Research in Developing Countries

RAPID RECONNAISSANCE GUIDELINES FOR AGRICULTURAL MARKETING AND FOOD SYSTEM-RESEARCH IN DEVELOPI.NG COUNTRIES by John S. Holtzman Working Paper No. 30 1986 M5U INTt!ItNATICNAl Dt!VElOPMENT PAPERS Carl 1<. Eicher, Carl Lledho!m, and Michael T. Weber Bdltors The MSU International Development Paper series is designed to further the comparative analysis of international development activities in Africa, Latin America, Asia, and the Near East. The papers report researr.:h findings on historical, as well as contemporary, international development problems. topics, such as The series includes papers on a wide range of alternative rural development strategies; nonfarm employment and small scale industry; housing and construction; farming and marketing systems; food and nutrition policy analysis; economics of rice production in West Africa; technological change, employment, and income distribution; computer techniques for farm dnd marketing surveys; farming systems and food security research. The papers are aimed at teachers, researchers, pOlic.y makers, donor agencies, and international development practitioners. Selected papers will be translated into French. Spanish, or Arabic. Individuals and institutions in Third World countries may receive single copies free of charge. See inside back cover for a list of available papers and their prices. For more information, write to: MSU International Development Papers Department of Agricultural Economics Agriculture Hall Michigan State University East Lansing, Michigan 48824-1039 U.S.A. r RAPID RECONNAISSANCE GUIDELINES FOR AGRICULTURAL MARKETING AND FeOD SYSTEM RESEARCH IN DEVELOPING COUNTRIES by John S. Holtzman . Department of Agricultural Economics Michigan State University 1986 This paper is published by the Department of Agricultural Economics, Michigan State University, under Food Security in Africa Cooperative Agreement DAN-1l90-A-OO-4092-00, jointly funded by the Bureau of Science and Technology (Office of Rural and Institutional Development) and the Africa Bureau (Office of Technical Resources), U.S. Agency for Interna tional Development, Washington, DC. An earlier version of this paper wa~ developed while the author was em ployed by USAID under the Small Farmer Marketing Access Project (No. 936-5313). MSU is an Affirmative Action/Equal Opportunity Institution , I ISSN 0731-34-38 @ All rights reserved by Michigan State University, 1983. Michigan State University agrees to and does hereby grant to the United States Government a royalty-free, nonexclusive and irrevocable license throughout the world to use, duplicate, disclose, or dispose of this publication in any manner and for any purpose and to permit others to do so. Published by the Department of Agricultural Economics, University, East Lansing, Michigan 48824-1039 U.S.A. ,, II Michigan State ACKNOWLEDGEMENTS These guidelines for rapid reconnaissance have evolved out of my work on food systems and agricultural marketing with Michigan State University and USAID over the past eight years. As a doctoral candidate at Michigan State, I conducted studies of livestock and food systems in Cameroon and Somalia 1 using both rapid appraisal methods and longitudinal, cost-route formal surveys. Upon completing my studies at Michigan State, I became a technical advisor in agricultural marketing to USAID (1982-85). In this capacity I had numerous opportunities to use rapid appraisal technique:; in doing feasibility studies, identifying and evaluating projects, and initiating programs of applied marketing research. I also reviewed many agricultural marketing studies carried out in developing countries. Many of these studies are conducted under limiting time and resource cons traints, and their quali ty ,varies greatly. The findings of shor t term studies often seem to reflect the interests or preconceptions of the researchers. Moreover, biases associated with the timing of the rapid appraisal exercise, as wen as the selection of places to visit and informants to be interviewed, are often evident. • This paper represents an attempt to improve the quality of marketing research conducted in developing countries under time and resource constraints. It is hoped that the paper will stimulate discussion of rapid appraisal methods among developing country analysts, researchers at universities, international agricultural research centers and other institutes, and technicians in donor agencies and consulting firms. W~i1e my experience in longer term programs of research has instilled an appreciation of the limits, potential misuses and biases of rapid assessments, I am convinced that rapid reconnaissance methods are useful for learning a lot of policy relevant information about marketing systems in a short period of time, and for identifying system problems, .. I constraints and opportunities. Rapid appraisals are especially useful as an input into the design of longer term programs of marketing and food systems research, in the same way tha t explora tory surveys are invaluable in ini tia ting programs of farming systems research. This paper draws heavily on earlier work done at the Institute of Development Studies at Sussex University, on papers presented at conferences at Sussex on rapid appraisal in 1978 and 1979, on papers by farming systems researchers at CIYMMT and the University of Florida, on books on the systems approach to agribusiness management by Harvard Business Sch, ~ 1 researchers, and on efforts of applied researchers in many fields, including agricultural econ mics~ anthropology, geography, development iii administration, and agribusiness management. lowe an intellectual debt to colleagues at USAID and in the Department of Agricultural Economics at Michigan State University. Merle Menegay provided important substantive and conceptual input at an early stage in the preparation of the paper. Harold Riley reviewed several drafts of the paper and made many useful suggestions and comments. Nicholas Minot reviewed the paper at a late stage in its development and offered valu"ible suggestions on how to improve its organization and substance. James Shaffer, Stephan Goetz, Michael Morris, Thorn Jayne, David Smith and David CampbeU provided critical input in the final stages of the paper's development. Michael Weber offered invaluable critical and edi torial assistance in moving the paper from draft to publication stage. FinaJIy, I gratefully acknowledge constructive and critical reviews of earlier drafts of this paper by Edgar Ariza-Nino, John Abbott, Lehman Fletcher, Jerry Martin, David Atwood, and Michael Burton. • , iv TABLE OF CONTENTS Chapter ~ . -I ACKNOWLEDGEMENTS.. . . . . . . . •.• .. .. •. . . . .. .• . . . . . .. ... •. •. .•. .. iii LIST OF TABLES. . . . . . . . . .. . .. . . . . .. . . . . . . . . . .. .. . . • . .. . . . . . . . . . . .. . . . .. . . v LIST OF FIGURES I. ; II. ix . IMPOR TANT CHARACT!:.RISTICS OF RAPID RECONNAISSANCE IN A FOOD SYSTEM CONTEXT. .. • ... . .. .. . .. . . •. .. . . •. . .• . . . .•. . ... . . .. 5 1 2 2.1 Rapid Reconnaissance Focus. • • • • • • • • • • • • • • • ... • • • • • • • • • • • • • • • • 5 2.2 Ge<>graphic Scope. .. . • • • . • • • • . .. . .. • • • • • • • • • • • • • • • • • • • .. • • • • . .. • . 6 2.3 Dura tion and Timing of Rapid Reconnaissance...... • • • • • • • • • • • • • Composi tion of Rapid Reconnaissance Teams ••••••••••••••••••• 6 7 AN ANALYTICAL FRAME WORK TO GUIDE FOOD SYSTEM APPLIED RESEARCH AND DEVELOPMENT ACTIVITIES. • • • • • • • • • • • • • • • • • • • • • • • 8 Need for an Analytical Framework •••••••••••••••••••••••••••• The Concept of a Food System and Strategies for Food 8 System Development ••••••••.•.•.••••••.•••.•••••••••••••.•. 8 The Subsector Approach: Evolution and Principal Character is tics .....,........................................ 11 KEY AREAS OF INVESTIGATION IN RAPID RECONNAISSANCE OF COMMODITY SUBSECTORS •• • •••••• • •••• ••••• •• • • •• •• •• • • • •• •. •• • • 18 ANAL YSIS OF PRICES AND MARKETING COSTS AND MARGINS....... 22 5.1 Data Collection and the Quality of Price Data..................... 5.1.1 Who collects the data? ••••••••••••••••••••••••••••••••• 5.1.2 What types of data are collected? ••••••••••••••••••••••• 5.1.3 Where are data collected? •• •• •• • • • •• •• •• •• • • • • • • • • • • • • • ~.1 5 When are data collected? •• •• • • • •• •• •• • • • • • • • • • • • • • • • • • 5.1.5 How are the da ta collected? • • • • • • • • • • • • • • • • • • • • • • • • • • • • 5.1.6 How are the data analyzed? •••••••••.•••••••••••••••••• 5.2 Recommended Typp.s of Price Analysis.. • • •• • • •• • •• •• ••• • • • • • • • 5.3 Analysis of Price Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ~.5 Analysis of Marketing Costs and Margins •••••••••••••••••• :.... 22 23 23 23 23 3.1 3.2 3.3 VI. . .. .. .. • .. .. 3 2.l~ V. .. . . .. . . . .. .• .. Increasing Interest in Rapid Reconnaissance Techniques •••••••••• The Appropriate Uses and Context for Rapid Reconnaissance •••••• Organization of the Paper " ~ . 1.1 1.2 1.3 IV. .. . . . INTRODUCTION IIIi III. .~ 2~ 2~ 2~ 29 30 KEY INDICATORS AND PROXY VARIABLES IN RAPID RECONNAISSANCE............................................... 31 6.1 6.2 31 31 Degree of Commercialization of an Area or Region.............. Vehicle Ownership and Availability............................ v Chapter 6.3 VII. The Relationship Between Administered and Parallel Prices •••••••••••••••••••••••••••••••••••••••••••••••••• of • • INSTITUTIONAL AND NON-MARKET FACTORS IN FOOD SySTEMS.... 34 7.1 Transactions Costs and Market Improvement •••••••••••••••••••• 7.2 7.3 Property Rights ••.••••••••••••.•.•••••.•.•••••••••.••.•••••. 34 35 36 Non-Economic Factors Ideology 7..3.1 7..3.2 Religion Cultural Endowments 7..3.3 . IJ • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • VIII. PREPARING FOR . . 41 41 42 44 IMPLEMENTING RAPID RECONNAISSANCE SURVEyS................ 47 Elements of Rapid Reconnaissance Field Work.................. Where to Begin the Rapid Reconnaissance •••••••••••••••••••••• Processes, Functions and Facilities to Observe During Rapid Reconnaissance..... • • • • • • • • • • • • • • • • • • • • • • • • • • • • Selection of Key Informants •••••••• ,......................... Informal Interviewing of Key Informants ••••••••••••••••••••••• Interviewing Techniques. •• • •• •• •••• ••• • • •• • • • • •• • •• •• • 9.5.1 9.5.2 Building in Consis tency Checks.. • • • • • • • • • • • • • • • • • • • • • • • 9.5.3 Repeat Interviews. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Interviewing Village Headmen and Other Local Informants. • • • • • • • 47 48 9.6 50 51 55 55 56 57 57 9.7 Group In terv iews ," . 58 9.8 9.9 9.10 9.11 Informal Delphi Techniques....... • • • • • • • • • • • • • • • • • • • • • • • • • • • • Recording Ra'pid Reconnaissance Findings •••••••••••••••••••••• Speeding up Information Gathering in the Field.. •• •••• •• ••• •• •• • Periodic Meeting of Rapid Reconnaissance Teams and Exchange of Preliminary Findings ••••••••••••••••••••••••••••• 59 60 PRESENTING RAPID RECONNAISSANCE FINDINGS.. •• • • • •• • • •• •• • • • 63 10.1 10.2 "0 61 Writing up the Results of the Rapid Reconnaissance Survey. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 Presenting the Principal Findings of the Rapid Reconnaissance • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 64- Follow-up to Rapid Reconnaissance. •• •• •• • • • •• •• •• •• • • • •• •• •• • 64 LIMITAnONS OF RAPID APPRAISAL ••••••••••••••••••••••••••••••• 66 11.1 11.2 Time-Boundedness and Bias. • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • Limitations of the Subsector Framework....................... 66 66 CONCLUSION. . . . . • . . . . . . . . . . . . • . . . . . . . . • . . . . . • . . . . • . • . • . . • . . . • . . 69 BIBLIOGRAPHY.. .•.• . • .• .. •. .. •.. .• •• .. ..•. . •• •. •. •.• .• .. •. •. • .• 71 10.3 XII. 39 Prepara tory Visi t and Planning.. • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • Selecting the Rapid Reconnaissance Team •••••••••••••••••••••• Planning the Rapid Reconnaissance Field Work... • • • • • • • • • • • • • • • Review of the Literature and Analysis of Secondary Data. • • • •• •• • 9.4 9.5 VI. .36 37 41 9.1 9.2 9.3 X. RAP~D ., RECONNAISSANCE FIELD WORK........... 8.1 8.2 8.3 8.4 IX. .32 vi LIST OF TABLES Table 1. Senega! Food System Matrix ...................•.................•. 10 2. Key Areas of Investigation in Rapid Reconnaissance •••••••••••••••••• 20 3. Price Analysis and Potential Problems •••••••••••••••••••••••••••••• 25 4. Key Informants in Food System Research: Advantages and Disadvantages as Informants ••••••••••••••••••••••••••••••••••• 53 .. vii LIST OF FIGURES 1. A Schematic Representation of the Structure, CO;lduct, Performance Paradigm as Applied to the Commodity Subsector Approach •••••••••••••••••••••••••••••••• ix 13 I. INTRODUCTION This paper develops rapid reconnaissance guidelines for conducting research on agricultural marketing components of food systems. The term rapid reconnaissance (RR) will be used interchangeably with the designations rapid appraisal, rapid assessment and rapid diagnostic assessment in this paper. A rapid reconnaissance survey is a broad and preliminary overview of the organization, operation and performance of a food system or components thereof, designed to identify system constraints and opportunities. It can be used as a tool for identifying system dynamics, linkages and overall esp cia~ly problems, which can then be examined more intensely during follow-up programs of research. 1.1 Increasing Interest in Rapid Reconnaissance Techniques Interest in rapid appraisal techniques has increased during the past decade for several reasons. First~ long term programs of farming and food systems research have become increasingly, and in some cases prohibitively, costly. These studies absorb significant financial and manag,ement resources and effective techniques are needed to help keep costs down. There is also frustration with many data-intensive, long term studies that fail to generate usable research findings within a time frame suitable to policy-makers. Such studies may be useful over the long run in generating knowledge about agricultural production and marketing systems. Policy-makers must make decisions on the basis of limited yet best available information, however, and cannot wait interminably for research results. At the same time, entry, processing and analysis of longitudinal survey data are becoming far more rapid as the use of microcomputers in field research expands. A second and perhaps most important reason is that many researchers who have carried out both short and long term studies have found that rapid reconnaissance techniques, when used judiciously, can yield valuable insights about subtle and dynamic factors in food systems. Researchers who have undertaken longitudinal surveys often find that the milnagement requirements are so great that they have little tim€' and few resources available for informal interviews and case studies. Many who have reserved time for these less formal data collection methods report that they are often a quicker and sometimes more effective means (as compared to data-intensive surveys) for learning about the interrelationships and linkages in farming and marketing systems, understanding system constraints and opportunities, and identifying potentially viable interventions. Farming systems researchers have successfully used rapid reconnaissance techniques to identify farm-level constraints and guide allocation of applied research resources to help allevia te these constraints. Researchers such as Byerlee and Collinson report that accurate qualitative and quantitative information can be' generated by informal surveys which serve as a baseline against which later comparisons can be made (Byerlee, et. al., 1980; Collinson, 1982). Notable contributions in the development of rapid reconnaissance methodologies applied at the farm' level are the work of the Institute of Development Studies of Sussex University in the U.K. (Chambers, 1980, 1981), papers by farming systems researchers such as Hildebrand (19""9, 1981), Collinson (1981, 1982) and Byerlee (1980), and papers on rapid reconnaissance tor developm.ent administration by Honadle (Honadle, 1979 and 1982). Entire issues of two journals (Agricultural Administration, 1981 and Institute for Development Studies Bulletin, 1981) were devoted to rapid reconnaissance. More recently, in October 1985, a con ference on rapid appraisal was held at the Univ ersi ty of Khon Kaen in Thailand. 1.2 The Appropriate Uses and Context for Rapid Reconnaissance As will be discussed in this paper, RR has several potential uses in marketing and food system research. First, it can be used by governments or donor agencies in the identification and design of agriCUltural marketing projects. Second, RR methods can be used by private firms to evaluate the feasibility of making particular types of investment, such as a new storage or processing facility. Third, a rapid appraisal can generate knowledge about an agricultural commodity subsector needed for making a policy decision or for evaluating the effectiveness of a policy or set of policies. A fourth potential use of RR is in monitoring and evaluating the effect of an agricultural marketing project or policy. A fifth use for RR methods is as a descriptive and diagnostic tool in the design of a program of longer term agricultural marketing and food system research. Despite these potentially varied uses, RR surveys are often inspired by policy initiatives, wherein host country governments and donor agencies require help in identifying possible projects or need for policy reform. In responding to such short term 2 r..l planning needs, RR risks being a subjective exercise, resulting in strong advocacy of j:redetermined positions. In these cases rapid appraisals provide economic or technical justificatior;s for project proposals that host country governments and donor agencies intend to implement. Researchers should strongly resist pressures from research sponsors to simply devise justifications for predetermined food system interventions. At the sam e time, researchers do need to be responsive to their client groups. When the cHent is a government organization or a development assistance agency, researchers are expected to prescribe project proposals or policy reforms. Hence, the applied research will have an action orientation. In doing rapid reconnaissance this poses a dilemma. While research sponsors expec! policy prescriptions, rapid reconnaissance techniques seem most useful as a tool for learning more about commodity marketing or other aspects of the food system through observation and analysis, for identifying system constraints and opportunities, and for identifying promising areas of further applied research. In most circumstances, it should not serve as a basis for generating definitive policy prescriptions, although policy-makers are free to use the RR prescriptions as they wish. Policy prescriptions will often be ter.tative and preliminary. In some cases rapid reconnaissance surveys may generate robust and conclusive enough findings so that policy-makers can choose with considerable confidence to implement particular policies or to discontinue policies with adverse consequences. Yet recommer.ded interventions emerging from rapid assessments, particularly improved technologies, marketing institutions, and management methods, should be tested as pilot projects and regarded as experiments. 1•.3 Organization of the Paper ThE' re~ap is organized into two major parts. The first part examines the substance of rapid reconnaissance in agricultural marketing research--its important characteristics (Chapter 11), the analytical framework used in rapid reconnaissance of commodity marketing systems (Chapter appraisals ~Chapter 110, key areas of investigation during rapid IV), analysis of prices and marketing margins (Chapter V), proxy variables and key indicators (Chapter VI), and nonmarket or institutional economics, and noneconomic factors (Chapter VII). The second part of the paper discusses the process of rapid reconnaissance (Chapters VIII and IX), including sections on the preparation for RR field work (Chapter VIII) and implementation of RR surveys (Chapter IX). These include selection of the RR team, research planning, review of the literature and analysis of 3 secondary data, the elements of RR field work, selection of key informaflts, and information gathering techniques. Report prep",ration, presentation of RR findings, and followup to rapid reconnaissance surveys are discussed in a secdon on wrapping up rapid reconnaissance (Chaptt:r X). Limitations of rapid appraisal methods are addressed in Chapter XI, and concluding remarks are made in Chapter' XII. 4 II. IMPORTANT CHARACTERISTICS OF RAPID RECONNAISSANCE IN A FOOD SYSTEM CONTEXT 2.1 Rapid Rt;<:onnaissance Focus Doing rapid reconnaissance of an entire food system is an unmanageable agenda. Yet keeping relevant system relationships and linkages as part of a narrowed-down focus is important. Classifying commodities into related groups, such as staple grains, legumes, vegetables, fruits and livestock products is a first step in narrowing down that agenda. The technical characteristics of related groups of commodities are often similar, which generally leads to similar organization and operation of commodity subsectors. RR teams may wish to focus on one commodity in each group, because it is representative, the most important in terms of volume or value, perceived to be the most problematic, or regarded as having potential to generate the most income and foreign exchan'ge earnings or savings. Many RR surveys will focus only on one commodity production-distribution subsystem and in any case no more than two or three related commodities, such as grains or ruminant livestock. In a relatively short period it is impossible to look at aU agricultural commodities in a food system in any depth. Restricting the commodity focus allows researchers to interview several participants at each major stage of each subsector, as well as to observe marketing processes and activi ties, wi thin the time constraints of RR. Alternatively, researchers can focus on food distribution in one or more urban markets. The focus may also be on staple commodities which provide most of urban consumers' calories, particUlarly where expenditures on those staples claim a high proportion of household budgets or where urban consumers have difficulty meeting basic daily calorie requirements. In addition, analysts might focus on commodities for which the urban marl<et is growing most rapidly. As another alternative to a commodity focus, researchers can do rapid assessments of particular industries in the food system, such as grocery stores, vegetable canners and processors, grain millers, or agricultural equipment manufacturers and distributors. The need for such a delimited study might arise if particular functions in the food system are performed poorly and at high cost, or if poor performance of those functions constitutes a system bottleneck. 