On Krugman Geographical Economics

Paul Krugman's Geographical Economics and Its Implications for Regional Development
Theory: A Critical Assessment

Author(s): Ron Martin and Peter Sunley
Source: Economic Geography, Vol. 72, No. 3 (Jul., 1996), pp. 259-292
Published by: Taylor & Francis, Ltd.
Stable URL: https://www.jstor.org/stable/144401
Accessed: 02-09-2019 20:37 UTC
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
Taylor & Francis, Ltd. is collaborating with JSTOR to digitize, preserve and extend access to
Economic Geography
This content downloaded from 161.116.100.129 on Mon, 02 Sep 2019 20:37:06 UTC
All use subject to https://about.jstor.org/terms
Paul Krugman's Geographical Economics and Its
Implications for Regional Development Theory:
A Critical Assessment*
Ron Martin
Department of Geography, University of Cambridge,
Cambrige CB2 3EN, U.K.
Peter Sunley
Department of Geography, University of Edinburgh,
Edinburgh EH8 9XP, U.K.
Abstract: Economists, it seems, are discovering geography. Over the past decade,
a "new trade theory" and "new economics of competitive advantage" have
emerged which, among other things, assign a key importance to the role that the
internal geography of a nation may play in determining the trading performance
of that nation's industries. Paul Krugman's work, in particular, has been very
influential in promoting this view. According to Krugman, in a world of imperfect
competition, international trade is driven as much by increasing returns and
external economies as by comparative advantage. Furthermore, these external
economies are more likely to be realized at the local and regional scale than at the
national or international level. To understand trade, therefore, Krugman argues
that it is necessary to understand the processes leading to the local and regional
concentration of production. To this end he draws on a range of geographical
ideas, from Marshallian agglomeration economies, through traditional location
theory, to notions of cumulative causation and regional specialization. Our
purpose in this paper is to provide a critical assessment of Krugman's
"geographical economics" and its implications for contemporary economic
geography. His work raises some significant issues for regional development
theory in general and the new industrial geography in particular. But at the same
time his theory also has significant limitations. We argue that while an exchange
of ideas between his theory and recent work in industrial geography would be
mutually beneficial, both approaches are limited by their treatment of
technological externalities and the legacy of orthodox neoclassical economics.
Key words: Krugman, trade, external economies, regional industrial concentra-
tion, regional industrial policy.
The relationship between economic drawn freely on the concepts and perspec-
geography and economics has long been tives of different schools of economics;
an asymmetric one. In constructing their but, for their part, economists have
theories and explanations of regional tended to accord little if any attention to
development, economic geographers have the role of geography in the economic
process. The case of trade theory admira-
bly illustrates this point. Regional devel-
*This is a revised version of a paper
presented at the Special Session on Economic
Geography held at the Annual Conference of tried to respond to as many of these as
the Institute of British Geographers, Newcas- possible. We also wish to thank the three
tle, January 1995. The authors gratefully anonymous referees for their suggestions,
acknowledge the many useful and constructive which have also helped to sharpen our
comments made on that occasion, and we have argument.
259
260 ECONOMIC GEOGRAPHY
opment theory has always been con- mists, it seems, are discovering geogra-
cerned with the question of interregional phy. In particular, Paul Krugman, the
trade, because a region's ability to export leading and extraordinarily prolific expo-
goods and services is one of the founda- nent of the "new" trade theory,2 has
tions of local economic growth and em- sought to show how trade is both influ-
ployment (Erickson 1989). The typical enced by and in turn influences the
approach to the study of interregional process of geographical industrial special-
trade has been to borrow and adapt the ization within nations (for example, Krug-
ideas and models of comparative advan- man 1991a). In his view, the importance
tage (factor endowment) trade theory of regional industrial specialization and
from economics. Trade economists, how- concentration is such that economic geog-
ever, have invariably regarded the na- raphy should be accepted as a major
tional economy as spaceless, and even subdiscipline within economics, "on a par
international trade typically has been seen with or even in some senses encompass-
as an exchange system devoid of any ing the field of international trade"
geography, a world where goods and (Krugman 1991a, 33). Likewise, from a
services move between dimensionless different, but ultimately related perspec-
tive, Michael Porter, the eminent busi-
points at zero or uniform transport costs.
This lack of a sensitivity to geographyness
by economist, has argued that the
trade theorists partly explains why there
degree of geographical clustering of in-
is no overall theoretical framework guid-
dustries within a national economy plays
ing geographical research on international
an important role in determining which of
trade (Grant 1994). The absence of such a
its sectors command a competitive advan-
tage within the international economy
framework is particularly evident at a time
when the "globalization" of economic (Porter 1990). In a similar vein to Krug-
man, Porter also argues that there are
relations and the continental regionaliza-
strong grounds for making economic
tion of trade are challenging the territorial
and regulatory significance of national geography a "core discipline in econom-
economic spaces and giving greater prom- ics" (Porter 1990, 790).
inence to the nature and performance of Paul Krugman's work, especially, is
individual regional and local economies worthy of closer interest by geographers.
within nations (Dunford and Kafkalas Krugman has written on a wide range of
1992; Anderson and Blackhurst 1993; issues that impinge on the regional
Gibb and Michalak 1994). development question: trade, externali-
Recently, however, there have been ties, the localization of industry, strategic
developments within economics which industrial policy, globalization, the role of
may mark the beginning of a closer history and "path dependence," and the
relationship with economic geography in implications of economic and monetary
general and regional development theory integration for regional growth. One of
more particularly. Over the past decade, a the key thrusts of his work is that in order
"new" trade theory and a new economics to understand trade we need to under-
of competitive advantage have emerged stand the process of regional development
which, among other important features, within nations. A number of his writings
assign a key significance to the role that have thus sought to explain why industrial
the internal geography of a nation may development is likely to be geographically
play in determining the trading perfor-
mance of that nation's industries.1 Econo-
Lancaster (1980), Krugman (1979, 1980, 1981),
Ethier (1982), and Helpman (1984).
2 Such has been Krugman's influence within
1 The set of ideas referred to as the "new the economics profession that Paul Samuelson
trade theory" was originally expounded in a (1994, vii) refers to him as "the rising star of
series of papers by Dixit and Norman (1980), this century and the next."
KRUGMAN'S GEOGRAPHICAL ECONOMICS 261
uneven. To this end, he draws on a range economics.3 However, this methodologi-

of economic and geographical ideas, from cal and theoretical disposition is unlikely
Alfred Marshall's account of localization to appeal to many economic geographers,
who have abandoned formal models and
economies, through traditional location
theories, to notions of cumulative causa-rigorous exegesis for a more discursive
approach, in which broad master concepts
tion. For Krugman, economic geogra-
(like "flexible specialization" and "post-
phy-by which he means uneven regional
Fordism") are mingled with anecdotal
development-is a central part of the spatial stereotypes ("industrial spaces"
process by which national economic pros- and "industrial districts").
perity and trade are created and main- These differences probably largely ex-
tained. plain why Krugman's writings have thus
Our aim in this paper is to provide a far had a limited impact on economic
critical assessment of Krugman's "geo- geography, and why they have been
graphical economics" and its implications summarily dismissed by certain geogra-
for contemporary economic geography. phers. Johnston (1992) dismisses Krug-
An exchange of ideas between his theory man's Geography and Trade (1991a) as
"highly simplistic" in its treatment of
and recent work in economic geography
"geography" and "patronizing" in its
would be mutually beneficial. Such an comments on the work of economic
exchange is not easy to engineer, how- geographers.4 In a somewhat similar vein,
ever, as there are several significant
obstacles, on both sides. First, Krugman's
ideas are far from static. Indeed, his views
seem to change continuously over time- 3This location-theory orientation has if
sometimes in a self-criticizing way-so it anything become even more pronounced in his
two most recent books on "spatial economics,"
is important to base any evaluation on a
Development, Geography and Economic The-
range of his works. Second, throughout
ory (Krugman 1995) and The Self-Organising
Krugman's writings there is a strong Economy (Krugman 1996). In the former,
distinction between what is theoretically lamenting economic geographers' "retreat"
possible and what is empirically and from quantitative models into Marxist and
practically important, so that his conclu- regulationist concerns with "post-Fordism,"
sions have to be read carefully and Krugman resurrects what he calls the five
"exiled traditions" of economic geography:
closely. Most important, however, Krug-
location theory, social physics, cumulative
man's geographical economics and con- causation, land use modeling, and Marshallian
temporary economic geography are very local external economies. In the latter, von
different academic genres, with different Thunen's model and central place theory
methodological styles and conventions of occupy a key role in his theorization of the
analysis and writing (Krugman 1993a). "self-organizing" space economy.
4 It is not difficult to see how economic
Krugman's method is to start with a real
geographers might take offense at Krugman's
world problem and then build a model to
view of their work. In Geography and Trade
capture the "essence" of that problem (1991a, 3-4) he writes: "The decision by
(Krugman 1989, 1992). The model, which international economists to ignore the fact that
is usually mathematically specified, is they are doing geography wouldn't matter so
made as simple as possible to remove much if someone else were busy ... looking at
localization and trade within countries. Unfor-
unnecessary clutter, although in most
tunately, nobody is. That is, of course an unfair
cases he also gives a highly readable
narrative account of the model. The statement. There are excellent economic geog-
raphers out there.... However, ... economic
mathematical aspect of his methodology
geographers proper are almost never found in
may well explain the strong location- economics departments, or even talking to
theory flavor to much of his geographical
economists .... They may do excellent work,
in his review of the same book, Hoare cance of history, "lock-in," and path
(1992, 679) criticizes the particular eco- dependence for regional development.
nomic geography used by Krugman as The subsequent section examines these
"dated, historically and intellectually" and ideas in closer, more critical detail, and
his analysis as based on the "flimsiest of compares Krugman's theories with those
empirical support." However, Krugman's that have emerged from the "new indus-
remarks are leveled primarily at his own trial geography" in the past few years. We
colleagues' failure to admit that "space then examine his arguments about the
matters" (Krugman 1991a, 8), and he impact of economic integration on re-
should at least be congratulated for gional development, especially his prog-
wanting "to bring geography back into noses of the regional implications of
economic analysis," even if the particular integration within the European Union
form of geography he uses-essentially a and his views on regional stabilization and
form of regional science-is open to industrial policy. We conclude the paper
criticism. Furthermore, Geography and by drawing together the main strengths
Trade gives only a partial glimpse into and weaknesses of Krugman's approach to
Krugman's analyses, and any considered economic geography.
judgment as to the significance of Krug-
man's work for economic geography must
also be based on his numerous other Trade, Externalities, and
Industrial Localization:
writings in the field.
We too have criticisms to make of The Bases of Krugman's
"Geographical
Krugman's treatment of economic geogra- Economics"
phy, although we also believe that his
The New Trade Theory and Location
work raises some interesting issues for
contemporary regional theory. We begin Krugman's geographical economics and
by outlining what we take to be the theorization of uneven regional develop-
essential arguments and components ment of are firmly rooted in his contribu-
his "geographical economics," focusingtions
on to the "new trade theory." Conven-
his interpretation of the relationships tional trade economics is based on
between location and trade, the role of Ricardian comparative advantage theor
(especially in its Heckscher-Ohlin-Sam
increasing returns and externalities in the
uelson versions), which argues that und
localization of industry, and the signifi-
conditions of perfect competition, and
given the relative immobility of one o
but it does not inform or influence the
more factors of production, nations wi
economics profession." It could equallyspecialize
be in those industries in whic
argued, of course, that it is the economists who
have failed to talk to economic geographers
they have comparative factor advantag
(favorable resources of raw material
and that, as a result, like Krugman they are
cheaper labor, and so forth). The relati
largely ignorant of the major developments
that have taken place in economic and factor endowments of different nations is
industrial geography over the past decade or thus the main reason for international
so. Equally irritating is Krugman's comment, trade and specialization. The principle of
in Development, Geography and Economic comparative advantage, then, predicts
Theory (1995, 88), that "in the end, we [i.e., that countries with dissimilar resource
economists] will integrate spatial issues into endowments will exchange dissimilar
economics through clever models (preferably
goods. The theory does not and cannot,
but not necessarily mine) that make sense of
however, predict what sort of goods will
the insights of the geographers in a way that
meets the standards of the economists." be exchanged by countries that have
similar resource endowments. But much
Whether economists have any such monopoly
over analytical or theoretical standards may of world trade, and most of Organization
most certainly be questioned. for Economic Cooperation and Develop-
ment (OECD) trade, is between countries accident. The specific location of a partic-
with similar factor endowments, and they ular microindustry is to a large degree
exchange predominantly similar products. indeterminate, and history-dependent.
Such intraindustry trade has been ex- But once a pattern of specialization is
panding rapidly in recent decades, even established, for whatever reason, that
though countries have been converging in pattern gets "locked in" by the cumulative
skill levels and per capita endowments of gains from trade. There is thus a strong
capital (OECD 1994). The "new trade tendency toward "path dependence" in
theory" is an attempt to account for this the patterns of specialization and trade
form of trade. The new trade theory between countries: history matters.
acknowledges that differences among Third, the patterns of demand for and
countries are one reason for trade, but it rewards to factors of production under
goes beyond the traditional view in four conditions of imperfect competition and
main ways (Krugman 1990).5 intraindustry trade will depend on the
First, it argues that much trade be- technological conditions of production at
tween countries, especially intraindustry the micro level, and nothing can be said a
trade between similar countries, repre- priori about the evolution of factor de-
sents specialization to take advantage of mands. Fourth, whereas under the Ricar-
increasing returns to scale rather than to dian model free trade is assumed to be the
capitalize on inherent differences in na- appropriate policy stance, the new trade
tional factor endowments. Contrary to the theory argues that the existence of imper-
assumptions of perfect competition and fect competition and increasing returns
constant returns to scale that underpin opens up the possibility of using trade
the basic Ricardian theory of comparative policies strategically to create compara-
advantage and trade, according to the new tive advantage by promoting those export
theory imperfect competition and increas- sectors where economies of scale-and
ing returns are pervasive features of particularly external economies-are im-
contemporary industrial economies.6 If portant sources of rent. In other words,
specialization and trade are driven by strategic trade policy may enable a nation
increasing returns and economies of scale to shift the pattern of international eco-
rather than by comparative advantage, the nomic specialization in its own favor
gains from trade arise because production (Krugman 1980).
costs fall as the scale of output increases. In Krugman's view, these develop-
Second, with this view of the world, ments in the "new trade theory" both
specialization is to some extent a historical necessitate and facilitate a rapprochement
between trade theory and location theory.
In recent work he has compared the
5 There are in fact several different versions contrasting assumptions underlying these
of the new trade theory, but the various two, hitherto largely separate, sets of
strands all subscribe to the basic elements economic literature (Krugman 1993a). His
elaborated by Krugman in his Rethinking geographical economics is a hybrid of the
International Trade (1990). two. It combines the models of imperfect
6 Of course, the idea that increasing returns competition and scale economies used in
and economies of scale could be alternatives to
new trade theory with location theory's
comparative advantage as explanations of emphasis on the significance of transport
international specialization and trade goes costs. The interaction of external econo-
back to Ohlin (1933), if not to Adam Smith. But
mies of scale with transport costs is the
while their importance has been recognized in
principle, they invariably have been assigned a key to his explanation of regional indus-
trial concentration and the formation of
subsidiary or supplementary role in formal
trade theory. The novelty of the "new trade regional "centers" and "peripheries"
theory" is that increasing returns and econo- (Krugman 1991a; Krugman and Venables
mies of scale are moved into the mainstream. 1990). His model suggests that high
transport costs will act to prohibit the industries or clusters of industries to

