Economic Development Institute Mqqg 3W5> of The World Bank Toward the Rural-Based Development of Commerce and Industry Selectecd Experiences froiim East Asia Elditedt by Yujiro Hayami EDI LEARNING RESOURCES SERIES EDI LEARNING REsouRcEs SERIES Toward the Rural-Based Development of Commerce and Industry Selected Experiences from East Asia Edited by Yujiro Hayami The World Bank Washington, D. C. Copyright 0 1998 The International Bank for Reconstruction and Development / THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America First printing May 1998 The Economic Development Institute (EDP) was established by the World Bank in 1955 to train officials concerned with development planning, policymaking, investment analysis, and project implementation in member developing countries. At present the substance of the EDI's work emphasizes macroeconomic and sectoral economic policy analysis. Through a variety of courses, seminars, and workshops, most of which are given overseas in cooperation with local institutions, the EDI seeks to sharpen analytical skills used in policy analysis and to broaden understanding of the experience of individual countries with economic development. Although the EDI's publications are designed to support its training activities, many are of interest to a much broader audience. EDI materials, including any findings, interpretations, and conclusions, are entirely those of the authors and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. Because of the informality of this series and to make the publication available with the least possible delay, the manuscript has not been edited as fully as would be the case with a more formal document, and the World Bank accepts no responsibility for errors. Some sources cited in this paper may be informal documents that are not readily available. The boundaries, colors, denominations, and other information shown on any map in this volume do not imply on the part of the World Bank Group any judgment on the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. Requests for permission to reproduce portions of it should be sent to the Office of the Publisher at the address shown in the copyright notice above. The World Bank encourages dissemination of its work and will normally give permission promptly and, when the reproduction is for noncommer- cial purposes, without asking a fee. Permission to photocopy portions for classroom use is granted through the Copyright Clearance Center, Inc., Suite 910, 222 Rosewood Drive, Danvers, Massachusetts 01923, U.S.A. The complete backlist of publications from the World Bank is shown in the annual Index of Publications, which contains an alphabetical title list with full ordering information. The latest edition is available free of charge from the Office of the Publisher, 1818 H Street, N. W., Washington, D. C. 20433, U.S.A. Yujiro Hayami is a professor of international economics at Aoyama-Gakuin University and Visiting Lee Teng-hui Professor of World Affairs Chair at Cornell University. Library of Congress Cataloging-in-Publication Data Toward the rural-based development of commerce and industry: selected experiences from East Asia / edited by Yujiro Hayami. p. cm.-(EDI learning resources series, ISSN 1020-3842) Includes bibliographical references (p. ). ISBN 0-8213-4026-3 1. Rural industries-East Asia. 2. Rural development-East Asia. I. Hayami, Yujiro, 1932- . TI. Series. HC460.5.T68 1998 338.095'09173'4-dc2l 97-34148 CIP Contents Foreword vii Preface ix Contributors xi Abbreviations and Acryonyms xv 1 Toward an Alternative Path of Economic Development: An Introduction Yujiro Hayami Relational Contracting and Vertical Integration 4 Conditions for Relational Contracting 7 A Historical Perspective 10 Major Findings of the Case Studies 13 Conclusions 18 References 18 PART I. Historical Experiences of Japan 2 Technical and Institutional Innovations in Rice Marketing in Japan 23 Toshihiko Kawvagoe Overview of Prewar Japanese Agriculture 24 The Traditional Rice-Marketing System in Japan 26 Marketing Innovation in Modem Economic Development 30 The Role of Rural Entrepreneurs 37 Government and Farmners' Cooperatives 40 Conclusions 43 Notes 44 References 45 3 Rural Entrepreneurs in the Cotton-Weaving Industry of Japan 47 Motoshige Itoh and Masayutki Tanimoto Overall Historical Picture of the Cotton Textile Industry of Japan 49 The Early Form of the Local Production System: Iruma, 1850-80 53 Development of the Putting-Out System 58 The Contemporary Putting-Out System 63 iii iv Con1ten2ts Conclusions 66 Notes 67 References 68 4 The Formation of Toyota's Relationship with Suppliers: A Modern Application of the Community Mechanism 69 Kazuto Wada Historical Background of Automobile Manufacturing in Japan 70 Supplier Characteristics in the Early Years 77 Tightening the Relationship 79 Reorganization of the Kyoho-kni 81 Conclusions 83 Notes 84 References 85 PART II. Current Developments in East Asia 87 5 Export-Oriented Garment Industries in the Rural Philippines 89 Masao Kikuchi The Garment Industry in the Philippines 91 Garment Subcontractors for Exports in Laguna Province 99 Profile of Firms and Entrepreneurs 104 Garment Production by Subcontractors 113 Sub-Subcontracting Arrangements 116 The Role of Government 122 Conclusions 123 Notes 125 References 128 6 The Rural Garment and Weaving Industries in Northern Thailand 131 Akihiko Olno and Benja Jirapatpimnol History of the Thai Textile Industry 133 Scope of Analysis 134 The Garment Sector 136 The Weaving Sector 138 Characteristics of Rural Entrepreneurs and Workers 139 Contractual Arrangements 141 Support for Rural Enterprises 143 Conclusions and Policy Recommendations 146 Appendix 6-1: Workshops and Sewers in the Garment Industry 148 Appendix 6-2: Workshops and Workers in the Weaving Industry 152 Notes 157 References 158 Conttents v 7 Township-Village Enterprises in the Garment Sector of China 161 Deqiang Liu and Keijiro Otsuka Development of Rural Enterprises and the Garment Sector in China 163 Changing Organizational Structures of TVEs in the Garment Industry 165 Contractual and Institutional Arrangements Between Enterprises 169 Marketing Channels and the Performnance of Enterprises 175 Characteristics of Rural Entrepreneurs 178 Support Policies of Local Government 181 Conclusions 183 Notes 185 References 186 8 Rural Entrepreneurship and Industrial Development in Korea 187 Jtung-Hivan Lee and Chong-Hyutk Suh Historical Sketch of Rural Industrialization Policy in Korea 189 Macro Perspective on the Reasons for Unsuccessful Rural Industrialization 191 A Case of Successful Rural Industrialization in Korea 195 Conclusions 207 References 209 9 Political Bases of Rural Entrepreneurship: Korea and Taiwan, China 211 David W. Lane The Puzzle of Rural Industrialization in Taiwan, China, and Korea 212 Policy Incentives and Politics 219 Financing Small Industry 222 Political Institutions and Social Organization 231 Conclusions 234 Notes 235 References 238 PART III. Comparative Perspectives 241 10 Organizational Characteristics of Rural Textile Industries in East Asia 243 Akihiko Ohno anid Masao Kikuchi Analytical Framework 243 Rural Textile Industries in an East Asian Perspective 249 Dualism in Legislation and the Relational Contracting System 254 A Long-Term View 256 The Rural Garment Industry in the Philippines and Thailand 258 Notes 264 References 265 vi Contents 11 On the Relevance of East Asian Experiences: A South Asian Perspective 267 V V Bhatt The Relational Contracting Model, East Asian Experiences, and Economic Rationale of Coordinating Mechanisms 267 Rationale for Nonfarm Development in South Asia 274 Cases of Rural Industrialization in India 276 Comparative Evaluation of East Asian and Indian Experiences 283 Conclusions 287 References 290 Foreword A s part of its mission to help share the experiences of economic growth among countries and regions, the World Bank's Economic Development Institute ad- Ii xministers the Program for the Study of the Japanese Development Management Experience, financed by the Human Resources Development Trust Fund established at the World Bank by the Government of Japan. This program, which sponsors important research on East Asia's development experi- ence, has published four books on topics ranging from corporate governance to the role of the civil service in economic development. The program was managed from its inception in 1989 by the late Hyung-Ki Kim, whose vision and intellectual leadership were instru- mental in shaping the program and its products. Before joining the Bank, Hyung-Ki had a distinguished career in public service and held several senior positions in the Korean gov- einment, including vice minister of education. Hyung-Ki was highly respected by his colleagues for his breadth of knowledge and intellectual integrity. This book and the other titles published to date are part of his legacy to the field of international development. Vinod Thomas, Director Economic Development Institute vii Preface T he explosive growth of developing economies, although it has been decelerating since the 1970s, has translated into a high growth rate of the labor force. Because this growth has far outpaced the increases in land available for cultivation, the rural sector has experienced increasing incidences of landlessness and fragmentation of landholdings, which has pushed dispossessed rural people into urban slums. To cope with this problem, maximum effort must be expended to increase the demand for labor in rural areas. Agriculture and small- and medium-scale industries are two obvious areas to inves- tigate for possible expansion of work opportunities. In agriculture, the so-called Green Revolution technology that has contributed greatly to increased food production and farm employment in Asia is not considered sufficient to solve the problem. The second option must be explored. For rural industries to develop to be of more than local significance, it is necessary to establish marketing channels that connect the production of small, scattered enterprises in rural areas with the large urban and foreign sources of demand. In developing economies, however, markets are insufficiently developed to serve this purpose. Some form(s) of per- sonal contract mechanisms are required as a substitute for impersonal markets in order to save transaction costs and to reduce risks amid severe information imperfection and mar- ket segmentation. The purpose of this study is to investigate the production and trade contracts now being tried at the grass-roots level in support of rural-based industrialization. The points of ref- erence in this investigation are historical explorations of the early stage of industrializa- tion in Japan and field surveys of current developments in East Asia. Special attention is given to the conditions that allow the relational contracting system to become the dominant production organization over a vertically integrated system. The term relational contracting is used here to encompass the long-term, continuous contract relationships between small, ix x Preface rural-based manufacturers and traders or agents of large, urban-based firms, as well as similar contracts between rural-based manufacturers and their workers, which are en- forced and maintained primarily by personal ties and community obligations. This con- tract mechanism is vital-it is the channel that conveys the demands of large markets, both foreign and domestic, to the small, isolated manufacturers in rural areas that would other- wise be subject to severe imperfections of market information. The absence of formal documents and official data on organizations and the practices of informal economic agents, such as cottage manufacturers and petty traders, made it neces- sary to rely on intensive case studies for both the historical and the field investigations. Because of the ephemeral nature of the production and trade contracts of small- and me- dium-size enterprises in developing economies, as well as the presence of sensitive issues such as taxation and labor codes, the investigation relies largely on personal observations and informal conversations with entrepreneurs rather than formal questionnaires. To verify the responses of a given entrepreneur, efforts were made to cross-check the information with the contracting parties, and repeated visits were needed until consistent answers could be established. The time-consuming nature of such an investigation limited the num- ber of cases that could be included in the study. The approach came to resemble that of anthropologists and sociologists, although the theoretical framework is based on standard economics. The nature of this intensive, but small-scale, investigation yielded results that are more illustrative (sometimes even anecdotal) than analytical. Nevertheless, the findings clearly demonstrate that rural people in developing areas of Asia possess sufficient entrepreneur- ship-if they are not held back by undue regulation, taxation, and macroeconomic insta- bility-to organize a network of relational contracts that link their production activities with reservoirs of national and international demand. Although we recognize the limitations of this approach, the illustrative examples col- lected in the case studies can provide the basis for the development of hypotheses that can be further investigated in larger, more analytical research studies in the future. Such con- tinuing work would clarify the vital role of rural-based development of commerce and industry in the alleviation of poverty and inequality in developing economies. The authors gratefully acknowledge the contributions to this volume of the helpful com- ments and discussions during the planning and interim workshops for this study. We thank, in particular, Masahiko Aoki (Stanford University), Gershon Feder (World Bank), J. C. R. Geaun (National Taiwan University), An-Pang Kao (National Chengshi University), Masahiro Okuno-Fujiwara (University of Tokyo), Jean-Philippe Platteau (University of Namur), and Alberto Valdes (World Bank). We were very fortunate to have excellent sup- port from John Didier and Latifah Alsegaf of the World Bank. On behalf of the authors, I would like to express our deep grief on the passing of Hyung- Ki Kim, who initiated this project and guided us with his broad perspective and solid judgment to the end of his life. This book is dedicated to his soul. Yujiro Hayami Contributors V. V. BHATT V. V. Bhatt is a consultant to the World Bank, where he has held a number of positions, including chief of the Public Finance Division. Previously he was the chief executive of the Industrial Development Bank of India, an adviser to the Reserve Bank of India, and has served as a director of several other financial institutions. He has written exten- sively on devlopment strategies; his most recent book is Financial Systems, Innovations, and Development. YuiRo HAYAMI Yujiro Hayami is a professor of international economics at Aoyama-Gakuin University and Visiting Lee Teng-hui Professor of World Affairs Chair at Cornell University. His pub- lications include Agricultural Developrmenit: An Initernationial Perspective (with V. W. Ruttan; Johns Hopkins University Press, 1971); Centriry ofAgricultural Growth in Japan: Its Relevance to Asian Development (University of Tokyo Press and University of Minnesota Press, 1975); Anatomy of A Peasant Economy: A Rice Village in the Philippines (ed., with V. W. Ruttan and H. M. Southworth; University Press of Hawaii, 1979); Asian Village Economy at the Crossroads (with M. Kikuchi; University of Tokyo Press and The Johns Hopkins University Press, 1981); Agriculttral Development: An International Perspective, revised and expanded edition (with V. W. Ruttan; The Johns Hopkins University Press, 1985); and Development Econom7- ics: From the Poverty to the Wealth of Nations (Oxford University Press, 1997). MOTOSHGE ITOH Motoshige Itoh is a professor of economics at the University of Tokyo, where he teaches international economics. Recent English publications include Small and Medium Enterprise Stupport Policy of Japan (with S. Urata; World Bank, 1995); Wool in Japan: Structural Change in the Textile and Clothing Market (with Christopher Findlay; Harper Education, 1994); "The Japanese Distribution System and Access to the Japanese Market," in P. Krugman, ed., Trade wvith Japan: Has the Door Opened Wider? (University of Chicago Press, 1991); and Eco- xi xii Contributors nomic Analysis of Indutstrial Policy (with K. Kiyono, M. Okuno-Fujiwara, and K. Suzumura; Academic Press, 1991). BENA JIRAPAIPIMOL Benja Jirapatpimol is an assistant professor of sociology at Chiang Mai University. She received the B.A. from Thammasart University and the M.A. in sociology from Chulalongkorn University. Her areas of interest are urbanization, industrialization, social development, and the working conditions of women, with research projects in social de- velopment, women's working conditions, and the industrial and informal sectors. Publi- cations include several articles related to women's labor in northern Thailand. TosHInKo KAWAGOE Toshihiko Kawagoe is a professor of economics at Seikei University. He conducted exten- sive field surveys on local marketing systems in Java when he was an agricultural econo- mist at the UN/ESCAP CGRPT Centre, Bogor, Indonesia, from 1987 to 1989. He is the author, with Yujiro Hayami, of The Agrariant Origins of Comnmerce and Industry (Macmillan and St. Martin's Press, 1993) and "Farmers and Middlemen in a Transmigration Area in Indonesia," Btulletin of Indoniesian Economic Studies (vol. 25, no. 3, 1989). His work has been concerned with Japanese agricultural policies in a historical perspective, and recent publi- cations include "Land Reform in Postwar Japan," and "Deregulation and Protectionism in Japanese Agriculture," in Juro Teranishi and Yuaka Koai, eds., The Japanese Experience of Econonzic Reforms (Macmillan and St. Martin's Press, 1993), and "Postwar Land Reform in Japan: An Economic Evaluation," The Economic Review (vol. 46, no. 3, 1995). He is also the coeditor (with Sueo Sekiguchi) of East Asian Econonoies: Challeniges and Transfornmation (ISEAS, 1995). MASAO KIKUCHI Masao Kikuchi is a professor of agricultural economics at Chiba University. His major field of study has been agricultural and rural development in South and Southeast Asia. His recent publications include Asian Village Econiomy at the Crossroads: An Economic Ap- proach to Instituttional Change (with Y Hayami; The Johns Hopkins University Press, 1981); Irrigation Investment Trends in Sri Lanka: New Con7struction and Beyond (with P. B. Aluwihare; International Irrigation Management Institute, 1991); "Sustainable Agriculture in Asian Developing Countries: An Economist's Perspective," in Sutstainable Agriclulture Develop- ment in Asia (Asian Productivity Organization, 1994); "Investments in Irrigation: Issues and Priorities," in Agricultural Water Management Technology in Asia and thte Pacific (Asian Productivity Organization, 1995); and "Regional Diversity in Rice Technology Adoption in Sri Lanka," Japanese Jouirnal of Tropical Agricuiltnre (December 1995). DAVID LANE David Lane is a Ph.D. candidate in government at Harvard University, writing on the varying responses of small industrialists to capital scarcity in postwar Japan, the Republic Conttributors xiii of Korea, and Taiwan (China). He has received research grants from the Fulbright Com- mission, the Social Science Research Council, the Japan Foundation, and the Korea Foun- dation, among others. In addition to work on both China's politics and its artistic heritage, his past research emphasizes the political economy of industrailizaiton in East Asia, and includes "The New Competitors: Industrial Strategies of Korea and Taiwan," written jointly with Stephen Haggard and Robert Wade for the Office of Technology Assessment, Con pet- ing Economies: America, Euirope, and the Pacific Rinm (U.S. Government Printing Office, 1991). JUNG-HWAN LEE Jung-Hwan Lee is research director of the Korea Rural Economic Institute. His research has concentrated on the theory of production and technological change, and related pub- lications include "The Measurement and Sources of Technological Change Bias, with an Application to Postwar Japanese Agriculture," Economica (vol. 50,1983); "Review on Evalu- ating Returns to Public Investment and Clarification Using the Dual Theory: With an Ap- plication to Korean Agriculture," Korean Rutral Economic Revieztw (vol. 10, 1987); and "A Duality between Output and Input Market: With Respect to Producer's Surplus," Korean Rural Economic Revie7v (vol. 12,1989). His recent work has concerned the structural adjust- ment of the agricultural sector in the process of economic adjustment, and related publica- tions include "Economic Development and Transformation of Agricultural Structure in Korea: An International Perspective," Journal of Rutral Development (vol. 16,1993); and "Law of Aging in Agricultural Labor Forces," Korea Rutral Economic Review (vol. 16, 1993). DEQIANG LIU Deqiang Liu studied in the graduate school of economics at Hitotsubashi University. He is an assistant professor of economics at Tokyo Gakugei University, where he teaches devel- opment economics, international economics, and China's economic reform. His publica- tions include Micro-Economic Reform in China (with K. Otsuka and N. Murakami; Nihon Keizai Shinbunsha, 1995) and "Technical and Allocative Efficiency among Socialist Enter- prises: The Case of the Garment Industry in China," Journal of Comparative Economics (with K. Otsuka and N. Murakami; vol. 19, 1994). AKIHIKO OHNO Akihiko Ohno is an associate professor of economics at Osaka City University, where he teaches development economics, with a focus on Asian economies. His recent English pub- lications include "Modernizing Agents and Organizational Adaptation of Factory Work- ers in Thailand," The Developinig Econonmies (vol. 33, no. 3, 1995) and "Technological and Labour Absorption in the Indigenous Sugar Industry of India: Analysis of Appropriate Technology," Hitotsuibashi louirnal of Economics (with Yukihiko Kiyokawa; vol. 36, 1995). KEUiRO OTSUKA Keijiro Otsuka is a professor of economics at Tokyo Metropolitan University, where he teaches development economics. He received the M.A. from the Tokyo Metropolitan Uni- xiv Contributtors versity and the Ph.D. from the University of Chicago. He was visiting scientist at the Inter- national Rice Institute from 1986 to 1989 and has been visiting research fellow at the Inter- national Food Policy Research Institute from 1993 to the present. He is the coauthor of Comparative Techiology Choice in Development: Tlhe Indian anid Japanese Cottoni Textile Indus- tries (Macmillan, 1988); Thle Economics of Contract Choice: An Agrarian Perspective (Clarendon, 1993); Modern Rice Technology alnd Inconme Distribution in Asia (Lynne Rienner, 1994); and Microecononmic Reform in China (Nihon Keizai Shinburn, 1995). CHoNG-HYuK SUH Chong-Hyuk Suh is research director of the Division of Rural Development at Korea Rural Economic Institute (KREI). His publications include The Strtucture of Industrial Credit Mar- ket in R7ural Korea (KREI, 1985); Evaluation of Rutral Industrialization and Off-Farm Inicome Policy in Korea (KREI, 1991); A Study on the Regional Product Differentiation of Korean Rice (KREI, 1993); Impact of Agricuiltural Trade Liberalization on Regional Agricultuiral Productioni in Korea (KREI, 1994); and Analysis of Regional Economic Strtucture and Strategiesfor Rural Non- Farm Induistries in Korea (KREI, 1995). MASAYUKI TANMOTO Masayuki Tanimoto is an associate professor of economic history at Tohoku University, where he teaches Japanese economic history. Most of his research has concentrated on the development of indigenous industries in the nineteenth and twentieth centuries in Japan. His English publications include "The Evolution of Indigenous Cotton Textile Manufac- ture before and after the Opening of the Ports," Japanese Yearbook on Business History, vol. 9 (Japanese Business History Institute, 1992); "Japanese Industrialization and the Indigenous Economic Development in Modem Japan," Historical Review (no. 539, 1995); " The Refor- mation of Indigenous Industry," (with Osamu Saito) in M. Umemura and Y Ymamoto, eds., Japanese Economic History, vol. 3 (Oxford University Press, forthcoming); and "The Emergence of Modem Enterprises and the Development of Modem Industry, Indigenous Industry" (with T. Abe) in M. Miyamoto and T. Abe, eds., Japanese Business History, vol. 2 (1995). KAzuo WADA Kazuo Wada is an associate professor of economics at the University of Tokyo, where he teaches business history. He received the Ph.D. from the London School of Economics and Political Science (subject: the British electricity industry before World War II). He recently edited, with Haruhito Shiomi, Fordism Transformed: The Development of Production Methlods in the Autontobile Industry (Oxford University Press, 1995). He has also written "The Devel- opment of Tiered-Firm Relationships in the Automobile Industry: A Case Study of Toyota Motor Corporation, Japanese Yearbook on Business History (vol. 8, 1991), reprinted in Will- iam Lazonick and William Mass, Organizational Capability (Edward Elgar, 1995). Abbreviations and Acronyms AFDC Agriculture and Fisheries Development Corporation ASEAN Association of Southeast Asian Nations BMW Bonded Manufacturing Warehouse CTVE TVE operated in cooperation with an SE or STE CV Coefficient of variation DTI Department of Trade and Industry EPZ Export-processing zone EQ Export quota FOB Free on board GNP Gross national product GTEB Garment and Textile Export Board ILO International Labor Organization ITVE Independent enterprise form of TVE JV Joint venture KMT Nationalist Party, Taiwan MITI Ministry of Trade and Industry NDDC Nationalist Dairy Development Corporation QC Quality control RTW Ready to wear SE State enterprise STE State trading enterprise TVE Township-village enterprise YOE Year of enterprise establishment xv 1 Toward an Alternative Path of Economic Development An Introduction Yujiro Hayami T he traditional paradigm in development economics, as espoused by theorists from Marx to W. A. Lewis, assumed that the process of modem economic growth is necessarily associated with a major shift in labor from rural areas to urban indus- trial centers. In this process, manufacturing activities in rural households and small work- shops are outcompeted by modem factory production in the urban centers. As a result, the rural sector becomes increasingly specialized in the primary production of food for the growing urban population and in cash crops for export earnings to support domestic in- dustrialization, while the rural labor force is linked with the urban informal sector to form a reserve for the urban formal sector. Nevertheless, the logic of economic development does not dictate that industrialization and urbanization are necessarily intertwined and inseparable, as assumed in the tradi- tional paradigm. The studies reported in this volume examine whether there is an alter- nate route to economic development in which there is movement of the modem production base into the rural sector, rather than migration of the rural labor force into the urban sector. In this alternative path, widespread rural industrial activities could be organized in a decentralized manner by exploiting not only the physical labor but also the entrepre- neurial ability of rural people-two resources that were underutilized in the past. This rural-based development strategy, if found to be feasible, could alleviate the major diffi- culty in the tradeoff between growth and equity that confronts developing countries. Indeed, the strong population pressure now faced by developing countries has created serious consequences for urban-based development according to the traditional paradigm. 