5 2.2 Geographic Scope Rapid reconnaissance will often be limited in geographic scope to one region, or to a major urban market and the rural areas from which it draws its food (food shed). The geographic breadth of coverage will depend, in part, on resources available for the study. Area coverage can be expanded if several teams of applied researchers work concurrently in different regions. Geographic coverage will also be guided by judgements as to the need for visiting additional producing areas and markets. Researchers will have to e'valuate whether it is worth the extra co~t to expand coverage. For example, does a producing area appear, on the basis of limited yet best available information, to be sufficiently different from other areas (in terms of crop mix, marketed surplus or numerous other criteria) to justify RR field work? In a survey of grain marketing in Burkina Faso conducted in 1984-85, the Universi ty of Michigan did RR field work in several surplus and deficit regions (Sherman, Dejou, et. al., 1986). In an assessment of secondary crops marketing ifl Indonesia, USAID, the Food Crops Directorate of the Ministry of Agriculture, and several local universities have carried out RR surveys in several surplus producing regions which are characterized by different agroecological conditions, cropping patterns, and marketing systems (Menegay,1985). In some cases RR teams may wish to look at more than one major urban market and their respective food sheds for comparative purposes. Where food systems are driven by foreign demand, RR teams may decide to surV(~y demand and marketing opportunities in key importing countries. 2.3 Duration and Timing of Rapid Reconnaissance Rapid reconnaissance surveys are usually at least one month and less than three months in duration, depending on prior knowledge of the food system components and regions where the RR will be conducted. In countries where many commodity subsectors have been heavily researched, and where the objectives of the RR are quite delimited, a rapid appraisal could be conducted in a few weeks. On the other hand, if little is known about a region and subsector, and the objectives of the RR are ~omewhat diffuse, rapid reconnaissance has a more difficult time standing on its own, and becomes important as the first step in a longer term research program, requiring collection of a great deal of primary data. 6 Conducting RR during a period. of less than three months precludes observation and data collection over an entire annual agricultural production and marketing c:ycle. Yet this may not be necessary, as RR uses longitudinal data from previous studies and secondary sources to document patterns of secular and seasonal change. Participation of researchers in rapid reconnaissance who have followed commodity subsystems over time will also improve understanding of trends and changes in the food system, as well as seasonal variations in agricultural production, marketing and consumption. The dming of RR will depend upon the per<':eived need for observing particular marketing functions and processes. Farming s}'atems researchers typically plan rapid reconnaissance surveys to coincide with the, crop growing season, particularly during 'I critical periods when crops are flowering, or during pedoda of peak household labor utilization (e.g., weeding periods in Africa). Timing in marketing and food system studies will depend in large part on what are initially perceived to be the key food system constraints and opportunities. If policy-makers are concerned m0st about storage practices and losses and alleged speculative practices, RR could be undertaken during periods of storage and sales from ~torage. If it is suspected that producers receive unremunerative prices for produce sold during the pdst-harvest period, the RR could be carried out at this point to observe transactions and terms and condi tions of sales. RR surveys of particular industries in the food system that are judged to be bottlenecks or highly inefficient might be timed so that industry performance could be most easily observed. 2.4 Composition of Rapid Reconnaissance Teams The composition of the RR team will vary according to preliminary identified food system problems and needs. Rapid reconnaissance can be conducted by one researcher, typically a social scientist, or by several researchers with similar disciplinary training. If the objectives of rapid reconnaissance are limited, multi-disciplinary teams are not usually necessary. The quality of most RR surveys will usually improve, however, if researchers with complementary disciplinary skills participate. These teams may include an agricultural economist or economist: an agribusiness management specialist, an economic anthropologist, a post-harvest technician, a transport economist, institutional analyst, or commodity specialist. It is important that an analyst with a general background in agricultural marketing (typically an agricultural economist) lead the team, and edit and complete the final RR report to provide an integrated picture. 7 III. AN ANALynCAL FRAMEWORK TO GUIDE FOOD SYSTEM APPLIED RESEARCH AND DEVELOPMENT ACTIVITIES 3.1 Need for an Analytical Framework In conducting applied research under time and resource constraints, it is helpful to work from an analytical or conceptual framework that is cast 1n a development strategy perspective. This enables researchers to distinguish between information and factors, programs and policies which are 1) likely to be relevant and important for helping to improve performance of the system, as opposed to 2) those which arc inte ~ ting but nonessential or perhaps trivial. Without a framework of analysis on which to ''lang dispara te observa tions and findings, the researcher risks missing the big picture and focusing on narrower problem areas, which may not be the most critical ones to improving system productivitY'and performance. 3.2 The Concept of a Food System and Strategies for Food System Development The broad analytical framework underlying RR is the food systems approach, developed and elaborated by researchers at Michigan State University (see Harrison et• ......\:~ al., 1974; Shaffer, 1980; Riley and Staatz, 1981; Shaffer with Riley, Weber and Staatz, 1983). The food system incorporates the agriculture and livestock sectors, and industries within those sectors. The participants in the food system that produce, transform and distribute the full range of agricultural commodities include input producers, input suppliers, agricultural producers, domestic traders, importers and exporters, processors, retail firms, institutional buyers and consumers. A food system has both horizontal and vertical dimensions. The horizontal dimension refers to firms within a particular industry, or to a particular stage of the food system where a similar set of functions are performed. Examples of this include the transportation industry, the tomato canning industry, agricultural equipment manufacturers, and retail stores. The vertical dimension refers to subsystems of single commodities or relatively homogeneous groups of commodities. This dimension is conceptualized as vertical because it cuts across stages of the system, where different functions, such as input distribution, production, assembly, storage, transport, processing and product distribution are performed. Initially labelled as a subsector (see Shaffer, 8 1973), but better described as a subsystem, a commodity subsystem incorporates productive transformation and value adding at each stage as il:lputs are supplied and a product moves from the farm to the consumer. One tool for usefully thinking about the food system and its dimensions is a matrix of agricu1t.ural products and functions, as conceptualized by Shaffer•. Agricultural products are arrayed vertically, corresponding to commodity subsystems. Food system functions, are generally carried out by related groups of firms, or industries, are w~,ich arrayed horizontally. Table 1 illustrates this matrix as applied to the food system in Senegal (Ndoy~ research, it and Newman, 1984). In designing program:i of applied food systems i~, not possible to study each cell in the matrix, given funding and resource limitations. Subject to the approval of policy makers, researchers C\re forced to make choices about those parts of the system on which they intend to concentrate their efforts. These chokes will be governed by numerous possible criteria, including political factors, the importance of commodities as staples or in generating/saving foreign exchange, regional considerations, or the degree to which particular functIons or industries constitute system bottlenecks. Judgements as to the liI<ely payoff from doing the research, as well as the feasibility of carrying out the research, will of course come into play. Another way to think about food systems is to develop an input-output or food accounting matrix (see Hay, 1980). This approach emphasizes quantification of technical production and transformation coefficients at different stages of the system, as well as input and commodity flows between parts or cells of the system. While this mat~ix serves as a useful accounting tool, it requires voluminous data and it is essentially a static construct. It stresses quantification of interrelationships rather than describing and diagnosing system constraints, bottlenecks and opportunities, and striving to capture dynamic interactions and disequilibria. It is difficult for all the cells of such an accounting matrix to be filled with the data available in many developing countries. The food· systems framework as developed and applied by MSU researchers is a developmental approach (see Harrison et. al. and Shaffer et. al.). It seeks to identify promising opportunities for improving the productivity of food systems as well as to diagnose barriers to improved system performance. Developing stra tegies, programs and projects for both private and public sector participants is central to the approach. It also assumes that there are alternative wo.ys to constitute or organize the system so as to 9 . '; .'.; Table I SENEGAL FOOD SYSTEM MATRIX Commodity Subsystems Prodl!ction/Distribution Functions Millet Sorghum Maize Rice Wheat Peanuts Vegetables Fruits Cotton Sugar Livestock Fishing Wood Input Distribution Extension Production Transformation o Storage Transport Exchange. Transactions Financing Coordinadng Functions Prices Information Grades and Standards Regulations Property Rights Exchange Arrangements Risk-Sharing Mechanisms Consumption Source: Ndoye. Ousseynou and Mark Newman. -Approches Methodologiques pour J'Etude de la Commercialisation des Prodults Agricoles et Alimentaires au Senegal." Actes de l'Atei~r de Kahone. 8-13 Mars 1984. Institut Senegalais de Recherches Agricoles. Bureau d·....nalyses Macro-Economiques. Document de Travail 84-2. Dakar. December 1984. .... improve productivity and performance. These alternatives will vary from country' to country and political/economic setting to political/economic setting. It also recognizes the possibility of alternative outcomes of different :;pts of policies and programs, rather than promoting canned solutions or universally applicable blueprints, such as simply getting the government out of the food system, or "modernizing" of marketing institutiolls so they more closely ressemble those found in the US, Western Europe or Japan. In contrast, the food system approach argues that the government is able to play, a.'1d in many cases must play, an important regulatory and facilitating role in food system development•.The mix of pubBc/private sector roles and activities will vary from country to country, depending upon historical and political factors, the stage of economic development, and the level of human capital development. An important part of a food system analyst's task is to examine the economic viability of alternative institutional arrangements, policies and technologies and their actual or probable effects on 'food system performance. Analysts then present the likely costs and benefits of these alternatives to policy-makers. Researchers are also encouraged to prescribe that set of policies or interventions that achieves government policy objectives at lowest cost. and is most likely to foster growth and increased productivity. The extent to which linkages between different participants in the food system and among the agricultura I, indus trial and service sectors can be strengthened is also an important considera tion. In the interest of making agricultural marketing and food system research in developing countries manageable, the next section will describe the important features and emphases of subsector or subsystem studies. While not all agricultural marketing studies in developing countries will follow a subsector format, many of them will, and the subsector framework is a useful method of examining commodity marketing subsystems. 3.3 The Subsector Approach: Evolution and Principal Characteristics Given the above strategy for promoting food system development, the approach to rapid reconnaissance presented in this paper is grounded in the commodity subsector (or subsystem) analytical framework. This approach has been best articulated by researchers of NC Project 117, partially funded by USDA, who have examined commodity subsectors in a comprehensive study of the U.S. food system (see ~arion and NC Project 117, 1986), and by researchers at Michigan State University (see Harrison et al., 1974 and Shaffer, 11 Riley, Weber and Staatz, 1983). A schematic representation of this approach, developed by Bruce Marion, is shown in Figure 1 (Marion, 1976 and 1986). The subsector framework builds on the structure, conduct, performance framework, developed by industrial organization theorists such as Bain (1968) and Scherer (l980). Industrial organization scholars examine the structure (organization), conduct (operation) and performance of industries, defined as a group of firms which produce similar cor:nmodities, (e.g., the meat-packing industry, the automobile manufacturing industry, the micro-computer industry). The focus of industrial organization is on horizontal linkages among firms. It is essentially static, taking industry structure as a given and anticipating certain types of conduct and performance characteristics from alternative structures. The macroeconomic environment and institutional landscape (rules and regulations, distribution of resources) are generally assumed to be exogenous. The subsector approach is similar to industrial orgarllzation theory in focusing on the performance consequences of alternative forms of industrial or economic organization. Yet it differs in several fundamental respects. First, a subsector is a vertically linked set of participants (firms or organizations) which produce a related output or group of outputs. Rather than adopting a conventional dichotomy of on-farm being "production" and beyond the farmga te being "marketing", the subsector approach emphasizes transformation, adding of value, and transactions at every stage in a subsector from input supply through production, marketing and consumption (Shaffer, 1973). In agriculture, subsectors are organized around particular commodi ties or commodity groups, such as grains, ruminant livestock, fruits, tuber CI"OPS, or anyone of many individual commodities. Participants in agricultural commodity subsectors include input suppliers, farmers, first handlers, wholesale traders, retailers and consumers. Consumers are considered as subsector participants, because their demand for agriculturai commodi ties in the aggrega te influences production and marketing decisions of all other participants in the subsector. A second fundamental difference of the subsector approach is that it closely examines underlying supply and demand conditions for commodities and likely trends and changes in commodity output and use. The subsector approach is demand oriented. Demand drives commodity subsectors or, alternatively, it pulls commodities through these subsectors. Shifts in demand induced by changes in relative prices, in purchasing power of consumers (per capita income), and in consumer tastes and preferences, affect 12 Fipe I. A Schematic Represent.-tion of the Structure, Conduct. Perfcwmance Paracllgm as Applied .to the Commodity Subsector Approach BASIC CONDITIONS STRUCTURE Indus tr)' Struetwe Production trlnds IIId leDlrApIllc diatrlbullon Consumption dIu&cterinlcs - Crowlh or decline On domestic &lid forelln /DUlceu) - Priee.lncome&lld crass e1Anicllies of demand - Differences b, socioeconomic &lid income goup - RurAl/_ban differences nme chArACterlslics of production and muket c)'Cles Trpe And delree of uncenAintiesl commodlly price pAlIerns, covernment policies, w..lI:ec/climAte pAlIerns, aCCClI 10 IIId ImporlAnce of ellterRAl muleets Laws and aovernment poUcies &lid reaulAtlons Macroeconomic v.iabla AS incentiva or dWncenllYl:1I ucbMCe rAte,lnterest And WAce rAta,lnfiatlon rAte &lid di!ferentJallmpACt I Firm Declalon En"lrQlUllmt _• • • • ·-'1 • AJtemallvCl • "enllva • Ccnll'OllIId lnQuenc:e - ...." CONDUCT htultr)' Product ItrAteu PrlclnC behA"ior Adverlilln& Reseuch and U-Yltion Mercerl and di"estitwa Risk _aement pr.cticel FunctionAlltruetura Ioca..... tImJn& and clustlll'lniof functions No. of IlA&CI No. of parallel chInneIs informAtion l1Item • Trpc of informAtion (andes. market conditions. eteJ - Distribution • Cost Structure of audlorll)'. r"ts and conlrol: decllion an.tlDmr ElldIAnae institutions (auctions. buylna IUlions. etcJ Trpes of udIAn&e (spot. cantr.etl, l)'InC acreementl. ele.) IUs:: shArln& innltutlonl and atran&ements Intlll'-slAle differences Ciocatlon, Ilze of enterprlle. ICIIOftIUty. production chAr.cterlsticl); nature of 1SIefllb1,. sortln& and 1)'I\CIlronIzin& tuIcs PElU'ORIiANCI! I W SubMcIDr ClrpnJution No.1IId lize of buyen And Iellen Entry IIId ellt conditions Product dIu.cterlstlcs - PerlshAblUty - QuAUt, requirementl - Differentiation Tedlnoloar dluJCOlt functions - CaplullntenaltYJ mlnImlllll efficient firm alze - IlAte of dlAnCe CapAcll)' 5peciallzatlon/diverllflatlon VeruClIllntecrAtlon finAnclnC and credit d1ArACterlllia Collective orpnJzatlonl • Cooperatives • TrAde ASSOCiatlonl BuIlnaa objectl,,", attitudes And ClIpablUtla Fr~ of pwc:bua IIId ales TechnlcAl and operatlonAl efficiency Prieln& efficiency (proDt And Sublec:lOr EffC:011 10 IhIft CDntrol • Type of e.dIAn&e UHd CoordinAtion .cti"Uia • Prediction of futwe aupply. demancf.and price - information communicated - QuaUtrlPecUiClition • Schedulln& and tlaoin& synchronization output leYels) Product chArACteristics • QuAUtr/wtlolao_ • V.iely Proarasl"enCII (proceu and . product) Sell1n& .ctlvllies • Eapense • Influence on consumption patterns and social VAIua ....lcet access and/or Ioreclosure - Efforts 10 influence IntenlAae cooperatlon/CDnfUct Process of deterllllnln& terma of eKeNnae (private, truty. Admlnistered.llIcf-offer. .cceplAnce, eteJ Response 10 chance forces A*ptecl frOIIi Bruce W. Marlon and He 117 Commltt... The OrnnlZ&tlon &lid Pcr'or!lW!C! o'.me U.s. Food SntCfft. D.C. He.v. and Coaopan,. Lellinl_. u.ssadlUHtts. ItI'- Subseclor A11oCl1t1"e ACCUnCJI eZlent to which suppl, oUerillp _tela cIcmAnd preferences with re&ard tID quantity. quaUty. tialinC and Ioation Stablnl)' of output. prlcea and proDts TedlnIal and operAtionAl efflcicncy • At eiCh Ita&e and ill IInIodnc Iuses (transaction costs) I!quIty with reprd to dlstrlbutloh • • RetW'nl"I. Ift"estnaents and risks • RI&bts and control " .. InYCllIlIents IIId risk _ AIt:curacy. Adequaq and equll)' of informAtion distributed SubMcIOr ....plablU.' LeYd ancItype of elllfllorlllellt w..te and spo1iaCe - Product wute • Re_ce ClIIlICrYation • Cap4citr utillz4t1on in a significant way the set of incentives facing participants throughout the subsector to supply inputs, produce, tr:ansform, store and transport agricultural commodities, and to supply products to consumers or end users in forms, at times and at places that they desire. While changing demand is often an important dynamic force in transformation of commodity subsectors, shifts i'"1 supply can '11so be an engine for change. Increased productive capacity, at the farm level or at other stages of the subsector, can lead to reorganization of commodity subsectors and affect subsector performance in important ways. As an example, heavy investments by the government of Senegal in two large scale dams on the Senegal River will dramatically increase the land area under irrigated cultivation. The subsequent expansion in output will likely lead to significant changes in the organization and operation of the rice subsector in Senegal, inducing entry of traders and processors into the sub ec~or, decreased flows of imported rice from Dakar to markets in Northern Senegal, and increased flows of locally produced rice to markets outside Northern Senegal (Morris, 1985). By emphasizing ongoing and potential changes in supply and demand, the subsector approach is dynamic in orientation and goes beyond the sta tic ef ficiency analysis common to many neoclassical studies of marketing systems or industrial organization studies (Riley and Staatz, 1981). A third important difference between the subsector approach and industrial organization studies is the emphasis in the former on the nature of coordinating institutional arrangements, such as contracts and vertical integration, and the role of selected firms, cooperatives, industry associations, and other active coordinating agents. Subsector studies examine vertical linkages between firms that contribute to the production, transformation and marketing of related commodities. In the agricultural sector the level of agricultural output is partially dependent on climatic factors, and aggregate output and agricultural commodity prices are uncertain from one year to the next. Producers, processors, traders and institutional buyers have an incentive to devise institutional arrangments for coordinating fluctuating supplies so they match demand without excessive price fluctuation. Coordinating mechanisms also facilitate specification of quality expectations and requirements. • Coordination of food systems is an active process (see Shaffer with Weber, Riley and Staatz, 1983). It will generally not be achieved in developing countries. without active intervention of the government, coordinating private and public institutions, and private firms with ample resources and a system perspective. Effective economic coordination is not assured by benign public sector laissez-faire, unsupported and 14 unfacilitated "private sector initiatives,1i and simplistic prescriptions of "getting the prices rights." This is not to prescribe heavy-handed and direct state involvement in agricultural prod;.Jction, marketing, processing, storage and distribution of food. Rather, ensuring effective coclrdination results from clearly defining public and private roles in tile food system, government provision of public good type facilitating services or functions, such CiS market information, communications, funding of production and marketing resesarch and extension, regulation which minimizes abuses and unfair play, and a policy environment which removes disincentives and fosters entrepi~neurship and the undertaking of risk. Through such developmental efforts food systems can move from low levels of productivity, plagued by small scale operations unable to innovate and achieve scale economies, high costs and minimal specializa tion, to systems where increasing specialization and exchange and achievement of scale economies lower food production and marketing costs and improve the wf~! are of producers and consumers alike. The subsector approach is most productive when it is used to describe salient elements of the subsector, diagnose key subsystem constraints to improved performance, and prescribe policies and projects to overcome these constraints. The subsector framework also emphasizes the search for opportunities to . improl/~ the system, as well as to take :::I.dvantage of unexploited and . under ~ploited the performance of opportunities to tap new markets, generat· significantly higher levels of 9\jtput, and improve food system efficiency. Improving food system performance is an incremental and iterative process. In evaluating performance the analyst should consider a set of diverse performance dimensions, which ma.y 'lot all be attainable in any given country context. Food system performance is evaluated broadly in terms of the effectiveness of coordinating mechanisms, technical and operational efficiency of the system, system progressiveness, and equity of returns to system participants given the distribution of risks, costs and responsibilities. A more detailed list of performance attributes is provided below. These mayor may not be of interest in a given situation. The list draws heavily on earlier' work by marketing researchers such as Sosnick (1964), Marion (1976), French (1977), Heimberger, Campbell and Dobson (198 I), and Shaffer (1980). 