of R&D con-
ducted by the latter being used mainly to support production elsewhere in the world
(Stoneman, 1989). Others have preferred to see this as a positive development,
strengthening the skill-base and providing an example to local firms (Cantwell and
Hodson, 1990). Wherever the truth lies, the real increase in British industrial R&D
spending during the 1980s would have been smaller still without the contributions of
foreign firms.
Patel and Pavitt have provided evidence of the relatively high proportion of the
U. K.'s innovative output that comes from foreign firms (Patel and Pavitt, 1990). In
1981-1986, foreign multinationals accounted for a much higher proportion of U. K.
patents taken out in the United States than they did for other European countries,
Belgium apart (19% against a European average of 6%). They have also shown that
British firms are themselves comparatively highly internationalized in their R&D
activities. One-third of the patents registered in the United States by large British
firmsthe European average was one-fifthcame from R&D conducted outside the
home country. Both statistics may be partly explained by the relatively high propensity
of U. S. and British multinationals to locate R&D and production in one anothers'
countries, due to the common language among other reasons.
Sectoral Shifts in R&D Spending: The Growth of Electronics
The sectoral distribution of expenditure on industrial R&D is shown in Table 5.12.
The proportion spent on chemicals held roughly constant between 1975 and 1986, and
was not far out of line with other advanced countries. The most striking feature of
Table 5.12 is, however, the very marked growth of R&D in electronics, and the appar-
ently higher proportion of R&D allocations to electronics in Britain than in the other
cited countries. This is a puzzle, given the evidence that individual firms such as GEC
and Plessey have not been high R&D spenders by international standards (Morgan et
al., 1989) and that electronics has not been an area of competitive advantage for Brit-
ain. Table 5.13 shows that not only did Britain's revealed technological advantage in
electronics slip over the period when R&D expenditures were increasing, but its
advantage in the fastest growing areas of technologies had deteriorated sharply.
Table 5.12. Expenditure on Industrial R&D, by Sector (

United Kingdom United
France Germany Japan States
1975 1981 1986 (1985) (1985) (1986) (1985)
Chemicals 18 16 18 17 21 16 11
Mechanical 7 6 5]
engineering
42
Electronics 21 33 34 f 33 37 39 36
Electricals 5 3 3j
Motor vehicles 6 5 7 10 15 14 9
Aerospace 22 20 17 19 7 2 23
Other 21 17 16 21 20 31 22
Total 100 100 100 100 100 100 100
Source: Cabinet Office (1988); NSF (1991).
Table 5.13. Britain's Revealed Technological Advantage in Fast-Growing
(FG) Patenting Areas Compared with Performance Overall (All), in
Chemicals, Electrical, and Mechanical Technologies
1963-1968 1969-1976 1977-1984
FG All FG All FG All
Chemical 1.02 1.07 1.14 1.04 1.44 1.16

Electrical 1.13 1.05 0.83 0.97 0.57 0.84
Mechanical 0.80 0.96 0.95 0.99 0.65 0.96
Source.' Patcl( 1988).
It should be noted, however, that in 1985 electronics accounted for close to one-
half (237m) of multinational R&D spending in Britain, and that as much as one-
third of all electronics R&D in Britain in that same year (5 50m) may have been con-
ducted for defense purposes (Walker, 1988). This being the case, only just over
one-half of the electronics R&D may have been carried out by British firms orienting
themselves toward civil markets. Moreover, this civil expenditure seems to have
yielded a low return in terms of exports and economic output. The largest civil item,
the telecommunications switch System X, has not been exported, and British firms
have generally been unsuccessful in high-volume areas such as semiconductors and
consumer electronics.
In summary, five main points emerge from this look at R&D expenditures:
1. Among the industrial countries, Britain is an average spender on R&D. Its position
in the rankings has, however, been falling, and is substantially owed to the very high
expenditure on defense technology.
2. In terms of industry-financed R&D, Britain now counts as a low spender among
the major industrial countries.
3. Except for defense, the role of government in the support of industrial innovation
diminished sharply in the 1980s.
4. R&D has become increasingly internationalized. A growing proportion of R&D
carried out in Britain is funded by foreign firms, and British firms have become tied
into European cooperative programs. Except for defense, they have, however,
tended to be junior partners.
5. The most rapid growth of R&D expenditure has been in the field of electronics. But
as little as one-half of civil electronics R&D may be carried out by British firms,
and in a number of areas the return on R&D investments appears to have been low.
SOURCES OF WEAKNESS IN THE INNOVATION SYSTEM
How can the persistent weakness of the British innovation system, and the variations
within it, be explained? One important factor has already been discussed. This is the
relationship between the manufacturing and financial sectors, and the pattern of
incentives that the latter imposes on the former. The lack of any strong desire to devise
compensating mechanisms in turn reflects the diminished status of manufacturing
within the contemporary British political economy.
Three other general explanatory factors deserve attention: the heavy commit-
ment to defense technology, the shortcomings of education and training, and what
may broadly be termed the problems of coordination.
Distortions Caused by Heavy Defense Spending

As noted above, the development of defense technologies occupies a large part of Bri-
tain's technological resources. Only the United States surpasses it among Western
nations. More precisely, defense procurement absorbs a large proportion of high tech-
nology engineering resources. By comparison, chemical and other industries have
been little involved in defense markets, in the postwar period at least, except as pro-
viders of materials that mostly differ little from those supplied to civil markets. Here
we therefore have one possible explanation for the relatively poor performance of Bri-
tain's engineering industries.
Heavy spending on defense technology may be harmful to performance for three
main reasons (Kaldor et al., 1986). First, it can have high opportunity costs, especially
in an economy such as Britain's, which has a weak skill-base. In the British context,
there also appears to have been little "spinoff" into the civil sector, partly because of
the rigorous separation of civil from defense activities within the firm, as in govern-
ment. Unlike the Pentagon and the Delegation Generale pour rArmament, the Min-
istry of Defence has consistently refused to take any responsibility for the development
of technologies that are not tied to specific defense requirements (ACOST, 1989).
Second, involvement in defense markets can influence the "style" of technolog-
ical activity in large firms. The emphasis is placed on product rather than process inno-
vation, with the result that dynamic learning effects may not be realized to the same
degree as in civil areas of production. Moreover, there is a tendency toward excessive
elaboration in product specifications (baroque technology), partly because there is a
lack of "market discipline" when major wars are not being fought (Kaldor, 1982).
The third argument is that protected defense markets lure the large firms that act
as prime contractors away from activities where risks are higher and competitive pres-
sures more pronounced, and where sales cannot be "fixed" through negotiation with
politicians and the military bureaucracy. As such, patterns of behavior are established
that are not conducive to success in open civil markets.
Among economists, there is broad agreement that defense spending has sapped,
rather than strengthened, Britain's industrial economy. However, there are large dif-
ferences of opinion over the weight to be ascribed to this factor. It cannot provide a
general explanation of Britain's long-run industrial decline, since defense procure-
ment in peace-time only began to absorb a significant proportion of industrial output
in the 1950s. Moreover, it is difficult to judge what might have happened if the defense
market had not been there to prop up high technology firms.
It is also debatable whether the traditional criticisms of defense spending are as
valid, in the British context, today as in the 1960s and 1970s. Under the "Levene"
reforms of procurement practices, the government has forced large firms to accept a
larger share of the financial risk in defense contracts, collaboration in defense projects
with Germany and other countries may have brought learning and greater industrial
discipline, and recently a French-style policy of export maximization has been
adopted that discourages product complexity. The British defense industry had
remarkable success in international markets in the 1980sdefense being the only
engineering sectors where an improvement in the trade balance was recorded. Britain
is now third to the United States and USSR as a defense exporter, and has overtaken
France.
However one judges the historic effects of defense spending on industrial perfor-
mance, the important problem now is that very substantial technological resources are
locked into a market that is expected to decline, the Gulf War notwithstanding. The
question is whether large parts of the aerospace and electronics industry can be trans-
formed into capabilities able to withstand competition in crowded and highly com-
petitive civil markets.
The Poverty of Education and Training in Britain

British industrialization did not rely on mass education. Nor did it rely on the formal
training of managerial or engineering personnel. Until the late nineteenth century
(perhaps even later), an education "system" did not exist in Britain, although propos-
als to establish one extended back to the eighteenth century. In contrast to Germany,
Japan, and even the United States where organized education was the springboard for
industrial advance, in Britain education was disorganized and lacked the strong asso-
ciation with the aims of economic development.
The deficiencies in British education have long been recognized. There have been
many attempts to reform it, particularly to widen access and increase the priority given
to technical education. Despite the efforts, there is broad agreement that the British
education system remains one of the poorest in Europe. The following features stand
out:
1. A smaller proportion of young people move into higher education, or experience

any kind of further education, in Britain than in any other major industrial coun-
try. Nearly two-thirds abandon full-time study at the age of 16, one-half of which
have no formal education or training thereafter (Financial Times, 31 October
1989).
2. Although in quantitative terms a minor part of the education system, private
schools carry exceptional prestige and are much better resourced than state-sup-
ported schools. Closely connected to the Universities of Oxford and Cambridge,
they still provide a high proportion of the country's economic and political elite.
Education is in this respect as elitist as it is in France. However, access to private
schools and then to the most prestigious universities is determined in Britain more
by wealth than ability, and the graduates of this "system within a system" do not
have the rigorous technical training of the French Enarques. Within the economic
sphere, the British private system is oriented toward the service more than to the
productive sector, and especially toward the financial sector. The French phenom-
enon of bright young Polytechniciens developing careers that span both govern-
ment service and industrial management has no parallel in Britain.
3. Although Britain's output of scientists and engineers receiving higher education is
not far out of line with that of other countries, the quality of education they receive,
and their general standing within the education system, appears to be compara-
tively low. This applies especially to engineers, reflecting the low standing of the
engineering profession in the country at large. A number of studies have compared

the pay, status, and career structures of engineers in Britain and in other European
countries, and especially in Germany. They have found that the engineering pro-
fession is held in much higher regard in Germany than in Britain; that in Germany
there is greater involvement of academic engineers in industry, and of engineers in
industrial management; and that there is generally a stronger engineering "culture"
in Germany (Fores and Bongers, 1975). The lack of the specialist institutions of
technical education that are found in Germany and most other European countries
(the Technische Hochschule etc.) is both symptom and cause of the shortcomings
of the engineering profession in Britain.
A further claim is that the concept of "engineering" in Britain retains the nine-
teenth century association with strictly practical endeavor, and with learning-by-
doing. It implies tacit rather than formal knowledge. There is no equivalent of the
German concept of "Technik," which combines practical activity with the system-
atic application of technical and scientific theory.
There have been periodic attempts to reinvigorate the engineering profession
in Britain, and to give it a better educational grounding. However, debates within
the engineering institutes in the early 1900s, over the creation of a centralized
administrative structure for technical education after 1945, and over the Finniston
Inquiry's recommendation that an "engineering dimension" should be introduced
at national and company levels (Finniston, 1980) all came up against two insur-
mountable obstacles: the conservatism of the established engineering institutes and
their desire to preserve the tradition of professional self-regulation, and the lack of
sufficient interest groups inside and outside government to impel reform (McCor-
mick, 1991).
4. The most rigorous comparative studies of technical education have been con-
ducted by Prais. He found that except at the level of the University doctorate, the
output of skilled manpower is lower at all levels in Britain than in France and Ger-
many (but less clearly below that of Japan and the United States). The greatest dis-
parity occurred, however, at the level of the craftsman (see Table 5.14):
Both the Germans and the French have twice as many qualifying each year as
craftsmen as they have qualifying as technicians or with university degrees in
engineering: whereas in Britain . . . the number qualifying as craftsmen is less
than the number qualifying at higher levels. (Prais, 1988)
Table 5.14. Numbers Qualifying in Engineering and Technology, c. 1985"

United United
Kingdom France Germany Japan States
Doctorates 0.7 0.3 1.0 0.3 0.5

Master and "enhanced" degrees 2 6+ 4+ 5 4
Bachelor degrees 14 15 21 30 19
Technicians 29 35 44 18-27 17
Craftsmen 35 92 120 44 n.a.
"Numbers for Japan and the United States have been reduced in proportion to the U.K. population. All units in 1000s.
Source: Prais (1988).
The position appears to have worsened during the 1980s in this respect. The num-
ber of apprenticeships in British manufacturing fell by two-thirds between 1979
and 1989 (Financial Times, 14 March 1990).
5. Prais's findings can be generalized. The problem for Britain rests less with the edu-
cation of the "top" 15% of the labor force than with the lack of skills evinced by the
remaining 85%. This is the result of an inadequately funded general education sys-
tem, of insufficient commitment to industrial training, and of a culture that does
not set great store by intellectual achievement or technical proficiency. It is a
demand- as well as a supply-side phenomenon: there is inadequate supply of skills
because employers have not recognized the need for a more highly skilled labor
force, and have not been prepared to pay for it (Senker, 1988); and in prestigious
parts of the service sector, systematic education or training has not hitherto been a
requirement (Ingham, 1984).
The poverty of engineering skills and of craftsmanship seems an important source

of weakness in Britain's engineering industries. British science shows greater strength.
Despite the relatively low funding levels, Britain's share of world scientific literature is
comparatively high (8.2% of papers in 1986, versus 7.7% for Japan, 5.8% for West Ger-
many, and 4.9% for France), as is its citation ratio (NSF, 1991). This may again go
some way to explaining the relative strength in chemicals and Pharmaceuticals, where
scientific research forms a more integral part of the innovation process than in the
engineering industries.
Problems of Coordination
One is here on more difficult analytical ground. How can the extent and form of eco-
nomic coordination be compared across countries, and its results measured? And
when does coordination become a source of rigidity rather than dynamic efficiency?
In a number of respects, however, there are grounds for believing that the British sys-
tem of innovation, and the industrial economy more broadly, suffer from a lack of,
and often the wrong kinds of, coordination. This has frequently been observed in the
following contexts.
Integration of Scientific and Technological Communities
Only in a few areas, such as the nexus of medical research, Pharmaceuticals and the
national health service, does the relationship between scientific research, technological
development, and diffusion appear to work well in Britain, partly for reasons already
discussed. The connection between the physical sciences and engineering seems espe-
cially weak. One feature of Britain is that it is comparatively poorly endowed with
"bridging" institutions, such as the Fraunhofer Gesellschaft in Germany (Rothwell et
al., 1988).
Organization of'R&D and Product Development
Much R&D expenditure in Britain has been devoted to the development of large sys-
tems technologies in aerospace, telecommunications, power generation and other
areas. Yet each has a long history of waste and disappointment: TSR-2, Concorde,
Nimrod, and the air-defense system (UKADGE); System-X and its electromechanical
predecessors; the advanced gas-cooled reactor (AGR); and the advanced passenger
train to mention some of the more notorious examples. Although comparisons are
difficult, the record seems less impressive than in other European countries. In Britain,
there appear to be special problems coping with high degrees of organizational com-
plexity in R&D, and in settling conflicts over technology choice.
Producer- User Relations
It was noted earlier that close and mutually sustaining producer-user relations tend to
be the exception rather than the rule in Britain. Where they are in evidence, as in the
case of the retailer Marks & Spencer and its suppliers, they become the subject of great
curiosity but not, by and large, imitation. In the retail as in other sectors, this is one
reason why the British market is so easily penetrated by foreign producers. The British
tradition is for the consumer to have complete freedom of choice, and to have no spe-
cial responsibility toward, or common cause with, indigenous suppliers.
Managerial Coordination
From a large literature, three themes are worth drawing out. First, British management
tends to be hierarchical rather than participatory, as is being demonstrated by com-
parisons with managerial practices in the Japanese firms setting up production facili-
ties in Britain. The distinction between "gentlemen" and "players" that characterized
nineteenth century managerial relations (Coleman and MacLeod, 1986) still has some
relevance today. The German, Japanese, and Swedish traditions of industrial consen-
sus building are notably absent from Britain. Second, Tylecote and others have argued
that British industry is most successful where it is not faced with complex problems of
cross-divisional coordination, such as between engineering design, production, and
marketing (Tylecote, 1987). Tylecote suggests that this may be an important reason
for the comparative success of British pharmaceutical companies, where decisions on
production and marketing follow naturally from success in development and testing.
And third, Prais has demonstrated that the larger the production facility, the less effi-
cient are British enterprises by comparison with their foreign counterparts (Prais,
1981). Again, this appears connected with problems of handling social and technolog-
ical complexity.
Banks and Industry
The close coordinating relationship between banks and industrial enterprises that has,
for instance, been a feature of German and Swedish industrialization, does not exist
in Britain. Banks are seldom represented on the Boards of British companies. In rela-
tion to industrial restructuring, the effective reorganization of the German aerospace
industry by Deutsche Bank in the late 1980s could not have happened in Britain.
Industrial restructuring in Britain occurs largely through the "market for corporate
control" (i.e., through mergers and acquisitions), without the strong regulatory hand
of banks, the state or other agentsexcept where national security or competition pol-
icies are affected.
This lack of "collective integration" in the British economy may be attributed, in
some degree, to inherent features of British society: its racial and cultural heterogeneity
that deny it the natural cohesion of, say, Japan or Sweden, the strong tradition of per-
sonal and institutional individualism, the long history of conflict between labor and
capital, and the mistrust it has engendered, and the cultural bias against systematic
thought and planning that, although sometimes a source of flexibility, can inhibit the
development of organizational capabilities required for large-scale industrial devel-
opment.
During the 1980s, however, the prevailing opinion in Britain was that the econ-
omy had suffered from too much, rather than too little coordination, and of the wrong
sort:
The ability of the economy to change and adapt was hampered by the combi-
nation of corporatism and powerful unions. Corporatism limited competition
and the birth of new firms whilst, at the same time, encouraging protectionism
and restrictions designed to help existing firms. (DTI, 1988)
Under Thatcherism, the "free market" therefore came to be regarded as the para-
mount form of coordination. In advocating it, the government sought to make the
economy more rather than less individualistic, conforming to its vision of a market as
a dynamic array of autonomous, competing entities. An enhancement of collective
performance would, the government hoped, be the paradoxical outcome of increasing
the strains on, and rewards to, individual economic agents.
THE STATE AND THE INNOVATION SYSTEM
Britain has an unusual political structure. It is a multinational state, a union of three

"kingdoms" (Scotland, Wales, and England) and one "province" (Northern Ireland).
In each kingdom there is a strong north-south divide: the Highlands and Lowlands of
Scotland, the agrarian (Welsh-speaking) north and industrial (English-speaking) south
of Wales, and the north and south of England, the north predominantly industrial and
the south the heart of the service economy. Despite (or perhaps because of) wide eco-
nomic and cultural disparities, Britain has, along with France, the most centralized
political administration in Western Europe. Regional and municipal government is
weak and became weaker still in the 1980s.
In modern history, the state in Britain has not acted as the catalyst of industrial
and technological development as it has in France, Japan, and several other nations.
Between the mid-eighteenth and last quarter of the nineteenth centuries, its economic
role was confined to some regulatory functions (financial markets, property law etc.)
and to the advancement and military protection of foreign trade. The next hundred
years brought a gradual rising tide of state intervention in the economy. This said, it
was often hesitant, usually resisted, and seldom as determined and coherent as in other
countries. Moreover, it tended to be reactiveto decline, to perceived threats from
other nations, to protectionism, and to the political pressures to redress past social
wrongs. Unlike in France or Japan, for example, the state has not generally seen itself
as a creator of new modes of production, as an entrepreneur in its own right.
The two World Wars and the Cold War were important in bringing a more active
stance on technical change. They initiated and subsequently gave greater legitimacy
to state sponsorship. They brought the first direct funding of R&D, the establishment
and expansion of government R&D laboratories, the use of procurement as an instru-
ment for creating new production capabilities, the creation of industries (notably
nuclear and aerospace) under the wing of the state, and the use of industrial planning
in energy and other areas. In general, the new technologies emerging from wartime
activity gave rise to the notion that the state could, in addition to its broader economic
functions, play a part in accelerating the development and diffusion of new technol-
ogy.
Between the wars, the state also became an advocate of higher industrial spending
on R&D. This it sought to achieve especially through the formation of industrial
research associations that would allow firms to pool technical resources while sharing
experience, the government providing matching funds by way of inducement. The
research associations were most prominent in the 1930s, 1940s and 1950s, but have
subsequently declined, although some, such as the Welding Institute, are still active
(Sharp and Cook, 1988).
The 1960s, and particularly the period 1964-1970 when the Labour Party was in
office, were the apogee of state intervention in the economy. Emphasis was given to
import substitution, to the achievement of greater scale economies through industrial
concentration, to the expansion of the education system, and to regaining technolog-
ical leadership through the support of R&D and other measures. This was all intended
to be achieved through increased partnership between government, the trade unions,
and industrial management, which met together in the National Economic Develop-
ment Council (NEDC).
The 1970s, during which there was a see-saw between Conservative and Labour
administrations, can be seen with hindsight as a transitional period when the pre-
sumptions that had guided economic policy after the war began to unravel. It brought
disillusion with the government's economic philosophy (and not least with Keynesi-
anism), its managerial abilities, and its powers of omniscience. Despite efforts to revive
economic fortunes, the domestic economy did not prosper, international trade shares
continued their decline, and many interventionist policies came to be regarded as fail-
ures. In the industrial sector, technological resources became excessively focused on
high-technology projects (Concorde, the Advanced Gas-Cooled Reactor, etc.) to the
detriment of sectoral performance; and the demise of British Leyland and other large
firms discredited policies that supported industrial concentration and the formation
of national champions.
In part, Thatcherism was therefore a response, and naturally an opportunistic
response, to perceived failure in policy. This background goes some way toward
explaining the three main prongs of government policy in the 1980s: the restoration
of a market economy based on competition, financial incentives and private owner-
ship; the use of monetary instruments as the foundation of macroeconomic policy,
allied to constraints on public expenditure; and the reining in of trade union power,
bringing an end to the attempt to manage the economy through a partnership among
labor, management, and the state.
The Market, Enterprise, and Value for Money
In contrast to the earlier period, technological performance ranked low among the
government's list of concerns in the 1980s. During the first term of Mrs. Thatcher's
administration, there was nevertheless considerable continuity with innovation poli-
cies laid down previously. There was even some increase in R&D support, partly in
reaction to the scare over Britain's failure to keep up with U. S. and Japanese achieve-
ments in information technology, and partly because of the expansion of the defense
budget as East-West relations deteriorated. The Alvey program was launched in 1981,
and became one of Britain's largest efforts to strengthen national technological capa-
bilities. It also broke new ground by encouraging cooperation between firms, and
between industry and universities (i.e., better coordination was a central objective).
The Alvey program was, however, short-lived. The change of course in the mid-
1980s occurred for two main reasons. The desire to reduce public expenditure led to
pressure to cut government spending on innovation, as on other things. And efforts
were made to bring innovation policyand the general handling of state-industry
relationsinto line with the neoliberal economic principles that the government
increasingly espoused. This was accompanied by an increasing centralization of con-
trol within government over budgetary allocations, with the Cabinet Office (which
serves the Cabinet and Prime Minister) in particular assuming responsibility for policy
coordination. Its Advisory Committee on Science and Technology (ACOST) became
an important focus for discussion of the government's priorities, even if its advice was
not always heeded.
We have seen that there followed, over a comparatively short period, a substantial
reduction in the state's role in the innovation system. Government R&D expenditures
were cut as was employment in R&D establishments, and by privatizing high-tech-
nology producers (mainly in aerospace) and users (utilities) the government lessened,
by choice, its direct influence over technological decisions. This can be seen as part of
a general international trend, whereby responsibility for technological development
was increasingly assumed by private agents, acting alone and in concert. The British
government nevertheless stands out for the zeal with which it set about abandoning its
former roletestimony to the strength of its neoliberal convictions, and to the com-
paratively low value it had come to place on technological achievement.
The comparative disregard for technological activity was not perceived by the
government to be inconsistent with its campaign to create a more dynamic economy.
From the mid-1980s onward, an increasingly ideosyncratic, populist view of the
sources of economic dynamism came to be propounded (Redwood, 1988). It appeared
in its most unabashed form in the government's 1988 White Paper, "DTIthe depart-
ment for Enterprise."
There were three central concepts. One was the "open market," which was
regarded as the natural, most efficient, stable yet dynamic regulator of economic activ-
ity. Governments should stay out of the market, while taking vigorous action to ensure
that it remained open and competitive. Thus "competition policy" came to form the
core of the government's industrial policy. Any state activity that potentially distorted
firms' relations with the market was frowned on.
The second concept was "enterprise," which gave the market its energy and cre-
ativity. It brought in new actors, challenged the old, and was the basic source of output
and employment growth:
Enterprise is fundamental to a dynamic and growing economy. Lack of enter-

prise played a major part in the relative decline of the British economy; its return
has played a major role in the recent economic revival. The key to continued
economic success lies in the further encouragement of the enterprise of our peo-
ple. (DTI, 1988, p. 1)
This concept of enterprise was distinctive. It denoted a broad cultural movement: it

was a romantic vision of the natural condition to which "our people" should return.
At the same, it was highly individualistic. The principal actors were conceived to be
individuals and individual firms (notably small firms), whose separate and competitive
activities formed the market. The individual entrepreneur was also not necessarily, or
even primarily, a progressive force, in the sense that he or she exploited new scientific
or technical knowledge, or brought novel organizational approaches to economic
activity (Edgerton and Hughes, 1989). The essential qualification was that the entre-
preneur should operate a new and expanding business, whether it be a retail store, soft-
ware house, manufacturing company, or removal firm. The "Enterprise Initiative"
that was launched in the second half of the 1980s thus did not involve the implicit or
explicit prioritization of innovative activity, broadly defined. It thus ran counter to the
hierarchical and temporal assumptions that have lain behind innovation policies in
most other advanced countries, with their stress on moving into "higher technology"
areas of production.
The third guiding assumption was that the government's support for innovation
should be constrained in expenditure terms. The emphasis should be placed on achiev-
ing greater "value-for-money" by raising the efficiency with which resources were
used, not least by making their allocation conditional on recipients satisfying strict
performance criteria. In the absence of normal market selection mechanisms, other
carrots and sticks were required to act as incentives and prevent institutional sclerosis.
In relation to education, to science, to R&D, and to all other areas where the state
played a part in the innovation system, the requirement for demonstrable returns on
investment became the lynchpin of government policy, with an effect that was often
equivalent to that of high real interest rates. Long-term developments with uncertain
payback were inevitably rationed.
As such, the government's approach to science and technology mirrored priori-
ties within the economy at large. Productivity improvements and cost reductions came
before expansion and the creation of new capabilities even where, as in education,
there was a serious historic tendency toward underinvestment.
The Neglect of Capabilities
In practice, the government's policies remained more pragmatic than those

announced in the DTI White Paper. Nevertheless, it adopted a markedly less active
role in relation to technical change than its predecessors or its foreign counterparts
(including the U. S. government), from which it has not departed since. The new ori-
entation was an understandable reaction to the intrusive and ultimately wasteful pol-
icies of earlier governments. The "mission-oriented" approach that Britain shared
with France and the United States, and that involved the heavy subsidy of R&D pro-
grams in aerospace, nuclear energy, and telecommunications allied to support for
national champions, was unsustainable and inappropriate for a country in Britain's
economic position. Instead, Britain has moved toward adopting the "diffusion-ori-
ented" approach of more successful countries such as Germany and Sweden, even if
there has as yet been insufficient commitment to creating the strong decentralized
institutions that it requires (Ergas, 1984).
What has been lost on the way, however, has been recognition of the central
importance of building durable technological capabilities. Although firms have pri-

mary responsibility for developing and marketing new technologies in the modern
economy, the state retains an important supportive role especially if, as in Britain,
firms consistently underinvest in R&D and in training. Misconceptions of what these
capabilities should comprise have also become widespread. In particular, the preoc-
cupation with individualistic enterprise has diverted attention away from the large-
scale, collective, and resource-intensive nature of much contemporary technological
activity. Innovation seldom comes cheap, and is always risky. Particularly in electron-
ics (and notably semiconductors, computers, and telecommunications), the state has,
like the private sector, shown little appetite for the heavy investments required to
maintain strong indigenous industries.
Another aspect of the state's diminished concern for nurturing capabilities has
been its growing reluctance to play a part in identifying and supporting the technolo-
gies that may have strategic value, whether in terms of supply security or their potential
economic importance in the future. Two examples are symptomatic: the hesitancy in
providing significant funds to support research on superconductors and the unconcern
shown over the fate of Britain's remaining semiconductor and computer capacities
(viz. the government's acquiescence in the sale of INMOS to SGS-Thomson, and ICL
to Fujitsu). Even in the defense field, the government has greatly reducedto nuclear
warheads, cryptography, and a few othersthe list of technological capabilities that it
considers must survive in British hands.
Partly because so little strategic significance is attached to innovation policy, this
is one of the few areas where the government has willingly ceded authority to the Euro-
pean Commission (while frequently criticizing its policies), and where it has allowed
regional bodies to assume greater responsibilities. As a result, it has presided over the
transfer of decision making to the European center at the same time as to the British
periphery. In the former context it has, as we have seen, tended to fall in behind the
lead provided by its European partners. An interesting example has been the attitude
taken toward the Action Plan launched in 1988 to integrate the Western European
armaments market. Although the British government strongly supported the liberal-
ization of defense trade that the Action Plan sought to achieve, it resisted the creation
of an accompanying European defense R&D program (EUCLID). It was mainly on
French insistence that EUCLID was established (Walker and Gummett, 1989).
The dispersal of responsibility to the regions has occurred in three main ways.
First, the regional development agencies (those in Scotland and Wales being the most
notable examples) have tried to stimulate the development and diffusion of technol-
ogy, particularly to encourage the growth of small firms and the renewal of existing
sectors. By and large, their main function has been to mobilize the resources (land,
skilled labor, and capital) that will attract foreign investment to their respective areas.
Thus the Scottish and Welsh agencies have helped pave the way for the expansion of
Japanese investment in electronics production in Scotland (mainly semiconductors)
and Wales (mainly consumer goods).
Second, Local Enterprise Agencies have been established in various parts of the
country, funded by combinations of local government, business, and financial insti-
tutions, to provide support facilities for small firms. By 1989, 300 had been estab-
lished. And third, science parks have been created in Britain as in other countries. By
1988, 33 science parks had been constructed in Britain. The initiatives to set them up
have come mainly from universities eager to find new sources of revenue when their
budgets were being cut, and from regional development agenciesand not from cen-
tral government (Monck et al., 1988).
Although locally important, international comparative studies of similar devel-
opments abroad have not always been flattering to Britain. In France, decentralization
has been more forthright, and better coordinated and funded, whereas in Germany it
has involved the extension and deepening of already strong regional infrastructures
(Rothwell and Dodgson, 1989). A comparison of the development of electronics
industries in Scotland (Silicon Glen) and around Grenoble shows that although sub-
stantial production capacity has been established in Scotland, it has largely involved
product assembly. There is less evidence of new technologies taking root in the local
soil than in the case of France (Dunford, 1989). And the technology transfer benefits
claimed for science parks may have been exaggerated. Quintas (1988) found that firms
located in British science parks did not have noticeably stronger ties with higher edu-
cation institutes than firms located elsewhere. Although a great deal has been said in
praise of these regional initiatives, their effects on British technological capabilities
may therefore still be slight.
Instability in Government Policies

Looking back over the past 30 years, one is struck, when comparisons are made with
other countries, by the instabilityand lack of true cooperationthat have marked
the state's relationship with the industrial sector in Britain. Policies have veered
between the excessive managerialism of the 1960s and the equally excessive disen-
gagement of the 1980s. In no other advanced country has the government department
responsible for industrial policy so frequently changed its name, its internal organi-
zation, or its Minister (six times in the 1980s alone, against twice in the Treasury). In
no other country has it set itself such ambitious tasks in one decade (the Ministry of
Technology under Wedgwood Benn in the late 1960s), or willed its own disbandment
in another (DTI under Ridley in the late 1980s).
In part this reflects the violent doctrinal swings that have occurred in British gov-
ernance in recent decades, and the particular character of relations between the state
and the production system in Britain. But it is also a symptom of economic decline,
and of the evident difficulties of arresting it. Actions in one period have tended to be
reactions, and often overreactions, to the perceived failure of policies during the pre-
ceding period. This is one of the vicious circles that have made it so difficult to reverse
economic decline since 1945: decline has engendered policy instability, which has
reduced the state's ability to orchestrate a sustained revival; and instability may itself
have contributed to the downward spiral.
CONCLUSIONS
Among the countries discussed in these pages, Britain has the oldest industrial econ-
omy. For much of the eighteenth and nineteenth centuries, its innovation system had
no match, generating revolutionary changes in the techniques of energy and material
transformation (the coal, iron, and steam nexus), in the organization of production
(the factory system), and in transportation (railways and the steam ship). Indeed, it
can be argued that the modern era's systematic pursuit of technological advance orig-
inated in Britain.
The twentieth century has seen a gradual erosion in the country's industrial
standing, to the extent that today it is no longer counted among the technological lead-
ers outside a few niches. At each stage, the British economy has adapted to changing
consumption patterns and production possibilities, to the extent that it now has one
of the most "modern" industrial structures. However, the new industries have not, by
and large, proved successful in international markets. Typically, they have been inven-
tive in their early stages but have failed to consolidate their positions as technologies
and markets have developed. How to operate as an effective follower, and to avoid
becoming a laggard, has become the main industrial task.
Among the various factors that have contributed to decline in the twentieth cen-
tury, four seem to stand out:
1. The nature and influence of Britain's capital markets, with their unusual dedication
to short-term gain, and to trading in rather than developing productive assets;
2. The persistent underresourcing and undervaluation of education and training
(except in the military field, as the Falkland and Gulf Wars have demonstrated);
3. Weakness in coordination, due inter alia to the strong tradition of individualism
(at institutional as well as personal levels), and to social conflicts that have been
exacerbated by decline;
4. The loss of a strong technological "culture," which is particularly evident in the low
status of engineers in contemporary Britain.
Hitherto, attempts to achieve reform in Britain have been pursued in a setting of sub-
stantial national autonomy, even insularity. During the twentieth century, the main
economic exemplar has been the United States because of its domination of new
industries and its primacy among foreign investors in Britain, and recently because of
its advocacy of neoliberal economic policies. This said, the search for new approaches
has been largely internalized, choices being made with reference more to Britain's par-
ticular historical experiences and administrative traditions than to foreign practices,
and often being tightly constrained by institutional inertia.
If this were to continue, there would be few grounds for believing that the above
structural impediments would be any more surmountable than they have been in the
past. During the 1990s, however, three notable discontinuities will occur in the context
in which British economic development will have to be approached. The first is the
implementation of the Single European Act and of European monetary integration,
which will tie Britain more tightly into a European system of economic exchange and
regulation. It will place substantial limits on the British government's abilities to
implement its ideosyncratic industrial and macroeconomic policies, while exposing
more cruelly any deficiencies in Britain's productive infrastructure, broadly defined.
The second change comes from the build-up of Japanese foreign investment in
Britain, particularly in the electronics and automobile industries. In several areas, Bri-
tain's fortunes will depend on Japanese corporate strategies in relation to the enlarged
European market, and on Japanese contentedness with Britain as a location for
advanced industrial production. This will force Britain to become more open to Jap-
anese models of industrial organization than to the U. S. models that have been influ-
ential in the past, while again exposing infrastructural inadequacies.
The third discontinuity is the ending of the East-West conflict and the need for
massed armaments that accompanied it. This robs Britain of its important politico-
strategic position in the defense of Northwest Europe, and in providing a bridge
between the United States and Western Europe within NATO. Despite the Gulf War,
this is likely to weaken Britain's economic and political ties with the United States
among external relationships, that with Japan could become as important as that with
the United Stateswhile binding Britain more tightly into the European framework.
And within Britain, it will require a substantial reconfiguration of high technology
engineering resources as demands for weaponry diminish.
Each of these changes poses threats to Britain's international status. From the
point of view of economic development, however, they also bring significant oppor-
tunities. In particular, they will require Britain to be more open to Japanese and Con-
tinental European influences where the management of productive resources is con-
cerned. Above all, Britain will benefit if it learns to place greater emphasis at all levels
on the nurturing of capabilities, rather than on the extraction of quick economic
returns.
One of the great ironies of Britain's current approach to Europe is that market
integration is regarded as a great boon, and social integration a great threat. Yet it is
the archaism and inefficiency of Britain's social institutions that seem to lie at the root
of so many of its industrial weaknesses. This being the case, the most important ques-
tion is whether European integration will bring modernization in Britain, not through
the further extension of market mechanisms, but through the diffusion of more effec-
tive institutional practices in the educational and other fields.
REFERENCES Dunford, M. (1989). "Technopoles, Politics and

Markets: The Development of Electronics in Gre-
noble and Silicon Glen." In M. Sharp and P.
Advisory Council on Science and Technology Holmes (eds.), Strategies for New Technologies.
(ACOST). (1989). "Defence R&D: A National New York and London: Philip Allan.
Resource." London: HMSO. Edgerton, D., and Hughes, K. (1989). "The Poverty
Bolton Committee: Report of the Committee of of Science." Public Administration 67(4):419-33.
Inquiry on Small Firms. (1971). London: HMSO. Ergas, H. (1984). "Why Do Some Countries Inno-
Buxton, A. (1990). "Technology and Structural vate More Than Others?" CEPS Papers No. 5.
Change." UK Economy Papers. London: Brussels: Centre for European Policy Studies.
National Economic Development Office. Feinstein, C., and Matthews, R. (1990). "The
Cabinet Office. (1990). "R&D 1988: Annual Review Growth of Output and Productivity in the UK:
of Government Funded Research and Develop- The 1980s as a Phase of the Post-War Period."
ment." London: HMSO. National Institute Economic Review 132, 78-90.
Cantwell, J., and Hodson, C. (1990). "The Interna- Finniston Report: (1979-80). "Engineering Our
tionalisation of Technological Activity and Brit- Future." Report of the Committee of Inquiry into
ish Competitiveness." Mimeo, University of the Engineering Profession, Session Cmnd 7794,
Reading, Department of Economics. January 1980.
Coleman, D., and MacLeod, C. (1986). "Attitudes Fores, M., and Bongers, N. (1975). "The Engineerin
to New Techniques: British Businessmen. 1800- Western Europe." Unpublished mimeo Decem-
1950." Economic History Review 39(4): 588-611. ber, (in SPRU Library).
Department of Trade and Industry. (1988). "DTI Freeman, C. (1989). "R&D, Technical Change and
the Department for Enterprise."London: HMSO. Investment in the UK." In F. Green (ed.), The
DTI: (1988-89). Memorandum to the House of Restructuring of the UK Economy. London: Har-
Commons Select Committee on Trade and Indus- vester Wheatsheaf.
try's Inquiry into Information Technology. Ganguly, P., and Bannock, G. (eds.) (1985). UK
Small Business Statistics and International Com- mens Bid for Plessey: The wider European
parisons. London: Harper & Row. Issues." PICT Working Paper 2, SPRU, January.
Henderson, P.O. (1977). "Two British Errors: Their Mowery, D., and Rosenberg, N. (1990). "The US
Probable Size and Some Possible Lessons." National Innovation System." Draft paper pre-
Oxford Economics Papers July, 159-205. pared for the Columbia University project on
Hobsbawm, E.J. (1987). The Age of Empire, 1875- national innovation systems, November.
1914. London: Weidenfeld and Nicolson. National Science Foundation (NSF) (1991). Inter-
House of Lords. (1990). Select Committee on Sci- national Science and Technology Data Update.
ence and Technology. Report on Definitions of NSF 91-309, Washington, D.C.
R&D. Session 1989-90. London: HMSO. Oulton, N. (1990). "Labour Productivity in UK
Ingham, G. (1984). Capitalism Divided? The City Manufacturing in the 1970s and in the 1980s."
and Industry in British Social Development. Lon- National Institute Economic Review May, 71-91.
don: Macmillan. Patel, P. (1988). "The Technical Activities of the
Irvine, J., Martin, B., and Isard, D. (1990). Investing UK: A Fresh Look." In A. Silberston (ed.), Tech-
in the Future: International Comparisons of Gov- nology and Economic Progress. London: Macmil-
ernment Funding of Academic and Related lan.
Research. Aldershot, England: Edward Elgar. Patel, P., and Pavitt, K. (1990). "Large Firms in the
Kaldor, M. (1982). The Baroque Arsenal. London: Production of the World's Technology: An
Andre Deutsch. Important Case of Non-Globalisation." Journal
Kaldor, M., Sharp, M., and Walker, W. (1986). of International Business Studies 22( 1): 1 -21.
"Industrial Competitiveness and Britain's Pavitt, K. (1989-90). Evidence Given to Select
Defence Commitments." Lloyds Bank Review Committee on Science and Technology's Report.
October, 31-49. "Definitions of R&D." House of Lords, Session.
Landes, D.S. (1969). The Unbound Prometheus, Redwood, J. (1988). Popular Capitalism. London:
Cambridge, England: Cambridge University Routledge.
Press. Quintas, P. (1988). "Science Parks: Image and Real-
Layard, R., and Nickell, S. (1989). "The Thatcher ity." Paper prepared for the Study Group confer-
Miracle?" Discussion Paper No. 315. London: ence on Aspects of Industrial Policy, Centre for
Centre for Economic Policy Research. Business Strategy, London Business School, June.
Lazonick, W. (1990). "Controlling the Market for Prais, S.J. (1981). Productivity and Industrial Struc-
Corporate Control: The Historical Significance of ture. Cambridge, England: Cambridge University
Managerial Capitalism." Paper presented to the Press.
Third International Joseph A. Schumpeter meet- Prais, S.J. (1988). "Qualified Manpower in Engi-
ing, June. neering: Britain and Other Industrially Advanced
McCormick, K. J. (1991). "The Development of Countries." National Institute Economic Review
Engineering Education in Britain and Japan." In February, 76-83.
H. Gospel (ed.), Industrial Training and Techno- Rothwell, R., and Dodgson, M. (1989). "Regional
logical Innovation: A Comparative and Historical Technology Policies: The Development of
Study. London: Routledge. Regional Technology Transfer Infrastructures."
Mackinder, H. (1904). "The Geographical Pivot of Paper Prepared for the Third International Work-
History." Geographical Journal 23(4):421-44. shop on Innovation, Technical Change and Spa-
Mayes, D.G. (1987). "Does Manufacturing Mat- tial Impacts, Cambridge, September.
ter?" National Institute Economic Review Rothwell, R., Dodgson, M., and Lowe, S. (1988).
November, 47-58. "Technology Transfer Mechanisms, Part I: The
Metcalf, D. (1988). "Trade Unions and Economic United Kingdom." London: National Economic
Performance: The British Evidence." Discussion Development Office.
Paper 353. London: Centre for Labour Econom- Scnker, P. (1988). "International Competition,
ics. Technical Change and Training." SPRU/Impe-
Monck, C, Porter, R., Quintas, P., Storey, D., and rial College Papers in Science, Technology and
Wynarczyk, P. (1988). Science Parks and the Public Policy No. 17, January.
Growth of High Technology Firms. London: Sharp, M. (1989). "European Technology: Does
Croom Helm. 1992 Matter?" SPRU/Imperial College Papers on
Morgan, K., and Davies, A. (1989). "Seeking Science, Technology and Public Policy No. 19,
Advantage from Telecommunications: Regula- February.
tory Innovation and Corporate Information Sharp, M., and Cook, P.L. (1988). "R&D Coopera-
Networks in the UK." Information Networks tion among Firms, Universities and Research
and Competitive Advantage, vol. 3 of Compar- Organizations in the United Kingdom." SPRU,
ative Reviews of Telecommunications Policies February.
and Usage in Europe, Paris: OECD, pp. 267- Stein, A. (1984). "The Hegemon's Dilemma: Great
317. Britain, the United States and the International
Morgan, K., Harbor, B., Hobday, M., von Tunzel- Economic Order." International Organization
mann, N., and Walker, W. (1989). "The GEC-Sie- 38: 355-386.
Stoneman, P. (1989). "Overseas Financing for van Ark, B. (1990). "Comparative Levels of Labour
Industrial R&D in the UK." Paper presented to Productivity in Dutch and British Manufactur-
the Annual Meeting of the British Association for ing." National Institute Economic Review
the Advancement of Science, Sheffield, Septem- 131:71-85.
ber. Walker, W. (1988). "UK Defence Electronics: A
Surrey, A.J., and Cook, P.L. (1983). "The Off-Shore Review of Government Statistics." PICT Policy
Petroleum Supplies Industry: British Govern- Papers No. 4. London: Economic and Social
ment Policy Compared with Norwegian and Research Council.
French Policies." SPRU Occasional Paper No. 21, Walker, W., and Gummett, P. (1989). "Britain and
October. the European Armaments Market." International
Svennilson, I. (1954). Growth and Stagnation in the Affairs 65(1):419-42.
European Economy. Geneva: United Nations. Wiener, M.J. (1987). "English Culture and the
Tylecote, A. (1987). "Time Horizons of Manage- Decline of the Industrial Spirit." Cambridge,
ment Decisions: Causes and Effects." Journal of England: Cambridge University Press.
Economic Studies 14(4):51-64.
6
The French National System of Innovation
FRANgOISCHESNAIS
The French system is essentially a creation of the post-World War II period. The higher
education sector, with its dual component (the Universities and the "Grandes Ecoles")
dates back to the late eighteenth century and to subsequent developments at given
periods of the nineteenth century. But otherwise today's institutions and mechanisms
have all evolved out of those that were first built just after the Liberation from 1945 to
1949 and again from 1958 to 1966 during the first phases of the Fifth Republic. The
system has several features that are quite specific to France: (1) the organization and
funding of the largest part of fundamental research through a special institution, the
CNRS, distinct from the higher education sector entities, which are funded by the
State and governed by scientists in an uneasy relationship with public authorities; (2)
a dual higher education sector producing at least one type of senior technical person
little known elsewhere, namely the "Grandes Ecoles" technical experts elite of engi-
neers cum industrial managers, cum high level political and administrative personnel;
and (3) a pervasive element of State involvement in the production not just of general
scientific and technical knowledge, but often of technology per se in the form of pat-
entable and/or immediately usable products or production processes.
Special attention will be paid to this last feature. The French national system of
innovation consists to a large extent of a set of vertically structured and fairly strongly
compartmentalized sectoral subsystems often working for public markets and invari-
ably involving an alliance between the State and public and/or private business enter-
prises belonging to the oligopolistic core of French industry. The most important sub-
systems are those that concern electrical power production (conventional and
nuclear), telecommunications, space, arms production, and electronics. But the State-
enterprise relationship also exists in petroleum, railroad equipment and transport sys-
tems, civil engineering, and marine technology. More generally it is present in a latent
form in every field where the State has built, at some period or another, an R&D capac-
ity and has looked, in particular, since the 1970s, for industrial partners to whom to
transfer the technology and knowledge produced: this is now also true in the areas of
medicine and agriculture. Even in the case of industrial sectors where public involve-
ment in science and technology is low as in the chemical-pharmaceutical complex,
the role of the State remains strong with respect to industrial restructuring and the
192
THE FRENCH NATIONAL SYSTEM OF INNOVATION 193
provision of capital. In the last analysis, there are few "truly private" firms, enjoying
little or no direct State support, among the top French firms or organizations patent-
ing in the United States (Pavitt and Patel, 1990). L'Oreal (fine chemicals), Michelin
(tires), Peugeot (cars), and Valeo (automobile parts) are among them but remain
exceptions.
Large or very large firms (at France's level of course) belonging to the oligopolistic
core of French business are the State's partners. In some instances, the firms concerned
are classic nationalized enterprises [as EDF(Electricite de France) in conventional and
nuclear electrical power]; in others, they are firms with all or a majority of their capital
owned by the State, but otherwise totally normal business enterprises in their strategy
and behavior (e.g., Elf-Sanofi or Thomson); in other cases again they are private cor-
porations in the sense of being owned by private asset holders (CGE or Dassault), but
they behave much in the same way as firms with public capital. For reasons of oppor-
tunity and/or political necessity (internal and external) the distribution of R&D, man-
ufacturing, and marketing capacity between the public and private sectors and the way
responsibilities are shared between a given State agency and its industrial partners
have often evolved considerably over time. But the changes have never been of a mag-
nitude such as to sever completely the State-capital relationship and bring the alliance
to an end. The chapter examines some of the reasons for this and gives priority to two
explanations: first the inherent historical weakness of French industrial capitalism
along with the need it has of receiving State support and second the important, highly
original role played by the elite of the "Grandes Ecoles" and the Grands Corps in cre-
ating particularly strong links between the State apparatus stricto sensu, the public or
quasipublic enterprise sector, and the private industrial and financial sector. This
implies that although the system was built after 1945, its configuration has nonetheless
been shaped by the whole legacy of French social history.
As in the case of other countries, the French system of innovation has many com-
ponent elements, is divided into different segments, and includes strata dating back to
different historical phrases and forms of technological accumulation. In an old coun-
try such as France these may have roots in earlier peasant and artisan "technical cul-
tures" going back in some instances to before the industrial revolution and the appli-
cation of science and technology to production (e.g., wine making and artisan food
processing) and in others to the nineteenth century (e.g., fashion and beauty products).
These components of the system of innovation should not be overlooked.1 However,
despite the potential inherent opportunities they offer for competitiveness and the
cohesion of the country's economic and social fabric, they have never been the object
of the same attention as the large technology system-oriented parts of the innovation
system. This is true of the support they receive but also of the studies they have
prompted. France is a country in which there is continuous talk about the importance
of SMEs and innovation, but where, barring a few exceptions, one finds only limited
evidence of their role as active components of the innovation system.2
For an innovation system as distinct from an R&D system, the "bottom line" is
the capacity to compete successfully in world markets. Consequently the chapter con-
cludes that the French system's numerous spectacular achievements cannot hide its
serious weaknesses, expressed today in the endemic vulnerability of the French trade
balance as well as in the strong rigidity of the system in the face of contemporary
requirements for technological change. Technology per se is not, however, the sole nor
even the most important factor at work. Technological innovation stems from within
given national economic, political, social, and cultural environments. The perfor-
mance of a national system will be shaped by these factors. This is particularly true of
France (Salomon, 1991).
THE NINETEENTH-CENTURY RECORD AND LEGACY: AN ADVANCED BUT

UNBALANCED SCIENCE SYSTEM: FEW LINKAGES WITH PRODUCTION: A
DUAL HIGHER EDUCATION SYSTEM
When one is examining the development of a national system of innovation, one is

reading a country's economic, social, and political history through the particular prism
of the conditions surrounding the use of technology in capitalist production and the
choices made by the ruling class or dominant elites regarding the production and
application of scientific and technological knowledge. In our case, after a brilliant start
in the eighteenth century, French scientific production had to wait until the mid-twen-
tieth century before the political and industrial conditions that would allow it to make
its full contribution to military and foreign policy and of course to growth.
French Science from Colbert to Napoleon

The French Academic Royale des Sciences was founded in 1676 by Colbert with the
explicit aim of fostering scientific capacities and fitting them into the machinery of
government. Thus basic science was immediately synonymous with expert science
seeking industrial and military applications. If the institutional establishment of sci-
entific research in France was from the very beginning an act of State, so of course was
much of manufacturing industry, in particular all the "manufactures" created under
Louis XIV. Members of the Academic often provided scientific leadership for the
model factories. Thus, in the second half of the eighteenth century, Berthollet and
Chaptal were active in research on the chemistry of dyes connected with the manu-
facture of tapestries at the Gobelins, Macquer worked on porcelain at Sevres, and
Lavoisier on gunpowder at the Arsenal (Grassland, 1975; Gillipsie, 1980).
The scientific community was a strong beneficiary of the absolute Monarchy, but
also from the 1740s onward under the leadership of d'Alembert, Diderot, and the
group that published the Encyclopedia, a pivotal component in the political struggle
against the Ancien Regime and in the programmatic preparation of the French Rev-
olution. The thrust of the Encyclopedia's campaign was that for "progress" to take
place and the fulfillment of science's economic and social promise to materialize, the
fetter of the Ancien Regime on French nascent capitalist society had to be broken. This
explains the speed at which the Revolutionary and Napoleonic governments were able
to mobilize science for military and industrial purposes (Dhombres, 1988). Toward
1810 Paris housed a scientific population of a size unprecedented up to that time. This
"critical mass" of talent was the start of one of the most creative periods in French
science and certainly the only phase in which it exercised international scientific lead-
ership. During the first half of the nineteenth century Paris was, in the words of von
Humboldt, the "true metropolis of science to which scientists flocked from all over
Europe to associate with such great scientific figures as Cuvier, Lavoisier, and Laplace"
3
(Gillipsie, 1980). During this period the College de France, the Ecole Polytechnique,
and the Museum d'Histoire Naturelle were the world's leading scientific institutions.
The economic policy measures rendered necessary by the British naval blockade
(the Continental system) forced the Napoleonic governments to try and root science-
based innovation in industry. This led in particular to the birth of a chemical industry
based on the research of Chaptal, Berthollet, and Leblanc. It included one firm that
has survived until today, Saint-Gobain, and was the leading chemical industry in
Europe until 1830 or so. Once the impetus of the Napoleonic State-led and State-
supported policies had worn out, French private industry did nothing to pursue the
necessary investment and maintain the ties with research. There were a few exceptions
as in protestant Mulhouse, but they served only to prove the rule.4 Though it was
Lavoisier who founded the modern theory of chemistry, the development of the sub-
ject took place outside France, especially in Germany. As Papon has aptly put it, the
relations between science and industry are those of "a very precocious divorce"
(Papon, 1978, p. 20).
Inevitably French scientific leadership came to an end because of this. Though
France produced later in the century leading scientific figures such as Claude Bernard
and Louis Pasteur, and at the end of the century, the mathematician Poincarre and
the Curies, scientific leadership passed rapidly to Germany. This process did not
escape the attention of the greatest scientific figures of the time. Pasteur, who was par-
ticularly active in the defense of French science, published as early as 1868 a pamphlet
with the revealing title, "Le budget de la science." After France's first humiliating mil-
itary defeat by Germany in 1870, he wrote a number of articles in which he argued
that one of the causes was the lack of support given to French science and technology.
However, Pasteur is a fascinating example of the very peculiar relationships that grew
up between science and industry in France in the second half of the nineteenth cen-
tury. Pasteur worked repeatedly on problems raised by industrial activity (animal dis-
ease as in the case of silkworms, intensive production of chicken and pork, food con-
servation, scientifically controlled wine and beer production). Yet, he refused to move
close to capitalist production and struggled to maintain the idea of fundamental, "dis-
interested" science (Salomon-Bayet, 1986).
Some Aspects of France's Slow Industrial Growth
A study of one of the major scientific institutions of nineteenth century France ends
with the overall assessment that "the slower growth and lower productivity of French
science, when viewed in comparison with German science, can be seen as a conse-
quence of the relatively slower expansion of the French economy, especially in the
chemical and metallurgical industries which grew much more rapidly across the
Rhine. In short, the French economy got the level of science it needed and could sup-
port" (Zwerling, 1980, p. 59). The slow and uneven development of industrial capi-
talism in the nineteenth century has been the object of considerable study and debate.5
I will select the features of French economic and social development in the nineteenth
century that seem both the less well known and the most important.
The first point concerns the particular strength of conservative forces vested in
the agrarian structure. Because of its direct intervention in the revolutionary process
in 1790, a large section of the peasantry won the right to own the land they had worked
on the large estates. Agrarian reform occurred to the extent that the Church's lands
were divided and sold, but many aristocratic large holdings survived the gale. In the
aftermath of the Revolution and the Restauration, the agrarian base of French society,
both peasant and bourgeois-aristocratic, emerged as a very strong economic and social
force. Some sections of the peasantry were fairly radical at some moments (in 1847-
1848 and again in the early twentieth century), but on the whole the peasantry pro-
vided one of the social foundations of political conservatism.
Rent and profit from agriculture were canalized by a strong, well-organized, and
advancedfinancial system, characterized by early centralization of the monetary sys-
tem under the Bank of France and a fairly concentrated banking system characterized
by the presence of a small number of powerful merchant banks. Rent from the land
cleverly managed, helped the growth of the financial rentiers. The high and generally
stable income provided by financial investment aided French capital owners to rapidly
develop strong "rentier" features with a propensity to invest in government loans, later
in railway stocks, then in colonial trading companies and banks (the Bank of Algeria,
of Indochina, etc.), which had the prized advantage of being private ventures vested
with regalian powers (this combination has really been the "must" of the French bour-
geois at all times) and finally in safe loans to foreign States. From the mid-nineteenth
century onward, it became frequent for contemporaries to draw a contrast between
Germany, which was strong through its industries, and France, which was strong on
account of its banks. On the eve of the first world war, France was the second largest
exporter of capital in the world (Britain was the first). Since it lacked Britain's industrial
base, France became the first large advanced capitalist country to balance a permanent
trade deficit with dividend and interest earnings from capital invested abroad.
Industrialization was handicapped by a combination of factors that made the rate
of return low and the banks uninterested in investing. France's coal and iron resources
were poor compared both to England and Germany. In contrast with England, French
industry was not faced with any large external demand but it did nothing to create it.
Later it accepted without much difficulty a very reactionary approach, socially but also
economically, to the management of France's new colonial possessions. Since demo-
graphic growth was very weak and the peasantry offered only a limited market for
industry, domestic demand was not inherently dynamic either. The small and rather
conservative businessman studied by David Landes6 is largely a natural outcome of
this overall situation. Consequently, industrialization came about in successive bursts
on the basis of exogenous market pull in the form of government garanteed and bank
consortium financed demand, notably for railroad building (both at home and abroad),
ships, and arms. The feats of French engineering were principally those of large proj-
ects, involving large or very large amounts of capital (e.g., the Suez canal) and so
dependent on the banks. Many projects furthermore were already negotiated in con-
ditions where foreign alliances and the successes of French foreign policy were of
importance in securing markets for French financial-industrial consortia. The heart
of concentrated French industry was almost from the outset situated in the iron and
steel industry (cf. the de Wendel family group) and in products for the railways and
the Army (cf. the Schneider family group). In these critical areas French industrialists,
pressed for time, went abroad to England and Belgium and later to Germany to buy
their technology. They even recruited their foremen and skilled operators in these
countries.
Higher Education Priorities for the Training of Engineers and Experts
Although the responsibility for the divorce between science and industry, from the
1840s onward, and a weaker endogenous industrial demand for science and technol-
ogy in France than in the United Kingdom, Germany, and the United States, lies prin-
cipally with French industrial capitalism, the legacy of the Napoleonic period with
regard to the organization of teaching and of research in science and technology
proved in time to be a further obstacle (Gilpin, 1968; Fox and Weisz, 1980). When
Napoleon undertook the reorganization of French higher education between 1806 and
1811, he largely reestablished the centralized structure fashioned in the Ancien
Regime in keeping with his increasingly conservative policies in many areas. This
structure gave primacy to the training of experts as distinct from researchers and cre-
ators. This was provided in the professional schools, which have come to be known
collectively as the "Grandes Ecoles." The Ecole Polytechnique founded in 1794 pro-
vided a grounding in engineering and science, which was then built on the more
advanced ecoles d'applicationsuch as the Ecole de I'Artillerie et du Genie at Metz
(for artillery officers and military engineers) and the Ecole des Mines or the Ecole des
Fonts et Chaussees (for civil engineers). By the same principle, the Ecole Veterinaire
at Alfort produced veterinary surgeons while the Ecole Normale Superieure, resur-
rected by Napoleon in 1808, prepared the elite of the secondary teaching profession
for service in the lycees; subsequent institution building took place on the same model,
even when private industry took the lead. This was the case for the Ecole Centrale des
Arts et Manufactures, the Ecole Municipale de Physique et de Chimie Industrielles
(founded in 1882), and the Ecole Superieure d'Electricite (1894).
In contrast to the German polytechnic schools (Technische Hochschulen), the
French mode of engineering schools lacked in general the spirit of modern scientific
research. Until well into the twentieth century, most of them suffered from severe
parochialism. Though one had to have an extensive and broad mathematical educa-
tion to be selected for one of the engineering Grandes Ecoles, the training and curric-
ulum at each school were designed to train experts, civil servants, and managers for a
particular ministry. The student at the Grandes Ecoles learned the results of science
(and even here often with considerable delay since the curriculae divorced from
research had difficulties in keeping up with progress), and not the methods of science.
They became either abstract mathematicians or production engineers who applied
existing knowledge, rather than research engineers who could make substantial
advances in the state of the art. The gradual expansion of enrollment could not com-
pensate for the weakness of the science base.
Not surprisingly France's nineteenth century strengths in technology, notably
large-scale civil engineering, mirror the priority given to higher level technical educa-
tion. Throughout most of the nineteenth century, French engineers were among the
best and played an important part in the industrialization of Europe and other parts
of the world. Kindleberger's remark that "the products of the Ecole des Mines per-
formed their engineering feats in Algeria, Spain, Russia and elsewhere outside the Hex-
agon" (1976, p. 251) is of wider pertinence than the sole Ecole des Mines. The Suez
Canal, perhaps the foremost technological feat of the nineteenth century, stands as a
monument to the French engineering education of the time. However, the success of
French engineers abroad did not stop the divorce of large parts of industry at home
with science, nor did they guarantee the use of technology in the less progressive part
of the manufacturing system.
The Weakness of the Research Base and of the Universities
Initially, the Ecole Polytechnique had been a research center; its laboratories were the
best equipped in Europe and the best scientists were assembled to teach and do
research. Their number included such men as Monge, Fournier, Laplace, Berthollet,
and Chaptal (Dhombres, 1988). During the Revolution and the Napoleonic Empire
Polytechnique was a leading center of scientific research and the first to do systematic
basic research. Under the Bourbon Restauration, the Ecole Polytechnique then
became a military engineering school separate from other institutions of higher edu-
cation and run by the Ministry of the Army. It has kept this status to this day (see
Shinn, 1981,1984). Though it continued to be an important technological institution,
it disappeared as a center of scientific research and has only started to become one
again during the last 30 years.
The Ecole Normale Superieure was initially set up to train the teachers required
by the newly established system of secondary education. It passed through a precarious
existence during much of the first half of the nineteenth century but was able to build
up its research potential and develop its ties with the university in Paris. After Pasteur's
nomination as Administrator (1857-1967) and the changes he managed to impose
regarding the status of research, the Ecole finally emerged in the latter part of the cen-
tury as the best training ground of French scholars and scientists (Zwerling, 1980).
There the most eminent of French scientists and mathematicians were educated and
taught. However, the Ecole Normale with its 30 science graduates each year repre-
sented m uch too narrow a base on which to build a sound scientific edifice. Throughout
the nineteenth century and well into the 1930s the universities were stopped from pro-
viding the other indispensable component of this edifice.
The university system of the Ancien Regime had been severely condemned by
the Encyclopedia and abolished by the Revolution. Once he had brought the Revo-
lution to a halt Napoleon basically restored the old Ancien Regime University. His
centrally state-controlled system incorporated faculties of sciences and letters
roughly the functional equivalent of the old faculties of artsas well as the schools of
law, medicine, pharmacy, and Protestant and Catholic theology. The faculties of sci-
ence and letters, which might in principle have provided a liberal education in the
traditions of the Encyclopedic, scarcely functioned at all. By the time of the Bourbon
Restoration in 1815, fewer than half of the projected faculties of science had actually
been opened. Insofar as the faculties of science and letters did have a role, it was above
all that of providing examiners for the hierarchy of qualifications that regulated teach-
ing and other professional careers. The duties of a faculty professor were largely those
of a high-level functionary. And, unlike the "professional" faculties of law and medi-
cine, which were well attended, the faculties of science and letters had very few stu-
dents. Still less were they conceived as centers of research in the manner of the German
faculties of philosophy (Fox and Weisz, 1980).
Attempts to reform and strengthen the universities took place from the 1880s
onward as part of a wider policy of strengthening through education, the political and
social basis of the Third Republic. Student enrollment increased significantly;
throughout France new buildings and facilities were constructed; in an attempt to cre-
ate some degree of decentralization, university professors were granted a significant
degree of autonomy in the handling of internal affairsa transformation that was con-
secrated symbolically in 1896 by the creation of universities, composed of groups of
faculties in the same town. Although they never fulfilled the hopes invested in them
by the reformers, these universities reflected contemporary trends by cutting across
existing institutional divisions and embracing all areas of human knowledge. For a
brief period, at least, they seemed to pose a real threat to the dominance of the Grandes
Ecoles. However as Terry Shinn (1981, 1984) and other recent scholars have shown
these had sufficiently strong social and political support to offset the challenge and
maintain their domination.
Consequently, the universities played only a small part in the production of sci-
entific and technical personnel compared with the Grandes Ecoles. In the early 1900s,
for example, all the universities combined awarded only some 500 science degrees (i.e.,
only twice the number of admissions to the Ecole Polytechnique alone). The provin-
cial universities often found it difficult to survive, as Paris attracted both teachers and
students and the competition of the Grandes Ecoles attracted a good proportion of
students away from the Universities. As a result, the universities rarely offered a base
for research of any magnitude. Well into the twentieth century, the typical R&D lab-
oratory was the small personal laboratory that came with the professor's chair. There
the professor could pursue his personal inclinations with a few assistants, though the
research might not be at the frontiers of scientific advance and the laboratory might
be too small, ill-equipped, and isolated to be efficient (see Prestre, 1984 for the area of
physics). In some cases, even a scientist of renown might not even be lucky enough to
have such minimal conditions. Pierre Curie, for instance, had no research funds, no
personal laboratory, not even an office of his own; his important work on magnetism
was carried out primarily in a corridor, and his work with his wife Marie on radium
was conducted under extremely adverse conditions. On being proposed for the Legion
d'Honneur, Pierre Curie wrote to a friend: "Please be so kind as to thank the Minister,
and inform him that I do not feel the slightest need of being decorated, but that I am
in the greatest need of a laboratory." The Paris Radium Institute was established only
in 1910, 4 years after his death. Only with the establishment of the CNRS was this
situation modified.
Such is the nineteenth-century legacy regarding the organization of higher edu-
cation that France has battled with throughout the twentieth century and for which no
other solution has been found than expanding the number of engineering schools and
thwarting repeatedly the attempt of the universities to strengthen their position along
the model of other countries.
The Contradictory Balance Sheet of the 1920s and 1930s

The divorce of science from industry is expressed by the almost total absence of the
kind of industrial R&D laboratory developed from the 1890s onward in the United
States and Germany and so a weak French position in "science-push industries." By
contrast in sectors where technological development took the form of pragmatic, step-
by-step innovations as in automobiles and aeronautics, French inventors and entre-
preneurs were very active. Up to World War II the French automobile industry was
the second largest world producer. Panhard (today only a military firm producing
tanks) and Peugeot date back to 1890 and Renault to 1899. Michelin produced the
first air-tube tire in 1895. In aviation again, Frenchmen flying French planes held
world records on a par with U. S. rivals up to World War II. Farman and Breguet were
large world exporters of planes and Gnome and Rhone, Hispano, and Renault of air-
plane engines between the two wars. Technological accumulation by French firms in
these areas is now secular and account of course for the rapid recovery after 1950. Dur-
ing the 1920s and 1930s, however, the progress and growth of these two industries went
hand in hand with very slow overall economic growth and a malthusian climate largely
strengthened by the negative demographic consequences of the huge human losses of
the first world war.
But the 1930s were also characterized by a radical and dramatic divorce between
France's well-developed automobile and aeronautics industries and the military doc-
trine upheld by the dominant ultraconservative French officer corps. While Breguet
planes were competing with the Wrights for transatlantic records, French generals
were building the Maginot line and a large fraction of the French bourgeoisie showed
itself much more concerned with the dangers of bolshevism than those of nazism. The
crushing defeat of 1940 followed by the close collaboration of the Vichy government
with Hitler were the inevitable outcome.
PHASES IN THE BUILDING OF THE SYSTEM
Along with economic planning, an active industrial policy, and large public invest-
ments in basic infrastructures, France's vigorous institution building in science and
technology and large investments in R&D must be seen as one of the central instru-
ments of the country's fairly spectacular economic and political recovery following the
end of World War II.
The Context: Recovery as a Second Tier World Power
In 1945, France, having represented a major political and military power for three cen-
turies, had suffered a crushing defeat in May-June 1940 at the hands of one of its two
major historical rivals, Germany. France owed its quite marginal presence in the camp
ofthe 1945 victors to the decision taken in 1940byajuniorgeneralandjuniorminister
of the governmentCharles de Gaulleto break with authorities and proclaim the
establishment of a provisional government that had initially been devoid of almost any
political and social support. Although France was granted the status of an occupation
power in Germany, after 1945, was granted a seat in the Security Council ofthe United
Nations, and was allowed to retrieve most of her rights as a colonial power (a mixed
blessing at the very least as the Indochina and Algerian wars of independence were to
prove), she was excluded from the negotiations and agreements at Yalta and Potsdam
and was not a party either to the discussions leading up to the all important Bretton
Woods agreement.
In 1945, France's industrial base was small and often extremely backward tech-
nologically. The industrial base, but also the coal and iron mines and the basic eco-
nomic infrastructures bore the scars of two earlier decades of chronic underinvest-
ment, the impact of the Slump of the 1930s and the destructions of the war. But the
state of industry in 1945 also reflected secular Malthusian tendencies on the part of a
large fraction of the owners of capital and landed property.7 The defeat in June 1940
had been the result of political and military conservatism of the French bourgeoisie,
which had also taken the form of a major technological gap with Germany. France
had had a number of brilliant scientists, but up to 1939 they had generally been almost
completely deprived of adequate resources to carry out their research. Whatever
French industry knew about industrial management was the result of a study of Tay-
lorism and Fordism. Finally 40% of the French population was still engaged in agri-
culture. Although farmers carried out their activity on a fertile soil, with the backing
of multisecular traditions in food and wine artisanship, the backward features of much
agricultural production meant that productivity and output were very low.
Thirty years later (i.e., by the mid-1970s) France had succeeded in attaining the
rank and attributes of a modern industrial power. Considered as a fraction of the total
1973 OECD GNP, France with 8% had long overtaken Great Britain (5.4%). The
active population in agriculture was down to less than 10%, while yields in key prod-
ucts such as wheat or milk production had tripled from 1945 levels. French industry
had in many sectors been restructured and had experienced quantitative but also qual-
itative growth. France's infrastructure in terms of railroads, electricity supply, urban
transport, and telecommunications system had been overhauled and modernized and
made similar infrastructure in many other countries look shabby, outdated, and
underequipped. Deprived of U. S. and U. K. nuclear technology by decisions dating
back to 1943-1944 and confirmed by Roosevelt in 1945-1946, France had succeeded
in building her own nuclear industry including nuclear weapons, however useful or
meaningful this might be. She had built one successful commercial aircraft, the Car-
avelle, and with the United Kingdom the supersonic plane Concorde, which was an
unsuccessful commercial venture, but nonetheless a technological feat. She had taken
the initiative of assembling a number of European partners to build Airbus and stay
in the business of large transatlantic airliners. She had, again unwisely, moved back
into large-scale level albeit second tier production (as compared with the United
States) of weapons, missiles, and military aircraft. In tandem with Germany she had
established a position as a driving force within the European Community and was
demonstrating that the United Kingdom had made a mistake in not signing the Rome
Treaty. In space she was actively rallying the Europeans within the ELDO and ESRO
organizations and persuading them that they could not let the United States establish
a monopoly for satellites and launchers. France would be a major founder of the Euro-
pean Space Agency and subsequently of Arianespace.
So, in almost every way the comparison between 1945 and 1975 shows a very
different situation and a tale of considerable success. The process of growth and trans-
formation behind this performance was closely geared to large investments in R&D
and two phases of intensive S&T institution building.
Institution Building Immediately after the War
The first phase of institution building took place immediately at the end of the war.8
In a significant and spectacular manner it began with the creation of a capacity for
R&D and production in nuclear energy and subsequently for military purposes,
lodged in a major agency, the Commissariat a FEnergie, CEA (October 1945). It also
included the reorganization and expansion of the CNRS (November 1945) and the
creation (1945) under the Ministry of Post, Telegraph and Telephone of the National
Centre for the Study of Telecommunications (CNET). Among the numerous techni-
cal agencies also established at the time under the Ministry of Defence, the most
important was the National Office of Aeronautical Studies and Research (ONERA),
which was given a mandate both for military and civil R&D. The major public agen-
cies in the industries which had just been nationalized in energy and basic infrastruc-
tures followed suit.
The same period saw the creation of the Office of Overseas Scientific and Tech-
nical Research (ORSTOM) (1944) responsible for doing research of interest to the
French colonial empire, primarily of course in tropical agriculture; the reestablish-
ment of the National Institute of Health (INH) by the Ministry of Public Health and
Population (1945) with the task of ensuring "the direction, organisation, and co-ordi-
nation of scientific medical research," and the reestablishment of the National Insti-
tute of Agricultural Research (INRA), originally set up in 1921, suppressed in 1934,
and finally reorganized in 1946 with two large centralized agricultural research com-
plexes near Paris, in addition to several regional centers.
The most portentous step was of course the decision to move into nuclear
research and production. This was subsequently to lead France into one of the largest
nuclear energy production programs in the world. The responsibility for the initial
move limited in principle to "civilian" industrial objectives, was shared mainly
between the Gaullists and the French Communist Party (acting in this as in all cases
after consultation with the Soviet Stalinist party). The alliance of these two forces pro-
vided the political but also the scientific foundation of France's entry into the military
nuclear field. This was decided around 1955 after the first commissioner Joliot Curie
had been ousted, before being fully implemented by de Gaulle after 1960. It was
opposed by the CP for reasons related to the Cold War, but the conjunctural character
of this opposition was shown later when the "force de frappe" received CP support at
the end of the 1970s. The building of the CEA's central R&D laboratory and pilot
plant capacities at Saclay from 1947 onward symbolized the start of a transformation
of French scientific institutions. In place of the small, poorly equipped laboratories of
individual professors that had characterized French science, large scientific resources
were brought together in a complex of modern laboratories with teams of researchers
and supporting technicians. In a country where no large firm had yet set up a major
industrial R&D laboratory based on the U. S. and German model, the building of
Saclay was France's first real step into twentieth-century fundamental and applied sci-
ence. It is only the more regrettable that it should have been for the atom.
The CNRS
Although it was formally founded in 1939 as a belated result of the political interest
for science shown by the 1936 Popular Front government led by Leon Blum, the exis-
tence of the National Centre for Scientific Research (CNRS) must really be dated from
its reorganization in 1945. Its initial mandate included the responsibility to develop,
orient, and coordinate all French science. Although this objective was never to be
achieved, the CNRS was to have a profound impact on the organization and devel-
opment of basic and long-term research, the availability of scientific and technical per-
sonnel, and the overall support of science.
Through the establishment of numerous laboratories and research facilities that
it administers, the CNRS has provided France since 1945 with an infrastructure of
research institutes similar to that created in Germany after 1911 by the Kaiser Wil-
helm Society (today the Max Planck Society). In particular, the CNRS has been able
to establish and administer laboratories in newer fields of research that could not be
placed within the French university structure. Within the French system the CNRS
has played the role assigned in other systems to industry and private foundations, that
is, support of university research. It has supported the otherwise very weak university
research by seconding researchers to university laboratories. Although these scientists
remain attached to CNRS in terms of promotions and salaries, they are fully integrated
into the university laboratory and its research program. The CNRS also supports uni-
versity research by providing the numerous services, assistants, and equipment
required by scientists, which neither the Ministry of Education nor the university bud-
gets has supported adequately. It supports colloquia on scientific subjects and finances
the attendance of French scientists at international conferences. It subvenes scientific
publication and the purchase of instrumentation and provides scientists and techni-
cians with a wide range of services including documentation, specialized training, and
assistance on patentable inventions.
Institution Building in the First Phases of the Fifth Republic (1958-1966)
During the 1945-1958 period, the production and diffusion of technology were driven
almost exclusively by the State and innovation capacity lodged principally in nation-
alized or publicly owned firms. In the second phase of institution building, which took
place after the Algiers military coup, the fall of the Fourth Republic, de Gaulle's return
to power, and the setting up of the Fifth Republic, innovation continued to be driven
strongly by the State. But policies start to be enacted to lodge at least a part of the
innovative capacity within the industry's national champions (e.g., the large public or
private firms with which the State has decided to build up in order to work in close
partnership with them). This is the period during which the overall framework of the
State-industry relationship is established by the Commissariat au Plan, which reached
its highest point of power and influence under de Gaulle in the 1960s.
Major institutional decisions in S&T first concerned space research with the cre-
ation in 1959 of a Committee for Space Research, which made proposals for a 6-year
program. While all the other European countries were hesitating over embarking on
a space program, political considerations, including de Gaulle's personal vision of
France's and Europe's place in the world, led to the decision in 1962 to launch the
proposed program and to set up a national organization for space research, the
National Centre for Space Studies (CNES). The decision was also made, however, to
adopt an organizational set up different from the CEA or the CNET and to involve
public and private firms in the program from the outset by contracting out a large part
of the R&D to the business sector.
The same pattern of State-industry relationship, based on procurement and often
involving the same firms, was adopted for the arms industry, following the full scale
reorganization of military R&D undertaken in the early 1960s (Kolodziej, 1987). A
very strong body with quasiministerial autonomous power, the Ministerial Delegation
for Armements (DMA), which later became the DGA, was set up in 1961 within the
Ministry of Defence, along with a Directorate for Military Research and Testing
(DRME). Military R&D was moved out of the State sector and reorganized on the
basis of R&D procurement to industry. The only exception was the military atomic
program, which retained a high degree of autonomy within the CEA's Directorate for
Military Applications and did not use firm-based R&D procurement. Although there
is considerable uncertainty about the exact cost of the nuclear military program (the
"force de frappe nucleaire"), it is generally considered that between 1959 and 1966 the
major atomic military, aeronautic, and space programs accounted for about 65% of
public R&D expenditure. These programs were launched and were all in answer to
political objectives coupled in the case of nuclear sectors with a subordinate industrial
objective (Papon, 1975, 1978).
After 1965, the problems of the French computer and data-processing industries
brought about a further development and yet another pattern of the State-industry
alliance. Faced with the difficulties of the French-computer industry and spurred once
again by a hopeless U. S. embargo decision, the French government, which had let the
Bull Machine Company come under U. S. control, launched a new "Major Program"
in the field of data processing: the Plan Calcul. This plan took shape in 1966 with the
set-up of a new private data-processing company, the International Data-Processing
Company (CII), which received huge financial aid from the State. In addition, the State
set up an Institute for Research into Data-Processing and Automatism (IRIA) and
gave further financial assistance to the French components and peripheral equipment
firms. Although IRIA is still in existence, the firms that the State has attempted to
upgrade into viable competitors on world markets have suffered countless misadven-
tures and undergone innumerable metamorphoses.
Developments in the 1970s and 1980s and the 1982 Reforms9

With respect to the overall structure and working of the French innovation system, the
1970s and 1980s have essentially brought about only shifts in emphasis in the area of
overall R&D resource allocation and the location of entrepreneurial capacity, along
with a clearer spelling out of features that were already contained within the system as
it had been built in the two previous phases. Two developments warrant special atten-
tion.
The first has been the development, based on institutions built during the earlier
periodsthe DMA-DGA and the DRET within the Defense Ministry and also the
CEA, the CNES, and the CNET, of a large military-industrial complex, which encom-
passes those parts of the space program that fall outside the European programs man-
aged by the European Space Agency and the operations of the Arianespace consor-
tium, a part of the activity in telecommunications, and the efforts made to maintain
a computer and components industry. The industrial elements of the complex now
represent France's most powerful and at least in appearance most successful high-tech-
nology corporations, in particular Thomson, Aerospatiale, and Matra. These firms are
almost indifferently "private" firms (as in the case of Matra) or "public firms" (as with
the other two).
The second novel but totally logical development concerned the steps taken first
to build new links or "bridges" between the research capacity accumulated within the
public sector and all firms that are ready to take the innovations to market, and later
to authorize and even force public research centers to move downstream toward the
market and to become "technological entrepreneurs" in their own right. The first cat-
egory of measures included the creation of ANVAR, which is a fairly classical type of
agency for technology transfer from government and university research laboratories
to industry. The second category of measures, developed during the Pompidou (1971-
1974) and Giscard (1974-1981) administrations, had to wait the Socialist-CP govern-
ment of 1981 to be written up and pushed through. They entailed the introduction of
a number of breaches in public sector characteristics of major fundamental and/or
applied government research laboratories in CNRS, INRA, INSERM, and so on.
The changes are still far from a full scale privatization of public sector R&D, but
they certainly represent a step in that direction. The status of the R&D laboratories
was changed (in 1982 for CNRS, in 1983 for INSERM and INRA) from administra-
tive public institutions to a new generic type of status with some attributes of private
law, the "etablissement public scientifique et technique." Under this new status lab-
oratories have been empowered to establish subsidiaries, acquire shares, and seek
cooperation around specific projects with scientific and industrial partners in public
interest groups (GIF) and scientific groupings (GS). These possibilities give the major
agencies more incentive to involve themselves in exploiting and marketing their inno-
vations. In practice, marketing will generally be undertaken by private law subsidiaries
that can more appropriately act as entrepreneurs than the research agencies them-
selves. An example of this was the creation, in 1983, of a firm called "Midi-Robots"
to develop and market products originating directly from the work of the CNRS auto-
mation and systems analysis laboratory (LAAS) in Toulouse. The firm set up business
partnerships with the government aircraft laboratory ONERA and the aerospace
firms. The 1982 arrangements can also take practical form in multipartite cooperation
contracts, or in the formation of an embryonic industrial undertaking and marketing
activity. The GIF is particularly suitable for setting up technology transfer centers and
joint CNRS/industrial teams cooperating to develop industrial prototypes. One of the
first examples was the "time frequency" technology GIF, in which three leading CNRS
research laboratories came together with Thomson to develop the very high-precision
timepieces required for airborne navigation and telecommunications networks.
THE OVERALL STRUCTURE OF THE R&D SYSTEM
An R&D system differs on several scores from an innovation system. Reported formal
R&D expenditures are only a part of the innovation-related outlays made by firms.
Formal R&D data ignore the complex processes of technological accumulation
whereby tacit, partly uncodified knowledge is built up and transmitted from one gen-
eration to the next within institutions, firms, and sometimes whole industries. Formal
R&D captures nothing of the linkages between organizations, the feedback processes,
and also the alliances and relationships of power between agencies and firms. An R&D
system is at best a poor proxy to an innovation system, but since the R&D data are the
only reasonably coherent and comprehensive data we possess, we have no choice but
to use them before examining in the following sections some of the systemic relation-
ships that give a better idea of the reality of the innovation system.
In the case of France the overall structure can be approached through a look at
the main aggregates in terms of funding and execution shown in Figure 6.1 published
by the Ministry of Research and Technology.10 This can be considered only a first
approximation, since funding arrows cannot capture the transfer of readily applicable
technology, which occurs on a large scale between the State sector laboratories and
business enterprises as a result of the policies just discussed.
The main structural characteristics are as follows:
1. The government funds approximately 50% of R&D and industry about 44%, the
rest coming from foreign sources.
2. About 55% is executed in industry and 43% in the "public sector" (the lower left
hand block), which includes both government research laboratories (about 27%)
and higher education research (about 16%), which includes funds provided by the
CNRS and general university funds as reported in OECD statistics, which are of
course basically aimed at teaching and only subordinately at research.
The Allocation of Nonmilitary R&D Funds to the State Laboratories

The way in which the official budget documents and other reports are presented makes
it extremely difficult to understand the precise pattern of resource allocation of funds.
In 1985, a serious estimate was made by one of the Associations of engineers and is
given in Figure 6.2. Several observations are required for a full understanding of the
data.
Figure 6.1. Sources of funding and sectors of performance.

Figure 6.2. The breakdown of the "budget civil de R-D" administered by the ministry of
research and technology: appropriations to government laboratories and support schemes to
industry. Source: Bulletin de I'Ademast, No. 10, January 1985.
1. The data cover the appropriations received through the Ministries of Industry and
of Research and Technology. Following the moves made under J. P. Chevenement
toward a unified system outside the military sphere, this is in fact very comprehen-
sive and misses only the appropriations provided to the CNET in telecommuni-
cations. In 1985, they represented 3.5 billion francs (i.e., approximately the same
asforCNES).
2. The appropriations made to CNRS include the expenditures it makes in its own
laboratories and the funds it allocates through its own commission to university
associated laboratories.
3. The CEA remains the single largest government R&D center, with a civil R&D
subsidy 40% higher than that of CNES or CNET, to which a large military appro-
priation must be added.
4. If the appropriations made for the CEA, the CNET, the civil aeronautics program
administered by the ONERA (which includes France's contribution to the R&D
support given to the Airbus consortium), and the informatics/electronics sector are
lumped together, then even independently of the military R&D allocation pattern
that has of course exactly the same thrust, one finds an overwhelming bias in favor
of the nuclear, aeronautics, space, telecommunications, and electronics sectors to
the detriment of the chemical, biological, and life science based sectors (INRA,
INSERM, IFREMER) as well as to that of the machine tool and robotics industries
and other small firm dominated sectors [which get at best a little scattered support
through the FRT (Fonds de Recherche Technologique)].
industrial R&D: The Available Data and What it Shows
We must now turn to the lower right-hand block in Figure 6.1. Industrial R&D, or
R&D carried out within firms, remains significantly weak. Whereas French GERD
represented 2.31% of GDP in 1987, R&D carried out in the business enterprise sector
(both publicly and privately owned) represented only 1.38% (as compared with 2.11 %
in Japan, 2.25% in the United States, and 2.49% in Germany). As far as the R&D actu-
ally financed by firms is concerned the percentage is even lower1.05% of GDP in
1987.
These figures are confirmed by survey data. In 1987 the Ministry of Research and
Technology reported that only 1990 firms were carrying out R&D as defined by the
OECD Frascati Manual. Since the Ministry includes in its survey a further 50 technical
research centers that are financed collectively by firms in different industrial branches,
data are provided covering 2040 firms and research centers. As a point of comparison,
for the same year the industrial census reported 90,000 firms in manufacturing and
services employing 10 or more people.
According to the survey the 1990 firms reporting formal R&D account for a third
of industrial employment and over half of industrial output. The survey reveals, how-
ever, a dual situation in this regard: in one group of industries, which includes food
processing, building materials, metallurgy, and textiles, firms doing R&D account for
no more than 20% of industrial branch output; in another group, which includes
energy production, electronics, data processing, aircraft, automobiles, chemicals, and
pharmaceutics, such firms account for over 75% of industrial branch output. These
branches account for about 87% of all R&D (Fig. 6.3).
Figure 6.3. The origin by ministerial departments of direct R&D support to firms.
Figure 6.4. The breakdown by recipient industries of direct R&D support to firms.
Furthermore, within the group of 2040 firms and centers reporting formal R&D
expenditure, effective R&D activity is concentrated within a very small group of firms.
In 1987, only 7% of the organizations concerned had an R&D staff with more than 50
science and engineering research workers. A group of no more than 150 firms accounts
in fact for the bulk of French R&D. According to the survey these firms carry out 75%
of R&D and receive over 90% of direct government support for R&D. The direct sup-
port for industrial R&D is highly concentrated. Two Ministries, Defense and Post and
Telecommunications, alone account for 85% of the funds channeled to firms and two
sectors, aerospace and electronics, receive 83% of the total (Fig. 6.4). Here we are talk-
ing exclusively of reported R&D. procurement and not of the other, possibly large,
sums (which go unrecorded publicly and one quite probably very hard to really keep
track of seriously), which reach industry indirectly through the partnerships that the
large State laboratories all establish with the firms that they deem to be their opposite
number in industry.
At the other end of the spectrum, among the 1990 firms reporting formal R&D,
some 1450 firms employ less than 10 fully trained scientists and engineers. These firms
carry out less than 8% of the recorded formal R&D and are the beneficiaries of no more
than 3% of government R&D subsidies. As a result, the concentration of R&D by
industrial branch is necessarily extremely high. In 1987, eight branches accounted for
over 85% of total R&D expenditure: electronics, 23.2%; aircraft, 17.8%; automobiles,
10%; chemicals, 10%; Pharmaceuticals, 7%; energy production, 7.2%; data processing,
5.2%; and heavy electrical material, 4.7%. By contrast industrial branches that account
for a significant part of French exports (agriculture and food processing) or still rep-
resent fairly important components of French industrial GDP (metallurgy and metal
working, textiles, machinery) account only for a very small fraction of industrial R&D.
Research in the Higher Education Sector
The third component of the R&D capacity is the one lodged within the higher edu-
cation sector, with its dual structure of universities and separate engineering schools.
Within universities R&D is in general the privileged and closely guarded domain
of specialized laboratories that have tended again to become partly divorced from
teaching. Significant research is a feature of a handfull of universities only; the Louis
Pasteur University in Strasbourg, which builds on the legacy of the German university
system, Orsay (Paris IX) with its close relationships with Saclay and many CNRS lab-
oratories, Grenoble, and Toulouse. Despite the proclamation by the 1968 and later
reforms of formal university autonomy, the universities as such still have little say
about research. The professors who are also heads of laboratories are jealous of their
independence. They take it on themselves to ensure that teaching posts, premises, and
operating budgets are allocated by University Councils along lines that represent
essentially a continuation of the status quo. The more enterprising among heads of
laboratories will negotiate directly with outside sources (public bodies and the CNRS
in particular and now increasingly with industry) for the extra funds needed to develop
high-quality research. The swamping of the universities by the massive surge of student
enrollment under the two-fold action of demographic factors and of a call for democ-
ratization accompanied by totally insufficient resources for staff or premises has accen-
tuated the tendency for the larger and better connected laboratories to try and isolate
themselves as best they can from the situation in which the universities have been
placed since the late 1970s.
The situation today is critical. Universities are part of the public sector, salaries
are aligned with those in the civil service, and in a period of fiscal crisis for the State
university buildings, their construction, repair, and day-to-day running costs are deter-
mined by public budgets. On other vital matters relating to the organization of teach-
ing and research, the so-called autonomy of universities is purely formal; everything
related to the administration of human and material resources falls under the strictest
rules commanding public administration. The traditional conception of "public ser-
vice" forbids universities to establish any kind of student entry selection; the net out-
come is to negate the notion of "public service" by dramatically lowering the quality
of teaching and achieving "selection through failure" after 1, 2, or 3 years.
The larger and better connected laboratories can count on the financial and mate-
rial assistance of the CNRS and also increasingly on resources coming through Euro-
pean Community (Esprit, BRITE, RACE, etc.) programs and Eureka projects. But in
many cases they are affected by an insufficient supply of young researchers stemming
from the overall decline in the university system; they often consider themselves lucky
when they can fill some gaps with foreign students on European or foreign country
grants. Unlike the CNRS's own laboratories, the university laboratories (even those
having the "associated with CNRS" status) have not been granted the EPST status
discussed above so as to keep them under the Ministry of Education's strict authority
and control. They remain hampered in their efforts to establish R&D contracts with
industry. Discussion in the 1970s and 1980s about the need for closer university-
industry relationships has not yet freed university research from the fetters of tradi-
tional state accounting procedures. The establishment of R&D contracts with firms
forces laboratories to circumvent the law and to hide part of their resources from offi-
cial university accountants and tax inspectors.
The other component of the education system, the engineering "Grandes

Ecoles," compensate for some of the weaknesses of the university system as far as pre-
serving the level of education is concerned, but not with respect to the needs of indus-
try in terms of numbers of trained personnel and less still with respect to those of long-
term basic research. The small size of the engineering schools and their very tough
selection procedures have guaranteed the level of education, in keeping with the over-
all elitist requirements of the system. Lower level or "less noble" engineering schools
were created during the 1960s and 1970s, but the assessment is that with a supply of
about 12,500 engineers a year the system does not produce the supply required by
French industry. This is thought to represent one reason for the plight of many seg-
ments of small sized industry. The laboratories of the Grandes Ecoles are much better
endowed than those of the universities and the relationships with industry are natu-
rally much stronger, but the laboratories are nearly empty. In 1982 an official report
found that of 10,500 diplomas awarded, only 500 were engineering Ph.D. R&D, in
particular basic or long-term research, remains weak, and in some instances is still
marginal within the engineering schools. The entry selection process is still based
mainly on mathematics and this discipline continues to determine the teaching cur-
riculum in many schools. In 1987, a survey on scientific and technical personnel found
28,600 trained full-time scientists and engineers doing research within the higher edu-
cation sector, but attributed only 1400 to the engineering schools. A 1985 study found
5600 people doing research within these schools (not full time), but noted that two-
thirds were either from other laboratories or agencies or under contract and that only
one-third were members of the normal teaching staff.
Agencies and Institutional Mechanisms for Technological Diffusion
An overview of the R&D system would be incomplete without reference to the agen-
cies and institutional mechanisms created to disseminate scientific and technical
knowledge and the technologies produced in the large public laboratories, the CNRS,
and the universities.
One mechanism was discussed at the end of the previous section, namely the attri-
bution of EPIC states (public agency with private law attributes) to certain research
centers and the possibility that they could set up joint ventures with industry. But tech-
nology transfer is also the mission of several specialized agencies, both at the national
and the regional level. The most important of these is ANVAR (Agence nationale pour
la valorisation de la recherche), which manages a portfolio of patents (it files 1000-
1200 applications at home and abroad on average a year) and finds industrial partners
for CNRS and university laboratories. In 1982 a new institution was created, the
CRITTs (Centres regionaux d'innovation et de transfer! de technologies), which are
joint venture organizations with private and public (mainly regional) financial partic-
ipation and the job of enhancing regional innovation-related networks between lab-
oratories, firms, and local governments. They can be specialized (as in Alsace in the
area of new materials) or general. One finds a wide mix of financing and effective par-
ticipation, which evolves over time as a firm network starts to get rooted. These insti-
tutions are obviously of great potential importance. The finance they command is,
however, still absolutely marginal (see the amounts for ANVAR, the FRT (Fonds de
la recherche technologique), and "other" in Fig. 6.2) in comparison with the funds
channeled to the core of the system, which we will now examine.
QUALITATIVE FEATURES OF THE HIGH-TECHNOLOGY SUBSYSTEMS
The analysis of the formal R&D data yields few insights on how the innovation system
really works. Given the way the system was established historically, the State occupies
a pivotal position in most R&D intensive industries. Generally it relies on industrial
partners to take the innovations to the market, but in some cases this task is assigned
to the commercial arm of a government department. The influence of distinct and
possibly rival government departments explains why the State influenced innovation
system is divided vertically into strongly compartmentalized sub-systems.'' Before dis-
cussing a number of these separately, we must examine the social foundations of the
State-industry alliance.
Business and State: The Blurring of the Public-Private Distinction
An important feature of French political and social history since the end of World War
II has been the progressive establishment between the State and the oligopolistic core
of public and private industry of a common view of the ways of attaining economic
growth, modernization, and military independence through autonomous arms pro-
duction, so preserving France's "rank in the world." The events of the 1930s and 1940s
brought about an understanding by the most determined group within the French rul-
ing class (of which de Gaulle became the figurehead) that the page should be turned
with respect to the classical liberal division of tasks between the State and industry.
The new approach was that private capital should rally round the State, accept its help,
and use it as an instrument for industrial restructuring and the channeling of financial
and human resources to priority areas.12
The process was prepared during the Fourth Republic by setting up the Planning
Commissariat, the concentration of responsibility for Finance and the Economy in
one extremely powerful ministry, the nationalizations, and the setting up of the large
R&D agencies. Under the Fourth Republic, the process was still hampered by two
main factors: (1) the lingering of an anachronistic approach to imperial status and
colonial power and the support of vested colonial interests backed by the Army, and
(2) a political system based on the supremacy of the Parliament and the ploy of tra-
ditional conservative political parties and leaders (Pinay, Laniel) who were still in a
position to defend the interests of the colonial lobbies and of small family business and
the "shopkeepers." Between 1958 and 1961 de Gaulle put an end to this with the clos-
ing of the Algerian war, the purging of the procolonial faction within the Army and its
reorientation to nuclear "force de frappe" strategical thinking, and, of course, the
establishment of the Fifth Republic, which, under the label of "democracy," is today
the nearest thing to "Republican Monarchy," marked by an unequaled concentration
of power in the executive.
The decade that followed de Gaulle's return to power was marked by an energetic
drive toward industrial restructuring and concentration. Given the highly fragmented
character of most sectors of French industry, the building of much larger industrial
groups and the concentration of industrial plants were, from the standpoint of French
capitalism, a necessity. At the end of the 1960s, acute U. S. observers could already
note that the "important force shaping corporate strategies and strategic planning is a
close relationship between business and the State" (McArthur and Scott, 1989, p. 8).13
S. Cohen (1969, p. 51) reached much the same conclusion: "The economie concertee
is a partnership of big business, the state, and, in theory though not in practice, the
trade unions. The managers of big business and the managers of the state run the mod-
ern core of the nation's economymostly the oligopoly sectors."
Subsequent performance, in the late 1970s and 1980s, in particular in the area of
industrial exports, suggests that the "national champion" policy also contributed to
strengthening other important parasitic traits of French industry and finance. The task
of remodeling French industry undertaken by Gaullist governments in the 1960s was
completed in 1982-1983 by the additional restructuring and reorganization of cor-
porate frontiers, which took place in 1982-1983 at the time of the so-called "nation-
alizations" of the first Mitterrand government. Aside from their purely ideological
purpose of allowing the first socialist and CP government to seem to be "on the left,"
the basic purpose of 1982-1983 "nationalizations" was to give the French State and
administration renewed authority and power to undertake the widespread restructur-
ing of several major industries and, if necessary, to tread on the toes of recalcitrant
capitalists while offering others (in steel and parts of petrochemicals) extremely, not to
say scandalously, favorable financial exit conditions. In the course of the 1970s and
1980s, the oligopolistic core of French industry gained greater and greater ascendency
in the State-industry partnership, as well as greater freedom of action. Price controls
were suppressed in the late 1970s and early 1980s; the first Mitterrand governments
planned and pushed through a full scale modernization, reorganization, and deregu-
lation of the financial system including the Stock Exchange (the Bourse), which has
proved to be particularly beneficial to the oligopolistic core of industry and banking;
legislation was passed to allow public sector and nationalized firms to internationalize
and to form joint ventures in France with foreign partners.
France's Unique "Power Elite": The "Grandes Ecoles"

One of the most effective instruments of this near symbiosis between a State apparatus
receptive to the arguments of oligopolistic industry and large firms marked by the
reflex of turning to the State for support has been the system of elite production
through the "grandes ecoles" [in particular the Ecole Nationale d'Administration
(ENA) and the Ecole Polytechnique] and the "grands corps." "The key fact in French
planning," as noted by Granick (1964, p. 147) in the 1960s, "is that the same type of
men are sitting in the management and civil service posts in this cartel: men of the
grandes ecoles, present and former civil servants who consider themselves techno-
crats." The large industrial enterprises in France, the nationalized industries, and the
public sector are to a very large extent run by people who come from the same schools
and are invariably members of the "grands corps," notably the Inspecteurs des
Finances and the engineering corps. By "corps" is meant a highly trained expert per-
sonnel who have successfully entered Polytechnique and go on to one of the select
engineering schools, in particular the Ecole des Mines, the Ecole des Ponts et Chau-
sees, the Ecole Nationale des Telecommunications, the Ecole Nationale Superieure de
1'Aeronautique et de 1'Espace (SupAero), and the specialized schools where the Ingen-
ieurs de l'Armement receive their last phase of training.14
A study of industrial managers in France made in the 1970s concluded that "the
most characteristic trait in the careers of the sample studied is the frequency of the
PDG (President Director-General) coming from the public sector" (Monjardet, 1972).
The study concludes that this type of career cannot be viewed any longer as simply the
seduction by the private sector of an elite whose sense of public service has been weak-
ened. Rather, it has to be seen as "an essential element in the training (acquisition of
competence, or relations) of the industrial managers in France." A further recent study
of the same phenomena has concluded unambiguously that this is the case: today the
normal path to the top banks and industrial firms involves final schooling at Polytech-
nique or ENA; entry into a "grands corps" and finally a passage through the top
administration or the private office of one or several Ministers (Bauer, 1987). Thus, at
the heart of each of the major innovation subsystems is a group of managers, research
directors, and private office Ministerial advisers belonging to the same "corps":
"Mines" and "Fonts et Chaussees" in the case of electronuclear power, the Corps des
ingenieurs des "telecom" in the case of telecommunications and space, the graduates
from SupAero in the aerospace industries, and the Corps des ingenieurs de 1'armement
elsewhere in the arms industry. These people possess what Salomon (1989) calls a "life-
long passeport" to the highest and best paid jobs, within a system in which severe busi-
ness failure almost unvariably goes unpunished.
SOME CHARACTERISTIC HIGH-TECHNOLOGY SUBSYSTEMS
In this section, we turn to some characteristic examples of the partnership between the
State and the oligopolistic core of large public and private firms operating in the high-
technology sectors other than Pharmaceuticals and the new branches of the chemical
sector. Figure 6.5 sets out the government agencies and their main industrial counter-
parts. Two sets of government agencies are indicated: those involved in technological
activity either as major laboratories (CEA, ONERA), as R&D procurement agencies
(DRET, CNES), or as both simultaneously (CNET), but also those politically respon-
sible for technological strategies. The Delegation Generate a 1'Armement (DGA)
appears at every point in the top line, indicating the key role played by the military
and the increasing influence they are now likely to exert even in telecommunications
and nonmilitary space as a result of the current post-Gulf war reorientation of mili-
tary-strategic priorities to space observation and telecommunications systems.
Figure 6.5. Some key relationships at the heart of the high tech innovation system. Source:
Adapted from Serfati (1991).
The Military High-Technology Innovation Subsector
Although this subsector is more recent than the one in nuclear technology, its strategic
position implies that we should begin with it. Expenditure on military R&D is esti-
mated as representing at least a third and probably well over 35% of French public
expenditure on R&D, meaning that the military subsystem of innovation is one of the
largest. Military R&D is divided into three fractions: (1) the nuclear R&D undertaken
at the CEA (approximately 30% of the total), (2) the R&D carried out in the State-
owned traditional arsenals and arms manufactures (now less than 10%), and (3) the
R&D commissioned to industrial firms through R&D contracts and procurement
(over 60%), which amounts (as shown in Fig. 6.3) to about two-thirds of the total of
the State's direct support to industrial R&D.
The high-technology arms industry is organized through a tightly knit relation-
ship between private and public industry and the General Delegation for Armement
(DGA). The head of this institution is a "nonpolitical" and hence a stable appointee,
who is Considered to be as powerful as many Ministers. The firms with which the DGA
works as prime contractors are all among the largest in French industry. These firms
do not compete among themselves. R&D contracts and arms procurement take place
without tendering, on the basis of a functional division of tasks established by the DGA
as part of its "industrial policy for arms." As a result, the French arms industry is in
fact organized around a set of monopolies or duopolies: for planes and helicopters,
Dassault and Aerospatiale; for aircraft engines, SNECMA; for missiles, Aerospatiale
and Matra; for electronics, Thomson and Dassault-Electronique.
Given the resources at the disposal of the Ministry of Defence for R&D and arms
purchases, the gradual enfeeblement of the Plan as an institutional mechanism and
the persistent weakness of the Ministry of Industry, the DGA is now the strongest body
in French industrial policy. As Kolodziej (1987, p. 274) has shown, the DGA is in a
position through its "multiple opportunities in allocating its contracts and deploying
its administrative organs to mark French industrial planning indelibly with an arms
imprint" Our own study of the relationships between military expenditures and com-
petitiveness (see Chesnais and Serfati, 1990, 1992) suggests that a large part of French
high-technology industry (perhaps really all of it outside the medical sector and phar-
maceuticals) has been shaped by the pervasive influence of defense markets and mil-
itary demand, notably the highly customized, nongeneric features of military tech-
nologies and their very low degree of transferability to civilian uses. Serfati (1991) has
shown that the disastrous balance of the French electronics industry, despite the atten-
tion and financial support it has received, cannot be dissociated from the fact that the
military has had priority in fixing the industry's R&D and industrial objectives. More
generally in the case of a country the size of France, the effects of the arms industry
and military R&D on industrial competitiveness cannot be analyzed simply in terms
of alternative opportunity costs or "crowding out" processes. Because of its position
at the heart of the electronics, electrical, and mechanical engineering interlocked com-
plex of industries, arms production also affects interindustry flows of technology and
shapes the whole process of technological learning and accumulation in these indus-
tries. In instances where new technologies emerge in the defense sector, as in laser tech-
nology, the transfer to civilian user has proved a complete failure.15
The power of the DGA, but also the organization of the arms related innovation
subsystem and the political strength of the French military-industrial complex, can-
not be dissociated from the role played by the corps of armament engineers.
Drawn largely from the Ecole Polytechnique . . ., the army engineers form a
unique and cohesive corps throughout their careers that lead almost inevitably
to the highest posts within the Ministry of Defense, the DGA and, increasingly,
to leadership positions throughout the nationalized and private sectors of the
arms industry. . . . Sharing a common schooling and formation, they have an
engineer's and technocrat's way of looking at issues. . . . The network of corps ties
goes deeper and is more extensive than the organization charts of the arms indus-
try or of the DGA can capture. The arms engineers who are liberally distributed
throughout the weapons complex are the glue that holds the system together.
Increasing numbers of the corps have no difficulty rationalizing their service to
the State from the perspective of an industrialist or functionary since both activ-
ities are viewed as different aspects of a single national policy to make and sell
arms. (Kolodziej, 1987, p. 278)
The overall result of these processes has been the transfer of vital decision making
from the political institutions nominally in charge to this group and the organizations
they work for. In economic as well as political terms (cf. the arming of Sadam Hus-
sein), the results of "what is good for the DGA and the large arms producers is good
for France" have at the best been debateable and at the worst disastrous.
The Electronuclear Subsystem: CEA, EOF, and Their Industrial Partners
Systemic relationships involving two major public organizations and large firms from
the semipublic and private sectors, cemented by the Polytechnique-trained engineers
belonging to the Fonts et Chaussees "Corps," also characterize this subsystem (see Fig.
6.6).
The CEA, as noted above, still has the single largest public R&D budget. Half is
for military purposes; the other half (the part shown in Fig. 6.2) is for industrial objec-
tives, where the CEA still receives nearly twice the resources as the CNET and the
CNES. EDF, the large public corporation dating back from the 1945-1946 national-
izations, is the world's second largest producer of nuclear-based electrical power. It
exports electricity through the European electrical grid system. Whenever (e.g., more
and more rarely) tenders are called for the building of nuclear power plants a French
consortium generally composed of Framatome and Alsthom/CGE will compete.
France is a world leader in patents relating to nuclear production and now has the
doubtful honor of being a world leader in nuclear waste disposal and/or recycling.
Much of this takes place at the Hague near Cherbourg, now a major "nuclear dustbin."
The history of the nuclear subsystem falls into two main periods (Gilpin, 1968;
Papon, 1975,1977;Debeiretal., 1989). The first is that of the natural uranium-graph-
ite-gas reactor system set up under the sway of a dual purpose military-civilian strat-
egy with its R&D and industrial programs.16 The second period, coinciding with de
Gaulle's departure from office in 1969, was built on a clear separation between military
and industrial objectives. A full-scale reorganization of the CEA was undertaken, the
EDF was established with responsibility for industrial strategy, and the decision was
Figure 6.6. The configuration of the nuclear subsystem in the mid-1980s. Source: G. Donna-
dieu, La mise en valeur des acquis de I'industrie nudeaire, p. 34. Report to the Conseil Econo-
mique et Social, 1984.
taken to adopt the dominant U. S. enriched-uranium technology. Westinghouse's

technological partnership for PWR reactors was sought and obtained on favorable
conditions. The reorganization of the civil industry-oriented segment of the CEA her-
alded the process of semiprivatization of the government sector laboratories previ-
ously discussed. From 1970 onward a number of departments and laboratories were
transformed into CEA industrial subsidiaries with private commercial law status, in
particular the Compagnie generate des matieres premieres (COGEMA) set up in 1976,
which controls, through a number of affiliated firms, the whole French nuclear fuel
cycle from prospecting for and extracting natural uranium through uranium enrich-
ment (EURODIF) to spent fuel processing via the fabrication, design, and marketing
of fuels (FBFC and FRAGEMA). In 1983 all private law CEA subsidiaries were inte-
grated in a holding company, CEA-Industries.
After an initial phase of sharp oligopolistic rivalry involving two industrial groups
each allied with separate U. S. partners (Westinghouse and GE) the heavy equipment
side of the innovation subsystem was organized around two poles, following one of the
numerous "industrial Yaltas" (a pleasant expression for State-blessed cartel creating),
which most of France's high-technology sectors have experienced at some point. The
Schneider-Empain group and their lead firm in electronuclear technology, Frama-
tome, were given the EOF contracts for nuclear reactors based on the Westinghouse
PWR technology, while CGE/Alsthom, which had previously acquired the number
one position in conventional heavy electric power equipment, got the orders for the
complementary "standard" heavy electric equipment surrounding the nuclear plant
(Debeir et al., 1989).
From the mid-1970s onward, CEA was considered to have grown well beyond
the size that was required. Because of the political power of the "Grands Corps," the
answer was not to slim it down but to empower it to do R&D in fields other than
nuclear energy. "Innovation and industrial development" thus joined "nuclear devel-
opments" as a mission of the CEA's Technology Research and Industrial Develop-
ment Institute. Nonnuclear activities include electronics and computer technology,
robotics, medical instrumentation, agrofood technology, new materials, and even
renewable energies. The technologies developed by the CEA are transferred to industry
in a variety of ways: notably the creation of R&D partnerships with firms in the indus-
tries concerned.
CNES and the Space Industry
There is a wide interface between the military subsystem of innovation and production
and the one for space R&D and related operations and space industrial production.
The CNES was set up at the end of 1961. It has a legal status somewhat similar to that
of the CEA, but has built up its network of technological and industrial relationships
differently. The first difference is that the CNES is not a public sector laboratory since
a government R&D capacity in aeronautics dates back to the 1920s and some space-
related questions existed before CNES, within ONERA (civil and military aviation
R&D) and ISI (Institut de Saint-Louis), which works on missiles. The second signifi-
cant difference is that French space-related technological investments began once the
industrial base had been rebuilt. CNES policy has always been to organize R&D pro-
curement in industry rather than developing space technology in-house. Given the
overall thrust of French industrial policy and the influence of the DGA just discussed,
the firms that have been assigned prime contractor responsibilities for space programs
are the same as the DGA's main partners. Matra is the lead firm in propulsion systems
for the launchers and shares with Aerospatiale work on satellite platforms as well as
satellite components regarding flight altitude, thermal controls, and solar generators.
Thomson-CSF is the principal contractor for internal workings of the satellites, includ-
ing electronic circuits and their components.
Another major difference with the CEA is that in the case of space, the scale of
the resources required for R&D and operations was such that from the outset, the
French projects were not conceived in purely national terms but involved interna-
tional cooperation, bilaterally with the United States and even the USSR, but mainly
with European States within ESRO and ELDO, and after 1975 within ESA. In contrast
with nuclear energy, defense, or telecommunications, in the area of civilian space the
French subsystem of innovation has provided the overall structure and represented the
backbone of Europe's involvement in space. It is sure that without France the European
commitment would not be of the scale it now is, nor Arianespace a viable competitor
of the main U. S. launchers. The success has brought new problems, in particular those
of France's European partners wanting to increase their industrial participation in the
development and manufacture of the launchers, satellites, and space cabins. Asset
holding by CNES and the four main French contractors in the Arianespace industrial
consortium has progressively been brought down from over 60 to about 48%, bringing
cooperation under strain, in particular since the last phases coincided with the fall in
arms exports. The French government has just announced that military satellites will
now represent a major priority of the DGA's investments and that closer coordination
with CNES on military objectives must result. This also will put European cooperation
under pressure unless it is extended to the military area, but this is not necessarily
acceptable to other members of ESA and raises all the issues that have put a brake on
R&D and production cooperation for military projects.
France Telecom, the CNET, and the Telecommunications Oligopoly
Prior to World War II, the equipment required by the rather underdeveloped French
telephone system was supplied by the affiliates of foreign multinationals, in particular
ITT and Eriksson. Since 1945, one of the permanent objectives of French policy has
been to endow France with its own equipment industry and make it a leader in tele-
communications. This began with the establishment of a large government laboratory,
CNET, which later took over other facilities including the military radioelectricity lab-
oratory. The Defense Ministry has always financed a fairly substantial share of the
CNET's R&D budget.
Along with the building up of an R&D and innovation capacity, two other steps
were taken. First a large-scale and dynamic system of public procurement managed
by the DGT (Direction Generate des Telecommunications) was established. Within
the P&T Ministry, the DGT plays a role somewhat similar to that of the DGA with
the Ministry of Defense. As in other sectors the commercial arm of this policy is a
publicly-owned private law subsidiary, French Telecomm, which has been active in
equipping industry and homes with phones and now with advanced electronic equip-
ment, in particular the Minitel videotex machine. The other step was the creation of
an endogenous industrial capacity that included the acquisition of the ITT and Eriks-
son affiliates and their incorporation through mergers with French firms into the core
nucleus of the French high-technology oligopoly. Today the manufacturing sector is
highly concentrated, with CGE-Alcatel as the leading group. Alcatel now has by far the
largest R&D budget in civil telecommunications.. Its French R&D outlays are about
twice as large as those of the following four equipment manufacturers (SAT, TRT,
Matra, and Jeumont-Schneider) together.
The R&D and innovation budget for telecommunication is divided in two.
About a third is allocated to the CNET, which employs a staff of 4200 people, includ-
ing 1500 engineers and researchers, working in six locations. The rest goes to public
or private industry in the form of R&D contracts. Again with respect to the results of
CNET's R&D, the policy is to find industrial partners as early in the development pro-
cess as possible. As a result, CNET holds less than 10% of all French patents in the area
while the manufacturers hold almost 80%.
The telecommunication subsystem has been marked by successes as well as fail-
ures (Salomon, 1986, is highly critical). In the case of the Minitel videotex system suc-
cess was built not only on correct technological decisions, but also on the large scale
use of public procurement by the DGT (over 5 million terminals were bought by the
DGT and offered free of charge to users), the initial subsidization of service suppliers,
and the centralization of supervision and development in the hands of a single body
(OECD, 1988; Cohendet and Llerena, 1991, Chapter 2). Today this system is under
pressure on two scores. First, the internationalization of Alcatel (now Alcatel NV) fol-
lowing the acquisition by CGE/Alcatel of ITT's European industrial assets in 1987 has
substantially shifted the apex of the group's economic interests to European markets.
In 1987, Alcatel NV had 75% of its turnover outside of France as against 40% in 1985.
Alcatel NV has the potential to become a truly global corporation with strongly inter-
nationalized R&D activities, but it will be forced to aim its R&D strategy to the needs
of all its European customers and not only those of France Telecom. Second, DGT
and France Telecom are coming under increasing pressure from the European Com-
mission to liberalize their procurement policy and reduce their direct involvement in
industrial development activities.
PATENTING AND EXPORTS AS INDICATORS OF SUCCESS AND FAILURE
In the case of France, use of the term "innovation system" innovation appears legiti-
mate. It refers to a set of identifiable relationships established among major political
institutions, research organizations, and business enterprises, which has been repro-
duced (along with the adaptations and adjustments made necessary by changes in the
economic and political context) over several decades. The mechanisms of reproduc-
tion include the role of State funding for R&D and industrial investment and the part
played by an original type of political and economic elite.
The system also has identifiable features with respect to the type of innovation it
produces. It is, to use Ergas's terminology (Ergas, 1984, 1987), one of the clearest
examples of a "mission-oriented" type of innovation system. The model is one where
"bigis beautiful." A premium has been constantly given to large technology intensive
systems (as in the military area, in electrical power, and in rail transport) or to products
that are inherently systemic (e.g., aircraft or space products). As a result, markets are
almost always invariably conceived by project leaders as being (or having to be) public
markets. At home these are created through public procurement. Abroad there are
invariably the outcome of political negotiations with governments, leveraged by
French diplomacy, very often directed to Third World countries or erstwhile "Social-
ist" States and backed by a widespread use of bribery and corruption.17
This system of innovation does not cover the complete spectrum of French pat-
enting and exports. One major sectorthe chemicals, Pharmaceuticals, agrochemi-
cals complexfalls partly outside the pattern of relationships we have just discussed.18
Several other activities and products for which France is successful and well known
abroad (wine and brandies, luxury consumer goods, beauty products) fall completely
outside the system. But it is quite definitely the system of innovation that the majority
of Frenchmen and practically the whole educated class know and with which they gen-
erally identify themselves. It is also the one that commands to a considerable extent
the external balance as well as industrial growth.
Today the system is in crisis. Its highest moment of success occurred somewhere
between 1975 and 1980. This was the heyday of nuclear power plant building in
France and of exports of the same large capital goods to a number of Third World
countries. It was the period of spectacular growth in arms exports, leading to a situa-
tion in which military production came to represent the second most important and
stable part of French engineering goods exports after automobiles and arms produc-
tion became "trapped" into working over 40% for export markets, which had to win
at any cost (cf. the contracts with Iraq). The space launcher Ariane-I was successfully
put into orbit with a commercial mission for the first time in 1979, and the Airbus
consortium finally consolidated the same year after a period of crisis with the U. K.
partners. But the worm was in the fruit. One of the longstanding features of French
economic performance has been the congenital weakness of industrial exports. During
a transitory period, this weakness was masked by a number of successes in high-tech-
nology systemic products and the good performance of the car industry. Now that
arms exports have fallen and that automobile sales have begun to collapse, the vul-
nerability of the system can no longer be masked.
French Technological Performance as Measured by Patenting in the United States
Today patents are considered by many scholars to represent, despite numerous cave-
ats, a reasonable proxy for innovative activity. One drawback, which is evidenced by
the French situation, is that although the granting of a patent recognizes technical nov-
elty and may thus be one measure of technological output, the possession of patents
does not guarantee subsequent competitive commercial application, sales, and exports.
In a context where nuclear plants have practically stopped being built in the world,
this is the case, for instance, for nuclear power technology. Despite this, it is interesting
to look at the French data, which show the areas where France produces internation-
ally recognized technical novelty even if this does not necessarily lead to commercially
successful products.
Pavitt and Patel (1990) collected information on the national origin of U. S. pat-
ents granted over a century (1883-1987) and compared the long-term trend of French
patenting with that of the United Kingdom and Germany. Outside the disruption
caused by two World Wars, France has tended to improve its position compared to
the United Kingdom. This trend was particularly marked from 1900 to 1910 and in
the 1920s, and has occurred again since 1970: in 1987, for the first time, France was
granted more U. S. patents than the United Kingdom. The comparison with Germany
is also revealing. Earlier this century German patents were about four times the French
level. Since the late 1950s, France has reduced this gap. German patents in the United
States have stabilized to about 2.8 times the French level.
Pavitt and Patel have also calculated an index of "revealed technology advan-
tage" (RTA), which confirms quite logically France's relative strength in sectors dom-
inated by public procurement and State funding of technical activities. In 1981-1986,
three of the first five sectors with the strongest RTA in France were in this category, as
were six of the eleven with an RTA of 1.10 or more: nuclear reactors, aircraft, telecom-
munications, other transport (in particular, railways and railway equipment), electri-
cal devices (in particular electrical transmission to vehicles), and normal electrical
power plants.
Using the patent data in the U. S. office for 1981-1987 and for 1982-1987, the
data provided on a comparable basis by patenting at the European Patent Office as
well as the results of current research on patent citation, Archibugi and Pianta (1992)
calculated their own index of specialization and ranked the major OECD countries
according to their top five fields of technological excellence. In the case of France,
Table 6.1 lists these results.
Using SIC classes instead of IPC classes and extending the data to include cita-
tions, Archibugi and Pianta present the following picture of French specialization
(Table 6.2).
Table 6.1. French Technological Performance as Measured by Patenting
Patents Granted in the United States (1981-1987)
Nuclear physics Medical preparation Mining Weapons Engineering
3.14 1.64 1.48 1.40 1.36
Patents Granted at the EPO (1982-1987)
Nuclear physics Building Lighting and heating Transport Agriculture
2.63 1.65 1.60 1.59 1.54
Pavitt and Patel have also examined the relative performance in U. S. patenting
of large firms (French and foreign controlled), State agencies, and smaller firms. Large
French and foreign-based firms are particularly important in the R&D-intensive sec-
tors (chemicals, electrical-electronics, aerospace) and in automobiles. The data con-
firm the strong contribution of State agencies, in particular in technologies related to
energy (EOF, CEA). The overall percentage of patenting associated with smaller
domestic firms is high, 37.2%, and is hardly lower than the one for the large domestic
firms. This does not fit with the findings on formal R&D and points to a much greater
amount of innovative activity by small firms than that reflected in the formal R&D
data. It points to a largely unresearched area of the French innovation system.
Pavitt and Patel have also identified the French-owned organizations appearing
in the list of the top 20 U. S. patentors for 33 sectors. Very few rank in the top U. S.
10; most of those that do are either State agencies or firms heavily dependent on State
markets and State R&D funding: Creusot-Loire (now Franmatome) and CEA in the
electronuclear branch; SNIAS (now Aerospatiale) and SNECMA in aerospace; and
Thomson in defense-related telecommunications. Only 1'Oreal (soaps, detergents,
body care products) and Michelin (tires) belong completely to the private sector.
The Commercial Performance of French Innovation: Exports and Imports
In an international economic system where the social validation of production-related

activities occurs mainly at the level of the market through successful commercializa-
tion, the foreign trade performance of an economy may reflect the efficiency of its
innovation system better than patents.
In the case of France, foreign trade indicators reveal a serious situation (Econ-
omic et Statistiques, 1989;OECD, 1990). Despite favorable conditions regarding price
competitiveness French industrial exports have experienced a long downward trend
since the late 1970s, while imports have risen continuously. Recent studies stress a
number of structural weaknesses; in particular, aging productive capacities, an insuf-
ficient rate of investment, a propensity (shaped by decades of selling to protected polit-
ical markets) to impose higher profit margins on exports than on domestic sales, con-
siderable difficulties on the part of executives and managers in abandoning the
traditional Fordist-Taylorist model of corporate management, and significant obsta-
cles to horizontal interindustry and intersectoral transfers of technology. These are due
to the vertical organization of innovation in many sectors and to the barriers that char-
acterize economies where arms industries occupy a central role within the high-tech-
nology complex. The system is ill-equipped to satisfy the requirements of the generic
technologies. Barriers to interindustry flows of technology have been further accen-
Table 6.2. French Specialization, 1975-81 and 1982-88
Top 5 SIC Classes With the Highest Specialization Indexes for Patents and Patent Citations
Indicator 1 2 3 4 5
Pat.75-81 Drugs & Medicines Agric. & Other Chem. Guid Miss.Space Veh. Aircr,Oth.Mech.Parts Ship,Boat Building
Pat.82-88 Guid Miss.Space Veh. Railroad Equipment Soaps,Detergents Aircr,Oth.Mech.Parts Drugs & Medicines
Cit. 75-81 Guid Miss.Space Veh. Motorcycles & Parts Aircr.Oth.Mech.Parts Drugs & Medicines Ship,Boat Building
Cit. 82-88 Guid Miss.Space Veh. Railroad Equipment Soaps,Detergents Ship,Boat Building Aircr,Oth.Mech.Parts
Table 6.3. Commodity Breakdown of the Trade Balance: Customs Basis, GIF-FOB, as a
Percentage of GDP
1982 1983 1984 1985 1986 1987 1988 1989
Agrofood products 0.4 0.6 0.6 0.7 0.5 0.6 0.7 0.8
Energy -4.9 -4.2 -4.3 -3.8 -1.8 -1.5 1.2 -1.4
Industrial goods 0.9 1.6 2.3 1.9 0.7 -0.2 -0.7 -0.9
Consumer goods -0.6 -0.4 -0.3 -0.3 -0.5 -0.8 -0.8 -0.7
Intermediate goods -O.I 0.1 0.2 0.2 -0.2 -0.3 -0.4 -0.7
Producer durable goods 0.4 0.7 0.8 0.6 0.2 -0.1 -0.4 -0.3
Land transport equipment 0.6 0.6 0.8 0.7 0.6 0.4 0.4 0.3
Military hardware and other 0.6 0.6 0.9 0.8 0.7 0.5 0.5 0.5
Total -3.6 -2.0 -1.4 -1.3 -0.6 -1.1 -1.2 -1.5
Sources: Direction generate des douanes et des droits indirects, INSEE, and OECD Secretariat estimates.
tuated by a strong element of secrecy stemming from the important military compo-
nent of technology production: this is at least one reason for the weakness of the
French electronics industry.
As shown in Table 6.3, the weaknesses in French industry are particularly pro-
nounced in the capital and equipment goods industries. This is both a consequence of
several of the factors listed above and a source per se of structural weakness. This pivot
function of the capital goods industries in the diffusion of new or best practice tech-
nology is well established: when these industries begin to collapse, the performance of
the entire manufacturing sector will be affected. Figure 6.7 shows that the deterioration
of the French trade balance is particularly marked in the case of low-technology indus-
tries, where performance was satisfactory until the early 1980s. The deterioration has
Figure 6.7. France: trade by R&D intensity groups.

also concerned the medium-technology industries, as well as the high-technology sec-
tor.
As shown in Table 6.4, the overall positive balance in the high-technology sector
is really due to only two industry groups, drugs and medicines and aerospace (civil and
military). Now that the car industry has begun to collapse in the face of German and,
despite trade barriers, of Japanese competition, the structural deficiencies of French
manufacturing have become evident.
The current failure of industrial exports is the inevitable outcome of what Pavitt
calls the "myopic" traits of the industrial and technological system. It cannot be attrib-
uted to an insufficient level of French R&D expenditures in industry. As Barre (1988)
has shown, this is comparable to that of France's major competitors, once account is
made of the industrial structure by branches. But it does have a lot to do with the struc-
ture of industrial R&D, the priorities chosen, and the institutional context in which
R&D is undertaken and its results taken to the market. The failure of exports is the
combined result of the inherent difficulties of selling large system-like products in an
economic context marked by strong monetary instability and high levels of public debt
(even outside Third World countries) and of an industrial and technological structure
particularly not prone to horizontal linkages between industrial branches, the exploi-
Table 6.4. France Export/Import Ratio

1980 1981 1982 1983 1984 1985 1986 1987
Food, beverages, and tobacco 1.19 1.28 1.18 1.12 1.17 1.15 1.16 1.20
Textiles, leather, and shoes 0.86 0.85 0.77 0.79 0.81 0.81 0.72 0.68
Wood, cork, furniture 0.47 0.52 0.48 0.52 0.54 0.55 0.49 0.47
Paper and printing 0.60 0.60 0.58 0.60 0.62 0.64 0.62 0.62
Chemicals 1.03 1.11 1.00 1.16 1.19 1.19 1.13 1.12
Drugs and medicines 2.17 1.98 3.76 1.95 1.97 1.99 1.90 1.80
Petroleum refineries 0.83 0.85 0.56 0.55 0.58 0.57 0.52 0.37
Rubber, plastic 1.27 1.29 1.23 1.22 1.26 1.24 1.16 1.07
Pottery, china, glass 0.98 1.06 1.06 1.22 1.22 1.19 1.01 0.97
Ferrous metals 1.36 1.61 1.38 1.42 1.51 1.52 1.37 1.30
Nonferrous metals 0.66 0.71 0.65 0.75 0.77 0.73 0.70 0.72
Metal products 1.22 1.29 1.10 1.18 1.17 1.00 0.86 0.77
Nonelectrical machinery 1.14 1.19 1.11 1.18 1.15 1.11 0.98 0.87
Office machinery, computers 0.75 0.74 0.58 0.65 0.69 0.66 0.70 0,71
Electrical machinery 1.26 1.30 1.22 1.26 1.26 1.22 1.08 1.00
Electronic components 0.85 0.80 0.78 0.94 0.93 1.00 0.90 0.87
Other transport 1.81 2.55 2.18 2.31 2.19 2.49 1.95 1.60
Shipbuilding 2.05 3.32 1.99 3.20 3.67 2.05 2.94 2.21
Vehicles 1.74 1.58 1.32 1.35 1.49 1.41 1.33 1.21
Aerospace 1.33 1.19 1.72 1.50 2.20 2.06 1.56 1.62
Scientific instruments 0.80 0.78 0.75 0.79 0.87 0.94 0.81 0.80
Other manufacturing 0.73 0.85 0.85 0.85 0.95 0.96 0.88 0.78
NEC 2.94 3.41 4.00 4.26 4.48 4.96 3.37 2.79
Total manufacturing .08 .12 1.02 1.06 .11 1.08 1.00 0.95
Sum of above .08 .12 1.02 1.06 .11 1.08 1.00 0.95
High-technology industries .03 .01 1.09 1.04 .12 1.11 0.99 0.96
Medium-technology industries .16 .21 1.09 1.18 .21 1.19 1.11 1.05
Low-technology industries .01 .08 0.94 0.96 .01 0.97 0.91 0.86
tation of externalities, and the flexibilities required for a proper user within manufac-
turing of information technology. The critical assessment that several U. S. groups (cf.
the MIT study of U. S. industrial competitiveness or the studies at BRIE) have recently
presented regarding U. S. industrial organization (the emphasis on hierarchy, the dif-
ficulty of cooperating, the antagonistic relationships between management and work-
ers) is pan passu applicable to French industry, with the particular rigidities of a State-
led system added for good measure.
NOTES
1. For details of the way traditional production of the Bordeaux wines was established on
proper scientific and technological foundations after 1945, see Ribereau-Gayon (1972) and Pey-
naud(1988).
2. In the Competitive Advantage of Nations Michael Porter (1990) points to the software
service industry, in particular for customized products not controlled by the large manufactur-
ers, as being an industry in which France has a recognized position. For a French view, see Hor-
aist(1986).
3. Gilpin(1968, p. 106) offers the following overall assessment:
The heroic period of French science, from 1800 to 1830, was the product of two important
factors: Revolutionary and Napoleonic reforms had established institutions and an envi-
ronment where French genius could flourish; and knowledge in many fields had advanced
to the point where it lent itself to mastery by the peculiar strengths of the French mind.
Around 1800, in chemistry, natural history, physiology, and other areas, someone was
needed to bring order to the disarray of conflicting opinions and positions. Such a task
required the patience, brilliance, and individuality of men like Lavoisier who founded the
use of precise measurements in chemistry and systematized the subject. French genius fash-
ioned the paradigms, or revolutionary new theories, that guided scientific research for much
of the nineteenth century. As John Merz has pointed out in his History of European
Thought "in France during the early part of the century the foundations of nearly all the
modern sciences were laid. Many of them were brought under the rule of strict mathemat-
ical treatment."
4. Important protestant names in French industry based in the Mulhouse, Belfort, Mont-
beliard region include Schlumberger, Dollfus, Koechlin, and Japy with whom the Peugeot broth-
ers later had ties by marriage (see Caron, 1987).
5. See inter alia Palmade (1961), Cameron (1961), Kindleberger (1964), and Caron (1979).
6. See Landes (1949, 1951). For counterarguments and qualifications of Landes' positions
see Levy-Leboyer (1974).
7. See Stanley Hoffman's assessment of France as a "stalemate society" marked by (1) a
preference for stability and protection over growth and competition, (2) a Malthusian fear of
overproduction of material goods and of educated people, (3) the burden of social, religious, and
political conflict, (4) the fragmented structure and conservatism of French industry, and (5) the
domination of agrarian and colonial interests over domestic industrial interests (Hoffman,
1963).
8. For this section see Gilpin (1968, Chapter 6), Papon (1978, Chapter 2), OECD (1966),
and Rouban (1974). A somewhat unimaginative history of the CNRS is also available; see Picard
(1990).
9. For a detailed presentation of these reforms, see OECD (1986) and for a critical assess-
ment, Salomon (1986).
10. See the lengthy annual Rapport sur I'etal dc la recherche et du developpemenl techno-
logique, annexed each year to the Projet de Loi de Finances presented to tne Parliament before
the discussion and vote of the French budget. This report is available each year in October.
11. The compartmentalized character of the French innovative system has recently been
recognized by one of the few critical studies on technological policy produced within the French
political machinery, the "Rapport Purge": see Commissariat General du Plan (1989, p. 22).
12. Research by Richard Kuisel (1981) shows that this position cannot be attributed solely
to left wing ideologists and the governments they inspired in 1936-1937 or 1946-1948, but was
prepared in the 1930s within very conservative industrial circles to which men such as Joan
Mounet belonged.
13. McArthur and Scott (1969) add that in the course of their research "it soon became
evident that the state-company relationship and not the planning process per sc was the most
important determinant of corporate strategic planning in France."
14. A first class analysis in English of the "Ecoles," the "Corps," and the organization and
power of the French industrial-political-financial elite can be found in Suleiman (1978).
15. The French set-up based on DGA procurement and the organization of R&D in a
highly specialized subsidized firm (Compagnie Industrielle des Lasers) belonging to CGE was
very successful in a purely military context but a complete failure once civil demand became
important. See Cohendet and Llerena (1991, Chapter 1).
16. Although it arose from the need to avoid depending on imports of foreign-enriched
uranium along with possible U. S. embargoes, the choice of a graphite-gas reactor system was
really the logical outcome of the decision taken in 1956-1957 to construct plutonium-generating
reactors in order to obtain plutonium for military purposes. By 1965-1966, it had become clear
that the achievement of the political aim of possessing at an extremely high price nuclear weap-
ons had led France up a blind alley with respect to the production of electricity, where cost con-
siderations could not be waved aside as in the military field. For a detailed discussion see Papon
(1975,1979), Gilpin (1968), and Salomon (1986).
17. Through this particular way of planning for markets and winning them in arms, con-
ventional and nuclear power production, telecommunication equipment, or urban transport
systems (as well as in the large scale civil engineering operations that go along with these where
some of France's most powerful industrial groups such as Bouygues are now located) corruption
has crept into the pores of the French administration, political life, and society. The process has,
of course, been powerfully aided by the constitutional structure of the Fifth Republic, which lays
almost total and uncontrolled power into the hands of the Executive. France is not a land of
Watergates or even of Colonel North type congressional and judicial investigations. The French
politicians, bankers, industrialists, and generals can sleep in peace: behind a facade of formalistic
legal control of legislation and administrative procedures by the Conseil Constitutionnel and the
Conseil d'Etat, corruption is rampant and goes unpunished. The Parliament has no powers and
the rule is that the "party of the President," be it Gaullist, Giscardian, or Socialist, must smother
any difference it may have with the President and vote obediently to a man.
18. During this project there was no time to undertake the detailed work required to under-
stand the exact configuration of the innovation system in the chemicals-pharmaceuticals-agro-
chemicals complex. The weight of the business enterprise component is obviously much more
important than in those related to electronics and arms, but the State is present in many signif-
icant ways. The most important is the role it plays from time to time in industrial restructuring,
the definition of new corporate boundaries, and the provision of finance. The last time this
occurred was in 1982, with the redistribution of industrial assets between three major groups,
Rhone-Poulenc, Elf-Sanofi, and CDF-Chimie, the last two of which have stale capital. While
R&D capacity is strongly lodged within these groups, they can also count on technology transfer
from the nonprofit and public sectors. A good example concerns the results of the Institut Pas-
teur, which are commercialized through Diagnostics-Pasteur, a joint venture in partnership with
Elf-Sanofi and Pasteur-Merieux-Serums et Vaccins, where the partnership is with Rhone-Pou-
lenc.
REFERENCES Ergas, H. (1987). "Does Technology Policy Matter."