1 2 Towvard the Rutral-Based Development of Comm1lerce and Indutstry Although population growth in many developing economies began to decelerate in the 1970s, the rapid growth of the labor force has continued as the population from the years of high birth rates has matured and entered the workplace. In most developing econo- mies, increases in the rural labor force have not been matched by increases in available farmland. The significant decline in the farmland available to each worker has aggravated rural poverty, increasing the fragmentation of landholdings and leading to growing num- bers of landless agricultural laborers. The result has been large-scale migration from rural to urban areas. Only a small fraction of the migrants, however, have found productive employment in modern industries, even in the economies of Southeast Asia that have recently recorded significant industrial development. This is partly because the modern industrial sector is still a minor part of the total economy, and partly because industrial technology has been imported from developed countries that have biases toward high capital intensity. The majority of city dwellers are thus forced to seek their subsistence from informal economic activities while living in urban slums. How are nations to cope with growing poverty and inequality in both rural and urban areas? If the core of this problem is determined to be population pressure that reduces the returns to labor relative to capital and land, maximum effort must be allocated to the expansion of demand for labor. Two obvious areas for the expansion of the demand for labor are agriculture and small- and medium-scale industries. But the Green Revolution technology that has contributed greatly to increases in food production and farm employ- ment in Asia is not considered sufficient to solve the problem. The second option must be explored. The evidence indicates that rural nonfarm activities carried out mainly in small-scale enterprises (including farm household enterprises) are a very important source of em- ployment and income in developing economies. As much as 30 to 50 percent of the rural labor force is either primarily or secondarily engaged in a wide range of nonfarm activi- ties, which generate 20 to 40 percent of rural household income. Particularly significant is rural manufacturing. Employment in this sector often exceeds that in urban manufactur- ing establishments; rural-based, small-scale industries are generally not only more labor- intensive, but also more productive per unit of scarce capital than their large-scale counterparts (Chuta and Liedholm 1979; Kilby and Liedholm 1986). These observations suggest a hypothesis that the promotion of rural industries is one of the most effective ways to alleviate rural poverty and pathological urban growth. To prosper, these indus- tries require the support of a network of commerce and other service activities. For opti- mal efficiency, local commerce and industry should be organized to utilize human resources in rural areas with low opportunity costs. The historical experience of Japan has proved that the development of labor-intensive, small-scale industries is the most effective way to achieve the dual goals of growth and equity. In the initial stage of modern economic growth in Japan, a large number of entre- preneurial peasants participated actively in trading and manufacturing. This process had already begun in the feudal Tokugawa period (1600-1868) in response to the gradual de- velopment of the market economy, and it accelerated with Japan's opening to foreign trade Toward ani Alternative Path of Economic Development 3 and national unification by the Meiji Restoration in 1868 (Smith 1956, 1959, 1988). The contributions of these small-scale, rural-based enterprises to the national economy were no less significant than those of large, modern corporations developed by the urban entre- preneurs who emerged from the preindustrial merchant class (such as Mitsui and Sumitomo) and the former warrior class (such as Mitsubishi). This development of rural- based enterprises made it possible for Japan's industrialization to be less capital-intensive than that of other latecomers to modem economic growth (Smith 1988). In the past, small-scale enterprises have been seen as being exploited by their larger counterparts in an effort to take advantage of cheap labor under factor market dualism, characterized by differentials in wages and interest rates between large and small enter- prises. These small enterprises have also been thought to tend to fail when the economy reaches maturity and factor market dualism is eliminated. This has not, however, been the experience of Japan's economic development. Even today, long after the demise of factor market dualism, Japan's industrial strength is largely based on numerous small-scale en- terprises, many in rural areas (Kiyonari 1980; Patrick and Rohlen 1987). It appears that this experience has been replicated in Taiwan, China. Is there a prospect for developing countries today tb follow a similar pattern of rural- based development in commerce and industry? If sufficient entrepreneurship can be mo- bilized in rural communities, then perhaps this is possible. A recent pilot study for Indonesia (Hayami and Kawagoe 1993) found that local marketing systems (consisting of a large number of petty traders and processors based in rural villages) work efficiently in econo- mizing on the use of scarce capital and educated manpower, while making intensive use of local resources, especially labor that has a low opportunity cost. The peasant entrepre- neurs who organized this system were found to be ingenious in designing and applying innovative trading practices and contracts for sharing risks and minimizing transaction costs through the effective use of community relations. The findings of this pilot study for Indonesia, when compared with the historical expe- rience of Japan, suggest the possibility of rural-based economic development in develop- ing countries and offer a model for postsocialist states that are working to create and organize markets. To explore the possibility further, this research project attempted to fill the gap between the observations from the Indonesian case study and conjectures from Japan's historical observations through in-depth case studies of historical experiences and current developments in East Asia. The studies in this project focus on the social and institutional context for the creation and development of rural entrepreneurship, although quantitative information was sought as well. Special emphasis is given to conditions that allow "relational contracting" to be- come the dominant production organization in the vertically integrated system. The term relational contracting is used here to represent the long-term, continuous contract relations between small, rural-based manufacturers and traders and/or agents of large, urban-based firms, as well as between the rural-based manufacturers and their workers, that are en- forced and maintained prirnarily through personal ties and community obligations. This contract mechanism is critically important as a channel to convey demands from large 4 Twovard the Rural-Based Developnment of Comnierce and Indutstry markets, both domestic and foreign, to small manufacturers in rural areas, who are subject to severe imperfection of market information. Development of the relational contract is expected to increase the access of rural labor to nonfarm economic activities. Thus, "the alternative path of development" that we are going to explore is not the self-contained development of the rural sector. Instead, it is the expansion of industrial and commercial activities based on rural labor and entrepreneurship, which are inseparably linked with the activities of urban entrepreneurs. The alternative path is thus envisioned as the bal- anced growth of the rural and urban sectors, which is a contrast to the disproportionate growth of the latter in the traditional paradigm. Relational Contracting and Vertical Integration As a basic working hypothesis, it can be postulated that the key to the wide diffusion of industrial and commercial activities in rural areas is development of the relational con- tracting system relative to the vertical integration system. Alternative Modes of Organization In vertical integration, the production and marketing activities of a commodity (or com- modities), from production of raw materials to delivery of the final product to consumers, are integrated, either partially or wholly, and under the control of a centralized manage- ment hierarchy. In contrast, these activities are decentralized in the relational contracting system under the generally autonomous management of independent entrepreneurs, co- ordinated by trade contracts, and usually organized by a principal who controls a link in the marketing chain. For example, a cotton-spinning company may establish a weaving division within the firm to ensure an outlet for its yarn; alternatively, it may contract out the cloth-making process to outside weaving houses. The latter is "relational" if it is developed as a long- term contract relationship that is enforced by accumulated personal ties and mutual trust. Contracts tend to be interlinked in the relational contracting system. For example, in a "putting-out" contract, a principal (for example, a spinning company) advances materials (yam) for processing to an agent (a weaving house), and later collects the finished prod- ucts (cloth) at the piece-rate payment. This system thus combines contracts for the supply of materials and the purchase of products with credit contracts for financing part of the agent's working capital and, if looms are loaned, fixed capital in the processing machines. Even in simpler "advance-order" contracts, in which the principal guarantees the pur- chase of finished products but does not advance materials, the provision of extra services such as credit guarantees and technical guidance is usually involved if the contract is to be considered relational. Relational contracting and vertical integration can be considered alternatives to an effi- cient market. If the market is efficient, with no transaction costs and no risk, inpersonal Towvard an Alternative Path of Economic Development 5 spot transactions based solely on prices are sufficient to coordinate an economywide divi- sion of labor, and neither vertically integrated firms nor interfirm cooperation through relational contracting should be needed. In the real world of incomplete information, however, transaction costs are significant and risk is high; this is especially the case in developing countries, where markets are small and segmented, the supply of needed market information--such as product stan- dards, grading, and brand names-is not well established, and insurance and credit insti- tutions are not sufficiently developed to cope with market risk. Thus, a trader who finds that a commodity is in great demand in an urban market or abroad will face the difficulty of procuring adequate quality supplies in a timely manner from the spot market. In order to exploit the profit opportunity, he may have to build and manage a factory for the mass production of standardized commodities; in this way he vertically integrates manufactur- ing and trading activities. Alternatively, he may organize relational contracts with autono- mous producers to assure a timely supply of the needed commodities of the quality and quantity required. If vertical integration is commonly chosen, the resulting industrial structure will in- clude a small number of large enterprises, which are likely to be located in urban-indus- trial centers to take advantage of economies of urban agglomeration, especially access to market information. There is little doubt that only with the development of the relational contracting system will industrial and commercial activities be spread widely over the rural sector. While the economies of agglomeration may also work among entities clus- tered in an urban site under a relational contract, this advantage can be minor compared with the merits of mobilizing rural labor and entrepreneurship, with their low opportu- nity costs, if the efficient design and practice of relational contracting can effectively re- duce information imperfection in the rural areas. Development of relational contracting is not based solely on the initiative of urban busi- ness principals-who hold information on demand in wide national and international markets-but also can be promoted by rural-based entrepreneurs. It is quite possible for rural entrepreneurs to approach urban traders and offer contracts for their mutual benefit. Irrespective of who takes the initiative, the adoption of relational contracting by rural entrepreneurs is an innovation in marketing organization that allows access to new mar- kets. This innovation is often associated with the production of new commodities (such as nylon cloth) based both on new materials (nylon yarn) and production processes (power looms). In that sense, the rural entrepreneurs who enter relational contracts with urban principals can legitimately be called "innovators" in the definition of Schumpeter (1961, p. 66). The mobilization of such latent entrepreneurial ability within the rural population is the key to success in the rural-based development paradigm. Spectrutm of Contracts It must be emphasized that both the vertical integration and subcontracting systems are not homogeneous within each category; they include a wide variety of organizations. A 6 Towvard the Rtral-Based Development of Comnierce antd Iniduistry vertically integrated firm can be structured in a rigid management hierarchy with central- ized decisionmaking. It can also be organized in a more decentralized way-in the ex- treme, each division can be structured in a nearly autonomous management unit, with interdivisional transactions (such as those for parts supply and assembly) negotiated among the divisions and accounted for in market or imputational prices. Similarly, the relational contracting system can be tightly structured in a hierarchy. Or- ganized from the parent or core firm to the higher-tier and lower-tier subcontractors, it involves intensive guidance, consultation, and assistance in managerial, financial, and tech- nical aspects, as exemplified by the subcontracting arrangement commonly used in the automobile industry in Japan today. Relational contracting in this form of subcontracting is not very different from the "divisionalized" form of vertical integration. At the other extreme, however, the relational contracting system can be organized loosely through short- term contracts in a way that differs little from spot-market transactions. Thus, various forms of vertical integration and relational contracting range along a con- tinuum. In this sense, the popular dichotomies in the theory of industrial organization- such as "make or buy" and "hierarchy or market"-are considered to represent the polar cases of this spectrum. One extreme is to make all materials and parts within a hierarchi- cal, internal organization, and the other is to rely completely on purely impersonal spot- market transactions. Merits and Drawbacks of the Alternative Systenms The major goal of this study is to identify the social and institutional conditions that gen- erally prompt the selection of the relational contracting system, thus identifying the poli- cies that will strengthen these conditions in support of the rural-based development of commerce and industry. The elements of this critical issue will be detailed in the next section. In preparation for this presentation, the merits and drawbacks of relational con- tracting relative to the vertical integration system will be identified through risk and trans- action costs according to orthodoxy in the theory of industrial organization (Coase 1937; Alchian and Demsetz 1972; Williamson 1975,1985). Vertical integration is essentially the system of hiring labor to work within a firm under the command of its management, whereas subcontracting is the system of contracting out the work to outside management. Naturally, the risks and transaction costs associated with contracts "hiring-in labor" and "contracting-out work" are distinctly different. In the vertical integration system, both capital and labor costs are largely fixed for the firm, at least in the short run. Therefore, risks arising from fluctuations in both the product and material input markets are predominantly shouldered by the principal (firm management) relative to the agent (hired labor) in the vertical integration system. In the relational con- tracting system, the schemes for risk-sharing can be determined by the principal (parent firm) and the agents (subcontractors). The transaction costs associated with hire-in labor contracts, especially the cost of moni- toring the work effort of hired laborers, are known to be very high. Indeed, how to pro- Tozvard an Alternative Path of Econiomic Development 7 mote work morale and prevent shirking by employees through incentives such as promo- tions and bonuses have been central issues in the theories of industrial management and organization. There is no cost to the principal in monitoring the work effort of contracted- out labor, and the costs are relatively modest for the agents, whose employees are typi- cally few in number and easy to monitor. The monitoring cost is especially small where the labor is supplied by family members and friends, which is usually is the case in rural- based enterprises. The principal or parent firm management in the relational contracting system, however, may face severe difficulties in forcing its agents to observe the terms of a contract, such as the quality of products to be supplied and the date of their delivery. For example, an auto- mobile assembler may be greatly harmed by a loss of reputation in the market if he uses parts supplied by a subcontractor that are later found to be inferior to the contractually agreed quality standard. He may also suffer losses by operating his plant at less than full capacity if a sufficient supply of parts is not delivered at an agreed-upon date. The cost of preventing these moral hazards, by such means as a formal quality inspection system, could be very significant when subcontractors are shrewd in exercising opportunism. At the same time, relational subcontractors are also not free from transaction costs. For example, if the parent firm refuses to accept the delivered supplies or requests a price reduction for a false claim on a defect in product quality, the subcontractor may face bank- ruptcy. Thus, whether relational contracting will come into common use will depend on how many risks and transaction costs are associated with this system relative to the verti- cal integration system, and how the risk and transaction costs are distributed between principal and agents. Conditions for Relational Contracting What, then, are the social and economic conditions under which the relational contracting system becomes relatively more advantageous, and thus more commonly used, than the vertical integration system? Technology and Commnodity Characteristics An overriding condition for use of relational contracting is the divisibility of a production process among accountable units-accountable in the sense that the inputs and outputs of one unit can be distinguished and measured separately from those of other units. This characteristic is often associated with scale economies that stem from the use of indivisible capital. For example, in the production of liquor for export, which requires a relatively large plant, it is difficult to break down the entire process into accountable components- material preparation, fermentation, distilling, bottling, and the like-that can be managed separately in different workshops. In general, production processes that flow continuously within a single plant, such as the production of chemical materials, are managed under the 8 Tozward the Rutral-Based Development of Contmerce and Itidurstry vertical integration system. In contrast, industries such as machine manufacturing and metalworking, which involve processing beyond the stage of raw material production (including low-level processing, such as rolling iron plates while pig iron from a furnace remains heated), can be broken down into many accountable components, and thus are more amenable to relational contracting systems. The demand for a commodity also has an important influence on the system chosen. One drawback of the relational contracting system is the relative difficulty of producing a large number of homogeneous products. Therefore, when there is a large demand for a commodity at a standard quality, the vertical integration system is likely to be preferable. In the early period of modern economic growth in Japan (from the late nineteenth to the early twentieth century), local traders who assembled woven textiles for domestic mar- kets commonly practiced the putting-out system. Traders advanced yarn to small cottage weavers, and executed contracts that provided for woven cloth to be collected by the trad- ers at a piece-rate service fee. In contrast, those who specialized in the collection of se- lected textiles for export markets that required assembly of a large amount of homogeneous products generally built their own workshops and employed their own workers. Factor Market Duialisnm Developing economies are often characterized by factor market dualism-labor wage rates are positively correlated with firm size, and large firms have much easier access to low- interest loans than small firms. Under this structure, there is a strong incentive for a large firm to contract work out to small firms rather than to hire labor within the firm to mini- mize wage costs. The large firm would thus be induced to make the effort to develop relational contract relationships with the small firms so that agents' moral hazards arising from information asymmetry could be effectively curbed. At the same time, the credits that large firms can mobilize at a relatively modest cost are often advanced to small producers, who suffer from severe capital constraints. One such example is the advance of materials for processing in putting-out contracts. These credit arrangements are commonly used to enforce assured commodity delivery and other con- tractual terms. Comnmuinity and Market The practice of creating "credit ties" represents one approach to strengthening enforce- ment of the relational contract by interlinking transactions. It is also common for the par- ent firms to provide technical assistance and other services to subcontractors. Related to this approach is the attempt to develop a long-term contractual relationship or to develop an expectation that a contract will continue to be repeated as long as contractual agree- ments are observed. Because establishing a stable, long-term contract covering many trans- actions entails significant time and effort, the expected gains from the maintenance of such a relationship are likely to exceed the short-term gains of exercising opportunism for both Toward an Alternative Path of Economic Development 9 contract parties. In this way, the likelihood of a "prisoner's dilemma" can be minimized (Hayami and Otsuka 1993). The power of long-term, interlinked contracts to prevent moral hazards can be strength- ened if they are embraced in a community relationship that includes not only the transac- tion of economic goods and services but also the exchange of personal favors and obligations. Anthropologists, sociologists, and political scientists commonly call such a relationship a patron-client relationship-"a special case of dyadic (two-person) ties in- volving a largely instrumental friendship in which an individual of higher socioeconomic status (patron) uses his own influence and resources to provide protection and/or benefits for a person of lower status (client) who, for his part, reciprocates by offering general support and assistance, including personal services, to the patron" (Scott 1972: 8). The ability of entrepreneurs to develop an efficient relational contracting system with low transaction costs while exploiting the merits of this system, such as risk-sharing and strong work incentives, may largely depend on the social tradition of communal relations. In a society where this tradition is strong and people are accustomed to patron-client rela- tions, relational contracting may be structured and operated relatively more efficiently than elsewhere. Rural economies in developing Asia may thus have a good chance of develop- ing efficient relational contracting arrangements based on their community tradition. It is important to recognize that the incorporation of such community relations into a relational contract does not weaken, but strengthens, the market mechanism. When there is asymmetry in information about product quality, the market is bound to be inefficient, or vanish altogether (Akerlof 1970). The effective use of community relations can be con- sidered an institutional innovation to correct this market failure (Hayami 1989). In this sense, the relational contract is a market institution and, in the absence of a market, it is likely to be replaced by the vertical integration system. This hypothesis is consistent with the organization of state enterprises in the former Soviet Union and other socialist econo- mies, which are geared toward vertical integration in its extreme. Governmenet Policies Unless there is competition in the market, relational contracting may turn out to be an instrument of monopolistic exploitation. Both economic efficiency and equity are endan- gered in the absence of a relative balance in bargaining power between the contracting parties. This balance is possible only when the parent firms compete with each other and when subcontractors have the option to shift from one parent firm to another, even if such shifts are rare. Therefore, government policies to limit market entry or to increase the advantage of large-scale operations-through concentration of import licenses and institutional credit in a few large firms, for example-tend to suppress the development of efficient relational contracting. Capital subsidization, using either direct subsidies or indirect subsidies such as tax exemptions and tariff escalation, which are usually captured disproportionately by large firms, will discourage decentralization based on relational contracting, thus encour- 10 Towvard the Rumral-Based Development of Conlmmerce an1d Inidu stry aging vertical integration (Little, Scitovsky, and Scott 1970). It might not be unreasonable to hypothesize that such government interventions are behind the failure of programs to develop cottage industry in some developing economies under the regime of import-sub- stitution industrialization. This hypothesis calls for empirical testing. At the same time, government services for small-scale enterprises, such as technical development and extension, dissemination of market information, legal consultation, and management training, will encourage decentralization by promoting the relative advan- tage of small-scale operations and by strengthening their bargaining position (Staley and Morse 1965). Above all, public investment in infrastructure such as rural electrification, transport, and telecommunication systems provide critical support for extension of the industrial and commercial network based on relational contracting. A Historical Perspective The theories and hypotheses advanced in the previous sections may be understood more concretely through a few examples from the history of modern economic growth in Japan. Developnient of the Putiting-Outt Systeni One interesting example is the development of the putting-out system in the textile indus- try. A popular presumption is that this system is a premodern form of industrial organiza- tion that is usually replaced during the course of modern industrialization bv the centralized workshop and hired-in labor of the factory system. In Japan, however, the putting-out system became common in textile weaving after the modem period of economic growth was initiated by Japan's opening to international trade in the late nineteenth century. A case study from Saitama Prefecture shows that before the Meiji Restoration (1868), farmers wove cloth from yam they spun from their harvested cotton; the woven cloth was collected by guild merchants (kablu nakama) in towns for transshipment to Edo (Tokyo) (Kandachi 1975: 87-100). The opening of international trade brought a large increase in domestic demand for cotton cloth, partly because of general increases in farm product prices and incomes and partly because of declines in the prices of cotton products brought about by import competition (Tanimoto 1987). This opportunity was exploited by rural- based traders outside the guild, who leased looms to farm households and advanced them imported yarn in piece-rate, service-fee contracts; the woven cloth collected by these trad- ers was mainly transshipped not to the traditional Edo market, but to remote rural areas such as Hokkaido and Tohoku (Kandachi 1975; Tanimoto 1986). This example seems to indicate that the putting-out system is an efficient mechanism for meeting dynamic demand expansion by mobilizing low-opportunity-cost labor. Farm fam- ily labor, idle between farm tasks and household chores, is particularly appropriate for this use, and carries a minimum labor-monitoring cost, in addition to other merits such as risk-sharing and low capital cost. Towvard an2 Alternative Path of Economic Development 11 This system continued to predominate in the Saitama area. Part of the reason may have been that community ties between rural-based traders and farmer-weavers worked as an effective brake on the latter's moral hazards, such as using less yam than required for a standard cloth quality or using a loom leased from one trader for production for other traders. In more urbanized areas, such as southern Osaka, these moral hazards were re- ported to be serious, and many traders moved to workshops based on hired-in labor (Abe and Saito 1987; Abe 1990). Such moral hazards are known to have been serious in the proto-industrialization stage of Western Europe (Landes 1969: 44). The continued dominance of the putting-out system in the Saitama area may be the product of the predominance of domestic markets as the destination for the striped cot- ton cloth woven in this prefecture, a market that did not require the quality standardiza- tion needed for exports. In contrast, it has been said that the factory system was preferred for the production of commodities for export, such as calico and glossy silk (hablitae). A major puzzle, however, is that the two major areas that produced glossy silk, Ishikawa and Fukui Prefectures, had very different organizations. In 1905, as many as 60 percent of the looms in Ishikawa were operated by cottage weavers and leased from traders under putting-out contracts; only 13 percent were under contracts of this kind in Fukui (Kandachi 1975: 138). The technological progress represented by the development of power looms appears to have had significant impact on the system chosen. With the adoption of power looms, local traders who had formerly leased out looms to cottage weavers generally moved to operate the machines in their own shops, but this did not bring about the demise of the putting-out system. These trader-weavers usually entered putting-out contracts with larger traders that included clauses on the provision of trade credit. These rela- tionships of small local traders with large traders had been common even when the small traders were principals in putting-out contracts with farmer-weavers. Such a hierarchy of traders can also be observed in developing economies today (Hayami and Kawagoe 1993). It is important to recognize that the development of power looms did not significantly increase the optimum scale of operation. Until the beginning of this century, when auto- matic looms were based on water and steam power, their installation required relatively large factories. As electrification progressed and small motors and low-cost power looms became available, however, even small weaving houses began to introduce power looms (Minami, Ishii, and Makino 1982). A case study of a village in Fukui Prefecture showed that the weaving houses that en- tered the business in the 1920s, when electricity became available, were generally much smaller than the houses that had begun work earlier. Correspondingly, these new enter- prises were initiated by middle-class peasants with relatively small land-asset holdings; the older businesses were operated predominantly by landlords and upper-class peasants (Kandachi 1975: 222-68). This example clearly indicates that modem industrial technol- ogy is not necessarily biased in favor of large-scale operations, but can be developed to increase the relative advantage of small manufacturers. 