1. Market coordina tion effectiveness, or the matching of supply and demand at each level of the production/marketing system. 15 • 2. Technical efficiency, which translates into maximum output per unit of inputs (in a narrow economic engineering sense), and operational efficiency, which refers to minimal cost/price relations (in a broader economic sense). 3. Equity of returns to participants in ~ht subsector, in light of the distribution of investments made, risks faced, costs assumed and responsibilities undertaken. 4. "s ~nevis ergo P ~o the ability of the subsector to adapt improved technical, management and institutional innovations which enhance productivity. 5. Minimal degradation of resources (soils, 'forests, water resources) in agricultural production and processing, i.e., minimiza tion of external dbeconomies. 6. Broad participation in food system activities, particularly in countries where employment opportunities are limited, and where unemployment or underemployment are an important SOdHI and economic problem. 7. Wholesome and nutritious diets for a broad range of urb,an and rural consumers, particularly those most likely to be nutritionally vulnerable. It is important to note that these attributes may be conflicting, and that tradeoffs will often be involved in attempting to attain (or maximize/minimize) one set of attributes as opposed to another. Policy-makers' perceptions of the de3irability and feasibility of emphasizing one objective or group of objectives will be important in asse.ssing what constitutes good system pe.rformance. As an example, policy-makers might concentrate on achieving increased technical efficiency and system progressiveness, which might adversely affect equi ty anJ resource management. One of the challenges of food system research is to develop better defini tions of performance attri'butes and to relate these attributes to precise and unambiguous performance norms. By reference to these norms, performance in a particular system can be assessed. But caution needs to be exercised in defining performance norms. They may vary in different sociocultural, political and institutional settings, and in different agroecological environments, where resources and endowments differ. When expatriate researchers become heavily involved in food systems research, it is essential to involve local policy-makers and researchers in defining performance norms and measures. 16 rr While evalua ting the performance of a particular com modi ty subsystem or the food system against a comprehensive set of performance attributes and norms is desirable, the overriding emphasis of RR marketing studies should be on identifying barriers to improved performance and unexploited or underexploited economic opportunities. The research challenge is to discover and focus on critical points of intervention which could not only alleviate known constraints but also lead to increased opportunities and system productivity. 17 IV. KEY AREAS OF INVESTIGATION IN RAPID RECONNAISSANCE OF COMMODITY SUBSECTORS Table 2 lays out ten broad areas of investigation for rapid subsector appraisals, lists components of each area, and notes methods of inquiry for obtaining necessary information. RR teams will not attempt an exhaustive study of all of the areas outlined below. The purpose of RR is not to describe food systems in great depth or to diagnose in full detail all system problems. Critical system, wide Jinkages to specific subsector components must be identified, however. In addition, the payoff from examining in depth anyone research area may be marginal in some rapid reconnaissance studies. This may not always be the case, however. For exampJe, agricultural policies may nave sign fi~ant disincentive effects, and researchers may wisely elect to devote most of their attention to the effects of such policies. As a second example, particular stages of the food system, such as wholesaling or retaiJing, may be perceived as especially problematic. Payoff from examining special problem areas or constraints may be judged sufficiently high ~ to justify allocating most or all of the RR research resources to an in-depth analysis. Such a judgement will depend upon researchers' skill and experience, and preliminary identification of researchable problems and needs. In addition to listing key areas of investigation, Table 2 briefly notes reasons for considering those key areas. Exar,~in g public sector intervention and performance .in . the food system (key area No.8) is critically important in nearly all studies of food' systems. On the other hand, for.using research resources on detailed analysis of the operation of firms and organikations in the marketing system (No.6) will often be beyond the scope of a RR survey. Furthermore, while an exhaustive examination of the organization of a commodity subsector or marketing system will generally not be possible in RR, a broad understanding of food system organization is usually essential to evaluating performance. Initial judgement, based on preliminary knowledge of a subsector, will be required to assess the importance of gathering information in anyone . of the key areas. This will depend on what are perceived to be the most likely barriers to improved performance. In addition, researchers will often find that certain areas of investigation are well-researched, while others are under researched or not at all researched. This may be because researchers have jUdged ~ that the Jikely payoff from examining the latter was not worth investing research resources. It may also reflect biases and preconceptions of previous researchers. 18 In carrying out most RR surveys, researchers are generally recommended to focus on estimating orders of magnitude rather than focusing on obtaining precise estimates of marketing variables. The cost of obtaining precise estimates is generally high in doing survey research in developing countries. In carrying out rapid appraisals under time and resource constraints, th( r)~nefits of increasing precision wi1l1il<ely not be justified by the high opportunity costs of gaining sl";h precision. Researchers who undertake rapid recon aiS ~ 'ICe need to ask themselves continually whether it is worth the time and effort to gather particular types of delta. 19 .. , T.... ! KEY AREAS OF INYESTIGATIOH IN RAPID U ~ coMPOHl!lnS AREAS OF IHYESnGATIOH I. Commodity ChArACteristics 2. Consumption PAll«ns A) II) c) d) DiU«ent &fAdes. end uscs. Decree bulkincu. p«iWbiUty. Physial hAndline requlremenu. Decree/type of proccssln&. 0' Al ScAsorW And seculAr trends In domatlc and .xpon mArkets. II) DisAuresAteel consumption PAttecna b, soc~ m~ande~c r~ c' Future _k.t prospects. METHOD OF INQUIRy n Review commodity 1IW\U&Js. atuclles. 2) ObsecvAtion of Iwldlin& And procaalnl)) newelop commodit, calcndus showinc periods of production And trUlSfonnAtioA. J) RC'tlew consumption atuclla. food bAlAnc. slice... and dCl1lAlld projectloni. 2) Consuuet food bAlance sheeU If data a"ailAble. )) Intecvlew nutrhkln!CDIISUIIlpdoD r_dlera. Mleeteel commodit, Importera and exportera And .institutlonAl buJcrs. and .elected rural and urban consumers. ). Supply SitUAtion N 0 •• Price RelAtionships and SuSONUt, ,. Food Syarena PutlciPAftU And Orpniution A) Product1oA by yur and b, r.&Jon for recent ,ura, nolln& trends and "ulabllity. b) SlllCks for trAnsfontlAtIon and COIIIUIlIpdan II)' I a - l and r.aJon. c) Flows from mAjor ~Iy V.U 10 major markeU, lncJudlnC Imporu And esporu. a). SccuIu trends in rcaJ prlca at the f.armpt.. n IlEoUONS FOR INWSnCAUfG Al CommodIty c:Ilar cterlat - ~ can InfIucAcc operation of lUbs)'ltcm. wtIich functk1lll ue p..~ baw ·dIc, ... performed. and relAdv. COlt at wbidl perlanD'" II) Natur. of productioft procaa inJIuctlc.. tImin& and _&nIrucle of producer aala and ~ r c 4 fJowa. Al Demand driva for puI" e-modid.. throup) IUbs)'Itcma. II) Strenclh And _ U t , of dcaland affect produl:tIon and allln&e Incenti"es. u well u dluctiDn and -&nItude of a:arkcted 00... L1Ift&er N l trenda and Oi'portun.l~ affect Investment decbiDna of panIclpanu In 1Ubs)'ltaa. Review commodit, .tudles. 2) Intervl.w Iu&e wholcalcn. puuutal - C e n , crop proclx:lIon r.seArcher.. Importera. uponera. processor.. cooperAtiv. And trade auocIation ofliclA'" ) Ute _p to show flo... and appuCRt surplus and deficit :leu. • ) Describe ~ I A vviAtion In.atDdca:nd ilowa. a) Supply and demand ~e bUk: c1emenu of ccoaoaaic anal)'lia.,. . II) Production levcla and "arIaII1IIty affect pricca (dcpcAdinc Oft eIUtlcJda). rcrurna "ia prk. mcchan1am. and rlalc pc.'CeptJana of procioccra. c) Level of a10dca \bin& dllfcnnt perk:cls affect:l. .-sonaJ "ariatlon In pricca and COCIliIIOCIlty J) Catllcr aecondarr price data far COllIIIlOcIlt)' and cJoae a) Prkca ar.......... of Ial:entlva Iacin& food I)'Ilcm wbalaalc And r.taIllev.... II) ScuonaJ and qcllal trends In prJcea. c) OIAnCa o"er time in relAtI,. prk. r.latlonlh1pa. aubstitutCi/compJ_U lor ten ex _rc )'Car pcrlocL 2) Deflate prkca or apr... In constant price terma. )) An&Iru aecuIar. qcllcaJ and...-al price trends, and dlan&a III relative price r.latlonahlpa. .) Eatilute suppI, and demand relAtIonsIl1pa If dAta puaalt. a) MArlfctln& dIInncIa and commodity IllIIHctar ataC_ b) TJPea. /ldlllbcn And &cocrapllJc: d1atrlbutlon of firma at ke, IUbsectoe ataCca. e) lmpoxtant assembl,. redlstrlbudon And terminal lllarkets. U Review prcyloua ClOIIIIIIOdIty atudies. 2) Chedc U ....dn& enumerations or' aample frallla in perlllllent a&enda fit-&-. UccnaIn& olJlcca). )) Interview IcnowledpUle obaervcn of aubsectora and aelecrc4 partlcipanu. .) Dr... aubsectar map Wow chard ahowin& principal . StACel and markclln& chAM.... J) lJH _p to show Important lDArketplaces. r ."a1lablUty• If) Shltu In suppl, oyer tlme _ , Inclicate rcspanse to poUcies. rcdlnolo&iaJ dlan&e. InatltutionaI Cft"u--cnt and aJternadYc lasti1Utiona1 arrancemenu. partIcipanu. It) CIIIn&in& relatlv. price rclatlonsblps _,lPdIcate shifts III procb:tloft ..... lUIIcetInC .~vltmc>k especIaJJ, U cauplcd witll acante eDIt of production data. c) Prkln& atructure provides Insipt Inta rqionaJ and national ClOIIIpar&tlve .....tap. a) FoocIl)'ltea orpnlzadoa for atrueture) InfIuenccs c:onCb:t of pArtIcIpants. wIlictlln rum affects per-b) Hlp IeYcla of CDnCeIltradlln of 1InDa at jlIrdcular lta&a of food 1 ) ' 1 - . ., Iud ta lli&flcr pro6IIcdan/ markctln& CDIU.1Ilan under condldoas of Iow« c:ancentration. c) P..-.alenee of llI"lad smaU IIrala wtu IaII to apccIaJIz. . at _ or more levela of food 1 ) ' 1 - _ , lad ta sca.le 41lt : IOlIIla aIld Nih coati. .. v. ANALYSIS OF PRICES AND MARKETING COSTS AND MARGINS 1 Price analysis is critically importan'c in agricultural marketing and food system research. Prices are a mechanism by which signals are transmitted in market economies. These signals provide incentives to food system participants tf.7 supply inputs, produce commodities, and market, transform and consume these commodities. Prices also serve as guideposts in essentially administered economies, although they may not reflect scarcity values and social opportunity costs, and hence encourage nonoptimal allocation of resources. In many developing countries with administered pricing systems, parallel markets emerge in which prices rllore closely reflect scarcity values, although there is typically some risk premium for in the parallel market. This risk parti~ ng premium will vary in direct proportion with the frequency and severity of punishments or sanctions for non-compliance with administered prices and participation outside formal marketing channels. 5.1 Data CoUection and the Quality of Price Data In an idea! research situation reliable price data are available at the farm, wholesale and retail levels of the marketing system on a weekly or monthly basis over a ten or more year period for a broad range of commodities (agricultural and nonagriculturaI), as well as inputs, in rural and urban areas. This is often not the case, however. When available, agricultural price data are more likely collected at higher levels of the marketing system (terminal and urban retail markets) than in rural areas. Moreover, price data are frequently of unknown reliability because little real analytical use has been made of them. Careful attention needs to be paid to analyzing existing data and examining the way in which the data are collected. An important part of analyzing price data is to ask a series of basic questions that can be remembered by reference to the interrogative pronouns who, what, when, where, and how." IFor a detailed training guide to the use of fundamental price analysis tools and to the use of a microcomputer software program to facilitate this analysis, see Goetz and Weber, 1986. 22 5.1.5 How are the data collected? If enumerators are given money to buy commodity samples, they may report exaggerated prices and pocket the difference between the inflated and actual prices. If produce is purchased by data collectors, do they purchase it as anonymous, impersonal buyers and weigh it away from the marketplace? Or do they merely ask marketing agents how much they are asking for conventional lots, which may not be standardized, without ever weighing them? In this case, data will be collected for offered as opposed to actual transacted prices. How often are data collected and how many observations are taken per collection period? 5.1.6 How are the da ta analyzed? If da ta are not reported for periods corresponding to the collection interval or frequency, how are they aggregated? Is there any attempt to adjust, say, weekly average prices in calculating monthly averages? If inflation rates are extremely high, as in many South American countries, are adjustments made to average weekly prices obtained at different periods of the month before arriving at monthly averages? In calCUlating monthly or annual average~ how are missing weekly or monthly values handled? If adjustments in the data are made, are these adequately noted? Are price data in any way weighted to reflect volumes moving through different channels? While it is important to understand how price data are collected and prepared, researchers need not be paralyzed if data collection methods fall short of the ideal. Rather, analysts need to view prices as approximations which provide insights into the workings of markets, the relative scarcity of resources, and incentives facing food system participants. A good understanding of how data are collected, what they represent, and likely orders of magnitude of data error will help researchers appreciate data limitations. Using existing data and asl<ing these questions is an important step in helping to improve the quality of price information collected over the longer run. 5.2 Recommended Types of Price Analysis Important types of price analysis are listed in Table 3, and supplemented wi th brief notes about potential data collection and methodological pitfalls. This list draws heavily on work by Tomek and Robinson (1981), Timmer, Falcon and Pearson (1983), Timmer (1985) and many others. The interested reader is advised to see Newberry and 24 5.1.1 Who collects the data? Have they received any training in data collection? How are they compensated? Do they have other responsibilities, such as agricultural extension? If enumerators are poorly paid or collecting data part-time as a secondary activity, they may have little incentive to collect data accurately. 5.1.2 What types of data are collected? Are farm level, assembly market, wholesale market (consumption/redistribution), retail, FOB or elF prices measured? Are buying or selling prices (or both) observed? Are the types of data collected by government agencies well-defined? Have the definitions of the data collected changed, or have the actual market participants (levels in the system) from which data are collected varied from one period to the next? If so, then data from different time periods may not be comparable. 5.1.3 Where are data collected? Are data collected only in urban areas or in rural areas as well? In many African countries, for example, price data are only collected in large urban markets, and sometimes only with any frequency and accuracy in capital cities. If prices are colJected in rural areas, at what level of the system are they colJected (farmgate, rural assembly markets, consumption/redistribution markets)? If prices are collected in urban areas, are they gathered at municipal markets or in shops or supermarkets? If both, are the prices reported separately or averaged? 5.1.4 When are data collected? During which periods are data collected (throughout the year, after the harvest)? With what frequency are data collected (daily, weekly, monthly, quarterly)? Are they collected on days of greatest buying and selling activity (large market days as opposed to ordinary days)? Is there consistency across data collection intervals and in the timing of da ta collection a t particular markets? 23 .. • T~le), PRICE ANALYSIS AND POTENnAL PROBLEMS Type of Price Analysis I. Trends in Real Prices D&ta Requirements a) b) c) d) Fum&ate prices Wholesale prices Retail prlc'!ls Deflator D&ta Collection Pitt.alb n Definition of fumlate price. 2) Which side of wholesale transaction (buyer, seller). 3) Reported 'Is. tRnsacte~ prices. •• These points apply to other types of price anAlysis as welL 2. Relative Price Relationships 3. international/Domestic Price Comparison a) Prices for key substitutes and complements. I) Are data available for key a) Import pulty prices, Jncludln& internAtional transport costs. b) Export puhy prices, Jncludin& domestic transport costs. c) Exchange rates. I) A r ~ N \.n ,. Interspatlal Price Variation a) Average monthly, weekly, or dally prices at same level of marketing Jystem. a, Detailed price data for at least several locations, collected at same points in time, preferably for same level of market system. ) a) Cbanle in nature and cbaracterbtles e"f product over time. b) Deflator only aval1able for wban area (consumer price index). c) Representativeness of basket of soods, accuracy of welpts in construetlnl deflator. sub ~Uutes? 2) 3) .) ,) ,. Seasonal Price Variation lIe1hoc1oJock:al Problems domesllcally produced and internationally traded commodities close substitutes? Which international prIce? Quality differences. Are actual transport costs known (e.g.. If transshipment)? Official exchanle rate may diverge Ireatly from shadow exchanle ~ate. I) High rates of inflation and strong trends may distort results of analysis. n MultIlocatIonal price data are not a) Assumes domestically produced commodity traded or potentially tRdc'~Ie. b) Poor comparablUty of domestic and international product an confuse analyals. d May be no direct transport between intern&tIonaI npocter and COlllltry. a) Price data are typically orJy available for wban areas. b) Price seasonaBty faclnl urban ~ers _y differ from Yarlatlon facinl rural households. a) Hlp intermarlcet correlations may be often available in tlme-serles. 2) Data may only be available for major towns and not for rural markets. evidence of effective competition or collusion/oligopoly. Need more Information for determining which Is the cue. b) Correlation may be spurious and no ~lcIen of c:ausaDty. 6. Marketin& Margins 7. Commodity/input Price Ratio I. Processed Product/Raw Material Price Rallo a) Prices at different levels of the food system for same commodity. b) D&ta on manetlnl costs If wish to analyze net marlins. I) Prices at different levels of a) MarlcetIng costs vary by scale of system must be collected durln& same period. 