In B. R. Guile and H. Brooks (eds.), Technology
and Global Industry. Washington, D.C.: National
Archibugi, D., and Pianta, M. (1992). The Techno- Academy Press.
logical Specialization of Advanced Countries: A Fox, R., and Weisz, G. (1980). "The Institutional
Report to the EEC. Dordrecht: Kluwer Academic Basis of French Science in the 19th Century." In
Publishers. R. Fox and G. Weisz (eds.), The Organisation of
Barre, R. (1988). "La faiblesse de la recherche indus- Science and Technology in France, 1808-1914.
trielle francaise: de quoi parle-t-on?" Culture Cambridge, England: Cambridge University
Technique 18 (March): 210-19. Press.
Bauer, M. (1987). Les 200: comment devient-on un Gillispie, C. C. (1980). Science and Polity in France
grand patron? Paris: Le Seuil. at the End of the Old Regime. Princeton: Prince-
Gallon, M. (1987). "La recherche francaise est-eile ton University Press.
en bonne sante." La Recherche 186 (Avril-Mai). Gilpin, R. (1968). France in the Age of the Scientific
Cameron, R. E. (1961). France and the Economic State. Princeton: Princeton University Press.
Development of Europe, 1800-1914. Princeton: Granick, D. (1964). The European Executive. New
Princeton University Press. York: Doubleday.
Caron, P. (1979). An Economic History oj Modern Hoffman, S. (1963). "Paradoxes of the French Polit-
France. New York: Columbia University Press. ical Community." In S. Horfman (ed.), In Search
Caron, P. (1987). "La capacite d'innovation tech- of France. Cambridge MA: Harvard University
nique de 1'industrie francaise." Le Debat 46 (Sep- Press.
tembre-Novembre). Horaist, J. (1986). "Les societes francaises de service
Carter, E. C, Foster, R., and Moody, J. N. (eds.) et d'ingenierie en informatique: historique et
(1976). Enterprise and Entrepreneurs in Nine- perspectives." Economic et Previsions, 72: 33-
teenth and Twentieth Century France, Baltimore, 48.
MD: Johns Hopkins University Press. Kindleberger C. P. (1964). Economic Growth in
Chesnais, F., and Serfati, C. (1990). "L'industrie France and Britain. Cambridge, MA: Harvard
d'armement: une locomotive du developpement University Press.
economique francais?" In F. Chesnais (ed.), Com- Kindleberger C. P. (1976). "Technical Education
petitivite Internationale et depenses militaires. and the French Entrepreneur." In E. C. Carter, R.
Paris: Economica. Foster, and J. N. Moody (eds.), Enterprise and
Chesnais, F., and Serfati, C. (1992). L'armement en Entrepreneurs in Nineteenth and Twentieth Cen-
France: gene.se, ampleur et couts d'une Industrie. tury France. Baltimore: Johns Hopkins Univer-
Paris: Nathan. sity Press.
Cohen, S. (1969). Modern Capitalist Planning: The Kolodziej, E. J. (1987). Making and Marketing of
French Model. Cambridge, MA: Harvard Univer- Arms: The French Experience and Its Implica-
sity Press. tions for the International System. Princeton:
Cohendet, P., and Llerena, P. (eds.) (1991). Modes Princeton University Press.
of Use and Diffusion of New Technologies and Kuisel, R. F. (1981), Capitalism and the State in
New Know/edge: The Case of France. BETA, Uni- France: Modernisation and Dirigism in the 20th
versity of Strasbourg, Report for the EC MONI- Century. Cambridge, England: Cambridge Uni-
TOR-FAST programme, FOP No. 228, Bruxelles. versity Press.
Commissariat General du Plan. (1989). La Science, Landes, D. S. (1949). "French Entrepreneurship and
la Technologic, /'Innovation: une politique glob- Industrial Growth in the 19th Century." Journal
ale, rapport de I'intergroupe Recherche et Devel- of Economic History 9( 1): 1 -22.
oppement Technologique. Paris: La Documenta- Landes, D. S. (1959). "French Business and Busi-
tion Francaise, Juin. nessmen in Social and Cultural Analysis." In E.
Conseil Economique et Social. (1989). L'economie M. Earle(ed.), Modern France. Princeton: Prince-
francaise souffre-t-elle d'une insufftsance de ton University Press.
recherche? Avis et rapport, Journal officiel de la McArthur, J. H., and Scott, B. R. (1969). Industrial
Republique francaise, vendredi 2 juin, No. 13. Planning in France. Boston: Harvard Graduate
Crossland, M. (1975). "The Development of a Pro- School of Business Administration.
fessional Career in Science in France." Minerva Monjardet, D. (1972). "Carriere des dirigeants et
13(1): 38-57. controle de 1'entreprise." Sociologie du Travail,
Debeir, S. C., Deleage, J. P., and Hemery, D. (1989). No. 2.
Les servitudes de la puissance: une hisloire de I'e- OECD (1966). Reviews of National Science Policy:
nergie en France. Paris: Flammarion. France. Paris: OECD.
Dhombres, N. (1988). Les savants en revolution, OECD (1986). Innovation Policy: France. Paris:
1789-1799. Paris: Editions de la Cite des Sciences OECD.
et de 1'Industrie. OECD (1988). New Telecommunications Services:
Ergas, H. (1984). "Why Do Some Countries Inno- Videotex Development Strategies. Paris: OECD.
vate More than Others?" Brussels: Centre for OECD (1990). Economic Surveys: France 1989-
European Policy Studies, CEPS Paper No. 5. 1990. Paris: OECD.
Palmade, G. P. (1961). Capitalisme et capitalistes Salomon J. J. (1986). Le Gaulois, le Cowboy et le
franfais au XIXe siecle. Paris: Armand Colin. Samourai: la politique francaise de la technologie.
Papon, P. (1975). "The State and Technological Paris: Economica.
Competition in France or Colbertism in the 20th Salomon J. J. (1991). "La capacite d'innovation." In
Century." Research Policy 4: 214-45. M. Levy-Leboyer (ed.), Entre I'Etat et le marche:
Papon, P. (1978). Lepouvoir et la science en France. I 'economie de la France au XXe siecle. Paris: Gal-
Paris: Le Centurion. limard.
Pavitt, K., and Patel, P. (1990). "L'accumulation Salomon-Bayet, C. (1986). Pasteur el la revolution
technologique en France: ce que les statistiques de pasteurienne. Paris: Payot.
brevets tendent a montrer." Revue d'Economie Serfati, C. (1991). "Primaute des technologies mili-
Industrielle. 51 (Spring): 10-32. (Also available in taires, faiblesse des retombees civiles et declin de
English as a SPRU working paper.) competitivite: Le cas de 1'industrie electronique
Peynaud, E. (1988). Le vin et les jours. Paris: Dunod. francaise." Paper for the Conference on the Social
Picard, S. F. (1990). La Republique des Savants. Mastery of Technology, Maison Rhone-Alpes des
Paris: Flammarion. Sciences de 1'Homme, Lyon, September.
Porter, M. (1990). The Competitive Advantage of Shinn, T. (1981). L'Ecole Polytechnique en France,
Countries. London: Macmillan. 1794-1914. Paris: Fondation Nationale des Sci-
Prestre, D. (1984). Physique etphysiciens en France ences Politiques.
1918-1940. Paris: Editions des Archives Contem- Shinn, T. (1984). "Reactionary Technologists: The
poraines. Struggle over the Ecole Polytechnique 1880-
Republique Francaise. (1991). Rapport sur I'etal de 1914." Minerva 22(3-4): 329-45.
la recherche et du developpement technologique en Suleiman, E. N. (1978). Elites in French Society: The
1989 et 1990, annexed to the Projet de Loi de Politics of Survival. Princeton: Princeton Univer-
Finances 1991. sity Press.
Ribereau-Gayon, J. (1972). Problemes de la Zwerling, C. (1980). "The Emergence of the Ecole
recherche scientifique et technologique en oenolo- Nationale Superieure as a Centre of Scientific
gie. Paris: Dunod. Education in France." in R. Fox and G. Weisz
Rouban, L. (1988). L 'Etat et la Science: La politique (eds.), The Organisation of Science and Technol-
publique de la science et de la technologie. Paris: ogy in France, 1808-1914. Cambridge, England:
Edition CNRS. Cambridge University Press.
7
The National System of Innovation: Italy
FRANCO MALERBA
Italy represents one of the success stories of postwar economic growth. Over the past
40 years, GNP growth has been higher in Italy than in most other industrialized coun-
tries. Similarly, productivity and income per capita have risen rapidly and manufac-
turing exports have increased considerably. In a relatively short period of time Italy
has been transformed from an agricultural and semi-industrialized country to an
advanced industrial economy. In addition, during the 1980s Italy experienced high
growth rates in R&D, although Italian international specialization remains mainly in
traditional products such as textile and shoes, as well as in mechanics and industrial
equipment.
What kind of national system of innovation lies at the base of the economic suc-
cess story of the past 40 years? Why is it that the R&D growth of the 1980s did not
translate into successful performance in high technology products?
Without considering the distinction between the developed north and center and
the underdeveloped south, a full understanding of the Italian system of innovation
during the 1980s and the 1990s has to start from the recognition that not one, but two
innovation systems are present in Italy: a small firms network and a core R&D system.
These two systems are quite different in terms of capabilities, organization, and per-
formance. The small firms network is composed of a large population of small and
medium size firms (in some cases located in industrial districts), which interact inten-
sively at the local level. The core R&D system is composed of large firms with indus-
trial laboratories, small high-technology firms, universities, large public research insti-
tutes, and the national government, linked through a complex organizational system
at the national level.
This chapter claims that the small firms network, grown up historically on a local,
regional, and vocational basis and characterized by capabilities accumulated through
productive experience, has worked effectively and performed successfully during the
past decades up until now. Firms in the network are engaged in rapid adoption of tech-
nology generated externally and in the adaptation and continuous improvement of
this technology. The success of the system is based on the atomistic interaction of a
large number of firms bound to each other by economic, local, cultural, and social
factors. Firms incrementally innovate through learning by doing, by using, and by
interacting with suppliers and users. They communicate in a formal as well as an infor-
mal way, share tacit knowledge, and are characterized by high labor mobility. The role
230
THE NATIONAL SYSTEM OF INNOVATION: ITALY 231
of regional agencies, local public authorities, and local professional schools is effective
in supporting the needs and requirements of small firms in the area.
The core R&D system, much more recent than the small firms network and
developed at a much later stage than those of countries such as Germany, the United
Kingdom, France, and the United States, is not characterized by advanced technolog-
ical capabilities and does not perform satisfactorily in terms of innovation and inter-
national competitiveness. In Italy, in spite of relevant quantitative growth of R&D
during the 1980s, some of the qualitative elements needed for an effective and suc-
cessful working of such a complex system are still missing or are not fully developed.
First, several industrial sectors do not have advanced research and technological capa-
bilities. Second, public policy of R&D support still exhibits major flaws. Third, an
advanced national infrastructure of services for R&D and an overall coordination of
public policies are still lacking. Fourth, advanced basic research performed in univer-
sities and public research centers is very unevenly distributed across institutions. Fifth,
shortages of skilled scientists and engineers are present. Finally, there is no tradition
of successful industry-university cooperation in research.
It must be noted that within and across these two systems, dynamic vicious and
virtuous cycles have reinforced the characteristics of each specific system. In the small
firms network, horizontal and vertical linkages have generated virtuous cycles of learn-
ing and incremental innovation, which have been at the base of the successful perfor-
mance of Italian small and medium firms over the past decades. In the core R&D sys-
tem, the lack of advanced capabilities in several components has generated vicious
cycles that have blocked the full development and the successful performance of high
technology industries in Italy.
This chapter will discuss and analyze the following areas: the history and the dual-
ism of Italian industry, the main features of the small firms network, and the core R&D
system. Finally, the virtuous and vicious cycles that have taken place between and
within the two systems are analyzed.
HISTORY AND BASIC FEATURES
A Recent Industrialization
Italy was a late industrializer (Graziani, 1979). Although industries such as steel, auto,
electrical machinery, and chemical were in existence before World War II, Italy did
not develop a modern industry until the 1950s. This meant that advanced technolog-
ical and productive capabilities, managerial skills, and an infrastructure typical of
other industrialized countries began to emerge only in the last 40 years (Malerba and
Falzoni, 1991).
A Lack of Tradition in R&D

In Italy there is no tradition of industrial R&D that dates back to the end of last century
and the first half of this century. Some centers of excellence existed in a few firms (such
as Montecatini), but in general until the second half of the twentieth century Italian
firms spent very little on R&D (Sapelli, 1989). As Table 7.1 shows, during the 1950s
Table 7.1. Share of R&D Expenditures on GDP"

1955 1963-1964 1970s 1985-1987 1989
Italy 0.2 0.6 0.7-0.8 1.2 1.3

Germany 0.6" 0.3 2.3-2.4 2.7 2.9
France 0.8' 1.9 1.8 2.3 2.3
United Kingdom 1.6* 2.3 2.2 2.2 n.a.
United States 3.0" 3.4 2.3-2.5 2.8 2.8
Japan n.a. 1.4 2.0 2.8 n.a.
"The data for the 1955 column are approximate values; *l 954;' 1960; d\ 956-1957; ''1959-60.
and 1960s, Italy was a low R&D intensive country and a technological follower. Dur-
ing the 1950s, 1960s, and 1970s, a large part of the technical change introduced by
Italian firms was through licenses obtained from foreign firms.
A High R&D Growth during the 1980s
Between 1980 and 1987, R&D expenditures grew at an annual rate of 9.9% in real
terms, a value higher than most OECD countries. This was due both to an increase in
R&D done by business enterprises (9.4%) and by the public sector (10.6%). The public
sector became a major source of funds for R&D: between 1980 and 1987 R&D fund-
ing from the public sector grew at an annual rate of 12.6% in real terms, while funding
from business enterprises grew at an annual rate of 6.5% in real terms.
A Dualistic Country in Terms of Firm Size and North-South Differences
The rapid post-World War II economic growth occurred within a dualistic economy
in terms of firm size and geography. Small firms are quite numerous in traditional and
specialized supplier sectors, which constitute a major part of Italian industry (see Table
7.2). In 1981 employment in plants of less than 100 employees represented 59% of
total employment in industry, a share much higher than the one of the other main
industrial countries with the exception of Japan. On the other hand, the core of Italian
Table 7.2. Number of Employees in Manufacturing Industry According to Pavitt's Taxonomy

Absolute Values Percent Values
Sectors 1981 1971 1961 1951 1981 1971 1961 1951
Science based 329,134 237,383 180,200 97,528 5.41 4.48 4.01 2.79
Specialized 703,363 520,232 429,437 284,370 11.55 9.81 9.55 8.13
suppliers
Scale intensive 1,895,526 1,764,353 1,378,429 1,048,569 31.14 33.28 30.66 29.97
Suppliers 2,714,848 2,452,909 2,292,345 1,957,486 44.60 46.26 50.99 55.96
dominated
Others 444,162 326,969 215,152 1 10,267 7.30 6.17 4.79 3.15
Total manufac- 6,087,033 5,301,846 4,495,563 3,498,220 100 100 100 100
turing
industry
Source: Malerba and Falzoni (1991).
industry is made up of large firms, active mainly in scale-intensive and high-technol-
ogy sectors.
Most of the large Italian firms and the bulk of technologically advanced industry
is located in the northern regions. The south of Italy is characterized by a limited
degree of industrialization, by a low R&D intensity, and a limited diffusion of
advanced technologies (Romano, 1989), as Table 7.3 shows.
The Role of Public Enterprises in R&D Growth
Another feature of the Italian industry is related to the presence of public enterprises
in scale-intensive and high-technology industries such as electronics, steel, food, aero-
space, and military. During the 1960s and 1970s public enterprises played a major role
in maintaining an indigenous capability in sectors such as electronics and aerospace.
Furthermore during the 1980s they made a major contribution to the impressive
growth rate of R&D (they had a 10.8% annual real growth rate, compared to 8.6% of
Table 7.3. Italy: North-South Differences

1987 Center-North South Italy
Population 70 30 100
Value-added 82 18 100
Exports 90 10 100
Public sector expenditure in R&D 91 9 100
Private sector expenditure in R&D 97 3 100
Source: 1ST AT
R&D/ Value Added of Industrial Firms 1983

Northwest 1.4
Northeast-center 0.4
South 0.2
Italy 0.7
Source: Santarelli and Sterlacchini (1989).
Center-North South
Telephone density 100 59
Bank branches density 100 53
Railway electrification rates 100 43
Source: OECD (1990).
Share of Total Share of Total

Flexible Automation Production Systems CAD Systems
Northwest 63.9 52.5
Northeast 28.4 32.9
Center 5.6 9.7
South 2.1 4.9
Italy 100.0 100.0
Source: MIPPolitecnico of Milan (1989).
private enterprises). In several circumstances their role as engine of technological

change in Italy has been impaired by political obligations to save industries in crisis
and to be active in the south.
The Beginning of the Internationalization Process of the 1980s
Until recently Italian industry has not shown a high degree of internationalization.
Italian firms in traditional sectors such as textiles, shoes, and furniture have been
major exporters, but foreign direct investments by Italian industry have been scarce
(9% of total employment in 1989) and the presence of foreign multinational corpo-
rations in Italy has been limited (13% of total employment in Italy in 1989). Only
recently the beginning of a move toward greater internationalization of activities and
cooperative agreements has taken place in Italian industry (Onida and Viesti, 1987;
Mariotti et al., 1986; Cominotti and Mariotti, 1990).
THE SMALL FIRMS NETWORK
A large part of Italian industry is composed of a large group of small and medium size
firms operating in traditional industries (such as textile and clothing, shoes, furniture),
in mechanics, and in equipment supplier industries. These firms are specialized in the
supply of custom made products and of fashion items. Most of them assemble and
integrate existing components and parts into systems for special applications or spe-
cific customers. These firms are highly profitable and quite successful internationally.
Over the past decades, this group of firms has given extreme flexibility and high adapt-
ability to Italian industry during the business cycle.
These small and medium size firms form a highly dynamic atomistic learning
network. They are characterized by advanced capabilities of absorbing, adapting,
improving, and tailoring new technologies (developed externally) to specific market
needs.
Innovation originates not from formal R&D, but from informal learning by
doing, by using, and by interacting. Engineering skills, product know-how, and under-
standing customers' requirements are the major sources of incremental innovations
and product customization by this group of firms. These characteristics emerge clearly
from Table 7.4, constructed from a major survey carried out by ISTAT in 1988 and
concerning innovation in more than 8000 firms. Table 7.4 identifies the cost of inno-
vation, the sources of innovation, and the relevance of collaboration in R&D for small
firms compared to large and medium size firms.
Three types of firms can be identified in this network: firms in the industrial dis-
trict, equipment producers, and traditional firms.
The Industrial District
In industrial districts, characterized by both cultural and social homogeneity and

developed historically on a vocational basis, technical change occurs through horizon-
tal linkages among a large number of small and medium size firms (Becattini, 1987,
1989). These districts are active in several industries and are located in various Italian
Table 7.4. Main Differences in the Innovation Process between Small, Medium, and Large
Firms"
Cost of Innovation (share of total cost)
Design Productive
R&D Engineering Investments Marketing Percent
Small firms 8.5 16.3 70.7 4.5 100
Medium firms 12.9 17.0 65.0 5.2 100
Large firms 21.4 29.5 43.5 5.6 100
Total firms 18.0 25.2 43.5 5.4 100
Sources of Innovation (values 1 to 6)
Inside Purchase of Raw Intermediate
R&D Design Proposals Technology Materials Goods
Small firms 1.7 2 .8 2.0 0.3 1.2 0.8
Medium firms 2.8 3 .5 2.9 0.7 1.2 0.8
Large firms 3.9 4 .2 3.5 1.5 1.2 0.9
Total firms 2.1 3 .1 2.3 0.5 1.2 0.8
New Human Professional Customers
Machinery Skills Training Needs Suppliers
Small firms 4.0 1.0 2.0 2.3 1.5
Medium firms 4.1 1.4 2.5 2.4 1.7
Large firms 3.9 1.7 3.1 2.7 2.1
Total firms 4.0 1.2 2.2 2.3 1.6
Joint Industrial Public
Ventures Exhibitions Institutions Consultants Imitation
Small firms 0.2 1.5 0.2 0.6 1.8
Medium firms 0.3 1.5 0.4 0.8 2.2
Large firms 0.7 1.4 0.9 1.0 2.7
Total firms 0.3 1.5 0.3 0.7 2.0
Collaboration in R&D (share of the total number of innovative firms)
Public
Total Institutions University Customers Suppliers
Small firms 8.8 16.8 23.6 39.4 61.2
Medium firms 20.9 24.4 37.6 37.3 55.4
Large firms 45.2 50.3 72.6 35.7 52.2
Total firms 13.9 28.2 37.9 39.4 57.4
"Small firms, 20-99 employees; medium firms, 100-499 employees; large firms, 500 employees and more.
Source'1ST AT (1988).
regions: textile in Prato, Como (silk), Biella, Carpi (knit); shoes in Vigevano, Barletta,
and Casarano; furniture in Brianza and Udine; ceramic tile in Sassuolo; gold jewelry
in Valenza Po and Arezzo; and household products in Lumezzane.
Some of these districts have been in existence for decades: such is the case of the
textile districts in northern Italy. Other industrial districts grew up during the 1960s
and 1970s, such as Sassuolo (Russo, 1985), Prato (Rullani and Zanfei, 1988a), and
Valdarno Inferiore (Gandolfi, 1988).
In these districts the division of labor among small and medium size firms is high
and the productive flexibility and adaptability to changing market demand at the final
product level are substantial (Becattini, 1987; Nuti, 1988). Most firms specialize in
only one stage of the production process: only a few firms eventually internalize more
than one stage of the production process and eventually sell the final product. As these
Italian districts are involved in a variety of technologies and industries, the organiza-
tion of production varies from case to case. For example, in the shoe districts, design
is developed externally, the sole and the heel are purchased from large firms, the var-
ious production stages are done by small specialized firms, and the sale and distribu-
tion by another firm (which eventually also internalizes some strategic stages of the
production process such as cutting). In the textile districts, in addition to small firms
specialized in a specific production stage, the fashion designer and the converter are
also present. Similarly, in the ceramic tile district, vertically integrated medium size
firms are present side by side with specialized small firms (CESPRI, 1990).
Local institutions and local associations play a major role in the working of the
organization of the district. Regional and local governments, banks, and professional
schools provide public support, financial resources, and a qualified labor force to firms.
Export and distribution associations help overcome the problems faced by small firms
in selling their final products on international markets. In some districts associations
among firms have been created for the sharing of complex and costly production
equipment.
Recently in most districts leading firms or local industrial groups have emerged,
In the first case some firms, strategically located at the commercialization stage, coor-
dinate the whole production process of the district. This is the case of the weaver at
Prato, the trader in Carpi, and the converter in Como. In the second case, through
acquisitions and participations, some industrial groups have been able to control (at
the strategic and financial level) the production of the district without, however, inter-
fering with the daily production of the small firms of the group.
Diffusion of process technology within the district is quite rapid. Technical
change is rapidly diffused within the district through the widespread transmission of
information among a large number of producers that share a common culture, have
the same level of capability, and, because they are similar, are also able to transmit and
assimilate tacit and noncodified knowledge. Personal contacts and the mobility of
technicians among the firms play a major role in this respect (Bellandi, 1989). Inter-
estingly enough, within the firms of the district, the diffusion of new capital equipment
has been more rapid than the diffusion of electronic information systems. Firms were
able to quickly adopt new machinery because they already had the technical capability
to insert and adapt the new machinery in their productive organization. On the other
hand, most new information systems required the drastic modification of the firms'
organizations and the creation of new in-house capabilities: therefore they met resist-
ance and delays in their diffusion within the district.
In these districts, both product and process innovations are of the incremental
type. Product innovations are the result of skills in product design and ability in focus-
ing on specific market requirements and consumer needs. Process innovations stem
from learning by doing in single specific production stages.
As it will be shown later, in most cases firms in the industrial district constitute
the major market for upstream equipment producers (usually located near or even
within the industrial district) that introduce new innovative equipment as a result of
close and continuous interaction with the downstream district firms.
Equipment Firms
The Italian industry is characterized by the presence of a large number of small and
medium size equipment producers, which are highly innovative and internationally
competitive. This type of firm, present in various regions in the north of Italy, also
includes machine-tool and robotics producers. Table 7.5 shows that the share of pat-
Table 7.5. Italy: Size Distribution of the Firms Patenting in the United States
(Percentages)
Employees
0-100 100-1,000 1,000-10,000 Over 10,000
Chemicals
1969-1974 8.2 26.6 65.2
1975-1979 0.3 11.2 23.6 64.9
1980-1984 2.9 11.5 16.0 69.6
Pharmaceuticals
1969-1974 31.8 31.8 26.4
1975-1979 29.0 21.8 49.2
1980-1984 0.5 16.4 18.0 65.1
Electronics
1969-1974 0.0 1.3 0.7 89.0
1975-1979 0.0 2.5 13.3 84.2
1980-1984 0.4 4.5 13.2 82.0
Electrical machinery
1969-1974 0.0 9.7 6.7 83.6
1975-1979 0.0 18.3 17.5 64.2
1980-1984 1.4 15.9 10.4 72.4
Mechanical machinery
1969-1974 9.5 18.7 13.3 58.5
1975-1979 10.1 27.2 12.7 50.0
1980-1984 16.1 26.4 10.6 46.9
Road vehicles
1969-1974 5.9 8.8 85.3
1975-1979 2.9 5.9 91.2
1980-1984 1.4 2.8 2.8 93.0
Other transports
1969-1974 3.4 38.0 10.3 48.3
1975-1979 6.7 26.6 6.7 60.0
1980-1984 13.8 17.2 7.0 62.0
Specialized industrial equipment
1969-1974 4.6 22.7 22.7 50.0
1975-1979 2.9 8.7 53.6 34.8
1980-1984 7.3 15.6 43.4 33.7
Metals
1969-1974 4.9 8.8 3.9 82.4
1975-1979 12.5 23.6 8.4 55.5
1980-1984 21.3 28.7 6.5 43.3
Source: Malerba and Orsenigo (1991).
Table 7.6. Share of Total Cost for Technological Innovation (Percentages)

Design
R&D Engineering Investments Marketing Total
Sectors 1 2 3 4 5
All sectors 17.9 25.2 51.5 5.4 100.0
Pharmaceuticals 41.2 15.5 31.8 11.5 100.0
Electronics 24.3 33.4 38.4 3.9 100.0
Auto 15.3 23.5 59.4 1.8 100.0
Mechanical products 7.3 17.1 71.9 3.7 100.0
Nonelectrical equipment 16.0 33.1 46.3 4.6 100.0
Machine tools 21.0 36.9 37.4 4.7 100.0
Textile machinery 8.0 44.3 44.2 3.5 100.0
Textiles 6.4 7.3 83.3 3.0 100.0
Clothing 5.8 17.6 68.9 7.7 100.0
Source: \STAT (19SS).
ents held by small firms (with less than 100 employees) is substantial and is increasing
over time.
The dynamics of entry and exit in this group of firms are high: new firms have
been set up by technicians and engineers who have left other established equipment
firms, or by some large users such as FIAT.
Because they are characterized by a lively entrepreneurship, a longstanding tra-
dition of advanced technical and design skills in mechanical equipment, an effective
understanding of users' needs, as well as a relevant capability of absorbing new elec-
tronics technologies in their products, equipment producers generate a continuous
stream of incremental innovations in equipment. This is achieved by tailoring prod-
ucts to users' needs, focusing equipment to specific market segments, and improving
and modifying existing equipment. Often innovations consist in system integration
aimed at specific applications or at the solution of specific technical problems of the
Table 7.7. Major Sources of Innovation"

Inside Purchase of Raw Intermediate New
R&D Design Proposals Technology Materials Goods Machinery
Sectors 1 22 33 4
4 5 6 7
All sectors 2.1 3.1 23 0.5 1 .2 0.8 4.0
Pharmaceuticals 4.3 3.5 33 2.1 1.4 0.7 3.3
Electronics 3.1 4.5 29 0.6 1 .3 1.0 3.7
Auto 4.8 4.2 30 0.7 1.4 1.2 4.3
Mechanical 1.6 2.9 21 0.4 1.0 0.7 4.1
products
Nonelectrical 2.7 4.2 25 0.5 1 .1 1.0 3.4
equipment
Machine tools 2.4 3.7 21 0.5
1
.1 1.0 3.6
Textile 3.4 4.4 22 0.5 ().9 1.1 3.7
machinery
Textiles 1.1 1.8 18 0.3 1 .0 0.5 4.5
Clothing 1.3 2.1 19 0.3 1 .2 0.6 4.1
"Values from 1 to 6.
Source-. 1ST AT (1988)
users (see the nonelectrical equipment, machine tools, and textile machinery indus-
tries in Table 7.6).
Several of these firms do not have a formal R&D laboratory: their knowledge is
mostly noncodified, tacit, and idiosyncratic, because it is embodied in technicians and
engineers. Some firms (mainly flexible automation and robotics producers) maintain
close links with engineering laboratories at the university (Camagni, 1984). Most firms
use consultants for the solution of specific problems (see the nonelectrical equipment,
machine tools, and textile machinery in Table 7.7).
Learning by interacting through the user-producer relationship plays a key role
in the innovation process (see Table 7.8 for nonelectrical equipment, machine tools,
and textile machinery). Vertical links with sophisticated users are extremely important
in providing an innovative stimulus and a continuous feedback on the use of the
machinery. Two main types of users may be identified: large firms (such as FIAT, Oli-
vetti, and Zanussi) and firms in industrial districts (such as the Biella district for textile
machinery and the Sassuolo district for equipment for ceramic tile production). As will
be shown later, these links have generated vertical virtuous cycles in Italian industry.
Firms in Traditional Sectors
A final group of firms is composed of the large population of small and medium size
firms operating in traditional industries, but not located within an industrial district.
Innovative firms are not very numerous in this very heterogeneous group of firms.
New product introduction is driven by marketing and production capabilities linked
to fashion design, product tailoring, and market segmentation. As Table 7.7 shows for
the textile and clothing industries, these firms greatly profit from the innovativeness
of upstream equipment producers. Embodied technical change in terms of invest-
ments in new machinery represents the major source of change in production pro-
cesses.
Human Professional Customers Joint Industrial Public

Skills Training Needs Suppliers Ventures Exhibition Institutions Consultants Imitation
8 9 10 11 12 13 14 15 16
1.2 2.2 2.3 1.6 0.3 1.5 0.3 0.6 2.0

1.6 2.4 1.7 1.2 0.7 1.2 1.5 1.1 2.5
1.8 2.7 2.8 1.8 0.4 1.7 0.5 0.7 2.4
1.6 3.7 1.1 2.7 1.4 2.2 1.3 1.0 1.9
1.1 2.0 2.2 1.5 0.2 1.3 0.2 0.5 1.7
1.5 2.4 2.8 1.6 0.3 2.0 0.3 0.7 2.1
1.8 2.7 2.8 1.7 0.2 2.0 0.3 0.6 2.0

1.8 2.6 3.2 1.9 0.3 2.2 0.3 0.7 2.3
0.8 1.8 2.1 1.5 0.1 1.2 0.1 0.5 1.6

0.9 2.2 1.7 1.4 0.2 1.5 0.0 0.6 1.6
Table 7.8. Relevance of External Cooperation of Innovating Firms (Number of Firms)
Suppliers
Number of Number of Percent of Public Other Industry Research
Innovative External External Institutions University Customers National Foreign Firms Associations Consortia
Firms Cooperation Cooperation (%) (%) (%) (%) (%) (%) (%) (%)
Sectors 1 2 2/1 3/2 4/2 5/2 6/2 7/2 8/2 9/2
All sectors 8220 1140 13.9 29.1 37.9 37.9 41.4 15.9 18.1 13.5 8.5
Pharmaceuticals 154 90 58.4 27.8 108.9 14.4 6.6 6.6 30.0 18.9 21.1
Electronics 658 159 24.2 37.1 37.1 45.3 38.9 16.9 8.2 13.2
Auto 9 6 66.7 33.3 83.3 33.3 83.3 66.6 16.7
Mechanical 1051 66 6.3 19.7 33.3 27.3 48.4 12.1 18.2 19.7 3.0
products
Nonelectrical 1404 255 18.2 19.6 24.3 57.3 48.6 13.3 15.7 10.2 5.1
equipment
Machine tools 287 50 17.4 16.0 22.0 48.0 64.0 12.0 18.0 6.0
Textile 97 32 33.0 3.1 15.6 68.8 59.3 12.5 21.9 6.3
machinery
Textiles 714 44 6.2 22.7 9.1 27.3 59.0 11.3 9.1 22.7 11.4
Clothing 412 19 4.6 5.3 42.1 63.1 10.5 5.3 15.8 15.8
Source: ISTAT( 1988).
It is interesting to note that some large Italian firms active in traditional sectors
have maintained some of the attributes of the industrial district by using a very decen-
tralized productive organization, in which a large number of small local firms are spe-
cialized in a specific stage of the production process or in the supply of a specific input.
Benetton (with more than 1330 employees and active in the wool and cotton gar-
ment industry) has been able to match a decentralized production organization (typ-
ical of Italian textile firms) with an advanced electronic sales network. It created a hier-
archical system composed of independent medium size firms assembling and
controlling the production of a large number of specialized independent subcontrac-
tors. Almost 80% of Benetton's production is handled by 350 external firms and arti-
san shops (accounting for up to 10,000 employees) specialized in labor-intensive and
nontechnologically progressive operations. Benetton's distribution, on the other hand,
is characterized by an advanced telematic network linking Benetton with a decentral-
ized sales structure composed by 75 independent agent firms and 4200 selling points.
This network allows Benetton to keep in close contact with customers, and to main-
tain control over information and market demand in different countries (Rullani and
Zanfei, 1988b; Belussiand Festa, 1990).
Some large firms in traditional industries have also been able to successfully
change the organization of production through the introduction of new electronics
technologies. For example, Miroglio (a textile producer with 3400 employees) has
been able to improve its productive efficiency through a policy of investment in new
electronic equipment. Miroglio has also become competitive internationally by spe-
cializing in a production geared toward well-defined market segments (and by devel-
oping a widespread commercialization structure). Similarly, Gruppo Finanziario Tes-
sile (GFT) (7500 employees operating in the garment industry) has chosen to
specialize in the high quality segment of the market by utilizing advanced electronic-
based technologies.
The Success of Policies in Favor of the Small Firm Network
Two types of policies have affected the atomistic learning network: policies aimed at
fostering the adoption of capital equipment (launched by the central government) and
policies aimed at developing a local technological infrastructure (launched both by the
central government and by regional and local authorities).
Policies in Favor of the Adoption of New Capital Equipment

Fiscal instruments have been used in policies aimed at increasing the adoption of new
capital equipment. These policies have been successful in reaching a large number of
potential adopters.
The earliest provision was Law No. 1329 of 1965 (the Sabatini Law), which has
proven to be a flexible instrument well suited to the needs of small and medium size
firms. The Sabatini Law calls for deferred payment of the entire expenditure up to 5
years in the case of machine tool purchases, obtained through the financial institute
Mediocredito Centrale. The bureaucratic procedures for obtaining these benefits are
simple. In 1987, Mediocredito Centrale distributed about 444 billion lire in subsidies
for more than 30,000 operations involving capital outlays of over 4000 billion lire.
To supplement the function of the Sabatini Law, another measure was passed in
1976 (Presidential Decree 902) whose purpose was to allow for the construction of new
facilities or the expansion and modernization of existing ones. This provision grants
low-interest loans with rates that vary according to two parameters: whether the firm
is located in northern or southern Italy and what the level of economic development
is in the specific area in which the firm operates. The implementation of the provision
has been extremely difficult because of the highly complicated nature of the program
as well as the difficulty in identifying those areas considered insufficiently developed.
The result has been that the law has financed modernization projects of small and
medium size firms in northern-central Italy, without achieving what the legislators
intended: the articulated differentiation of the intervention over the whole country.
It should be noted that a special program for southern Italy was launched by Law
64 of 1986. This law provided low-interest loans and subsidies for research projects,
purchases of real services, technological investments, and the establishment of new
research centers.
A third provision giving incentives to diffusion has been Law 696 (1983), which
concerns the purchase or leasing of advanced equipment such as automation equip-
ment. It grants an additional VAT deduction of 6% off the taxable base. It should be
stressed that this provision was passed as a temporary measure scheduled to lapse in
1984, but due to its success, the term was extended until 1985. Over 16,000firmsben-
efited. Decree law No. 318 (1987) reintroduced the same regime.
Policies for Information and Technology Diffusion

Policies for information and technology diffusion have been characterized by the
blending of national policies with regional and local initiatives, which have arisen in
a decentralized and bottom up manner.
Broad programs have been aimed at developing centers and structures for
research, experimentation, personnel training, production upgrading, and technical
consultancy. The most significant initiatives are the Tecnotex Program at Biella (in
the textile industry), the Trieste Research Area (which also includes the international
genetic engineering center promoted by UNIDO), and Tecnopolis in Bari (the major
science park existing in Italy).
Centers of support for diffusion such as CESTEC in Lombardy, ASTER in Emi-
lia, and DITEL in Liguria aimed at providing information and training, and organiz-
ing pilot projects and demonstrative activities to stimulate industrial awareness in the
use of new technology.
Satisfactory results have been obtained by support policies focused on the crea-
tion of sectoral centers for technology transfer and general technical assistance aimed
at local industry. These centers conduct experimental research, engage in design and
engineering, and provide consultant services. In general, they are promoted and man-
aged jointly by regions, regional financing companies, chambers of commerce, and, in
some cases, research institutes, together with private firms and industrial associations.
There are currently some 40 of these in Italy, most of them located in Emilia, Marches,
and Lombardy (Lassini, 1990).
With similar objectives, the "Sprint" project in the Prato industrial district aimed
at developing and installing a data transmission network. This project involved not
only manufacturers but service companies, banks, government offices, and associa-
tions, so that the cost of transferring technical, financial, and commercial information
for the firms in the district could be reduced. The project also included technical assis-
tance for the introduction of process innovations in textile production.
Cooperation among firms involved in technology diffusion has also taken place.
At the local level these initiatives have been quite successful. In Lombardy, for exam-
ple, there are provisions designed to stimulate exchanges of technological expertise for
the creation of new products, applications of new technologies, and utilization of
advanced technical services such as quality and reliability testing. From 1985 to 1987,
public support was granted for around 50 of these initiatives involving a total of 400
firms, whose aim was to promote the search for partners and the cooperation among
firms. In general, in Italy, around 20 intercompany technical service centers have been
recently established, with the direct involvement of 800 firms as partners or promoters.
In addition to these initiatives, there are around 50 formal agreements involving
research institutes, public agencies, industrial associations, and major public enter-
prises, whose purpose is to promote the transfer of knowledge and technology. One
example is Law No. 34 (1985) of the Lombardy Region. In addition to supporting
innovation in small enterprises, legislation provides incentives for research contracts
and consultance between universities and small firms. Another example is provided
by the Autonomous Province of Trento, which has promulgated a law that supports
cooperation between industry and universities (Lassini, 1990).
THE CORE R&D SYSTEM
The other side of the Italian national system of innovation is the core R&D system.
This system is highly affected by technological opportunity and demand conditions.
It includes a number of different segments: large oligopolistic firms, small high tech-
nology firms, universities, public research centers, central and local government.
This section argues that the Italian core R&D system grew considerably in quan-
titative terms during the 1980s, but at the end of the decade still had several qualitative
weaknesses related to low opportunity and demand conditions. There were few large
firms, few small high technology enterprises, an underdeveloped industry-university
interaction, and a limited degree of internationalization. All these issues will be
addressed in the following sections.
The High Quantitative Growth of the R&D System during the 1980s
During the 1980s the relative distance between Italy and the other major OECD coun-
tries (with the exception of Japan), as far as total R&D expenditures and business
enterprise R&D expenditures are concerned, was reduced (see Table 7.1). The share
of Italian gross domestic expenditures on R&D in OECD countries increased from
2.85% in 1981 to 3.14% in 1987, with a yearly rate of growth of 8.9% (1979-1982),
11.7% (1983-1985), and 6.6% (1986-1989) (Archibugi-Pianta, 1990).
This quantitative growth has been accompanied by major sectoral changes in the
composition of business enterprise R&D, as a result of the difference in growth of pro-
duction and of R&D intensity of various sectors. High-technology sectors such as elec-
tronics and aerospace significantly increased their relative importance (reaching 24
and 15%, respectively, in 1987-1988) while other sectors such as transportation (auto,
ships, etc.) and chemicals reduced their share (declining to 16 and 8%, respectively, in
1987-1988).
Public support for business enterprises R&D has also significantly increased.
During the 1980s the yearly compounded growth rates of government-financed R&D
has been higher than that of most OECD countries. The share of government financed
business enterprise R&D over total business enterprise R&D increased from 4.5%
(1971) to 22.4% (1987). Government support, however, was not spread evenly across
all industries. As in the other major OECD countries, this support was concentrated
in aerospace, computers, and electronics.
The Limited R&D Intensity of the Italian R&D System and the Unsatisfactory
International Performance in R&D Intensive Sectors
In spite of their quantitative growth during the 1980s, in 1988 Italian R&D expendi-
tures in absolute terms and as a percentage of GDP are still much lower than those of
other major OECD countries, such as the United States, Japan, Germany, France, and
the United Kingdom (see Table 7.1).
Major differences between the north and the south persist. The south continues
to lag far behind the north and the center (in particular the northwest Milan-Turin-
Genoa triangle) in terms of R&D and technological innovation. For example, the
south produces approximately 18% of the total Italian value added, and the northwest
34%. Still the south, which has a 0.2% R&D intensity, performs only 7% of all national
R&D and has obtained only 0.7% of total Italian patents registered in the United
States. The northwest of Italy on the contrary has 1.4, 70, and 83% respectively (San-
tarelli and Sterlacchini, 1989). In addition, in the south approximately 90% of R&D
employees are in the public sector (mainly in universities) (Romano, 1989) and most
of these employees are involved in basic research, acting independently of the local
productive system.
The growth of Italian R&D expenditures did not translate into a satisfactory per-
formance in terms of technological innovation or international competitiveness in
high-technology industries. As Table 7.9 shows, the share of Italian patents registered
in the United States declined from 3.4% in the early 1960s to 2.9% in the late 1980s,
while the Italian share of overall OECD exports of high-technology industries declined
from 4.5% in 1970 to 3.5% in 1987, mainly as a result of the high growth of Japan.
Similarly, the Italian share of OECD exports in most high-technology sectors declined
during the 1970s and 1980s (see Table 7.9), while there has been a deficit in the Italian
balance of trade in most of these products (except for telecommunication equipment
and helicopters-aircraft).
Factors Limiting the Full Development of the R&D System in Italy

What factors limit the full development of the R&D system in Italy?
Six major factors can be identified: limited endogenous generation of advanced
technological opportunities, weak demand conditions, a small oligopolistic core, few
small high-technology firms, an underdeveloped industry-university-research orga-
nization interface, and a still limited degree of internationalization. These six factors
will now be examined in detail.
Table 7.9. Indicators of Italian Technological Performances

Technological Balance of Payments (TBP)
1980-1981 1982-1983 1984-1985 1986-1987
Expenditures/R&D 0.30 0.27 0.22 0.16
Expenditures/revenues
Total 2.80 3.80 3.50 2.90
Manufacture 3.30 4.50 3.40 3.16
Sora?.Falzoni(1990).
Patents and International Competitiveness

Share of Total OECD
Italian Share of Total Foreign Patents in the Exports of High R&D
United States Intensive Sectors"
1963-1968 1976-1981 1982-1988 1970 1980 1987
Italy 3.4 3.1 2.9 4.5 3.9 3.5
"Aerospace, office machines, computers, electronics and components, drugs, instruments, electrical
machinery.
Italian Share of OECD Exports

1970-1972 1983-1985
Aerospace 2.3 4.1
Office machines 8.3 1.6
Computers 5.9 4.4
Electronic components 3.4 2.2
Telecommunications 5.0 3.8
Drugs 6.4 5.4
Instruments 2.8 2.8
Electrical machinery 5.0 4.3
SoKrvMalerba(l988).
Limited Endogenous Generation of Advanced Technological Opportunities

The endogenous generation of advanced scientific and technological opportunities in
Italy is characterized by a fragmentation of efforts and a high variance of scientific
output.
The level of research at Italian universities varies greatly across scientific fields.
Areas of excellence exist in physics (particularly particle physics), space, lasers, syn-
thetic chemistry, and optoelectronics. In several other areas, on the other hand,
research has not reached advanced international standards and the presence of islands
of scientific excellence is more the result of the efforts of single scientists working in
isolation at the frontier than the work of teams of scientists (Dosi, 1989). This situation
is aggravated by the scarce funds available for the purchase and use of advanced
research equipment, the lack of advanced multidisciplinary research, and the still lim-
ited access of academicians to industry.
Recently, improvements in the organization of Italian university have occurred.
In 1980, departments (around which major research areas are grouped), research doc-
torates, and the possibility for part-time work for university professors were intro-
duced. In 1989, with the passage of the Italian university under the newly formed Min-
istry for University and Research, university research was more closely tied to teaching
and training.
The introduction of the first level diploma in 1990 also aimed to change the trend
of a decrease in the supply of industrial researchers, engineers, and technicians. In fact,
university graduates in the science fields decreased from 11,912 in 1977 to 11,043 in
1988, while graduates in engineering decreased from ll,313in 1977 to 10,524 in 1988
(Centra Studi IBM, 1989).
Unlike the university system, the Italian public research network does mainly
mission-oriented research rather than basic research.
The Italian National Research Council (CNR) has a wide range of research activ-
ities distributed amongst quite a large number of institutes and centers (289). There is
no evidence of an evaluation mechanism for internal research and there is no selection
mechanism for the opening of new institutes in new scientific areas or for the closing
of old ones in old scientific fields. Recently, in a wide range of research fields, the CNR
has successfully launched "Finalized Projects," targeted to specific national objectives
and aimed to improve cooperation in research. In addition, the CNR has focused some
of its internal research activities around "Strategic Projects" concerning information
technology, environment, biological systems, infrastructure, and services.
The rest of the public research network is composed of ENEA, ASI, INFN, ISS,
and the experimental stations among others (see Table 7.10 for an overview). ENEA
(the National Board for Nuclear and Alternative Energy Sources, under the control of
the Ministry of Industry) has shown considerable dynamism by expanding its range
from nuclear energy to renewable energy sources and energy conservation, to envi-
ronment and health protection, to process technologies, biotechnologies for agricul-
tural applications, advanced materials, lasers, optoelectronics, and robotics. In addi-
tion, ENEA stimulates the diffusion of new technologies among small firms. ASI
Table 7.10. The Organization of Scientific and Technological Research in Italy"
CTPE
The Council of Ministers- CIPI
Central Statistical OfficeISTAT (S 2832, R 415)
Ministers
ISCO
Budget and planning
ISPE
Agriculture and forestry Agriculture Research Centres
CNR (S 5317, R 2522)
ASI
University and research~ErEE~ Universities
INFN(S 1420, R 394)
Defense Research Centres of Navy, Army, Air Force
ENEA (S 5000)
industry and Trade=^
Experimental stations
Labor CENSIS
State enterprises R&D of State enterprises
ilc-dlin
TJ~,.J.L,
ISS(S 1399, R246)
Transportation and civil aviation Centre for Automotive Research and Testing
"S, staff; R, researchers.
(Italian Space Agency) is active in space research, receives considerable public support,
and collaborates with industry, universities, and CNR. The INFN (National Institute
of Nuclear Physics) does research in nuclear and particle physics, has extensive inter-
national research collaboration, and has a high research reputation. The ISS (Higher
Institute of Health, under the control of the Ministry of Health) conducts health
research and drug testing. The Experimental Stations of the Ministry of Industry do
material and product controls, testing, and certifications. These stations are small and
in some cases have inadequate equipment and facilities. They are characterized by
limited turnover: some of these stations have been in operation since the end of the
last century, while the opening of new ones for ceramics and plastics has been rejected.
It must be noted that limited basic research is performed by the few large firms
that have central corporate laboratories such as Ansaldo, FIAT, STET, ENEL, ENI,
and Montedison.
Weak Demand Conditions
In Italy, demand conditions have hindered innovation in the R&D system in well
denned ways.
Large firms have rarely provided innovative stimulus to domestic producers sim-
ply because in several cases they themselves have not shown a high degree of innova-
tiveness. Until recently large firms have preferred to buy state-of-the-art components
or equipment abroad when they were not available domestically, rather than stimu-
lating a potentially innovative domestic supply.
Public procurement has rarely been used as a conscious stimulus for innovation
(Pontarollo, 1986). A successful case of development of advanced capabilities con-
cerns civil engineering firms that grew through public contracts for public works. But
in general public administration normally purchases existing products from existing
producers: in several cases it has, de facto, maintained unchanged historical quotas
directed to domestic firms. Public procurement as a means of change has been
impaired also by norms regarding public contracts, difficulties and delays in financial
payments by the public sector, cumbersome bureaucratic procedures, and political or
social goals.
Similarly, Italian military demand (with few relevant exceptions) has been gen-
erally less technologically progressive, smaller, and more open to imports than mili-
tary demand in other European countries. In 1986, Italian expenditures on military
equipment was $2500 million, versus $7100 million in the United Kingdom and
$4400 million in the Federal Republic of Germany. It must be noted that 21.5% of the
Italian military demand was satisfied by imports (Nones, 1988). Some cases of suc-
cessful development exist, however. They are in most cases linked to participation in
international programs such as Alenia in the Tornado fighter plane and Agusta in heli-
copters.
A Small Oligopolistic Core
One of the peculiarities of the Italian system of innovation compared to the other main
advanced countries is the reduced number and the limited size of the large oligopolistic
core. Italian concentration ratios in terms of R&D as well as of patents (Malerba, 1988;
Malerba and Orsenigo, 1991) are higher than those of the main advanced industrial-
ized countries (see Table 7.11). The core of industrial R&D in Italy is concentrated in
a few large industrial groups: FIAT, Ferruzzi-Montedison, ENI, STET (IRI), Olivetti,
Table 7.11. R&D Concentration in Italy"

Italy United States Japan United Kingdom France
1985 1985 1984 1978 1983
C5 46 16 19 41 30
C7 52 18 24 n.a. n.a.
"R&D Concentration ratio: share of the major 5 and 7 R&D spenders; share of
total business enterprise R&D.
Source. Malerba(1988).
and Pirelli (see Table 7.12). These six industrial groups operate in the auto, chemicals,
Pharmaceuticals, electronics, and tire industries.
Table 7.7 shows that the sources of innovation and the organization of the inno-
vative process in these large companies are quite similar to the ones present in other
large companies around the world. Innovation is driven mainly by internal R&D,
engineering, design, and suggestions coming from other internal functional areas.
External R&D contracts are mostly given to firms within the same industrial groups.
R&D cooperation with public research institutions (such as CNR and ENEA) and
with the university is quite common (see Table 7.8). It must be emphasized that in
terms of R&D intensity at the product level, Italian firms are not at major disadvan-
tages with respect to their international competitors. The low R&D intensity at the
industry and at the country level is the result of the fact that within an industry, Italian
firms are in general active in those products that have a low R&D intensity.
The new technological dynamism and R&D growth of Italian large firms during
the 1980s represent the continuation of a process of technological accumulation begun
by some of these firms during the 1950s and early 1960s, and interrupted during the
1970s. The period of rapid industrial reconstruction and fast growth of the 1950s and
1960s based on low labor costs and foreign licenses, and centered on mechanics, tra-
ditional sectors, and scale-intensive sectors (such as basic chemicals, auto, and steel)
was characterized by indigenous technological developments in computers by Olivetti,
drugs by Lepetit, nuclear energy by CNEN and INFN, lasers and electronics by CISE
Table 7.12. Italian Firms With the Largest R&D Expenditures and Patents
(B) (C)
(A) R&D Expenditures/ Share of Total Patents
R&D Expenditures Sales (Percentage) Granted to Italian Firms
(Million Dollars) from the Late 1 960s to
1987 1982 1987 the Early 1980s
FIAT 1050 2.4 3.5 5.4
STET (IRI) 422" 2.9 3.2" 4.6
Italtel 141.2 7.8 12.4
Montedison 290 2.6 2.7 11.9
Olivetti 330.8 4.9 5.8 5.1
ENI 149.4 0.6 1.3 5.8
Pirelli* 133.1 6.2 6.8 3.6
"SOS-Thomson is excluded.
* All firms of the group.
Sources: (A) and (B) Malerba (1988) and AIRI (1988); (C) Archibugi (1987).
and Politecnico of Milan, and chemicals by Donegani (Montedison). Original research

and development, however, was not linked to large-scale manufacturing and to sup-
port activities such as design, engineering, and marketing, because of the lack of real
commitment to a significant activity in high technologies and financial difficulties
(because of increasing labor costs). Several Italian firms discontinued their research
efforts and followed strategies centered on the pursuit of static efficiency (such as
reduction of production costs through decentralization and rationalization of what
was existing) instead of focusing on dynamic efficiency and technological innovation.
Olivetti decided to remain a producer of mechanical typewriters and sold its electron-
ics operation to General Electric in 1963; CNEN discontinued its activities in the sec-
ond half of the 1960s; Lepetit was purchased by Dow Chemical; Montedison discon-
tinued research on advanced new materials of the Donegani Institute (Malerba, 1988;
Soria, 1979; Antonelli, 1984). During the period of financial crisis and labor disputes
of the 1970s, although in some firms groups of researchers still continued work in
advanced technological areas, the focus of the larger Italian firms moved away from
innovation and R&D at the technological frontier. The extensive productive ratio-
nalization and the return to profitability in the early 1980s allowed Italian firms to
invest more in R&D.
The increase in the level and intensity of R&D in the 1980s was concentrated in
specific technological areas, and was linked to productive specialization (Fornengo
and Silva, 1989). In the information-processing industry Olivetti followed a successful
strategy in several key hardware and software areas; in microelectronics SGS (STET)
closed a considerable gap in semiconductor technologies, and merged with Thomson
(France); in telecommunications Italtel and Telettra developed advanced products; in
robotics, firms such as DEA, Jobs, and Comau introduced a large number of innova-
tive products; in the aerospace industry firms such as Aeritalia, Agusta and Fiat Avia-
zione developed specific capabilities in well-defined technologies; in the chemical and
pharmaceutical industries Montedison consolidated its position in selected technolog-
ical areas.
During the 1980s these firms developed vertical and horizontal cooperative agree-
ments with other firms and institutions. Olivetti developed a network of alliances,
acquisitions, and participations with hardware and software companies (Malerba,
1988); Montedison participated in or controlled several other firms; Italtel and SGS-
Thomson developed a range of cooperative agreements with foreign firms. FIAT dras-
tically changed its policy of subcontracting, reducing the number of suppliers (from
1200 in 1980 to 850 in the late-1980s) and pushing decentralization of component
design and innovation through development contracts. Through these contracts,
FIAT supports part of the cost of the development of a new component by its suppliers.
In case of successful development, FIAT becomes the owner of the technical docu-
mentation of the component and may establish a long-term supply contract (3 to 5
years) with the same supplier. These types of development contracts increased from
28 in 1980 to 48 in 1984, while the share of FIAT purchases covered by long-term
supply contracts following development contracts has increased from 3.7% in 1981 to
7% in 1984 (Enrietti and Fornengo, 1989). Despite this increase, however, the share is
still quite limited as a result of the lack of advanced development capabilities by com-
ponent producers, which were used for too much time to supply FIAT with products
that would passively meet FIAT requirements.
Notwithstanding the high growth rates of R&D expenditures, the oligopolistic
core still shows limited absolute values of R&D expenditures on an international scale
and experiences difficulties in performance at the technological frontier. For a long
time, large Italian firms have been accustomed to compete on a cost basis, to be active
in protected domestic markets, and to maintain a technological follower strategy. In
the near future, international technological competition and European integration
will act as a selection mechanism that will compel big Italian industrial firms either to
be innovators on a continuous basis or to assume the role of technological followers.
Few High-Technology Small Firms
Another weakness of the Italian R&D system is the limited number of new high-tech-
nology firms operating in electronics, software, biotechnology, and services. Some of
these firms are closely linked to the oligopolistic core of large Italian companies, such
as the small electronics firms in the Canavese region surrounding Olivetti, or the soft-
ware and service companies in the Milan area. Others are linked to the few scientific
parks that exist in Italy.
The organization of the innovative process in these firms is centered around
design and research activities, not always formalized into an R&D laboratory. Inter-
estingly enough, those firms that have developed internal technological capabilities are
more open to external cooperation in research.
Within this restricted group, several firms innovate by integrating components,
hardware, and software into systems. In most cases system integration is directed to
specific final applications or to specific customers.
In Italy, demand and dynamic interdependencies have been the critical factors
behind the establishment of new firms. In a sector such as software, new firms have
been successful by offering a specialized, customized, or segmented product that sat-
isfies a specific demand and by utilizing existing technology and adapting it to new
applications or to potential users. These firms have grown by entering into product-
related market segments, but have not reached a large size. On the other hand, cases
in which new firms offer new products based on a technological innovation are rare
(Raffa and Zollo, 1988). This last type of entry, however, is increasing with time, as a
consequence of the growth in the number of electronics engineers and the spin-off of
engineers and technicians from large corporations.
Dynamic interdependencies meant that advanced capabilities in an established
industry became a major factor in the development of a new industry, as in the case
of the relationship between the advanced capabilities in machinery and machine tools
and the development of the robotic industry.
Weak Interfaces between University, Public Research Organizations, and Industry
Another weakness of the Italian R&D system concerns the interfaces among univer-
sity, research organizations, and industry. The efficiency and effectiveness of these
interfaces have been impaired by the limited number of centers of excellence in Italian
universities, the limited mobility in and out of the university system, and the bureau-
cratic and institutional structure of universities. A survey questionnaire involving 14
universities, 25 research organizations, and 44 firms in Italy, and 49 universities, 50
research organizations, and 41 firms in the other European countries in 1988 showed
that compared to European firms, Italian firms do recognize that the difficulties pre-
sented by bureaucratic constraints represent a major obstacle to cooperation with uni-
versities and other research organizations (Fornari et al., 1989).
The Still Limited (albeit Growing) Internationalization of the Italian System

As previously mentioned, the limited degree of internationalization of the Italian sys-
tem has acted as a protected environment, which generated a not particularly pro-
gressive demand to Italian firms. The globalization of international technological
competition and the increasing cost and complexity of R&D in several sectors have
compelled Italian firms to increase their limited degree of internationalization (Com-
inotti and Mariotti, 1990) and to follow an articulated policy of cooperation in R&D
(Vacca', 1986). Allowing for a margin of error in the sources used for the calculations
(newspapers and economic magazines), it is possible to claim that the number of inter-
national cooperative agreements of Italian firms has increased during the 1980s from
96 in 1984-1985,to 181 in 1986, and to 202 in 1987, with R&D being one of the main
motivations for international cooperative agreements (Malerba, 1988).
These newly established networks of international cooperative agreements in
R&D broadened the knowledge base of Italian firms and provided them with comple-
mentary technological competences. Because they are still limited in number and rep-
resent a quite new phenomenon for Italian firms, however, these networks of coop-
erative agreements have not been able to exert major influences on the international
performance of Italian firms.
The Role of Public Policy
Public policy in support of technological innovation does not have a long tradition in
Italy. During the 1950s and 1960s there was no policy at all. Whereas other advanced
countries began to support the electronics industry, in fact, Italian public policy
focused its support on sectors with standard technologies and economies of scale, such
as steel and basic chemicals, and aimed to increase the productive capacity of the coun-
try in these sectors. In most cases this policy resulted in inefficient or duplicative large
plants. On the other hand, the Italian government did not intervene in high-technol-
ogy sectors. In 1963, for example, only 0.8% of business enterprise R&D was funded
by the state. The Government did not support Olivetti's R&D and production of com-
puters. Nor did it intervene in the purchase of Olivetti's computer operations by Gen-
eral Electric in 1964. Similarly, it did not adopt any policy of support for the new semi-
conductor industry during those years (Malerba, 1987). Even the first attempts of
policies of support of high-technology industries (Law 1089 in 1968 and Law 675 in
1977) were characterized by limited resources, and by the lack of a precise policy
model.
The timing of public policy of support of high-technology sectors has been deter-
mined to a large extent by the emergence of industrial and scientific lobbies in
advanced sectors, as has been the case in the decisions to support the electronics indus-
try during the late 1970s and early 1980s and the biotechnology industry during the
second half of the 1980s (Adams and Orsenigo, 1988).
In Italy, the contemporary public policy supporting innovation is implemented
at three different levels. At the basic research end of the innovation process there are
the Finalized Programs of the National Research Council (CNR), originally intended
to guide basic research toward economic applications and to stimulate the transfer of
basic results from the universities and research organizations to industry. Further
downstream in the innovation process there are the National Research Plans of the
Ministry of Scientific and Technological Research (MRST), designed to stimulate
cooperation in R&D in high-risk projects. Finally, at the applied and development
stage of the innovative process, industrial R&D is supported through the Applied
Research Fund and the Technological Innovation Fund. The former aims to support
applied industrial R&D, while the latter focuses on development and prototype pro-
duction.
The Finalized Programs of the CNR
CNR's first Finalized Programs were launched in 1975: by 1986 they totalled 38. They
included a wide range of research fields: food processing, health care, land and envi-
ronment, advanced technologies, energy, and so on. Between 1976 and 1986, the
financing of Finalized Programs reached 1063 billion lire: 12.3% went to the CNR
institutes, 36.2% to university laboratories, and 44% to companies and other research
organizations. In 1987, 10 additional Finalized Programs were approved with a pro-
jected duration of 5 years and an estimated cost of 690 billion lire for CNR, and of 300
billion lire for private partners. These Finalized Programs also include an educational
component, represented by more than 1200 fellowships granted in the fields consid-
ered (Ginebri, 1987).
CNR's Finalized Programs have followed a bottom up approach. Initially, the
leadership of the academic community pushed the programs toward academic
research, with rather confused guidelines concerning implementation. The focus then
gradually moved toward more applied research, a higher level of funding given to
firms, and a greater emphasis on the coordination of the various operating units
involved. It should be stressed that delays in defining objectives, selecting projects, and
choosing partners created serious problems of obsolescence in the Finalized Programs.
The most significant problem that has emerged so far is the development of a
cooperative approach. It is evident that cooperation between companies, universities,
and other research organizations is closely related to the ability to identify unifying
objectives and the capacity of industrialists or project directors to assume leadership
roles. In the numerous cases where these conditions were lacking, the result was that
the individual research units continued to do what they did beforehand, but with
greater financial resources than before.
The National Research Programs (NRPs) of the Ministry for Scientific and
Technological Research (MRST)
The NRPs were introduced in 1982. They are defined in a top-down way by the Min-
ister of University and Research: cooperation is stimulated by defining objectives that
require interdisciplinary and complementary expertise.
By the end of 1987, 9 NRPs had been approved for a total of 58 contracts and
714 billion lire in funding, of which 15% has gone to universities and government
research agencies. In the case of electronics, steelmaking, and building construction
(where basic research is now of limited importance) the share of these institutions has
not exceeded 5%. In a field like biotechnology the share has been higher, between 5
and 10%. The share has risen even further (to almost 50%) in pharmaceutical-related
projects.
The Applied Research Fund
The Applied Research Fund introduced in 1968 (Law 109) and modified in 1982 (Law
46) grants low-interest loans for research projects (up to a maximum of 40% of total
expenditures in general, to 60% for projects concerning sectors included in the applied
programs of the National Board for Industrial Planning-CIPI, and up to 70% for high-
priority sectors) and subsidies (up to 20% of R&D expenditures). For relevant R&D
projects the coverage of costs by the Applied Research Fund may reach up to 90%. The
Fund is managed by the public Industrial Credit Institute IMI under the guidance of
the Ministry for University and Research. The Applied Research Fund granted 4179
billion lire between 1968 and 1989. The support of the Applied Research Fund cov-
ered 5.1% of business enterprise R&D in the period 1970-1987 (7.6% between 1981
and 1987). The support has been highly concentrated in the oligopolistic core: FIAT,
Olivetti, and IRI (Italtel, SGS-Thomson, Aeritalia, Selenia). Small and medium size
firms have obtained a very limited share of total funds, much less than the 20% min-
imum s.hare allocated by the Fund. Similarly, while a 40% share was set aside for com-
panies located in southern Italy, a much lower share was actually granted to them.
The Technological Innovation Fund
This Fund, launched in 1982 (Law .46) and managed by the Ministry of Industry, aims
to promote technological innovation and development in products and processes. It
is based on low-interest loans and subsidies, and permits the funding of programs
already begun. The Fund, initially designed for five industries (automotive, electron-
ics, chemicals, steel, aeronautics), has also included support for the mechanics and
agrofood industries. It has faster bureaucratic procedures and less discretion in the
selection of the program than the Applied Research Fund. Until 1988, the Techno-
logical Innovation Fund granted 2314 billion lire. Also the Technological Innovation
Fund has concentrated its support on the Italian oligopolistic core. The support to the
activities of small and medium size firms, however, is much higher than the Applied
Research Fund.
The Applied Research Fund and the Technological Innovation Fund have
undoubtedly contributed to increase the technological and research capabilities of Ital-
ian industry over the past 20 years, but they still exhibit several major flaws. First, no
overall coherent framework and coordination exists between the two Funds. This is in
part a consequence of the political genesis of Law 46, which divided the power of man-
aging government support for innovation among two different Ministries (Ministry of
University and Research for the Applied Research Fund and Ministry of Industry for
the Innovation Fund), which then aimed at total independence of action. Second, the
two Funds have supported already existing activities and projects already under way,
rather than stimulating totally new projects. Third, the majority of the projects are of
a medium level of innovativeness rather than at the technological frontier. Fourth, no
explicit policy in favor of the natality of small Schumpeterian firms in high-technology
industries has been included. Fifth, the Funds use direct financial support of firms'
activities (easiest to manage for the public sector), and do not rely on a broader range
of policy tools (Momigliano, 1986). Sixth, the spectrum of sectors (not necessarily at
the technological frontier) chosen for support is very broad, while an ex-ante selection
of specific technological areas is absent. Seventh, the decision-making process and the
bureaucratic procedures of the two Funds (particularly for the Applied Research
Fund) are still too long.
Policies Favoring Cooperation
During the 1980s Italian public policies and major public organizations (such as CNR,
ENEA, INFN and ISS, see Table 7.10) have increasingly supported cooperation in
R&D. Law 240 (1981) managed by the Mediocredito Centrale (a public financial insti-
tution) provides credit in terms of low-interest loans for a period of 10 years and
amounts no greater than 1 billion lire, in favor of consortia composed of small and
medium size enterprises, public research agencies, and local government. The purpose
of cooperation may range from scientific and technological research and technical
experimentation, to the updating of managerial techniques, and to the assistance and
technical consultation to cooperative member firms. The operational results of Law
240 have been meager because the procedural and bureaucratic difficulties and the
organizational forms contemplated by the law (consortia of at least nine companies
and multiyear duration) have proven too restrictive and complex for small and
medium size enterprises. Among the major public research centers (and in addition to
the previously mentioned Finalized Project launched by the CNR) since 1982 ENEA
has been increasing its cooperation with industry and universities. In energy, ENEA
has sought to improve the interaction with industry in terms of participation in joint
research projects and of the diffusion of results to small and medium size firms. In
addition, ENEA has launched "industrial promotion projects" (in microelectronics,
biotechnology for agricultural applications, process technologies, advanced materials,
lasers, optical technologies) concerning component testing and the creation of joint
ventures with participation of universities, CNR, local governments, and firms. Also
INFN has recently begun a specific program designed to commercially exploit scien-
tific discoveries in collaboration with companies that operate in advanced technolo-
gies. Similarly, the ISS has increased interaction with industry not only in the usual
phases of testing and new drugs authorization, but also in the Finalized Projects of
CNR and the National Research Plans of MRST. Finally, liaison offices have been
promoted with the purpose of stimulating transfer of scientific results to the industrial
system. During the period 1985-1987 20 programs of this type were started, especially
on the initiative of ENEA and CNR.
European Public Policy
The participation of Italian firms to European programs has helped overcome some
of the weaknesses of the Italian R&D system, by opening up international networks
of information exchanges and collaborations and by allowing Italian firms or research
centers to cooperate with centers of excellence around Europe. It must be noted that
the Italian participation in ESPRIT (12% of total ESPRIT funds) and in other pro-
grams (such as EUREKA) reflects the strengths and weaknesses of Italian industry.
Large firms such as Olivetti, STET, and FIAT are present in areas such as computer-
integrated manufacturing, office systems, and lasers, while small high-technology
firms have a limited presence.
The specific experiences of the Italian institutions involved in European coop-
eration in space technologies and nuclear physics also indicate that in the long run
international R&D collaboration has strongly benefited Italian firms. Italian partici-
pation in the European Space Agency and in CERN during the 1970s and the 1980s
(Kluzer, 1989) shows that given advanced technological capabilities by Italian firms
and institutions, the involvement in international programs has stimulated research
at the frontier and has contributed to further develop skills and competences of Italian
firms.
THE DYNAMICS OF THE SMALL FIRMS NETWORK AND THE CORE R&D
SYSTEM: VIRTUOUS AND VICIOUS CYCLES
One of the relevant aspects of the Italian case has been the presence of interdependen-
cies and virtuous and vicious cycles within and across the two systems of innovation.
Virtuous Cycles
Virtuous vertical innovative cycles took place between equipment producers and
users. The dynamics are as follows. Technologically progressive and highly competi-
tive users requested new advanced capital equipment to upstream producers that were
therefore stimulated to satisfy users' demand with innovative equipment. With the
availability of the new equipment, users were able to improve their own technological
capabilities and competitiveness, and, in turn, generated new demand for additional
improvements in capital equipment, and so on. A virtuous dynamic vertical cycle was
then set in motion.
In Italy one of the most relevant virtuous vertical cycles of this kind existed
between producers of manufacturing equipment and firms in the industrial district, as
a consequence of intensive learning by doing, learning by using, and learning by inter-
acting taking place between producers and users through formal and informal com-
munication, share of tacit knowledge, on the spot interaction, and personnel mobility.
As previously noted, in industrial districts (such as the textile one of Biella in Piedmont
or the ceramic tile one of Sassuolo in Emilia), a very innovative and dynamic group
of equipment producers linked to the production of the district is also present (such as
textile equipment firms in Biella and producers of equipment for ceramic tile produc-
tion in Sassuolo).
It must be noted that virtuous cycles have greatly affected the rate of diffusion of
new technologies. Given the advanced capabilities of flexible automation producers
and the proximity of several producers and users in the north of Italy, the diffusion of
flexible automation has been relatively faster in Italy compared to other European
countries. On the other hand, information technologies and EDP, not related to a
competitive strength of Italian producers, linked to different functions within firms
and acting on a different set of firms' capabilities, had a slower rate of diffusion in Italy
compared to other European countries (see Table 7.13).
Virtuous cycles also took place when upstream producers with advanced tech-
nological capabilities faced a large and technologically progressive user. This is the case
of the relationship between the robotics and laser industries and FIAT. In the case of
robotics, FIAT had a clear perception of its needs, an advanced knowledge of its pro-
duction process, and a willingness to invest a large amount of resources in new pro-
duction processes; robotics producers, on the other hand, disposed of accumulated
Table 7.13. Indicators of Diffusion of Various Technologies

Federal
Republic of United United
Italy Germany France Kingdom States Japan
Number of robots"
1988 8300 17700 5658 5034 32600 176000
Number of flexible manufacturing
systems*
1984 19 23 37 37 81 103
1988 69 117 65 82 118 190
EDP expenditures per worker' ($)
1988 541 888 862 743 1103 947
1990 630 951 959 772 1262 1088
Personal computer d (millions)
1988 1.2 1.7 1.7 1.9 28.0 6.1
1990 2.1 3.7 3.0 3.6 43.3 9.1
Personal computers per 100 workers'
1990 9.9 13.0 13.5 13.2 36.6 14.6
Digital switching penetration^
(share of total switching)
1987 11.9 1.5 69.7 48.4 76.2 36.1
Digital local lines places in services*'
(millions)
1988 I'.l .5 1.7 2.5 7.6 5.0
Sources: "''CERIS, MIPPolitecnico di Milano;<'~ ''ASSINFORM and NOMOS/* IEFEUniversita Bocconi.
capabilities in machinery development and production and an advanced international

competitiveness. Similarly, in the case of lasers FIAT promptly began a focused and
interdisciplinary R&D program on the utilization of lasers in car production, which
in turn benefited the production of high power lasers for industrial applications
(Kluzer, 1989).
Vicious Cycles
Vicious cycles have been present when the lack of competitiveness or advanced tech-
nological capabilities in one agent hindered the introduction of innovations or the
development of advanced technological capabilities by a second agent linked to the
first. This in turn reinforced the noncompetitive situation or the limited capabilities
of the first agent, and so on.
A clear example of vicious cycles is represented by the interactions among the
various participants in the core R&D system. The uneven level of scientific research
performed by the university and the public sector, the lack of technologically progres-
sive public procurement, and the limited technological capabilities of several large
firms have in fact hindered the full development of the national R&D system.
Dynamic interdependencies have also negatively affected the introduction of new
technologies or the birth of new industries in Italy. The lack of a population of
advanced small and medium size electronics firms highly impaired the development
of small power lasers in the late 1970s (Kluzer, 1989). The limited capabilities and
international competitiveness of final electronics goods producers negatively affected
the development of the electronic components industry during the 1960s and 1970s
(Malerba, 1987). Similarly, the limited experience and technological capabilities of the
electromechanical industry during the 1950s had negative consequences on the R&D
efforts of semiconductor firms during the 1960s (Malerba, 1987). Finally, the limited
R&D capability in Pharmaceuticals of the Italian chemical and drug companies and
the fragmented R&D of the Italian food processing industry were a major cause of the
late and difficult start of R&D in biotechnology during the 1980s (Galimberti et al.,
1989).
CONCLUSIONS
This chapter has shown that two systems of innovation with quite different organiza-
tions may coexist within the same country. In onethe small firms networksmall
firms compete and interact at the local level, in most cases inserted in a homogeneous
vocational, social, and cultural atmosphere. In the otherthe core R&D system
large firms, new small high-technology firms, large public institutions, universities,
and the national government operate in a complex web of relationships.
The very same factors that allowed for the success of the small firms network in
Italy seriously impaired the effectiveness of the core R&D system, namely, vivacious
individualism and entrepreneurship, strong family links, high localism, a limited oli-
gopolistic industrial core, an inability to manage large organizations, and a rather inef-
ficient public sector.
The case of Italy has also shown that a country can enter the club of advanced
industrial economies, and even prosper and grow, without a developed formal R&D
system.
These features of the Italian national system of innovation lead to wonder, how-
ever, if this situation will be sustainable in the future in the presence of a globalization
of competition as well as European integration.
To survive, the small firms network will have to modify some of its basic char-
acteristics and evolve toward a more concentrated structure, in which a limited num-
ber of leading firms or firm associations will perform a key role in both strategic and
commercial terms. This is the direction that the small firms network seems to have
taken in recent years.
Different predictions may be made with regard to the core R&D system. The Ital-
ian R&D system faces the choice of either remaining emarginated (as in the past) from
international technological competition among advanced industrialized countries, or
making the move from being a follower to a leader in some high-technology sectors.
The only way Italy can move toward a leadership position in even a limited num-
ber of fields is if the existing core is able to invest more in R&D, and if new big indus-
trial R&D segments and new small high-technology firms enter the scene. National
public policy will have to become more coherent and coordinated, universities will
have to be able to supply a greater number of skilled personnel, and a more effective
interface between industry and university will have to develop. In this respect global-
ization, while greatly increasing the competitive pressure in Italian firms, will at the
same time open up new opportunities for innovation and growth by stimulating the
generation, transmission, and acquisition of advanced technological knowledge and
expertise.
REFERENCES menti strutturali e strategic verso il 1993." IV

Rapporto CESPRI-Bocconi, Camera di Commer-
cio, Milan
Adams, P., and Orsenigo, L. (1988). "Tecnologie Cominotti, R., and Mariotti, S. (1990). Italia multi-
emergent! e politica Industrial in Italia." Rivista nazionali. Milan: F. Angeli.
di Scienza dell'Amministrazione 2:113-138. Consiglio Nazionale Ricerche (CNR) (Various
Amendola, G. (1989). "La diffusione dei material! years). Annual Report.
polimerici nell'industria automobilistica." In F. Contini, B., and Revelli, R. (1986). "Natalita e mor-
Onida and R. Malaman (eds.), Industria italiana talita delle imprese italiane: risultati preliminari e
ed alle technologie, 310-89. Milan: F. Angeli. nuove prospettive di ricerca." LTndustria 1,
Antonelli, C. (1984). "Element! per un'analisi della 2:195-232.
politica dell'innovazione in Italia: il processo di Dosi G. (1989). "Imprese ed istituzioni nello svi-
formazione di un potenziale scientifico e tecnol- luppo di nuovi paradigmi tecnologici in Italia. In
ogico nazionale." Economia Pubblica. 195-522. F. Onida and R. Malaman (eds.), Industria ital-
Antonelli, C. (ed.) (1988). New Information Tech- iana edalte tecnologie, 67-82. Milan: F. Angeli.
nology and Industrial Change: The Italian Case. Dosi, G., and Moggi M. (1989). "Diffusione delle
Dordrecht: Kluwer. tecnologie elettroniche ed evoluzione della strut-
Archibugi, D. (1987). "La struttura settoriale tura industriale contemporanea." In G. P. Bar-
dell'innovazione industriale in Italia." Rivista di betta and F. Silva (eds.), Trasformazioni struttur-
Politica Economica. February 132-76. ali delle imprese italiane, 207-30. Bologna: II
Archibugi, D., and Pianta M. (1990). The European Mulino.
Technological Specialization. Rome: CNR. Emrietti, A., and Fornengo, G. // Gruppo Fiat:
Battaggion, M. R. (1990). "R-S e brevetti." In F. dall'initio degli anni ollanta alle proslettive del
Onida and F. Malerba (eds.), La ricerca scienti- mercato unificato. Roma: Nuova Italia Scienti-
Jioz, 97-145. Rome: SIPI. fica, 1988.
Becattini, G. (1987). Mercato e forze locali: il dis- Falzoni, A. (1990). "La bilancia tecnologica dei
tretto industriale. Bologna: II Mulino. pagamenti italiana." In F. Onida and F. Malerba
Becattini, G. (1989). Modelli locali disviluppo. Bolo- (eds.), La ricera scientifica, 147-66. Rome: SIPI.
gna: II Mulino. Fornari, I., Malaman, R., and Morawetz, A. (1989).
Bellandi, M. (1989). "Capacita innovativa diffusa e The Relations between Public Research Organi-
sistemi locali di impresi". In G. Becattini (ed.), zations and Firms: The Case of Italy. Milan.
Modelli locali di sviluppo, 149-72. Bologna: II Working Paper CESPRI Universita Bocconi.
Mulino. Fornengo, G., and Silva, F. (1989). Growth Strate-
Belussi, F., and Festa, M. (1990). "L'impresa retedel gies of the Largest Italian Groups (1981-85).
modello veneto". Oltre ilPonle, 31:3-81. Manuscript: Turin.
Brusco, S. (1982). "The Emilian Model: Productive Galimberti, I. (1989). "Le communicazioni in fibra
Decentralization and Social Integration." Cam- ottica." In F. Onida and R. Malaman (eds.),
bridge Journal of Economics 6:167-184. Industria italiana ed alte tecnologie, 197-326.
Cainarca, G., Colombo, M., and Mariotti, S. Milan: F. Angeli.
(1989a). "An Evolutionary Pattern of Innovation Galimberti, I., and Klutzer, S. (1988). "I programmi
Diffusion. The Case of Flexible Automation." tecnologici comuni e il loro impatto sul sistema
Research Policy 18:59-86. ricerca-innovazione in Italia." In R. Malaman
Cainarca,G.Colombo, M., ana Mariotti, S. (1989b). and P. Ranci (eds.), Le politiche industrial! della
"Sentieri di automazione cd evoluzione della CEE. Bologna: II Mulino.
struttura industriale." In G. P. Barbctta and F. Galimberti, I., Gola, C., and Orsenigo, L. (1989).
Silva (eds.), Trasformazioni strutlurali delle "Le biotecnologie." In F. Onida and R. Malaman
imprese italiane. Bologna: 11 Mulino. (eds.), Industria italiana ed alte tecnologie, 27-
Camagni, R. (1984). II robot italiano. Milan: II Sole 112. Milan: F. Angeli.
24 Ore. Gandoln, V. (l98X).Aree sistema. Milan: F. Angeli.
Censis - IMI (1990). Ricerca e industria in Italia. Ginebri, S. (1987). "I progetti fmalizzati del CNR e
Vent'anni di fondo IMI. Rome: IMI-I1 Sole 24 le loro ricadute in termini di impulsi innovativi.
Ore. CESPE papers
Centro Studi ConfindustriaCSC (various years). Graziani, A. (1979). L'economia italiana dal 1945
La spesa dell'indimtria privata per la ricerca scien- adoggi. Bologna: II Mulino.
tijica. Rome. Gros Pietro, G. M. (1985). "L'innovazione
Centro Studi IBM ItaliaAssolombarda (1989). / neU'industria italiana, situazioni e problemi."
neolaureali in azienda. Milan: II Sole 24 Ore. L'Industria 6(2):2\7-233.
Cesaratto, S., Mangano, S., and Sirilli, G. (1991). ICE (1987), Dossier made in Italy: I beni ad alia tec-
"The Innovative Behaviour of Italian Firms: A nologia. Rome.
Survey of Technological Innovation and R&D." Istat (1988). "Indagine statistica suU'innovazione
Scientometrics 21:207-33. tecnologica nell'industria Italiana." December.
Cespri (1990). "L'integrazionc internazionale del Istat (Various years). "Indagini sulla ricerca scienti-
sistema di industria e servizi dellTtalia: rnuta- fica," Rome.
Kluzer, S. (1989a). "Alcune osservazioni sull'attivita Nencini, G. (1989). La ricerca scientifica in Italia.
di ricerca scientifica e tecnologica sui laser in Ita- Rome: La Nuova Italia Scientifica.
lia." In F. Onida and R. Malaman (eds.), Indus- Nones, M. (1988). La struttura del sistema difesa-
tria italiana edalta tecnologie, 85-193. Milan: F. industria in Italia. Manuscript: Genua.
Angeli. Nuti, F. (1988). "Trasformazioni strutturali delle
KJuzer, S. (1989b). "II settore dei laser commerci- piccole e medie imprese nei distretti industrial)."
ali." In F. Onida and R. Malaman (eds.), Indus- In the Conference "Contesto competitivo e tras-
lria italiana edalta tecnolognie. Milan: F. Angeli. formazioni strutturali delle piccole e medie
La Noce, M. (1990). Governmental Support to Ital- imperse nei distretti industrial! e nei rapporti di
ian Industry's R-D and Technological Innovation subfornitura." Milan, February.
Expenditures During the 1980s. Manuscript: OECD (1991). Review of National Science and Tech-
Stockholm. nology Policy. Italy." Paris: OECD.
Lassini, A. (1985). Gli inlerventi regionaliper i serv- Onida, F., and Malaman, R. (1989). Industria ital-
izi alle imprese. Milan: F. Angeli. iana edalte tecnologie. Milan: F. Angeli.
Lassini, A. (1990). "I centri di servizio Onida, F., and Malerba, F. (1990). La ricerca scien-
all'innovazione per le piccole e medie imprese." tifica. Rome: SIPI.
In F. Onida and F. Malerba (eds.), La ricerca Onida, F., and Viesti, G. (eds.) (1987). Italian Mul-
Scienticica. Rome: SIPI. tinationals. London: Croom Helms.
Malaman, R. (1989). "Lo sviluppo dei nuovi mater- Piore, M., and Sabel, C. (1984). The Second Indus-
iali ceramici." In F. Onida and R. Malaman trial Divide: Possibilities for Prosperity. New
(eds.), Industria italiana ed aha tecnologie, 132- York: New Basic Books.
226. Milan: F. Angeli. Pontarollo, E. (1986). Domanda pubblica e politica
Malerba, F. (1987). Dalla dipendenza alia capacita industriale: FS, SIP, ENEL. Milan: Marsilio.
tecnologica autonoma. Milan: F. Angeli. Raffa, M., and Zollo, G. (1987). "Innovazione e
Malerba, F. (1988). "La dinamica di lungo periodo crescita delle piccole e mdie imprese del settore
della R-S dell'industria italiana." Rivista di poli- automazione, strumentazione e sistemi." L'lm-
tica economica April. 467-96. presaN.2:197-226.
Malerba, F., and Falzoni, A. (1991). "Tecn'ologia e Raffa, M., and Zollo, G. (1988). "Dinamiche e stra-
dinamica settoriale nello sviluppo economico ital- tegic delle piccole imprese innovative." L'lndus-
iano." In C. Filippini (ed.), Innovazione tecnolo- tria 3:419-69.
gica e servizi alle imprese. Milan: F. Angeli. Riva, A. (1983). "Sviluppo e innovazione nella pic-
Malerba, F., and Orsenigo, L. (1991). "L'accumu- cola e media impresa." IRER. Milan: F. Angeli.
lazione delle capacita tecnologiche nell'industria Romano, A. (1989). Mezzogiorno 1992. Milan: F.
italiana (1969-1984)." In C. Filippini (ed.), Angeli.
Innovazione tecnologica e servizi alle imprese. Rullani, E., and Zanfei, A. (1988a). "Networks
Milan: F. Angeli. between Manufacturing and Demand: Cases from
Mariotti, S. (1989). "II riordino della funzione tec- Textile and Clothing Industry." In C. Antonelli
nica di produzione dell'industria." In V. Balloni (ed.), Information Technology and Industrial
(ed.), Processi di aggiustamento delle Industrie Change: The Italian Case, 57-95. Dordrecht:
negli anni 80, 233-65. Bologna: II Mulino. Kluwer.
Mariotti, S., and Colombo, M. (1987). "L'innova- Rullani, E., and Zanfei, A. (1988b). "Area Net-
zione di processo nell'industria italiana." L'ln- works: Telematics Connections in a Traditional
dustria 2:27-66. Textile District." In C. Antonelli (ed.), Informa-
Mariotti, S., and Onida, F., (1986). L'ltalia Multi- tion Technology and Industrial Change: The Ital-
nazionale. Milan: II Sole 24 Ore. ian Case, 97-113. Dordrecht: Kluwer.
Momigliano, F. (1985). "Determinant;, tipologiaed Russo, M. (1985). "Technical Change and the
effetti dell'innovazione come fattore di competi- Industrial District." Research Policy 3:329-43.
tivita." In F. Onida (ed.), Innovazione, competitiv- Santarelli, E., and Sterlacchini, A. (1989). "Patterns
itd e vincolo energetico, 197-286. Bologna: II regionali di innovazione tecnologica in Italia. R-
Mulino. S, brevetti sull'estero, imprese innovative." L'ln-
Momigliano, F. (1986). Le leggi della politico indus- dustria 1:25-56.
triale in Italia. Bologna: II Mulino. Sapelli, G. (1989). Modelli di crescita eprogresso lec-
Morandini, C. (1990). "II ruolo delle piccole e medie nico. Riflessioni sull'Italia. Manuscript: Milan.
imprese." In F. Onida and F. Malerba (eds.), La Sirilli, G. (1987). "Patents and Inventors: an empir-
ricera scientifica, 249-91. Rome: SIPI. ical study." Research Policy. 16:157-74.
Mussati, G. (1988). "II ruolo della piccola e media Soria, L. (1979). Injormalica: un 'occasions perduta.
impresa nella struttura industriale italiana." Cen- Turin: Einaudi.
tro F. Cicogna, Working Paper No. 6. Vacca', S. (1986). "Internazionalizzazione delle
Napolitano, G., and Sirilli, G. (1990). "Technical imprese: passaggio obbligato per lo sviluppo tec-
Change in the Italian Manufacturing Industry: A nologico o veicolo di dipendenza?" Kconomia e
User Oriented Innovation Model." ISRDS-CNR Politica Industriale 49:3-24,
Working Paper, Rome.
PART II
SMALLER
HIGH-INCOME
COUNTRIES
Part II is concerned with the innovation systems of four "small" high-income
countries. Their levels of gross domestic product per capita are comparable to
those of the countries considered in Part I, and stand high above the income
levels of the countries considered in Part III. In terms of literacy rates, basic
enrollment in secondary education, and in third level education, these coun-
tries look a lot like those in Part I, with Canada being very similar to the United
States. However, in terms of population these countries are much smaller.
Even the largest of them, Canada, has less than half the population of the large
European countries, and is far smaller than the United States or Japan. The
populations of the other three countries are even smaller. Thus their internal
markets are much smaller than those of the countries considered in Part I.
Two of these economies, the two largest ones physicallyCanada and
Australiagrew up as colonies of Great Britain (with Canada also having an
old French connection), which significantly shaped their economic roles and
international economic connections, at least to the close of World War II. The
other two are on the European continent, which also has shaped their pattern
of development and trade.
Aside from its small population, Sweden in many ways now looks like
the large industrialized nations in Part I. She is in their league regarding man-
ufacturing output as a fraction of Gross Domestic Product, manufacturing
exports account for most of her total exports, and her ratio of R&D to GNP
is close to that of the highest of the large industrial nations. However, reflecting
her small internal market, Sweden exports are a much larger fraction of her
GDP than the countries considered in Part I, except for Germany.
In the other three countries manufacturing output as a fraction of Gross
Domestic Product is significantly below that of the United States, the large
European countries, and Japan, and manufacturing exports account for a
smaller fraction of total exports. Agriculture and resource exploitation
account for a larger share of economic activity than in the six large countries
considered in Part I, and a larger share of exports are based on these activities.
And the ratio of R&D to GNP in these countries stands significantly lower
than that of the countries considered in Part I, except for Italy. This largely is
associated with their smaller share of manufacturing activity in total GNP,
and in particular their minor participation in some of the most R&D-intensi ve
industries.
With the exception of Sweden, in these countries business R&D accounts
for a significantly smaller fraction of total R&D. A larger fraction goes on in
universities and government laboratories (Table II. 1).
263
264 SMALLER HIGH-INCOME COUNTRIES
Table H.l. Comparison of the Four "Small" High-Income Countries

Sweden Denmark Canada Australia
GDP/capita, 1 989 official exchange 19,300 18,450 16,960 12,340
rates
GDP/capita, 1988 purchasing 14,772 13,555 18,446 13,412
power parity
Population, 1988 8,438 5,130 25,950 16,538
Average growth rate GDP/hour 1.8 1.8 2.7 1.7
average 1965-1988
Gross domestic investment GDP 15 15.5 24 22
average 1965-1988
Manufacturing output/GDP 30 25 23 18
Manufacturing exports/GDP 26 19 16 5
Total exports/GDP 31 31 25 10
Literacy rate >95 >95 >95 >95
Secondary level enrollment rate 91 107 104 98
Third level enrollment rate 31 30 58 29
Scientists and engineers/population 0.25 0.17 0.22 0.21
R&D/GNP 3.0 1.3 1.5 1.4
Private R&D/total R&D 61 47 42 61
Business R&D/total R&D 66.8 55.6 55.0 37.4
Private business R&D/total business 87.0 85.9 71.8 93.3
R&D
8
Comparing the Danish and Swedish
Systems of Innovation
CHARLES EDQUIST
BENGT-AKE LUNDVALL
SIMILARITIES AND DIFFERENCES BETWEEN DENMARK AND SWEDEN1
Most comparative work on national systems supporting technical change has focused
on the differences between the United States and Japan. Such comparisons may be
very useful as first approximations; they demonstrate, lucidly, how differences in
national economic structures and institutional setups are reflected in differences in the
rate and direction of technical change. In this chapter Sweden and Denmark, i.e.,
countries that are much more similar, are compared.
From a non-European perspective, Sweden and Denmark might seem almost
identical.2 But the closer the observer gets the more visible the differences become. In
this chapter it is proposed that the Danish and the Swedish systems of innovation have
quite different characteristics. It is argued that these characteristics are embedded in
the economic structure and in the socioinstitutional setup and that they have strong
and deep roots in the economic history of the two countries.3
Proximity and Common Characteristics
The distant observer is certainly right in perceiving striking similarities between Den-
mark and Sweden. It takes less than half an hour to cross the Sound between them,
and on arrival in the neighboring state you can read its newspapers and, with a minor
effort, take part in conversation with its citizens. There are few legal difficulties limiting
movements across the borders. You do not have to show your passport, and you have
the same right as the indigenous population to enter the labor market.
The living standard, life-style, and consumption pattern do not seem to vary
much between the two societies. In both countries the public sector is in charge of sys-
tems for education and health service, which are of a comparably high quality. This is
reflected in high tax rates and generally, of course, in a large public sector. Important
factors behind the expansion of the public sector have been the strong Social Demo-
cratic presence in politics and the strong trade unions.
265
Looking back in history, we find important parallels that might partly explain the
similarities of today. Both countries were industrial latecomers. In 1979 the two coun-
tries were ranked among the 10 countries having the highest GNP/capita (Sweden as
number 6 and Denmark as 8). Apparently, both countries have been successful in
exploiting technology to stimulate economic growth. Thus, it is easy to understand
why a distant observer would tend to speak about these two small and rich countries
as reflecting one single Scandinavian, or Nordic, model. One point to be made in this
chapter is that such similarities in overall aggregate performance might coexist with
and even concealradical differences between national systems of innovation. Also,
it implies that quite different institutional setups may be effective in exploiting tech-
nological opportunities. One question, however, is whether the Nordic systems of
innovation, so successful historically, will remain so in the present and prospective
global context.4
Conspicuous Differences in indicators on Innovative Activities
Although the long-term growth records look rather similar, the most up-to-date indi-
cators on innovative activities put Denmark and Sweden into two different leagues.
We can see from Table 8.1 that Sweden is a country investing heavily in R&D whereas
Denmark is characterized by an internationally very modest R&D effort. Turning to
output indicators we find that Sweden is one of the countries with the highest numbers
of patents per capita in the United States, whereas Denmark has a very weak record
in this respect.5
What is the background for these dramatic differences between the two countries?
We shall argue that they reflect qualitative differences in the national systems of inno-
vation and that these differences can be understood only if we take into account spec-
ificities of the historical process of industrialization in the two countries.
We will also propose that one important, contemporary, factor behind the
observed differences in registered innovative activities is the degree of economic con-
centration, and the related role of domestically based multinational capital in the two
economies.
Table 8.1. Comparison of R&D Intensity (1985) and Number of United

States Patents Per Million Inhabitants in Denmark, Sweden, United
Kingdom, United States, and Japan, Annual Averages 1980-1985
United States Patents
R&D Expenditures/GDP Per Million Inhabitants
Innovation Indicators (%) 1980-1985
Denmark 1.25 27
Sweden 2.78 89
United Kingdom 2.33 41
United States 2.78 158
Japan 2.81 79
Source: R&D expenditures from OECD (1988a, p. 13). United States patents for Denmark (the average
for 1980-1982) from Mjoset (1986, p. 81) and for the other countries from Pate! and Pavitt (1989, p.
20).
DANISH AND SWEDISH SYSTEMS OF INNOVATION 267
Again, the two countries present themselves as very different as shown in Table
8.2. In Denmark the small and medium-sized enterprises dominate in manufacturing
and only a few internationally small multinational firms have their home base in Den-
mark. In Sweden the degree of concentration in terms of ownership and control is very
high, even in international terms, and the amount of capital abroad controlled by
Swedish firms is large, relatively speaking.
The Structure of This Chapter as Reflecting Conceptual Matters

The rest of this chapter is divided into four main sections. The first gives a brief his-
torical sketch of the processes of industrialization in Denmark and Sweden. The sec-
ond describes the anatomy of the two systems of innovation in terms of the socioin-
stitutional setup, the structure of production, and the R&D system. The performance
of the systems is discussed in the next section. Finally some prospective problems are
addressed. This structure reflects our understanding of what constitutes a national sys-
tem of innovation, as outlined below.
The national system of innovation6 is constituted by the institutions and eco-
nomic structures affecting the rate and direction of technological change in the society.
Obviously, the national system of innovation is larger than the R&D system. It must,
for example, include not only the system of technology diffusion and the R&D system
but also institutions and factors determining how new technology affects productivity
and economic growth. At the same time, the system of technological change is, of
course, less comprehensive than the economy/society as a whole.7
We assume that technical competence is built in a cumulative process. This is the
reason why we find it useful to give a brief historical sketch of the industrialization
process. The cumulative character of the process of technical change reflects that inter-
active learning is at the center of the process of technical change. Individual agents and
organizations increase their knowledge in technical matters, not in isolation from each
other, but in a process of interaction, involving learning from each other as well as
producing new knowledge and innovations in cooperation.8
We also assume learning to be a process strongly based in routine economic activ-
ities. At the firm level, it involves workers and technicians (engaged in production rou-
tines), marketing people (engaged in selling to the ordinary customers), as well as lab-
oratory personel (engaged in routine experiments). One result from learning is that
new problems are registered, and entered on the agenda of organizations. But it also
Table 8.2. Indicators of Economic Concentration and of

Domestically Based Multinational Firms
Foreign Employment in
Share of Value Added in Domestic Large Firms,
Manufacturing by the 10 1984 (Absolute
Biggest Firms, 1982 (%) Number)
Denmark 14 51,000
Sweden 33 256,000
Source: Data on value added from Nordic Growth (1984, p. 185) and on foreign
employment from Industriokonomisk Institut (1986, p. 61).
produces important inputs to the process of solving the problems faced by organiza-
tions engaged in searching (R&D departments, etc.).
This is the general background for the section on the anatomy of the national
system of innovation. The socioinstitutional setup focuses on corporative arrange-
ments forming a framework for interactive learning. The development blocks define
areas of specialization giving such processes their content and direction. The R&D
efforts, representing attempts to expand knowledge through searching rather than
learning, will, as well as technology policies, be rooted partly in the production system.
Having stressed interactivity and learning, it is analytically useful to make certain
distinctions with regard to the process of technical change. In the section on perfor-
mance, two such distinctions are used. First, it is fruitful to distinguish between prod-
uct technologies and production technologies (or process technologies) in the sphere
of production. Product technologies are or are included in products (goods or services)
and production technologies are used in the process of production to produce goods
and services.
Second, it is also useful to make an analytical distinction between development
of technologies and their diffusion.9 The development of technology involves the pro-
duction 'of technically feasible prototypes (inventions) and their modification and
development into economically feasible technologies (innovations).10 A breakthrough
innovation is here assumed to be made only once in time and space. Everything else
is diffusion, which includes absorption or assimilation of technologies developed in
other countries.'' Diffusion also includes adaptations and incremental innovations. In
other words, technologies are gradually modified and changed during the whole pro-
cess of technical change (i.e., also during the process of diffusion).12
It may sometimes be difficult empirically to make a clearcut distinction between
development and diffusion. And the same artifact may sometimes be a product tech-
nology and sometimes a process technology.13 Sometimes, but not always, a radically
new product technology will also involve a need for new process technology; they may
mutually condition each other. But at the same time, the distinctions prove to be very
useful, as analytical tools, when it comes to analyze and characterize the performance
of national systems of innovation.
The fact that a country is strong in the development of new technology, for exam-
ple, does not rule out that it has great difficulty in terms of diffusionreflecting a weak
ability to absorb new technologies (products or processes) into its production system.
And the fact that a country is extremely strong in terms of the diffusion of process
technology does not rule out that it has great weaknesses in the diffusion of new prod-
uct technologiesreflecting a weak ability to take up the production of new products,
characterized by strong growth, and developed abroad. As we will see, there are exam-
ples of both these kinds.
In the section on performance, Denmark and Sweden are compared regarding
their capacity to develop new technologies. We then focus on the diffusion of new pro-
cess technologies and, finally, we look at the diffusion of new product technologies.
The analyses demonstrate that the capability profiles of Sweden and Denmark differ
in these dimensionswhich confirms the usefulness of the distinctions made.
The understanding of technical change as a cumulative process based on inter-
active learning implies that national systems tend to be geared to follow specific tra-
jectories. The close relation between learning and searching is crucial when it comes
to moving rapidly ahead along given technological trajectories. However, the close
relation might also result in locking the system of innovation into a specific trajectory
in a period characterized by a change in technology paradigm. In the long run, this
may cause problems in terms of stagnation, especially as it may involve a weakness in
relation to the diffusion of product technologies (i.e., a weak capability to absorb new
product technologies developed abroad). This problem is discussed briefly with regard
to the two Nordic systems of innovation in the final section on structural and insti-
tutional problems.
There are several important elements of the national systems of innovation which
we have not included in this analysis of Sweden and Denmark. Examples are the finan-
cial system and the institutional characteristics of the R&D system. It is especially true
for the education and training system and the work organization.14 We believe that the
interface between technology and the development of human resources is critical for
the prospective survival and growth of national systems of innovation but this per-
spective will have to be elaborated elsewhere.
HISTORY: THE PROCESSES OF INDUSTRIALIZATION IN SWEDEN AND

DENMARK
The present profile of the two systems of innovation illustrates the cumulative char-
acter of processes of technical change and the stubborness of historically established
institutional setups and economic structures. The two economies have taken part in a
rapid, and far-reaching, process of internationalization. Nevertheless, it is to an impor-
tant extent possible to trace their present patterns of specialization and current specific
strengths and weaknesses in their systems of innovation back to how their respective
economies were industrialized.
From Grain to Butter and Meat: The Establishment of the Danish Agroindustrial
Complex
In the 1870s, the Danish economy was predominantly agrarian and linked to the world
economy mainly through its grain exports to the United Kingdom. In this decade new
competitive grain producers (Russia and the United States) entered the British market.
The Danish response to the resulting dramatic fall in grain prices (40% price reduction
1880-1890), and to the ensuing crisis in Danish agriculture, was a reorientation of
production from grain to pigs and cattle and, also, gradually toward more processed
products.15
It is interesting to note that the cream separator (a continuous centrifuge), which
was to play a crucial role in making the dairy industry a modern process industry, was
invented almost simultaneously in Sweden and Denmark in the 1870s.16 The rapid
diffusion of this technical innovation through the Danish economy in the following
decades was combined with an important social innovationthe introduction of
cooperatively organized ownership of dairies. This form of organization proved to be
an efficient framework for the modernization of both primary agriculture and the sec-
ondary industries refining milk and cattle into butter and meat; by 1910 90% of total
exports were processed agricultural products such as bacon, butter, and cheese.17
In the wake of the modernization of agriculture and as farmer's income level rose,
the domestic market for industrial products expanded. In the 1890s, and again in the
1930s, new home market industries developed, but it was not until the 1960s that an
increasingly export-oriented manufacturing sector developed. As late as 1958, 61% of
commodity exports emanated from agriculture (including agroindustrial products
such as butter, cheese, and bacon) and 39% came from manufacturing. In 1968, this
had been reversed: 62% came from manufacturing and 38% originated in agriculture
(Andersen, 1972, p. 62ff).
As we shall see, the history of the transformation of the agroindustrial complex is
reflected both in the pattern of specialization of Danish technical competence and in
the organization of innovation. The cooperatively owned agroindustrial sector has
been very successful in developing, producing, and marketing food products of a uni-
form, and high, standard. State-supported development work and quality control have
interacted with the efforts of the cooperatives aiming at a high and stable quality. The
success is reflected even in the present export specialization of Denmark. Butter,
cheese, ham, and bacon are still dominating export products earning an important
part of foreign currency. The export specialization indices presented in Table 8.3 illus-
trate this fact.
One strength of cooperative ownership has been its stimulation of rapid technol-
ogy diffusion. A weakness has been the one-sided focus of innovative activities on pro-
cess technologies.'8 The R&D efforts have generally been small in the Danish agroin-
dustrial sector19 and this is especially true for R&D aiming at new product
technologies.
Table 8.3. Export Specialization for Main Product Groups

and for Some Dominating Natural Resource-Based Export
Commodities in Denmark 1973-1987"
Product* and SITC Number 1973 1979 1987
Products based on natural resources 1.3 1.5 1.6