geographical concentration of production. concentrate in space" (Krugman 1993b,
With some reduction in transport costs, 173). This tendency, he argues, provides a
however, firms will want to concentrate in decisive refutation of the competitive
one site to realize economies of scale both model of economic equilibrium, for when
in production and in transport. In Krug- one turns to the location of production in
man's words, "Because of the costs of space the "irrelevance of equilibrium
transacting across distance, the preferred economics" is compelling and there are
locations for each individual producer are multiple possible equilibria. An econo-
those where demand is large or supply of my's form is determined by contingency,
inputs is particularly convenient-which path dependence, and the initial condi-
in general are the locations chosen by tons set by history and accident. Forward
other producers" (1991a, 98). If transport and backward linkages mean that once an
costs continue to fall, the model suggests initial regional advantage is established it
that the need to locate near to markets may become cumulative. There is there-
will disappear and production may dis- fore no automatic tendency toward an
perse. However, given that some trans- optimum solution, as apparently "irratio-
port costs will remain, the circular rela- nal" economic distributions may be
tion, or positive feedback, between "locked in" through increasing returns. So
production and demand means that re- that while Krugman associates economic
gions which have a head start in manufac- geography with path dependence, or what
turing, typically as a result of accidental he calls "the economics of qwerty,"7 he
good fortune, will attract industry and does not neglect the possibility of reversal
growth away from regions with less and change. Rather, he argues that when
favorable initial conditions. Krugman change in regional fortunes occurs it will
(1991a, 1991d) argues that this model be sudden and unpredictable. He repeat-
explains the rise of the manufacturing belt edly uses the example of Massachusetts as
in the Northeastern United States during a regional economy that has gone into a
the nineteenth century. It has also been tailspin. Krugman (1991c) suggests that,
applied to the discussion of the likely fate under certain conditions, self-fulfilling
of peripheral regions in the European expectations may outweigh accumulated
Union (Krugman and Venables 1990). advantages, and pessimistic expectations
Recently, Krugman (1993d) has developed about a region's prospects may become
a new variant of the approach which self-justifying.
argues, not that successful regions system-
atically attract industry away from periph-
eral areas, but that trade and external Increasing Returns and
economies produce more-specialized re- Imperfect Competition
gions, which are then more vulnerable to According to Baldwin (1994), Krug-
random "shocks."
man's analysis represents a genuinely new
On the basis of this location model,
location theory. In actuality, not only does
Krugman (1993b) argues that large-scale
regions are more significant economic
units than nation-states. He writes that a
7 This term derives from the first line of
satellite image of the world at night shows
regional agglomerations rather than na- letters (QWERTYUIOP) on the keyboard of a
typewriter or word processor. That this order
tional concentrations. Furthermore, in his
is the same today as it was on the first
view, "The best evidence for the practical mechanical typewriters of the nineteenth
importance of external economies is so century, even though more efficient sequences
obvious that it tends to be overlooked. It are possible, represents a form of "lock-in" and
is the strong tendency of both economic persistence that has analogous parallels in the
activity in general and of particular economy.
it echo Ohlin (1933), Hirschman (1958), external economies (see, for example,
and Myrdal (1957), it strongly resembles Romer 1990) have given them a new
Weber's (1929) model of the overlaying of tractability. He argues, in one paper
transport costs on agglomeration econo- (Krugman 1981), that external economies
mies. Whereas Weber identified spatial at a national level are the key to the
overlaps, the Krugman-Venables model uneven development of countries. Yet,
adds the general level of transport costs as increasingly, Krugman has been reluctant
a variable that can fluctuate over time. to treat nations as economic units and has
Given these predecessors, we shouldemphasized the significance of external
consider whether there is anything really economies at a local and regional scale.
new in Krugman's geographical econom- Indeed, in Geography and Trade his
ics. In several places he himself states thataccount of the localization of industries
he is simply retelling an old story in aand agglomeration at a relatively small
more rigorous way. It would be tempting scale is based on Marshall's three types of
to conclude, as some critics have done, external economy: labor market pooling,
that there is nothing new in this. How- the availability of specialist suppliers, and
ever, this conclusion overlooks the way in the presence of technological knowledge
which Krugman's reading of agglomera- spillovers. However, he places greater
tion has been shaped both by the emphasis on the first two of these and
developments in trade theory and by deals only briefly with local technological
recent models of industrial organization. externalities. This might seem strange
One of the main reasons for trade theory's given that elsewhere he has argued that,
traditional neglect of the advantages empirically, the most plausible source of
which arise from increasing returns and positive externalities from trade is the
economies of scale was the difficulty of inability of innovative firms to monopolize
modeling market structure. In one sense, the knowledge they create (Krugman
recent developments in modeling market 1987a, 137). But the arguments are not
structure with nonconstant returns have incompatible, for Krugman argues that
facilitated the new trade theory (Helpman these externalities are difficult to measure
1984; Krugman 1983a; Buchanan and and track and that many of them are
Yoon 1994; Smith 1994). Hence the best national or international in scope (see also
place to start, in order to understand Ethier 1982). So while local technological
Krugman's interpretation of increasing externalities are important in some high-
returns, is with these models. Two ap- tech districts, he considers their general
proaches are particularly relevant to locational significance to be limited.
Krugman's account of geographical con- Further reasons for Krugman's lack of
centration, namely the Marshallian and emphasis on technological spillovers be-
Chamberlinian models. come apparent when we turn to the
The Marshallian approach to under- second model of market structure that has
standing increasing returns is already been influential in new trade theory,
familiar in economic geography. It is namely the Chamberlinian model (Cham-
based in a long tradition that sees berlin 1949). This model of market equi-
economies of scale as primarily external, librium envisages competition among sim-
as arising from the specialization of the ilar firms producing differentiated
social division of labor (Young 1928; products which are close but not perfect
Stigler 1951). Typically, economies of substitutes. Each firm faces a downward
scale have been taken to be purely sloping demand curve and has some
external, so that the assumptions of monopoly power. The entry of new firms
perfect competition may be retained producing slightly different products
(Chipman 1970). While Krugman is aware eliminates monopoly profits and means
of this long tradition, he suggests that that there are many little monopolists.
recent advances in the modeling of such Many explanations of intraindustry trade
by new trade theorists have been devel- divide between "technological" and "pe-
oped from this model, with the assump- cuniary" externalities is misleading and
tion of economies of scale that are internal unhelpful.9 In the competitive equilib-
to firms.8 According to Helpman and rium model, technological externalities
Krugman (1985), these internal economies are defined as those consequences of
are easy to justify. They argue that firms activity which directly influence the pro-
could both achieve economies of scale and duction function in ways other than
meet a demand for differentiated products through the market. They have real
from other producers and consumers by welfare and efficiency consequences
locating at one site and engaging in (Meade 1952; Mishan 1971). In a situation
intraindustry trade (Krugman 1989). of perfect competition and constant re-
Krugman sees this approach as especially turns, however, pecuniary externalities
relevant to intermediate products and which arise through buying and selling in
components, where the scope for differen- the market are scarce (Scitovsky 1954).
tiation is high and the market often too Krugman argues that this type of distinc-
small for an exhaustion of scale econo- tion is misleading. In his words, "It is
mies. Moreover, he argues that valid only when there are constant returns
where intermediate goods produced with
and perfect competition; in a world of
economies of scale are not tradeable, the increasing returns and imperfect competi-
result will be to induce the formation of tion, the range of significant external
"industrial complexes"-groups, of industry economies is much larger. In particular,
tied together by the need to coANentrate allthere are true external economies associ-
users of intermediate goods in the sameated with a variety of market-size effects"
country. In this case the pattern of special-(Krugman 1993b, 166). In summary, there
isation and trade in the Chamberlinian
are increasing returns in production, so
world will come to resemble the pattern in the size of the market matters and
that
the Marshallian world described previously.
pecuniary economies also have real wel-
(Krugman 1987c, 319; compare Losch 1967,
109) fare significance. Elsewhere Krugman
writes,
Both of these models of competition are
Over the past decade . . . it has become a
implicit in Krugman's discussion of re-
familiar point that in the presence of
gional and local externalities. imperfect competition and increasing re-
turns, pecuniary externalities matter; for
The Role and Implications example, if one firm's actions affect the
of Externalities demand for the product of another firm
Krugman uses Marshall's theorization

of external economies to explain geo- 9 The term "pecuniary externalities" was
graphical clustering at a relatively small used by Scitovsky (1954) to refer to externali-
scale, that of urban specialization and city ties arising from market imperfections of both
agglomerations. However, the Chamber- demand and supply. Market-size effects are an
linian approach has been equally impor- important form of pecuniary external econ-
tant to his account of externalities. In omy; the larger the market, the more individ-
ual firms can increase their output without
Krugman's view, the presence of increas-
having to cut prices. Increasing market size
ing returns implies that the orthodox
permits increasing returns. Such market-size
effects may operate at various geographical
scales, from the international to the local.
8 Hanink (1988, 1994) describes these ap-Technological externalities refer to the situa-
tion where there are spillovers from the
proaches as the theory of differentiated mar-
kets, and uses an extended Linder model to production function of one firm into those of
explain the consequences of geographical
other firms, for example when a firm makes an
product differentiation. innovation that other firms can imitate.
whose price exceeds marginal cost, and this competitive advantage of the constituent
is as much a "real" externality as if one firms. In this sense, then, Krugman's work
firm's research and development spills over conveys a sense of regional competitive-
into the general knowledge pool. At the ness.11 At the same time, his recent
same time, by focusing on pecuniary
externalities, we are able to make the
writings on the international economy
analysis much more concrete than if we have criticized certain popular definitions
allowed external economies to arise in some and uses of competitiveness, and it is
invisible form. (Krugman 1991b, 485) important to set his regional work within
the context of his more general under-
This focus on pecuniary externalities standing of the consequences of trade
shapes Krugman's interpretation of Mar- (Krugman 1994a, 1994b, 1994c, 1994d).
shall (see also David and Rosenbloom The primary issue here is that Krugman
1990). For instance, he claims that both sees all forms of international economic
labor market pooling and the availability integration, including trade and capital
of nontraded intermediate goods are and labor mobility, as essentially benefi-
examples of market-size effects.10 More- cial. For example, the specialization pro-
over, he emphasizes "pecuniary" external- duced by trade raises the efficiency of the
ities which are derived from both external
world economy as a whole and produces
and internal economies of scale. As he
mutual benefits to the trading nations
notes, "Even if economies of scale are (Krugman 1994d). This view is partly
internal to firms, internal economies in founded on his belief that comparative
the production of intermediate inputs can advantage still remains important and
behave like external economies for the useful. It is also based on his belief that
firms which buy them" (Krugman 1981, the "new" trade theory's recognition of
151). Furthermore, in a recent paper on
externalities and imperfect competition
metropolitan location (Krugman 1993c),
highlights the potential gains from eco-
he demonstrates that the "centripetal nomic integration. Increased trade may
forces" which hold a city together can allow
be greater economies of scale through
derived from the interaction of economies
rationalization and, in other situations, it
of scale at the plant level with transport may have a beneficial effect on oligopolis-
costs. Thus, he argues, local external tic markets by increasing competition.
economies do not have to be assumed; The complexity of much of Krugman's
instead, they are again derived from work also reflects the fact that the
market-size or market-potential effects. existence of significant externalities and
The key point to note about these nonconstant returns also opens up possi-
arguments is that internal economies of ble ways in which increased trade and
scale, by increasing the incentive for firms integration may have adverse effects.
to concentrate on one site, intensify the Krugman (1989) highlights two main
tendency toward the geographical concen- sources of adverse effects. The first is the
tration of production. Thus Krugman possibility of the uneven distribution of
associates the rise of the North American benefits associated with the existence of
manufacturing belt with the rise of the excess returns in imperfectly competitive
Chandlerian corporation. industries. A country that gains a dispro-
Krugman's analysis emphasizes that portionate share of high-returns industries
externalities operating within and be- can gain at others' expense, raising the
tween industries in these regional ag- possibility that trade policies designed to
glomerations make a difference to the
l In a similar fashion, Porter (1990) argues