12 Towvard the Rutral-Based Development of Comnnerce anld Indilstry The Origins of Modern Suibcontracting Of major interest for our study is how the traditional putting-out system that was orga- nized mainly by local traders in the early stage of modem economic growth has been transformed into the modem subcontracting system of Japanese industry today. The his- tory of Japan clearly shows that the putting-out system is not limited to the stage of devel- opment characterized by factor market dualism and low-capital-intensive technology, as hitherto assumed (Shinohara 1968). The weaving industry in Ishikawa and Fukui, for ex- ample, continues to be dominated by small and medium-size enterprises and is pervaded by electronic control and other high technologies, although factor market dualism disap- peared in the mid-1960s. These enterprises, which today are engaged mainly in weaving chemical textiles, are organized in a putting-out system that is essentially the same as the system that evolved before the era of high economic growth beginning in the mid-1950s. Putting-out contracts in Ishikawa and Fukui are now organized not only by local trad- ers but also by large chemical fiber manufacturers and large trading houses (s6g6 sh6sha) based in Tokyo and Osaka. Optimal combinations of contracting parties and contractual forms are deliberately chosen to best accommodate the production of different products for a variety of markets (Itoh and Urata 1993). The large chemical companies have pre- ferred this system to vertical integration, not because of factor market dualism, but be- cause of such advantages as low labor-monitoring costs, strong work incentives for management, and flexibility in employment and staffing. These observations are consistent with the hypothesis that the putting-out system is not an organizational form destined to be replaced by the factory system in the process of modernization (Pollard 1965; Saito 1984). Rather, the advantage of this system seems to increase as the economies advance to a stage characterized by high per capita income and a diversification of consumer demands. Indeed, in Japan today, the putting-out system for weaving chemical textiles is only one of many variations of the subcontracting system practiced by modem high-technology industries, such as automobiles and electronics. These industries have developed highly sophisticated, long-term, and interlinked contracts be- tween an assembler and many parts suppliers that minimize moral hazards and opportu- nistic behavior between the contracting parties. One major force commonly used by a parent company to control subcontractors is the establishment of quasi-community rela- tions of the patron-client model. At the same time, other devices even stronger than mar- ket competition have been designed to enhance competition within a relatively closed corporative group (keiretst) (Asanuma 1985, 1988; Wada 1991). With these efforts, coordination between a parent company and subcontractors has been made extremely precise, as can be seen in Toyota's famous just-in-time (kanban) system. The improved subcontracting system is now considered a major organizational innovation that underlies the strength of Japanese industries (Abegglen and Stalk 1985; Aoki 1988). One central question addressed in this volume is how much of this modem subcontract- ing system can be explained as an inheritance from the traditional putting-out system and how much is an innovation in response to modem industrial needs. The likely answer is that Torward an Alternatizve Path of Economic Devielopmlienzt 13 these two factors have reinforced each other to create the present form. Thus, our investiga- fion must focus on the mechanism and the process of their interactions. It is especially im- portant to our study to identify the role of traditional community relations. It is said that industrial subcontracting is more tightly and effectively organized in Japan than in Europe and North America. How much of this modern success in subcontracting is based on the intensive community relations nurtured in the rice-based village communities that relied heavily on communal gravity-irrigation systems in the mountainous topography of Japan, which required strong cooperation and tight coordination among villagers? In one of the hypotheses that considers modern subcontracting a relatively recent prod- uct of management efforts rather than the result of historical tradition, it has been argued that the current, highly sophisticated, and tightly structured subcontracting system has its roots in government guidance under the Resource Mobilization Program during World War II (Okazaki and Okuno-Fujiwara 1993, chapter 1). This hypothesis is plausible, con- sidering that close coordination between assemblers and parts suppliers was critical to the assured supply of high-quality parts needed to produce weapons and other strategic goods that required great precision. Nevertheless, a major puzzle remains. Why did the-government try to strengthen the subcontracting system during the war, rather than consolidating assemblers and parts suppliers into a vertically integrated production unit of the Soviet model? The Resource Mobilization Program was designed, in many respects, in accordance with the methodol- ogy of Soviet central planning. Is it not unreasonable to hypothesize that social traditions in Japan made decentralized organizations based on relational contracting more efficient relative to vertical integration, even in the autocratic regime during the war? The main task of this project, therefore, is to investigate the mechanism and process of the mutual reinforcement exerted by institutional traditions, modem management efforts, and government guidance that has produced an optimum industrial organization in each stage of economic development. For this task it is indispensable to identify the entrepre- neurs who have accomplished such institutional innovations and how much potential they have for reaffirming the rural-based development paradigm. Major Findings of the Case Studies In order to approach this problem, eight case studies were undertaken, and the results are reported in this volume. The major findings of these individual studies are summarized here in their contributions to the research goal. The three chapters following this introduc- tion are case studies on the historical experiences in Japan. Next, five chapters deal with current developments in East Asia. Chapter 2: Kawagoe on Rice Marketing in Prewar Japan Kawagoe's study is focused on the impact of railway development on rice-marketing orga- nizations. Popular views, such as "populism" in the United States, have presumed that the 14 Tow7ard thte Ru ral-Based Development of Coimnerce and Indishfry modem technologies represented by the railways were used by large capitalists to exploit small farmers, artisans, and petty traders. Contrary to this presumption, Kawagoe demon- strates that the construction of the railway network in the early Meiji period gave rise to active participation by peasant entrepreneurs in rice marketing, which had been monopo- lized by large, privileged merchants in port cities since the feudal Tokugawa period. The high risks and large capital requirement associated with shipping rice by sea had made it difficult for small traders to engage in rice marketing across regions. When railways re- moved this constraint, a large number of small traders, many of whom came from the middle peasantry, began to cut into the business of shipping rice to metropolitan areas. Further, growing competition in the nationwide market, which was integrated by national unifica- tion and the railway network, induced group actions at a local level to establish institutions that could reduce transaction costs through such measures as grading and use of tradable rice drafts. The system worked efficiently, with little government intervention in the market. Kawagoe clearly shows that a large reserve of entrepreneurial talent does exist in rural ar- eas, and it can be mobilized with adequate support from public infrastructure and govern- ment avoidance of direct market intervention. The high responsiveness of rural people to new marketing opportunities demonstrated in his study represents a sharp contrast to the traditional image of peasants in developing economies, which granted them neither the will- ingness nor the capability to undertake commercial and industrial enterprises. Chapter 3: Itoh and Tanimioto on Cotton Weaving in Japan Based on a detailed historical analysis of the Iruma area, Itoh and Tanimoto provide strong evidence in support of the hypothesis that the putting-out system practiced by rural entre- preneurs was an efficient form of relational contracting, well adapted to meet the increased market demands for cotton cloth brought about by national unification and the opening of international trade in the early Meiji period. The advantage of this system in mobilizing rural labor, with its low opportunity costs, through such measures as employing females in farm households during the off-season in farming, is clearly illustrated. Of particular interest is the description of highly complex, intertwined financial flows corresponding to commodity flows among rural-based weaving entrepreneurs and town-based wholesal- ers engaging in shipments to distant markets. Many financial transactions were made with drafts and money orders that were purely private and had no official endorsement. The efficiency in relational contracting between traders and weavers should have been sup- ported by community ties in a relatively narrow area of high population density. The au- thors' hypothesis that high population density was a major factor in the persistence of the putting-out system during the period of modem economic growth in Japan, compared with its early demise in England, has important implications for the applicability of this system to developing economies, and thus needs further verification. Itoh and Tanimoto illustrate several interesting examples of how the relational contract arrangements used in Iruma in the early Meiji period are still being used by modern textile enterprises in Japan as a basis for their organizational efficiency. Towvard an Alteriative Patlh of Economic Development 15 Chiapter 4: Wada on Toyota's Relationship with Suppliers Toyota's subcontracting relationship with parts suppliers can be considered a prime ex- ample of the successful application of relational contracting to modem industry. Wada traces the evolution of this relationship from the beginning of Toyota's automobile pro- duction in 1933. It is interesting to learn that Kiichiro Toyoda, the founder of the Toyota Motor Company, intended to build "a pastoral factory" in Koromo (today's Toyota City), which was then a typical rural town with few industries. Kiichiro's idea was not only to locate the factory in a pastoral environment, but also to build a production organization in which Toyota's assembly plant would be surroLnded by trusted parts suppliers who were tied together by a common community spirit. To this end, Toyota tried as much as possible to enter contracts with manufacturers located near the Koromo factory who were recep- tive to Toyota's guidance, even if they had no previous experience in the precise, sophisti- cated work needed for automobile production. Kiichiro's idea of the "pastoral factory" may have stemmed from his knowledge of the applicability of community relationships to industrial production, which he may have learned through experience as a supplier of automatic looms for rural-based weaving entrepreneurs-before he began automobile manu- facturing. In any case, it is fascinating that the traditional community relationship became a guiding principle for a modem big business such as Toyota. Chapter 5: Kikitchi on Garnlent Manufacture in the Phiilippines Based on the wealth of information collected from his own field survey, Kikuchi identifies the mechanism that allowed the putting-out contract to become an effective instrument for mobilizing rural resources to meet the rapidly expanding demands of garments for export. The contract is customarily closed between an urban-based wholesaler (called "ex- porter"), who collects garments of a given standard for foreign buyers, and a rural entre- preneur (called "subcontractor"), who has a small workshop for garment processing. The subcontractors are generally women from the rural middle class who have some educa- tion, but no special skill in garment making. The case reported by Kikuchi represents an extreme of how far the division of labor can be pushed within the putting-out system: the process of garment production is subdivided into many simple tasks, such as cutting, sew- ing, and button placing, so that work efforts are easily measurable for piece-rate pay- ments. This system seems to show that a division of labor to the high degree illustrated by Adam Smith's famous example of a pin factory can be implemented in a decentralized manner through coordination across small workshops. In addition to this device for the effective use of piece-rate payment incentives, the urban principals try to strengthen per- sonal ties with reliable subcontractors by giving favors. Similar efforts are made by sub- contractors on behalf of their workers. The most important qualification for subcontractors is the development and maintenance of good relations with both exporters and workers. These ties allow the subcontractors to secure a steady flow of orders from the former and conscientious work from the latter. An interesting finding is that a majority of these sub- 16 Towvard thte Rutral-Based Development of Commerce anid Itiditstry contractors have prior experience in petty trading, which seems to support the hypothesis that commerce is the training ground for industrial entrepreneurship. Chapter 6: Ohno anid Jirapatpinmol on thle Garnment and Weaving Indulstries in Thailand Rural-based developments in the garment and weaving industries in Thailand have many similarities to those seen in the Philippines. Work is carried out mainly by women who work as subcontractors of urban wholesalers or factories. The women who act as rural entrepreneurs in the garment industry usually belong to the lower segment of village com- munities and used to be sewing workers in urban factories, whereas those in the weaving industry came from the upper segment. A very important finding is that putting-out con- tracts are common in the garment sector, whereas advance-order contracts with no provi- sion by the principal of materials for processing are used in the weaving sector. This difference is explained largely by the much stronger reliance of garment subcontractors on urban principals for the supply of external market information on the specification of prod- ucts and, hence, materials. The difference in the ability to finance working capital between garment and weaving entrepreneurs, however, could well be another significant reason. The comparative advantage of rural over urban entrepreneurs in the management of rural labor through community relationships is certainly a crucial factor for the urban entrepre- neurs who contract work out to rural entrepreneurs. Another major reason is the inappli- cability of minimum wage laws and other labor regulations to small rural enterprises, which makes their effective labor costs much lower than those of large urban factories, as demonstrated clearly in the Thai case study, as well as in the Philippine study by Kikuchi. Chapter 7: Liui and Otsuka on Village-Township Enterprises in China This study focuses on the contract mechanism that enables village-township enterprises in postreform China to be associated with state enterprises and foreign businesses so that they can engage in garment production for export. Despite progress in economic liberal- ization, foreign trade is still largely monopolized by state exporting companies, which are difficult for village-township enterprises to access directly. Their export-oriented produc- tion thus must take the form of subcontracting with state enterprises or joint ventures with foreign firms. A major advantage for state enterprises that enter a relational contract with village-township enterprises is the escape from the high labor costs created by the strong welfare program in the state enterprises. The advantage for the village-township enter- prises is that they receive transfers of technical information and trained manpower. A distinctive feature in the organization of the garment industry in China is that enterprises are very large-not only state enterprises but also village-township enterprises are much larger than those usually found in market economies such as Japan, the Philippines, and Thailand. This comparison seems to support the general hypothesis that the absence of Towvard an Alternative Path of Economic Development 17 market promotes vertical integration. A major puzzle is why the village-township enter- prises do not practice subcontracting with small workshops and households, as is com- mon among rural entrepreneurs in the Philippines and Thailand. Chapter 8: Lee and Siuh on Rutral Enterprises in the Republic of Korea Lee and Suh try to show why, despite the remarkable success of industrialization in Korea, rural-based industries have largely failed to develop, especially in comparison with the record of Taiwan, China. Part of the failure may be explained by an insufficient allo- cation of public resources to the rural sector. Increased government support in recent years, however, has not yielded significant results. Government support has included such measures as government provision of needed infrastructure for rural industrial es- tates and generous credit and tax treatment for the establishment of factories, which are often underutilized. Lee and Suh made a field survey in a local township where rural entrepreneurs have been exceptionally active in promoting the weaving industry for both urban and export markets through de facto relational contracting with urban wholesalers. In this exceptional case, the local entrepreneurs made good use of the industrial estate provided by the government. In contrast, several neighboring estates that invited urban entrepreneurs to establish factories do not show comparable success. Lee and Suh con- clude that in order to achieve genuine rural industrialization, it is necessary to promote local rural entrepreneurs rather than to transfer urban enterprises to rural sites. The major question to be answered is, of course, how to mobilize the entrepreneurship rooted in the rural areas. Chapter 9: Lane on Comparisons Between the Repuiblic of Korea and Taiwan, China Lane offers a political science perspective on the major contrast in rural industrialization between the Republic of Korea and Taiwan, China. First, he documents the major dispari- ties in public infrastructure in rural areas to explain the different performances in rural industrialization between these two economies. He then tries to explain the differences through their polities. In Korea, the autocratic, centralized governing structure is consid- ered to have hindered the allocation of public resources to rural areas. In particular, the system of appointment of key government officials in local administrations by the central government instead of their election has closed the channel that conveys the demands of local people in the allocative decisions concerning public resources. This Korean system is contrasted with Taiwan, China's policy of electing local administrators. It is plausible to identify this difference in local administrative structure as a major factor in the perfor- mance of rural-based industries. Yet it remains to be explained how crucial this factor is in comparison with many others, including initial conditions such as the colonial develop- ment strategies of Japan, the destruction of the Korean War, and the policies of the govern- ments following independence, especially with respect to foreign direct investments. 18 Tovard the Rural-Based Developmiienit of Conmmerce antd Inidustry Conclusions These case studies provide more questions than answers, and that is what we intended. Because of the highly elusive nature of informal economic activities, our investigations were necessarily based on intensive personal observation in the field, as well as the exami- nation of microhistorical data, rather than fornal questionnaire surveys. Findings from such case studies are bound to be more illustrative (or even anecdotal) and suggestive than they are analytical and conclusive. Yet comparisons of our case study results can produce many interesting empirical regularities, a basis for the construction of meaning- ful hypotheses. Indeed, Ohno and Kikuchi's comparison of the five case studies of textile industries (chapter 10) produced plausible propositions on the choice of contract in the course of integration of local markets into wide national and international markets. It may not be an unreasonable expectation that the efforts to replicate the case studies further, both within and beyond East Asia (as suggested by Bhatt in chapter 11), as well as to process their results for more formal testing of hypotheses, will identify the conditions that will allow developing economies to be guided into the "alternative path of develop- ment." It is hoped that the studies in this volume will be a step toward such a research accomplishment. References Abe, Takeshi. 1990. "Menkigy6" ("Cotton Textile Industry"). In Shunsaku Nishikawa and Takeshi Abe, eds., Sangyoka n1o jidai (The Era of Inidiustrializationi). Vol. 4, Nihoni keizai-shi (Japaniese Eco010noic History). Tokyo: Iwanami Shoten. Abe, Takeshi, and Osamu Saito. 1987. "Chinbata kara rikishokki k8j6 e: Meiji koki ni okeru menorimonogy6 no baai" ("From Fee-Contract Looms to Power Loom Factories: The Case of Cotton Textile Industry in the Late Meiji Period"). In Ryoshin Minami and Yukihiko Kiyokawa, eds., Nihoni n1o kogy6ka to gijuttsnt shinipo (Industrializationi anid Techntological Development in1 Japanl). Tokyo: T6y6 Keizai Shinposha. Abegglen, James C., and George Stalk, Jr. 1985. Kaisha: The Japaniese Corporation. New York: Basic Books. Akerlof, George A. 1970. "Market for Lemons: Quality Uncertainty and the Market Mechanism." Quiarterly Jouirnal of Economics 84(3): 488-500. Alchian, Armian, and Harold Demsetz. 1972. "Production, Information Costs, and Economic Orga- nization." Americani Economic Reviewv 62(5): 777-95. Aoki, Masahiko. 1988. Information, Incentives and Bargaining in the Japanese Economy. Cambridge, U.K.: Cambridge University Press. Asanuma, Banri. 1985. "Organization of Parts Purchases in the Japanese Automobile Industry." Japanese Economic Stuidies 13(1): 32-53. . 1988. "Manufacturer-Supplier Relationships in Japan in the Concept of Relation-Specific Skill." Journal of the Japanese and In tern1ational Economies 3(1): 1-30. To7vard an Alternative Path of Economlic Developnient 19 Chuta, Enyinna, and Carl Liedholm. 1979. "Rural Non-farm Employment: A Review of the State of the Art." MSU Rural Development Paper No. 4. East Lansing: Michigan State University, Depart- ment of Agricultural Economics. Coase, Ronald H. 1937. "The Nature of the Firm." Economica 4(16): 386-405. Hayami, Yujiro. 1989. "Community, Market, and State." In Allen Maunder and Alberto Valdes, eds., Agricultuhre and Goverinmenits in an Interdepentdent World. Aldershot, U.K.: Dartmouth. Hayami, Yujiro, and Toshihiko Kawagoe. 1993. The Agmrian Origins of Commerce and Indutstry: A Sturdy of Peasanit Marketing in Indoniesia. London: Macmillan. Hayami, Yujiro, and Keijiro Otsuka. 1993. The Economics of Contract Choice: An Agrarian Perspective. Oxford, U.K.: Clarendon. Itoh, Motoshige, and Shujiro Urata. 1993. "Small and Medium Enterprise Support Policy in Japan." University of Tokyo, Faculty of Economics. Photocopy. Kandachi, Haruki. 1975. Meiji-ki n6son orimoniogyO no tenkai (Development of the Rutral Textile Indutstry in the Meiji Period), 2d ed. Tokyo: Ochanomizu Shobo. Kilby, Peter, and Carl Liedholm. 1986. "The Role of Nonfarm Activities in the Rural Economy." EEPA Discussion Paper No. 7. Washington, D.C.: U.S. Agency for International Development. Kiyonari, Tadashi. 1980. Chfish kigyo dokuthon (Textbook of Small and Medium-Size Indutstries). Tokyo: T6y6 Keizai Shinp6sha. Landes, David S. 1969. The Unbountid Promethieuts. Cambridge, U.K.: Cambridge University Press. Little, Ian, Tibor Scitovsky, and Maurice Scott. 1970. Industry and Trade in Some Developing Countries. Oxford, U.K.: Oxford University Press. Minami, Ryoshin, Tadashi Ishii, and Fumio Makino. 1982. "Gijutsu fukyQi no sh6j6ken: rikishokuki no baai" ("Conditions of Technology Diffusion: The Case of Power Looms"). Keizai kenkytl 38(3): 216-30. Okazaki, Tetsujiro, and Masahiro Okuno-Fujiwara, eds. 1993. Gendai Nihon keizaishisutenmiu nogenryfl (The Origins of the Contenmporary Econonmic System of Japan). Tokyo: Nihon Keizai Shinp6sha. Oshima, Harry T. 1986. "Off-Farm Employment and Incomes in Postwar East Asian Growth." In Richard T. Shand, ed., Off-Farm Employnment in the Development of Rural Asia. Canberra: Australian National University National Center for Development Studies. Patrick, Hugh T., and Thomas P. Rohlen. 1987. "Small-scale Family Enterprises." In KozoYamamura and Yasukichi Yasuba, eds., The Political Econonmy of Japan. Vol. 1, T7he Domestic Transformation. Stanford, Calif.: Stanford University Press. Pollard, Sydney. 1965. The Genesis of Modern Management: A Study of the Indutstrial Revoluttion in Great Britain. London: Arnold. Saito, Osamu. 1984. "Zairai orimonogyO ni okeru k6j6sei k6gy6ka no sh6y6in" ("Factors Underly- ing the Factory-System Industrialization in Japan"). Shakai keizai shigaku 49(6): 114-31. Schumpeter, Joseph A. 1961. The Tlheory of Economic Development, R. Opic, trans. New York: Oxford University Press. Scott, James C. 1972. "The Erosion of Patron-Client Books and Social Change in Rural Southeast Asia." Journal of Asian Studies 33(1): 5-37. 1976. The Moml Economy of the Peasant. New Haven, Conn.: Yale University Press. Shinohara, Miyohei. 1968. "A Survey of the Japanese Literature on Small Industry." In Bert F. Hoselitz, ed., The Role of SnIall Industry in the Process of Economic Development. The Hague: Mouton. Smith, Thomas C. 1956. "Landlords and Rural Capitalists in the Modernization of Japan." Jouirnal of Econonmic History 16(2): 165-68. 20 Tozward the Rutral-Based Dezelopment of Coninnerce anid Indlustry - 1959. The Agrarianr Origins of Modernt lapani. Stanford, Calif.: Stanford University Press. . 1960. "Landlords' Sons in the Business Elites." Economiiic Developmentt and Cultural Chiange 9(1): 93-107. 1988. Native Souirces of Japanese Inidutstrializationi, 1750-1920. Berkeley: University of Califor- nia Press. Staley, Eugene, and Richard Morse. 1965. Modern Sm1?all-Scale Indiustry for Developin?g Couintries: Evi- dencefrom? Gha1na. New York: Praeger. Tanimoto, Masayuki. 1986. "Bakumatsu-Meiji zenki menorimonogyo no tenkai: Saitama-ken Iruma- gun o chushin ni shite" ("Development of the Cotton Textile Industry in the Late Tokugawa and Early Meiji Period: Centering on Iruma County in Saitama Prefecture"). Shakai keizai shigakut 52(2): 151-84. - 1987. "Bakumatsu Meiji-ki menpu kokunaishijo no tenkai" ('Expansion in the Domestic Market for Cotton Cloth in the late Tokugawa-Meiji Period"). Tochi seido shigaku 29(3): 54-67. Wada, Kazo. 1991. "The Development of Tiered Inter-firm Relationship in the Automobile Industry: A Case Study of Toyota Motor Corporation." Japan?ese Yearbook on1 Butsiness History (August): 23-47. Williamson, Oliver E. 1975. Market and Hierarchlies: Analysis and Antitrust Implications. New York: Free Press. 