Bacon, ham, etc. (012.1) 49.4 46.9 30.4
Prepared preserved meat (013.8) 33.6 28.7 28.8
Butter (023.0) 10.3 6.8 4.7
Oil and gas 0.7 0.8 0.8
Chemicals 0.7 0.7 0.8
Engineering 0.7 0.7 0.6
Other industry 1.2 1.1 1.4
"The export specialization figures in Tables 8.3, 8.4, and 8.8 are calculated as
"revealed comparative advantage" indices:
where Ry is the export specialization index of country / in commodity j\ A# is

exports from country i of commodity j, Xj is total exports from country /', Xj is total
OECD exports of commodity), and JIT are total OECD exports (Daium et al., 1988,
p. 134). If the value is 1.0 there is no specialization. The more the values deviate
from 1.0, the stronger is the specialization. If the value is larger than 1.0 there is a
large export of the product (group) in question. If the value is below 1.0, this is a
reflexion of a small export of the relevant product.
''The exact content of the product groups is specified in Dalum {1989, Appendix
1).
Source: IKE database (Bent Dalum).
From Agriculture, Forestry, and Mining toward Engineering in Sweden
Sweden was, of course, also a predominantly agrarian country in the latter half of the
nineteenth century. Agricultural products were also exported. From about 1830 Swe-
den began exporting oats to England.20 During some years in the late 1870s, oats was
the second largest export product (after wood). The export of oats decreased sharply
after 1880, but was immediately replaced by a large butter export.
However, in contrast to Denmark, a large part of Swedish exports had its origin
in forestry and mining rather than agriculture. And this export did not consist of plain
raw materials, but of refined products. From the seventeenth century to the mid-nine-
teenth century iron was the main mining-based export product.2' During the latter half
of the nineteenth century new processes made possible production of cheap steel
also from the phosphor-rich iron ore abundant in Sweden. Somewhat later iron and
steel began to be transformed into machinery products before export. The forestry-
based export commodities were from the mid-nineteenth century "plank and
boards"processed by the sawmill industryas opposed to timber. Later, wood was
further refined and exported as pulp and paper.22
The fact that the export products emanating from mining and forestry were
already refined products created, in the case of iron and steel, an important metallur-
gical capability. Hence, an important part of the nineteenth-century transformation
in Sweden took place outside agriculture. In particular, the engineering industry
becameand still isa strategic sector for the modernization of the Swedish econ-
omy around the turn of the century. Employment in the engineering industry
increased from 23,000 in 1890 to 63,000 in 1912, while employment in mining and
iron works decreased from 35,000 to 32,000 during the same period (Gardlund 1942,
p. 279).
This expansion of the engineering industry was gradually reflected in exports. In
1880 it constituted 3% of total exports, growing to 10.5% in 1911-1913 (Senghaas,
1982, p. 131). In the interwar period the proportion of machinery and transport equip-
ment kept growing and in 1950 its share of total export had passed 20% (Svennilson,
1954, p. 180ff). Its share of manufacturing exports was now getting close to 50% and
only in the United States was this share higher among OECD countries.
The development of the Swedish engineering industries reflected a small number
of specific technical innovations introduced by Swedish inventors and entrepreneurs.
Among early important Swedish technical innovations we find an automatic machine
for cutting matches (Lagerman invented it around 1860), the separator already men-
tioned above (invented in the 1870s), and new methods for processing pulp (developed
by Ekman in the beginning of the 1870s). It is important to note that these three early
inventions were all closely related to the export-oriented process industries.
Several important Swedish innovations in mechanical and electromechanical
systems took place around the turn of the century. Several of the largest Swedish mul-
tinational firms may, actually, be regarded as originally based on single product inno-
vations. This is certainly true for Ericsson (innovations exploiting low current tech-
nology by Lars-Magnus Ericsson), SKF (the development of ball bearing technology
by Wingquist), ASEA (the development of a three-phased motor by Wenstrom), and
Alfa Laval (the separator). The focus on single inventions and inventors might be mis-
leading to a certain degree. All of these technological breakthroughs were, obviously,
cumulative in the sense that they were built on a competence reflecting several decades
of experience with metal working and with technical development in engineering.
In Sweden, an important outcome of the process of industrialization was the
combination of exports based on refined and processed raw materials on the one hand
and the multinational engineering firms on the other. The Swedish system of inno-
vation still reflects this combination both in its institutional setup and in its pattern of
specializationas indicated by the export specialization indices presented in Table
8.4. In particular, it is the small number of multinational firms in the engineering
industry that plays a decisive role in this system.23
ANATOMY: THE TWO NATIONAL SYSTEMS OF INNOVATION
Introduction
In this section we will discuss the anatomy of the Swedish and Danish systems of tech-
nological change. We begin with a discussion of the socioinstitutional setups in the two
countries. The focus is on the corporative interaction between the labor unions, the
welfare state, and private capital.
We then take a closer look at the patterns of specialization and the production
structure in the two countries. Among other things, we point out a few strategic devel-
opment blocks in each country.
Finally, we address the R&D systems and the role of the state in relation to tech-
nological change (i.e., technology policy). We end by briefly characterizing the two
systems of innovation.
Social and Political Similarities and Differences
Both Denmark and Sweden are modern welfare states. The size of the public sector
and, especially, of the parts engaged in social security, health care, and education put
them in a class by themselves. And, since the 1930s, both countries have been char-
Table 8.4. Export Specialization for Main Product Groups

and for Some Dominating Natural Resource-Based Export
Commodities for Sweden 1973-1987"
Products* and SITC Number 1973 1979 1987
Product based on natural resources 1.1 1.1 1.1
Mechanical wood pulp (25 1 .2) 20.0 31.0 20.5
Kraft paper (64 1.3) 18.5 24.6 24.9
Iron and steel powder (67 1 .3) 14.8 16.9 16.7
Oil and gas 0.2 0.5 0.7
Chemicals 0.5 0.5 0.6
Engineering 1.1 1.1 1.0
Other industry 0.9 0.9 0.9
"For definition of export specialization index, see note a in Table 8.3.
''The exact content of the product groups is specified in Dalum (1989, Appendix
1).
Source: IK.E database (Bent Dalum)
acterized by strong trade unions and by a strong political presence of Social Demo-
cratic parties.
However, there are also important differences of a social and political character
between the two countries. In Sweden a stable modernist social liberal norm system
dominates the political culture and institutions while political culture is less homo-
geneous and stable in Denmark. For example, the room for political deviations to the
left and right has been greater in Denmark than in Sweden. The Social Democratic
dominance of the political scene and inside the trade unions has been less continuous
and hegemonic in Denmark than in Sweden.
Differences in the structure and institutions of private economic power, rooted in
the history of industrialization, are also important influences on the differences in
political culture in the two countries. As we have seen in Table 8.2, a small number of
private firms dominate Swedish manufacturing. Also ownership and financial control
are much more concentrated in Sweden than in Denmark. In Denmark the farmers'
movement and the small owners have played a much more important role. The Swed-
ish corporative interaction between big capital and labor has contributed to the sta-
bility of political development while it has been more difficult to establish long-term
and corporatist social compromises in Denmark because none of the parties has been
strong enough to guarantee the fulfillment of such a contract.24
The Labor Market

The more "modern" character of the Swedish system and its more stable context for
corporatist cooperation are reflected in the institutions around the labor market. In
Denmark the influence of preindustrial organizational forms is reflected in the trade
union structure, as well as in the systems for training workers. The trade unions in
Sweden are centralized and organized along industry lines, while the Danish unions
are organized according to profession and competence. The training system in Den-
mark is a dual system where one of its elements may be regarded as a remnant from
preindustrial artisan training.
Labor market conditions differ greatly between the two countries. In Sweden the
registered unemployment rate has never in the postwar period reached 4%, while it has
reached and permanently stayed above 8% in Denmark since the 1973 oil crisis. In
1990 it is about 1.5 and 10%, respectively.25 The very low level of unemployment in
Sweden means that most sectors and regions experience a severe shortage of labor.
Both countries spend considerable amounts of money on labor market policy. In Swe-
den most of the money is used to retrain or to reemploy hit or threatened workers. In
Denmark almost all of the expenditures go to cash payments to the many unemployed
workers.
The Swedish Model and Its Impact on the Process of Technical Change
In Sweden a fundamental historical compromise on rationalization and technical
change was established in the 1920s and 1930s. In the late 1920s and early 1930s, the
leadership of the trade union movement changed its attitude to the introduction of
new technology, and to rationalization of the labor process, into a much more posi-
tive one.
Originally, the trade unions had been extremely negative to Taylorism and ratio-
nalization, regarding them as threatening the workers with technological unemploy-
ment. One of the reasons for the change in attitudes was the beliefrightly or
wronglythat the Social Democratic government, established in 1933, could guar-
antee that new production technology would not result in widespread unemployment.
Another reason was their belief that unions would be capable of appropriating what
they considered their rightful share of the productivity gains. Hence, the very positive
attitude toward technical change of the Swedish unions has been around for many
decades now. This is one sense in which the term "Swedish model" can be used.
Johansson (1989) shows in detail how agreements between labor and capital
with regard to rationalization, technical change, and related mattershave been
reached through a complex process. In the majority of conflicting issues, the unions
have, in negotiations with the employers, threatened to use their privileged links with
(Social Democratic) state power and to "solve" the problems through legal means.
However, this threat has, normally, led to an agreement through negotiations between
the labor market partners, without any formal state interference. But, indirectly, the
welfare state and its commitment to full employment policies has been an important
precondition for the lasting viability of this compromise. Hence, the emergence of the
"Swedish model" in this sense has been characterized by an interplay between three
parties: (efficient and concentrated) capital, (strong and centralized) trade unions, and
the (Social Democratic) state.
The permanent shortage of workers in the postwar period means that firms com-
pete with each other in attracting workers. One means of competition is salary.
Another possibility is to make jobs more interesting and stimulating. Firms are pushed
to redefine jobs in order to enlarge the scope of tasks allocated to each worker and
thereby make the jobs more attractive. This means that the shortage of labor has
important consequences for the quality of jobs (Edquist and Glimell, 1989, p. 43).
The shortage of labor may also influence management strategies and work orga-
nization ideas and thereby the sociotechnical and organizational design of new fac-
tories. One example, although atypical in its radicalism, is the Volvo Uddevalla fac-
tory. The shortage of labor, combined with inferior working conditions in traditional
automobile assembly, has contributed to a high turnover of workers and to absentee-
ismwhich means large costs for the employer. This has been one of the reasons why
Volvo has recently designed the factory in Uddevalla according to new principles. In
Uddevalla there is no moving assembly line at all. Instead a small group of workers
"builds" (which is the new word for assemble) the whole car in a 2-hour cycle26 (Edqu-
ist and Glimell, 1989, pp. 43-44).
The term "Swedish model" can also be used to denote the Swedish model for
economic policy, including union-managed solidaristic wage policy, which developed
after World War II. One important aspect of the Swedish model, in this sense, was
acceptance of and even active support of structural change of the trade unions. If an
industrial sector was ailing it was assumed to reflect low relative productivity and/or
stagnating demand. General economic policies pursued by the state (through invest-
ment funds and labor market policies) and by the centralized trade unions (through
solidarity wage policies, securing the largest wage increases for the lowest paid worker)
combined forces in stimulating a transfer of resources from the ailing sectors to growth
industries with better long-run prospects. Gosta Rehn and Rudolf Meidner were the
architects behind the Swedish model of wage formation.
The Danish Model and Its Impact on the Process of Technical Change
The dominance of small scale production, and the historical importance of the coop-
erative movement in agriculture, has limited the influence of big capital in Denmark.
One important consequence of the success of the cooperative movement in gaining
control in the dominating export sectors was that it closed these strategic sectors to
private capital. This is reflected in the weak position of financial capital in relation to
the Danish production system.
The survival of small scale and artisan-like production has fostered a kind of cor-
poratism, very different from the Swedish. Small, independent, entrepreneurs in Den-
mark will often be quite negative to central trade union power, but at the same time,
often willing to cooperate, locally, with their workers and their representatives. They
will often be driven by incentives less oriented toward firm growth and more toward
the maintenance of an independent, and reasonably comfortable, life-style.
This small-scale corporatist model often involving a flexible use of reasonably
advanced production equipment and a continuous development of incremental prod-
uct innovations has its strength in flexible adaptation. But a great weakness seems to
be that it does not give enough incentives to use and develop human resources inside
firms. In the 1980s Danish firms have invested heavily in advanced process technol-
ogies and the disappointing outcome in terms of poor productivity performance
reflects, primarily, problems with labor organization and skill development (Gjerding
etal., 1990).
The workers' attitude to rationalization seems to be different in Denmark than
in Sweden. Some years ago, a questionnaire sent to a representative sample of Danish
citizens asked about attitudes toward "new technology." Two-thirds of the respon-
dents expected mainly a negative impact on employment, as well as on work satisfac-
tion. The negative expectations were strongest among the unskilled workers but, even
among small owners, the negative expectations dominated (Petersen et al., 1983).
Economic policies at the national level reflect the fact that the corporative parties
are weak and heterogeneous. Although incomes policies, at least until recently, was a
trade union responsibility in Sweden it is the state in Denmark that has repeatedly tried
to regulate incomes and wages. It has proved difficult to make incomes policies effi-
cient without a high level of unemployment, however.
On the Structure of Production as Reflected in Development Blocks
When comparing the structure of production in Sweden and Denmark, it is useful to

isolate a small number of development blocks, in each of the two countries. This con-
cept was developed by the Swedish economist Erik Dahmen. Dahmen's interest is in
the transformation of industry and trade. Therefore he focuses on changes over time
within and between microentities. According to Dahmen such changes are the essence
of industrial dynamics. They imply disequilibria, which should not, however, be
regarded as disturbances because they are essential in transformation processes (Dah-
men, 1970, 1988, p. 4).
A development block is a sequence of complementaritiesbetween technologi-
cal, economic, and other related factorswhich by way of a series of disequilibria (or
structural tensions) may result in a balanced situation.27 Dahmen talks about devel-
opment blocks and structural tensions at different levels of aggregationfrom the fac-
tory floor over the interplay between enterprises to a broader socioeconomic context.
We will here use the concept at a level of aggregation lower than the national economy
but still not focusing on distribution of production and employment only at the level
of industrial sectors.28
A development block will, typically, involve several domestic sectors, coupled by
strong quantitative and qualitative linkages. We assume that development blocks
form frameworks for interactive learning and that the specific combination of such
blocks at the national level will have an important impact on the direction of processes
of technological change.
Development Blocks in the Danish Economy

In the historical review we saw how an export-oriented development block had already
formed around the agroindustrial activities at the beginning of this century. Still, at
the end of the 1970s slightly more than 42% of total nonservice employment and non-
service exports emanated from this block (Lundvall et al., 1984, p. 23). Its present
importance is also reflected in the export specialization figures for the postwar period
presented in Table 8.3. The successful transformation of the Danish agriculture in the
last part of the nineteenth century is still strongly reflected in the pattern of export
specialization.
Actually, the block encompasses two different modes of development and two
different modes of innovation. The dominating one has at its center the farmer-owned
cooperative process industries. The other mode is dominated by a few large (according
to a Danish standard) private food-producing firms. Some of these firms are, or have
been, semipublic firms, started as licensed monopolistic producers of, for example,
liquor or sugar and several of them are controlled by the old merchant and shipping
firm 0K.
We have already commented on the strengths and the weaknesses of the coop-
erative model. The one-sided focus on standardization and rationalization of the labor
process in slaughteries and dairies combined with the neglect of product innovation
reflects the mode of competition and ownership in the cooperative sector. With prices
given through EC regulations the competition between the different cooperative com-
panies has been directed toward gaining access to a maximum amount of raw mate-
rials (milk, pigs, or cattle). This competition has put strong upward pressure on input
prices, and the individual owners of the cooperatives have not had any strong incen-
tives to engage in costly and uncertain long-term projects. Actually the cooperative
movement has been so successful that it has eliminated most private Danish slaughter
firms and dairies. Also, the process of concentration of production in the cooperative
process industries has gone much further than in the privately owned sector. Now,
more than two-thirds of all milk is passing through one dairy company and at present
strong efforts are being made to join the last handful of cooperative slaughter firms into
a single one.
Among the private companies involved in food processing, we find some of the
technically most advanced Danish firms. These firms are often conglomerates, with
activities in related fields of chemical production or engineering. For example, the
sugar refining company developed advanced know-how in filter technology to support
its main line of business. This technology was later applied by affiliated engineering
firms, first to refine milk, and later to wastewater treatment, and was built into envi-
ronment technology. By international standards these firms are not big, however, and
they operate, typically, in specialized niches.
Until the beginning of the 1970s, two other development blocks played an impor-
tant role in stimulating the growth of the economy. One was oriented toward shipping
and ship-building and the other toward construction. With the international crisis in
shipping, following in the wake of the oil crisis, this development block became a struc-
tural problem in Denmark as in many other countries.
The same was true for the development block oriented toward domestic construc-
tion. This block had expanded very rapidly, especially in the 1960s, and it involved
not only construction as such but also the cement industry. Again we find a strong
Danish technological position in cement processing machinery and production sys-
tems (F.L. Schmidt). With the oil crisis a drastic reduction in the activities inside this
block took place.
Of course, not all parts of Danish production are integrated in such development
blocks. In the export-oriented industrial growth, in the 1960s, a number of technically
rather advanced firms in electronics expanded and became internationalized. Several
of them had developed their basic, and original, competence in an interaction with
domestic usersoften to be found inside the public sectorbut gradually they devel-
oped into enclaves.29 These firms are engaged in medical technologies, scientific instru-
ments, and measurement, for example.
Another, much larger group of Danish firms, in the engineering sector, produces
traditional machinery or components for engineering products for larger Danish or
foreign firms. The production takes place on the basis of skilled workmanship often
with the help of computer-controlled machinery. It is often very flexible in terms of
the products, while the connection to R&D and science is quite weak.
Development Blocks in the Swedish Economy

As in Denmark, it is possible to locate Swedish development blocks around the basic
export commodities (e.g., iron and steel and wood, pulp, and paper). In these areas
there has developed an infrastructure of smaller firms, service organizations, equip-
ment producers, and research organizations, according to Stenberg (1987). In partic-
ular, the relations between specific raw material sectors and suppliers of equipment
are sometimes quite important. Examples could be paper pulp producers and paper
pulp machine producers or mining and mining equipment producers. This argument
means, of course, clustering together specific parts of the engineering industry with its
customers, implying that the relations between producers and users of various prod-
ucts may constitute an important relation of interactive learning.
Another development block, discussed in Stenberg (1987), is one grouped around
production technology for the engineering sector. However, the engineering industry
as a whole would be a very large and heterogeneous block, and it would therefore be
useful to divide it into subsectors. Such an analysis might reflect the interaction
between (groups of) engineering firms and their suppliers and/or customers. Such a
division of the engineering industry into subblocks would be motivated also by the fact
that the suppliers and customers mentioned are also, often, firms in the engineering
industry.
Such an analysis would result in a number of development blocks centered
around the large Swedish engineering firms or parts of them. Examples could be an
electricity block (ABB Electricity, Vattenfall, etc.), an automobile block (Volvo and
SAAB automobile production and their subcontractors), a telecommunication block
(Ericsson, Televerket, etc.) and an aereoplane block (SAAB aeroplane production, its
suppliers and the Swedish defense procurement). The Swedish development blocks
might, because of the dominant role of a single or a few large firms, work differently
than the Danish ones.
The relative strength of these blocks can be indicated through figures on export
specialization. In Table 8.4 we have given these numbers for products based on domes-
tic natural resources (agricultural products, wood, pulp, paper, textiles and iron and
steel), oil and gas, chemicals, engineering, electronics, and transport equipment, and
other industrial products. Table 8.4 illustrates the strong position still held by the sec-
tors based on the domestic raw materials wood and iron. It also points to a relatively
strong position in engineering and to a very weak position in chemicals.30 Actually,
one might argue that engineering, historically, and in a complex manner, has grown
out of the raw material development block and, of course, especially the part oriented
toward iron and steel. This perspective is supported by the fact that Swedish compe-
tence and R&D activity in nonferrous metals are quite weak, while it is strong both in
engineering and in iron and steel (IVA, 1979a, p. 52).
The R&D System and Technology Policy

We will not try to give a detailed description of the institutions and their activities in
the field of R&D and public technology policy in Sweden and Denmark in this section.
For detailed presentations we refer to various national sources as well as to the OECD
Reviews of National Science and Technology Policies for Sweden (OECD, 1987) and
for Denmark (OECD, 1988a). Instead we will make an effort to bring forward the most
characteristic aspects in a comparative perspective.
The R&D Effort

In the introduction we noted that Sweden is a big spender on R&D (2.8% of GDP) and
that Denmark spends much less (1.3% of GDP).31 Business enterprise expenditure on
R&D was in 1985 1.05% of domestic product of industry in Denmark and 3.02% in
SwedenSweden having the highest figure of all OECD countries (OECD, 1988b,
Table 17). Sweden is like the large economic powers and Denmark is at a low level also
compared to her fellow small countries in this respect.
Another'difference is that in 1985 46.5% of all R&D was financed by public
sources in Denmark but only 34.0% in Sweden. Hence more than 65% of the R&D
was funded by the private sector in Sweden (OECD, 1988b Table 5). To the extent that
a high proportion of private funding of R&D is an indicator of technological "matu-
rity," Sweden is more advanced, in this respect, than Denmark. The Swedish firms
have transformed themselves to R&D-based activities to a larger extent than the Dan-
ish ones.32
It is also important who executes R&D. Roughly a quarter of all Swedish R&D
work is performed within the higher education system. The other three-quarters is per-
formed by industry, national authorities, public, private, and cooperative research
institutes, and independent consultants. However, the higher education system plays
a much more prominent role in the performance of research (R) than in development
(D). The basic and applied research (R) in industry is estimated at some 12% and devel-
opment (D) at 88%. About half of the R&D undertaken by public authorities, insti-
tutes, etc., is research (R) and half is development (D) (OECD, 1987, p. 41). Most of
the firm-based research (R) in Sweden is carried out by a handful of firmsincluding
the pharmaceutical companies, which dominate in this respect. Most of the other cor-
porations concentrate on more or less qualified development work (D).
The much more modest efforts in R&D in Denmark reflect several factors. To a
certain degree, the low rate of investment in R&D in Denmark reflects the fact that
small manufacturing firms tend to invest proportionally less in R&D than larger com-
panies. Hence the small size of firms and the limited role of big multinational firms
are important. Many of the small and medium sized Danish firms are, however,
involved in development activities. For example, two-thirds of a sample of small and
medium sized firms reported that they had introduced at least one new product in the
market 1984-1989. However, most of these activities are not separated clearly from
routine activities and do not get registered in R&D statistics.
In addition, the production structure of Danish manufacturing is a crucial deter-
minant behind the low R&D expenditures. Several of the strategic export sectors have
weak connections to a science base, and limited technological opportunities (e.g., in
terms of prospects of development of radically new products such as furniture and
food products). For example, in relation to value added the Danish food, drink, and
tobacco industry (ISIC 31) spent 0.6% on R&D in 1987. The corresponding figure for
Finland and Sweden was 1.6 and 1.7%, respectively (Nordisk FoU-statistik, 1987 og
1981-1987, 1990) Hence R&D expenditures in this sector are small in all countries.
In this sense, the low overall R&D intensity in Denmark is to a large extent a conse-
quence of the type of specialization of the economy." However, the factors mentioned
here cannot fully account for Denmark's limited R&D efforts. As we shall see later
there seems to be a different "mode of innovation" in Denmark, based less on for-
malized R&D efforts.
Technology Policy
Technology policy is often identified with public R&D policy. Here, the term will be
given a wider content. Technology policy includes public intervention in the process
of technological change as a whole. It embraces both technological development and
encouragement of the application of new technology regardless of its origin (i.e.,
absorption or diffusion). We will here briefly address several elements of technology
policy: state support to R&D, national technology programs, support of technology
diffusion, and public technology procurement.
Public support for technical R&D in Sweden is channeled through agencies such
as STU (The Swedish Board for Technical Development) and is allocated, mostly, in
response to initiatives from firms, institutions, and so on. Support measures consist of
advisory services, grants or loans, commissions or contracts, and fellowships or prizes
(for inventors). When the technology is developed, other agencies exist with the objec-
tive of providing support for later stages in the process of commercialization (OECD,
1987).34
During the 1980s a new element was added to the instruments of technology pol-
icy in Sweden. In late 1983 a decision was taken in Parliament to create a national
program on microelectronics components, financed by government and the partici-

pating firms jointly.35 It has later been followed by a similar program on information
technology.36 And in 1990 a decision was taken to launch a national program to give
state support to technology development in small and medium sized firms (mainly
subcontractors) in the engineering industry.
Danish technology policy reflects the dominance of small and medium-sized
enterprises in the economy. Public efforts to promote the application of new technol-
ogy in industry strongly emphasize two objectives (OECD, 1988a, p. 59). The first is
to ensure that enterprises have access to information and advice on the application of
new technology. The second objective is to ensure that enterprises can be supplied with
solutions in areas where they have neither the personnel nor the equipment to pursue
investigations themselves. Such areas may include analyses or testing, experiments or
development, investment planning involving the procurement, commissioning, and
operation of a new production plant, organizing quality and production control sys-
tems, and other specialized activities related to technology (OECD, 1988a, p. 24).
Generally the emphasis has been on technical service organized through a net of
public or semipublic technological institutes rather than on support schemes for R&D
inside firms. The ambitions regarding industrial policy have been shifting radically
during the past decades. Until the middle of the 1980s there were few attempts to inter-
vene besides the technical service net. In the middle of the 1980s began a short period
with a rather active policy effort.
In 1984 a technology policy initiative, with special reference to information tech-
nology, was initiated as the Technological Development Programme. This was a siz-
able national technology program under the Ministry of Industries, worth 1500 mil-
lion Danish crowns, which was to be used over 5 years (1984-1989). The main aim of
the program was to promote the use of internationally available information technol-
ogy in Danish industry as a whole, but it was also intended to stimulate independent
Danish development efforts in IT areas where there are special circumstances and
prospects (OECD, 1988a, p. 59).
This program wasofficially because it had already succeeded in its goalsinter-
rupted before 1989. Thereafter the level of ambition as regards an active technology
policy has been reduced. Except for limited programs for new materials and biotech-
nology, the government has moved to a policy where cost reductions, through tax
reductions and restrictive wage policy, are assumed to be sufficient to support indus-
trial development and international competitiveness.
Obviously, Danish technology policy is biased toward diffusion rather than sup-
porting the development of new knowledge and products. This is in sharp contrast to
the Swedish technology policy, which is very much a policy of R&D support to the
private sector. It seems as if the policy efforts in both countries tend to support the
already established mode of technological change rather than trying to correct its weak
sides. The interaction between state intervention and the market mechanism is
accordingly characterized by duplication rather than by complementarity.37
The technology policy initiatives mentioned above work mainly from the supply
side. There are also possibilities to intervene from the demand side.38 One interesting
way to influence technical change from the demand side is to use public demand as a
means. Public procurement can, of course, influence the choice between existing tech-
nologies as well as which new ones are to be developed. Important consequences for
technology development might emanate from public demand when no such inten-
tions lie behind the demand.
When analyzing the differences in procurement policies between Denmark and
Sweden it is important to take into account their respective roles in international pol-
itics. While Denmark, as a member of NATO, can be quite modest in terms of inde-
pendent military strength the Swedish armed neutrality is partly built on indigenous
military technology. The cooperation between the state and the big private firms in
ambitious projects, developing large and complex technological systems, was origi-
nally developed in areas of military importance. Typical examples were the different
generations of military aircraft developed by SAAB; but there were several other large
projects where the Ministry of Defense closely cooperated with the largest Swedish
firms (Annerstedt, 1976). Denmark has never had any large-scale high-technology
development programs in defense, space, or similar areas.
The Swedish nuclear power research program was originally oriented toward a
potential Swedish atomic bomb. But when ASEA-ATOM, the development enterprise
(which joined ASEA, Uddeholm, and the state) was established in 1969, the respon-
sibility was to develop civil nuclear power. Another development enterprise, LM-Tel,
joined the state (Swedish Telecom) with Ericsson in the development of telecommu-
nication technology. In the 1950s, we find other similar cooperative enterprises focus-
ing on, respectively, computer technology, and space technology. In the civilian
sphere, public procurement in Sweden has also led to important results in the electric-
ity field (ASEA-Vattenfall) and in railway equipment (ASEAThe State Railway
Company).
Such large technology projects have not developed in Denmark. Neither the mil-
itary demand, the firm structure, nor the strong noninterventionist ideology behind
industrial policy has allowed such activities. The Danish state has played a role in tech-
nology areas relevant to manufacturing but through more indirect mechanisms. The
most important function of the Danish state in this context has been, mostly uninten-
tionally, to support innovation through organizing and financing professional and
advanced demand for specific products. Among important examples of products, orig-
inally developed for public induced demand but later becoming hi-tech niche products
with big international market shares, we find electronic hearing aids, mobile com-
munication technology, medical instruments, and windmill technology.
This implies quite different kinds of relationships between the state and the pri-
vate sector. In Sweden, a few large firms (e.g., Volvo, ASEA, SAAB, Ericsson, and
Bofors) have been involved in quite intimate cooperation with state agencies. Such a
close cooperation has contributed to mutual understanding between the state and the
most important fraction of private capital in Sweden. In Denmark some small and
medium-sized niche firms have profited from the development of an advanced public
demand but most of the firms have not been involved in any formalized cooperation
with the state.
An important part of technology development in Sweden takes place in a few
large multinational corporations, strongly involved in big cooperative activities with
professional government agencies. The working of the Danish system is quite different.
The process of technical change in Denmark is organized neither by big firms nor by
the state. It is quite self-organized. The only reasonably strong coordinating block in
the economy has been the export-oriented, and cooperatively organized, agroindus-
trial sector. This sector has played a certain coordinating role as a demanding user of
technologies developed by small and medium-sized firms. Even some of the strategic
competencies in the small Danish high technology sector in engineering, pharmaceu-
ticals, and chemicals can be traced back to experiences and learning made in relation
to the agroindustrial complex.
The much weaker R&D effort in Denmark reflects the crucial role played by user-
producer interaction in product development and the correspondingly weaker links
between technical change and science. A recent study of 166 Danish product inno-
vations in manufacturing (Christensen and Valentin, 1989) shows that almost all of
the product innovations were based on knowledge already accumulated within the
firms rather than on external knowledge sources. Many innovations were stimulated
by users and they were also often involved during the process of development.
PERFORMANCE: THE DANISH AND SWEDISH SYSTEMS OF INNOVATION
In this section the performance of the two systems of innovation will be discussed.
Performance will be expressed, first, in terms of economic growth and other macro-
economic indicators. Second, the focus is on output indicators relating more directly
to the process of technical change and the national systems of innovation.
Macroeconomic Performance
Long-term economic growth rates in Denmark and Sweden were at internationally

high levels during the century up to the 1960s. Since the middle of the 1960s the growth
rates have become more moderate in relative terms, however. In both countries the
average for the 1970s and the 1980s is clearly below the OECD average (Table 8.5).
Since 1986 the Danish economy has been characterized by stagnating GNP and, pres-
ently, there are many signs pointing toward a period of zero growth also in the Swedish
economy. There is also a serious productivity problem in Sweden.
These similarities in long-term trends is somewhat misleading, however. In the
last decades the rythm of production and employment growth in manufacturing has
been quite different in the two countries (Table 8.6). Immediately after the first oil
crisis, the Danish industrial system reacted by dismissing many employees while the
firms in Sweden kept their workers, absorbing the crisis in a marked slowdown in pro-
ductivity growth. In the 1980s, the Danish economy had a brief period of strong
growth in production and employment (1984-1987) while productivity growth was
extremely slow. In this period employment in manufacturing started to fall more rap-
Table 8.5. GNP growth in Denmark, Sweden, and the OECD Average for 1960-1984
1960-1964 1965-1969 1970-1974 1975-1979 1980-1984 1985-1986 1987-1989"
Denmark 5.6 4.2 2.6 2.5 1.6 3.4 -0.4

Sweden 5.2 3.6 3.4 1.5 1.5 1.8 2.2
OECD 5.1 5.0 3.8 3.2 2.0 2.8 2.7
growth
Preliminary figures.
Sources: 1960-1984, Mjoset (1986, p. 329)andafter 1984 DanskOkonomi (1988).
Table 8.6. Annual Rate of Growth in GNP, Labor Productivity, and
Employment in Manufacturing 1973-1985 in Denmark and Sweden
1973-1979 1979-1985
y z n u y z n u
Denmark 1.6 3.6 -2.0 5 2.8 1.8 1.0 9
Sweden 0.5 1.0 -0.5 2 1.6 3.1 -1.5 3
Sources: production (y), productivity (z), and employment growth (n) from Gjerding et al. (1988, p. 52);
unemployment figures, averages for the periods (u) from Kongshoj Madsen (1989, p. 13).
idly in the Swedish economy. And this in spite of the two large devaluations in Sweden
1981(10%) and 1982 (16%).
Another important difference reflected in Table 8.6 is, of course, the rates of
unemployment. The Danish rate is now assumed to stabilize close to 10% while there
is an extremely high demand for labor in Sweden in 1990. One result of this constel-
lation is that some of the big Swedish firms now, actively, recruit skilled workers in
Denmark. However, there are signs indicating that the Swedish rate of unemployment
will increase in the near future.
The aggregate growth rates as indicators of performance imply that the prolonged
relative economic success of Denmark and Sweden might now be coming to an end.
The slowdown in growth to be observed since the last part of the 1960s has become
more serious in the last years. In the following sections we shall look at some perfor-
mance indicators more narrowly connected to technical innovation and technology
diffusion. In this context it will be argued that the slowdown in economic growth, at
least to a certain degree, reflects weaknesses in the national systems of technological
change. Specifically, we shall look first at the capacity of the two systems to contribute
to the global pool of generic technology. But, especially for small countries, the capa-
bility to absorb and diffuse new process technology and new product technology devel-
oped abroad is crucial.39
Development of Technologies
Technology development may involve a scale of technical advances from small incre-
mental innovations to radical innovations and crucial contributions to generic tech-
nology. Here the focus will be on radical innovations and generic technology rather
than on incremental change.40
Taking number of patents in the United States as an indicator of contributions
to the common pool of technical knowledge we found (in Table 8.1) the Swedish con-
tribution to be substantial (89 patents per million inhabitants) and the Danish to be
very modest (27 patents per million inhabitants).
A closer study, based on Delphi methods, analyzing the specific areas in which
Sweden has contributed to the global pool of technical knowledge found that the most
important were electrical power, iron ore, special steel, and ship building and that the
focus was almost always on developing the process technology (I VA, 1979b, p. 1431T).
In Denmark no such study has been made but it is well known that original con-
tributions have been made in filter technology, extending the use of ultrafiltration to
food processing and Pharmaceuticals (artificial insulin).
But, generally, Swedish and Danish contributions have been modest and most
generic technologies used in the two countries have been developed elsewhere. None
of them has, for example, made any substantial contributions to the new engineering
process technologies combining mechanical machines and electronics. In none of the
main "mechatronic" engineering automation techniquesComputer Numerical
Control-machinery (CNC), Industrial Robots, Computer Aided Design (CAD), and
Flexible Manufacturing Systems (FMS)do we find substantial Swedish or Danish
innovations. Denmark is not a major producer of any of them and a partly Swedish
firm (ABB) is active as a large producer only in one of the four areasindustrial
robots. It is especially interesting to note that the Swedish engineering industry, which
was characterized by a series of quite distinct and radical innovations almost a century
ago and still presents itself as a core sector in the Swedish economy, did not contribute,
more substantially, in this area (Edquist, 1989b).
The most important contributions from Denmark and Sweden to the global pool
of knowledge might, however, be socioorganizational rather than technical. The coop-
erative form of agroorganization in Denmark and the technoeconomic system for
food production based on it is an example with possible applications in many devel-
oping and eastern European countries. There is also a potential for diffusion of social
and organizational forms from Sweden to other OECD countries. Some of the most
visible and important experiments in this area have been developed by Volvo. Fifteen
years ago, a new teamwork organization was introduced at Volvo's factory in Kalmar.
And the new Volvo plant in Uddevalla represents another radical step in terms of work
design (Edquist and Glimell, 1989).
Diffusion of New Process Technologies
How rapidly can a national system of technological change absorb new process tech-
nologies developed abroad? This is an important question because rapid absorption
and domestic diffusion would, normally, result in a more productive and competitive
national economy. Therefore, the degree of technology diffusion is a very important
indicator of the performance of a national system of technological change. In this sec-
tion we shall use the densities of specific automation techniques (CNC machine tools,
industrial robots, CAD, and FMS) in the engineering industry as indicators for the
absorption capability regarding process technology. Density is measured as the num-
ber of units per million employees in the engineering sector.
According to Table 8.7, Sweden is characterized by the most intensive use of CNC
machine tools together with Japan. In industrial robots, it is number two after Japan,
while it has the most intensive use of CAD systems. When it comes to the most com-
plex and advanced technology included in the tableFMSSweden has a strong
lead, as compared to the other countries. Sweden and Japan are well ahead of the
United Kingdom, the United States, and West Germany as regards the degree of dif-
fusion of automation technology in the engineering industry. The numbers presented
in Table 8.7 refer to the end of December 1984. For industrial robots there are figures
for the same set of countries also for 1987. These figures show that Sweden remains
second after Japan, in terms of density, but they also indicate that the growth has been
more modest in Sweden than in the other countries from 1984 to 1987 (Edquist,
1989b,Table6).
Table 8.7. Diffusion of Flexible Automation Technologies in the Engineering
Industries of Six OECD Countries (Number of Systems and Density at the
End of 1984")
CNC Machine Tools Robots CAD FMS
Denmark
Stock 3,000 100 1,000 0
Density 19,230 641 6,410 0
Sweden
Stock 6,010 1,900 1,900 15
Density 22,177 7,011 7,011 55
United States
Stock 103,308 13,000 59,400 60
Density 11,728 1,475 6,743 7
Japan
Stock 118,157 64,657 7,300 100
Density 22,399 12,257 1,384 19
West Germany
Stock 46,435 6,600 11,000 25
Density 11,376 1,617 2,694 6
Great Britain
Stock 32,566 2,623 9,000 10
Density 10,505 846 2,903 3
"Density reflects the total number of systems (end of 1984) divided by million of employees (1980) in
the national engineering sector. For all countries except Denmark the source is Edquist and Jacobsson
(1988, p. 104). The number of systems in Denmark has been estimated by methods differing from the
ones used for the other countries (Kallehauge, 1989). This may imply problems of comparability. For
example, the definition of CNC machine tools is wider in the Danish casewhich means some overes-
timation of the Danish figure.
The figures demonstrate that Swedish firms are advanced users of automated
engineering techniquesand this in spite of the fact than none of the flexible auto-
mation technologies was originally developed in Sweden. Comparative case studies
confirm that Swedish firms have been more successful than British firms in introduc-
ing FMS systems, the success reflecting close interaction with producers, involvement
of workers, and a step-by-step strategy when upgrading the technology (Haywood and
Bessant, 1987).
The figures for Denmark have been gathered through many different sources and
are quite uncertain, as indicated in the note to Table 8.7. They point to a high density
for CNC and CAD but show extremely low values regarding industrial robots and
FMS in 1984. In the years after 1984-1986, the use of CNC and CAD expanded rad-
ically. This is true also for industrial robots. But in robots the density is much lower
than in Sweden and, still in 1990, there is no sophisticated FMS system in Denmark.
The Danish figures point to a late, but rapid, absorption of process technology and to
a specific weakness when it comes to the most advanced and complex new production
systems (as FMS). Recent studies point to great difficulties in Danish firms when intro-
ducing new process technology, the difficulties reflecting a lack of technical compe-
tence but, especially, problems with qualifications and work organization. Actually,
the introduction of new techniques seems, in a number of instances, to affect produc-
tivity negatively in the period 1984-1987 (Gjerding and Lundvall, 1990).
The general conclusion with regard to the technological level of process technol-
ogy (in the engineering industry) is that Sweden is very advanced and that Denmark
is weak in the more complex production systems.
Diffusion of New Product Technologies

An economy, or a firm, that continuously absorbs new process technology but neglects
adjusting its product program to the change in the market and demand structure will
sooner or later run into problems in terms of stagnation.41 An inability to develop new
products or absorb new product technologies developed abroad keeps the firm or
country outside the growth sectors and products.42
To analyze the capacity to absorb new and advanced product technology from
abroad we shall use the relative weight in exports of products with a high ratio between
R&D expenditures and salesso-called R&D-intensive products. We assume that the
relative weight of such products in exports reflects the capability to absorb new product
technology and to develop a production of new products and product families. One
reason for focusing on R&D-intensive products is that the demand grows considerably
more rapidly for these products than for other products (Table 8.10).
Table 8.8 shows that both Denmark and Sweden have a weak export specializa-
tion in R&D-intensive products. This is, however, true for almost all the small coun-
tries represented in the table. Hence, the problem is a small country problem to a cer-
tain degree. One of the two exceptions is Irelandnow a host country for
multinationals producing microelectronic products but performing only the low-end
activities in Ireland. It illustrates how just a few elements of the total production pro-
Table 8.8. Export Specialization Indices" for R&D-Intensive

Products6 in 16 OECD Countries 1961-1987
1961-1965 1973-1979 1983-1987
Denmark 0.4 0.6 0.6

Sweden 0.7 0.9 0.8
United States 1.2 1.4 1.6
Japan 0.9 1.1 1.2
Great Britain 1.4 1.2 1.1
West Germany 1.4 1.1 1.0
Switzerland 1.6 1.8 1.4
Ireland 0.1 0.7 1.2
Netherlands 0.7 0.9 0.8
France 0.9 0.9 1.0
Belgium 0.7 0.6 0.6
Canada 0.3 0.5 0.5
Italy 0.9 0.7 0.7
Finland 0.1 0.3 0.4
Norway 0.3 0.4 0.4
Austria 0.5 0.7 0.9
"For definition of export specialization index, see note a in Table 8.3.
''The R&D-intensive products are inorganic chemicals, Pharmaceuticals, plastic
materials, power generating machinery, computers and peripherals, semiconduc-
tors, telecommunications equipment, machinery for production and distribution
of electricity, scientific instruments, photographic supplies, watches, and clocks,
and aircraft. The SITC numbers of the product groups are given in Dalum (1989,
Appendix).
Source: The IKE databank and Dalum (1989).
Table 8.9. Share of Exports of Engineering Products Emanating from
Industries with a High R&D Intensity" (%)
1971 1975 1980 1985 1986
Sweden 23.7 23.8 25.3 29.7 28.1
United States 42.4 37.5 44.9 52.5 55.6
Japan 35.2 29.4 35.9 41.8 42.5
West Germany 24.2 24.4 27.6 30.4 29.3
France 28.7 27.2 30.0 40.1 38.4
United Kingdom 29.7 32.9 39.9 50.9 51.9
Switzerland 45.0 42.5 44.4 45.1 45.2
Netherlands 50.0 44.9 48.6 52.4 51.8
Finland 19.1 18.5 26.1 25.1 27.3
Norway 19.3 13.2 22.0 22.1 23.4
Denmark 27.5 24.4 31.2 35.6 36.3
OECD 31.2 29.5 34.1 39.3 39.1
"Industries with a high R&D intensity are ISIC (Rev 2): 3825 (Office machinery and Computers), 383
(Electrical Machinery), 3845 (Aerospace), and 385 (Scientific Instruments).
Source: OECD Industrial Outlook Data Base (January 1989) (COMTAP) and Edquist (1989b, Table 7).
cess may be transferred. The other more genuine exception is Switzerland, successful
both in pharmaceutical products and in microelectronic based engineering products.
Large countries such as the United States, Japan, the United Kingdom, West Germany
and France are relatively specialized in their exports of R&D-intensive products.
Given its importance it is relevant to take a closer look at the engineering sector
and its specialization in terms of R&D-intensive products. Table 8.9 demonstrates
that the proportion of engineering products exported having a high R&D intensity is
low in Sweden (as in West Germany, Finland, and Norway). Denmark has a stronger
position in this respect, but is far behind countries such as Japan, the United States,
Netherlands, and Switzerland. Accordingly, two of the small countries perform quite
well on this indicator.43
We find thaton the wholeboth Denmark and Sweden have a relatively weak
position in R&D-intensive products. Taking into account that the world markets for
these products grow considerably more rapidly than for other products (Table 8.10)
this constitutes a long-term structural problem for both Denmark and Sweden and
may, partially, explain why the growth rates in these countries are becoming low, as
compared to the OECD average.
It is not too surprising that Denmark, given its weak R&D effort, has a weak posi-
tion in R&D-intensive goods. But it is remarkable that Sweden, in spite of its very
substantial investments in R&D, its high number of patents per million inhabitants in
the United States, and its strong multinationals in engineering, has been so slow in
absorbing R&D-intensive products.
PROBLEMS: STRUCTURAL AND INSTITUTIONAL STRAINS ON THE TWO

SYSTEMS OF INNOVATION
In the previous section we pointed to one structural problem common to the two sys-
tems. This problem refers to the product composition of production and exports. The
Table 8.10. Differences in Annual Growth Rates (%) in the

Volume of the OECD Market, in Current Prices
Annual Growth Rates R&D-Intensive
in the OECD Imports Products" Other Products
1961-1973 19.1 15.0
1973-1983 14.0 9.4
"For specification of R&D-intensive products, see note b in Table 8.8.
Source: Dalum (1989).
internationally most rapidly growing product groups are weakly represented in both
economies. Not even Sweden, in spite of its very major efforts in R&D, has succeeded
in overcoming this structural problem. But, of course, the problem takes on different
forms in the two countries and we shall discuss some of these specificities separately.
The other family of problems refers to the institutional framework of the econ-
omy. Again, we find one common problem, reflecting changes in how the relationships
between the national and the global economy are regulated. And again, the problem
takes on quite different forms in the two economies. We shall end this chapter with a
discussion of the character of this institutional problem in the Danish and the Swedish
systems of innovation, respectively.44
Structural Problems in the Danish System
We want to point to two fundamental structural problems in the Danish economy.

The first problem is that demand for the end products of the traditionally strong, devel-
opment blocks around construction, shipping, and agriculture is falling or stagnating.
The new hi-tech product groups in the field of electronics are still small and they have
not established anything like the coherence characteristic of the old development
blocks. On the contrary, after a period of innovative activities, strongly rooted in the
domestic economy, they tend to establish themselves as enclaves, importing at least
90% of their components and exporting at least 90% of their final output.
The agroindustrial and food-industrial sectors are still extremely important for
the health of the Danish economy and, at present (1990), a series of mergers are taking
place both inside the cooperative part of the economy and in the private sector. It is
an open question if this kind of regrouping will give the necessary impetus to a more
offensive strategy in terms of innovative behavior. Actually, the cooperative block
tends, increasingly, to play a conservative role in the Danish economy and, as long as
the cooperative firms can secure acceptable prices and incomes for their owners, the
incentives to develop a new strategy are quite weak. For example, we have seen that,
relatively speaking, Denmark spends less on R&D in the food industry (in relation to
value added) than Finland and Sweden. There might be a potential in biotechnological
applications in these product areas and there is, as a matter of fact, a quite strong bio-
technological competence in some Danish firms. (NOVO is the best-known example.)
But this competence has been oriented mostly toward the Pharmaceuticals applica-
tions and it might take some strong policy and other efforts to establish the necessary
coupling and interaction between the agroindustrial complex and the biotechnology
firms.
Table 8.11. The Distribution of Europe's 500 Largest Manufacturing Firms

Among a Number of Small Countries
Number of Firms Number of Firms/Million Inhabitants
Denmark 2 0.4
Sweden 31 3.7
Norway 4 1.0
Finland 6 1.2
Austria 8 1.1
Belgium 12 1.2
Switzerland 22 3.4
Netherlands 29 2.0
Source: Forum for industriel udvikling (1988, Figure D-5).
The second structural problem is illustrated in Table 8.11. It refers to the fact that
the Danish economy is extreme in its dependence on small and medium-sized firms.
The two big Danish firms referred to in the table are to be found in, respectively, the
dairy industry (Danish Dairy, a cooperative firm) and breweries (Carlsberg, a private
firm). Outside the food industry, there are few units large enough to organize scale-
intensive R&D activities. This problem is reinforced by the institutional setup of the
financial sector, to be discussed in connection with the institutional problems.
It is remarkable that a small country so weak in terms of most traditional tech-
nology performance indicators (R&D intensity, process diffusion and product com-
position) has been able, for so long, to sustain a high level of income per capita. One
interpretation might be that these indicators do not catch very well all forms of incre-
mental innovation important to competitiveness. Another, less optimistic, interpre-
tation is that the present stagnation signals the end of success of this model, based on
modest incremental indigenous technical change.
Structural Problems in the Swedish System
Also in the case of Sweden, we shall address two structural problems, both of which
point toward vulnerability and a risk for secular stagnation. The first relates to the pro-
spective demand for its raw materials and its semimanufactured goods. The exports
of wood products, pulp, paper, iron ore, steel, and so on is still very important to the
dynamics of the economy. The rapid development in biotechnology and material
technology might, however, in the long run result in new materials. The demand for
some of the classical Swedish export products might dwindle, and the result might be
stagnation and a need for a drastic transformation of the economy. We have pointed
to the fact that the engineering industry focuses on the production and export of tra-
ditional, mechanical, low-R&D products with a limited growth prospect.
We have seen that the natural resource bases of Sweden and Denmark have
shaped the two national systems of innovation to a very important degree.45 However,
Sweden also developed a strong mechanical engineering industry in the early parts of
this century. It would, of course, be interesting to find out in detail why this happened
in Sweden and not in Denmarkor Canada and Australia. At the same time the
Swedish engineering industry is still mainly mechanical. It has not managed to trans-
form itself to also produce electronics products. In this sense Sweden is locked into the
mechanical trajectory in the engineering industry and still quite tied to the raw mate-
rial base in other sectors of industry. Again, it would be very interesting to analyze, in
detail, why Sweden did not manage to develop an IT capability (except in the case of
Ericsson), while firms in other countries succeeded. In other words, why is Sweden
(with a few exceptions) not producing in the new industries, which are not based on
raw materials'?4''
Another kind of vulnerability emanates from the concentration in the core sector
in the economy engineering. The five largest engineering firms (Volvo, Electrolux,
Ericsson, ASEA, and SKF) produced around one-fifth of the total value added in man-
ufacturing in 1982 (Economic Growth, 1984, p. 193) and in 1976 they organized more
than a fourth of total exports (IVA, 1979c, p. 82). The concentration of economic
activities is even more extreme than these figures indicate. The large firms have built
networks of domestic suppliers directly dependent on the performance of the large
firms. And everything points toward increasing ratios of concentration since these data
were produced. If these large firms do not succeed in developing efficient production
in the new high-growth product areas, the problems of stagnation will become very
serious for the Swedish economy. And if one or two of these large firms run into prob-
lems, it can have devastating consequences for the whole economy.47
A comparative advantage of the Swedish system has been its combination of
advanced production methods in traditional product areas on the one hand, and cor-
poratist institutional setups, making it easy to develop and introduce advanced process
technology, on the other hand. Paradoxically, this comparative advantage might be an
important factor in explaining the rigidities in the product structure. The whole insti-
tutional setup, the whole innovation system (including state policy), seems to have
become geared toward movement along one trajectory. The ease with which advances
have been made in this direction may have weakened the incentives to take on the
more difficult task of developing the technological infrastructural basis necessary for
radically new product areas.48
The average low-R&D character of Swedish production is a severe problem for
the Swedish system of technological change. And this problem is certainly not solved
spontaneously by the market. Therefore there are reasons to consider whether state
intervention could mitigate the problem. A recent evaluation (Edquist, 1989b) of the
activitites of The Swedish Board for Technical Change (STU) in the field of engineer-
ing technology showed that STU, in the 1980s, spent most of its resources in this field
on support to process technology and a smallerand decreasingshare on product
technology. Hence, STU supports an area where Sweden performs extremely well and
neglects an area that is a major problem for the Swedish system of innovation. Accord-
ingly, a reallocation of the efforts of STU may be called for. The currently launched
national program to give state support to technology in the engineering industry was
partly based on the evaluation. In the new program, a more balanced allocation of the
state resources between production technologies and product technologies is pre-
scribed (Regeringens Proposition, 1990, p. 107).49
Institutional Problems in the Two Systems of Innovation

One fundamental characteristic of the Danish system of innovation is the absence of
strong coordinating agencies. Neither the state nor concentrated capital has been pow-
erful enough to take the lead in the process of restructuring the economy. The single
strongest organized power has been the cooperative farmer's movement and, for insti-
tutional reasons, this movement has become quite myopic in relation to strategies of
innovation. In the earlier period of rapid growth in demand, this absence of a coor-
dinating force was not a serious problem. Today, with the increasingly systemic char-
acter of technology, the vulnerability of the model becomes more apparent. It is illus-
trated by the fact that foreign capital now, rapidly, is taking over many of the Danish
niche firms characterized by advanced product technology, in, for example, electron-
ics (Aaen, 1986).
Lately, there has been a growing awareness of this vacuum that leaves structural
change to market forces and initiatives joining private firms, banks, and pension funds
have resulted in a series of mergers, especially in the food industry. At the same time,
the trade unions have pointed to wage earner's funds as potential candidates for orga-
nizing a more coordinated industrial restructuring. It is an open question if these dif-
ferent efforts will succeed in defending a reasonable degree of coherence in the Danish
production and innovations systems. The prospective project of a single European
market is regarded by many policy agents as the ultimate response to the structural
and institutional problems of Denmark.
In Sweden there is a corresponding growing awareness of a weakening of the foun-
dation for the Swedish model. "The Swedish model" canas we previously saw
mean different things. If it implies the mode of wage formation developing after World
War II, it is clear that it has partly disintegrated during the past 10-15 years. The dif-
ference between nominal wage increases and productivity increases has, for a consid-
erable period, been much larger than in competing countries, which has led to a rela-
tively high inflation in Swedenand several devaluations around 1980. One reason
for the high wage increases has been the shortage of labor.50
Shortage of labor and large wage increases create incentives for rationalization of
production and the introduction of labor-saving process technologies (e.g., automa-
tion). And there is widespread acceptance of this in the union movement.51 If "the
Swedish model" is given the other meaning (i.e., a consensus on issues of rationaliza-
tion and technical change between capital, trade unions, and the state), the model is
still alive.
One fundamental problem, in this context, is the fact that the home-based mul-
tinational firms are now becoming, increasingly, footloose. ASEA's merger with
Brown-Bo very from Switzerland into ABB in 1989 has now been followed up by merg-
ers involving, for example, SAAB (with General Motors), Volvo (with Renault), and
several others.52
In addition to the fact that the nationality of several of the largest "Swedish" firms
can be questioned, the process of internationalization is also proceeding in other
spheres. For example, a very important change in the 1980s was the development of a
very free flow of capital across borders. This means that the possibility of pursuing an
independent monetary policy in a country such as Sweden is severely restricted.53 Fur-
ther, the Swedish flow of direct investments abroad is enormous. It has increased from
5 to 55 billion Swedish crowns per year between 1980 and 1989. During the same
period the foreign direct investments in Sweden remained around 5 billion per year.54
The trend toward internationalization has definitely weakened the structural
position of the labor unionswhich are still quite nationalin Sweden.55 But it has
also weakened the position of the state. And capital has strengthened its positionin
particular the internationalized capital. Accordingly, the power relations between the
three main parties of the Swedish model have been fundamentally changed.
In other words, the process of internationalization changes, drastically, the basic
workings of the Swedish model. The old model was based on the fact that the Social-
Democratic state and the large firms were mutually dependent. If this dependency goes
only one way in the future, the basis for corporatism will erode. This problem is
referred to in a report to the 1989 congress of the Swedish Metal Workers Union:
The problem today is that these multinationals do not any longer need Sweden,
but the Swedish economy needs them. The "third way politics" was the last
major contribution that the Swedish society could make to big business, and still
expecting something in return. (Solidarisk arbetspolitik, 1989, p. 93)56
This brings us to an end, pointing to the possibility that the two small national
systems of innovation studied here, Denmark and Sweden, are now in a critical phase.
NOTES
1. This chapter has been written within the Columbia University project on National Sys-
tems Supporting Technical Advance in Industry. We have greatly profited from discussions at
the conferences organized within this project and in particular from Richard Nelson's com-
ments. Swedish and Danish colleagues have also commented on earlier drafts. We have received
valuable comments from Esben Sloth Andersen, Boel Berner, Ted Bradshaw, Martin Borjeson,
Tarja Cronberg, Erik Dahmen, Bent Dalum, Mark Elam, Hans Glimell, Lars Herlitz, Mikael
Hard, Bjorn Johnson, Maureen McKelvey, Lennart Stenberg, Elisabeth Sundin, Lars Svensson,
and Hans Vallentin. The paper was written in 1990 and completed in January 1991. Accordingly
it does not cover events and changes occurring thereafter (see also note 56).
2. Comparing countries with many features in common allows us to keep some variables
constant. This might be effective when it comes to establishing the most crucial elements and
relationships in national systems of technical change. At the same time, this chapter brings into
focus some of the specific problems facing small national systems of innovation.
3. The comparison presented here confirms that it still is useful to assume that national
borders matter when it comes to the workings of the process of technical change. If two distinct
and different systems of innovation can be isolated in two countries, so close in terms of culture
and space, this should be true, a forteriori, for countries more distant in these respects.
4. This issue will be discussed in the final section on structural and institutional problems.
5. Hence the conduct of Sweden looks very much like that of the large economic powers
in this respect, while the conduct of Denmark is similar to other small OECD countries. The
figures presented in this section also show that a country can be quite successful in terms of
growth and have a high GNP/capita in spite of low R&D expenditures and a poor patenting
record.
6. Alternative terms to denote the same thing are, for example, "national systems sup-
porting technical advances" and "national systems of technological change."
7. One wayamong othersto try to determine the limits of a national system of inno-
vation would be to try to causally explain the invention, innovation, and diffusion of technolo-
gies and let the explanatory factors define the limits of the system. An attempt in this direction
but only with regard to diffusion of process technologies in the engineering industrywas made
in Edquist(1989a, pp. 10-11).
8. This theoretical perspective has been developed in Andersen and Lundvall (1988), in
Johnson and Lundvall (1988), and in Lundvall (1992).
9. One reason why it is useful to distinguish between development and diffusion of tech-
nologies is that the determinants differ between these stages of technological change. The ana-
lytical distinctions discussed here are dealt with in more depth in Edquist (1989b, 1991, 1992).
10. For a product technology an innovation is the first economically successful production
of a product. A failed innovation is when a product has been marketed and disappeared from
the marketand thereby from production. For a production technology an innovation means
that new means of production are efficiently used for the first time in productionwhich, of
course, presupposes that they have been produced earlier.
11. For process technologies diffusion means, of course, that additional producers intro-
duce the technology in their process of production. For product technologies diffusion means
that additional producers begin to produce the new product. (Hence, we do not mean diffusion
to users or consumers when we talk about diffusion of product technologies.) Failed diffusion of
a product technology might reflect a technological monopoly based on patents, secrecy, or other
barriers to entry.
12. In small economies, the development of generic technologies and radical innovations
is currently a rare phenomenon, and the most important activity in the national system of tech-
nological change is often diffusion, mainly the absorption and adaptation of technologies devel-
oped abroad.
13. For example, an industrial robot is a product technology as output and a production
technology when it is used in the process of production.
14. The role of education is, for example, dealt with in Ahlstrom (1982) and Berner (1991).
15. There are few systematic attempts to write the history of the Danish system of inno-
vation. In this part we have been inspired by several not very widely circulated contributions by
our colleague Esben Sloth Andersen (1973, 1978a,b).
16. In Denmark the innovation was made in 1878 by Burmeister and Wain, but the sep-
arator was invented by N.J. Fjordone of the first professors in dairying at the Danish Agricul-
tural College, established in 1874. In Sweden the separator was invented by Gustav de Laval.
Later it became one of the core products for AB Separator, a firm that later developed into the
multinational firm Alfa-Laval.
17. The transformation of the Danish economy in the last quarter of the nineteenth cen-
tury is an interesting illustration of the importance of social mobilization and social innovation.
Actually, Grundtvig might be regarded as the most important single Danish innovator, as one
who has put his marks on the modern Danish system of production and innovation. His suc-
cessful ideological mobilization of the Danish peasants toward self-reliance, for example,
through organizing local education in the popular "hojskoler," prepared the ground for the local
organizational undertakings that gradually developed into a coordinated national movement
toward farmers' cooperatives.
18. This means, for example, automation and reductions of direct labor costs in dairies and
slaughterhouses.
19. Some data on this are presented in the section on the R&D effort.
20. In addition to horses used in mines many thousands were used by transport companies
in London and other cities.
21. The large Swedish export of iron ore developed later.
22. Foreign demand for wood products and iron was important in the Swedish process of
industrialization (Schon, 1985, p. 35).
23. This includes production of ships (until the 1980s), automobiles, mining equipment,
and aeroplanes, except for the product groups already mentioned.
24. In his analysis of small country corporatism Katzenstein (1985) characterizes Denmark
as having a weak partner at the capitalist side and a strong one at the labor side. We would rather
294 SMALLER HIGH-SNCOME COUNTRIES
point to the fact that in Denmark none of the parties is as centralized, homogeneous, and united
as their Swedish counterparts.
25. However, unemployment in Sweden is in the process of increasing in 1990.
26. If this design turns out to be unprofitable compared to automobile assembly according
to traditional principles in, for example, West Germany and Japan, the conclusion may be that
car assembly cannot be carried out in Swedenfor structural reasons. On the other hand, if it
turns out to be profitable, the Uddevalla example may show the way for the general development
of assembly workwhich is one-fourth of all work in the engineering industryin the future
(Edquist and Glimell, 1989, p. 44). Another possibility is, of course, that the structural condi-
tions changethrough increased unemployment, etc.and traditional assembly becomes via-
ble again in Sweden. The Uddevalla factory has been addressed in more detail in Berggren (1990)
and in Clarke (1989).
27. A "development block" has also been defined as a broad set of interconnected produc-
ers and users of products, developing in close interaction with each other and often supported
by knowledge-producing private, or public, organizations (Stenberg, 1987, p. 46).
28. In its most complete form, a development block would include whole chains of pro-
duction, final users of the products, producers of means of production, as well as independent
knowledge resources. Only the United States and Japan would today contain complete
development blocks in this sense within important future areas (Stenberg, 1987, pp. 46-47).
Dahmen uses "development blocks" in both an ex ante and an ex post sense (Dahmen, 1988,
pp. 6-7).
29. Their export and import ratios lie typically around or over 90%.
30. If sales from the Swedish-owned engineering companies abroad had been included the
figure for this industry would have been much higher.
31. In relation to value added, the figures were 7.6% (Sweden) and 2.9% (Denmark). And
in relation to fixed investments the percentages were 57 and 26%, respectively (Edquist, 1989b,
Table 11).
32. However, the R&D expenditures (in Sweden) of Swedish firms decreased for the first
time between 1987 and 1989 by 7% in real terms.
33. This particular example also indicates, however, that Denmark has a lower R&D inten-
sity than other Nordic countries in one of her most important export sectors.
34. STU has an annual budget of about 1 billion Swedish Crowns and is the main tech-
nology policy agency in Sweden.
35. The emergence of the program has been analyzed in Glimell (1988, 1989). It has been
evaluated ex post in Wennerberg(1989).
36. The first phase of the IT program has been evaluated in Arnold and Guy (1989).
37. The mechanisms behind this process are dealt with in Edquist (1989b, 1990).
38. In Edquist (1989c, 1990) supply-side as well as demand-side technology policy instru-
ments relevant for technology development as well as for technology diffusion are discussed in
more detail.
39. In this sense, the "system of technology diffusion"including absorbtion, assimila-
tion, and incremental changeis much more important than the R&D system for small coun-
tries. There are, therefore, strong reasons for spending as much energy on analyzing the mech-
anisms, institutions, and policies for acquiring foreign technology and disseminating it
domestically as on the R&D systemalthough this is seldomly done. (To do so would be a major
task, since we do not know the detailed features of the systems of technology diffusion in Den-
mark and Sweden.) One reason for this neglect may be that the diffusion system is less often the
object of public policy intervention. Diffusion is normally left to be handled spontaneously by
the market forces (although Denmark has been an exception in this respect). As we will see this
may also be true when the diffusion system does not function satisfactorilyas in the case of the
the diffusion of new product technology in Denmark and Sweden.
40. Incremental innovations are, in this chapter, considered to be part of the process of
technology diffusion.
41. Such a combination of labor-saving technical change with lack of presence in growing
markets may lead to "technological unemployment." This is the basic idea behind the Pasinetti
(1981) analysis of economic growth and structural change. The argument has been adjusted and
applied to small open economies in Lundvall (1987).
42. When a new product is developed in one country its diffusion to another country is
often presented in the framework of product life cycle theory. However, the product life cycle
model tends to understate the difficulties involved. The reason is that it may be quite complex
to move complete production processes between countries. The knowledge base may involve
tacit elements that are difficult to transfer or it may involve extremely complex combinations of
scientific knowledge. Infrastructure and institutions in the receiving country may also not be
adapted to the product family.
43. The indicators used in Tables 8.8 and 8.9 are somewhat blunt because of the rather
high level of aggregation. This is because of lack of more detailed R&D data. The analysis of the
diffusion of R&D-intensive product technologies in Sweden has later been developed in Edquist
and McKelvey( 1991).
44. In the section on problems we will address structural and institutional problems with-
out distinguishing between those problems that can be solved through (further) political inter-
vention in the systems of innovation and those that cannot. Such a distinction would be neces-
sary if the policy implications of the present analysis were to be explicitly and systematically
drawn out.
45. In a country such as Japan the resource base may have been less important, and civilian
technology policy a more important determinant of the direction of technological change. In the
United States the government-financed military R&D and procurement may have been the most
important factors.
46. There are many challenging questions in this context! How do national systems of
innovation use their resource base without being constrained by it (e.g., by neglecting new prod-
uct areas) in later stages of development? Except for liberating the system from the domestic
natural resource base (like the Japanese and Koreans), can a country also "liberate" its system
of innovation from socioeconomic and cultural conditions? How do transfers between techno-
logical trajectories and paradigms actually take place?
47. As a matter of fact, both the automobile producers (SAAB and Volvo) seem to start
trembling in mid-1990, having announced drastic profit reductions and considerable layoffs.
48. The large devaluations around 1980 have most probably contributed to the conserva-
tion of the product structureby making traditional products more profitable and leaving little
room for offensive ventures (in an economy of full capacity utilization).
49. In the process of implementation of this program during 1990, the resources have, to
a large extent, been allocated to subcontractors within the automobile industry. This is probably
related to the emerging problems for the Swedish automobile industryand looks more like
crisis management than an offensive strategy of the future.
50. However, the sum of inflation and rate of unemployment is still in 1990 relatively
lowwhich means that the Swedish model, in this sense, has not broken down completely.
51. As we have shown earlier Sweden is good at the introduction of advanced process tech-
nology but bad at increasing productivitywhich is not necessarily contradictory.
52. If only one or two of the very large and internationalized "Swedish" firms decide to
move their R&D abroad the ratio between R&D expenditures and GDP would dwindle. How-
ever, this has not happened so far.
53. Previously the Swedish Central Bank could determine the domestic interest rate and
adjust it to the business cycle. Now the interest rates in the surrounding world determine the
Swedish one. Hence, the interest rate is no longer controlled from the political level in Sweden
but by the global market. The same seems to be becoming increasingly true for taxationat least
taxation of incomes emanating from capital.
54. And during the first half of 1990 Swedish corporations bought firms in Western Europe
for a total of 72.2 billion SEK. This is to be compared to the acquisitions of U. S. firm of 32.7
billion and of Japanese of 10.5 billion. This exceptional record may, of course, have been related
to the insecurity among Swedish firms concerning the future relations between Sweden and the
European Community. However, during the autumn of 1990 the Swedish government declared
its intention to apply for membership in the EC. A large majority in Parliament supports this.
55. This structural weakening concerns issues such as industrial democracy and codeter-
mination. In the field of wage formation it is balanced by the shortage of laboras long as that
persists. A labor shortage may lead to problems of competitiveness (i.e., full employment con-
tains the seeds of its own destruction). On the other hand, unions tend to accept labor-saving
technical change if full employment is the case.
56. As mentioned in note 1, this chapter was written in 1990 and completed in January
1991. Hence the analysis does not address changes occurring thereafter. However, we want to
mention a few events that have since occurred in Sweden. In July 1991 the Social Democratic
government submitted Sweden's application for membership to the EC. In the election of Sep-
tember 1991 the Social Democrats were replaced by a center-right coalition government. Fur-
ther, the rate of unemployment increased to about 4% in 1991. These changes are not unrelated
to each other. Together they seem to strengthen the observations expressed at the end of this
chapter (e.g., on the gradual disintegration of "the Swedish model," on internationalization, and
on the changes in the power balance among the labor unions, the state, and private capital).
REFERENCES Berner, B. (1991). "Professional or Wage Worker?

Engineers and Economic Transformation in Swe-
den." In C. Smith and P. Meiksins (eds.), Engi-
Aaen, F. (1986). "Hvor dansk er dansk elektronikin- neering Class Politics. Verso.
dustri?" Aalborg: Aalborg Universitetsforlag. Christensen, J. F., and Valentin, F. (1989). "Produk-
Ahlstrom, G. (1982). Engineers and Industrial tinnovationer i dansk industri." Arbejdsnote 4,
Growth. London: Croom Helm. Institute for industrial research and social devel-
Andersen, E. S. (1973). "Noter om industri, stat og opment, Copenhagen school of economics and
teknisknaturvidenskabelig forskning." Koben- social science.
havn: Danske Studerendes Fsllersrad. Clarke, T. (1989). "Imaginative Flexibility in Pro-
Andersen, E. S. (1978a). "Eksport, teknologi og duction Engineering: The Volvo Uddevalla
forskning." Aalborg: Institut for Produktion, Plant." Paper presented at the conference "A
mimeo. Flexible Future," Cardiff" Business School, Sep-
Andersen, E. S. (1978b). "Eksportudvikling, tember 19-20.
erhvervsstruktur og forskningspolitik." Aalborg: Dahmen, E. (1970). Entrepreneurial Activity and the
Aalborg Universitetsforlag. Development of Swedish Industry 1919-1959.
Andersen, E. S., and Lundvall B. -A. (1988). "Small Homewood, IL: Published for the American Eco-
National Systems of Innovation Facing Techno- nomic Association by R. D. Irwin.
logical RevolutionsAn Analytical Frame- Dahmen, E. (1988). '"Development Blocks' in
work." In C. Freeman and B. -A. Lundvall (eds.), Industrial Economics." Scandinavian Economic
Small Countries Facing the Technological Revo- History Review 1:3-14. Also published in Carls-
lution. London: Pinter Publishers. son, B. (ed.) (1987). Industrial Dynamics. Boston:
Annerstedt, J. (1976). "Niigra drag i den svenska sta- Kluwer Academic.
tliga forskningspolitikens framvaxt." In V. Ander- Dalum, B. (1989). "Export Specialization, Compet-
sen and R. Kalleberg (eds.), Forskningenstaten itiveness and National Systems of Innovation."
og kapilalen. Kobenhavn: Nordisk Sommeruniv- Aalborg: Institut for Produktion, mimeo.
ersitet. Dalum, B., Fagerberg, J., and Jorgensen, U. (1988).
Arnold, E., and Guy, K. (1989). "Evaluation of the "Small Open Economies in the World Market for
IT4 Programme. Phaze I." Stockholm. Electronics: The Case of the Nordic Countries."
Berggren, C. (1990). Del nya bilarbetetKonkur- In C. Freeman and B-A Lundvall (eds.), Small
rensen mellan olika produktionskoncept i svensk Countries Facing the Technological Revolution.
bilindustri 1970-1990. Lund: Arkiv. London: Pinter publishers.
Dansk 0konomi. (1988). December, Kobenhavn, Institute of Production, Aalborg University Cen-
Del Okonomiske Rad. ter.
Edquist, C. (1989a). "The Realm of Freedom in Gjerding, A. N., Johnson, B., Kallehauge, L. E.,
Modern Times? New Technology in Theory and Lundvall, B. -A., and Madsen, P. T. (1988). Pro-
Practice." Inaugural Lecture, Tema T Report 18, duktivitet og international konkurrenceevne. Aal-
1989, Department of Technology and Social borg: Aalborg Universitetsforlag.
Change, University of Linkoping, Sweden. Gjerding, A. N., Johnson, B., Kallehauge, L. E.,
Edquist, C. (1989b). "Utvardering av statligt stod till Lundvall, B. -A., and Madsen, P. T. (1990). Den
verkstadsindustriell teknik." Linkoping, Tema Forsvundne Produktivitet. Kobenhavn: PJOF:s
Teknik och Social Forandring, mimeo (also pub- forlag.
lished as Tema T Report 22, 1991). Glimell, H. (cd.) (1988). Industrifdrnyelse i Norden.
Edquist, C. (1989c). "Teknisk forandring och tek- 80-talets programsatsningar pa mikroelektronik.
nikpolitik." In 'Energin, makten och framtiden Roskilde: Forlaget Samfundsokonomi og Plan-
samhallsvetenskapliga perspektiv pa teknisk for- laggning.
andring." Statens energiverk 1989-.R16, Stock- Glimell, H. (1989). "Atererovra datapolitiken. En
holm. rapport om staten och informationsteknologi
Edquist, C. (1990). "Technology PolicySocial, under fyra decennier." Tema T Report 20, 1989,
Economic and Political Aspects." Tema T Work- Department of Technology and Social Change,
ing Paper 75, Department of Technology and University of Linkoping, Sweden.
Social Change, University of Linkoping, Sweden. Haywood, B., and Bessant, J. (1987). "FMS in Swe-
(This report will also be published in a book enti- den and the UK." Brighton: Innovation Research
tled Technology PolicyTowards an Integration Group, Brighton Polytechnic.
of Social and Ecological Concerns, Institut fur Industriokonomisk Institut. (1986). "Industri i Nor-
Hohere Studien, Wien, 1992.) den: Konkurranse og samarbeid." Stavanger:
Edquist, C. (1992), "Technologocial and Organiza- Universitetsforlaget.
tional InnovationsA Conceptual Discussion IVA (1979a). "Kunskap som grund for industriell
and Some Notes on Consequences for Productiv- utveckling." Stockholm: Ingenjorsvetenskapsak-
ity and Employment." Working Paper. World ademien.
Employment Programme, International Labour IVA (1979b). "Teknik och industristruktur." Stock-
Organization, Geneva. holm: Ingenjorsvetenskapsakademien.
Edquist, C., and Glimell, H. (1989). "Swedish Fron- IVA (1979c). "Svensk industris langsiktiga utvec-
tiers of ChangeA Guide to the Impact of New kling." Stockholm: Ingenjorsvetenskapsakade-
Technologies, Work Design and Management mien.
Practises." Report prepared for the Directorate Johansson, A. (1989). Tillvdxt och klassamarbete
for Social Affairs, Manpower and Education, en sludie av den svenska modellens Uppkomst.
OECD, mimeo (to be published by OECD). Stockholm: Tiden.
Edquist, C., and Jacobsson, S. (1988). Flexible Auto- Johnson, B., and Lundvall, B. -A. (1988). "Institu-
mationThe Global Diffusion of New Technol- tional Learning and National Systems oflnnova-
ogy in the Engineering Industry. Oxford: Basil tion." Paper presented at the conference on
Blackwell. "Strategies of Flexibilization in Western Europe,"
Edquist, C., and McKelvey, M. (1991). "Hogteknol- Roskilde University, April, mimeo.
ogiska produkter och produktivitet i svensk indus- Kallehauge, L. E. (1989). "Spredningenafnumerisk
tri." A study for the Swedish Productivity Dele- styrede VEerktajsmaskiner i dansk 0konomi i per-
gation, in Forskning, Teknikspridning och ioden 1980-87." PIKE-smaskrift No. 3, August.
Produktivitet. Expert Report number 10 of the Aalborg: Institut for Produktion, mimeo.
Productivity Delegation. This report has also been Katzenstein, P. J. (1985). Small States in World
published in English as a working paper at the Markets. Ithaca: Cornell University Press.
Consortium on Competitiveness and Coopera- Lundvall, B. -A. (1987). "Technological Unemploy-
tion (CCC), the Center for Research in Manage- ment in a Small Open Economy." In R. Lund, et
ment, University of California at Berkeley, 1992. al. (eds.), Studies in Unemployment. Copenhagen:
The English title is "The Diffusion of New Prod- New Social Science Monographs.
uct Technologies and Productivity Growth in Lundvall, B. -A. (ed.) (1992). National Systems of
Swedish Industry." InnovationTowards a Theory of Innovation
Forum for Industriel Udvikling. (1988). "Har Dan- and Interactive Learning. London: Pinter Pub-
mark en framtid som industrination?" Koben- lishers.
havn, mimeo. Lundvall, B. -A., MaarbjergOlesen, N., and Aaen, I.
Gardlund, T. (1942). Industrialismens Samhdlle. (1984). "Det landbrugsindustrielle kompleks."
Stockholm: Tidens forlag. Aalborg: Aalborg Universitetsforlag.
Gjerding, A. N., and Lundvall, B. -A. (1990). "Tek- Madsen, P. K. (1989). "Orienten begynder i Malmo!
nisk fornyelse og produktivitetsudvikling i danske Beskasftigelsespolitik i Sverige, Finland og
industriverksamheder." PIKE-smaskrift No. 8, Norge." Arbejdsnotat No. 3, Specialarbejdernes
beskgeftigelsesudvalg. Kobenhavn: Specialarbej- Europaeiske lande." Arbejdsnotat No. 13, Speci-

derforbundet, mimeo. alarbejdernes vskstkommission. Kobenhavn:
Mjoset, L. (ed.) (1986). Norden dagen derpd. Oslo: Specialarbejderforbundet.
Universitetsforlaget. Petersen, E., et al. (1983). "Danskernes tilvarelse
Nordic Economic Growth. (1984). Economic under krisen." Psykologisk Skriftserie 8(3).
Growth in a Nordic Perspective. Helsinki: Frenck- Regeringens Proposition. (1990). Vissa Naringspol-
ell. itiska Fragor, Proposition No. 1989/90:88, Sver-
Nordisk FoU-statistik 1987og 1981-1987(1990). iges Riksdag. Stockholm: Mars.
OECD (1987). Reviews of National Science and Schon, L. (1985). Industrialismens FiJrutsdttningar.
Technology Policy: Sweden. Paris: OECD. Malmo: Liber.
OECD (1988a). Reviews of National Science and Senghaas, D. (1982). Von Europa lernenEntwick-
Technological Policy: Denmark. Paris: OECD. lungsgeschicht-liche Betrachtungen. Frankfurt am
OECD (1988b). Main Science and Technology Indi- Main: Suhrkamp.
cators 1982-88. Paris: OECD. Solidarisk Arbetspolitik. (1989). Huvudrapport fran
Pasinetti, L. (1981). Structural Change and Eco- Programkommitten om valfarden, industrins
nomic Growth. Cambridge, England: Cambridge omvandling och det goda arbetet. Stockholm,
University Press. Svenska Metallindustriarbetarforbundet, Kon-
Patel, P., and Pavitt, K. (1989). "Large Firms in gress 3-9 September.
Western Europe's Technological Competitive- Stenberg, L. (1987). "Utvecklingsblock i fornyelse
ness." Revised version of paper prepared for the av svensk industri, Ds I 1987:3." Stockholm:
Prince Bertil Symposium, November 1988, Industridepartementet
Stockholm. Svennilson, I. (1954). Growth and Stagnation in the
Pavitt, K.., and Patel, P. (1988). "The International European Economy. Geneva, United Nations,
Distribution and Determinants of Technological Economic Commission for Europe.
Activities." Oxford Review of Economic Policy Wennerberg, S. (1989). "Nationella Mikroelektron-
4(4). ikprogrammet (NMP) och Industriell Utvec-
Pedersen, J. (1986). "Afskedigelsesbeskyttelse og kling" (Utvarderingsrapport).
ansartlelsestryghed for den enkelte arbejder i 7
9
The Canadian System of Industrial
Innovation
DONALD G. MCFETRIDGE
The purpose of this chapter is to describe the set of institutions and institutional link-
ages that has evolved for the generation, diffusion, and application of knowledge in
Canada. The building blocks of this innovation system are the knowledge-generating
organizations including research and development performing firms, universities,
other nonprofit organizations, and government laboratories. As important but less
readily observed are the organizations that apply knowledge and the interactions
among users and suppliers of knowledge. As Freeman (1988, p. 2) puts it:
The national system of innovation is not just a set of laboratories but is a cumu-
lative process of learning by producing, learning by using and learning by the
interaction of producers and users.
The organizations comprising an innovation system can interact in a variety of

ways:
1. Joint support or performance of research, development, and demonstration proj-
ects.
2. Cooperation between users and suppliers to improve intermediate inputs.
3. Acquisition of knowledge and skills by employees of one organization that are
transferable to other organizations or to new ventures.
Interaction can occur in a variety of dimensions or spaces. It can occur in geo-
graphic space wherein individual organizations derive benefits from other organiza-
tions operating in community or region or it can occur among firms making use of the
same underlying technology. In this case a discovery by one organization contributes
to the solution of problems in a variety of industries. Interaction also occurs between
users and suppliers along the production chain in a single industry.
The importance of interaction is that it makes an innovation system more than
the sum of its parts. Each organization contributes to the effectiveness of the others.
Some of these contributions are priced and reflected in normal market transactions.
Some are not priced and some are only vaguely perceived or defined. The existence of
299
these cumulative external effects makes it very difficult to apply the concept of allo-
cative efficiency to the analysis of innovation systems.
Implicit in the concept of a national system of innovation is the assumption that
the interaction of domestic organizations is both more important than and to be pre-
ferred to international linkages. In the Canadian system international linkages domi-
nate even in sectors where relatively strong domestic linkages exist. Relatively little
innovation occurs in Canada without foreign participation at some stage in the pro-
cess. The question in Canada has always been whether there are benefits to be derived
from substituting domestic for foreign linkages. The Canadian experience shows that
under certain circumstances this can be done. What is virtually impossible to resolve,
even ex post, is whether this is wealth improving.
National systems of innovation are shaped by national characteristics. The Cana-
dian system has a number of features that can be traced both to the relatively small
size of its domestic market and to such unique characteristics as its natural resource
endowment and vastness. Canada's small country characteristics appear to have been
accentuated by its proximity to a very much larger but technologically similar country.
This has contributed to the attenuation of domestic innovative linkages between, for
example, the resource industries and machinery producers. On the other hand it has
facilitated Canadian participation in the U. S. system, which appears to have improved
technology adoption in some cases (motor vehicles) and resulted in world scale R&D
mandates for affiliates in other cases. Thus the Canadian situation is unique both in
the constraints on and opportunities for domestic innovation.
The next section of the chapter describes both the characteristics of the Canadian
economy and how they have shaped its innovation system. The following sections
describe the innovating organizations that are the building blocks of the Canadian sys-
tem and the prominent linkages that exist and the resulting innovation complexes and
clusters. The question of whether these systems, dominated as they are by interna-
tional (largely transborder) linkages, are sufficient to generate future growth in
national wealth is addressed in the concluding section.
THE CHARACTERISTICS OF THE CANADIAN ECONOMY AND THEIR IMPACT

ON ITS INNOVATION SYSTEM
Canada ranks seventh among OECD members in Gross Domestic Product. A number
of studies (Walsh, 1987, 1988; Soete, 1988; OECD, 1989b) have classified Canada as
the smallest of the "large" OECD economies.
Industries based on natural resources have played a significant role in Canadian
economic development (Canada, Royal Commission on the Economic Union, 1985,
pp. 401 -413). Resource industries (including agriculture) presently account for nearly
9% of Canadian GDP and 17% of Canadian exports. Resource industries account for
a larger fraction of Canadian GDP than is the case in other large OECD economies
(Charette et al., 1986, Table II-3).
Manufacturing accounts for a correspondingly smaller share of Canadian GDP
than is the case in other large OECD economies. Canadian manufacturing activity has
historically been highly protected. With some notable exceptions such as pulp and
paper and nonferrous metals it has been oriented principally toward the domestic mar-
THE CANADIAN SYSTEM OF INDUSTRIAL INNOVATION 301
ket and has been burdened by the diseconomies of small-scale production (Baldwin
andGorecki, 1986).
A major departure from this pattern occurred in 1965 with the institution of free
trade in automobiles between the United States and Canada. The "Auto Pact" resulted
in the rationalization of motor vehicle production on a North American basis. Cana-
dian production of motor vehicles and parts presently accounts for 9% of Canadian
manufacturing GDP and exports of these products to the United States account for
25% of Canadian merchandise exports.
The Canadian economy is closely integrated with the U. S. economy. Seventy-
three percent of Canadian exports are to the United States and 69% of Canadian
imports are U. S. origin. U. S. and Canadian capital markets are closely linked. Nearly
24% of the assets of nonfinancial Canadian corporations are foreign controlled with
71 % of this being in U. S. hands. The figures for manufacturing are 44 and 68%, respec-
tively. Safarian (1985, p. 9) concludes that the incidence of foreign control is higher in
Canada than in any other advanced market economy with Australia, Belgium, and
Ireland being the closest rivals.
Canadian and U. S. labor markets are not integrated. Immigrants to Canada have
tended historically to come from the United Kingdom and more recently from Europe
in general and southeast Asia. The United States accounts for roughly 8-10% of the
immigrants to Canada. The United States has been the most important destination for
Canadian emigrants. Migrants in general have tended to be relatively highly qualified.
Canada devotes a relatively large fraction of its resources to education. Canada
ranks well ahead of the United States, Japan and the United Kingdom in terms of
public expenditures on education as a proportion of GNP (Ontario Premier's Council,
1988b, p. 224). Canada ranks less highly in terms of both the rate of participation in
formal education and the proportion of students completing secondary school.
According to one study some 31% of Canadian students fail to complete their second-
ary education. This is true of only 25% of U. S., 15% of U. K., and 14% of German
students (Ontario Premier's Council, 1988b, p. 221).
A more detailed study (Easton, 1988, pp. 11-13) shows that while a smaller pro-
portion of Canadians complete secondary school than is the case in the United States,
the proportion of Canadians completing university is roughly the same as in the
United States and the proportion of Canadians completing some form of postsecond-
ary education is higher.
Canadian secondary education is regarded as being somewhat weak, especially
relative to far eastern countries, in science and mathematics. Canada's requirements
for both scientific and entrepreneurial talent have historically been met, to a consid-
erable degree, by immigration.
The real public rate of return on a 4-year university degree in Canada is estimated
(as of 1980-1981) to lie in the 6 to 9% range depending on the region of the country
(Vaillancourt and Henriques, 1986). It is 7% in Ontario and 9% in Quebec. Given that
this is an average rate of return and includes returns to labor market signaling (which
are private but not social benefits) it likely overstates the true marginal public rate of
return. The implication is that Canada is not presently underinvesting in higher edu-
cation as a whole. The chapter on Australia in this volume concludes that the same is
true in that country.
Canada has a small and widely dispersed population. There are (as of 1986) only
three metropolitan areas with populations in excess of one million. Urban areas in
Canada tend to be widely separated frequently being in closer proximity to large U. S.
cities than to each other. Some analysts (Lacroix and Martin, 1987; Britton and Ger-
tler, 1986) concluded that the potential economies of urban agglomeration are
unlikely to be fully realized outside of Toronto, Montreal, and Vancouver. They are,
as a consequence, pessimistic regarding the ability of other Canadian cities to attract
major R&D facilities and to generate sustained new business spinoffs.
In its historic natural resource orientation, small dispersed population, and inte-
gration with a much larger economy Canada is unique. These characteristics helped
to shape an innovation system that is also unique in a number of respects.
The development of the national innovation systems of countries with small
domestic markets is subject to a variety of constraints not faced by larger countries. As
a result the characteristics of national systems of innovation should differ between
large and small countries. Ergas (1987) has argued that large size continues to yield
advantages even among the largest countries. Specifically, the United States has
advantages in innovation that are not available even to Germany, France, and the
United Kingdom. The Canadian innovation system has a number of small market
characteristics which are accentuated by its proximity to the U. S. system.
The interaction of users and suppliers has become recognized as an important
factor in the innovative process (Freeman, 1982). For small country suppliers (and
users for that matter) this interaction is necessarily international. It may be burdened
by the costs of communicating over long distances and of overcoming national differ-
ences in language, culture, or technological infrastructure. As a result small countries
are likely to be at a disadvantage in mass market innovation.
There is some evidence that Canadian innovative activity has tended to focus on
the "custom design" end of the market. De Bresson and Murray (1984) find that a
disproportionate share of Canadian innovations in the machinery and equipment sec-
tor is of the custom produced or small batch variety and conclude that in this sector
there is a "customized bias" in technological development. Polese and Verrault (1989)
attribute the very considerable shares of international consulting engineering contracts
won by Canadian firms, in part, to the nonstandardized nature of the product. The
locational and technical characteristics of each job differ and this reduces the advan-
tage accruing to firms based in large countries.
The disadvantage that small countries face in mass market innovation may result
in their allocating a smaller portion of their R&D resources to applied industrial
research and development and a larger portion to basic research (performed in uni-
versities or government laboratories) than large countries. Walsh (1987, pp. 107-108)
observed that small countries tend to be more oriented to basic research than large
countries. Canada exhibits this tendency. Canada's share of world research papers
averaged 4.3% (in all fields) over the period 1973-1984. The Canadian share of exter-
nal OECD patents over the period 1970-1983 was 1.4%. In contrast, Germany
accounted for 6.3% of world research papers and 19.7% of external OECD patents.
Inferences regarding the tendency of small countries to perform a smaller portion
of their R&D activity in their business sectors can be drawn from OECD data (OECD,
1989b, Table 7). There is a rough correspondence between GDP and the proportion
of R&D performed in the business sector. Canada is an outlier in this relationship. For
reasons that will be explored later, the proportion of Canadian R&D performed by the
government is higher than in a number of smaller countries.
In their efforts to reduce the cost of interacting with users in other countries inno-
vating firms based in small countries may perform more of their R&D abroad than
firms based in large countries. This tendency is likely to be particularly pronounced if
export sales are concentrated in a single foreign market as is the case with Canada.
Pavitt and Patel (1989, Table 8) provide evidence that this tendency exists. They find
that a larger proportion of the U.S. patents granted to large firms based in small coun-
tries are of foreign (i.e., outside the home country) origin than is the case with large
firms based in large countries. In this respect Canada bears a fairly close resemblance
to the Netherlands, Switzerland, Belgium, and Sweden. For all these countries a sig-
nificant fraction of U. S. patents granted to their large firms is of either U. S. or German
origin.
A second problem facing small countries is that given a minimum required scale
of effort for successful innovation in most industries, they will be under greater pres-
sure than large countries to specialize. Soete (1988) confirmed that small countries do
tend to be more specialized in their innovative activity. He also finds that Canada
exhibits this small country characteristic (Soete, 1988, Table 5.4).
This tendency to specialize also implies that industrial R&D spending is more
concentrated (less evenly distributed among firms) in small countries than in large
countries. A few world class players may account for a large fraction of a small coun-
try's industrial R&D. Van Tulder and Junne (1988, pp. 178-180) conclude from their
examination of large firm R&D spending in the United States, Japan and Europe that
country size and concentration of industrial R&D are negatively related. Canada con-
forms to this pattern. Canadian industrial R&D spending is more concentrated than
that of the United States, Japan, Germany, France, the United Kingdom, or Italy but
less concentrated than that of the Netherlands, Switzerland, or Sweden (see later for
more detail on Canada).
It might also be conjectured that because they are less able to support specialized
or "full-time" inventors, small countries might also have proportionately more part-
time or individual inventors. The finding of Seguin-Dulude and Desranleau (1989)
that the share of individual inventors in Canadian patents is roughly twice the share
of individual inventors in U. S. patents is consistent with this reasoning.
Even with specialization, the innovative sectors of a small country are likely to
be relatively thin in the sense that their fortunes ride on a few firms and, perhaps, on
a few technologies. This lack of diversity is a characteristic of Canadian high-tech sec-
tors and regions (Amesse et al., 1989). A consequence of this lack of diversity is that
the failure of one firm or project can be highly disruptive. For example, it is said that
the cancellation of the Avro Arrow interceptor program in 1958 resulted in the dis-
persal of a considerable portion of Canada's aeronautical design and engineering talent
largely to the United States (Lukasiewicz, 1986). Similar concerns have been raised
regarding the CANDU nuclear reactor, which has not sold well internationally.
Smaller failures can be accommodated. Nichol (1985) traces the formation of many
successful Ottawa information technology firms to the failure of Microsystems Inter-
national, a local microchip manufacturer.
Successful imitation or adoption of new technologies developed elsewhere also
requires resources. Walsh (1987, 1988) notes that industries in which there is relativel)
little domestic R&D spending may have difficulty assimilating technological advances
made abroad and may, therefore, have difficulty remaining competitive. Small market
size can handicap imitation as well as innovation. As a result small countries may be
less able, hence slower, to assimilate new technologies. The rate of diffusion has been
a long-standing Canadian concern (Economic Council, 1983). There has also been a
recognition that one of the costs of world-scale innovative capability in a few sectors
may be a reduced rate of diffusion and adoption in others.
GOVERNMENT AND UNIVERSITY R&D
Just under 20% of Canada's R&D is presently performed by the federal and provincial
governments. The proportion of R&D performed in the government sector has fallen
by some 47% over the past 20 years.
The decline in the share of R&D performed in government facilities is a conse-
quence of the relatively rapid growth of business financed R&D. The Canadian expe-
rience in this regard is similar to that of the OECD as a whole (OECD, 1986, pp.
25-29).
The proportion of the national R&D effort performed within the government sec-
tor in Canada continues to be much higher than in both the large OECD economies
and a number of advanced smaller economies such as the Netherlands, Sweden, Bel-
gium, and Switzerland.
The relatively large fraction of R&D performed within the government sector
may be, in part, a matter of bureaucratic choice, which can be altered over time by
directions to contract-out (Senate Special Committee on Science Policy, 1972; Senate
Committee on National Finance, 1984). There may, however, be other forces, most
importantly the nature of the R&D activity funded by governments in Canada, at
work.
The composition of government funded R&D in Canada differs from the OECD
norm in a number of respects. Relatively little government R&D is devoted to defense.
Relatively large amounts of R&D are devoted to agriculture, energy, and mining
(OECD, 1989b, p. 44). This emphasis on the resource sectors goes back to the turn of
the century (Smith, 1983). A number of provincial governments also regard the
resource sectors as their top R&D priority (OECD, 1988, pp. 47-50). There may well
be benefits in the form of both the accumulation of experience and project comple-
mentarity from centralizing the performance of this type of on-going process improv-
ing generic R&D. Given centralized performance there may be further benefits in
terms of reduced contracting costs from the government performing the R&D itself.
Canada's most important R&D performing, supporting, and coordinating insti-
tution is the National Research Council (NRC). The NRC engages in both basic and
applied research, the latter in support of both government and industry. The NRC,
which opened its first laboratory in Ottawa in 1932, has also operated combined tech-
nical advisory and industrial R&D support program (Industrial Research Assistance
Program, IRAP) with considerable success since 1947.
Individual provinces also maintain research councils. These organizations are of
modest scale with combined R&D spending in 1987 being roughly one-quarter of the
NRC's spending. The provincial research councils are oriented toward the servicing of
local industry with jointly funded R&D projects becoming more common.
Although much of the R&D performed within the government sector in Canada
is intended to assist industry, doubts continue to be expressed regarding its relevance
(Task Force on Federal Policies and Programs for Technology Development, 1984;
Clarke and Reavely, 1988). Although there are clear exceptions, such as agricultural
R&D, the basic problem appears to be a lack of communication between government
laboratories and their potential clients. Numerous attempts have been made to inte-
grate government laboratories more closely into industry innovation systems (Smith,
1983).
Universities play a number of roles in an innovation system, the education and
basic research roles being the most prominent. Universities also engage in applied
research. Sometimes this is on behalf of or in cooperation with industry. In other cases
technologies developed in universities are licensed or are exploited in new businesses
started by faculty members (academic entrepreneurs).
Canadian universities appear to have been relatively effective in fulfilling their
basic research function. Canada ranks seventh in its share of world research papers
(Statistics Canada, 1988, Table 41). Similarly, Canada ranked third and fourth among
OECD countries in its share of publications in the fields of computers and electronics
and electrical engineering respectively in 1984 (OECD, 1989a, Table 6).
Although they have been effective in their education and research functions,
Canadian universities appear to have been less closely linked to the domestic industrial
innovation system than are universities in other countries. The business sector funded
3.2% of Canadian university R&D in 1988 while in the United States the figure was
6.5% (National Science Foundation, 1989, Table 8-1). In Germany, Sweden, and Nor-
way, business funds approximately 5% of university R&D (OECD, 1989b, p. 68).
While there are numerous examples of successful transfers of technology from
universities to industry and of successful academic entrepreneurs (see, for example,
Clifford, 1990), the impression left by very limited evidence is that Canadian univer-
sities have not served the incubator function that U. S. universities have and that tech-
nologies developed in Canadian universities are frequently exploited abroad rather
than domestically (McMullan and Melnyk, 1988; Stuart-Haile, 1989; OECD, 1988;
Doutriaux, 1989). Attempts have been in recent years to develop better university-
industry linkages (Science Council, 1987b). Examples of linkages that have existed or
have been formed in recent years are cited later.
INDUSTRIAL R&D, PATENTING, AND TECHNOLOGY ADOPTION
Major R&D Performing Firms

Canadian industrial R&D spending amounted to $4.4 billion (Cdn.) in 1988. Canada
ranks tenth of the 12 largest OECD economies in its industrial R&D:GDP ratio. Can-
ada ranks ahead of Australia and Spain but has slipped behind Italy in recent years.
Canadian industrial R&D spending is concentrated within a few firms as Table
9.1 indicates. The four largest R&D performers account for one-third of the Canadian
total. The identity of the largest R&D performers is also instructive. The largest is BCE,
Table 9.1, Major Canadian Industrial R&D Performers (Self-Financed R&D Performed in
Canada in Millions of Canadian $)
Cumulative
1989 R&D 1988 Percentage of
Company Industry Ownership (est.) R&D Canadian Total
BCE Inc. Telecommunications Canadian 813 768 19.5
equipment
Pratt and Whitney Aeronautics United States 247 219 25.1
Canada
IBM Canada Information technology United States 181 151 28.9
Atomic Energy of Nuclear Government 166 163 33.1
Canada
Ontario Hydro Electric power Government 150 117 36.1
Hydro Quebec Electric power Government 136 129 39.4
Alcan Aluminum Nont'errous metals Canadian 111 105 42.0
Canadair Aircraft Canadian 106 91 44.3
Boeing of Canada Aircraft United States 91 83 46.5
CAE Industries Aeronautics Canadian 75 60 48.0
Imperial Oil Petroleum products United States 70 67 49.7
Allied Signal Canada Electronics United States 38 37 50.6
Inco Ltd. Nonferrous metals Canadian 37 33 51.5
BC Telephone Telecommunications United States 37 34 52.3
Digital Equipment Information technology United States 36 38 53.3
of Canada
Shell Canada Petroleum products Netherlands 33 31 54.1
Mitel Telecommunications United 29 29 54.8
equipment Kingdom
CIL Industrial chemicals United 29 27 55.5
Kingdom
MacMillan Bloedel Forest products Canadian 25 24 56.1
DuPont Canada Industrial chemicals United States 21 18 56.6
Source: Financial Post (October 26,1989).
which is comprised of Bell Canada, Northern Telecom and Bell-Northern Research.

Pratt and Whitney is owned by United Technologies and has a world product mandate
in small turbine aircraft engines. IBM Canada does software R&D in support of IBM's
worldwide operations. AECL is owned by the federal government. Its principal prod-
uct is the CANDU nuclear reactor and much of its research budget is spent in the sup-
port and continuing development of it.
There is a preponderance of U. S. subsidiaries, government enterprises, and util-
ities or utility-linked companies among the largest R&D performers in Canada.
Although its spending dwarfs that of other Canadian firms, Northern Telecom R&D
is quite modest in comparison with its international competitors. For example, AT&T
spent over $2 billion on R&D in 1988 (Ontario Premier's Council, 1988b, p. 147).
An examination of patenting activity reveals a somewhat broader distribution of
innovating activity. The largest corporate patentees of inventions of Canadian origin
are listed in Table 9.2. Northern Telecom still dominates being responsible for 10% of
corporate patents over the period 1982-1986.
Table 9.2. Major Corporate Recipients of Canadian Resident Patents
Patents Patents
Firm Industry Ownership 1978-1982 1982-1986
Northern Telecom Telecommunications equipment Canadian 59 74
Canadian General Electrical appliances and United States 25 15
Electric industrial equipment
Company
DuPont Canada Chemicals United States 15 11
Inco Mining Canadian 15 8
Domtar Paper and products Canadian 11 9
Imperial Oil Petroleum products United States 10 10
Alcan Aluminum Canadian 11 7
Polysar Petrochemicals Canadian 10 8
CIL Chemicals United Kingdom 10 8
AECL Nuclear Government 8 6
Corporate total 791 774
Source: Statistics Canada ( 1 989b, Tables 4.4 and 4.5).
Major R&D Performing Industries
The distribution of R&D spending by industry is summarized in Table 9.3. The struc-
ture of R&D in Canada differs from that of the large OECD economies in some impor-
tant respects. First, Canada has a much higher proportion of its R&D in the service
sector (including utilities, communications and engineering, and scientific services)
than do the large OECD economies (OECD, 1986, p. 82).
Second, a much smaller proportion of Canadian industrial R&D occurs in the
motor vehicles industry than is the case in both the large and the small, advanced
OECD economies.
Third, a much smaller proportion of Canadian industrial R&D is in the machin-
ery sector than is the case in both the large and many small, advanced OECD econo-
mies.
Fourth, a larger proportion of Canadian industrial R&D occurs in the basic met-
als and "other industries" (the largest being paper and wood products) than is the case
in most other OECD economies.
The distribution of industrial R&D spending in Canada reflects both the distri-
bution of industrial activity and the characteristics of innovative activity in Canada.
Resource sectors including forest products, metals, and energy are relatively important
in Canada as is R&D in support of them. Innovation in support of the resource indus-
tries has historically been oriented toward solving local process problems. Product
innovations have tended to be either of the custom design or small batch variety. Large
market technologies have tended to come from the United States.
As a result, domestic innovative activity in the machinery and surface transpor-
tation industries has been concentrated in fields such as logging equipment, hard-rock
drilling equipment, and off-road vehicles. Canadian mass market product innovations
such as the snowmobile are regarded as exceptions (De Bresson and Lampel, 1985).
Although Canada has a relatively large motor vehicle industry, it is completely
integrated with that of the United States. Research, development, and engineering are
Table 9.3. Distribution of Intramural Business R&D Spending by Industry, 1988

1988 R&D Average Annual Rate 1986 R&D
Spending Cumulative of R&D Growth Intensity
Industry ($Mil Cdn) Share (%) 1980-1988 (%)
Telecommunications 720 18.7 16.8 17.8
equipment
Aircraft and parts 533 30.1 13.9 13.9
Engineering and 363 36.8 20.4 22.8
scientific services
Business machines 298 43.6 20.2 3.5
Other electronic 293 50.3 20.7 12.8
equipment
Electric power 242 55.8 14.4 0.9
Other manufacturing 232 61.1 35.0 0.9
industries
Computer services 216 66.0 32.5 15.1
Other 185 70.2 9.3 1.4
(nonpharmaceuticals)
chemical products
Primary metals 157 73.8 4.9 0.9
Petroleum and coal 136 76.9 -2.1 0.5
products
Transportation and 134 79.9 13.1 0.4
other utilities
Pulp and paper and 128 82.8 8.3 0.3
wood products
Other transportation 105 85.2 9.8 0.3
equipment
Source. Statistics Canada (1989c, Cat. No. 88-001, July).
centralized in the United States. The natural tendency of research, development, and
engineering to locate in the larger market (and near corporate headquarters) was rein-
forced by the Auto Pact in 1965 under which Canada effectively traded local research
development and engineering activity for more assembly jobs (Acheson, 1989).
industrial Technology Adoption

In its 1983 report the Economic Council of Canada concluded that new technology
diffuses more slowly both to Canada from abroad and within Canada than is the case
in other developed countries. In their subsequent examination of the evidence
McFetridge and Corvari (1986) concluded that Canada tended to be among the earliest
recipients of new U. S. technologies over the period 1966-1979. They attributed this
to the proximity of Canada to the United States and to the prominence of U. S. mul-
tinationals in Canada. They also noted that Canada's position in the transfer order of
U. S. technologies had slipped over time. With respect to domestic diffusion the evi-
dence was that it had been slower or less complete than in the United States in some
manufacturing industries. The evidence for the service industries showed some lags by
comparison with the United States but was generally inconclusive.
More systematic evidence is provided by the surveys of advanced manufacturing
technology conducted by Statistics Canada (1989a) and the U. S. Department of Com-

merce (1989). The published results of these surveys are not comparable. They differ
with respect to the sizes of establishments surveyed and the industrial groupings. Sta-
tistics Canada has prepared and will ultimately publish tabulations that match the U.
S. survey more closely. These tabulations were made available to the author for statis-
tical analysis.
Statistical analysis of the percentage of establishments in five major groups (metal
fabricating, machinery, transportation equipment, electrical and electronic equip-
ment, and instruments and related products), using each of 17 advanced manufactur-
ing technologies, reveals that, on average, the incidence of use is greater in the United
States than in Canada by some 5.6 percentage points or 44%. This difference is statis-
tically significant. It holds across major groups and technologies.
The greater incidence of advanced manufacturing technology use in the United
States is due in part to differences in both the scale of industrial establishments and the
mix of industries in the two countries. Industrial establishments in the United States
tend to be larger and thus more able to absorb the fixed cost of adopting new technol-
ogies than Canadian establishments. Industries that are intensive users of advanced
manufacturing technologies tend to be more prominent in the United States than in
Canada. An important example is the defense-related industries.
If Canadian plants were of the same scale as U. S. plants, the incidence of use of
advanced manufacturing technologies in the United States would be 3.5 percentage
points or 24% greater than in Canada. Correcting for the difference between the two
countries in the importance of defense production would further narrow but would
not eliminate the gap between them.
The remaining gap between U. S. and Canadian advanced manufacturing tech-
nology usage may be a consequence, in part, of the relative weakness of certain aspects
of the Canadian innovation system. The importance of skilled personnel in facilitating
the adoption of new technology has been cited in a variety of studies (see for example
Julien et al., 1988). Supplier linkages may be more tenuous in Canada than in the
United States, perhaps because a greater proportion of the advanced manufacturing
technology used in Canada is imported. A possible remedy might be the further lib-
eralization of restrictions on trade and migration.
LINKAGES AMONG INNOVATING ORGANIZATIONS
Horizontal Linkages
Horizontal linkages are defined here as relationships among organizations at the same
stage in the innovative process. Horizontal linkages exist for two reasons. First, they
are a means of internalizing externalities. That is, they are a means by which the ben-
eficiaries of a technological improvement can contribute toward the recovery of its
cost. Second, horizontal linkages are a means of realizing the benefits of specialization
in research and development.
Horizontal linkages can be embedded in a wide variety of institutional arrange-
ments. Among the more familiar are technology joint ventures and industry technol-
ogy institutes. Horizontal arrangements can also embrace governments and univer-
310
SMALLER HIGH-INCOME COUNTRIES
sities. Government subsidies and tax incentives are a means of channeling support to
innovators from spillover beneficiaries. Government and university laboratories may
also contribute their special skills to cooperative research arrangements.
Some indication of the relative importance of various forms of formal and infor-
mal, horizontal and vertical linkages is provided by Longo (1989). His findings are
reported in Table 9.4. Foreign affiliates, suppliers, and universities are the most prom-
inent external sources of R&D. Taken together government R&D agencies and hori-
zontal domestic linkages (with or without alliances) are also important. Customers and
trade associations are relatively unimportant.
Longo also investigates the incidence and nature of technology alliances among
R&D performers. He finds that over half the firms surveyed were involved in tech-
nology alliances. Foreign and domestic alliances were of roughly equal frequency.
Trade association research, development, and diffusion activities in Canada have
been examined by Litvak (1985). He finds that in general there are relatively few trade
associations involved in R&D or with any technical staff. The largest and oldest indus-
try research institute is the Pulp and Paper Research Institute of Canada (Paprican),
which was founded in 1925. It is supported largely by contributions from member
companies. Paprican engages largely in applied process research. It has been respon-
sible for a number of process innovations and has played an important role in solving
common production problems.
There are two active research institutes in the energy industry. The Canadian Gas
Research Institute conducts research on gas using appliances, gas distribution, and
safety on behalf of member utilities from whom it derives 80% of its budget. It is also
an associate member of the (U. S.) Gas Research Institute and benefits from research
conducted by that organization. The Canadian Electrical Association conducts R&D
into problems in electricity generation, transmission, and distribution on behalf of
member utilities and equipment suppliers from whom it derives most of its revenue
(the balance comes from the federal government).
Some research associations serve more to coordinate than to perform R&D. An
example is the Canadian Steel Industry Research Association, which provides a link
between the steel companies and metallurgical R&D, which is conducted largely in
Table 9.4. External Sources of R&D for Ontario R&D

Performers, 1989
Percentage of Respondents
Ranking the Source in the
Source Top Three
Affiliated foreignfirms 24
Suppliers 23
Universities and colleges 20
Other government R&D agencies 16
National Research Council 14
Other unaffiliated firms, same industry 13
Unaffiliated firms in alliances 12
Purchasers 10
Industry research associations 10
Source: Longo (1989, p. 9).
universities and provincial research councils. Support for this R&D comes largely
from federal and provincial governments and from the American Iron and Steel
Institute.
A government agency that has played a prominent role in supporting, coordi-
nating, and disseminating research in the area of heavy oil extraction is the Alberta Oil
Sands Technology Research Agency (AOSTRA). AOSTRA has funded research,
development, and demonstration projects jointly with petroleum companies that have
evolved into commercially viable heavy oil recovery plants. AOSTRA has also funded
a good deal of academic research in petroleum geology and engineering.
A more general government program that supports industrial R&D and serves as
a clearinghouse for specialized technological capabilities is the National Research
Council's Industrial Research Assistance Program (IRAP). The program is staffed by
a network of 250 industry technology advisors operating out of 160 technology centers
across Canada. It is regarded as Canada's most successful innovation support program
(Supply and Services, 1990).
Other federal government programs serve largely to provide financial assistance
to R&D-performing firms. Of note are the Defence Industry Productivity Program,
which subsidizes R&D in the aerospace industry (Ontario Premier's Council, 1988a,
Exhibit VIII.21), and the R&D tax credit provisions, which have been among the most
generous in the world (United Kingdom, Inland Revenue, 1987).
Total government financial support including R&D tax credits, subsidies, and
contracts amounts to approximately 20% of Canadian industrial R&D spending. It is
frequently pointed out that the U. S. government directly funds some 30% of U. S.
industrial R&D and that, by this standard, the level of government support in Canada
is inadequate (Ontario Ministry of Industry, Trade and Technology, 1988).
The validity of this type of comparison depends on whether tax measures, sub-
sidies, and contracts provide equivalent support to industrial R&D performers. There
are reasons to believe that R&D contracts, which tend to be the means by which the
government funds R&D in the United States and a number of other countries, are not
equivalent, dollar for dollar, with subsidies and tax credits, so that intercountry aggre-
gate R&D support comparisons are not meaningful.
More fundamentally, the question arises as to whether it is fruitful for govern-
ments in Canada to contract for more industrial R&D. This issue is addressed later.
To conclude, although there are numerous examples of cooperative research
projects, there appear to be relatively few examples of significant institutionalized
industry-wide R&D activity in Canada. There are a number of possible explanations.
First, both the NRC and the provincial research councils perform and support the type
of applied R&D and technology acquisition activity that might otherwise be per-
formed by industry associations.
Second, Canada has a well-developed engineering services industry that can pro-
vide, perhaps at a lower cost, the type of specialized expertise in applying technology
that an industry association might accumulate.
Third, there is the small country problem. Industrial R&D in Canada has tended
to address local problems. The more broadly applicable technologies tend to come
from abroad. It may be the case that in many Canadian industries, especially those
outside the resource sector, the type of industry-wide yet uniquely Canadian techno-
logical problems that might be addressed by an industry association are relatively rare.
Vertical Linkages
A vertical linkage is a relationship between organizations at different stages in the inno-

vative process. The importance of these user-supplier linkages in the innovative pro-
cess has been emphasized by Rosenberg (1982), Freeman (1982), and Anderson and
Lundvall(1988).
Anderson and Lundvall (1988, p. 11) have observed that it is a characteristic of
small countries that the specialization pattern and technological competence of the
engineering sector reflect, in part, the history of the primary sector. The Canadian
experience is consistent with this observation. As will be demonstrated, much of the
indigenous Canadian innovative effort has been in the service of the agriculture, min-
ing, energy, forestry, and telecommunications industries.
The influence of user-supplier linkages on innovation in Canada has been inves-
tigated by De Bresson and Murray (1984). These authors rank Canadian industries by
their number of world first innovations and adoptions over the period 1945-1978.
They also rank capital goods industries according to their capacity relative to domestic
demand. They find that two industries, miscellaneous ground transportation (off-road
vehicles) and services to mining (air survey, exploration, and mapping) stand out as
being world first innovators without being important adopters of foreign innovations.
The unique requirements of Canadian geography bred a unique Canadian innovative
effort in these two cases (p. 101).
Two capital goods industries are cited for their record of innovation and success
at exporting and/or competing with imports. The logging equipment industry has a
strong record of innovation and is highly export oriented (p. 131). The pulp and paper
machinery industry has also recorded a number of world first innovations and has sup-
plied a significant portion of domestic demand in the past (p. 133).
More generally, De Bresson and Murray find that the major innovating industries
are also the major adopters of foreign technologies. The most prominent innovating
industries are characterized by strong links with resource sector or government (some-
times acting on behalf of the resource sector) or public (electricity, gas, telecommu-
nications) utility customers but not with the manufacturing sector. Prominent
resource sector users of innovations or adoptions include farming (chemicals), petro-
leum and gas recovery, pulp and paper, smelting and refining, and hard rock mining
(p. 181). The major innovating industries and their most prominent customers are
listed in Table 9.5.
There are a number of Canadian examples that illustrate the essential role that
users can play in defining the innovative task and in testing and refining innovations.
De Bresson and Murray cite the role of the forestry company MacMillan Blodel in the
design, testing, and use of a self-unloading barge for log transport. Litvak and Maule
(1982) note that Imperial Oil was both the first customer and the source of the design
for off-road vehicles built by Canadian Foremost.
The role of government, government utilities, and regulated utilities as users of
innovation in Canada also speaks to the role of procurement in encouraging the devel-
opment of domestic innovation systems and, perhaps, in strategies to "engineer com-
parative advantage." Polese and Verrault (1989) and the Ontario Premier's Council
(1988a) argue that the prominence of Quebec-based engineering consulting firms in
electricity and dam construction both in Canada and in world markets can be traced
Table 9.5. Major Innovating Industries and Their Customers, 1945-1978
Rank by Number
Innovating Industry of Innovations Using Industries
Miscellaneous machinery and 1 Wood products, mining, milling,

equipment smelting and refining
Scientific and professional 2 Health and welfare services,
equipment government, mining services
Aircraft and parts 3 Aircraft and parts, government
Communications equipment 4 Communications utilities, government
electrical products
Engineering and scientific services 5 Utilities, government
Industrial chemicals 6 Industrial chemicals
Electrical industrial equipment 7 Government, utilities
Source: De Bresson and Murray (1984, Tables 3 and 3.3 and List 5.3).
not only to Quebec's massive hydroelectric developments but also to Hydro-Quebec's

policy of contracting out engineering requirements to local firms. Attempts by Hydro-
Quebec to pursue similar policies with respect to suppliers of capital equipment (gen-
erators, transformers, cable, etc.) are described by Faucher (1989). The prospects of
levering local purchases into an export industry seem less promising in this case.
Dalpe (1988) argued that Northern Telecom's status as a major exporter of tele-
communications equipment (and Canada's largest R&D performer) is a result of its
integration with and preferred access as a supplier of the Bell system in Canada. It is
certainly the case that Northern Telecom's emergence as an innovative force began
with the severance of its technological cooperation agreement with Western Electric
by U. S. antitrust authorities in 1956. There is also agreement that the assurance of a
domestic market was an important factor in Northern's decision to engage in the R&D
projects in the 1960s and 1970s, which provided the foundation for its current success
(Restrictive Trade Practices Commission, 1983, pp. 201-224). There is considerably
less agreement, however, regarding the importance of full vertical integration for tech-
nological development in this industry (Globerman, 1980).
Although user-supplier linkages have contributed to innovation in Canada, it has
been suggested that domestic linkages are weaker and less prevalent in Canada than
elsewhere. With respect to the forestry industry in Canada, Hayter (1988) concludes:
Although forest-product firms are the principal customers of equipment sup-

pliers, in Canada there is no systematic co-operation between forest-product
firms and equipment suppliers in the development and exploitation of technol-
ogy (such co-operation exists in Scandinavia). Technological liasons between the
in-house R&D groups of forest-product firms and of equipment suppliers are
sporadic. R&D by equipment suppliers generally leads to the development of
prototypes that are tested in the facilities of the forest-product firm (which may
obtain lower prices for its willingness to be a guinea pig), by its in-house R&D
group, where one exists. Similarly forest-product firms that develop a new tech-
nology normally seek a close working relationship with an equipment supplier.
(P. 76)
Similarly, De Bresson and Murray (1984) find that the incidence of sustained
user-supplier interaction is lower in Canada than in the United Kingdom. They con-
clude that there are
few clusters of interacting industries which appears to be, in comparison to the