2) Cost data difficult to colJecq may be misrepresented. ) May fall to ellUlllerate key cost components. enterprJsc. resultin& in different mar&lns for different firm sizes. b) SIze of mar&1A and percent of retwn to producer vary by commodity, reDectIn& decree of value added and marketIn& costs. a) Consistent series of commodity prices and Input prices or cost Index series. U Need to Identify relevant Input a) Cash Inputs (ferts, pests) may only a) Prices of processed product. b) Consistent prices of raw materiaL I) Nature and quality of procesRd (fertilizer. insecticide, pes tlclcle type). product may chanle over time. 2) Quality differences In raw material. ,.. be used bylarle farmers. ~ fmportane of~tmayY r considerably over 1en&th of tIme-serles. a) Comparln& ratios with other countries with different factor proportions, costs of production, and prices may be mbleadins· Stiglitz, 1981, for an advanced treatment of price analysis and commodity stabilization programs. In some developing countries, price data are only available for capital cities and large towns (and not for the region of interest or rural areas). When data are available, they are frequently of poor quality, available over a short period, or incomplete (missing observations). In some assessments of marketing systems, it may be necessary to generate price data at different levels of the commodity subsector, although these very short term snapshots of price patterns cannot substitute for longer term information on prices. In most developing countries agricultural production, trade and price data are most complete for exported cash crops, such as cotton, cocoa and coffee, and for staple grains. Data on tuber crops, legumes, oilseeds, and domestically consumed fruits and vegetables are often less complete and reliable, if available at all, or they are highly aggrega ted (averaged across quarters or years). Even when reliable price data are available for the commodity in question, it is difficult to do all th~ a.nalyses listed above in most rapid reconnaissance surveys. Examination of spatial prke variation (No.5) is most useful when detailed, frequent observa tions of agricultural prices are available for a wide range of markets, which is rarely the case. Analyzing the price ratio of a processed commodity to its raw commodity material (No.8) is an imperfect measure of technical efficiency, which may , be better quantified through raw product/processed product transformation ratios. Nevertheless, attempting price analyses No.1 - 4 and Nos. 6 and 7 is useful for rapid reconnaissance purposes. Potential difficulties and pitfalls in doing each type of price analysis are discussed below. Examining trends in real farm level and urban wholesale and retail prices for related commodities (No.l) can be very useful in understanding how incentives to produce and consume particular crops change over time. The validity of the analysis will depend critically, however, on the quality and appropriateness of the deflator used in generating real prices. The only deflator available in many developing countries is typically an urban consumer price index, which is constructed using a basket of goods and services (e.g., food, electrici ty, transport) consumed by urban residents. This may not be the most appropriate for the analytical task at hand. In some cases, such as where producer supply response is being estimated, an index of prices paid by producers may be more appropriate. An urban consumer price index is better suited to estimating demand functions which use urban wholesale or retail price data. 26 Since urban consumer price indices are generaUy all that is available, it is useful to comment on potential problems in using these indices as deflators. Urban and rural consumption patterns differ markedly in many developing countries, so the appropriateness of the basket of goods used in generating the deflator can be questioned. Certain goods consumed in urban areas are not widely available (or available at all) in rural areas. Moreover, these goods may not be important in rural consumption patterns. Including these goods in a deflator to calculate real farmgate prices may introduce bias. In addition, the data from which consumer price indices are generated are urban household and expenditure surveys. Relative prices, per capita incomes, tastes and preferences, and hence consumption patterns, as well as the composition of the urban population are likely to change over time. As a result, the basket of goods used in generating the deflator and the weights assigned to each type of good may not even accurately reflect urban consumption patterns ten or perhaps five years after the budget and expenditure survey has been completed. The differential effect of inflation on the relative prices of goods and services in the basket can also affect the reliability and accuracy of the deflator. Certain goods, particularly luxury goods, may drop out of the basket or carry lower weights several years after the survey data were generated. Other goods, particularly necessities, may assume greater importance 'as a result of inflation or declining real incomes. . A second difficulty in examining trends in real commodity prices over long periods is that the commodities may have changed fundamentally in nature and quality. The commodity may have greater value added in recent years, through improved sorting and grading or better packaging and hygiene, than in earlier years. Hence data may not be readily comparable over the entire time series. Exercise care in ascertaining whether the definitions of the commodity form and grade are consistent over time. Examining changes in relative prices facing producers (No.2), as well as relative costs of production, is extremely useful in analyzing changes in farmers' crop mix and changing regional/national patterns of resource aUocation. Researchers should note if prices for commodities are official prices or actual prices, particularly in cases where parallel markets are important. Analyzing changes in the terms of trad~ facing agricultural producers presupposes availability of time-series price data for rural areas for a broad range of agricultural inputs and outputs, as well as consumer goods. Data are rarely available in long enough time-series or reliable in rural areas of developing countries, particularly in Africa. Using urban price data in analyzing rural-urban terms 27 of trade is misleading, because urban prices of consumer goods (particularly manufactured consumer goods) are generally far lower than prices paid by producers in rural areas. Similarly, food prices in urban areas overstate returns to producers, because they are composed of farm level prices plus marketing margins. Comparing border prices with domestic prices (No.3) is not appropriate when the commodity is not traded (e.g., tuber crops in Africa). This is not to say that production and consumption of non traded staples are not affected by shifts in international prices of traded commodities, such as grains, which may be sll;Jstitutes. Monthly price data are not always available at the farm level or for rural markets, so analyses of price seasonality will often rely on urban retail or wholesale data (No.4). These urban price data will show seasonality of prices facing urban consumers but not necessarily seasonality of prices received by rural producers or faced by rural consumers. This will depend upon the deb "ee and volume of storage of com modi ties after purchase from rural producers, the magnitude and variation over time of transport costs between rural ~nd urban areas, the degree of integration of rural and urban markets at different times of the year, and the possible existence of bottlenecks in the provision of marketing services. Seasonal price trends may also vary between rural and urban areas due to seasonal demand shifts, caused by cultural events or availability of income at parti,cular times of the year (e.g., post-harvest period in rural areas). • Analyzing commodity/input price ratios (No.7) will not be appropriate when little or no fertilizer or other cash inputs are used in crop (or livestock) production.· This is certainly the case for grains such as sorghum and millet, tuber crops and some legumes and oilseeds produced in developing countries, particularly in Africa. Even when a particular group of farmers, such as large commercial farmers, uses cash inputs in agricultural production, the input/output price ratio may not be useful in analyzing the responsiveness of small farmers who do not use cash inputs. The price ratio will, of course, be a useful indicator of incentives to use the inputs for larger, commercially oriented producers. Examination of available price data can provide analysts with important insights in to incentives producers and consumers face in making decisions about agricultural production and consumption patterns. Without a good understanding of this incentive structure, a limited study of one or more related commodities risks overlooking 28 interrelationships among commodities and patterns of comparative advantage. Subsector studies are partial, which can be a serious drawback when the commodity (or commodi ties) under examination is not analyzed in a broader context of substitutes and complements in production and consumption• .5.3 Analysis of Price Data When price data for a broad range of agricultural commodities, inputs and consumer goods are available, and do not have to be obtained from many government agencies, one researcher in the RR team may profi tably spend several weeks to a couple months analyzing these data. Since improved data analysis software packages, such as Micro TSP (Time Series Processor), the price analysis modules of Microstat and Ab-stat, and the relevant subprograms of MSTAT, have become available for microcomputers, a surprising amount of price analysis can be accomplished in a short period of time. Analysis can be expedited if the data have been collected and entered into a com puter ized data base. If the data are ready for analysis, the RR researcher can rapidly analyze seasonal and secular tr~nds in prices of key commodities and their . . substitutes, compare changes in their relative prices ('·ier time and in production and consumption patterns, calculate marketing margins, and perhaps estimate supply and demand functions if data on prices and quantities of inputs, outputs, imports and exports and per capi ta incomes are available. In addi tion, constructing food balance sheets can give the analyst a simple overview of supply and .demand condi tions and food consumption patterns. As implied in section 5.1, it is advisable to check the accuracy and reliability of secondary data sources. If time permits, researchers might also collect supplementary price data at levels of the marketing system, such as the farm level and rural markets, for which secondary data are not available. In longer RR exercises, the analyst can distill, summarize and further examine useful quantitative information from earlier farm level, marketing and commodity studies, and from rural or urban consumption surveys. Typically, however, a researcher requires more time than available during rapid reconnaissance to process and analyze large data sets. 29 '.4 Analysis of Marketing Costs and Margins For the purposes of RR, practitioners are generally advised to examine gross marketing margins rather than trying to collect detailed information on each cost component so as to estimate net margins. High gross marketing margIns in developing countries often reflect real marketing costs, due to inadequacies in transporr.ation systems and/or isolation/dispersion of producers in rural areas. In such cases investment in marketing infrastructure is likely to lower marketing costs. High gross margins may also indica te geographic areas or levels of the marketing system where returns to particular marketing activities are high, and where improved access to markets and information or greater competitiveness might lower returns and hence marketing costs. If researchers are able to collect data on the prices at different levels of the marketing system and costs of different marketing system participants, net marketing margins can be estimated. Collecting such cost data is very time-consuming v however, and marketing agents may not report costs accurately. Cost and return data are sensitive information, and many informflnts wiJI not divulge complete or accurate information during an initial interview. They usually fear that these data are being collected for tax purposes and will be used to increase their tax burden. This attitude has evolved over years of typically antagonistic rela tions between developing country governments, particularly regulatory agencies, and priva te traders. In-depth case studies, which use multiple interviews, are usually a better tool for obtaining information about marketing agents' costs than single visit interviews carried out wi th a Jimi ted number of informants during rapid appraisal. It is possible, however, to estimate approximate net margins of participants from data gathered during RR field work. One way to do this is through informal surveys, designed to help construct enterprise budgets using data obtained from a purposive sample of relatively homogeneous firms. In most cases researchers will find that it is easier to collect some but not all cost data from individual informants during RR interviews. Certain informants are asked about particular cost components, while others are asked about other costs. What emerges is a composi te or synthetic enterprise budget. Tht: budgets are synthetic in that they incorporate data pieced together from different sources but representative of the marketing costs of a relatively homogeneous sample of informants. 30 VI. KEY INDICATORS AND PROXY VARIABLES IN RAPID RECONNAISSANCE Rapid reconnaissance techniques developed by other researchers, such as Chambers (1980, 1981) and Honadle (1982), strive to capture essential features of rural and agricultural systems through use of key indicators and, where appropriate, proxies for a series of variables or complicated phenomena. These researchers emphasize that use of key indicators can provide valuable insights and substitutr. in part for massive data collection. These indica tors include soil color as a proxy for soil type and quality, birth weight as a proxy for nutritionally adequate diets, and vehicle'S (and their contents) at marketplaces, ferry crossings or along important arteries, as indicators for marketed surplus in a rural area and direction and magnitude of commodity flows. This paper makes only modest contributions to the development of key indicators and proxies for agricultural marketing research. Several of these are discussed below. Considerably more work needs to be done in this area. 6.1 Degree of Commercialization of an Area or Rer;ion One commonly used proxy of the degree of commercialization of an area or region in a developing country, and the purchasing power of people living in that area, is inventories of particular goods. Inventories of consum,er goods, such as soap, kerosene and mirrors, in village shops are one example (Chambers, 1980 and 1981, ~nd Honadle, 1982). The proportion of dwellings in rural areas having tin roofs or doors, glass windows or wood-burning stoves is another proxy for village wealth and degree of commercialization. A similar indicator is an inventory of the proportion of households in a village possessing consumer durables such as bicycles or radios. CalcUlating the ratio of the number of functioning units to the number of purchased units is also a useful indicator of the availability of spare parts and hence a proxy for the effectiveness of input and consumer goods distribution networks, as well as maintenance services (see Pruitt, 1984- and Food Studies Group, 1985). 6.2 Vehicle Ownership and Availability Another set of indicators which provides insight into the organization and performance of the agricultural marketing system is patterns of truck and pickup ownership and utilization. Transportation of agricultural commodities is typically a high 31 cost component of marketing in developing countries, especially in Africa, making up a large part of the gross marketing margin. Inefficient utilization of transport and any. concentration in ownership or distinctive patterns of ownership and use are likely to affect transport costs. If truck ownership is highly concentrated in the hands of relatively few entrepreneurs based in urban markets, as in the case of Fulbe traders in Northern Cameroon, there may be opportunities for collusion in provision of transport services, which may increase the costs of food crop marketing. In many countries entrepreneurs·who provide transport serviCes are, for the most part, a different group from those engaging in agricultural marketing. Another useful indicator in analyzing transport effectiveness and the performance of input supply networks is the ratio of functioning vehicles to vehicles owned and operated in a region. We shall refer to this as the effective vehicle availability rate, which is defined as the ra tio of opera ting vehicles to the total number of vehicles in a regional fleet, including non-opera ting vehicles tha t are potentially repairable and operative. In rural areas of some African countries effective vehicle availability rates are reported to be as low as 25-33% This cripples the ability of rural-based traders to compete against trader/transporters based in urban areas, who have far better accp.ss to . spare parts and maintenance services. It may also raise marketing costs, as competition among transporters may be reduced, increasing the opportunity for monopoly pricing. 6.3 The Relationship Between Administered and Parallel Prices Another example of a key indicator in agricultural marketing research is the disparity hetween official or administered prices and parallel market prices in developing countries characterized by direct state intervention in agricultural pricing. Parallel markets develop as alternative channels to official markets where legal monopolies are granted to parasta tal agencies. A large disparity can indicate one of two possibili ties. If parallel prices greatly exceed official prices for an agricultural commodity, which is the more common case, it is likely that a) most of the com ~.,dity moves outside of official channels, and that b) the effects of any effort to liberalize prices will be minimal. Since most of the commodity moves in the parallel market already, parallel market prices likely approximate quite closely market-clearing prices that would prevail without government intervention. Participants in the food system have adjusted in part to the distortion of administered prices. Yet uncertainty and transactions costs may stiJl he high, affecting the dynamics of the system, participants' incentives, prices (which will 32 reflect a risk premium), and participation in the food trade. The extent and severity of these impacts is partly dependent upon the enforcement capacity of the government and its ability to administer sanctions. Nevertheless, less short-run adjustment may accompany liberalization when the formal market captures a small share of the commodity trade (see Berg, 198.5). In the second case, where official prices greatly exceed market-clearing price levels, liberalization wil1likely have a profound effect on the organization and operation of the marketing system (see Timmer, 1985). In order to support an official price above market-dearing levels, the government must stand ready to procure the commodity at the high support price. This can distort incentives of producers and marketing agents, who will shift resources from al terna tive enterprises to production and marketing of the price supported commodity, for which returns have become artificially ~igh. Once the government is no longer capable of maintaining the support price, and the price drops toward market-clearing levels, producers and marketing agents no longer have the same incentives to grow and market the once-supported commodity. 1 Such a scenario actually unfolded in Liberia in t.he early i980s. The parastatol agency, the Liberian Produce Marketing Corporation (LPMC), was compelled to buy paddy rice from all comers at a price above the world price beginning in late 1982. LPMC bought large volumes of paddy at the high support rice until its storage facilities were fi!led and it no longer had the financial resources to maintain the support price. Local producers sold a larger than ." normal marketed surplus, and paddy streamed into Liberia from neighboring countries, until LPMC went heavily into debt by August 1984- and ceased buying paddy. As more RR techniques are applied in food systems research, it will be important to develop and use other key indica tors and proxies. The discussion in this chapter is only a beginning. IThis argument is essentially static and dismisses a possible infant industry policy objective. That is, a developing country government might promote production of a particular crop through offering a guaranteed price above international market levels. The government might do this if it believed that production costs would decline or international prices would rise over time, thereby resulting in a shift in that country's medium or long term comparative advantage. 33 VII. INSTITUTIONAL AND NON-MARKE.T FACTORS IN FOOD SYSTE.MS During the past couple decades the literature on institutional, public choice, and non-market econom ics has burgeoned. This literature has ar isen in recognition of the limitations of neoclassical economic analysis in addressing common problems of externalities, where private costs and benefits do not correspond to social costs and benefits, of h1gh transactions costs, of the critica.l role of property rights in determining who has access to particular sets of resources or opportuni ties, and whose preferences count in economic decision-making, and of pervasiveness of uncertainty in economies ~he of all types. The literature is broad and diffuse, and there will be no attempt to summarize it here. Rather, we will identify several institutional issues which merit a tten tion during rapid reconnaissance surveys. 7.1 Transactions Costs and Market Improvement A fundamental problem of agricultural market development is transactions costs. In some cases 'transacti0n costs may be so high that no market exists. Assembly of agricultural commodities from small producers in developing countries is plagued by high transactions costs. Small farmers often produce Ii ttle for the market. Small lots of marketable commodities are widely distributed and assembled at very high cost. Given di fferences in agroecological condi tions across space, even from year to year (e~g., in levels of rainfall, pest and disease incicl"nce and prevalence), and poor information about com modi ty supplies and surpluses, traders may incur signficant search costs in identifying areas from which marketable surpluses can be obtained a t costs which may not exceed likely returns. High transactions costs in developing countries have encouraged entrepreneurs to create a wide range of both formal and non-formal institutional arrangements which significantly lower these costs as well as uncertainties in relying on spot market mechanisms. Forward contracts negotia ted between producers and either assemblers or processors are one means of lowering transactions costs and uncertainty associated with fluctuating supplies, prices and producer incomes. These contracts are most common in the production and marketing of high value commodities, such as fruits and vegetables, dairy products, or export commodities, such as coffee, cocoa, palm products, and copra. Vertical integratIon of processing firms with suppliers is another institutional arrangement for reducing risk and uncertainty that gives the firm that integrates 34 .. verticaJly greater control over the timing and flow of inputs (e.g., raw unprocessed commodhics) and/or the distribution of outputs. I In conducting rapid assessments, investigators need to be alert to these types of institutional arrangements. If most of a commodity flows through such an institutional channel, researchers need to evaluate the effect on returns to producers, processors and handlers, conditions of entry to production, marketing and processing, consumer satisfaction with commodities flowing through these channels, ~nd responsiveness of participants working under non spot market coordination arrangements to changing demand, technology and relative prices. 