10 Krugman also praises Fujita (1989) for his that the geographical concentration of leading
emphasis on market-size effects as explana- industries often reinforces and intensifies their
tions of urban agglomeration. competitive advantage.
foster these industries will lead to trade Krugman's "geographical economics"

conflict. Hence, "While the possibility of shows the positive effects of agglomera-
actual losses from trade is probably purely tion on productivity, it is also particularly
academic, there is a real issue of conflict well suited to explaining these possible
over the division of the gains" (Krugman adverse consequences.
1989, 361). (For a global example, see
Krugman and Venables 1994.) There is
clearly also a regional dimension to this Krugman's Geographical
problem of uneven distribution. As a Economics and Economic
result of the importance of external Geography: A
economies and the accumulated, path- Critical Comparison
dependent advantages of certain regions,
it is possible that these leading regions Clearly, Krugman shares an interest in
will capture a disproportionate share of regional agglomeration and the geograph-
the benefits of increased integration. A ical consequences of trade with many
major obstacle to integration, in this view, economic geographers. At the same time,
is that its benefits are not equally shared his treatment of these issues has been
across regions within countries. significantly different from the approaches
This forms the basis of the second set of pursued in economic geography in recent
adverse consequences identified by Krug- years. In this part of the paper we shall
man, namely adjustment costs. He argues examine the most important of these
that although it is costly for capital and differences and consider the lessons that
labor to shift into new industries, these Krugman and economic geographers can
costs represent a type of investment. They learn from each other. As we have noted
may be deserving of compensation, but already, a fundamental difference be-
they are not reasons to prevent or delay tween Krugman's geographical economics
change. On the other hand, where these and the various schools of contemporary
adjustments involve significant social economic geography is one of method.
costs, most importantly unemployment, Krugman's reliance on formal models
he concedes that this may provide a case means that his work is rigorous and
against moving too fast. The possibility of supported by mathematical proofs. In his
adjustment costs becoming real social view, the dependence of many of these
costs should not be dismissed lightly. One models on unrealistic assumptions is not a
ameliorating factor that Krugman notes is grave problem nor a serious limitation.
that the growth of trade between the Instead, he appears to regard these
industrialized nations since the Second models as rough metaphors or representa-
World War, especially within Europe, has tions of the core of real world problems
largely been intraindustry trade. "Thus (Krugman 1995). When the models' re-
sults are found to be inadequate, their
the specialisation that took place as trade
in manufactured goods grew tended to assumptions can be modified. In contrast,
involve concentration on different niches most contemporary economic geography
within sectors rather than wholesale con- has abandoned the use of formal modeling
centration of different countries on differ- and is dominated instead by various types
ent industries" (Krugman 1989, 364). of political economy, which aim, above all,
There is clearly a tension in Krugman's to be "realistic." From this perspective,
work between his positive evaluation of Krugman's models have an inadequate
trade and integration in general and his sense of geographical and historical con-
demonstration that significant adverse text. Knox and Agnew (1994), for example,
consequences are possible. As we have argue that Krugman's core-periphery
seen, the question of uneven distribution model in Geography and Trade differs
of the benefits and costs is probably the from other location models in that it does
most important of these. Indeed, while not suggest a long-term process of conver-
gence. Instead, "the long term never ing of Krugman's approach remains un-
arrives" (Knox and Agnew 1994, 83). clear and clouded by ambiguity. What is
There are multiple equilibria as concen- clear is that his emphasis on continuity in
trations persist for long periods of time the forces responsible for capital's agglom-
but may then be unraveled by new eration contrasts with economic geogra-
patterns of concentration. However, Knox phers' focus on historical patterns of
and Agnew insist that restructuring. However, the relative mer-
concentration somewhere . . . is the its of this more historical approach de-
perpetual rule. So though apparently atten-
pend on precisely how change is theo-
tive to historical change, this model isrized and explained. It is impossible here
static
to talk
in its assumptions about the operation of about economic geography as a
economic-locational principles. The whole;
same we have therefore selected two
relevant areas of work, namely the recent
principles of increasing returns, imperfect
competition, and agglomeration are atliterature
work in industrial geography on
in the same way all the time. Fromregional
this agglomeration and recent writing
point of view, geographical outcomes can
on theorizing the geography of trade.'3
change but the process driving them does
not. (1994, 83; original emphasis)
The Resurgence of Regional Economies
While this phrasing may be too
strong,12 Krugman (1991a) clearly states During the past decade, the most
that the patterns of concentration that he influential approach to industrial organiza-
describes are typical only of some indus- tion within economic geography has been
tries under certain conditions; neverthe- the notion of a fundamental transition
less, it does identify an important weak- from Fordist mass production to more
ness in Krugman's work. He claims that flexible production methods, such as
the same broad locational forces which flexible specialization. Scott and Storper
explain the growth of nineteenth-century(1992a, 1992b), Scott (1988), Storper and
concentrations also underlie the contin- Walker (1989), and others have argued
ued tendency to agglomeration. Indeed,that internal economies of scale and scope
this is one reason why he is reluctant to
have been undermined by increased
emphasize technological spillovers as market
a uncertainty and technological
key determinant of contemporary clus- change. They argue that the response has
ters. At the same time, however, Krug- been horizontal and vertical disintegra-
man makes several passing references to tion, or an externalization of production,
the way in which the nature of agglomer-which enables a greater ability to meet
ation has changed over time. Thus he
differentiated demand and a greater
suggests, in one paper, that the railway adaptability to market forces. Where a
and steamship were responsible for themultiplicity of linkages are created which
emergence of core-periphery distinctions have geographically sensitive transaction
and that the age of this type of divergence
costs, externalization is positively related
may have passed. But such a "throw-to agglomeration. In this view, "Agglom-
away" suggestion requires a great deal eration is a strategy whereby producers
more explanation. The historical ground- ease the tasks of transactional interaction
because proximity translates into lower
12 After all, most of the main schools of

economic geography and political economy 13These are two of the leading fields in
may be criticized on similar grounds. All, for
contemporary (post-Marxist) economic geogra-
example, assume that the basic laws of phy. It would require another paper to
economic development (as they perceive consider Krugman's work in relation to the
them) remain essentially unchanged as capital- complete corpus of geographical work on
ism evolves over historical time. uneven regional development.
costs and wider opportunities for match- disintegration and therefore a movement
ing needs and capabilities" (Scott and back toward perfect competition. How-
Storper 1992b, 13). In summary, the shift ever, the idea that corporate disintegra-
to flexible specialization has been respon- tion is a necessary response to uncertainty
sible for the rise of new industrial districts can be criticized (see, for example, Lover-
and for the new, or renewed, significance ing 1990; Phelps 1992). Moreover, as
of regional agglomeration (Sabel 1989). Phelps (1992) has argued, Scott's analysis
While there are many contrasts between of the causes of agglomeration pertains
this "new industrial geography" and primarily to situations approximating to
Krugman's geographical economics (see that of perfect competition. In Phelps's
Table 1), we shall focus on three issues: view, "The assumption of near-perfect
the treatment of industrial and market competition is otherwise implicit in an
structure, of externalities, and of nonmar-analysis which applies to single plant firms
ket transactions and relations. and neglects considerations of differential
As we have seen, Krugman tends to economic power embodied in linkage
rely on several abstract models of monop- structures" (1992, 41). This is especially
olistic and oligopolistic market structure. problematic when the analysis is applied
The assumptions of these models are in to international trade. As Markusen (1993,
some ways unrealistic, but the underlying 287) writes, "Most important internation-
rationale is that they are useful because of ally traded industries are now multinucle-
the pervasive presence of imperfect com- ated, with large national firms thrust into
petition. In contrast, the flexible special- more spirited competition with similarly
ization approach has envisaged a new type sized and politically well-endowed firms
of competition involving downsizing and from other nations." Hence, "the number
of players is relatively small, their sizes
and clout are varied, and none of them is
able 1
A Comparison of Krugman's "Geographical Economics" with the "New Industrial

Geography"
New Industrial
Krugman Geography
Externalities Marshallian, especially Marshallian trio
labor pooling, specialist Labor market
suppliers Specialist suppliers
"Pecuniary" market-size Technological and
effects knowledge spillovers
Agglomeration Local clusters Industrial districts
Interregional center- Craft-based
periphery pattern High-tech
Financial centers
Competition Imperfect: monopolistic Competitive flexible
and oligopolistic; economies specialization;
of scale economies of scope
Transfer costs Transport, including Transactions costs
trade barriers
Technological Not typical, but important Local and fundamental
spillovers in some industries; local to innovatory success
and international in high-tech clusters
Labor market pooling Strategy of insurance Form of local social
against risk (both employers embeddedness
and employees)
Social and cultural Difficult to formalize and Key preconditions
characteristics of assumed a priori; best left to for successful
clusters sociologists localization
unaware of the behaviour of its neighbors. rates of growth and occur with imperfect
These are characteristics of oligopolized competition. Again, there are examples
markets, not perfectly competitive ones." from endogenous growth theory: Romer
Several arguments have been used to (1990) sees investment in R&D in a
support the association of near-perfect situation of monopolistic competition as
competition with agglomeration. One is generating externalities in disembodied
the finding that in some industries and knowledge. The third type of externality
places larger producers are located away arises from demand spillovers between
from local clusters of industry (Hoare sectors and industries. Murphy, Schleifer,
1975; Scott 1986), and another is the and Vishny (1989), for example, argue that
observation that the decline of some there are low-level equilibrium traps,
industrial districts has been associatedwhere industrialization remains unprofit-
able. Industrial production only becomes
with the concentration of production into
larger firms (Steed 1971). These areprofitable for individual firms in the
contingent findings, however. For exam- context of more general demand linkages.
ple, Scott (1992a) notes that large produc-This framework provides a means of
ers are integral to the Southern Califor-
comparing the two approaches to agglom-
eration (Table 2). Krugman's focus on
nian computer districts. Even where large
firms are not found in local industrial market-size effects is clearly closest to,
clusters, they may be central to theand draws most heavily on, pecuniary
externalities.14 As we have seen, his
regional and metropolitan concentrations
modeled by Krugman. It cannot be explanation of local clustering also invokes
assumed that internal economies of scale
certain types of Marshallian external
and scope act against agglomeration. economy. Importantly, he has tended to
Indeed, the intraindustry trade literaturedownplay the significance of externalities
implies that, with increasing productbased on spillovers in technological
diversity, internal economies and agglom- knowledge. Krugman (1987c) describes
eration become more closely linked. these as "elusive," preferring to concen-
There is clearly a need to research thetrate on externalities that can be modeled.
relations between market structure and The difference between his approach and
locational dynamics in more detail. that of the "new industrial geography" is
This difference on the issue of competi- apparent. In accordance with the reliance
tion has important consequences for un-on situations of near-perfect competition,
derstanding externalities. In order toMarshallian external economies have
compare Krugman's representation of been at the forefront of the industrial
externalities with that used in the "new districts literature. As Phelps has argued,
industrial geography," it is helpful to setand as Krugman's account demonstrates,
both approaches within a general frame-the external economies that can be used
work. De Melo and Robinson (1990) arguein this approach are only a subset of the
that three main approaches to externali-range available (Phelps 1992). To a certain
ties are apparent in recent economic extent this limitation has been weakened
literature. The first is the Marshallian by those revisionist studies of industrial
externalities approach, which we dis-
cussed above. They suggest that some
parts of endogenous growth theory fall
within this approach. For example, in an14 Krugman (1993b, 1995) describes this
type of external economy as similar to those
article on the externalities arising from
envisaged in "Big Push" interpretations of
human capital formation, Lucas (1988)
industrialization. He argues that a large-scale
talked about increasing returns at an
program of industrialization can take advan-
economywide level. The second type of tage of external economies and complementa-
externality that De Melo and Robinson rities and so reduce the risk of investment (see
identify are those that result in uneven
Rosenstein-Rodan 1943).
Table 2
Comparison of the Treatment of Externalities in Krugman's

"Geographical Economics" and the "New Industrial Geography"
Application to Agglomeration
Type of Externality Krugman New Industrial Geography