United Kingdom, a feature peculiar to the Canadian system. An obvious factor
of the lack of clusters of interacting industries in Canada is the reliance on for-
eign technological inputs, (p. 187)
Although sustained and successful horizontal and vertical innovative linkages

can be identified in Canada, they coexist with and are often dominated by foreign link-
ages. This appears especially true of innovations involving capital goods. It also
involves links with foreign research associations and foreign "branch plants" as both
innovators and suppliers. For some (Hayter, 1988, pp. 74-75), the problem is the
branch plant itself with its mandate limited to the pursuit of anything but uniquely
Canadian problems. Whether this is the only form of branch plant participation in
innovation in Canada and whether foreign ownership can be regarded as a causal fac-
tor in the analysis of innovation systems will be discussed further later.
Employee Mobility
Organizations that provide their employees with the knowledge and skills required to
found new innovative organizations are called incubators. Employees may acquire
knowledge and skills in the performance of their duties (learning by doing) or through
formal in-service training. To the extent that some forms of learning necessarily occur
on the job, incubator organizations are an essential complement to the formal edu-
cation system. Incubators may be especially important in providing managerial skills.
The role of large firms and nonmarket organizations as sources of technological entre-
preneurship has been widely discussed (Miller and Cote, 1987).
The sources of technologically oriented entrepreneurship in Canada have been
investigated by Litvak and Maule (1972,1980, 1982) and Nichol (1985). These studies
emphasize the international linkages in this aspect of Canada's innovation system. A
significant fraction of "high-tech" entrepreneurs in Canada is either foreign born and/
or foreign trained. Many entrepreneurs have also either worked abroad or been
employed by foreign-owned firms.
The studies by Nichol and Litvak and Maule differ in their conclusions regarding
the nature of the knowledge entrepreneurs take with them from their previous employ-
ers. Litvak and Maule (1972, pp. 4-30) find that 64% of the entrepreneurs they survey
left their previous employment to exploit a specific idea or product. Nichol (1985,
Table 4.20) finds that the entrepreneurs' initial product lines were competitive with
those of previous employers in 30% of the cases she examines. She concludes that in
the remaining 70% of the cases entrepreneurs made use of technology management
skills acquired from previous employers.
Nichol provides some interesting evidence on the previous employment of high-
tech entrepreneurs in the Ottawa area. Her results are summarized in Table 9.6. These
results are instructive in a number of respects. First, there are two universities and a
major R&D organization (the National Research Council) in Ottawa. No entrepre-
Table9.6. Sources of High-Tech Entrepreneurship in Ottawa, 1982

Organization Type Percentage
Computing Devices of Canada Foreign-owned corporation 18.9
Leigh Instruments Canadian-owned corporation 13.2
Northern Telecom, Bell-Northern Research, Canadian-owned corporation 28.3
Microsystems International
Telesat Canada Government corporation 3.8
Other corporate 7.5
Federal government departments and 28.3
regulatory agencies
Total 100.0
Source: Nichol (1985, Table 4.18).
neurs came from these organizations. Second, Ottawa is home to the bulk of federal
government employment including scientific employment. Although federal depart-
ments were the source of 28% of the technological entrepreneurs surveyed, this is a
disproportionately small contribution given the federal government share of Ottawa
employment.
Third, although the Bell-Northern group is an important source of technological
entrepreneurship having spawned some 50 smaller companies (Ontario Premier's
Council, 1988b, p. 167), much of its contribution to new high-tech entrepreneurship
in Ottawa is the result of the bankruptcy of its subsidiary Microsystems International.
Fourth, Computing Devices of Canada, which is proportionately the most impor-
tant incubator in the Ottawa area, has been U. S. owned since 1956. It has been a suc-
cessful exporter of electronic surveillance and guidance systems to military customers
for 30 years. Its employees have founded such companies as Leigh Instruments (itself
an important incubator), Gandalf, and Lumonics. Thus, foreign affiliates can serve as
incubators and may be superior in this regard to government scientific organizations.
Whether they are superior to local firms (to the extent that the latter are an alternative)
is an as yet, unanswered question.
Innovation Systems
Interaction among innovators may take place in a geographic, a technological, or an

industry context. Innovative activity is frequently concentrated geographically. The
benefits of geographic concentration appear to lie in the economies derived from large
pools of skilled personnel and from personal communication between users and sup-
pliers of innovation and among researchers pursuing alternative approaches to the
same problem (Malecki, 1987).
The geographic distribution of innovative activity in Canada has been examined
by Amesse et al. (1989). These authors report that, as of 1983, some 47% of Canadian
high-tech employment was in Toronto with 22% in Montreal and 12% in Ottawa. In
addition to being the largest, the Toronto high-tech sector has been established the
longest and is the most diversified with significant representation in telecommunica-
tions, aerospace and information technology. Montreal is primarily aerospace, led by
Pratt and Whitney, and telecommunications. Ottawa's development is the most
recent. It is concentrated in the telecommunications (Bell-Northern) and information
technology sectors.
In the opinion of Amesse et al. (1989) and of Lacroix and Martin (1987),
Toronto, Montreal, and Ottawa qualify as Canada's only high-tech agglomerations.
The OECD (1988) concludes that the "critical mass" necessary for the sustained devel-
opment of new products and new technology-based firms has been reached in the west-
ern provinces in some industries. In particular critical mass has been achieved in Brit-
ish Columbia in telecommunications (remote terminals, signal processors, data
transmission, image processors, chip design), in Alberta in electronic devices related
to geophysical and seismic applications, and in Saskatchewan in electronic instru-
mentation and transmission devices (satellite receivers, electronic tracking systems).
Interaction among innovators also occurs in a technological and an industrial
context. Technological interaction occurs when an innovation in one sector makes
possible innovations in other not necessarily related sectors. The importance of this
type of "innovation clustering" has been emphasized by Rosenberg (1979). Interac-
tion can also occur among industries linked in a user-supplier relationship. De Bres-
son and Murray (1984) identify two complexes of frequent and sustained user-sup-
plier linkages in metallurgy and electric power. They identify two clusters of more
modest linkages in wood and chemical products.
The metallurgy complex is comprised of the nonferrous metal mining and mill-
ing (at the center) and the iron and steel and the smelting and refining industries. These
industries are linked with each other and with the instruments, mining services, and
industrial chemicals industries. Links with the machinery industry are relatively weak.
The metallurgical complex also involves government and industry supported univer-
sity research (Litvak, 1985; Hutchison et al., 1987).
The electric power complex is comprised of the power generation and distribu-
tion industry, which is a strong user of electrical industrial equipment, miscellaneous
electrical products, and engineering services. This complex would also include uni-
versities such as the Ecole Polytechnique in Montreal. There is evidence of successful
collaboration among universities, engineering consultants, and utilities (Hutchison et
al., 1987) and between utilities and equipment suppliers (Faucher, 1989; Faucher and
Fitzgibbons, 1990).
The wood cluster has pulp and paper at the center with links to logging and mill-
ing. All three are linked strongly back to the machinery industry, while some are users
of innovations supplied by the miscellaneous transportation equipment and industrial
chemicals industries. This cluster would also include industry research associations
and the forestry institutes and laboratories of the Canadian Forestry Service as well as
university forestry schools. The industry research associations have focused on local-
ized process improvements and have been relatively successful (Hayter, 1988, Ch. 3;
Miller and Blais, 1990; De Bresson and Murray, 1984). There have also been at least
some cases of successful university-industry collaboration (Hutchison et al., 1987).
Although there are well-defined and relatively successful innovative linkages in
the forestry industry, its critics claim that innovation in this sector has fallen well short
of its potential in the areas of new forest product development and machinery (other
than logging equipment) for use in the forestry industry (OECD, 1988; Science Coun-
cil, 1987a). The sources of these failings are said to be inadequate R&D in the forest
products companies (Hayter, 1988; Miller and Blais, 1990) and weak domestic link-
ages with universities and with machinery manufacturers:
There is little research and development carried out in British Columbia by the
forestry companies. These companies do not contract out research at the uni-
versities. Forestry research at the universities is academic. Forintek . . . has an
excellent laboratory facility and could contribute significantly to a turnaround
in the forest sector. Some innovative products and processes have been devel-
oped at Forintek but the mechanisms for commercializing them seem inade-
quate. (OECD, 1988, p. 58)
A measure of the relative strength of domestic linkages in metallurgy and forestry

is the proportion of patents on technologies used in these industries, which is of domes-
tic origin. Information on the source of patents used by the pulp and paper and non-
ferrous metal industries is presented in Table 9.7. The process orientation of Canadian
innovative activity is evident. It is also evident that although these industries generally
source a larger fraction of their technology domestically than the Canadian average
(approximately 7% of Canadian patents are of Canadian origin) the bulk of their tech-
nology continues to come from abroad.
The chemicals cluster is comprised of industrial chemicals, miscellaneous chem-
icals, and the petroleum and plastics industries. According to De Bresson and Murray
(1984) these industries interact weakly with each other and strongly, as users, with
engineering, machinery, and communications equipment and as suppliers with agri-
culture, mining, and pulp and paper. Among the types of interaction that occur in the
petroleum industry are joint research, development, and demonstration projects on
in situ recovery of heavy oil coordinated by the Alberta Oil Sands Technology
Research Agency (AOSTRA) and joint petrochemicals research at the Institute for
Chemical Science and Technology at Sarnia, Ontario. The petroleum industry has also
been responsible for the development of petroleum engineering firms (Canadian firms
ranked sixth in the world market in 1982) and seismic instrument firms in Alberta
(Polese and Verreault, 1989; OECD, 1988).
Table 9.7. Sources of Patented Inventions for Use in Canadian Industries,

1978-1986
Type of Canadian U.S.
Industry Invention Share (%) Share (%)
Pulp and paper Product 8.5 42.4

Process 14.6 32.7
Nonferrous metal Product 8.5 37.2
Smelting and refining Process 15.5 38.0
Crude petroleum and natural gas Process 13.0 72.0
Refined petroleum products Process 8.0 64.0
Electrical power systems Product 7.0 57.0
Communication Product 8.0 56.0
Source: Statistics Canada, Science and Technology Indicators, 1988 (1989b, Tables 4.9,4.10, and 4.13).
Imperial Oil (an Exxon affiliate) serves as an example of the role that a foreign
owned firm can play in a national innovation system. Imperial's R&D effort includes
an Exxon world research mandate in lubricating oils as well as research on domestic
problems including enhanced recovery of conventional oil, recovery of heavy oil, and
Arctic drilling (Clarke, 1980). Through the Exxon link foreign technology is brought
to bear on Canadian problems and Canadian technology is applied internationally.
There are several other concentrations of innovative activity in Canada that are
worthy of note. These are the aeronautics, telecommunications and information tech-
nology industries, and agriculture.
The aeronautics industry is a largely self-contained innovative system (De Bres-
son and Murray, 1984; Lambert, 1989). Among the major R&D performers are Pratt
and Whitney, which has a world mandate from its U. S. parent in small turbine
engines, Canadair and DeHavilland, which build airframes, and CAE Industries,
which makes flight simulators. Government participation in this industry has taken
the form of subsidization either directly through the Defence Industry Productivity
program or indirectly through losses absorbed during the period of government own-
ership of DeHavilland and Canadair and through procurement.
Pratt and Whitney provides another illustration of the contribution a foreign
owned firm can make to a national innovation system (De Bresson et al., 1990). In
contrast, the history of Canadair and DeHavilland is one of financial difficulty and
changing ownership. DeHavilland was formerly a successful builder of bush planes
and is now a less successful builder of short take-off and landing passenger aircraft. It
has gone from British to Canadian government to U. S. ownership. Canadair formerly
built jet fighters and now builds business jets. It has gone from U. S. to Canadian gov-
ernment to Canadian public ownership. The history of these companies illustrates the
opportunities and constraints associated with foreign ownership, the difficulties
encountered by firms based in small countries in competing in world markets, and the
problems posed by direct government involvement in proprietary (as opposed to
generic or industry-wide) R&D decisions (Nelson, 1982).
The telecommunications industry is dominated by Bell Canada enterprises,
which includes the Bell System, Northern Telecom, and Bell Northern Research.
There are, in addition, a large number of other companies, both foreign and domestic,
producing components, turnkey systems, and information systems (McNeil, 1986).
These companies are located in western Canada as well as in the eastern high-tech
agglomerations. Indeed, the largest R&D performer in western Canada is Novatel a
subsidiary of Telus, formerly Alberta Government Telephones (Donville, 1991). Can-
ada has been a major innovator in the use of satellites for telecommunications pur-
poses (Canadian Research, 1986, p. 46). Government research is regarded as having
been particularly important in this area (Dalpe, 1988).
Among the major participants in the information technology industry are the
Canadian affiliates of IBM and Digital Equipment. IBM does software R&D in Can-
ada for worldwide use. IBM is said to have been an important incubator. One of Can-
ada's largest and most successful information technology consulting firms, DMR, was
founded by former IBM employees (Pitts, 1990). There are also a large number of
smaller Canadian-owned software companies that have entered the market over the
past 20 years (Amesse et al., 1989; Ontario Premier's Council, 1988a, Ch. IX, X).
These firms tend to make specialized products and export a large fraction of their out-
put. Among the more important linking institutions are the institutes of computer
research at the universities of Waterloo and Toronto.
At the heart of the agricultural innovation system is the network of federal gov-
ernment experimental farms that was begun in 1886. Its early contributions were to
popularize summer-fallowing, a dryland farming technique, and to develop the "Mar-
quis" strain of early-ripening wheat (Fowke, 1946). Both innovations increased the
returns to prairie agriculture. The system has evolved to include government labora-
tories, agricultural colleges, and joint business-government R&D projects. The result-
ing crop and livestock improvements have yielded high rates of return (Nagy and Fur-
tan, 1984;Ulrichetal., 1986; Zentner and Peterson, 1984; WidneretaL, 1988;Dotto,
1987).
Innovation systems in Canada can be separated into two categories. The first cat-
egory is comprised of "traditional" innovation systems in the fields of energy, metal-
lurgy, forestry, and agriculture. These systems are oriented toward domestic process
improvements. The return to innovation in these sectors comes in the form of higher
rents to unique Canadian resources. Innovation in these sectors increases the profit-
ability of selling staples on world markets.
The second category is comprised of the emergent innovation systems in aero-
nautics, telecommunications, and information technology. These systems have more
tenuous links with Canada's natural resource endowment and more closely related to
its human resource endowment. They are oriented toward product innovation. The
return to innovation in these sectors comes from identifying and occupying niches in
world markets for differentiated products.
The two categories of innovation systems have the common characteristics of
supporting export activity and of being integrated with the innovation systems of other
countries.
CONCLUSIONS
Canada's system of innovation has been shaped by its position as a thinly populated,
resource-rich country adjacent to a very much larger but technologically and culturally
similar country. Canada has an open innovation system with international (largely
transborder) linkages dominating domestic linkages. Manifestations of the dominance
of international linkages include the following:
1. A significant fraction of Canadian high-tech entrepreneurs are foreign born or
trained or have foreign work experience.
2. Even in industries characterized by well-developed innovating organizations and
linkages the vast bulk of the technology in use continues to come from abroad.
3. User-supplier linkages frequently involve foreign organizations as users or sup-
pliers.
a. Foreign firms are frequently involved in the commercialization of technologies
developed in Canadian universities and research institutes.
b. Canadian firms in the aerospace, information technology, and telecommuni-
cations equipment industries generally make a large portion of their sales
abroad.
c. The capital goods used by most Canadian industries tend to be of foreign origin
or design.
4. A significant fraction of Canadian industrial R&D is performed by affiliates of for-
eign companies. Affiliate R&D may be limited to solving local process problems or
it may support either a world product mandate (Pratt and Whitney, General Elec-
tric, Du Pont) or a world research mandate (IBM, Esso, Xerox).
Innovation in Canada has focused historically on adding value to the domestic
natural resource base. It is in the resource sectors that innovation systems are the most
visible. There are a number of ex post rate of return studies that conclude that this
innovative effort has been highly successful. Most observers agree that the historic
approach has been appropriate at least as a first step.
The Canadian system has been criticized in a number of quarters for its perceived
failure to go beyond solving local process problems in the resource industries. It is
argued that by comparison with the Scandinavian countries, for example, the indige-
nous innovative capability of the engineering (machinery) industries supplying the
resource sector is quite meager. It is further argued that there has been little in the way
of resource-upgrading innovation. According to this view, the innovation system has
failed to reduce Canadian reliance on natural resources as has been the case in other
formerly resource-oriented countries.
These failings are attributed to a lack of indigenous innovative capability in man-
ufacturing especially in the machinery (engineering) industries. This may have
reduced the productivity of other elements of the innovation system. Government and
university research organizations have reported difficulties in generating local interest
in their work. Similarly, threshold high-technology firms have difficulty finding
domestic strategic partners.
Foreign ownership is said to be at the root of these problems. This raises the ques-
tion of the extent to which foreign ownership inhibits the development of a national
innovation system and, indeed, whether a national innovation system is necessarily
preferred to integration into an international system.
Foreign owned firms have contributed to the development of innovation systems
in Canada. They contribute toward the realization of agglomeration economies. They
have served as incubators. They have served as users and suppliers at least with respect
to local or small batch innovations. They have facilitated Canadian participation in
the U. S. system both as users and suppliers. Large market innovative capability has,
however, tended to be centralized usually in the United States. The question is whether
this locational choice reflects the underlying reality of a small market next to a large
market or whether it is a mere quirk of ownership.
There is reason to believe that a small country, even one in Canada's position,
does have some leverage (Steed, 1989). The emergence of a strong innovative capa-
bility in the telecommunications equipment industry began with the severance of the
technology cooperation agreement between Northern Electric (now Northern Tele-
com) and Western Electric in the United States. The wheel had to be reinvented and
in this case it was. Similarly, Hydro-Quebec has successfully used purchasing policy as
a means of encouraging the development of an engineering consulting industry in that
province. Ontario Hydro has sourced its engineering internally and this is said to be
one reason for the failure of this industry to develop in Ontario (Ontario Premier's
Council, 1988a, pp. 285-300).
A large and sustained local demand can be used strategically to create domestic
user-supplier innovative linkages. This type of policy is not without its costs even in
the limited area in which it may be applicable. The requisite ex ante comparison of
external benefits against the cost of local preference is all but impossible. The Cana-
dian experience with two world-scale proprietary technologies that were heavily sup-
ported by government (CANDU reactor, Lermer, 1987; Challenger business jet, Bor-
ins, 1986) militates in favor of caution in pursuing this type of policy.
The alternative is to attempt to make the best of foreign linkages that will, in any
event, continue to dominate the Canadian system. This would involve the following
types of policies:
1. The maintenance of an open door to foreign technological and entrepreneurial tal-
ent.
2. The maintenance of market conditions and infrastructure conducive to timely
domestic adoption of foreign technology.
3. The encouragement of domestic research-performing organizations to search for-
eign markets for compatible industrial partners where domestic partners are not
available.
REFERENCES Technology Advancement: An Overview. Ottawa:

Queen's Printer.
Canada, Supply and Services (1990). R&D Bulletin
Acheson, K. (1989). "Power-Steering the Canadian No. 204, March.
Automotive Industry." Journal of Economic Canada, Task Force on Federal Policies and Pro-
Behavior and Organization 11: 337-51. grams for Technology Development (1984). A
Amesse, F., Lamy, P., and Tahmil, B. (1989). "L'axe Report to the Honourable Edward Lumley, Min-
40/401: Notre 'Silicon Valley' du Nord." Gestion ister of Statefor Science and Technology. Ottawa:
14: 15-22. Ministry of Supply and Services.
Anderson, E., and Lundvall, B. (1988). "Smaii Charette, M., Henry, R., and Kaufman, B. (1986).
National Systems of Innovation Facing Techno- "The Evolution of Canadian Industrial Structure:
logical Revolutions: An Analytical Framework." An International Perspective." In D. G.
In C. Freeman and B. Lundvall (eds.), Small McFetridge (ed.), Canadian Industry in Transi-
Countries Facing the Technological Revolution, tion, 61-134. Toronto: University of Toronto
9-36. London: Pinter Publishers. Press.
Baldwin, J., and Gorecki, P. (1986). "Canada-U. S. Clarke, P. (1980). "Research: Imperial's Long His-
Productivity Differences in the Manufacturing tory and More of the Same." Imperial Oil Review
Sector: 1970-1979." In D. G. McFetridge (ed.), 64(6): 26-29.
Canadian Industry in Transition, 211-60. Clarke, T., and Reavely, J. (1988). "Problems Faced
Toronto: University of Toronto Press. by R&D Managers in Canadian Federal Govern-
Borins, S. (1986). Investments in Failure: Five Gov- ment Laboratories." R&D Management 18(1):
ernment Investments That Cost the Taxpayer Bil- 33-44.
lions. Toronto: Methuen. Clifford, E. (1990). "Investors Put Faith in Quadra
Britton, J., and Gertler, M. (1986). "Locational Per- Logic." Globe and Mail Report on Business May
spectives on Policies for Innovation." In J. 31, B9.
Dermer (ed.), Competitiveness through Technol- Dalpe, R. (1988). "Innovation and Technology Pol-
ogy: What Business Needs, 159-75. Toronto: icy in a Small Open Economy: The Canadian
Lexington. Case." In C. Freeman and B. Lundvall (eds.),
Canada, Royal Commission on the Economic Small Countries Facing the Technological Revo-
Union and Development Prospects for Canada lution, 250-61. London: Pinter Publishers.
(1985). Report Volume Two. Ottawa: Supply and De Bresson, C., and Lampel, J. (1985). "Bopmbar-
Services. dier's Mass Production of Snowmobiles: The
Canada, Senate Special Committee on Science Pol- Canadian Exception?" Scientia Canadiensis 29,
icy (1972). Targets and Strategies for the Seven- 133-49.
ties, Vol. 2. Ottawa: Queen's Printer. De Bresson, C., and Murray, B. (1984). "Innovation
Canada, Senate Standing Committee on National in Canada." New Westminister, Co-operative
Finance (1984). Federal Government Support for Research Unit on Science and Technology.
De Bresson, C, Niosi, J., Dalpe, R., and Winer, D. quences de la Decentralisation Regionale des
(1990). "Technological Capability, Innovation Activites de R-D. Quebec: Conseil de la Science et
and Linkages of Domestic and Foreign Owned de la Technologic.
Firms: The Case of the Aircraft Industry." Paper Lambert, D. (1989). "Analyse de 1'industrie aeron-
presented at a Conference on Foreign Investment, autique canadien et quebecoise: perspective de
Technology and Economic Growth, Ottawa, Sep- developpement." Cahiers de Recherche, 89-107.
tember 6-7. Montreal: Ecole des Hautes Etudes Commer-
Donville, C. (1991). "Novatel Pressured by Industry ciales.
Boom." Globe and Mail Report on Business Jan- Lermer, G. (1987). AECL:An Evaluation of a Crown
uary 7, Bl. Corporation as a Strategist in an Entrepreneurial,
Dotto, L. (1987). "Canada-A Canadian Success Global Scale Industry. Ottawa: Economic Coun-
Story." University of Toronto Alumni Magazine cil of Canada.
15: 16-18. Litvak, I. (1985). "Canadian Trade Associations and
Doutriaux, J. (1989). "High Technology Entrepre- the Promotion and Diffusion of Innovation."
neurship and Academia: Are They Compatible in Ottawa: Department of Regional Industrial
Canada." Ottawa: University of Ottawa, Faculty Expansion, Technological Innovation Studies
of Administration, Working Paper No. 89-9. Program, Research Report 100.
Easlon, S. (1988). Education in Canada. Vancouver: Litvak, I., and Maule, C. (1972). "A Study of Suc-
The Fraser Institute. cessful Technical Entrepreneurs in Canada."
Economic Council of Canada (1983). The Bottom Ottawa: Department of Industry, Trade and
Line: Trade, Technology and Income Growth. Commerce, Technological Innovation Studies
Ottawa: Supply and Services. Program Report No. 11.
Ergas, H. (1987). "The Importance of Technology Litvak, L, and Maule, C. (1980). "Entrepreneurial
Policy." In P. Dasgupta and P. Stoneman (eds.), Success or Failure - Ten Years Later: A Study of
Economic Policy and Technological Performance, 47 Technologically Oriented Enterprises."
51-96. Cambridge, England: Cambridge Univer- Ottawa: Department of Regional Industrial
sity Press. Expansion, Technological Innovation Studies
Faucher, P. (1989). "Politique d'achat et developpe- Program, Report No. 80.
ment technologique: le cas d'Hydro Quebec." In Litvak, L, and Maule, C. (1982). "Canadian Entre-
Le Marche public et le developpement technolo- preneurship and Innovation: Six Case Studies."
gique ait Quebec: Six rapports d'etude. Quebec: Ottawa: Department of Regional Industrial
Conseil de la Science et de la Technologic. Expansion, Technological Innovation Studies
Faucher, P., and Fitzgibbons, K.. (1990). "The Polit- Program, Report No. 86.
ical Economy of Electric Power Generation: Pro- Longo, F. (1989). "Some New Findings on the Costs,
curement Policy and Technological Development Returns and Organization of R&D in Ontario."
in Quebec, Ontario and British Columbia." Mimeo, Toronto.
Working Paper 90-03, CREDIT, University of Lukasiewicz, J. (1986). "Canada's Encounter with
Quebec at Montreal. High Speed Aeronautics." Technology and Cul-
Financial Post. October 26, 1989. ture 2T. 223-61.
Fowke, V. (1946). Canadian Agricultural Policy: Malecki, E. (1987). "The R&D Location Decision of
The Historical Pattern. Toronto: University of the Firm and 'Creative Regions'A Survey."
Toronto Press. Technovation 5: 205-22.
Freeman, C. (1982). The Economics of Industrial McFetridge, D., and Corvari, R. (1986). "Technol-
Innovation. London: Pinter Publishers. ogy Diffusion: A Survey of Canadian Evidence
Freeman, C. (1988). "Introduction." In C. Freeman and Public Policy Issues." In D. G. McFetridgc
and B. Lundvall (eds.), Small Countries Facing (ed.), Technological Change in Canadian Indus-
the Technological Revolution, 1-8. London: Pin- try. Toronto: University of Toronto Press.
ter Publishers. McMullan, W., and Melnyk, K. (1988). "University
Globerman, S. (1980). "Markets, Hierarchies and Innovation Centres and Academic Venture For-
Innovation." Journal of Economic Issues 14, mation." R&D Management 18(1): 5-12.
977-98. McNeil, R. (1986). "Telecommunications: A Cana-
Haytcr, R. (1988). Technology and the Canadian dian Overview." Canadian Research 19: 38-
Forest Products Industries. Ottawa: Science 46.
Council of Canada Background Study No. 54. Miller, R., and Blais, R. (1990). "Les Logiques de
Hutchison, W., Miller, P., Baird, N., and Bevelan- ITnnovation dans Six Secteurs Industriels." Ges-
der, D. (1987). R&D Links Between Firms and tion 15:9-16.
Universities: Six Case Studies. Ottawa: Science Miller, R., and Cote, M. (1987). Growing the Next
Council of Canada. Silicon Valley. Lexington: Lexington Books.
Julien, P., Carriere, J., and Hebert, L. (1988). "La dif- Nagy, J., and Furtan, H. (1984). "Economic Cost
fusion des nouvelles technologies dans trois sec- and Returns from Crop Development Research:
teurs industricls." Quebec: Conseil de la Science The Case of Rapeseed Breeding in Canada."
et Technologic. Canadian Journal of Agricultural Economics 1:1-
Lacroix, R., and Martin, F. (1987). Lcs Conse- 14.
Nelson, R. (1982). Government and Technological Individual Canadian Inventor. Ottawa: Cat. No.
Progress. New York: Pergamon. 88-510.
Nichol, L. (1985). "Spin-offs and New Firm For- Soete, L. (1988). "Technical Change and Interna-
mation: Entrepreneurship and High Technology tional Implications for Small Countries." In C.
in the Ottawa Area." Ottawa: Unpublished M. A. Freeman and B. Lundvall (eds.), Small Countries
Thesis, University of Ottawa, Department of Facing the Technological Revolution 98-112.
Geography. London: Pinter Publishers.
Ontario, Ministry of Industry, Trade and Technol- Smith, M. (1983). "The Role of Federal Laborato-
ogy (1988). A Commitment to Research and ries in the Technological Development of Cana-
Development: An Action Plan. Toronto: Queen's dian Industry." Journal of Canadian Studies 17:
Printer for Ontario. 10-19.
Ontario, Premier's Council (1988a). Competing in Statistics Canada (1988). The Provincial Research
the New Global Economy, Vol. 2. Toronto: Organizations. Ottawa: Cat. No. 88-201.
Queen's Printer for Ontario. Statistics Canada (1989a). "Survey of Manufactur-
Ontario, Premier's Council (1988a). Competing in ing Technologies 1989." Ottawa: Cat. No. ST-89-
the New Global Economy, Vol. 1. Toronto: 10.
Queen's Printer for Ontario. Statistics Canada (1989b). Science Indicators 1988.
Organization for Economic Co-operation and Ottawa: Cat. No. 88-201.
Development (1986). OECD Science and Tech- Statistics Canada (1989c). Industrial Research and
nology Indicators. Paris: OECD. Development Expenditures, 1980 to 1989.
Organization for Economic Co-operation and Ottawa: Cat. No. 88-001, Vol. 13, No. 3.
Development (1988). Innovation Policy: Western Steed, G. (1989). Not a Long Shot: Canadian Indus-
Provinces of Canada. Paris: OECD. trial Science and Technology Policy. Ottawa: Sci-
Organization for Economic Co-operation and ence Council of Canada.
Development (1989a). Major R&D Programs for Stuart-Haile, M. (1989). "Technology Transfer: The
Information Technology. Paris: OECD. Science Business." Canadian Research 22: 14-22.
Organization for Economic Co-operation and Ulrich, A., Furtan, H., and Schmitz, A. (1986).
Development (1989b). OECD Science and Tech- "Public and Private Returns from Joint Venture
nology Indicators. Paris: OECD. Research: An Example from Agriculture." The
Pavitt, K., and Patel, P. (1989). "Do Large Firms Quarterly Journal of Economics 101: 103-30.
Control the World's Technologies?" Mimeo, Uni- United Kingdom, Inland Revenue (1987). Fiscal
versity of Sussex, Science Policy Research Unit. Incentives for R&D Spending: An International
Pitts, G. (1990). "DMR's Success in Past Efforts Cre- Survey. London.
ates Bullish View of Newly Expanding European United States Department of Commerce (1989).
Market." Financial Post (June 6), 14. Manufacturing Technology 1988. Washington,
Polese, M., and Verrault, R. (1989). "Trade in Infor- D.C.
mation-Intensive Services: How and Why United States National Science Foundation (1989).
Regions Develop Export Advantages." Canadian National Patterns of R&D Resources. NSF 89-
Public Policy 15(4): 376-86. 308, Washington, D. C.
Restrictive Trade Practices Commission (1983). Vaillancourt, F., and Henriques, I. (1986). "The
Telecommunications in Canada///, Vertical Returns to University Schooling in Canada."
Integration. Ottawa. Canadian Public Policy 12: 449-58.
Rosenberg, N. (1979). "Technological Interdepen- Van Tulder, R., and Junne, G. (1988). European
dence in the American Economy." Technology Multinationals in Core Technologies. Chichester:
and Culture 20: 25-50. John Wiley.
Rosenberg, N. (1982). Inside The Black Box: Tech- Walsh, V. (1987). "Technology, Competitiveness
nology and Economics, Cambridge, England: and the Special Problems of Small Countries" STI
Cambridge University Press. Review (Summer) 81-133.
Safarian, A. E. (1985). "Government Control of For- Walsh, V. (1988). "Technology, Competitiveness
eign Business Investment." In J. Whalley (ed.), and the Special Problems of Small Countries: A
Domestic Policies and the International Economic Review." In C. Freeman and B. Lundvall (eds.),
Environment, pp. 7-56. Toronto: University of Small Countries Facing the Technological Revo-
Toronto Press. lution, 37-66. London: Pinter Publishers.
Science Council of Canada (1987a). "A Sectoral Widner, L., Fox, G., and Brinkman, G. L. (1988).
Approach to Innovation: The Case of the Forest "The Rate of Return to Agricultural Research in
Products Industries." Ottawa: Supply and Ser- a Small Country: The Case of Beef Cattle
vices. Research in Canada." Canadian Journal of Farm
Science Council of Canada (1987b). University- Economics 36: 23-36.
Industry Research Centres: An Interface between Zentner, R., and Peterson, W. (1984). "An Eco-
University and Industry. Ottawa: Supply and Ser- nomic Evaluation of Public Wheat Research and
vices. Extension Expenditures in Canada." Canadian
Seguin-Dulude, L., and Desranlau, C. (1989). The Journal of Agricultural Economics 32: 325-53.
10
The Australian Innovation System
ROBERT G.GREGORY
The distinctive characteristics of the Australian national innovation system are a low
level of science and technology expenditure, a high level of government involvement
in financing and undertaking research, a low level of private sector research and devel-
opment, and exceptionally high dependence on foreign technology. The system has
evolved in response to three dominant structural features of the economy, which can
be detected very early in Australian history and have changed little.
The first structural feature is the ability of a small population to produce high
living standards from the production of primary products. As an indication of these
living standards, it has been estimated that a young adult convict sent to Australia dur-
ing the first 40 years from the founding of the colony in 1788 could expect to live sub-
stantially longer than if he remained in the United Kingdom. The early wealth of the
colony was generated by domestic production of goods and services and the export of
wool, then the discovery of gold during the 1850s, and finally the export of wheat from
the 1870s. Each wave of new rural production continued the pattern of combining
large quantities of land, little labor, and simple technologies to further increase living
standards. By the beginning of the twentieth century GDP per capita was perhaps 10-
30% higher than the United States (Butlin, 1962). As a result of the efficient export of
primary products there was no obvious and important direct association between eco-
nomic development and the systematic application of new and sophisticated scientific
knowledge (Schedvin, 1987).'
The second structural feature is that the small population and high wages led to
small scale and simple manufacturing, which, as a result of high costs, found it difficult
to export and compete against imports. The manufacturing sector was directed almost
exclusively to the small home market, large segments were protected by tariffs, and
most technologies were imported. As a result, there was a low level of private sector
R&D.
The third feature is the high degree of government provision of business and
social services. Economic historians talk of "colonial socialism,"2 the evidence of
which can be found in many facets of Australian life (Butlin et al., 1982).3 Given the
dominance of small scale rural production, and the tradition of "colonial socialism"
a high degree of government involvement in research financing and performance was
to be expected.
These structural features of the economy have been remarkably stable over Aus-
324
THE AUSTRALIAN INNOVATION SYSTEM 325
tralia's 200 year history. At the end of the 1980s, about 80% of exports are still primary
products, approximately half from the rural and half from the mineral sector. GDP
per capita is still relatively high, although now about 10 to 15% less than the United
States. The manufacturing sector continues to be small scale, accounts for 16% of
GDP, and less than 10% of output is exportedmainly processed mineral and pri-
mary products. Manufacturing is protected by an average effective tariff rate of 19%
and continues to import most technology. The government remains important and
dominates tertiary education, R&D expenditure, and telecommunications and is a
large provider of transport and banking services.
The innovation system has worked reasonably well in that Australians still enjoy
high living standards. Nevertheless, attitudes and policy have been changing slowly
over the last decade. There is now a strong demand for adjustments, driven by a desire
to change the structure of the Australian economy, which is increasingly seen as one
too dependent on a small range of primary products, prone to balance of payment
problems, and not generating sufficiently fast growth rates of per capita income. The
demand for adjustment is focused on increasing private sector R&D, targeting gov-
ernment sector research more toward national priorities, increasing competition in the
research community, and expanding the education system.
The broad outline of the national innovations system is first discussed, with spe-
cial attention given to the role of government and the effects of industry structure on
R&D expenditure. Then the recent history of the manufacturing sector is used to
explain the strong desire to change the system, particular aspects of new policies and
directions are discussed and assessed, and we comment on the current fashion of link-
ing the national innovations system to the balance of payments. Concluding com-
ments are then presented.
THE NATIONAL INNOVATION SYSTEM AND THE GOVERNMENT-PRIVATE

SECTOR BALANCE
Most of our focus is on R&D expenditure. Although, as discussed in the introductory

chapter, this is only a small part of the innovation system that has been attracting sig-
nificant policy interest and the economic and political forces leading to a reassessment
of R&D policy are being felt in other parts of the national innovation system. Indeed,
a thorough discussion of R&D policy issues should encompass most of the important
conceptual issues that relate to other parts of the national innovation system.
Australia spends a relatively small fraction of GDP on R&D. Among the sub-
group of OECD nations listed in Table 10.1 Australia is ranked twelfth, just ahead of
New Zealand (column 1). If the same set of countries is ranked by government sector
R&D expenditure (Civil) Australia is fifth (column 2). The prime reason for Australia's
low R&D is private sector expenditure, which, as a fraction of GDP, is just over 20%
of large spenders (column 3). When classified by sector of performance the importance
of government is again evident with Australia being ranked third (column 4).
These international comparisons suggest a range of questions. Do the same influ-
ences explain each of the key facts; high government R&D, low private sector R&D,
and low R&D expenditure in aggregate? Does government crowd-out the private sec-
tor? Are Table 10.1 ratios appropriate for the structure of the Australian economy?
Table 10.1 R&D Expenditure as a Ratio of GDP: Selected OECD Countries (Latest Available Data)
R&D Expenditure Total % Source of R&D Funds Source of R&D Funds R&D Performance
ofGDP Government (Civil) % of GDP Business % of GDP Government % of GDP
(1) (2) (3) (4)
1 . Switzerland 2.88 i. Netherlands 1.05 1 .Switzerland 2.27 1 . France 0.57
2. Japan 2.87 2. Germany 0.83 2. Japan 1.96 2. New Zealand 0.57
3. Sweden 2.82 3. Norway 0.74 3. Germany 1.81 3. Australia 0.40
4. Germany 2.78 4. France 0.73 4. Sweden 1.77 4. Netherlands 0.38
5. United States 2.66 5. Australia 0.71 5. United States 1.26 5. United Kingdom 0.37
6. Netherlands 2.40 6. Sweden 0.65 6. Netherlands 1.21 6. Germany 0.36
7. United Kingdom 2.36 7. New Zealand 0.58 7. United Kingdom 1.17 7. United States 0.3!
8. France 2.29 8. Japan 0.57 8. France 0.93 8. Ireland 0.30
9. Norway 1.83 9. Canada 0.55 9. Norway 0.93 9. Japan 0.28
10. Finland 1.64 10. Ireland 0.55 10. Canada 0.58 10. Norway 0.27
1 1 . Canada 1.35 1 1 . Switzerland 0.54 1 1. Australia 0.45 1 1 . Canada 0.26
12. Australia 1.20 12. United States 0.51 12. Ireland 0.43 12. Switzerland 0.18
13. New Zealand 0.97 1 3. United Kingdom 0.36 13. New Zealand 0.39 13. Sweden 0.13
Source: "Science and Technology Budtget Statement, 1989-90," Australian Government Publishing Service, Canberra, 1989.
Why Is Government R&D Expenditure and Performance So High?
It has become commonplace to assert that government expenditure is too high and
private sector expenditure too low, as if government has crowded-out private sector
R&D. It is probably a mistake to link the two sectors in this way. First, even if crowd-
ing-out exists it is too small to explain low business sector R&D. A significant reallo-
cation of government expenditure to the private sector, say 30% (Table 10.1), is suf-
ficient to move the Australian government ranking from fifth to twelfth but would
change the private sector ranking by only one place. Second, the Australian economy
has a unique structure that might be expected to generate R&D that is above average
for government and below average for business. Indeed, a case can be made that R&D
expenditure and performance in both sectors are about right if OECD averages are
used as benchmarks.
Table 10.2 lists government R&D expenditure by socioeconomic objectives for
Australia and other OECD nations. The large differences are low Australian expen-
diture on energy research (Australia is energy abundant and, with the exception of
electricity generation, energy is produced by the private sector) and high expenditure
on agricultural research, which is about three times greater than the OECD average,
excluding Australia and New Zealand. If government R&D expenditure on agricul-
ture matched the OECD average then Australian government expenditure in total
would be ranked about sixth in Table 10.1 (column 4) rather than third. Australia's
position would not be exceptional.
Rural R&D expenditure by government is about half as large as all R&D under-
taken in the private sector. The sector produces 4-6% of GDP but accounts for 26%
of government R&D. At the state level, the sector accounts for 73 cents in every $1
spent (see Table 10.3).
Government dominance of rural R&D is quite exceptional; it accounts for 97%
of all expenditure. This dominance is to be expected if the community wishes to
achieve an optimal allocation of resources. Small scale rural production units make it
difficult for individual producers to finance R&D and to capture its benefits to the
exclusion of others. If the free market were left to determine rural R&D levels the
extent of market failure would be relatively great. Other important influences that
encourage government expenditure are the world dominance of Australian wool,
which enables the nation to capture most of the benefits of wool R&D,4 and the
unusual physical environment, which limits the scope for rural technology imports but
enables research benefits to be captured by Australian producers.
The major institution undertaking rural research is the Commonwealth Scientific
Table 10.2. Government Research Budgets by Socioeconomic Objective (1982 or Nearest

Year) (cents per $ 100 of GDP)
Advancement
Agriculture Industry Energy Environment Health Social of Knowledge Other
Australia 15.5 8,.4 3.3 1.8 3.0 2,.3 29. 2 3..7
OECD (excluding 4.2 9,.7 8.9 1.4 4.5 4 .4 31 .7 5,.8
Australia and
New Zealand)
Source.' OECD, Science Resources Newsletter, No. 7, Summer 1983.
Table 10.3. Government R&D Expenditure by Socioeconomic Objective, 1986-1987
Total Commonwealth State Universities
Socioeconomic Objective Percent Percent Percent Percent
National security 9 22
Economic development 45 55 80 24
Rural 26 24 73 10
Manufacturing 7 15 0 3
Community Welfare 22 9 17 30
Health 14 8 12
Advancement of knowledge 24 13 3 45
Total 100 100 100 100
Source: Australian Bureau of Statistics, "Research and Experimental Development," All Sector Summary, Australia 1986-
87, Cat. No. 8112.0, Table 9.
and Industrial Research Organization (CSIRO), which was created as an independent

statutory authority in 1949 and succeeded the former Council of Scientific and Indus-
trial Research (CSIR), established in 1926 to serve the research needs of the rural sec-
tor.5 The CSIRO budget accounts for 16% of Commonwealth government expendi-
ture on civil R&D, around 70% of rural research, 70% of mining research, and 10% of
all research in Australia. Research expenditure within CSIRO is about half of that of
Australian universities. Most funding is by block grant and CSIRO chooses research
priorities.6
Rather than finance virtually all rural research from general revenue, government
could provide an institutional framework to facilitate collection of private sector R&D
monies to prevent free rider problems. Recently government has begun to do this. For
wool, meat, wheat, and many other rural industries government now collects com-
pulsory industry levies, and in most instances, matches them $1 for $1 to an amount
equivalent to 0.5% of gross value of production. Although government is placing
increased importance on moving research closer to the market place, and encouraging
the private sector to finance and control more of their research needs, the sums of
money collected from the rural sector are still relatively small, $37 million in 1988-
1989.
The arguments for government financing R&D are different from arguments as
to where R&D should be performed. Why should government undertake research
rather than finance research in the private sector? There was little choice at the time
decisions were made to set up government research agencies. When CSIR was estab-
lished in 1926 there were few private researchers outside universities (the first Ph.D.
awarded by an Australian university occurred in 1946). Immediately after World War
II, when the Australian government began to plan for economic development, it could
have acted to foster private research centers but the tradition of "colonial socialism,"
the satisfactory experience of war-time planning, and the lack of a technologically
sophisticated private sector naturally led to the belief that government should take the
lead. Hence the creation of CSIRO, the Institute of Advanced Studies at the Australian
National University, and a number of government research organizations. Once cre-
ated, these institutions showed no tendency to use government contracts to foster pri-
vate R&D companies.
Australia therefore spends proportionately more government funds on R&D
than the OECD average because rural production and rural exports are more impor-
tant (Table 10.4). But should government expenditure be increased or reduced? The
structure of the economy is quite different from the typical OECD country, but over
the last 30 years rural sector dominance has declined. The rural export share has fallen
from around 85% during the early 1950s to around 35% at the end of the 1980s and
the GDP share from 15 to 4%. Assuming expenditure ratios were appropriate 20 or 30
years ago, these trends might be expected to lead to less rural R&D as a share of the
total and less government sector R&D.7 Neither has occurred.
Why Is Private Sector R&D Expenditure so Low?
If private R&D is classified by enterprise of product field then over 90% of expenditure
is in the manufacturing sector and the reasons for low private sector R&D are to be
found there. Among the more important reasons are the following.
First, the manufacturing sector is small. The sector peaked at 27% of GDP in
1968-1969 but since 1974-1975 has been contracting and is currently producing
slightly more than 16% of GDP. Australia is among those OECD countries with the
smallest manufacturing sector as a proportion of GDP.
Second, the typical manufacturing establishment is small scale and orientated
toward serving the domestic market. Manufacturing exports, mainly low level pro-
cessing of primary products, are less than 20% of manufacturing value added and most
technology needs are easily imported.
Third, the manufacturing structure is biased away from high technology. Most
OECD countries export m ore high technology products than they import (Table 10.5).
Australia, however, imports seven times more high technology products than it
exports.
Fourth, most large manufacturing companies are foreign multinationals that
have not undertaken large quantities of R&D within Australia. In 1982-1983, 32% of
manufacturing value added was under foreign control.
The influence of these factors can be grouped into the effects either of industry
structure or of R&D intensity within each industry. The Bureau of Industry Econom-
ics (1990) has compared Australian private sector R&D in 1981-1982 as a proportion
of manufacturing value added (0.8%) with the average of six OECD countries (4.5%).
Table 10.4. Export and Manufacturing Shares Selected Countries"

Manufacturing
Share of Exports
Export Share Food/Raw Share of Change
of GDP Manufacturing Materials Energy GDP 1968-1987
Country (%) (%) (%) (%) (%) (%)
Australia 16 34 46 20 17 -32
United Kingdom 24 84 9 8 21 -21
Germany 29 93 6 1 32 -14
France 31 78 19 2 22 -22
Italy 20 89 9 2 23 -16
United States 9 82 15 3 19 -30
Japan 11 98 1 0 29 -8
"Export shares, 1988; manufacturing shares, 1987.
Sources: National Accounts, Main Aggregates, Vol. 1, 1960-1988, OECD 1990 (export share ofGDP) and OECD Data Bases
(share of exports).
Table 10.5. Trade in Selected Technology Based Products,

OECD Ratios, 1983
Itatio of Expon s Per Capita
Country t o Imports Exports US $
1 . Japan '1.62 571

2. West Germany 1 .95 1133
3. Switzerland 1 .83 2426
4. United States 1 .25 340
5. Netherlands 1 .24 1313
6. Ireland 1 .21 986
7. Italy 1 .18 363
8. United Kingdom 1 .06 568
9. France 1.06 551
10. Sweden 11 .02 1060
1 1 . Belgium 1 .00 1159
12. New Zealand ().18 110
13. Australia ().15 73
Source: "Science and Technology Statement" 1985-86, Department of Science,

Australian Government Publishing Service, Canberra, 1986, Table 14.
It concludes that if Australia possessed the same manufacturing industry structure as

each of the six OECD countries R&D would increase to 1.2% on average (row 2),
accounting for 11% of the Australian-OECD gap (Table 10.6). R&D intensity within
industries is far more important, accounting for 60% of the gap, the remaining 29%
being due to interaction effects. Australian R&D intensity has increased since 1981-
1982 but the Australian-OECD gap is much the same.
There are differences of "within industry R&D intensity" between Australia and
the OECD. For medium- and low-technology industries Australian R&D intensity, at
current expenditure levels, is only slightly less than other OECD countries. The largest
gaps occur for high technology industries, dominated by foreign multinationals,
where, in almost every instance, Australian R&D is very much lower (DITAC-OECD,
1989).
Given the small manufacturing sector and industry structure, the low private sec-
tor expenditure on R&D may not be inappropriate, unless it is firmly believed that
Table 10.6. R&D Intensities and Industry Structure"

A B C D E F
Australia 0.8 0.8 0.8 1.5 1.5 1.5

Six OECD countries'1 4.5 1.2 3.0 5.3 2.0 3.0
Gap 3.7 0.4 2.2 3.8 0.5 1.5
"A, The ratio of actual R&D to value added 1981. B, Australian intensity, other countries industry struc-
ture 198 1. C, Australian industry structure, other countries research intensity 1981. D, Actual R&D to
value added 1986-1987 (1986-1987 Australia, 1985 OECD countries). E, Australian intensity, other
countries industry structure 1986-1987. F, Australian industry structure, other countries research inten-
sity 1986-1987.
^Japan, United States, West Germany, United Kingdom, Sweden, and Canada.
Source: "Manufacturing Investment," Bureau of Industry Economics Research Report No. 33, Austra-
lian Government Publishing Service, Canberra, 1990, Table 6.10.
R&D expenditure determines industry structure or that it is more important for mul-
tinationals to produce research in the host country rather than import technologies.
Changing Trends
Over the past two decades there have been a number of noticeable changes in Austra-
lian attitudes and the pattern of R&D expenditure. First, there has been a dip in R&D
spending from the mid-1970s until the early 1980s, most of which was concentrated
in the business sector where R&D expenditure, as a proportion of GDP, approxi-
mately halved (Table 10.7). The government sector also reduced expenditure margin-
ally, but localized to defense R&D.
Second, business sector expenditures increased from the mid-1980s and the gov-
ernment share began to fall. In 1978-1979 government contributed 77 cents in every
research dollar. By 1987-1988 the contribution had fallen to 60 cents.8
Third, irrespective of the source of finance, there is a feeling that R&D should be
more applied and nearer the product market. Block funding has been reduced and
directed grants increased. Research agencies have been encouraged to raise private sec-
tor money. The change has been particularly marked for CSIRO, which during the
early 1970s accounted for more than a quarter of the government research budget and
has now been reduced to 16.0%. The share of the CSIRO budget from nondirect appro-
priations has increased to 30% and to change incentives CSIRO is allowed to keep roy-
alties, which were previously paid to government general revenue. The same strategy
has been applied to other government research institutions.
THE RECENT HISTORY OF MANUFACTURING
Over 90% of private sector R&D expenditure classified by industry of product field
occurs in the manufacturing sector and so much current discussion and motivation
Table 10.7. Trends in Research and Development as a Ratio of GDP

Source of Funds Performance
Government CSIRO% of
Government Commonweatlh
Year GERD/GDP Business Total Civil Defense Agencies Unis R&D Support
1968-1969 1.33 0.47 0.81 0.66 0.15 0.53 0.31
1973-1974 1.21 0.39 0.79 0.67 0.12 0.50 0.32
1976-1977 1.02 0.22 0.77 0.67 0.10 0.48 0.29
1978-1979 1.00 0.20 0.77 0.69 0.08 0.44 0.32 27
1981-1982 1.00 0.23 0.75 0.68 0.07 0.46 0.31 29
1984-1985 1.14 0.31 0.78 0.71 0.07 0.45 0.34 23
1986-1987 1.25 0.45 0.77 0.71 0.06 0.44 0.36 20
1987-1988 1.20 0.45 0.72 0.66 0.06 0.40 0.35
1989-1990 16
Sources: Australian Science and Technology Data Brief Update, Department of Industry and Commerce, Australian Gov-
ernment Publishing Service, Canberra, June 1989. Science and Technology Budget Statement 1989-90, Australian Govern-
ment Publishing Service, Canberra, 1989.
for policies to increase private sector R&D stems from dissatisfaction with the perfor-
mance of manufacturing. To understand the dissatisfaction it is useful to begin at
World War II when the Japanese occupied land within a hundred miles of the Austra-
lian coast. The fact that the United Kingdom was unable to direct significant resources
to the defense of the Pacific, and fear of a possible future invasion, led to the decision
to pursue economic development with more vigor. The development program was to
be based on augmenting Australian endowments with immigrants to provide labor
and foreign companies to provide capital and technology. The major policy instru-
ments were a liberalized immigration program, including paid passages, and import
protection for companies that began Australian production.
The impact of the development strategy changed Australian society. Since 1945
more than half the population increase has been immigrants and by 1981 one-fifth of
the population were born overseas and 40% were either born overseas or a child of at
least one overseas born parent. A large proportion of immigrant labor, especially the
low skilled, was employed in the manufacturing sector behind high levels of tariffassis-
tance. Foreign investment responded to fast population growth and automatic tariff
protection. By 1968-1969 the average nominal tariff protecting manufacturing was
24% and the average effective tariff rate 35%.
A profile of manufacturing fostered by this strategy is presented in Table 10.8
where industries are grouped into quartiles according to the effective rate of assistance.
Industries in the highly protected quartile were producing output behind an average
effective tariff rate of 85%, which implied a net subsidy equivalent of the tariff that
exceeded the wage bill; 42% of the workforce was female, 42% born overseas, and aver-
age wage, labor productivity, and capital intensity was low. The second most highly
protected quartile produced behind an effective tariff rate of 48%. The postwar devel-
opment policy had fostered and developed a manufacturing sector that was unecon-
omic, fragmented, and a drain on national resources. By the late 1960s criticism of
protection policy had become widespread and there were demands for a new direction.
There was another factor lending support for a change in policy direction. It had
been believed, at least since the 1930s depression, that in the face of falling relative
prices rural exports would not grow at a pace sufficient to avoid balance of payments
crises. It was hoped that manufacturing would develop an export capability and fill the
balance of payments gap that would otherwise develop with rising living standards.
The development policy, however, failed to generate large increases in manufactured
exports, which were less than 7% of manufacturing turnover for 1968-1969, and
largely confined to processing of foodstuffs and basic minerals. Concern for the bal-
ance of payments, however, began to dissipate with the rapid development of mineral
exports, which increased their share of total exports from 11 to 34% over the period
1967-1968 to 1982-1983 (Table 10.9). It was widely believed that this new source of
exports would avoid future balance of payments deficits (Gregory, 1976).
Australia therefore embarked on a new industry policy of lower tariffs. Since 1973
the average tariff has been reduced by 50% but manufacturing has not prospered. Its
share of GDP has declined by 40% and since 1970 imports have increased from 17 to
25% of market supplies of manufactured goods. In addition, the share of manufac-
tured goods in total exports has not changed substantially and Australia has not shared
the experience, common among OECD countries, for import and export shares of
each manufacturing industry to increase. Furthermore, the balance of payments has
Table 10.8. Characteristics of Manufacturing Industries: Australia 1968-1969
Exports as a
Wage and Females as a Persons Employed Proportion of Imports Indirectly
Salary Proportion of Born Overseas as a Total Exports Allocated as a
Effective Averge Wage Share of Total Proportion of by Proportion of Total
Rate of Capital Labor Per Person Value Number Total Number Manufacturing Imports of
Quartile Assistance Intensity Productivity Employed Added Employed Employed Sector Manufactured Products
Ranking (%) ($) ($) ($) (%) (%) (%) (%) (%)
1 1 8581 6691 3030 45 21 31 24 1!
2 28 7916 6997 3348 48 17 38 38 32
3 48 4155 5973 3088 52 24 35 15 43
4 85 3622 4963 2717 55 42 42 22 14
Manufacturing 35 5774 6049 3023 50 27 37
sector
average
Source: Industries Assistance Commission, "Annual Report 1973-74," Australian Government Publishing Service, Canberra, 1974.
Table 10.9. Composition of Exports by Sector: 1967-1968 to
1987-1988
Agricultural Mineral Manufactures Services
(%) (%) (%) (%)
1967-1968 54 11 18 17
1972-1973 49 18 20 13
1977-1978 39 30 17 14
1982-1983 33 34 18 15
1987-1988 34 27 21 18
Source: "Australian Exports, Performance, Obstacles and Issues of Assistance,"
Australian Government Publishing Service, July 1989.
not moved into surplus, apart from a few years during the early 1970s, but seems to
be moving towards higher deficits (Figure 10.1).
In 1984-1985, the Australian terms of trade fell 20%, the nominal exchange rate
fell 40% and foreign debt continued to increase from 4 to 30% of GDP. Government
began to talk more seriously about the need for structural changes, manufacturing
development, and new manufacturing exports. It was generally believed that the real
exchange rate depreciation would be maintained and lead to a restructuring of the
Australian economy toward increased production for import replacement and
exports. As might be expected, given the lags in the system, the current account con-
tinued to deteriorate but the real exchange rate depreciation dissipated without a sub-
stantial increase in manufactured exports. Recently the current account has deterio-
rated further and on August 25, 1989, Moody's Investment Service downgraded its
country ceiling on the ratings of Australian long term debt from Aal to Aa2. In Octo-
ber 1989, Standard and Poors also downgraded Australian debt. Among industrialized
nations Australia's debt/GDP ratio is exceeded only by NZ, and no OECD country is
accumulating foreign debt as fast.
Against this depressed manufacturing environment and rapid accumulation of
Figure 10.1. Net overseas borrowing/GNP; 1962-1963 to 1988-1989.

debt more is being asked of the National Innovation System. We now turn to a dis-
cussion of some recent initiatives.
THE NATIONAL INNOVATION SYSTEM AND IMPORTANT ISSUES OF THE DAY
With the exception of tariff protection Australian governments have been reluctant to
intervene directly into industries. The dominance of tariff assistance is evident in
Table 10.10 where it accounts for 92% of assistance for outputs and inputs of the man-
ufacturing sector. There has, however, been a limited number of recent direct grants
to specific industries. One major grant has been a 20% computer bounty for Australian
computer production, research, design, and development, which is currently running
at $47 million per annum and has grown quickly from $13.2 million in 1985-1986.
Another grant is $ 150 million for R&D over 5 years to foreign owned motor vehicle
producers. In addition, in 1989, before the election, there was $64 million grant to
Kodaka leading exporter from the manufacturing sectornot to close its Austra-
lian plant.
There is also some generalized export assistance to Australian companies. In
1988-1989, the budget of Austrade for export information, promotion, and assistance
to small firms was $243 million. For motor vehicles, there is a specific program that
for each export dollar earned a dollar of motor vehicles or parts can be imported duty
free. This export subsidy can be substantial and the export of motor vehicle parts have
increased significantly.
As in most countries there has been some concern as to the shortage of venture
capital and the Australian response was to establish the Management Investment
Table 10.10. Assistance to Manufacturing, by Form: 1984-

1985, 1988-1989"
1984-1985 1988-19896
Assistance to outputs
Tariffs 8,994 10,880
Quantitative import restrictions 1,945 524
Bounties^ 214 174
Export incentives'7 133 176
Assistance to materials
Tariffs 3,760 4,470
Quantitative import restrictions 663 263
Excise taxes 43 110
"Figures are in current prices. The figures for assistance to outputs and materials
are, respectively, the sum of the gross subsidy equivalents and the tax on materials
for individual industries, classified according to form of assistance. The summa-
tion of these amounts across industries will exceed the total for the sector due to
the outputs of many industries being used by other industries as intermediate
inputs.
'Preliminary estimates.
c
The estimates presented in this table do not represent the actual bounty and
export incentive payments in each year. The estimates measure the assistance
afforded by the current rates of bounty and export incentives in each year using
fixed 1983-1984 production patterns.
Source: "Industries Assistance Commission, Annual Report 1988-89," Australian
Government Publishing Service, Canberra, 1989, Table A7.5.
Company Program in 1984. There are 11 Management Investment Companies
(MICs) that can raise a limited amount of capital, fully tax deductible in the year of
investment. The MICs raised $190 million in the early years of operation, but since
then the program has not expanded. Most of the MIC investment portfolio went to
activities in electronic equipment, robotics, computer software, scientific and techni-
cal services, and biomedical products.
There are three recent changes to the system that warrant special attention: the
150% R&D tax concession, the Partnerships for Development, and the expansion and
reorientation of the education system.
The 150% Research and Development Allowance
In May 1985 the government announced a 150% taxation concession for R&D expen-
diture. The concession is to operate until June 1993, after which it will be phased down
to 125%. At the present company tax rate of 39% the concession provides for an addi-
tional 19.5 cents of R&D expenditure to be recouped from normal tax liability.' Cur-
rent cost is estimated to be $200 million, about one-seventh as large as all private sector
R&D expenditure. For companies that cannot fully avail themselves of the concession
there is a Grants for Industrial Research and Development Scheme (GIRD) that can
grant up to 50% direct funding for R&D. Budget expenditure in 1989-1990 is esti-
mated at $35 million.
Before discussing the 150% tax concession in depth there is a need to provide a
little more background as to the structure of the manufacturing sector. Within man-
ufacturing 54.8% of R&D expenditure takes place within three industries that produce
3.1% of GDP: appliances and electrical equipment (ASIC 3351/3352), chemical,
petroleum, and coal products (ASIC 27), and motor vehicle and parts (ASIC 323).
Appliances and electrical equipment, which produces 0.3% of GDP, accounts for just
over one-quarter of all private sector R&D (Table 10.11). This concentration of R&D
illustrates the significance of industry structure. If appliance and electrical equipment
produced 1.0% of GDP rather than 0.3%, and R&D expenditure increased propor-
tionately, then the aggregate R&D/GDP ratio would increase 50% and Australia
would no longer rank so low on OECD league tables.
These three industries illustrate other important features of the economy. First,
a significant fraction of R&D is under foreign control. In motor vehicles foreign own-
ership is 70.4%, chemicals, petroleum, and coal is 64.2% and appliances and electrical
equipment is 23.4%. Consequently, effectiveness and desirability of R&D subsidies
and policies need to be assessed against the behavior of multinationals.
Second, these industries do not possess any clear comparative advantage. The
import share of market supplies in 1981-1982, for example, varied between 19.7% for
motor vehicles (although subject to import quotas) and 65.4% for appliances and elec-
trical equipment. Since then import shares have increased. The demand for products
of these three industries account for about one-third of all Australian imports. Another
indication of lack of comparative advantage is the small proportion of turnover
exported, ranging between 4 and 11 %, well below average for the manufacturing sec-
tor. The largest export market is New Zealand, illustrating their difficulty in exporting
further afield.
Third, with the exception of chemicals, petroleum, and coal products each indus-
Table 10.11. Research and Development Expenditure and Industry Characteristics of Three Major Industries
Import Share of
Share of Effective . i .0
Market i-
Supplies r-
Export _Share
01. of < Foreign
Share of Private GDP Tariff Rate ,,, T .0, . Major Export M arket Control of
<\S1C Sector R&D 1%^ '%) (%) urnover ( o) ^ Expert Share ) (%) R&D (%)
Code Definition (86/87) (86/87) (86/87) (68/69) (81/82) (68/69) (81/82) (81/82) (86/87)
3352/3351 Appliances and 30.6 0.3 30 37.9 65.4 3.0 6.0 NZ(22) 23
electrical
equipment
27 Chemical, 13.6 1.7 12 25.3 32.3 3.0 11.0 NZ(27) 64
petroleum, and
coal products
323 Motor vehicles and 10.6 1.1 86 21.6 19.7 3.0 4.0 NZ(41) 70
parts
Total manufacturing 90.0 19.0 19 17.0 25.4 8.0 12.0 JAP(14) 33
Sources: "Australian Trade Classified by Industry: 1968-69 to 1981-82," Industries Assistance Commission, Working Paper No. 85, Canberra (import share of market supplies, export share of turnover,
and major export market). "Annual Report 1986/87," Industries Assistance Commission, Australian Government Publishing Service, Canberra (effective tariff rate). "Foreign Control in Research and
Experimental Development Business Enterprises, Australia 1986-87," Australian Bureau of Statistics, Cat. No. 5330.0, Canberra (share of private sector R&D and foreign control R&D).
try is highly protected against imports. For motor vehicles, the current effective tariff
rate is 86%, a reduction from a peak of 250% when the industry was protected by
import quotas over the 1975-1988 period. For appliances and electrical equipment
the effective tariff rate is around 30%.
Each of these features suggest a number of conjectures as to the role of R&D as a
policy instrument.
How Does the Magnitude of the RAD Subsidy Compare with Other Subsidies for
Industry Development?
Governments provide subsidies for other economic activities primarily in the form of
import tariffs and quotas and it may be of interest to place the R&D subsidy against
those of other activities. We can see from Table 10.12 that the community subsidizes
transport equipment and textiles $0.48 and $0.68, respectively, for each $ 1 of the wage
and salary bill. Those who produce clothing and footwear receive a subsidy of $0.89.
At $0.19 it is clear that the government does not assist private sector R&D activity to
the same extent as some manufacturing activities. Indeed, if the data are disaggregated
further the tariff assistance to some activities such as cardigan and pullover production
is 15 times the subsidy rate of R&D.
Do Any Special Considerations Arise from the Fact the Most of the R&D Subsidy Is
Acquired by Industries with High Levels of Industry Assistance?
For an export industry without government assistance or externalities, facing fixed
world prices and undertaking R&D that reduces production costs, industry and com-
munity gains from R&D are identical. The cost reduction from R&D increases invest-
Table 10.12. Government Assistance to Various Activities

Assistance Per Dollar of
Wage and Salaries $ Net Subsidy Equivalent
ASIC Code Definition 1987-1988 $ Million 1986-1987
21 Food, beverages, and tobacco 0.13 448
23 Textiles 0.68 458
24 Clothing and footwear 0.89 1005
25 Wood, wood products, and furniture 0.34 435
26 Paper, paper products, and printing 0.24 573
27 Chemical, petroleum, and coal 0.32 462
products
28 Nonmetallic mineral products 0.09 85
29 Basic metal products 0.18 372
31 Fabricated metal products 0.30 595
32 Transport equipment 0.48 1188
33 Other machinery and equipment 0.33 884
34 Miscellaneous manufacturing 0.41 493
Total manufacturing 0.32 6997
1 50% Research and allowance concession" 0.19 170
"The estimate of the value of 150% taxation concession was calculated by assuming that all companies paid 39 cents in the
dollar as company tax and fully claimed the concession (0.19).
Sources: "Annual Report 1987-1988 Industries Assistance Commision," Table A9.5, A9.7. Australian Government Publish-
ing Service, Canberra, 1988. "Science and Technology Budget Statement 1989-1990." Australian Government Publishing
Service, Canberra, 1989.
ment profitability and national income. What is good for this industry is good for the
economy. This need not be so for an industry protected by import tariffs. This industry
is larger because of tariffs, and given no externalities, community resources are being
utilized inefficiently. In these circumstances a paradox arises.10 Although R&D expen-
diture may be worthwhile for the industry, because it reduces costs and increases prof-
its, the R&D will make the community worse off and depress national income. In
response to cost reductions, imports will reduce, the industry will expand and utilize
more resources uneconomically than before, the increased inefficient resource utili-
zation from the output expansion exceeding the resource cost reduction from the sup-
ply curve moving downward. The industry cost reductions, from the community
viewpoint, are a transfer of import revenue previously collected, some of which will
be used for inefficient resource utilization by the industry and some for greater eco-
nomic rents.
This argument seems to establish a strong prima facie case against R&D subsidies
for industries heavily protected against imports. Can anything be said therefore, in
favor of the subsidy going to highly protected industries? One possibility is that an
R&D subsidy, which increases the industry's ability to cope with lower assistance lev-
els, may be part of the political requirements of a tariff reduction strategy. However,
where subsidy effects are linked to matching tariff reductions there is no industry gain
from undertaking R&D and it is quickly evident that analysis of the interaction
between industry R&D and government tariff strategy can become very complicated.
Where an industry is protected by a substantial tariff, the case for an R&D subsidy
must rest on a much larger externality than for industries without a tariff. It may be
difficult to satisfy such a requirement in Australia since most highly protected indus-
tries are dominated by oligopolies that should internalize most R&D benefits and
thereby reduce the size of the externality. The case for R&D concessions seems much
stronger where industry assistance is low and industry structure is atomistic, as in the
rural export sector.
Do Any Special Considerations Arise from the Fact That Most of the R&D Subsidy
Is Acquired by Multinationals Rather Than Australian-Owned Companies?
Much of R&D expenditure is undertaken by foreign owned companies. This may pose
special problems. Suppose the concession fails to increase R&D expenditure. For an
Australian company the concession becomes an income transfer from the taxpayer, in
the form of company tax foregone, to the Australian shareholder who receives extra
profit equal to the concession. For a foreign company the income transfer is a gift from
Australian taxpayers to foreign citizens. This undesirable feature remains even when
the subsidy increases R&D expenditure. Furthermore, suppose laboratories are moved
to Australia to undertake work that would have been undertaken in the home country.
Does Australia gain anything from this? It is difficult to see what the externality could
be. There is not a well-developed private sector research capability in Australia to
enable research workers to carry ideas from one laboratory to another or even from
the laboratory to the factory floor."
Has the Subsidy Affected the Level of R&D Expenditure?
As indicated earlier the last two decades have been difficult for manufacturing. The
history of manufacturing profitability is evident in Figure 10.2, which presents indices
of gross operating surplus to wages, salaries, and supplements. Profitability declined
Figure 10.2. Profitability and business R&D expenditure. Source: Australian Bureau of Statis-
tics, Australian National Accounts Cat. No. 5204-0 (various issues).
gently during the 1960s then, in 1974-1975, suddenly falls to 60% of the 1968-1969
level and remains there for 9 years. Since 1982-1983 profitability has increased,
almost doubling in 8 years. We have also included an index of private sector R&D/
GDP. The two lines are remarkably close, which suggests that R&D expenditure is
associated with profitability.12 There are not many data points, but the association
between R&D expenditure and profitability was not disturbed to a significant degree
when the concession was introduced in 1985 (the last two data points). When R&D is
deflated by manufacturing value added the impression created changes considerably
and research intensity has increased toward the end of the period but the increase
began before the subsidy was introduced.
On the basis of a survey of concession respondents the Bureau of Industry Eco-
nomics estimates that 2 years after the beginning of the program, and for each $ 1 spent
on the subsidy, business increased R&D in real terms by 28 cents.13 This calculation
suggests that if the subsidy is judged worthwhile and the government is allocating its
budget correctly at the margin, then an extra research $1 spent in universities or
CSIRO is worth only 28 cents of the $1 spent in the private sector. If we assume that
the BIE estimate is correct (they suggest that it is an overestimate) and, on average,
manufacturing industries that receive the subsidy are protected by tariffs, then on the
basis of earlier analysis the gap between the sectors may be even wider. Since much of
R&D is undertaken by multinationals the gap will widen further and the return for the
concession may fall to around 15 cents, which suggests that to be a good use of public
money, additional private sector R&D has to be at least five or six times more valuable
than public sector research. This suggests either an inordinate premium to private sec-
tor R&D or government research is overexpanded.
Government Procurement and Partnerships for Development

In 1985-1986, federal and state governments purchased high-technology products at
a cost of $3.6 billion, mainly imported from multinationals. For some time govern-
ment has used this purchasing power, within an offsets program, as an industry devel-
opment instrument "to gain access to internal markets and marketing facilities of large
world corporations which dominate these products and leading edge technologies
which otherwise would be closed to Australian companies" (Australian Civil Offsets
Program, Annual Report, 1988-1989, p. 1). Under the Offsets Program major over-
seas suppliers trading with the Australian government must direct activities of tech-
nological significance to Australian industry, including manufacturing, export mar-
keting, and R&D.14
Aerospace, telecommunication, and computer industries account for around
90% of offsets. In response to dramatic increases in the import of information tech-
nology and some general dissatisfaction with the offset program, the government
introduced in 1986-1987 voluntary Partnerships for Development for the informa-
tion industry. A Partner for Development would be exempt from offset requirements
in return for an undertaking to increase R&D to 5% of total sales in Australia, to
achieve a 50% export-import ratio within 7 years, and to achieve an average Austra-
lian value added content of 70% for exports. Partners would have access to generally
available assistance programs to help them engage in product development and export
activities.
The aim of the scheme is to link Australian activities into the Partner's global
operations. Some idea of the potential effects of Partnerships is given in Table 10.13,
which lists prepartnership exports and R&D expenditure and an estimate of the
annual commitment 7 years hence.15 It appears as though the first 15 partners have
committed themselves to a 7-fold increase in exports and a 6-fold increase in R&D
expenditure.
The change in performance implied by these data is extraordinary. If commit-
ments eventuate, and a significant fraction of the increase in R&D would not have
occurred without the scheme, then this program will have a much greater impact on
R&D than the 150% concession. For example, the BIE estimated the 150% concession
increased R&D by $54 million in 1986-1987 (Bureau of Industry Economics, 1990).
The Partners are pledged to increase R&D by $220 million per annum. Another way
to illustrate the importance of this scheme is to place the data against those of Table
10.14, which indicates that 12 enterprises in Australia spend at least $10 million on
R&D. '6 If Partners meet their targets then, within 7 years, perhaps as many as another
10 companies will be added to the group, doubling the number of large R&D spenders
and adding 20% to current R&D expenditure.
These data raise a number of interesting questions. First, we know that each year
the 150% concession costs approximately $200 million and according to the BIE gen-
Table 10.13. Partners for Development

Exports R&D
SMillion SMillion
Prepartnership performance 178 46
target levels 7 years hence 1202 266
Increase (%) 675 578
Source: "Australian Civil Offsets Program, Annual Report 1988- i 989" Australian
Government Publishing Service, Canberra.
Table 10.14. R&D Expenditure by Size of 1986-1987 Business Enterprise
Share of Total
Sire of Expenditure Number of Enterprises BERD Expenditure
($000) (All Industries) Percent of Enterprises (All Industries) (%)
Less than 100 1692 57.7 5.5
100 to 499 899 30.7 16.5
500 to 999 183 6.2 10.6
1000 to 1999 109 3.7 12.6
2000 to 4999 67 2.2 17.4
5000 to 9999 20 0.7 12.2
10,000 or more 12 0.4 25.3
Total 2932 100.0 100.0
Source: "Research and Experimental Development, Business Enterprises, Australia," Australian Bureau of Statistics, Cat. No.
8104.0.
crates an extra $54 million of R&D. Each extra $1 of R&D carried out in the private
sector, in response to the program, costs the government at least $4. What is the eco-
nomic cost of the partnership program? It has minimum budgetary cost. Is this large
increase in R&D free?
There seems to be two responses to this question. First, as proponents of the
scheme have argued, it may be that Partners R&D can be located in most countries at
much the same cost and consequently the company may be willing to respond to a
mild amount of political pressure and relocate some of its R&D activities. Under these
circumstances, the additional resource cost to Australia is minimal, apart from the
education cost of research scientists and the subsidy. Indeed Australia has avoided
imposing on itself the offsets cost. Second, it may cost the company quite a lot to meet
these R&D commitments but less than the cost imposed by the Offsets program.
Under these circumstances political pressure alone is not sufficient for location of
R&D and the true motivating factor is the avoidance of the extra burden of the offset
costs.
Expanding and Reorientating the Education System

Since its election in 1983 the Labor Government has sought to change the education
system so that it is attuned better to serve the needs of the economy. There are a num-
ber of forces for change but the more important are a concern for increasing teenage
unemployment and a belief that the economic and education system was not suffi-
ciently encouraging skill formation.' 7 The government has actively pursued a policy
of increasing school participation and introduced an income support scheme for
school children in their final 2 years of high school, subject to a parents income and
asset test. About 35% of school enrollments over the age of 16 receive a payment which
can be as large as $58 to $102 per week, depending on whether the student lives at
home. The rise in school participation since 1983 has been dramatic; retention rates
to final year of high school have increased from 35 to 63% and teenage labor seeking
full-time employment has fallen from 10.9 to 5.1% of the cohort.
Policy has also been directed toward increasing enrollments in tertiary institu-
tions. Available places have been increased as have student subsidies. Currently 42%
of tertiary students are assisted by income support (Austudy), which at the full rate
varies between $69 and $243 per week. Income support for postgraduate students has
been increased to $ 12,000-$ 16,000 nontaxable per annum, which is about 60-80% of
average weekly earnings after tax.
There has been little analysis as to whether, on the basis of economic criteria, the
tertiary education system should be expanded. However, there must be some doubt as
to whether it is a good thing to expand production of a commodity that is priced so far
below cost.18 Is there any evidence that expansion of the system is a good idea? First,
we look at rates of returns to graduates taken from the Income Distribution Surveys
(Table 10.15). The real rate of return has been falling over the last two decades,
although it remains quite high. If additional students could earn these real rates of
return then expansion would seem to be warranted.
There are two qualifications. First, these rates of return have been calculated
before the expanded number of students have entered the labor force and refer to aver-
age rather than marginal returns. Second, a more appropriate calculation is the rate
of return at the margin, that is, what would be the average rate of return for additional
entrants to the system? To address this question we use 1986 Census Data.19 We
divided university graduates and high school leavers into quintiles according to their
income in 1986 and calculated a range of education returns for each quintile, on the
assumption that their actual 1986 relative income level would stay the same through
their working life, adjusted for cohort income growth. The alternative income stream
for additional university graduates is drawn from the average of each income quintile
of high school graduates who did not obtain tertiary qualifications. Thus, for each per-
son there are five rates of return (Table 10.16). For example, we have calculated a rate
of return for a person in the top income quintile of high school leavers, H1, on the
basis that if he or she had acquired a degree he or she would have been in the top quin-
tile of income earners among graduates, Dl, then the calculation is repeated on the
assumption that he or she would be in the second quintile, D2, and so on. If instead
of dropping out after high school and being in the top quintile of high school graduates
this person had completed a degree and received income in the top quintile of univer-
sity graduates the rate of return would be 6.99%.20 If the alternative income stream was
in the second highest quintile of university graduates the rate of return would be neg-
ative. The overwhelming impression created from these calculations is that for a uni-
Table 10.15. Real Pretax Rate of Return to a

Degree"
Year Males (%)
1968/1969 24.0
1973/1974 22.4
1978/1979 19.4
1981/1982 18.3
1985/1986 16.8
"Assumptions: pretax total income of full-time full-year
males, takes 3.3 years to complete degree, and only cost of
higher education included is the opportunity cost.
Source: Chia (1990), Income Distribution Surveys (various
issues), Australian Bureau of Statistics.
Table 10.16. Rate of Return to Higher Education Implied by Matching

Different Parts of the High-School and Degree-Holders' Income Distribution
(Posttax Income)
High School Dl (%) D2 (%) D3 (%) D4 (%) D5 (%)
HI 6.99 Negative Negative Negative Negative

H2 17.48 8.15 3.92 Negative Negative
H3 23.12 13.50 9.38 4.42 Negative
H4 29.04 18.94 14.63 9.78 Negative
H5 50.83 38.85 33.67 28.02 12.51
Source: Chin (1990).
versity degree to provide a positive rate of return someone drawn from the income
distribution of high school leavers must do as least as well, relative to degree holders,
as he or she does relative to high school leavers. Given that most students who do not
attend university are not qualified to enter it seems unlikely that those who gain addi-
tional places as a result of expansion will do as well as the average of all those who
enter. On economic grounds, therefore, the expansion of the system must be a doubt-
ful proposition. These calculations are obviously very crude but they do serve to high-
light the important principle that it is marginal rates of return that are important rather
than average returns.
Science undergraduates have maintained their share within the larger system but
there has been a large reduction in demand for science places and university entry
requirements have fallen considerably and are now among the lowest. At the post-
graduate level the number of science Ph.D.s and Masters degrees has fallen consider-
ably, primarily in response to poor job prospects. The proportion of Australian stu-
dents within the group has also fallen. In 1973 science accounted for 50% of Ph.D. and
higher degree graduations. By 1985 the proportion had fallen to 36% of the total. Engi-
neering Ph.D.s have also declined. The growth areas have been arts and education.
Some idea of the poor job prospects for Ph.D.s in science, and the importance of
the government sector as an employer, can be found from an examination of the des-
tinations of Ph.D. graduates (Table 10.17). For those physical scientists who graduated
Table 10.17. Distribution of Higher Degree Graduates by Field of Study,

1982
Employed in Australia
Government
Employed
Teaching Other Private Overseas
(%) (%) (%) (%)
Physical sciences 24.5 15.3 10.2 43.9

Biomedical sciences 30.6 22.6 8.1 29.8
Engineers 14.8 8.2 32.8 39.3
Humanities, education,
and economics 48.6 12.8 2.8 16.5
Source: Department of Industry, Technology and Commerce, "Measures of Science and Innovation,"
Australian Government Publishing Service, Canberra, 1987.
in 1982, 43.9% left for employment overseas. In biomedical sciences the proportion
was 29.8%. In both disciplines the government sector, including teaching, accounted
for 4 to 6 times more jobs than the private sector.
Despite clear market evidence of oversupply of scientists the view prevails among
science policy advisors and the government that the country needs more scientists.21
As this is obviously not true in current circumstancesin the sense that employment
prospects are poorthe only legitimate arguments must focus on future require-
ments. There are two strands to the arguments. The first is that scientists will be needed
for forthcoming structural changes in manufacturing. The second is that the average
age of science researchers in universities and research institutions is increasing and,
at the time that Australia will need to replace scientists, there will be a worldwide
shortage.
Will there be a shortage of scientists in the future? It all depends on government
policy as government employs 70% of scientists and thus effectively controls
demand.22 It also controls supply, as it determines the number of tertiary education
places. In response to low job opportunities and the turning away from science by pro-
spective graduate students the government has responded by introducing policies to
increase the supply of scientists and to provide them with employment. The Australian
Research Council has created new graduate scholarships and the government has
recently announced The Cooperative Research Centre Scheme, which is to create 50
new research centers with an average budget of $4 million each; $2 million to be sup-
plied by the government and $2 million by the participating institutions.23 With
respect to the private sector it is unlikely that the manufacturing sector will change
sufficiently to create shortages.
Finally, we look briefly at on-the-job training. Australian wage and salary earners
do not seem to have a very steep age earning's profile. For example, in 1981 in the
United States a 45-year-old male earned approximately 42% more than a 25-year-old
male. For Australia the gap was 13%. Earnings peak very early in Australia. After
adjusting for the education level of each cohort the gap between the two countries nar-
rows but is still substantial (Daly, 1991). The flat earning's profile has been interpreted
to suggest that on-the-job training is low in Australia and the government and union
movement have begun to address the issue. As from June 30, 1990 each corporation
with a payroll exceeding $200,000 will be required to spend 1 % of payroll on approved
training courses for personnel. The unions have also sought award wage restructuring
to pay greater rewards for skill.
THE NATIONAL INNOVATION SYSTEM AND THE BALANCE OF PAYMENTS
The speed at which foreign debt has recently increased in Australia, and some skep-
ticism as to the efficacy of macro-policy instruments, has led to a debate on the inter-
relationships between manufacturing, science and technology policy, and the balance
of payments. The debate focuses on whether governments should intervene more into
the economy to foster high technology and export orientated industries. A good exam-
ple of current discussions is the following, which is taken from a joint discussion paper
by the Department of Industry, Technology and Commerce (DITAC) and the Orga-
nization for Economic Cooperation and Development (1989).
6.1 Australia needs to restructure and revitalize its manufacturing and service
industries if it is to maintain its standard of living. The long term decline in
prices for agriculture and mineral commodities together with rising domes-
tic demand for imports of manufactures has led to a large and persistent
trade deficit. This has been particularly pronounced in high technology
goods. Improved export performance in high technology goods since 1981 /
82 has been overwhelmed by stronger growth in imports. Australian foreign
debt has grown rapidly over the last seven years, rising to 30 percent of GDP
in 1986-87, a level which threatens future economic growth.
6.2 Stabilizing Australia's foreign debt will require substantial structural change
in manufacturing industry. The existing structural profile is skewed towards
low technology and low growth industries compared to other OECD coun-
tries. At least until 1984 (the last year for which detailed data are available)
Australia was generally contracting in the high technology sector, where
other countries and world markets were expanding rapidly.
6.3 For the high technology industries in Australia to grow .. . will require
increased expenditure on R & D and technologically sophisticated machin-
ery, as well as innovations in organizational structure, marketing and dis-
tribution. (DITAC/OECD, 1989, p. 43)
Another example is taken from the 1990-1991 Science and Technology Budget
Statement (p. 1), which states,
The governments approach to realising the benefits of science and technology

has been to develop a positive strategic framework which recognizes that science
and technology are essential elements in addressing two of the major problems
facing Australia today, namely:
the need to develop a more internationally competitive economy based on
innovation in order to solve our balance of payments problems;
the need to integrate economic development with environmental consider-
ations to ensure that development is ecologically sustainable.
Should the recommendation for increased expenditure on R&D and the need to
foster development of high technology industries to improve export performance be
accepted? Consider the following well known identity:
where G is government expenditure, T government tax revenue, / private sector

investment, 5* private sector savings, M imports, and X exports. The DITAC/OECD
paper focuses on the (M X) gap as being determined by a poor industry structure
leading to a slow growth of exports (agricultural and mineral commodities) and a fast
growth of imports (particularly high technology manufactured goods). To change (M
- X) it advocates fostering technological development in export and import compet-
ing industries to improve competitiveness. Since Eq. (1) is an identity, it follows, if this
argument is correct and describes the only factors to be considered, that the right-hand
side of Eq. (1) determines the left-hand side, that is, government and private sector
savings and investment balances are determined by the economic performance of

export and import competing industries.
The Treasury has also been concerned about the growing (M X) gap and the
mounting foreign debt/GDP ratios and attributed them not to poor industry structure
and inadequate technology and research and development, which might directly affect
the right-hand side of Eq. (1), but to factors that determine the left-hand side, the bal-
ance between domestic savings and investment. The Treasury argues that excessive
government deficits, (G T), have been largely financed from overseas borrowings.
Hence, the current account deterioration during the early 1970s, and its failure to
improve, is not attributed to poor industry structure leading to increased imports and
a slow growth of exports but to government deficits creating economic conditions that
encourage imports and discourage exports. The change in industry structure is a nec-
essary part of financing the government deficit. Similarly, private net savings (/ S)
have fallen over the last decade and also contributed to the current account outcome
(M X) and deterioration of the performance of the manufacturing sector. From this
viewpoint, the change in international competitiveness and what appears to be an inef-
ficient industry structure is not a cause of our economic difficulties but a symptom.
The cause lies in savings and investment imbalances and, given governments desire
for economic growth, must be solved by increased savings [for a discussion of these
issues see Pitchford (1989), Forsyth (1990), and Gregory (1991)]. Forsyth (1990, p. 1)
puts the argument most forcefully:
Attempts to achieve a solution to the current account balance of payments prob-
lem by enhancing the competitiveness of Australian industry, increasing pro-
ductivity, fostering an export culture in manufacturing and other industries, and
industry restructuring are doomed to fail. This is because the outcome of the
current account of payments is not determined by how competitive Australian
industries are. Rather, the reverse causation is truein the main, the current
outcome determines the competitiveness of the country's industries.
Only reforms directed toward eliminating the savings and investment distor-
tions will reduce the deficit.
There is obviously a conflict between the pure structuralist view that (M X) is deter-
mined by industry structure and the pure macro view that (M X) outcomes are best
analyzed in terms of savings and investment decisions. Since Eq. (1) is an identity
either side may drive the other; the exogenous factors may lie elsewhere, or, at different
times, different parts of the identity may predominate.
This conflict lies at the heart of current attitudes toward government policy and
the national system of innovation. If there is a desire to improve the current account
deficit then those who emphasize factors that directly affect the left-hand side of Eq.
(1) are likely to recommend policies to increase the private savings ratio, reduce gov-
ernment deficits, and so on. This is the policy stance of the Treasury, Reserve Bank,
and main stream economic advice. Those who emphasize the right-hand side of the
identity are likely to recommend policies that directly encourage exports and discour-
age imports and tend not to see the industry structure as evolving from saving and
investment decisions within a general equilibrium model. These advisors often tend
not to be economists and are grouped around industry and science centers in the
bureaucracy.
As might be expected government policy has adopted strands of both ideas and
although there are a number of innovations in the national innovations system the
prevailing ethos has remained one of nonintervention into the activities of particular
industries and greater emphasis on increasing savings, especially those of the govern-
ment.
CONCLUDING REMARKS
There is currently an exceptional amount of questioning of the national innovation

system. The effects are evident in the changing finance of R&D between government
and private sectors, reforms of the tertiary education system, the encouragement of
scientific research, and a greater degree of direct government intervention into the pri-
vate sector through training requirements and the creation of Partnerships for Devel-
opment. The announced object of these changes is to improve the performance and
structure of the economy.24 When these changes are put against the evolution in the
economy, however, it is difficult to believe that they will have a noticeable effect at the
macro level, at least in the foreseeable future. Three examples will illustrate this point.
First, over the 6 years 1983-1989 Australia experienced the fastest job growth of
all OECD countries. A preliminary analysis of that job growth is given in Table 10.18.
The economy is evolving toward part-time and service sector workers. Those parts of
manufacturing involved in substantial R&D expenditure are shrinking as a share of
total employment. To date the private sector is not moving toward high technology
industries or generating exceptional demands for scientists, except in information
technology.
Second, it does seem strange that a country that has never had a comparative
advantage in high-technology manufacturing, and enjoys high standards of living, is
discussing solutions to balance of payments difficulties in terms of R&D expenditure
to stimulate high-technology manufacturing. Most export growth still appears to be
occurring around traditional exports and new industries of tourism and service exports
rather than high-technology manufactures. If manufacturing is to make a better con-
tribution to the balance of payments it is difficult to escape the conclusion that most
Table 10.18. Employment Population Changes 1983-1990

Share of Employment Population Change
Growth Rate Full Time Part Time
Aggregate E/P Full Part
p. a Time Time Male Female Male Female Male Female
1983/1985 1.59 60 40 30 70 28 34 3 38
1985/1988 1.18 25 75 4 96 -9 35 12 62
1988/1990 1.31 31 69 19 81 -11 39 29 46
Sources: Australian Bureau of Statistics, Labour Force Historical Summary 1966 to 1984. Cat. No. 6204.0, Canberra 1987.
Australian Bureau of Statistics, The Labour Force Australia, Cat No. 6203.0, various annual issues August 1985, August 1988,
August 1990.
of that contribution will be made on the import replacement side rather than in terms
of large increases in manufacturing exports.25 It may be best to think of the national
innovation system in terms of micro reforms, that is doing things better rather than
thinking of the system as an instrument of macro policy that will significantly change
the structure of the Australian economy.
Third, the concentration of R&D within a few firms and industries suggests that
R&D policy will not have sufficient leverage on the economy as a whole.
To conclude, the current shaking up of the national innovation system, as it
relates to the government sector, seems to be a good thing, although such judgments
are exceptionally difficult to make. The expectations that are held as to the level of
benefits to flow from the changes, however, do seem to be misplaced. There has been
a significant overselling of the economic benefits of a government-led science and tech-
nology push.
NOTES
Helpful comments have been received from E. Klug, M. Teubel, K. Pavitt, Bengt Ake Lundvall,
and S. Pearce.
1. It is often said that science and technology are somewhat alien to the Australian way of
life. For example, an OECD Science and Technology review team stated, "we were struck by
what seems to be a widespread Australian view of technology as in some sense external to
national life" (OECD, 1986, p. 13). This view can be explained to a large extent by Australian
history of economic development. However, it is interesting to note, by way of contrast, that
establishment of the first European settlement was a direct result of support given to a scientific
project proposed by the British Royal Society. Captain Cook was sent to the Pacific to observe
the transit of Venus and search for the southern continent. In response to a report from the bot-
anist on board, the British government settled New South Wales with convicts 18 years later.
The new environment did lead to some adaptations and inventions and the impact of new
technologies on rural products should not be underrated. For example, many European com-
mercial plants did not flourish in Australia. British wheat varieties were late maturing so that in
many areas summer droughts set in before the crop was ready for harvesting. William Farrer
solved this problem by crossing Indian and English wheat varieties in 1901. He also began
research that led to rust resistant wheat varieties. Native grasses were also not the best for grazing
and imported grasses did not prosper. Eventually scientists discovered the value of subterranean
clover and superphosphate.
Other well known innovations are John Ridley's grain stripper in 1842of which it is said
that four men could now do in one day the work of two men in a complete harvesting season
and R. B. Smith's stump jump plough that reduced the cost of clearing as small stumps could be
left in the ground to decompose. In 1884, H. V. McKay invented the combine harvester, which
could strip, thresh, and clean grain in one continuous operation. A successful company was set
up to manufacture and to become Australia's largest farm machine producer before being taken
over by the Canadian firm of Massey-Harris (Science and Technology in Australia 1977-78,
1978).
2. In their introductory chapter Butlin et al. (1982, p. 4) say, "we have attempted to bring
the interaction of public and private decision making to the forefront. In doing so, we have had
to recognize that much Western historical experience (and its related theory) has only limited
value to Australia because of the prominence of governments throughout virtually the whole of
Australian history in all decision making processes."
3. The first Labor government in the world was elected in the 1890s in the state of Queens-
land and at the Federal level in 1910. Old age and invalid pensions were introduced in 1908 and
1910, respectively, at the federal level, and even earlier in some states.
4. Of course some benefits are lost through terms of trade effects that deliver some of the
cost reduction to the foreign buyer by way of lower prices.
5. Before CSIR government research was primarily financed by state government depart-
ments of agriculture but they neglected pastoral research. Research was hampered by lack of
equipment, small projects, and demands of administration. In response to these factors the Com-
monwealth entered the research field. Rural policy, resource management, and some industry
research are supported in part by a number of research bureaus within the government sector.
These include the Australian Bureau of Agricultural and Resource Economics, the Bureau of
Industry Economics, and the Bureau of Transport and Communications Economics.
6. In 1983 the CSIRO budget was three quarters that of the German Max Planck Gesells-
chaft, which accounted for 1.8% of German R&D expenditure and twice as large as the German
Fraunhofergesellschaft, established specifically for the support of industry.
7. Of course, the correct calculation is to look at the rate of return to rural research. When-
ever this is done the return seems quite high.
8. There is a difficulty with these figures in that the 150% R&D allowance may have
changed accounting practices in the private sector to increase the proportion of expenditure eli-
gible for the concession and thus inflate private sector R&D expenditure relative to the past.
9. There are a number of difficult conceptual problems involved in calculating this figure.
Australia has now moved toward an imputation system and as the tax rates are not the same for
companies and individuals the calculation can become very tricky depending on whether the
company decisions are thought of as fully reflecting the tax liabilities of their shareholders.
10. This analysis applies to an industry that could not survive without protection against
imports. If part of the industry could exist without the tariff then the community could gain from
R&D-induced reductions in the supply curve, but the gain will be less than that measured from
the cost reduction from the industry viewpoint and less than that from equivalent profitable
R&D for an unassisted industry.
11. Many of the above issues are clearly illustrated by the Motor Vehicles and Components
Development Grants Scheme established in 1985. The scheme provided for up to $ 150 million
to be allocated over 5 years to assist the highly protected and foreign owned companiesFord,
General Motors Australia, Toyoto, Mitsubishi, and Nissanto develop new or substantially
improved vehicles and automotive components of local design. The scheme has subsequently
become less generous when the 150% allowance was introduced but to date $63 million has been
paid from government revenue. Similarly $2.4 million of the Computer bounty, in addition to
the 150% R&D allowance, was paid to IBM in 1989-1990.
12. This result is supported by interviews of Australian businessmen who usually indicate
that R&D expenditure is first to be reduced if profitability falls.
13. Doube and Deaton (1989, p. 86) examined the rate of growth of newspaper advertise-
ments for scientists before and after the introduction of the allowance and found no significant
break in trend. They conclude that "there is no clear evidence that the scheme has increased the
number of R&D jobs in Australia." This is perhaps a better test than looking at expenditure
trends that are subject to reclassification in response to the tax concession.
14. To incur an offset obligation the threshold contract with the government must be above
$2.5 million with an import content of more than 30%. The value of the offset obligation is 30%
of the imported content of the government contract.
15. The data provide only an estimate because commitments are conditional on actual
sales and in some instances the data are given as cumulative totals over the 7 years.
16. Although we do not know the current expenditure of individual partners it is clear that
the maximum number of partners that could be currently spending more than $10 million
is two.
17. Throughout the 1950s and 1960s Australia had one of the lowest unemployment rates
among OECD nations, averaging about 1.5% of the labor force and 5% for teenagers. The average
interrupted unemployment spell was about 5 weeks. During the 1970s, but particularly from
1974-1975 onward, unemployment began to increase and in 1983 peaked at 10% of the labor
force. Teenage unemployment exceeded 20% and the average interrupted unemployment spell
for all males increased to 43 weeks. Since 1983, strong employment growth has managed to
reduce unemployment only to 6% but the interrupted unemployment spell for adult males has
continued to increase and is currently 57 weeks. Over 1990-1991 unemployment will increase
again.
18. Between 1973 and 1988 university places were free but rationed. To help finance the
expansion a charge of about 20% of the course cost has been introduced for all students. It can
be paid concurrent with the course or a tax liability accepted, indexed for inflation, which must
be paid back at the rate of 2% of taxable income after the student's taxable income reaches aver-
age weekly full-time earnings. There are higher rates of repayment as income increases about
this level. This system appears to have had a minimal impact on the demand for places, which
are still rationed in most disciplines.
19. These data give a posttax rate of return of 9%.
20. For these data the average rate of return calculated after tax is around 9%.
21. There have been a large number of highly publicized changes in the science bureau-
cracy that seem designed to highlight the importance of the area. They are listed in various Sci-
ence and Technology statements by the Prime Minister, and the Minister for Science, Customs
and Small Business. The major changes include incorporating the Science and Technology Min-
istry in the economic portfolio of Industry Technology and Commerce, the creation of a Science
Council, chaired by the Prime Minister, and the appointment of a Chief Scientist who reports
directly to the Prime Minister. Each year there is a special Science and Technology Statement
listing the government new expenditure proposals on science.
22. Among Ph.D. graduates only about 10% take their first job in the private sector (Table
10.17).
23. It is hoped that these centers will bring researchers together and create scale economies
in discovery.
24. "Since the early 1980s the Governments Science and Technology policies have placed
an increasing emphasis on improving research agencies' responsiveness to national economic
objectives, particularly the need to promote a more open and internationally competitive econ-
omy and to facilitate structural adjustment" (P3.315 Budget Statements 1990-1991 Budget
Paper No. 1).
25. This judgment may need to be qualified. The government has been paying significant
subsidies for the export of motor vehicles and components as part of the Motor Vehicle Plan and
the industry has been responding. Aluminium exports have also been growing.
REFERENCES ^ Cambridge, England. Cambridge University

Press.
Butiin, N. G., Barnard, A., and Pincus, J. J. (1982).
Australian Export Performance; Obstacles and Government and Capitalism. Sydney: Allen and
Issues of Assistance. (1989). Report of the Com- Unwin.
mittee for Review oj Export Market Development Chia, T. T. (1990). "Returns to Higher Education in
Assistance. Canberra: Australian Government Australia." Ph.D. Thesis, Australian National
Publishing Service. University.
Bureau of Industry Economics. (1990). Manufactur- Daly, A. (1991). "The Labour Market in Three
ing Investment. Research Report No. 33. Can- Countries: An International Comparison of Rel-
berra: Australian Government Publishing Ser- ative Earnings in Australia, Great Britain and the
vice. United States of America," Ph.D. Thesis Austra-
Butlin, N. G. (1962). Australian Domestic Product Han National University.
Investment and Foreign Borrowing 1861-I938/ Department of Industry, Technology and Com-
merce and the OECD (1989). Industrial R & D Gregory, R. G. (1976). "Some Implications of the
Investment and Structural Change in A ustralian Growth of the Mineral Sector." Australian Jour-
Manufacturing. Canberra: Australian Govern- nal oj Agricultural Economics 20:71 -91.
ment Publishing Service. Gregory, R. G. (1991). "The Current Account and
Department of Industry, Technology and Com- Economic Policy." In The Pacific Economy:
merce (1990). Australian Civil Offsets Program, Growth and External Stability. Sydney: Allen and
Annual Report 1988-89. Canberra: Australian Unwin.
Government Publishing Service. OECD (1986). Review of National Science and Tech-
Doube, B. ML, and Deaton, C. B. (1989). "Has the nology Policy: Australia. Paris: OECD.
150 per cent Tax Incentive Scheme Created New Pitchford, J. D. (1989). "Does Australia Really have
Jobs in R & D?" Search 20(3). a Current Account Problem?" Policy Winter 2.
Forsyth, P. J. (1990). "Competitiveness, Microeco- Schedvin, C. B. (1987). Shaping Science and Indus-
nomic Reform and the Current Account Deficit." try. Sydney: Allen and Unwin.
Centre for Economic Policy Research, Discussion
Paper No. 228, March.
PART III
LOWER INCOME COUNTRIES

Part III is concerned with five countries, all with levels of gross domestic prod-
uct per capita significantly lower than those considered in Parts I and II, but
otherwise very diverse. There is considerable intragroup variation even in cur-
rent productivity and income levels, with Taiwan and Israel having (as of
1989) levels of GDP per capita roughly half of the mean of the high income
countries considered earlier, but with Korea, Argentina, and Brazil having lev-
els of GDP per capita less than half of those in Israel and Taiwan. With the
exception of Israel, the countries considered in Part III all are "large" in terms
of population, with Brazil being very large. However, because their income
levels are low, the countries of this group, with the possible exception of Brazil,
have relatively small internal markets for manufactured goods, thus putting
them in the same position as the countries considered in Part II.
The countries in this group have experienced very different rates and pat-
terns of economic growth in the period since 1960. Prior to World War II the
level of per capita income in Argentina was roughly comparable to the levels
of income in continental Europe. Korea, Taiwan, and Brazil had much lower
levels of productivity and income. Israel (which was born only after World
War II) obviously is a special case. Argentina has been marked by virtual eco-
nomic stagnation over much of the post-World War II period, and today her
productivity and income do not stand much higher than they did in 1960. Bra-
zil and Israel have seen periods of very rapid growth, but have grown quite
slowly in recent years. Korea and Taiwan have experienced extraordinarily
rapid and sustained growth since the 1960s.
In terms of gross investment as a fraction of GDP, Taiwan stands even
above Japan. Argentina has the lowest investment rate of all the countries con-
sidered in this study. The other countries are roughly in the middle.
The countries stand surprisingly close to each other in terms of manu-
facturing output as a fraction of GDP, and, interestingly, that fraction is
roughly comparable to most of the countries of group I. However, the coun-
tries differ significantly in terms of the nature of national ownership of man-
ufacturing companies, and in their success in generating and sustaining man-
ufacturing exports. There is very little direct foreign ownership of firms in
Korea and Taiwan. Foreign ownership is significantly higher in Brazil, and
higher still in Argentina. Korea and Taiwan have been very successful in gen-
erating manufacturing exports, and in both countries manufacturing exports
account for a very large share of their total exports. This also is the case for
Israel: however because Israel is such a small country one would expect that
exports would account for a large fraction of GDP. The exports of Brazil and
Argentina account for a significantly smaller fraction of GDP, and manufac-
turing exports account for half or less of total exports.
The countries differ significantly in human resource development. The
355
356 LOWER INCOME COUNTRIES
literacy rate in Argentina and Israel was, as of 1988, in the league with that in
the high income countries. Taiwan and Korea were rapidly closing the gap.
Brazil lagged behind. Regarding enrollment at the secondary school level
Korea, Taiwan, and Israel are at the levels of the high-income countries or
rapidly approaching them. Argentina is somewhat behind, and Brazil far
behind. A similar statement holds for enrollment in postsecondary education.
The stock of scientists and engineers as a fraction of the work force in
Israel may be the highest in the world. In Korea and Taiwan that ratio is sig-
nificantly lower than in the industrialized high-income countries, but is rap-
idly growing. In Brazil and Argentina the ratio is much lower. The ratio of
R&D to GNP in these countries roughly follows the pattern of the figures for
the stock of scientists and engineers as a fraction of the total population.
In Korea the fraction of total R&D that is privately funded is similar to
that in Japan. The figure for Taiwan is roughly comparable to that for France
and Italy, with government support of R&D significantly higher than in
Korea. In the other three countries by far the majority of total R&D is publicly
funded, but for different reasons. In Israel the reason is that a very large share
of total R&D is directed to military concerns. In Argentina and Brazil the rea-
son is that very little R&D goes on in the manufacturing sector, with most of
it at universities and laboratories that are largely funded by government (Table
III.l).
Table III.l. Comparison of the Five Lower Income Countries

Korea Taiwan Brazil Argentina Israel
GDP/capita, 1989 official 3600 8000 2160 2520 8650
exchange rates
Population (World 42.4 16.2 147.3 31.9 4.5
Development Report,
except Taiwan)
Average growth rate, GDP/ 6.8 6.5 3.6 0.0 2.7
hour average 1965-1988
Gross domestic investment/ 22.5 32 21.5 16.5 23
GDP average 1965-1988
Manufacturing output/ 32 29 31
GDP
Manufacturing exports/ 33 5 3 18
GDP
Total exports/GDP 42 55.7 10 10 32
Literacy rate 88 92 78 95 95
Secondary level enrollment 88 39 74 83
ratio
Third level enrollment ratio 36 11 39 34
Engineers as percent of 0.11 0.04 0.06 0.47
population
R&D/GNP 1.0 1.16 0.4 0.5 3.7
Private R&D/total R&D 81/29 40/60 20/80 8/92 22/78
Direct foreign investment/ 1.9 2.0 11.1 23.9
GDP
11
National System of Industrial Innovation:
Dynamics of Capability Building in Korea
LiNSU KIM
South Korea (hereafter Korea) has emerged as one of the fastest growing economy and
has been transformed from a subsistent agricultural country into a newly industrial
one during the past quarter century. As late as 1961, Korea suffered from almost all
difficulties facing most poor countries today. But beginning in 1962, the Korean econ-
omy grew at an average annual growth rate of almost 9%, raising GNP per capita in
current prices from $87 in 1962 to $4936 in 1989. Exports increased from a mere $40
million in 1953 to $62 billion in 1989. The share of manufactured goods in exports
increased from 14.3 to 95.0% during the same period. In the mid-1960s, Korea began
exporting textile, apparel, plywood, and other labor-intensive products. Ten years
later, ships, steel, and construction services from Korea challenged established sup-
pliers from the advanced countries. By the mid-1980s, computers, semiconductor
memory chips, video cassette recorders, and automobiles were added to the list of
Korea's major export items. By the end of the 1980s, Korea's R&D investment as a
fraction of GNP and the number of scientists and engineers as a fraction of the work
force were approaching the levels of some of the highly industrialized countries of
Europe. What made it possible for Korea to achieve such a phenomenal growth in
industrialization? Will Korea be able to sustain the growth in the future?
As a way to answer these questions, this chapter presents Korea's national system
related to industrial innovation, which influences to a great extent the direction and
pace of technological capability development that in turn lead to industrialization.
The main focus of this chapter will be concerned with describing the industrial inno-
vation system as of this new era in Korea's industrial history. It is, however, also
important to examine how the national system in the past has worked for Korea to
acquire technological capabilities that brought the country where it is today.
HOW KOREA GOT WHERE IT IS TODAY
Thirty-six years of Japanese colonialization left some industrial base, mostly in the
northern part. But political and economic vacuums and chaos caused by the fall of
Japanese colonial rule, arbitrary division of the nation into North and South, and
357
ensuing civil war, which all took place between 1945 and 1953, flattened Korea as a
nation with little left of its past and facing a bleak future.
Particularly, the Korean War destroyed the majority of industrial and infrastruc-
ture facilities Korea inherited from Japanese colonialization. When the Armistice was
signed in 1953, the net commodity product decreased by 26% compared to the figure
10 years before, decreasing net commodity product per capita by 44% (Kim and Roe-
mer, 1979). Ironically, the Korean War also served positively to a certain extent for
the subsequent economic development, having completely transformed a traditional
rigid society into a highly mobile one by forcing geographical mobility and contributed
to the rapid formation of basic skills among the male labor force, most of who served
compulsory military service.
Given virtually no domestic savings, about 70% of all reconstruction projects had
to be financed by foreign aid. American economic assistance averaged about 10% of
Korean GNP between 1953 and 1960, enabling the country to recover back to the
prewar period by the time the aid was completely phased out and Korea launched its
economic development program in the early 1960s.
Korea's rapid economic progress may be attributed to many economic, social,
and technical factors. The most important of all may be technological capabilities,
which is a combined outcome of various economic, social, and technical inputs. It is
the ability to make effective use of technical knowledge in production, investment, and
innovation. Technological capability enables one to assimilate, use, adapt, change, or
create technology and to develop new products and processes in response to a changing
economic environment. The process of acquiring technological capability is so diverse
and complex that it defies a simple analysis. But this section attempts to identify both
macro/general and micro factors that influenced the acquisition of technological capa-
bility in Korea.
Macro/General Factors
Macro/general factorsthe process of human resource formation, the inflow of for-
eign technologies, the government's industrial, trade, and science/technology policies
that shaped industrial structure and the direction of growth, and so onset the stage
for the rapid acquisition of technological capability.
First, human resource development may be a most basic and crucial determinant,
as technological capability is embodied in people. Korea's greatest resource is its
human resources. Modern education was first introduced to Korea by American mis-
sionaries and later expanded by the Japanese colonial government. But at the end of
the Japanese colonial rule, only 2% of the Korean population over 14 years of age had
completed secondary school, and the illiteracy rate stood at 78%.
The importance of education in the Korean government's modernization strat-
egy can be seen in the growth of government investment in education: the share of
education in the total government budget, for instance, rose from 2.5% in 1951 to over
22% in the 1980s. Government expenditures, however, accounted for only one-third
of the total expenditures in education, the remainder being borne by the private sector
and parents, reflecting the high commitment for education held by the Korean society.
This commitment was the strongest in Korea among eight industrialized (Denmark,
Germany, Italy, Japan, Sweden, Switzerland, the United Kingdom, and the United
CAPABILITY BUILDING IN KOREA 359
States) and two semiindustrialized (Singapore and Korea) countries Porter (1990)
studied.
As a result, enrollment in the various levels of the formal education system has
increased rapidly since 1953. In the case of elementary school, enrollment has
increased by over five times. Even faster growth is seen in the secondary and tertiary
education levels. Secondary school enrollment increased by more than 28.5 times
between 1945 and 1980, while enrollment in colleges and universities increased by
almost 150 times during the same period. Table 11.1 shows that school enrollment as
a percentage of the corresponding age group rose to over 100% in 1970 for the ele-
mentary school level, to over 95% in the early 1980s for the middle school level, to
over 70% for the high school level, and over 25% for the tertiary level by 1980.
Several other developing countries also attained as rapid growth in elementary
education as did Korea. But what was unique in Korea was the well-balanced expan-
sion in all levels of education early enough to support its economic development. Har-
bison and Myers (1964) show that given GNP per capita from a regression line of 73
developing countries, Korea's level of educational achievement was furthest of that
predicted by its level of economic development. That is, with a per capita income of
$90, Korea's educational achievement stood fairly close to the normal pattern of
human resource development for a country with a mean per capita GNP of $200.
Korea's level of human resource development, when its per capita GNP reached $ 107,
was equivalent to that of countries with a GNP per capita of $380.
Overseas training and observation were also characteristic of human resource
development in Korea. Overseas training traces back to the foreign aid period in the
1950s, when a high proportion of senior personnel in government, business, and aca-
demia were exposed to foreign training, mainly in the U. S., under economic assistance
Table 11.1. Indicators of Human Resource Development in Korea