1985. The Economic Institutions of Capitalism. New York: Free Press. Part I. Historical Experiences of Japan 2 Technical and Institutional Innovations in Rice Marketing in Japan Toshihiko Kawagoe T- he marketing of agricultural commodities plays an important role in the econo- T mies of developing countries with dominant agricultural sectors. Limited access to knowledge and information about marketing activities, however, often hampers planners and policymakers in these countries in their efforts to set up appropriate devel- opment strategies. It is widely believed that, in developing countries, monopoly or usury by traders is common and the market is inefficient. The traditional, but still influential, pessimistic view is that peasants in developing countries have neither the capabilities nor the willingness to undertake commercial and industrial enterprises (Boeke 1953; Geertz 1963). Such a view often rationalizes improper government intervention in the market. Therefore, it is critical to better understand the nature, roles, and mechanisms of the mar- keting system. It is also important to identify possible roles for marketing agents in the process of economic development.' Several important questions arise in this context. How can a marketing system be gener- ated in the early stages of economic development? When technological and economic en- vironments change, how is the market system transformed and how do entrepreneurs respond to the change? What kinds of entrepreneurs appear and decline? What sorts of roles can the government play in forming a modern marketing system? In order to answer I gratefully acknowledge helpful comments from the other participants in this project, anonymous referees, as well as the insights of Kunio Ohokama and Yoshiro Miwa. 23 24 Towvard the Rutral-Based Developmiient of Commnerce and Ibdustry these questions, this chapter analyzes the rice-marketing system in Japan, with special emphasis on the prewar period. ("Prewar" in this chapter refers to the era from the Meiji Restoration of 1868 to the 1930s, the beginning of the Pacific War.) Information on these issues is highly applicable to developing countries and transition economies, where mar- ket formation is critically important to economic development. The prewar period can be divided into three phases, according to the nature and basic structure of the rice-marketing system: (1) the early Meiji period (1868 to around 1890), which was a transitional period from the traditional to the modem system; (2) the era of marketing innovation from the mid-Meiji to the Taisho period (about 1890 to the 1920s), when the structure of the marketing system for rice changed drastically in the course of modern economic growth; and (3) the Taisho to the early Showa period (1920s to 1930s), when government intervention in the market was introduced and the market was gradu- ally suppressed. In addition, a brief investigation into the rice-marketing system of the late Tokugawa period (before 1868) may offer a useful comparison. Overview of Prewar Japanese Agriculture Japanese agriculture has traditionally been carried out by the peasantry and was com- prised of a large number of small family farms engaged in rice farming. This tradition, inherited from the Tokugawa period (1600-1868), was retained during the period of mod- em economic growth after the Meiji Restoration. Throughout the postwar period, the num- ber of farms in Japan remained roughly constant, at about 5.5 million households. The agricultural land area totaled about 6 million hectares, of which half was wet paddy fields, while the rest was cultivated upland. Accordingly, the average operational land area of each farm was very small, approximately 1 hectare. Labor was relatively abundant and land was the major constraint on agricultural pro- duction. Efforts at agricultural development were geared to facilitate the substitution of labor for land. An intensive rice culture was nurtured under a well-developed infrastruc- ture. A public agricultural research and extension system was established by the end of the nineteenth century. An industrial sector capable of supplying fertilizer and chemicals was also developed. Agricultural output grew at an annual compounded rate of about 1.6 percent in the late nineteenth century, which accelerated to 2.0 percent from the beginning of the twentieth century until the 1920s (Hayami and others 1991: 15-19). Although the Meiji government attempted to catch up with Western nations through industrialization, agriculture remained a major sector in the prewar period. Primary industry accounted for about 70 percent of the labor force and about 40 percent of the net domestic product in 1880 (Hayami others 1991: 14). Gross output from rice produc- tion represented 55 percent of total farm output in 1880 (see table 2-1). Although the share of rice in total output had gradually declined to 42 percent by 1930, rice continued to be dominant in Japanese agriculture throughout the prewar period. Rice was also the Tech1nical antd Institu(tional Itnovations in Rice Marketing in Japan 25 Table 2-1. Rice Production, Import, and Consumption in Prewar Japan Year 1880 1890 1900 1910 1920 1930 Production (million yen)a Agriculture 327 398 468 584 719 767 Rice 181 209 227 270 315 323 Share of rice (percent) 55 52 48 46 44 42 Rice (thousand metric tons)" Production 5,093 5,861 6,372 7,588 8,838 9,070 Import 4 96 298 442 936 1,462 Consumption per capita (kg/year) 117 144 147 160 169 154 Note: Data refer to five-year averages. a. Farm value, 1874-76 prices. b. Brown rice equivalent. Source: Japan, Ministry of Agriculture, Forestry, and Fisheries, Food Agency (1990: 448-55); Umemura and others (1966: 148-49). most important staple in the Japanese diet. In the early 1930s, 61 percent of the daily intake of calories and 44 percent of the daily intake of protein came from rice (Kayo 1977, tables J-c-3 and J-c-4). Until shortly after the Restoration, self-sufficiency in rice had been maintained. Acceler- ated population growth,2 however, coupled with increased per capita consumption, soon forced the country to import rice. In the 1870s, annual per capita consumption was about 100 kilograms (milled rice), which increased to 130-150 kilograms by the end of the nine- teenth century and peaked at 170 kilograms by the mid-1910s. Accordingly, the share of imported rice gradually increased from around 1 percent of total consumption in the late nineteenth century to about 10 percent in the 1930s (table 2-1). In most Asian countries, farmers sell rice to traders in the form of paddy-that is, rough, unmilled rice. It is then processed into polished rice by rice mills for further marketing. The function of the rice mills is not limited to processing, but also includes storage and shipment. The rice mills thus play a key role in the marketing of rice. In contrast, in Japan the paddy is immediately husked by farm producers and sold to traders in the form of brown rice; it is stored and transported in this form. Brown rice is milled into polished rice in the final stage of marketing, generally by urban wholesalers or retailers. Thus, the major distinguishing characteristic of rice marketing in Japan is the absence of rice mills in the countryside and small towns. It is not clear why rice marketing is done in brown rice only in Japan (as well as the Republic of Korea and Taiwan, China). Paddy stores better than brown rice, and polished rice is unsuitable for storage because it deteriorates rapidly. At the same time, paddy is much bulkier than brown or polished rice. The cool climate of Japan may have allowed brown rice to be stored or shipped without major deterioration in quality, while storage and transport capacity were conserved. 26 Toward the Rufral-Based Developmenrt of Comnierce and Ilndutstry The Traditional Rice-Marketing System in Japan The Feudal Tokugawa Period (before 1868) In the Tokugawa period, rice occupied an overwhelmingly important position in the economy. It was not only the major staple in the Japanese diet and the most dominant crop in production, but also a major source of income for the government. Land was administered by the Tokugawa shogunate and numerous daimyo. Their financial base rested largely on the rice tax, which was levied in kind from the peasants. In principle, peasants were bound to their daimyo and to their land. Landlord-tenant relations were prohibited. Later, as intensive cultivation based on small family farms became more efficient than large farms with attached labor, a de facto landlord-tenant system was developed between legal peasants and attached laborers. This process was accelerated when the land tax system shifted from variable levies based on crop yield assessments (keni) to fixed levies in kind (jnien), and daimyo became less eager to enforce tenancy regulations (Hayami and others 1991: 61-63). The political and military power of the daimyo was measured by the taxable capacity of the rice, known as kokudaka, obtained from their domains. Each daimyo had his own army, which was based in the castle town of his domain. Although the Tokugawa shogunate had some power to control the daimyo, each daimyo retained sovereignty in his domain. Com- munication between domains was often limited, and agricultural technologies and im- proved crop varieties were not easily transferred among regions. In times of crop failure, daimyo often imposed embargoes on the shipment of rice.3 As discussed below, although urban-to-local marketing was well developed in the late Tokugawa period, the interre- gional marketing network did not yet function well. Rice marketing in the late Tokugawa period was characterized by the dual structure of marketing channels (figure 2-la). The major route was the channel for government-con- trolled rice. Part of the rice collected by a daimyo was paid to his retainers as in-kind salary, and the rest was sold on the market to raise cash revenue. The local markets in the daimyo's domain, however, were not always large enough to absorb the supply. There- fore, bulk shipments were sent to urban markets for sale by the daimyo or by privileged merchants. They shipped rice to selected consignees or to warehouses owned by the daimyo in the urban market. Throughout the Tokugawa period, rice production increased steadily. In 1600 the aver- age yield of rice per hectare was 1.4 metric tons of brown rice; total production was about 3 million tons. By the end of the period, the yield had increased to about 2 tons, and total production reached 6 million tons (Nakamura 1968: 168-73; Hayami and Miyamoto 1988: 44). At the same time, rice tax rates and the taxable standard of the rice yield changed only gradually, especially after fixed levies in kind were imposed. As a result, peasants could retain the increase in production, which enabled them to sell rice through commercial channels. Part of the rice sold by peasants was consumed locally and part was shipped to urban markets by local traders. Techntical and Institutional Innovations int Rice Marketing in Japan 27 Figure 2-1. The Traditional Rice-Marketing System in Japan (a) The Late Tokugawa Period (before 1868) Urlban wholesale market :< . ... ... ... ... ... ... ... ... ....................... Consignee Sailboats Overlord (Huidasashi/Beikaisho) SibasOelr i, : Sailboats/ Local port ...... horseal : Urban trader < .: SLtcl trad Vassals (Beikokni don ya) (Sanchi donya) I ~~~~~. . . . . . . . . . . . . . . .. .. . . . . . . . . . . . . . . . . . . . . . ..... ....... .... Resalers ................ - (Waki ten) * Collectors 1 > (Nakagai) < Farmers Retailers Retailers .. .- Urban consumers Local consumers < Commercially marketed rice < Government-controlled rice (b) The Early Meiji Period (1868 to around 1890) Urbant wholesale market Local seaport . Consignee: (Kaimai donya) < Local trader (Saiiclii donya) Landlords I ~~Sail/steam-* ' ~~~boats '--. Resalers b v (Waki ten) . - Farmers Collectors( Tenant-farmers Retailers (Nakagai) < Owner-farmers Retailers ............... Urban consumers Local consumers Although few quantitative data are available at the national level on rice production and distribution in this period, we can draw a conjectural sketch (Miyamoto 1988: 2-25).4 In the early nineteenth century, the total population of Japan was 31 million; 85 percent were members of the peasantry and 6 percent belonged to the warrior class. Rice produc- tion was 27 million koku (about 4 million metric tons in brown rice), of which 27 percent, 7.2 million koku, was collected by daimyo as rice taxes.5 They consumed 2.2 million kokui, and the rest was sold to city dwellers through commercial channels. The peasants con- 28 Torvard the Rtural-Based Development of Commerce and Indutstry sumed around 18 million koku and supplied 1.4 million kokui to consumers through private marketing channels. Thus, after deducting 1.4 million koku for marketing and postharvest losses, the total amount of commercially marketed rice was estimated to be 5 million koku, or around 18 percent of total production. It should be noted that this 18 percent share shows the lower bounds; it is assumed that warriors and peasants did not purchase rice from the market. Rice was normally shipped in large sailing vessels with capacities of at least 300 to 500 koku.6 Coastal transportation by chartered vessels required full loads of rice, which only large traders could afford. It was a slow and risky operation-rice often deteriorated in quality, and there was a chance that prices might decline before the destination was reached. The risk of shipwreck was also high.7 Long-distance land transportation, however, which allowed only 2 bags (=120kg) of rice to be carried on each pack animal, was prohibitively costly because of poor road conditions. Let us consider, for example, the route for rice shipped from Sakata. One of the major ports in northern Japan, Sakata was a marketing center for rice produced in the Shonai plains and other neighboring regions. Rice collected in Sakata was transported by boat to Osaka by the westbound coastal route. It was then shipped to Edo, which took more than two months (see figure 2-2) (Yamagata-ken Sanmai Kairy6 Ky6kai Reng6kai 1958: 6). The eastbound route was also occasionally used, but it involved a higher risk of shipwreck. Figure 2-2. Major Cities and Coastal Transportation Routes for Rice from Sakata to Osaka/Edo in the Late Tokugawa Period <:SK-,i f AtWf .XW;:: ::577 Technlical and Inistituttioial innovations in Rice Marketing in Japan 29 The strong economies of scale and the high risk of transport by ship were critical factors in the organizational structure of the rice-marketing system in this period. It was an effi- cient system, given the prevailing physical and economic environments. Bulk shipment by the daimyo or their agents, who were large merchants with sufficient financial resources to take the risk, was considered an efficient system that minimized the costs under the resource constraints of the time (Miyamoto 1988: 436). The Tranisitional Period (1868 to about 1890) After 1868 the new Meiji government dismantled the system of feudal daimyo and built a new, centralized administrative system. Drastic institutional reforms were adopted by the Meiji government. The traditional rice tax was replaced by a modern land tax based on land value and paid in cash, and interprovincial trade and communication were expanded. The reforms caused significant changes in local and interregional marketing systems. In the early Meiji period, however, the marketing system for rice remained largely the same as that inherited from the Tokugawa period (Mochida 1970). The rice-marketing channels in this period are shown in figure 2-lb. The medium of marketing was still characterized by large shipments on sailing vessels, although they were soon replaced by steamships. Local traders shipped rice in bulk to certain consignees in urban wholesale markets. Steamships sped up the operation and decreased the risk, but it was still an uncertain effort. Deterioration in the quality of the rice and the loss caused by unforeseeable price declines at the destination were still factors. All risk in the transac- tion was borne by the large merchants in the local port cities, while the consignees in the urban markets sold rice on commission. Therefore, fixed and long-term relationships were maintained between local traders and consignees in order to minimize the moral hazard of consignee cheating. Changes in the marketing system in the early Meiji period were caused by the shift in the land tax system. Because agriculture was the most important source of national revenue, the Meiji government implemented a land tax reform (Chiso Kaisei) in 1873. Under the modem land tax system based on land value, landowners, mostly farmers, had to sell their rice locally in order to pay the tax in cash. Thus, village-based marketing was developed. The role of large local traders (sanchi donya) at local ports, who could bear high risk and mobi- lize substantial working capital, became critical in the shipment of rice to urban markets. During the Tokugawa period, the development of local markets varied among the do- mains. For example, in Sakata domain, the government promoted the activities of private merchants, and the interregional trade of rice was handled extensively by these private traders. As a result, the Sakata port was a local marketing center for this domain, as well as for neighboring domains. In Mori domain, the government built depots at the port and provided low-interest loans to traders who shipped merchandise through these depots (Suzuki 1941: 47). In these provinces, local markets were well developed and local traders could thus man- age the rapid expansion in the market supply of rice brought about by institutional change; 30 Tozward the Rtral-Based Dezvelopnment of Comminierce and fidiustry that is, the tax reform in the early Meiji period. In contrast, before the Meiji Restoration, in Sendai domain the shipment of rice to urban areas was handled solely by state trading and few private merchants were activated. Therefore, after the Meiji Restoration, local agents had insufficient capacity to handle the new commercialization of rice, and the local market in Sendai was in disarray. Marketing Innovation in Modern Economic Development The Meiji Restoration of 1868 brought about major institutional reforms. The new, central- ized administrative system removed communication barriers among the regions. The land tax reform forced farmers to sell their rice to local markets, which enhanced commercial- ization. The activities of local collectors were promoted through this process. Institutional reform, however, did not necessarily induce marketing innovation. In the early Meiji pe- riod, marketing relied on the infrastructure inherited from the Tokugawa era. Marketing innovations occurred with innovation in infrastructure in the mid-Meiji period. Thus, in- stitutional reform was a necessary condition-but not a sufficient condition-for market- ing innovation. Let us now examine the process of institutional change in marketing in relation to innovation in infrastructure. The Period of Marketing Innovation (1890 to the 1920s) Toward the end of the nineteenth century, the structure of the marketing system for rice changed markedly in the course of modem economic growth. It was a period of institu- tional innovation and development of the marketing system. New marketing institutions emerged outside the traditional market channels controlled by privileged merchants. Gov- errnment intervention in the market was limited.8 Locally segmented markets were inte- grated into a nationwide market. Two major underlying factors affected the organizational structure of the rice-marketing system. First, the network of railroads was extended to local areas. Rice transport shifted from ships to railroads, bringing change in marketing channels, as well as in the characteristics of marketing agents and the mode of transactions. Second, through severe competition among rice-producing areas, quality controls for rice were enhanced by qualitative and quantitative standards (san-mai kairy6) in each region. This resulted in reductions in trans- action costs and in the risk involved in long-distance transport. After about 1890, changes in the technological environment led to a transformation of the organizational structure of marketing (figure 2-3). The railroad network was quickly extended to local areas throughout the period. The Japanese railroad system began in 1872 with 29 kilometers of track between Tokyo and Yokohama. It was extended to 2,200 kilo- meters in 1890,6,000 kilometers in 1900, and exceeded 20,000 kilometers in the mid-1920s. Rolling-stock kilometers-the distance between railroad stations multiplied by the num- Tednical anid Inistituttioinal Innoz7ations in Rice Marketing in Japan 31 Figure 2-3. The Modern Rice-Marketing System in the Mid-Meiji to Taisho Period Urban wh1olesale market Local town ..... (Shoomai shijyo) Local rice traders (Beikokut shio) Urban traders :alrod. (Beikokut sho) Railroad..... Rice-grading: : * \ : : warehouse: Landlords : Resalers \: : (Beiken Soko) : (Nakatsiigi) , . * . .: ' / ,. . . . . . . . . . . . . . . . ...,.. ./ kz/ \ . Farmersm Retailers Collectors < : Tenant-farmers (Nakagai) < Owner-farmers Retailers ............... Urban consumers Local consumers ber of freight cars that pass-increased from 18 million kilometers in 1890 to more than 2 billion kilometers in the 1920s (Minami 1965: 204-5). In 1910 less than half the rice arriving at the Tokyo wholesale market was shipped by rail (figure 2-4), but the share of railroad transportation increased to almost 99 percent by the end of the Taisho period (1926) (Tokyo Kaimai Donya Kumiai 1937:271-74). Although railroads were owned by the state, there was no evidence that monopolistic pricing was applied to the freight of rice (Kimura 1936, chapter 3). There were several advantages to railroad transportation: loss during loading and unloading was reduced, transportation time was much shorter, and loading was easy (Japan, Ministry of Railroads 1925: 391-92). In addition to these direct advantages, the shift to railroad transportation brought about many changes that led to marketing innovations, including changes in marketing chan- nels and their structure, mode of transportation, and the risk involved in marketing rice. Two marketing channels, traditional water transportation and modem railroads, are de- scribed in figure 2-5. MARKETING CHANNELS AND THEIR STRUCTURE. As marketing channels shifted to the railroad lines and each railroad station became a local marketing center, the local coastal and river ports lost their importance as marketing centers. These changes can be seen as analogous to computer networks. If we consider traditional marketing channels as hierarchical net- works under mainframe host computers, then the modem marketing channels connected by railroads were a peer-to-peer LAN system comprised of many small workstations. In the traditional system, rice produced in a village was carted to a nearby rural market- ing center on a riverbank, then loaded into a small riverboat and transported to a large 32 Toward the Rural-Based Development of Conimerce and Industry Figure 2-4. Extension of the Railroad and Shipment of Rice to the Tokyo Market Billion ton-kilometers 15 Railroad ton-kilometers! 10 5 0 , ~~~~~~~~~~~~~~~ ~~Million bags 0 Water l l 1900 1910 1920 1930 Source: Minami (1965: table 12), Tokyo Kaimai Donya Kumiai (1937). local trader at a seaport. Large local traders collected sufficient amounts of rice to send on to urban markets by boat. Thus, local markets were connected to urban markets only through local traders who served as local hubs for marketing. The role of riverports was limited to their use as transit points for transshipments. Under the modem system, local marketing centers were decentralized. Each railroad station was connected directly to ur- ban markets by rail and became a local marketing center. Furthermore, the extended rail- road network enabled each region tobe connected with other regions, and marketing among local regions became easier. MODE OF TRANSPORTATION. Economies of scale in transportation were reduced under the modern system. The minimum unit of transactions was significantly reduced-the 500 koku (about 75 metric tons of brown rice) required for a shipment by boat was replaced by 40 koku (6 tons), the maximum needed to fill a single freight car (Suzuki 1941: 96). Freight trains were serviced frequently, and the traders could easily arrange their trans- shipments. Thus, economies of scale in transportation became much smaller. Small trad- ers with limited financial capacities could engage in local-to-urban shipments under the modem system. Tecdnical and Institutfionial Inniovations in Rice Marketing in Japan 33 RISK INVOLVED IN RICE MARKETING. The high risk of transactions-price risk, quantity loss, and quality deterioration-was another critical factor that characterized the organizational structure of the traditional system. With railroad transportation, the time for long-distance transport was significantly reduced. Once rice was loaded onto freight cars at a local rail- road station, it was sent directly to an urban station without further transshipment and reached its destination within a few days. Quantity loss or quality deterioration, which was inevitable in lengthy transportation and frequent transshipments, could be avoided. The price risk was reduced, because the price fluctuation at the urban market within a few days was unlikely to be critical, while more distant future prices (beyond a few months) could not be foreseen. Risks that had been inevitable under the traditional system were thus reduced. Because consideration of economies of scale in transportation disappeared and related risks were drastically reduced with railroad transportation, large traders at local ports lost their comparative advantage and declined in importance. Small traders were able to en- gage in rice marketing with small shipments from many small local marketing centers. Under the traditional system, the important ability was the capacity to mobilize large amounts of capital and coordinate many agents to collecf merchandise for large shipments. Under the modem system, this ability was no longer critical; the ability to respond to market conditions in urban and local markets became more important. Thus, the charac- teristics required for marketing agents changed. Figure 2-5. Changes in Marketing Channels Induced by Innovation in Infrastructure: Railroads Traditional marketing channels Seaport Villages Seaport Horses Urbanr Large local Boats River ml arket Sail/steamboats trader port Horses Villages Villages Villages Modem marketing channels Railroad station Villages Villages Urban < Freight car Horses market Horses Horses Villages Villages Villages 34 Towtard thze Rutral-Based Delelopment of Commnerce and Induistry Standardization of Rice Under the modem system, local-to-urban and local-to-local marketing were more strongly integrated by railroads. The size of the integrated commercial market grew, and the share of shipped merchandise in each region increased. Integration of the local and urban mar- kets into a nationwide market reduced interregional differences in rice prices.9 In 1880 the coefficient of variation (CV) of local rice prices over the prefectures was 13.3, which de- clined to 4.5 in 1910.10 A number of reasons caused the decline of the CV, including the extension of the railroad and telegraph networks and nationwide standardization of rice varieties and qualities. The values of CV and the total rail distance showed a high negative correlation of -0.85 over the last three decades of the Meiji period. In the Higo region (see figure 2-1), the share of rice exported to other regions increased sharply throughout the prewar period (table 2-2). Before the spread of the railroads in this region, only 8 percent of rice was exported to other regions, but the export share increased with the extension of the railroad. Just after the opening of a tunnel that connected Kyushu to the main island, Honshu, in 1911, the share of exported rice reached 40 percent of total production. Through the process of market integration, competition among rice-produc- ing regions was strengthened, and market competition induced inncvation in agricultural production and marketing. Another innovation in marketing and agricultural production was standardization, in- cluding improvement in the quality of rice (san-mai kainyO) brought about by increased competition and expanding market size. It reduced transaction costs and the possible risks of marketing. Groups of progressive farmers, landlords, and traders and farmers' associa- tions (n6kai) played important roles in improving the quality of rice. Rice-grading ware- houses (beiken soko) established local standards for marketed rice. With severe competition among the rice-producing regions, the quality of the rice from. a given region directly reflected its price in the urban market. Standardization also af- Table 2-2. Production and Export of Rice in the Higo Region (thousand metric tons) Prodtuction Export Export ratio Year (A) (B) (B/A) Railroad extended 1887 157 12 0.08 1893 141 23 0.17 Kumamoto (1892) 1897 188 43 0.23 Yatsushiro (1896) 1902 185 59 0.32 1907 228 82 0.36 Hitoyoshi (1907) 1913 227 90 0.40 Shimonoseki-moji connected by tunnel to Honshu (1911) Note: Higo is currently Kumamoto Prefecture on Kyushu. Data refer to three-year averages, except 1893 and 1913. 1892 = 1893-94 average. 1913 = 1912-13 average. Source: Japan, Ministry of Railroads (1916: 691-92). Technzical anrd Instituttionial Innovations in Rice Marketing in Japan 35 fected price. Not only better quality, but also standardized and stable quality ensured higher prices in urban markets. Efforts were made to improve and control the quality of rice in many rice-producing regions. Postharvest controls, such as drying, husking, and packing, were enhanced. Improved technologies for postharvest processing were devel- oped and producers were encouraged to adopt them. De facto qualitative and quantita- tive standards were established locally in each region. As early as the 1880s, groups of farmers, landlords, and traders took the initiative in improving the quality of rice in various regions.1' Around 1900, these voluntary efforts were transformed into official regulations by local governments. For example, in Yamagata Prefecture, the local government introduced an ordinance on the paddy-drying standard based on recommendations from the farmers' associations in 1907. Many farmers' associa- tions paid subsidies to farmers for improving postharvest controls on rice. The local gov- emrnent also set standards for the packaging of rice shipped to other regions (Yamagata-ken Sanmai Kairy6 Kyokai Reng6kai 1958: 380-418). Rice-grading warehouses (beiken s6ko) were rice-marketing institutions and issued rice exchange notes (beiken). They were managed by private companies or cooperatives and built in the mid- to late Meiji period. In 1910 there xwere thirty-three beiken soko in the country (Suzuki 1941: 137-38). Their function was not limited to storage but also included grading and standardization of rice. A beiken s8ko normally had three sections: inspection/ grading, repackaging, and storage. A financial section was also included, or the ware- house cooperated with commercial banks. Since they issued rice exchange notes, a form of bill of exchange, finance was a critical function in the warehouses. Because a beikenl soko had to maintain large storage facilities, the operation involved economies of scale. Moreover, because of strong competition among the regions, it could not be profitable unless it could collect a large volume of rice. Many attempts to build beikeni s6ko failed because they could not collect a sufficient amount of rice. Warehouses that could not establish cooperation with commercial banks were not sustainable. The Sankyo (beiken) soko in Sakata, Yamagata Prefecture, was one of the most successful warehouses and had an excellent reputation. It was attached to the Sakata rice auction market, which was formally reorganized in 1893.12 It used storage facilities from the ware- house owned by the daimyo in Sakata. The former daimyo and the Honma family, a very large landlord, supported the warehouse, which ensured its reliability. The major functions of beiken s6ko are shown in figure 2-6. Farmers and collectors brought ungraded rice to the warehouse, which inspected and graded the rice and issued a rice slip indicating the grade, volume, and date of receipt. The minimum unit of accepted volume was one go (180 cubic centimneters). The warehouse unpacked the rice, mixed it for each grade, and packed it into standard packages. Each package was set to contain 4 to (60 kilograms) of brown rice. Through this process, quantity and quality were standardized for each grade. Better storage facilities ensured a stable quality over the seasons. Thus, the warehouse could supply rice of consistent, standard quality. The warehouses prompted innovations in marketing as well as finance. The warehouse issued a rice exchange note in exchange for rice slips when the total volume of the rice 36 Toward the Rural-Based Development of Comnterce antd lindustry Figure 2-6. Functions of Beiken Soko (Rice-Grading Houses) Rice-grading warehouse (Beiken Soko) Ungraded Farmers/ rice Grade Blending collectors 4 .....I...... PIweighing 7) packaging Rice slip Traders .............