Marshallian Local clusters of Districts associated with
external economies industry associated vertical distintegration
with market-size effects and transaction costs;
(labor pooling and technological spillovers
specialist suppliers) important
and with internal
economies
Knowledge and Important in some (Not typical;
technological spillovers industries, but not where present large
under imperfect typical and difficult producers tend to adopt
competition to model- "too decentralized and flexible
fashionable" organizational forms)
Pecuniary externalities Regional (Typically regarded as
(demand and supply specialization and Marshallian; much more
spillovers) concentration on a grand emphasis on nonmarket
scale (center-periphery) conditions)
through interaction of
market size, demand,
and transport costs
districts which argue that large producers those of economic geographers is the
can imitate decentralization. However, manner in which the increasing power of
the contradiction between a commitment larger producers has been related to
to perfect competition and the depen- contemporary localizations of industry. In
dence of Schumpeterian models of "cre- economic geography there has been some
ative destruction" and local technological dissatisfaction with the way in which the
spillovers on imperfect competition can- flexible specialization literature has ig-
not be resolved easily.15 nored the increasing internationalization
The differences between Krugman's of firm structures and globalization more
geographical economics and the recent generally (Amin and Robins 1990; Gertler
work in economic geography on regional 1992). Consequently, there has been an
development are not confined to indus- interest in the way in which large firms
trial structure and externalities, but also interact with industrial districts. In con-
extend to the question of nonmarket trast to Krugman's market-size effects,
transactions. Thus another important con- however, the main emphasis has been on
trast between Krugman's approach and the intermingling of firm and local net-
works (Amin and Thrift 1992; Grabher
1993). Networks have usually been de-
15 Schumpeterian models of "creative defined as types of organizational relation
struction," technological spillovers, and endog- that are neither market transactions nor
enous growth depend on imperfect competi- hierarchies, and the term has been used
tion. Typically, the incentive for firms to
to refer to cooperative and mutually
develop new products and processes stems
from the temporary monopoly profits which
beneficial relationships among producers
they can earn (Grossman and Helpman 1991; (Cooke and Morgan 1993). Using this
Aghion and Howitt 1993). This sits uneasily definition, the boundaries of firms be-
with the new industrial geography's emphasis come blurred, and firms and districts
on near-perfect competition. become intermingled. On the one hand,
Krugman's contrary emphasis on pecuni- tance to envisage nonmarket linkages

ary relations is a reminder to geographers seems to conflict with his commitment to
not to lose sight of market effects. But on new-Keynesian economics, where expec-
the other hand, Krugman's neglect of tations and conventions are central. He
externalities that are intangible and leave himself shows (Krugman 1991c) that,
no paper trail appears too restrictive. As under certain conditions, expectations
Jaffe, Trajtenberg, and Henderson (1993) may affect the course of regional develop-
have pointed out, knowledge flows do ment. But, if they are to be understood,
sometimes leave a paper trail, in the form expectations cannot be treated as exoge-
of citation of patents. nous "animal spirits." Rather, they are an
The interest in network forms of integral part of social conventions and
organization in economic geographymeanings,
re- and their formation should be
flects a more general concern to examine
an important area for regional research.
Our
the ways in which economic activities conclusion, then, is that there is a
are
"embedded" in, and made possibleneed by, for a greater exchange of ideas
social and cultural conditions. This has
between Krugman's work and the geo-
been applied with particular force bygraphical literature. But this applies not
Storper (1992a) to high-technology dis-only to research on regional agglomera-
tricts. As Harrison (1992) notes, this tion but also to that on trade more
interest in embeddedness has been the generally.
distinctive contribution of the recent
geographical literature on industrial dis-
The New Political Economy of Trade
tricts.16 This stands in complete contrast
to Krugman's rejection of invisible exter-Recent years have seen a growing
nalities. However, as Storper (1992b) interest by economic geographers in the
argues, in the context of increasing market spatial patterns of international trade.
While this work has lacked a comprehen-
contestability, it is difficult to explain the
continuing competitive advantage of cer- sive theoretical framework, it has never-
tain districts if their conventions, rules theless
of been characterized by shared
behavior, and implicit accords are not themes revolving around the inability of
taken into account. Conversely, the de- conventional geographies, based on Ricar-
dian comparative advantage, to explain
cline of other regions appears to be partly
a result of the "lock-in" of outmoded fully the complex character of contempo-
conventions and rules of behavior (Grab- rary international patterns. In accord with
her 1993). Krugman's rejection of non- the "new trade theory," this geographical
revival has stressed the importance of
market linkages seems to be made primar-
ily on the grounds that if externalities
shifts in the world economy and the rise of
cannot be modeled then they have tointraindustry
be and intracorporate trade.
assumed a priori, so that the analyst can
One of the defining features of this revival
say anything she or he likes about typeshas
of been a call to study the ways in which
spillover. But this rules out other methods
the geography of trade is shaped by states
of research and more sociological ap-
and by trade regimes. This emphasis on
proaches. Moreover, Krugman's reluc-
state policy, and the interpretation of
trade on which it is based, contrasts with
Krugman's approach in ways that raise
fundamental questions about the effects of
1Amin and Thrift (1994) describe this
trade and its policy implications.
embedding as best summed up by the phrase
"institutional thickness." This is defined by Some
a years ago, Johnston (1989) called
for trade to be explained as part of a
strong institutional presence in a local area,
holistic theory of uneven development
high levels of interaction among these institu-
that combines the logics of capitalism and
tions, strong social structures, and a collective
awareness of common enterprise. the policies of states. To some degree, his
plea for an enlarged research agenda has view, factor endowments can be used to
begun to be recognized. Grant summa- explain trade in minerals, agricultural
rizes recent developments as follows: goods, and some labor-intensive con-
sumer goods, but the majority of trade
The unifying theme in newer approaches is
their study of the interactions between between developed market economies
governments and firms and their connec- can only be explained by a "dynamic
tions to trade and industrial policy within revisionist" theory. This has four major
the context of a politically and economically tenets that contradict orthodox trade
competitive world economy, one in which theory. First, the mix of sectors matters,
governments attempt to "create" the most as some industries have greater growth
advantageous environment for national and productivity differentials. Second,
business. Accordingly, approaches recon- growth is not constrained by factors but
textualize comparative advantage to include
by demand for the product. Third, in
an understanding of developments in the
some industries rapid growth leads to
trade-industrial policy arena. (1994, 301)
continuing success due to increasing
In line with this theme, Grant focuses returns. Fourth, because of the existence
on the role of governments, especially the of increasing returns, comparative advan-
formation of regional blocs, and the role of tage may conceivably be created by
firms as the bases of a more comprehen- strategic intervention on the part of
sive theory. Moreover, he argues that nation-states and regional authorities. On
high-technology trade occupies a key this basis they argue that the orthodox
place in any new theory, as success in view that free trade means growth for all
regions is mistaken; instead, "there is
high-technology bestows national benefits
on productivity and high-wage job cre- some danger that the unfettered pursuit
ation (see also Drache and Gertler 1991). of free trade will actually depress wages
Likewise, in their recent study of trade in and employment and lower world living
textiles and clothing, Glasmeier, Thomp- standards" (Howes and Markusen 1993,
son, and Kays (1993) contend that it is 35). Furthermore, Markusen (1993) ar-
necessary to understand how the actions gues that in the United States free trade
of the state influence the structure of and laissez-faire strategies have produced
global competition. Indeed, they conclude persistent unemployment and a waste of
that state actions have superseded market infrastructure.
forces as the regulator of the industry'sWhile this "dynamic revisionist" theory
geographical evolution. shares an emphasis on "new trade theory"
The conceptual movement away from with Krugman, it more closely resembles
orthodox comparative advantage explana- the strategic trade views of authors such
tions has been most fully spelled out in as Tyson (1992) and Reich (1991), whom
Trading Industries, Trading Regions, ed- Krugman (1994a, 1994c) has recently
ited by Noponen, Graham, and Markusen criticized.17 As we have noted, Krugman
(1993). Here again it is argued that remains convinced that the mutual bene-
success in trade is fundamentally shaped fits of greater international trade outweigh
by government intervention. In a chapter the costs. Moreover, in his opinion,
in that volume, Howes and Markusen comparative advantage is not just a
claim that governments have played a key sector-specific theory, it remains a general
role in creating and maintaining industrial principle that explains the beneficial
leadership, and that "in a world with
governments successfully conducting such
industrial and trade policies, open econo- 17 For a debate on Krugman's critique see
mies without such efforts will find them-
the discussion on "The Fight over Competi-
selves the targets of import penetration tiveness" in Foreign Affairs (1994a), Friedman
and potential export market shrinkage" (1994), and The Economist ("The Economics of
Howes and Markusen 1993, 4). In this Meaning" 1994).
consequences of trade. The concept overs which some authors have suggested
makes clear that absolute productivity (for example, Hanink 1994).
advantage in some areas is not necessary The second question is whether Krug-
for a country to gain from economic man underestimates the significance of
integration. Trade, therefore, is not a adjustment costs and the obstacles to
zero-sum game, so that concerns about regional adjustment. On the one hand,
national competitiveness are misplaced Krugman is committed to a nonequilib-
and unfounded. Krugman (1987b) con- rium view of economic geography in
cedes that the intellectual case for free which there is no process of convergence
trade has been weakened and that it is not to a spatial equilibrium where all factors
are equally rewarded. He rejects the
an absolute ideal, but he believes that it is
still the best general policy or rule of neoclassical faith in the efficiency of
thumb. But Krugman's position faces markets on the grounds that the collective
several key questions. The first is the result of individual choices may be to
extent to which this continuing use of "lock-in" a bad result. On the other hand,
comparative advantage is compatible with in a methodological sense, Krugman
his own emphasis on the pervasive (1993a) insists that all economic models
presence of increasing returns. Kaldor should contain a well-specified equilib-
(1985), for example, argues that the rium. By this he means that they should
presence of increasing and diminishing specify how individuals behave and show
returns conflicts with the basic tenets of how market outcomes emerge from the
Ricardian comparative advantage. Simply interaction of these individual behaviors
put, he contends that diminishing returns (Krugman 1993a, 115-16). He holds these
may mean that the resources released by two opposing convictions together, it
trade will not necessarily be employed in seems, through a commitment to a "new
other sectors, so that there is a real Keynesian" brand of economics. Accord-
possibility of absolute loss (a "negative ing to this, economic trends and patterns
sum" game). Conversely, increasing re- are the products of innumerable individ-
turns in some industries may inhibit the ual decisions, but these decisions are not
transfer of resources elsewhere. Krug- perfectly rational and informed. Instead
man's economic geography pays insuffi- they are frequently both near-rational and
cient attention to these problems. This is individually reasonable and sensible.
reflected by his insistence that it is However, in imperfectly competitive mar-
pointless to try to identify high-return kets the aggregate result will be unstable
sectors, so that the mix of sectors does not and irrational. In his words, "What look
really matter.18 Given his insistence on like highly irrational outcomes in the
the importance of productivity, it is marketplace are caused by the interaction
surprising that he devotes little attention between imperfectly competitive markets
to the extent to which high-technology and slightly less than perfectly rational
sectors do generate the productivity spill- individuals" (Krugman 1994c, 213). But if
emphasis is placed on the second of these
factors, then the position is readily reas-
similated into a neoclassical view of the
18 Krugman's (1994c) argument is that it is economy. It lends itself to the view that
wrong to assume that high-technology sectors markets would adapt efficently and rap-
such as computers and aerospace are the idly if only people would behave ration-
sectors with highest value added per worker.
In fact, he notes that in the United States the ally. This is exemplified, perhaps, by
real high-value industries are extremely capi- Krugman's (1993e) argument that Euro-
tal-intensive sectors, such as cigarettes and oil sclersosis, or the problem of a persistently
high level of unemployment in Europe,
refining. This says nothing, however, about the
possibility of positive spillovers from the has been caused by the effects of welfare
high-technology sectors. states on labor markets. The whole
question of adjustment to the effects of European Commission (Commission of