1953 1960 1970 1980 1987
Literacy rate (%) 22,0 72.1 89.4 a a

Enrollment as percentage of
corresponding age group
Elementary school 59.6 86.2 102.8 101.0 100.2
(ages 6-11)
Middle School (ages 21.1 33.3 53.3 94.6 98.8
12-14)
High School (ages 15- 12.4 19.9 29.3 68.5 82.8
17)
College/university 3.1 6.4 9.3 14.9 25.5
Graduates of vocational NA NA 31.6 104.5 50.9
training centers (1000)
Number of graduates from NA 10* 11 27 54
tertiary schools (per
10,000 population)
Number of scientists and 4,157 16,436 65,687 174,832 361,330
engineers'
"Illiteracy rate after the mid-1970s was so insignificant that the government ceased to collect data on it.
*For 1965.
'Cumulative number of science and engineering graduates since 1945.
Source: Ministry of Education and McGinn et al. (1980).
programs (Mason et al., 1980). The tradition of overseas training continues to this day.
The ratio of third-level students training abroad to all third-level students has been
twice as great in Korea as in Argentina, Brazil, and India and higher than in Mexico
(Westphal et al., 1985). As a result, Korea surpassed the aforementioned newly indus-
trializing countries by almost all indices of educational attainment. The rapid expan-
sion of education resulted in serious unemployment problems for the educated in the
1950s and 1960s, but the formation of educated human resources, despite low per
capita income, laid an important foundation for the subsequent adaptation of
imported technology and development of indigenous technology. The rapid expan-
sion of the Korean economy soon absorbed the surplus of educated workers.
Second, Korean's hard working habits in long hours may be another factor
accounting for the fast acquisition of technological capability. Korea's land area is
scarcely larger than New Jersey or Hungary, crossed by mountains with a relatively
small portion of the land arable for more than 40 million people. In terms of popula-
tion density, Korea trails only Bangladesh and Taiwan. The cramped conditions and
severely cold winters appear to have forced Koreans to work hard and long whenever
possible in order to survive in an unfavorable environment. The average working
hours per week in manufacturing in 1985 was 53.8 hours in Korea compared to 33.1
to 42.9 in OECD countries including Japan, 44 to 48.0 hours in other Asian NICs, and
46.0 hours in Mexico (Kim, 1988). Besides, the older generation is motivated by the
memory of deprivation and hard times under Japanese occupation and during the
Korean War. A sense of Geug-il or "beat Japan" to settle old scores and national eco-
nomic competition with North Korea are also major forces motivating Koreans (Por-
ter, 1990).
Third, lacking technological capability at the outset of its economic development,
Korea had to rely on foreign technology imports. However, Korea's policies concern-
ing direct foreign investment (DPI) and foreign licensing (FL) were quite restrictive in
the early years of industrialization, when technology was not a critical element and the
mature technologies needed could be easily acquired through mechanisms other than
DPI or FL (e.g., reverse engineering). Table 11.2 presents data on technology transfer
through DPI and FL to Korea.
Consequently, the size of DPI and its proportion to total external borrowi ng were
significantly lower in Korea than in other NICs. For example, Korea's stock of DPI in
1983 was only 7% of the size of that in Brazil, 23% of that in Singapore, and less than
a half the size of that in Taiwan and Hong Kong. The proportion of DPI to total exter-
nal borrowing was only 6.1% in Korea compared with 91.9% in Singapore, 45% in
Taiwan, and 21.8% in Brazil (KEB, 1987). The comparative figure reflects Korea's
explicit policy of promoting its "independence" from multinationals in management
control.
In contrast, Korea promoted technology transfer through the procurement of
turnkey plants in the early years. For example, the chemical, cement, steel, and paper
industries, established in the 1960s and early 1970s, all resorted to the turnkey mode
for their initial setup. But Korean firms assimilated imported technologies so rapidly
that they managed to undertake subsequent expansions and upgrading with little assis-
tance from foreigners. The rapid growth of the Korean economy has called for com-
mensurate growth in investment for production facilities. Government policy had,
however, been biased in favor of the importation of foreign capital goods and against
Table 11.2. Foreign Technology Transfer to Korea (US$ Million)
Source 62-66 67-71 72-76 77-81 82-86 Total
1 . Direct foreign investment (DPI)
Japan 8.3 89.7 627.1 300.9 875.2 1,901.2
United States 25.0 95.3 135.0 235.7 581.6 1,072.6
All others 12.1 33.6 117.3 184.0 309.7 658.7
Total 45.4 218.6 879.4 720.6 1,766.5 3,632.5
2. Foreign licensing (FL)
Japan 5.0 58.7 139.8 323.7 527.2
United States 0.6 7.8 21.3 159.2 602.7 791.6
All others 0.2 3.5 16.6 152.4 258.5 431.1
Total 0.8 16.3 96.5 451.4 1,184.9 1,749.9
3. Technical consultancy
_
Japan 12.1 7.7 20.8 89.2 129.8
United States 3.1 6.0 16.7 159.1 184.9
All others 1.6 4.8 17.2 84.0 107.6
Total 16.8 18.5 54.7 332.3 422.3
4. Capital goods imports
Japan 148 1,292 4,423 14,269 20,986 41,118
United States 75 472 1,973 6,219 12,394 21,133
All others 93 111 2,445 7,490 53,338 64,143
Total 316 2,541 8,841 27,978 86,718 126,394
Source: Ministry of Finance on DPI and FL; Ministry of Science and Technology for technical consultancy data; and Korean
Society for Advancement of Machinery Industry for capital goods data.
the development of a rudimentary local capital goods industry as a way to strengthen

international competitiveness of the capital goods user industries. Such a policy led to
massive imports of foreign capital goods, which became a major source of learning
through reverse engineering by Korean firms (Kim and Kim, 1985). Of the four cat-
egories of technology transfer listed in Table 11.2, capital goods imports far surpass
other means of technology transfer in terms of value. Capital goods imports were
worth 34 times the value of DPI and 72 times the value of FL. The total value of capital
goods imports is 21 times that of the other two categories combined, suggesting that
Korea may have acquired more technology from advanced countries through the
importation of capital goods than through any other means. Among NICs, the pro-
portion of capital goods imports to total technology transfer was highest in Korea
(Westphal et al., 1985).
In short, Korea restricted DPI but promoted instead technology transfer through
other means such as capital goods imports. Capital was acquired in the form of foreign
loans. Such a policy, designed to maintain Korea's independence from foreign mul-
tinationals, has been relatively effective in acquiring technological capability in the
mature industries, as Korea's well-trained human resources and their entrepreneur-
ship enabled the country to learn quickly from foreign capital goods.
Fourth, continued relations with the United States in the national defense has
served an important role in enabling Korean firms to accumulate experience and capa-
bility. The U. S. military procurement program afforded many producers in a number
of sectors with occasions for assisted learning by doing to meet exacting product spec-
ifications. Construction contractors, plywood producers, and tire makers, among oth-
ers, built up their initial capability through the U. S. military contracts and later
became major exporters (Westphal et al., 1985). For example, through U. S. military
contracts, local construction firms learned, under the direction of the U. S. Engineer-
ing Corps, specific details of stringent U. S. specifications, the preparation of bidding
documents, and management of construction projects, enabling local firms to incre-
mentally accumulate increasingly more sophisticated construction capabilities. By the
early 1970s, Korean construction firms became important contractors for the U. S.
military bases in Guam and Vietnam. Capabilities accumulated in these foreign coun-
tries led Koreans to launch an aggressive expansion to the Middle East in the wake of
oil crises. The U. S. military presence also reduced the defense expenditure burden of
the Korean economy, enabling the country to divert more resources for industrial
development.
Fifth, despite a strong backup of the U. S. military presence in Korea, Korea still
had to maintain one of the largest military forces in the Free World by allocating over
36% of its annual budget to national defense. Such a national security concern, par-
ticularly during and after the Carter Administration, which expressed a possibility of
the U. S. military pullout from Korea, had some impact on the development of par-
ticular industries and their technological capabilities. For example, one of the major
factors the Korean government has considered in pushing hard heavy machinery
industries in the late 1970s appears to be its concern to acquire capability for defense
production. Although such an ambitious investment in heavy machinery industries in
the absence of adequate local capability has resulted at the outset in extremely low
capacity utilization and resource allocation problems, the investment over local mar-
ket and local capability has put enormous pressure on both the government and pri-
vate entrepreneurs to upgrade technological capability for survival. As a result, Korea
became an important contender in exporting large scale plants in the Third World.
The major impact of national security concern in technological capability building
may be the significant improvement in precision in the machinery industry. The pro-
curement demand of the Korean military after the Korean War also helped local con-
sumer electronics producers to diversify into industrial electronics.
Sixth, to overcome a disadvantage of a small-sized domestic market and to take
advantage of the stable nature of mature technologies on which industrialization strat-
egy was based, the Korean government designated so-called "strategic" industries for
import substitution and export promotion. Plywood, textiles, consumer electronics,
and automobiles in the 1960s and steel, shipbuilding, construction services, and
machinery in the 1970s are examples. These industries, which were created in viola-
tion of its static comparative advantage, had to suffer from high costs in addition to
"infant industry" growing pains. To enable the industries to overcome the problems,
the government used various policy mechanisms; assistance to mitigate risk and to
secure stability, market protection from foreign competition, export promotion, and
tax incentives and preferential financing for products and producers designated by the
government. The average effective rate of protection was atypically high for the "stra-
tegic" industries. In some industries, protection quickly became redundant with firms
experiencing a rapid rite depassage from an infant to an exporter, but in others, where
technology was complex and marketing was more elaborate, protection had been rel-
atively long lasting in order to provide a longer period of incubation (Kim, 1987).
Seventh, in contrast to Taiwan, which restricted the emergence of large enter-
prises, the Korean government intentionally created large firms, chaebols, as an instru-
ment to bring about the economy of scale in mature technologies and in turn to
develop these "strategic industries" and to lead exports and economy. The govern-
ment helped the capital formation as well as the subsequent diversification of these
chaebols. Unlike Taiwan, which did not privatize Japanese enterprises confiscated
after World War II (Hou and Gee, 1990), the Korean government sold Japanese colo-
nial properties and state-owned enterprises to selected local entrepreneurs with a
favorable long-term installment payment during the inflationary period, giving local
entrepreneurs a windfall capital gain. Owning all commercial banks, the government
helped more significantly the formation of chaebols by allocating scarce foreign
exchange and preferential financing at the official rates both of which were only one-
half the real market rate (Jung, 1989). The state also gave them large import-substi-
tution projects, for which these entrepreneurs imported production technology on
turnkey basis with foreign loans, which was guaranteed by the government. Then, for-
eign debt burdens due to currency devaluations were compensated with increased low
interest loans, reducing market risks for these businesses.
Where Korea differs from other developing countries in promoting big businesses
was in the discipline its state exercised over these chaebols by penalizing poor perform-
ers and rewarding only good ones. Good performers were rewarded with new licenses
in other industries, leading to further diversification. For those entering risky indus-
tries, the government rewarded entrants with licenses in more lucrative sectors, pro-
viding a cushion to nurture risky infant industries. In contrast, the government refused
to bail out relatively large scale, badly managed, bankrupt firms in otherwise healthy
industries, instead selected better managed chaebols to take over them (Amsden,
1989). The government as the controller of commercial banks was in a powerful posi-
tion to punish poorly managed firms by freezing bank credits. As a result, only three
of the largest 10 chaebols in 1965Samsung, Lucky-Goldstar, and Ssangyong
remained on the same list 10 years later. Similarly, seven of the largest 10 in 1975
remained on the same list in 1985.
Chaebols' rapid growth and diversification have enormously affected industrial
structure and market concentration in Korea. By 1977, 93% of all commodities, or
62% of all shipments, were produced under monopoly, duopoly, or oligopoly condi-
tions, under which the top three producers accounted for more than 60% market share.
The 10 largest chaebols accounted for 48.1% of GNP in 1980, indicating that Korean
industry was even more highly concentrated than that of Taiwan or Japan. However,
Chenery et al. (1986) concluded that total factor productivity as well as output grew
faster in Korea's highly concentrated economy than in that of almost any other coun-
try they studied.
Eighth, the Korean government also set forth exports as something of a life or
death struggle to achieve economic growth goals with the small domestic market.
Thus, the government had pushed and pulled firms with threats and promises. Firms
were granted unrestricted and tariff-free access to imported intermediate inputs. Firms
were also granted automatic access to bank loans for working capital for all export
activities, even when the domestic money supply was being tightened. These firms also
had unrestricted access to foreign capital goods, as mentioned earlier, and were
encouraged to integrate vertically in order to sustain international competitiveness.
These incentives operated automatically constituted the crux of the Korean system of
export promotion. Furthermore, the rationing of longer term bank loans was used as
a carrot to draw down to new paths of exporting, encouraging diversification, and to

export more than otherwise. These incentives were applied to all exporting firms, but
particularly helped large firms with greater organizational, financial, and political lev-
erage to grow even larger.
That the Korean economy was primarily export oriented appears to have affected
the fast acquisition of technological capability at least in three ways. First, lump-sum
investments for capacity beyond local market size to bring about the economy of scale
forced local businesses to acquire technological capability fast in order to maximize
capacity utilization through improving international competitiveness. Second, as pro-
ducers entered the international market, the keen international competition forced
them to invest significantly in technological efforts, mainly in learning by doing and
reverse engineering, in order to be competitive in both quality and price. Third, infor-
mal technical assistance offered by foreign OEM (original equipment manufacture)
buyers to ensure Korea-made products meeting their technical specifications provided
invaluable help to Korean firms in acquiring necessary capability.
Ninth, while industrial and trade policies affected significantly the emergence and
growth of local industries, particularly chaebols, science and technology infrastructure
played little role in promoting the development of industries with mature technolo-
gies. In light of the absence of R&D capability in universities, the government estab-
lished in 1966 Korea Institute of Science and Technology (KIST) as an integrated tech-
nical center and subsequently its spin-offs to meet industry's technical needs.
However, the institutes suffered from poor linkages with industry at least through the
mid-1970s. In these institutes, most of the overseas trained Korean researchers came
from either academic fields or from R&D organizations of highly industrialized coun-
tries that undertook advanced research. Expertise was particularly lacking in manu-
facturing know-how and the development of prototypes, which were in great demand
in the early years. Furthermore, Korean researchers could not compete with foreign
licensors in supplying detailed blueprints and other manufacturing know-how, as well
as being unable to assist industry in solving the problems in the crucial initial stages
(Kim, 1989).
Under the aforementioned economic environment, Korean firms developed
strategies at the microeconomic level to sustain growth by acquiring technological
capability fast, which is the subject to be discussed below.
Microeconomic Behavior of Korean Firms
The microeconomic behavior of Korean firms in acquiring technological capability

varies significantly by industry. First, industries using small batch and unit production
(hereinafter small batch) such as shipbuilding and machinery focus more on devel-
oping capability to design and manufacture products than on setting up production
processes, as they usually produce highly differentiated products. As a result, large
firms in these industries relied heavily on the formal transfer of foreign technology
primarily in the form of foreign licensings and consultancies for the initial erection of
production systems and for the design of their products. These industries accounted
for a disproportionately large share of Korea's total foreign licensings through the
1970s. These firms have, however, exerted increasingly their own technological efforts
to strengthen their own capability and in turn to reduce their reliance on foreign licen-
sors, as they accumulated experience in production and product design. In contrast,

small firms in these industries (plus small-sized firms in the electronics and automobile
industries, whose large firms deploy large batch operations) resorted primarily to their
own imitative efforts to evolve organically over a long period of time, establishing their
initial production facilities with primitive technologies developed by themselves and
then gradually upgraded their production system and product quality through the imi-
tative reverse engineering of foreign products and processes (Kim and Lee, 1987).
Technology diffusion within these industries took rapidly through the mobility of
experienced engineers from existing firms to new entrants. Particularly, engineers
experienced at state-owned enterprises in the early 1950s have spun-off and played a
pivotal role in emerging private firms.
Second, industries using large batch and mass production systems (hereinafter
large batch) such as electronics and automobiles produce less differentiated products.
Large firms in these industries were also dependent on foreign licensings but to a lesser
extent than those with small batch operations. In the advent of an import substitution
policy, local firms entered technical assistance agreements with leading foreign firms
to import "packaged" technology (including assembly processes, product specifica-
tions, production know-how, training local technical personnel, and component
parts) so as to establish the initial production system. The immediate technological
task was the implementation of transferred foreign technology for the assembly of
imported components and parts, whose final products had been tested and proven
elsewhere, requiring only engineering (E) efforts. Most late entrants poached experi-
enced personnel from existing firms to set up and begin their operations, resulting in
a rapid diffusion of imported technologies. Then, competition stemming from new
entrants, together with increasing export promotion, called for indigenous efforts to
assimilate foreign technology, thus placing emphasis not only on engineering but also
on limited development (D&E). Capabilities acquired through assimilation of foreign
imported technology also enabled local firms to develop, through imitative engineer-
ing, related products without foreign assistance (Kim, 1980).
Third, industries using continuous processes such as chemicals, cement, paper,
and steel produce the least differentiated products in highly capital-intensive processes.
Since the final product and the equipment are relatively well known, but the propri-
etary know-how that lies in the details of the production process can make a significant
difference in the productivity of these industries, the initial production system of large
chemical, cement and steel plants were established on turnkey basis by Western firms.
Initial production capability to operate and maintain such process-oriented plants
stemmed largely from extensive training by foreign suppliers before, during, and after
the erection of the initial production processes. The local firms, however, used delib-
erate strategies with considerable efforts to acquire capability, leading to a series of
minor improvements with significant productivity increase, and to an increased tak-
ing-charge of engineering tasks in the subsequent expansions. A steel mill, for instance,
relied completely on the Japanese in the first phase in 1971 for annual capacity of one
million tons. But by 1981 capacity had been increased to 8.5 million tons in three
expansions that were increasingly under the direction of Korean engineers. The pro-
gressive substitution of foreign engineering by local engineering is well indicated by
the fall in the cost of foreign project engineering per ton from $6.13 to $0.13 during
the same period (Kim, 1987). Unlike other process-oriented industries, local phar-
maceutical firms in Korea started as small importers/dealers of packaged finished

drugs and later entered the drug manufacturing business by packaging imported bulk
drugs. Then, they gradually extended into more intricate operations, first from for-
mulating imported raw materials and later to backward integration into chemical
components production. Through this process, they grew in size and technological
capability (Kim etal., 1989).
In short, given the large size investment and lack of experience and capability,
large firms across industries have resorted heavily to foreign sources in order to ensure
quick erection and smooth start-up. In contrast, small firms underwent a long process
of imitative reverse engineering, as they lacked financial resources and organizational
capability to identify and negotiate with foreign technology suppliers. Nevertheless,
both large and small firms deployed deliberate, aggressive technological strategies to
acquire their own technological capabilities through exerted efforts to assimilate for-
eign technologies from the very outset. Korean firms have acquired technological
capabilities in piecemeal fashion, as successively more sophisticated capabilities have
been acquired and put into practice. The process of acquisition has clearly been one
of purposive efforts involving a succession of incremental steps, with production capa-
bilities being developed somewhat in advance of engineering and innovation capabil-
ities. The selectivity of import substitution for the elements of technology has meant
continued reliance on imports for at least some elements in almost all industries, but
the pattern of imports has continually shifted as local capabilities have replaced foreign
ones (Westphal et al, 1985). R&D in the formal sense of the term was not important
for Korea during this stage of imitating mature technologies. Industries in fact reversed
the sequence of R,D&E: it started with engineering (E) for products and processes
imported from abroad, and then progressively evolved into the position of undertak-
ing substantial development (D). But research (R) was not relevant to Korea's indus-
trialization through the 1970s. Several studies conducted in other countries also pro-
vide a similar evidence. Studies on Japanese industrial history show that its industries
went through a similar pattern (Shishido, 1972; Yakushiji, 1977; Ozawa, 1974). The
United States also reversed the sequence of research, development, and engineering,
that is, it started with practical technologies imported from Britain, and then slowly
evolved into the present position of being afforded substantial research (NAS, 1973).
Brazilian and Argentinian firms began manufacturing with assembly and packing
operations and eventually extended into more intricate operations. Engineering was
the initiating portion of the R,D&E spectrum leading gradually to more advanced
development and research efforts (Baranson, 1973; Dahlman and Frischtak, 1990;
Katz and Bercovich, 1990).
The rapid acquisition of capability through such a process may be attributed to
many factors. The continuous inflow of foreign technology (through formal and infor-
mal mechanisms) to learn from, the formation of highly trained human resources who
were willing to learn, entrepreneur's "can do" spirit exercised freely under restricted
equity participation of multinationals, and the government's orchestral role in direct-
ing chaebols and selectively allocating resources to them to achieve ambitious growth
objectives under the pressure of the competitive international market have worked
together for Korea to grow rapidly through the 1970s.
Chaebols particularly played an important role in this process, as they were in the
most advantageous position in drawing best quality human resources, in identifying,
negotiating, and financing foreign technology transfer, in obtaining business licenses
and preferential financing from the government, in applying experience gained in one
field of business to another, and in risk-taking in new businesses with existing busi-
nesses providing cushion therefor. As a result, by the end of 1970s, Korea had the larg-
est textile plant, the largest plywood plant, the largest shipyard, the largest cement
plant, and the largest heavy machinery plant in the world.
CONTEMPORARY NATIONAL SYSTEM FOR INDUSTRIAL INNOVATION
The economic environment for Korea has changed significantly in the 1980s for sev-
eral reasons. First, the world economy has generally slowed down in the 1980s, affect-
ing particularly outward-looking economies such as Korea. Second, in the wake of
rising trade imbalance, North America and Europe have moved toward protectionistic
policies, making it increasingly difficult for Korea to sustain export growth in indus-
tries that have led Korea's export-oriented strategy in the past. Third, Korea has lost
its competitiveness in low wage based labor-intensive industries, as its real wage has
risen at the average annual growth rate of 5.8% in the 1960s and 7.5% in the 1970s.
Concomitantly, other developing countries with much lower wage rates are rapidly
catching up with Korea in these industries. Fourth, advanced countries, particularly
Japan, are increasingly reluctant to transfer technology to Korea as Korea attempts to
enter industries that have been dominated by these countries. Fifth, Korea was forced
to change its copyright and patent laws, preempting the imitative reverse engineering
of foreign products. Such an unfavorable environment makes it imperative for Korea
to enhance its own technological capabilities in order to innovate independently and
to strengthen its bargaining power against foreign technology suppliers. This section
focuses on Korea's national system in the 1980s that are designed to cope with these
problems.
The Shift of Public Policies
In the face of such an unfavorable environment, the Korean government set out on
an economic liberalization program, making a series of policy shifts in an attempt to
introduce market mechanisms by reducing government intervention and to undertake
a structural change towards the development of relatively more technology-based
industries. First, in the face of chaebols' increasing economic power and its resultant
monopolistic abuses such as creating scarcities, price gouging, and ruining smaller
competitors, the government adopted a policy of "economic democratization" to mit-
igate monopolistic abuses by chaebols. The recently introduced Fair Trade Act along
the lines of American antitrust legislation includes, among other things, the prohibi-
tion of unfair cartel practices and mutual investment among chaebol's affiliate com-
panies, sets a ceiling of credit flow going to predesignated chaebols, and regulates their
vertical and horizontal integration. Nevertheless, chaebols grew even bigger and eco-
nomic concentration increased further: the combined sales of the five largest chaebols
as percent of GNP increased from 12.8 in 1975 to 35.0 in 1980 and to 52.4 in 1984.
Lee et al. (1986) report that the number of chaebols designated by the government as
dominating their respective markets increased from 105 in 1981 to 216 in 1985, but
only 10 were accused of having abused their economic power. Of 1172 applications
for vertical or horizontal integrations, only two were rejected by the government.
Second, as part of a financial liberalization program, the government reduced reg-
ulation of nonbank financial intermediaries and denationalized commercial banks.
The deregulation of nonbank financial intermediaries, many of which had long been
controlled by chaebols, resulted in their share of total deposit liabilities from 27% in
1980 to 42% in 1984. The denationalization of commercial banks resulted in the fall
of their ownership from government hands into those of chaebols. With increased
financial resources, chaebols went further to buy state enterprises that were being pri-
vatized and to buy financially troubled firms. In short, the financial liberalization pro-
gram enhanced the growth of chaebols,'1 economic concentration. Thus, Amsden
(1989, p. 137) concludes, "in the presence of large concentrations of market power,
reliance on the market mechanism for reform appear to produce some perverse results,
not only in Latin America but also in the Far East."
Third, since 1983, the government commenced a program to liberalize imports,
raising import liberalization ratio (defined as the ratio of the number of unrestricted
items to the total) from 51% in 1973 to 85% in 1984 and to 95.2% by 1988. The gov-
ernment also brought down the average tariff rate from 26% in 1984 to 16% in 1988.
As a result, imports increased by 20.1% in 1989 compared to 2.8% in export increase,
making the domestic market significantly more competitive than before.
Fourth, Korea gradually liberalized its technology transfer policies in the 1980s:
Korea needs progressively more sophisticated foreign technologies than in the earlier
years in order to sustain its international competitiveness in export-led industrializa-
tion. The liberalization raised the share of Korea's 999 industrial subsectors open to
foreign investment from 44% in the 1970s to 66% in 1984 and to almost 90% in 1988.
Foreign licensings has been completely open for all industries and for all terms and
conditions. Table 11.2 presents data on technology transfer through DPI and FL to
Korea. Over 48% of total DPI and 67% of royalty payments in FL since 1962 has taken
place in the last 5 years (1986-1991), reflecting the liberalized public policy environ-
ment as well as the private sector's aggressiveness in acquiring more sophisticated for-
eign technologies. In the 1960s and 1970s, foreign companies invested in Korea to reap
Korea's cheap labor costs. Now foreign companies invest to collaborate with Korean
companies in relatively more technology-intensive areas.
Fifth, the government began promoting small and medium-sized industries
(SMI), particularly technology-based small firms, to remedy the imbalance between
the large and small business sectors. The government established "sanctuaries" for
SMIs, designating 205 business territories where neither the large corporations nor
their affiliates can intrude. A program called The Compulsory Lending Ratio provides
that commercial (nationwide) banks should allow more than 35% of total loans to the
SMIs. In addition, local or regional banks have to provide more than 80% of their total
loans to these SMIs. The government also took initiatives in establishing the venture
capital industry as a way to promote the emergence of technology-based small firms,
when the private sector had no interest. Specifically, the government enacted a special
law to establish the first venture capital firm, jointly funded by the state and a group
of private firms. The government took a further step with the enactment of the Small
and Medium Enterprise Formation Act in 1986, leading to the emergence of 12 ven-
ture business firms, all jointly invested by the state and private sector.
Sixth, the focus of industrial policy has shifted from the promotion of "strategic"
sectors to that of innovation-related activities. That is, in the 1960s and 1970s, special
incentivestax exemptions, custom rebates, access to foreign exchange, and other

forms of protection or enhancementswere granted to "strategic" industries in order
to make them competitive at a world level. In contrast in the 1980s, the government
abolished all industry-specific promotion acts, which had been introduced in the
1970s, and instead legislated a new industrial promotion act that ties all incentives
with specific industrial activities such as the promotion of R&D and the development
of human resources. However, in the late 1980s, the government designated several
high-technology industries for promotion.
Finally, the government, as a way to democratize economy and society, liberal-
ized labor movements that had been tightly controlled in the 1960s and 1970s. Given
the opportunity, unionists' demands suddenly exploded disorderly, violently, and in
many cases unlawfully, resulting in work stoppages, missed export delivery deadlines,
lowered product quality, and wage hike beyond productivity increase. The average
company suffered from 19.2 days of labor strikes in 1989, up from 5.3 days in 1987.
In 1989 alone, labor disputes cost a total of about $6.2 billion in production as well as
$1.36 billion in reduced exports. Korea's real wages have increased a cumulative
62.5% since 1987. When adjusted in U. S. dollar terms, the increase amounts to 91.1 %,
which is 2.8 times that of productivity (Business Korea, 1990). In short, labor disputes
are a new variable that affects significantly the nation's efforts to innovate in the 1980s
and 1990s.
Under these policy environments, both the government and private sector inten-
sified their technological efforts to sustain international competitiveness. The follow-
ing sections present various aspects of these technological activities.
Structural Change in R&D Activities

Since many skills and activities required in reverse engineering are also the same in
R&D, activities that had been called reverse engineering have easily been transformed
into activities called R&D, as Korea approached the technological frontier. Reverse
engineering involved activities that sensed the potential needs in market, activities that
located knowledge or products that would meet the market needs, and activities that
would infuse these two elements into a new project. Reverse engineering also involved
purposive search of relevant information, effective interactions among technical
members within a project group and with marketing and production departments
within the firm, effective interactions with other organizations such as suppliers, cus-
tomers, local research institutes, and universities, and trial and error in developing a
satisfactory result. Skills and activities required in these processes are in fact the same
in innovation process in R&D.
Korea's R&D efforts may well be reflected in aggregate R&D investment trends.
Table 11.3 shows that the total R&D investment increased from W. 10.6 billion (U S$
28.6 million) in 1971 to W. 1878 billion (U S$ 2.37 billion) in 1987. Though the
Korean economy recorded one of the world's fastest growth rates, R&D expenditure
rose faster than GNP, increasing its share of GNP (R&D/GNP) from 0.32 to 1.93%
during the same period. A drastic increase was seen in the late 1970s and 1980s, reflect-
ing the shifts of government policy from the promotion of the light/mature industries
to that of heavy/technology-based industries and from the assimilation of production
technology to a strong drive for technological innovation. However, Korea's total
Table 11.3. Major R&D Indicators in Korea (Billion Won)
1971 1976 1981 1987
R&D expenditures 10.67 60.90 293.13 1,878.0

Funds from government 7.29 39.18 121.73 383.0
Funds from private sources 3.38 21.72 171.40 1,495.0
Government vs. private 68:32 64:36 42:58 20:80
Manufacturing NA 16.70 111.77 1,114.7
R&D/sales NA 0.36 0.67 1.83
Gross national products 3,376 13,881 45,126 97,532
R&D/GNP 0.32 0.44 0.65 1.93
Number of researchers (total)" 5,320 11,661 20,718 52,783
Government/public institutions 2,477 3,592 5,065 9,184
Universities 1,918 4,811 8,488 17,495
Private sector 925 3,258 7,165 26,104
R&D exp/researcher (W. 1000) 4,306 5,223 14,149 35,580
Researcher/ 1 000 population 0.08 0.33 0.54 1.27
Number of corporate R&D 1 12 65 455
laboratories
"The figures do not include research assistants, technicians, and other supporting personnel.
Source: MOST, Science and Technology Annals, various years.
R&D expenditure amounts only to about 54% of that of the General Motors Corpo-
ration.
Government statistics indicate that basic research accounted for 16.6%, applied
research for 19.6%, and development for the remaining 63.8% of the nation's total
R&D expenditures in 1987. The statistics also show that the private sector accounted
for 40% of the nation's basic and applied research, while universities accounted for
only 28% and public R&D institutes for 32% in the same year (KIRI, 1989). These
figures for basic and applied research, particularly the share of the private sector
therein, appear to have been much exaggerated, as leading chaebols have just started
a limited investment in applied research in the field of semiconductors and informat-
ics, let alone basic research.
There has been structural change in R&D investment. The private sector has
gradually assumed a larger role in the country's R&D efforts in response partly to
increasing international competition and partly to policy environment conducive to
private R&D activities. For example, the public sector accounted for 68% of the
nation's total R&D expenditure in 1971 but for only 20% in 1987. In short, the gov-
ernment took initiatives in local R&D efforts when the private sector had neither the
capability nor market incentives to undertake such activities. But as industrialization
progressed and Korea lost its comparative advantage in labor-intensive industries, the
private sector gradually took the leading role in local R&D in order to sustain its inter-
national competitiveness.
The Korean government has also promoted corporate R&D activities by offering
various programs to induce the private sector to set up formal R&D laboratories. The
programs include tax incentives and preferential financing to set up new laboratories
and exemption from military service obligation for key R&D personnel. Spurred by
these programs, the number of corporate R&D laboratories increased from one in
1970 to 122 by 1983 and to 604 in 1987, reflecting the seriousness with which Korean
firms are pursuing high-technology development. Although small and medium-sized
firms account for over 50% of total number of corporate R&D centers, chaebols dom-
inate R&D activities (to be discussed in detail below). Table 11.3 also shows that R&D
spending in the manufacturing sector has increased from W.I6.7 billion (U S$ 22.25
million) in 1976 to W.I, 114.74 billion (U S$ 1.41 billion) in 1987. Even in constant
prices, R&D spending has more than quadrupled during the period. This is faster than
the growth of sales: R&D spending as a ratio of total sales increased from 0.39% in
1976 to 1.83% in 1987. The machinery and electronics industry spent over 4% of sales
for R&D activities in 1987. However, manufacturer's R&D spending as a proportion
of total sales is less than half of that of manufacturers in the United States and Japan.
One of the major mistakes made by the Korean government in developing a
national system for industrial innovation has been its underinvestment in higher edu-
cational institutions. The number of universities has increased significantly from 69
in 1966 to 100 in 1985. The number of university students has increased from 131,354
to 931,884, with the proportion of high school graduates going on to universities grow-
ing from 28.2 to 38.3% during the same period. The environment for quality educa-
tion and research has, however, deteriorated over this period. For example, the stu-
dent-professor ratio has retrogressed from 22.6 in 1966 to 35.8 in 1985, making all
universities primarily undergraduate teaching-oriented rather than research-oriented.
Thus, the "publish or perish" principle is not applied.1
As a result, Korea has failed to develop a stock of highly trained scientists and
engineers who will be necessary in the 1990s for Korea to sustain its international com-
petitiveness. Compared to education at other levels, graduate education in science and
engineering has been much neglected in terms of quantity and quality. The govern-
ment's recent move to correct this discrepancy has resulted in across-the-board expan-
sion of graduate programs in all universities in an attempt to meet future needs at least
in quantity. It will take a great deal of time and effort, however, before such a move
can produce quality output.
Realizing the difficulty of reforming old undergraduate teaching-oriented
traditions in education, the government adopted a dual system: since all universities,
public or private, under the Ministry of Education (MOE) are basically teaching-ori-
ented, the Ministry of Science and Technology (MOST) founded a research-oriented
S&T school and plans to have another one soon, establishing a new research tradition
in university education.
Consequently, research activities in universities have been relatively underdevel-
oped. For example, though university R&D expenditures increased significantly from
W.572 million ($1.5 million) in 1971 to W.198 billion ($250 million) in 1987, they
account for only 5.4% of the nation's total R&D expenditures in 1971 and 10.5% in
1987, while universities account for 33.1% of the nation's R&D manpower and 78.4%
of its Ph.D. level R&D manpower in 1987.
The Korean government, however, is on the move in strengthening Korea's basic
research capability. For example, the government established a joint semiconductor
research institute in a state university, which will be used jointly by other universities
and corporate R&D centers. The government recently designated five university R&D
centers in 1989 that will receive increased financial assistance in order to strengthen
their R&D capabilities. The government is also in the process of enacting a "Basic
Research Promotion Law" as a way to promote the strengthening of basic research
capability as one of the nation's top technological goals.
In view of the absence of R&D capability in universities, public R&D institutes
have been a backbone of advanced R&D in Korea. They have also been major instru-
ments for national R&D projects, as they are recipients of over 90% of research grants
awarded by the government in "new" technology areas.
The Funding Source of Industrial R&D
Table 11.3 shows that the private sector has increasingly assumed a bigger role in the
nation's R&D, accounting for 80% in 1987. The Korean government has played a sig-
nificant role in helping the corporate sector secure funds for R&D activities mainly
through three mechanisms: direct R&D subsidy, preferential financing, and tax incen-
tives.
Direct R&D Subsidy
The government had no direct R&D subsidy programs through the 1970s, when
Korea relied primarily on light industries for economic development. It was in the
1980s when the government introduced two schemes for direct funding of private
R&D: one is "National R&D Projects" (NRP) administered by MOST in "new" tech-
nology areas focusing primarily on future problems and the other is "Industrial Base
Technology Development Projects" (IBTDP) administered by the Ministry of Trade
and Industry (MTI) in "existing" technology areas focusing primarily on current prob-
lems. Both schemes designate target technologies and offer direct R&D subsidy to
R&D organizations.
Criteria for NRP designation include projects involving "new" (to Korea) tech-
nologies with a high risk of failure or with high economic externalities, thus warranting
public support. MOST identified several target areas: localization of machinery parts
and components, new materials development, semiconductor design, super-mini-
computer development, energy conservation technology, localization of nuclear
energy fuel, new chemical development, biotechnology development, and basic
research in universities. In contrast, MTI undertakes a survey each year to identify
urgent R&D projects in industrial firms and provides R&D subsidy (IBTDPs) to
research organizations (public, private, and university) to undertake the projects. In
1989, 174 technologies were identified, 146 of which were designated as projects to be
funded.
The government's total investment in NRPs, as shown in Table 11.4, increased
significantly from W. 13.3 billion (U S$ 17.7 million) in 1982 to W.55.0 billion (U S$
69.4 million) in 1987. The government's R&D subsidy to the private sector through
NRPs also increased from W.5.0 billion (U S$ 6.7 million) to W.I8.1 billion (U S$
22.8 million). However, even in the case of 1987, this amounts only to approximately
U S$ 61,700 per project (equivalent to about two R&D man years) and U S$ 91,000
per participating firm. For IBTDPs, the government earmarked W. 11.5 billion (U S$
17.2 million) in 1989, about U S$ 117,936 per project. The amounts are not substan-
tial enough to solve critical problems.
Preferential Financing for Corporate R&D
The most important mechanism for funding corporate R&D is preferential financing
offered by state-controlled banks and public funds. Through the 1970s, the interest
Table 11.4. National R&D Project Investment (Billion Won)
1982 1983 1984 1985 1986 1987

Public and industrial projects
Projects with public leadership 8.2 13.9 16.1 17.3 27.1 28.2
Public R&D institute-industry 4.6 6.7 4.5 8.1 15.1 18.1
joint projects
(matched by private sector) 5.0 12.6 9.0 13.2 46.1 49.5
Number of projects 66 106 106 186 296 370
Number of corporates involved 86 131 134 212 240 250
International joint projects _ 1.6 2.3 2.9
Basic research 1.0 0.9 1.5 5.0 5.0
Technical assistance to small/ 0.4 0.4 0.5 1.0 1.5
medium firms
R&D evaluation projects 0.5 0.6 0.8
Total NRPs" 13.3 22.0 22.0 30.0 51.7 55.0
"National R&D Project.
Source: Science and Technology Annals (1988).
rate for R&D loans was one of the highest, reflecting the low priority of R&D in gov-
ernment policies. Table 11.5 shows that preferential financing amounted to W.671.6
billion (U S$ 848.0 million) in 1987, accounting for 94.3% of total corporate R&D
financing (W.712.4 billion equivalent to U S$ 899.5 million) by the state that year. In
contrast, direct R&D subsidy through NRPs and IBTDPs accounts for only 4% and
direct investment through venture capital firms for 1.7% of the total. The total amount
of public financing (W.712.4 billion) mostly in the form of preferential loans, in fact,
accounted for 64% of total R&D expenditure in manufacturing in 1987. In short, the
government plays a major role in funding corporate R&D through the allocation of
preferential financing.
Table 11.5. Public Source of Funding for Corporate R&D (W. 100 Million)
1982 1983 1984 1985 1986 1987
Direct R&D subsidy
NRPs0 51 71 51 81 151 185
IBTDPs" 100
Inventor's prototype development 0.5 0.7 1.0 1.0 1.1 1.2
Subtotal 51.5 71.7 52.0 82.0 152.1 286.2
Direct investment through venture capital 2 23 31 44 60 122
firms
Preferential financing
Venture capital firms 221 499 601 827 905 1,000
State-controlled banks 521 879 920 880 1,352 1,730
National investment funds 9 33 4 193 525
Industrial development funds 95 64 91 139 195 307
Industrial technology improvement 316 2,654
funds
Subtotal 837 1,451 1,717 1,942 3,252 6,716
Total 890 1,545 1,800 2,068 3,464 7,124
"NRPs, National R&D Projects administered by MOST in "new" technology areas; IBTDPs, Industrial Base Technology
Development Projects administered by MTI in "existing" technology areas.
Source: Science and Technology Annals (1988).
374
LOWER INCOME COUNTRIES
Tax Incentives
Tax incentives are a major indirect mechanism in making funds available for corpo-
rate R&D. In Korea, tax incentives may be classified into five categories according to
objectives associated with such incentives. Most important is tax incentives that are
aimed at promoting corporate R&D investment: reduced tariffs on import of R&D
equipment and supplies, the deduction of annual noncapital R&D expenditures and
human resource development costs from taxable income, accelerated depreciation on
industrial R&D facilities, and the exemption of real estate tax on R&D-related prop-
erties. The incentives also include a tax reduction scheme, called the Technology
Development Reserve Fund, whereby an enterprise can set aside up to 20% (30% for
high-technology industries) of profits before tax in any one year to be used for its R&D
work in the following 4 years. The private sector did not take advantage of this scheme
in early years in the absence of needs for technological activities. Other tax incentives
are aimed at reducing the cost of acquiring foreign technology, promoting technology-
based small firms, reducing the cost of commercializing locally generated technolo-
gies, reducing the cost of introducing new products, and promoting the venture capital
industry and venture investment.
Interorganizationaf Cooperation
Public R&D Institutes-Industry Cooperation

In the 1980s, the government used various means to bring about effective linkages
between the public R&D institutes and industry. For example, NRPs and IBTDPs,
albeit insignificant in their total monetary value, were used as major means to facilitate
public R&D institute-industry cooperation to develop the aforementioned target
technologies. The schemes were also aimed at facilitating university involvement in
the projects, but have produced only limited results. In some cases, large firms, how-
ever, tend to shy away from receiving public support in recent years in order to keep
their R&D activities confidential; thus the programs now largely support small and
medium-sized firms. Nevertheless, both mechanisms have led to significant joint R&D
results. For example, a public R&D institute and four private firms have successfully
developed and commercialized a low-density electronics switching system (TDX-1)
and are now in the process of designing a high-density system (TDX-10). The same
public R&D institute and three semiconductor producers completed jointly designing
a 4-mega dynamic RAM (to be elaborated further below).
University-Industry Cooperation
Given the teaching orientation of almost all universities, there is fundamental lack of
interplay between the universities and the private sector. This is one of the greatest
weaknesses of Korea's national system for industrial innovation. Unlike universities
in advanced countries that play a major role in basic as well as applied research that
have important bearings on industrial R&D, almost all Korean universities suffer from
insufficient facilities and their faculty members have neither the time nor the incen-
tives to undertake serious research. Consequently, universities have lost the confidence
of the private sectors, drawing almost no research funds from business circles. The
government, as a way to promote university-industry cooperation, opened its NRPs
to universities but universities participated in only 6 of 255 projects in 1984, involving
W.I 94 million, or less than 1% of total NRP expenditure (KAIST, 1986). In short,
university-industry R&D cooperation in the formal sense of the term has been insig-
nificant. But informal collaborations in the form of consultancy by individual faculty
members are quite prevalent. This may be symptomatic of Korea's position as a catch-
up country, where technological tasks are not yet at the frontier.
Intercorporate Cooperation
A major form of intercorporate R&D cooperation is industrial research cooperatives.
The technology Development Promotion Act of 1977 promotes the formation of such
cooperatives, but it was not until 1982 that the first such cooperative emerged in
response to direct R&D subsidies offered by NRPs. Eleven cooperatives emerged in
the first year. By mid-1989, 46 industrial research cooperatives were founded involv-
ing 986 firms, 759 of which are small and medium-sized firms. The government offers
various tax incentives to promote R&D cooperatives. Most existing cooperatives,
however, are in name only with little substance. NRPs and IBTDPs are used as instru-
ments to bring about effective intercorporate R&D cooperation by inviting R&D
cooperatives to undertake joint research with public R&D institutes, but they neither
have adequate R&D facilities nor full-time researchers to take advantage of subsidies
to R&D cooperatives.
Technological Strategy of the Private Sector
As chaebols have been a driving force in expanding production and exports in the
1960s and 1970s, they also play a pivotal role in developing high-technology industries
in Korea in the 1980s. In doing so, chaebols have aggressively diversified their sourcing
of technologies. First, several major chaebols have set up outposts in Silicon Valley in
California to "leapfrog" into state-of-the-art technologies by monitoring technological
changes and to acquire advanced semiconductor and computer technologies. Califor-
nia also gave another advantage to chaebols by solving the most obvious obstacle to
Korea's high-technology adventureslack of experienced scientists and engineers for
high-technology industries. California is populated by thousands of top-notch scien-
tists and engineers who are Korean-Americans. Leading chaebols have lured away
some of the best. Many of them left Korea more than a decade ago, earned Ph.D.s in
America's best universities, and rose through the ranks of such leading U. S. concerns
as IBM, Fairchild, Intel, and National Semiconductor. The well-financed Korean
chaebols gave them challenging jobs and attractive compensation packages without
their having to leave their American life-styles and with considerable independence.
For example, Hyundai's first step on entering industrial electronics was establishing a
subsidiaryModern Electrosystems in Sunnyvale, Californiamanned mostly by
Korean-American holders of Ph.D.s and research veterans of top U. S. semiconductor
makers. Samsung also established two subsidiaries in Silicon Valley to undertake
R&D activities in microelectronics and biotechnology fields, led by several Ph.D.-
holding Korean-Americans, with the rest being native American engineers. Another
Korean-owned start-up, ID Focus of Santa Clara, undertook market research and
industrial design for Daewoo Electronics and Daewoo Telecom with a staff composed
entirely of Americans. Lucky-Goldstar also established two subsidiaries for semicon-
ductor/computer and biotechnology R&D. These subsidiaries are engaged mainly in
R&D activities for mass production in Korea. These outposts in California also serve
as an "antenna" for information on research activities in advanced countries and as
training posts for scientists and engineers from R&D centers and manufacturing
plants in Korea (Kim, 1987).
Second, it is also chaebols that have successfully developed ties with multination-
als, which provide important inputs in developing Korea's high technologies. Multi-
nationals from advanced countries are looking at Korea to form consortiums for tech-
nology cooperation, as Korea is very weak in basic design and basic science but has an
international competitive edge in manufacturing and process technologies. Multina-
tionals in high-technology areas such as IBM, Hewlett-Packard, Honeywell, ATT,
Monsanto, Hitachi, and Toshiba, have recently established joint ventures with Korean
chaebols in high-technology industries. Chaebols have also entered extensive licensing
ties with foreign high technology firms.
Third, chaebols developed closer ties with local public R&D institutes in the
1980s than the previous decades. Capabilities available in the public R&D institutes
became relevant to chaebols' efforts to develop high-technology industries.
Fourth, chaebols have also invested aggressively in developing in-house R&D
activities in order not only to absorb, assimilate, and adapt imported technologies
from their outposts in Silicon Valley and from multinational partners but also to
strengthen their own innovative activities. For example, Samsung set up 12 new R&D
centers in the country during the 1980s, including Samsung Advanced Institute of
Technology and Samsung Integrated R&D center. In 1989 alone, Samsung spent $900
million, Hyundai and Lucky-Goldstar spent $600 million each, and Daewoo spent
about $300 million in R&D. Approximately 80% of the investment went for electric
and electronics, about 18% for chemicals, and the remaining 2% for other fields. These
four largest chaebols accounted for most of the nation's total industrial R&D. Leading
chaebols have relative advantage over small or medium-sized firms in recruiting
scarce, high-caliber human resources by offering attractive compensation and the best
opportunities for career development among Korean firms. It is these chaebols that
lure foreign-educated Korean scientists and engineers to their R&D centers in Korea
and that recruit the top graduates from Korea's best universities. Leading scientists
and engineers in chaebols' R&D centers and technology planning posts at headquar-
ters are almost all foreign-educated Ph.D.s often with extensive experience at major
multinational firms in advanced countries. Chaebols also invest most ambitiously in
training to upgrade their human resources.
With the technological network, high-caliber scientific and engineering man-
power, and ambitious R&D investment, chaebols took a leading role in moving
toward the technological frontier. Illustrations in some of high-technology industries
such as semiconductors, telecommunications, robotics, computers, and aircraft might
shed light on how chaebols have managed to acquire technological capabilities to start
and grow in these industries in the 1980s.
First, Korea has long participated in the semiconductor industry, involving
merely simple packaging processesbonded assembly processing by wholly owned
foreign subsidiaries or joint ventures, with all parts and components imported from
the parent companies and reexported back to the consignors. Then, it made a technical
leap to the production of VLSI (very large-scale integrated circuit) in 1983. Large chae-
bols started the heavy betting to build a semiconductor industry and the odds were
stacked in their favor: a ready access to funds syphoned from cash-cow industries, a
motivated and highly skilled work force recruited from abroad, and a low labor-cost
advantage. But most important of all, there were a number of distressed small semi-
conductor companies in the United States that were ready to sell what Korea's chae-
bols needed mostchip designs and processesin attempts to fuel cash for survival.
Samsung Semiconductor's 64K DRAM (dynamic random access memory) technol-
ogy was licensed from Boisie, Idaho-based Micron Technologies, enabling Samsung
to hit the market with 64K DRAM some 18 months after the first Japanese ones
became commercially available. Design for Samsung's high-speed CMOS process was
picked up for $2.1 million from Zytrex of California. Samsung then turned to Mostek
for its 256K DRAM design. Goldstar Semiconductor signed a licensing agreement
with Advanced Micron Devices in Silicon Valley for chip designs. ATT, the joint ven-
ture partner, was a source of design technology for 1M DRAM. Hyundai turned to
Vitelic of San Jose for designs of 1M and 256K DRAM as well as for a variety of SRAM
(static) designs. Daewoo, the fourth member of the chaebol chip club, got its window
on the business in 1985 by injecting some $18 million into the ailing Zymos Corp,
acquiring 51% equity. Silicon wafers, the most important input for chipmaking, are
produced by two local firms: Korsil, a joint venture with Monsanto, and Lucky
Advanced Materials, which licensed wafer processing technology from Siltec of Cali-
fornia (now owned by Mitsubishi Metals). In short, chaebols relied mainly on foreign
firms for design and process technologies.
Another important factor that made Korea possible to take root in the semicon-
ductor industry was export restraints imposed on Japan after trade friction with the
United States and the subsequent move upmarket by Japanese firms, opening up
opportunities for Korean firms to penetrate the U. S. market. This was also true of
consumer electronics and subcompact cars (Porter, 1990).
The chipmaking firms made enormous investment in both production facilities
and R&D. By the end of 1988, four semiconductor producers invested over $1.3 bil-
lion in new facility investment. Total R&D expenditures increased dramatically, from
$13 million in 1983 to $95 million in 1987, accounting for 10 and 22% of the total
investment, respectively. Industry's aggressive R&D spending cut down Japan's lead.
For example, Korea was just 6 months behind Japan in introducing the 1M DRAM.
But the road ahead was getting bumpier and is even more so today. Texas Instru-
ments filed a suit against Samsung and eight Japanese chipmakers for infringing on its
patents for DRAM designs in 1986, while Intel Corp filed a similar suit against Hyun-
dai and its American design suppliers. Both Samsung and Hyundai ended up paying
royalties on past and future sales of their memory products. Furthermore, work on the
next generation of chipsthe 4M DRAMmeant exploring the frontiers of semicon-
ductor technology but also competing neck-in-neck with Japanese and U. S. compa-
nies. As the stakes have risen in the chip game, the field of players has grown smaller
worldwide, meaning that few, if any, of those left in the game can be counted on to sell
state-of-the-art chip design technology to Korean chaebols. So, Koreans had to tackle
the 4M DRAM design alone. To avoid duplicate research and investment, the govern-
ment stepped in and designated the R&D of the 4M DRAM as a national project. A
public R&D institute played the coordinating role with three chaebols' participation.
Samsung was the first among three chaebols that announced its completion of design-
ing 4M DRAM in 1989 and 16M DRAM in 1990, only a few months after Japan.
Second, electronic switching system (ESS) and optical fiber/cable may be high-
technology products in telecommunications. The Korean government (the sole buyer)
decided in 1981 to shift public telephone switching from an electrical direct dialing
(EDD) system to ESS in 1981. Lacking the technological capability to enter ESS pro-
duction, three chaebols and one medium-sized firm entered foreign collaborations
two in the form of joint venture and the other two by licensingto locally assemble
foreign-designed ESS equipment. While the local producers had progressively
increased the local content ratio for ESS, raising as much as 76% for some of the mod-
els by 1986, a public R&D institute organized a consortium with the four firms to
develop Korea's own rural (low-density) ESS system (Time Division Exchange or
TDX-1). Production experience of foreign systems plus high-caliber Korean engi-
neering manpower recruited from abroad blended together. On completion, the
design technology was transferred to the participating firms for production. TDX-1
has replaced a Swedish model since 1987. The consortium produced an improved
version (TDX-1B) and exported it to the Philippines. At the same time, the con-
sortium is undertaking a project to develop a high-density system (TDX-10) for
urban use.
When Corning Glass refused to transfer optical fiber production technology to
Korea in 1977, two chaebol copper cable producers entered a joint R&D project with
a public R&D institute. After 7 years of R&D, the locally developed optical cable was
tested successfully on a 35-km route in 1983. Although this local effort eventually
grounded to a halt due mainly to slow progress in R&D, it nonetheless helped local
firms gain bargaining power in acquiring foreign technology on favorable terms. Four
chaebols entered licensing agreements with multinationals in 1984: Samsung with
IT&T, Goldstar with AT&T, Daewoo with Canada's Northern Telecom, and Tainan
with Japan's Sumitomo.
Third, Korea took an independent, reverse engineering route in developing
robotics in early years. Again four leading chaebolsHyundai, Samsung, Goldstar,
and Daewoohave dominated the industry. The first batch of welding robotics for
chaebol's automobile assemblers, shipbuilders, and aerospace firms were all imported.
The first batch of automatic part inserting robotics for largest electronics firms were
all imported. But the users of these imported robotics were active assimilators of
imported technology. Based on operating experience of foreign models and capability
gained from in-house R&D efforts, they developed imitative models first to meet the
subsequent needs within the plant and then to meet increasing demand within their
chaebol groups. In contrast, some smaller chaebols recently entered the industry as a
local dealer of foreign models as a way to gain experience. Nevertheless, these chaebols
stepped up their R&D efforts (over $28 million in 1987 alone) in the robotics industry
with the aim of cashing in on fast-growing demand both at home and abroad, giving
rapid rise to technological capability and leading to the introduction of their own mod-
els. For example, Daewoo's five-axes multiple process NOVA-10 introduced in 1984
had both quality and price competitiveness to export to the United States. As in other
chaebols, U. S.-trained Ph.D.s played a central role in initiating and expanding the
robotics business at Daewoo. The government is also mapping out a comprehensive
program to boost automation-related industries such as robotics production. How-
ever, robotics technology in Korea is still confined mainly to the production of robotics
for simply repetitive welding works, making it difficult to turn out highly integrated
products. In response to the need for more sophisticated robotics, chaebols such as
Hyundai and Daewoo are now pursuing technology tieups with multinationals, while
public R&D institutes are undertaking advanced research in areas such as artificial
intelligence, vision-sensor, six-axes robotics, and a voice-recognition system.
Fourth, the computer industry has also undergone a similar route in Korea.
Chaebols with experience in consumer electronics entered the industry through imi-
tative reverse engineering, leading to the introduction of IBM PC (personal computer)
clones. These chaebols rapidly expanded their businesses as OEM (original equipment
manufacture) producers. Smaller chaebols without consumer electronics background
entered first as a local dealer/service agent as a way to accumulate experience in com-
puter technology and gradually backward integrated to the design and production of
their own models (Kim et al., 1987). With a good grip on the chip business and micro-
computers, Korea now tackles another national project: the design and manufacture
of Korea's own super-minicomputers. Again, a team of software and hardware engi-
neers from a public R&D institute and a number of chaebols have been formed to
work on the project. The basic idea of the super-mini project is to buy and transplant
foreign technology, in this case a super-mini design from one of the U. S. venture firms,
and to progressively localize design capability.
Fifth, chaebols also dominate the aircraft industry. Korea's rudimentary experi-
ence in aircraft technology stemmed largely from international subcontracting and
maintenance. Boeing and other U. S. aircraft makers that used Japan as a manufac-
turing base for bodies, parts, and components have shifted their sourcing to Korea in
the facing of rising Yen. Three chaebols entered into close relations with U. S. man-
ufacturers. For instance, Daewoo has supplied outer fuselages, side cockpit panels, and
central fins to General Dynamics. Daewoo has also supplied upper decks for Boeing
747s and inspar wing ribs, and wing part components for Boeing-737s and wing and
tail parts for a new U. S. fighter to Lockheed. Korean Air, which has accumulated air-
craft technology from maintenance service of such U. S. military fighters as the F-15,
F-16, and F-5 Phantom, has supplied wing tip extensions for Boeing and sheet metal
assemblies for McDonnel-Douglas. In addition, Daewoo entered a joint venture with
Sikorsky and Samsung exchanged a memorandum of understanding with Bell to pro-
duce helicopters in Korea for Korean military and for exports under OEM basis. These
chaebols will participate in manufacturing the next mainstay fighters for the Korean
Air Force. The U. S. government proposed that Korea buys 12 fighters as finished air-
planes, produces 36 through kit assembly, and manufactures the remainder under
license. Through this process, Korea hopes to acquire enough technological capability
to develop its own aircraft. Chaebols have intensified their R&D efforts in order to
assimilate foreign technologies. The Korean government also takes an active role in
promoting the industry: the government legislated in 1987 Aerospace Industry Devel-
opment Law and established Aerospace Technology Institute as a way to support
R&D activities of the private sector.
In short, chaebols have deliberately taken a pathfrom reverse engineering to
international tieups and to own research and developmentto expedite and cost-min-
imize in acquiring technological capability in high-technology industries. When tech-
nological tasks were relatively simple, they opted to reverse-engineer foreign products.
The robotics and computer industries in early years may be good examples. When
technological tasks were beyond their capacity, they quickly turned to multinationals,
which were willing to share technologies with Korea for various reasons. The semi-
conductor, electronic switching, aircraft industries in the beginning, and the robotics,
computer, and optical fiber industries in the later stage may fit this case. When Korea
approached the technological frontier, where few multinationals were willing to share
their technologies, Korea consolidated its efforts to develop its own products. The
recent development of the 4M DRAM may be a case in point. During this process,
public R&D institutes have played a pivotal role in giving rise to the bargaining power
of chaebols in acquiring foreign technologies and in coordinating joint research efforts
with chaebols.
Although a complaint is mounting that chaebols dominate high-technology
industries like they did in many other sectors in the past, this dominance is considered
inevitable at the initial stage owing to the great deal of risk and investment and to the
absence of research-oriented universities that may incubate technical entrepreneurs
who might start small technology-based venture firms. There are, however, some indi-
cations that a few venture businesses have recently appeared in high-technology fields.
Expanding investment in basic research at universities, increased availability of ven-
ture capital, increasingly conducive economic environment created by the govern-
ment, and the fact that chaebols cannot dominate a wide range of application tech-
nologies will hopefully lead to a strong appearance of small-sized venture businesses
in Korea in the near future.
Other Issues
There are two other issues worth covering here: the effects of military R&D on indus-
trial innovation and the formation of hi-tech "valleys." First, given the threat of hos-
tilities with North Korea, national security has been one of the major concerns in
Korea, making the home market unusually sophisticated and demanding. As a way to
mitigate the effects of constraints imposed by foreign weapon suppliers on national
security and to ensure military independence, the Korean government in the 1970s
launched an ambitious program to build up local capability to develop weaponry. Its
budget size was, however, $ 114 million in 1988, only 0.2% of that in the United States
and 4% of that in France. The percent of military R&D budget to total military expen-
ditures was only 1.5% in Korea compared to over 10% in France and the United States.
The nature of projects, however, has been kept strictly confidential. And R&D has
been undertaken almost strictly within the military: only 1.48% of military R&D bud-
get was allotted to universities in 1988 (NDC, 1990). Such isolated military R&D
activities appear to have suffered from lack of interaction with the rest of Korea's S&T
infrastructure and also from competitive disadvantages vis-a-vis public and corporate
R&D institutes in recruting highly qualified scientists and engineers. As a result, the
ambitious program has been drastically scaled down in the 1980s. The private sector
is involved in manufacturing some traditional weaponry, but the spillover effects of
such operations on industrial innovation appear to have been quite negligible.
Second, the government created two high-technology centers: Seoul Science Park
and Daeduck Science Town, 200 km south of Seoul. Seoul Science Park started with
three R&D institutes and three economic research institutes but has failed to attract
private R&D centers in the vicinity. Furthermore, one of the three R&D institutes has
relocated, while the second one will be relocated within a few years.
In contrast, the government has actively promoted the development of Daeduck

Science Town, where it has located nine public R&D institutes and three tertiary edu-
cational institutions and attracted four private R&D laboratories. About 20 major
firms plan to locate their central R&D laboratories in the Town within a few years,
making the Town the first high-technology "valley" in the country.
In addition, there are signs in different parts of the country that corporate R&D
laboratories in some fields tend to cluster together in the same vicinity. For example,
semiconductor R&D centers and biotechnology R&D centers cluster near Seoul, tele-
communications R&D laboratories near Kumi, which is known as an electronics
industrial zone, and new materials R&D centers near Pohang, where POSCO has its
first steel plant and its own research-oriented private university. It may be premature
to regard them as high-technology "valleys" at this point but they appear to be in the
incubation stage.
CONCLUDING REMARKS
Korea has been transformed from a subsistent agricultural economy into a newly
industrializing one in a very short span of time. During the transformation process,
Korea's economic environment has changed drastically. Protectionist moves in North
America and Europe are making it increasingly difficult for Korea to sustain export
growth in the industries that have led Korea's outward-looking economic strategy in
the past. These countries have also put pressure on Korea to liberalize its domestic
economy, forcing Korea to compete against multinationals not only in international
markets but also in its domestic market. At the same time, Korea's competitiveness in
labor-intensive industries had been rapidly eroded by rising real wages. Other devel-
oping countries with much lower wage rates are rapidly catching up with Korea in
labor-intensive industries. Furthermore, industrialized countries, particularly Japan,
are increasingly reluctant to transfer technology to Korea (Kim, 1988).
The rapid changes in economic environment have not given Korea enough time
to make and consolidate a major shift in designing an appropriate national system for
industrial innovation. The government has introduced in the 1980s various new pol-
icy instruments in an attempt to develop a new national system, which is yet at the
formative stage. Furthermore, a consensus on the method of redesigning the institu-
tional system and on drastic increases in resource allocation for R&D expenditure is
difficult to be made. At the same time, the traditionally teaching-oriented universities
could not easily be transformed into research-oriented institutions, giving gloomy
implications for generating high-caliber scientists and engineers, let alone university-
industry cooperation. If the formation of skilled workers in the 1950s were an impor-
tant leading indicator for the subsequent development in the 1960s and 1970s, then
the poor health of universities and the resultant lack of high-caliber human resources
in the 1980s may be a critical warning signal for the 1990s and on.
The government took the initiatives in the 1960s and 1970s in directing and driv-
ing the private sector to achieve growth by fostering the formation of large chaebols
through the allocation of projects and resources. This scheme appears to have helped
in bringing about an economy of scale in the stable, mature industries and achieving
outward-looking objectives in the past. At the same time, however, the same scheme
gave rise to highly concentrated economic power within a limited number of'chaebols
and stifled the growth of small and medium-sized industries. These chaebols may be
in an advantageous position in the 1980s in developing such industries as semicon-
ductors and aircraft, which require a large investment. The highly concentrated mar-
ket structure has, however, curtailed the market dynamism. Furthermore, highly
bureaucratized large chaebols, which effectively functioned in the stable industries in
the past, now find themselves rigid and slow in adapting to rapid changes in technology
and market. Some chaebols have, nonetheless, managed to build up their capability to
cope with the changing economic environment. They began investing heavily in R&D
to give rise to existing capability in order to design and manufacture "next stage" prod-
ucts. But most others are highly dependent on foreign sources for new technologies.
Another problem in Korean economy is the vital missing link in the presence of
related and especially supporting industries. Chaebols have largely focused their efforts
in end products with heavy dependence on imported parts and machinery. The prob-
lem, as Porter (1990) notes, is not the cost of inputs but the effects on the innovation
process. Without fluid supports from capable domestic suppliers, Korean firms will
remain behind in product and process innovations.
Korea is, indeed, at the crossroad. While the government and private sector exert
their efforts to redesign the national system for industrial innovation and in turn to
sustain the nation's economic growth by strengthening Korea's technological capabil-
ities, there are several critical signs, as mentioned above, that have led Korea to a seri-
ous economic crisis since 1989 and that may impede its future development.
NOTE
1. Exceptions are Korea Advanced Institute of Science and Technology (KAIST), a grad-
uate school of applied science, and Korea Institute of Technology (KIT), an undergraduate
school within the KAIST system, both of which are under the jurisdiction of MOST. All other
universities and colleges are under the jurisdiction of the Ministry of Education (MOE). Another
exception is POSTEC, a new research-oriented school of science and technology established by
Pohang Iron and Steel Corporation (POSCO), one of the most cost-competitive steel companies
in the world.
REFERENCES Industrialization and Growth: A Comparative

Study. New York: Oxford University Press.
Dahlman, C. J., and Frischtak, C. R. (1990).
Amsden, A. H. (1989). Asia's Next Giant: South "National Systems Supporting Technical
Korea and Late Industrialization. New York: Advanced in Industry: The Brazilian Experi-
Oxford University Press. ence." Paper presented at National Technical Sys-
Baranson, J.(1973). "The Drive Toward Technolog- tem Conference at Stanford University, October
ical Self-Sufficiency in Developing Countries." 18-19.
Paper presented at the Conference on Latin Harbison, Frederick, H., and Myers, C. S. (1964).
America-United States Economic Interactions: Education, Manpower, and Economic Growth.
Conflict, Accommodation, and Policies for the New York: McGraw-Hill.
Future, at the University of Texas at Austin, Hou, C. M., and Gee, S. (1990). "National Systems
March 19. Supporting Technical Advance in IndustryThe
Business Korea (1990). "Balancing Workers' Inter- Case of Taiwan." Paper presented at National
ests against the Nations." February, p. 31. Technical System Conference at Stanford Univer-
Chenery, H. Robinson, S., and Syrquin, M. (1986). sity on October 18-19.
Jung, K. H. (1989). "Business-Government Rela- Strategy of Local Pharmaceutical Firms in Korea:
tions in Korea." In K. H. Chung and H. C. Lee A Multivariate Analysis." Technology Analysis &
(eds.), Korean Managerial Dynamics, 11-26. Strategic Management l(l):29-44.
New York: Praeger. KIRI (Korea Industrial Research Institutes) (1989).
KAIST (Korea Advanced Institute of Science and Major indicators of Industrial Technology. Seoul:
Technology) (1986). Urinara Kwahak Gisul KIRI.
Gebal Systemeo Jeonge Gwajungkwa Jeonmang. Lee, K. U., Urata, S. and Choi, I. (1986). "Recent
(A Study on Science and Technology Develop- Development in Industrial Organizational Issues
ment System in Korea and Its Future Direction). in Korea." Mimeo, Korea Development Institute
Seoul: Ministry of Science and Technology. and the World Bank, Washington, D.C.
Katz, J., and Bercovich, N. A. (1990). "Science, Mason, E. S., Perkins, D. H. Kim, K. S. Cole, D. C.
Technology and Socioeconomic Re-structuring: and Kim, M. J. (1980). The Economic and Social
The Case of Argentina." Paper presented at Modernization of the Republic of Korea. Cam-
National Technical System Conference at Stan- bridge, MA: Council on East Asian Studies, Har-
ford University, October 18-19. vard University Press.
KEB (Korea Exchange Bank) (1987). "Direct For- McGinn, N. F., Snodgrass, D. R., Kim, Y. B., Kim,
eign Investment in Korea." Monthly Review S. B. and Kim, Q.-Y. (1980). Education and
October, 3-13. Development in Korea. Cambridge, MA: Council
Kim, K. S., and Roemer, M. (1979). Growth and on East Asian Studies, Harvard University Press.
Structural Transformation. Cambridge, MA: MOST (Ministry of Science and Technology) (vari-
Council on East Asian Studies, Harvard Univer- ous years). Science and Technology Annals. Seoul:
sity Press. MOST
Kim, L. (1980). "Stages of Development of Indus- NAS (National Academy of Sciences) (1973). U. S.
trial Technology in a LDC: A Model." Research International Firms and R,D&E in Developing
Policy 9:254-77. Countries. Washington, D. C.: National Academy
Kim, L. (1987). Technological Transformation in of Sciences.
Korea: Progress Achieved and Problems Ahead. NDC (National Defence College) (1990). Gukbang
Helsinki: World Institute for Development Eco- yungu gebalkwa mingan yungu gebaleo gyunge
nomic Research, mimeo. ganghwa bangane kwanhan yungu. (A proposal to
Kim, L. (1988). "Korea's Acquisition of Technolog- strengthen the link between military and indus-
ical Capability for Internationalization: Macro trial R&D). Seoul: KIST Center for Science and
and Micro Factors." Business Review 22( 1): 183- Technology Policy.
97. Ozawa, T. (1974). Japan's Technological Challenge
Kim, L. (1989). "Science and Technology Policies to the West, 1950-1974: Motivation and Accom-
for Industrialization in Korea." In J. Suh (ed.), plishment. Cambridge, MA: MIT Press.
Strategies for Industrial Development, 255-272. Porter, M. E. (1990). The Competitive Advantage of
Kuala Lumpur: Asia and Pacific Development Nations. New York: The Free Press.
Council. Shishido, T. (1972). Japanese Policies for Science
Kim , L., and Kim, Y. (1985). "Innovation in a and Technology. Tokyo: Nikko Research Center,
Newly Industrializing Country: A Multiple Dis- mimeo.
criminant Analysis." Management Science 31 (3): Westphal, L. E., Kim, L. and Dahlman, C. (1985).
312-22. "Reflections on the Republic of Korea's Acquisi-
Kim, L., and Lee, H. (1987). "Patterns of Techno- tion of Technological Capability." In N. Rosen-
logical Change in a Rapidly Developing Country: berg and C. Frischtak (eds.), International Tech-
A Synthesis." Technovation 6:261-76. nology Transfer: Concepts, Measures, and
Kim, L., Lee, J. and Lee, J. (1987) "Korea's Entry Comparisons 167-221. New York: Praeger.
into the Computer Industry and Its Acquisition of Yakushiji, T. (1977). Dynamics of Policy Interven-
Technological Capability." Technovalion tions: Government and Automobile Industry in
6(4):277-93. Japan, 1900-1960. Ph.D. dissertation, MIT,
Kim, Y., Kim, L., and Lee, J. (1989). "Innovation Cambridge, MA.
12
National Systems Supporting Technical
Advance in Industry: The Case of Taiwan
CHI-MING HOU
SAN GEE
The Taiwan model of economic development has attracted worldwide attention. Tai-
wan's record of economic growth has been phenomenal and the fruits of economic
growth have been widely shared by all income groups on the island. Furthermore, the
improvement of the material well-being of the people was accomplished in a climate
of high consumer sovereignty without undue government control, high economic sta-
bility without serious inflation or unemployment, and financial solvency without for-
eign debt. Also, a high degree of structural transformation took place and caused a shift
from primary to secondary industries. Some countries have done as well as Taiwan in
some of these areas but few, if any, have done as well as Taiwan in all of them.
It is not our intention to propagandize Taiwan's performance, but we wish to
explore the role of technology in assisting Taiwan's economic achievement as we
described above. As a newly industrializing country in Asia, Taiwan has been pro-
moting technology development, which is considered one of the most important strat-
egies for her overall economic development. However, unlike the developed countries,
in Taiwan, an immediate difficulty blocking the promotion of technology develop-
ment has been, and still is, the lack of skilled technicians and qualified researchers.
Furthermore, it is well known that Taiwan's economy mainly consists of small
and medium enterprises (SMEs). Under such circumstances, the immediate questions
that the government has to face are: Will the SMEs be able to survive given the rapid
technological development in the rest of the world? Will these SMEs be capable of
conducting enough research and development (R&D) to sustain their long-term com-
petitiveness? What are the SME's major sources of technology in Taiwan? What are
the possible policy options that the government has in promoting technological devel-
opment among SMEs?
Because of the rapid development in science and technology, which in turn
improves telecommunications and transportation, the world's economies have been
quickly integrated into one unit. As such, the development of science and technology
is critical to the international competitiveness both of firms and of the country. In this
384
TECHNICAL ADVANCE IN INDUSTRY: TAIWAN 385
chapter we look at some of the technological-development experiences of Taiwan,
which we hope could be useful to other developing countries as well as to further
research in this field.
SOURCES OF ECONOMIC GROWTH
What has contributed to the fast growth in Taiwan? Undoubtedly a great deal of
research remains to be done before an answer can be advanced, if indeed such an
answer can be found at all. Work ethics, the high propensity to save, the emphasis on
education, the early adoption of an outward-looking development strategy, land
reform, the crucial role of foreign aid in the early stages of development, and the favor-
able international economic environment have all been emphasized, in various
degrees, as among the most important factors contributing to Taiwan's economic suc-
cess.1 In our view, however, technological improvement, especially by the SMEs, has
played the most crucial role.
We recognize it is extremely difficult to substantiate any hypothesis on the sources
of economic growth, but if the Solow-type method is used, it may be computed that
nearly 54% of the growth in gross domestic product from 1952 to 1979 was due to
technological improvements (or the "residual"), and only 29% has been due to labor
growth and 18% to capital growth.2
Another way of looking at this matter is through an analysis of labor productivity.
For manufacturing, average labor productivity in Taiwan grew at 7.38% a year in
1953-1961,9.15% in 1962-1971, and 3.84% in 1972-1981. For the entire period from
1953 to 1981, it grew at an average rate of 6.77% a year.3
Undoubtedly, the increase in the capital-labor ratio contributed to a part of the
increase in labor productivity. For during the period 1953-1980, it increased at 2.8%
a year.4 Admittedly, data on capital and employment are weak, but the increase in
labor productivity cannot be totally accounted for by the rising capital-labor ratio.
The improvement in the quality of the labor force must also have contributed to
the increase in labor productivity. Education has played a crucial role in the improve-
ment of the quality of the labor force. The cognitive and affective skills produced by
education have contributed to efficiency in the production process, increased the labor
force participation rate of women, and facilitated labor mobility, all contributing to
improvement in labor productivity.5 Available evidence suggests that the earnings of
employees and workers are positively correlated with educational attainment and that
earnings are basically determined by labor productivity as labor market is, or at least
has been, highly competitive.6
The increase in the educational level of the labor force in Taiwan has been
remarkable. In 1952, 42.1% of the population aged six or older were illiterate. The
proportion dropped to 14.7% in 1970 and 7.4% in 1988. The proportion of the pop-
ulation 6 years of age or older with a secondary education was 8.8% in 1952, 26.5% in
1970, and 44.2% in 1988. The proportion of the population 6 years of age or older with
a higher education was 1.4% in 1952, 3.7% in 1970, and 10.1% in 1988.7
Technology alone can never make a contribution to production, however crucial
it may be. It has to be embodied in the labor force or capital before it can become
effective. It is the sharp improvement in the quality of the labor force and the fast accu-
mulation of capital that has enabled technology to play such a key role in the economic
growth of Taiwan.8
But what has brought about the improvement in technology? And who has intro-
duced this new technology to Taiwan?
THE INITIAL BASE
It is sometimes suggested that a principal reason Taiwan has been able to achieve what
it has economically was the solid economic foundation that Japan had laid down
before withdrawing from Taiwan in 1945. This view is, however, unfounded.
During the period of Japanese occupation from 1895 to 1945, Taiwan was always
regarded as a colony that supplied raw materials and agricultural products to Japan
and as a market for Japanese manufactured goods. There was never any intention on
the part of Japan to industrialize Taiwan. Culturally, Japan aimed at replacing local
and Chinese values, institutions, and traditions with Japanese ones.
Thus, the Chinese people as well as the native aborigines had to learn the Japanese
language, and very few of them were allowed to receive schooling beyond the high
school level. Those who were lucky enough to go to a college or a university were
allowed to study only agriculture or commerce; they were riot allowed to study engi-
neering or the natural sciences. It is obvious that the Japanese rulers did not want the
Chinese or the aborigines to acquire advanced industrial knowledge or technology.
Consequently, for the few industries that there were, they were completely under
Japanese management and control. When Japan was defeated in 1945, virtually all
the Japanese soldiers and citizens left Taiwan for Japan, and took with them their
industrial technologies.
This is not to say that Japan did not make any contributions to Taiwan during
the period of their colonial rule. Japan did set up transportation, electrical, and public
health facilities and improved agricultural technology. Furthermore, under Japanese
rule, primary education was widespread.
The fact that the Chinese people were forced to learn the Japanese language has
also had a positive effect on Taiwan's economic development since 1945. Many indus-
trial leaders in Taiwan grew up during the Japanese occupation. Their ability to speak
and read the Japanese language and their knowledge of Japanese culture have greatly
helped to facilitate Japanese investment in Taiwan and Taiwan-Japan trade, and to
acquire Japanese technology in the course of Taiwan's economic development over
the past 40 years or so.
It should be noted that Taiwan got a great deal of help from those Chinese who
left the Chinese mainland for the island, especially after 1949. More than 2 million
Chinese mainlanders came to Taiwan with the Chinese National Government. Many
of them were soldiers, but a number of them were teachers, professors, businessmen,
industrialists, and government officials. They were the cream of the crop of their
respective groups and had been trained and had acquired experience in industrial
management and technology, education, administration, and the like while in the
mainland. There is little doubt that these people made a great contribution to the
increases in industrial technology and development in Taiwan.
But still, the industrial base in Taiwan in the 1950s was extremely weak. All that
Taiwan was capable of exporting was principally sugar, rice, and bananas. Per capita
GNP was only U S$ 145 in 1951, much less than that of many contemporary devel-
oping countries in Asia and Latin America. In 1988, per capita GNP was U S$6333,
and, in 1989, it was estimated to be U S$7571, much higher than that of many coun-
tries that were ahead of Taiwan in 1951.
THE ECONOMIC ENVIRONMENT AND THE CHARACTERISTICS OF THE FIRMS

IN TAIWAN
After the nationalist government retreated from mainland China to Taiwan, the gov-
ernment took over those enterprises established by the Japanese. Taiwan's economy
was thus dominated by publicly owned enterprises, encompassing the power genera-
tion, petroleum, shipbuilding, sugar and agriculture products sectors, and so on. Fur-
thermore, the statistics show that publicly owned enterprises accounted for 56.2% of
total industrial value-added in 1952. There is no doubt that in those days, the publicly
owned enterprises in Taiwan had capital, manpower, and technology.
By the late 1950s to the early 1960s, the situation began to change. The import
substitution policies adopted in the 1950s had proven that they could not sustain rapid
economic growth in a limited local market, like that of Taiwan. The efficiency problem
of the publicly owned enterprises together with low factory utilization rates also called
for readjustment. To react to this, the government adopted an export-promotion pol-
icy. Under this policy, the sluggish economy was not only revitalized but the policy
also changed the industrial structure in Taiwan's economy.
The transformation of the industrial structure in Taiwan can be understood from
the following statistics: the share of agriculture in net domestic product declined from
38.3% in 1953 to 18.0% in 1970, 9.2% in 1980 and 6.1% in 1988. The corresponding
share accounted for by industry (manufacturing, construction, electricity, gas, and
water) increased from 17.7% in 1953 to 34.5% in 1970, 44.7% in 1980, and 46.2% in
1988. In terms of employment, the share of agriculture was 55.6% in 1953, 36.7% in
1970, 19.5% in 1980, and 13.7% in 1988. The corresponding share of industry
increased from 17.6% in 1953 to 28.0% in 1970, 42.4% in 1980, and 42.6% in 1988.
From an international trade point of view, it was in Taiwan's comparative advan-
tage to develop labor-intensive industries, as they can be operated on a smaller scale
with less capital investment. Therefore, under the export promotion policy, the small-
scale private enterprises concentrating on labor-intensive industries began to flourish
and quickly became the backbone of Taiwan's economy. The importance of the pri-
vate sector in Taiwan can be seen in the following statistics: the share of private-enter-
prise value-added was 56.2% in the manufacturing sector in 1960; the comparable
ratios were 84.7% for 1971 and 89.8% for 1987. In 1987, the publicly owned enter-
prises accounted for only 10.2% of the total value-added in the manufacturing sector,
in dramatic contrast to what it was in the 1950s.
But the role of the publicly owned enterprises was also critical during the above
development process. Generally speaking, publicly owned enterprises have played a
supportive role in the development of private enterprises. These enterprises have often
served as upstream suppliers providing stable, reliable, and low-cost raw materials to
the downstream private manufacturers. Therefore, the upstream publicly owned
enterprises are actually integrated with the predominantly private downstream man-
ufactures whereby supporting their development.
The strength of the SMEs in Taiwan can be understood from the following sta-
tistics: in 1985 there were 119,073 enterprises in the manufacturing sector in Taiwan,
however, 98.8% of them were considered to be SMEs for their paid in capital was less
than 1 million U. S. dollars; furthermore, these SMEs accounted for 28.9% of total
domestic sales and 71.1% of total earnings from export sales. Clearly, SMEs in Taiwan
are highly export oriented.
The implications of SMEs on technology development are as follows: first,
because most of the SMEs in Taiwan were developed initially based on some simple
labor-intensive processing technologies, their technology levels are generally rather
low. But as the unemployment rate dropped to less than 3% in 1971 and the economy
reached the full-employment level, the rapid increase in wages has gradually weakened
the comparative advantage of the SMEs in the labor-intensive industries. In fact, the
situation is getting worse as the shortage in unskilled labor became obvious in the
1980s. Therefore, the shortage of unskilled labor accompanied by soaring wages has
exerted tremendous pressure on the SMEs either to upgrade their technological level
or to produce higher value-added products, or both. Clearly, the SMEs in Taiwan are
now facing greater technological challenges than ever before.
Second, most of the SMEs in Taiwan have devoted relatively little resources to
research and development (R&D) due to their limited amount of capital and man-
power resources. The question then is how can these SMEs compete with the other
major players in the world? Furthermore, given their limited technological capability
and limited endowment of resources, they have very little to offer to major foreign
players in the forming of strategic alliances in order to gain access to advance tech-
nology.
There is no doubt that both the entrepreneurs and the government in Taiwan
must react positively to these problems given the rapid technological development
throughout the world. In the next several sections, we wish to discuss various policies
adopted by both enterprises and the government to react to these challenges.
SOURCES OF TECHNOLOGY AND THE ROLE OF FD1 ENTERPRISES
As we mentioned above, due to the various difficulties that SMEs might encounter
during the process of technological development, one of the easiest ways for them to
develop technology is to become affiliated with a major (foreign) manufacturer thus
becoming their OEM (original equipment manufacturing) supplier. Many of the
SMEs in Taiwan arc satisfied with such an arrangement because they not only make
an acceptable level of profit but also avoid the risk involved in developing the tech-
nology by themselves.
The SMEs who are engaging in OEM or "technological cooperation" with foreign
firms in Taiwan frequently form alliances with Japanese firms. Taiwan's historical
background, as we discussed before, is one good reason for such a phenomenon. In
terms of foreign direct investment (FDI) in Taiwan, beginning in the early 1980s,
Japan topped the United States and became the largest FDI country investing in Tai-
wan.
By looking at FDI statistics, it can be seen that during the period 1952 to 1987,
electronics, metal products, chemical products, and machinery and instruments were
the top four industries that attracted the majority of FDI in Taiwan. In terms of the
average amount of investment, the European firms invested the most, followed by
U. S. and Japanese firms.
FDI enterprises in Taiwan can diffuse their technologies to domestic firms
through various direct and indirect channels. One of the most typical channels for
technology diffusion, from the FDI to the domestic firms, is labor mobility. One study
showed, among the 161 surveyed firms in Taiwan in 1987, that hired former FDI-firm
workers working as managers or technicians believed that the worker's former expe-
rience in FDI firms made a real contribution to strengthening management technol-
ogy, product design, and marketing. More specifically, among these surveyed firms,
96.3% considered that those workers can make a contribution to improving their man-
agerial technology, 85.1% of them considered that FDI workers make a contribution
that improves product design, and 75.8% of them considered that FDI workers can
make a contribution by improving marketing technology.9
Besides FDI and technical cooperation with foreign firms, there are many other
sources whereby domestic firms can acquire technology. A large-scale survey was con-
ducted by the Directorate-General of Budget Accounting and Statistics (DGBAS) of
the Executive Yuan (the government) in Taiwan to study the major sources of tech-
nology of 4226 firms in manufacturing in 1985. As we can see from Table 12.1, there
are 9 possible sources of technology that were listed in the survey. Of the 4226 firms
62.96% of them considered that the firm's own R&D was their major source of tech-
nology, whereas 30.64% of them considered that their technology came mainly from
abroad by way of purchasing formulas and authorization or physical plants, foreign
technological cooperation, foreign consultants, improving products of other coun-
tries, etc.
In Table 12.1 we organized the firms by industry and by the degree of foreign/
domestic ownership. Domestic firms are denoted by DOM, firms with foreign own-
ership of less than 50% by FL50, and firms with foreign ownership of over 50% by
FG50. As it is shown in this table, only 25.61% of DOMs consider foreign technology
sources as their major channel through which to acquire technology; instead, they are
more dependent on their own R&D.
Table 12.1. Percentage Distribution of Major Sources of Technology of the Firms by Industry
and by Foreign Ownership"
Domestic
Technology Foreign Technology
Total
N (%) A B C Subtotal D E F G H I
Total 4226 100 62.96 1.17 5.22 30.64 2.59 1.15 11.13 7.30 8.47
DOM 3699 100 67.35 1.28 5.73 25.61 2.32 0.96 8.05 5.39 8.89
FL50 207 100 38.08 0.87 3.78 57.26 3.20 0.58 31.10 14.82 7.56
FG50 320 100 35.95 0.22 0.89 62.90 4.94 3.59 28.09 21.57 4.71
"A, firm's own research and development; B, purchasing domestic patents; C, joint research with local research institutes; D,
purchasing formula and authorization; E, purchasing the entire plant; F, foreign technology participation; G, consultancy by
foreign organization; H, improving products of other countries; I, others,
Source: San (1988).
However, it is important to note that among the various sources of foreign tech-
nology, item H, namely improving products of other countries, was the most impor-
tant source of foreign technology for these domestic firms. As it was vaguely defined
in the survey, we suspect that activities such as imitating, copying, or limited improve-
ments on the existing foreign product (i.e., various reverse engineering tactics), were
their major sources of acquiring foreign technologies. Furthermore, it is reasonable to
suspect that many of the surveyed firms consider their reverse engineering efforts as
genuine R&D efforts, which undoubtedly overly exaggerates their own R&D efforts.
In contrast, firms with foreign capital participation such as FL50 or FG50 were more
heavily dependent on foreign technology through the means of foreign technology
cooperation and foreign consulting organizations and less on reverse engineering. It is
particularly significant for those FG50 firms, as their share of reverse engineering tech-
niques under item H was very low.
In terms of industries, for those more labor-intensive and less technology-inten-
sive industries, such as food processing, garment, bamboo and wood products, and
paper and printing, the DOMs had lower foreign technology participation. In contrast,
for industries with higher technology intensity or more capital intensity, generally the
DOMs had higher foreign technology participation. For instance, over 30% of the
DOMs in industries such as chemical products, machinery, electronic machinery,
transportation equipment, and precision instruments reported that their major tech-
nological sources came from abroad.
The above survey results can also be examined by looking at the size of the firm.
If we consider the firms with total employees less than 100 as small firms, denoted as
Small, firms with total employees between 100 and 500 as medium-size firms, denoted
as Medium, and firms with total employees over 500 as large firms, denoted as Large,
the statistics in Table 12.2 show that the larger the firms, the more capable that they
were in attracting foreign technology, as the ratios of firms who considered foreign
technology as their major technology sources are. 13.39, 27.34, and 34.45 for Small,
Medium, and Large, respectively. It should be noted that Small firms acquired tech-
nology from abroad least frequently for all of the manufacturing industries except for
the chemical products, metal, machinery, electronic machinery, transportation equip-
ment, and precision instruments industries. Small firms in these industries had higher
foreign technology source ratios, but these industries are also more technology-inten-
sive or capital-intensive.
Table 12.2. Percentage Distribution of Major Sources of Technology of the Firms by Size of
the Firm."
Domestic
Technology Foreign Technology
Total
N (%) A B C Subtotal D E F G H 1
Total 4226 100 60.82 1.13 5.04 33.00 2.50 1.11 10.75 7.05 8.18 3.04
Small 2277 100 32.42 0.68 2.23 13.39 1.12 0.40 3.34 1.86 4.38 2.30
Medium 1597 100 42.76 0.63 3.89 27.34 2.13 1.06 10.16 6.45 5.98 1.56
Large 352 100 34.05 0.65 4.98 34.45 1.57 1.05 14.14 12.70 3.54 1.44
Source: this study.
"For definitions of A-I, sec note a to Table 12.1.
By examining Small firms in the metal, machinery, electronic machinery, trans-
portation equipment, and precision instruments industries, it is clear that item H (i.e.,
improving products of other countries in terms of various reverse engineering tech-
niques) is the major source of foreign technology. Clearly, for those Small firms in the
more technology-intensive industries in Taiwan, reverse engineering is still the key to
acquiring technology. These firms are still far away from becoming an inventor of
technology.
The larger firms have better capability in finding foreign technology cooperation
partners and in inviting foreign experts as technology consultants, therefore they rely
more on these two channels to acquire foreign technology. As it was found in the sur-
vey, Medium and Large firms in the manufacturing industry in Taiwan were the most
common entities that either engaged in technology cooperation with foreign partners
(item F) or hired foreign consultants as their major source of foreign technology.
Among the three major domestic sources of technology, namely, the firms' own
research and development (A), purchasing domestic patents (B), and joint research
with local research institutes (C), it is clear that the firms' own R&D in reverse engi-
neering efforts was their most important source of technology. However, it is also clear
that due to limited resources and/or diffusion channels, Small firms relied more on
their own R&D efforts whereas Medium or Large firms more on conducting joint
research projects with local research institutions to acqurie needed technology.
It is thus clear from data in Tables 12.1 and 12.2 that domestic firms in the man-
ufacturing sector in Taiwan rely more on their own efforts in acquiring technology,
whereas FDI firms in Taiwan rely more on foreign technology sources. In terms of the
size of the firm, small firms rely more on their own R&D as their major domestic tech-
nology source, or, alternatively, on reverse engineering techniques such as imitation,
copying, or improvement of existing foreign products. The larger firms by virtue of
better contact with domestic research institutions and in forming alliances with foreign
corporations depend on their own R&D efforts, conducting joint research projects
with domestic research institutions, engaging in technology cooperation with foreign
partners, and/or the hiring of foreign consultants or experts as their major acquisition
channels for domestic and foreign technology.
From the above discussion, it can be seen that the firms' own R&D efforts, assis-
tance offered by local research institutions, technology cooperation with foreign part-
ners, foreign consultants, and reverse engineering techniques such as imitation or
improvement of existing foreign products are the five major sources of technology for
manufacturing firms in Taiwan. Under such circumstances, the role that the govern-
ment plays in technical advance is critical, for government policies are closely involved
with R&D activities, government-sponsored research institutions, regulation of tech-
nology cooperation with foreign partners, and the protection of intellectual property
rights. We shall now turn to the role of the government in supporting industrial tech-
nology advancement.
THE ROLE OF THE GOVERNMENT
The government has played a very active role in the upgrading of industrial technology
in Taiwan. In 1959, a "Plan of National Long-Term Development of Science" was
adopted as the basic guideline of development regarding science and technology. To

carry out the plan, the Council on National Long-Term Development of Science was
founded with the President of Academia Sinica as its Chairman and the Minister of
Education as its Vice-Chairman. The chief function of the Council is to design and
carry out long-term research in the sciences. In 1967, this Council was expanded and
reorganized as the National Science Council, which has become, and still is, a principal
government organization that designs strategy and promotes scientific research.
But the National Science Council is only one of the government agencies that
does research and promotes science and technology. There are many other research
organizations as shown below. Many of the research organizations are directly under
the Ministry of Economic Affairs, the Ministry of Communications, and the Ministry
of Defense (development of agricultural technology is not discussed in this chapter).
Broadly speaking, the objectives of the various government plans and organiza-
tions to upgrade science and technology are 3-fold: (1) to improve the quality of life of
the people especially in the area of health and to strengthen environmental protection;
(2) to transform Taiwan's economic structure from being labor intensive to capital
intensive and technology intensive; and (3) to develop industries necessary for a
national defense system independent of foreign control.
As for the areas of scientific research, the following eight fields have been chosen
as keys: (1) energy, (2) materials, (3) information, (4) automation, (5) biotechnology,
(6) electron-optics, (7) hepatitis control, and (8) food technology.
At the policy level, the government has adopted (1) a strong education policy; (2)
fiscal and financial policies to encourage and indeed to require business enterprises to
do R&D; (3) a policy to establish government-controlled research organizations and
institutes, and to support the development of high-tech industries; (4) a policy to
engage in technological alliance with foreign partners to develop needed technology;
and (5) technology policies to allow Taiwan to better react to the globalization of the
world's economy.
EDUCATIONAL POLICY
There is little question that there has been very rapid educational growth in Taiwan;
that is a reflection of not only government policy but also a strong desire for education
on the part of the people that is a deep-rooted traditional value. The annual rate of
increase of total student enrollment for all age groups was 4.87% for the 30 years from
1950 to 1979, the crucial years in Taiwan's development. This was substantially higher
than the rate of increase in the school-age population. As a result, the enrollment rate
(number of students as a percentage of the school-age population) went up from 36%
in 1951-1955 to 57% in 1976-1979.10
It should also be noted that the growth rate became higher as the level of educa-
tion became higher. That is, the growth rate of higher education (in terms of total
enrollment) was higher than that of secondary education, and the growth rate of sec-
ondary education was higher than that of elementary education. This holds true both
before and after the adoption of the 9-ycar free-education system in 1968.
Total cost per student (as measured by total expenditures per student) increased
substantially in real terms for all levels of education (except private colleges and uni-
versities), indicative of the improvement in educational quality. Higher education's

per student cost went up by the greatest amountexpenditure per student at public
colleges and universities in 1980 was nearly 3 times as much as in 1952.
The percentage of total government expenditure for education went up from
about 10% in the early 1950s to about 15% in the late 1970s. The percentage of public
education expenditure in GNP went up from less than 2% in the early 1950s to more
than 4% in the late 1970s.
It also should be noted that secondary education in Taiwan may be divided into
three types of schoolsgeneral high schools, normal schools, and vocational schools.
Promoted actively by the government, the role of vocational schools grew quickly in
importance. In the 1983-1984 school year, 202 vocational schools were imparting
technical skills to 404,549 students. The percentage of vocational school students was
also growing as well. In 1970, the ratio of general high school students to vocational
school students was 1:1, but in 1980 it was 1:2.
Over the years, the demand for higher education in Taiwan has been extraordi-
narily strong. It remains so today. No one reason exists to explain adequately this
strong demand for higher education, or more accurately, a university or college
diploma. The perceived financial returns, gains in social status, and upward mobility
associated with a diploma are reinforced by an important set of social and cultural
factors. A diploma is seen as the first link on the path to power, glory, and wealth or,
in other words, to joining the elite of society.
The heightened expectations associated with college diplomas create difficulties
in the government's efforts to promote vocational schools and in helping plug the
growing requirements of economic growth. To induce more junior high school stu-
dents to enter into vocational schools, rather than the regular high school, a technical
college that offers a BA degree was established in 1974. Thus students in this college,
who come from the vocational track, can also obtain the coveted BA degree.''
A strong education policy, of course, represents investment in human resources
or human capital. It means that the number of people who are capable of doing useful
research on industrial technology increases and the number of people who can apply
research results in industrial production increases. This is of course what successful
economic development requires.
While the government is actively engaging in promoting the quality of its human
resources, Taiwan is also encountering a "brain-drain" problem, that is, a number of
those who received education, especially higher education, have gone abroad for
advanced studies and never returned. This is particularly true for students in the fields
related to industrial technology such as engineering. For instance, in 1979, the total
number of students in the field of engineering who returned to Taiwan after studying
abroad was less than that of students who majored in the same field and went abroad
for advance study by a margin of 1285 persons. This number accounts for 20.6% of
the total number of students who graduated in the field of engineering in 1979 in Tai-
wan. If we take the above ratio to gauge the brain-drain problem (hereafter we shall
call it the brain-drain ratio), then the brain-drain ratios for students in the fields of
education, natural science, engineering, medicine, and agriculture were 10.0, 15.5,
3.8, 27.9, 20.6, 19.2, and 31.9%, respectively, in 1979.
In response to the brain-drain problem, the government in Taiwan has adopted
various policies to attract those who study abroad to return home to work and live in
Taiwan. These policies include providing for visiting professorships at universities,
and providing more flexible employment channels for Ph.D. degree holders, which
allow for the circumvention of the rigid public-servant regulations required in order
to join the government in Taiwan. However, the preferential treatment has not proved
to be very effective.
As Taiwan rapidly transformed herself into a newly industrialized economy, the
brain-drain problem seems to have improved significantly. The brain-drain ratios for
the fields of law, social science, education, natural science, engineering, medicine, and
agriculture were -0.5, 9.4, -1.6, 16.2, 16.8, 8.2, and 19.2%, respectively, in 1984.
The most fundamental factor behind the rapid improvement in the brain-drain prob-
lem was not the government policies we described above, since the public sector and
the universities can provide only a limited number of positions, but the strong demand
for high-level manpower in the private sector. In the 1980s, as Taiwan was rapidly
losing her comparative advantage in low-wage labor-intensive industries, the country
gradually moved into technology-intensive or higher value-added industries. Under
such circumstances, the demand for better technologies became fierce, which led to a
stronger derived demand for a higher level of engineers who were educated and trained
abroad. Furthermore, many overseas Chinese, after having accumulated many years
of work experience and technological capability, began to set up their own enterprises
in Taiwan because it is their homeland. Therefore, beginning in the mid-1980s, we
observed a reverse in the brain-drain situation in Taiwan.
FISCAL AND FINANCIAL POLICY TO PROMOTE TECHNOLOGY DEVELOPMENT
High expenditure requirements together with the high risk of development often has
become the major obstacle of firms in conducting R&D in many technology-intensive
industries. This problem is particularly significant as Taiwan's various R&D indica-
tors are significantly lower than those of many developing and developed countries.
For instance, the total civilian R&D expenditure to GNP was about 0.53% during the
period of 1978 to 1982, which was significantly lower than the comparable ratios of
other countries such as 2.22% for Japan, 2.42% for the United States, 2.45% for West
Germany, 1.90% for the Netherlands, and 0.68% for South Korea. Furthermore, from
the statistics on R&D expenditure sources, it can be seen that in 1979 the government
share accounted for 52% of total ROC expenditures; in 1984 it was 51.5%. Obviously,
compared to many developed and developing countries, the government has shoul-
dered too much of the burden for conducting R&D. Therefore, how to promote R&D
in the private sector has become an important policy issue.
The government has adopted a number of measures to encourage business firms
to intensify their R&D efforts. On the taxation side, the Statute for Encouragement of
Investment (SEI) stipulates that the R&D expenses of a firm shall be deducted from
the taxable income of the current year, and for equipment bought for the purpose of
conducting R&D with a service life exceeding 2 years, accelerated depreciation is
allowed. Furthermore, if the amount of R&D expenses of a firm in a tax year exceeds
the highest amount of R&D expenditure in the five preceding years then 20% of the
amount in excess thereof may be deducted from the firm's income tax payable for that
year; however, the deductible amount shall not exceed 50% of the firm's income tax
payable for that year.
On the financial side, the government introduced the Assistance Program for
Strategic Industries (APSI) in 1982. The government initially selected 151 products as
strategic products for development. By December 1987, the number of products had
increased to 214. Almost half of the selected items were electronic or information
products. To implement this program, the government put aside NT$ 20 billion,
which is given to firms in the form of loans. These loans have been used for the instal-
lation of machinery necessary for the production of the items selected. Any qualifying
firm can receive a 10-year loan with a maximum loanable amount of 80% of the total
capital needed or 65% of the total expenditure required for the investment. Further-
more, a preferential interest rate, 1.75% below the prime rate of the Bank of Com-
munications, is also charged.
Besides the APSI and SEI, there are many other assistance programs that have
been introduced in the 1980s, such as the "600 million U S dollar Low-Interest Loan
to Promote Export Programs," the "Assistance Program to Small and Medium-Sized
Enterprises under the Sino-American Fund," and the "Assistance Program of Low-
Interest Loans for Production Industries to Set Up Satellite Plants." Furthermore,
there are other nonfinancial assistance programs that aim at providing consultancy
services for the upgrading of managerial technology. These programs have been intro-
duced by various government sponsored institutions (e.g., the China Productivity
Center).
In addition to the "carrot" intended to attract firms to conduct more R&D, as
described above, the SEI also specifically laid down minimum R&D levels for both
domestic and FDI enterprises that enjoy substantial preferential tax treatment under
the statute. Under the minimum R&D stipulations, if an enterprise's R&D expendi-
ture to annual total sales ratio is lower than the prescribed standard then the enterprise
is required to contribute the difference to a government controlled research and devel-
opment fund for financing collective research and development projects. Basically, the
required ratios vary according to the characteristics of the business and by the annual
sales value of the enterprises. The range of the required ratios is from 0.5 to 1.5%.
Enterprises in traditional sectors, such as the food and garment industries, have lower
ratios than those of the more technology-intensive industries.
Currently, the government is actively engaged in introducing a new Industrial
Upgrading Statute (IUS) to replace the SEI. The major difference between IUS and
SEI is that under IUS, firms will qualify to enjoy preferential treatment only if they
have shown that they meet the subscribed standards in certain areas. This includes
conducting R&D, compliance with environmental protection standards, engaging in
manpower training, and setting up of international marketing channels, whereas,
under the SEI, a capital investment of either a newly established or an expanding firm
is all that was necessary to enjoy tax-exempted preferential treatment. Clearly under
the IUS, the accumulation of capital is no longer considered the only factor in pro-
moting industrial upgrading in Taiwan.
How effective are these government policies in enhancing R&D activities in the
private sector? A study was made of 1406 firms in the shoe, leather, machinery, and
electronic machinery industries in Taiwan in 1987. The firm's opinions about the
most effective way in which the government should respond to promote a satisfactory
level of technology was studied.12 In this study, 9 possible policy options were listed,
namely (1) transferring technology through government sponsored research institu-
tions, (2) coordination among firms to do joint research, (3) introduction of new tech-
nology from abroad, (4) educating more R&D people, (5) helping firms to establish
their own brand names, (6) offering low-interest loans for R&D activities, (7) stan-
dardization of parts and components, (8) giving tax credits for R&D expenditures, and
(9) others. It was found that only 7.6% of the 1406 surveyed firms regarded the offering
of low-interest loans for R&D activities as the most effective way in which the govern-
ment should respond to promote a satisfactory level of technology in the economy.
Similarly, only 7.8% of the surveyed firms regarded giving tax credits for R&D expen-
ditures as the most effective measure. Clearly, the survey results have shown that the
majority of the surveyed firms in the sample do not consider that the tax or financial
incentives, which are provided by the government, are effective in promoting tech-
nological advancement.
What, then, are the most effective ways by which the government can promote
technology? The study found that 18.8% of the surveyed firms (which is the highest
percentage) considered educating more R&D people to be the most effective method
to upgrade technology, followed by coordination among firms to do joint research
(18.6%), the introduction of new technology from abroad (17.2%), transferring tech-
nology through government-sponsored research institutions (15.6%), and helping the
firms establish their own brand names (8.8%).
Clearly, the study shows that government policies geared toward educating more
R&D people, coordinating joint research efforts, and introducing and transferring
technologies are all critical to technology development in Taiwan, whereas financial
or tax incentive policies, as discussed above, do not prove to be particularly helpful.
In the next two sections, we shall take the information industry as an example to show
how the government has succeeded in promoting technology development.
GOVERNMENT-SUPPORTED RESEARCH INSTITUTIONS AND THE STRATEGY

TO DEVELOP THE HIGH-TECH INFORMATION INDUSTRY
Structurally, Taiwan's economy was built on a large number of SMEs, as mentioned

previously. Due to their limited amount of resources for conducting R&D, SMEs rely
heavily on the efforts of the government to develop technology and/or on government-
sponsored research institutions to transfer technology to them. Therefore, govern-
ment-sponsored research institutions are critical to technology development and dif-
fusion in Taiwan.
To help SMEs in the information industry to overcome the various problems
associated with conducting R&D, training, and marketing, the government has set up
two institutions to assist firms in the industry: the Industrial Technology Research
Institute (ITRI) and the Institute for the Information Industry (III). Ill was established
in 1979; its major mission includes introducing and developing software technology,
assisting government agencies and public enterprises in their computerization proj-
ects, training and educating information professionals, supplying market and tech-
nology information related to the information industry, and promoting the develop-
ment and usage of computer-related technologies. In sum, the major function of III is
to support activities that are undersupplied, much like the case for government-sup-
plied public goods having positive external effects, to the member firms in the industry.
More importantly, these activities, such as manpower training and the dissemination
of market information, can make up for disadvantages resulting from the economies
of scale that prohibit SMEs from conducting such activities.
The III programs complement ITRI's hardware development program for the
electronic/information industry in Taiwan. ITRI was set up in 1973 under the direc-
tion of the Ministry of Economic Affairs. There are five major divisions in ITRI that
support the development of various industrial technologies. As for the electronics
industry, the Electronic Research and Service Organization (ERSO) is critical to its
technology development. Basically ERSO has two major tasks: to develop needed
technology for the industry and to diffuse the developed technology among the indus-
try's firms. There are various mechanisms that can be used to diffuse the new tech-
nology, including issuing technical documents and organizing conferences for elec-
tronic firms in Taiwan; furthermore, the new technology can be transferred to
individual firms through licensing agreements where royalty charges are levied on the
recipient firm(s), once they employ the technology. If, however, the developed tech-
nology has marketing potential and there is the desire to set up a new joint venture to
disseminate the technology then a new spin-off venture company would be established
by ERSO. It is important to note that although the spin-off company is supported by
ERSO's engineers and its funding is provided by the government, venture company is
cautiously organized as a privately owned company. Both the technology endowment
received from ERSO and the capital endowment supplied by the government aim at
attracting investment from the private sector. It is hoped that the private sector can
account for at least 60 to 70% of the total shares of the newly established venture com-
pany. There are at least five advantages to this kind of setup. First from an engineers
point of view, the newly established venture company can provide immediate pecu-
niary rewards and/or professional positions, which allow them to further develop such
technologies; therefore, the incentive to develop new technologies is very strong. Sec-
ond, since the established venture company's new technology will be supported by the
same group of engineers who developed it, the company does not have to start from
scratch and can be in full operation within a very short period of time. Third, the com-
pany remains basically a privately owned company; therefore, it can be flexible and
can be responsive to rapid changes in the external environment. Fourth, one of the
main purposes for developing new technology is to help the private sector in Taiwan
to upgrade its technological position. The government has no intention of monopo-
lizing the technology. As a result, privately owned venture companies can serve as a
conduit to diffuse new technology. Fifth, and most importantly, many entrepreneurs
in the traditional sectors of the economy would like to diversify their investments to
include the electronic/information industry for various reasons such as risk sharing
and industrial upgrading. A venture company investment project would certainly be
an ideal investment and the easiest one to invest in. The mechanisms that allow ERSO
to diffuse its technology can thus be summarized in Figure 12.1.
As a result of the successful spin-off model developed by ERSO, described above,
since the early 1980s, six 1C chip companies began operation in Taiwan, namely UMC
(1982), Advanced (1982), TSMC (1987), Coin Tek (1988), Winbord (1988), and Ten
Figure 12.1. Mechanism for ERSO to diffuse its technology.
Tech (1989). They were set up in succession. Taiwan's major 1C chip manufacturers
in 1989 and their capacity can be summarized in Table 12.3.
It is important to note that prior to the DRAM shortage in 1988, most key indus-
trial policy makers and entrepreneurs were in agreement on the fact that Taiwan
should fully capitalize on their advantage of having a large number of SMEs in the
information industry. The development of ASICs would be an ideal strategy that
would fit in Taiwan's industrial structure. The relevant factors that favor development
of ASICs are the fact that the design of the integrate circuits can be handled efficiently
by SMEs and many talented and independent engineers would excel in the SME type
of 1C design firms. However, 1C design firms must be incorporated into the 1C fabri-
cation network in order to complete the industrial network demanded in the produc-
tion of custom chips. To facilitate this TSMC, a joint venture between ITRI and Phil-
lips was established in 1987. TSMC is not personally involved in 1C design but
fabricates custom 1C chips for various 1C design companies. As the result of TSMC
operations, the 1C design firms in Taiwan have increased from just a few in 1987 to
more than 50 in 1989. Clearly, prior to 1988 Taiwan's strategy to cope with the rapid
Table 12.3. Major 1C Producers in Taiwan in 1989
Production Sources of Sources of Manufacturing Monthly
Company Began in Capital Technology Technology Capacity Note
UMC January Domestic ERSO 1.2-5 ^m 4", 50,000
1982 CMOS 6", 1 5,000 The third
factory is
under
planning
TSMC July Taiwan and ERSO 1.25 Mm 6", 12,000 The third
1987 Netherlands CMOS factory is
under
planning
Coin Tek July Domestic ERSO 1.2-5 Mm 5", 30,000 The second has
1988 CMOS begun to
build in
1989
Winbord October Domestic ERSO 1.2-5 M m 5", 20,000 The second
1988 CMOS factory is
scheduled to
build in
early 1990
Advanced January Domestic ERSO 4-5 /um 4", 8,000
1986 CMOS
Ten Tech October Domestic ERSO 2-5 Mm 4", 8,000
1988 CMOS
Source: ERSO.
change in the 1C industry was to adopt an economy of scope strategy, and not an econ-
omy of scale strategy. The country concentrated her efforts on the production of ASICs
rather than standard memory chips such as DRAM. However, the worldwide DRAM
shortage problem of 1988 changed this strategy.
As the shortage of DRAMs became more and more severe in mid-1988, discus-
sions (or debates) on whether or not Taiwan should build up her own DRAM industry
were intensified. A whole series of debates and discussions dragged on for several
months in Taiwan. The potential risk or inherent difficulty in setting up a DRAM fab-
rication plant has not discouraged DRAM production proponents and some fruitful
results have been produced by the discussions. First, the government has agreed to put
DRAM on the list of important high-tech products. As a result of this policy, manu-
facturers of DRAMs can enjoy favorable tax treatment. Furthermore, non-DRAM
producers can now hold more than 40% of the total shares of a company that produces
DRAMs without any legal restrictions. Second, the president of TSMC has agreed to
devote part of its second plant, which is scheduled to be in operation in 1990, to the
production of DRAM, provided that the downstream firms can make a commitment
to purchase DRAMs from TSMCwhich they have. Third, two private information
corporations, Vitelic and Acer, have decided to invest heavily in the production of
DRAMs, because they feel that there is a potential market for DRAMs. As a non-
DRAM producer, Acer's strategy to get into the production of DRAMs is to make a
foreign alliance with TI. The Acer-TI DRAM fabrication plant joint venture is sched-
uled to be in operation by 1991. The forthcoming players in the 1C industry in Taiwan
can be summarized in Table 12.4.
Table 12.4. The Forthcoming 1C Manufacturers After 1989 in Taiwan

Monthly Major
Company Capacity Major Product Technology
Mosel 6", 6,000 BiCOMS Product, ASIC, BiCMOS
SRAM CMOS 1.0 M m
Holtek 5", 10,000 Custom 1C CMOS 2.0 ^m
Vitelic 6", 10,000 DRAM CMOS 1.0 Mm
Silicon Integrated System 6", 30,000 Mask DRAM
Acer-Tl 6", 30,000 DRAM
Macronix 6", 8,000 Communication ICs
Utic 4", 8,000 Linear 1C Bipolar
FORMOSA 6", EPROH
Source: ERSO.
Furthermore, to actively respond to rapid technological development in the

information industry, joint efforts similar to the SEMATECH project on submicro
technology in the United States have also been launched under the coordination of the
Ministry of Economic Affairs in Taiwan. In August 1989, ITRI, UMC, TSMC, Win-
bord, Ten Tech, and Vitelic, Taiwan's major 1C chips manufacturers, and other com-
panies in Taiwan decided to form a "technology development alliance" to develop
16M DRAM and 4M SRAM chip technology over the next 5 years. To be better pre-
pared to handle the project, a micro electronic laboratory that meets class 10 cleanli-
ness standards for 0.2 /urn, and that has the capability of developing the technology
needed to produce 0.8-jum chips, was completed. The 5-year development plan calls
for an investment of 5 to 6 billion NT dollars and requires more than 200 engineers
and technicians to be devoted to the project.
It is thus clear from the above discussion that Taiwan's information development
strategies in the 1990s will be a more balanced combination of economy of scale (such
as the development of DRAMs) and that of scope (such as the development of ASICs).
It is hoped that through this combination, Taiwan's information industry will become
one of the major world players in the industry.
TECHNOLOGICAL ALLIANCE WITH FOREIGN PARTNERS TO PROMOTE

COMPUTER SOFTWARE DEVELOPMENT: AN EXAMPLE
One of the most fundamental strategies to promote the country's software industry is
to build up its technological capabilities and develop experience in production. To do
this, technological alliances with the computer giants will be necessary. Under these
alliances, Taiwan can provide low-cost, quality, but relatively inexperienced, engi-
neers to foreign partners. In turn, the foreign partners can provide their vast experience
in software research and development. Clearly, both parties will mutually benefit from
the alliance. The establishment of the Neotech Development Corp. (NDC) is a typical
example of such a strategy.
NDC was established in 1983 with the support of III and IBM. Ill and IBM agreed
that NDC would provide exclusive software development and design services to IBM.
On assigned projects, IBM pays both the NDC staff s salaries and the bills for all R&D
equipment. During the project-development period, the NDC research staff frequently
are sent to IBM plants around the world to acquire needed expertise. At the same time,
IBM's engineers also are sent to NDC to guide research operations. It has been
reported13 that over the period of 1983 to 1988 that IBM paid 17 million U. S. dollars
to NDC to complete various projects for IBM, primarily in the categories of infor-
mation design and the development of application software (e.g., control boards for
the IBM 5216 daisy wheel printer, peripherals for banking computers, peripherals for
retail-house computers, and banking software). As for NDC engineers, they not only
have obtained hardware facilities support from IBM but also have obtained knowledge
on R&D procedures and testing techniques that could not have been easily mastered
through the engineers' own efforts, but that are vital to the success of product devel-
opment projects.
The success of NDC's relationship with IBM has led to another cooperation pro-
gramthe establishment of International Integrated System Inc. (IIS) in 1988. Under
an agreement with IBM, III put up 3.5 million U. S. dollars in capital and dispatched
70 software designers to work at IIS. IBM provided the necessary technical support for
the designing of IBM software. The major difference between the IIS and NDC agree-
ments is that IIS does not have to provide contract services to IBM exclusively, as NDC
does. IIS can also provide services, such as VAN, database, and artificial intelligence
services, to other independent businesses. However, the design project for the business
must solely utilize IBM systems.
Clearly, a strategic alliance between IBM and III is mutually beneficial to both
parties, since IBM can enhance its global competitiveness through the alliance, while
its Chinese counterpart can significantly upgrade its technological capability of its soft-
ware/design engineers. Also technology can be diffused to other private firms by the
inevitable turnover of engineers. The latest development of NDC and IIS is that IBM's
Taiwan branch has formally proposed to the board of III to merge NDC and IIS into
one single company called Integrated Systems Development Corp. (ISDC). IBM has
agreed to purchase 35% of ISDC shares. Ill will hold 45% of the shares and the remain-
ing 20% will be held by the firm's employees. The proposal has been approved by the
board of III.
Basically ISDC will engage in the kind of business that NDC and IIS have been
engaging in all along. However, by combining both companies' software/design engi-
neers, it is hoped that ISDC will have both economy of scale and the technical expertise
to become a significant software/design player in the international market.14
THE INTERACTIONS AMONG VARIOUS GOVERNMENT-SPONSORED

INSTITUTIONS AND UNIVERSITIES
Besides ITRI and III, which are directly related to industrial technology development
in Taiwan, there are in the administrative structure of the government a number of
institutes, councils, and the like that are charged with the responsibility of advancing
science and technology in general and industrial technology in particular. They are
briefly listed and described below.
I. Those that report directly to the President's office:

1. Guidance Committee on the Development of Science
This is an advisory body to the President.
2. Academia Sinica
The highest academic institution in the country, which does research in var-
ious fields.
II. Those under the Executive Yuan:
1. Committee of Advisors on Science and Technology
The members are now primarily scientists from other countries.
2. Committee on Research and Development of Applied Technology
3. National Science Council
Under the direction of this Council, there is a data center for science and
technology, a development center for precision instruments, and a bureau
in charge of industrial parks.
4. Council of Atomic Energy
Under this council, there is a research center for atomic energy.
5. Council of Agricultural Development
6. Ministry of Economic Affairs
Under this Ministry, there are 36 institutes that do research in various fields
of industrial technology.
7. Ministry of Communications
Under this council, there are four research institutes doing research related
to transportation and communications, including weather research.
8. Ministry of Education
The Ministry is in charge of all colleges and universities, many of which
have research institutes.
9. Administration of Public Health
Under the administration, there are two research institutes.
10. Ministry of Defense
Under this ministry, there is the important Chung-Shan Science College
that does research on technologies related to national defense.
11. Taiwan Provincial Government
Under provincial government control there are 18 research institutes doing
research largely on agriculture, public health, and transportation.
Of all the above research institutions, the National Science Council plays the key
role in the development of science and technology. It is charged with designing
research strategy and plans, promoting basic research, pioneering applied research,
improving the environment for research, cultivating and attracting research person-
nel, and coordinating research work undertaken by the various government agencies.
It is also the principal grantor of funds to researchers at Academia Sinica and in Tai-
wan's colleges and universities.
The division of labor structure in technology development in Taiwan is shown in
Table 12.5. In this table, it is shown that basic research is conducted mainly by Aca-
demia Sinica and-the universities, while applied research, technology development,
and commercialization of the technology are mainly the responsibility of various
industrial technology research institutions as well as the enterprises themselves. Under
Table 12.5. The Division of Labor to Promote Technology Development in Taiwan

The Implementation of the Policies
Schools and
Type of Research Government-Support
Research Policy Making Institutions Research Institution Enterprises
Basic Academia Sinica Academia Sinica
Applied research Ministry of Universities
Education
National Science
Council
Technology Various ministries Various direct ITRI Private enterprises
development under the ROC research III Public-owned
Commercialized government institutions Other government- enterprises
and applied under the sponsored research
different institutions that are
levels of not directly under
government the government
Source: National Science Council, Taiwan, ROC.
the above division of labor structure, one will find both advantages and disadvantages.
The major advantage of the system is that limited R&D resources are pooled and the
research effort can then be aimed at developing generic types of technology that can
be beneficial to all the firms in the industry. In contrast, the major disadvantage of the
system is that firms are more likely to rely heavily on the government as their major
source of new technology.
To overcome such a deficiency, as we discussed before, policies, which rigidly
require firms to meet the minimum R&D input levels if they wish to qualify for the
tax exemption provision as stipulated in the SEI, have been introduced. It is thus
hoped that given the SME economy in Taiwan, a more balanced R&D mixture
between the government and private enterprises can be achieved.
THE INTERACTIONS BETWEEN FIRMS AND THE GOVERNMENT ON

TECHNOLOGY DEVELOPMENT TO COPE WITH THE GLOBALIZATION OF THE
WORLD'S ECONOMY
Due to the rapid increase in the volume of trade and investment in the world and
owing to the tremendous development in science and technology during this century,
the world's economy is rapidly becoming more and more integrated into a single-econ-
omy structure. However, as the globalization of the world's economy increases in
momentum, both the competitive and complementary relationships, regarding tech-
nology, among firms and countries will be intensified. Therefore, how to properly react
to this inevitable trend becomes critical. From the entrepreneur's point of view, one
of the important strategies of firms in Taiwan, to cope with the globalization trend in
the world economy, is overseas mergers. This strategy is becoming more important
and is a common practice for many firms in Taiwan. There are several important eco-
nomic factors that contribute to the adoption of this strategy. Financially speaking,
huge foreign exchange reserves make overseas mergers feasible. Furthermore, with the
appreciation of the NT dollar over the period of 1986-1988, overseas mergers have
been made more affordable and attractive than exporting.
Besides the financial factors, acquiring needed technology is another major moti-
vation for overseas mergers. The overseas mergers taken by the Acer group with Coun-
terpoint Computers, a minicomputer manufacturing company, and PPL, a computer-
control printing system company, both of which are located in the United States, is a
typical example of a merger of this type. These mergers have helped Acer acquire mini-
computer technology and expand her personal computer functions. The Acer group's
overseas mergers are certainly not the only example of mergers being conducted by
firms in Taiwan; another example is Taiwan Microtex's overseas merger with the
U. S.'s Mouse System Company to enhance its technological capability.
It is frequently argued that one of the major disadvantages of SMEs in Taiwan is
that most of these firms are too small to have their own marketing channels let alone
have their own brand names. Therefore, the future competitive capability of the SMEs
is somewhat doubtful. Many entrepreneurs in Taiwan find that overseas mergers are
one of the most feasible ways to make up for this shortcoming. As a matter of fact,
some entrepreneurs have even considered that it is more effective and more conve-
nient to buy or to control a foreign marketing company that already has a marketing
network set up than to start a marketing network from scratch and try to build one by
themselves. Acer group's merger with DYNA, the third largest computer dealer in the
United States, is a typical example of this type of marketing consideration.
Besides building a marketing channel, many of the SMEs in Taiwan find that
overseas mergers with an existing company can also help them "purchase" a brand
name that they can use for their products. A good example of this kind of merger is
where one of the largest U. S. furniture companies, Stoneville, and her long-time OEM
supplier in Taiwan merged so that the Taiwan-manufacturer could take control of
both Stoneville's marketing channels and name. Clearly, Taiwan's SMEs have actually
engaged in building brand names for their products as well as marketing channels
through overseas mergers.
In sum, overseas mergers are appearing not only as a result of huge foreign
exchange reserves and the vast appreciation of the NT dollar but are also due to the
fact that it can help the domestic enterprises to vertically integrate themselves and take
over a foreign firm's marketing channels, brand names, and technology. This is, obvi-
ously, a complementary and feasible strategy for domestic firms in coping with the
globalization of the world's economy.
The government's policy toward overseas mergers is undoubtedly supportive, in
fact, actions to promote these activities have already been taken. One of the major
policy tools of the government is the providing of low-interest loans or part of the
investment capital needed so that private companies can engage in overseas mergers.
A typical example of this strategy is the case of WYSE Technology, a major terminal
manufacturer in the United States, founded by an overseas Chinese company, and
which dominates around 50% of the world's terminal market. Due to the company's
diversification but unsuccessful strategies in producing personal computers, WYSE
ran into serious financial trouble and looked for a strategic alliance with another firm
to help the company both in the production of personal computers and also to help
out with the company's finance. Mitac, one of the major personal computer manu-
facturers in Taiwan, was approached by WYSE and responded positively. To facilitate
the merger, a venture company was quickly set up. Private capital from Mitac, China
Trust, and Kuo Chiau was gathered and put in the venture company. The funds to
finalize the merger, which came from these companies, were roughly 80% of the total
needed to complete the merger and the remaining 20% of the funds came from the
government's Central Development Fund. The merger of WYSE was thus completed.
The successful experience of the WYSE case has certainly served as an important
example and has set a precedent for future government involvement in supporting
overseas mergers.
One must admit, however, that Taiwan overseas mergers are still in the prelimi-
nary stages of development. Since the investment banks in Taiwan are not developed
to support overseas mergers very well, there is a lack of experienced international legal
and accounting services and overseas mergers face many obstacles. Regardless, over-
seas mergers have become an important strategy of domestic firms in Taiwan. Also,
overseas Chinese, in some cases but not all, will continue to play an important role in
facilitating such activities.
In view of the significant role that the overseas Chinese had in promoting tech-
nology development and in response to the brain-drain problem that we have dis-
cussed earlier, the government in Taiwan set up the Hsinchu Science-based Industrial
Park in 1980. Unlike the tax and duty-free export-processing zones in Kaohsiung,
Nantze, and Taichung, which were designed to attract foreign investment for export
expansion and to transfer technology, the park was designed primarily to attract
investment in high-tech industries especially by returning students or overseas Chi-
nese.
In this park, plants, utilities, residential housing, schools, and the like are all built
with high standards. All an investor needs to do is move in and start a business. By
December 1988, there were already 96 high-tech firms established in the park. Total
sales of these firms went up from NT$3.3 billion in 1983 to NT$48.3 billion in 1988.
Total employees in 1988 (December) numbered 16,500 with 95 of them having a
Ph.D. and 696 a M. A. (M. S.) degree. Of all workers, 41% had a college education.15
It is increasingly apparent that the establishment of the park has attracted many
experienced and well-established Chinese engineers and technicians to come back
home and develop their professional careers. As a result of this "reverse" trend, a lot
of advanced technology has been brought into domestic firms. Also, domestic firms in
Taiwan now have more windows with which to internationalize their operations and
spread the efforts of their R&D, which can further enhance their competitiveness in
the world.
By recognizing the significant role of FDI in introducing technology, Taiwan has
changed her policy toward FDI. In the past, the policy toward FDI could be described
as "encouragement with caution." It has been stipulated in law that overseas Chinese
and foreigners can invest in only certain sectors of the economy in Taiwan such as
manufacturing, certain service enterprises, and scientific and technical research and
development. However, the law also states that all foreign direct investment must be
approved by the government. Therefore, the "apply for approval" type of administra-
tion or "positive list" scope of investment strategy was adopted by the government.
To further encourage investors and firms to invest in Taiwan, the government made
a bold move by changing the "positive list" in the scope of investment into a "negative
list" in 1988. The major difference between the "positive list method" and the "neg-
ative list method" is clear; so long as overseas Chinese and foreigners do not make
investments in the prohibited sectors listed by the government (i.e., on the negative
list), their investment in Taiwan will be approved automatically by the authority-in-
charge. This certainly enhances the open-door policy that is geared toward meeting
the challenge of the globalization of the world's economy.
As Taiwan rapidly rushes to become an advanced industrialized country, the gov-
ernment has become aware of the fact that the protection of intellectual property rights
must be strengthened in order to meet international standards. Furthermore, drives to
encourage domestic firms to establish their own brand names for their products in
international markets as well as drives to develop the software industry all have called
for stricter regulations on the protection of trademarks and intellectual property.
Therefore, trademark law, copyright law, and other relevant regulations are all under
revision in Taiwan. For instance, in 1985 the trademark law was revised, so that pen-
alties such as fines (with a maximum of NT$90,000), and/or imprisonment (usually
not more than one year), and/or damage compensation (based on an amount equiv-
alent to 500 to 1500 times the unit price of one community) have been introduced.
Furthermore, public education on what constitutes intellectual property as well as the
prosecution of violators have all been strengthened. The sincerity and determination
of the government in protecting intellectual property rights are without question; after
all the implementation of this policy will be beneficial to the industrial upgrading of
the local economy and also will be in accordance with the drive to internationalize
Taiwan's economy under the globalization of the world economy.
NATIONAL DEFENSE AND TECHNOLOGY DEVELOPMENT
It is generally believed that Taiwan has a fairly strong national defense, but little is
known about the relationship between national defense and technology development.
It is reasonable to assume that certain technology development is necessary for the
development of a weapons system, but whether such technology is transferable or is
transferred to civilian industries is quite a different matter.
Taiwan's development of military weapons has been largely determined by its
relationship with the United States. Prior to January 1, 1979 when the United States
and the Republic of China broke off diplomatic relations, Taiwan's national defense
and the island's military-hardware needs were primarily supplied by the United States.
National defense industries were hardly developed in Taiwan at that time. It is true
that a 6-year development plan for 1976-1981 was adopted by the government to
develop capital- and technology-intensive heavy and petrochemical industries, but no
significant progress was made.
A real push for the development of national defense related industries was not
made until after the severance of diplomatic relations in 1979 when the United States
decided to scale down the supply of military weapons to Taiwan. The Chung-Shan
Science Colleague has been the key research institute developing the technologies
needed to build strong weapons systems, especially for the Air Force and the Navy.
ITRI has also done research related to national defense. However, further research is
needed to study how national defense has affected the technological aspect of indus-
trial development.
R&D EXPENDITURES
The National Science Council started in 1980 to collect data and compile indicators
measuring the progress made on the development of science and technology (R&D)
in Taiwan. One of the indicators is expenditures on R&D. It is estimated that in 1979,
total R&D expenditures for science and technology were NT$9.9 billion, constituting
0.84% of GNP or 0.91% of national income. The amount increased to NT$ 19.2 billion
in 1983, or 0.94% of GNP or 1.02% of national income. In 1986, it increased to
NTS28.7 billion, or 1.04% of GNP or 1.13% of national income.16 But data from 1984
are not strictly comparable with data before 1984. From 1984 on, expenditures on
research and development also include those in the humanities and social sciences.
Such inclusion should not change the general conclusion, however, for expenditures
on research for humanities and social sciences are relatively small.
In real terms, expenditures on R&D in 1986 were twice as much as in 1979. The
annual growth rate averaged 9% from 1979 to 1983, and 11% from 1984 to 1986.
The amount of R&D expenditures per researcher in the science and technology
fields has been quite stable. In 1986, it amounted to NT$ 1,034,400 a year (or
US$28,730, at US$1 =NT$36).
In 1986, the government's share of total R&D expenditures was 60% whereas the
share of the private sector was 40%. The government's share includes the share of pub-
lic enterprises, which was 14.5% of total R&D expenditures.
The private sector includes private business enterprises, nonprofit research insti-
tutions, and foreign institutions in Taiwan. In 1986, their shares were 38.7, 0.7, and
0.5%, respectively.
In terms of trends, the government's share in total R&D expenditures has
remained the same, about 60% from 1980 to 1986. But within the government, the
share of public enterprises declined from an average of 20% in 1979-1981 to 15% in
1984-1986.
As for the private sector, the share of private business enterprises increased from
34% in 1979-1981 to 37% in 1984-1986. The share of nonprofit institutions declined
from 5.2% in 1979-1981 to 1.3% in 1984-1986. The share of foreign institutions
remained negligible.
As for the purposes of R&D expenditures, in 1986, 11 % of the total was for "basic
research," 38% for "applied research," and 51% for technology development.
As for private business enterprises in 1986, 3.7% of the total R&D expenditure
was for basic research, 22.5% for applied research, and 73.8% for technology devel-
opment. For public enterprises, 3.0% of total R&D expenditure was for basic research,
26.5% for applied research, and 70.5% for technology development. It is clear, then,
for both public and private enterprises, R&D expenditures are primarily for technol-
ogy development, as one would expect.
In terms of industrial distribution for both private and public enterprises, the
amount of R&D expenditure per NT$1000 of sales in 1986 was only NT$4.27 for
manufacturing, NT$1.06 for construction, NT$2.71 for transportation and commu-
nications, NT$6.45 for water, gas, and electricity, and NT$1.02 for mining. For all
fields, total R&D expenditure amounted to only NT$4.14 per NT$1000 of sales, or
0.41% of sales. The percentage was about the same for both public and private enter-
prises (0.399% for private enterprises).
In 1986, for the following industries in manufacturing, R&D expenditure per