- ....... A- Storage I I Rice exchange note (Beiken) Traders - - - - - - - - - - - - - - - - - - - - - I > Flow of rice *.... *> Flow of rice slip - - _ _> Flow of rice exchange note specified on the slips reached 10 koku (1.5 metric tons) or more for each grade. The rice exchange note was transferable, and its holder could receive the prescribed amount of rice in exchange for the note, plus storage and handling fees. These fees were a major source of income for the warehouses. The rice exchange notes were also used as collateral for bank loans or in pawn shops. Transactions between traders were completed by exchanging these notes. Small-scale transactions between farmers or collectors and traders were done with rice slips, since they were also transferable. Landlords preferred to receive rent from tenant farmers in the form of rice slips rather than in-kind payments of rice. Transactions by rice exchange notes provided important advantages for marketing agents. Traders and landlords did not need storage facilities, nor did they need to move rice for each transaction, which greatly re- duced marketing costs. Warehouses provided standardized rice, and thus marketing agents did not need to inspect the quality of the rice. Instead, they could purchase the desired rice by referring to the brand, grade, and number of bags. Transaction costs for marketing were also reduced, and because the notes could be used as collateral, capital constraints were eased for marketing agents. Quality controls and standardization had a great impact on marketing and agricultural production. Once a local standard was established, transactions in the markets could be done by referring to a brand name. Although local standards could not be established in all regions, rice from regions that were successful with standardization was used as a standard in urban wholesale markets, and its price was higher than for rice from other Techlnical and Institiutional Innovations in Rice Marketing in Japan 37 regions. It should be noted that the success in rice quality improvement and beiken s6ko was achieved by the initiatives of rural-based entrepreneurs, such as groups of progres- sive farmers and rice traders. The Role of Rural Entrepreneurs Village-Level Marketin1g Throughout the Tokugawa period, increased rice yields enabled peasants to diversify into a variety of cash crops. In advanced areas, such as on the outskirts of Osaka and Kyoto (the Kinai region), many peasants took advantage of this diversification. Rural-based cottage industries also became pervasive in these areas. In the seventeenth century, regional spe- cialization progressed.13 For example, of the merchandise produced in the country nearly 40 percent of the items (706 of 1,807 items) were produced primarily in the Kinai area (Miyamoto 1988: 39), while remote areas specialized in the production of rice and other staple crops. Interregional specialization was supported by interregional marketing. Spe- cialization or diversification in agricultural production enabled rural entrepreneurs to act as marketing agents. Many merchants emerged from the peasantry in advanced areas to become competitors to the privileged merchants who were given monopoly licenses by daimyo in the trade of rice and other major commodities. In the Meiji and Taisho periods, rice marketing at the village level was carried out by small collectors (nakagai). They collected rice directly from farmers and landlords and de- livered it to local traders or rice-grading warehouses in nearby towns (see figures 2-2 and 2-3). Most of these activities were taken on by peasants as side jobs. They shipped rice on commission or traded on their own. Landlords and large farmers often sold rice to large traders in local towns through these small collectors. The collectors also purchased rice for cash from producers and sold it to large local traders. Because they purchased ungraded rice at the farms, the ability to identify the quality of rice and foresee market prices was crucial for their activities. One farmer in Yamagata Prefecture, Zenji, recorded his daily activities in his diary,14 which covered over forty-two years, from 1893 to 1934. He acted as a collector while also engaging in rice farming. In the peak year of his activity as a collector, in 1916, he pur- chased 603 bags"5 of rice from 90 farmers in 18 nearby hamlets. He also took an additional 293 bags for transportation on commission. In that year, he handled 1,067 bags of rice, including 171 bags of his own rice (Toyohara 1976: 169-73). He sold it to local traders or delivered it to beiken s8ko in Sakata. Transportation was by a horse-drawn wagon or sled, each of which could carry eight to thirteen bags. He earned Y130 for this operation during the year. At that time, the value of 1 tan (0.1 hectares) was Y350 to Y400 (Toyohara 1976: 184). When he married in 1904, his family cultivated 1.2 hectares of owned land and 0.9 hectares of rented land, which was typical for an owner/tenant-farmer in the region. His activities as a collector enabled him to earn sufficient income to expand his landholdings. 38 Tohward the Rutral-Based Detelopmenit of Connnerce and I1ndiustry In the early 1920s, his operation was expanded to 1.6 hectares of owned land and 1.9 hectares of rented land. He also became the owner of 6 hectares of forest and 0.08 hectares of tenanted land (Toyohara 1976: 16, 18). His profitable operation as a collector, however, did not mean that he could exploit monopsonistic profit. He made his fortune from his extraordinary hard work and entrepreneurial ability. It should also be pointed out that nearly perfect information on prices in the urban and local town markets was available in the village. Farm producers had the alternative of selling their products to one of these village traders or selling directly to town traders nearby. Therefore, there was little room for traders' monopsony (Kimura 1936: chapter 3). According to Zenji's diary, there were at least one or two collectors in each hamlet. Their activities were stimulated by the land tax reform in the early Meiji period. In order to pay the land tax in cash, it became necessary for farmers to sell rice in the market. Zenji is representative of many peasant entrepreneurs who participated in marketing activities in response to new profit opportunities created by the modern institutional reforms of the Meiji period. The Origins of Rutral Entreprenieurs In a traditional subsistence-oriented economy, where peasants cultivate traditional crops in traditional ways, the marketable surplus of farm produce is limited. The market is seg- mented by limited infrastructure. Under these circumstances, there is little opportunity for entrepreneurs to emerge from the peasantry to play an important part in marketing (figure 2-7). When innovation in agricultural production is initiated, the marketable surplus of farm produce increases, which provides a marketing opporth.mity for rural entrepreneurs. Im- proved infrastructure enables them to engage in interregional marketing, resulting in the integration of segmented markets into a large regional or national market. If marketing is activated, this, in turn, stimulates innovations in farm production. The integration of mar- kets enables a region to specialize in the production of products in which it has compara- tive advantage. Interregional competition in the market leads to institutional innovations in marketing. Efforts toward quality improvement and standardization of products be- come necessary for each region to gain a share of the expanded market. As a result, mar- keting efficiency is improved, as observed in prewar Japan. A similar process was observed in corn and cotton production and marketing in the United States in the nineteenth cen- tury (Chandler 1977). The process involves the mutual reinforcement of innovations be- tween agricultural production and marketing. In a region of Java where traditional upland crops and rain-fed rice were produced, the introduction of a new cash crop, vegetables, created new marketing channels to the urban center (Hayami and Kawagoe 1993: 111-34). Establishment of the new marketing chan- nels enhanced vegetable production in the region and vice versa. We should note that the marketing of vegetables in this case was done by rural-based traders from the peasant class. In the late Tokugawa period in Japan, diversification in agricultural production also Techlical and Instituttional Innovations in Rice Marketing in Japan 39 Figure 2-7. Role of Rural Entrepreneurs in Marketing Innovation Institutional reform .s Innovation in infrastructure Marketing innovation Integrationt of the local markets Rural Expansion of the entrepreneurs market through * specialization * competition Innovation in Agricultural Production ( Standardization A locally segniented market Limited infrastructure Traditional marketing Traditional production enabled rural entrepreneurs to act as marketing agents and to gradually grow to be com- petitors of the privileged feudal merchants based in castle towns. A question now arises. In what segment of rural society do rural entrepreneurs origi- nate? One possible source may be landlords. A widely accepted view asserts that the land- lords had overwhelming power to control tenants as landowners, traders, and moneylenders. If this view is correct, it implies that landlords played an important role as entrepreneurs in rural marketing. Their role in local marketing in Japan, however, does not seem to have been vital. According to a survey conducted by the Ministry of Agricul- ture and Commerce in 1925, there were 3,176 landlords who owned more than 50 hectares of farmland."6 Among them, only 358 (11 percent) engaged in commerce (table 2-3). More than one-third engaged in farming, while another third who did not engage in economic activities other than rent collection may be legitimately called "parasitic landlords" (Tobata 1936: 78-87). Although the data are too limited to draw firm conclusions, Zenji's case and the case of vegetable marketing in Java suggest that it is highly possible that rural entrepreneurs are not limited to the landlord class, but that they can emerge from the peasantry when prof- itable opportunities become available. 40 Towuard the Rutral-Based Development of Commerce and Industry Table 2-3. Occupations of Landlords Holding More than 50 Hectares of Farmland, 1925 Occutpation Nutmnber Percentage Farming 1,122 35 None 952 30 Commerce 358 11 Official/office worker/director 206 6 Moneylender 147 5 Winery 147 5 Mining/forestry/fishery 56 2 Other 46 1 Subtotal 3,034 95 Owned by corporation 145 5 Total 3,179 100 Source: Japan, Ministry of Agriculture and Commerce (1925: 691-92). Government and Farmers' Cooperatives In order to consider the government's role and the functions of farmers' cooperatives in marketing, an examination of the rice-marketing system during wartime Japan is needed. Rice marketing during World War II showed a process of market distortion and suppres- sion of entrepreneurship caused by heavy government intervention in the market. Government Regulation in the Wartime Economy (1920s-30s) The Taisho to the early Showa period covers the interwar period from the 1920s to the 1930s. These years were characterized by government intervention in the market, which was introduced in the early 1920s and gradually increased throughout the period. The activities of private marketing agents were suppressed and marketing functions were re- placed by farmers' cooperatives, and market mechanisms were neglected. Two factors affected the marketing system. The first was direct government intervention in the mar- ket. The second was the government's promotion of farmers' cooperatives. Inflation during World War I caused the domestic rice price to more than double. The poor were hard pressed, and a rice riot broke out in 1918. The mob demanded rice and attacked rice stores and landlords' depots. The military was called out to suppress the riots, which became the turning point in the government's attitude toward the rice mar- ket. When the war ended, the prices of farm products were depressed by the postwar recession. The government began to intervene in the rice market under these economic and social conditions. In 1921 a rice law was enacted. Under this law, the government Technical anid Institittional Innoz7ations in Rice Marketing it? Japan 41 established a permanent system designed to regulate rice supply through the purchase and release rice, as circumstances required. Following the return to the gold standard in 1930 at pre-World War I parity, the Japa- nese economy was severely depressed, and there was a serious agricultural crisis. Although farm prices had already declined more than 20 percent in the late 1920s, they dropped an additional 30 percent during the next two years (Kawagoe 1993: 185). In response, the rice law was revised on several occasions. In the revision of 1931, a license system for rice import and export and a system that permitted the government to purchase or release rice at floor/ceiling prices were introduced. In the following year, it was arranged that the floor/ceiling prices would be calculated based on farm household expenses and produc- tion costs. Monopolistic power was given to the government to import and export rice. In 1933, the rice law was superseded by a rice control law, which introduced a system per- mitting the government to purchase and release rice in an unlimited manner within the floor/ceiling prices. Government intervention in the market gradually increased. It was initially intended to bring about price stability, but later intervention turned into rigid economic control, while the market mechanism was slowly suppressed, and domestic rice supply fell increasingly short of demand. At the same time, farmers' cooperatives were promoted under a govern- ment policy of protecting agriculture and rural society. Farmiers' Cooperatives There has been a myth that farmers' cooperatives could be a strong countermeasure against traders' monopsony. Many developing countries have thus tried to promote farmers' co- operatives as marketing agents. Of course, the functions of farmers' cooperatives are not limited to marketing. They could provide a variety of services to farmers, such as the exchange of farming knowledge, mutual crop insurance, the extension of new technolo- gies and education, or officially supported loans. If the market is competitive, however, as observed in the rice market in prewar Japan, there is no economic rationale to form coop- eratives focused on marketing."7 The story in the 1920s and 1930s tells us that government-guided farmers' coopera- tives lacked entrepreneurship as marketing agents. The government promoted the es- tablishment of farmers' cooperatives. In marketing activities, subsidies were provided to build farm depots (nogyo soko) managed by the farmers' cooperatives. The depots were facilities for the assembly, joint storage, and shipping of farm produce. In 1931 the Japan Federation of Rice Trading was established as the central organization of farmers' cooperatives for the sale and purchase of farm produce and raw materials. Farm depots numbered about 1,000 in 1921, and substantially increased to 5,000-6,000 by the late 1930s. We should note, however, that the dominance of the farmers' cooperatives in the marketing in this period does not imply their competence as entrepreneurs. Instead, the government policy that gave priority and privileges to cooperatives, while suppressing the activities of traders, resulted in the decline of traders and the dominance of farmers' 42 Towvard tite Ruiral-Based Development of Com1Zmerce and Indutstry Figure 2-8. Rice Marketing under Government Control in Japan: Taisho to Early Showa Period (1920s to 1930s) Govemment Local town. Cooperatives' Association Cooperatives' (Zenkokt hanrbai-ren) < Association (Ken hnnbrai-reni Cooperative (Sangyo kumiaf) Urball 7wholesale nmarket : J: (Shomai shijyo) ' I I I 4- Local rice traders Farm depots Urban traders (Beikok: sho) (Sangyo kumiai) (Itakut/Kaitsuke don ya) Railroad ....... *| 4 : Rice-grading:I : Resal:rs * . warehouse Landlords Resalers :X: /X : (Nakatsugi) .' w *1 # .. -_.... /. ...... _ Farmers . Retailers Collectors < Tenant-farmers (Nakagai) < Owner-farmers -to Retailers ............... Urban consumers Local consumers cooperatives. The farm depots were given priority over private traders in collecting rice for the government (figure 2-8). The marketing channels of farmers' cooperatives were inefficient compared with those of private traders. For example, a farm depot in Yamagata Prefecture reported that it sold rice to traders in Tokyo, not to its central organization in Tokyo, because the latter was slow to respond and offered lower prices (Yamagata-ken Sanmai Kairy6 Reng6 kai 1958: 28-29). Because of inexperienced administration and bureaucratic delays in operations, cooperatives could not provide services that were as good or as flexible as those of private traders."8 Similar inefficiencies have also been observed in the operations of farmers' co- operatives in Indonesia (Mears 1981: 133-37; Kawagoe, Ohkama, and Bayo 1992). After the war, farmers' cooperatives played an important role in the marketing of government- controlled rice. The functions of the marketing institutions, rice policies, and the role of private marketing agents were different from those seen in the prewar period. There was little functioning of the market mechanism and little room for entrepreneurs to play active roles in the marketing of rice. Rice marketing in the postwar period has been strictly regulated. It has resulted in the overproduction of rice, huge financial deficits, inefficient production, and rent-seeking by marketing agents."9 The rice market in postwar Japan was a clear example of market dis- tortion created by government intervention.'0 Techniical anid Instituttional innoz2ations in Rice Marketing in Japanl 43 Conclusions Stnimnary Before World War I, rice-marketing activities in Japan were undertaken by numerous pri- vate traders. The market generally operated under free competition, and government in- tervention was limited. In the course of modem economic growth in the Meiji period, the marketing system of rice changed drastically in structure. Institutional reform-that is, land-tax reform-promoted the activities of rural-based marketing agents. The railroad network was extended to local areas by the mid-Meiji pe- riod. The shift from water to land transportation brought a major change in marketing chan- nels through reductions in economies of scale in transportation and reduced marketing risk. The market was expanded significantly through the integration of hitherto segmented local markets. Market integration enhanced competition among rice-producing areas. Through severe competition, quality controls on rice were promoted by the initiative of ru- ral-based entrepreneurs, including progressive farmers and rice traders, and qualitative and quantitative standards were established in each region. Rice-grading warehouses were de- veloped as the key institution to establish standards and facilitate finance for marketing. The establishment of local standards, coupled with the extended telegraph network, enabled marketing agents to transact their business more easily by utilizing local brand names and grades. Transaction costs and the risks associated with rice marketing were greatly reduced by this institutional innovation. We should note that these marketing in- novations were initiated by rural entrepreneurs who originated not only from the land- lord class, but also from among the peasantry. The findings present strong counterevidence to the traditional view that peasants in developing countries have neither the capabilities nor the willingness to undertake commercial and industrial enterprises. Policy Inmplications Local marketing systems in developing economies are often regarded as inefficient and imperfect. Some views, such as traders' monopoly and usury, often rationalize govern- ment intervention in the market. The case of postwar Japan, however, indicates that im- proper government intervention is a source of market distortion and inefficiency. We should note that the dynamic changes in the marketing system of prewar Japan took place when the market was under free competition and government intervention was limited. Entrepreneurs emerged from the rural population and played a key role in creat- ing a modern marketing system-even standardization and grading were promoted at the local level through the initiative of rural entrepreneurs. Entrepreneurship is not unique to a certain social or ethnic group, but can emerge from the peasantry if market opportunities are available. Nevertheless, this does not mean that government should not do anything in the mar- keting sector. Marketing innovations can be initiated and accelerated by improvements in 44 Toward the Ruiral-Based Developnment of Comnunerce anid Iniduistry infrastructure. Therefore, public investment and deregulation that encourage the estab- lishment of infrastructure-such as railroads, roads, and telecommunication networks- are critically important. Government efforts to support innovation in agricultural production are also vital to marketing innovation. Changes in the marketing system lead to changes in agricultural production and vice versa. Therefore, research and extension in agricultural production, including the introduction of new marketable crops, are also important and are promising areas for effective government initiative. The standardization of agricultural commodities greatly reduces their marketing cost. Introduction of a unilateral official standard designed by bureaucrats, however, often ham- pers the natural development of the market. Careful observations of whether any de facto local standards are informally adopted by marketing agents will be important in the de- termination of useful official standards for commodities. Notes 1. Although the term "marketing" describes a wide range of activities, the major focus of this chapter is rice-distribution systems, from producers to the urban wholesale market. 2. The annual average population growth rate in the prewar period (1885-1940) was 1.2 percent, while that in the latter half of the Tokugawa era (1730-1872) was 0.02 percent (Ohkawa, Shinohara, and Umemura, eds., Vol. 2; Hayami and Miyamoto 1988: 43-44). 3. It was called tsu(dome, which often caused severe hunger in regions experiencing crop failure. 4. The Tokugawa government almost completely closed Japan to trade with foreign countries. Therefore, we do not need to take into account the international trade in rice. 5. The rice tax rate is assumed to be 40 percent, and the taxable standard is taken to be two-thirds of the total rice harvest (Miyamoto 1988: 21). 6. One kokh of brown rice is 150 kilograms. By the end of Tokugawa period to the early Meiji period, traditional vessels were replaced by much larger modem sailing ships and steamers. The ninimum transaction unit in these ships was 500 koku (= 75 metric tons of brown rice). 7. Under the national isolation policy of the Tokugawa shogunate, beginning in the early seventeenth century, the building of ocean-going ships was banned and the art of navigation degenerated. 8. Although import tariffs were occasionally imposed on rice from Korea and Taiwan, China, there was little government intervention in the domestic market during this period. 9. The role of local auction markets, which had formed local centers for the formulation of rice prices in each region since the Tokugawa period, was weakened through this process. 10. Koiwa (1981: 41). The data represent five-year averages centered on each of the years shown, respectively. 11. Similar movements for quality control and standardization initiated by local traders were observed in the cotton-weaving industry. See the chapter by Itoh and Tanimoto in this volume. 12. The discussion in this section is based on Yamagata-ken Sanmai Kairy6 Ky6kai Reng6kai (1958). Techntical and Inistitutional Innovations in Rice Marketing in Japan 45 13. See Smith (1959: chapter 6) for the development of markets in the Tokugawa period. 14. The following description draws on Toyohara KenkyCikai (1976: chapter 7; 1977). 15. A bag made from rice straw, called tawvara (Inyo), contained about 0.4 kokut, or 60 kilograms, of brown rice. 16. Only very large landlords held more than 50 hectares of farmland; the average holding of noncultivating landlords was only around I hectare. 17. If post-harvesting or marketing processes owned by producers reveal economies of scale, groups of farm producers, dealing with one specific area of activity, may organize cooperatives in order to seek scale economies for marketing through joint facilities for storage or processing. 18. Yamagata-ken Sanmai Kairy6 Kyokai Reng6kai (1958: 26-33). 19. See Kawagoe (1993) for further discussion. 20. Under the Uruguay Round of multilateral trade negotiations of the General Agreement on Tariffs and Trade, the Japanese government agreed to maintain minimum access opportunities for rice imports in compensation for tariffs. The government also planned to replace the obsolete food control system with a deregulated system based on a new food law in 1995. Under the new system, it is expected that private market channels will be activated. References Boeke, J. S. 1953. Econontics and Economtic Policy of DUIal Societies as Exemplified by Indonesia. New York: Institute of Pacific Relations. Chandler, A. D., Jr. 1977. Thre Visible Hanid: The Managerial Revolution in Americani Butsiness. Cam- bridge, Mass.: Belknap Press of Harvard University Press. Geertz, Cliford. 1963. Agricuiltural Involuttion: The Processes of Ecological Change in Indonesia. Berkeley: University of California Press. Hayami, Akira, and Matao Miyamoto, eds. 1988. Keizai shakai nto seiritsut: 17-18 seiki (Formationi of the Econonmy in the Seventeenth and Eighteenith Centuries), Vol. 1 of Nihon keizai-shi Uapanese Econonmic History). Tokyo: Iwanami Shoten. Hayami,Yujiro, and Toshihiko Kawagoe. 1993. The Agrarian Origins of Comnterce and Industry: A Stiudy of Peasant Marketing in Indoniesia. London: Macmillan. Hayami, Yujiro, and Saburo Yamada with Masakatsu Akino, Le Than Ngiep, Toshihiko Kawagoe, and Masayoshi Honma. 1991. The Agricuiltural Developmenit of Japan: A Century's Perspective. Tokyo: University of Tokyo Press. Japan, Ministry of Agriculture and Commerce. 1925. 50-chobui ij6 no daijinushi (List of Large Landlords Holding More than 50 Hectares). Tokyo: Ministry of Agriculture and Commerce. (Reprint: Nihon Nogyo Hatatsu-shi Ch6sakai, ed. 1955. Nihon nogyd hiatatsui-shi [History of Agricultutral Develop- ment in Japani], Vol. 7. Tokyo: Chu6 K6ronsha.) Japan, Ministry of Agriculture, Forestry, and Fisheries, Food Agency. 1990. Slhokury0 kanri tokei nenpo 1990 (Annuial Statisticsfor Staple Food Control, 1990). Tokyo: Food Agency. Japan, Ministry of Railroads. 1916. Honpo tetsud6 no shMakai-keizai 11i oyobosenr eikyo (Impact ofRailroads on1 the Society and Economy of /apan). Tokyo: Ministry of Railroads. . 1925. Kome ni kansuri keizai chOsa (Economic Sturvey on Rice). Tokyo: Ministry of Railroads. 46 Toward tire Ritral-Based Developmnent of Commerce and Industry Kawagoe Toshihiko. 1993. "Deregulation and Protectionism in Japanese Agriculture." In Juro Teranishi and Yutaka Kosai, eds., Thte Japanese Experience of Economfic Reforms. London: Macmillan. Kawagoe, Toshihiko, Kunio Ohkama, and Al Sri Bagyo. 1992. "Collective Actions and Rural Orga- nizations in a Peasant Economy in Indonesia." Developing Econonmies 30(3): 215-35. Kayo, Nobufumi, ed. 1977. Kaitei Nihon nogyo kiso tokei (Basic Statistics on Japanese Agricultuire, re- vised edition). Tokyo: N6rin T6kei Kyokai. Kimura, Wasaburo. 1936. Beikokut ryfitsfi hiyo kenktil (A Stuidy on7 Rice-Marketing Costs). Tokyo: Nihon Gakujutsu Shink6kai. (Reprint of 1980 edition published by N6san Gy6san Bunka Ky6kai.) Koiwa, Nobutake. 1981. "Meiji-ki ni okeru beika oyobi tashohin kakaku no chiikikan hend6 ni tsuite" ("Interregional Price Differences for Rice and Other Goods in the Meiji Period"). Keizai kenkyll (Hirosaki University) 4(0ctober): 30-41. Mears, Leon A. 1981. The New Rice Economy of Indonesia. Yogyakarta: Gadjah Mada University Press. Minami, Ryoshin. 1965. TetsWd6 to denryokui (Railroads and Electricity). Tokyo: T6yo Keizai Shinp6sha. Miyamoto, Matao. 1988. Kinsei Nihoni nio shiJ6 keizai (Market Econoniics in Modern Japan). Tokyo: Yuhikaku. Mochida, Keizo. 