trade is one that Krugman has recently the European Communities 1991, 1994).
considered explicitly in terms of the In contrast, others argue that economic
impact of economic integration on re- integration will intensify rather than
gional development, particularly in the reduce regional imbalances in growth and
European Union, and it is to this aspect of income across the European Union. In-
his work that we now turn. stead of leading to equalizing centrifugal
movements of firms and investment to-
ward depressed and peripheral regions
Krugman's Model of Economic within the European Union, economic
Integration and Regional integration is likely to stimulate a spatial
Development: The Lessons of the reconfiguration of economic activity in
United States for Europe? favor of growth regions precisely because
these are the areas that already enjoy
The regional consequences of European
greater comparative advantage in terms of
economic integration is an issue that has
access to markets, inputs, expertise, and
attracted surprisingly little attention from business infrastructure.20
economic geographers. At the heart of this
Krugman falls into the second of these
issue is the question of what the impact of
two camps, although he appears to sub-
progressive economic and monetary inte- scribe to two somewhat different models
gration in the European Union (EU) will
of regional divergence. In an earlier paper
be on regional patterns of economic
(Krugman and Venables 1990), he follows
growth, employment, and income across
member states. Economists have offered a core-periphery argument not unlike that
in Geography and Trade. Although the
two opposing answers to this question. On removal of barriers to trade and move-
the one hand, there are those who believe
ment of capital and labor within the
that the free movement of goods, services,
European Union will increase the inflow
and capital associated with European
of capital into, and the relative competi-
economic and monetary integration
tiveness of, the low-wage peripheral re-
(EMU) should lead to regional conver-
gions, given transport costs this centrifu-
gence, not only in factor returns and
gal process is on balance likely to be
economic performance but also in eco-
outweighed by further concentration of
nomic structure. To the extent that wages
and other costs are lower in the less industry and employment in the high-
wage core regions, because these areas
productive and slower-growing regions,
have the largest markets, well-developed
the removal of barriers to trade and factor
external economies and infrastructures,
movements, it is argued, should enable
and a comparative advantage in terms of
industries and services in these regions to
better exercise this comparative advan-
tage and to attract increased flows of
20 Because the gains foreseen from comple-
capital investment.19 This optimistic sce-
tion of the internal market are thought to be
nario is, on balance, the view taken by the
generated mainly endogenously, the various
processes of resource allocation are bound to
cumulate resources in the leading core re-
19 Additionally, economic integration repre-
gions. It is the historically established compet-
sents a major supply shock to such regions,
itive advantage of the growth regions which
since it exposes them to the full force of
enables them to capture a disproportionate
competition elsewhere in the system. Suchshare of the benefits of economic integration.
As for the depressed and lagging regions,
shocks, the argument continues, should (allow-
economic integration is seen as bringing
ing for adjustment lags) eliminate inefficient
firms, work practices, and products in de- prolonged problems of adjustment and the
pressed regions and improve their supply-side
need for greater levels of spending on regional
competitiveness and flexibility. policies.
relative accessibility. His second approach the increased factor mobility that integra-
is more emphatic, but different in its tion will promote, such region-specific
specific arguments. In his paper on the shocks will lead to divergent long-term
"Lessons of Massachusetts for EMU," he regional growth paths. Thus, fourth, given
supports the movement toward European that under EMU member states will no
economic integration as "a generally good longer be able to use the exchange rate
thing," but argues that it will lead to mechanism as a policy instrument (see
greater regional instability and divergence also Krugman 1989), the only way regional
of regional growth rates (Krugman 1993d, adjustment problems can be ameliorated
241). In developing this thesis he begins is by transferring a significant part of
by drawing on his earlier ideas on trade national budgets to the European Union
and the localization of industries that we to allow fiscal federalism to function as an
have discussed above: automatic stabilizer.
Thus, in contrast to his previous work-
For regional issues . . .in the EC, . the for example, in Geography and Trade
key aspect of regional specialisation is the
(1991a) and Krugman and Venables
dependence of regional economies on ex-
port clusters held together by Marshallian (1990)-Krugman argues that the process
external economies ... Are such regional of uneven regional development that
clusters more likely to form in a more EMU may be expected to produce will
integrated economy? The answer is defi- not be one of cumulative divergence into
nitely yes. (Krugman 1993d, 244) a core-periphery pattern. He believes that
the forces generating this form of uneven
These ideas are then used in a some-
regional development have probably
what different way from his earlier work
reached their limit in advanced industrial
to produce a theoretical account that not
nations; indeed, he suggests that in both
only carries over some of the problems we
the United States and Europe industrial
have already highlighted, but also intro-
activity is becoming much more evenly
duces additional elements of contention.
distributed geographically (Krugman
The gist of this second model may be 1993d). Rather, in his view the process
summarized as follows. First, given the
will be one of increasing regional export
existence of increasing returns, the expan-
specialization, with the result that the
sion of interregional trade that EMU will
pattern of regional growth and decline
bring about will lead to greater regional will be more unpredictable, dependent on
industrial concentration and specialization
the particular incidence of random de-
along essentially arbitrary lines. Once mand shocks. Hence, unlike the argument
under way, there will be a tendency for
in Geography and Trade, past regional
this regional specialization process to success need not be self-reinforcing, and
become "locked in" by the operation of
even prosperous regions may experience
location-specific external economies. Sec- sudden reversals of fortune.
ond, Krugman argues that this increased
Another distinctive feature of Krug-
regional specialization will render the man's exposition is the method he uses to
European regions much more subject to support his theory empirically. The
random, idiosyncratic demand and tech- United States is taken to be the sort of
nology shocks, so that region-specific integrated economic and monetary unit
recessions and crises will be more likely
which the European Union is seeking to
to occur.21 Third, when combined with
emulate, so that regional experiences in
the former are considered to be a good
guide as to what to expect in the latter.
21 In an earlier paper, Krugman (1989)
stressed that increasing interdependence in
Europe acts as a buffer against regional and acts against locally generated recessions such
national shocks, but this buffering effect only as those caused by investment slumps.
Using simple measures of the dispersion geographical scale at which local external
of economic structure, Krugman (1991a, economies and the processes leading to
1993d) finds that the broad regions of theindustrial clustering actually operate. The
United States are more industrially spe- basic point is that the analysis of localiza-
cialized than are European countries. tion economies requires an identification
Furthermore, a comparison of Belgium of the relevant regions as economic areas
with the state of Ohio is used to suggest and the relevant level of industrial disag-
that the regional employment growth gregation at which to measure the extent
rates in the United States are more of geographical concentration and special-
unstable than in the European Union. ization.
He The geographical literature on
"new" flexible industrial districts indicates
then examines the disparities in long-term
growth rates between certain states inthat
the such clusters are in fact quite
United States and among the main EU
localized, and far smaller than the broad
countries and finds that these disparities
spatial units used by Krugman. Certainly
are larger in the United States than in European Union, local differences
in the
Europe. In addition, he uses the recent
in economic structure and economic
economic slump in the New England growth rates within member countries
region of the United States as an illustra-
(for example, at the so-called NUTS1 and
NUTS2 level regions) are much larger
tion of how, in a monetary union, regional
than
industrial specialization can give rise to the disparities between countries
(Collier
pronounced local instability in the face of 1994; Dunford 1993; Dunford and
region-specific demand shocks, andKafkalas
how 1992). Likewise, as von Hagen
such shocks can lead to permanently and Hammond (1994) argue, the metro-
politan rather than the state or broad
lower levels of employment (Krugman
1993d). These various empirical results
regional level is the most meaningful one
are taken as lending support to his thesis
for analyzing geographical differences in
that increased market integration inindustrial
the concentration and localization
European Union will lead to more re-
economies in the United States, a view to
gional specialization and unequal growth.
which, as we have already noted, Krug-
man has elsewhere subscribed. These
In our view, however, Krugman's empiri-
cal examples and findings are not of
findings imply that Krugman's method of
themselves sufficient to prove his comparing
case, the European Union with the
and several features of his analysisUnited
are States will generate different
problematic. results according to the geographical scale
used to define economic regions in the
Economic Integration and two areas. Indeed, it may even be that at
Regional Specialization some geographical scales regional special-
ization is not in fact greater in the
The first problem concerns the evi- more-integrated economy of the United
dence on regional specialization. What is States than in the European Union.
the "regional" scale being referred to? In any case, is increasing regional
The "regions" used by Krugman in his industrial specialization an inevitable out-
comparisons of regional specialization and come of economic integration? While the
regional growth rate disparities in the existence of external economies and local-
United States and the European Union ization economies in the European Union
are extremely aggregate ones: the four could well lead to the increased regional
"Great Regions" and individual states in specialization that Krugman predicts
the former and whole countries in the (Baldwin and Lyons 1990; Cabellero and
latter. Krugman argues that these spatial Lyons 1990, 1991; Martin and Rogers
units are of roughly similar size, and thus1994a, 1994b), some observers have ar-
broadly comparable. That may be so, but gued that product market integration in
they do not necessarily represent the
the European Union will increase the
scope of intraindustry trade there still example, Bayoumi and Eichengreen 1993;
further, and that this is likely to render Palmini and Cray 1992; von Hagen and
regional industrial structures increasingly Hammond 1994). There seems to be no
similar over time (Commission of the simple relationship between economic
European Communities 1991; Eichen- integration, regional specialization, and
green 1993; Emerson, Anjean, and Cati- regional shocks. Both the pattern and
nat 1988). Indeed, possible evidence for severity of shocks will depend not only on
this effect is provided for the United the degree and geography of regional
States by Krugman: as he shows, U.S. industrial specialization, but also on how
statistics indicate that regional specializa- such shocks are transmitted between
tion there has actually been declining regions (for example, through interre-
since the Second World War (Krugman gional input-output linkages and the
1991a, Chap. 3). He suggests that this may impact of government policies) and on
be a statistical illusion, in that specializa- how flexible regional labor markets are in
tion may have become more difficult to adjusting to disturbances. In short, much
measure but may not necessarily be less more theoretical and empirical analysis of
in fact. However, there is also evidence regional industrial specialization within
from Europe that economic integration both the United States and the European
and increased trade lead to regional Union is required before the former can
industrial diversification rather than spe- be taken as a guide of what to expect in
cialization (Peschel 1982). Indeed, both the latter.
the definition of regional economic "spe-
cialization" and the question of how
specialization actually influences regional Economic Integration and Divergent
instability are not straightforward issues.22 Regional Growth
As a number of writers have shown, the
This last point links with the third
empirical patterns of regional shocks in
element of Krugman's thesis, that demand
both the United States and the European
shocks in an integrated Europe will have
Union appear to be more complex than
permanent regional growth effects, in the
those posited by Krugman (see, for
same way that temporary policies may
have long-term implications (Krugman
22 There is a sizable literature on this topic, 1987a). Suppose a region experiences a
decline in the demand for its clusters of
although it is not referred to by Krugman (for
example, see Barth, Kraft, and Wiest 1975; export industries. This would put down-
Conroy 1975; Brewer 1984; Jackson 1984; ward pressure on relative wages and other
Kurre and Weller 1989). Much of this is based factor costs in the region. If relative wages
on what is called a "portfolio" approach to the and other costs fall, this would help to
analysis of regional industrial specialization. restore the region's competitiveness vis-
This type of analysis, first applied to regional a-vis other, higher-cost regions, so that
economics by Conroy (1975), borrows the new industries would be attracted there
concepts of expected return and risk from
theories of the optimal diversification of
and demand and growth should be
financial portfolios developed by Markowitz restored. As Krugman puts it,
(1959). The regional industrial structure may Regions that have been unlucky in their
be conceptualized as a "portfolio" which heritage of industries from the past will
provides "returns" to the region in the form of have lower costs than lucky regions, and
employment, income, and tax revenues. These will therefore be more likely to break into
returns are associated with risk-arising from industries in the future. We would expect
demand and technology shocks-as repre- this process to put limits on the extent of
sented by the variance and covariance in the regional divergence in growth. (Krugman
returns. It is this measure of risk, "the portfolio 1993d, 248)
variance," which measures the degree of
instability of the region. Unfortunately, however, according to
Krugman labor mobility prevents the curious that Krugman is at pains to

wage flexibility mechanism from bringing distinguish his model of uneven regional
regional growth rates into balance in this development in the European Union not
self-correcting way. To the contrary: only from "core-periphery" models of
cumulative concentration but also from
An unfortunate region will not have lower
factor prices for very long: capital and "local endogenous growth" models (Krug-
labour will move to other regions until man 1993d). His own model implies a
payments are equalized. This means, how- similar cumulative divergent growth
ever, that there is no particular reason to mechanism, at least in the sense that
expect a region whose traditional industries interregional shifts in labor prevent the
are faring badly to attract new industries. It reequilibration of regional growth rates.
can simply shed people instead. The impli- The question mark over his analysis is
cation is that relative output and employ- exactly how far labor mobility will in-
ment of regions should look more like a crease in an integrated Europe. Although
random walk than like a process that interregional migration across national
returns to some norm. (Krugman 1993d,
248)
borders will in principle be unrestricted,
there are further reasons to doubt
In developing this argument, Krugman whether labor mobility will ever reach the
draws on Blanchard and Katz's (1992) levels found in the United States. The
study of patterns of growth among U.S. marked cultural and language differences
states. According to these authors, while across Europe will continue to be a
employment growth rates differ consis- significant barrier to migration for many
tently across U.S. states, unemployment groups of workers. But if this form of
rates and wages vary much less, suggest- adjustment to regional shocks remains
ing that when states are hit by demand slow, where does this leave Krugman's
shocks workers react by relocating (see view of regional long-term growth differ-
also Barro and Sala-i-Martin 1992). There ences in an integrated European Union?
is no discernible tendency for states to The implication of his model is that if
recover lost jobs: relative regional unem- labor mobility is low, then local (down-
ployment returns to normal through the ward) relative wage flexibility will serve to
out-migration of workers. This would restrain the degree of divergence be-
seem to be in contrast to the adjustment tween regional growth rates. Unfortu-
process in the European Union, where nately, wages in the European Union do
historically factor mobility has tended tonot seem to be particularly flexible:
be far lower than in the United States andEuropean labor markets appear to be
regional unemployment disparities appear more rigid or "sclerotic" than their
to be characterized by greater hysteresis American counterparts, a point high-
(Eichengreen 1993). Krugman draws the lighted by Krugman (1993e). In the
obvious implication that if Europe moves European regions, adverse sectoral de-
toward U.S. levels of regional specializa- mand shocks trigger greater unemploy-
tion and factor mobility, disparities in ment, without the equilibrating mecha-
economic growth rates among countries nisms of labor migration or downward
and regions may be expected to increase. relative wage movements (the rigidity of
Labor mobility is thus central to Krug- the latter possibly reflects the consider-
man's model of divergent regional growth. ably higher rates of institutionalized wage
In this respect his analysis is similar to setting among workers and the availability
local "endogenous growth" models, in of more generous unemployment benefits
which labor mobility intensifies local in the EU countries compared to the
disparities in the accumulation of human United States). The Commission of the
capital and hence long-term development European Communities (1990) argues that
(Grossman and Helpman 1991; Bertola EMU, by increasing the credibility of
1993). In this respect we find it somewhat fiscal authorities' commitment not to bail
out depressed regions, should encourage Trade and the Regional

workers in such areas to moderate their
Policy Issue
wage claims, thus imparting greater local
wage flexibility. In practice, little isStrategic Trade Policy
known about how far regional relative Like much of his other work, Krug-
wages would have to fall in order to man's views on the role of trade and
stimulate capital inflows and the restora-industrial policy have shifted over time.
tion of employment. Equally, we stillIn his early writings he reacted against
know little about interregional productiv- the idea of targeted industrial policies, on
ity and technology spillovers, which maythe grounds that they were based on
offset the need for wage reductions (Jaffe,crude misconceptions and that even if
Trajtenberg, and Henderson 1993; Au-more-sophisticated theorizations could be
dretsch and Feldman 1994). In short, it isfound such policies were still unlikely to
by no means obvious whether increasingbe effective in practice (Krugman 1983,
integration in the European Union will 1983b, 1984). Not long after, however, he
lead to convergence or divergence ofhad constructed a sophisticated theoreti-
regional growth. The evidence so far cal argument for "strategic trade policy"
would seem to suggest that "club conver-(Krugman 1986). One of the most conten-
gence" may be the most likely outcome, tious aspects of the new trade theory has
with convergence within the northern,been the debate it has generated over the
core regions, on the one hand, and withinquestion of strategic industrial policy.
the southern and peripheral regions, onWhereas conventional trade theory denies
the other, but little or no convergence there is any case for "activist" trade or
between these subsets (Button and Pente-industrial policies, the new trade theory
cost 1993; Chatterji 1993; Neven and directly challenges the conventional view.
Gouyette 1994). According to Krugman (1986) and other
Thus, though suggestive, Krugman'snew trade theorists (for example, Brander
arguments about the impact of economicand Spencer 1983, 1985), an "activist"
integration on regional trade, specializa-
trade policy can benefit a country relative
tion, instability, and long-term growthto free trade in two ways. The first is
disparities in the European Union arethrough "rent creation." If a government
problematic and limited. Comparison be-can promote a new or expanded monopoly
tween the United States and the Euro- position for domestic factors of production
pean Union in terms of "regions" and
in industries that trade internationally,
their structures, shocks and reactions then
to a targeted industrial policy can in
them is not, perhaps, as valid as Krugmanprinciple raise a country's income at
and others (such as Eichengreen) assume. foreign expense. Second, targeting can
We do not have a counterfactual history raise income if there are certain industries
for the United States-that is, a picture in
of which the resources committed by
what regional development would be like individual firms indirectly raise the earn-
if the United States was not an economic
ings of other firms' resources-that is,
and monetary union. Nor do we know where external economies can be gener-
what would have happened to the regions ated. In both instances, the argument is
of the European countries in the absence that it may well be possible to identify
of the formation of the European Commu- some "strategic sectors" that at the margin
nity and its recent movement toward are more valuable than others, and that
EMU. Finally, what of Krugman's views the promotion of these sectors through
on the regional policy implications of protection, export subsidies, support of
European integration? To assess this R&D, and so on could raise national
aspect of his analysis we need to look atincome.
the policy debate within the new trade More recently, however, Krugman has
theory more generally. reacted against strategic trade policy. In
Peddling Prosperity (1994c), he questions industrial policy." His initial skepticism of