NT$1000 of sales was higher than NT$5:
Chemical materials NT$9.87

Chemical products 7.42
Rubber products 6.58
Machinery 7.94
Electrical machinery and appliances 5.98
Miscellaneous industries 6.79
For these industries, R&D expenditure was overwhelmingly for the improvement
of production techniques.
Furthermore, by utilizing the survey data, which contain 246 domestic and 72
FDI electronic enterprises in Taiwan in 1986, comparisons between SMEs and large
enterprises, and domestic and FDI enterprises in terms of their R&D expenditure, are
made.17 The study found that no matter which of the different R&D indicators one
uses, annual R&D expenditure, annual R&D expenditure per employee (RDE), or the
annual R&D expenditure to total sales ratio (RDS), FDI enterprises devote more
resources to R&D than do domestic firms. However, in terms of the nationality of the
FDI enterprises, it is shown that European FDI enterprises in Taiwan have placed sig-
nificantly more effort on R&D than the rest of the FDI firms in Taiwan. American
FDI enterprise have devoted slightly more resources on R&D than domestic firms,
while Japanese FDI enterprises in Taiwan have consistently shown the least interest
in R&D.
It is also found in the study that the effort of smaller firms' R&D in terms of RDE
or RDS is comparable to that of the larger firms. Therefore, the R&D effort made by
the firm may not be positively related to the size of the firm. Finally, it was also found
that one of the most important factors in determining a firm's R&D effort is its level
of technology intensity embodied in the firm (which is measured by the ratio of the
total number of engineers and technicians to total number of employees in a firm),
and the significance of this factor in determining a firm's R&D effort is greater than its
size, government preferential treatment on R&D expenditure, or any other factor.
INDICATORS OF DEVELOPMENT IN SCIENCE AND TECHNOLOGY
It would be difficult for any country to come up with a set of indicators measuring the
progress made in science and development. Surely data of this sort in Taiwan are
extremely scarce. But available data suggest that a great deal of progress has been made
in the development of science and technology in Taiwan.
As regards human resources, the number of persons engaged in research on the
development of science and technology increased from 8345 in 1979 to 27,747 in
1986, that is, the number of researchers in science and technology in 1986 was 3.3
times as many as in 1979. In 1979, there were 12.8 researchers in science and tech-
nology per 10,000 persons in the population. In 1986, the number increased to 34.9
per 10,000 population.18
In terms of the educational background of the researchers engaged in research
science and technology, in 1986, there were 3146 persons with doctors' degrees (11.3%
of total researchers), 6514 persons with masters' degrees (23.5% of the total), 12,485
persons with bachelors' degrees (or 45% of the total), and 6786 persons with an edu-
cation in professional or vocational schools at the level of a junior college (19.7% of
the total).
Of the 27,747 researchers, 47.8% were in private business enterprises, 26.8% in
research institutes and organizations, and 25.4% in colleges and universities.
In terms of fields of research, 61.6% of these 27,747 researchers were in engineer-
ing, 12.6% in natural sciences, 11.0% in agriculture, 9.2% in medicine, 3.9% in social
sciences, and only 1.7% in the humanities.
It may be interesting to note that those who do research in natural sciences, engi-
neering, medicine, and agriculture were primarily B.A.s and M.A.s. On the other
hand, those who do research in the humanities and social sciences are primarily Ph.D.s
and M.A.s.
In absolute terms, the number of researchers in science and technology seems to
be quite small, but in terms of growth rates, the number of Ph.D.s increased at 11% a
year from 1979 to 1986. For M.A.s, the annual growth rate was about 12% a year. For
researchers with education in professional and vocational schools the growth rate was
23% from 1979 to 1986.
A tangible result of the research has been the number of published papers. In
1986, a total of 7395 papers were published in professional or academic journals, 74%
of them in Taiwan and the remainder in foreign countries. In addition, there were
4730 papers that were read but not published in journals in 1986.19
Another tangible result of research on science and technology has been the num-
ber of patents granted. In 1986, there were 29,511 applications for patents, of which
some 12,355 were approved. Of the patents approved, 53% were by Chinese nationals
and 47% by foreign nationals. This represents a substantial increase in the number
filed by Chinese Nationals. In 1979, there were 10,411 applications, of which 3,686
were approved. Of the patents approved, 41% were by Chinese nationals and 59% by
foreign nationals. It is evident that the number of patents approved has increased sub-
stantially, and the share accounted for by Chinese nationals has steadily increased,
whereas the share given to foreign nationals has declined.
Still another indicator of the result of research on science and technology is the
nature or the technology content of exports. In the 1950s, the main items of export
were sugar, rice, and bananas. These products were basically land-intensive. As time
went on, labor-intensive products such as textiles and shoes began to become the main
items of exports. In more recent years, more technologically advanced or capital-
intensive products, such as machinery in general and electrical machinery in particular
as well as other types of apparatuses, have assumed a greater and greater share in total
exports. By one estimate, the share of "technology-intensive" products constituted
21.9% of total merchandise exports in 1983. This proportion gradually rose to 24.3 in
1987.20
A NATIONAL PLAN
The government is well aware of the fact that Taiwan has to accelerate its efforts in the
development of science and technology in order to upgrade its economic structure,
Table 12.6. The Key Targets of Taiwan's 10-Year Science and Technology Plan
Annual
1986 1995 Growth Rate
Total number of researchers 27,747 43,000 6.1%

Ph.D.s (percentage in total) 3,146(11.3%) 6,600(15%) 8.6%
M.A.s (percentage in total) 6,514(23.5%) 15,400(35%) 10%
R&D expenditures (NTSbillion) 28.7 90.0
Annual growth rate 13.5%
As a percentage of GNP 1.04% 2.0%
Basic research as a percentage of 11.1% 12.0%
total R&D
Government's share in R&D 60% 40%
Private share in R&D 40% 60%
improve the quality of life of the people, and strengthen national defense. Conferences
at all levels have been held frequently to find ways to improve technology, especially
industrial technology. A 10-year plan for science and technology was adopted for the
period 1986-1995. The key targets of this 10-year plan are shown in Table 12.6.
It remains to be seen, however, whether the targets in Table 12.6 will be reached.
It may be noted that in Korea, the proportion of R&D expenditures constituted 1.82%
of her GNP in 1986. In the same year per capita GDP for Korea was US$2365, whereas
in Taiwan it was US$3688. Taiwan should achieve the 2% of GDP target by 1995.
AN APPRAISAL AND SOME CONCLUDING REMARKS
Throughout this chapter we have tried to suggest that Taiwan's economic performance
has been successful in many respects, and technology has been an important contrib-
uting factor to this. We have also tried to describe the efforts of the government and
the private sector to advance science and technology, especially industrial technology.
These efforts, however, are far from adequate.
It must be admitted that a great deal of research remains to be done before one
can understand exactly how industrial technology has been advanced in Taiwan, how
technology in industry may be denned or classified for the purpose of empirical
research, who introduced or applied the new or improved technologies and what they
consisted of, and, finally, who failed, who succeeded, and why.
Another set of questions that remains to be answered regard government policy.
Surely the government has done a great deal in advancing science and technology on
Taiwan. It will be a gigantic task to evaluate what the government has done in this area,
and this is something we would not attempt to do here. What follows is simply some
questions about certain government policies that have an important bearing on the
future development of technology.
First, we look at Taiwan's education policy. The government, in the past, has
emphasized vocational or employment-specific education. This, in our view, should
be modified if not totally changed. In an economy that has been, and still is, under-
going structural change, what the labor force needs is the ability to adapt to changing
conditions, changing technologies, and changing production processes. In our view, it

is the "general" or the "liberal arts" type of education that best prepares the labor force
with such flexibility. We are not downgrading the importance of specialized or
employment-specific knowledge; what we contend is simply that it is important to
have a proper dose of liberal arts education in the curriculum.
Now "compulsory" or, more properly, free education will be extended from 9 to
12 years in Taiwan. It is the government's policy to emphasize vocational or profes-
sional education for the last 3 years of this period at least in terms of the number of
vocational vs. general senior high schools. We suggest that serious consideration be
given to the question of whether there is enough "general education" in the curriculum
of vocational schools.
Second, we need to review the "pick the winners" policy. In the past the govern-
ment often chose certain industries or products as strategic and favored them with
preferential treatment, as described above. In our view, the industries or products that
are developed should be determined by market forces; any deliberate government pol-
icy to alter the comparative advantage that an economy enjoys will distort the allo-
cation of resources except in cases where there is market failure or externalities. To be
more specific, the technology that is appropriate for an economy to adopt should be a
matter that business firms decide, not the government.
It is true that the concepts of economic liberalization and internationalization
have now been adopted by the government as its guiding principles in the formulation
of economic policy, and government interference in the choice of technology is to be
minimized. But given that the government desires to upgrade technology and the eco-
nomic structure, we hope that it will resist pressure to grant special favors to any spe-
cific industries or technologies that are claimed by vested interest groups to be "stra-
tegic."
Third, a word on small and medium enterprises. Small and medium enterprises
have been, and still are, a vital force for Taiwan's economy, especially the export sec-
tor. They are being confronted, and indeed have been for some time, with some special
difficulties arising from the appreciation of the NT dollar and increasing competition
by other developing countries where unit labor costs are much lower. As a result, some
of them are being forced to move out of the country and invest abroad.
But this is not desirable for the economy as a whole, for Taiwan is still a devel-
oping economy and needs a lot of capital investment at home. Hence, it is not in a
position to invest abroad on a large scale. A more desirable policy is to upgrade tech-
nology so as to increase the competitiveness of small and medium enterprises.
One reason that small enterprises shy away from spending a great deal of money
to upgrade technology is that the kind of technology they need can often be imitated,
and cannot be effectively protected by patents. As a result, everyone is reluctant to be
an innovator, at least in regards to some types of technologies.
In our judgment, some of the government-supported research institutions such
as the Industrial Technology Research Institute should double and redouble their
efforts to work on the kind of technology that is appropriate for small and medium
enterprises, instead of devoting their efforts primarily to "high-tech."
Finally, government-owned enterprises in Taiwan are still significant, accounting
for 10.2% of total value-added in the manufacturing sector. They are often criticized
as being inefficient, but this is not to say that their managers and workers are less able.
The real difficulty is that their operation is under many restrictions, there is no effective
incentive systems, and, worst of all, their managers and employees have the status of
a civil servant. Being a civil servant, they have a number of restrictions imposed on
them, and hence they cannot be as innovative as their counterparts in private business.
In our view, the best way to make them innovative, especially in upgrading or choosing
appropriate technology, is to privatize those public enterprises that are not natural
monopolies or related to national defense.
NOTES
The authors wish to express their thanks to Ms. Yih-shiow Chen for her assistance.
1. See, for example, Chi-ming Hou. (1988). "Relevance of the Taiwan Model of Devel-
opment." In Conference on Successful Economic Development Strategies of the Pacific Rim
Nations. Taipei: Chung-Hua Institution for Economic Research. Reprinted in Industry of Free
China, February 1989, 9-32.
2. See Chi-ming Hou and Ching-hsi Chang. (1981). "Education and Economic Growth in
Taiwan: The Mechanism of Adjustment." In Conference on Experiences and Lessons of Eco-
nomic Development in Taiwan, 492. Taipei: The Institute of Economics, Academia Sinica.
3. Chi-ming Hou and Hui-lin Wu. (1983). "Wage and Labor Productivity in the ROC." In
Raising Productivity: Experience of the Republic of China. Tokyo: Asian Productivity Organi-
zation. Reprinted in Industry of Free China, May 1985.
4. Calculated from data in Duller, H. J. (1983). Technique in Taiwan: The Role of Tech-
nology in Taiwan's Past and Present Development. Taipei: Sun Yat-San Institute of Social Sci-
ence and Philosophy, Academia Sinica. Duller's data were provided by the Central Bank of
China.
5. Hou and Chang (1981, p. 486).
6. Hou and Wu( 1983, p. 65).
7. Taiwan Statistical Data Book, 1989, p. 7.
8. The data, as cited in note 4, imply the following rates of growth of capital in the non-
agricultural sector: 2.8% in 1953-1961, 8.1% in 1962-1971, and 12.6% in 1972-1980. For the
period from 1953 to 1980, the average rate of growth was 7.8% a year.
9. See San Gee. (1989). "Direct Foreign Investment and Technology Transfer in the
ROC."Proceedings of Conference on the Industrial Policies of the ROK and the ROC, Confer-
ence Series 89-01, KDI, Seoul.
10. Hou and Chang (1981, pp. 474-481).
11. San Gee and Chao-nan Chen. (1988). In-Service Training in Taiwan, ROC. CIER Eco-
nomic Monograph Series No. 20, CIER, Taipei, Taiwan.
12. San Gee. (1989). "Direct Foreign Investment and Technology Transfer in the ROC."
Proceedings of Conference on the Industrial Policies of the ROK and the ROC, Conference Series
89-01, KDI, Seoul.
13. See report by Liu, P. (1989). "Big Glue Benefits from Funding R&D in Taiwan." ASIA/
Pacific Business, May 15.
14. For a more detailed discussion on the information industry in Taiwan, see San Gee.
(1990). "The Status and Evaluation of the Electronic Industry in Taiwan." OECD Development
Centre Technical Papers Series No. 29, OECD June.
15. Science and Technology Yearbook, Republic of China, 1988, p. 416. National Science
Council, Executive Yuan, in Chinese.
16. Science and Technology Yearbook, Republic of China 1987.
17. See San Gee. (1989). "A Comparison on R&D Expenditure among Domestic and FDI
Firms and a Study on Factors Which Determine This ExpenditureA Study on the Electronic
Firms in Taiwan." Academia Economic Papers, the Institute of Economics, Academia Sinica,
17(1), March (in Chinese).
18. Science and Technology Yearbook, Republic of China 1987.
19. Science and Technology, Republic of China 1987.
20. Science and Technology, Republic of China 1987.
13
Nationai Systems Supporting Technical
Advance in Industry: The Brazilian
Experience
CARLJ.DAHLMAN
CLAUDIO R. FRISCHTAK
Brazil has the ninth largest GDP in the world, the sixth largest population, and the fifth
largest area. Its per capita income in 1987 was US$2241, slightly above Mexico but
below Argentina, South Korea, and Taiwan. The country is well endowed with natural
resources, has a productive agriculture sector, and a mature and diversified industrial
base. Manufacturing value-added in 1987 was US$79 billion, the ninth largest in the
world, about 15% larger than that of Canada, two times that of Mexico or Korea, and
almost three times that of Australia.
During the late 1960s and most of the 1970s Brazil was one of the fastest growing
economies in the world. Yet in the past decade, output growth of Brazilian manufac-
turing industry has slowed significantly (Table 13.1). In contrast with the East Asian
and several other newly industrializing countries, Brazil's average GDP and manufac-
turing growth were low, 2.9 and 2.2% p.a., respectively. In addition, Brazil's share of
manufactured exports in manufactured output, 11.8% in 1987, is small, even though
manufactured export growth has outstripped manufactured output growth since 1965.
This paper focuses on technological factors behind Brazil's faltering competitive
position and the role of Brazil's innovation system. The innovation system is here
defined as the network of agents and set of policies and institutions that affect the intro-
duction of technology that is new to the economy. Since in the vast majority of devel-
oping countries technology is imported, the innovation system is defined very broadly
to include policies toward arms-length technology transfer, intellectual property
rights, importation of capital goods, and direct foreign investment. The innovation
system also comprises the network of public and private institutions and agents sup-
porting or undertaking scientific and technological activities, including research and
development, diffusion, and creation of technical human capital.
The first section presents the historical and political context that gave birth to the
innovation system in Brazil. It suggests that it was not until the late 1800s, with labor
shortages in agriculture and depletion of mines, that there were strong enough incen-
414
TECHNICAL ADVANCE IN INDUSTRY: BRAZIL 415
Table 13.1. Selected Economic Indicators

Growth of Growth of
Manufactured Manufactured
Share of
GDP Output Exports
GDP Per Manufactured
Capita 1965- 1980- 1965- 1980- 1965- 1980- Exports in
1988 1980 1988 1980 1988 1980 1988 Output 1987
(US$) (%) (%) (%) (%) (%) (%) (%)
Brazil 2,241.07 8.8 2.9 9.8 2.2 22.1 6.2 11.8
Indonesia 476.09 8.0 5.1 12.0 13.1 13.3 35.4 44.8
Japan 23,195.02 6.5 3.9 8.2 6.7 11.9 5.3 32.4
Korea 4,078.81 9.6 9.9 18.7 13.5 31.2 14.7 103.1
Mexico 2,111.11 6.5 0.5 7.4 0.2 9.4 21.1 27.1
Thailand 1,063.30 7.2 6.0 11.2 6.8 21.3 19.3 53.3
Turkey 1,196.28 6.3 5.3 7.5 7.9 26.4 23.3 42.6
Sources: World Development Report (1990); World Tables (1989-90 edition). International Financial Statistics (December
1989).
tives for innovation and technological development. Those needs were expanded
greatly with twentieth-century industrialization. Yet it is only after World War II that
the government's focus on science and technology is explicit, and centered in the
build-up of an institutional infrastructure and in the development of human resources.
This process was formalized during the military period, particularly after 1970, when
scientific and technological development became a specific policy objective.
Although the enlargement of the S&T base helped improve the competitive posi-
tion of Brazilian firms in the 1970s, its impact was limited in view of the still modest
importance that domestic firms attributed to technological development. The second
section briefly discusses Brazil's competitive status in world markets and suggests that
its faltering position is related to the limited domestic technological efforts of industrial
producers. The third section shows that the modest nature of these efforts has not been
offset by substantial technology imports (which have been restricted by regulatory con-
straints on arms-length technology transfer), direct foreign investment, and capital
goods imports. Although some of these restrictions, in combination with government
financial support, induced firms to engage in technological activities, relatively few
producers were able to approach the international best-practice "frontier" in the
1980s.
In view of the modest technological activities undertaken by local private pro-
ducers and the limited inflows of foreign technology, development efforts fall on the
shoulders of the public sector. The fourth section notes that 80-90% of R&D expen-
ditures have been undertaken by the government. Unfortunately, dispersion of these
public sector R&D efforts, and weak linkages between public R&D institutes and the
productive sector, have meant that the flow of useful technological services has not
been commensurate with the volume of public sector resources allocated to techno-
logical activities. Finally, in human resource formationan area in which the public
sector role is criticalweak investment decisions have undermined the creation of
technological capabilities for the future. The final section presents the chapter's con-
clusions.
HISTORICAL AND POLITICAL CONTEXT
A complex interaction between economic forces and political regimes has shaped the
development of science and technology (S&T) in Brazil. This section highlights some
of the salient dimensions of such interaction in order to provide a historical context
for understanding many of the current features of Brazil's innovation system.
1500-1900: From Colonization to the Twentieth Century
During this phase, which spanned four centuries, Brazil was a source of agricultural
and mining products for the world. Brazilian economic history during this period can
be characterized as a period of boom and bust cycles of primary commodity exports.
These have included sugar, gold, cotton, cocoa, rubber, and finally coffee, whose
importance continued into the twentieth century and provided the economic foun-
dation for industrialization.'
The main innovations during the colonial period, which lasted until 1822 when
Brazil gained its independence from Portugal, were the result of agricultural technol-
ogies being transplanted to Brazil. They included new seeds (cotton), crop techniques
(for tobacco, sugarcane), and the use of manure for fertilizer. Since land and slave
labor were abundant during this period, there were few incentives to use land more
efficiently or to mechanize production. In mining, primitive technologies were trans-
ferred as minerals were readily accessible near the surface. During this early period the
country also experienced moderate development of cartography and military engi-
neering oriented to the construction of forts, canals, bridges, and the production of
weapons and munitions. 2
The migration of the Royal Portuguese Family to Brazil in 1808, fleeing Napo-
leon's armies, led to the establishment of a number of scientific and technological insti-
tutions in the Brazilian colony. These included the Medical School of Bahia, the Med-
ical School of Rio de Janeiro, the National Library, the Royal Farm, the National
Museum, a gunpowder factory, an iron foundry, the School of Mines of Ouro Preto
(1875), and the Agricultural Research Station of Campinas (1887).
It was not until the late 1800s, however, that there were strong economic incen-
tives for greater attention to research and development. In mining, labor shortages,
exacerbated by the abolition of slavery in 1889, combined with the depletion of the
surface deposits, called for improved mining technology, a greater degree of mecha-
nization, and a better understanding of chemistry and mineralogy. Similarly, low
yields and labor shortages created a need to increase productivity in agriculture and
led to the establishment of several agricultural colleges and research institutes. The
Agricultural Research Station of Campinas and the Luis de Queiroz Agricultural Col-
lege (both in the state of Sao Paulo), in particular, contributed in a major way to
improve agricultural productivity, especially of export crops such as coffee, sugarcane,
and cotton.3
1900-1946: Laying the Basis for Industry-Led Growth
The first three decades of the twentieth century saw the beginning of industrialization
in Brazil. In spite of some early efforts at import substitution in the 1890s, on the eve
of World War I Brazil's industrial structure was very primitive and only 3% of the labor
force was employed in manufacturing.4 The difficulty of importing industrial goods
during the war strengthened the need for import substitution. However, most import
substitutes were relatively simple products and did not demand much research capa-
bility or engineering manpower. Most of the technology was obtained by copying from
abroad and through firm-to-firm diffusion.
An important development during the early part of this period was the spread of
the railroad, which brought with it the need for increased capabilities in civil engi-
neering and greater technological sophistication in production and use of steel (includ-
ing foundry, forging, and machining operations) and steam power. The Ouro Preto
School of Mines was the first civilian institution to become involved in training pro-
grams for the establishment and expansion of the railroad network. Another was the
Sao Paulo Polytechnic, which set up a Laboratory of Material Resistance in 1926 to
test materials and equipment for railroads and power plants.
The Brazilian economy, like many other Latin American primary products
exporters, went into depression with the economic crash of 1929. Exports collapsed,
and the lack of foreign exchange stimulated domestic production of cement, chemi-
cals, paper, metals, and textiles. The import-substitution drive also received strong
impetus from World War II, which made it difficult to obtain key inputs such as steel
and capital goods. By the end of the war, the country's manufacturing sector had
become quite diversified though not very competitive by world standards. Still, the
technological complexity of many of the new industries increased the demand for engi-
neers and technical support services.5
1947-1964: Heavy Industrialization and a Preliminary Focus on Science and

Technology
This period was characterized by a strong drive toward industrialization from the sec-
ond Vargas (1951-1954) and the Kubitschek governments (1956-1960). The indus-
trial sector grew 262% in contrast to only 87% for the agricultural sector. Kubitschek's
"Piano de Metas" (1956-1961) was the first organized effort focused on heavy industry
and supportive infrastructure. It was built around ambitious projects in energy, trans-
portation, steel and petroleum refining, chemicals and petrochemicals, capital goods,
automobiles, and Pharmaceuticals.
One of the key elements of the industrial development strategy was to induce for-
eign firms to set up manufacturing facilities in Brazil. This consisted not only of pro-
tecting the local market but also of offering significant subsidies and special treatment
for foreign investors. By 1960, foreign subsidiaries accounted for more than 50% of the
capital goods producers, 70% of chemicals (except petrochemicals), 90% of pharma-
ceuticals, and 100% of the nascent automobile industry.6
At the same time, however, significant steps were taken in the science and tech-
nology area. The National Research Council (CNPq) was established in 1951 to pro-
mote research in all areas. However, one of its main purposes, as clearly articulated by
Admiral Alvaro Alberto da Mota Silva, CNPq's president from 1951 to 1955, was to
prepare Brazil to use its mineral resources for the production of atomic energy.7
The creation of CNPq has to be seen in the context of postwar polarization, com-
petition, and conflict. Harnessing atomic energy was perceived as the key to military
power and a crucial step for accelerating the process of economic development.8 The
emphasis on science and technology during this period was also a matter of national
prestige. There was an important demonstration effect from the scientific and tech-
nological activities of the advanced economies that drove countries such as Brazil to
try to keep up in order to gain international stature. Much of the effort was driven by
the military. Many of its personnel received training in the United States and came to
appreciate the importance of a strong technological base for military power. A number
of important institutions were set up by the armed forces during this period. Possibly
the most successful was the Aerospace Research Center (CTA) established in 1947,
which played a major role in the development of the Brazilian aviation industry as well
as in the gasohol program.
In the early 1960s the Brazilian economy lost steam as ambitious projects of the
Kubitschek era (which included the building of the new capital of Brasilia) and sub-
sequent poor economic management led to growing macroeconomic imbalances. The
deterioration in economic and political stability eventually led to a military coup in
1964 and to 21 years of military rule.
1964-1985: The Military Period and the Formalization of the Science and
Technology Infrastructure
Economic Development Plans

The period of military rule saw many changes in the structure of the Brazilian econ-
omy and in policy-making. Greater efforts were directed toward planning for devel-
opment and formalizing key programs, and increased resources were allocated to sci-
ence and technology. The first military government (1964-1967) focused essentially
on reestablishing macroeconomic stability and introducing important economic
reforms in the tax and financial systems, whereas the one that followed gave emphasis
to economic recovery and the resumption of growth. The third military government
produced the "First National Development Plan" (1972-1974). The objective of that
plan was to achieve sustained industrial and economic growth based on an expansion
of exports and the domestic market. As part of a strategy of increasing national power,
new areas (such as nuclear energy, electronics, and space research) were expected to
have rapid development. The plan also expressed concern over the growing presence
of multinational firms in Brazil.
The Second National Development Plan (1975-1979) set out to adjust Brazil's
economy to the oil shortage by deepening the process of import substitution and mov-
ing the country toward energy self-sufficiency. Yet changes in the real exchange rate
were avoided, and so was tighter fiscal and monetary control. At the same time there
was growing reliance on domestic and external borrowing to finance investment and
growth. Brazil entered a period of "debt-led growth."9 By the end of 1978 Brazil's
external debt had risen to $40 billion compared to only $3.5 billion in 1968 (for a more
detailed discussion, see the next section). Brazil's economic situation was worsened
further by the 1979 oil shock, which led to another spree of external borrowing, such
that by the end of S985 it had the largest foreign debt in the world at $ 100 billion. At
the same time, inflation accelerated from average annual rates of 20.7% in 1970/1974
to 45.9% in 1975/1979 and 141.7% in 1980/1984. Thus, although the country main-
tained relatively high rates of economic growth in the 1970s, it did so at the cost of
growing macroeconomic imbalances.
Science and Technology Policies and Institution-Building Efforts

Three important financing institutions relating to S&T were created during the first 3
years of the military period: FUNTEC, FINAME, and FINEP.10 Yet only in 1968, sci-
entific and technological development became a specific policy objective. The 1968-
1969 development plan (Programa Estrategico do DesenvolvimientoPED) defined
for the first time an explicit policy for S&T at the federal level. The PED proposed the
creation of a National System of Scientific and Technological Development
(SNDCT), of S&T Basic Plans that would spell out the actions foreseen in the National
Development plans, and of a National Fund for Scientific and Technological Devel-
opment (FNDCT) to finance the SNDCT.
Since the creation of the SNDCT, three Basic Plans have been issued, covering
the periods 1973-1985. The First Basic Plan of Science and Technology Development
(I PBDCT1973-1974) programmed an increase in the volume of resources for S&T
by strengthening the FNDCT and other financial mechanisms. The Second Plan (II
PBDCT1975-1979) set priorities that included the development of new technolo-
gies, with research on nonconventional sources of energy, space activities, and ocean-
ography. '' In the area of basic research, emphasis was placed on human resource devel-
opment through the National Graduate Education Plan. The Second PBDCT also
identified the growing presence of multinationals and the weakness of national firms
as obstacles to the development of local technology.
The Third Basic Plan of Science and Technology (III PBDCT1980-1985)
aimed at expanding the supply of S&T resources while reinforcing the technological
capabilities of national firms. The Third Plan differed from the previous two in that
rather than specifying government actions in the form of programs, projects, and pri-
ority activities, it established a set of policy directions that was supposed to orient the
actions of the public and private sector. Furthermore, it did not present any figures on
planned expenditures.
In the industrial area, the most important development was the creation of the
Secretaria de Tecnologia Industrial (STI) of the Ministry of Industry and Commerce
(MIC) in 1972. It was the first sectoral science and technology unit to be established
within a ministry. The STI took over various S&T institutions already existing in MIC,
and its activities included (1) carrying out R&D programs through its own institutes,
(2) funding for technological development in public and private enterprises, and for
the development of specific technologies,12 (3) supplying technological information to
firms, (4) administering a system of intellectual property rights (basically patent and
trademarks) and regulating technology transfer through INPI (National Institute of
Industrial Property Rights), and (5) acting as the executive secretariat of the National
Council for Metrology, Normalization and Quality Control (CONMETRO).
Overall, the military period can be characterized as a time when a planning sys-
tem was established for S&T and strong emphasis placed on institutional develop-
ment. Science and technology were perceived as important elements for enhancing
national power. The first military government represented the more liberal faction of
the armed forces. The three subsequent military governments represented more
nationalistic tendencies, with greater preoccupation with issues of national sover-
eignty and autonomy. These tendencies found expression in a strong emphasis on
strengthening national technological capability in military and strategic areas such as
informatics, telecommunications, the defense industry, aviation, and nuclear energy.
These tendencies also led to a restrictive attitude toward international technology
transfer and heavy emphasis on local technology development. The military period
also exhibited a strong statist tendency, which was expressed in the drive for state par-
ticipation in strategic sectors such as steel, petrochemicals and chemicals, heavy cap-
ital goods, and many important intermediate goods industries. Although it was often
necessary to work with multinationals as equity partners as a way of getting access to
foreign technology, the government generally sought to control the enterprises by
inviting minority private national participation in order to have majority local con-
trol.
The Brazilian "miracle" was short-lived. Debt and inflation-led growth proved to
be unsustainable in the longer term. As argued in the next section, a combination of
macroeconomic instability, lower rates of investment, and growing inefficiencies in
the use and allocation of resources led in the 1980s to a reversal of the competitive
gains achieved in the Brazilian economy in the previous decade. Yet the investments
made in the build-up of human resource and institutional capacity in science and tech-
nology during the military period may still prove to have far reaching consequences
for the competitive position of Brazilian industry when investment levels recover.
Although the technological efforts of Brazilian firms are still limited, there is growing
awareness that in a more open and competitive environment the ability to innovate
and produce low-cost, high-quality products will be the key to their survival. It will call
for a more effective use of existing S&T infrastructure, including a more balanced
development of the human resource base.
BRAZIL'S COMPETITIVE POSITION AND THE TECHNOLOGICAL EFFORTS OF

BRAZILIAN FIRMS
Brazil's competitive position in the 1970s improved significantly. Between 1970 and
1980 Brazil expanded its world market share in 83 of 93 industrial segments (three-
digit SITC categories) with positive manufactured exports (Table 13.2).13 Moreover,
Brazil had an impressive record of market diversification and product differentiation
in the 1970s. Between 1971 and 1980, in most machinery and transport equipment
segmentsand in shoes, textiles, and fibersthe number of countries importing from
Brazil increased, and the number of products exported in each goods category
expanded. In many categoriessuch as office machinery, boats and ships, aircraft, and
textilesthe export composition shifted to higher value-added products.14
Brazil's Lagging International Competitiveness in the 1980s

Since 1980, the competitive dynamics of the Brazilian economy have changed sub-
stantially, with constant or declining market shares in 46 of the 93 relevant SITC seg-
ments, including technologically sophisticated areas such as electric machinery and
instruments (Table 13.2). The only significant exceptions were aircraft, telecommu-
nications equipment, and electric distribution equipment. Moreover, in contrast with
the 1970s, when 20 of 26 segments in which Brazil had achieved a "revealed" com-
parative advantage (i.e., normalized market shares greater than one) increased their
shares, only nine segments did so in the period 1980-1987.15 Virtually all nontradi-
tional and nonnatural resource-based manufactured products had their gains
reversed.16 Surprisingly, this reversal even occurred in nontraditional products such as

textiles, garments, travel goods, and handbags, in which Brazil could have expected to
maintain its competitive position.
Brazil is now saddled with many internationally uncompetitive segments, their
number having grown significantly since the early 1980s. A study of 67 manufacturing
subsectors reveals that in the period 1983/1985,45% were internationally competitive
(that is, the ratio of domestic to international prices was less than one) and 7% were
marginally competitive (domestic to international price ratios between 1 and 1.1). A
similar comparison shows that in 1988 only 15% of the subsectors were clearly com-
petitive, while in 12% of the cases competitiveness was marginal. These studies suggest
that the proportion of fully or marginally competitive segments fell by nearly half
(from 52 to 27%) between 1983/1985 and 1988.17
Additional evidence of a deteriorating competitive position comes from a 1989
survey of the views of 550 major industrial producers. Although most industrialists
judged the Brazil's industrial sector to be more efficient than at the beginning of the
decade, nearly 65% of those interviewed also perceived the extent of modernization as
"insufficient" to keep up with the international price performance frontier, whereas
only 3.8% thought that it was "quite adequate."18
Several factors explain the country's inability to sustain its competitive position.
After the first oil shock in 1983, Brazil faced a further 38% decline in its terms of trade
in 1977-1981. Yet, macroeconomic adjustment was postponed, and as a result Bra-
zil's growth became increasingly debt dependent. By the late 1970s international inter-
est rates rose rapidly, and economies driven by debt-led growth strategies became
unacceptably risky for lenders once countries began to default on sovereign debt. In
the case of Brazil, the flow of capital was reversed, from plus to minus 4-5% between
the mid-1970s to the mid-1980s. These shocks led to severe macroeconomic instabil-
ity, two major recessions, and a strong loss of confidence in the longer term prospects
of the economy. Ultimately, they brought about a significant contraction in the rate
of investment. The ratio of investment to GDPafter reaching nearly 25% in 1975
fell to 15-16% in 1988.
Brazil's falling performance relative to other industrializing countries has been
due not only to a contraction in the rate of investment but also to inefficient resource
allocation and use. Brazil has a large and diversified industrial base, the result of a rel-
atively long process of natural and policy-induced import substitution. Beginning in
the 1950s, producers were attracted to invest in new industrial segments by a combi-
nation of trade barriers, entry regulations, and fiscal, credit, and other incentives. Dur-
ing this period, Brazilian industry went through substantial structural change (Table
13.3). The share in value-added of traditional consumer goods sectors such as textiles,
food products, beverages, and tobacco decreased significantly, whereas the shares of
machinery, electrical equipment, transport equipment, and chemicals/petrochemi-
cals increased.
In the 1970s, and within the framework of the Second National Development
Plan (1974-1979), the intermediates and capital goods industries were targeted and
capacity was created at very fast rates. Producers were attracted to invest in these sec-
tors by a combination of trade barriers, entry regulations, and fiscal, credit, and other
incentives. The goal was to shift resources to domestic firms so as to stimulate and
enable them to invest in targeted sectors. As part of an import substitution strategy,
Table 13.2. Shifts in Market Shares for Brazil's Manufacturing Exports, 1970-1987
Market Shares Ratios
SITC Commodities 1970 1980 1987 80/70 87/80
512 Organic chemicals 0.003 0.009 0.012 3.0 1.3
513 Inorg elemnts, oxides, etc. 0.000 0.004 0.008 9.3 1.8
514 Other inorganic chemicals 0.001 0.002 0.004 1.6 1.6
515 Radioactive etc. material 0.000 0.000 0.000 0.0 37.1
521 Coal, petroleum etc. chems 0.000 0.000 0.004 5.1 68.1
531 Synt dye, nat indgo, lakes 0.000 0.001 0.001 4.2 0.7
532 Dyes nes, tanning prods 0.040 0.041 0.052 1.0 1.3
533 Pigments, paints, etc. 0.000 0.002 0.002 11.1 0.9
541 Medicinal etc. products 0.002 0.003 0.003 1.6 1.1
551 Essentl oil, perfume, etc. 0.023 0.018 0.014 0.8 0.8
553 Perfume, cosmetics, etc. 0.001 0.005 0.003 4.8 0.6
554 Soaps, cleaning etc. preps 0.000 0.006 0.002 16.1 0.4
561 Fertilizers manufactured 0.000 0.000 0.001 31.9 4.2
571 Explosives, pyrotech prod 0.002 0.041 0.168 23.8 4.1
581 Plastic materials etc. 0.000 0.003 0.007 20.4 2.4
599 Chemicals nes 0.002 0.008 0.008 3.4 1.1
611 Leather 0.021 0.029 0.032 1.4 1.1
612 Leather etc. manufactures 0.005 0.025 0.044 5.1 1.8
613 Fur skins tanned, dressed 0.000 0.001 0.002 55.4 2.4
621 Materials of rubber 0.000 0.003 0.003 16.9 0.9
629 Rubber articles nes 0.003 0.011 0.016 4.1 1.4
631 Veneers, plywood, etc. 0.025 0.023 0.022 0.9 1.0
632 Wood manufactures nes 0.003 0.012 0.010 4.7 0.8
633 Cork manufactures 0.000 0.001 0.001 18.4 0.7
641 Paper and paperboard 0.000 0.006 0.009 78.5 1.5
642 Articles of paper etc. 0.000 0.004 0.006 16.1 1.4
651 Textile yarn and thread 0.004 0.020 0.018 5.1 0.9
652 Cotton fabrics, woven 0.006 0.016 0.015 2.6 1.0
653 Woven textiles noncotton 0.001 0.004 0.002 5.3 0.6
654 Lace, ribbons, tulle, etc. 0.004 0.003 0.003 0.7 0.9
655 Special textile etc. prod 0.003 0.020 0.008 6.0 0.4
656 Textile etc. products nes 0.003 0.028 0.034 8.1 1.2
657 Floor covr, tapestry etc. 0.000 0.001 0.001 5.2 1.2
661 Cement etc. building prod 0.000 0.006 0.004 17.3 0.7
662 Clay, refractory bldg prd 0.002 0.012 0.014 6.8 1.2
663 Oth nonmetal mineral mfs 0.001 0.007 0.005 7.5 0.7
664 Glass 0.010 0.005 0.006 0.5 1.0
665 Glassware 0.000 0.004 0.006 14.6 1.6
666 Pottery 0.001 0.007 0.009 7.0 1.2
667 Pearl, prec-, semi-p stone 0.004 0.002 0.005 0.6 2.1
671 Pig iron etc. 0.026 0.073 0.145 2.8 2.0
672 Iron, stl primary forms 0.015 0.009 0.064 0.6 7.5
673 Iron and steel shapes 0.009 0.007 0.023 0.7 3.4
674 Irn, stl univ, plate, sheet 0.004 0.011 0.022 2.9 1.9
675 Iron, steel hoop, strip 0.000 0.004 0.004 34.8 1.0
676 Railwy rails etc. irn, stl 0.000 0.005 0.000 25.0 0.0
677 Irn, stl wire excl w rod 0.002 0.005 0.007 2.8 1.5
Market Shares Ratios

SITC Commodities 1970 1980 1987 80/70 87/80
678 Iron, stl tubes, pipes, etc. 0.001 0.010 0.007 8.3 0.7
__
679 Irn, sll castings unworkd 0.002 0.005 2.8
691 Structures and parts nes 0.000 0.002 0.004 14.9 1.5
692 Metal tanks, boxes, etc. 0.001 0.009 0.007 10.0 0.8
693 Wire products non electr 0.000 0.009 0.012 19.4 1.3
694 Stl, coppr nails, nuts, etc. 0.000 0.003 0.003 10.9 1.1
695 Tools 0.004 0.007 0.005 1.6 0.7
696 Cutlery 0.006 0.020 0.018 3.6 0.9
697 Base mtl household equip 0.001 0.012 0.007 15.2 0.6
698 Metal manufactures nes 0.000 0.003 0.003 7.8 1.2
711 Power machinery non-elec 0.000 0.013 0.014 34.5 1.1
712 Agricultural machinery 0.001 0.016 0.016 23.2 1.0
714 Office machines 0.007 0.011 0.002 1.6 0.2
715 Metalworking machinery 0.002 0.006 0.003 3.1 0.5
717 Textile, leather machnry 0.002 0.005 0.005 3.2 1.0
718 Machs for spcl industries 0.002 0.007 0.005 2.7 0.8
719 Machines nes nonelectric 0.001 0.005 0.005 5.7 0.9
722 Elec pwr mach, switchgear 0.001 0.005 0.003 5.1 0.7
723 Electr distributing mach 0.001 0.002 0.006 2.2 2.3
724 Telecommunications equip 0.001 0.005 0.010 4.7 2.0
725 Domestic electric equip 0.001 0.004 0.005 2.7 1.5
726 Electro-medcl, xray equip 0.001 0.001 0.000 1.7 0.3
729 Electrical machinery nes 0.001 0.005 0.003 3.7 0.5
731 Railway vehicles 0.000 0.016 0.004 47.4 0.3
732 Road motor vehicles 0.000 0.008 0.009 18.3 1.1
733 Road vehicles non-motor 0.000 0.007 0.001 48.0 0.2
734 Aircraft 0.000 0.003 0.011 40.5 3.3
735 Ships and boats 0.001 0.007 0.010 6.2 1.4
812 Plumbg, heating, Ightng 0.000 0.004 0.003 9.0 0.7
equ
821 Furniture 0.001 0.003 0.002 1.9 0.9
831 Travel goods, handbags 0.000 0.013 0.012 41.9 1.0
841 Clothing not of fur 0.001 0.003 0.003 6.0 1.0
842 Fur etc. clothes, prod 0.000 0.017 0.004 299.8 0.2
851 Footwear 0.005 0.033 0.057 6.7 1.7
861 Instruments, apparatus 0.000 0.002 0.002 11.5 0.8
862 Photo, cinema supplies 0.001 0.007 0.010 6.7 1.4
863 Developed cinema film 0.000 0.001 0.000 2.3 0.1
864 Watches and clocks 0.000 0.001 0.001 15.1 1.5
891 Sound recorders, producrs 0.001 0.001 0.000 1.8 0.1
892 Printed matter 0.001 0.003 0.002 2.1 0.6
893 Articles of plastic nes 0.000 0.004 0.002 22.4 0.4
894 Toys, sporting goods etc. 0.001 0.006 0.003 5.7 0.5
895 Office supplies nes 0.000 0.002 0.005 13.0 2.1
896 Works of art etc. 0.000 0.000 0.000 0.7 0.1
897 Gold, silver ware, jewelry 0.004 0.002 0.003 0.6 1.3
899 Other manufactured goods 0.001 0.006 0.005 7.9 0.8
0.002 0.007 0.008 3.5 1.1
Table 13.3. Changes in Brazil's Industrial Structure 1949-1985 Gross Value-

Added (Percentage Distribution)
1949 1963 1975 1980 1985
Nonmetallic mineral 7.4 5.2 6.2 5.7 4.3
Metal products 9.4 12.0 12.6 12.0 12.2
Machinery 2.2 3.2 10.3 10.1 9.2
Electrical equipment 1.7 6.1 5.8 7.0 7.6
Transport equipment 2.3 10.5 6.3 7.5 6.4
Wood products . 6.1 4.0 2.9 2.6 1.6
Furniture ___ 2.0 1.8 1.4
Paper products 2.1 2.9 2.5 2.8 2.9
Rubber products 2.0 1.9 1.7 1.3 1.8
Leather products 1.3 0.7 0.5 0.5 0.6
Chemicals" 9.4 15.5 12.0 14.5 17.3
Pharmaceuticals 2.5 1.6 1.7
Perfumes, soap, candles 1.2 0.9 0.9
Plastic products _ _ 2.2 2.3 2.2
Textiles 20.1 11.6 6.1 6.2 6.0
Clothing and shoes 4.3 3.6 3.8 4.7 5.2
Food products 19.7 14.1 11.3 11.6 12.0
Beverages 4.3 3.2 1.8 1.3 1.2
Tobacco 1.6 1.6 1.0 0.8 0.8
Printing and publishing 4.2 2.5 3.6 2.5 1.9
Miscellaneous 1.9 1.4 3.7 2.7 2.7
Total 100.0 100.0 100.0 100.0 100.0
"Includes chemicals, Pharmaceuticals, perfumes, and plastic products for 1949 and 1963.
Source: Baer et al. (1987, Table 3, p. 278) and IBGE.
these policies were successful. In key industrial segments, import-output ratios fell
quite dramatically (Table 13.4).
However, efforts to deepen the industrial base, as well as to adjust the economy
to the adverse macroeconomic shocks, were based on restricting imports. In 1986, the
ratio of imports to GDP in Brazil was 5.8%, lower than Japan's (an economy nine
times as large as Brazil) and one-sixth of Korea's (Table 13.5). Correspondingly, Bra-
zil's industrial base became highly diversified. The extent of import-substitution
driven diversification and absence of intraindustry specialization is suggested by
extremely low ratios of manufactured imports over manufactured value-added (Table
13.5).
As noted above, Brazil is now saddled with an increasing number of internation-
ally noncompetitive areas. They are a product, inter alia, of structural rigidities intro-
duced by the instruments used to spur industrial development. Entry and exit barriers,
overdiversified production structures, and high mandated domestic content levels
slowed down mobility, tied-up resources, brought about low degrees of inter- and
intraindustry specialization, and led ultimately to high cost structures.
Such policies also shielded producers from domestic and international competi-
tion. As the industrial sector matured, these protective barriers solidified, and firms
became increasingly secure in their market positions. Protection from competition
made firms less resilient and management less responsive to the rapid shifts in the
international economy: an accelerated rate of innovation and intense technological
Table 13.4. Import-Output Ratios in Selected Industrial Segments
Industrial Segments" 1967 1973 1980
Mechanical equipment (66) 45.6 38.0 18.5
Electrical and telecommunications equipment (72) 13.2 22.5 15.2
Paper and paper products (25, 64) 7.5 9.2 4.8
Chemicals (23 1 .2, 266, 42, 43, 5) 15.3 17.4 10.6
"Numbers in parentheses are the corresponding SITC code(s).
Source: Fasano Filho (n.d., Table 2).
and commercial rivalry were accompanied by increased protectionism in developed

countries.
The Scope and Intensity of R&D Activities among Industrial Firms
The inability of Brazilian firms to keep up with the international frontier is a reflection
of a highly heterogeneous industrial structure, relatively few competitive producers,19
and limited technological efforts. Brazil's large and diversified industrial basethe
result of a long process of import substitutionis insufficiently "deep." Relatively few
firms are in a position to consolidate and expand international market share on the
basis of substantial gains in productivity, continuous improvements in product quality
and reliability, or development of new designs. Most producers have based their mar-
ket position more on extensive exploitation of natural resources and reliance on low-
wage labor than on the quality and productivity of labor itself and the introduction of
new or better products.
Evidence strongly suggests that R&D, as a formal activity, is undertaken by rel-
atively few firms. The number of firms declaring expenditures on R&D in their income
tax returns fell from 1050 in 1976 and 1977 to 780 at the end of the 1981-1983 reces-
sion, but recovered to 1090 by 1985.20 R&D expenditures by these firms were 0.2% of
their net revenues in 1983, but doubled to 0.4% by 1985.21 These R&D expenditures
are highly concentrated. Data from 1983 show that state enterprises accounted for
62.6% of budgeted R&D outlays, with eight producers responsible for over half of the
total expenditures.22 Only about 25 private industrial groups were responsible for an
additional 17.4% of expenditures (Table 13.6).
The smallness of industry-related R&D in Brazil also is reflected in the distribu-
tion of researchers (Table 13.7). In 1986 there were 52,863 researchers out of 3.5 mil-
lion college graduates, a relatively low proportion; the ratio of researchers to popula-
tion was 4 to 10,000, whereas in developed countries the number is about 40 to 10,000
Table 13.5. Share of Manufactured Imports in Manufacturing, Value-Added Share of

Merchandise Imports in GDP Selected Countries, 1986 (%)
Indicator Brazil Indonesia Japan Korea Mexico Thailand Turkey
Merchandise imports/GDP 5.8 17.8 6.5 32.2 9.4 22.0 21.0
Manufactured imports/ 10.8 102.4 31.0 74.0 32.3 83.0 51.4
manufacturing value-
added
Source: World Development Report (1988) and IENIN data base.
Table 13.6. R&D Outlays in Brazilian Industry, 1983"
Proportion of Total Outlays
Nature of Firm (%)
Public sector enterprises 62.6
Private groups 17.4
Private firms focused on individual markets 20.0
încluded were 1118 firms, of which 43 were public enterprises and 1075 private producers. Firms in
the electric machinery and made-to-order capital goods subsectors were not included in this study. As a
result, the percentage for public enterprises is biased upward.
Source: Paulinyi (1984).
inhabitants. It is striking that 91.6% of researchers worked in government, the majority

in public universities (62.0%), followed by specialized S&T institutions (20%), state
enterprises (3.4%), and other agencies (6.1%).
Only 8.4% of researchers were in nongovernmental jobs. Most of them worked
in private universities (6.5%); relatively few worked in private research institutions
(0.5%) or other agencies (0.8%). Particularly noteworthy is the minute share of
researchers in private firms (0.6%). Even including state enterprises, the percentage of
R&D personnel employed by the productive sector is modest, accounting for less than
4% of the total.23
Information compiled in the mid-1980s indicates that R&D continues to be
undertaken by a narrow set of firms; a core of 366 firms can be regarded as "R&D
active" (Table 13.8).24 As expected from international experience, "technology-inten-
sive" segments such as electronics, vehicles, and the chemical/petrochemical/phar-
maceutical cluster contain the largest share of firms active in R&D. Conversely, seg-
ments such as food and beverages, textiles, garments and footwear, and leather and
wood products have relatively few R&D-engaged firms, with capital goods and metal
fabrication somewhere in between.
Among R&D active firms, 88 are major exporters, most concentrated in steel,
vehicles, petrochemicals, chemicals, and autoparts. Some of these are world class pro-
ducers in sophisticated segments (such as aircraft, missiles, high-performance pistons,
Table 13.7. Distribution of Researchers According to Place

of Activity, 1986
Total Doctorates
Institution No. % No. %
Universities 36,112 68.3 9,952 86.6

(public universities) 32,775 62.0 n.a. n.a.
Specialized S&T institutions 10,856 20.5 1,358 11.8
Government agencies 3,203 6.1 82 0.7
State enterprises 1,811 3.4 38 0.3
Private institutions 727 1.4 60 0.5
Private firms 295 0.6 15 0.1
Other 111 0.2
Total 52,863 100.0 11,492 100.0
Source: Martins and Queiroz (1987).
Table 13.8. Sectoral Distribution of Export-Oriented and R&D Engaged Brazilian Firms,an Firms.
1985-1988
R&D R&D Active R&D Active
R&D Active With No Total* as % of
Industrial Sector Active Exporters" Exports Firms Total Firms
I Steel and other metals 24 17 7 517 4.6

II Metal fabrication 9 3 6
III Capital goods 49 7 42 419C 13.8C
IV Vehicles 17 12 5 70 24.3
V Auto parts 12 8 4 139 8.6
VI Rubber products 3 3 0 na na
VII Electronics 90 4 86 258 34.9
VIII Petrochemicals 34 12 22
IX Chemicals 47 8 39
X Fertilizer 8 0 8
XI Pharmaceuticals/fine chemicals 13 0 13 635^ 16.6d
XII Pulp and paper 9 4 5 131 6.9
XIII Nonmetallic minerals 12 0 12 231 5.2
XIV Textiles, garments, footwear 8 4 4
XV Leather goods, wood products 4 2 2 757'' 1.6'
XVI Food and beverages 24 5 19 712 3.4
XVII Others 3 0 3
Cl Metal-mechanics I-V 111 47 64 1145 9.7
C2 Electronics VII 90 4 86 258 34.9
C3 Chemicals/petrochemicals VIII-XI 102 20 82 635 16.1
C4 Traditional XIV-XVI 36 11 25 1469 2.5
Total 366 88 278 3869 9.5
"Drawn from the Cacex data of firms exporting over $6 million in 1985.
*The universe of medium to large manufacturing firms with published balance sheets, and drawn from Gazeta Mercantil,
Balance Annual, 1987.
Includes II and III.
''includes VIII-XI.
"Includes XIV-XV.
Source: Own compilation of R&D engaged firms based on information from ANPEI (National Association of Industrial R&D
Centers), the Institute of Industrial Economics of the Federal University of Rio de Janeiro, and newsmagazines/newspaper
accounts of R&D activity by individual firms.
and special alloys). Substantial production, design, and marketing experience, com-
bined with purposeful R&D efforts, result in their competitive strength. A number of
R&D active producers are multinational firms, mostly in the metal-mechanics indus-
tries (particularly vehicles), with some in rubber products and petrochemicals. Their
R&D efforts were not necessarily undertaken in response to export requirements
(although many were drawn to export markets in response to strong government
inducementssuch as the BEFIEX export program), but to local market demands
(e.g., for ethanol-based engines or large hydroelectric turbines). However, the majority
of R&D-engaged exporters are those that combine Brazil's traditional sources of com-
parative advantage (abundant natural resources and inexpensive labor) with a mea-
sure of product development and process improvements to stay competitive in world
markets. They are producers of steel, autoparts, pulp and paper, chemicals, textiles,
and food and beverages.
Most firms that carry out R&D but do not export are in electronics and capital
goods. There are segments with traditionally high R&D intensity, as suggested by
international experience, and that since the mid-1970s have expanded rapidly, behind
high protectionist barriers. Although government policies and institutional support
mechanisms have induced these producers to undertake substantial R&D, such activ-
ity has been insufficient to close in the technological frontier. Moreover, some efforts
probably have been misdirected and wasteful. Continuously high levels of protection
have allowed producers to fall behind best practice without being penalized by com-
petition. Across-the-board import substitution also led to excessive product fragmen-
tation, with the dispersion of technological resources. As a result, low levels of intrain-
dustry and intranrm specialization precluded these producers from reaping economies
of scale in production and development of new products.
R&D is just one dimension of firms' technological activities, however, and pos-
sibly not the most important. Production planning and organization, troubleshooting,
rescaling, raw materials adaptation, and quality control are other technological tasks
that have a direct bearing on firms' competitiveness. But the fact that so few firms were
engaged in R&D is significant, as firms tend to formalize R&D activities once com-
petitive pressures require new designs and major improvements in production pro-
cesses.
Moreover, despite the fact that many industrial firms are engaged in minor tech-
nological activitiesranging from the design of tools to improvements in plant layout
and production processesmost of those activities have had only limited impact in
bringing firms closer to best practice.25 A detailed study of these firms' technological
behavior has shown that raw material inventory was monitored by over 92% of sam-
pled producers, but more than 81% did so manually. Similarly, nearly 93% of firms
were engaged in quality control, while just 28.8% employed modern methods (such as
statistical process control).26 At the same time, only 23.5% were committed to devel-
oping new products in a systematic way. Imitationthrough reverse engineering or
hiring away competitorswas perceived to be the general rule in industry.27 These
findings suggest that most technology generated in Brazilian industry consists of minor
innovations and adaptationssufficient to cater to a relatively protected domestic
market but inadequate to sustain its competitive position in world markets.
In sum, the evidence assembled in this section points to very limited commitment
by Brazilian industrial firms to R&D and to technological development in general. As
shown in the next section, such limited efforts are not offset by a regime to facilitate
technology imports. The effectiveness of imported technology is hampered, in any
case, by the weak absorptive capacity of Brazilian industry.
TECHNOLOGY POLICIES
Explicit policies for industrial technology development in Brazil have followed the
broader industrial policies of the Government. The acquisition of foreign technology
through arms-length technology transfer, capital goods imports, and direct foreign
investment has been strongly influenced by the government objective of minimizing
the outflow of foreign exchange and promoting import substitution. Similarly, finan-
cial incentives to foster R&D and boost the engineering consulting capabilities of
national enterprises were focused on import-substituting activities or national pro-
ducers that were competing in areas dominated by multinational producers.
Arms Length Technology Transfer
Brazil's current technology transfer regime is designed to improve the bargaining posi-
tion of national firms and minimize foreign exchange outflow.28 Since 1958, the Cen-
tral Bank has controlled royalty payments and has established that the maximum roy-
alty should be 5% of net sales. The Government's main concern was that foreign firms
were using technology transfer payments as a way of remitting profits. Subsequently,
in 1962 the law governing foreign investments in Brazil29 established strict controls on
technology transfer, which are still in effect. The law required that foreign payments
for technology transfer be registered with the Central Bank. It also prohibited tech-
nology transfer payments between subsidiaries and their parent companies and
between joint ventures and foreign partners who held more than a 50% stake in the
venture. The rationale was that the parent firms already earned a return on the tech-
nology, through profits, and therefore should not be allowed to deduct technology
licensing payments. The law maintained the 5% maximum royalty rate, and also
restricted deducting technology transfer payments for fiscal purposes to a maximum
period of 5 years.30
In 1970 the National Institute of Industrial Property (INPI) was created. Among
other responsibilities, it was to regulate technology transfer (taking over from the Cen-
tral Bank). Technology transfer agreements were divided into five categories. Each
agreement must be registered with INPI as to terms and payments authorized, period
of validity, required Brazilian participation, and special provisions, depending on type
of contract. Royalties for patent licenses and trademark licenses may be paid only if
the licenses are registered in Brazil. Moreover, the intellectual property regime in Bra-
zil does not offer significant patent protection in the areas of chemicals, pharmaceu-
ticals, metal alloys and mixtures, and alloys in general. One of the principal rationales
of the Brazilian property rights system is to protect technologically less capable local
firms against potentially predatory behavior by foreign producers and to reduce the
cost of appropriating relevant foreign technology. However, a frail intellectual prop-
erty rights regime, including weak trade secret protection, possibly deters foreign firms
from transferring or using locally their most advanced or up-to-date technology.
Industrial technology license agreements are less onerous since such licensing
does not depend on INPI's registering the patent or trademark.31 However, INPI treats
such an agreement as a sale: after expiration of the contract, INPI regards the licensee
as the owner of the technology. In addition, the parties must satisfy INPI that the tech-
nology
is not available domestically and that its transfer to a Brazilian licensee is in line with
national development objectives;
brings real advantages for the development of the industrial sector;
improves product quality and allows for substitution of Brazilian products for
imports; and
allows the licensee to absorb and master the technology within the lifetime of the
contract.
Registration of technical and industrial cooperation agreements depends on convinc-

ing INPI that the services are not available locally and that there will be short-term
benefits to the sector. This might mean producing a quality product with export and/
or import substitution possibilities. For technical service agreements, INPI must be
satisfied that the services are not available domestically. It also must approve the
detailed schedule of payments for foreign technicians.32
Maximum approved royalty rates range from 2% for plastic and rubber articles
to 5% for electrical equipment." Those royalties are based on net sales, which can be
considerably lower than total sales.34 Royalties for trademarks have a maximum of 1 %.
In addition, royalty payments are taxed at a rate of 25%, and that tax is counted within
the specified maximum royalty limits. The maximum duration is 5 years, with a pos-
sible extension for another 5 years, at the discretion of INPI. Also, as noted above,
royalty payments are not permitted between subsidiaries and their parents or between
joint venture firms and foreign partners holding more than a 50% stake in the venture.
Technology contracts prohibit restrictive clauses involving limits on exports,
pricing guidelines, use of tied inputs, purchase of other technology, secrecy about the
technology after expiration of the contract, and obligatory transfer to the seller of
improvements made in the technology by the buyer. The seller, however, is obliged to
pass on improvements made to the technology after it was sold and is liable for any
legal action that originates because of defects in the technology or intellectual property
right infringements, even by third parties. In addition, full disclosure of all technical
data is required, and engineering drawings must be provided, as well as all information
necessary to update the know-how. The latter requirement is of extreme concern to
many technology suppliers who fear that technology information may be appropriated
by third parties.
The impact of INPI's attempts to control technology imports has not been eval-
uated systematically but may be quite significant. Between 1972 and 1987 more than
23,500 contracts were submitted to INPI for approval. Of those, about 18,000, or
roughly 75%, were approved. INPI's approval rate has followed a generally downward
trend, the result of its stricter control, especially regarding technical services (Table
13.9).35
Table 13.9. INPI: Approvals of Technology Transfer

Contracts, 1978-1987
Contracts Presented Contracts Percentage
Year for Approval Approved Approved
1978 1473 1451 98.5

1979 1456 1416 97.3
1980 1576 1332 84.5
1981 1426 1178 82.6
1982 1438 1135 78.9
1983 1296 969 74.8
1984 907 786 86.7
1985 1043 769 73.7
1986 1185 885 74.7
1987 1815 1213 65.8
Total 13615 11134 81.6
Source: Cardo/,o (1988), based on INPI data.
Fifty percent of all technology contracts approved between 1978 and 1987 were
concentrated in four of the 70 economic sectors that INPI uses to classify contracts
(mechanical, metallurgical, chemical, and mining). Although the shares of most of the
leading sectors have been falling, electronics and electrical material are expanding rap-
idly, with the proportion of approved contracts increasing from 4% in 1978 to more
than 15% in 1987. This partly reflects the increased importance of outright technology
purchases by the electronics industry, since direct foreign investment has been severely
restricted in a number of its segments.
Data on technology payments are available for 1979-1989 (Table 13.10), show-
ing payments of US$2.4 billion during that period. Roughly 79% of the total was for
specialized technical services even though their share fell markedly over the period.
About 18% was for unpatented industrial technology (including 7.3% for unpatented
industrial technology for made-to-order capital goods). The smallest share (just 3%)
went for patent and trademark licenses. These values understate the actual inflows
since payments by subsidiaries to their parent companies are not allowed. An almost
continuous decline in the values recorded is notable and this fall is not explained by
the economic recession alone.
It is not clear to what extent INPI has improved the bargaining position of domes-
tic firms and helped them obtain technology more cheaply, however. There are various
reasons to surmise that INPI's controls may have been deleterious to the national
interest in two respects. First, the regulations discriminate against national firms
because these firms must pass their contracts through a bureaucratic process for
approval. In contrast, subsidiaries of foreign companies have direct access to their par-
ent companies' technology and can transfer the technology even if they cannot deduct
the cost for tax purposes. Second, for foreign technology suppliers, including investors,
restrictive transfer controls as well as inadequacies in the direct foreign investment
regime (e.g., not being able to capitalize technology contributions or to deduct tech-
nology fees) make it not worthwhile to license or even use their best technology in
Brazil. This situation is aggravated by the lack of adequate protection from the intel-
Table 13.10. Payments for Technology Transfer Contracts 1979-1987 (in

Millions of Current Dollars)
Industrial Industrial
Patents and Cooperation Technology Technical
Year Trademarks Agreements Licenses Services Total
1979 9 6 11 287 313
1980 12 11 14 284 321
1981 12 18 12 234 276
1982 5 17 10 208 240
1983 12 10 14 182 218
1984 9 8 8 177 202
1985 5 21 41 108 175
1986 2 20 43 119 184
1987 3 40 27 123 193
1988 3 12 27 99 141
1989 3 15 39 112 169
Total 75 178 246 1,933 2,432
Source: Cardozo (1988), based on Annual Reports of the Central Bank.
lectual property regime. The net result is that Brazil may be receiving obsolete or out-
moded technology.
Imports of Capital Goods
Access to the most productive capital goods is critical if an economy is to benefit from
embodied advances in technology. Brazil has relied much less than fast growing East
Asian economies on imported capital goods for acquiring foreign technology (Table
13.11). Moreover, the ratio of capital goods imports to gross domestic investment
(GDI) has decreased since 1980. This resulted from the policy of promoting the local
capital goods industry through trade restrictions on imports and special financial
incentives for local production. Excessive protection may be hampering the interna-
tional competitiveness of local industry.
The restrictive capital goods import regime has had significant impact on mod-
ernization as well. A relatively large number of manufactures in many specialized
product areas (even more from the United States or Japan) has meant an overly diver-
sified capital goods industry lacking the economies of scale or specialization necessary
to reduce costs, keep up with new technologies, and stay competitive. Excessive depen-
dence on more expensive and often outmoded locally produced capital goods, espe-
cially in electronics and process control, has placed the downstream user industry at a
disadvantage with respect to Brazil's international competitors.
Foreign Investment
Brazil's foreign investment policy has reflected the government's objective of attract-
ing capital to specific industrial segments and (after 1972) of obtaining export com-
mitments from individual firms. Technology development has been of secondary
importance to import substitution and balance-of-payments considerations. Although
the inflow of direct foreign investment (DPI) has been considerable, Brazil may not
have been receiving the most advanced technology. The overall effect of restriction has
made foreign investment in Brazil, particularly investment with a high technology
content, relatively less attractive than in other locations.36
Table 13.11. Imports of Capital Goods" as Ratio of GDI GDI

1965 1970 1975 1980 1985 1986 1987
Hong Kong 1.949 3.76 2.962 2.224 4.073 3.826 3.848
South Korea 0.991 0.902 1.272 1.119 1.114 1.034 1.066
Malaysia 1.768 1.488 1.435 1.441 1.454 1.541 1.636
Singapore 5.878 3.357 3.592 4.42 3.489 3.801 4.160
Thailand 0.889 0.714 0.824 1.112 1.029 0.944 1.041
Brazil 0.251 0.327 0.409 0.455 0.376 0.299 0.259
Mexico 0.352 0.302 0.312 0.370 0.400 0.390
United States 0.181 0.270 0.442 0.575 0.559 0.574 0.591
"Capital goods are defined as the following SITC Revision 2 categories: 659, 711, 712, 714, 715, 716, 717, 718, 719 (minus
7194), 721, 722, 723, 7249, 726, 7295, 7296, 7297, 7299, 731, 7322, 7323, 7324, 7325, 7327, 7333, 734, 735, and 861. The
ratios in the table are greater than one for some of the Asian countries because many of the imported capital goods are rcex-
ported as part of new products.
Source: Own computations based on UN trade data and BESD system.
Profit remittances are limited to 12% of the original investment plus reinvested
earnings." In addition, the original foreign currency value of the investment is not
corrected for inflation, which means that the original base for remittances decreases in
real terms over time. Furthermore, parent companies may not receive royalty pay-
ments for patents and trademarks from their subsidiaries and may not capitalize their
technology as equity contributions to the investments.38
For minority joint venture partners, there are also arbitrary and strict limits on
maximum payments for technology transfer (as noted above). In addition, the foreign
investor who brings technology to the joint venture has to pay a 25% tax on this part
of his capital contribution to make it part of his remittance base.
Still, the foreign investment regime is quite open. In general, up to 100% foreign
ownership has been allowed in most areas. Until recently, the only sectors closed to
DPI were petroleum extraction and refining, certain segments of the informatics
industry, communications media, and most domestic transport, although there were
also some restrictions on mining, banking, insurance, and other financial activities.
The new Constitution adopted in 1988 has taken a more nationalistic posture toward
DPI. Its main restrictions were limiting foreign control in mining, mineral explora-
tion, and production of electric power, and prohibiting foreign oil companies from
engaging in oil exploration through risk contracts. However, the new Government that
took power in March 1990 has announced a more open policy toward foreign invest-
ment, and a possible reduction in the scope of restrictions in the informatics sector.
Brazil has the largest stock of foreign capital of any developing country. Regis-
tered foreign investment was US$27.9 billion in 1987. Almost 75% of all foreign
investment is in manufacturing, 20% in services, 3% in the extractive mineral industry,
less than 1% in agriculture, and 2% in other activities. Within manufacturing, the sub-
sectors accounting for the largest shares of foreign investment are automobiles (11%),
basic chemicals (9%), mechanical (9%), and electrical and communications (8%). This
particular concentration pattern resulted because most foreign investment was
attracted under specific sectoral programs (e.g., for automobiles, shipbuilding, capital
goods, and chemicals). Such programs provided fiscal and financial incentives (as well
as protection in the local market) in exchange for performance requirements pertain-
ing to investment volume, local content, and exports.39 At the same time, foreign
investment in key areas for industrial competitiveness has been limited or even totally
excluded.
The most comprehensive attempt to block foreign investment and promote
domestic firms' capabilities has been in "informatics," broadly defined to include
computers and peripherals, professional and industrial electronics, and microelec-
tronics. Brazil's informatics policies, set in mid-1976 and predetermined to last until
late 1992, reserved the domestic market exclusively for national producers, fostering
the establishment and growth of a sizable number of firms. These policies, managed
until recently by a powerful Special Secretariat of Informatics, have been sufficiently
focused to attract resources and create a new industrial segment in a relatively short
period of time. However, they have not brought adequate levels of productive effi-
ciency, while design efforts have resulted at most in incremental innovations within
well-known and relatively open technologies. The industry has yet to become inter-
nationally competitive in most product lines, even though domestic prices have come
down substantially for more mature goods. Particularly for products that have become
high-volume commodities or that have undergone radical improvements, the price-

performance differential between the international and domestic market has actually
widened since the mid-1980s.40
Financial Incentives for Local Technology Development

Financial incentives have been the main instrument for encouraging the development
of technological capabilities at the firm level. Since 1973, FINEP (Agency for Financ-
ing Studies and Projects) has used subsidized loans, risk-sharing instruments, and, to
a lesser extent, equity participation to foster national firms' technological activities. In
the period 1973-1989, FINEP contracted 1761 technology development support
operations for approximately US$810 million. The focus has been on the develop-
ment of import substitutes (as in chemicals and Pharmaceuticals) and products that
would allow national firms to compete with foreign-owned producers in the domestic
market (as in capital goods, electronics, vehicles, and autoparts). There has been a
growing emphasis on establishing R&D and quality control labs and developing and
improving products, processes and tools.
Although there is not enough information to evaluate the impact of FINEP's pro-
grams, it is likely that most national firms with significant R&D activity have bene-
fited from FINEP's assistance. Nonetheless, the effectiveness of its actions has been
limited not only by relatively narrow lending criteria, but by internal bureaucratic
obstacles to timely operation, and the difficulties small, technically based firms
encounter in qualifying for its programs. Larger firms, on the other hand, increasingly
find FINEP's financial capabilities too limited for their innovation finance require-
ments.41
In conclusion, the technology policy regime in Brazil has been characterized by
objectives other than the acquisition of technological capabilities that would allow
firms to become internationally competitive. Government interventions have been
oriented instead toward enabling domestic firms to operate in new areas, design
import substitutes with their own or acquired means, achieve a measure of technolog-
ical "autonomy," and displace multinational firms from certain key industrial seg-
ments. As a result, the country has failed to attract best-practice technology via direct
foreign investment or through arms-length transactions. A combination of weak
domestic technological efforts and restrictive access to the most valuable foreign tech-
nology appears to have hampered the modernization efforts of Brazilian firms. As
argued in the next section, substantial public sector involvement in technological
activities has been a weak substitute for greater involvement by the productive sector
and a more flexible technology policy regime.
INSTITUTIONAL NETWORK SUPPORTING TECHNOLOGY DEVELOPMENT
Background and Current Structure

Shortly after the inauguration of the new civilian government, in March 1985 a Min-
istry of Science and Technology (MCT) was created. Although initially MCT sought
to take control of all existing institutions working in the area of S&T, including STI
and EMBRAPA (the major government institution responsible for agricultural
research), it ended up with direct responsibility over only three key agencies: CNPq,
which still directly managed and oversaw a national network of research institutes,
FINEP, and SEI. Subsequently it acquired control over the National Institute of Tech-
nology (INT) as well as several other specialized institutes.
With the creation of the MCT, clearer priorities were established. Policy was
directed to the expansion of infrastructure in science and technology, principally
related to human resource development. There was also an attempt to define strategic
sectors for Brazil to keep up with international trends. The new policy emphasis
reflected a greater awareness of the weaknesses still inherent in the S&T base of the
country, especially in qualitative terms, and the view that a strategic focus on new tech-
nologies was necessary for a maturing economy seeking to close the gap with advanced
nations.
Throughout 1985-1986 there were many broad-based meetings, conferences,
and congresses on S&T, as many who had not had a chance to express their views saw
their opportunity during this opening up of the system. There was much euphoria on
the possible contribution of S&T to all aspects of economic and social development.
There were also ambitious proposals to increase R&D expenditure to 2% of GDP by
the 1990s. In addition there were proposals to include a section on science and tech-
nology in the new constitution being drafted.
The new National Development Plan placed special emphasis on S&T and
devoted a chapter to the subject. It recognized that in the past, government support to
technology suppliers was frequently based on academic criteria without considering
the relevance of the technology to market needs. It also recognized that the vast net-
work of publicly-supported R&D institutes suffered from excessive diversification and
lack of sufficient integration with the needs of industry. Furthermore, in the area of
regulation the plan recognized that numerous reforms were necessary in standardiza-
tion, technical norms, and testing laboratories, to improve the performance of the pro-
ductive sectors and reduce high rates of wastage.
The MCT hoped to promote the increase of total expenditures of R&D to nearly
1% of GDP in 1986 and to higher levels subsequently by encouraging the productive
sector to invest more in R&D through an improved economic environment, increased
financing for technology, and the introduction of fiscal incentives for R&D. However,
due to the difficult budgetary situation and increasing macroeconomic instability,
R&D expenditures remained flat, despite the fact that fiscal incentives for R&D were
enacted as part of the new industrial policy program in 1988.
Since 1989 the Special Secretariat for Science and Technology has been the apex
S&T institution in Brazil. In addition, 17 major federal implementing agencies for
S&T policy receive at least 1% of the S&T budget. Combined they account for over
90% of total allocations for S&T (Table 13.12). Approximately 40% of the allocations
finance research related activities (metrology, norms and standards, normative coor-
dination, for example), and the remainder goes to postgraduate training and actual
research.
Public R&D Institutes

A complex and differentiated government institutional network carries out public sec-
tor R&D activities in Brazil. In major areas, such as health and biological sciences, as
Table 13.12. Key S&T Agencies According to Budgetary Allocations"

Final 1988 Budgetary
Allocation
Agency Function Amount'' Percent
CNPq The National Council for S&T Development T,R 291,639 17.01
EMBRAPA Brazil Agricultural Research Corp. T,R 200,542 11.69
Secretariat of Sec. of S&T' N,R 183,903 10.72
CAPES Coordinating Agency for Postgraduate Training T,I 139,108 8.11
CNEN National Council for Nuclear Energy N,R,T 129,128 7.53
EM BRATER Brazil Rural Extension Corp. I 98,916 5.77
IFES Federal University and Other Teaching Inst. T,R 93,842 5.47
CSN National Security Council N 89,431 5.22
EMFA Joint Chiefs of Staff R 72,760 4.24
Secretariat of Min. of Mines and Energy'' N 65,617 3.83
INPE Institute of Space Research T,R 48,421 2.82
FIOCRUZ Foundation Institute Oswaldo Cruz T,R 33,137 1.93
Min. of Aeronautics R 31,931 1.86
INMETRO National Institute of Metrol., Stand. I,N 27,178 1.58
Secretariat of Min. of Justice 25,199 1.47
CEDATE Center for Tech. Supp. of Ed.'' 21,083 1.23
CTI Center for Informatics Technology R 19,918 1.16
INPA National Institute for Research of Amazon Reg. R 18,981 1.11
Subtotal 1,590,644 92.75
Total S&T budget 1,714,851 100.0
T, training; R, research; D, dissemination of information; I, institutional development; N, normative coordination.
'Amounts in 1987 dollars.
'Includes the FNDCT and the PADCT research funds.
Êighty percent is allocated for equity investment in NUCLEBRAS.
''Ninety percent is allocated for classroom scientific equipment and instruments.
Source: CNPq, "Recursos do Tesouro da Uniao: Dotacao Inicial, Dotacao Final, e Despesa Realizada 1980-89," May 1989.
well as in experimental physics and chemistry, most resources are in the federal and
State of Sao Paulo universities. Public enterprises' in-house R&D facilities cover
energy generation and conservation, oil exploration, telecommunications, and aircraft
development. Government institutions not attached directly either to federal or state
universities, or to public sector enterprises, play an important role in agriculture, food
technology, Amazon ecology, biotechnology, tropical disease research, physics and
nuclear technology, aerospace, and computer sciences. In addition, development
efforts in various areas of engineering and mining technology are undertaken by a few
state-level institutions (the most important of which is IPT of Sao Paulo).
Government-undertaken industrial R&D has not been very effective due to the
generally weak linkages with the productive sector. A survey of sources of technology
in Brazilian industry notes that less than 2.5, 5.1, and 3.9% of product designs, tool
designs, and manufacturing processes, respectively, has originated in research insti-
tutes.42
Although industry ties are strong in some cases, the dispersion of resources
through an excessive number of research projects and competing activities hampers
efforts. An example of the dispersion of a research program in the public sector is the
work program of CPqD, the research unit of Telebras, Brazil's public telephone hold-
ing company. In 1988 CPqD was the largest and most sophisticated applications lab-
oratory in Latin America. It employed 400 professionals directly engaged in R&D
work, in addition to sponsoring personnel from industry and universities. CPqD's
budget of US$60 million was allocated to a very broad research agenda, approximately
80 R&D projects in seven priority areas: electronic switching, digital transmission,
optical communications, data and text communications, satellite communications
equipment, tools, and materials. Even adjusted for skilled labor cost differences, a bud-
get of US$60 million is minute for the scope of CPqD's research agenda. Budgetary
allocation would average less than US$8 million per program or US$1 million per
project, whereas each program could in itself justify the whole budget. The absence of
research focus and specialization has led to systematic delays in the CPqD research
chronogram and, therefore, to delays in the market launching of its applications.
These delays have had a high cost for the economy because in many cases, such as the
digital exchange program, government policy prevented the use of existing, less costly,
foreign technology while waiting for the local technology to be developed.43
In two areas, however, government R&D efforts have been quite effective: agro-
technology and aerospace. In agrotechnology, the government agency EMBRAPA
(Empresa Brasieira de Pesquisa Agropecuaria) has coordinated and partly executed a
complex and far-ranging national agricultural research program. This has resulted in,
among other things, the incorporation of previously infertile land through a better
understanding of soil biology and the successful introduction of new plant varieties.
In aerospace, Brazil (with China and India) probably has the most advanced
research program among industrializing countries. CTA (Centra Tecnologico da
Aeronautica) and INPE (Institute Nacional de Pesquisas Espaciais) have been the core
R&D institutes in this subsector. CTA, in particular, has generated and effectively
transferred aircraft, rocket, fuel, and other aerospace-related technology to industry.
In addition, it has had an important role in the development of gasohol-based engines.
Among industry-oriented institutes, IPT (Institute de Pesquisas Tecnologicas) of
the State of Sao Paulo is possibly the one that has forged the closest links with the pro-
ductive sector. Other than IPT (and two more state-level R&D institutesthose serv-
ing Minas Gerais and Bahia), government-led, industry-directed efforts have been
small and not very effective.
Public and Aggregate R&D Expenditures
Public sector R&D has been dominant not only in scope but in resources allocated.
Although there are no firm estimates of total R&D expenditures in Brazilprivate
sector outlays are not known with certaintymost sources estimate that public sector
expenditures account for between 70 to 90% of total R&D outlays. Based on actual
public expenditures, a series of imputed private and total R&D expenditures can be
constructed (Table 13.13).
Although in absolute terms R&D outlays in Brazil are not modest compared to
other industrializing countries, they are small in comparison to developed economies
(Table 13.14). Moreover, after a significant increase in the 1970s, R&D expenditures
as a proportion of GNP remained basically flat throughout the 1980s, whereas they
expanded substantially in the fast growing East Asian economies. As a result, absolute
levels of R&D expenditures in South Korea, for example, are at least twice as large as
Table 13.13. R&D Expenditures in Brazil, 1981-1988 (in Constant 1987 Dollars)
Public
Year Expenditure" GDP PE/GDP TE1/GDP* TE2/GDP''
1981 966,133 253,785,250 0.38 0.54 0.42

1982 1,184,970 257,988,623 0.46 0.66 0.51
1983 938,207 247,500,345 0.38 0.54 0.42
1984 907,447 253,138,254 0.36 0.51 0.40
1985 1,242,271 282,347,174 0.44 0.63 0.49
1986 1,455,239 314,477,537 0.46 0.66 0.51
1987 1,485,365 323,573,075 0.46 0.66 0.51
1988 1,468,595 322,597,503 0.46 0.65 0.51
"For 1981 -1987 these are actual expenditures; for 1988, it is the final budget allocation.
*TE 1 is total imputed expenditures on the assumption that public expenditures make up 70% of R&D outlays in Brazil.
C
TE2 is total imputed expenditures on the assumption that public expenditures make up 90% of R&D outlays in the country.
Source: CNPq, "Recursos do Tesouro da Uniao: Dotacao Inicial, Dotacao Final, e Despesa Realizada 1980-89," May 1989.
in Brazil, although the Brazilian economy is more than twice the size of the Korean
economy.
Is Brazil devoting an insufficient amount of resources to R&D? It should be
emphasized that even if R&D outlays are stagnant, it certainly does not imply that
Brazil is spending too little on scientific and technological development. In the absence
of knowledge about rates of return in R&D compared to other economic activities,
not much can be said about the optimal level of R&D expenditures. On the one hand,
a number of countries that are allocating larger amounts of resources to R&D are not
necessarily reaping greater benefits. For example, between 1977 and 1985, India dou-
bled its relative outlays to R&D without major gains in terms of technological capa-
bilities in the productive sector. On the other hand, countries that have improved their
competitive position have also been increasing their R&D efforts on a systematic basis.
More important, most of these efforts have been undertaken by the productive sector.
In contrast, not only the level and intensity of Brazil's R&D expenditures have
Table 3.14. R&D Expenditures in Selected Countries (R&D Expenditures in Absolute Terms
and as a Proportion of GNP)
Absolute Expenditure" R&D Intensity
Country Amount Year 1970 1977 1982 Latest Year

Brazil 1448 1982 0.24 0.70 0.59 0.59(1987)''
Argentina 1087 1981 n.a. 1.80 0.20 0.20(1982)
Mexico 881 1984 0.20 0.30 0.20 0.60(1984)
India 1482 1984 n.a. 0.50 0.76 1.0 (1985)
South Korea 1307 1983 0.39 0.60 0.90 1.8 (1986)
Taiwan 505 1985 n.a. n.a. 0.90 1.06(1985)
Japan 39 1 1 7 1987 1.90 2.00 2.40 2.9 (1987)
United States 100823 1987 2.60 2.10 2.50 2.6 (1987)
"Expenditures in millions of 1982 dollars.
''Average between the low and high estimates of Table 13.13.
Source: UNESCO, Statistical Yearbook 1988, Paris, 1989.
been flat since the beginning of the decade, but more important, most R&D is both
financed and undertaken by government (Table 13.15). At the beginning of the dec-
ade, government was responsible for 67% of R&D financing and industry for only
20%. Estimates for 1988 indicate that with an additional 10% transferred from gov-
ernment and other sources, industry's share of innovation financing may have reached
30%.
This pattern of government dominance of R&D expenditures is also apparent in
other industrializing economiessuch as Argentina, Mexico, and Indiathat have
not been able to sustain their competitive position in world markets for more sophis-
ticated goods. In South Korea and Japan, in contrast, industry both finances and car-
ries out most R&D. In the United States, financing is shared equally by government
(a good proportion of R&D being defense related) and industry, although industry ulti-
mately undertakes most R&D.
Human Capital Formation

Brazil's education system is one of the main obstacles to the country's modernization
and technological upgrading. Although major deficiencies characterize all compo-
nents of the system, more fundamental weaknesses are observed in primary education
and at the top end in science and engineering.
Primary and Secondary Schooling
Technical progress in industry is increasingly dependent on an educated labor force.
Low levels of basic education and skills are not compatible with the complexity, pre-
cision, and consistency of modern industry. Yet in 1980, 73% of the Brazilian labor
force either had no education or had not completed primary school, a figure among
the highest for middle-income countries. In 1985 Brazil's total secondary enrollment
represented only 35% of secondary school-age population, well below the average for
middle income countries (Table 13.16).
Brazil's heavy investment in vocational and technical training has compensated
to a limited extent for the weakness of the formal school system. Although the pro-
portion of students enrolled in vocational training as a proportion of the working age
population is less than in South Korea or Taiwan, it is comparable to that of Mexico
Table 13.15. R&D Funding and Expenditures in Selected Countries"

Source of Funding Source of Expenditure
Country Year G I F O I HE GA
Brazil 1982 67 20 5 8 30 17 53
Argentina 1981 95 1 4 41 22 37
Mexico 1984 15 1 1 83 30 51 19
India 1984 87 13 26 74
South Korea 1986 19 81 67 11 22
Japan 1985 21 79 67 20 13
United States 1986 47 50 3 73 12 15
a
G, government; I, industry; F, foreign; O, other; HE, higher education; GA, government agencies.
Source: Unesco, Statistical Yearbook, Paris 1988.
Table 13.16. Indicators of Investments in Human Capital in Selected NICs and Japan
South Korea Taiwan Hong Kong Singapore Brazil Mexico India Indonesia Japan
Percent age group enrolled in

Primary education
(1965) 101 97 103 105 108 92 74 72 100
(1985) 96 100 105 115 104 115 92 118 102
Secondary education
(1965) 35 38 29 45 16 17 27 12 82
(1985) 94 91 69 71 35 55 35 39 96
Tertiary education
(1965) 6 7 5 10 2 4 5 1 13
(1985) 32 13 13 12 11 16 9 7 30
No. of tertiary students per 100,000 population
(latest year) 3606 2080 1410 1406 1140 1508 776" 600 2006
No. of tertiary students in CSE* (1000) 585 207 36 22 535 563 1443 235 707
(Year) (1987) (1984) (1984) (1983) (1983) (1986) (1980) (1985) (1986)
As % of population
Total 1.39 1.06 0.67 0.89 0.40 0.70 0.21 0.14 0.58
Urban 2.02 1.36 0.72 0.89 0.58 1.02 0.97 0.53 0.77
No. of students in SME' ( 1 000) 320.6 151.7 27.5 16.2 323.3 336.9 1269.9 137.3 486.9
As % of population
Total 0.76 0.78 0.51 0.73 0.24 0.42 0.19 0.09 0.40
Urban 1.10 1.00 0.55 0.73 0.34 0.59 0.86 0.33 0.53
No. of students in engineering only (1000) 227.6 128.7 21.1 i5.4 164.6 281.8 397.0 109.5 418.9
As % of population
Total 0.54 0.68 0.41 0.61 0.13 0.35 0.06 0.07 0.34
Urban 0.78 0.85 0.42 0.61 0.17 0.50 0.27 0.27 0.45
No. of students enrolled in vocational
training (1000) 814.5 404.6 31.7 9.4 1481.0 853.6 397.7 1061.3 1415.4
(Year) (1986) (1984) (1984) (1984) (1985) (1986) (1981) (1986) (1986)
As % of population of working age 3.06 3.24 0.86 0.54 1.83 2.0 0.1 1.14 1.71
"1980.
General science and engineering fields: natural science, mathematics and computer science, medicine, engineering, architecture, trade, craft, transport and communication, agriculture, forestry, fishery.
^Natural science, mathematics and computer science, engineering.
Source: Lall (1992). Original data from World Development Report, 1988; UNESCO. Statistical Yearbook 1988. Pans, 1989; and Government of Republic of China. Statistical Yearbook of Republic of
China 1988, Taiwan. Government of Republic of China, Ministry of Education, Educational Statistics of Republic of China, 1984.
and considerably above other comparator countries (Table 13.16). Yet the trainability
of students at vocational and technical schools increasingly depends on the quality of
their basic school education. More generally, the technological content of industrial
and other economic activities requires growing levels of formal education for the labor
force, which the Brazilian education system has failed to provide.
Higher Education
Between 1960 and 1985, the total number of tertiary students went from 93,202 to
1,437,232, a rapid expansion. As a percentage of their age group, students enrolled in
higher education comprised 11%, up from 2% two decades earlier (Table 13.16).
Although that percentage does not compare poorly with other industrializing coun-
tries (with the exception of South Korea), certain features of the higher education sys-
tem tend to undermine its effectiveness as a breeding ground for technical labor and
innovation.
First, the 3-fold expansion of undergraduate enrollment during the 1970s, com-
bined with a decline in full-time faculty, has had an adverse effect on the quality of
university education across the board. Many postgraduate courses have been rede-
signed to remedy the insufficiencies of undergraduate training. The low status and
poor qualification of undergraduate teachers are the core of the problem. About
45,000 full-time teachers, hired initially on a provisional basis, without formal pro-
cedures or evaluation, are now tenured. Most lack academic training and have not
gone beyond a B. A. degree. They serve some 450,000 students in free, public uni-
versities.44 In addition, around 60,000 teachers, also not well qualified and many
with a large teaching load in several institutions, serve 850,000 students in private
schools.
Second, the postgraduate system also faces severe quality problems. Although in
1970 there were 57 doctoral programs in Brazilian universities, in 1985 there were
more than 300, with another 800 providing training at the M. A. level. Combined,
they were graduating 5000 students at both levels each year.45 Yet there is a wide
quality variance among postgraduate programs, and according to CAPES' (Coor-
denacao de Aperfeicoamento do Ensino Superior) evaluations, only about one-
fourth of new graduate programs are academically satisfactory. Most programs also
face uncertain financial support from FINEP and CAPES, and very few have been
able to diversify their funding by linking up with or supporting innovative efforts in
the productive sector.
Third, the proportion of students in science, mathematics, and engineering is rela-
tively small compared to other industrializing countries, being considerably below
fast growing East Asian economies and Mexico (Table 13.16). In engineering alone
the disproportion is even greater: there are more than four times as many engineer-
ing students as a percentage of the population in South Korea, Hong Kong, Taiwan,
and Singapore as there are in Brazil; and there are three times as many in Mexico
(Table 13.17).
The combination of low-quality undergraduate teaching, few strong graduate
programs, a relatively small pool of students in science, mathematics, and engineering,
and underuse of existing educational capabilities outside formal institutions of higher
learning has contributed to Brazil's lagging R&D manpower. Brazil has fewer scientists
Table 13.17. Human Resources Devoted to R&D According to Major Reids,

1986
Total Doctorates
Field No. % No. %
Exact and earth sciences" 6,651 12.6 2,192 19.1
Biological sciences 5,117 9.7 1,900 16.5
Engineering 7,765 14.7 1,180 10.3
Health sciences 6,107 11.6 1,939 16.9
Agrarian sciences 7,607 14.4 1,340 11.7
Applied social sciences 4,543 8.6 842 7.3
Human sciences 4,993 9.4 1,516 13.2
Linguistics, letters and arts 1,471 2.8 583 5.1
_
No information 8,609 16.3
Total 52,863 100.0 11,492 100.0
Încludes physics, mathematics, and other "exact" disciplines.
Source.' Martins and Queiroz (1987).
and engineers engaged in R&D in relation to population (256 per million) than Tai-
wan (1,426 per million), Singapore (960 per million), South Korea (804 per million),
and Argentina (360 per million).
Moreover, the distribution of human resources allocated to R&D is concentrated
in biological and health sciences, as well as applied social sciences, human sciences,
and the arts/literature (Table 13.17). In two key fields, engineering and agrarian sci-
ences, the proportion of researchers is relatively small. Their secondary position in the
distribution of R&D researchers suggests that much R&D undertaken in Brazil is not
closely related to productive activities.
The System at Work

Despite certain fundamental weaknesses in the institutional support network for tech-
nological advancement, there are many instances of successful cooperation with
industry and other segments of the productive sector. In these cases strong and effec-
tive links have been forged, and producers have achieved international levels of com-
petitiveness. We shall provide three examples to illustrate different ways in which this
has occurred.
An impressive case of a top performing public enterprise with strong roots in the
public research infrastructure is that of Embraer. In 1946 the Aeronautics Technology
Institute (ITA) was established to train high-level human technical resources. Drawing
heavily on ITA graduates, the Aeronautics Technology Center (CTA) was created in
1947.46 Among other tasks, CTA was given the mission of designing an airplane suit-
able to Brazilian conditions. The Bandeirante turboprop had its first successful flight
in 1968. The original team was then transferred from CTA to set up Embraer in 1969
and produce the Bandeirante on a commercial scale.47 Five hundred Bandeirantes
were produced and most sold internationally, where it found a niche in the commuter
airplane market. Embraer followed the Bandeirante with several other models, again
focused strongly on the international market.48
Embraer progressed from a heavily protected and subsidized public enterprise to

a dynamic world-class competitor. Embraer has also been proficient at using public
research institutions and universities as well as tapping finance from FINEP. Its strong
focus on the export market has been crucial in offsetting development costs by per-
mitting larger production scales, in bringing new ideas for further technical change,
and in demanding exacting performance standards.
Metal Leve is a case that illustrates the trajectory of a very dynamic domestic pri-
vate firm that has effectively used the public support infrastructure. Metal Leve started
to operate in 1950 producing pistons for local subsidiaries of multinational companies
manufacturing vehicles in Brazil (Ford, Volkswagen, GM, and Mercedes Benz). Its
initial focus was on quality control, combined with well structured technology transfer
programs, involving engineering of piston manufacturing processes. Metal Leve
entered the international market in 1965 as a way to use up its excess capacity, and
induced by fiscal and credit incentives. Export-output ratios rose to 8% in 1970, 17-
20% in the 1980s, and 35% in 1988. The driving force for technical change during the
early period was demanding foreign buyers. Efforts at quality improvement at Metal
Leve were part of a managerial culture that continuously emphasized the importance
of best practice techniques and methods. Effective absorption of imported technology
was the result of prior and detailed studies of the processes that were to be transferred,
intense training in-house and at the premises of the technology supplier, and system-
atic adaptation of the technology to local conditions.
In the late 1970s the firm began developing new products to compete in markets
that required Metal Leve to supply designs.49 In 1979 it set up its own research center
with financial assistance from FINEP (on the order of US$2.2 million), spurred largely
by the challenge posed by increased rivalry in export markets. By 1988 the staff of the
R&D center had grown to 230, and Metal Leve spent 2.7% of sales on R&D. In addi-
tion, it contracted out research to local R&D centers and universities. In 1988 Metal
Leve set up an Advanced Technology Center in Ann Arbor, Michigan, as an extension
of its Sao Paulo R&D center, motivated by the need to be better informed about new
developments and future plans of American buyers.50 In September 1989 Metal Leve
opened a manufacturing facility in South Carolina to produce articulated pistons for
diesel engines (a Metal Leve innovation) to equip Caterpillar's new diesel engines (to
be introduced in 1990) and also cater to demand from Volvo and Cummins.
The case of Metal Leve, like that of Embraer, illustrates the systematic, slep-by-
step technology efforts needed to become an internationally competitive producer. It
also emphasizes the importance of export markets as a source not only of technical
information (supplied by buyers or gleaned from competitors) but of continuous pres-
sure to improve performance. Just as Embraer, Metal Leve also was quite successful
in using Brazil's institutional system supporting innovation. It maintains close inter-
action with FINEP, as well as its collaborative ventures with universities and technical
and research institutions. These arrangements suggest that a motivated management
with long-term commitment to technological excellence and with a record of tech-
nological and commercial accomplishments is able to tap existing technology-oriented
institutional, human, and financial resources effectively.
In agrotechnology, the Government agency EMBRAPA (Empresa Brasieira de
Pesquisa Agropecuaria) has coordinated and partly executed a complex and far rang-
ing "national agricultural research program," resulting, among other things, in the
incorporation of previously infertile land through a better understanding of soil biol-

ogy and the successful introduction of new varieties.
In 1987, agriculture accounted for 10% of Brazil's GDP, 25% of employment, and
43% of total exports. Prior to 1980 agriculture production growth was led by nontra-
ditional exports (mainly soybeans and citrus), based almost entirely from new land
expansion. Since the early part of the 1980s, the expansion of the agricultural frontier
has come to a virtual halt (with the exception of areas in the north). To a large extent,
it has been compensated by the introduction of new varieties, made possible by tech-
nological advances in agricultural research, which has led to significant increases in
yields on traditional export crops (cotton, tobacco, and coffee) and many food crops
(rice, maize, and wheat). In this and other agriculture-related innovations, the role of
EMBRAPA has been critical.
EMBRAPA was created in 1972 as a public sector corporation attached to the
Ministry of Agriculture, with a large degree of financial and administrative autonomy.
It is the coordinating agency for the country's cooperative agricultural research system,
which, besides EMBRAPA, comprises (1) state-level research companies and institu-
tions, (2) universities, (3) the private sector, and (4) some 40 research institutions
throughout the world with which EMBRAPA entertains technical cooperative
arrangements.
EMBRAPA executes its own research program through a national network of
research stations that comprises 6 regional centers, 21 national centers for commodity
research, and 8 state-level research centers. The regional research centers concentrate
and coordinate research on specific problems of the major agroecological regions in
which they are located. At the commodity research centers, multidisciplinary teams
work on priority programs related to agronomic problems and production systems for
each specific commodity. State-level research stations (either managed by EMBRAPA
or by state-owned companies) are responsible for taking the findings and technological
components generated by the regional and national centers and incorporating these
into adaptive research at the local level.
Finally, in support of all these research programs, EMBRAPA operates a number
of special service centers: the Center for Genetic Resources and Biotechnology (Cen-
argen), the Service for Soil Surveys and Soil Conservation (SNLCS), the Basic Seeds
Production Service (SPSB), the National Research Center for Agroindustry and Food
Technology (CTAA), the National Center for Plant Protection and Environmental
Research (CNPDA), the National Center for Irrigation and Drainage Research
(CNPA), and the National Center for Soil Biology Research (CNPBS).
EMBRAPA's activities have been backed by strong and consistent Government
commitment to agriculture research since the early 1970s, in contrast to the more frag-
mented and far less effective approach to industrial research. EMBRAPA's first year
(1974) budget (in 1986 dollars) amounted to US$25 million. Funding increased stead-
ily to about US$300.6 million by 1980 and an estimated US$324 million for 1989.
Most other research institutions have not fared so well, having suffered frequent and
unexpected budget cuts since the early 1980s due to a worsening fiscal crisis.
Although EMBRAPA has been successful in expanding agricultural production
in less fertile areas and in introducing new seed varieties (for cotton, tobacco, coffee,
rice, maize, and wheat), the diffusion of research results has been limited mostly to
producers in better endowed regions. Many of the technologies generated by research
need to be adapted and screened through on-farm tests before they can be dissemi-
nated to resource-poor, medium and small-scale producers, particularly in the less
developed regions of the country. Formal mechanisms for coordination between
research and extension, including consultation with the farming community, were ini-
tiated only in 1986. Still, research centers generally are endowed with only modest
human and physical resources for technology transfer because technology diffusion
has yet to be recognized as an explicit function of the research establishment.
CONCLUSIONS
This chapter's major contention is that a combination of limited technological

involvement by domestic producers, regulatory and policy restrictions on both
embodied and disembodied forms of technology imports, and weak institutional sup-
port to industrial firms has increased Brazilian firms' distance from the price-perfor-
mance frontier. In addition, major gaps in the educational system, particularly low
enrollment levels in secondary school and in science and engineering, compromise the
supply of technical labor force and the acquisition of technological capabilities in the
future.
As in most discussions of this kind, these propositions are subject to a number of
caveats. First, the suggestion that relatively few firms are actively engaged in technol-
ogy development omits the fact that not long ago, say in the early 1970s, far fewer than
the present core of 350 R&D-active firms (with a contour of 1200 producers) were
technologically active. In many cases, improved technological performance came as a
result of competitive pressures from the international market. For national firms, par-
ticularly those in capital goods and electronics, support from government technology
policies and institutions was critical.
This chapter suggests that much of the underlying motivation for the Brazilian
Government's technology policies was not the improvement of the country's technical
base but the more short-sighted consideration of saving or earning foreign exchange.
A second caveat is now in order. Although that general statement is on the whole cor-
rect, the Government's technology policies also improved the bargaining position of
local firms in negotiating arms-length technology transfer deals. In addition, they pro-
moted entry of domestic firms in areas that normally would be precluded through pat-
ent protection or the exercise of overwhelming market power (pharmaceuticals and
electronics), while stimulating the creation of a fairly sophisticated capital goods sec-
tor. Moreover, such policies effectively steered multinational firms that traditionally
had been domestic market-oriented toward exporting, thereby forcing the technolog-
ical upgrading of their domestic operations. Finally, the policies created for the first
time a financing mechanism for the technological needs of Brazilian industrial and
engineering consulting firms.
What is implicitly argued in this chapter, however, is that these government pol-
icies were carried too far and became outdated. Restrictive arms-length technology
transfer policies, for example, assumed the presence of relatively unsophisticated
domestic producers and the considerable eagerness of foreign suppliers to market their
latest technological wares. Nearly two decades after these policies were first imple-
mented, the assumptions are much less justifiable. Similarly, capacity creation in
sophisticated industrial segments, such as capital goods and electronics, was regarded
as double benefits for economic development: it allowed for import substitution in
areas of heavy foreign exchange outflow while creating capabilities in strategic seg-
ments for the formation of skills and diffusion of knowledge. But across-the-board
import substitution led to excessive diversification, fragmentation of efforts within
firms, and substantial waste in government-supported technological pursuits.
A third and final caveat concerns the institutional support system for technical
advance in industry. This chapter argues that despite a number of important excep-
tions (such as EMBRAPA, CTA, and IPT), the public R&D network is not very
responsive to the productive needs of the economy. Moreover, overlapping and frag-
mentary efforts dissipate scarce resources. What was left unsaid, however, is that the
faults of the institutional network are not unique to Brazil nor to industrializing coun-
tries, for that matter. A more demanding industrial sector, challenged by competitive
markets and focused on fewer product areas, would be a far better user of existing insti-
tutional resources (as the experiences of Metal Leve and Embraer show) and an effec-
tive force for reform.
NOTES
1. See Furtado( 1984).