7.2 Property Rights Neoclassical economic analysis generally assumes tha t the distribution of resources, rights and privileges is given and rarely questions the effect of this distributi?n on access to income streams, participa tion in economic decision-making, and responsiveness to new economic opportunities. Yet resource distribution in many developing countries is highly unequal. Urban-bas~d people typically have better access to policy-makers and administrators of government programs than do rural people. This enables urban interests to influence policy formulation and facilitates their access to credit, licenses, inputs, information and other resources. Urban based people are able to extract favors which defend and enhance their interests. In some countries rUf'=1llanded groups control access to land, particularly higher quality, fertile, well-watered and level land. They own large holdings, where economies of size mal<e economically viable and attractive adoption of improved technologies such as mechanization and irrigation. Rural land holders, especially the larger ones, are sometimes able to capture most of the gains from improved technology, inflation (which increases the value of land), and other economic factors affecting rural income streams. Last, particular "ethnic groups may dominate trade in certain commodities, particularly at higher levels of marketing systems, restric t entry, and perhaps capture oligopoly rents. In effective food systems research, it is important to understand the evolution of the existing social structure and ISee Minot, Nicholas, Contract Farming and its Impact on Small Farmers in Less Developed Countries, MSU International Development Working Paper No. 31, Department of Agricultural Economics, Michigan State University, East Lansing, Michigan, 1986. 35 the power ard prerogatives of urban, rural or ethnic groups in developing countries. Members of interest groups which control particular types of commerce or trade in certain com modi ties may compete effectively among themselves, but this is not always the case. Control of resources, such as land, non-recoverable or depletable resources (e.g., minerals or petroleum), irrigation rights, water resources or rights to import/export, obtain credit, technology, etc., has a critical effect on patterns of resource aIJocation and income streams. The existing distribution of property and rights determines whose preferences count. Groups that benefit from this distribution wiH most certainly oppose proposed changes in rights and privileges that wiJJ weaken their positions and preroga tives. Certain types of research may be opposed by influential interest groups, whether governmental, urban, rural or commodity based, who stand to lose from any change in the distribution of resources and rights. Researchers who conduct rapid reconnaissance need to be sensitive to this. Proposing a thorough and critical review of the accounts and activities of a parastatal organization wiJJ, for example, not be wellreceived in a country "'here that parastatal dominates trade in a particular commodity and represents a firmly entrenched and special bure"aucratic interest. 7.3 Non-Economic Factors Political economists, sociologists and anthropologists have long been aware of the importance of non-economic factors which affect the performance of food systems in developing countries. In this section we will examine the role of three types of noneconomic factors: ideology, religion and cultural endowments. 7.3.1 Ideology By ideology we mean shared sets of beliefs, attitudes and values in a society which affect the ways in which people interact and engage in political and economic activities. Rigid adherence to an ideology can preclude certain types of economic activity (e.g., gambling, smuggling), lead to prescription of others (collective organization of economic enterprises, administered prices, free markets), and undermine or encourage other types. 36 .. One example of how ideology has affected the organization and operation of agricultural markets is the People's Republic of China. The communist ideology of sharing in sacrifice and benefits of work were translated into a particular set of policies before the economic reforms of the late 1970s and early 1980s. Grain production was communally organized, the state controlled disposal of grain surpluses, and prices were administratively fixed. Everyone was guaranteed access to a minimally acceptable quantity of grain. This set of policies affected the incentives of producers organized into collective units. Individuals met the requirements of cooperative units in providing labor' to grain production, but they also devoted a lot of time and attention to private production of vegetables and small livestock to meet cash needs. Small vegetable plots or ownership of small stock were means of supplementing meager incomes from communal activi ties. A second example of how ideology can affect the organization and operation of an economic system is in the historical evolution of capitalist economies. Capitalist . economies are characterized by specialization, exchange, innovation and search for , means of reducing costs. This often leads to impressive increases in product.ivity, but there is a tendency toward increasing scale and concentration over time in inany industries, including the food system (see Conner et al., 198.5). As concentration increases, opportunities for oligopoly pricing may also increase. Product differentiation may also advance to the point of bewildering proliferation. This is usually accompanied by large outlays for advertising, which do not add value to products but which facili ta te capturing of larger market shares by individual companies. Finally, private firms are not always required to internalize the costs of external economies in their production and marketing decisions, thereby imposing heavy costs on society at large or on groups lacking voice in particular. 7.3.2 Religion Religion is related to and influences ideology, but it will be treated separately here. There is little doubt that religion often has a profound effect on the motivation and incentives of participants in economic systems. The economic dynamism of protestant countries, such as Germany, Switzerland, u~land and Holland, can be in part attributed to religious attitudes and predispositions toward hard work, savings and investment, characteristics which are not always present in developing countries. In protestant societies commercial success was encouraged and viewed as evidence of 37 I divine blessing. On the other hand, commercial failure and unemployment may be viewed as just desserts for the unfortunate, and redistributive mechanisms may be poorly developed in such societies. In contrast, Moslem societies are noted for their emphasis on social responsibili ty of wealthy individuals toward their less fortunate brethren. Moslem societies are often more committed to redistributic,i' of wealth than other western or capitalist societies. There are strong pressures for the wealthy to provide for poor relativ ~ and the indigent. In African countries such as Niger and Senegal, muslim clerics and brotherhoods play an important role in agricultural marketing and redis tribution of surplus grain. This does not discourage the accumulation of fortunes, but it may shift relatively more resources to meeting current consumption requirements than to savings and investment. Moreover, it may contribute to fatalistic attitudes which undermine i,"!dividual initiative in some Moslem countries. "In sha'allah" or "If God is willing" is a phrase which is often uttered. On the other hand, powerful Muslin clerics have the authority to require followers to pr<>vide labor for collective agricultural production. This phenomenon, which is practiced by the Mourides in ~lagenS is a highly disciplined contractual arrangement for increasing agricultural production and assuring' subsistence of the labor force. Another characteristic of Moslem countries which affects economic activity is the Koranic stricture against charging interest on loans. In principle, interest is never demanded or paid. In practice, profits from use of a loan are sometimes divided between lenders and borrowers. In addition, lenders may not require interest payments, but they may expect other benefi ts, such as preferential exchange arrangements, tied or close business relations with the borrowers, and special services or favors in providing inputs or outputs. Moslem businessmen are not incognizant of the time value of money. Rather, forms of payment for use of money are often indirect and varied. Fatalistic attitudes often also pervade animist societies, where rural people attribute economic success or failure to magic, the occult, the blessing or displeasure of ancestors, or the envy and ill-will of neighbors. These attitudes may discourage innovation and receptiveness of producers to new technology and farming practices, as weJl as to agricultural marketing opportunities. On the other hand, animist societies have often devised redistributive mechanisms for coping with agricultural production shortfalls and periods of exceptional hardship (see Campbell, 1984 and Campbell and Trechter, 1982). 38 7.3.3 Cultural Endowments Ruttan and Hayami refer to particular cultural traits, group practices 'and communal forms as cultural endowments (Ruttan and Hayami, 1984). They cite communal efforts by the Japanese and Taiwanese to maintain irrigation channels as examples of a cultural endowment which has contributed significantly to high rates of agricUltural growth in those countries. Another example of a cultural endowment or characteristic is coping mechanisms developed in resource poor, drought prone African societies. ;"n years of drought, pestilence or other disaster, many African societies redistribute staple foods so that the subsistence needs of as many rural people as possible are met. This redistribution is effected outside of marke~ing channels. There are strong social pressures to £;ive food, livestock or money to less wealthy relatives and neighbors in times of need. In s'" me cases reciprocity is later expected in the form of agricultu;'allabor, political allegiance or help during crisis periods. In other cases there is ~l t or no expectation of reciprocity. Although there may be no expectation ot'economic gain or eventual benefit, these typ~s of coping mechanisms can be viewed as a form of social insurance. In addition, in many African societies collective labor is mobilized to maintain rural roads and to meet peak season labor requirements in food production and harvesting, which have important implications for performance of agricultural co.mmodity systems. A last example of a cultural endowment which afbcts food systems is the commercial acumen of particular ethnic groups who sometimes dominate marketing and international trade of some agricultural commodities in some developing countries. Examples from Africa include the Hausa in the long distance cattle trade in West Africa and the Somali in the East African livestock trade. The importance of the Bamilike of Cameroon and the Mandingos of Liberia, Guinea and Sierra Leone in domestic agricultural marketing, as well as the predominance of Lebanese or Indian international traders in other countries, are other African examples. An Asian example is the ethnic Chinese who domina te trade in certain commodities in Indonesia, Thailand, Malaysia and other Southeast Asian countries. Numerous examples from Latin America are also available, including regional groups within given countries like Colombia and Brazil. 39 These ethnic groups are often alleged to wield considerable market power, colluding to force agricultural prices low and to propel prices of consumer goods high to rural producers, and creating artificial shortages through speculation and hoarding. While most empirical studies do not support such allegations, it is sometimes true that trade in particular commodities is dominated by certain ethnic groups, and that entry may be difficult for outsiders. This may have important implications for the organization and performance of marketing systems. Barton (1977), Landa (1981), Cohen (1965), Shaffer et al. (1983), and others have noted that trading within an ethnic group significantly lowers transactions costs and generally decreases risks. Informal social obli:.;ations and trust cC\n substitute for formal regulations such as legislated grades and standards in many coun i.ries. Doing business with members of one's kinship or ethnic group can reduc~ risk a.Ild uncer ~:ainty, decrease the importance of P(~, :e' stabilize returns, lower marketing costs, and incentives in transactions. 40 VIII. PREPARING FOR RAPID RECONNAISSANCE FIELD WORK 8.1 Preparatory Visit and Planning Before planning for the RR field work begins, the RR team leader should discuss the objectives of the RR with study sponsors and policy-makers, and negotiate to modify the terms of reference if necessary. Several initial, exploratory interviews with selected participants in the commodity subsector will also prove useful as a crosscheck on policymakers' perceptions and objectives. The team leader can also identify, and if necessary contract, researchers and institutions to collect, compile and analyze secondary data. If these data are entered into a microcomputer data base, further analysis is facili ta ted. Finally, the team leader can begin to make administrative arrangements for the RR field work. Generally, planning for the RR and preparation of secondary data will have a gradual buildup, unless an expatriate is asked to leap the team. In this case she/he will usually make a preliminary visit to the country where the survey will be done. 8.2 Selecting the Rapid Reconnaissance Team RR findings are often most illuminating when the RR is conducted by multidisciplinary teams. Selection of analysts with different disciplinary backgrounds will be guided by preliminary identification of marketing problems and the expertise most effective in addressing particular problems. In some cases the RR team will have to depend on the perceptions and demands of organizations and agencies which fund the rapid reconnaissance. In carrying out the rapid reconnaissance surveys, the RR team may want to divide into two or more groups of two to three researchers each, preferably having different disciplinary skills. RR surveys will generally be stronger if all of the team participants have conducted rapid appraisals in the past, as well as year or longer studies of food systems in developing countries. By having done longer studies~ members of the RR team will be more sensitive toseasonal factors and the potential dangers of making inferences about production/marketing systems from information gathered from a limited number of participants, observations and areas during one short and not necessarily representative time period. 41 When the RR team is composed of researchers who have not all done longer term studies or rapid appraisal, it is advisal)le to mix experienced with ineXj)erienced analysts. If expatriates are involved in RR work, it is critically important for host country researchers to participate in the exercise as well. Their knowledge of commodity production and marketing systems can of ~et potential bias from observing the system at one point in time. Their understanding of the sociocultural and political context is also invaluable in approaching key informants and structuring informal interviews. Local researchers will generally be more skilled in approaching participants in systems and handling more sensi tive issues than outside consultants. gnitekr~m Knowledge of local customs, institutional and polltical organization and behavior, and agricUltural and economic development in historical perspective will also improve interpretation and analysis of data generated by the RR. While outsiders may come in with fresh persp.ectives, they lack detailed knowledge of local conditions. They may also come with preconceptions or inappropriate comparisons. In a few cases, however, expatriate researchers may be able to obtain· inforrna tion tha t local researchers could no t obtain. In these instances informants perceive foreign researchers as outsiders who do not represent the government and will not use sensitive information against them, as a local researcher may be perceived as capable of doing. By developing the capability to undertake rapid assessments, local researchers and research organizations can use the RR process to continue monitoring and evalua ting the performance of commodity subsectors. They may also examine the organization, operation and performance of other subsectors. Familiarity with RR methods, and a critical awareness of their strengths and limitations, will also help to instill a healthy skepticism of the findings of short-term consultants and other RR teams. 8.3 Planning the Rapid Reconnaissance Field Work During the first week or so of a rapid reconnaissance study, the team will need tCl define RR objectives, review available studies and secondary data, identify data gaps and needs, develop information gathering strategies, and define clearly the objectives of the RR and the roles of each of the team members in the survey. It is often useful to do a preliminary out1in~ of the RR report (see Chapter X). This helps the team to focus on information needs and priority topics. 42 During this period it 1s necessary to define clearly the objectives of the RR and the roles of each of the team members in the survey. In cases where some members of the team have not done RR, it is useful to spend at least one day introducing them to the RR methodology. Experienced team members can give examples of problems they encountered in RR in earlier studies or of particular interviewing strategies or time saving techniques for data gathering that they have used successfully. The review of the RR methodology is not only useful for the uninitiated but also for those who have done RR before, so that every team member has a clear idea of how RR will be used in that particular context. After the discussion of the RR methodology and review of the literature and secondary data, team members should be encouraged to jot down their hypotheses about the food system components under examination. Developing hypotheses about the organization and operation of the system will help to focus data gathering efforts. It will also hopefully make the researchers more conscious of possible sources of bias in their forthcoming informa tion gathering. As a result, they must plan to offset these biases and preconceptions consciously in their research. Before actually beginning the RR field work, it is useful to develop a research itinerary and activity lists. These lists should note tentative research plans for each day, including towns, agricultural processing plants, rural markets, and producing areas to be visited and government officials, types of marketing agents, and producers to be interviewed. If researchers plan to visit rural periodic markets, it is advisable to find out beforehand which days or how often particular markets are held. It is also useful to plan around government holidays, communal work days, religious festivals, days of worship and rest, or particular times of the day when potential informants are praying, working or otherwise engaged. An activity list is a useful tool in focussing data gathering on critical and necessary activities. It enforces discipline in planning research logistics. When travel and protocol requirements are taken into account, there are often strict limits on the number of places that can be visited.and the number of informants interviewed. However, the researchers do not need to be slaves to an activity list, and it is wise not to overload it. Unanticipated opportunities to observe marketing processes or interview system participants may arise which can justify deviation from a fixed itinerary. It is advisable to allow time for improvisation, possible repeat visits or in terviews, and other contingencies. 43 3.4 Review of the Literature and Analysis of Secondary Data Although it may seem obvious to many that a logical place to begin research is by reviewing earlier work, there are many cases where this is not done, or not done very systematically. There is often a wealth of useful information and data in scholarly studies (including student theses), annual reports of government agencies or barastatals, I ministry of agriculture data banks, records of cooperatives and private firmd, project documents, trade and professional journals, and consulting reports. Althougd, many researchers like to think that earlier studies are inadequa te or unfocussed 10l.· their particular purposes, this work often contains useful information and insight:J,t I \ While planning the RR survey, the team should review the literature: nd available secondary data. Not every team member has to review every document or d,ta set, thus division of labor along disciplinary or subdisciplinary lines is usually appropriate. Each team member should summarize the principal findings from the literature and preliminary analysis of available data for other team members. During the first week of the RR, team members can make oral presentations, or draft a series of annotations or memoranda. In preparing important data for others, team members may wish to tabulate secondary data so that it is readily accessible and usable. Critically important papers that every team member should read before beginning the RR surveys need to be iden tified. The objective of this review is not to burden team members with busy work and supplementary writing assignments. Rather, it is intended to prepare all team members for the RR field work in as rapid and systema tic a way as possible. Types of secondary data that are usually readily accessible include: l. Wholesale and retail prices for agricultural com modi ties, usually collected in capital cities and other major urban areas. Farrngate prices are often not collected or may only be official producer prices. (See Chapter V). 2. Price indices, usually consumer price indices, constructed for a basl<et of com modi ties purchased by urban consumers in large ci ties. Serious attention needs to be paid to how representative the basket of commodities and the weights used in constructing the indices are for different groups of consumers. Consumer purchasing patterns and price relationships among commodities change over time. (See Chapter V). 