the theoretical validity of strategic indus- the theoretical credentials and practical
trial policy and goes on to berate leading applicability of the external economies
American politicians (especially President rationale for targeted industrial policies
Clinton) and their economic advisors was on the grounds that only technologi-
(especially Thurow and Reich) for misap- cal, not pecuniary external economies are
propriating strategic trade theory and of concern, that technological externalities
applying it in a "simple-minded way." in any case are of limited significance, and
Both Reich (1991) and Thurow (1994) are that they are international rather than
criticized for peddling the view that if the national in scope (Krugman 1983b, 1984).
United States is to compete in the global But, as we have seen, he now believes
economy, the government should aban- that external economies associated with
don its notions of free trade and minimal- market-size effects are substantial and
ist industrial intervention and instead demonstrable (and often indistinguishable
from technological external economies),
pursue a more activist stance aimed at
promoting the shift of American industryand this means that targeted industrial
policies have a potential role after all
into "high value" (Reich) and "sunrise"
(Thurow) sectors. Krugman believes that(Krugman 1987a, 1993b). Moreover,
such views are based on fallacious theory,
impractical politics, and an erroneous many of the important market-size effects
obsession with the idea of "competitive- apply not at the level of the international or
even the national economy, but at a
ness": "While competitive problems could
regional or local level. The argument that
arise in principle, as a practical, empirical
the gains from support of industries that
matter the major nations of the world are generate external economies will be dissi-
not to any significant degree in economic pated abroad is therefore mostly wrong.
competition with each other" (Krugman (Krugman 1993b, 167)
1994a, 35). In his view, competitiveness
relies on the metaphor of a country being In this revised view of industrial policy,
a big corporation, when, in fact, countriesnot only are regional and local industrial
(and regions) are nothing like corpora- clusters considered to provide empirical
proof of the importance of external
tions. Hence it is very difficult to establish
a meaningful definition of national or economies, such clusters help to define
regional competitiveness. Furthermore,what industries should be supported.
he argues that it is wrong to see Geographical clustering provides the jus-
international trade as competition-as a tification for industrial intervention, and
sort of zero-sum game-when it is a the aim of that intervention should be to
process of exchange involving mutual foster local externalities. In effect, what
benefit. By the early 1990s, then, Krug- Krugman seems to be suggesting, though
man had come to denounce strategic trade he does not use the term explicitly, is that
policy as "bad economics." Yet, while the only justifiable form of industrial
Krugman has vigorously attacked the (trade) policy is in fact regional industrial
whole ensemble of policies that have development policy. In line with strategic
come to be labeled as "strategic trade trade theory, the underlying premise is
policy," it appears that he now sees a role that national industrial comparative ad-
for a limited and focused industrial policy. vantage can be created through support-
ive and targeted industrial policies which
aim to create and facilitate key sectoral
Geographical Clustering and Strategic
specializations. The twist in Krugman's
Industrial Policy
argument, however, is that the most
In a recent paper Krugman (1993b, 160) effective scale at which to create that
states that he has "now changed his mind advantage is at the level of regional
and . . . gone, at least slightly, soft on clusters. Essentially the same argument is
implicit in Porter's (1990) major work on industrial districts celebrated by Alfred

national competitive advantage. Indeed, Marshall early this century (for example,
he now sees local and economic develop- see Sunley 1992). An equally persuasive
ment policies as having an instrumental case can be made that industrial diversifi-
role in fostering national industrial com- cation rather than specialization is the
petitiveness (Porter 1994). most appropriate regional development
It is only a small step from this to argue policy route, that diversifying the regional
that the promotion of specialized export industrial "portfolio" reduces the suscep-
clusters is also the most viable approach tibility of the regional economy to adverse
to reviving and regenerating old industrial demand shocks and localized structural
regions. This is, in fact, what certain crisis (this is the general conclusion of the
writers on flexible specialization and portfolio studies referred to earlier; see
industrial districts have implied. These also Geroski 1989).
authors use the success of certain well- Krugman, on the other hand, appears to
known specialized (usually export-orien- believe that the most important policy
tated) industrial districts as a model for response to the possibility of regional
"indigenous" local economic regeneration instability in more-specialized regions is
more generally (see, for example, Hirst fiscal stabilization. In the case of Euro-
and Zeitlin 1989; Pyke, Becattini, and pean economic integration, for example,
Sengenberger 1989; Sabel 1989; Stohr Krugman suggests that national budgets
1989; Cooke 1990; Scott 1992b). The path will have to be substantially centralized so
to the reindustrialization of economically that automatic federal European fiscal
and structurally depressed regions is seen transfers can perform the required stabili-
to be via the promotion and support of zation role when asymmetric regional
neo-Marshallian small-firm, flexibly spe- shocks occur.23 He notes the way in which
cialized production complexes involving the U.S. federal budget tends automati-
dense local networks of cooperation, cally to redistribute resources toward
competition, and horizontal interdepen- regions affected by negative economic
dencies. In some ways this support for a shocks (via compensating adjustments in
local industrial strategy based on multiple the tax-take and in welfare payments
externalities is similar to Krugman's ratio- across regions):
nale for industrial policy. However, the While the US does not cope with the
advocacy of regional export specialization
problems perfectly (as the current travails
either as a local economic development not only of New England but of the North
strategy or as a form of trade policy is East, in general, and increasingly of Califor-
contentious. nia, demonstrate), a highly federalized fiscal
The key question over the promotion of system helps a good deal. The lack of such a
regional industrial specialization is system in Europe therefore is a real
whether the potential advantages are problem. (Krugman 1993d, 258)
outweighed by the likelihood of greater
regional instability and shocks, and the
risk of structural depression. As Krugman 23 In the European case, although fiscal
(1993d) notes for the case of Massachu- federalism is indeed a natural corollary to
setts, regional industrial specialization is a EMU, national governments in a future
double-edged sword: it can be the basis of European monetary union would not lose all of
their instruments of economic policy. National
a high rate of export-led local economic
budgetary policies would continue to have
growth in one period, but the source of some, even if constrained, automatic stabiliz-
prolonged local economic depression if ing role (see Boonstra (1991) on the limits that
that demand subsequently collapses or is EMU will impose on national budgetary
captured by other competing regions autonomy). In this sense member states in a
(often in other countries). This is precisely European EMU would be somewhat different
what happened to many of the specialized from the individual states in the United States.
Krugman is at pains to distinguish this tion and specialization leads to regional

form of regional policy response from that economic instability and divergent long-
needed to ameliorate "core-periphery" term growth paths. The response of the
patterns of uneven regional development new industrial geography to this dilemma,
or the regional decline that stems from of course, is to insist that flexibly special-
specialization in outmoded industries and ized industrial districts are more adapt-
products. The policy response to these able to economic and technological
sorts of regional issues, he says, "is much change by virtue of the dynamism and
less related to EMU than the stabilization networking of the small enterprises of
problem" (Krugman 1993d, 259). which they are (invariably assumed to be)
However, while automatic fiscal trans- composed. However, this claim remains
fers may well help to alleviate and far from proven. In addition, the number
of flexibly specialized districts remains
stabilize the income and growth problems
associated with economically depressed small, and their origins and dynamics are
regions, they are not an adequate re- matters of debate (see Markusen 1993;
sponse to uneven regional development. Markusen and Park 1993). This is not to
By themselves, they are not sufficient to dismiss the new "indigenous" approach to
recast the structure and dynamics of regional policy based on arguments of
regional development so as to improve (flexibly) specialized industrial develop-
the long-term economic performance and ment; but it is to signal that this approach
wealth of the regions concerned. This is is no more of a general panacea for
why, of course, historically many Euro- uneven regional development than was
pean countries have developed elaborate the old model of redistributive regional
systems of region-specific developmental policy. Nor do we wish to imply that
aid and assistance, and why the European increasing returns and external economies
Union has been strengthening and re- are unimportant in the regional policy
forming its own centrally administered debate. To the contrary, not only is there
regional structural funds in the context of evidence from Europe that increasing
the movement toward increasing eco- returns industries are more concentrated
nomic integration (and enlargement) of in regions with better infrastructures,
the Union (see, for example, Martin 1993; especially technological and educational
Collier 1994). Krugman's distinction be- infrastructures (Martin and Rogers 1994a,
tween regional instabilities due to idio- 1994b), endogenous growth theory also
syncratic demand shocks and regional suggests that external economies and
problems of a more "core-periphery" and technological spillovers are likely to play a
"structural" nature is surprising and ques-key role in the local growth process in an
tionable. For if, as he argues, short-term integrated Europe. But in our judgment,
regional shocks have long-term effects on there is an urgent need for much more
regional growth, then interregional fiscal thought on how local and regional policies
stabilization is an inadequate policy re- can foster and support these externalities
sponse, and other, more strategic forms of without simultaneously narrowing the
regional policy are required. industrial base and increasing the vulner-
In our view, the proposal of fiscal ability of regions to demand shocks.
federalism does not temper the worry that
regional clusters of specialized industry Conclusions
will be unstable and fragile. The basic
tension in Krugman's argument re- A few years ago, Neil Smith (1989)
mains-namely, how to reconcile his argued for a rebuilding of regional theory
suggestion that the aim of industrial policy within geography based on a synthesis of
should be to promote industrially special- ideas from location theory and uneven
ized regional clusters with his thesis that development theory. More recently,
increased regional industrial concentra- Krugman (1993a) has argued for a similar
synthesis of location theory and trade is to pursue a closer exchange between

theory, for using economic geography as a Krugman's "geographical economics" and
key component in the construction of a the new industrial and economic geogra-
new "geographical economics" of trade. In phy. Neither can claim to have a monop-
this paper we have sought to provide a oly of insight, but an exchange of ideas
critical assessment of Krugman's ideas on between the two would, we believe, be
economic geography and his attempts to beneficial.
use these ideas to forge a "geographical One strength of Krugman's work, with-
economics." Because of the enormous out doubt, is that his linking of external
volume and breadth of his writings we economies and regional industrial agglom-
have had to skate across many of his ideas,eration with trade provides an important
and as a result we have no doubt failed to corrective to the flexible specialization
accord some of them the full attention thesis of the new industrial geography, in
they deserve. In addition, Krugman's which regional industrial development is
tendency to constantly revise and even viewed overwhelmingly as an indigenous
reject his earlier ideas renders the task ofprocess and the role of trade is typically
assessment akin to tracking a moving either subordinated or neglected alto-
target. Nevertheless, we believe we have gether. Furthermore, Krugman's empha-
succeeded in isolating the core compo- sis on imperfect competition and pecuni-
nents of his arguments sufficiently to be ary externalities likewise exposes the
able to identify some of their main limitations of the conceptions of external-
strengths and weaknesses, particularly as ities now prevalent within the geographi-
they relate to the question of regional cal literature. The thrust of flexible
growth and development. In many ways, specialization ideas in economic geogra-
Krugman's approach to economic geography is that agglomeration is associated
phy is a regional science one, a reworking with the shift from vertical integration to
of traditional location theory concepts and the horizontal integration of related activ-
models. The new economic and industrial ities among small, competitive firms
geography, of course, has moved well which cluster together to minimize trans-
away from that tradition. For this reason, action costs. Williamsonian transaction
it might well be questioned whether costs economics-itself a neoclassical-
Krugman's work contains anything that is oriented form of institutional economics-
new or useful for economic geographers. has been used to give a new theoretical
It would be wrong to be so readily underpinning to Marshall's notion of
dismissive, however; Krugman's work is industrial localization. Krugman's focus on
not as simplistic as Johnston and others pecuniary externalities, especially market-
have suggested (nor for that matter is size effects, and the role that large,
geographical work always as sophisticated oligopolistic producers can play in indus-
as its practitioners appear to believe). For trial agglomeration suggests that indus-
it is perhaps less the specific results of trial geographers need to reassess their
Krugman's analyses that are important for theoretical accounts accordingly. Yet, at
economic geography than the general the same time, one of the most important
stimulus they provide for further inquiry. limitations of Krugman's geographical
In this respect we concur with the view economics is his stubborn concentration
that Krugman's work "is rich in ideas, only on those externalities that can be
seductive in taking us though simple mathematically modeled, and thus his
logical arguments to surprising conclu- reluctance to discuss the geographical
sions and so self-confident in the discus- impacts of technological and knowledge
sion of its assumptions and its premises spillovers. Although the recent geograph-
that reading it is at the same time great ical literature has begun to assign key
fun and a continuous challenge" (Casella importance to technical change and tech-
1993, 261-62). The challenge, as we see it, nological externalities in shaping and
transforming the space economy, and Krugman is right to stress the role of
hence to some extent holds some impor- geography in the historical, path-depen-
tant lessons for Krugman's analysis in this dent nature of the economic process, but
respect, it too has yet to explore fully the he fails to explicate the nature of that role.
cumulative and spillover effects associated A third aspect of Krugman's geographi-
with technological change. cal economics that we want to highlight,
A second significant aspect of Krug- and which also has both strengths and
man's geographical is the recognition that weaknesses, is his analysis of the way that
regional economic development is a his- region-specific shocks can have long-term
torical, path-dependent process. His re- growth consequences. How regions re-
peated exhortation that "history matters," spond and adjust to demand and supply
both in terms of the arbitrary initial shocks, both in the short term and in the
conditions and accidental events that set long run, in an increasingly deregulated,
in motion particular patterns of industrial market-propelled, and uncertain world is
development over time and space and inan important research issue, but one that
terms of the subsequent "locking in" ofhas been neglected by the new industrial
those patterns via self-reinforcing effects,geography. Krugman's analysis for the EU
is not of course particularly novel toregions, using the U.S. regions for com-
economic geographers. Geographers haveparison, provides a useful basis for devel-
long recognized that a given pattern of oping this research agenda. However, as
uneven regional development, once es-we have seen, his analysis is far from
tablished, tends to exhibit a high degreeunproblematic. Apart from being too
of persistence or."inertia" over time, andAmerican-centered (as exemplified by his
that this inertia can operate either toemphasis on the central role of labor
foster regional growth or to retard it. The mobility, which is considerably higher in
more recent interest by geographers inthe United States than in Europe), his
the local socio-institutional "embedded- models do not adequately explain why a
ness" of economic activity also bears upon successful regional economy (like Massa-
the issues of path dependence and lock-in. chusetts, for example) can suddenly go
In Krugman's view the role played by into reverse, or why the geography of
geography in determining "lock-in" is uneven regional development can and
strictly an increasing returns phenome- does undergo significant reconfigurations
non, in the form of the Marshallian ("spatial switching"), or why some regions
externalities associated with local indus- seem better able than others to withstand
trial agglomeration (or, under certainor adjust to negative external shocks.
Krugman singles out industrial specializa-
circumstances, in the form of self-fulfilling
expectations). What he fails to consider istion as the main factor shaping the
the influence exerted by local institu-relative stability of different regions and
tional, social, and cultural structures inthe labor market as the key determinant
facilitating or constraining local economic of the regional adjustment process. But a
development. This neglect would seem tofull account must surely also consider
stem in large part from Krugman's com-other reasons for regional crisis and
plaint that noneconomic or "social" factors restructuring and mechanisms other than
are not easily modeled and that theylabor market flexibility in order to explain
should therefore be left to sociologists.the degree of and differences in regional
But as recent studies in the new industrial adjustment.
and economic geography have begun to There is, then, considerable scope for a
show, the "thickness" and nature of such potentially fruitful cross-fertilization of
socio-institutional "externalities" are fun- ideas between Krugman's geographical
damental to the initial emergence, trajec- economics and the new industral-eco-
tory, and adaptability of industrial dis- nomic geography, and for the elaboration
tricts and regional economies. Thus of each. Both draw heavily on a Marshal-
lian view of industrial localization. But