2. See Pastore( 1974).
3. See Pastore (1974) and S. Motoyama (1984).
4. See Fishlow( 1980).
5. The enlarged demand for these services led to the expansion and transformation of the
Laboratory of Material Resistance of the Sao Paulo Polytechnic into the Technological Research
Institute (IPT) in 1934. IPT focused on testing key materials for the construction and mechanical
industries, as well as providing support for the further development of civil and soil engineering
for large industrial projects, including Volta Redonda, the first integrated flat steel plant, estab-
lished in 1941. IPT remains the largest and one of the most respected industrial technology
research institutes in Brazil. Another key industry-oriented research institute, the National Tech-
nology Institute (INT), was established in 1933, as an outgrowth of an experimental research
station on combustibles and materials founded in 1921. The need for greater professional and
technical skills and some research capability also led to the creation of the University of Sao
Paulo (USP) in 1934. One Brazilian science historian has argued that the creation of USP was
Sao Paulo's reaction to its defeat in the Constitutional Revolution of 1932. It sought its redemp-
tion by investing heavily in human resource development and betting on the increasing impor-
tance of science and development, a bet that seems to have paid off very well. See Motoyama
(1984).
6. Sec Pastore (1974).
7. In this early period there was a strong emphasis on the development of human resources,
especially in physics, because of the preoccupation with atomic energy. The focus on higher level
technical human resources was institutionalized through the creation in 1951 of the Coordinat-
ing Agency for Advanced Training of High Level Personnel (CAPES). Its primary objective was
to stimulate and finance the development of higher level technical resources in order to
strengthen S&T capability, particularly in the universities and in R&D institutions. However, as
suggested, compared to other middle-income developing countries, Brazil has underinvested in
general education.
8. The Brazilian push for an independent development of atomic energy led by Admiral
Alvaro Alberto conflicted with U. S. interests at the time, and he was forced to resign in 1955.
The atomic energy program was then redirected from an effort to gain greater autonomy over
that important new technology to that of being a supplier of uranium and other strategic min-
erals. Furthermore, support for nuclear issues was transferred out of the CNPq to a new insti-
tution the National Nuclear Energy Commission (CNEN) established in 1956.
9. See Fishlow( 1980).
10. FUNTEC was a special fund created within the National Bank of Economic Devel-
opment (BNDE) to finance the training of specialized technical personnel for research and
related activities in the universities. In 1964, BNDE also established the Fund for the Acquisition
of Machinery and Industrial Equipment (FINAME), which was to play an important role in the
development of the local capital goods industry. FINER is the Agency for Financing Studies and
Projects. It started out in 1965 as a fund in the Ministry of Planning and Coordination to finance
feasibility studies and project development for investments in sectors and activities that were
considered of priority for the country's economic and social development. In the administrative
reforms that took place in 1967, the Government transformed FINEP into a public enterprise.
11. The plan also sought to prepare Brazil for more extensive use of nuclear energy (up to
10 million kW by 1990) and to continue working on other applications of nuclear energy such
as isotopes in agriculture, medicine, and industry. It also included a massive effort for prospect-
ing for nuclear minerals, and a large program for absorbing technology for uranium enrichment
and for building nuclear reactors with local materials and components.
12. The STI was the promoter of the national alcohol program, which was supposed to
supply 10.7 billion liters of alcohol by 1985 to substitute for 45% of the projected gasoline con-
sumption for that year. This was the most ambitious program developed by the STI. The pro-
gram involved not only the technological problems of blending and using various mixtures of
gasoline and alcohol, but also adapting gasoline and diesel engines to run on pure alcohol, search-
ing for different raw material sources of alcohol, increasing crop yields, and developing new pro-
cesses to produce alcohol.
13. Market share is defined as XJXlv where Xt is the country's exports of good ; and X& is
the world's exports of good /.
14. See Tavares de Araujo (1982, Table III).
15. The normalized market share index is denned as (Xi/X)/(X^/Xv), where X-, is the coun-
try's exports of good i, X is total country exports, Xivl is the world exports of good i, and Xv is
total world exports; hence, the share of a particular good in the country's exports is normalized
by the share of world exports of that good in total world exports. The index is quite sensitive to
the competitive position of individual subsectors in the national economy, and changes in the
index are indicative of shifts with respect to the world prices.
16. The only major exception is "explosive pyrotechnical products," which continued to
increase their normalized shares in the 1980s and showed in 1987 the highest "revealed" com-
parative advantage. This category is believed to contain many arms exports.
17. See Tavares de Araujo et al. (1989), Kume (1988). Kume's 1988 price comparisons
were for 88 subsectors on the basis of a representative sample of 715 products (eight-digit BTN
classification) of the IPEA-FUNCEX data base. International price comparisons tend to present
well-known problems of choice of an appropriate exchange rate, and comparability of products
differentiated by quality and performance. Yet such price data can be suggestive of broad trends
in a country's competitive position in world markets.
18. See Confederacao Nacional da Industria (1990).
19. In 1985 the top 254 exporters of manufactured goods were responsible for 70 to 80%
of total manufactured exports. These producers comprise approximately 6% of all medium and
large firms with published balance sheets and less than 3% of all establishments with over 100
employees. In addition, a number of smaller firms export through trading houses; these firms are
concentrated in shoes, cast iron, steel, and autoparts. Overall, however, exports are undertaken
by that narrow set of large producers in metal-mechanics, chemicals and petrochemicals, and
traditional segments.
20. The number of firms filing income tax returns fell from 570,000in 1977 to just 266,000
in 1985.
21. See Paulinyi( 1990).
22. These are Petrobras (oil and petrochemicals), Telebras (telecommunications), Electro-
bras (electricity), Nuclebras (nuclear energy), Siderbras (steel), CVRD (minerals), Embraer (air-
planes), and Cobra (informatics).
23. The limited participation of R&D personnel in the productive sector is even more
accentuated in terms of researchers with doctoral degrees. The share for the productive sector is
0.46%, most in public enterprise R&D labs.
24. "R&D active" firms are defined as those that are known to have formal R&D labora-
tories or to undertake systematic R&D efforts. They are a subset of about one-third of all firms
that declare R&D expenditure on their income tax returns.
25. See Braga and Matesco (1989). The study reports the answers to a detailed question-
naire on the technological activities of the Brazilian industry from a 1980/1981 sample of 4309
industrial firms operating 7156 plants.
26. Braga and Matesco (1989, pp. 12-14).
27. Braga and Matesco (1989) report that over 67% of respondents pointed out that copy-
ing of product lines from competitors is a common practice in their sector.
28. Substantial changes in the technology transfer and intellectual property rights regime
are being considered, but have not been put into effect yet.
29. Law 4131.
30. Cardozo(1988).
31. For more information on the regulations covering the different types of agreements,
see Rosenn( 1988).
32. In 1983 INPI issued Normative Act 64 whose principal objective is to strengthen
national technological capability. Under this Act, INPI's approval for a technology transfer
agreement is conditioned on whether the receiving firm has an adequate program to assimilate
the technology and carry out R&D to gain greater technological autonomy. The amount of the
investment required depends on the financial situation of the Brazilian contractor and the
amount to be paid for the imported technology.
33. However, it appears that since the "New Industrial Policy" was initiated in 1988, roy-
alties as high as 10% are being permitted for some high-technology sectors such as software. See
Rosenn(1988).
34. Net sales is calculated as total sales minus duties, taxes, imported inputs and compo-
nents, commissions, transport, insurance, and other deductions.
35. Data from income declarations corroborate this downward trend in foreign technology
contracts. The number of firms paying foreigners for royalties, and technical assistance has fallen
from between 500 and 600 in the mid-1970s to just over 200 in the mid-1980s (see Paulinyi,
1990).
36. See Rosenn (1988).
37. Profit remittances above 12% of registered capital are subjected to a 40-60% supple-
mental income tax.
38. This causes tax problems for U. S. companies because Sections 367 and 382 of the
U. S. Internal Revenue Code force companies to include imputed royalties on technology trans-
fers in their U. S. income for tax purposes. See Rosenn (1988).
39. The main program was BEFIEX, established in 1972. This program, initially devel-
oped for the auto industry, gave firms incentives in the forms of reduction of imports duties and
other fiscal exemptions as well as financial incentives in return for specific export commitments.
The program was expanded to other sectors and also to national firms. In 1989 it was estimated
that 50% of all Brazilian manufactured exports were covered under this program.
40. Take the example of printers. In the case of dot matrix, international-domestic price
differentials were in the range of 3:1 at the early stage of production in Brazil (and have remained
so since), whereas for laser printers this lag has grown to 5:1.
41. Note in this respect that the value of contracted technology development operations in
1989 was less than US$ 18 million, little over one-tenth of what was contracted 2 years earlier in
1987.
42. Braga and Matesco (1989, Table 2.1).
43. CPqD, however, has been fairly successful in transferring many of the products it has
developed to industrial firms for large-scale production. By the end of 1987,75 different products
developed singly by CPqD or in association with universities (which were in charge of carrying
out most applied research) and industrial firms (generally focused on the later stages of devel-
opment, such as prototyping) were being manufactured by 25 producers. See Frischtak (1989).
44. See Schwartzman (1988, pp. 99-119).
45. Schwartzman (1988, p. 104).
46. By 1988 ITA had trained more than 3000 engineers, 800 of which were in the aero-
nautics field.
47. The ITA/CTA graduate engineering, education/research nexus led to the spontaneous
development of Brazil's first and largest industrial high technology park around San Jose dos
Campos. Along with Embraer it has led to the establishment of many high-technology compa-
nies including Avibras (missiles), Orbita (missiles), Engesa (military equipment), Tecnasa (elec-
tronic communication equipment), Composite Tecnologica (composite materials), and Quan-
tum (software).
48. As of February 1989 it had an accumulated production of 3983 aircraft.
49. Between 1970 and 1980, 31 new types of pistons and 59 types of sleeve bearings were
introduced, including "heavy duty" pistons for large diesel engines and a special line for alcohol
vehicles.
50. The reported cost of the new center was US$3 million, 90% of which was financed by
F1NEP. See Gazeta Mercanlil, February 1, 1988.
REFERENCES '-^ The Wor'd Ban k, Industry and Energy

Department.
Furtado, C. (1984). The Economic Growth of Brazil:
Baer, W., da Fonseca, M., and Guilhoto, J. (1987). A Survey from Colonial to Modern Times. West-
"Structural Changes in Brazil's Industrial Econ- port, CT: Greenwood Press.
omy: 1960-1980." World Development 15(2): Government of Republic of China, Ministry of Edu-
277-86. cation. (1984). Educational Statistics of Republic
Braga, H., and Matesco, V. (1989). "Desempenho of China.
Tecnologico da Industrial Brasileira: uma Analise Government of Republic of China. (1988). Statisti-
Exploratoria." Textos Para Discussao Interna cat Yearbook of Republic of China 1988. Taiwan.
No. 162, INPES/IPEA, February. Kume, H. (1988). "A Politica Tarifaria Brasileira no
Cardozo, A. C. (1988). "A Implantacao de Leis e Periodo 1980-85: Avaliacao e Reforma." Mimeo.
Regulamentos Sobre Transferencia de Tecnolo- Lall, S. (1992). "Explaining industrial Success in
gia: A Experiencia do Brasil." Mimeo. Developing Countries." In S. Lall and V. N. Bal-
CNPq. (1989). "Recursos do Tesouro da Uniao: asubramanyam (eds.), Current Issues: in Develop-
Dotacao Inicial, Dotacao Final, e Despesa Reali- ment Economics. London: Macmillan (in press).
zada 1980-89." May. Martins, G., and Queiroz, R. (1987). "O Pertil do
Confederacao Nacional da Industria. (1990). Com- Pesquisador Brasileiro." Revista Brasileira de
petitividade e Estrategia Industrial: a Visao de Tecnologia 18(6).
Lideres Industrials Brasileiros. Rio. CNI. Motoyama, S. (1984). "Ciencia e Tecnologia e a His-
Fishlow, A. (1980). "Brazilian Development in toria da Dependencia do Brasil." Revista Brasi-
Long-Term Perspective." American Economic leira de Tecnologia 15(3).
Review May, 102-8. Pastore, J. (1974). "Science and Technology in Bra-
Frischtak, C. (1989). "Specialization, Technical zilian Development." In W. Beranek, Jr. and G.
Change and Competitiveness in the Brazilian Ranis (eds.), Science, Technology, and Economic
Electronics Industry." Industry Series Paper, No. Development. New York: Praeger.
Paulinyi, E. (1984). "Empresa NacionalPan- Machado, J. B. (1989). "Protecao, Competitivi-

orama do Setor Empresarial em 1983." Revista dade e Desempenho Exportador da Economia
Bmsileira de Tecnologia 15(3). Brasileira nos Anos 80." Revista Brasileira de
Paulinyi, E. (1990). "Padroes de Capacitacao e Comercio Exterior. 26.
Atualizacao Tecnologica da Empresa Brasileira." Tavares de Araujo, J. (1982). "Mudanca Tecnolo-
Informe Estatistico do CNPq. 1(4). gica e Competitividade das Exportacoes Brasi-
Rosenn, K. (1988). "Regulation of Foreign Invest- leiras de Manufaturados," IEI/FEA/UFRJ, Texlo
ment in Brazil." Mimeo. para Discussao n. 8. Table III
Schwartzman, S. (1988). "Brazil: Opportunity and UNESCO. (1989). Statistical Yearbook 1988, Paris.
Crisis in Higher Education." Higher Education U. N.
17:99-119. World Development Report. (1988, 1990).
Tavares de Araujo, J., Haguenauer, L., and
14
National Systems of Innovation Supporting
Technical Advance in Industry: The Case of
Argentina
JORGE M. KATZ
NESTOR A. BERCOVICH
As a result of its successful integration into world markets for foodstuffs and agricul-
tural raw materials Argentina was, at the beginning of the present century, a relatively
rich nation with a fast growing economy. From 1895 to 1914 the number of acres
under cultivation increased from 5 to 25 million. Wheat and maize exports multiplied
by 3- and 4-fold, respectively, and the local population expanded from 3.9 to 7.8 mil-
lion, due to heavy immigration from Europe. In 1914 one-third of the population had
been born abroad, with nearly one million Italians and 800,000 Spaniards having
taken up local residence; 90% of exports were primary products, most of which were
directed toward western European countries (Diaz Alejandro, 1970; Rock, 1985).
Welfare indicators such as income and calories per capita clearly place Argentina
among the richest nations in the world in those early years of the century.1
Contrary to the above picture, and in spite of a major process of industrialization,
structural change, and social mobility, Argentina exhibits today a high level of mac-
roeconomic instability, long-term decay in aggregate economic performance, and
extreme forms of institutional fragility. Output per head is currently some 20% lower
than two decades ago. The recent transition from a constitutional government to its
freely elected civilian successor was the first to take place in many decades, pointing
to the strong military presence in the country's political life, as well as to the weakness
of the Argentine civilian institutions.
Thirty years of technological stagnation in its primary sector, a concentrated and
oligopolistic industry that developed in an overprotected environment that relied basi-
cally on direct foreign investment as a source of technological dynamism, an entre-
preneurial community that has increasingly involved itself in bureaucratic lobbying,
opportunistic behavior, and rent-seeking activities rather than in technological and
innovative efforts, a scientific and technological sector that has lacked in sense of pur-
pose and direction as well as in incentives and accountability, and bureaucratic gov-
ernments that have recurrently failed to incorporate the idea that a coherent network
of institutions and policies supporting the process of technological change and mod-
451
ernization is needed if any given country is to adequately exploit its long-term oppor-
tunities constitute the main explanations of why the industrialization process has grad-
ually lost strength and dynamism, and today faces major difficulties for its long-term
consolidation.
Many of these difficulties relate to the weaknesses of Argentina's institutions and
policies supporting technological change and innovation in industry and agriculture.
In this chapter we shall examine the structure and performance of the country's
National System of Innovation, its changes through time, and its relationship to the
nation's long-term industrial and agricultural performance.
The first section of the chapter deals with the national system of innovation's sup-
port for technical change in industry. We begin by identifying various subperiods in
the country's long-term industrialization process and examining the sources and
nature of technological change and innovation as well as the functioning of the
national system of innovation in each one of these. In the last of these periodsfrom
the mid-1970s to the presentthe country's industrial sector has been undergoing a
major restructuring process, which has deeply affected the organization of work at the
individual plant level as well as the functioning of markets and regulatory institutions.
This process is also having a major impact on the national system of innovation sup-
porting technological change and innovation in industry, the impact of which we will
examine. The next section examines the network of agents, institutions, and policies
influencing technological change and innovation in agriculture.
Having looked at agents and institutions related to innovation and technical
change in agriculture and industry, we then proceed with the examination of R&D and
knowledge-generating activities performed by public sector research agencies, univer-
sities, decentralized public enterprises, and the Armed Forces. We pay attention to the
various ideologies, political and military lobbies, that in one way or another influenced
the somewhat chaotic evolution of this major part of Argentina's national system of
innovation.
The chapter concludes with an overall evaluation of the organization and perfor-
mance of this system. It is quite clear that we cannot truly argue that the country has
been able to develop a coherent and effective set of policies and institutions capable of
inducing a successful process of generation, transfer, and diffusion of scientific and
technological knowledge throughout the production structure. Nor can we say that
Argentina has had anything close to a technological strategy reflecting its long-term
development needs or its opportunities in the international marketplace.
Much could surely be achieved in this field if a major institutional build-up effort
could be undertaken in the years to come. A less regulated macro policy environment,
more respectful of the country's resource endowments and of its opportunities in
world markets, a major reorganization of the educational system with the introduction
of new mechanisms for the upgrading and recycling of the labor force and for the
development of scientific and technological manpower; new ways of interaction and
collaboration between manufacturing firms, universities, public R&D laboratories,
and regulatory agencies; and better choice of priorities as far as public R&D and engi-
neering efforts are concerned seem to all be different pieces of a complex puzzle to
which Argentina will no doubt have to pay closer attention in the years to come, if it
wants to regain the capacity for growth and the patterns of social equity that are cur-
rently being lost at a rapid pace.
TECHNICAL ADVANCE IN INDUSTRY: ARGENTINA 453
In many of these fields we can identify powerful reasons for which decentralized
market decisions might not necessarily be enough to ensure socially acceptable out-
comes. Different types of externalities and public goods are involved, suggesting that
new forms of intervention and institutional build-up efforts will probably have to be
explored in the future even in spite of the fact that public intervention has had a clear
record of failure as far as Argentina is concerned. How to work out an adequate bal-
ance between regulatory mechanisms and decentralized market decisions in the field
of knowledge generation and diffusion or in relation to the training, upgrading, and
recycling of human resources is of course a difficult question for which the country
will have to work out imaginative new answers in the future.
THE NATIONAL SYSTEM OF INNOVATION SUPPORTING TECHNOLOGICAL

CHANGE AND INNOVATION IN INDUSTRY
Argentina's industrial sector is today significantly different from what it was, say, two
decades ago. The organization of production at the individual firm level, the structure
of industry, and the behavior of markets have all been experiencing drastic changes
during the course of the last decade. The national system of innovation supporting
technical advance in industry has also been changing, as we shall see throughout the
present section.
The Argentine industrialization process began in the early years of the present
century. It acquired momentum in the 1930s and during World War II, when lack of
foreign supplies induced local entrepreneurs into import substitution industrialization
efforts.
Since the early 1960s, the massive arrival of foreign firms has significantly affected
the rate and nature of the import substitution industrialization process, as well as tech-
nological change and innovation in industry. In the late 1970s the "foreign-led"
expansionary phase came to a halt and since then a major industrial restructuring pro-
cessreflected in the contraction of metalworking industries and the concomitant
expansion of raw material processing industrieshas been taking place.
We shall now examine this industrialization process in some detail.
The Pre-1950s
The Argentine industrialization process of Argentina began with the establishment of

food-processing industries linked to its primary production. Meatpacking houses, tan-
neries, flour mills, wineries and so on came into being and fared rather well before
1930.
A completely open economy and strict respect for Gold Standard principles, on
the domestic side, along with a rapidly expanding foreign demand for foodstufs and
raw materials allowed Argentine industries to prosper in world markets.
Capital goods for the locally established plants were at that point entirely
imported. Their repair and maintenance, however, induced an early expansion of
domestic technological capabilities in areas such as stamping, forging, and machining.
Such technical skills were to become important some years laterin the 1940swhen
the inconvertibility of the British pound, on the one hand, and the outbreak of World
War II, on the other, induced the military government in power into a "domestically
oriented" import substitution industrialization strategy that was to have long lasting
consequences on the country's social and economic performance.
Central to the new macroeconomic policy environment, we find two new policies
that strongly affected the process of industrialization. We refer to the introduction of
exchange controlwhich began to be applied late in 1931 and to import tariffs that
soon followed. From 1930 onward, the economy became much more regulated and
closed itself off to foreign trade, with imports and exports falling in their percentage of
the GDP. In particular, tariff protection was to play a major role as a determinant of
the structure and nature of the industrialization process that followed. The average
firm size, market structure and behavior, and degree of vertical integration were all
significantly affected by the new industrial policy that was adopted by the country.
In the late 1930s and 1940s, textiles and chemicals and later a whole range of
electromechanical products began to be domestically produced under the incentives
of a rapidly expanding internal market, tariff protection, and subsidized financing
from public agencies, in particular from the Industrial Bank, which was created for
such purposes in 1944. From 1945 onward, the nationalization of foreign trade and
the expropriation of agricultural profits allowed the government to transfer resources
from rural landlords to the rapidly growing urban-industrial sector that developed in
those days under the incentives of an expanding real wage rate and of an industrial
policy that promoted various so-called "industries of national interest." The role of
Argentina's highly nationalistic military forces became central to the country's mac-
roeconomic policy formulation mechanism during those years. A fortiori it also
exerted a major impact on the way the national system of innovation behaved
throughout the war period and during the 1950s.2
Whereas the pre-1930s industrialization processinvolving mostly the local pro-
cessing of primary raw materialswas carried out in modern and updated facilities
and was highly competitive by international standards, the import substitution indus-
trialization efforts of the 1940s and afterwards entailed the erection of a "second best"
industrial sector exclusively catering to local demands. Foreign exchange shortages
originated in the stagnation of primary exportsintroduced a major barrier to the
expansion of manufacturing production.3
On the external front the situation was by no means better. The end of the war
and the rapid technological reconstruction of DCs during the 1950s brought about an
increasing technological gap between local manufacturing industries and their inter-
national counterparts.
Even before the military coup d'etat of 1955 that ousted the Peronist government
in power, the local administration had begun to revise its antagonistic attitude toward
foreign capital and had accepted the idea that foreign investment and technology could
become major "sources of economic growth." In April 1953 the Administration
passed Argentina's first foreign investment law and started negotiations with a number
of MNCs, trying to attract them to Argentina. It is important to note the weak bar-
gaining position from where Argentina carried out these negotiations if we are to
understand the extremely high level of protectionalong with other advantagesthat
many of these firms managed to obtain from the local authorities in exchange for their
promise of setting up domestic production facilities.
Be that as it may, the massive arrival of foreign firms in the late 1950s and early
1960s clearly signals the beginning of a somewhat different period in the country's
industrialization history as well as in the functioning of the agencies and institutions
supporting technical change and innovation. This period covers the years between
1960 and 1975, which we now examine.
1960-1975: The Establishment of the Automobile Industry and the Expansion of

Metalworking Industries in General
The investment boom that took place between 1957 and 1961 clearly constitutes the
starting point of a period of rapid expansion, both in manufacturing output and in the
productivity of labor. Close to 200 foreign corporations established domestic produc-
tion facilities during those years, bringing about a major change in the country's indus-
trial culture.
The share of MNCs in manufacturing production grew quite significantly in those
years until it reached about one-third of the total. Central to this overall episode is the
establishment of the automobile industry that represented only 2.5% of the manufac-
turing GDPin 1951 and grew to 10.3% of the manufacturing GDP in 1965, exhibiting
a growth rate of nearly 25% per year between 1958 and 1965.
The industrialization process we are describing had a number of highly idiosyn-
cratic features in relation to both the size and nature of firms as well as in terms of
market structure and performance. On the one hand, plant sizes were never bigger
than 10% of comparable plants in European countries. On the other hand, and due to
the low degree of standardization and normalization prevailing in local markets, as
well as to the absence of reliable subcontractors, manufacturing plants started their
local operation with a degree of vertical integration much higher than companies in
Europe.
Given the small size of the domestic marketand the fact that very few of these
firms originally expected to engage themselves in export activitiestheir output mix
rapidly became wider than that of comparable plants elsewhere. Thus, the organiza-
tion of work and the social division of labor did not develop as a copy of those pre-
vailing in more mature societies. As a consequence of the above, many of the newly
erected firms found themselves needing to gradually develop "in house" technological
capabilities in areas such as product design, production engineering, industrial plan-
ning, and organization where "from the shelf technology was not particularly well
suited, given the highly idiosyncratic nature of the local production structure.
In other words, given the highly "localized" nature of their domestic operation,
many of these newly created companies found themselves needing to establish "in
house" engineering departments whose basic mission was to adapt foreign product
designs and production processes to the local working and regulatory environment.4
This process of industrial growth had a major impact on the functioning of the
national system of innovation supporting technological change on industry, as we
shall now show.
First, manufacturing output and labor productivity grew quite rapidly through-
out the industrial structure. Between 1960 and 1974, Argentina's industrial output
grew at about 8% per annum while employment and labor productivity grew at 2 and
6%, respectively.
Second, in spite of the domestic orientation of the industrialization process, man-

ufacturing exports gradually began to expand in the late 1960s and early 1970s.
Third, foreign firms clearly acquired a leading role within the local manufacturing
sector and had a major influence on the way the national system of innovation was to
be developed. There are various reasons for this. On the one hand, these firms intro-
duced quality control practices, tolerance limits, and subcontracting practices, which
simply had not been available before within the local industrial environment. On the
other hand, their new production facilities served as training grounds for local tech-
nicians and engineers. In spite of the obvious beneficial impact obtained by this pro-
cess it should also be taken into account that it clearly restricted the spectrum of the
technological "search" activities carried out by engineers and technologists, limiting
them to "adaptive" technological efforts within the boundaries of the technology orig-
inally imported from abroad.
Three different studies document our previous points.
The first of these studies (Katz, 1974) constitutes an attempt to measure the flow
of R&D and engineering expenditure carried out by domestic manufacturing firms
throughout the 1960s, and their impact on individual firm productivity growth.
Taking the largest 200 industrial firms in Argentina, a survey was conducted that
showed that some 20 million U. S. dollars per year were spent by the firms under exam-
ination in "adaptive" R&D and engineering activities as well as in production plan-
ning and organization. Many of these activities were nonroutine in the local context
even though some of them would not have been covered by OECD or NSF standard
definitions of R&D expenditure.
About half of the firms in the sample employed somewhere between 5 and 50
people in product design activities and/or in process engineering departments, whose
function was "debugging" and adapting product designs and production processes
brought from abroad to the local environment, under some form of licensing contract.
Even in those cases in which no formal R&D departments could be identifiedsuch
as in the case of many medium-size family enterprisesknowledge creation activities
were carried out by ad hoc groups or even by production personnel on a part-time
basis. Pilot plant work and the construction of prototypes were quite frequently under-
taken by the companies examined during our field work.
For each one of the companies in the sample, a total factor productivity growth
rate was estimated covering the 1960s. Also output growth, employment, R&D, and
engineering expenditure and royalties were measured for each firm over the same
period and their relationship to labor productivity growth was examined using stan-
dard econometric techniques (Katz, 1974).
Our results show that labor productivity growth was "explained" both by a
"scale" factormeasured by the expansion of outputand by a "technological" fac-
tor that in the present case was represented by the accumulated R&D and engineering
expenditure carried out by each company throughout the 1960s. As mentioned before,
this decade was, for many of these firms, one of rapid physical expansion in the domes-
tic environment and of technological adaptation to its highly particular atmosphere.
Both "scale" and "technological learning" effects can be reasonably expected to
underlie their observed performance. Concerning the impact the industrialization pro-
cess had on the country's pattern of exports, two additional studies carried out a few
years later (Ablin and Katz, 1985) provide evidence that industrial exports of an
increasing degree of technological sophistication gradually expanded in the early

1970s. Manufacturing exports increased from some 100 million U S$ in 1959less
than 10% of total exportsto nearly 1500 million U S$ in 1974close to 25% of a
much higher figure of total exports. Among the newly exported products we find
machine tools, agricultural equipment, electromechanical instruments, industrial
boilers, and automobiles. Examining the experience of the largest 20 industrial export-
ers, we found that product design and process engineering efforts were by no means
negligible as explanatory variables of these companies' gradual success in foreign mar-
kets.
On the other hand, we notice that during the 1970s the industrialization process
resulted in not just manufacturing exports of an increasing degree of technological
sophistication, but also an incipient export trade of pure technologyboth under the
form of licensing activities and through the delivery of complete manufacturing plants
to enterprises of neighboring countries. Meatpacking houses, industrial dairies or bak-
eries, bottling and canning facilities, and machining plants appear on the list. Con-
comitant with such developments we also found evidence indicating that large domes-
tic engineering firms engaged in the construction of hospitals, airports, pipelines, and
so on to be delivered on a turnkey basis to other Latin American countries (Ablin et
al., 1985).
Beyond doubt, export subsidies played a significant role in inducing such devel-
opments, but we cannot rule out the fact that technological "learning" and domestic
engineering efforts gradually acquired significance as explanatory forces of the newly
emerging trends.
Similar results were reported by other researchers looking at other NICs such as
India (Lall, 1982), Brazil, Mexico, Hong Kong, and Taiwan (World Development,
1984), confirming the fact that technological "learning" from peripheral societies
could be expected to have dynamic consequences that had not been previously exam-
ined in the development literature.
Direct foreign investment activities from local entrepreneurs also increased
throughout those years (Katz and Kosacoff, 1983) suggesting that revenues could be
captured in third markets through the export of goods and technology and also by
means of "joint venture" agreements involving the direct investment of local entre-
preneurs in neighboring societies.
The development process we have so far been describing reached a plateau in the
mid-1970s.
Both domestic and external reasons account for the limitations encountered by
local industries at that point. On the domestic side it is clear that many consumer dura-
ble markets became highly "saturated." Consider the case of automobiles. Whereas in
the early 1950s Argentina had one automobile per 50 people, by 1975 the figure had
fallen to around 6 people per vehicle. A similar case can be made for many other dur-
ables. The import substitution industrialization strategy became increasingly criti-
cized for its overprotected nature as well as for its lack of a stronger export drive.
Exports continued to be a small fraction of the total industrial output and the country's
balance of payments fell into a recurrent 3 to 4 years cycle. A generalized perception
developed concerning the fact that the import substitution industrialization process
was leading to a dead end and that structural reforms were badly needed.
On the external side, the rapid diffusion of microprocessors and microelectronic
technology quickly eroded the competitive advantage many local firms had managed
to attain in Latin American markets for machine tools, electrical instruments, and
capital goods in general. Local entrepreneurs were not able to follow the pace of the
international technological frontier, which was introducing digital and numerical con-
trol devices and miniaturization, and rapidly lost external markets with their relatively
less sophisticated electromechanical designs.
In addition to these economic circumstances, it should also be taken into consid-
eration that the domestic institutional atmosphere became rather tense in the late
1960s and early 1970s with militant trade unions and student groups increasingly chal-
lenging the political status quo. The degree of social unrest increased considerably
throughout the country in those years opening up major questions concerning the
capabilities of the country's institutional fabric to survive under such pressure. Argen-
tina was at that point entering into a turbulent period from which it has not as yet
completely recovered. A major process of structural change began in the early 1970s
with a civilian government in power and continuedin a more dramatic and repres-
sive wayafter the military coup d'etat of March 1976. We turn now to the exami-
nation of this period.
1975 to the Present: Deregulation and Opening up of the Local Economy: From
Mechanical Engineering to Raw Material Processing Industries
The perception that the economy was heading for a dead end and running into
stronger balance of payment difficulties induced the Peronist government in office in
the early 1970s to press import substitution industrialization efforts even further. With
this purpose, it introduced a new set of fiscal incentives hoping to attract investment
for industries such as petrochemical, aluminum, and pulp and paper, which were
heavily dependent on foreign supplies. In the space of approximately one decade, a
significant number of new, capital-intensive factories came into being with as much
as 60-70% of the required capital investment coming from public sources under the
form of tax concessions (Azpiazu, 1985). A small group of large domestic holdings
benefited from such policy actions.
Shortly after, in early 1976, a new military takeover occurred, and in the context
of a massive process of social repression, the authorities attempted to deregulate and
open up the local economy to foreign competition. On the sociopolitical front the
Armed Forces tried to curb worker activism through the physical intervention of trade
unions and the imprisonment of labor leaders; on the economic front their disciplin-
ary action came under the form of a drastic return to orthodox market principles.
Among their initial economic policy actions we find (1) a major reduction in tariff
protection: the average ad valorem rate was brought down from 95 to 55%, (2) a reval-
uation of the local currency of approximately 40% in real termsintended as an
instrument for inducing foreign competition in local markets and price restraints on
the part of domestic entrepreneurs, and (3) the deregulation of the financial sector of
the economy. On account of this measure the rate of interestnegative during pre-
vious decadesmoved into positive levels.
These policiestogether with the fiscal incentives for raw material processing
industries approved in the early 1970s and not removed by the military authorities,
probably because of the strong lobbying pressure exerted by large domestic corporate
groupshad a major impact on the rate and nature of the industrialization process as
well as on the behavior of the national system of innovation supporting technical
advance in industry. Let us consider both these aspects. Concerning the process of
industrialization we notice the following.
First, industrial production as a percentage of the GDP has fallen significantly
during the course of the past 15 years. Industrial output has fallen from about one-
third of the GDP in the mid-1970s to approximately 23% at present.
Second, the structure of industry has changed drastically in recent years. The pro-
duction of metalworking products and capital goods in general has contracted, while
resource-based industries producing steel, petrochemicals, aluminum, edible oil, and
so on have expanded considerably. In the mid-1970s Argentina produced some
350,000 automobiles, 25,000 machine tools, and 60,000 tractors per annum. At pres-
ent, it produces only 150,000, 6000, and 5000, respectively, in factories that have
changed significantly as far as production organization, import content, subcontract-
ing practices, and employment absorption, are concerned. Contrary to the above, pro-
duction has increased from 865 to 1800 thousand tons worth of petrochemical prod-
ucts and from 2250 to 3670 thousand tons of steel products during the course of the
same period. As far as edible oil is concerned, production rose from 630 thousand tons
in 1975-1977 to nearly 2 million tons in 1984-1986.
Third, the share of MNCs in domestic manufacturing production has fallen from
32 to 25%. The local production facilities of major MNCs such as General Motors,
Olivetti, and Citroen have been closed down in recent years.
Fourth, with takeovers, mergers, and exits, the degree of business concentration
has increased quite dramatically in the automotive, steel, petrochemical, textile
apparel and shoe production, and tobacco sectors.
Fifth, the consolidation of a small number of large domestic holdings can now be
observed. These groups have expanded, taking advantage of the industrial promotion
legislation of the early 1970s (Schvarzer, 1983). They have gradually diversified their
activities from textiles to oil exploration and extraction and from automotive to
atomic energy component production in a rapid process of horizontal and vertical
expansion.
Sixth, together with the structural transformation, the pattern of industrial
exports has experienced substantial changes. In the mid-1970s about one-third of the
country's manufacturing exportssold basically in the Latin American regionwere
electromechanical products such as vehicles, machine tools, and agricultural equip-
ment. These products account for less than 10% of total industrial exports today.
Instead, we now register large and growing exports of steel plate, edible oil, aluminum,
and pulp and paper, all being industrial commodities where local firms behave as
"price takers" in strongly competitive international markets. Companies frequently
sell abroad at marginal cost, with domestic prices for the same products being signifi-
cantly higher than their international quotations.
Seventh, the organization of work and labor relations at the individual firm level
as well as the overall functioning of the labor market have also undergone major
changes. On the one hand, the contraction of the metalworking sector has brought
about a sharp reduction in Argentina's industrial labor force. The auto industry
employs today some 70,000 workers, less than it did one decade ago, and a similar
situation can be observed in the production of tractors, agricultural equipment, and
capital goods in general. On the other hand, the production of industrial "commodi-
ties" in capital-intensive plants has scarcely required new workers. Moreover, the mas-
sive military repression among trade union members carried out between 1976 and
1979 significantly reduced worker activism and trade union bargaining strength within
the society at large.
These aspects add up to a major "change of regime" in which the structure of
industry, the organization of production at the individual firm level, the market struc-
ture, the degree of business concentration, the participation of foreign companies, the
functioning of the labor market, trade union bargainig strength, and regulatory insti-
tutions all seem to be part of a global economic and institutional restructuring process
of far-reaching dimensions. Firms, markets, and institutions all seem to be involved
in this process. When compared with the old "mechanical engineering" period of the
1960s and early 1970s, it is clear that the import substitution efforts of the 1980s
involve technologically more updated production facilities, a much higher degree of
integration into world markets, and a different entrepreneurial constituency, insofar
as large domestic corporate groupsand not foreign firmsnow dominate the local
industrial landscape.
A major question now emerges in relation to the impact of the previously
described process on the country's national system of innovation (i.e., on agents, insti-
tutions, and policies supporting technical change and innovation in industry). A num-
ber of comments follow.
Consider first the impact that macroeconomic stabilization policies and the open-
ing up of the domestic economy are having on the rate of new capital formation, entre-
preneurial spirit, and locally performed R&D activities. Both the rate of savings and
investment have fallen quite sharply relative to the 1960s and 1970s. A much higher
degree of macroeconomic uncertainty and volatibility, together with a completely
open financial sector, account for the fact that capital flights from Argentina increased
quite dramatically during the course of the 1980s. The reduced tendency to invest
locally (Escude and Guerberoff, 1990) is reflected in the low level of imports of
machinery and equipment that has prevailed throughout recent years (Katz and Kosa-
coff, 1989).
Direct foreign manufacturing investment has also fallen sharply throughout the
1980s reflecting the scarce interest and low expectations held by foreign firms con-
cerning the future of the Argentine economy.
This behavioral pattern (i.e., the reduction in the rate of new capital formation
and in the average propensity to invest on the part of the entreprenurial community)
appears to be associated with an expansion of opportunistic and rent-seeking activities
and with a general fall in spirit that permeates the production structure.
In addition to the above, yet another significant new structural element concerns
the impact that the industrial restructuring process is having on the rate and nature of
locally performed R&D and engineering efforts. Metalworking firms have reacted to
the reduction in their level of activity by cutting down "in house" product design and
production organization efforts. Raw material processing industries, on the other
hand, have not yet significantly expanded their R&D commitments. Thus the fact that
Argentina now produces about half the number of cars, machine tools, and tractors
than it did 10 years ago and that capital-intensive factories producing industrial com-
modities have taken over their role as leaders within manufacturing production seems
to have produced an absolute contraction in industrial R&D and engineering efforts.
The reasons for this are as follows. From an accounting point of view, activities
such as product design engineering efforts, production planning and organization,
plant layout balancing actions, and time and motion studies constitute an indirect cost
of production. Therefore with the contraction of sales, many metalworking firms pro-
ceeded to reduce their engineering departments, particularly those concerned with
product design activities.
On the other hand, the production of industrial commodities in new and tech-
nologically more updated facilitiesmany of which were purchased abroad on a turn-
key basishas so far required very little in the form of R&D and/or process engineer-
ing efforts on the part of local companies.
In addition to the previous explanation, we should also notice that the more tur-
bulent and uncertain macroeconomic scenario of the last decade has probably mili-
tated against long-term technological and innovative commitments on the part of the
local entrepreneurial community. Firms have turned more to the search for subsidies
and special privileges than to investment and technological change. Moreover, the
exchange rate overvaluation that existed during the period between 1979 and 1981
induced many firms to incorporate imported machinery and equipment with the
expectation that the recession that was going on in the country at that time was to be
a short one and that demand would soon return to normal levels. These expectations
never materialized and idle production capacity was apparent throughout the indus-
trial spectrum during the mid- and late 1980s.
So much for capital formation, entrepreneuric spirit, and domestic R&D activi-
ties. Furthermore, it is also evident that a new set of regulatory institutions is gradually
gaining ground within the local environment and that its impact on the national sys-
tem of innovation is going to be far from negligible. The deregulation of markets and
the opening up of the economy to foreign competition, the privatization of public
enterprises, and the capitalization of the country's external debt in exchange for
domestic public utilities and industrial assets have now become central parts of the
newand implicitindustrial policy. It is still premature to evaluate the impact that
this new regulatory regime is going to have on local industry, and particularly on the
spirit and functioning of the national system of innovation supporting technical
advance in industry, but there is no doubt that the impact will be a major one in the
years to come. Recent investments in the petrochemical, steel, and automotive indus-
tries carried out with the purpose of establishing a strong export platform in Argentina
suggest that large companies are gradually beginning to adapt themselves to the new
regulatory environment and that a more open and competitive industrial structure,
better integrated to world markets, might eventually emerge out of the ongoing struc-
tural transformation. It is still too early to determine whether the country is taking a
good or a bad route in its industrialization process, but it is clear that the old route on
the "mechanical engineering" front was running into a dead end and that fundamental
changes in industrial policy were badly needed.
Having so far examined the functioning of the national system of innovation sup-
porting technical advance in industry, we now proceed with the study of the way in
which this system has worked in the case of the agricultural sector.
THE NATIONAL SYSTEM OF INNOVATION SUPPORTING TECHNOLOGICALGical
CHANGE IN AGRICULTURE
Argentina's agricultural sector presents us with one of the more intriguing puzzles in
the country's long-term development history; Figure 14.1 presents the case.
After nearly four decades of rapid expansionwhich brought the area under cul-
tivation to nearly 20 million acres from 1938 to 1940, and the total grain output to 20
million tonsthe sector entered into a long period of decay and stagnation that lasted
nearly 30 years. In the early 1950s the area under cultivation had fallen to less than 15
million acres and total agricultural output was roughly equivalent to the 1919/1920
level. It was only in the late 1960s that the agricultural sector of Argentina returned to
its 1939/1940 output levels, with 20 million acres under cultivation and production
at about 23 million tons.
The process of expansion then proceeded quickly, with productivity per acre dou-
bling between the late 1960s and the present. This improvement in productivity
resulted initially from a more rapid rate of introduction of agricultural equipment and
production organization technologies, and subsequently from the massive diffusion of
maize hybrids, the use of new agrochemicals, and the introduction of new crops such
as soybeans and sorghum. Public sector agencies (see the discussion of this topic in our
next section) have played an important role with regard to the development of new
product varieties, mainly in wheat and soybeans, and in the diffusion of production
Figure 14.1. Total production of grain for the country: 1911-1964.

organization technologies. On the other hand, private firms, in particular, local sub-
sidiaries of large MNCs, are now playing an increasingly important role in the pro-
duction and commercialization of agricultural machinery, agrochemicals, and
hybrids.
How do we explain the poor performance of Argentine agriculture between the
1930s and 1960s? Which patterns of interaction prevailed among agents, institutions,
and policies that locked Argentina into a vicious circle of agricultural stagnation and
decay that lasted nearly 30 years? Which events unlocked this vicious circle in more
recent times? We proceed with an analysis of these major questions.
Obviously there is not a single and universally valid view as to what happened to
the rural sector of Argentina in the 1930s, when the agricultural frontier came to an
end and the country had to face the limits of its natural resource endowments. The
following considerations, however, should be taken into account:
First, in spite of the fact that prices for primary products suffered an important
downfall during the 1930s and 1940s and world demand for grains, wool, and beef
remained rather weak throughout those years, these should not be regarded as the prin-
cipal explanations for what happened to the Argentine agricultural sector at that point
in history. Rather, there were domestic reasonsa major increase in local consump-
tion, a downfall in the rate of mechanization, a high degree of uncertainty concerning
agricultural prices and profits, and political antagonism between the government in
power and rural landlords that prevented a more entrepreneurial behavior on the part
of the latterthat should be taken into consideration when trying to explain what
actually happened to the sector during the war years and the 1950s.
The need to emphasize the "internal" explanation is obvious if we observe the
dissimilar export performance of Argentina compared to that of other agricultural
giants in the same designated period.
Wheat and maize exports can be used as examples of what happened on this front.
In the prewar years Argentina's annual exports of wheat reached some 3.3 million tons
and represented about 20% of total world exports. Canada with 4.8 million tons
accounted for 28% of the total, Australia with 2.8 million tons absorbed 16%, and the
United States with 1.3 million tons represented about 7%. From 1950 to 1954, with a
much higher world demand for wheat, Argentine exports fell to 2.2 million tons, which
constituted only 9% of total world exports, whereas Canada had doubled its exports,
and the U. S. exports increased from 1.3 to 8.9 million tons (CEPAL, 1959).
Concerning maize, in the prewar period Argentina exported 6.3 million tons,
absorbing as much as 64% of the world trade. From 1950 to 1954, exports fell to just
over 1 million tons per annum whereas U. S. exports increased 3-fold from 800,000
to 2.6 million tons.
A similar situation can be found with regard to wool and beef exports where the
Argentine share of the world trade fell from 12 to 9% in the first case and from 40 to
19% in the latter comparing the 1950/1954 period to the prewar years.
Summarizing: falling prices and a weak world demand for primary products obvi-
ously had a negative impact on Argentine exports in the late 1930s and 1940s, but the
main explanation for its stagnation and decay is to be found in the fact that Argentina
produced less grain and beef in those years than in the 1920s and simultaneously
increased its domestic consumption of such goods, as a consequence of public policies
aimed at benefiting urban working classes and import substitution industrialization
efforts. As a result of this Argentina gradually lost its share of the world trade to the
United States, Canada, Australia, and New Zealand.
Second, as far as domestic consumption is concerned, we should notice that
whereas from 1925 to 1929 local consumption of agricultural products accounted for
48% of the total production, by 1950-1954 this proportion had expanded to 69%. Dur-
ing the war years wheat and maize were burned as substitutes for fuel, which could not
be imported.
Third, although domestic consumption of wheat and maize expanded, the pro-
duction fell in real terms. Labor shortages caused by heavy rural migration induced by
the rapid expansion of industry led rural land owners into open-pasture cattle raising
instead of agricultural production. On the other hand, the nationalization of foreign
trade and the setting up of low agricultural prices for various consecutive years resulted
in a climate of uncertainty and unrest among rural landlords. This antagonistic atmo-
sphere created a barrier against mechanization and a more rapid diffusion of fertilizers.
Consider the following evidence (CEPAL, 1959). The United States employed
less than 50 kg worth of fertilizers per acre in the prewar years and increased to around
170 kg per acre from 1950 to 1954. In contrast, Argentina employed less than 5 kg of
fertilizers per acre from 1950 to 1954. Of course, relative prices were highly different
among both countries, providing a good explanation for the observed differences in
the use of fertilizersconsider that 100 kg worth of wheat was needed to purchase 11
kg of fertilizers in Argentina from 1950 to 1954 while the same amount of grain man-
aged to obtain 44 kg of fertilizers in the United States. In addition to relative prices we
should also notice the antagonistic atmosphere in which the rural sector operated in
those days as a consequence of government policies.
A similar case can be made concerning the degree of mechanization. Between
1930/1934 and 1945/1949 the number of operating tractors decreased by about 40%
in Argentina, whereas they multiplied 3-fold in the United States. Between 1940 and
1955 the United States increased its stock of tractors from 1.5 to 4.3 million. By 1955
the United States had 30 tractors per 1000 acres under cultivation, which was about
15 times more than Argentina, which at that point had approximately 2 tractors per
1000 acres under exploitation. Again, relative pricesas well as import difficulties
during the war periodare good explanations for the observed differences in mecha-
nization, although we should not disregard the political aspects and the antagonism
between rural landlords and Peronist authorities that exacerbated risk-aversion atti-
tudes on the part of the former.
Beginning in 1952, a number of major policy changes were implemented by the
government, many of them directed at improving the position of the rural sector. A
major currency devaluation and a drastic change in relative prices favoring primary
production over industrial goods were particularly relevant in this respect. Slowly, the
agricultural sector gave signs of reaction. The area under cultivation began to increase
once again, reaching 18 million acres in 1956 and 1957.
Throughout the 1960s, the agricultural output grew by 30% whereas in the fol-
lowing 20 years the production of primary products multiplied 3-fold, and the pro-
ductivity of labor increased nearly 4-fold, being at present at their all time historical
peak.
Four different sources can be identified behind the observed trends (Obschatko,
1988). First, "disembodied" technical changenew agricultural practicesplayed an
important role during the 1960s. The National Institute for Agricultural Technology
(INTA)created in 1956was instrumental in the dissemination of new agricultural
practices and patterns of production organization. Second, the expansion of the trac-
tor industry, which took place during the 1960s, had a major impact on the degree of
mechanization employed by the rural sector. The country's total stock of tractors
expanded from around 77,000 in 1960 to 140,000 in 1970 and to 162,000 in 1980
(Huici, 1988). Four domestic subsidiaries of large multinational corporations entered
the local market for tractors during the 1960s and various forms of tax concessions
were used by the authorities to induce a rapid expansion in the degree of mechaniza-
tion of the rural sector.
The third major source of technological change and productivity growth as far as
agriculture is concerned, involved the diffusion of hybrids (in maize, sorghum, and
sunflower), new varieties of wheat, and the introduction of soybean into production
during the late 1970s.
Finally, the fourth and last set of innovative actions is related to the recent expan-
sion in the use of agrochemicals, herbicides, and pesticides.
An important change in the institutional scenario underlying these processes
should be noted at this point. Whereas in the 1950s and 1960s it is the public sector
mainly through the INTAthat carries out research activities in the field of new agri-
cultural practices and hybrids, particularly in the case of maize, the most recent trend
toward the more rapid diffusion of hybrids, new product varieties, herbicides, and pes-
ticides seems to be increasingly in the hands of domestic subsidiaries of large multi-
national corporations. Whereas in the former period technological change had a more
"disembodied" natureinvolving agricultural practices and production organization
technologiesin the latter it seems to have more of an "embodied" nature with
hybrids and agrochemicals now being the major way of introduction. Yet another
major technological innovation of recent yearsthis time of an institutional nature
relates to the diffusion of subcontracting practices as far as harvesting is concerned.
New forms of market organization are now allowing medium size and small farmers
to operate with capital-intensive equipment that the scale of their plots would not have
allowed them to own.
Having examined the nature and behavior of the national system of innovation
supporting technological change in industry and agriculture, we now turn to the study
of other parts of this system, in particular to those related to public R&D efforts.
PUBLIC R&D EFFORTS IN ARGENTINA
The Aggregate Picture

Public R&D expenditure and knowledge generation efforts came close to 380 million
U. S. dollars per annum in 1988. We arrive at this figure by adding up 330 million
U. S. dollars accounted for under the relevant chapter of the National Budget plus
some 50 million U. S. dollars spent by decentralized public enterprises (the National
Petroleum CompanyYPFbeing the largest one in this group), not included in the
previously mentioned Budget chapter.
Public R&D activities are highly concentrated. Eight agencies and institutions
absorb nearly 90% of the total budget, as Table 14.1 indicates. Most of these agencies
Table 14.1. Major Public Agencies and Institutions Carrying or Financing

R&D Activities in Argentina (Million U.S. Dollars), 1988
Total
Agency or Institution Expenditure
National Council for Science and Technology (CONICET) 135.4
Atomic Energy Commission (CNEA) 56.3
Institute for Agricultural Technology (1NTA) 67.2
Institute for Industrial Technology (INTI) 14.2
National Petroleum Enterprise
Yacimientos Pctroliferos Fiscales (YPF) 15.0
Armed Forces
Instituto de Invesligaciones Cicntificas y Tecnicas dc las 12.2
Fuerzas Armadas (CTTEFA)
Secretary of Science and Technology (SECYT) 8.2
Universities 30.0
Others 41.5
Estimated total public R&D expenditure 380.0
Sources: National Budget Law, complemented with direct information from YPF, CITEFA, and the
University of Bs.As.
were created after World War II and at least some of them experienced a period of
rapid expansion and institutional consolidation throughout the 1960s and early 1970s.
Ideologies, academic as well as corporate, and military "lobbies" have played a major
role as determinants of the budgetary allocation process as far as public R&D expen-
diture is concerned. We should therefore keep in mind the following aspects when
dealing with the individual history of some of these agencies:
1. Argentina's traditionally strong military forces exerted major pressure in favor of

developing domestic nuclear power capabilities and supporting the evolution of the
Atomic Energy Commission, which thus received special treatment within the bud-
getary process for quite a long period of time.
2. INTA's late creation, in 1956, came about after a drastic rethinking of the role of
agriculture in the country's long-term development process. As we have seen
before, such rethinking came about after nearly three decades of neglect, which had
dramatically negative consequences.
3. Almost as a matter of policy, Argentine universities have been kept as rather weak
institutions, just marginally contributing to the national system of innovation. This
was particularly so during authoritarian periods of military rule.
4. Research for the sake of new knowledgerather than technologically oriented
effortsand little tendency to collaborate with the production sector of the econ-
omy appear as distinctive ideological features of the local academic community.
This is particularly noticeable in the early history of CONICET.
Keeping in mind these ideological and political factors, let us now proceed with
a brief presentation of the evolution of some of the agencies and institutions.
Brief Description of Some of the Main Public R&D Activities
The National Council for Science and Technology (CONICET)

CONICET was created in 1958 with the purpose of "promoting, coordinating and car-
rying out research both in applied and pure sciences." Whereas in 1971, CONICET
had 490 research fellows and 13 affiliated institutes, by 1988 the figures had gone up
to 198 affiliated institutes and 11 regional research centers with a total payroll of close
to 7500 people. Such a dramatic institutional build-up took place in about one decade
and in a somewhat chaotic and unplanned fashion, more as a consequence of a delib-
erate action by the military authorities to weaken the institutional position of the
National Universities after the 1966 military coup d'etat than as a result of serious
long-term planning of public R&D activities. Needless to say, CONICET was by no
means administratively prepared to handle this build-up and is still suffering from seri-
ous bureaucratic disorder.
In 1988 approximately 40% of CONICET's research personnel belonged to the
medical and biological sciences, while the humanities absorbed 26% and the techno-
logical sciences 14% of the total payroll.
A few points stand out concerning CONICET's current situation. On the one
hand, it is the largest and fastest growing R&D agency in Argentina. On the other hand,
it exhibits a somewhat large interinstitute and interprogram variance in research qual-
ity and productivity and has not yet been able to introduce systematic quality control
and peer evaluation actions.
A special comment should be made concerning technology transfers toward the
production sector of the economy. A Technology Transfer Office was opened up by
CONICET authorities in 1985 with the purpose of strengthening the technological
linkages between its institutes and firms producing goods and services. Between 1985
and 1988, some 225 contracts were signed between both public and private firms.
About half of these contracts involve R&D activities jointly performed and financed
by both parts of the agreement, while another 40% of them correspond to services pro-
vided by CONICET personnel in areas such as quality control and metrology. Nearly
one-half of the contracts relate to chemical and petrochemical firms, while a surprising
10% is accounted for by companies dealing in biotechnology. In most cases, medium
size and small firms without their own R&D infrastructure are the ones engaged in
these contracts. The average size of the agreements is rather small with just a few of
them having funds of over 100 thousand U. S. dollars annually (Nivoli, 1990).
The Atomic Energy Commission (CNEA)
The Atomic Energy Commission was founded in 1951. It has a total labor force of
nearly 6000 people, 1600 of whom perform R&D activities. Although during the war
years the Argentine Military Forces entertained the idea of developing technological
self-sufficiency in the atomic energy field, it was only in the mid-1960s that the country
decided to build up nuclear power stations for electricity generation. The first such unit
was put into operation in 1973 and the second one in 1983. Although both of these
plants were purchased on a turnkey basis from international subcontractors, the
degree of local participation and technological "unpackaging" increased considerably
from the first to the second project (Tanis, 1986).
In May 1980, CNEA signed contracts with a German firm for a third nuclear
power station. This time CNEA went much further into the basic engineering speci-
fication of the technology it wanted and took responsibility for the overall supervision
of the project, thus indicating that the "learning" component had by no means been
negligible in this particular field of activity.
Ever since its creation, CNEA has had an explicit policy aimed at upgrading both
its technical personnel and its many subcontractors. A study carried out by Sabato et
al. (1978) indicates that nearly 1000 scientists and technicians have been trained
abroad in physics, chemistry, radiobiology, metallurgy, and material sciences. Also,
and as part of its efforts in the development of human resources, various pilot plants
and experimental research reactors were constructed during the late 1950s and 1960s,
thus providing the infrastructure for further applied R&D efforts.
Besides developing "in house" technological capabilities, CNEA generated a sig-
nificant flow of externalities by upgrading and developing an extensive number of sub-
contractors.
During recent years, CNEA signed various "joint venture" agreements with pri-
vate firms. As in the case of subcontractors, CNEA's impact on the private nuclear
industry has been far from negligible. It has also developed an engineering design
capacity that is currently being offered both locally and in foreign markets through
INVAP, an R&D and engineering firm presently employing more than 500 people and
carrying out research both for CNEA and for private firms in fields such as electronic
instruments, and biomedical equipment.
CNEA has recently suffered major reductions in its overall budget as well as the
loss of scientific personnel. The fact that the country has, in recent years, discovered
huge reserves of natural gasit is now estimated that approximately 50 years of gas
reserves are stored undergroundcasts strong doubts on the likelihood of the future
expansion of Argentina's nuclear power sector.
The National Institute for Agricultural Technology (INTA)

This public R&D agency was created in 1956 with the purpose of strengthening agri-
cultural research and extension activities. It is interesting to note that in a country
where primary raw materials have always been a major source of economic growth, so
little importance had been given before to R&D efforts related to the primary sector.
Obviously, this is in sharp contrast to the early interest such activities received in sim-
ilarly endowed countries such as Australia or Canada, where the upgrading of rural
life, institutions, and production capabilities appeared as a major public commitment
much before they did in Argentina (Fogarty, 1977).
INTA's budget is financed with a 1.5% ad valorem tax on agricultural and live-
stock exports. With a total personnel of close to 5000 people, INTA employs today
some 1900 professionals and technicians, a figure that is about three times higher than
its initial level in 1956/1957. Approximately 230 of these people pursued postgraduate
studies at a Masters level while 54 completed Ph.D. training. Just over 100 students
currently receive local training at a Masters level, while nearly 40 more are enrolled as
Ph.D. candidates at various educational institutions around the world.
INTA runs a large number of regional experimental farms and extension agen-
cies. Nearly 30 research programs have been organized and pursued by INTA's staff
throughout the years. Obviously, not all of these programs have attained the same
degree of success, but there is evidence that some of them have been highly successful
in the generation and diffusion of genetic materials and in the design and diffusion of
production organization technology. In particular Penna et al. (1983) have shown very
high social returns for INTA's R&D efforts in relation to new wheat varieties derived
from Mexican germoplasm. Similarly, successful results have been reported in yet
another recent study (Katz and Bercovich, 1989) with regard to biotechnology. In spite
of this, however, expert opinion suggests that in terms of molecular biology and genet-
ics, INTA is presently lagging well behind the international technological frontier.
As previously explained Argentine universities do not constitute an important source

of new technology. Basic research is normally performed for the "advancement of sci-
ence." Only 30.0 million $US5% of the total 640 million $US that make up the
budget for the 26 National Universitieswent to R&D activities in 1988. It is believed
that just one of four university professors performs research efforts, and nearly one-
third of those that do belong to the biomedical fields.
The University of Buenos Airesthe largest in the country with 110,000 stu-
dentshas recently set up a Technology Transfer Office and signed some 20 contracts
for the transfer of technical knowledge to the production sector. These contracts are
of a somewhat modest size. Furthermore, this University has just completed building
a company, UBATEC, jointly with private firms and the Municipality of Buenos
Aires, with the objective of facilitating the transfer of technology to the productive sec-
tor.
University activities in general and R&D efforts in particular seem to be going
through a period of decay and lack of financing that might prove difficult to overcome
and damaging in the long run, from the point of view of the country's scientific capa-
bilities.
Defense R&D Expenditure

Chapter 8 of the Budget Law indicates that defense R&D activities absorbed about 6%
of the total public R&D expenditure in 1988.
Roughly 60% of these expenses were carried out by CITEFA (Instituto de Inves-
tigaciones Cientificas y Tecnicas de las Fuerzas Armadas), the largest research agency
within the military sector. CITEFA employs close to 900 people, 200 of which are
highly trained military engineers. The agency's annual budget is about 10 million
U. S. dollars, 25% of which has been normally assigned to R&D projects.
Argentina has developed a certain degree of autonomy as far as the development
and construction of weapons is concerned. Medium-range missiles, jet fighters,
medium size tanks, and cannons are locally designed and constructed, largely in mil-
itary factories.
Exports of military equipment and technology have taken place in the past both
within Latin America and also to Middle Eastern countries such as Iraq and Iran, but
Argentina has had a far less significant experience in this respect than Brazil, which
has managed to develop a highly successful export sector in armaments, largely based
on active R&D and engineering collaboration between state-owned enterprises and
private firms. This kind of collaboration is not to be found in Argentina.
We conclude this overview of public R&D activities in Argentina as yet another
part of the national system of innovation. In spite of the potential social value of public
R&D activities, many of the approximately 20,000 professionals now involved are
grossly underutilized and could become important agents of modernization and tech-
nological change if only the country could find the institutional and economic mech-
anisms and the incentives for this to be so. We shall explore such matters in the last
section of this chapter.
ARGENTINA'S NATIONAL SYSTEM OF INNOVATION: OVERVIEW AND POLICY

IMPLICATIONS
A careful examination of Argentina's national system of innovation shows that the

country is far from having an integrated and coherent network of agents, institutions,
and policies dealing with questions of scientific and technological knowledge genera-
tion, diffusion, and utilization. Neither does it have a technological strategy that would
reflect a successful exploitation of its opportunities in the international market place.
Consider first the situation of the Argentine manufacturing industry. Having
relied on direct foreign investment and on external transfers of technology as a major
source of technological dynamism and due to the fact that domestic production facil-
ities are of a highly idiosyncratic nature as far as size, production organization, and
degree of vertical integration are concerned it can scarcely be surprising to know that
most R&D efforts performed by manufacturing firms have taken the form of "adap-
tive" engineering activities that, in many cases, would not fall under internationally
accepted definitions of R&D activities. The lack of suitable technical know-how and
engineering routines, differences in raw materials, and a highly particular legal and
institutional environment frequently induced local subsidiaries of large MNCs to use
domestic engineering and technological manpower with the purpose of recreating
product design and/or process engineering technolgies that are simply not available
"from the shelf." Locally owned companies have normally followed a similar pattern
of behavior. Although "adaptive" R&D efforts can and have systematically been per-
formed by firms, it is seldom the case that companies pass this stage and engage them-
selves in more complex state-of-the-art R&D activities, as we see is happening, for
example, in the case of many Korean firms (Kim, 1990).
There might be various reasons for this phenomenon. On the one hand, the his-
torically overprotected nature of the domestic industrialization process probably mil-
itated against a more dynamic outward-going attitude on the part of local entrepre-
neurs. Firms have never really felt the challenge of external competition domestically,
nor have they been coached by policy makers to search for export markets. Tax con-
cessions have not been handed out on a quid pro quo basis for R&D and export com-
mitments as we see happening in many countries around the world.
Concomitant with the above, it is also true that macroeconomic instability has
played a major role in this field, exacerbating risk-aversion attitudes and oportunistic
behavior on the part of local entreprenurs. The search for subsidies and special privi-
leges has normally been better rewarded than innovative behavior and investment in
new production capacity. In actual fact investment in new capital equipment has dra-
matically been reduced during the course of the last decade and so have imports of
machinery and equipment, R&D activities, and entrepreneurial spirit in general.
Yet as previously explained, the industrial sector of Argentina is presently under-
going a major transformation. The old overprotected industrial sector is gradually los-
ing ground and a new breed of technologically more updated firms, involved in the
production of intermediate industrial commodities and resource-based products, is
taking its place. It is certainly much too early to make predictions as to the future suc-
cess of these newcomers, as well as to their likely technological strategies in the years
to come. A more stable macroeconomic environment, a higher degree of exposure to
external competition, and a more active government policy on the technological front
might well induce them into a stronger technological commitment in the years ahead.
So much for the industrial sector. As far as agriculture is concerned, we have
pointed out that after many years of technological stagnation the sector is presently
undergoing a period of rapid technological change associated with the diffusion of new
product varieties and hybrids as well as with the massive utilization of herbicides, fer-
tilizers, and new production organization techniques. Sunflower and soybean produc-
tion have expanded massively in recent times. However, in spite of the fact that both
the area under cultivation and yields per acre are currently at their all time historical
peaks, we cannot fail to notice that the sector is still far from exploiting its real growth
potential and available opportunities in the international market place. In this respect,
the national system of innovation supporting technical change and innovation in agri^
culture is not performing an entirely successful task.
Let us now turn to public scientific and technological activities. Lack of priorities,
of ex ante evaluating mechanisms, of ex post peer review missions, and of institutional
coordination account for low research productivity, a scarce sense of purpose and
direction, frustration, and a great deal of slack and inefficiency. Many of these short-
comings, however, can be found in other spheres of activity in contemporary Argen-
tina and are by no means exclusive to public R&D agencies.
As explained before, much seems to be out of place in present day Argentina, with
per capita income some 20% lower than a decade ago and a continuing uncertain sit-
uation as far as macroeconomic equilibrium is concerned.
It would be wrong to assume that the present state of affairs is exclusively due to
lack of resources and that by providing more money to the public R&D systemwith-
out simultaneously acting on aspects of organization, coordination, and accountabil-
itythings could be improved significantly. No doubt research personnel is being
grossly underpaidthe average monthly income of a professional graduate working
in any of the previously mentioned agencies is presently somewhere in the order of
300 U. S. dollars per month, but overlapping programs, research groups that are just
"empty boxes," and institutions that do not have the minimum critical scale in terms
of experimental equipment or qualified personnel required to attain sensible results,
suggest that a great deal has to be reorganized if the expansion of public R&D expen-
diture is to become socially worthwhile in the future. The Secretary of Science and
Technology clearly has a long and complex agenda ahead of it, in terms of coordina-
tion, evaluation, and reorganization of R&D institutes and programs.
Part of such reorganization efforts could probably involve a new and different set
of principles in relation to collaborative ventures and risk-sharing activities between
public agencies and private firms producing goods or services. As we have seen before,
both CONICET and the University of Buenos Aires are already exploring new mech-
anisms in this respect and it would be advisable to further proceed along these lines in
the future.
Let us now turn to a few final comments concerning the national system of inno-
vation as a whole, its priorities, and future possibilities.
Clearly, one of the major challenges Argentina faces today is how to regain the
capacity for growth. After nearly two decades where output per head and the rate of
new capital formation have been steadily declining it is quite clear that the question of
how to return to a viable long-term growth path is a very major one.
Let us assume that the industrial restructuring process we have previously
described will continue in the immediate future and that as a consequence of this,
Argentina will eventually settle down to a new "industrial regime" in which the struc-
ture of industry, the organization of work, the social division of labor, and industrial
relations will lose many of the idiosyncratic features associated with the "old" domes-
tically oriented regime.5 In such a scenario, manufacturing and agricultural exports
will probably continue to grow.
Under such circumstances (i.e., large positive foreign trade balances) domestic
fiscal and macroeconomic policy issuesobviously related to questions of fiscal equi-
librium, social equity, and income distributionare necessarily going to acquire
increasing importance in the future. The transition from the old overprotected and
domestically oriented regime to the new onemore deregulated and open up to mar-
ket signalsis proving to be far from costless. Quite the contrary, high unemployment
rates and a deteriorating pattern of social equity and income distribution seem to be
undesirable consequences of the ongoing transformation. In a country where there is
not yet an ex ante public debate of the national budget, where tax evasion is dramat-
ically high, and where public authorities are not capable of managing a credible fiscal
policy, there are obvious institutional shortcomings that need to be corrected if the
country's growth potential is to be realized and if acceptable levels of social equity are
to be maintained. The national system of innovation could significantly contribute to
such goals in more than one direction. On the one hand, it could certainly collaborate
in the modernization and upgrading of the nation's badly deteriorating social sec-
torssuch as health or educationwhich are now lagging behind as a consequence
of the drastic reduction in social public expenditure and investment (Dieguez et al.,
1990). On the other hand, it could become an important source of new ideas and tech-
nology with which to undertake the country's transition to the new world of microe-
lectronicsincluding flexible production automation and telecommunications
which is rapidly developing internationally and which Argentina will have to incor-
porate in the years ahead. In both of these spheresas well as in many othersArgen-
tina will probably have to face difficult decisions in the immediate future concerning
redeployment of human resources, upgrading of its labor force, deciding how much
foreign "from the shelf technology it should use given its particular resource endow-
ments and regulatory environment, and how much such technology is to be comple-
mented by domestic R&D efforts to be adequately absorbed by the local society. As
the previous examples suggest, there is a long list of issues related to the functioning of
its national system of innovation to which Argentina will have to pay close attention
in the future, if it is to successfully adapt itself to ongoing changes in the world's insti-
tutional, technological, and economic scenarios. In many of these fields, we can iden-
tify powerful reasons for which decentralized market decisions might not necessarily
be enough to ensure socially acceptable outcomes. New forms of intervention will
probably have to be explored in the future even in spite of the fact that government
intervention has a clear history of failure as far as Argentina is concerned. How to work
out an adequate balance between public regulation and decentralized market deci-
sions in fields where there are obvious reasons for market failure constitutes, of course,
a difficult problem for which the country will have to work out imaginative new
answers if it is to regain the capacity for growth and to maintain acceptable levels of
social equity.
NOTES
Both authors belong to Argentina's National Council for Scientific and Technological
Research (CONICET). The former is Professor of Industrial Economics at the University of Bue-
nos Aires and Consultant to the Economic Commission for Latin America (ECLA). The ideas
hereby presented are their exclusive responsibility. The present chapter has been prepared as a
contribution to the project "National Systems Supporting Technical Advance in Industry" and
is based upon drafts presented by the authors at the seminars held in November 1989 in Maas-
tricht, Holland, and in October 1990, at Stanford, California. The authors would like to
acknowledge detailed comments from the participants at both such workshops.
1. It is usual among social scientists to compare the case of Argentina with that of Australia,
which underwent a similar process of early integration into the world economy, but whose long-
term performance has been significantly better than the one attained by Argentina. A wide vari-
ety of explanations of the observed differences have been put through in the literature. On the
one hand, a richer natural endowment of gold and minerals allowed the Australian population
to enjoy a higher income per capita than Argentina, even before the nineteenth century agricul-
tural export boom had actually started (Diegues, 1969). On the other hand, various writers have
argued (Fogarty, 1977) that the earlier exhaustion of the agricultural frontier induced the Aus-
tralian government to press for a rapid and widespread diffusion of fertilizers and new varieties
of wheat, this having a major and earlier impact on land productivity and agricultural profits.
In addition to the above we should notice that a different set of institutions, for example,
highly different land leasing contracts, induced a much less careful land exploitation system in
Argentina than in Australia as well as a longer lag in the adoption of new agricultural technol-
ogies. At a more general level, it is important to realize that in spite of its economic success,
Argentina failed in those early years of the century to develop and consolidate civilian and polit-
ical institutions of the sort Australia managed to establish together with its rapid economic
expansion. Such weaker institutional fabric was to remain as a central feature of the Argentine
society and still today recurrently appears as a major barrier to a more successful socioeconomic
performance.
2. Political historians have shown that the Argentine Armed Forces have had a strong polit-
ical and institutional presence in Argentina ever since the 1920's, or even earlier (Potash, 1980).
It is, however, after the 1930 right-wing nationalistic coup d'etat headed by General Uriburu
against the civilian government of President Yrigoyen that the political role of the Armed Forces
and the nationalist ideas underlying military education became much more influential in rela-
tion to the country's long term development process. Notions of economic self-sufficiency, of a
strong military production industry capable of designing and constructing different types of
defense goods, etc, gain preeminence at that point (D. Rock, 1985). Such military presence has
had a significant impact upon the national system of innovation thereafter, inducing patterns of
institutional behavior with long lasting consequences for the country's development process.
Consider, for example, that for many years the production of steel was regarded as an issue of
'national security' and that private entrepreneurs were not allowed to enter the market or to
introduce new production technologies. 'Down stream' industries producing consumer durables
and capital goods were negatively affected by such policies. Similar examples can be found in
other areas of the production structure.
3. The structural model relating industrial growth, import requirements, foreign exchange
shortages, devaluation, and income distribution effectsgiven the fact that Argentina's main
exports are wage goods whose domestic prices immediately reflect changes in the value of the
foreign currencyhas been extensively discussed in the local and international professional lit-
erature since the early contribution by Diaz Alejandro (1970). See also Canitrot (1975).
4. The neoclassical methaphor of complete specification and perfect availability of "from-
the-shelf' production functions just does not seem to be very useful if one is to make sense of
the present case. Rather, evolutionary ideas based on notions of incomplete specification of pro-
duction knowhow, imperfect information of domestic entrepreneurs, "adaptive" R&D efforts,
"technological learning," bottlenecks, and disequilibrium seem to fit much better the set of
issues we need to examine. In relation to such topics the reader could see Nelson and Winter
(1982) and Dosi et al. (1988).
5. It is outside the scope of the present chapter to examine various potential difficulties that
might in the future hamper the development of this new industrial regime. High structural
unemployment, falling terms of trade in the field of industrial commodities, a continuously dete-
riorating pattern of income distribution, and a high level of social conflict could very well
increase in the future, opening up major questions as to the likelihood of a steady consolidation
of present trends.
REFERENCES Escude, G., and Gucrbcroff, S. (1990). "Ajuste

maeroeconomico, deuda exlerna y ahorro en la
Argentina." In Ahorro y formation de capital.
Azpiazu, D. (1985). La promotion industrial en la CEPAL/ONUDI, Buenos Aires: Grupo Editor
Argentina 1973-1983. Buenos Aires: CEPAL. Lationamericano.
Ablin, E., and Katz, J. (1985). De la Industria Inci- Fogarty, J. (1977). "Difusion de technologiaen areas
piente a la Exportation de Tecnologia. La exper- de asentamiento reciente: el caso dc Australia y de
iencia Argentina en la venta international de plan- la Argentina." Desarrollo Economico, Buenos
las industrials y obras de Ingeneiria. Cepal/ Aires, April/June.
Eudeba, Bs.As. Huici, N. (1988). La inditstria de maquinaria agri-
Ablin, E., Gatto, F., Katz, .!., Kosacoff, B., and cola en la Argentina. Buenos Aires: C1SEA.
Soifer, D. (1985). Internationalization de empre- Katz, J. (1973). Patentes de invention, Convenio de
sas y tecnologia de origen Argentina. Buenos Paris y paises de menor grado de desarrollo rela-
Aires: Cepal/Eudeba. livo. Mimco, Institute DiTella, Buenos Aires.
Bercovich, N., and Kalz, J. (1988). "Innovacion Katz,.!. (1974). Importation de tecnologia, aprendi-
genetica, esfuer/o publico de I&D y la frontera zaje e industrialization dependienle. Mexico:
tecnologica internacional. Nuevos hibridos en el Hondo de Cultura Economica.
INTA." Desarrollo Economico, Buenos Aires, Katz, J., & Kosacoff, B. (1983). "Multinationals
July-September. from Argentina." In S. Lall (ed.), The New Mul-
Canitrot, A. (1975). "La experiencia populista de tinationals. The Spread of Third World Enter-
redistribution de ingrcsos." Desarrollo Econom- prises. New York: Wiley.
ico, Buenos Aires, October/December. Katz, J., and Kosacoff, B. (1989). El proeeso de
CEPAL (1959). Analisis y proyecciones del Desar- industrialization en la Argentina: Evolution,
rollo Economico. El desarrollo economico de la retroeeso v prospectiva. Buenos Aires: CEAL/
Argentina. Parte II. Naciones Unidas, Mexico. CEPAL.
Diaz Alejandro, C. (1970). Essays in the Economic Kim, L. (1990). Korea's National System for indus-
History of the Argentine Republic. New Haven: trial innovation. Mimeo, presented at Stanford
Yale University Press. University, (Final version in this book.)
Dieguez, II. (1969). "Argentina y Australia: algunos Lall, S. (1982). Developing Countries as Exporters of
aspectos de su desarrollo economico comparado. Technology. A First Look at the Indian Experi-
Desarrollo Economico. Buenos Aires, January/ ence. London: Macmillan.
March. Nelson, R., and Winter, S. (1982). An Evolutionary
Diegue/, H., Llach, J., and Petrecolla, A. (1990). "El Theory of Economic Change. Boston, MA: The
gasto Publico social." Centra de Investigaciones Belknap Press of Harvard University Press.
Economicas, Institute Torcuato DiTella, August, Nivoli, A. (1989). "Balance de la experiencia de la
Buenos Aires. Oficina de Transferencia de Tecnologia." Revista
Dosi, J., Freeman, C., Nelson, R., Silverberg, G., and del Derecho Industrial, Edicioncs DEPALHA,
Socte, L. (1988) Technical Change and Economic- Buenos Aires, January-April.
Theory. London: Pinter Publishers. Obschatko, E. (1988). Transformation economicay
tecnologica de la agricultura pampeana 1950- trial Argentine en el periodo de Martinez de

1984. Buenos Aires: Ediciones Culturales Argen- Hoz." Desarrollo Economico, Buenos Aires,
tinas. October/December.
Penna, ,1. (1983). Difimon de las variedades de trigo Tanis, S. (1986). "Evaluation de lacapacidad indus-
con germoplasma mexicano y su impacto en la trial Argentina y desarrollo de los proveedores
production national. Buenos Aires: INTA. para instalaciones nucleares." Paper submitted at
Potash, R. (1969). The Army and Politics in Argen- the Seminar on Supporting Industrial Infrastruc-
tina. 1928/1945 Yrigoyen to Peron. Stanford: ture Requirements and Development for Nuclear
Stanford University Press, 1980. Power, organized by the International Atomic
Rock, D. (\WS) Argentina 15/6-1987. From Span- Energy Agency, Vienna, April.
ish colonization to the Falkland^ war. Berkeley, World Development. (1984). Exports of Technology
CA: University of California Press. by Newly-Industrializing Countries. Case studies
Sabato, J., Wortman, O., and Gargiulo, G. (1978). on Hong Kong, Taiwan, Korea, India, Egypt, Bra-
Energia Atomica e Industria National. Washing- zil, Mexico, Argentina, Vol. 12. Oxford: Perga-
ton, D. C: OEA. mon.
Schvarzer, J. (1983). "Cambiosen el liderzgo indus-
15
The Innovation System of Israe!:
Description^ Performance, and
Outstanding Issues
MORRIS TEUBAL
The peculiarities of the Israeli innovation system stem to a large degree from the envi-
ronment surrounding and the forces shaping Jewish immigration to Palestine during
the first decade of the century. Specifically, the tradition of Jewish scholarship, the
determination to create a modern state, and the Zionist ideology were driving forces
that created functioning, institutionalized bases for modern scientific research, tech-
nological development, and the training of new generations of scientists. Both the
Hebrew University and the Technion of Haifa were established almost a quarter of a
century before the creation of the State in 1948. (The Weizmann Institute began oper-
ations in 1934.) So did the appearance of significant applied research in agriculture
and in medicine, as well as the beginnings of industrial research.
The process of expansion and innovation in the science and technology system
continued during the first decades after the creation of the State (the 1950s and 1960s),'
a fact that explains some of its salient features: the high fraction of GNP devoted to
R&D, the strength of the university system, and the dominant roles of both defense
and government finance of R&D. This was also a period of significant economic
growth by international standards.
The 1970s and 1980s represent a break from historical trends, both in terms of
economic growth performance and in terms of innovation in the science and tech-
nology system. The first section of this chapter focuses primarily on economic perfor-
mance and the second section on policies (primarily technological) and institutions of
the 1970s and 1980s. The former section looks at the connection between structural
change and economic growth. After briefly describing the 1950s and 1960s, it presents
the "paradox" of the 1970s and 1980s: a continuation of deep structural change
facilitated by abundant skills and leading to the emergence and growth of high-tech
industryand the onset of economic decline and stagnation. Possible "crowding out"
effects of activities in the defense sector during the eighties are considered. Finally, the
second section describes the stagnation occurring in science and technology policy
during the 1980s, one that parallels economic stagnation. Its maximum expression is
476
INNOVATION SYSTEM OF ISRAEL 477
absence of the new technological infrastructures, which, given the abundant stock of
skills, could have propelled the economy back into reasonable growth. This institu-
tional-organization-political (rather than market) failure takes place across the board:
in relation to SMEs and conventional industry, in relation to defense and high-tech,
and in connection with strategic science. In the final section I relate it to the inability
to create new patterns of cooperation and networking within the economy, and with
the incapacity of Israel's government in the 1980s to plan ahead and to crystallize a
vision of its place in the world community.
GROWTH, STRUCTURAL CHANGE, AND SCIENCE/TECHNOLOGY

BACKGROUND
In this section I present essential background and motivation for an in-depth analysis
of technology policies and related institutions during the 1970s and 1980san impor-
tant aspect, in our opinion, of Israel's declining economic performance.
Economic Development and Research Patterns during the 1950s and 1960s
Economic Structure
The Israeli economyfollowing the behavior of the Jewish economy during the Man-
date period that preceded the establishment of the State in 1948has a long tradition
of rapid structural change actively promoted by government policy. Although the
share of manufacturing (including mining) has not changed substantially since the
1920sif anything, it has slightly declined during the period, standing at around 20%
of GNP during 1986there has been considerable change in the structure of manu-
facturing, particularly during World War II and after the establishment of the State.
During the 1950s, a number of industrials sectors were targeted by the government:
food processing, clothing and leather, and metals. These sectors, together with the agri-
cultural sector, which also received preferential support, contributed to supply the
needs of the almost 700,000 immigrants who came to the State during the 1948-1951
period, most of them Jewish refugees from Arab countries.2 The industrial sectors sup-
ported were largely labor intensive, and the methods for support involved both import
tariffs and quotas and investment grants and loans. During the 1960s, government
targeting focused on another set of industrial branches, more capital intensive and pos-
sessing a larger export orientation. These sectors included textiles, chemicals, cement
and mining, and, toward the end of the 1960s, high-tech industry. The methods
involved both those mentioned above together with export subsidies and the begin-
nings of subsidization of R&D in industry. The general evaluation of such policies,
which had only a minor explicit technology promotion component, was positive. Pack
(1971), in his book on the Israeli economy, suggests that this was the case despite the
fact that the growth path followed was not necessarily the efficient one from the ben-
efit-cost analysis viewpoint. Thus, despite the inefficiencies of the government in exe-
cuting what was largely an investment, population dispersal, and employment policy,
the aggregate growth rate of the economy was one of the highest in the world, possibly
second only to Japan during 1950-1970.3 One can understand the contribution of gov-
ernment to this process, given market failures in the early stages of "modern" indus-
trial development: infant industries and technologies; scarcity of entrepreneurs;

underdeveloped capital markets; the coordination requirements of industrial devel-
opment, immigration absorption, and defense; and other factors. Moreover, the tasks
confronting the young state were relatively clear at the time, and a set of quite remark-
able individuals filled the ranks of government during what was, in fact, the "heroic"
period of Israel's statehood.
Science and Technology
Scientific activities during the period of the Mandate (1918-1948) were mainly con-
cerned with the study of characteristics of the landclimate, soil, water resources
plant and animal life, plagues and illnesses, geography, and geology. Basic research,
conducted in accordance with the western tradition of academic freedom, and agri-
cultural research achieved a high level relative to the standards of the period. There
was practically no industrial research, except for some connected with the Dead Sea
Works. After independence, Prime Minister D. Ben-Gurion himself headed the
Research Council founded in 1949; its objective was to extend further the institutional
structure of scientific work in Israel. The new government established a number of
governmental research laboratories during the 1950s, for example, the Fibers Institute,
the aim of which was to support the textile industry being developed to provide
employment to the more than half a million Jewish refugees from Arab countries, and
the National Physics Laboratory. It also established several new universitiesTel-
Aviv, Bar Ilan, Negevand founded other institutions such as the National Council
for Research and Development (NCRD) in 1959 and the Israel Academy of Sciences
in 1961.
By the mid-1960s, institutional arrangements for scientific work were well under
way; additional efforts were made to reinforce existing institutions rather than to estab-
lish new ones. This period shows also the first attempts at a more systematic approach
toward research and development. Thus among the objectives of the National Council
for Research and Development we find planning governmental policy toward research
and development and defining "national research needs" in various fields. The most
significant event, however, was the nomination in 1966 of a committee for the orga-
nization and administration of government researchthe Kachalsky committee. The
committee's main recommendation was that bureaus of chief scientists be created in
ministries such as the Ministry of Commerce and Industry to coordinate their activi-
ties in research and technology and to stimulate applied research. This is an essential
background to the growth of civilian R&D expenditure in industry during the 1970s.
Agricultural Research
Research in agriculture played an important role during the Mandate period and dur-
ing the two first decades of the State of Israel.4 Scientific knowledge was perceived as
instrumental for a shift from the traditional dry-farming methods (involving both sep-
arate animal and plant husbandry and manual work) to a mixed crop pattern that
would be combined with the raising of cattle and with irrigation. Organized agricul-
tural research began in the 1920s with the establishment of the Jewish Agricultural
Research Stations (A.R.S., now the Volcani Institute), founded on the pattern of
American agricultural and mechanical colleges. The knowledge required in the early
period fell into two categories: knowledge concerning the natural characteristics of
land and knowledge that could be directly applied. A significant part of the activities
of the A.R.S. consisted of research into the botany, zoology, and geology of the terri-
tory. This "generic research" was both of scientific interest and directly relevant for
the local agricultural communityit provided basic information needed for the selec-
tion of areas of settlement; and without surveying the conditions of the soils, climate,
plants, and animals the possibility of applying the results of new experiments done in
the country or elsewhere would have been extremely limited. In spite of the freedom
of agricultural research workers from bureaucratic controls and political pressures,
there emerged close informal contacts between them and the agriculturalists with the
result that the contribution of agricultural research to productivity growth in agricul-
ture was, and is still, substantial (e.g., for ideological reasons scientists originally also
did extension work and provided essential services to agriculturalists such as seed test-
ing and artificial insemination). During the 1930s, the relationships between the two
groups began to take a more institutionalized form with the emergence of "branch
committees" (field crops, citrus, cattle, etc.) where both researchers, users, and admin-
istrators were represented.
In the 1940s and 1950s, with the creation of the Faculty of Agriculture at the
Hebrew University, agricultural researcheven at the Vulcani Institute, where sci-
entific publications increasingly became the basis for promotionbecame increas-
ingly academized. Moreover a functional separation of research from services/exten-
sion has gradually emerged. The contribution of agricultural research, however, to
agricultural productivity seems to have continued throughout the 1980s, although few
systematic studies have attempted to make such an assessment.
The Rise of High-Tech Industry and Declining Growth Performance During the
1970s and 1980s
The relationship between rapid structural change and fast aggregate and per capita
income growth was reversed in some sense after 1973-1974. This was precisely the
period of emergence and development of high-tech industry in what is widely regarded
as a successful result of industrial and technological policy. The most salient feature
of the past 20 years is the rise in the relative importance of this group of skill- and
R&D-intensive industries. Table 15.1 shows an industry categorization based on skill
intensity at a three-digit industrial branch level. Skill intensity is defined as the pro-
portion of scientists and engineers in the labor force (see Halperin et al., 1988; Teubal,
1989). This seems to be the most crucial variable in defining degree of sophistication,
at least in Israel where R&D data exclude defense R&D. The three clearly defined cat-
egories of industry are Category "H" (high-tech), which includes electrooptics and fine
machinery (SIC numbers 280, 281), aircraft and ships (262, 263), and electronics and
communications equipment (254, 253); Category "OS" (other sophisticated), which
includes various categories of metal products, heavy and specialty chemicals, and elec-
trical goods; and Category "C" (conventional), which includes all the rest, such as food
and textiles, pulp and paper, and wood and furniture. The skill intensities for the three
categories were 16,6, and 2%, respectively, in 1984. These three categories largely cor-
respond to the categories proposed by Pavitt (1984)science-based, production-
intensive, and traditionalwith some differences that are largely specific to Israel (e.g.,
the pharmaceutical industry is not yet as high-tech in Israel as it is in most advanced
countries).
Table 15.1. Israel: Industry Categories and Skill Intensity, 1984"

Subsector/Category SIC Skill Intensity (1984) %
Category "H"
(280,281) Electrooptics and fine machinery 12
(262,3) Aircraft and ships 16
(254,3) Electronics and communications equipment 18.5
1984 16.0
1987 (18.5)
Category "OS"
Basic metals (220,1) Iron and steel castings 4.1
(223) Pipes 6.4
Metal products (233,4) Heating and cooking equipment 4.8
Machinery 240, 1 ,3 For agriculture, industry, and construction, pumps and 7.0
compressors
242 Used in services and in households 4.9
Electrical 250 Electric motors 6.4
machinery 251,2,3 Batteries, household consumer durables, illuminating 4.3
equipment
Mining 100,3,4,8 Mining of salts and minerals 7.2
Chemicals 200 Basic chemicals 8.6
201 Pharmaceuticals 5.7
202,3,4,8 Soaps and detergents, paints and other chemical 5.6
products
205 Disinfectants and insecticides (and herbicides) 8.4
Average Category "S" ( 1 984) 6.0
Category "C"
All other 3-digit (Food, textiles, wood products, leather and shoes, etc.)
subsectors
Average Category "C" (1984) 2.0
"Share of qualified scientists and engineers in total sector/category employment.
Table 15.2 describes structural change within the Israeli industrial sector between
1968 and 1983, in terms of percentage shifts in shares of industrial outputs and inputs
among the various categories mentioned above. The increase in the share of both
industrial product and industrial exports in favor of Category "H" (and largely against
Category "C") is truly dramatic: from 6 and 5%, respectively, in 1968 to 24 and 28%
in 1983. The shift in the share of inputs is also quite dramatic: a more than doubling
of the percentage share of skills, physical capital, and labor allocated to Category "H"
Table 15.2. Structural Change in Israeli Industry,0 1968-1983

Share of
Share of Product Share of Export Share of Skills Capital Share of Labor
"H" "OS" "C" "H" "S" "C" "H" "S""C" "H" "S" "C" "H" "S" "C"
1968 6 20 74 5 42 53 ~30 ~70 4 30 66 6 15 79
1983 24 20 56 28 32 40 -70 -30 11 32 57 17 14 69
Change in share +18 0 ^18 23 ^\Q ^13 +40 ^40 + 7 +2 -9 +11 -1 -10
"Excluding diamonds (all figures in percentages).
Sources: Halperin et al. (1988) and Teubal (1989).
Table 15.3. Growth Performance of Israeli Industry and

Business Sector (Average Annual Percentage Change)
1968-1972 1973-1979 1980-1984
Industry
Product 15.1 4.9 2.5
TFPG" 7.0 1.4 0.0
1961-1972 1973-1981
Business sector performance

Gross domestic product 9.7 3.4
GDP* per hour of work 5.8 2.4
TFPG 4.2 0.6
"TFPG, total factor productivity growth.
*GDP, gross domestic product.
Sources: Industry: Bregman (1987); business: Metzer (1986).
between those two dates. In particular, note that the share of scientists and engineers
employed in high-tech increased from 30 to 70% of the total pool employed in indus-
try.
The process of structural change occurring in industry was profound and far
reaching. Prima facie and following Kuznetz (1971) and the structuralist perspective
more generally, this should have prevented a sustained decline or stagnation of aggre-
gate manufacturing growth and of manufacturing total factor productivity growth
(TFPG).5 These have been declining sharply during the 1970s and early 1980s as can
be seen in Table 15.3 with industrial TFPG declining to zero during 1980-1984 (Breg-
man, 1987).6 The same trend can be seen when looking at the business sector (which
involves much more than industry)see the lower part of the table. Similar tendencies
can be shown when observing the economy as a whole (see Syrquin, 1986).
Growth of Skills and Quantitative Science and Research Indicators
The high growth of skills is a central factor explaining the rise of high-tech industry
since the Six-Day War. Tables 15.4-15.6 show the growth of skills, principally in
Israeli industry, since the 1960s. In Table 15.4, which is based on data from the Min-
istry of Industry and Trade (MI&T), we see that scientists and engineers in manufac-
turing rose from 3400 in 1968 to almost 20,000 in 1987 (an annual rate of growth of
Table 15.4. Israel: Employment, Skills, and Skill Intensity in

Manufacturing (Aggregate)
Year Total Employment Total Skills Skill Intensity (%)
1968 227,000 3,400 1.3

1971 272,000 4,184 1.5
1975 292,000 6,749 2.4
1978 308,000 8,860 2.9
1981 313,500 12,158 4.0
1984 331,000 17,450 5.3
1987 341,150 19,735 5.8
Source: Ministry oflndustry and Trade.
Table 15.5. Skilled Labor in Israel

1961 1972 1983
Scientists and engineers
Industry 1,470 3,560 10,180
Total 6,870 15,985 33,070
Percent industry 21.4 22.3 30.8
Technicians
Industry 9,225 15,315
Total 30,295 45,990
Percent industry 30.5 32.6
Sources: Central Bureau of Statistics, Census 1961, 1972, 1983; and additional
publications on the Labor Force (No. 13). A comparison of these figures to those
from the Ministry of Industry and Trade data shows them to he consistently
smaller. We believe that this reflects a more compact definition of both industry
and engineers in the census, as well as a possible bias in the reporting of firms in
the Ministry of Industry and Trade data base.
9.2%). The table also shows a very significant increase in the skill intensity of manu-
facturing industry during the periodfrom 1.3 to 5.8%. Table 15.5 also provides data
on skills in the manufacturing industry, not only our narrow definition including sci-
entists and engineers but also numbers of technicians (including practical engineers).
In addition, there exists information on the share of total skills employed in industry
(e.g., the share of Israel's scientists and engineers employed in industry increased from
21.4% in 1961 to 30.8% in 1983). Note that the increase in the share of technicians
and practical engineers employed in industry increased relatively little between 1972
and 1983 (it stands at around 1/3). Finally, Table 15.6 provides an international per-
spective of the growth in skilled manpower in Israel; it shows what was already becom-
ing clear, namely, that the growth of scientists and engineers in Israeli industry has
been phenomenal since 1967, probably one of the highest recorded (due provision
should be given to the fact that the base of skills in industry during 1967 was very low,
even less than one-half on a per-capita basis compared with the United States). It has
certainly outstripped the growth of skills for the country as a whole. This fact is related
to the growth of high-tech industry since 1967, to the university system, and to immi-
gration. If we accept the "structuralist perspective" viewpoint that skills are the critical
factor in economic growth, then the potential growth of the manufacturing sector
could have been enormous during the period in question. Why hasn't it materialized?
One aspect relates to the role of defense in industry, and, in particular, in high-tech
industry. Before considering this, let us present some additional quantitative indica-
Table 15.6. Scientists and Engineers in IndustryOECD and Israel

United West
United States Kingdom Germany France Japan Israel
1967 495,500 49,900 61,000 42,800 117,600 3,400
1980 573,900 80,700 1 1 1 ,000 68,000 272,000 12,150
(1981)
Percent increase 16 62 82 59 132 257
Source: Halperin ( 1 986).
tors of science and research in Israel, part of which are related to the growth of skills
(see Table 15.7).
Israel possesses one of the highest ratios of total R&D to GNPan estimate of
over 3% in the mid-1980s, compared to ratios approaching this number for three of
the major powersthe United States, Germany, and Japan (this ratio was also one of
the highest of the world during the mid-1970s); probably the highest ratio of defense-
related R&D to total R&D in the world. Our estimate is probably 2/3, from which it
follows that the non-defense-to-GNP ratio for Israel is most certainly surpassed by
those of Japan, West Germany, and other countries. Another feature is the extremely
high share of total and civilian R&D that is paid for by government. In the mid-1970s,
the proportion of civilian R&D financed by government was 62%, a number attained
only by Norway at the time (the figures for the other countries in the seventies range
from 27% for Japan to 47% for Belgium). Presumably, Israel preserves a very high
ranking in this respect even today. Some figures on skill intensity of the population
and skills involved in R&D (with skills defined as numbers of scientists and engineers)
are also known from other sources: during the mid-1970s, Israel possessed one of the
highest ratios of skills to total population40 per 10,000compared with the highest
ratio of 42 reached by the United States at the time. For the mid-1980s, we have some
figures (Table 15.7C) showing the numbers of recipients of academic degrees in the
Table 15.7. Israel: Some Basic Science and Technology-Related Indicators

A. Aggregate R&D Ratios"
Government
Defense R&D Financing of
Total R&D X 100 X 100 Total Civilian R&D
GNP R&D (%)
Year 1974-1976 1987 1970s 1981 1970s 1984
1974-1976, 1987 2.2 >3.0 40 65 62 797
B. The Distribution of Civilian R&D (percent)

Year University Industry Government
1974-1975 60.0 22.0 18.0
1978 45.0 43.0 12.0
1982-1983 47.0 37.0 16.0
various sources
C. Recipients of Academic Degrees in the Natural Sciences and Engineering per 10,000
Participants in the Labor Force
First Degree Advanced Degree
Natural Natural
Year Sciences8 Engineering Sciences Engineering
1986-1987 9.5 8.1 4.2 2.0
" Ministry of Science and Development (1988).
Sources: UNESCO, Statistical Yearbook 1987, Paris, 1988; and for Israel: Israel Central Bureau of Sta-
tistics, Recipients of Degrees from Universities 1986/87, Jerusalem, 1988; Council for Higher Educa-
tionPlanning and Grants Committee, Higher Education in IsraelStatistical Abstract 1986/87, Jeru-
salem, 1988.
natural sciences and engineering per 10,000 participants in the labor force; both for
advanced and for first degrees, Israel ranks very high in a sample of advanced countries
(only surpassed by Japan or the United States, depending on the specialty and on the
degree).
A second important aspect is the distribution of civilian R&D among broadly
denned performing sectorsindustry, universities, and government. The situation in
the mid-1970s is strikingalmost 60% of all civilian R&D was being performed at
universities, while the figures for all of the other countries ranged between 8% (United
Kingdom) and 33% (Norway). Correspondingly, the share being performed in indus-
try was very low (22%) compared to the next in rank51% for Norway. This clearly
points out to a structural disequilibrium in Israel at the timestrong universities and
a weak industrial base. This imbalance has since then partially corrected itself, for
example, the share of industry in total civilian R&D rose to 43% already in 1978 (see
Table 15.IE). It probably is more than this now, although the share of R&D performed
by universities is probably still relatively high by international standards.
The growth of skills and the general strength of science and technology in Israel
suggest a paradox in Israeli growth: rapid structural change (fueled by these factors)
but declining growth performance. I will address it in two stages, first in terms of mac-
roeconomic factors and the role of defense and then in terms of inefficient policies and
institutions (particularly during the 1980s).
Explaining the israeii Paradox: Macroeconomic Factors and Defense R&D
The most important factor mentioned by neoclassical economists in explaining

declining overall performance is the extremely high and accelerating rates of inflation
experienced during 1982-1983 and 1985-1986the yearly rate reached a level of
about 400% in 1983-1984. Such levels and acceleration of inflation are usually asso-
ciated with violent changes in relative prices, increased uncertainty that affects invest-
ment decisions, increased time and effort devoted to financial management relative to
productivity growth and efficiency improvements, and so on. For an attempt at a
quantitative estimate of the effects see Bregman (1987). A related factor is government
investment policy, which led to extremely high levels of capital subsidization during
periods of rapid and increasing inflation: during most of the period after the Yom Kip-
pur War of 1973 either none or only part of the interest paid on government loans was
indexed. The resulting excessive investment led to substantial capacity underutiliza-
tion and to choice of highly capital-intensive techniques of production in industry.
The effects of these policies are analyzed in Metzer (1986), Bregman (1986), and May-
shar (1986). They are the result of an industrial policy that, if not misconceived, was
certainly not adapted to the macroeconomic disorder of the late 1970s and 1980s.
Effects of the Military Sector

The Six-Day War led to a perception that a large measure of autonomy or autarchy in
the major pieces of military hardware was required, both for military and political rea-
sons. It was also thought, in line with prevailing thinking in the United States at the
time, that the indirect effects and spinoffs of military industries to the national econ-
omy would be significant (e.g., in the area of industrial exports). Thus the perception
was that the military industries to be developed were strategic also from the economic
Table 15.8. The Military Industrial Sector" in Israel, 1984
Sector Number
1. Employees 65,000
2. Engineers and scientists 11,000-12,000
3. Technicians 14,000
In millions of US$
4. Sales 2,250
5. R&D 750
6. Export 850
"Military industrial sector includes 1AI (Israel Aircraft Industries); Rafael (the
Ministry of Defense's armaments development authorityA.D.A.); IMI (Israel
Military Industries); internal activity in the IDF, in the fields of R&D, upgrading,
and production (not including maintenance); national laboratories for nuclear
research; and firms and factories that deal in military R&D and production and
have not been included in the previous list (Tadiran, Elta, Elbit, El-Op, Soltam,
Urdan, Elisra, and some smaller firms).
Source: Hal perin( 1986).
point of view. Table 15.8 provides a quantitative definition of the military-industrial

sector of Israel while Table 15.9 sets some skill-related aspects of the sector in an indus-
try-wide perspective. This at least suggests the possibility of strong competition
between military and civilian uses of skills.
A central point to be highlighted is that high defense expenditures that resulted
both from the security problem facing the country and from the above perspective
concerning the strategic importance of military industries and military R&D have
caused both inflationary problems and the absorption of a significant fraction of the
growth of skills.9 The latter effect has dampened the growth of civilian high-tech during
the early eighties through a "crowding out" phenomenon similar to that mentioned
in connection with the science/technology system of countries such as the United
Table 15.9. Military Industry and Civilian Industry:

Competition for Critical Resources
Military-Industrial Industry a/b

Skills System (1984) (a) (b) (%)
Scientists and engineers 11,000 17,450" 63
Technicians 14,000 15,315* 91
Skill Intensity 1978 1984 1987
c
Industry Category "S"
Civilian 3.1 8.6 9.1
Military 5.6 9.3 11.0
Industry Category "H"'
Civilian 6.3 12.6 14.5
Military 9.0 17.6 19.4
"Data from Ministry of Commerce and Industry (1984).
''Data from Central Bureau of Statistics (1983).
C
"H" includes the two-digit sectors 25 and 26 rather than (254,3): (262,3); (280,81)
which compose "H." It thus includes electrical machinery (250-3) but excludes
electrooptics and precision machinery. The share of "H" in total skills for 1987 is
approximately 60%.
Kingdom and France (Ergas, 1986). I will attempt to delineate the basis for such an
argument by looking at the data of both the military industries and civilian high-tech
during the 1970-1985 period.
The 1970s were a period of fast growth of domestic sales of the military indus-
trieswhich include both hardware and R&D procurement to the Ministry of
Defence (MOD)and of military exports. Although domestic sales experienced an
acceleration during the first half of the 1970s followed by a reduction in the rate of
growth (and even absolute decline in some years), the growth in military exports was
rapid and sustained throughout the 1980s. A significantly divergent behavior of the
two occurred during the first half of the 1980s: whereas domestic sales of military
industries increased sharply, military exports experienced a sharp decline with the sub-
sequent recovery enabling the level achieved in 1980 to be attained only in 1985. One
of the main reasons for this divergence in behavior was the beginning of large military
R&D projects during the period such as that of the ill-fated Lavi fighter plane. This
implied sharp increases in domestic procurement that, together with the slump in mil-
itary export markets, explains the declining export performance of the military-indus-
trial sector.1" During the 1970s both the level of civilian high-tech exports and their
rate of growth were lower compared to that of military exports. However, civilian high-
tech exports continued to grow throughout the 1980s, despite the crisis in the sector
during 1985-1986, and briefly exceeded military high-tech exports during 1983-1985.
This is one indication of the increasing maturity of the civilian high-tech industry and
the increasing contribution to the national economy.
The relative "success" of the infant high-tech sector to achieve maturity in the
1980s was compromised by, among other factors, the sharp expansion of military
R&D during this decade.
This is for a number of reasons. First, average wages in the military high-tech sec-
tor invariably exceeded those of civilian high tech"; second, and more important, mil-
itary high-tech wage raises pace those occurring in civilian high-tech industry. Thus
the sharp increases in salaries of scientists and engineers employed in civilian high-tech
during the early 1980s were significantly influenced by increased salaries paid then in
the military high-tech sector (and despite declining military exports). These raises
partly explain the problems facing civilian high-tech during the period since at that
time many firms were planning and executing significant investments directed to pen-
etrate or expand their share in export markets. The sharp increases in the cost of
employing skilled labor blocked a more vigorous expansion of civilian high-tech
exports during the period and contributed to the crisis of the mid-1980s.12
The second, and no less important, reasons for declining economic performance
(especially during the 1980s) that we focus on relates to technological policies and
institutions. To these we now turn.
TECHNOLOGICAL POLICIES AND INSTITUTIONS
The stage is set for a more explicit discussion both of technology policy and of the
Israeli innovation system. During the late 1960s, Israel pioneered the establishment of
a mechanism for direct state subsidy of industrial R&D, one that focused on the activ-
ity performed within an industry that was almost wholly composed of small firms. In
the 1980s, however, a fundamental weakness of the Israeli innovation system has been
revealed, in particular a striking failure to establish the new technological infrastruc-
tures required by the technological revolution. This will be analyzed in the next-sec-
tion, where we will also consider national failure in creating new patterns of cooper-
ation and new institutions and networks, both for policy formulation and for
implementation.
Direct Support of Civilian R&D in Industry"
The Israeli system for the promotion of civilian technology is centered at the Office of
the Chief Scientist, Ministry of Industry and Trade (MI&T). The system was estab-
lished in the late 1960s as part of a shift in focus in national R&D policy toward applied
research and development. Most applied R&D at the time was centered in the agri-
cultural and defense areas, and the objective was to extend this activity to other areas
as well. For this purpose a number of Chief Scientist offices were established, the most
important of which was centered at the MI&T.
The objectives of the promotion system were (1) promotion of exports of high-
tech products, more specifically, exports resulting from local R&D efforts; and (2)
more broadly, promotion of structural change, in the direction of high-tech, R&D-
intensive industries. The specific goals were promotion of R&D performed directly by
industrial firms rather than by research institutes; and, related to this, stimulating the
development of innovation capabilities and high-tech entrepreneurship.14 The target
population for the newly established promotion scheme (around 1967-1968) con-
sisted, first, of small- and medium-sized enterprises (SMEs) including those that did
not yet existthe so-called new start ups (NSUs)rather than large industrial firms,
few of which were in existence at the time.
Instruments of Policy and Locus of Support
Implementation involved the allocation of grants supporting the R&D performed
directly by industrial firms. The locus of support was not government laboratories (in
fact, the relative allocation of subsidies for R&D to these laboratories declined in the
aftermath of the new scheme), nor was there an attempt at that time to support a ven-
ture capital segment of financial institutions. The approach was quite revolutionary at
the time, since the prevailing philosophy for supporting technological development in
industry was first, to support R&D performed at government laboratories, and second,
to stimulate the transfer of the technology so developed to industrial firms (for exam-
ple, the case of India, and even today, the emphasis given in Taiwan).15
Concerning the policy instrument usedgrantsit seems to me that at the time
few countries, if any, used this method. The United States and Canada, which sup-
ported R&D, favored the use of tax concessions rather than grants. Grants are more
adapted to a small firm environment, especially to those that are not currently prof-
itable.
The formulators of the policy rightly understood that the critical bottleneck to
innovation was R&D: the result was support of only this stage in the innovation pro-
cess. Preexisting schemes would supposedly provide incentives for other steps such as
physical investments and export marketing. Experience has shown that this was prob-
ably not the right design, at least in the longer run. There is a feeling today that prob-
ably too much R&D may have been done, which means that an unduly small propor-
tion was effectively applied. Part of the problem was the enormous investments
associated with the post-R&D stages, particularly export market penetration invest-
ments. These would normally exceed investments in R&D, and Israeli financial insti-
tutions seldom extended loans for these nonconventional purposes. In the last few
years, the definition of R&D has been broadened to take care of at least part of the
problem (it now includes what is called Beta-Site prototype testing). There is still, how-
ever, no integrated mechanism to support the innovation process as a whole. This is
probably one of the major weaknesses of Israeli direct support of industrial R&D.
Extent of Support and Policy Approach
The 50% rate of subsidization of R&D was considered a reasonable division of effort
and risk between government and the firm. This basic rate is even today the standard
one used in most projects.
Support was and continues to be universal (i.e., any R&D project satisfying some
minimal criteria of honesty, technical feasibility, and lately a checklist of marketing
factors could be supported). It did not matter to which industrial sector or technology
the project belonged. Moreover, the 50% rate applies to all projectsprovided they
were aiming at exportsand this assures at least formal neutrality in the R&D pro-
motion scheme.16
Quantitative Trends and the Effects of Policy
A statistical overview is found in Table 15.10. More specifically, Table 15.11 presents
information on civilian R&D expenditures in Israeli industry between 1966 and 1985.
The nominal amount in 1966 was a mere $4.6 million, and this rose to $335 million
in 1985. The real rate of growth of R&D between 1969 and 1984 was 18% per
annuma substantial growth rate. After 1986, industrial R&D declined considerably
for a number of reasons, including a reduction in government support (which accom-
panied a switch back to R&D grants to large firms after the failure of the tax concession
promotion systems introduced in the early 1980s). Parallel to the significant increase
in R&D, Table 15.11 also shows the significant increase in the ratio of civilian R&D
Table 15.10. Growth of Industrial R&D" in Israel19

1969-1970 1985-1986 Change(%)
Current expenditures on R&D (million constant 26 347 1230
1984-1 985 US$)
Current expenditure on R&D as a percentage of 0.45 2.2 390
total sales of industry (%)
Scientists and engineers* engaged in R&D
Absolute numbers 890 4300 380
Share in total industrial employment (%) 0.45 1.4 220
Practical engineers and technicians
Absolute numbers 671 3260C
Share in total industrial employment (%) 0.34 1.72C
Establishments conducting R&D 210 370r
^Civilian industrial firms only.

*Not in full-time equivalents.
Tor 1984/1985.
Source: Israel Central Bureau of Statistics, Survey of Research and Development in Industry, 1985/86. Jerusalem, 1988.
Table 15.11. Israel: Civilian R&D in Industry, 1966-1985

R&D*
(Millions of Total R&D/Sales
Year Nominal US$) Industry (%) "S"17 (%) "C" (%)
1970 14.7 0.45 1.45 0.08

1973 22.8
1974 30.5 0.54 1.59 0.09
1976 64.0
1977 99.0
1979 1.14 3.21 0.18
1984 292.0 1.91 5.29 0.22
1985 335.0
1986 460.0
"Real annual rates of growth 1969-1984 = 18%. Real reduction during 1985-
1987 period = 36%.
''R&D Industry Surveys, CBS.
'Two categories of industry are used; the "S" category includes electronics and
electrical machinery, transportation equipment, miscellaneous manufacturing,
machinery, chemicals and oil, and metal products. The "C" category includes all
others.
to industry salesfrom 0.45% in 1970 to 1.91 % in 1984. A simple industry breakdown

into the sophisticated group (which roughly covers our "H" and "S" categories) and
the conventional group of industries essentially highlights the tremendous difference
in R&D intensity between these sectorsa dual economy-type situation, possibly with
the intensity increasing much faster in the first group than in the second. Table 15.10
also shows the parallel development of technological manpower employed in civilian,
industrial R&D. Using our broad definition of skilled manpower (i.e., scientists, engi-
neers, practical engineers, and technicians) the numbers employed in R&D increased
from 1557 in 1969 to around 7000 in 1984. The number of establishments involved
in civilian R&D has not, however, increased so substantiallyfrom 210 in 1969 to
370 in 1984-1985. This reflects both birth and death of firms and an increase in the
average size of the R&D-performing establishment. Finally, Table 15.12 shows both
the growth in R&D support from the government and in R&D-related exports"out-
puts" of government support of R&D. Despite data and identification problems, there
is a strong presumption that R&D-intensive exports have increased very fastat least
during the first 15 yearsthus contributing both to the high rate of growth of indus-
trial exports and to the change in the structure of exports previously mentioned.
Evaluation oj System
The Israeli promotion system of civilian technology in industry is generally considered
to have been successful in promoting R&D expenditures in industry, in promoting the
employment of qualified manpower in industry (including absorption of immigrant
scientists and engineers) and in industrial R&D, and in promoting industrial exports.
More importantly, it was certainly one of the two or three important factors helping
to explain the profound structural transformation occurring since 1967 in the direc-
tion of high-tech and, more generally, of R&D-intensive industry. Its success during
the first decade or two derives from (1) it being a grant system, particularly adapted to
SMEs under conditions of imperfect capital markets, and (2) the universality of sup-
Table 15.12. Growth in Governmental Support of Research and Development in Industry
and in Related Exports
(Nominal) R&D Grants Exports Resulting from High Technology
to Industry from Grants for R&D Industrial Exports
Ministry of Commerce (Current US$ in (Category H) (Millions
Year arid Industry 20 Millions) 21 of 1983USS) 22
1968 1.5 3.7 44

1970 3.0 8.0 54
1972 3.7 20.8 89
1974 9.0 233.4 144
1975 10.0 289.9 176
1 976 20.0 283,6 265
1977 25.2 416.3 373
1978 27.0 550.0 4.38
1979 32.0 750.0 574
1980 643
1981 1000.0 816
1982 60.0 1400.0 811
1984-1985 922
1986 1299
port and its neutrality, which fits well with the "innovation capabilities" perspective.
The system, however, has problems derived from (1) supporting R&D rather than the
whole innovation process (which results in "excessive"even duplicateR&D and
low-rate of application of R&D results), (2) the lack of evolution of the policies, par-
ticularly the need of reducing universality (e.g., focusing on SMEs and/or high-risk
R&D only) without loss of neutrality, and (3) insufficient incentives to cooperative
R&D." We should also mention that direct support of R&D in industry is the main
direction of Israel's direct policy support of technological promotion in industry.
There is no policy for explicitly promoting technology adoption and diffusion by busi-
ness firms such as that increasingly applied in Europe.18
Technological Infrastructure and Related Institutions23
We separately consider SMEs and conventional industry, defense and high-tech indus-
try, and industry-university relationships.
SMEs and Conventional Industry
Problems of Industrial Research Institutes. Most of them belong to the Industrial
Research Authority (IRA), which depends on the Ministry of Industry and Trade
(MI&T).24 These include Israel Fibers Institute, Israel Institute of Metals, National
Physics Lab, Israel Ceramics and Silicate Institute, Institute of Plastics, and the Rubber
Research Association. Most of these institutes were founded during the 1950s and
1960s. They cater to the needs of traditional industry with "services" or consultancy
probably comprising a large fraction of their activity. Turnover has been declining
since the 1970s relative both to conventional industry turnover and to R&D expen-
ditures (see Table 15.13)(Toren, 1989).
The general impression is that these institutions are not effective in catering to the
27
Table 15.13. Share of the Budgets of Government Laboratories (G.L.) in
Total Civilian Industrial R&D (CR&D) and in CR&D Support by
Government
1976 1980 1985 198921
Budgets of G.L. (X 100)

CR&D support (%) 16.0 7.4 4.2 2.2
Total CR&D (%) 4.5 1.6 0.5
CR&D of non-R&D-intensive sectors (%) 37.8 9.6 7.2
Source: Adapted from Table 7 in Beni Toren (1989): "R&D in non-R&D intensive industrial sectors
and in Government Labs," Industrial Policy Development Group (1PDG), The Jerusalem Institute,
September.
needs of the small and medium-sized enterprises (SMEs) comprising a significant part
of conventional industry. On the one hand, we hear that the labs undertake too little
long-term research25; on the other hand, that they perform too little R&D commis-
sioned by industry (e.g., in relation to the largest lab, The Fibers Institute). This appar-
ent paradox can be solved once we recognize that the problem is not insufficient longer
term research per se, but rather insufficient research of the kind that is relevant to the
longer term needs of industry and that would enable the gradual build-up of capabil-
ities for the associated industrial sectorsmore specifically, development of an engi-
neering and technological support network for SMEs and more generally for industrial
categories "C" and "S." In the plastics industry, for example, the skills and capabilities
required, and rarely supplied, include an understanding of raw materials and of the
relationships between their characteristics and those of the final product, control of
the production process, material compounding capabilities, and design.26
Defense and High-Tech Industry
I will focus here on some structural aspects of the high-tech-oriented innovation sys-
tem, particularly that which is military oriented. The emphasis will be development
during the past 10-15 years affecting infrastructure development (mainly technolog-
ical capabilities) and patterns of firm cooperation within the military-industrial com-
plex of Israel. These aspects provide a partial explanation for the "skills crowding out"
argument previously mentioned, but also have independent effects on economic per-
formance.
Defense R&D of the late 1970s and early 1980s involved a number of projects of
significantly larger size and complexity compared to the past. Some of these involved
linking and integrating a number of scientific and technological areas such as com-
munications, computers, Optronics, software, and propulsion systems. Moreover the
ambition of developing a local version of each major piece of military hardware led
also to some attempts at developing a new "platform" (e.g., such as that for the now-
defunct Lavi fighter plane) and not only the electronics and avionics attached to exist-
ing platforms. The increased complexity and cost of these new military systems would
naturally have dictated new patterns of cooperation and division of labor at least
within the military-industrial complexboth in relation to infrastructures and in
relation to specific projects. There is clear evidence, however, that this was not the case,
at least for a wide array of technologies and projects. Rather than cooperate, defense-
related firms attempted to add new capabilities to those they historically possessed.29
The results involved (1) considerable duplication of effort in establishing new tech-
nological infrastructures (e.g., optoelectronics), (2) failed and/or delayed projects, as a
result of less than critical size of effort, (3) inadequate exploitation of export potential
due to excessive competitiona striking example being the mutual price undercutting
in the sale of "drones,"30 and (4) insufficient "market creation" in connection with the
assembly for export of complex projects and systems, the success of which depends
critically on creating a network of (long-term relationships connecting) supplier
firms.31 Since the above-mentioned failures occurred within the defense-related indus-
trial sectora sector with heavy government involvement and even ownership of
firmsthey are of a "governmental-institutional-organizational" nature rather than
traditional failures in the working of market forces. They provide, in my opinion, a
central characterization of the military-oriented innovation system of Israel (one part
of the overall national industrial innovation system), one that was shaped by the pol-
icies of the Ministry of Defense (MOD). The main causes seem to be (1) MOD's desire
to promote competition among suppliers in order to reduce the cost of procuring mil-
itary equipment, (2) absence of a clear policy assigning new technological infrastruc-
tures and capabilities to specific firms, and (3) the special treatment received by the
Armaments Development Authority (ADA, or Rafael).32'33 Thus policy focused on
advancing competition rather than in achieving an appropriate competition-cooper-
ation configuration. It also distinguished only imperfectly between specific innova-
tions and technological infrastructure.
University-Industry Linkages
Israeli universities, which developed very early an awareness of the importance of link-
ing with industry, pioneered during the 1960s the establishment of research authorities
and science parks. These fostered the commercialization of research results and the
establishment of new high-technology firms headed by entrepreneurs spun-off from
the ranks of university staff. Both should be considered at least moderately successful.34
One aspect of this success is captured by Herskovic's (1988) data on university pat-
enting (see Table 15.14). The relative "success" of Israeli universities in becoming
more industry oriented in the past 10 years does not hide the fact that there exists a
"structural imbalance" between the university system and industry (or application of
university research more generally). One possible indication of this is the share of uni-
Table 15.14.3' A Comparison of University Patenting Activity in the United States, Canada,
and Israel in 1984
United States Canadian Israeli
Universities Universities Universities
Domestic patents granted 520 21 34
Domestic patents granted to universities as a 1 1.4 1 1.0 14.7
percentage of total domestic patents
granted (%)
Domestic patents granted per million US$ 0.08 0.02 0.18
expenditure on R&D in universities in
1981(%)
Source: S. Herskovic, University Patenting Activitythe Case of Israel, Ministry of Science and Development, Jerusalem,
1988. This paper was presented at a workshop, Science Indicators: Their Use in Science Policy and Their Role in Science
Studies, Leiden University, Leiden, Netherlands, November 14-15, 1988.
versities in national R&D compared to that of other countries. Table 15.2 showed that
despite the significant decline in the past 10-15 years (and the corresponding rise of
R&D performed in industry) this share is still very high by international standards
47% in 1982/1983.
The broad-based scientific infrastructure, however, may be a significant facilitat-
ing factor in the continued movement to ever-increasing industrial sophistication,
although the contribution need not flow from university research directly. Material-
izing such potential requires new interfaces between the traditional areas of basic
research at universities and applied research in industry, such as those which emerged
in a number of OECD countries during the 1980s.35 It has been increasingly recognized
that the effective transfer to industry of basic research performed at universities
requires, prior to specific attempts at application, additional (1) "translation" into
engineering terms and (2) bundling with complementary capabilities. All of this may
also imply some need for priorities in science36: thus, the importance of generic
research and the emergence of what has been termed strategic science."
The Israeli system for financing universities is totally neutral vis-a-vis strategic
areas of research. Moreover, there is very little awareness of the need to experiment
with the emerging trends in this type of research. This derives from a number of fac-
tors. First, there is no effective separation between the financing of student education
and training on the one hand, and the financing of research, on the other.38 Second,
the system is viewed by academicians as effectively blocking outside political interfer-
ence into the university system and therefore as a bulwark of academic freedom.
Third, there is no forum of experts or of leaders of academia where strategic issues
concerning the future of the university system are discussed. PGC not only does not
attempt to establish research priorities or to identify strategic areas of science in coop-
eration with users in industry, defense, and so on, but it also does not make a serious
attempt at rationalizing the allocation of resources within the university system, pro-
moting cooperation among parties, or constructing future growth scenarios under
alternative assumptions. Fourth, the crisis of the universities is perceived as a basic
research crisis rather than as reflecting, among other things, an insufficiently developed
interfacing with other segments of society, (e.g., industry) (see Israel Academy of Sci-
ence, 1988).
We conclude that the Israeli university system, despite its successes, has failed to
adapt to the changing needs dictated both by the dynamic external environment (new
trends in cooperation and in industry-university relationships) and by internal devel-
opments (economic stagnation and massive immigration). This failure to adapt par-
allels the failure mentioned in relation to the defense and civilian industrial sectors
a general inability to develop new forms of cooperation and infrastructure both within
the sector and with outside organizations and institutions.
SUMMARY AND CONCLUSIONS
Gaps in the Israeli Innovation System

The industrial innovation system of Israel has failed to develop in the 1980s a number
of institutions, mechanisms, and infrastructures that increasingly have been found
useful in other countriesboth large and small, advanced and newly industrialized.
These include the following:
1. Branch-specific organizations to house new technological capabilities for conven-
tional industryincluding management reorientation of existing institutions
(from simple shift to demand pull to "strategic" capability development) and to
promote technology diffusion.
2. University-related Centers of Excellence in multisector functional areas of science
and technology such as biotechnology, microelectronics, optoelectronics, and new
materials. Existence of these centers would have eliminated fragmentation and
duplication of effort in that part of overall national effort that should essentially
remain curiosity oriented. They also could play a role in generating new technolog-
ical infrastructures.
3. A Technological Council to assist in the identification of desired technological
infrastructure, to coordinate the establishment of the various infrastructures (e.g.,
CIM for a particular sector with a nationwide, multisector CIM program), to
develop criteria for evaluating the economic and other contribution of cooperative
infrastructure-type R&D programs, and to represent a national locus of capabilities
in the technological infrastructure policy area.
4. A system for the support of small and medium-sized enterprises (SME), both on
technological and business matterssuch as the DTI (Department of Trade and
Industry) schemes in the United Kingdom, the Technological Service Network in
Denmark, and the ITRI network of institutions of Taiwan.
5. Secretariats and temporary joint labs in high-tech areas, such as those involved in
the Esprit program of the EEC and the microelectronics and optoelectronics pro-
grams of Japan.
6. Vision exercises and policies, at the branch, sectorial (e.g., high-tech industry), and
national levels. These are essential ingredients for effective development of new
technological capabilities.
In most cases the above institutions are critical nodes in important networks linking
firms and other agents. Their absence implies insufficient networking in science and
technology and between them and industry.
Anomalies and Distortions in Technological Effort
There is no obvious underinvestment in R&D relative to the international norm if we

look at the aggregate R&D/GNP ratios (maybe the opposite), although the situation
is less clear when we consider civilian R&D or civilian R&D performed in industry.
The major problem, however, relates to the structure of R&D in industry: too little
generic R&D or cooperative infrastructure type R&D,40 and probably excessive project
R&D (due to duplication and relative to actual incorporation of R&D results into
innovations). A second important problem is insufficient engineering (non-R&D)
effort, especially of the cooperative type required for the development of capabilities
in conventional industry. This is related to the issue of diffusion, a central aspect of
technology and technology policy of countries such as Italy (see Malerba, 1992), Ger-
many, and Switzerland (see Ergas, 1986), and other countries.
It is pretty clear that Israel is not a case of a diffusion-oriented innovation system,
and for several reasons. First, there is very little explicit policy that promotes diffu-
sionnot in connection with the technologies developed within the military indus-
tries or in connection with the adoption and diffusion of civilian (specifically,
imported) technology4'; and second, there is little actual diffusion, from the military
to civilian industries and from abroad to small- and medium-sized firms within con-
ventional industries. In addition, insufficient diffusion of technologies has character-
ized the situation within the defense industries as evidenced by excessive project and
technological infrastructure duplication. Having said this we should add that the
mechanism for promoting civilian R&D within industrial firms should, to some extent,
be regarded as a mechanism of diffusionin the sense of adapting and applying for-
eign technology (e.g., new electronic components) to numerous industrial applica-
tions. This mechanism, however, favors R&D-intensive firm rather than firms in con-
ventional industries. In addition, it is a direct-support rather than an infrastructure-
support method of diffusion. On both counts, this diffusion system is imperfect and
incomplete.
A Network Interpretation of Transition Problems

The design and implementation of the new institutions, mechanisms, and infrastruc-
tures required for the 1980s necessitate the development of new sorts of relationships
among the major players involved: government, industry, and universities. More con-
cretely, what is required is a shift from an atomistic, noncollaborative pattern that also
emphasizes control, power, and secrecy to a spontaneously interactive and collabo-
rative patternone that recognizes the pervasiveness of potential positive game situ-
ations opened up by the technological revolution. Israel has also been extremely short
of what may be termed network-entrepreneurs,42 whose main function is to create the
new networks of agents (firms, universities, individuals, etc.) required to generate these
new collaborative patterns.43
In fact, I would like to postulate that the fundamental problem has been the non-
creation of these new networks. The cause of this failure is not, first, market failure (as
could be the case in relation to component producers-assembler networks or with
innovator-user networks), since we are concerned with public/private mechanisms
and with institutions associated with technological infrastructure rather than with spe-
cific innovations. Rather, the failure is institutional-organizational-governmental.
Since network creation may have to overcome significant indivisibilities,44 and since
success in this effort may require a PBX-entrepreneur to perform the necessary coor-
dination, we may fairly safely state that a root cause of the malaise probably lies in a
political system that has not nurtured such (public sector) entrepreneurs.
Government Objectives, the Bureaucracy, and Politicians