44 • 3. Data on quantities of commodities marketed, transported, and imported or exported. a) Extension agents or agricultural statistics enumerators sometimes attempt to collect data on the volume of commodities marketed in rural areas, particularly at rural markets, which capture part of total marketed output. Whlle absolute volume figures should not be taken too llterally, year to year changes in marketed output may be reasonably accurate indicators of significant changes in production and marketing. It is important to note, however, that policy changes may shift the location of apparent market surpluses, as well as the direction and magnitude of marketed flows. Furthermore, a change in government policy may encourage more officially recorded marketings through formal channels (transfer of sales from informal to formal markets), even though the total quantities sold may not have changed from one year to another. b) Interregional transport data are less common and may be highly inaccurate, depending upon government controls and taxes and whether commodities are transported in small or large lots. D;;lta may be collected at entr"y points to large cl ties, at shipping and receiving points on rail, air and water lines, and at water crossings (ferries). Origin and destination traffic surveys provide accurate and detalled information but are often carried out at only one point in time, which mayor may not coincide with the periods of major commodity flows. When traffic surveys are conducted at intervals over the course of one year, researchers need to assess the representativeness of the periods of data gathering. c) 1m port ar.d export da ta (quan ti ties and value) are usually more accura te than other types of quantity and flow data, but they may understate actual volume of imports and exports if government restrictions, quotas, taxes or overvalued exchange rates encourage smuggling or underlnvoicing. 4. Data on the volume of processed or transformed commodities are sometimes collected by government agencies for taxation purposes. Processing firms are often asked to submit records of the quantities of produce processed to 45 government agencies. Since taxation encourages evasion, government data may significantly underestimate processed output. For example, livestock slaughter statistics are usually quite accurate for large ruminants (cattle, camels, buffaloes) slaughtered in urban areas, but typically incomplete for smaller stock (goats, sheep, pigs, poultry). During rapid appraisals investigators should collect and analyze only secondary data which can be obtained with a minimum of difficulty. If RR teams can only obtain secondary data through extensive digging in government archives, or if aggregation of voluminous records is necessary, then these activities may be best reserved for later in-depth studies. 46 IX. IMPLEMENTING RAPID RECONNAISSANCE SURVEYS 9.1 Elements of Rapid Reconnaissance Field Work The three critical elements of RR field work are: J) direct observation of food system facilities and processes, 2) examination of firm and organization records, and 3) informannterviewing of subsector participants and knowledgeable observers of commodity systems. There is no 5ubsti tute for observing how food system components functions firsthand. This includes inspecting marketing infrastructure, such as marketplace, storage and processing facili ties, as weB as the transportation network and transport equipment. Observation of physical handling and packaging of commodities, wear and tear on produce in transport, and transactions at differen t levels of the system (a t the farmgate, at assembly and redistribution markets, in retail stores) is also necessary and instructive. By directly observing marketing processes and flinctiolls, investigators are able to identify marketing problems and evaluate what key informants say about the organiza tion and opera tion of the system against wha t is actually observed. Examination of detailed firm and organization records can provide data on the volume of the target com modi ty bought, processed, stored and sold during particular time periods, the value of organization assets, purchase and sale prices for inputs and outputs, and costs and returns of marketing operations. Yet in many developing countries records for individual firms are incomplete and inaccurate, particularly for smaller firms. Managers of many smaB firms are illiterate or do not keep records. Even in firms or organizations where data are recorded, these data may not always be organized or recorded in a form which can be readily used by outsiders. Or the records may be organized for the convenience of outside regulators, but they are likely inaccurate and designed to misinform or misrepresent. For example, wholesale grain traders in the Peanut Basin of Senegal keep one se~ of record for government regulatory officials in order to document tha t they trade grain at official prices, and a second set for recording actual purchase and sale prices (Newman, Sow, Ndoye, personal 47 communication, 1985). Some larger firms and marketing organizations may deliberately under record transactions, understate sale prices, underinvoice or in some other way misrepresent financial transactions so as to evade taxation or close government scrutiny. Interviewing key informants in the food system is essential for understandIng theIr perceptions of what constitutes good system performance, problems and constraInts at the level of the fIrm and in the overall food system, and unexpioited or underexploited I marketing opportuni ties. It is preferable to Interview key informants alone and confidentiaJ1y, under which circumstances they are more wilHng to discuss their marketing activities and sensitive topics such as price informatIon, credit arrangements and relations with other firms. Group interviews and informal Delphi methods, which .. encourage in tense and frank interaction among participants, can generate much useful information in a Sllort period of time and provide insights into the interaction of participan ts, but such methods also risk introducing bias into RR findings. 9.2 When~ to Begin the Rapid Reconnaissance In abbreviated supreys, more so than in Il:>nger term studies, whcLt one often finds depends heavily on when the surveys are carried out, who is interviewed, what is observed, and where research is conducted. Whc~re an investiga tion begins is usually a function of the rapid reconnaissance objectives. If RR studies are funded as an input into the design of a project for improving urban food distribution, surveys will begin in urban areas. Researchers will need to analyze demand patterns and prospects and urban food distribution channels before surveying rural producing areas. If RR pn~ced s design of a project which will promote production of particUlar commodities, surveys will usually begin in rural producing areas. When a particular rural area is targeted for production increases, the researchers will visit that area early in the RR to examine constraints to increasing production and marketed output. If the objectives of the RIR are broad and the organization funding the research does not demand that RR focus on any particular segment 0 f the marketing system or a particular geographic area, researchers will usually find it useful to interview wholesalers based in large markets and secondary towns. They are typically more Imowledgeable about the organization and operation of the entire marketing system than other participants, and they often have a vantage point acting as "channel captains" in the marketing system (Harrison et al., 1974). They may be reluctant to divulge much information, especial1y if they perceive that the study is to b(;: used in developing additional regulations and taxes. Special efforts are often required to rela te to key system informants. Where the RR survey begins will suggest where the investiga tors wish to go next. Studies which begin in urban areas will usuaJJy work bad: through marketing channels to rural producing zones. Selection of zones is not determined by hard-and-fast criteria. In some cases researchers will go first to the producing area which ships the largest quantity of produce to the urban market. It may also be useful to visit areas where there is agronomic potential for producing the commodity in question, or for producing significantly more of the com modi ty, in order to examine why production and marketed output are low or nonexistent. When there are several important producing areas, the RR team will need to develop criteria for selecting particular zones and subzones for examination. Random selection is usually sui table for subzones. Selection of zones may sometimes be guided by political objectives or regional equity considerations. Selection of particular villages or subsector participants presents other problems. In RR selection is generaUy purposive and rarely defined clearly, so selection criteria 'have to be established in each case. RR does not include large sample surveys and does not generate estimates which are representative in any statistical sense. In order to sample purposively, RR teams should know something about the population of villages, trading firms, processors, transporters, etc. This can be determined through earlier studies, interviews with I<ey wholesalers or knowledgeable observers of commodity subsectors, and existing enumera tions of firms by government agencies. If the general characteristics of the popUlation are known and different strata can be identified, individuals and firms can be selected randomly from each stratum. For example, RR teams may stratify agricultural prodllcers into five groups: small farmers who do not produce the target commodity; small farmers who produce it but who sell little or none; small farmers who sell a significant proportion of wha t they produce; medium-sized farms which sell most of what they produce; and large farms which sell aU or nearly all of the commodity produced. As a second example, wholesalers can be stratified by volume of the commodity they handle, the approximate value of their assets (vehicles, warehouses, storage facilities), or commodity mix. Purposive sampling becomes more difficult when there are no existing studies or enumerations, if existing enumerations are incomplete or inaccurate, or if knowledgeable observers cannot provide information about numbers of different size/type firms at 49 diflerent levels of the food system. RR teams are then encouraged to select strategically placed such as large scale wholesalers or processors, for informants~ in-depth interviews. They can then proceed to retail firms, first handlers or producers linked vertically to these informants. Alternatively, researchers might interview a second or third informant at the same level of the system. One RR team can proceed backward through the subsector toward producers, while a second can move forward toward consumers (or institutional buyers). Through this process the teams are able to gain a better understanding of the organization of the food system, marketing proce~s and vertical1inkages in the system. They are also able to crosscheck information provided in 'earlier interviews with informants at the same stage or at adjacent stages of the system. Different perceptions of problems and opportunities can also be elicited. These will vary, of course, depending on where firms are placed in the subsystem. 9.3 Processes, Functions and Facilities to Observe During Rapid Reconnaissance The following processes, functions, and facili ties are important to observe and inspect during the rapid reconnaissance: 1) Transae tions for both inputs and outputs a t the farm level (1I possible), at assembly and wholesale marketplaces, and at retail outlets. 2) Handling, weighing and measuring, sorting, grading, packaging, processing, transport, and storage of commodities at different stages of the marketing system, including the farm, assembly markets, processing plants, terminal markets, storage facilities and retail stores. 3) racili ties for buying/selJing, processing, transporting, grading and storing corn modi ties in rural and urban areas. A useful technique for observing facilities, functions and processes is to follow agricultural commodities from the farm to the terminal market, or at least through pa,rt of the production-marketing chain. This is a form of wha t anthropologists call "participant observation." Accompanying wholesale traders or their agents to rural areas and then back to urban markets is one method of observation. Investigators can observe 50 transactions, note costs and losses for a particular marketing trip, asl< the wholes<lier how representative these costs and losses are for his/her business, and ask the wholesaler questions along the way about marketing processes and functions. Researchers are also usually able to spot inefficiencies and problems in the system during these trips, as well as to inspect marketing Infrastructure and facili ties. During RR field work It is often useful to purchase retail lots of the commodIty in question in markets and towns visited during the RR. It is Important to bring small scales along so that these purchased quantities can be weighed, the price per unIt of measure calcula ted, and comparisons made with values obtained in other loca tions. If there are unexpected or unexpectedly large differences, the investIgators can interview traders and retailers on the spo t in order to ascertain reasons for these dif ferences. It may also be useful to bring hanging scales (and accompanying hool<s and ropes for suspending the scale) for weighing bags of produce (or quarters of beef, cra tes of vegetables, etc.) or other units sold wholesale. By weighing produce sold wholesale, the weight of local units of measure, and any variation therein, can be determined. Since produce is often bought and sold in the same units (e.g., sacks) at the farm, in rural assembly markets, and in urban wholesale markets, prices per unit can be calcula ted for produce sold at different levels of the marketing system. Adjustments may need to be made as the commodity moves along the marketing chain for shrinkage, loss or addition of foreign matter. Gross marketing margins can thus be established. 9.4 Selection of Key Informants There are two sets of key informants: subsector participants and knowledgeable observers of subsectors. In the rapid reconnaissance literature subsector participants are not usually considered key informants, who are designa ted as "key" due to their special know lege and understanding of agricultural production and marketing systems (see Beebe, 1985). In this paper the term key informants designa tes selected subsector participants, as well as outsiders with broad and less parochial perspectives. Subsector participants are linked forward and backward to other participants in the production-distribution system. Some participants, particUlarly wholesalers and processors, have a systems perspective about the interrelationships among the parts of the system and reSUlting system performance. They are able to identify both system- 51 wide problems and potentials as well .:'s stage-specific constraints. Other participants have long years of experience in the commodity subsector and in-depth knowledge of particular problems. They often have parochial views and attitudes and may not be able to identify system-wide problems. Their perceptions of stage-specific marketing problems are important, however, and need to be tapped by RR researchers. Table 4 lists key informants who can be interviewed during RR surveys. The advantages and disadvantages of each type of informant are noted. The types of informants who will be interviewed during a rapi rl appraisal will depend on the objectives and focus of the study. It is clearly not necessary to contact all the types of informants listed in Table 4. Moreover, other researchers will doubtless be able to add to the list. In most RR surveys, more than one set of key informants will need to be purposively selected in order to elicit multiple views of subsector performance, gdvernment policies, and subsector potential and problems. Different informants will emphasize different performance attributes and have different ideas of what constitutes desired performance. This reflects different perceptions and priorities, which are often parochial, as well as concerns about the existing and potential distribution of resources and income, which are unstated and implici t. The types and numbers of informants selected at anyone stage of the subsector will depend upon the degree of heterogeneity of the following characteristics of firms: size, patterns of input use, product mix, geographic distribution, resource endowments, management, technology, and standard operating procedures. The more diverse firms at each stage are with respect to the above characteristics, the larger the sample of informants will need to be. Sampling of disadvantaged and less vocal groups, such as landless laborers or nutritionally vulnerable consumers, as well as systema tic contacting of women or their associations, may be necessar} to offset common biases of many rapid appraisers. It is noteworthy that women play an important role in staple food crop production, processing and retailing in many developing countries, particularly in Africa. While agricultural production and marketing interventions affect female participants in the food system and may exclude women from gaining access to certain resources, some analysts have failed to anticipate these impacts. Given recent interest in disaggrega ted welfare effects of food policies, most analysts will probably devote some attention to examining consumption patterns of disadvantaged groups in food systems (Tim mer et al., 1983). .52 ... Table. KEY INPORMANTS IN FOOD SYSTEM RESEARCH. ADVANTAGES AND DISADVANTAGES AS INFORMANTS KEY INPORMANT Wholesalers ~GATNAVDA AS INPORMANTS DISADVANTAGES AS INPORMANTS a) Located at system node which offers vanCAae point and system perspective. b) Knowledae of production, stocks, flows, prices and strenath of demand In different rural and ll:rban areu. a) Extremely busy and often difficult to Interview for more than a shor t period. b) Given typical hostility of government, they may be uncooperatl':'; Informants. I. a) Detailed knowledae of exchange arrangementa with producers and wholesalers. b) Knowledge of market opportunities, productlon, stocks, and prices In particUlar rural area•• a) Knowledge rarely extend. outside circumscribed rural areas. b) May have parochial perceptions and attitudes. Managers of Processing Firms a) Located at system node which offers vantage point and system perspective. b) Knowledge of production and prices In selected rural areu, and demand for processed products In urban markets. c) Oetalled knowledle of exchange arrange,:" ments and rlik-sharlna mechanisms with producers or producer aroup•• a) Given typical hostility Cor government, may be uncooperative 11I1'0rmants. b) May be unwilling to diVUlge details of exchange arranlements vdth producers. c) Will often underreport throughput In order to evade taxation. Transporters a) Knowledge of direction and magnitude of commodity flows. b) Familiar with structure of commodity trade. Can often-Identify large volume traders. .) Do not actually participate In trade, a) Knowledge of magnitude, timing and price. of Imports and exports. b) Detailed knowledge of Import/export practices, procedureJ and regulatlc'I::I. a) May know little about how commodities are assembled for export, or how they are distributed after ImportatIon. b) Since smuggling and underlnvoiclng are common practices in many countries, they may be unWilling to report volumes or prices. c). If rights to Import/export are obtained through privileged access or rentsharlna, they may be unwllllna to discuss buslnes. practices. First Handlers Importers/ Exporters II so lack knowilldle of trading practices, prices and stratcgies. Representatives of Cooperatives, Trade Associations a) Knowledge of numbers and size. of member firms, and their output. b) May effectively represent membership and its perceptions of constraints, opportunities. a) If repre.entatlves are l:ppolnted by government, may not effectIvely represent membership. b) Membership may bl!! restricted to larger fIrms and producers. Bank Loan Officers a) May possess Inform&tlon about the operations, throuKhput and returns of larger wholesalers, processors and retailers. b) Access to information about composition of commercial bank loan portfolios. a) ,. "y not possess systems perspective. May ekIl~ judgements on basis of narrow rate of retu~n criteria. b) May be unwilling to diVUlge confidential information about borrowers' operations. a) Often ma/ot buyers of high value commodities, such as fruits and vegetables, livestock products. b) May have negotiated contractual arranaements with large volume wholesalers, processors or Importers• a) As buyers of final prodUCts, may have limited knowledge of system organization and opera tion. b) Usually constitute small proportion (If final demand for staple commodities. Institutional and Private Sector (Large Supermarket) Buyers .53 itt- IF" .;;. .- .. firIE- 54 9•.5 Informal Interviewing of Key Informants 9.5.1 Interviewing Techniques Researchers will rarely have the luxury of developing formal questionnaires (which are pre-tested, translated and back-translated, etc.) for each type of participant interviewed. Rigorous and intensive questionnaire development might be necessary if the researchers wish to focus attention and resources of the inquiry on one or a small number of stages, or if information obtained in initial interviews is deemed inaccurate, inappropriate or ambiguous. Nevertheless, it is useful to develop interview strategies and topic gUidelines for different groups of participants, including specific questions, desirable sequences of questions, and types or ranges of questions for initial and followup interviews. Informal interviews will be structured in the sense that the intention is to cover important topics in a preferred sequence. Yet they will be unstructured in the sense that interviewers will be able to vary the leng'th and format of an interview, probing promising lines of inquiry in depth, where feasible, or adhering to noncontroversial or less sensi tive topics, where necessary. For example, in one instance the interviewer might encourage a respondent to focus on subsector problems, government policies or marketing opportunities. A parallel interview with another informant at the same stage of the subsector might focus on tha t firm's management, sources and uses of credi t, standard opera ting procedures for carrying O'Jt particular marketing functions, and rela tions wi th other firms. While it is useful to develop informal interview guidelines for different types of • participants before beginning RR field work, it is important to realize that interviewing busy marketing system participants is an art. Investigators rarely have the time to asl< even the most cooperative of informants everything they would llke to ask, unless it is possible to arrange a followup interview or two. So investigators have to focus the interview on particular issues and problems. Allowing informants enough flexibility to discuss issues and topics which interest them or problems which they find especially bothersome can have high payoff. Investigators can often uncover unexpected insights in this way. In addi tion, it can be very effective to challenge informants on particular issues, if only to stimulate discussion and compel them to articulate their views more clearly. Informants who are bored or annoyed by interviews will sometimes offer incomplete or 55 unsatisfactory answers to questions, hl)ping that the inv~stigator will accept those responses uncritically and continue toward completion of the interview. It is very ". important t;o challenge such responses and to demonstrate to the informant that the research ~r unJel':itands enough about the mcrketing system to realize that his/her answer is incomplete or unsatisfactory. To do this in a humorous or clever way can liven up an otherwise routine interview, improve rapport, and facilitate the information gathering task. Informal interviews in rapid reconnaissance are best used to elicit information on informants' perceptions of commodity system problems and opportunities, ideas of how the system can be improved, views of the effect of particular government policies, and the need for policy changes. The emphasis should not be placed on accumulating detailed informa tion on the organization and opera tion of the subsector at each particular stage. 