There would be much to be gained, in our
whereas the new industrial geography hasview, if both Krugman's geographical
sought to reinterpret the Marshallianeconomics and the new industrial and
account in terms of transaction cost economic geography revisited the method
economics, Krugman instead has tried and
to the message of Kaldor's work. But
link Marshallian industrial localization that, as Krugman would say, is another
with the economics of imperfect competi-
story.
tion, increasing returns, path depen-
dence, and cumulative causation. These
concerns were at the center of Nicholas References
Kaldor's (1978, 1981, 1985) earlier seminal
Aghion, P., and Howitt, P. 1993. A model of
work on trade, endogenous growth, and
growth through creative destruction. In
regional development, a debt that Krug-
Technology and the wealth of nations: The
man acknowledges:
dynamics of constructed advantage, ed. D.
This clear dependence on history is the Foray and C. Freeman, 145-72. London:
Pinter.
most convincing evidence available that we
live in an economy closer to Kaldor's visionAmin, A., and Robins, K. 1990. Industrial
of a dynamic world driven by cumulative districts and regional development: Limits
processes than to the standard constant and possibilities. In Industrial districts and
returns model. (1991a, 9-10) inter-firm co-operation in Italy, ed. F. Pyke,
G. Becattini, and W. Sengenberger, 185-
Krugman even goes so far as to admit that 219. Geneva: Institute for Labour Studies.
in a sense his own work is only "a Amin, A., and Thrift, N. 1992. Neo-Marshal-
repetition" of Kaldor's ideas. There are lian nodes in global networks. International
significant differences between the two, Journal of Urban and Regional Research
16:571-87.
however. In his quest for economic rigor,
, eds. 1994. Globalization, institutions
Krugman's mathematical formalization of
and regional development in Europe. Ox-
the processes of industrial agglomeration ford: Oxford University Press.
and uneven regional development has Anderson, K., and Blackhurst, R., eds. 1993.
taken him away from the richness of Regional integration and the global trading
Kaldor's original approach toward the system. London: Harvester Wheatsheaf.
limited abstract landscapes of regional Audretsch, D., and Feldman, M. 1994. Knowl-
science. Indeed, in Development, Geogra- edge spillovers and the geography of inno-
phy and Economic Theory (Krugman vation and production. Discussion Paper
1995) and The Self-Organising Economy 953. London: Centre for Economic Policy
Research.
(1996), his role model seems to be that
Baldwin, R. 1994. Towards An integrated
doyen of regional science Walter Isard,
Europe. London: Centre for Economic
rather than Nicholas Kaldor, who, one Policy Research.
suspects, would have been extremely Baldwin, R., and Lyons, R. 1990. External
skeptical of the unrealistic, deductive economies and European integration: The
model-building that is the hallmark of the potential for self-fulfilling expectations. Dis-
regional science tradition. And despite
Krugman's apparent agreement with Kal-
dor's argument for the "irrelevance of theories were in fact anticipated by Kaldor.
equilibrium economics," the ghosts of But whereas he eschewed the deductive and
constrained maximization and equilibrium mathematical for the inductive and realistic (as
solutions still haunt much of his analysis.24 expressed in his emphasis on "stylized facts"
and nonequilibrium), the new theorists have
deliberately sought to "systematize" his ideas
through mathematical formalism and appeal to
24 The same is true of the new trade theory all-embracing principles of optimizing eco-
and the new endogenous growth theory more nomic behavior (see the assessments by Kitson
generally. Many of the ideas found in these and Michie 1995; Skott and Auerbach 1995).
cussion Paper 471. London: Centre for "Lessons of Massachusetts for EMU." In
Economic Policy Research. Adjustment and growth in the EMU, ed. F.
Barro, R. J., and Sala-i-Martin, X. 1991. Torres and F. Giavazzi, 261-66. Cambridge:
Convergence across states and regions. Cambridge University Press.
Brookings Papers on Economic Activity Chamberlin, E. H. 1949. The theory of
1:107-82. monopolistic competition. 6th ed. London:
Barth, J.; Kraft, J.; and Wiest, P. 1975. A Oxford University Press.
portfolio-theoretic approach to industrialChatterji, M. 1993. Convergence clubs and
diversification and regional unemployment. endogenous growth. Oxford Review of Eco-
Journal of Regional Science 15 (1):9-15. nomic Policy 8:57-69.
Bayoumi, T., and Eichengreen, B. 1993. Chipman, J. 1970. External economies of scale
Shocking aspects of European monetary and competitive equilibrium. Quarterly
integration. In Adjustment and growth in Journal of Economics 84:347-85.
the EMU, ed. F. Torres and F. Giavazzi, Collier, J. 1994. Regional disparities, the
193-229. Cambridge: Cambridge University Single Market and EMU. In Unemployment
Press. in Europe, ed. J. Michie and J. Grieve
Bertola, G. 1993. Models of economic integra- Smith, 145-59. London: Academic Press.
tion and localized growth. In Adjustment Commission of the European Communities.
and growth in the EMU, ed. F. Torres and 1990. An empirical assessment of factors
F. Giavazzi, 159-79. Cambridge: Cambridge shaping regional competitiveness in problem
University Press. regions. Main report. Luxembourg: Office
Blanchard, O. J., and Katz, L. F. 1992. for Official EC Publications.
Regional evolutions. Brookings Papers on . 1991. The regions in the 1990s.
Eonomic Activity 1:1-59. Luxembourg: Office for Official EC Publica-
Boonstra, W. W. 1991. The EMU and national tions.
autonomy on budget issues. In Finance and . 1994. Competitiveness and cohesion
the international economy, Vol. 4., ed. R. trends in the regions. Luxembourg: Office
O'Brien and S. Hewin, 208-24. Oxford: for Official EC Publications.
Oxford University Press. Conroy, M. 1975. The concept and measure-
Brander, J., and Spencer, B. 1983. Interna- ment of regional industrial diversification.
tional R and D rivalry and industrial Southern Economic Journal 41:492-505.
strategy. Review of Economic Studies 50: Cooke, P. 1990. Manufacturing miracles: The
707-22. changing nature of the local economy. In
. 1985. Export subsidies and interna- Local economic policy, ed. M. Campbell,
tional market share rivalry. Journal of 25-42. London: Cassell.
International Economics 18:83-100. Cooke, P., and Morgan, K. 1993. The network
Brewer, H. 1984. Regional economic stabiliza- paradigm: New departures in corporate and
tion: An efficient diversification approach. regional development. Environment and
Review of Regional Studies 14 (1):8-21. Planning D: Society and Space 11:543-64.
Buchanan, J., and Yoon, Y., eds. 1994. TheDavid, P., and Rosenbloom, J. 1990. Marshal-
return to increasing returns. Ann Arbor: lian factor market externalities and the
The University of Michigan Press. dynamics of industrial localization. Journal
Button, K., and Pentecost, E. 1993. Testingfor of Urban Economics 2:349-70.
convergence of the EC regional economies. De Melo, J., and Robinson, S. 1990. Produc-
Economics Research Paper 93/5. Loughbor- tivity and externalities: Models of export-led
ough: Department of Economics, Loughbor- growth. Discussion Paper 400. London:
ough University of Technology. Centre for Economic Policy Research.
Cabellero, R., and Lyons, R. 1990. InternalDixit, A. K., and Norman, V. 1980. Theory of
versus external economies in European international trade. Cambridge: Cambridge
manufacturing. European Economic Review University Press.
34:805-30. Drache, D., and Gertler, M., eds. 1991. The
. 1991. External effects and European new era of global competition: State policy
integration. In The effect of 1992 on tradeand market power. Toronto: McGill-Queens
and industry, ed. L. A.Winters and A. University Press.
Venables, 34-53. Cambridge: Cambridge Dunford, M. 1993. Regional disparities in the
University Press. EC. Evidence from the REGIO data bank.
Casella, A. 1993. Discussion of Krugman's Regional Studies 27:727-43.
Dunford, M., and Kafkalas, G., eds. 1992. Helpman, E. 1981. International trade in the
Cities and regions in the new Europe. presence of product differentiation, econo-
London: Belhaven. mies of scale and monopolistic competition.
Economics of meaning, The. 1994. The Econo- Journal of International Economics 11:305-
mist, 30 April, 13-14. 40.
Eichengreen, B. 1993. European monetary . 1984. Increasing returns, imperfect
unification. Journal of Economic Literature
markets and trade theory. In Handbook of
31:1321-57. international economics, Vol. 1, ed. R. Jones
Emerson, M.; Aujean, M.; and Catinat, M.and P. Kenen, 325-65. Oxford: North
1988. The economics of 1992. Oxford: Oxford Holland.
University Press. Helpman, E., and Krugman, P. 1985. Market
Erickson, R. 1989. Export performance and structure and foreign trade: Increasing
state industrial growth. Economic Geogra-
returns, imperfect competition and the
phy 65:280-92.
international economy. London: MIT Press.
Ethier, W. 1982. National and international
Hirschman, A. 1958. The strategy of economic
returns to scale in the modern theory of
international trade. American Economic Re- development. New Haven: Yale University
Press.
view 72:389-405.
Friedman, B. 1994. Must we compete? New Hirst, P., and Zeitlin, J., eds. 1989. Reversing
York Review of Books, 20 October, 14-20. industrial decline? Oxford: Berg.
Fujita, M. 1989. Urban economic theory: LandHoare, A. 1975. Linkage flows, locational
use and city size. Cambridge: Cambridge evaluation, and industrial geography: A case
University Press. study of Greater London. Environment and
Geroski, P. A. 1989. The choice between scale Planning A 7:41-58.
and diversity. In 1992: Myths and realities, . 1992. Review of P. Krugman's "Geog-
ed. E. Davis, 29-45. London: Centre for raphy and Trade." Regional Studies 26:679.
Business Strategy, London Business School. Howes, C., and Markusen, A. 1993. Trade,
Gertler, M. 1992. Flexibility revisited: Dis- industry and economic development. In
tricts, nation-states and the forces of produc- Trading industries, trading regions, ed. H.
tion. Transactions of the Institute of British Noponen, J. Graham, and A. Markusen,
Geographers n.s. 17:259-78. 1-44. London: Guilford Press.
Gibb, R., and Michalak, W., eds. 1994. Jackson, R. 1984. An evaluation of alternative
Continental trading blocs. Chichester: measures of regional industrial diversifica-
Wiley. tion. Regional Studies 8:103-12.
Glasmeier, A.; Thompson, J.; and Kays, A. Jaffe, A.; Trajtenberg, M.; and Henderson, R.
1993. The geography of trade policy: Trade 1993. Geographical localization of knowl-
regimes and location decisions in the textile edge spillovers as evidenced by patent
and apparel complex. Transactions of the citations. Quarterly Journal of Economics
Institute of British Geographers n.s. 18:19- 108:577-98.
35. Johnston, R. J. 1989. Extending the research
Grabher, G. 1993. The embeddedfirm: On the agenda. Economic Geography 65:338-47.
socioeconomics of industrial networks. Lon- . 1992. Review of P. Krugman's "Geog-
don: Routledge. raphy and Trade." Environment and Plan-
Grant, R. 1994. The geography of international ning A 24:1006.
trade. Progress in Human Geography 18: Kaldor, N. 1978. The case for regional policies.
298-312. In Further essays on economic theory,
Grossman, G., and Helpman, E. 1991. Innova- 139-54. London: Duckworth.
tion and growth in the global economy. . 1981. The role of increasing returns,
London: MIT Press. technical progress and cumulative causation
Hanink, D. 1988. An extended Linder model in the theory of international trade and
of international trade. Economic Geography economic growth. Economic Appliquee 34
64:322-34. (4). Reprinted in The essential Kaldor (1989),
. 1994. The international economy: A ed. F. Targetti and A. Thirlwall, 327-50.
geographical perspective. London: Wiley. London: Duckworth.
Harrison, B. 1992. Industrial districts: Old . 1985. Interregional trade and cumula-
wine in new bottles? Regional Studies tive causation. In Economics without equi-
26:469-84. librium, 57-79. New York: Sharpe.
Kitson, M., and Michie, J. 1995. Conflict, . 1991d. History and industrial location:
cooperation and change: The political econ- The case of the manufacturing belt. Ameri-
omy of trade and trade policy. Review of can Economic Review (Papers and Proceed-
International Political Economy 2 (4):632- ings) 81:80-83.
57. . 1992. Currencies and crises. London:
Knox, P., and Agnew, J. 1994. The geography MIT Press.
of the world economy. 2d ed. London: . 1993a. On the relationship between
Edward Arnold. trade theory and location theory. Review of
Krugman, P. 1979. Increasing returns, monop- International Economics 1:110-22.
olistic competition and international trade. . 1993b. The current case for industrial
Journal of International Economics 9:469- policy. In Protectionism and world welfare,
79. ed. D. Salvatore, 160-79. Cambridge: Cam-
. 1980. Scale economies, product differ- bridge University Press.
entiation and the pattern of trade. Amercan . 1993c. First nature, second nature and
Economic Review 70:950-59. metropolitan location. Journal of Regional
. 1981. Trade, accumulation and uneven Science 33 (2):129-44.
development. Journal of Development Eco-_ . 1993d. The lessons of Massachusetts
nomics 8:149-61. for EMU. In Adjustment and growth in the
. 1983a. The "new theory" of interna- European Monetary Union, ed. F. Torres
tional trade and the multinational enter- and F. Giavazzi, 241-69. Cambridge: Cam-
prise. In The multinational corporation inbridge University Press.
the 1980s, ed. C. Kindleberger and D. . 1993e. Inequality and the political
Audretsch, 57-73. London: MIT Press. economy of Eurosclerosis. November. Lon-
. 1983b. Targeted industrial policies: don: Centre for Economic Policy Research.
Theory and evidence. In Structural change . 1994a. Competitiveness: A dangerous
and public policy. Kansas City: Federal obsession. Foreign Affairs (March-April):
Reserve Bank of Kansas City. 28-44.
. 1984. Foreign industrial targetting and . 1994b. The age of diminished expecta-
the US economy. Brookings Papers on tions: US economic policy in the 1990s. Rev.
Economic Activity 1:77-121. ed. London: MIT Press.
, ed. 1986. Strategic trade policy and . 1994c. Peddling prosperity: Economic
the new international economics. Cam- sense and nonsense in the age of diminshed
bridge: MIT Press. expectations. New York: Norton.
_ 1987a. The narrow moving band, the . 1994d. Does Third World growth hurt
"Dutch disease" and the competitive conse- First World prosperity? Harvard Business
quences of Mrs. Thatcher. Journal ofReview (July-August):113-21.
Development Economics 26:42-55. . 1995. Development, geography and
____ . 1987b. Is free trade passe? Economic economic theory. Cambridge: MIT Press.
Perspectives 1 (2):131-44. . 1996. The self-organising economy.
. 1987c. Increasing returns and the Cambridge, Mass.: Blackwell Publishers.
theory of international trade. In Advances in Krugman, P., and Venables, A. 1990. Integra-
economic theory, ed. T. Bewley, 301-28. tion and the competitiveness of peripheral
Cambridge: Cambridge University Press. industry. In Unity with diversity in the
1989. Economic integration in Eu- European Community, ed. C. Bliss and J.
rope: Some conceptual issues. In The Braga de Macedo, 56-75. Cambridge: Cam-
European internal market-trade and com- bridge University Press.
petition, ed. A. Jacquemin and A. Sapir, . 1994. Globalization and the inequality
357-80. Oxford: Oxford University Press. of nations. Discussion paper 1015. London:
. 1990. Rethinking international trade. Centre for Economic Policy Research.
Cambridge: MIT Press. Kurre, J., and Weller, B. 1989. Regional
_ . 1991a. Geography and trade. Leuven: cyclical instability. Regional Studies 24:318-
Leuven University Press. 29.
. 1991b. Increasing returns and eco- Lancaster, K. 1980. Intra-industry trade under
nomic geography. Journal of Political Econ- perfect monopolistic competition. Journal of
omy 99:483-99. International Economics 10:151-76.
_ 1991c. History versus expectations. Losch, A. 1967. The economics of location.
Quarterly Journal of Economics 106:651-67. New York: John Wiley Science Edition.
Lovering, J. 1990. Fordism's unknown succes- gence or divergence? A study of regiona