1970. Beikoku shijo n1o tenkai katei (Transformiiationi of the Rice Market). Tokyo: Univer- sity of Tokyo Press. Nakamura, Satoru. 1968. Meiji Ishi no kiso kozd (Basic Strlctulre of the Meiji Restoration). Tokyo: Miraisha. Ohkawa, Kazushi, Miyohei Shinohara, and Mataji Umemura, eds. Various years. Choki keizai t6kei (Long-term Economic Statistics of Japan [LTES]). Tokyo: T6y6 Keizai Shinp6sha. Smith, Thomas C. 1969. The Agrarian Origins of Modernt Japan. Stanford, Calif.: Stanford University Press. Suzuki, Naoji. 1941. Beikoku haiktpi no kenkyfi (Stuidy on Rice Marketing). Tokyo: Shozan-bo. Tobata, Seiichi. 1936. Nilion n6gy6 no tenkai katei (Developmiienit of Japanese Agricututitre). Tokyo: Iwanami Shoten. Tokyo Kaimai Donya Kumiai. 1937. Tokyo kainmai donya kuimia Fuikugawa shlmai-shiJ6 50-nen-shi (Fifty- Year History of thte Futkutgawva Rice Market in the Tokyo Rice Wholesalers' Union). Tokyo: Tokyo Kaimai Donya Kumiai. Toyohara Kenkyuikai. 1976. Zenji nisshi kaidai (Commiiienetaries on Zenji's Diaries). Tokyo: National Re- search Institute of Agricultural Economics. - . 1977. Zenji nisshi (Zenji's Diaries, 1893-1934), 2 vols. Tokyo: National Research Institute of Agricultural Economics. Yamagata-ken Sanmai Kairy6 Reng6kai. 1958. Yamagata-kent beikoku njfitsut keizaishi (History of Rice Marketing in Yamagata Prefecthre). Yamagata: Yamagata-ken Sanmai Kairy6 Ky6kai Reng6kai. Umemura, Mataji, Saburo Yamada, Yujiro Hayami, Nobukiyo Takamatsu, and Minoru Kumazaki. 1966. Noringyo (Agricuiltutre and Forestry). Tokyo: T6y6 Keizai Shinp6sha. Rural Entrepreneurs in the Cotton- Weaving Industry of Japan Motoshige Itoh and Masayuki Tanimoto Tl Nhe textile industry has been important to the Japanese economy for a long time. The industry had already assumed a crucial position in the Japanese economy in the Tokugawa period (1603-1867). National distribution networks were function- ing well for a variety of textile products during this period, and both local and national wholesalers played important roles in distributing products, making it one of the indus- tries of the time with the most sophisticated distribution systems. There were many estab- lished local production areas for textile products in the Tokugawa period, and the putting-out system, which is one of the main subjects of this chapter, could be observed. The change from autarky to free trade at the end of Tokugawa period (1859) created a great shock for the industry. On the one hand, competitive foreign textile products, both yarn and fabrics, came from foreign countries, and severe competition restructured the domestic industry. On the other hand, international trade offered domestic producers the opportunity to sell their products in foreign countries. Japan established itself as one of the largest exporters of textile products in the world in the 1920s, and it became the world's largest exporter of cotton textile products in the 1930s. The textile industry continued to be a major export industry for Japan until the beginning of the 1970s. Table 3-1 shows the share of the textile industry in the total exports and outputs of Japan from 1882 until 1970. The shares, both in exports and in outputs, were very high (note that the cotton textile industry was in the majority in these figures). Although the structure of the industry changed many times during this period, the in- dustry retained many of the characteristics it had more than 100 years ago, when Japan opened its market to the rest of the world. Active use of the putting-out system and the role played by wholesalers are typical examples of the traditional business practices. 47 48 Tozvard the Rutral-Based Development of Comn2erce and Iiididstry Table 3-1. The Share of the Textile Industry in Total Manufacturing Imports and Output, 1882-1970 (percentage) Period Imltports Ourtputt 1882-91 35.2 27.8 1892-1901 52 34.7 1902-11 50 26.8 1912-21 51.3 30.1 1922-31 55.5 30.1 1930-39 43.7 23.7 1951-55 41.5 18.3 1966-70 14 8.2 Souirce: Yamazawa (1984). A study of the historical development of the Japanese textile industry at the turn of the century is helpful in understanding how the industry has developed and achieved its present form. Such a historical study will also provide useful insights on the general topic of this volume, "the role of rural entrepreneurs in Asia," because the textile industry is typical of rural industry and its development and has been supported by the activity of many local traders and farmers. There is a vast literature on the history of the textile industry in Japan. Both semi- mac- roeconomic and microeconomic analyses of the industry are available for a variety of pro- duction regions. Since most of the studies are reported solely in the Japanese language, it will be useful to review the literature. Although a comprehensive survey is far beyond the scope of this chapter, we offer a rough summary of the literature here as background for our discussion. The major portion of this chapter is based on Tanimoto's (1986) previous study of one particular production area, Iruma. The study is a microeconomic analysis of production and distribution activities in this narrow region, based on the detailed documents of a few local merchants in the area in the period from 1850 through 1925.1 By using these docu- ments, we can draw a detailed picture of the structure of transactions in the region. Iruma is an interesting case for our project because it is located in a rural area and its products were primarily intended for the domestic market. The cotton-textile industry was greatly affected by international trade, and some production areas in Japan became heavily involved in exports. lruma, however, was completely dependent on the domestic market. This dependence was supported by the network of wholesalers, local merchants, local wholesalers, national wholesalers, and financial agents. Without such sophisticated networks of merchants it would have been impossible for the small, independent produc- ers of the area to meet the demand of consumers throughout the nation. An analysis of the structure of transactions among these wholesalers, as well as producer-farmers, is essen- Rural Entrepreneutrs in the Cotton-Weaving Industry of Japan 49 tial for the study of the evolution of rural entrepreneurs in the cotton-textile industry in Japan. The putting-out system is one of the most important elements in our study. In its histori- cal as well as in its contemporary form, the textile industry has had many kinds of putting- out systems. This contrasts with the industry's development in England, for example, where the factory production system replaced the putting-out system in an early stage of the industry's development. Putting out is still very visible in many places in the present tex- tile industry in Japan, not only in the cotton-weaving industry but also in the production of other textile materials, although the styles of putting out vary among materials. Iruma's case is useful for its illustration of the emergence of the putting-out system: it came into use only after 1890 in this region, although it was observed much earlier in other regions, and data are available for both the periods before and after the putting-out practice emerged, so it is possible to compare the two periods. Quality control in its broad meaning is quite important for the emergence of putting-out system in this area, although this is only one of several reasons we can think of as factors that generated the overall putting-out system; major reasons for the emergence and development of the putting-out system differ among regions. Overall Historical Picture of the Cotton Textile Industry of Japan By the latter half of Tokugawa period, there were already many established local cotton- production areas, and distribution networks were well developed. In this period the pat- tern of procurement of cloth by ordinary people took three forms: purchase of second-hand cloth, home production of cotton apparel from cotton material, and purchase of new cot- ton cloth from the market. Although most of the people in rural areas purchased second- hand cloth or made their own cloth from raw materials, some segments of the population purchased new cotton cloth. According to a study in Yamaguchi Prefecture (Chosyu), about 23 percent of the relatively high-income population purchased new cotton cloth (Saito and Tanimoto 1989). Japan was under autarky until 1859, and opening of the market to the rest of the world in 1859 greatly affected the structure of the industry. The export of silk grew rapidly right after the opening of the market, but the export of cotton fabric products did not show much activity in the early years. Both yarn and fabrics were imported from foreign coun- tries, and imported yarn replaced handmade domestic yarn. The availability of inexpensive imported yarn changed the competitive positions of pro- duction areas. Iruma took advantage of imported yarn and expanded its share in the na- tional market. Some of the production areas, which not only wove textiles but also spun their own yam, lost their competitive position as the cheap yarn was imported from abroad.2 Import of cotton yarn far exceeded domestic production soon after the opening of the market. Figure 3-1 shows the magnitude of imports and domestic production of machine- spun cotton yarn from 1874 until 1914. According to figure 3-1, domestic production reached 50 Towvard the Rural-Based Development of Commerce and industry Figure 3-1. Output and Trade of Cotton Yarn, 1874-1914 Million yen 200 100 100 ~~~~~~ Production A. 50 Production B 1n , o ~~~~~~~~~~~~~Export 10 1~~~~~ , 5 _,s* / / Import r 1874 1880 1890 1900 1910 1914 Production A is based on the estimate by Shinohara. Production B is based on data of "Kindai-teki Boseki Goisha." Source: Abe (1990). the level of imports only in the mid-1880s. The large volume of imported yam from the opening of the market until the 1890s played an important role in the development of new production areas such as hruma, as will be explained in the next section. The domestic production of cotton yarn by large factories started in the 1880s. As shown in figure 3-1, production expanded rapidly in the 1880s and 1890s. Growing use of cotton yam by domestic weavers was the most important force supporting the expansion of do- mestic yarn production. Although exports of cotton yarn started in 1890, the share of ex- ports held by spinners of domestically produced yarn was less than 6 percent in 1895. The use of their own cotton yarn by the textile division was negligible in 1885, and most of the yarn produced was supplied to independent weavers in Japan. In this sense, fabric pro- duction in localities throughout Japan played an important role in sustaining the growth of domestic yarn production. Figure 3-2 shows the trends in cotton fabric markets from 1860 until 1895. Import of cotton fabrics expanded rapidly from the opening of the market until the early 1870s. At Rural Entrepreneturs in the Cotton-Weaving Indutstry of Japani 51 the peak, imports accounted for about 34 percent of domestic consumption. Domestic de- mand also expanded rapidly in this period, increasing threefold from 1875 to 1895. Behind this high growth of domestic demand was the changing pattern of consumer cloth pro- curement. As previously mentioned, before the opening of the market, second-hand use and home production were the dominant forms of clothing procurement, and purchase of new fabrics was restricted to relatively wealthy families. In the latter half of the nineteenth century, however, procurement of cloth shifted from second-hand use and home produc- tion to the purchase of new fabric. Reflecting this changing pattem of demand, the growth rate of demand for fabrics in rural areas exceeded the national average. According to Saito and Tanimoto (1989), the increasing income of farmers and the fall in the price of cotton fabrics relative to the prices of other goods were crucial factors in this growth in demand. As we can see in figure 3-2, both domestic demand and domestic production grew rap- idly. As we will see below, there was a clear difference between imported fabrics and do- mestically produced fabrics in this period. It is important to note that the domestic distribution system played an important role in meeting the expanding demand and the Figure 3-2. Cotton Fabric Demand and Supply in Japan, 1860-95 10 thousands (volume) (log) 20,000 10,000 5,000 Domestic demand Domestic production 2,000- 1,000 _ 500 Import 200 I ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~ l I I II 1860 1865 1870 1875 1880 1885 1890 1895 Source: Saito and Tanimoto (1989). 52 ToTvard the Rural-Based Development of Commerce and Induistry needs of domestic production. Newly rising wholesalers in Tokyo, such as Chogin and Benichu (Chu Itoh), were actively engaged in expanding their sales territories into such rural areas as Tohoku, Hokkaido, Chugoku, Shikoku, and Kyusyu, all located far from centers such as Tokyo and Osaka. We will see the role of these large wholesalers in the next section. There were many production regions for cotton fabrics in Japan. After the opening of the market, there was substantial restructuring in these regions. Roughly speaking, the production regions were classified into two large categories: one included the regions en- gaged in mass production of white cotton fabrics, and the other encompassed the regions that produced small volume, differentiated cotton fabrics. In the first category, large firms with both spinning and weaving divisions contributed a large portion of the national pro- duction in more recent years (this analysis is based on Tanimoto 1986). There were also several production areas where the white cottons were produced under the putting-out system. There were numerous local production areas, each engaged in the production of differentiated products, and each specialized in its own product list. Iruma, which we will discuss in detail in the next two sections, was typical of this category. The products of local production networks were generally for domestic use. Differen- tiated textiles produced in small volume lacked international competitiveness, but the products were accepted by the domestic market. Only the products of large spinning- weaving firms were exported in the 1890s, but export of the products of local networks became established, gradually expanded, and became an important part of Japanese exports in the 1920s. Because the focus of this chapter is the rural production system, attention will largely be restricted to local production networks. As will be discussed in greater detail later, the putting-out system played an important role in these local production networks. Accord- ing to Abe (1990), among the 448,609 weaving firms (or houses) registered in the 1905 statistics of the Ministry of Agriculture and Commerce, 35 percent were independent weav- ers and 65 percent were under putting-out contracts. Among the independent weavers, 1 percent were relatively large factories that employed more than ten workers; 31 percent were small, independent weavers with fewer than ten workers; and 3 percent were whole- salers. (Note that these shares were calculated in number of firms, or houses). Putting out was thus a common practice of the industry in this period, at least in the number of the producers involved. The putting-out system was in place in the Tokugawa period, and two arrangements could be observed. One was Watagae Sei (cotton offer system) and the other was Dashibata Sei (machine-rental system). Under Watagae Sei, wholesalers provided farmers with raw cotton and received the outputs. The farmers' fees were often paid in raw cotton (or ginned cotton). Under Daslhibata Sei, wholesalers provided farmers with looms and cotton yam and collected the finished fabrics. Although there were many areas where wholesalers engaged in the simple sale of materials and the purchase of the outputs (thus, no putting- out), there were also some regions where the putting-out systems were observed. As we will discuss below, putting-out was not active in Iruma until the 1890s. Rutral Entreprenteuirs in the Cotton-Weavfing Indtustry of Japani 53 The Early Form of the Local Production System: Iruma, 1850-80 The Iruma area offers an excellent illustration of the functioning of a local textile produc- tion network in the early period (the latter half of the nineteenth century).3 Cotton prod- ucts in Iruma gained expanded markets through the activities of local merchants, who were supplied by the farmers of the region. The construction of a sales network with a local center was crucial to gaining access to the national market for the products of Iruma. The distribution network, supported by a sophisticated payment system, promoted the expansion of sales. Hachioji City, the local center for Iruma, was successful in collecting products of the region, and newly emerging large wholesalers in national centers such as Tokyo and Osaka considered Hachioji City as one of the important sites for their procure- ment of textile products. Distribution of the products was supported by these traders, and production processes depended on the work of wives or daughters of the farmers and were carried out in the farmers' "houses" as domestic industries. Iruma, in Saitama Prefecture, is a typical rural production area of the cotton textile indus- try. It is located in a rural agricultural area, about 30-40 kilometers northwest of Tokyo. The production of cotton textiles in the Iruma area started in the beginning of nineteenth century, and it had become one of the leading new cotton textile production areas of Japan by the 1890s. This area produced differentiated cotton fabric products, primarily for domestic uses. Thus, this production area is in the second category in our classification system. The structure of the production and distribution of the products produced in this region in the period from 1850 through 1880 is illustrated in figures 3-3, 3-4, and 3-5. Figure 3-3 illustrates the production relationships in a village, and figures 3-4 and 3-5 illustrate the distribution network for the products. The main players in this system are (1) local mer- chants near the village, (2) local wholesalers near the local city, (3) farmers, (4) yam mer- chants, (5) financial intermediaries, and (6) national-level wholesalers. Local merchants played the role of channel leaders in this local production network. Channel leaders connected the producers, who were widely scattered among several vil- lages, with the distribution system that had been organized by local and national whole- salers. They purchased cotton fabrics at local markets or directly from farmers and sold them to local wholesalers. The local market (ichi) was the center for local transactions, and it opened regularly on predetermined dates in each town.4 Some local merchants who collected cotton fabrics from farmers brought the fabrics to the markets for sale; others purchased the fabrics at the market and sold them to local wholesalers. Tanimoto (1986) studied the transactions of a typical local merchant, the Hosobuchi family. The record reported an interesting pattern of change in the transactions this family organized. In the early period (the 1850s), the family purchased the products from farmers and sold them in the local markets (see figure 3-4); the family's transactions have expanded considerably since then. Figure 3-5 illustrates the family's pattern of transactions in the 1880s, when they expanded considerably and the family purchased their products not only from farmers but also in local markets. The family sold the products to the local wholesalers living in Hachioji (the neighboring large, local city). The Hosobuchis expanded 54 Towvard the Ruiral-Based Development of Contmerce and Indutstry Figure 3-3. Production Process of Cloth in Iruma District, 1850-80 Local merchant Cloth Died yarn Farmer-weaver _ Dyeing White yam White yam Yam merchant White yam White yarn Imported Domestic yam yarn transactional relations with large local wholesalers in Hachioji City, and through these transactions the family expanded its business. Note also that there were established routes to the national market through these large local wholesalers during this period. Let us next look at the relationships between farmers and local merchants (figure 3-3). In this period we cannot yet observe the putting-out system. Local merchants simply pur- chased finished cotton fabrics from farmers with cash. Farmers purchased wooden, handheld looms and yarn, and they made fabrics at their own risk. They then sold the fabrics to local merchants. Note that the problem of incentives, which will be discussed in the next section, was not serious in this system-farmers accepted the risk. The price of a wooden, handheld loom was about 1.5 yen, while the output of cotton fabric is about 1 yen per tan (3.3-3.5 square meters); thus, the prices of loomns were not expensive compared with the prices of fabrics. Although we cannot calculate cost of in- vestment with such limited information, we can say that the farmers could afford to buy looms by themselves. The low price of a loom was an important factor in the establish- ment of rural production networks. Rtural Entreprenieuirs in the Cotton-Weaving Iidustrsty of Japan 55 Figure 34. Distribution System of Cloth in Iruma District, 1850s Local market Local market Local merchant Xt Farmer-weavers In textile production we can observe a typical division of labor among family members. One family usually owned one or two looms. Male workers were engaged in agricultural production, and only female workers were involved in weaving. We can observe some seasonal fluctuation (figure 3-6) in the production of textile products, which reflects the cycle of agricultural activities.5 Female workers were involved in weaving activities when they were not asked to supply agricultural labor service. Figure 3-5. Distribution System of Cloth in Iruma District, 1873-74 Local market Local merchants Farmer- Farmer-weaver weavers . Local merchant Local wholesaler Large wholesaler Large wholesaler (Tokyo) (Osaka) 56 Tozward the Rutral-Based Developmenit of Commerce and Inidutstry Figure 3-6. Seasonal Fluctuation of Transactions Between Takizawa and Farmer-Weavers Index 110 100 - 90 80 - 70 60 18 50- 191 40 -* 30 . 20 : : | | 1 *- ~~~~~~~1907 20 10 Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. Farmers purchased yam from yarn merchants (figure 3-3), who lived in the local village and purchased domestic yarn from neighboring yam-production villages and imported yam from importers.6 The transactions of yam between yam merchants and farmers were usually accomplished through cash or draft, but deferred payments were also observed. Deferred payment made it easier for farmers to purchase yam for their production. This area originally purchased yam from a near-by (but not in the same village) production area through the channel of yarn merchants, but gradually local yam was replaced by imported yarn, in contrast with other production areas where both weaving and spinning were conducted in the same village. In the latter case, replacement by imported yam was delayed, and the delay hindered the development of the production region, since imported yam was much cheaper. Let us now discuss briefly the position of local wholesalers in the local commercial cen- ter. As mentioned above, Hachioji was the local center of this production area. It began as one of the local markets for textile products, but it gradually became the center in the region. Not only the cotton products of Iruma, but also other products, including silk, were collected in Hachioji and shipped to large markets such as Tokyo. Large wholesalers in Hachioji had branches in Tokyo and were involved in active transactions with large national wholesalers there. As mentioned before, new national wholesalers such as Chogin and Benichu (Chu Itoh) were very anxious to expand their sales territories. Some large Rutral Entrepreneuirs in the Cotton-Weazvinig Inditstry of Japan 57 merchants in Tokyo had strong ties with local wholesalers in Hachioji and purchased the cotton fabrics of Iruma. Through such a network, the products collected in Hachioji were shipped to the whole nation. In the beginning, cash was used for the transactions within the distribution system, but as the amount of transactions expanded, commercial credits were gradually introduced. Figure 3-7 illustrates the flow of payments among wholesalers in 1871. In this picture we have four examples: local wholesalers in the local center (Hachioji), local merchants col- lecting products from small traders and selling to the large local wholesalers, small local merchants collecting products from farmers and selling them to large local merchants, and financial agents. The transactions between the local wholesalers and the local merchants were paid by tegata (draft promissory notes issued by buyers), which were discounted by the financial agents. There is no commodity transaction between local merchants and the financial agents. The transactions between large and small local merchants were made by kazase (money orders issued by local merchants), much like a check using the account of the financial agent. Commercial credits were extended by local merchants to the local wholesalers. These credits were an important method for local merchants to explore new markets for their products. Nevertheless, active use of commercial credits also caused financial shortages for local merchants. The role of financial agents was important in this area, because local merchants could discount drafts through these agents. The financial network thus sup- ported the expansion of sales of products in Iruma to the national market. Financial transactions became more sophisticated later. Figure 3-8 illustrates the finan- cial flow in this region in the period from 1877 through 1882. Although the basic structure of the financial flow did not change from that shown in figure 3-7, there are more mer- Figure 3-7. Payment System (I), 1871 Draft Local wholesaler P Local merchant (Local center) , s ~~~~~~~Fabrics/ Draft Cash Draft Fabrics Money order Cash 4 Financial agent < Smaller merchant Money order 58 ToWard the Rtral-Based Developnienit of Conmmerce and Induistry Figure 3-8. Payment System (II), 1877-82 Draft Draft Local wholesaler A Local merchant > Smaller merchant Fabrics Fabrics Draft Cash Fabrics Money order Financial agent _ Farmer-weaver Money order chants involved in the transactions between small local merchants and local wholesalers. Financial credits by draft were used to facilitate the transactions. Development of the Putting-Out System Although putting out was not observed during the period of 1850-80, when Iruma be- came one of the large local production regions of cotton textile products, the practice of putting out gradually spread in this region after that time. The reason for the emergence of the putting-out system differs among the various pro- duction regions. In the case of Iruma, the quality control problem was a crucial factor that led to the emergence of the putting-out system. As production areas developed in Japan, competition among local production areas increased. In such a competitive environment, the quality of products became the crucial factor in their success. (By quality of products we mean not only such simple quality considerations as flaws, but also such elements as design and style of the products.) The cotton fabrics of Iruma became recognized as a local brand in the national market. Wholesalers, both local merchants and local wholesalers, became more conscious of these factors. In order to establish the Iruma brand, it became increasingly important for local mer- chants to standardize the quality of their products. To encourage this, the merchants started to give farmers dyed and sized yarn (not undyed yam) and to receive fabrics of specified pattern and color. The traditional transaction relation, in which farmers purchased yam, wove fabrics, and sold to the merchants, is unsuitable for standardization of products. The farmer does not have much information about the final consumer market or the capability Ruiral Entrepreneeurs in the Cottoni-Weavinig Inidutstry of Japan 59 to make quick adaptation to market demand. Without someone coordinating the quality control of a large number of independent farmers, it would have been impossible to estab- lish the local brand. The putting-out method was a natural choice by the merchants to achieve the needed standardization. Note that the putting-out system is not the only way to achieve standardization-fac- tory production may be a more desirable method. The comparison between Japan and England is quite interesting in this respect. In England, the factory production system was chosen. Hiring workers and monitoring their behavior in factories are certainly better av- enues to control quality, but this was not what happened in Iruma. To understand the reason why putting out, not factory production, was chosen in Iruma, it is useful to consider the ways in which the putting-out system is superior to factory production, and the ways it is not. One of the strengths of the putting-out system is its flexibility. By organizing a putting-out system, wholesalers can be more flexible about the size of their outputs. When the market is active, they can increase transactions with farm- ers, and when the market is weak, they can diminish transactions. In the case of factory production, however, the owner of the factory must make a commitment to the size of output because of the large amount of fixed capital. The other advantage of the putting-out system is the availability of cheap labor in the farm family. Only female workers were involved in the production of cotton products in Iruma, and they engaged in cotton production only when they did not have agricultural or household tasks. For these female workers, it was important to be able to stay at home because they could use their spare time for weaving. The opportunity costs for working in factories