The Israeli government has lost its capacity for long-term strategic decision making
especially on economic matters. This probably is a result of the evolution of politics
and the political system during the past 10-15 years, symbolized by the transition from
traditional dominance by the center-left Labour Party to a two-party system effectively
dominated by the center-right Likud. Such a long-term strategic capability did exist in
the past, and served the needs of the country very well in its first two decades of exis-
tence. The loss of such capability is in part conceptual, not only a problem of imple-
mentation (e.g., lack of awareness of the nature and importance of technological infra-
structure). Moreover, economic growth does not figure centrally or explicitly in the list
of objectives of the Israeli government. The main nonpolitical objectives beyond social
welfare (including housing) are employment, price stability, and others of a short-term
macroeconomic nature. Moreover, the government's (and most of academia's) think-
ing concerning economic growth is neoclassical rather than structuralist (and evolu-
tionary) with lip service adherence to the virtues of the free market implicitly preclud-
ing long-term, strategic thinking. 45 Finally, Israeli governments essentially act to what
amounts to the political economy of maintaining the existing system rather than as a
developmental state.46 This relates to the other side of the governmentits significant
powers and extensive interference at the microeconomic level, one that generates an
enormous amount of (neoclassical) inefficiencies. The sources of political power
derive, apparently, from this political and distributional micromanagement rather
than from leadership and courage in promoting change.47
Professor Y. Dror (1989) asserts that the main problems affecting Israel's capacity
to govern (including long-term decision-making) relate to (1) the nature and capabil-
ities of the bureaucracy, (2) excessive political interference in decision making (vis-a-
vis the bureaucracy), and (3) incapable ministers who cannot restrain some excesses
of the bureaucracy.48 These are very important points, which bear directly on Israel's
capacity to adopt long-term policies for the cooperative development of new capabil-
ities.
First, it is clear that Israel has not yet developed a self-conscious, well-educated
civil service, and especially one whose education has been strengthened to the extent
of understanding the importance of, and mutual relationships between, technology
and society in this day and age, and moreover, one where its long-term interests as a
group (or subgroup) can predominate over partisan, short-term benefits. We should
mention in this connection the lack of awareness and misperceptions concerning
generic R&D and infrastructure, in all ministries including Treasury and (even more
strikingly) the MI&T. Second, it is clear that coalition politics has frequently led to the
nomination of ministers lacking sufficient capabilities and that short-term political
considerations figure prominently in their agenda. The lack of implementation of the
1984 R&D Laws calling for the creation of a Technology Council within the MI&T
may be ascribed both to this factor and (possibly) to partisan opposition from other
bureaucrats within the Ministry (whose relative power on the subject matter of the
Council would presumably diminish). Third, "uneducated ministers" mean that the
political leadership is ineffective in controlling partisan bureaucratic interests. Y. Dror
mentions the lack of control of MOD over Israel defense forces (concerning profes-
sional matters only).49
The weaknesses mentioned above block both the emergence and the activity of
potential network entrepreneurs wishing to coordinate various parts of the bureau-
cracy in relation to initiatives in the area of generic R&D-technological infrastructure.
Interministerial coordination is a serious problem in Israel, and it is known that inter-
ministerial committees have frequently not led to the application of the policies agreed
on. The Treasury, despite its potential coordinating function, is a weak force in this
respect.50
Final Thoughts
The transition to cooperative technology infrastructure development programs may

require a greater measure of political leadership than that seen in the past. This
includes a number of components: taking a longer, more strategic view of the economy
and technology; a more capable political leadership and especially, one more open to
the need for change and to new approaches to experiment with; and a will to coordi-
nate the various parts of a fragmented bureaucracy. Within this context, public sector
network entrepreneurs may conceptualize and coordinate specific programs. How-
ever, even then, initial obstacles are great, due to the generalized absence of trust pre-
vailing in society as a whole.51
The result of Israel's case is that abundant skills are not sufficient for steady eco-
nomic growth, since failures of an institutional, organizational, and governmental
character may create strong obstacles in the way of successfully adapting to changes in
the external environment. In Israel, these failures are not only static neoclassical dis-
tortions but failures of a dynamic, strategic nature starting, in my opinion, from lack
of a clear vision of Israel's future in the world community. More specifically, the bril-
liant successes of Israel's entrepreneurs (market forces) to a very large extent have been
achieved despite the existence of a vast array of government policies and semipublic
mechanisms and institutions, at least in the last decade and a half.
The above is intimately related, in my opinion, to Israel's nonadaptive reaction
to the Palestinian question and to the Palestinian uprising of the past 4 years. The
mechanisms are subtle, but as with a lot of those insinuated in this chapter, they do
not seem to be less real. First, the day-to-day dealing with the uprising puts a great deal
of pressure on the government's "management time," and this enhances the already
existent short-term bias of government policy, one inimical to infrastructure and new
network creation. Second, absence of a dialogue with the Palestinians exacerbates the
noncooperative, noninteractive traits developed in the past, which also block the tran-
sition to the new approach to policy. And finally, the associated international friction,
especially with Europe, makes the whole area of international networking more diffi-
cult. This is serious for a small country searching for its place in an increasingly con-
nected world. It directly relates to technological infrastructure once one realizes that
an enormous share of capability and knowledge inputs must come from abroad
from the "external network." Therefore, its potential impact on the central message
of this chapter concerning Israel's inability to adapt is very serious indeed.
NOTES
Previous drafts of this chapter have benefited from Richard Nelson's comments and suggestions.
1. See Ministry of Science and Development, Scientific Research in Israel (1989, Ch. 16).
For a useful discussion of the attitude towards science of the prestate Zionist leadership, see Katz
and Ben David (1975, pp. 152-54).
2. The share of agriculture in GNP rose somewhat during the 1950s, probably in response
to government incentives.
3. See Syrquin( 1986).
4. The main points in what follows were taken from Kat/ and Ben David (1975).
5. For the structuralist perspective on economic growth and development see Justman and
Teubal(1987, 1989, 1991b).
6. A reduction in economic growth and in productivity growth has also characterized other
industrialized countries since the energy crisis; however, there are numerous indicators that
show that Israel's decline has been sharper and more persistent than that of many other coun-
tries. See Bruno (1989) for the brief resuscitation after 1988, followed by subsequent stagnation.
7. Total expenditure in R&D.
8. Includes mathematics and computer science.
9. The share of defense expenditure in GNP (plus import surplus) has grown from 7-9%
during 1958-1966 to almost 18-21% during 1968-1980 (Berglas, 1986). This share probably did
not decline, at least during the first half of the 1980s.
10. The Lavi plane's contribution to national security was also probably negative, the pro-
ject having been discontinued in 1987 after a long political debate and after immense invest-
ments in critical resources.
11. There are indications that this is not fully explainable by higher average skills, but
rather it has a lot to do with the management environment associated with a large public-sector
organization operating on a cost-plus basis.
12. Thus, large military R&D programs of the early 1980s may have delayed the take-off'
of civilian high tech.
13. See also Teubal (1983).
14. Development of innovation capabilities at the early stage of high-tech development is
of crucial importance, much more than at a more mature stage of the sector. It implies the desir-
ability of supporting the R&D activity of large numbers of firms with a significant probability of
failingthe reason being the very significant learning process taking place so that while direct
private profitability might be negative, total social profitability might be positive. Thus, pro-
motion of R&D is a target in itself at the early stage of high-tech industry. The usual "market
failure" arguments for supporting R&D apply even more to the generation of capabilities. Exter-
nalities are rampant, due to the significant flow of knowledge and experience going from firms
that go broke to newly established and other firms. Capability creation is a social or collective
phenomenon, rather than a process occurring exclusively within the realm of the single firm, so
the public good and indivisibility/critical mass argument are applicable here. Finally, the need
to finance costly learningan activity not existing previously, at least in the high-tech context
almost inevitably leads to capital market problems, since the required financial institutions or
mechanisms have not yet had the opportunity to develop.
15. The relative lack of emphasis on government laboratories can be explained by a num-
ber of factors: first, there were numerous opportunities for small projects whose R&D could be
carried out by small firms (e.g., attaching a minicomputer and even more so later on, a micro-
processor, to a measuring or medical instrument); second, the high quality of the engineers and
scientists involved, which also relates to the high quality of the institutions of higher learning;
third, related to the first reason, the fact that the promotion scheme did not deal with the "stra-
tegic level" of technology promotionone normally involving the development of technologi-
cal infrastructures that are inevitably "lumpy"but rather, with the current support level deal-
ing with specific, commercially oriented products or process development. At the strategic level,
many more indivisibilities will be present, so cooperative R&D with the participation of govern-
ment labs is essential (this was the case to some extent within the military sector).
16. This means that the system did not formally discriminate among projects belonging to
different sectors, branches, or technologies. 1'he fact that the share of electronics in total indus-
trial R&D and in total MI&T support for industrial R&D has increased is not an indication of
an explicit preference given io this sector vis-a-vis others. It is rather the result of a process of
natural selection of products, entrepreneurs, and firms that took place as a result of the neutral
policy environment (sec Teubal, 1983). This process expresses the main advantage of neutrality
in support.
17. Despite the world trend toward cooperation in R&D, Israel has lagged in this respect.
Although lately some incentive to undertake cooperative industrial R&D projects has been
introduced, the main thrust of the Office of the Chief Scientist is overwhelmingly in the direction
of supporting individual projects in individual firms. However, a new scheme, BIRDIsrael-
U. S. Bi-national Industrial Research and Development Foundationhas been instrumental in
promoting international cooperation in innovation, including strategic partnering between
small Israeli firms and larger foreign ones.
18. Israel's Law for the Encouragement of Capital Investment, in fact, has traditionally
favored the "hardware" over the "software" component of capital investment. It thereby implic-
itly discriminates against the diffusion of new information technologies in industry, such as com-
puter-integrated manufacturing (CIM), which possess a significant software component (see
Justman and Teubal, 1988). Discussions are underway for formulating a more rational diffusion
of technology policy.
19. From Ministry of Science and Development (1989).
20. Current US$ in millions. Central Bureau of Statistics, Industrial R&D Surveys, various
years. Grants for 1989 approximate US$ 100 million.
21. Internal information, Office of the Chief Scientist, MI&T.
22. Data Bank of the Industrial Development Policy Group (IDPG), Jerusalem Institute.
For a definition of category "H" see Table 15.1.
23. Some of the issues dealt with here were considered in Justman and Teubal (1990) and
in Brodet et al. (1990)Summary and Conclusions.
24. The exceptions include some non-high-tech institutes belonging to the Technion such
as the Building Materials Institute and the Food Industry R&D Center.
25. See the report of the Yiftach Committee (1984), which also proposes a better balance
between longer term R&D (which they identify as "infrastructure R&D") and shorter term prob-
lem-solving, consulting, and performance of R&D commissioned by industry; and also the ear-
marking of a certain proportion of the overall budget for the purpose of purchasing equipment.
26. See Yinnon and Meron (1990).
27. Government Laboratories include The Institute for Applied Research, The Fibers
Institute, The Metals Institute, The Ceramics Institute, The Institute of Rubber Research, The
Plastics Institute, and the Physics Laboratory.
28. Approved budgets.
29. An ideal example: a firm traditionally involved in communications would attempt to
add a radar capability in order to compete in the supply of certain new military systems with the
firm possessing a long-standing radar capability. The latter in turn would attempt to add a com-
munications capability in order to be able to compete in the next generation of such systems.
Critical mass problems indicate that the national economy would, in principle, have benefited
from a joint venture in system development and production and from a division of labor in rela-
tion to capabilities.
30. The competing firms were Israel's largest electronics firmTadiranand Israel's larg-
est companyIsrael Aircraft Industries (IAI). At some stage, the Ministry of Defense (MOD)
forced the firms into a joint venture that was subsequently dissolved. A "drone" is an unmanned
light reconnaissance plane developed in Israel in the early 1980s.
31. The unexploited potential in the export of military systems and turnkey projects has
been emphasized by close observers of the military-industrial scene such as U. Galil and B.
Peled. There also exists significant unexploited potential in relation to the supply of civilian sys-
tems and turnkey projects (e.g., in the areas of education, health, and agriculture).
32. MOD policies of the late 1970s and early 1980s that aimed at enhancing competition
among alternative suppliers were probably justified during the 1970s when systems were rela-
lively small and simple. However, they were inappropriate for a small country such as Israel
during the 1980s when the supply of enhanced system complexity necessitated the prior estab-
lishment of a spectrum of indivisible new R&D and production capabilities.
33. Absence of an explicitly civilian technological infrastructure and insufficient coopera-
tion and networking among firms (at least until 1988; see Teubal et al, 1989) also characterize
civilian high-tech industry. No government ministry has assumed a role in generating techno-
logical infrastructures for civilian industry, nor has there been any consideration of civilian
industry's needs (and no consultation regarding such needs) when designing technological infra-
structures for the military.
34. Spinoffs from the Weizmann Institute, for example, include a number of highly inno-
vative biotechnology firms.
35. See Mowery and Rosenberg (1992) and OECD (1984).
36. See Martin and Irvine (1984, 1989). A well-documented example is biotechnology
research in the United Kingdom organized and directed by the Biotechnology Directorate of the
Science and Engineering Research Council (SERC) (Senker, 1989). Additional examples are the
Alvey and Esprit programs on information technology for the United Kingdom and the EEC,
respectively.
37. The nature and significance of generic research were first analyzed by Nelson et al.
(1967). Further analyses of their importance for growth and innovation can be found in Nelson
(1983, 1984). Effective application of strategic science for capability development requires indus-
try relevance, multidisciplinarity, integration with engineering capabilities, excellence, and the
creation of tailor-made institutions and mechanisms (e.g., the Engineering Research Centers in
the United States. See Chemical Engineering (1989)).
38. The Planning and Grants Committee (PGC) of the Council of Higher Education per-
forms in Israel most of the role performed in the United Kingdom both by the University Grants
Committee (UGC) and the Science and Engineering Research Council (SERC).
39. Ministry of Science and Development (1988).
40. This problem has been mentioned also for the United States. See Nelson (1983, 1984)
and Nelson and Winston (1982).
41. Both Malerba (1992) and Ergas (1986) have emphasized the role in diffusion policies
played by collective institutions and even spontaneous cooperative effort. We have seen that
these have been particularly absent in Israel.
42. See Imai and Baba (1989) and Teubal et al. (1990).
43. For a survey of the various modes of interaction and collaboration of OECD countries
in information technologies, see Arnold and Guy (1987, pp. 67-68). Networksinvolving pat-
terns of long-term relationships and spontaneous interaction among many agentsare partic-
ularly required in the new collaborative forms of systemic technological development (see Imai
and Baba, 1989). They are required in order to ensure orderly patterns of division of labor and
cooperation among firms, to exploit new (unhidden) opportunities arising from spontaneous
interaction, and, more generally, as an organizational framework assuring technological system
evolution. In the cooperative development of new technological infrastructure, they are also
essential in order to establish priorities and for formulating a joint, efficient, and mutually
acceptable R&D agenda.
44. Network creation (or new network creation) may be difficult for a number of reasons:
first, the fixed costs of establishing new channels and new codes of communication among
agents, a fact that may cause existing relationships to persist despite the need for fundamental
change [see Lundvall's (1989) "stubborn" user-producer relationships and his "incomplete" net-
works case, as applied to the Danish agroindustrial complex); second, critical mass problems that
may hinder network take-off and lead to stagnation within a low-level-equilibrium trap [Teubal
et al.'s (1990) case of inventor-user networks]. I suggest that this also partly explains problems
facing the creation of the new cooperative, public/private, technological infrastructure networks
in Israel.
45. The ideology and interests of Israel's political establishment are to some extent inimical
to systematic screening of world trends, to the building of alternative economic, political, and
social scenarios based on Israel's strengths, and to open public discussion of alternative policy
trajectories.
46. For the notion of "developmental state," see Johnson (1982).
47. Prevailing economic thinking is thus right in assessing the enormous costs of govern-
ment intervention in the economy. It is wrong, however (even when abstracting from political
economy considerations), to judge that simply reducing government's role is sufficient for the
renewal of economic growth. They may be right in periods of "normal" growth but are definitely
wrong at the nodes of structural change (see Justman and Teubal, 1990).
48. See Dror (1990). This summarizes a book that also reflects, among other things, the
author's participation in, and in the conclusions of, the Kubarsky Committee (calling for the
professionalization of Israel's public service).
49. I emhasized lack of control of MOD ministers and bureaucrats over the military-indus-
trial complex.
50. A possible coordination failure involving both the Ministry of Science and Technology
and the PGC concerns the application of the policies agreed on concerning biotechnology.
51. The short-term perspective within segments of the business sector was nurtured by
years of macroeconomic instability and by some negative effects of the (generally successful)
Stabilization Program of 1986 (e.g., long periods of extremely high real interest rates). Moreover
the 1980s also witnessed the partial destruction of the "trust fund" of society (e.g., the demise of
a large number of cooperative arrangements involving mutual insurance between members such
as those within the moshav and kibbutz movements).
REFERENCES Ergas, H. (1986). "Does Technology Policy Mat-

ter?" Center for European Policy Studies.
Halperin, A. (1986). "The Military Buildup and
Arnold, E., and Guy, K. (1987). Information Tech- Economic Growth in Israel." Paper presented at
nology and Emerging Growth Areas. SPRU, the Falk Institute Conference on Economic Issues
August. and Policy in Israel, June.
Berglas, E. (1986). "Defence and the Economy." In Halperin, A., Herman, E., and Teubal, M. (1988).
Y. Ben-Porath (ed.), The Israeli Economy "The Trade-Off Between Economic Growth and
Maturing through Crisis, Ch. 8. Cambridge: Har- Military Industries." Draft,
vard University Press. Herskovic, S. (1988). "University Patenting Activ-
Bregman, A. (1986). Industry and Industrialization ityTheCaseofIsrael."Paperpresentedatwork-
Policy in Israel. Bank of Israel, Research Depart- shop on Science Indicators: Their Use in Science
ment (in Hebrew). Polity and Their Role in Science Studies. Leiden,
Bregman, A. (1987). "Technological Progress, Netherlands: University of Leiden.
Structural Changes and Productivity in Manufac- Imai, K., and Baba, Y. (1989). "Systemic Innova-
turing: The Case of Israel." Journal of Develop- tion and Cross-Border Networks." Paper pre-
ment Economics 25. sented at the International Seminar on the Con-
Brodet, D., Justman, M., and Teubal, M. (1990). An tributions of Science and Technology to Economic
Industrial and Technological Policy for Israel. Growth, at the OECD, Paris, June.
Jerusalem, Israel: The Jerusalem Institute for Israel Academy of Sciences. (1988). The Basic
Israel Studies (Hebrew). Research Fund, 13th Annual Report, 1988-89.
Bruno, M. (1989). "Economic Recuperation in His- Johnson, C. (1982). Miti and the Japanese Miracle:
torical Perspective." Typescript, May. The Growth of Industrial Policy 1925-1975. Stan-
Chemical Engineering (1989). "Engineering ford: Stanford University Press.
Research Centers: The Jury Is Still Out." May. Justman, M., and Teubal, M. (1987) "A Framework
Dror, Y. (1989). A Grand Strategy For Israel. Jeru- For an Explicit Industrial and Technological Pol-
salem's Academon, The Hebrew University stu- icy for Israel and Some Specific Policies." In C.
dents printing and publishing house. Freeman and B. Lundvall (eds.) Small Countries
Facing The Technological Revolution. London: nical Progress: Lessons from History." Journal of
Frances Pinter. Policy Analysis and Managements 2(4): 499-514.
Justman, M., and Teubal, M. (1989). "The Struc- Nelson, R. (1984). High Technology Policy: A Five-
turalist Perspective on Economic Growth and Nation Caparison. Washington D. C.: American
Development: Conceptual Foundations and Pol- Enterprise Institute of Public Policy Research.
icy Implications." Reprinted in Evenson, R., and Nelson, R., and Winston, S. (1982). An Evolutionary
Ranis, G. (1990). The Role of Science and Tech- Theory of Economic Change. Cambridge, MA:
nology in Development. Boulder, CO: Westview Harvard University Press.
Press. Nelson, R., Peck, M., and Kalachek, E. (1967).
Justman, M., and Teubal, M. (1991). "A Structur- Technology, Economic Growth and Public Policy.
alist Perspective on the Role of Technology in Washington, D. C.: Brookings Institution.
Economic Growth and Development." Type- OECD. (1984). Industry and University: New Forms
script. World Development, 19(9) 1167-1183. of Cooperation and Communication. Paris:
Justman, M. and Teubal, M.( 1990). "A Framework OECD.
for an Industrial and Technological Policy for OECD. (1988). Reviews of National Science and
Israel." In D. Brodet, M. Justman, and M. Teubal Technology PolicyDenmark. Paris: OECD.
(eds.), An Industrial and Technological Policy for Pack, H. (1971). Structural Change and Economic
Israel, Ch. 3. Jerusalem, Israel: The Jerusalem Policy in Israel. New Haven, CT: Yale University
Institute for Israel Studies (Hebrew). 79-96. Press.
Katz, S., and Ben-David, J. (1975). "Scientific Senker, J. (1989). Evaluating the Funding of Strate-
Research and Agricultural Innovation in Israel." gic ScienceSome Lessons from the British
Minerva 13(2), 151-182. Experience. SPRU, typescript, September.
Kuznetz, S. (1971). Economic Growth of Nations. Syrquin, M. (1986). "Economic Growth and Struc-
Cambridge, MA: Belknap Press of Harvard Uni- tural Change: An International Perspective."
versity Press. Reprinted in Y. Ben-Porath, (ed.), The Israeli
Lundvall, B. A. (1989). "Innovation, the Organized EconomyMaturing through Crisis. Cambridge,
Market and the Productivity Slow-Down," type- MA: Harvard University Press.
script. Teubal, M. (1982). "The R&D Performance
Malerba, F. (1990). "Italy: The National System of Through Time of an Electronics Firm."
Innovation." This volume. Reprinted in M. Teubal (ed.), Innovation Perfor-
Martin, B., and Irvine, J. (1989) Research Fore- mance, Learning, and Government Policy. Madi-
sightPriority Setting in Science, Netherlands son, WI: University of Wisconsin Press, 1987.
Ministry of Education and Science. Teubal, M. (1983). "Neutrality in Science Policy:
Martin, B., and J. Irvine (1984). Foresight in Sci- The Development of Sophisticated Industrial
encePicking the Winners. London: Pinter Pub- Technology in Israel." Minerva. Reprinted in M.
lishers. Teubal (ed.), Innovation Performance, Learning,
Mayshar, J. (1986). "Investment Patterns." In Y. and Government Policy. Madison, WI: University
Ben-Porath (ed.), The Israeli EconomyMatur- of Wisconsin Press, 1987.
ing through Crisis. Cambridge: Harvard Univer- Teubal, M. (1987). Innovation Performance, Learn-
sity Press. ing, and Government Policy. Madison, WI: Uni-
Metzer, J. (1986). "The Slowdown of Economic versity of Wisconsin Press.
Growth: A Passive Phase or the End of the Big Teubal, M. (1989). "Israel's High Tech Industry,
Spurt." In Y. Ben-Porath (ed.), The Israeli Econ- Where To?" Published in 1988. Proceedings of the
omyMaturing through Crisis. Cambridge: Har- Israel Economists' Association, December.
vard University Press. Teubal, M., Yinnon, T., and Zuscovitch, E. in press.
Ministry of Science and Development. (1989). Sci- "Networks and Market Creation." Research Pol-
entific Research in Israel, 8th ed. N. Greenwald icy, 20:381-92.
and S. Herskovic. Toren, B. (1989). "R&D in Non-R&D Intensive
Mowery, D. C, and Rosenberg, N. (1992). "The Sectors and in Government Labs," IDPG, Jeru-
U. S. National Innovation System." This volume. salem Institute, Israel
National Agency for Technology. (1983). A Tech- Yiftach Committee Report (1984). R&D, Technol-
nological Development Program. Denmark: Min- ogy, and Science-Based Industry in Israel. Report
istry of Industry. of the Committee for the Evaluation of Govern-
National Council for Research and Development. ment Research Organization and Management,
(1988). Policy for the Promotion of Biotechnology Ministry of Science and Development.
R&D in Israel. Report of the National Committee Yinnon, T.,and Meron, M. (1990). Industrial Capa-
for Biotechnology (Katzir Report). bility in the Israeli Plastics Industry. Draft
Nelson, R. (1983). "Government Support of Tech- (Hebrew).
PART IV
NATIONAL INNOVATION SYSTEMS

16
A Retrospective
RICHARD R. NELSON
It is customary for a concluding chapter to summarize the findings presented in the

text. However, here such an objective is impossible. The 15 country studies that make
up the heart of this book cover an enormous amount of material. No other project has
come remotely close to treating the range of countries considered here. Many of the
individual studies stand as major contributions in their own right to the understanding
of the innovation systems of particular countries going far beyond anything written on
that subject before. To try to summarize would be absurd.
However, it is possible to try to assess the extent to which the project has had
success in achieving the goals that were foremost in the minds of the researchers. This
concluding chapter offers such a retrospective on the project.
The principal motivations and objectives of this study were laid out in some detail
in Chapter 1. Therefore they need only brief recapitulation here. The first objective
was, simply, to describe, to compare, and to try to understand the similarities and dif-
ferences across countries in their innovation systems. We noted that until this project,
studies of national innovation systems almost always focused on one country, in
recent years commonly Japan, in earlier ones the United States, with comparisons
with other countries left mostly implicit. Thus, prior to this one, there existed no sys-
tematic comparative study covering a sizable range of countries, which probed explic-
itly for what were the common elements, and what were the apparently significant
differences, and what were the factors that seemed to lie behind these. Such a study
seemed highly worthwhile in itself.
Yet these earlier studies had spawned considerable argument about what made
an innovative system effective, in the sense that it contributed positively to national
economic performance. As we noted in Chapter 1, the originators of this project were
bothered by the tendency of various analysts to argue, on the basis of very little evi-
dence or analysis, that this feature, or that one, was a major factor behind country
performance differences, a tendency checked neither by strong conceptual under-
standing of what is and what is likely not a causal factor, nor by the requirement that
the proposed causal connections be consistent with a wide range of country observa-
tions. We hoped that significantly expanding the comparison group would constrain
such practice, and that the proposed causal arguments that survived the stronger
empirical constraints might be worthy of serious analytic consideration. This was the
second major objective of this project.
505
To a considerable extent, the arguments about the role of innovation in national

economic performance explicit or implicit in the normative studies characterized
above were drawing or rationalizing attention to a small set of industries and technol-
ogies marked by high R&D intensityspecifically semiconductors, computers and
telecommunications, aerospace, and modern biotechnologyand calling for or jus-
tifying significant government support for national firms engaged in these areas.
Although the arguments struck the project organizers as analytically loose, and empir-
ically unsubstantiated, national governments were heeding them. Thus, trade in
"high-technology" products, the behavior of companies in these industries, and in par-
ticular national policies in support of these industries had become major issues of con-
tention among nations. At the same time, companies in these industries were becom-
ing increasingly transnational. A third objective of this project was to throw some light
on this complex of issues.
From another point of view the project as a whole can be considered a test of a
set of intellectual commitments that guided the work. In particular, the basic notion
of a "national innovation system" that oriented the project is a complex and some-
what problematic one. Our concept of innovation was broad and not necessarily tied
to leadership in a technology but rather to effective competitive performance in a
dynamic context. The participants in the project felt that the key questions demanded
that the studies go beyond narrow examination of research and development, patent-
ing, and the like, and had the faith that our wider exploration could be kept analytically
compact. But this was an open question.
Also, using either the broad or the narrower concept of innovation, the partici-
pants in the project were well aware that innovation systems are not neatly divided by
national borders. The transnational aspects of technological advance, and of the key
actors involved, has grown increasingly prominent in recent years. Nonetheless, our
belief was that nations continue to be meaningful units of observation. Another objec-
tive of this study was to explore in what ways this was the case, while at the same time
considering how internationalization has affected how one ought to think about
national systems.
COUNTRY DIFFERENCES AND WHAT LIES BEHIND THEM
To compare means to identify similarities as well as differences. Certainly the broad

view of technical innovation that we laid out in Chapter 1 and that guided this study
implies certain commonalities. That view applies to economies in which profit-ori-
ented firms are the principal providers of goods and services, and where central plan-
ning and control arc weak. These conditions hold in all of the countries in our set,
although in some a certain portion of industry is nationalized, and in some govern-
ments do try to mold the shape of industrial development in at least a few economic
sectors. In all of the countries in our set, the bulk of education, including university
education, is conducted in public institutions. In all, the government is presumed to
have major responsibility for the funding of basic research, although there are major
differences across countries regarding how much of that they do, and where basic
research is mostly carried out, and in the kinds of applied research and development
A RETROSPECTIVE 507
the governments finance. From one point of view, what is most striking about the
country comparisons is the amount of basic similarity. Had the old Soviet Union been
included in the set, or China, or Nigeria, the matter would have been different. But, as
it is, the differences across our set of countries must be understood as differences of
individuals of the same species.
Within our group of countries, it would appear that to a considerable extent the
differences in the innovation systems reflect differences in economic and political cir-
cumstances and priorities. First, size and the degree of affluence matter a lot. Countries
with large affluent populations can provide a protected market for a wide range of
manufacturing industries and may engage in other activities that "small" countries
cannot pursue, at least with any chance of success, and their innovation systems will
reflect this. Low-income countries tend to differ from high-income ones in the kinds
of economic activities in which they can have comparative advantage, and in internal
demand patterns, and these differences profoundly shape the nature of technical inno-
vation that is relevant.
The 3-fold division of our countries into large high-income industrial nations,
small high-income countries, and low-income countries thus turned out to be a useful
first cut analytic separation. By and large the economies in the first group had a sig-
nificantly larger fraction of their economies in R&D-intensive industry, such as aero-
space, electronics, and chemical products, which require large sales to be economic,
than economies in the second group. There are some anomalies, at the surface at least.
Thus Sweden in the second group and Israel and Korea in the third have higher R&D
to GNP ratios than several of the countries in the first group. Some of the mystery
disappears when Israel's ambitious military R&D is recognized, and Sweden's and
Korea's strong presence in several R&D-intensive industries that live largely through
export. Both of the latter two countries also have strong defense programs and this also
undoubtedly affected their R&D intensities. There are certain interesting similarities
of countries in different groupsJapan and Korea for example. However, by and large
there were strong intragroup similarities and strong intergroup differences. Thus the
United States and Japan look much less different than advertised, once one brings
Australia and Israel into the comparison set. And much of the United States-Japan
difference can be seen to reside in differences in their resource bases and defense pol-
icies.
Whether a country had rich natural resources or ample farming land clearly is
another important variable influencing the shape of its innovation system. It turns out
that all our "small" high-income countries also were well endowed in this respect.
Among the large high-income countries the United States was by far the best endowed
here. Countries that possess resources and good farm land face a different set of oppor-
tunities and constraints than countries without these assets.
Countries that lack them must import resources and farm products, which forces
their economies toward export-oriented manufacturing, and an innovation system
that supports this. One sees this strikingly in the cases of Germany, Japan, and Korea.
On the other hand, countries with a rich resource base can support relatively high liv-
ing standards with these and the affiliated industries providing exports to pay for
imported manufactured goods. The countries that have been able to do thisDen-
mark, Canada, and Australia stand out in our sethave developed significant publicly
supported R&D programs to back these industries. So also has the United States.
Although effective agriculture and resource exploitation does require R&D, compared
with "high-tech" industry the R&D intensity here is low.
The discussion above suggests that to some extent at least, a nation's innovation
system is shaped by factors such as size and resource endowments that affect compar-
ative advantage at a basic level. But it also is true that a nation's innovation system
tends to reflect conscious decisions to develop and sustain economic strength in cer-
tain areas, that is, it builds and shapes comparative advantage.
Some of the project members were surprised to find that in many of our countries
national security concerns had been important in shaping innovation systems.
First, among high-income countries defense R&D accounts for the majority of
the differences among the countries in government funding on industrial R&D, and
the presences of large military programs thus explains why government industrial
R&D spending in the United States, and the United Kingdom, and France is so much
greater than in Japan and Germany. Second, the industries from which the military
procures tend to be R&D intensive, whether the firms are selling to the military or to
civilians. The study of Japan shows clearly that the present industrial structure was
largely put in place during an era when national security concerns were strong. This
structure, now oriented to civilian products, is one of the reasons for Japan's high
R&D intensity. It is possible that to some extent, this argument also holds for Ger-
many.
Interestingly, every one of the low-income countries in our study has been influ-
enced by national security concerns, or a military government, or both. Thus much of
high-tech industry in Israel is largely oriented toward the military. The broad eco-
nomic policies, industrial structures, and innovation systems of Korea and Taiwan
were molded in part by their need to have a capable military establishment. The pock-
ets of "high tech" atop the basically backward Brazilian and Argentine economies
clearly reflect the ambitions of their military elites.
As noted, all of the countries in our set are, basically, ones in which firms are
mostly expected to fend for themselves in markets that are, to a considerable extent,
competitive. However, all are marked by significant pockets of government overview,
funding, and protection. In our countries with big military procurement programs, the
defense industries are the largest such pocket. However, in many of our countries gov-
ernment support and protection extend into space, electric power, and telecommu-
nications, and other areas of civilian "high tech." Although these extensions are most
significant in the big high-income countries, Canada has large public programs in elec-
tric power and telecommunications, and so does Sweden.
There clearly are significant differences across the nations regarding beliefs about
the kind of role government should play in shaping industrial development. The role
of military concerns clearly is a powerful variable influencing this. But a relatively
active government also is associated with "late" development, along the lines put forth
by Alexander Gerschenkron (1962). Aside from the arena of national security and
related areas, Britain and the United States arc marked by restrained government. On
the other hand all of our low-income late-developing countries have quite active gov-
ernments. However, there certainly are exceptions to this rule. France's Etatism goes
way back in history, and although Italy is a late developer, except during the Fascist
era her government has been weak.
A RETROSPECTIVE 509
The above discussion suggests that one ought to see considerable continuity in a
nation's innovation system, at least to the extent that the basic national objectives and
conditions have a continuity. Although this proposition clearly has only limited bear-
ing on the countries in our set that were formed or gained independence only in recent
yearsIsrael, Taiwan, and Koreaeven here one can see a certain consistency within
these nation's short histories. All of these countries have experienced dramatic
improvements in living standards from the 1950s, and their industrial structure has
changed markedly. Their innovation systems have changed as well, but as our authors
tell the story, in all of these countries today's institutional structures supporting inno-
vation clearly show their origins in those of 30 years ago.
For countries with longer histories, the institutional continuity is striking, at least
to the study authors. Thus one can see many of the same things in 1990 in France,
Germany, and Japan that were there in 1890, and this despite the enormous advances
in living standards and shifts in industrial structure all have experienced, and the total
defeat of the latter two nations in World War II and the stripping away of their mili-
tary. Britain in 1990 continues many the institutional characteristics of Britain in
1890, although they seemed to work better then than now.
Indeed, it seems that of the countries with long histories the one that has changed
most institutionally is the United States. The governmental roles in funding university
research, and defense R&D, that came into place only after World War II, had little
precedent prior to the war, and profoundly changed the nature of the innovation sys-
tem.
WHAT IS REQUIRED FOR EFFECTIVE INNOVATIVE PERFORMANCE?
We have denned innovation broadly so that the term basically stands for what is
required of firms if they are to stay competitive in industries where technological
advance is important. Such industries span a large share of manufacturing, many ser-
vice sectors such as air transport, telecommunications, and medical care, and impor-
tant areas of agriculture and mining. Staying competitive means different things in
different national contexts. For firms located in high-wage countries, being competi-
tive may require having a significantly more attractive product or a better production
process than firms in low-wage countries. For the latter, being competitive may not
require being at the forefront. Indeed much of innovation in low-income countries
involves the learning of foreign technology, its diffusion, and perhaps its adaption to
local circumstances of demand or production. But in either kind of country, if tech-
nological advance in the industry is significant, staying competitive requires continu-
ing innovation.
We, the group that has produced the country studies, think we can discern several
basic features that are common to effective innovative performance, and that are lack-
ing or attenuated in countries where innovation arguably has been weak. First, the
firms in the industry were highly competent in what mattered to be competitive in
their lines of business. Generally this involved competence in product design and pro-
duction, but usually also effective overall management, ability to assess consumer
needs, links into upstream and downstream markets, and so on. In most cases signif-
icant investments lay behind these firm capabilities. All this enabled firms to master
the relevant technologies and other practices needed to compete and to stay up with
or lead with new developments.
This observation does contain a hint of tautology, but is better regarded as con-
firmation of a point stressed in Chapter 1, that the bulk of the effort in innovation
needs to be done by the firms themselves. Although they may draw on outside devel-
opments, significant internal effort and skill are needed to complement and imple-
ment these. One cannot read the studies of Japan, Germany, Italy, Korea, and Taiwan,
all arguably countries in which firms have displayed strong performance in certain
industries, without being impressed by the authors' description of the firms. On the
other hand, one is impressed the other way by the authors' commentary on the weak-
nesses of firms in certain industries in Britain, France, Australia, Argentina, and Israel.
Being strong did not necessarily mean that firms were large. Economists long
have understood that although in some industries a firm has to be large to be a capable
innovator, in other industries this is not the case. Many of the strong Italian, Tai-
wanese, and Danish firms are relatively small. Nor does it mean that the firms spend
heavily on formal R&D. In some fields such as electronics generally it did, at least for
firms in our first two groups of countries; however in Korea and Taiwan electronics
firms often were doing well with technical efforts mostly oriented toward "reverse engi-
neering." The Italian textile industry is strong on fashion and design, but little of that
work is accounted as R&D. In Chapter 1 we noted that scholars of technological inno-
vation now understand that there are significant interindustry differences in the extent
to which resources allocated to innovation, in the broad sense in which we are using
the term, get counted as R&D. Nor does it imply that the firms were not benefiting
from publicly funded R&D programs, or favored procurement status. However, as our
authors describe it, the bulk of the inputs and direction for innovative activity were
coming from the firms themselves.
Although our concept of strong firm entails ability to compete, in all of our cases
becoming strong involved actually being exposed to strong competition and being
forced to compete. As Michael Porter (1990) noted, in a number of cases the firms
faced strong rivals in their own country. Thus the Japanese auto and electronics com-
panies compete strongly with each other, and American pharmaceutical companies
and Italian clothing producers compete. However, it is not clear that this generaliza-
tion holds for small countries, where there may be only one or a few national firms as
Ericson in Sweden and Northern Telecom in Canada. For these firms most of their
competition is with foreign rivals.
Porter (1990) and Bengt-Ake Lundvall (1988) proposed that firms in industries
where a country is strong tend to have strong interactive linkages with their upstream
suppliers, who also are national firms. Our studies show many cases in which this prop-
osition is verified. The supplier networks of Japanese automobile firms, and the
upstream-downstream connections in Danish agricultural product processing, are
good examples. The cooperation of Italian textile producers with each other and with
their equipment suppliers is another. However, there arc a number of examples in
which the proposition does not seem to hold. Pharmaceutical companies, strong in
Germany and the United States, do not seem generally to have any particularly strong
supplier connections, international or national. In aircraft production, the producers
of components and subcomponents increasingly are located in countries other than
that of the system designer and assembler.
A RETROSPECTIVE 511
A similar observation is obtained regarding the proposed importance of a

demanding set of home market customers. In many cases this holds. But in small
countries or for industries that from their start have been export oriented, the main
customer discipline may come from foreign customers.
Although "strong firms" are the key, that only pushes the question back a stage.
Under what conditions do strong firms arise? As the discussion above suggests, to some
extent the answer is "spontaneously." However, our studies do indicate strongly that
aspects of the national background in which firms operate matter greatly.
One important feature distinguishing countries that were sustaining competitive
and innovative firms was education and training systems that provide these firms with
a flow of people with the requisite knowledge and skills. For industries in which uni-
versity-trained engineers and scientists were needed, this does not simply mean that
the universities provide training in these fields, but also that they consciously train
their students with an eye to industry needs. The contrast here between the United
States and Germany on the one hand, and Britain and France on the other, is quite
sharp, at least according to the authors of our studies. Indeed these studies suggest
strongly that a principal reasons why the former two countries surged ahead of the
latter two, around the turn of the century, in the science-based industries emerging
then is that the university systems of the former were much more responsive to the
training needs of industry.
Although strength in "high tech" depends on the availability of university-trained
people, industry more generally requires a supply of literate, numerically competent
people in a wide range of functions outside of R&D, who are trained to industry
demands either by the firms themselves (as in Japan) or in external training systems
linked to firms (as in several German and Swedish industries). Countries differed in
the extent to which their public education and training systems combined with private
training to provide this supply, and the differences mattered. Thus among high-
income countries Germany, Japan, and Sweden came through much stronger in this
respect than Britain and Australia. Among developing countries the contrast is equally
sharp between Korea and Taiwan on the one hand, and Brazil on the other.
The examples of Korea and Taiwan, and the other Asian "tigers," can be read as
remarkably successful cases of education led growth. As the authors indicate, the abil-
ity of firms in these countries to move quickly from the relatively simple products they
produced in the 1950s and 1960s to the much more complex and technologically
sophisticated products they produced successfully in the 1980s was made possible by
the availability of a young domestic workforce that had received the schooling neces-
sary for the new jobs. On the other hand, the cases of Argentina and Israel suggest that
the availability of an educated workforce is not enough by itself. The economic incen-
tives facing firms must be such as to compel them to mind the market and to take
advantage of the presence of a skilled work force to compete effectively with their
rivals.
Another factor that seems to differentiate countries in which firms were effec-
tively innovative from those in which they were not, is the package of fiscal, monetary,
and trade policies. Where these combined to make exporting attractive for firms, firms
have been drawn to innovate and compete. Where they have made exporting difficult
or unattractive, firms have hunkered down in their home markets, and when in trou-
ble called for protection. As I shall indicate later, in some cases at the same time firms
were competing abroad, they were working within a rather protected home market, so
the argument is not a simple one for "free trade." Rather, it is that export incentives
matter significantly because for most countries if firms do not compete on world mar-
kets they do not compete strongly. Until recently the United States possibly was an
exception to this rule. The U. S. market was large enough to support considerable com-
petition among domestic firms, which kept them on their toes and innovative. No
other country could afford the luxury of not forcing their firms to compete on world
markets. Now the United States cannot either.
Of course much of the current interest in national systems of innovation reflects
a belief that the innovative prowess of national firms is determined to a considerable
extent by government policies. Above I have identified two features of the national
environment in which firms live that seem to affect their ability and incentives to inno-
vate profoundly, and which are central responsibilities of government in all of the
countries in our sample: the education of the work force and the macroeconomic cli-
mate. But what of government policies and programs more directly targeted at tech-
nological advance? This is where much of the contemporary interest is focused. How
effective have these kinds of policies been?
In assessing this question in the light of the 15 country systems studied in this
project, one strong impression is the wide range of policies targeted at technological
advance. Thus in recent years government policies toward industrial mergers and req-
uisitions, interfirm agreements and joint ventures, and allowable industry-wide activ-
ities often have been strongly influenced by beliefs about the effects of such policies on
innovative performance. Many countries (and the E. C.) now are encouraging firms
to cooperate in R&D of various sorts. Similarly, in recent years a number of govern-
ments have worked to restructure or augment financial institutions with the goal of
fostering industrial innovation; thus several have tried to establish their analogue to
the "venture capital" market that exists in the United States. As suggested, these pol-
icies arc a very diverse lot and differ from country to country. Our case studies do pro-
vide scattered evidence on them, but, simply because they are so diverse, I cannot see
any strong generalizations that can be drawn.
Of course our country study authors were primed to look at government pro-
grams directly supporting R&D, and here 1 think the evidence collected is more sys-
tematic. It seems useful to distinguish between government programs that largely pro-
vide funds for university research or for research in government or other laboratories
not tied to particular business firms, and government programs that directly support
R&D done in firms. I consider each in turn.
Scholars of innovation now understood that in many sectors, publicly supported
research at universities and in public laboratories is an important part of the sectoral
innovation system. A substantial share of the funding of such institutions goes into
fields directly connected with technological or industrial needs, such as agronomy,
pathology, computer science, materials science, and chemical and electrical engineer-
ing.
Do our country studies support the proposition that strong research at universi-
ties or public laboratories aids a country's firms in innovation, defining that term
broadly as we have? Not surprisingly, the answer seems to differ from field to field, and
to be sensitive to the mechanisms in place to mold and facilitate interactions with
industry. All the countries that are strong and innovative in fine chemicals and phar-
A RETROSPECTIVE 513
maceuticals have strong university research in chemistry and the biomedical sciences.
A strong agriculture, and a strong farm product processing industry, is associated in
all of our cases with significant research going on relevant to these fields in national
universities, or other types of public research institutions dedicated to these industries.
In contrast Argentine agriculture is surprisingly weak, despite favorable natural
endowments. The authors of the chapter on Argentina lay the blame on Argentina's
failure to develop an adequate agricultural research system.
Where countries have strong electronics firms, for the most part there is some
strong research in university departments of electrical engineering, and this would
appear to include Japan. Government laboratories have been important sources of
new designs later taken over by firms in Taiwan. On the other hand, university
research does not seem of much importance to technical advance in automobiles and
aerospace.
Where universities or public laboratories do seem to be helping national firms,
one tends to see either direct interactions between particular firms and particular fac-
ulty members or research projects, as through consulting arrangements, or mecha-
nisms that tie university or public laboratory programs to groups of firms. Thus in the
United States agricultural experimentation stations do research of relevance to farm-
ers, and seed producers, and have close interactions with them. Various German uni-
versities have programs designed to help machinery producers. Taiwan's electronics
industry is closely linked to government laboratories. In all of these cases, the relation-
ships between the university or government labs and the industry are not appropriately
described as the universities or public laboratories simply doing research of relevance
to the industry in question. The connections were much broader and closer than that,
involving information dissemination and problem solving. They were co-partners in
a technological community. Although not important in all industries, a strong case can
be made that such technology and industry-oriented public programs have made a big
difference in many fields.
These programs do not cost a great deal, and are far less politically visible than
government programs that directly support industrial R&D. Countries differ signifi-
cantly in the extent to which the government directly funds industrial R&D. And
although most of these programs tend to be concentrated on a narrow range of "high-
tech" industries, programs of this sort vary significantly and have been put in place for
different reasons.
I noted above that in most of our countries, military R&D accounts for the largest
portion of government funding of industrial R&D. Analysts have been divided as to
whether military R&D and procurement has been a help, or a hinderance, to the com-
mercial competitiveness of national industry. Of the major industrial nations, the
United States spends by far the largest share of industrial R&D on military projects. A
strong case can be made that in the 1960s this helped the American electronics and
aircraft industries to dominate commercial markets, but that since the late 1960s there
has been little "spillover." Britain has the second largest defense R&D budget among
our set of nations, but most of the companies receiving R&D contracts have shown
little capability to crack the nonmilitary markets. The same can be said for most of the
French companies. Although until recently civilian commercial spillover seldom has
been a central objective of military R&D, except in the sense that it was recognized
that selling on civilian markets could reduce the public costs of sustaining a strong
military procurement base, it is interesting to try to understand where military R&D

did lend civilian market strength and where it did not.
Analysis of the U. S. experience suggests that civilian strength is lent when mili-
tary R&D programs are opening up a broad new generic technology, as contrasted with
focusing virtually exclusively on procuring particular new pieces of fancy hardware
wanted by the military. Increasingly, the U. S. military effort has shifted from the for-
mer to the latter. A much smaller share of military R&D now goes into research and
exploratory development than during the 1960s, and a larger share now goes into
highly specialized systems development. And the efforts of the other countries in our
set that have invested significantly in military R&DBritain, France, and Israel
have from the beginning focused largely on the latter.
Space programs and nuclear power programs have much in common with mili-
tary R&D and procurement. They tend to involve the same kind of government
agency leadership in determining what is done. They also tend to be concentrated on
large-scale systems developments. Spillover outside the field has been quite limited.
Government programs in support of company R&D in telecommunications,
other civilian electronics, and aircraft may overlap the technical fields supported by
military and space programs, and in some cases the support may go to the same com-
panies. These programs also tend to involve the same blend of industrial R&D support
and protection from foreign competition. However, there are several important dif-
ferences. One is that compared with military R&D, the public funds almost invariably
are much smaller. Indeed programs such as Eureka, Esprit, Jessi, Fifth Generation,
and Sematech are small relative to industry funding in the targeted areas. Second, the
firms themselves largely determine the way the public monies are spent, and the proj-
ects are subject to far less detailed public management and overview than are defense
projects. Third, these programs are targeted to firms and products in civilian markets,
and although their home base may be protected through import restriction or prefer-
ential procurement, the hope is that the firms ultimately will be able to stand on their
own.
Thus although they involve a commitment to high R&D spending, these pro-
grams have much in common with other "infant industry" protection programs,
many of which have grown up for reasons with no connections to national security or
a belief in the importance of "high tech," but simply because of the desire of a govern-
ment to present or create a "national" industry. Infant industry protection, subsidy,
and government guidance are policies that have been around for a very long time.
They mark French policy since Colbert. During the nineteenth century and through
World War II the United States was protectionist. The Japanese and Korean steel and
auto industries, which were highly protected up until the 1980s, are more contempo-
rary examples.
Do the infants ever grow up? Some do and some do not. The Japanese auto and
electronics companies and the Korean chaebol-based enterprises are well-known
examples of presently strong firms that grew up in a protected market, but it also
should be recognized that the American computer and semiconductor industries grew
up with their market shielded from foreign competition and with their R&D funded
to a considerable extent by the Department of Defense. After a period of such shelter
and support, these firms came to dominate the world's commercial markets. Airbus
may or may not be another successful example. On the other hand, the country studies
A RETROSPECTIVE 515
in this project give many examples of protected and subsidized industries that have
never got to the stage at which the firms can compete on their own. France's electronics
industry is a striking example, but so also are the import-substituting industries of
Argentina and Brazil.
What lies behind the differences? If I were to make a bet it is that the differences
reside in two things. First, the education and training systems that in some cases did
and in others did not provide the protected firms with the strong skills they needed to
make it on their own. Second, at least in today's world, the extent to which economic
conditions, including government policies, provide strong incentives for the firms to
quickly start trying to compete on world markets, as contrasted with hunkering down
in their protected enclave.
The picture of government policies supporting industrial innovation that I have
been presenting highlights the diversity of such policies and programs, and their gen-
erally fragmented naturesome supporting research and other activities aimed to
help industry in universities or public labs, others connected with defense or space or
nuclear power, and still others aimed directly at supporting or protecting certain indus-
tries or industry groups. This is what I conclude from the country studies of this pro-
ject. These studies play down the existence of active coherent industrial policies more
broadly. The interpretation they present of the policies of nations widely believed to
have them is closer to that of modern day infant industry protection with some R&D
subsidy than to a well-structured and thought through general policy.
Some readers will dispute this conclusion, arguing that the failure of the studies
in this project of countries well known to have active coherent industrial policies to
highlight them and their successes reflects a serious misjudgment of the authors. And
it is fair to say that it is not yet clear whether programs such as Esprit and Jessi will in
fact be successful in enabling European companies to become competitive in semi-
conductors. The authors of those studies respond by arguing that in fact government
policies in their countries are highly decentralized, and by pointing, with the case of
Airbus an exception, to the very small fraction of industry R&D accounted for by gov-
ernment programs.
The skeptics rejoin that although the policies did not involve massive public mon-
ies, they had a lot of leverage on private decisions and investments. The authors
respond that government leverage has been exaggerated and that where strong policies
have been executed, they as often lead to failure as to success. This clearly is the posi-
tion taken by our Japanese authors on MITI. Without a better understanding of tech-
nological innovation than we now have, there is no way of resolving this debate in a
way that will persuade all people.
THE DISPUTE OVER "HIGH-TECH" POLICIES
Above I stressed that the bulk of government R&D support, particularly support of
industrial R&D, goes into "high tech," a portion of it through programs expressly
designed to lend their firms a commercial edge. Where these latter programs exist, they
tend to be complemented by various forms of protection and, sometimes, export sub-
sidy. They are motivated and justified by the argument that if an economy does not
have considerable strength in "high tech" it will be disadvantaged relative to countries
that do.
But does this seem to be the case? The logic of the case and the evidence sup-
porting it is not strong.
For a firm or industry to be competitive in a high-wage country certainly requires
that it make effective use of skills, and technological and managerial sophistication,
that are not readily available in low-wage countries. The "high-tech," high-R&D
intensity industries are of this sort, but there are many others as well. The definition
of "high" tech used by statistical agencies is directly tied to R&D intensity. However,
we have stressed that an industry can be characterized by considerable innovation and
not have a high R&D intensity. If firms are relatively small, or if there is significant
design work aimed at particular customers or market niches, although considerable
innovation may be going on, the firms may not report much R&D.
Further, although national programs have tended to focus on areas such as semi-
conductors, computers, and new materials, where technical advance clearly is dra-
matic, much of the economic value created by these advances occurs downstream, in
the industries and activities that incorporate these new products into their own pro-
cesses and productsautomobiles, industrial machinery, financial services, shipping.
To do this effectively often involves significant innovation and creative innovation
here may generate major competitive advantage, but not much in the way of large-
scale formal R&D may be involved. On the other hand, it can be argued that active
government policies often can be more effective when aimed to help an industry take
advantage of new upstream technologies than when oriented toward subsidizing major
breakthroughs. A large portion of the clearly effective public programs discussed in the
various country studies of this project were or are focused on bringing an industry up
to world practice (this certainly characterizes many of the successful Japanese pro-
grams) or to spread knowledge about new developments (American agriculture and
several of the government programs in Germany and Denmark and Sweden).
Of course, the lure of "high tech" to countries that know they must be highly
innovative if they are to compete with lower wage countries is not based solely on sta-
tistical illusion. The discussion above acknowledges the special place of innovation in
semiconductors, computers, new materials, and the like in the contemporary pattern
of industrial innovation more broadly. Advances in these fields provide the building
blocks, the key opportunities, for technical innovation in a wide range of downstream
industries, from high-speed trains to cellular telephones to commercial banking. Many
observers noting this have proposed that a nation that wants its firms to be strong over
the coming years in the downstream industries had better not let foreign firms control
the key upstream technologies. This argument is prevalent in some newly developing
countries, such as Brazil, Korea, and Taiwan, as well as today's high-income ones.
Another argument seems to square the circle. It is that a nation needs to have
strength in the downstream industries to provide a market for the key component
industries. Thus, nations are supporting firms working on high definition T. V., and
telecommunications, partly on the argument that without a home market a nation's
semiconductor and computer firms will be disadvantaged. Similarly, public support
of aerospace is justified partly on alleged stimulation to upstream technology.
The authors of our country studies clearly have different, and perhaps mixed,
minds about this matter. There is a certain plaintiveness expressed in the studies of the
A RETROSPECTIVE 517
major European countries that, although doing well in some other areas, national
firms are not doing well in these critical "high-tech" fields. The authors of the studies
of Australia and Canada, on the other hand, seem to regard electronics envy as silly
and expensive fadism.
Although our country studies cannot resolve the issues, they can at least bring
attention to three matters that ought to give pause to the zealots. In the first place, there
does not seem to be strong empirical support for the proposition that national econ-
omies are broadly advantaged if their firms are especially strong in high tech and dis-
advantaged if they are not. Thus the United States continues to be strong (and a major
net exporter) in a wide range of "high-technology" R&D-intensive industries, but its
economic growth has been lagging badly for nearly 20 years. Italy has very limited
capacity in these industries, but its overall productivity and income levels have been
growing briskly for many years. One can argue that France has had broad economic
success more despite her efforts to nurture and subsidize her high-technology indus-
tries than because of them. Japan is strong in DRAMS, but also in automobile pro-
duction, which accounts for much more employment and export value, and her effi-
ciency in producing cars seems to have little to do with "high tech." And Canada,
Australia, Denmark, and the United States all continue to be strongly competitive in
industries based on agriculture or natural resources.
Also, as we have noted, the record of national policies expressly aimed to help
high-tech industries through support of industrial R&D is very uneven. Indeed, the
strongest positive examples occurred long ago, when the U. S. government provided
broad support for advances in electronics and aircraft, and the American edge here has
not proved to be durable. The European success with Airbus may or may not be a
recent positive example. Other successful cases are largely "infant industry" cases (e.g.,
Japanese electronics during the 1960s and 1970s, and Korea during the early 1980s)
where, as the companies became strong, the active and protective role of government
diminished.
And of crucial importance, firms and projects in the aircraft and electronics
industries are rapidly becoming transnational. Partly this is because of a need to share
very high up-front R&D costs, which can be met by joining with other firms. Tradi-
tional intranational rivalries tend to make firms look for foreign partners. And this
tendency, of course, is increased to the extent that governments try to keep the prod-
ucts of foreign firms out of domestic markets and to channel subsidy to national firms.
Unless the home market is very rich and the subsidies very high, firms have strong
incentives to somehow form links with other firms so that they have a chance at other
markets.
Today, there probably is no other matter that so forces one to step back, and con-
sider the contemporary meaning of a "national innovation system." To what extent
are there really "innovation systems," and to the extent that there are, in what ways
are they defined by nation states?
WHAT REMAINS NATIONAL ABOUT INNOVATION SYSTEMS?
There obviously are a number of difficulties with the concept of a "national innovation
system." In the first place, unless one defines innovation very narrowly and cuts the
institutional fabric to that narrow definition, and we did neither, it is inevitable that
analysis of innovation in a country sometimes would get drawn into discussion of
labor markets, financial systems, monetary fiscal and trade policies, and so on. One
cannot draw a line neatly around those aspects of a nation's institutional structure that
are concerned predominantly with innovation in a narrow sense excluding everything
else, and still tell a coherent story about innovation in a broad sense. Nonetheless,
most of our authors were able to tell a pretty coherent story about innovation in their
country focusing largely on institutions and mechanisms that fit the narrow definition,
with discussion of country institutions more broadly serving largely as a frame.
Second, the term suggests much more uniformity and connectedness within a
nation than is the case. Thus, one can discuss Canadian agriculture pretty indepen-
dently of Canadian telecommunications. R&D and innovation in the American phar-
maceutical industry and R&D of aircraft by American companies have little in com-
mon. And yet, one cannot read the studies of Japan, Germany, France, Korea,
Argentina, and Israel, to name just a few, without coming away with the strong feeling
that nationhood matters and has a persuasive influence. In all these cases, a distinctive
national character pervades the firms, the educational system, the law, the politics, and
the government, all of which have been shaped by a shared historical experience and
culture.
I believe that most of us would square these somewhat divergent observations as
follows. In the spirit of the discussion above, if one focuses narrowly on what we have
defined as "innovation systems" these tend to be sectorally specific. But if one broad-
ens the focus the factors that make for commonality within a country come strongly
into view, and these largely define the factors that make for commonality across sectors
within a country.
And from the start of this project we recognized that borders around nations are
porous, and increasingly so. Indeed, one of the questions that motivated this study was
whether the concept of national innovation systems made sense anymore. I suspect
that many of us come out on this as follows.
It is safe to say that there will be increasing internationalization of these aspects
of technology that are reasonably well understood scientifically. Efforts on the part of
nations, and firms, to keep new understandings won in research and development
privy increasingly will be futile. Among firms with the requisite scientific and technical
people, the competitive edge will depend on the details of design, production process,
firm strategy and organization, upstream-downstream connections, and so on.
Today, this is quite clearly the case in fields such as semiconductors, aircraft, com-
puters, and automobiles. In these fields, there are no broad technological secrets pos-
sessed by individual countries or particular firms. On the other hand, strong firms have
a good deal of firm specific know-how and capability.
It is also safe to say that differences across firms stamped into them by national
policies, histories, and cultures will diminish in importance. Partly that will be because
the world is becoming much more unified culturally, for better or for worse. Partly it
will be because firm managers and scholars of management increasingly are paying
attention to how firms in other countries arc organized and managed. And cross-coun-
try interfirm connections are likely to grow in importance. Firms in industries where
there are large up front R&D design and production engineering costs increasingly are
forging alliances with firms in other countries, to share some of the costs, and to get
A RETROSPECTIVE 519
over government-made market barriers. The establishment of branch plants in pro-

tected countries or regions is another mechanism. Thus, increasingly, the attempts of
national governments to define and support a national industry will be frustrated
because of internationalization.
What will remain of "national systems"? The firms that reside in the country will
remain, but people and governments will have to get used to dealing with plants whose
headquarters are abroad. The countries of Europe have been struggling with this mat-
ter for some time, and many of the Latin American countries, too. The United States
is now having to try to deal with this, and Japan and Korea are beginning to. As yet,
no large country seems to have made its peace with the problem, however. Although
in most countries, resident firms will be largely national, the presence of "foreign"
firms in important industries is something that nations will have to learn to cope with
better.
We noted earlier the striking continuity of a nation's basic institutions bearing on
industrial innovation. A good example is national education systems, which some-
times seem never to change in their basics. Although top level scientists and engineers
may be highly mobile, and some high level students will continue to take training
abroad, below the Ph.D. level, by and large, countries will be stuck with their nationals
who are trained at home.
The nations system of university research and public laboratories will continue
to be, largely, national, particularly the programs that are specifically keyed to advanc-
ing technology or otherwise facilitating technical progress in industry, and with built
in mechanisms for interacting with industry. These programs will have to work with
foreign branch firms as well as domestic ones in certain fields. But the notion that uni-
versities and public laboratories basically provide "public goods" and that therefore
there are no advantages to firms that have close formal links simply does not fit the
facts in many industries.
The nation's other public infrastructure, and laws, its financial institutions, its
fiscal monetary and trade policies, and its general economic ambiance still will be a
major influence on economic activity, including innovating, and these are very dura-
ble. For large high-income countries at least, the majority of private investment will
continue to be domestic, and constrained by domestic savings. And nations will con-
tinue to have their own distinctive views of the appropriate relationships between gov-
ernment and business.
And these will strongly influence a nation's policies bearing explicitly on science
and technology. From the evidence in this study, these must be understood as an
agglomeration of policies directed toward different national objectives, each with a
somewhat special domain in terms of the fields and the institutions most affected,
rather than as a coherent package.
All can hope that there will be a significant diminution of defense programs, but
military R&D will probably continue to account for most government industrial R&D
spending in the United States, France, Britain, and Israel. It is likely, however, that
there will be little commercial "spillover."
Outside of defense and space, a nation's programs of R&D support will in all like-
lihood continue to reflect both the needs of industry and broad attitudes toward what
government should be doing and how. Although there will be exceptions, particularly
when a defense connection is argued, the United States will continue to resist programs
that directly fund industrial R&D, but will use the universities as the base for a variety
of programs including some directly targeted at certain technologies and industries.
European countries are likely to make much more use of programs that directly sup-
port civil industrial R&D, either in individual firms, or in industry wide research orga-
nizations. And in Japan, France, and various other countries, government agencies
and high-tech firms will continue to be quite close.
THE DIVERSITY OF NATIONAL SYSTEMS: DO WE NEED SOME STANDARDS

REGARDING WHAT IS FAIR?
At present nations seem to be conscious as never before of their "innovation systems"

and how they differ from those of their peers. This consciousness of differences is lead-
ing in two very different directions.
On the one hand, it is leading to attempts on the part of nations to adopt aspects
of other systems that they see as lending them strength. However, the experimentation
is far from systematic, and it is highly influenced by perceptions that may have little
contact with reality. Thus the United States and the European countries (and the E.
C.) have been loosening laws that restrict interfirm R&D cooperation, and establishing
programs to encourage and subsidize it in some areas. If the chapter on Japan has got
it right, this may be somewhat ironic in view of the argument that the role in Japan's
rapid postwar growth of cooperative R&D among firms in the same line of business
probably has been exaggerated, and in any case is diminishing.
The LDCs arc looking, with good reason, to Korea and Taiwan for models. But
aside from their strong support of education, high levels of investment in plant and
equipment, and their pressure on firms to go for exports, these two countries have
quite different innovation systems. In one, Taiwan, government research laboratories
have been an important source of industrial technology; in the other, Korea, appar-
ently they have not, at least until recently. Korea has encouraged the growth of large
industrial conglomerates, and resisted foreign ownership; Taiwan has not especially
encouraged the growth of large firms and has admitted foreign firms selectively. But
both have been successful in building innovative competitive manufacturing industry
based on foreign created technologies and other low income countries are trying to
learn from their experience.
Although today attempts at emulation are at a peak, they are nothing new. The
study of Japan shows how earlier in the century the Japanese tried to pick and choose
from European and American experience, and came out with something quite differ-
ent. The Americans earlier tried to adopt the German university system, and actually
built a very different one.
At the same time, perceptions of differences are leading nations to declare certain
aspects of their rival's systems as illegitimate. Prominent Americans have expressed
the opinion that MITI support and guidance of key Japanese industries, together with
the special connections between Japanese firms and their customers and their sources
of finance, amount to an unfair system, involving subsidy and dumping as well as pro-
tection. Similar complaints have been lodged against Eureka and Airbus. The Euro-
peans complain about Japan, and about U. S. programs such as the SDI, claiming that
such large-scale government R&D support, although aimed at a military target, is sure
A RETROSPECTIVE 521
to build commercial advantages, and that that requires response on their part. The
Japanese make similar complaints, but particularly about the import barriers being
imposed by other countries. Some have gone so far as to argue that presently there is
a war between competing national innovation systems that can be resolved only if
there are new accepted standards regarding what is fair and what is not (see, e.g., Ostry,
1990). Otherwise, nations will have to adopt the norm of managed trade in high-tech-
nology products.
These two aspects of the current concern about differences in national innovation
systemsattempts at emulation and expressions of hostilityare opposite sides of the
same coin. They reflect a combination of beliefs that a nations performance in "high
tech" is vital to its broader economic performance and security, real uncertainty
regardingjust how to achieve high performance, and lack of agreed on criteria for judg-
ing what are legitimate and illegitimate government policies.
In my view, which may not be shared by all of my colleagues, the current argu-
ment seems somewhat hysterical. There is no more reason to get upset over inter-
country differences in the government's role in industrial innovation than there is to
complain about many other areas where the government's role in economic activity
differs across nations. For one thing, governments' anguish that their economies are
fated to be surely disadvantaged if they do not have a "high-tech" industry of their own
probably is unwarranted. For another, beliefs that strength in high-tech is due largely
to promotional government policies seem grossly exaggerated.
At the same time, the studies in this project show that the institutional structures
supporting technical innovation are complex and variegated. Technology and science
interact in complex ways. Both private for profit and public institutions play roles in
virtually all arenas of technological advance and the efficient division of labor is not
obvious. Arguments that private enterprise does industrial innovation and that public
institutions have little useful role in it are simple minded.
In this area it is not totally clear what one should call subsidy or protection, as
contrasted with legitimate public spending or coordination or regulation.
Economists tend to draw the line in terms of whether government spending or
regulation or guidance can be justified by market failure arguments. If so, although
public action may give advantage to a particular national industry, such support can
be argued to increase economic efficiency. If not, it is considered naked subsidy or
protection, and is not to be condoned. Thus, although international trade theorists
have long known that a nation could enhance the well-being of its own citizens vis-a-
vis those in other countries by selected naked subsidy or protection, the argument was
that under the theory then in vogue, for nations taken as a group, this was a negative
sum game.
But the problem with this line of argument is that "market failure" is ubiquitous
in the activities associated with industrial innovation, and thus subsidy or protection
or guidance could be efficiency enhancing; hence the game of active industrial policy
need not be negative sum. What has come to be called "the new trade theory" recog-
nizes some of this, nervously. If there are large "up front" R&D costs, or significant
learning through doing or using, or major externalities in certain activities such as
research and training, the simple arguments that free trade is "Pareto Optimal" (in the
parlance of economists) falls apart.
Of course "market failure" is greater in certain activities than in others. Also, gov-
eminent competence and incentives are more likely to lead to productive programs in
certain arenas than in others. Further, it is apparent that competitive protection and
subsidy among nations can get beyond any level conceivably justified on grounds of
"efficiency." It is in the interest of all nations to reign in such tendencies.
However, it seems unlikely that simple rulesfor example, that government sup-
port of R&D on public sector needs and for "basic" research is efficient and fair and
direct support of industrial R&D aimed to develop products for a civilian market is
both inefficient and unfairwill carry the discussion very far. This argument certainly
can be used to attack Airbus. But Europeans rejoin that government help was needed
to overcome the huge headstart American companies had won in large part as a spill-
over from military R&D, and can be justified economically both on infant industry
grounds, and as a policy to avoid the development of a one company world monopoly.
And what of government support for telecommunications R&D where telecommu-
nications is a government service? Americans are prone to argue that telecommuni-
cations should be privatized, but there surely is limited agreement on that. And what
to one eye is blockage to competition in public procurement to another is a valuable
close relationship between customer and supplier.
Nor are there clean lines separating "basic" from applied research. No one seems
to object to government support for research on the causes of cancer (although a break-
through here may give the firms with close contact with the research a major advantage
in coming up with a proprietary product). But what about research to advance agri-
cultural productivity? To deal improve crops growing in a particular national soil and
climate? Research on superconductivity, or on surface phenomena in semiconduc-
tors, conducted in universities? Conducted in an industry-cooperative research orga-
nization? In a particular firm?
The argument about whether government funding of certain kinds of R&D is
appropriate and efficient or unfair subsidy of course gets intertwined with arguments
about protection, and about constraints in direct foreign investments. Here countries
clearly disagree regarding what they regard as appropriate. The disagreements can be
discussed, and agreements negotiated. However, it does not seem to me that the ques-
tion of whether a protected industry is "high-tech" changes the nature of the discus-
sion, or the stakes, that much.
All this is no argument against trying to establish some norms and rules regarding
government policies bearing on industrial innovation, and in certain areas aiming for
uniform or at least comparable policies. However, it is an argument against one nation
or another getting self-righteous that its ways are efficient, fair, and quite justified, and
the policies of other nations are not. And it is an argument against the belief that agree-
ing on ground rules will be simple, if only the advice of economists is heeded.
And finally, it is an argument against trying to impose too much uniformity.
Countries differ in their traditions, ideologies, and beliefs about appropriate roles for
government, and they will guard the differences they think matter. A central reason
why this project was undertaken was, by expanding the set of countries considered,
and by trying to enable comparisons where these seemed most interesting, to try to
determine what features of national systems seemed systematically to enhance inno-
vation performance, and what features seemed useless or worse. My colleagues and 1
like to believe that we have learned a good deal. But there still is a lot of room for
informed differences of opinion. Given that there is, it is not appropriate for our group
A RETROSPECTIVE 523
or another to argue for its preferred uniformity. We still can learn from observing
diversity closely. Although (as this project testifies) it is not easy to differentiate signal
from noise, potentially we all can learn from each other about what seems to be effec-
tive and what is not.
REFERENCES tive Process: From User-Producer Interaction to

the National System of Innovation." In G. Dosi,
C. Freeman, R. Nelson, G. Silverberg, and L.
Dosi, G., Freeman, C, Nelson, R., Silverberg, G., Soete (eds.), Technical Change and Economic
and Soete, L. (1988). Technical Change and Eco- Theory. London: Pinter Publishers.
nomic Theory. London: Pinter Publishers. Ostry, S. (1990). Governments and Corporations in a
Gerschenkron, A. (1962). Economic Development in Shrinking World. New York: Council on Foreign
Historical Perspectives. Cambridge: Harvard Uni- Relations Press.
versity Press. Porter, M. (1990). The Competitive Advantage of
Lundvall, B. A. (1988). "Innovation as an Interac- Nations. New York: The Free Press.
Index
ABB, 284 Agrotechnology, Brazil, 437, 443-44

Academia Sinica (Taiwan), 402 Airbus, 59, 145, 201
Academie Royale des Sciences (France), 194 Aircraft design
Acer group, 399, 404 cumulative, incremental, 8-9
Act on the Mining and Manufacturing Industry scientific development and, 7
Technology Research Association, 87 Aircraft industry
Adams Act (1906), 37 France, 201
Advanced Micron Devices, 377 Germany, 145
AECL, 306 Korea, 379
AEG, 129, 1-43 technical advance in, 13, 14
Aerodynamics, 7 Alberta Oil Sands Technology Research Agency
Aeronautical engineering, 7 (AOSTRA), 311, 317
Aeronautics industry, Canada, 318 Alcatel NV, 220
Aronautics technology Alfa Laval, 271
Brazil, 437, 442 Aluminum-smelting industry, Japan, 89
U.S., 35 Alvey Project, 173, 184
Aerospace industry, Great Britain, 167 "American system of manufactures," 31
Aerospatiale, 218 Antitrust legislation, U.S., 29, 32, 34
Aerospace technology, Brazil, 418, 437 start-up firms and, 49
Agrarian reform, France, 196 technological development and, 62-63
Agricultural cooperatives, Denmark, 270, 276 ANVAR, 205, 211
Agricultural extension agencies/activities Applied Research Fund (Italy), 253
Argentina, 468-69 Apprenticeships
U.S., 37, 52 Germany, 121
Agricultural industry Great Britain, 180
Argentina, 462-65, 471 Argentina, 451-73
Denmark's agroindustrial complex, 269-70, agricultural industry, 462-65, 471
284, 288 farm mechanization, 464-65
Great Britain, 161 arms exports, 469
Agricultural research automobile industry, 455, 457, 459
Argentina, 465, 468-69 deregulation and economic liberalization,
Brazil, 416 458-61
Israel, 478-79 economic characteristics, international
U.S., 36-38, 51-52 comparisons, 355-56
Agricultural research stations/farms, 12 foreign direct investment, 457, 460
Argentina, 468 foreign trade, 463, 464
Brazil, 416 history and present innovation status, 451-52
Canada, 319 industrialization process, 453-61
Israel, 478-79 multinationals and, 454-58, 459, 463, 465
U.S., 37 restructuring, 458-60
525
526 INDEX
Argentina (continued) Canada, 301, 307-08

labor market, 459 Germany, 129, 131
productivity in, 456 Great Britain, 168-70
manufacturing industries, 455-57, 470-71 Japan, 97-101
R&D activities, 460-61 Korea, 365
agricultural research, 465 Avro Arrow inteceptor program (Canada),
defense R & D, 469-70 303
government-supported, 465-70, 471
industrial research, 456, 470-71 Bandeirante turboprop, 442
university research, 469 Banks and banking
technology transfer, 467 France, 196
Arianspace, 201, 204 role in British industry, 181-82
Arms industry/exports BASF, 138
Argentina and Brazil, 469 Basic research. See also Scientific research
France. 215-16 applied research vs., value in innovation
Israel. 486, 487-90 systems, 522
ASEA, 271 British expenditures, 172-73
ASEA-ATOM, 281 Canada, 305
ASICs, Taiwan, 398 Israel, 478
Assistance Program for Strategic Industries Japan, 110-11
(APSI) (Taiwan), 395 Korea, 370, 371-72
Association of German Engineers, 131 U.S., 42
Association of the Beet Sugar Industry Basic Research Promotion Law (Korea), 372
(Germany). 126 Bayer, 126, 138, 139
Atomic Energy Commission (CNEA) BCE, 305-06
(Argentina), 467-68 Beccham, 167
ATT, 32 Beet-sugar industry, Germany. 125-26
Australia, 263, 324-49 Belgium, German technological development
Civil Offsets Program, 341 and, 116-17
"colonial socialism" in, 324 Bell Canada, 313, 315, 318
economic characteristics, 324-25 Bel! Northern Research, 318
education system, 342-45 Benetton, 241
higher education, .342-44 Bercovich, Nestor A., "National Systems of
employment population changes (1983-90), Innovation Supporting Technical
348 Advance in Industry: The Case of
foreign debt, effects of, 334-35, 345-48 Argentina," 451-73
foreign trade, 329-30, 332, 333, 334 Berlin Technical Institute, 120
government assistance to industries, 335-36 Berlin University, 118
150% R & D allowance, 336 Biotechnology
government technology procurement, 340-42 Denmark, 288
immigration policy since 1945, 332 U.S.
international economic and R & D postwar development, 49
comparisons, 264 start-up firms and, 57
manufacturing industries, 329-35 university/industry research cooperation,
multinationals in, 339 54
on-thc-job training, 345 Boeing, 14, 379
Partnerships for Development, 341-42 "Brain-drain," in Taiwan, 393-94
R & D expenditures, 325-31 Brazil, 414-46
government, 327-29, 331 aeronautics research, 442
private sector, 329-31 capital goods importation, 432
research centers, 345 economic characteristics, 414-15
rural R & D, 327-29 international comparisons, 355-56
tariff protection, 335 education system, 439-42
Auto Pact (1965) (Canada), 301, 308 higher education, 441-42
Automobile industry foreign direct investment, 432-34
Argentina, 455, 457, 459 history of industrialization, 416-30
Australia, 336, 337 industrial research, 419, 425-28, 442-45
INDEX 527
"informatics" industries, 433-34 R & D activities, 302-03
international competitiveness, 420-23 agricultural experimental stations, 319
local technology development incentives government funding, 304-05, 311
(FINEP programs), 434 industrial research, 303, 305-08, 310-11
manufacturing industries, 421-28 military R & D , 311
military R & D, 418 provincial research councils, 304-05
Ministry of Science and Technology (MCT), university research, 305
434-35 relationship with U.S. economy, 301, 307-09
multinationals in, 427 research organizations, 309-11
national agricultural research program, 443- Electrical Association, 310
44 Gas Research Institute, 310
National Development Plans, 418 Steel Industry Research Association, 310
piston-manufacturing industry, 443 steel industry, 310-11
R & D expenditures, 437-39 technology diffusion, 308-09
science & technology research development, telecommunications industry, 318, 320
417-20 U.S. patents, 303, 306-07, 317
institutions, 417, 434-37 Canadair, 318
technology policies, 428-34 CANDU nuclear reactor, 303, 306
technology transfer, 429-32 Capital-goods industry, France, 224
Breguet, 200 Cement industry, Denmark, 277
"Bridging" institutions, 180 CGE-Alcatel, 219
British Aerospace, 167 CGE/Alsthom, 216, 217
British Petroleum (BP), 161 Chaebols, 362-63, 366-68, 375-80
Bull Machine Company, 204 Chemical engineering, 7
Business schools, Germany, 121 Chemicals industry
Australia, 336, 337
CAD systems, diffusion in Sweden and Canada, 317
Denmark, 284-85 Germany, 126, 131
CAE industries, 318 Great Britain, 167
Canada, 263, 299-321 Korea, 365-66
aeronautics industry, 318 scientific development and, 6
automobile industry, 301 technical advance in, 14
chemicals industry, 317 U.S., 33
economic characteristics, 300-04 Chemistry, polymer, development of, 7
education system, 301 Chesnais, Francois, "The French National
higher education, 301 System of Innovation," 192-226
electric power complex, 316 Chung-Shan Science Colleague, 406
energy industry, 310 CITEFA, 469
forestry industry, 313, 316-17 CNES, 218-19
high-technology entrepreneurship (Ottawa), CNET, 219
314-15 CNRS, 192, 202-03, 210
hydroelectric development, 313 COGEMA, 217
information technology industry, 318-19 Combustion engine, 129
innovation failures, 303 Commissariat a I'Energie (CEA), 202, 216-18
innovation linkages, 299-300, 309-19 Committee for German Industry Norms (DIN),
user-supplier relationships, 312-14 131
international economic and R & D Committee for Space Research (France), 203
comparisons, 264 Commodities industries. See also Foodstuffs/
international linkages, 300 food-processing industries; Petroleum
logging-equipment industry. 312 and oil industries
manufacturing industries, 300-01, 309 technical advance in, 14-15
metallurgy complex, 316 Commonwealth Scientific and Industrial
natural resources industries, 300 Research Organization (CSIRO)
petroleum industry, 311, 317 (Australia), 327-28, 331
pulp and paper industry Communication-equipment industry. See also
machinery, 312 Telecommunications industry
wood innovation cluster, 316 Japan, 95-97
528 INDEX
Computer industry economic concentration and multinational

Brazil. 433 capital, 266-67
Canada, 318-19 employment rates, 273, 283
Great Britain, 168 food/drink/tobacco industries, 279
Korea, 379 foreign trade, 270
postwar U.S. development, 49 industrialization. 269-70
Computer Numerical Control (CNC) international economic and R & D
machinery, diffusion in Sweden and comparisons, 264
Denmark, 284-85 labor market, 273
Computer technology/software development lack of innovation coordination, 290-91
France, 204 macrocconomic performance, 282-83
Korean chaebol activities, 375-77 manufacturing industries, 279
Taiwan. 400-01 process of technical change, 275
Computing Devices of Canada, 315 production system, 275
Concorde. 201 small firms in, 289
Construction industry, Denmark, 277 socio-political aspects, 272-75
Cooperative research. See Research structural problems in the economy, 288-89
collaboration Sweden and. comparisons, 265-69
Cooperative Research Centre Scheme technological contributions, 283-84
(Australia), 345 Technology Development Programme, 280
Corporatism technology diffusion, 285-87
Denmark. 275 technology policy. 280-82
Great Britain, 182 U.S. patents, 266, 283
Sweden. 273 Design, in process of technical advance. 8
Council of Scientific and Industrial Research Deutsche Forschungsgeneinschaft. 132, 133,
(CSIR) (Australia), 328 141
Counterpoint Computers, 404 Development blocks
CPqD, 436-37 definition, 275-76
Craftsmenship, Great Britain, 179-80 Sweden and Denmark, 275-78
Cream separator, development of, 269 Digital Equipment Corporation (DEC)
CRITTs, 211 in Canada, 318
in Great Britain, 164
Direct foreign investment (DFI). See Foreign
Daeduck Science Town, 380-81 investment
Daewoo, 375, 376, 377, 378, 379 Direction Generale des Telecommunications
Dahlman, Carl .J., "National Systems (DGT), 219
Supporting Technical Advance in Directorate for Military Research and Testing
Industry: The Brazilian Experience," (DRME) (France). 204
414-46 Delegation Generale a I'Armament (DGA)
Daihatsu. 98, 100 (France), 214. 215-16
Daimler-Benz, 138. 145 DMR, 318
Dairy industry, Denmark, 269-70, 276-77 DRAM technology
Data-processing development. See also Korea, 377
Computer industry Taiwan, 398-400
France, 204 Du Pont, 32, 50, 62
"Datsun," 98, 100 Dwoty and Smiths Industries, 167
Defence Industry Productivity Program DYNA, 404
(Canada), 311
Defense electronics Eastman Kodak, 32, 50
Great Britain, 167 Ecole Normale Superieure, 198
Defense industry, France. 215-16 Ecole Polytechique, 198, 213
Defense Production Act (U.S.), 46 Edison effect, 8
Defense R & D. See Military R & D Edquist, Charles, "Comparing the Danish and
DeHavilland, 318 Swedish Systems of innovation," 265-92
Denmark, 265-92 Education systems. See also Engineering
agroindustrial complex. 269-70, 284, 288 education/training
development blocks. 276-77 Australia, 342-45
INDEX 529
Brazil, 439-42 "Enterprise," concept in Great Britain, 184-85

Canada, 301 Entrepreneurship
competitive innovation systems and, 511 Canada, 314-15
France, 197-99 Great Britain, 185
Germany, 147 Japan, 98, 101-02
19th cent., 117-23 Israel, 495
current status, 139-41 Equipment-manufacturing industry, Italy, 237-
Great Britain, 178-80 39. See a/so Electrical-equipment
international comparisons, 140, 440 industry
Japan, 78-80 Ericsson, 271
Korea, 358-59 ESPRIT, 254
Taiwan, 385, 392-94 EUCLID, 186
U.S., 48 European Community
Electric power complex, in Canada, 316 British research cooperation and, 174
Electric Research Institution (Japan), 83 defense R & D program (EUCLID), 186
Electrical-equipment industries German exports and, 136
Australia, 336, 337 innovation, Germany, and, 147
Japan, 93-95 Single European Act, 188
scientific development and, 6 European Space Agency, 201, 204, 255
Electronic Research and Service Organization European system of innovation, 16
(ERSO), 397 Export/import performance
Electronics industry. See also Microelectronics Argentina, 455, 456-57, 459, 463, 464, 469
industry Australia, 329, 332, 333, 334
Brazil, 433 ' Brazil, 422-24, 425, 426-27
Denmark, 277, 288 British (1978-88), 164-65
Great Britain, 168, 174, 175-76 Denmark. 270
Italy, 256-57 France, 222, 224-26
Korea, 365, 374, 378 Germany, 115, 128, 129-30, 133-37
Scotland, 187 large, high-income countries, 25
Taiwan, 397-400 promotion in Korea, 363
Electronics switching system (ESS), Sweden, 271-72, 278
development in Korea, 374, 378 Exxon, Canadian innovation system and, 318
Electronuclear industry, France, 216-18
Electrotechnical industry, Germany, 127-28, Fair Trade Act (Korea), 367
131 Farm mechanization, Argentina, 464-65
Embraer, 442 Farman, 200
EMBRAPA, 437, 443-44 Federal Ministry for Research and Technology
Emergency Association of German Science, (Germany), 142, 143, 146
131-32 Federally funded research and development
Employment, long-term, in Japan, 106, 107 centers (FFRDCs), 42, 47
Encyclopedia, 194, 198 FIAT, 249
ENEA, 254 Fibers Institute (Italy), 478, 490
Energy industries Filter technology, Denmark, 283
Canadian research institutes, 310 FINAME, 419
Great Britain, 161 FINEP, 419, 435
Japan, 88-89 Firms. See Industrial research; Large firms;
Engineering design, technical advance and, 8 Multinationals; Small firms; Small- and
Engineering education/training medium-sized firms (SMEs)
France, 197-98, 211, 213-14 Flexible manufacturing systems (FMS),
Germany, 119-21, 123 diffusion in Sweden and Denmark, 284-
Great Britain, 178-79 85
Japan, 79-80 Foodstuffs/food-processing industries
Taiwan, 393 Argentina, 453
U.S., 36 Denmark, 276-77, 279, 288
Engineering industries Great Britain, 161, 168
Denmark, 277, 284-86 Ford Company
Sweden, 271, 277-78, 284, 289-90 British auto industry and, 168
530 INDEX
Ford Company (continued) telecommunications industry, 219-20

German factory, 131 U.S. patents, 221-22, 223
Japanese auto industry and, 97, 99 vital statistics, 25-27
Fordism, U.S., 31 France Telecom, 219, 220
Foreign investment/foreign direct investment "Free market," in Great Britain, 182, 183
(FDI) Frischtak, Claudio R., "National Systems
Argentina, 457, 460 Supporting Technical Advance in
Brazil, 417. 432 Industry: The Brazilian Experience,"
Great Britain, 168-70. 174-75 414-46
Japan. 86, 102 Fuji Electric, 94
Korea, 360-62 FUNTEC, 419
Sweden and Denmark, 266-67 Furukawa Mine, 94
Taiwan, 388-91, 405-06
Foreign licensing (FI,), Korea, 360. 365 (iakushin, 84
Foreign trade. See also Export/import GEC, 163, 167
performance Gee, San. "National Systems Supporting
Argentina, 454, 463, 464 Technical Advance in Industry: The Case
Australia, 329-30, 332, 333, 334 of Taiwan." 384-412
competitive innovation systems and, 511-12 General Electric, 32, 50
Denmark, 270 Japanese electrical equipment industry and,
France, 222, 224-26 94-95
Germany (1913), 128, 129-30, 133-37 General Motors (GM)
Great Britain, 164-65. 168-70 British auto industry and, 168
Sweden, 271-72 Japanese auto industry and, 97, 99
technological competitiveness and, 60-61 Germany, 115-47
Forestry industry Airbus program, 145
Canada, 313, 316-17 beet-sugar industry, 125-26
Framatome. 216, 217 education system, 139-41
France, 192-226 university system, 117-20, 140-41, 147
aeronautics industry, 199-200 vocational training, 139-40
automobile industry, 199-200 engineering profession in, 179
data-processing industry, 204 European Community and, 136
defense industry, 215-16 Federal Ministry for Research and
financial system development, 196 Technology, 142, 143, 146
foreign trade performance, 222, 224-26 foreign trade, 128, 129-30
government's role in technological export performance, 115, 133-37
development, 192-93 future of innovation system, 146
"Grandes Kcoles," 197, 198, 2 1 1 , 213-14 government funding of research, 137, 138,
high-technology sector, 214-20 141, 142-45
higher education, 197-98 government laboratories in, 13
industrialization & economic development, history of industrialization, 116-33
195-96, 200-01 industrial research
large firms (nationalized enterprises) in, 193 19th century, 125-29
military-industrial complex development, 204 20th century, 124, 131-33, 138-39
R & D structure, 205-11 Japanese industry and trade, effects, 137
government funding, 206-08 military R & D, 132
industrial research, 208-09 research organizations, 123-24, 131-32
military R & D, 203-04, 215-16 science parks, 145
research laboratories, 13, 205 technology policy, 142-45
science & technology institution building, U.S. patents (1975-85), 137, 139
201-03 university research in, 12
scientific research, 194-95 vital statistics, 25-27
university research, 198-99, 210-11 G.I. Bill (U.S.), 48
revealed technological advantage (RTA). 221 Glass & rubber industries, U.S.. 33, 34
space industry. 218-19 Glaxo, 167
state/industry alliance, 212-13 Globalization (concept), 63
technology information dissemination, 211 Japan and, 110
INDEX 531
Goto, Akira, "The Japanese System of Grants for Industrial Research and
Innovation: Past, Present, and Future," Development Scheme (GIRD)
76-111 (Australia), 336
Gottingen University, 117 Great Britain, 158-89
Government funding/support. See also Military Advisory Committee on Science and
funding Technology (ACOST), 184
Argentine R & D, 465-70, 471 agricultural sector, 161
Australian R & D, 327-29, 331, 335-40, 345 coordination problems, 180-82
Brazilian R & D, 434-39 craftsmen and apprenticeships in, 179-80
British R & D , 172-74, 182-87 economic decline, 158-60
Canadian R & D , 304, 311 economic performance (1980s), 160
Danish R & D , 280 economic structure, 160-65
French R & D , 206-08, 212-13 education system, 178-80
Germany higher education, 178-79
higher education and science, 124-25 electronics industry, 174, 175-76
R & D , 117, 137, 138, 141, 142-45 engineering profession in, 179
specialized research organizations, 123-24 "enterprise" concept in, 184-85, 186
universities (19th cent.), 119 European Community and, 188
Israeli industrial research, 487-90 foodstuffs industry, 161
Italy foreign investment (multinationals), 164, 166,
R & D , 244 168-70, 174-75, 176, 188
small-firm networks, 241-42 foreign trade, 164-65, 168-169
Japan "free market" in, 182, 183
auto industry, 100-101 German technological development and,
electrical & communication equipment 116-17
industries, 96 high-technology industries, 165-66
industrial plants in Meiji, 80 innovation system weaknesses, 176-82
industrial research, 87-89 "Levene" reforms, 177
Korea manufacturing industries, 163, 165-70
education system, 358 petroleum industry, 161
R & D , 370-74 political structure, 182
R & D , fairness questions, 522-23 R & D activities
Swedish R & D , 279 government funding, 172-74, 183-87
Taiwanese industrial R & D , 391-92, 394- industrial research, 170-72, 174-75
400, 402-03 military R & D , 172-73, 177-78
U.S. R & D , 30 university research, 12
agricultural research, 37, 51-52 R & D expenditures (1970s; 1980s), 172-76
during World War II, 39-40 regional development agencies, 186
postwar, 40-48 revealed technological advantage (RTA)
prior to 1945, 34-36 (1978-86; 1988), 168-69, 176
university research, 47-48 science parks, 186-87
Government policies, for technology service industries, 162-64
development U.S. patents, 175
Australian high-technology, 345-48 vital statistics, 25-27
effects of, 512-15 Gregory, Robert G., "The Australian
Italy, 251-55 Innovation System," 324-39
Japan's industrial development, 103 Gross national expenditures and production
U.S.,58-61 Denmark and Sweden, 282-83
Government research laboratories, 12-13 Japan (1875-1940), 82
Germany, 144 Gruppo Finanziario Tessile (GFT), 241
Government's role, in national innovation
systems, 508, 512-13 Hakunetsusha, 93, 94
GTP, 168 Halle University, 117
"Grandes Ecoles," 197, 199, 2 1 1 , 213-14 Hatch Act (1887), 37
Ecole Polytechnique, 198, 213 Heavy industries
technical experts, 192 Brazil, 417-18
Grands corps, 213, 214, 218 Japan, 84
532 INDEX
Heavy oil extraction, Canada, 311, 317 industrial Base Technology Development
Hebrew University, 479 Projects (IBTDP) (Korea), 372-74
High-definition television (HDTV), U.S. industrial Parks, Taiwan, 405
initiatives, 59 Industrial policies. Sec also Government
High-technology sector/industries policies; Technology policies
Australia, 345-48 Japan, 102-03
France, 214-20 Industrial research, 10-11
Great Britain, 165-66 Argentina, 456, 470-71
Israel, 479-82, 491-92 Australia, 329-31, 336-40
Japan, 89 Brazil, 419, 425-28, 442-45
Korea, 375-79 British productivity and, 170-72
national innovation systems and, 515-17, 521 Canada, 303, 305-08, 310-11
U.S. development, 49 expenditures, international comparisons, 175
High-temperature superconductivity (HTS), 58 France, 208-09
Higher education. See also University research Germany, 124, 125-29, 138-39, 144
Australia, 342-43 government support and, 13
Brazil, 441-42 Great Britain, foreign investment in, 174-75
Canada, 301 Israel, 484
France, 197-99 Italy, 231-32, 247-50
Germany, 147 Japan, 84-85, 87-89, 105-09
19th century, 117-19, 124-25 Korea, 364, 370-71, 372-74, 375-80
business schools, 121 Sweden and Denmark, 278-79
current system, 140-41 Taiwan, 394-96
Great Britain, 178-79 U.S., 30
Japan, 79 antitrust policy and, 34
Taiwan, 392, 393 before 1945, 31-39
U.S., 35-36, 48 post-World War II, 39-52
Hino, 100 decline in 1970s, 50-51
Hitachi. 94, 95 1980s, 53-56
Hoechsl, 138, 139 industrial Research Institution (Japan), 83
Honda, 100 industrial research laboratories, 10-11
Hou, Chi-ming, "National Systems Supporting Germany, 144
Technical Advance in Industry: The Case Japan, 84-85
of Taiwan," 384-412 U.S., 33-34, 50
Hsinchu Science-based Industrial Park, 405 Industrial Upgrading Statute (1US) (Taiwan),
Human resources development. See Education 395
systems Infant-industry protection
Hydro-Quebec, 313 Japan, 99-100
Hydroelectric development, Canada, 313 national innovation systems and, 514-15
Hyundai, 375, 376, 377, 378 INFN, 254
Information technology industry
IBM, 14 Australia, 341-42
Canada, 306, 318 Canada, 318
Great Britain, 164 Taiwan, 396-400
PC clones, Korea and, 379 "Innovation," definition, 4
Taiwan, 400-01 "Innovation clustering," 316
IC chip technology, Taiwan, 397-400 Innovation systems. See National innovation
ICI, 167 systems; and names of individual
ID Focus, 375 countries, e.g. Korea, United States, etc.
IG Farben. 131 INPE, 437
IHI, 98 Institute for Chemical Science and Technology
Immigration policies, Australia, 332 (Canada), 317
Imperial Institute of Physics and Technology Institute for Research into Data-Processing and
(Germany), 123-24 Automation (IRIA) (France), 204
Imperial Oil, 312, 318 Institute for the Information industry (III)
Incubators. 314 (Taiwan), 396, 400-01
INDEX 533
Institute of Physical and Chemical Research university research, 484, 492-93
(Riken) (Japan), 83-84 university-industrial linkages, 492-93
Integrated Systems Development Corp (ISDC), ISS, 254
401 Isuzu, 98, 99, 100
Intel, 377 Italy, 230-57
Intellectual property rights electronics industry, 256-57
Brazil, 429 high-technology sector, 250, 251
Taiwan, 406 industrialization, 231-34
U.S. inter-institution research cooperation, 254
agricultural research, 51 North-South industrial differences, 232-33
protection initiatives, 59 public enterprises, 233-34
trade policies and, 60-61 public policy and technological development,
International competitiveness 251-55
Brazil, 420-23 public research network, 246-47
national innovative performance and, 509 R & D activities, 231-34
International research cooperation European research programs, 254-55
Great Britain, 174-75 government-funded research, 241-42, 244
Italy, 251, 254-55 industrial research, 247-50
U.S., 55, 56 large-firm R & D, 230, 231, 232-33, 235,
International Data-Processing Company (CII), 243-55
204 small-firm networks, 230-31, 232, 234-
International Integrated System Inc. (IIS), 401 43
Internationalization, 17-18 university research, 245-46
Germany and, 147 R & D expenditures, 243-44, 245
Italy, 234, 251 Sabatini Law, 241-42
national innovation systems and, 518-19 science parks, 242
Sweden and, 291-92 technological information dissemination,
IPT, 437 236, 242-43
Iron and steel industries technology diffusion, 255-56
Canadian R & D, 310-11 textile industry, 235, 236, 241
development in Japan, 90-92 U.S. patents, 237-38, 244, 245
France, 196 virtuous and vicious innovative cycles, 255-
Germany, origins, 127 57
Sweden, 271 vital statistics, 25-27
Israel, 476-97
Academy of Sciences, 478
defense R & D, 483, 484-86, 491-92 Japan, 76-111
early scientific activities, 478-79 automobile industry, 97-101
economic characteristics, international basic research in, 110-11
comparisons, 355-56 education system
economic development (1950s; 1960s), 477- early schools, 78-79
79 higher education, 79
entrepreneurship in, 495 public education, 78-80
high-tech industries, 479-82, 491-92 electrical & communication equipment
industrial research, 484, 487-90 industries, 93-97
innovation system gaps, 493-94 energy-saving R & D, 88-89
manufacturing industries, 481-82 External Trade Organization (JETRO), 103
Ministry of Industry & Trade, 487, 490 foreign investment in Great Britain, 188
research networking, 495 globalization of research and, 110
scientific & engineering personnel, 479-80, industrial policy, 102-03
481-82, 483, 489 industrial research
small- and medium-sized enterprises (SMEs), current status, 105-09
490-91 government support, 87-89
technology diffusion, 494-95 managers and, 106-07
total factor productivity, 481 new technology introduction, 108
U.S. patents, 492-93 since 1960s, 87-89
534 INDEX
Japan (continued) human resources development, 358-60
industrialization history education system, 358-59
Tokugawa Era and before (to 1868), 77-79, Institute of Science and Technology (KIST),
90 364
Meiji Era (1868-1911), 79-81 Intel-organizational research cooperation, 374
1914-1945.83-85 labor movement, 369
postwar era, 1945-1970s, 85-87 multinationals in, 376
since 1960s, 87-89 national defense and heavy-machinery
iron and steel industries, 90-92 industries, 362, 380
national model of innovation, 17 promotion of big business, 362-63
National Railways, 103 R & D activities, 369-75
national research institutions, 83-85 government support, 370-74
patents, 104-05 industrial research, 364, 370-71, 372-74,
in U.S., 104-05, 137 375-80
R & D expenditures (1989), 103-04 university research, 371, 374
total factor productivity (1960s-80s), 105 small- and medium-sized firms promotion,
Unequal Treaties, 80-8 1 368
university research, 110-11 "strategic" industry development, 362-63,
vital statistics, 25-27 369
JESSI project, 147 technological capability, large firm vs. small
Jitsuyo Jidosha, 98 firm behavior, 364-67
technology development policies, 367-69
Kachalsky committee, 478 tradition of overseas training and observation,
Kaiser-Wilhelm-Society, 124, 133, 141 359-60
government funding, 132 U.S. military presence and, 361-62
research institutes, 131 working habits, 360
Kaishinsha, 97, 98 Korean Air, 379
Kamaishi plant, 90-91 Korsil, 377
Karakuri masters, 78
Katz, Jorge M., "National Systems on Labor market, Sweden and Denmark, 273, 274
Innovation Supporting Technical Labor productivity, Taiwan, 385
Advance in Industry: The Case of Labor relations
Argentina," 451-73 Japan. 107, 108, 109-10
Kawasaki, 91, 92 Korea, 369
Japanese auto industry and, 98 Large firms
Keck, Otto, "The National System for France, 193
Technical Innovation in Germany," Italy, 230, 231, 232-33, 235, 243-55
115-47 Korea, 364-67
Kim, Linsu, "National System of Industrial Laser industry, FIAT and, 255-56
Innovation: Dymnamics of Capability LM-Tel, 281
building in Korea." 357-82 Logging-equipment industry, Canada, 312
Korea, 357-82 Louis Pasteur University, 210
aircraft industry, 379 Lucas Aerospace, 167
automobile industry, 365 Lucky-Goldstar, 375, 376, 377, 378
chaebols, 362-6.3, .366-68, 375-80 Luis de Queiroz Agricultural College, 416
chemicals industry, 365-66
economic characteristics, 357 Machine-construction industry, Germany, 127,
international comparisons, 355-56 129
economic liberalization progarm, 367-69 MacMillan Blodel, 312
effects of civil war (1945-53), 358 Malerba, Franco. "The National System of
electronics industry, 365 Innovation: Italy," 230-57
electronics switching system (ESS), .374. Management
378 Australia, 335-36
export promotion, 363-64 British, industrial innovation and, 181
foreign technology transfer, 360-62 France, 213-14
high-technology transfer, 375-79 role in Japanese firms, 106-07
high-tech "valleys." 380-81 Manhattan Project, 39
INDEX 535
Manufacturing industries Sweden and Denmark, 281
Argentina, 454-58, 470 Taiwan, 406
Australia, 324, 325, 329-35 U.S., .30, 37, 42-43, 46-47, 49-50, 514
Brazil, 421-28 Military/civilian technology spillovers, U.S., 43,
Canada, 300-01, 309 46-47, 49-50, 57, 58-59
Denmark, 270, 279 Mining industries
France, 194 Germany, origins, 127
Germany, 128, 129-30, 136, 137 Sweden, 271
Great Britain, 163, 165-72 Ministerial Delegation for Armaments (DMA)
Israel, 481-82 (France), 204
Italy, 232, 234-43 Ministry for Scientific and Technological
Japan, 86 Research (MRST) (Italy), National
R & D in, international comparisons, 329-30 Research Programs (NRPs), 252-53
U.S., 31-32, 33-34 Minitel videotex, 219
Marks & Spencer, 181 Miroglio, 241
Mass production, in U.S., 31 Mitac, 404-05
Massachusetts Institute of Technology (MIT), MITI (Japan)
40 information collection and diffusion, 103
Industrial Liaison Program, 54 Japanese auto industry and, 98, 100
Matra, 218 support of Japanese RAs and, 88
Max-Planck-Society, 133, 141 Mitsubishi, 84, 91, 94, 98
Mazda, 98 Mitsui, 98, 99
McFetridge, Donald G., "The Canadian System Modern Electrosystems, 375
of Industrial Innovation," 299-321 Moonlight Project (Japan), 89
Meikosha, 95 Morrill Act (1862), 37
Merck Company, 126 Motor-vehicle industry. See Automobile
Mergers industry
Taiwan, 403-05 Mouse System Company, 404
U.S. (1895-1904), 32 Mowery, David C, "The U.S. National
Metal Leve, 443 Innovation System," 29-64
Metal-processing industry Multinationals
Germany, origins, 127 Argentina, 454-58, 459, 463, 465
Sweden, 271 Australia, 339
Metallurgy complex, in Canada, 316, 317 Brazil, 427
Metalworking industry, Argentina, 460, 461 Great Britain, 164, 166, 168-70, 174-75,
Michelin, 200 176
Microelectronics industry Korea, 360, 361, 376
Brazil, 43.3 Sweden, 267, 281, 291, 292
U.S., 49, 57, 59
Micron Technologies, 377 National Advisory Committee on Aeronautics
Microsystems International, 303, 315 (NACA) (U.S.), 35
Microtex, 404 National Aeronautics and Space Administration
Midi-Robots, 205 (NASA), 35
Military funding National Center for Manufacturing Sciences
Japanese auto industry, 97-99 (NCMS) (U.S.), 58-59
U.S. commercial technologies, 58-59 National Centre for Scientific Research (CNRS)
Military R & D (France), 192, 202-03, 210
Argentina, 469-70 National Centre for Space Studies (CNES)
Brazil, 418 (France), 203
Canada, 311 National Centre for the Study of
effects on national innovation systems, 508, Telecommunications (CNET) (France),
513-14 202
France, 203-04, 215-16 National Cooperative Research Act (1984). 59-
Germany, 132, 142 60
Great Britain, 172-73, 177-78 National Council for Metrology, Normalization
Israel, 483, 484-86, 491-92 and Quality Control (CONMETRO)
Korea, 380 (Brazil), 419
536 INDEX
National Council for Research and National research institutions/laboratories

Development (NCRD) (Israel), 478 Japan, 83-85
National Council for Science and Technology Germany, 144-45
(CONICET) (Argentina), 467 National Science Council (Taiwan), 392, 402
National development plans, Brazil, 418 National Science Foundation (U.S.),
National Economic Development Council interdisciplinary engineering research
(Great Britain), 183 centers, 53
National Fund for Scientific and Technological "National system," definition, 5
Development (FNDCT) (Brazil), 419 National System of Scientific and Technological
National innovation systems. See also names of Development (SNDCT) (Brazil), 419
individual countries, e.g. Korea, United National technology programs, Sweden and
States, etc. Denmark, 279-80
advantages of large countries vs. small Nationalization, industrial, in France, 193, 213
countries, 302-04 Natural-resources industries
concept of, 15-17 Canada, 300, 307
defense R & D and, 508, 513-14 Nelson, Richard R.
definitions & characteristics, 4-5, 267-69 "A Retrospective," 505-23
education systems and, 511 "Technical Innovation and National
"emulation" and "fairness" concepts, 520-23 Systems," 3-21
government technology policies and, 512-13 Neotech Development Corp. (NDC), 400-01
government's role, 508, 512-13 New process/product technology
high-tech industrial development and, 515- diffusion in Denmark and Sweden, 284-87
17, 521 introduction, in Japanese firm, 108
infant-industry protection and, 514-15 Nippon Electric Co. (NEC), 95-96
international competitiveness and, 509 Nippon Kokan, 84, 91
internationalization and, 518-19 Nippon Steel, 91-92
"market failure" arguments and, 521 Nissan Motors, 98, 99, 100
"mission-oriented," 220 NKK (Japan), 84, 91
"national systems" concept, 15-17 Nordic countries. See Denmark; Sweden
national continuity and, 509 Northern Securities decision (1904), 32
natural and agricultural resources, Northern Telecom, 306, 307, 313, 315, 318, 320
significance of, 507 Notgemeinschaft, 131-32
structural change in highly-industrialized Novatel, 318
countries, 63 Nuclear industry
technological globalization and, 63 Argentina, 467-68
university research and, 512-13 France, 202, 216-18
National Institute for Agricultural Technology Nuclear power development
(INTA) (Argentina), 465, 466, 468-69 German government funding, 143, 145
National Institute of Agricultural Research national innovation systems and, 514
(INRA) (France), 202 Swedish, 281
National Institute of Health (INH) (France), 202
National Institute of Industrial Property (INPI)
(Brazil), 419, 429-32 Odargiri, Hiroyuki, "The Japanese System of
National Office of Aeronautical Studies and Innovation: Past, Present, and Future,"
Research (ONERA) (France), 202, 205 76-111
National Plan for Science and Technology Office of Overseas Scientific and Technical
(Taiwan), 409-10 Research (ORSTOM) (France), 202
National R & D Projects (NRP) (Korea), 372, Oki, 95
373 Olivetti, 25 1
National Research Council (CNPq) (Brazil), Omnibus Trade and Competitiveness Act
417, 435 (1988), 60
National Research Council (CNR) (Italy), 246 Ontario Hydro, 320
Finalized Programs, 252 Opel, 131
National Research Council (NRC) (Canada), 304 Ouro Preto School of Mines (Brazil), 417
Industrial Research Assistance Program Oxygen converter gas recovery system, 92
(IRAP), 311 Oxygen furnace method, steelmaking, 92
INDEX 537
Panhard, 200 Research & development activities. See also

Paris Radium Institute, 199 Industrial research; Military R & D;
Partnerships for Development, Australia, 341- University research
42 Argentina, 460-61, 465-70
Patents. See also Intellectual property rights; Canada, 302-03
Patents, U.S. France, 201-03, 205-1 1
international comparisons, 104-05 Italy, 231-34, 243-55
Taiwan, 409 Korea, 369-75
Patents, U.S. large, high-income countries. 27
Canada, 306-07, 317 Sweden and Denmark, 278-79
France, 221-22, 223 Research & development expenditures
Germany, 130, 137, 139 Australia, 325-31
Great Britain (1981-86), 175 Brazil, 425, 426, 437-39
Israel, 492-93 Germany, 138, 139, 142-45
Italy, 237-38, 244, 245 Great Britain, 172-76
Japan (1855-1902), 81, 137, 139 international comparisons, 438
large v.s. small countries, 303 Italy, 232
Petroleum and oil industries Japan, 103-04
Australia, 36, 337 Korea, 369-70
Canada, 311, 317 Taiwan, 407-08
Great Britain, 161 Research assodations/consortia
U.S., 33, 34 Canada, 310-11
Peugeot, 200 Japan, 87-88
Pharmaceuticals industry. Sec also Chemicals Microelectronics and Computer Technology
industry Corporation (U.S.), 59
Germany, 126, 127 Research centers
Great Britain, 167 Australia, 345
Physics, rise of research in U.S., 36 Brazil EMBRAPA, 444
Pilot plant, 7, 8 Italy, 242
Plan Calcul (France), 204 Korea, 371, 380-81
Plant Patent Act (1930), 5 1 U.S., interdisciplinary engineering, 53
Plant Variety Protection Act (PVPA) (1970), 51 Research collaboration/cooperation
Plessey, 163 Canada, 310-11
Polymer chemistry, 7 Germany, 144
Polytechnical schools, Germany, 120, 121 Great Britain, 174-75
Prato (Italy), Sprint project, 242-43 Israel, 495
Pratt (U.K.), 99, 318 Italy, 249, 251
Productivity rates Japan, 87-88
Great Britain (1980s), 170-72 Korea, 374
Israel, 481 Sweden, 281
Taiwan, 385 U.S., 53-54, 55, 56, 59-60
total factor productivity (TFP), international Research councils
comparison, 105 Canada, 304-05
Prototypes, 8 Israel, 478
Pulp and paper industry Research institutes
Canada, 310, 312, 316 Brazil, 435-37
Sweden, 271 Canada, 310
Pulp and Paper Research Institute of Canada, France, 202, 203
310 Germany, 131, 145
Israel, 490-91
Korea, 371, 372
Raw materials processing industries, Argentina, Taiwan. 406
460, 461 Research institutions/organizations. Sec also
RCA, 50 names of specific types of organizations,
Regional development agencies, 186 e.g. Research associations/consortia,
Renault, 200 Research laboratories, etc.
538 INDEX
Research institutions/organizations (continued) Scientist and engineers
France, 202-04 Australia, 344-45
Germany, 123-24 Brazil, 425-26, 442
Italy, 252-54 France, 197, 199
Japan, 83-85 Great Britain, 178-79
Taiwan, 392 in industry, international comparisons, 482
Research laboratories Israel, 479-80, 481 -82, 483, 489
France, 198, 199, 205, 210, 211 Japan, 104
Germany. 126, 144-45 large, high-income countries, 26
government, 12-13, 513 Taiwan, 408-09
industrial, 10-11 Scotland, electronics industries in, 187
Israel, 478 Sematech consortium, 47, 58-59
Japan, 83, 84. 110 Semiconductor Chip Protection Act (1984), 57
Korea, 370-71 Semiconductors
U.S., 33-34, 50 Korean chaebol activities, 375-77
Research personnel. See Scientists and engineers U.S., 43, 57
Revealed technological advantage (RTA) Seoul Science Park, .380
F'rance, 221 Service industries, Great Britain, 162-64
Great Britain, 168-69, 176 Shell, 161
Reverse engineering, 11 Shibaura Engineering Works, 94
Korea, 365, 366, 369, 378-79 Shibaura Seisakusho, 84
Riken, 83-85 Shipbuilding industry, Denmark, 277
Robotics Siemens, 94, 127-28, 138, 139, 143, 163, 168
development in Korea, 378-79 Single European Act, 188
diffusion in Sweden and Denmark, 284-85 Single European Market, 174
FIAT and, 255-56 SKF, 271
Rolls-Royce, 167 Skilled labor. See Scientists and engineers
Rosenberg, Nathan, "The U.S. National Small and Medium Enterprise Formation Act
Innovation System," 29-64 (1986), 368
Small- and medium-sized enterprises (SMEs)
Sabatini Law, 241-42 Israel, 490-91
Saint-Gobain, 195 Korea, 368
Samsung, 375, 377, 378, 379 Taiwan, 384, 388, 390-91, 396-400, 403-05,
research organizations, 376 411
Sao Paulo Polytechnic, 417 Small firms
Scandanavian countries. See Denmark; Sweden Denmark, 289
Schncider-Empain group, 217 Great Britain, 167
Science Council (Japan), 84 Italy, 230-31, 231, 234-43, 250
Science parks Korea, 364-67, 368
Germany, 145 U.S., 29, 48-49, 56-57
Great Britain, 186-87 South Korea. See Korea
Israel, 492 Space technology programs
Italy, 242 France, 201, 203, 204, 218-19
Scientific research, 402. See also University Germany, 145
research national innovation systems and, 514
British expenditures, 172-73 Sprint project, 242-43
Canada, 305 Start-up firms, U.S., 29, 48-49, 56-57
France, 194-95, 198-99 Statute for Encouragement of Investment (SEI)
Germany, 117-24, 131-32 (Taiwan), 394
Israel, 478 Steel industry. See Iron and steel industries
Italy, 245-46 Stifterverband, 132
Japan, 1 10-11 Stoneville, 404
Taiwan, 392 Sumitomo Metal, 91, 92
technological advance and, 6 Sunshine Project (Japan), 89
U.S., 35-36, 42 Suzuki, 100
INDEX 539
Sweden, 263, 265-92 R & D expenditures, 407-08

Board for Technical Changes, 290 research organizations
Board for Technical Development (STU), government-sponsored, 396-400
279 research institutes, 406
Denmark and, comparisons, 265-69 scientific research, 392, 402
employment rates, 273, 283 scientist and engineers, numbers of, 408-09
engineering industry, 290 small- and medium-sized enterprises (SMEs)
exports, 271-72 in, 384, 387-88,411
industrialization, 271-72, 273-74 overseas mergers and, 403-05
international economic and R & D technology development and, 388-91, 396-
comparisons, 264 400
labor market, 273-74 sources of economic growth, 385-86
macroeconomic performance, 282-83 technology transfer among firms, 389
multinationals in, 266-67, 281, 291, 292 Tanaka Seisakusho, 93
national technology programs, 279-80 Tariff'protection
natural resources, importance to economy, Argentina, 454
289 Australia, 335
product innovations, 271-72 TDX-l/TDX-10, 374, 378
production system, 273-74 Technical advance/change
development blocks, 275, 277-78 concept of "national systems," 15-17
R & D activities, 278-79 cumulative, incremental, 8-9
nuclear power research program, 281 Danish production system and, 275
socio-political aspects, 272-75 government support, 12-13
structural problems in economy, 288-89 interindustry differences, 13-15
technological contributions, 283-84 process of, 5, 267-68
technology diffusion, 284-87 R & D facilities and, 5-6
technology policy, 279-82 industrial research laboratories, 10-11
U.S. patents, 266, 283 rise of science and, 6-9
Synthetic dyestuffs industry, Germany, 126, start-up firms and, 48-49, 56-57
130 Swedish trade unions and, 273-74
System X (Great Britain), 176 transnational aspects, 17-18
"System(s)," concepts, 4-5 university research and, 11-12
Technical education/training. Sec also
Engineering education/training
Taiwan, 384-412 Brazil, 439-40
"brain-drain" problem, 393-94 Germany, 120, 121, 139-40
computer software development, 400-01 Great Britain, 179-80
economic characteristics, international Taiwan, 393
comparisons, 355-56 Technical innovation. See National innovation
education system, 385, 392-94 systems; Technical advance/change
engineering education, 393 Technische Hochschulen, 120, 123
higher education, 392, 393 "Technological community," concept, 15
foreign direct investment, 388-91, 405-06 Technological Innovation Fund (Italy), 253
government's role in industrial technology Technology, indigenous
development, 391-92 Great Britain, 167-68
assistance programs, 394-95 Japan, 77-78, 81, 87-88, 101
industrial base in 1950s, 386 Technology Development Promotion Act
industrial research, 394-96 (1977) (Korea), 375
intellectual property rights protection, 406 Technology Development Reserve Fund
Japanese occupation, 386 (Korea), 374
labor productivity, 385 Technology diffusion
national defense industries, 406 Canada, 308-09
National Plan for S & T, 409-10 France, 211
patents, 409 international comparisons, 256
publicly-owned enterprises in, 387 Israel, 494-95
540 INDEX
Technology diffusion (continued) UBATEC, 469

Italy. 236. 242-43. 255-56 Unequal Treaties, 80-81
Sweden and Denmark, 284-87 Union of German Engineers, 120
Technology policies Unions
Brazil, 428-34 Denmark, 273
effects on innovation, 512-13 Germany, 133
Germany. 142-45 Japan, 108
Israel. 487-88 Korea, 369
Italy, 251-55 Sweden, 273-74, 291-92
Korea. 367-69 United Kingdom. See Great Britain
Sweden and Denmark, 279-82 United States, 29-64
Technology transfer agricultural research, 36-38, 51-52
Argentina. 467 American model of innovation, 16
Brazil, 429 antitrust statutes, 29, 32, 34, 49, 59-60, 62-
Canada, 305. 308-09 63
France, 211 education system, 48
Germany, 116-17 engineering training, 36
Italy, 24.2-43 higher education, 35-36, 48
Japan, 81, 85-87. 92, 96, 100, 101 government R & D funding, 34-36
Korea, 360-62, 364-67. 368 industrial research
Taiwan, 389-91 before 1945, 31-39
"Technonationalism," 3 post-World War II, 39-52
Tekijuku, 78 decline in 1970s, 50-51
Telecommunications industry 1980s, 53-56
Canada, 313, 315, 318, 320 strength of, 30
France, 219-20 intellectual property rights protection
Great Britain, 168 initiatives, 59
Korea, 374, .378 Korean chaebol activities in, 375, 377
Terakoyu, 78 living standards, 52
Teubal. Morris, "The Innovation System of military R & D, 30, 42-43, 46-47, 514
Israel: Description, Performance, and military funding of commercial technologies,
Outstanding Issues," 476-97 58-59
Texas Instruments, 377 national R & D investment, 29. 30
Textile industry Northern Securities decision (1904), 32
Italy, 235, 236, 241 patents, 104
"Thatcherism." 159 public policies for technology development,
Thermodynamics. 7 58-61
Thomson-CSF, 218 relationship with Canadian economy, 301,
Tokyo Denki, 84 307-09
Tokyo Electric Power Company, 93, 94 research collaboration initiatives, 59-60
Tokyo Gas Electric, 98 start-up firms in, 29, 48-49, 56-57
Tokyo Ishikawajima Zosensho, 98 university/industry research cooperation, 53-
Tokyo Shibaura Electric, 94 54
Toshiba, 84, 93, 94 university research system, 12, .35-36, 40, 47-
Total factor productivity (T'FP) 48
international comparison (1960s- 80s), 105 vital statistics, 25-27
Israel, 481 U.S. Defense Advanced Research Projects
Toyota, 98, 99. 100 Agency (DARPA), 59
Trade associations, Canadian R & D and, U.S. Department of Health and Human
310 Services, basic research budget, 42
Transistor, development, 9 U.S./Japan Agreement on Scientific
Transnational aspects, of technical advance, 17- Cooperation, 60
18 U.S. Office of Scientific Research and
Transportation equipment industry, U.S., 34 Development (OSRD), 39
TSMC, 398-99 U.S. patents. See Patents, U.S.
INDEX 541
U.S. Pentagon, R & D development initiatives, Volksschule, 121

58 Volkswagen, 138
University of Buenos Aires, 469 Volvo Uddevalla factory, 274
University research
Argentina, 469 Walker, William, "National Innovation
Canada, 305 Systems: Britain," 158-89
France, 198-99, 210-11 Welding Institute, 183
Germany, 117-20, 140-41 West Germany. See Germany
industrial innovation and, 512-13 Western Electric, 40
Israel, 478, 484, 492-93 Japanese communications equipment
Italy, 245-46 industry and, 95
Japan, 110-11 Westinghouse, 94
Korea, 371 Wood and paper industry. See Pulp and paper
Taiwan, 402 industry
technical advance and, 7, 11-12 Work organization
U.S., 12, 35-36, 40, 42, 47-48 Danish and Swedish innovations, 284
Uruguay Round, 60 Fordist/Taylorist, 31-32
Worker rotation, in Japan, 107
Vitelic, 399 WYSE Technology, 404-05
Vocational/technical training
Brazil, 439-40
Germany, 120, 121, 139-40 Yawata Steel Works, 90, 91-92, 101
Taiwan, 393
Vulcani Institute. 478-79 Zaibatsus, 98, 101

Nelson - NIS, A Comparative Analysis

Uploaded by

Copyright:

Available Formats

Nelson - NIS, A Comparative Analysis

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Nelson - NIS, A Comparative Analysis

Uploaded by

Copyright:

Available Formats

What is the book about?

What is the book about?

What institutions funded the research project?

What institutions funded the research project?

National Innovation Systems

This page intentionally left blank

New York Oxford

Copyright 1993 by Oxford University Press, Inc.

Library of Congress Cataloging-in-Publication Data

Part I Large High-Income Countries, 23

Part II Smaller High-Income Countries, 261

Part III Lower Income Countries, 353

12. National Systems Supporting Technical Advance in Industry:

Part IV National Innovation Systems, 503

Nestor A. Bercovich Professor Jorge Katz

Francois Chesnais Professor Otto Keck

Professor Nathan Rosenberg Professor Morris Teubal

Dr. Gee San Professor William Walker

Claude Barfield Professor Nathan Rosenberg

WHAT IS THIS STUDY ABOUT

Technical Advance: An Overview of the Processes and Institutions Involved

To understand national innovation systems, it is essential to understand how technical

THE INTERTWINING OF SCIENCE AND TECHNOLOGY

Science as Leader and Follower

The Limits of Science, Learning by Trying, and Cumulative Incremental

It is insufficiently appreciated that successful innovation in high technology industries

THE MAJOR INSTITUTIONAL ACTORS

Firms and Industrial Research Laboratories

Other Institutional Actors

vision set or an automatic loom. In either case technical advance in system-product

TECHNOLOGICAL COMMUNITIES, BOUNDARIES, AND NATIONAL

Technological advance proceeds through the interaction of many actors. Above we

NATIONAL SYSTEMS AND TRANSNATIONAL TECHNOLOGY

A GUIDE TO THIS VOLUME

Allen, R. C. (1983). "Collective Invention." Journal vation." In G. Dosi, C. Freeman, R. Nelson, G.

LARGE HIGH-INCOME COUNTRIES

all of these countries manufacturing exports accounted for the majority of

Table I.I. Comparison of the Six Large High-Income Countries

A descriptive analysis of the U. S. national innovation "system" is a gargantuan task,

THE U. S. SYSTEM BEFORE 1945

The Origins of U. S. Industrial Research

adversarial atmosphere of labor-management relations and low levels of investment

The Growth of Industrial Research

Although recent historiography on U. S. industrial research has focused primarily on

Table 2.1. Employment of Scientists and Engineers in Industrial Research

Publicly Funded Research and the Universities

Table 2.2. Defense R&D as a Share of Federal R&D Spending, 1960-1990

THE POSTWAR SYSTEM

World War II and Its Aftermath

Postwar R&D Expenditures

Military R&D Funding

Table 2.4. Trends in Federal R&D Expenditures

1983 Federal R&D Expenditures18

University Research and Federal Funding

sities, government, and a number of private firms served as important "incubators"

The rapid growth in postwar agricultural productivity (measured either in terms of

POLICY CHALLENGES AND STRUCTURAL CHANGE IN THE 1980S

The Growth (Rebirth?) of University-Industry Research Cooperation

Industrial Research in an Era of Structural Change

Restructuring U. S. Industrial Research?