sor. International Journal of Urban and business cycle patterns among and withi
Regional Research 14:159-74. US census regions. Regional Science Per
Lucas, R. 1988. On the mechanics of economic spectives 22(2):30-52.
development. Journal of Monetary Econom- Peschel, K. 1982. International trade, integr
ics 22:3-42.
tion and industrial location. Regional Sci-
Markowitz, H. 1959. Portfolio selection and ence and Urban Economics 12:247-69.
efficient diversification of investment. New
Phelps, N. 1992. External economies, agglom-
Haven: Yale University Press. eration and flexible accumulation. Transac-
Markusen, A. 1993. Trade as a regional tions of the Institute of British Geographers
development issue: Policies for job and n.s. 17(1):35-46.
community preservation. In Trading indus-
Porter, M. E. 1990. The competitive advantage
tries, trading regions, ed. H. Noponen, J.
of nations. London: Macmillan.
Graham, and A. Markusen, 285-302. Lon-
don: Guilford Press. . 1994. The role of location in competi-
Markusen, A., and Park, S. 0. 1993. New tion. Journal of the Economics of Business
industrial districts: A critique and extension
1(1):35-39.
from the developing countries. Working Pyke, F.; Becattini, G.; and Sengenberger, W.,
paper, Project on Regional and Industrial eds. 1989. Industrial districts and inter-firm
Economics, Rutgers University. co-operation in Italy. Geneva: Institute for
Labour Studies.
Martin, P. J., and Rogers, C. A. 1994a.
Industrial location and public infrastruc- Reich, R. B. 1991. The work of nations:
ture. Discussion paper 909. London: Centre Preparing ourselves for 21st century capital-
for Economic Policy Research. ism. London: Simon and Schuster.
. 1994b. Trade effects of regional aid. Romer, P. 1990. Endogenous technological
Discussion paper 910. London: Centre for change. Journal of Political Economy 98:71-
Economic Policy Research. 102.
Martin, R. L. 1993. Reviving the economic Rosenstein-Rodan, P. 1943. Problems of indus-
case for regional policy. In Spatial policy in trialization of Eastern and South-Eastern
a divided nation, ed. R. T. Harrison and M. Europe. Economic Journal 53:202-11.
Hart, 270-90. London: Jessica Kinglsey. Sabel, C. 1989. Flexible specialization and the
Meade, J. 1952. External economies and re-emergence of regional economies. In
diseconomies in a competitive situation. Reversing industrial decline? ed. P. Hirst
Economic Journal 62:54-67. and J. Zeitlin, 17-70. Oxford: Berg.
Mishan, E. 1971. Externalities. Journal of Samuelson, P. 1994. Preface. In The age of
Economic Literature 9:1-28. diminished expectations, ed. P. Krugman,
Murphy, A.; Schleifer, A.; and Vishny, R. vii. London: MIT Press.
1989. Industrialization and the big push.Scitovsky, T. 1954. Two concepts of external
Journal of Political Economy 97:1003-26. economies. Economic Journal 62:52-67.
Myrdal, G. 1957. Economic theory and under-Scott, A. J. 1986. Industrial organization and
developed regions. London: Duckworth. location, divisons of labor and the firm's
Neven, D. J., and Gouyette, C. 1994. Regionalspatial process. Economic Geography 62:
convergence in the European Community. 215-31.
Discussion paper 914. London: Centre for . 1988. New industrial spaces. London:
Economic Policy Research. Pion.
Noponen, H.; Graham, J.; and Markusen, A., .1992a. The role of large producers in
eds. 1993. Trading industries, trading re-industrial districts: A case study of high-tech
gions. London: Guilford Press. systems houses in Southern California.
Ohlin, B. 1933. Interregional and interna- Regional Studies 26:265-75.
tional trade. Cambridge: Harvard Univer- . 1992b. The collective order of flexible
sity Press. production agglomeration: Lessons for local
Organization for Economic Cooperation and economic development policy and strategic
Development (OECD). 1994. The OECD choice. Economic Geography 68:219-33.
jobs study: Part 1. Labour market trends Scott, A. J., and Storper, M. 1992a. Regional
and underlying forces of change. Paris: development reconsidered. In Regional de-
OECD. velopment and contemporary industrial re-
Palmini, D., and Cray, R. F. 1992. Conver- sponse: Extendingflexible specialization, ed.
H. Ernste and V. Meier, 1-24. London: production: Learning and innovation in the
Belhaven. technology districts of France, Italy and the
. 1992b. Industrialization and regional USA. Regional Studies 27:433-55.
development. In Pathways to industrializa- . 1992b. The limits to globalization:
tion and regional development, ed. M. Technology districts and international trade.
Storper and A. Scott, 3-17. London: Rout- Economic Geography 68:60-93.
ledge. Storper, M., and Walker, R. 1989. The
Skott, P., and Auerbach, P. 1995. Cumulative capitalist imperative: Territory, technology
causation and the "new" theories of eco- and industrial growth. Oxford: Basil Black-
nomic growth. Journal of Post Keynesian well.
Economics 17(3):381-402. Sunley, P. J. 1992. Marshallian industrial
Smith, A. 1994. Imperfect competition and districts: The case of the Lancashire cotton
international trade. In Surveys in interna- industry in the inter-war years. Transactions
tional trade, ed. D. Greenaway and A. of the Institute of British Geographers n.s.
Winters, 43-65. Oxford: Basil Blackwell. 17:306-32.
Smith, N. 1989. Uneven development theory Thurow, L. 1994. Head to head: The coming
and location theory: Towards a synthesis. In economic battle among Japan, Europe and
New models in geography, Vol. 1, ed. R. America. 2d ed. London: Brealey Publish-
Peet and N. Thrift, 142-63. London: Unwin ing.
Hyman. Tyson, L. 1992. Who's bashing whom? Trade
Steed, G. 1971. Internal organization, firm conflict in high technology industries. Wash-
integration and locational change: The ington, D.C.: Institute for International
Northern Ireland linen complex, 1954-64. Economics.
Economic Geography 47:371-83. von Hagen, J., and Hammond, G. 1994.
Stigler, G. 1951. The division of labour is Industrial localization: An empirical testfor
limited by the extent of the market. Journal Marshallian localization economies. Discus-
of Political Economy 59:185-93. sion paper 917. London: Centre for Eco-
Stohr, W. B. 1989. Regional policy at the nomic Policy Research.
crossroads. In Regional policy at the cross- Weber, A. 1929. Theory of the location of
roads: European perspectives, ed. L. Al- industries. Chicago: University of Chicago
brechts, F. Moulaert, P. Roberts, and E. Press.
Swyngedouw, 191-97. London: Jessica Young, A. A. 1928. Increasing returns and
Kingsley Press. economic progress. Economic Journal 38:
Storper, M. 1992a. Regional "worlds" of 527-42.

On Krugman Geographical Economics

Uploaded by

Copyright:

Available Formats

On Krugman Geographical Economics

Uploaded by

Document Information

Original Title

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

On Krugman Geographical Economics

Uploaded by

Copyright:

Available Formats

Paul Krugman's Geographical Economics and Its Implications for Regional Development

Theory: A Critical Assessment

uneven. To this end, he draws on a range economics.3 However, this methodologi-

transport costs will act to prohibit the industries or clusters of industries to

Krugman uses Marshall's theorization

l In a similar fashion, Porter (1990) argues

foster these industries will lead to trade Krugman's "geographical economics"

12 After all, most of the main schools of

A Comparison of Krugman's "Geographical Economics" with the "New Industrial

Comparison of the Treatment of Externalities in Krugman's

Type of Externality Krugman New Industrial Geography

Krugman's contrary emphasis on pecuni- tance to envisage nonmarket linkages

question of adjustment to the effects of European Commission (Commission of

Krugman labor mobility prevents the curious that Krugman is at pains to

out depressed regions, should encourage Trade and the Regional

Peddling Prosperity (1994c), he questions industrial policy." His initial skepticism of

implicit in Porter's (1990) major work on industrial districts celebrated by Alfred

Krugman is at pains to distinguish this tion and specialization leads to regional

synthesis of location theory and trade is to pursue a closer exchange between

lian view of industrial localization. But

Lovering, J. 1990. Fordism's unknown succes- gence or divergence? A study of regiona

You might also like