were quite high. From the viewpoint of farmers, the putting-out system offered far more flexibility in working hours than being employed in a factory. It also provided opportunities to enter the industry without much risk. Since farmers got substantial portions of their income from agricultural activities, fluctuation of income from weaving did not cause serious problems. Furthermore, it was possible for farmers to have putting-out arrangements with more than two merchants, which helped them cope with demand fluctuation. The farmer- weavers were thus less dependent on merchants than workers are on their factory manag- ers in several important areas. Of course, the factory production system is better than the putting-out system in sev- eral respects. First, quality control is far easier under factory production than under putting out; this is particularly so in the issue of embezzlement, as will be discussed below. Second, it is much easier to achieve production stability under the factory system than under the putting-out system. Because the labor of subcontracting weavers fluctu- ates with the schedule of agricultural activity, merchants should have had difficulty achieving production stability, but flexibility of production, mentioned above as one of the merits of putting-out system, and instability of production are the two sides of the same coin. Third, factory production is better than putting out when technology is sub- ject to scale economies. When powered looms were introduced into this industry, fac- tory production replaced a large portion of putting-out production by farmers. Scale 60 Towvard the Rural-Based Development of Coninerce and Indutstry economy based on expensive machines makes factory production more suitable than putting-out arrangements.7 Both the factory production and the putting-out systems have strong points, and it is not possible to conclude which system was better without specifying the environments the industry faced. The more frequent use of the putting-out system in Japan than in En- gland implies that there were several factors in Japan that made putting out more suitable. The following points from our case study of Iruma offer insight into these differences. Table 3-2 shows how the merchants of Takizawa coped with fluctuation of demand. This table details the distribution of the amount of transactions for each farmer and the number of farmers among the categories of transactions. The merchants had a large vol- ume of transactions with the farmers who had long-term transactional relations with the merchant, while there were a large number of farmers with whom the merchants had only small transactions, and only during a given year during the period of 1896-1925. The local merchant could cope with fluctuation in demand in this fashion. The farmer-weavers who had short-term transactional relationships with this merchant had transactions with some other merchant or merchants as well. As will be discussed in the next section, that is the case even under the putting-out arrangement in the contemporary textile industry. The long-term transactional relation is crucial for the maintenance of quality. Under long-term transactions, farmer-weavers have reduced incentives to cheat.' The farmer- weavers might compare the short-run gain from cheating (cheating of many forms, from such obvious actions as embezzlement to more subtle cheating, such as less effort toward quality improvement) with the long-run loss of losing transactions with the merchant. The longer the transactional relationship was in place, and the larger the rent from the transac- tions, the less were the incentives to cheat. Table 3-2. Distribution of TransactionYears inTotalTransactions, 1896-1925 Term of trans- Volutme of trans- action between actions betwveen Share in total Takizawvas and Takizazvas and volume of each farmer- Number of farmer-wveavers transaction wveaver (years) farnmer-weavers (1,000 tan) (percent) 1 623 24 5.4 2 272 35 7.8 3-4 195 56 12.2 5-9 180 156 34.0 10-14 47 117 25.6 15-19 12 43 9.3 > 20 6 26 5.7 Total 1,335 457 100.0 Note: "Tan" is a Japanese unit of cloth, about 3.5 square meters: width 0.35 meters, length 10 meters. Source: Archives of Takizawas. Riral Enitrepreneuirs in the Cottoni-Weaving Induistry of Japan 61 Note that the issue of embezzlement was an important one for putting-out arrange- ments (see, for example, Styles 1983 and Randall 1990). When farmers purchased yarn by themselves, embezzlement could never arise. Under the factory production system it was difficult for workers to cheat by stealing yarn. Thus, embezzlement (that is, cheating of yarn) can occur only under a putting-out arrangement. When farmers were provided with yam from merchants, they had incentives to take some yarn for their own use. Thus, un- der the putting-out arrangement, merchants must prepare some mechanisms to discour- age cheating. The long-term transactional relationship is one such mechanism. By forming a combination of a large volume of transactions with a small number of farmers through long-term transactions and a small volume of transactions with a large number of farmers through short-period transactions, merchants could achieve two goals. They received products of good quality from the farmers with whom they had long-term transactional relations, and they gained flexibility in the amount of fabric they received through spot transactions with a large number of farmers. The question then arises of why the putting-out system was not used in England, in spite of the advantages it offered. We do not have a good explanation now, but population density may be one factor that explains the difference between Japan and England. Popu- lation was much denser in Japan than in England, and dense population allowed whole- salers to organize a putting-out system in a relatively narrow region. Geographical concentration permitted wholesalers to have better monitoring. Human relationships within the community and other social ties might provide better incentives for farmers not to cheat in the Japanese context. Table 3-3 provides some insight in this matter. It illustrates the transactions of one local merchant in 1897 and 1918 and the amount of transactions in each village and their share in total transactions. As the table demonstrates, this merchant's transactions be- came much more geographically concentrated in 1918 than they were in 1897. Thus, as putting-out practices expanded, the merchant made the geographical range of his trans- actions smaller. Such restriction of geographical distance gave merchants better control of quality. The role of public support is also of significance here. As in other local production areas, public support by local government played an important role in the development of the Iruma production area.9 For the case of Iruma, the roles of its industrial association and the technical school should be mentioned. In 1900 the Japanese government enacted Jutyo Buissan Dogyo Kurmiai Ho (Law on Indus- trial Associations of Important Products). Under this law, an industrial association would be established in a local area if there was consensus on the formation of the association by more than two-thirds of the producers in a given product category; all producers are then required to be members of the association. Under this law, the association must include members who serve as inspectors. These members check the products of the association members and punish any cheating. Based on this law, the Musashino Orimnono Dogyo Kurmiai (Industrial Association of Weav- ers in Musashino) was established; the association has been active since around 1910. 62 Toward the Rurral-Based Developmenit of Commerce and Inidutstry Table 3-3. Geographical Distribution of Transactions 1897 1918 Transactio7t Transaction volume between voliume betweent Takizazvas Share Ta kiza was Share andfarmiier- in total andfarmer- in total weavers tratsactions wveavers transactions Village or townti Section (tan)a (percent) (tan) (percent) North Irumagawa Unoki 200 1.8 East Iruma Iriso 60 0.3 Mizutomi Sasai 228 1.0 Toyo'oka Takakura 493 4.5 6,123 27.0 Machiya 386 1.7 Higashikaneko Koyatab 5,110 46.7 8,541 37.7 Araku 1,010 9.2 4,661 20.6 Motokaji Noda 590 2.6 Bushi 115 0.5 Kaji Iwasawa 38 0.3 Asu 13 0.1 117 0.5 Miyadera Ogayado 12 0.1 Omori 135 0.6 Motosayama Nihongi 142 0.6 13 km Kaneko Negisi 138 1.3 781 3.4 Nakagami 12 0.1 85 0.4 Hananoki 90 0.4 Yaganuki 26 0.2 564 2.5 Kamiyaganuki 77 0.7 Teratake 363 3.3 Mine 286 2.6 Mokurenji 1,231 11.2 Kasumi Nanokaichiba 169 1.5 Imai 986 9.0 Imadera 65 0.6 Daimon 286 2.6 Fujihashi 412 3.8 South Yano 13 0.1 West Chofu, Nogami 12 0.1 etc. 13 0.1 13 0.1 Total 10,953 100.0 22,643 100.0 Sum of top three sections 7,351 67.1 19,325 85.3 a. See the footnote to table 3-2. b. Takizawas (= putters-out) is located in Koyata (Higashikaneko village). Source: Archives of Takizawas. Ruiral Entrepreneutrs in the Cotton-Weavinig Indiustry of lapani 63 Seizogtlo-bu Seimien Butkai (Young Members' Group of Weavers) was formed within the as- sociation, and the members of this group visited weavers and examined the quantity of yam provided and the fabrics produced in order to check the possibility of embezzlement. In order to establish the standardization of the products in the area, the association prohibited farmer-weavers from using looms that did not pass its inspection. The associa- tion also required members to use only the dyeing factories authorized by the group. Another association (Hosho Sekinin Muisashi Orirnon o Shinyo Hanbai Koutbai Kurmiai) was also established. The purpose of this group was to obtain subsidized financial support from both the national and local governments to purchase yarn. The prefecture government (local government) started the Kazwagoe Senshokut Gakko (dye- ing school) in 1907. The people who were educated in this school became the core mem- bers of a new generation of local entrepreneurs in this area. The engineering teachers in the school were also involved in the inspection activities of the association (see Tokorozawa- shi 1989). It is not easy to evaluate how effective these local government supports were in the development of the Iruma production area. It should be noted, however, that the local government made an effort to nurture the local production area. This is not unique to Iruma, or to the weaving industry. Similar supports can be observed in many other areas and in many other industries in Japan. The Contemporary Putting-Out System Although our main concern in this chapter is historical forms of the putting-out system in the cotton textile industry of Japan, it is useful to review briefly the contemporary forms of the putting-out system. It is important to note that the putting-out system is common practice in the contemporary textile industry of Japan. Putting out is more commonly ob- served in Japan than in other countries, not only in the textile industry, but also in other industries such as automobile production and electronics (see Itoh and Urata 1994 for the structure of the putting-out system in the production of automobile parts, silverware, and synthetic-fiber textiles). Two cases of the putting-out practice in the contemporary textile industry are illustra- tive."' One is the case of synthetic-fiber textile industry, where extensive putting-out rela- tionships can be observed between large chemical companies (or wholesalers) and weaving houses. The other is the case of the weaving of wool fabric, where a large number of small weavers have putting-out relationships with parent weavers. In the synthetic-fiber textiles industry, a large portion of weavers have putting-out con- tracts, either with large chemical companies or local wholesalers. In a typical putting-out contract between chemical companies and weaving houses, the chemical companies, which are large, listed companies, produce synthetic-fiber yarn from chemical process and dis- tribute these yarns to weaving houses. Woven fabrics are returned to the chemical compa- 64 Touard the Ruiral-Based Development of Conmnerce and Indutstry nies for the finishing process, and chemical companies sell these products to the market. Weaving houses are the subcontractors for chemical companies; weaving houses do not engage in the marketing of their products, nor do they take any risk for selling them. The chemical companies not only play the role of manufacturers as the producers of synthetic- fiber yam, but also the role of marketing organizer. The role played by chemical compa- nies has some similarity to that of the merchants in our Iruma example. The local weaving houses depend heavily on chemical companies for services such as marketing, product development, and risk-taking. All the weaving houses in the synthetic-fiber industry are small and medium-size firms; employees number fewer than 300, and the majority of the houses employ fewer than 50. Through their relationship with the chemical companies, the small and medium-size local weaving houses can survive in a very competitive mar- ket environment. Large chemical companies also gain from the putting-out system. The chemical com- panies can enjoy the flexibility of the putting-out system, as in the case of Iruma, ex- plained earlier. They can adjust the amount of production by adjusting the orders they give to weaving houses. As in the case of Iruma, the subcontracting weaving houses are ranked according to the degree of their relationship with each chemical company. Stable orders are given to the weaving houses with the closest ties to a given chemical com- pany, while the orders to less firmly aligned weavers fluctuate considerably, depending on market conditions. The latter type of weaving houses have transactional relation- ships with a number of chemical companies as well as with local wholesalers so that they can spread their risk. T'his system of putting-out in the synthetic-fiber textile industry emerged in the postwar period. After the invention of new synthetic fiber materials, a group of leading local weav- ing houses asked large chemical companies to start a putting-out system. This request was based on the weavers' lack of adequate financial resources to purchase expensive syn- thetic-fiber yarn. This request from weaving houses for putting-out contracts also ben- efited the chemical companies, because they did not have to make an investment in the weaving process. The existence of the network of weavers in local areas made it possible for large chemical companies to enjoy rapid expansion in their production of synthetic fiber textiles. It is also important to note that this contemporary putting-out arrangement allowed traditional small-scale weaving houses to survive the modem technology of syn- thetic fibers. The leading local production areas of synthetic-fiber textiles are located on the Japan Sea side of Japan (Fukui and Ishikawa Prefectures), which was originally the production area of traditional silk textiles. This local production network started more than 100 years ago and has gradually changed its products. In that sense, we can observe historical con- tinuity. There had been substantial concentration of producers in this area-not only weav- ers, but also finishing companies, textile-machine producers, and textile wholesalers. This production area also nurtured the textile-machine industry and the dyeing and printing industry; some of the modem technology firms emerged from these industries. Rural Enitrepreneuirs in the Cotton-Weaving Industry of Japan 65 The next example of the use of the putting-out system is in the weaving of wool textiles (the following is based on Findlay and Itoh 1994). Ichinomiya is the largest production area in Japan for wool, and there are many weaving houses there. Contrary to the case of synthetic-fiber textiles, weaving houses of woolen cloth are independent from large spin- ning companies. Weaving houses purchase wool yarn from spinning firms at their own risk and sell their woven fabrics to apparel companies; putting out is not as common be- tween wool spinners and weavers. Although there are some weaving houses that are sub- contractors for spinning firms, they do not represent the majority of the industry. We can, however, observe another form of the putting-out relationship, that between large and small weaving houses. According to a study in the late 1980s, there were about 250 parent weaving houses and about 4,000 to 5,000 subcontracting weaving houses in the Ichinomiya area. We can observe extensive putting-out transactions between parent weav- ing houses and subcontracting weaving houses. The subcontracting weaving houses are usually family-run businesses, and each weaving house generally employs fewer than three workers. These small weaving houses accept orders from larger weaving houses, and they are given yarn and the specifications for the fabrics they are to weave and are paid at the piece rate-it is a typical putting-out relationship, with all the risks borne by the parent weaving houses."1 The structure of the putting-out relationship between large and small weaving houses is similar to the case of Iruma or that of contemporary synthetic-fiber textiles. For a large parent weaving house, subcontracting small weaving houses are classified according to the closeness of the relationship. Some weaving houses are completely dependent on a single parent weaving house, while others have transactions with more than two larger weaving houses. Parent weaving houses tend to offer stable orders to their dependent small weaving houses, while they adjust the amount of their orders to less dependent weaving houses. The latter category of weavers transact with more than two parent weav- ing houses, as well as with wholesalers, in order to stabilize the orders they receive. As in the case of cotton textiles in Iruma and synthetic-fiber textiles in Fukui and Ishikawa, this form of the putting-out arrangement allows parent weaving houses to cope with de- mand fluctuations. Long-term relationships between parent weavers and some small weav- ers are the basis for their cooperative relationships. It is important to note that there is a wide variety of products, even in the narrow cat- egory of wool fabrics. It included not only a wide classification of male and female prod- ucts, but also finer classifications, such as fashionable products and basic products, piece dying and yam dying, mono color products and complicated pattern products, and so on. The skills needed to produce each product are different, and there is some room for spe- cialization. We can observe specialization of each small weaver in some kind of product. Through such specialization and the putting-out arrangement, the production network can foster efficient skill accumulation and product diversity. The wool production area of Ichinomiya has a long history in textiles. It was originally an area of cotton and silk fabric production and gradually shifted to wool fabrics. We can 66 Towvard tlhe Ruiral-Based Development of Commerce anid indusbry thus observe historical continuity in this area, and the present putting-out system is the product of that history. Conclusions Comparing the two examples of the putting-out system in the contemporary textile indus- try with the historical case study of Iruma, we can observe some similarities. They can be summarized as follows. * The putting-out system allows flexibility for parent firms (wholesalers). By adjusting the orders to small weavers, parent firms (or wholesalers) can change the amount of total production as well as the allocation of products. * We can observe long-term transactional relations with parent firms and some (but not all) small weavers. The long-term relationship provided each side with incentives to be cooperative with the other, and the problem of embezzlement is mitigated. l The putting-out system often has historical corftinuity, and the use of agricultural labor during the seasons of low farm activity is an important factor in the emergence of the putting-out system. As time goes on, however, and as the production area is established, small weavers tend to specialize in selected weaving activities. - The putting-out relationship can continue while changing its contractual form sub- stantially in response to drastic technical change, such as the introduction of new materials. - The traditional networks of the putting-out relationship often become a nursery of modem technology firms, such as machine-manufacturing, dyeing, and printing firms. Many modern Japanese firms were nurtured by the textile industry-for example, Toyota was originally a loom-producing company. It is surprising that a traditional industry such as textiles can survive in a high-wage country such as Japan in a competitive international market environment. Findlay and Itoh (1994) point out the importance of the domestic market for the survival of the textile industry and the role the putting-out relationship can play in the industry. Flexibility and risk-sharing capacity supported by a subcontracting system are crucial if the industry is to supply differentiated, high-quality products. It is interesting to note that a very similar reason can be given for the emergence of production areas such as Iruma. It is dangerous to jump to the easy conclusion that the domestic market is an important driving force in the emergence of rural enterprises. We can, however, observe a similar pattern in the development of rural enterprises in other industries, such as silverware. More research should be carried out to study the distribution network in the domestic market. As explained in the discussion of lruma, not only the putting-out relationships of the local production area but also a nationwide network of wholesalers are crucial for the emergence and growth of rural production areas. Rutral Enitrepreneuirs in flte Cottoni-Weavintg Indutstry of Japan 67 Although the role of public support has not been examined in this chapter, it is impor- tant to note that public support, especially that of local government, played an impor- tant role in the development of local production networks. As discussed in detail in Itoh and Urata (1994), local government often played an important role in nurturing local production networks in Japan. Numerous policy measures were introduced by local government, including establishment of local industrial associations, recruitment of highly skilled engineers from other parts of Japan, and the purchase of machinery from abroad. Such local government supports did play some role in lruma. It is an important feature of Japanese rural enterprises that a concentration of similar producers can be observed. Local public technical centers and industrial associations were often channels for tech- nological information. There is another area that is not touched in the chapter-the role of export in the devel- opment of local production areas. The textile industry became an important export indus- try for Japan, and expanding exports most certainly affected the structure of local production networks. In the case of silverware reported in Itoh and Urata (1994), the structure of the industry changed substantially when the production area shifted from traditional local production areas to an export center for the products. The Inrma production area is not a good example for this kind of study, because it disappeared when powered looms were introduced. There are, however, several production areas of textile products in Japan that grew with expanding exports.'2 Notes 1. Because of limited availability of data, Tanimoto utilized a few different sources of business documents of the merchants and connects them. 2. Tanimoto (1992) compares two (fabric-making) production areas -one purchased cotton yam from other producers and the other purchased cotton material and spun the yam themselves. It was easier for the former kind of producer to switch from domestic to imported yarn; yam merchants established in the production area played an important role in the shift to imported yarn. In the latter type of production area, where cotton yarn was produced from raw cotton, incentives to switch to imported yarn were much weaker. According to Tanimoto, this difference in approach to imported yarn was reflected in the pattern of regional industrial growth. The former expanded their share while the latter declined toward the end of the nineteenth century. 3. Nevertheless, large firms that produced both yam and fabrics were not as important in the nineteenth century in the cotton textile industry in Japan. Their share in the total production of cotton textiles in Japan was about 5 percent in 1889. Their presence became more visible after the Japan- Russia War (1904-5), with a share of 22 percent in 1905 and 41 percent in 1914. Note that the remaining products were those of local production networks, which depended on the putting-out system. 4. For example, in a local town, Ogimachiya village market opened on the 3rd, 8th, 13th, 23rd, and 28th of every month. 5. Major agricultural products in this area were barley, wheat, tea, and sericulture, which re- quired intensive labor input (especially female labor input in the latter two) in May, June, and July. 68 Towvard the Riral-Based Developirent of Coninmerce and Indlisty 6. According to Tanimoto (1986), a local yarn merchant in Iruma had direct transactions with yarn traders in Tokyo. 7. As explained in the next section, however, even now we can observe various kinds of putting- out arrangements in the textile industry of Japan. Thus the shift to powered looms did not destroy the putting-out arrangement in this industry. 8. This behavior pattern is analyzed in various ways by the theory of repeated games. 9. See Itoh and Urata (1994) for the role of government support in the development of local production networks. 10. The following cases are based on Itoh and Urata (1994), for the synthetic-fiber textile industry, and on Findlay and Itoh (1994) for the wool textile industry. 11. Note that the putting-out relationship can also be observed between wholesalers and small weaving houses and between spinning firms and small weaving houses. Here, we do not discuss these putting-out systems. 12. Kikuchi (in this volume) offers an interesting case study of rural entrepreneurs who depend on the export market. References Abe, Takesi. 1990. "Menkogyo" ("Cotton Textile Industry"). In Shunsaku Nishikawa and T. Abe, eds., San qvoka no Jidai (Tlhe Era of Inidutstrialization), NihonI Keizaislii (japanese Economic History), Vol. 4. Tokyo: Iwanami Shoten. Findlay, Christopher, and Motoshige Itoh. 1994. Wool in1 Japan. Pymble, NSW, Australia: Harper Education. Itoh, Motoshige, and Shujiro Urata. 1994. Small and Mediuim Enterprise Suipport Policies in Japan. Wash- ington, D.C.: World Bank. Randall, Adrian J. 1990. "Peculiar Perquisites and Pernicious Practices: Embezzlement in the West of England Woolen Industry: 1750-840." Interniationtal Revieiv of Social History 35(2): 194-219. Saito, Osamu, and M. Tanimoto. 1989. "Zairai Sangyo no Saihensei "("Restructuring of the Indiginous Industries"). In Mataji Umemura and Yuzo Yamamoto, eds., The Opening of the Ports and the Meiii Restoration (Kaiko to Ishin), Nihon Keizaishi (Japanese Econiom0iic History), Vol. 3. Tokyo: Iwanami Shoten. Styles, John. 1983. "Embezzlement, Industry and the Law in England, 1600-1800." In Maxine Berg, Pat Hudson, and Michael Sonenscher, eds., Manufactutre in Tovn and Counltry before the Factory. Cambridge, U.K.: Cambridge University Press. Tanimoto, Masayuki. 1986. "Bakumatsu-Meiji Zenki Menorimonogyo no Tenkai: Saitama-ken irumagun o Chusin ni shite" ("Development of Cotton Textile Industry in the Late Tokugawa and Early Meiji Period: Centering on the Iruma County in the Saitama Prefecture"). Shakai Keizai Shigaku 52(2) (June): 151-84. Tanimoto, Masayuki. 1992. "The Evolution of Indigenous Cotton Textile Manufacture before and after the Opening of the Ports." Japan Year Book oni Businiess Histony, No. 9:29-56. Tokorozawashi-shi Hensan-shitsu (The Committee to Edit the History of Tokorozawa City). 1989. "Tokorozawa Orimono Sanchi no Keisei to Hatten" ("Formation and Development of Tokorozawa Area Weavers Production Network"). Tokorozawa: Tokorozawa Municipal City Office. Yamazawa, Ippei. 1984. Nihoon no Keizai Hattent to Kokiusai Bun jqvo (Economic Development of Japan and Initernationial Division of Labor). Tokyo: Toyo Keizai Shinposha. 