9.5.2 Building in Consistency Checks Informal interviews can be structured so that information about certain topics is obtained in more than one way, either in different sequences of questions or by approaching the topic from two or more angles. For example, interviewers can obtain information about producers' marketed output by first asking producers directly the rlumber of sacks of a commodi ty tha t have been sold since the harvest. An indirect way of obtaining the same information would be to ask the producer the numbers of sacks . I harvested, given and received, and consumed. The residual would then be the number of sacks sold. Information obtained from interviews with key informants about exchange arrangements, risk reducing and sharing mechanisms, credi t arrangements, com modi ty flows and other vertical linkages should be cross-checked wi th informants at adjacent stages of the subsector. The overall validity of rapid reconnaissance findings can be also checked with knowledgeable observers of commodity subsectors, including researchers, certain government technocra ts, selected agricultural project managers, and regionally important business people. A useful method of consistency checking is mirror-image interviewing. This technique involves asking informants at adjacent stages of a subsector the same set of questions. Major dif ferences in responses are generally an indica tor tha t one or both informants are misinforming the research team. Such differences may also indicate that one or both of the informants does not know or accurately recal! the answer to the question. 56 9.5.3 Repeat Interviews • Repeat interviews with cooperative and knowledgeable informants, or interviews with infor na t~ who substitute for uncooperative or less useful informants, are often necessary in order to: 1) Follow up on initial, more general interviews with questions about the organization and operation of particular firms (especially if these firms play an important role in the subsector or at a particular stage of the subsector). 2) Clarify statements or viewpoints presented during an earlier interview. 3) Cross-check information provided during an earlier interview or by another informant at the same stage of the subsector or an adjacent stage•. ia. 4) Discuss more sensitive topics, such as credit arrangements, price formation, relations with other firms, circumvention of government restrictions, parallel markets, etc. Fol w·~up interviews with selected key informants during the rapid appraisal can lay the base for case studies during later phases of research. Case studies entail multiple visits over a longer period, typically a year. 9.6 Interviewing Village Headmen and Other Local Informants One information gathering shortcut in rural studies is to interview traditional leaders, such as village leaders or headmen. In many African countries local notables meet periodically as a group to arbitrate between conflicting parties. These groups often discuss issues of land tenure and use, disputes between farmers and herders and other grievances, and issues of collective organization, such as constructing social infrastructure (SC~, ,015, clinics), road maintenance, and formation of cooperatives. While calling together all the members of a local deliberating body may take time, it may prove valuable if some of the above issues are addressed in an RR study. When the focus of a rapid appraisal is agricultural marketing, it will generally not be necessary to convene a large group of notables. An interview with a village leader or headman will 57 usuaJly suffice. Interviews with older members of the community who are producers or traders can also be very valuable in learning about change in agricultural production and marketing practices and policies over long periods. These informants possess a wealth of local knowledge, as weJl as a longer term historical perspective. Such perspectives are valuable and researchers can elicit perceptions of current IJroblems in historical context, as well as historical information on when particular types of agricultural production technology were first tried and adopted, when rural roads and wholesale trading networks penetrated rural areas, and shifts in agricultural production patterns over time in response to marketing opportunities and avalJablllty of inputs. As a general word of caution, it is advisable not to ask sub-district or vl11age headmen que5tions which demand detailed responses that they are unable to provide. In other words, it is best to tailor questions to respondents' frame of reference and level of knowledge. For example, a subdistrict official is unHl<ely to possess detailed information about the size distribution of farms in his jurisdiction or of marketed surplus 01 particular types of farms, whereas a village headman may well be able to answer questions about these topics. 9.7 Group Interviews Depending upon the cultura: context, interviews of relatively homogeneous groups of subsector participants can elicit views of subsector performance, the need for and effects of government policies and regula tions, and system bottlenecks and opportunities. Group interviews can also serve to legitimize the process of inquiry among potential participants. In many Asian and African countries, for example, agricultural producers are more likely to cooperate in survey research once the village chief or elders have met with the researcher and approved the research agenda. The main drawback of group interviews is that they can be dominated by especially articulate and forceful individuals, whose views and perceptions may not be widely shared. The findings of group interviews or of meetings with reprp-sentatives of producer, trade or industry associations must always be cross-checked with individual informants. Preferably some of these informants will not have attended the group meeting. 58 9.8 Informal Delphi Techniques Informal Delphi techniques for obtaining information about the views and perceptions of marketing system participants are sometimes appropria te in rapid reconnaissance field research (Honadle, 1982). Delphi methods are an itera tive form of information gathering that can involve several group interviews with brainstorming and intense interaction or iterative, private consultations with anonymous informants. Either form of Delphi is designed to elicit candid appraisals of participants' views, perceptions and ideas (in this case, about marketing system performance, problems, constraints and opportuni ties). The Delphi approach assumes tha t the group will move toward consensus and that false or misleading views and perceptions wiJJ be exposed and discredited. This approach is a potentially attractive information gathering shortcut for RR researchers in tha t reliable informa tion on sensi tive topics can supposedly be obtained in a short time span. Moreover, the difficult process of trying to separate out reliable information from unreliable information and misinformation supplied by individual informants can be largely avoided. There are several potentially serious problems with Delphi methods, however, particularly in group meetings. In heterogeneous groupings of Delphi participants, some individuals may not express their views and ideas candidly, deferring to participants with , more power in the marketing system or to representatives of government agencies. Less powerful participants avoid risks by being reticent and not openly criticizing participants who can apply sanctions or use confidential information in a way that harms them. Among groups of peers (participants at the same stage of the marketing system), Delphi methods are most likely to generate reliable information. Individual participants are less likely to refrain from criticizing the government or marketing agencies, or other groups' of participants (especially powerful participants) in the marketing system. Informal Delphi techniques can be quite useful in evaluating the performance of organizations and agencies wi thin the marketing system. For example, representatives of producer cooperatives could diagnose problems associated with the input procurement and distribution, crop storage, and crop marketing practices of the cooperative. It is not recommended to include government overseers of the cooperative or appointed cooperative officers, who are often not producers, in the discussio;'ls. Unequal status among participants will usually preclude frank discussion of problems. 59 9.9 Recording Rapid Reconnaissance Findings When recording information during RR interviews, time-saving techniques must be adopted. There are few things more disturbing to a busy marketing agent than to sit through long interviews where the investigator spends half or nearly half of the time writing longhand notes. Several shortcut techniques are possible. Researchers might only note quantities, prices, and other continuous variables, which are more difficult to recall than quaJi tative da ta, during interviews and reserve detailed recording of other information until after the interview is completed. Standardized formats for different types of data gathering, such as forms for recording prices and quantities in marketplaces, and for different types of informants, such as producers, can also speed up the note-taking process. And, of course, developing an effective shorthand for recording informa tion is another means of shortening the time required for interviews. A common mistake in conducting informalinter'tiews is to postpone writing down observa tions, perceptions and responses until long after interviews are completed. It is also easy to fall into the trap ot taking poor or incomplete notes. It is strongly recommended tha t researchers record the findings of informal interviews immedia tely after each interview is completed. In some cases, using a large informal questionnaire is a valuable tool in forcing analysts to record findings during or shortly after each interview. Having this recorded information will be essential at a later stage when the investigators write up the research results. 9.10 Speeding up Information Cathedng in the Field Depending on the circumstances, the two or three members of a RR team can work separately to increase the rate of information gathering. For example, one member of the team can interview wholesalers and retailers at a rural market, while a second can observe the market, counting the numbers of traders of each type, estima ting the quantities of the commodity brought to the market that day, noting the numbers and types (make, tonnage) of trucks at the market, and chatting with truckers to obtain informa tion about transport cos ts and the magnitude and direction of marketed flows. When the RR team is in rural areas interviewing producers, team members can individually interview producers to get a broader sample. The more sta'ndardized the informal interview format for producers, the less risk there is in having team members carry out individual interviews. When researchers are not following informal interview 60 ., guidelines but are extemporizing, reflecting the interviewers' dif er~nt findings may differ quite significantly, n~search interests and biases, or perhaps the asking of questions on similar topics in quite different ways. While splitting up RR teams will accelerate information gathering, this may not always be desirable. When two or more researchers participate in an informal interview, they may interpret the informant's responses in different ways. At the end of each day, or perhaps immediately after each interview, the team members can discuss informants' responses and their implications. By comparing interview findings and inferences within each RR team, possible bias in interpretation can be offset. Teaming up to do interviews can also speed up individual interviews. One researcher can pose questions in an informal, conversational style, while the other records the informant's responses. The two researchers can take turns asking questions in their areas of specialization during an interview. In the final analysis, researchers participa ting in rapid reconnaissance surveys will need to evaluate the tradeoff between breadth of coverage (numbers of participants interviewed) with depth and accuracy of coverage. Choices about interviewing strategies will be influenced by the time available for RR, the sl<ills and experience of researchers participating in RR, and the degree of variation in interview findings and interpretations of informant responses. !' 9.11 Periodic Meeting of Rapid Reconnaissance Teams and Exchange of Preliminary Findings Although it is recommended that RR groups divide into two or three person teams during the field research, the teams should not work in isolation. The ir,dividual RR teams should meet periodically to discuss preliminary findings during the reconnaissance surveys. The meetings may be infrequent (weekly or perhaps biweekly) due to logistical difficulties. Nevertheless, it is important that the different teams discuss preliminary research findings, tentative conclusions and hypotheses inferred from the findings, information gaps, and needed data gathering emphases during the RR. In some cases there may be disagreement among the participants. In the ensuir.g debate, the researchers may uncover preconceptions, assumptions, and unclear or un~ta ed unjustified inferences. Periodic meetings are also useful in helping the researchers to focus ir~c easingly on key research issues, which typically emerge during the course of 61 " the RR surveys, rather than to (;OntlnllC gathering Information In a broadt:!I", less directed way. 62 .J X_ PRESENTING RAPID RECONNAISSANCE FINDINCS 10.1 Writing up the Results of the Rapid Reconnaissance Survey It is useful to outline the written report that will be prepared before initiating the 'jllrvey research. The final organization of the report will probably be somewhat different, bu t outlining the report during the planning stage helps to identify key issues 8& well as who will be responsible for addressing these issues. When preparing the final report, separate papers along disciplinary Jines are discouraged. It is the responsibility of the team leader to integrate different disciplinary findings into a summary report. In many cases it is desirable to write a concise (less than 10 pages) summary which discusses key findings and marketing system problems, whiJe attaching the more detailed reports of subject matter specialists as annexes. If policy-makers are the principal audience for the document, it is recommended to prepare a cross-referenced summary that directs the interested reader to elaborations in technical annexes. A suggested format for the summary paper is as follows: 1. Executive summary 2. State RR objectives 3. Briefly discuss research methodology and its limitations ./1 4. Define poJicy objectives and performance goals 5.. Food (or commodity) system overview a) Commodity characteristics b) Commodity consumption patterns c) Supply situation for the target commodity d) Price analysis ~) Marketing system organization f) Marketing system opera tion g) Marketing infrastructure h) Institutions and poJicies affecting agricultural marketing 6. Identify and diagnose key problems and constraints, as welJ as untapped and underexploited opportunities to improve system performance 63 7. Suggest polley and program options 8. Identify further research needs 1,0.2 Presenting the Prlnclpal findings of the Rapid Recon ais~ In many cases RR Investigators will be asked to brief any policy or decision makers who commissioned the RR about the principal findings of the investIgation. Presentations should be kept relatively short (no more than one hour) and may need to, be considerably shorter for senior policy-makers. Ample time should be left for dIscussion of the RR findings. The RR team should try to elici t dIscussion of the identified marketing problems and constraints. Do the policy-makers agree that these are key problem areas? If 50, how would they rank order them? If not, which problems were missed, and why are these important? In addition, what areas do polley-makers view as most promising for further research? Do they support research in certain areas but appear reluctan t to approve research in other areas? Wha t are the reasons for the enthusiasm and/or hesitancy? 10.3 Follow-up to Rapid Reconnaissance A rapid reconnaissance study may serve as a one-time, stand-alone effort which requires no further research. Some policy-makers may feel that a RR survey provides enough information on which to make informed policy choices. Others may wish to .. commission further studies, which follow guidelines laid out in the RR summary report• RR surveys will identify food and marketing system problems and constraints which may be diagnosed in depth during later phases of marketing research. In moving from rapid reconnaissance to a longer term program of research, it is often difficult to reach a consensus as to the key problem areas and the most feasible and relevant research topics. It is very important to present the RR findings as clearly and lucidly as possible to help establish the rank ordering of research problems and priori ties for further research agenda. One tool for faci!1 ta ting this process could be a two-dimensional array having perceived subsector problems as row headings and subsec:tor participant groups and government agencies as column headings. Subsector problems could then be discussed and priori tized by the researchers in collabora tion wi th policy-makers. It may be useful 64 to distinguish between system-wide and stage-specific problems. Criteria for prioritizing problems and constraints wHl have to be specified clearly. Another tool might be a marketing constraint matrix with constraints as rows and selection criteria such as system-wide effect of the constraint, feasibiJity of doing research on each constraint, liI<eJihood tha t research findings wiJJ lead to polley change, and other factors as columns. Ordinal scales can be developed to rank order the potential payoff of doing further research on different constraints and taking action to reJieve constraints. , .. 6.5 • XI. LIMITATIONS OF RAPID APPRAISAL 1I.! Time-Buundedness and Bias Rapid reconnaissance should be used with caution as a tool for generating policy prescriptions because it has two fundamental limitations. First, it does not allow for observation of all the seasonal dimensions of agricultural production and marketing (see Chambers, Longhur:it, Bradley and Peachem, 1979). It is not a substitute for longer term, more data-intensive studies in obtaining a comprehensive understanding of agricultural production and marketing systems. Rapid reconnaissance techniques are particularly unsuited for obtaining reliable information on flow variables, such as agricultural production, quanti ties and prices of inputs and outputs purchased and sold, commodi ty flows, and labor inputs, over long periods. Yet RR methods are often useful as a first step in the design of formal surveys with which flow data are collected. Second, rapid reconnaissance can be subject to bias, because informants and areas or facilities for visits are selected non-randomly. Predispositions of researchers may also introduce biases (Chambers, 1980, 1981; Car ruther::; and Chambers, 1981). Policy prescriptions coming out of rapid appraisals may be premature. They may also fail to identify key marketing opportunities, or they may lead to the implementation of weak or invalid marketing strategies (Belshaw, 1981). Rapid reconnaissance is a very useful tool, however, for identifying food system problems, constraints and opportunities, and for informing longer term, focused studies of particular aspects of food systems. In some cases it may be useful to test on a pilot basis, during longer term research programs, technical or management innovations Identified during RR. The International Livestock Centre for Africa (ILCA) has found this willingness to experiment and transfer known technologies to a new research environment to be an important dimension of a longer term program of applied research (personal communication, Stephen Sandford, 1984). Such a research strategy will help to satisfy client groups that clamor for action and that are not inclined to support long term research programs without demons trable inter mediate results. 11.2 Limitations of the Subsector Framework An important strength of the subsector approach is its focus on how the system operates as a whole, and its emphasis on dynamic forces which drive and transform food 66 • systems. Yet subsector is often evaluated with reference to concepts which r~e formance have been criticized as elusive and difficult to quantify. Concepts such as workable competition and effective coordination have been attacked as imprecise and dependent upon researchers' jUdgement. The subsector approach has also been criticized as beIng overly descriptive and unfocused. Since the subsector .framework is a systems approach, albeit a pQrtial one (i.e., a subsystems approach), researchers can become bogged down in describing and analyzing many different elements of the subsystem. The temptation is always to colJect too much information without adequa te focus. This is a very real problem. Proposing working hypotheses about the system at the outset and tailoring da ta gathering ".fforts to test, disprove and modify these hypotheses can help to focus data collection. This approach is at the heart of inductive methods of research (Platt, 1964). At the same time, it is important to remember that rapid reconnaissance is essentialJy an empirical approach to initiating research. Overemphasis of formal hypothesis testing at an early stage of the research may effectively limit the scope and comprehensiveness of rapid reconnaissance. In particular, expatriate researchers who do applied research on food systems in developing countries may not know enough about these systems to formulate useful, relevant and testable hypotheses before th ~y begin doing research. It is important to know in general how the system is organized and how it functions within its particular institutional, social, and macroeconomic environment before focusing research on gnir~hta data to test a set of hypotheses. Acquiring this understanding requires an investment in analyzing secondary data, observing the system a t different levels and locations, conducting informal interviews with system participants and knowledgeable observers, and identifying system constraints and opportunities before finalizing hypotheses that guide intensive data collection. Another criticism of the subsector approach is that it is partial. In focusing on one or more related commodities, researchers may miss important interrelationships among commodities whkh are substitutes in production and consumption. Use of a broC'.der partial equilibrium framework, such as the Food Policy Analysis approach dev\~loped by Timmer, Falcon and Pearson (1983), is often advanced. This approach focus\~ on several key staple crops simultaneously. Heavy emphasis is placed on price analysis broadly construed; that is, analysts examine input and output prices) and interest, wage and exchange ra tes, and their effects on incentives facing producers, 67 .. marketing agents and consumers. The effects of macroeconomic policies and intcrna tional trade on agricultural production/marketIng systems are analyzed. There is much in the Food Policy Analysis perspective that can be usefully adapted to subsector studies. Clearly, the effects of macroeconomic policies and interna tiona I trade on incentives facing subsector participants (inc!uding input suppliers, producers, first handlers, wholesalers, processors, retailers, importers and exporters) are often underestimated if the analysis takes a too partial view of production/marketing systems. Moreover, the subsector approach needs to be expanded to Include examination of close substitutes of the commodity under examination in production and consumption. The challenge for researchers examining !;ubsectors is to know when and how to broaden the approach to include important factors without making the research agenda unmanageable. 68 • XIL CONCLUSION Rapid reconnaissance surveys are often sponsored by donor agencies or developing country governments seeking information for policy formulation or justifications for developing and fundIng projects. The rapid reconnaissance guidelines presented in this paper are designed to be a tool for generating knowledge about food and commodity marketirlg systems in a short period of time that may be used in policy formulatIon and project development, but tha t are best suited to informing longer term programs of applied research. The fundamental strength of rapid reconnaissance is that applied researchers can obtain a broad understanding of commodity marketing systems in a relatively short time and that system constraints and opportunities can be identified. Through this process it is hoped that the design and policy relevance of follow-on applied research will be improved. By identifying promising arp,as of further research, rapid reconnaissance can help to improve the alJocation of resources to programs of food and agricultural marketing research. Rapid appraisal can also play an important role in pointing out actual or proposed policies with adverse or potentiaUy adverse consequences. This can help policy makers to reform counterproductive policies or to avoid poor decisions. Rapid appraisals can also be conducted ir1termittentJy and . repeatedly in the policy and project formulation, monitoring and evaluation cycles. In this way rapid recon ai~sance can inform the policy-making process and project design and implementation on an ongoing basis in a timely and cost effective way. 69 ,. , BIBLIOGRAPHY Agricultural Administration, Vol. 8, No.6, entire issue, November 1981. Bain, Joe, Industrial Organization, Second edition, John WHey and Sons, New York, 1968. Barton, Clifton G., "Credit and Commercial Control: Strategies and Methods of Chinese Businessmen in South Vietnam," Unpublished Ph.D. Dis (~rta ion, Cornell University, 1977. Beebe, James, "Rapid Rural Appraisal: The Critical FIrst Step in a Farming Systems Approach to Research," Farming Systems Support Project, Networking Paper No.5, International Programs, Institute v: Food and Agricultural Sciences, Univeristy of Florida, Gainesville, 1985. Belshaw, Deryke, "A Theoretical Framework for Data-Eccmomising Appraisal Procedures wi th Applications to Rural Development Planning," Rapid Rural Apprais'!!" IDS Bulletin No.4, Vol. 12, Institute for Development Studies, University of Sussex, October 1981. Berg, Elliot, "Why Don't LDC Governments Liberalize Agricultural Markets?" Paper presented at the World Bank's Agricultural Marketing Seminar, May 6-17, 1985. Boomgard, James J., Stephen P. Davies, Steve Haggblade, Donald C. Mead, Subsector Analysis: Its ~ature, Conduct and Potential Contribution to Small Enterprise Development, MSU International Development Working Paper No. 26, Dept. of AgricUltural Economics, Michigan State University, East Lansing, Michigan, 1986. Byerlee, Derek, Michael Collinson et al., Planning Technologies Appropriate to Farmers: Concepts and Procedures, International Maize and Wheat Improvement Center (CIY MMT), El Batan, 1980. . Campbell, David J., "Response to Drought Among Farmers and Herders in Southern Kiljiado District, Kenya, Human Ecology, Vol. 12, No.1, 35-64. . Campbell, David J. and David T. Trechter, "Strategies for Coping with Food Consumption Shortage in the Mandara Mountains Region of North Cameroon," Social Science and Medicine, Vol. 16,2117-2127,1982. Carruthers, Ian, "A Mental Construct for Unstructured On-Farm Interviews for Use in Rapid Rural Appraisal," Agricultural Administration, Vol. 8, No.5, 1981. Carruthers, Ian and Robert Chambers, "Rapid Appraisal for Rural Development," Agricultural Administration, Vol. 8, No.6, November 1981. Casley, Dennis J. and Denis A. Lury, M,2nitoring and Evaluation of AgricUltural and Rural , Development Projects in Developing Countries, World Bank, 1982. _ _ _ _ _-', Data Collection in Developing Countries, Clarendon Press, Oxford, 1981. Chambers, Robert, Institute for Development Studies (IDS), University of Sussex, "Shortcut Methods 'in Information Gathering for Rural Development Projects," Paper for World Bank Agricultural Sector Symposia, January 1980• .. • 71 .. • , ' "Rapid Rural Appraisal: Rationale and Repertoire", Pub !.!£M.!rJ in is ~t and Development, Vol. I, 9.5-106, 1981. Published carlier as IDS Discussion Pape'r No. 1.5..5, September 1980. Chambers, Robert, Richard Longhurst, David Bradley and Richard Feachem, "Seasonal Dimensions to Rural Poverty: Analysis and Practical Implications," IDS Discussion Paper No. 142, Institute for Development Studies, February 1979. Cohen, Abner, "The Social Organization of Credit in a West African Cattle Madcet," Africa, Vol. 3.5, 1:8-20, 196.5• • •• CoJ1inson, Michael, "A Low Cost Approach to Understanding SmaJJ Farmers," Agricultural ,Administration, Vol. 8, No.6, 1981. '~ -=~ = Farming Systems Research in Eastern Africa: The Experience of CIYMMT and Some National Agricultural Research Services, 1976-81. MSU International Development Paper No.3, Dept. of Agricultural Economics, Michigan State University, East Lansing, Michigan, 1982. Connor, J.M., R.T. Rogers, B.W. Marion arid W.F. MuelJer, The Food Manufacturing Industr;' ~ Structure, Strategies. Performance and Policies, D.C. Heath and Company, Lexington r Massachusetts, 198.5. Ellman, Anthony, "Rapid Appraisal for Rural Project Preparation," Agricultural Administration. Vol. 8, No.6, 1981. Food Studies Group, Oxford, U.K., "Results of the SmalJ Farmer Survey, Bandundu Region, Zaire," Report under Contract for USAID/Kinshasa, 1985. French, Ben C., "The Analysis of Productive EffiCiency in Agricultural Marketing: Models, Methods, and Progress," in A Survey of Agricultural Economics Literature, Vol. 1, .fradi tional Fields of Agricultural Economics, edi ted Lee. R. Martin, Univet·si ty of Minnesota Press, for the American Agricultural Economics Association, Min ej~polis, 1977. Goetz, Stephan and Michael T. Weber, Fundamentals of Price Analysis i'l~ oping Countries' Food Systems: A Training Manual to Accompany the Microcom.p.uter Software Program "MSTAT", MSU International Development Working Paper No. ,M.S.U., Dept. of Ag. Economics., East Lansing, Michigan, 1986. Goetz, Stephan with Harold Riley and Michael Weber, "Diagnostic Assessments of FoodGrain Subsectors in the Context of Food Security in Africa," Forthcoming MSU Interna tiona1 Development Working Paper, Dept. of Agric. Economics, Michigan Sta te University, East Lansing, Michigan, 1986. 'Goldberg, Ray A. and Richard C. McGinity, Agribusiness Man gem~nt f~r Developing Countries: Southeast Asian Corn System and American and Japanese Trends. Affecting !!" Ballinger Publishing Company, Cambridge, Massat:husetts, 1979. Harrison, KeJJy, Do~ald Henley, Harold Riley and James Shaffer, ,Improving Food Marketing Systems in Developing Countries: Experiences from Latin America, LAMP Research Report No.6, Michigan State University, East Lansing, Michigan, 1974. 72 .. .. • Harriss, Barbara, "Th(~rc Is a Method in My Madness: Or is it Vice Versa? Measuring I\grlc:ultural Marl<ct Performance," Food Research Institute Stuc!ies, Vol. 16, No.2, 197213, 1979. Hay, Roger, "The Food Accounting Matrix: An Analytical Device for Food Planners," Oxford Bulletin of Economics and Statistics, No. 42, May 1980. Heimberger ~ Peter G., Gerald R. Campbell and William D. Donbson, "Organization and Performance of Agricultural Markets," A Survey of Agricultural Economics Literature, Volume .3, ,Economics .2LWJ!lfare, Rural Development and Natura! Resources in Agricutture, 19 110s to 19705. edited by Lee R. Martin, University of Minnesota Press, for the American Agricultural Economics Association, Minneapolis, 1981. Hildebrand, Peter E., "Combining Disciplines in Rapid Appraisal: The Sondeo Approach," Agrir,ultural Administration. Vol. 8, No.6, 1981. ___ ..J "Summary of Sondeo Methodology Used by ICTA," Instituto de Ciencla y Technologi.:\ Agrico/!1s, Guatemala, June 1979. Holtzman, John, John Stuatz and Michael T. Weber, ~n Analysis of the Livestock Production and Marketing Subsystem in Northwest Province. Cameroor, MSU Rurai Development Working Paper No. 10, Michigan State University, East Lansing, Michigan, 1980. Honadle, George, "Rapid Reconnaissance for Development Administration: Mapping and Moulding Organiza tiona1 Landscapes," World Development. Vol. 10, No.8, 1982. -=-_-;--__...J. Rapid Recon ais anc~proaches to Organizational Anal)!,sis for Development AdmJnistration. Working Paper No.1, Development Alternatl,ves, Inc., Washington, D.C., December 1979. Inst! tute for Development Studies, Universi ty of Sussex, Bulletin No. I~, Vol. 12, October 1981. Rapiq,~al Appraisal, IDS Landa, Janet T., "A Theory of the Ethnically Homogeneous Middleman Group: An Institutional Alternative to Contract Law," The Journal of Legal Studies. Vol. 10, June 1981. Marion, Bruce and NC '17 Committee, T~rganizatio. ! . ~Q. Eerfo mance Food System, D.C. Heath and Company, Lexington, Massachusetts, 1986. of the U.S. Marion, Bruce, "Application of the Structure, Conduct, Performance Paradigm to Subsector Analysis," N.C. Project 117 Working Paper No.7, Madison, Wisconsin, November 1976. , Menegay, Merle R., "Indonesian Assessment Development Report," USAID, Bureau of Science and Technology, Office of Rural and Institutional Development, SmaIJ Farmer Marketing Access Project, Washington, D.C., 1985. Minot, Nicholas W., Contiact Farming and Its Imoact on Small Farmers in Less Developed Countries, MSU International Development Working Paper No. 31, Department of Agricultural Econornics, Michigan State University, East Lansing, Michigan, 1986 • 7.3 • Morris, Michael L., "The I3AME Fleuve Cereals Marketing Study: Methodology," ISRA, Bureau d'l\nalyses Macro-Econorniques, Working Paper 8'-16, October 1985. Ndoy~, Ousseynou and Mark Newman, "Approches Methodologlques pour l'Etude de la Co.nrnercialisatlon des F(oduits Agricoles et Allrnentaires au Senegal," Actes de l'Atelier de Kahone, 8··13 Mars 1984, Instl tut Senegalais de Recherches Agricoles, Bureau d'Analysl.":s Macro-Economiques, Document de Travail 84-2, Dakar, December 1984. Newberry, D.M.G. and J.E. Stiglitz, .The Theory of Commodity Price Stabilization: A Study in the Econornics of Risk, Clarendon Press, Oxford, 1981. Newman, Mark D., A. Pap Sow, Ndoye Ousseynou, personal communication and "Regulatory Uncertainty, Government 'Objectives, Organization and Performance of Cereals Markets: The Senegalese Case," ISRA/BAME Worldng Paper 8.5-11, Dakar, 1985. Pagliaro, Jehn, Method for Conducting a Diagnostic Study uf Wholesale/Retail Food Distribution Systems in Developing Countries, Prepared for USAID, August 1984. Platt, John R., "Strong Inference," Science, Vol. 146, No. 3642, 16 October 1964. Pruitt, William, "Small Farmer Study, t:entral Bandundu, 1984: Notes in Progress," USAID/DEO, Kinshasa, Zaire, August 1984. Rhoades, Robert E., "Informal Survey Methods for Farming Systems Research," Human .Qr.ganizatlon, Vol. 44, No.3, 215-218, Fall 198.5. _~ , "The Art of the Informal Agricultural Survey," International Potato Center, Social Science Departrnent Training Document 1982-2, Lima, Peru, 1982. Riley, Harold and Michael T. Weber, "Marketing in Developing Countries," in Future Frontiers in Agricultural Marketing Research, edited by Paul L. Farris, Iowa State University Press, Ames, Iowa, 1983. ir. geveloping Riley, Harold and John Staa tz, FO'Jd SY5tern Organiza tion Pr~blerns Countries, Agricultural Development Council Report No. 23, December 1981. Ruttan, Vernon W., "Social Science Knowledge and Institutional Change," American Journal of Agricultural Economics, Vol. 20, No• .5, December 1984. Ruttan, Vernon W. and Yujiro Hayami, "Toward a Theory of Induced Institutional Innovation," The Journal of Development Studies, Vol. 20, No.4, July 198~. Sandford Stephen, ILCA, Livestock Policy Analysis Unit, personal communication, 1984. Scherer, F. M., Industrial Market Structure and Economic Performance, Second Edi tion, Rand McNally, Chicago, 1980. Scott, Chris, Sampling for Monitoring and Evaluation, Technical Supplement to Moni toring and Evaluation of Agriculture and Rural Development Projects, World Bank, 1985. Shaffer, James D., "On the Con~ept of Subsector Studies," Agricul tural Economics, Vol • .5.5, pp. 333-335, 1973. A~ican Journal of . 74 . ._ ' "Food System Organization ilnd Performance: Toward a Conceptual Framework," American Journc:d of Agricultural Economics, Vol. 62, No.2, May 1980. Sh,dfer, James D. with Michael Weber, Harold RlJey, and John Staatz, "Influencing the Design of Marketing Sy~tems to Promote Development in Third World Countries," Paper prelsented at the Interna tiona 1 Workshop on Agricultural Markets in the Semi-Arid Tropics, sponsored by ICRISAT, Hyderabad, India, October 24, 1983. Shaner, W.W., P.F. PhiHpp, and W.R. Schmehl, Farming Systems Research and .Q£velopment: Guidelines for Developing Countries, consortium for International Development Study, BOUlder, Colorado, 1981. ,namr~fhS Jacqueline R. and Chantal Dejou, Forthcoming Monograph on Rapid Reconnaissance Survey of Grain Marketing in Burkina Faso, Center for Research on Economic Development, University of Michigan, Ann Arbor, 1986. Sosnick, Stephen H., "Operational Criteria for Market Performance," in Paul L. Farris, editor, Market Structure Research: Theory and Practj.f~ in Agricultural E_S2nornics, pp. 81-125, Iowa State Universi ty Press, Ames, Iowa, 1964. Southworth, V. Roy, WilHam O. Jones and Scott R. Pearson, "Food Crop Marketing in A tebubu District, Ghana," Food Research Institute Studies, Vol. 17, No.2, 157-195, 1979. Swift, Jeremy, "Rapid Appraisal and Cost-Effective Participatory Research in D. J Pastoral Arens of West Africa," Agricultural Administration, Vol. 8, No.6, 1981. rimmer, C. Peter, "The Relationship Between Food Marketing and Price PoHcy," Paper p.esented at the World Bank's Agricultural Marketing Seminar, May 6-17, 1985. Timmer, C. Peter, Walter P. Falcon and Scott R. Pearson, Food Poli'£y Analysis, Johns Hopkins University Press, 1983• • Tomel<, William G. and Kenneth L. Robinson, Agrisultural Produc! Prices, Second Edition, CornelJ University Press, Ithaca, 1981. to • 75 .. ro!,' I • MSU INTeRNATIONAL oeveLOPMCNT PAPeR.S • lOP No. I Carl K. Eicher and Doyle C. Baker, "Research on Agricultural DovelojJ,nent In SubSaharan Africa. A Critical Survey," 19&2/346 pp.). $ lOP No. IF Carl K. Eicher at Doyle C. Baker, "Etude critique de la recherche sur 10 developpernent agrlcole en Afrique subsaharlenne," 198', (43' pp.). $I().OO lOP No.2 Eric W. CraWford, "A Simulation Study of Constraints on Traditional Farming Systems In Northern Nigeria," 1982 (l36 pp.). $ '.00 lOP No.3 M.P. CoJllnsoll, "Farming Systems Research in Eastern Africa. The Experience of· CIMMYl and Some National Agricultural Research Services, 1976-81," 1982 (67 pp.). $ 4.00 lOP No.4 Vincent Barrett, Gregory lassiter, Dayld Wllco<:k, Doyle Baker, and Eric Crawford, "Animal Traction In Eastern Upper Volta. A Technical, Economic and Institutional Analysis," 1982 (l n pp.). $ '.00 IDPNo. , John Strauss, "Socia-Economic Determlnant:s of Fo.>d Consumption nnd Production in Rural Sierra Leone. Application of an Agricultural Household Model with Seyeral Commodities," 1983 (91 pp.). lOP No.6 Beyerly Fleisher and Lindon J. Robison, "Applications of Decision Theory and the Measurement of Attitudes Towards Risk in FSlrm Management Research In Industrialized and Third World Setting:.," 198' (106 pp.). Ou~ .~O af Prinr $ '.00 MSU INTERNATIONAL DEVELOPMENT WORKINC PAPERS & ,. • WP No. I Daniel Galt, Alyaro Diaz, Mario Contreras, Frank feain, Joshua Posner and Franklin Rosales, "Farm 109 Systems Research (FSR) in Honduras, 1977-81: A Case StUdy," 1982 (48 pp.). Out of Print WP No.2 Edouard K. Tapsoba, IlCredit Agrlcole et Credit Informel dans Ie Region Orientale de Haute-Volta: Analyse Economique, Performance Institutionnelle et Implications en Matiere de Polltlque de Deyeioppement Agricole," 1982 (L2' pp.). Out of Print WP No. J W.P. Strassmann, "Employment and Construction: Multicountry Estimates of Costs and Substitution Elasticities for Small Dwellings," 1982 (48 pp.). Out of Print WP No.4 Donald C. Mead, "Sub-contracting in Rural Areas of Thailand," 1982 ('2 pp.). Out of Print WP No.' Michael T. Weber, James Pease, Warren Vincent, Eric W. Crawford and Thomas Stilwell, "Microcomputers and Programmable Calculators for Agricultural Research in Developing Countries," 198,' (113 pp.). WP No.6 Thomas Stilwell, "Periodicals for Microcomputers: An Annotated Bibliography," 198.3 (70 pp.). WP No.7 W. Paul Strassmann, "Employment and Housing in lima, Peru," 1983 (96 pp.). WP No.8 Carl K. Eicher, "Faire Face a la Crise Alimentaire de l'Afrlque," 198.3 (29 pp.). Free WP No.9 Thomas C. Stilwell, "Software Directories for Microcornputers: An Annotated Bibliography," 1983 (14 pp.). Sec IDWP 1122 WP No. 10 Ralph E. Hepp, "Instructional Aids for Teaching How to Use the TI-'9 Programmable Calculatort" 1983 (1)3 pp.). Out of Print WP No. II Michael L. Morris and Michcoel T. Weber, "Programmable Calculator (TI-'9) Programs for Marketing and Price Analysis in Third World Countries," 1983 (1o, ppJ. Out of Print WP No. 12 Valerie Kell!'/ Robert D. Stevens, Thomas StilwelI, and Michael T. Weber, "An Annotated Directory of Statistical and Related Micl'ocomputer Software for Socioeconomic Data Analysis," 1983 (16' pp.). $ 7.00 WP No. 13 Chris Wolf, "Guidellnu for Selection of Microcomputer Hardware," 1983 (90 pp.). $ '.00 $ 5.00 See IDWP H21 Out of Print , , M5U INTERNATIONAL OEVELOPMENT WORKING PAPERS - CONTINUED ~ ErIc W. Crawford, Tlng-Ing Ho, and A. Allan Schmid, "User'. Guide to BENCC)S-SUp4JrCalc Template for 8eneflt-Cost Analysis," 1984 (.3' pp.). $ Copy of BENCOS Template In IBM PC· DOS 1.1 Format, on single sided double density diskette (readable on most MS-OOS systems). $1'.00 WP No. U lame::! W. Pease and Raoul Lepage with Valerie Kelly, Rita Laker.OJok, Brian Thelen, and Paul Wolberg, "An Evaluation of Selected Microcomputer Statistical Programs," 1984 (187 pp.). $ 7,00 WP No. 16 Stephen Davies,. James Seale, Donald C. Mead, Mahmoud Badr, Nadia EI Sheikh, and Abdel Rahman Saldl, "Small Enterprises In Egyptl A Study of Two Governorates," 1984 (J 00 pp.). $ WP No. 17 Thomas C. Stilwell, "Microcomputer Statistical Packages for Agricultural Research," 1984 (23 pp.). $ 3.00 WP No. 18 Thomas C. Stilwell and P. Jordan Smith, "An Annotated Olrectory of Citation Database, Educational, System Diagnostics and Other Miscellaneous Microcomputer Software of Potential Use to Agricultural Scientists In Developing Countries," 1984 (34 pp.). $ 3.00 WP No. 19 Amalla Rinaldi, "Irrigation In Soull,,:rn Africa: An Annotated Blbllography," 198' (60 pp.). $ 4.00 WP No. 20 Danlf!1 C. Goodman, Jr., Thomas C. Stilwell, and P. Jordan Smith, "A Microcomputer Based Planning and Budgeting System for Agricultural Research Programs," 198' (n pp.). $ '.00 WP No. 21 Thomas C. Stilwell, "Periodicals for Microcomputers: An Annotated Bibliography," Second Edi tion, 198' (89 pp.). $ '.00 Tho·mas C. Stilwell, "Software Directories for Microcomputersl An Arlnotate'd Bibllography," Second Edition, 198' (21 pp.). $ 3.00 WP No. 23 Alan Hrapsky with Micha~1 Weber and Harold Riley, "A Diagnostic Prescriptive Assessment of the Production and Marketing System !'or Mangoes in the Easttlrn Caribbeafl," 198' (106 pp.). $ WP No. 24 Donald C. Mead, "Subcontracting Systems and Assistance Programs: Opportunities for Intervention," 198' (32 pp.). $ 3.00 WP No. 25 Carl Liedholm, "Small Scale Enterprise Credit Schemes: Administrative Costs and th(1 Role of Inventory Norms," 198' (23 pp.). $ 3.00 WP No. 26 James J. Boomgard, Stephen P. Davies, Steve Haggblade, and Donald C. Mead, "Subsector Analysis: Its Nature, Conduct and Potential Contribution to Small Enterprise Development," 1986 U7 pp.). $ 4.00 WP No. 27 Steve Haggblade, Carl Liedhclm, and Donald C. Mead, "The Effect of Polley and Pollcy Reforms on Non-Agricultural Enterprises and Employment in Developing Countries: A Review of Past Experiences," 1986 (133 pp.). $ '.00 WP No. 28 John T. Milimo and Yacob Fisseha, "Rural Small Scale Enterprises in Zambia: Results of a 198' Country-Wide Survey," 1986 (76 pp.). $ '.00 WP No. 29 Stephan GC'~tz and Michael T. Weber, "Fundamen:·,ls of Price Analysis in Developing Countries' Food Systems: A Training Manual to Accompany the Microcomputer Software' Program 'MSTAT, '" 1986 (148 pp.). $ 7.00 John S. Holtzm",n, "Rapid Reconnaissance Guidelines for Agricultural Marketing and Food In Developing Countries," !986 (7' pp.). System ;~es arch $ '.00 WP No. 14 WP No. 2:2 WP No. 30 3.00 '.00 5.00 Copies may be obtained from: MSLI International Development Papers, Department of Agricultural Economics, 7 Agriculture Hall, Michigan State University, East Lansing, Michigan 48824-1039, U.S.A. All orders must be prepaid in United States CII'" ~cy. Pleoase do not send cash. Make checks or money orders payable to Michigan State University. There . ~ discount on all <...'ders of 10 or more sale copies. Individuals and institutions in the Third World and USAID ,.. ,: '. ,', lay receive single copies free of charge. "'-:~t r'.·~ ~ . " • ,