4 The Formation of Toyota's Relationship with Suppliers: A Modern Application of the Community Mechanism Kazuo Wada Jn 1980 Japan produced over 11 million vehicles and became the largest car-producing country in the world, surpassing the United States. The Machinle That Changed the World (Womack, Jones, and Roos 1990) revealed that Japanese assemblers also generally bested their American and European counterparts in both quality and production costs in the 1980s. Many researchers have tried to identify the factors that contributed to the superior performance of Japanese automobile assemblers. Some have pointed to the unique ap- proach to research and development of the assemblers and parts suppliers; others thought the "noninventory" production method was chiefly responsible for the better performance. The findings were all closely examined; some were implemented in overseas assembly plants and contributed to lower costs and improved workmanship. The Japanese automo- bile industry thus has provided fertile subject matter for journalists, engineers, businesspersons, and academics. Although the topics raised by the researchers were varied, most of the research seemed to suggest a common conclusion: producing better automobiles at less cost requires a smooth and coordinated flow of information, parts, and materials between assemblers and sup- pliers. It is not an easy task to establish such a relationship, but most researchers admit that Japanese assemblers, especially Toyota, have better relationships with their suppliers than assemblers in other countries have managed to form. Japanese assemblers transact directly with a few suppliers for the long term. They did not build their relationships with suppliers "primarily on trust, but on the mutual interdependence enshrined in the agreed- 69 70 Towoard the Rural-Based Developmenrt of Commerce and industry upon rules of the game" (Womak, Jones, and Roos 1990: 155). Nevertheless, mutual inter- dependence does not mean a lack of competition among suppliers, who strive to improve their own performance in order to obtain more orders from assemblers. The Japanese as- semblers appear to adopt a policy of "cooperation and competition among the few." This "improved subcontracting system is now considered a major organizational innovation that underlies the strength of Japanese industries, especially the automobile industry" (Hayami and Kawagoe 1993: 172.) As is the case for Toyota, this improved subcontracting system originated in manage- ment efforts of relatively recent years, mainly since the 1950s (Wada 1991). If Toyota had not established close personal relationships between the company and its suppliers in earlier years, however, it would have been more difficult for it to undertake organizational innovation after the 1950s and to create the present, highly sophisticated subcontracting system. This chapter argues that in the early years of Toyota, personal connections facilitated the establishment of close relationships between Toyota and its suppliers, especially up to 1955, when Toyota marketed Crown, "Japan's first completely domestically produced passenger car" (Kato 1981: 62). Historical Background of Automobile Manufacturing in Japan The Nascent Years In 1902 the first attempt was made to manufacture a motor vehicle in Japan. Shintaro Yoshida, a bicycle merchant, and Komanosuke Uchiyama, an engineer, collaborated to produce a passenger car and a bus, using foreign-made engines. Several other attempts followed.' Despite these efforts, as late as the 1920s, Japanese engineers could not produce automobiles on a solid commercial basis. The use of automobiles also remained very lim- ited in Japan. This significantly changed in 1923, after a major earthquake devastated the Tokyo area and destroyed railway and other transportation networks. The City of Tokyo imported 800 Ford Model T trucks to serve as buses. The number of motor vehicles im- ported then jumped from 1,938 in 1923 to 4,063 in 1924 (see table 4-1). This was, in a true sense, the first acquaintance of ordinary Japanese people with motor vehicles. The large order for trucks prompted Ford to establish a fully owned subsidiary in Japan, so that in 1925, Ford of Japan began to assemble knockdown parts in Yokohama. General Motors followed suit, setting up an assembly plant in Osaka in 1927. Most of the motor vehicles were used as buses or taxis in urban areas; in major cities such as Tokyo and Osaka, they provided indispensable transportation services. By 1930 the demand for motor vehicles had increased, and the existing Japanese companies suffered from the domination of the market by American subsidiaries. The growing demand for automobiles, including three-wheel vehicles, laid the ground- work for the emergence of automobile parts suppliers. After Ford and General Motors had The Fonrmationi of Toyota's Relationship with Suippliers 71 Table 4-1. Imports of Automobiles and Automobile Parts Auttonmobiles Automobile Cash anioliunt parts Total Nuimber (A) A/C (B) B/C (C) Year of units (thouisand yen) (percent) (thousanid yen) (percent) (thoutsanid yen2) 1914 94 421 62.0 258 38.0 679 1915 30 71 42.8 95 57.2 166 1916 218 387 54.2 327 45.8 714 1917 860 1,570 58.8 1,098 41.2 2,668 1918 1,712 4,525 59.1 3,137 40.9 7,662 1919 1,579 5,532 49.0 5,751 51.0 11,283 1920 1,745 4,866 46.4 5,614 53.6 10,479 1921 1,074 3,262 40.4 4,806 59.6 8,068 1922 752 2,262 30.7 5,094 69.3 7,355 1923 1,938 2,956 25.7 8,528 74.3 11,483 1924 4,063 8,773 41.4 12,414 58.6 21,187 1925 1,765 4,601 39.4 7,062 60.6 11,662 1926 2,381 5,325 33.9 10,392 66.1 15,717 1927 3,895 8,064 44.1 10,219 55.9 18,282 1928 7,883 13,771 42.7 18,475 57.3 32,245 1929 5,018 9,546 28.4 24,063 71.6 33,609 1930 2,591 4,897 23.6 15,877 76.4 20,774 1931 1,887 3,379 20.7 12,952 79.3 16,330 1932 997 2,895 19.5 11,928 80.5 14,822 1933 491 11,865 49.7 12,007 50.3 23,872 1934 896 3,358 10.4 28,946 89.6 32,303 1935 931 3,203 9.8 29,388 90.2 32,590 1936 1,117 3,578 9.7 33,459 90.3 37,037 Source: Amaya (1982). established plants for assembling knockdown parts, they placed some orders for parts with Japanese companies. Both firms were so eager "to develop local sources of supply that they already had begun to turn over, at nominal fees, detailed specifications of de- sired parts to Japanese producers" (Mason 1992:81). The annual production of three-wheel automobiles reached nearly 10,000 vehicles in the late 1930s. Parts suppliers mushroomed. The production of automobile parts rapidly grew from Y14 million in 1931 to Y53 million in 1935 (Amaya 1982: 67). According to a recent study, these new parts suppliers "provided an invaluable infrastructure for Nissan and Toyota" (Mason 1992: 81). The war with China broke out in 1937, bringing significant changes to the Japanese economy. The motor vehicle industry, like other industries, fell under tight government control. In 1938 the Ministry of Commerce and Industry forbade the production of passen- ger cars other than for military use, and vehicle production concentrated on trucks, which were needed for the transport of soldiers and military goods in China (Mason 1992: 81). Automobiles were included on a list of fifteen major items targeted by the governiment in 72 Towvard thte Rutral-Based Development of ComM1erce and Indiistry 1938 for increased production under a four-year plan. As the war front widened, however, and especially after the confrontation with Great Britain and the United States began in 1941, the strategic status of motor vehicles declined. Aircraft, rather than trucks, became a major priority, especially in the fight against the United States. Toyota's Strutggle to Becomiie Established EXPERIMENTAL DAYS. The growing demand for automobiles in the late 1920s and early 1930s also stimulated some entrepreneurs to consider entering the business of automobile manufacturing. One such entrepreneur was Kiichiro Toyoda, the founder of Toyoda Auto- mobile Company. Kiichiro was the son of Sakichi Toyoda, a famous automatic loom inven- tor, who established the Toyoda Automatic Loom Works, Ltd., in 1926. Sakichi was the son of a carpenter and received no systematic engineering training. He carefully investigated imported looms; at first he imitated them, but then he moved on to invent the high-stan- dard power looms. Even Platt Brothers in Great Britain, one of the largest loom-makers of the time, wanted to purchase the patent rights to the automatic loom Sakichi invented, and it started to negotiate with the Toyoda Automatic Loom Works, Ltd. In 1929, when he visited England to conclude patent negotiations with Platt Brothers, Kiichiro Toyoda was impressed by the large number of automobiles he saw on the streets. Platt Brothers paid £100,000 to Toyoda Automatic Loom Works, Ltd., for the patent rights to manufacture and sell the automatic loom invented by Sakichi Toyoda.2 This fee gave Kiichiro the financial base he needed to begin manufacturing automobiles. Even before obtaining the fee, Kiichiro had purchased many precision machine tools that were not required to manufacture tex- tile machinery. After obtaining the fee, he was able to amass the tools and learn the techno- logical skills needed to manufacture automobiles at a greater speed than his contemporaries. The Toyota Automobile Company started out as the Automobile Department of the Toyoda Automatic Loom Works in 1932, and it became an independent company in 1937. Even before the board of directors of Toyoda Automatic Loom Works formally endorsed the plan for manufacturing automobiles in 1933,3 Kiichiro Toyoda had started to carry out research. In 1930 he had begun experimenting with the manufacture of gasoline motors. He and his staff learned by taking motorcycles apart. In the summer of 1933 they at last succeeded in manufacturing a four-horsepower gasoline engine (Toyota Motor Company History Editorial Committee 1958: 25). Kiichiro organized a group of engineers and stocked up on machine tools at the com- pany, but he needed engineers with automotive experience and invited, for example, Takatoshi Kan to join the company. Kan had participated in a pioneering automobile manu- facturing project, the "Chukyo Detroit Project," which was an attempt to establish the Chukyo area around Nagoya as a center of automobile production, just as Detroit was in the United States. The project was abandoned because it was not economically feasible. Kan was then one of only a handful of engineers with experience in manufacturing auto- mobiles. The Formationt of Toyota's Relationiship with Suppliers 73 Kiichiro still faced an uphill struggle. He had to produce a prototype car as soon as possible. At the same time, a plant for manufacturing automobiles had to be set up. Kiichiro carried out two surveys of parts suppliers for automobiles in 1931 and 1933. Although there were established parts suppliers, and their production schedules had grown, he de- cided that the quality of parts made in Japan was poor, primarily because of inferior steel. This prompted him to set up a steel manufacturing department within Toyoda Automatic Loom Works in 1934.4 In May 1935 the company managed to produce a prototype passenger car. It had taken at least two years to produce the prototype from the time the company had formally em- barked on the business of manufacturing automobiles. Nevertheless, the prototype was not "purely domestic-made"-it was constructed using many Chevrolet parts. Kiichiro and Kan intentionally took this approach because the company could not, either finan- cially or technically, produce every part. Because their survey of parts-makers had re- vealed the poor quality of parts made in Japan, in order to shorten the time required to make the prototype car and to make it possible to use reliable, genuine Chevrolet parts, they intentionally designed their prototype to have the same size parts as those used in Chevrolets and Fords. This design strategy smoothed the way for them to start their auto- mobile business. The flaw in their plan, however, was that some of the parts in their cars had been made in Japan, and these caused breakdowns in the early days. As a result, Toyoda had to begin producing parts in its own plant.5 This experience later prompted Toyoda to be demanding when it began training its suppliers. THE "PASTORAL FACTORY": KoRoMo PLANT. When Kan moved to the Toyoda Company in 1933, Kan and Kiichiro started to draw up plans for a pilot plant, which later was known as the Kariya plant. The obstacles seemed to be insuperable: they had little detailed knowl- edge of the automobile production process, or the processes or materials involved in mak- ing parts. Kan went to the United States to purchase machines for the Kariya plant in January 1934. In accordance with instructions from Kiichiro that he visit a number of factories, he visited the Ford, Chrysler, and Graham Paige factories during his five-month stay. Kan observed their operations carefully, analyzing the manufacturing process and the flow of production. In March of 1936, Kiichiro Toyoda directed Kan to design a second plant. The result was the Koromo plant, which began operations in November 1938 with the capacity to produce 2,000 cars each month. This factory encapsulated Kan's observations of the American manu- facturers' production methods and Kiichiro's ideas about auto manufacturing. Koromo, where Toyota decided to build this new assembly plant, was far from being an industrial area with a large number of parts suppliers or engineering and machining firms (see figures 4-1 and 4-2). At the end of 1936, Aichi Prefecture, where Koromo was situated, had 9,677 factories that employed 5 or more workers (not including 11 public works). A breakdown by industry shows that Aichi Prefecture specialized in the spinning and weav- ing, machine-tool, and ceramic industries (see table 4-2). Nagoya, the prefectural capital, had 4,265 of these factories, or almost half the total, in 1936. In Koromo there were just 74 Tozoard the Rutral-Based Development of Commerce and Indtustry Figure 4-1. Population Density of Aichi Prefecture, 1915 Ichinomi_a People per 15.42 kM2 Under 1,000 t 2/ 1,000 to 2,000 2,000 to 2,500 4,000 toS,000 5,00O to 8,000 ~ 8,000 to 10,0OO 10,000 to 15,000 25,000 to 50,000 50,000 to 70,000 over 70,000 IBRD 29578 Figure 4-2. Distribution of Factories in Aichi Prefecture, 1936 (factories employing fewer than 5 people are not counted) 427 62C/5 135 y \ 264 <\\0 t 21 ( 25 4 46 ) _N~~agoya i ry -~ ~'C" OKoromo t :) 373 270") 146 k /J' f 474 XJ /\ ,=-1'V 5 Note: In this map, figures are not given for Seto City (345 factories) and Handa City (229 factories). These two cities are rather close to Nagoya. IBRD 29579 The Formationz of Toyota's Relationship wvith Suppliers 75 Table 4-2. Aichi Prefecture's Industrial Structure at the End of 1936 Factories Employees Percentage Percentage Indiustry Nuimber of total Nuimber of total Spinning and weaving 4,211 43.6 131,298 49.8 Metal 454 4.7 8,330 3.2 Machine tools 1,001 10.4 50,872 19.3 Ceramic 1,007 10.4 25,366 9.6 Chemnical 194 2.0 13,271 5.0 Woodworking 766 7.9 10,542 4.0 Printing 243 2.5 3,772 1.4 Food 784 8.1 9,232 3.5 Gas and electricity supply 15 0.2 473 0.2 Other 992 10.3 10,717 4.1 Total 9,667 (100) 263,873 (100) Source: Aichi Prefecture, Department of Statistical Survey (1937). seven factories (Aichi Prefecture 1937:286-88), which were engaged in traditional occupa- tions such as brewing sake. After the 1880s, sericulture began to develop in this area, and Koromo became a cocoon-trading center. Before the Koromo plant came into operation in 1938, the town's main products were related to agriculture and sericulture (see table 4-3). Why did Kiichiro choose Koromo as the location of an assembly plant? The town clearly offered attractive conditions to Toyota (Toyoda Automatic Loom Works 1967: 215-16). As the development of artificial silk began to depress the sericulture industry and bring down the price of cocoons, the town of Koromo had an incentive to attract new plants to the area. It was easier and less expensive to purchase a large piece of land in such a rural area than in a more densely populated area. In addition, the infrastructure had been greatly im- proved. In 1930 the Ministry of Railways opened Japan's first bus route in the area, and the route ran through Koromo. A local railway also began operating from Okazaki to Koromo Table 4-3. Changes in Products at Koromo Category 1933 1934 1935 1936 1937 1938 1939 Product (percent) Agricultural 31.6 29.8 29.6 30.6 31.8 6.4 3.1 Sericultural 51.2 47.9 54.9 53.4 46.1 9.3 1.8 Manufacturing 7.6 9.8 6.3 6.2 10.7 81.5 93.8 Total products (1,000 yen) 2,779 2,359 3,248 3,643 3,437 18,894 56,513 Value per inhabitant (yen) 189 158 222 245 217 1,019 2,685 Source: Watanabe and Koide (1953: 131). 76 Toward the Rtral-Based Development of Conimerce and Indutstry in 1934. Such improved transportation facilities could easily have persuaded Kiichiro to build his assembly plant in Koromo. The idea of establishing an assembly plant in such a rural area was surely based on economic calculation-on the possibility of hiring cheap labor as well as being able to purchase a large tract of land at a low price. Nevertheless, setting up an assembly plant in such an isolated area meant that Toyota could not employ the specially trained manual workers easily found in urban areas. It is possible that Kiichiro preferred to employ un- trained workers at his plant, rather than workers who already had been trained. Kiichiro called the Koromo plant a "pastoral factory," one that encapsulated his ideal of "training unspoiled persons who had been brought up in rural districts, and producing inexpensive and good-quality cars on a large scale" (Toyoda 1937). He may have recognized the ad- vantages of a rural base after to his experience in Kariya, the site of the Toyoda Automatic Loom Works plant. It is possible that Kiichiro made room in his plans for taking Koromo- trained workers. Kiichiro also had a ready supply of workers for the automobile manufac- turing operations in his Kariya plant; because Koromo was not far from Kariya, workers were transferred to the Koromo plant. The establishment of the plant in this rural area also provided nonfarm employment to the local labor force, and later stimulated an influx of labor from other areas. Koromo's population increased rapidly after the opening of the plant, from 15,821 in 1937 to 18,541 in 1938 (a 17 percent increase in one year; see figure 4-3). The industrial structure at Koromo also shifted from agriculture to (automobile) manufacturing (see table 4-3). Figure 4-3. Population Growth at Koromo, 1925-39 Population 25,000 20,000 - 15,000 - 10,000 _ 5,000 - O 0 I I I I I 1925 1927 1929 1931 1933 1935 1937 1939 Year The Forniationi of Toyota's Relationship with Sutppliers 77 Supplier Characteristics in the EarlyYears Setting up the assembly plant far out in the countryside meant there were no parts suppli- ers around the plant; those nearest supplied the Toyoda Automatic Loom Works. As a result, Toyota relied mainly on parts purchased from firms in the Tokyo and Osaka areas. In addition to those suppliers, Toyota also placed orders with existing suppliers of its Loom Works. Further, some entrepreneurs tried to seize the opportunities created by the new Toyota project. In 1939 Toyota's suppliers formed the Kyoriyoku-kai (Cooperative Association),6 which was an annual, convivial gathering of Toyota and its suppliers. Toyota's official history lists the names of eighteen suppliers at that time.7 There is some doubt whether all were original members of the Kyoryoku-kai, but being listed in the company history means they were all regarded as important Toyota suppliers. Most continue as parts suppliers in the Aichi area. Nagoya City's 1936 directory of local manufacturers lists five of these firms. (Directories published in later years may list more firms, but they do not contain rich information on individual firms because of the secrecy imposed on such information in the wartime economy.) Apparently these five firms did not produce only automobile parts. Hamakichi Kojima was a good example of the entrepreneurs who jumped at the op- portunity to enter the automobile-related business. According to an anecdote, he de- cided to manufacture automobile parts soon after the newspapers reported that the government had passed the automobile industry law of 1936. At that time he success- fully ran a small factory manufacturing traditional small household wares such as por- table body warmers. The manufacture of a portable body warmer required the stamping of tin plates, but his machines could not stamp harder metals such as steel or iron. Also, the manufacture of body warmers did not require making parts with high precision. The stamping capabilities of his firm obviously fell far short of the standards needed to manu- facture automobile parts. In spite of his factory's poor technical capability, Kojima was determined to enter the automobile-related business because he believed automobiles would have a big market in the near future, given the way their use was spreading in Tokyo. He managed to meet Kiichiro, who gave him an order for buckets (Kojima Stamping Company 1988). The buck- ets were used to store sand on the running boards of vehicles, because sand had to be spread in order to increase friction when vehicles were driven up steep, slippery slopes. In this way, automobiles with poor horsepower managed the steeply inclined roads so preva- lent in Japan at the time. The bucket itself had nothing to do with the mechanical functions of automobiles, but the item was very important in giving Kojima a toehold in the busi- ness. The next item Toyota ordered from Kojima was washers. According to the company's history, Kojima could not understand the design drawings of automobile parts, and he even found it difficult to understand some English technical terms. The local directory for 1936 described Kojima's firm as a manufacturer of portable body warmers and automo- bile parts that employed fifteen workers and was able to boast of a one-horsepower elec- 78 Tozvard the Rutral-Based Dezvelopment of Conunierce and Indutstry tric motor. This shows that by the end of 1936, Kojima had managed to enter the business of making automobile parts. Gradually Kojima's firm expanded the range of its products, mainly into small parts such as washers. It would seem that Toyota was training Kojima and enabling him to upgrade his company's technological capability. When the Koromo Plant opened in 1938, Kojima also built a new and much larger factory and shifted its product lines substantially toward automobile parts. Kojima's determination was bearing fruit, thanks to some assistance from Toyota. The Kojima Stamping Company typifies Toyota's attitude toward newcomers that wanted to enter the automobile parts business and were willing to follow Toyota's technical assis- tance. There is no doubt that Toyota could have placed orders with specialized, city-based engineering firms. This may have been much easier, and the parts less expensive, because there would have been no need to render technical assistance to an established firm. Nev- ertheless, Kiichiro continued to place orders with Kojima and to nurse his company along as an automobile parts maker. This suggested that Kiichiro preferred to develop tight rela- tionships with unestablished and, hence, trainable manufacturers with a strong willing- ness to follow his advice and guidance. New suppliers with the determination and openness of Hamakichi Kojima did not ap- pear in large numbers. His was an exceptional case. The other parts suppliers whose names appeared in the directory of 1936 were suppliers to Toyoda Automatic Loom Works or makers of parts for looms, airplanes, and bicycles. Meido Iron Company, a screw manufacturer, traded with the Toyoda Automatic Loom Works through an agent at first, but in 1932 began trading directly with Toyoda Automatic Loom Works. Meido Iron's official company history says: "In response to a rapidly in- creasing demand for screws for automobiles, the company started direct dealings with Toyoda Automatic Loom Works" (Meido Iron Company 1974:34). In 1932 Meido Iron also decided to establish a new plant, completed in 1935, the same year Toyoda Automatic Loom Works established the Kariya plant as a pilot plant for manufacturing automobiles (Meido Iron Company 1974: 34). Other companies mentioned in Aichi Prefecture's directory of 1936 were the Hayashi Spring Manufacturing Company, which produced springs for looms as well as for auto- mobiles and employed eighty-four; Niwa Iron Company, which employed forty-one people and manufactured parts for airplanes, looms, and automobiles; Kato Iron Company, which manufactured parts for bicycles and automobiles and employed twenty; and Tsuda Iron Company, which, according to the directory, only produced parts for looms but was situ- ated in Kariya, where the Toyoda Automatic Loom Works was located. The above six firms did not have any experience in manufacturing parts for the Ford and General Motors subsidiaries in Japan. Some members of the Kyoryoku-kai were forced to enhance their technological capabilities, as in the case of the Kojima Stamping Com- pany. It is also worth noting that none of these parts-suppliers was a large firm. Once the production of automobiles at Toyota started, they had to expand their manufacturing ca- pacities or build new plants, as the Meido Iron and Kojima Stamping companies did. Toyota often had to lend technical support to these firms because of its decision not to rely on The Formation of Toyota's Relationship with Suippliers 79 larger engineering companies with more sophisticated technical capabilities. But provid- ing technical assistance to its suppliers meant that Toyota was able to accustom them to Toyota's way of doing things and its demanding specializations without having to exert strong pressure or sanctions. This made it easier to improve the flow of information couched in a common terminology, and to build strong personal ties between Toyota and its suppli- ers. It should not be overlooked that the relations between Toyota and its suppliers took shape on the basis of very close personal dealings, especially in the nascent years. Tightening the Relationship In 1940 Toyota made a significant decision regarding purchasing policy. The company decided to reduce the number of parts in order to reduce costs by "giving consideration to the geographical distribution of suppliers" (Toyota Motor Company History Editorial Committee 1967: 180). It greatly reduced the number of items ordered from the Tokyo and Osaka areas, reducing orders from the Nagoya area only slightly (see table 4-4). This would suggest Toyota's disappointment with the technological levels of its existing suppliers. Toyota's official history cites the internal report on purchasing policy, which explains why Toyota had to launch its new purchasing policy. It was, the company explained, entirely based on its own detailed estimate of costs, and was adopted so that Toyota could make a profit from its automobile manufacturing business. Even though the Kyoryoku-kai had been formed only a few months before, and this policy change would inevitably cause some tension among suppliers, the company still went ahead with it. The Kyoryoku-kai was organized for the purpose of fostering close personal ties, but Toyota also needed inexpensive and good-quality parts if it was to place the business of manufacturing auto- mobiles on a solid basis. Table 4-4. Proposed Changes in Purchasing of Parts Jan iary 1940 Proposed target Proposed cianges Value Value Reduced per unrtit per unit Reduiced value Nuimber of Valute of part NuImber of Valute of part Redutced valute per unit purchased per car (yen) puirchased per car (yen) items per car of part items (yen) (B/A) items (yen) (E/D) (percenit) (percent) (percent) Area [A) IBJ [C] [DI [El [Fl (D/A) (E/B) (F/C) Nagoya 366 764.12 2.09 300 500 1.67 82.0 65.4 79.9 Osaka 104 273 2.63 50 100 2 48.1 36.6 76.0 Tokyo 100 1,888.4 18.89 30 400 13.34 30.0 21.2 70.6 Total 570 2,925.52 5.14 380 1,000 2.64 66.7 34.2 51.4 Source: Toyota Motor Company History Editorial Committee (1967: 180). 80 Toward thte Riural-Based Development of Conuiinerce and Industry Table 4-4 reveals an interesting fact: Toyota purchased more than half the items for as- sembling automobiles in the Nagoya area, but the total value of the parts purchased from that region was far below the value of parts from the Tokyo area. Even the proposed changes to its purchasing policy meant that Tokyo-area suppliers were receiving more money per part than those from the Nagoya and Osaka areas. As of January 1940, the average value per part for each car was about Y19 for Tokyo-area suppliers; Y2.6 for Osaka-area suppli- ers; and Y2.1 for Nagoya-area suppliers.8 The technical capabilities of the Nagoya-area suppliers were not as great as those of their counterparts in the Tokyo area, and hence they were not involved in producing expensive parts. The purchasing policy proposed in 1940 meant that the Tokyo-area suppliers' share of Toyota's business would be greatly reduced in both items and value. Nagoya-area suppli- ers would gain an increased share, although the value per item would still remain around Y2. What the proposed purchasing policy showed, therefore, was Toyota's intent to rely on in-house manufacturing of expensive items, for the more inexpensive items Toyota re- mained relatively reliant on out-sourcing, especially with Nagoya-area suppliers. Government policy also had a significant effect on Toyota's dealings with suppliers. After the Ministry of Commerce and Industry forbade the production of passenger cars for nonmilitary use in 1938, the government more or less coopted the suppliers producing good-quality parts, and in 1939 it persuaded these suppliers to organize themselves into an association. In 1940 the government circulated a "Notification Regarding the Organiza- tion of the Automobile Parts Industry," which resulted in the establishment of the Auto- mobile Parts Association of Japan in 1942, with a membership of 305 firms. The government tried to force each supplier it had certified as qualified to produce parts at an acceptable level of quality to specialize in manufacturing one kind of part. This chain of events forced Toyota to deal with a smaller number of suppliers. The shortage of steel, however, would have an even greater impact on supplier relation- ships at Toyota. The government decided to allocate steel in accordance with its priorities in the war effort. Motor vehicles had originally been included among the fifteen priority items targeted for increased production under a four-year plan announced in 1938, but as the battlefront expanded, especially after 1941, aircraft took over as a more strategic item in the war effort, especially for use against U.S. forces in the Pacific. The government or- dered aircraft production to be increased as much as possible, giving planes priority over trucks. As a result, suppliers at Toyota found it increasingly difficult to get the steel and other materials needed to manufacture truck parts. This resulted in a reorganization of the Kyoryoku-kai. In 1943 the Kyoryoku-kai changed its name to the Kyoho-kai, and through this organization Toyota distributed materials to individual affiliates. Hamakichi Kojima was elected the first chairman. Suppliers, in cooperation with Toyota, allocated the avail- able steel. As a result, the Kyoho-kai itself came to play an important role in automobile production, although obtaining steel became more and more difficult. Toyota and its sup- pliers were motivated to cooperate closely under the difficult conditions caused by the steel shortage. This experience built close personal ties among the suppliers and between the suppliers and Toyota, which in turn made it easier for them work closely within the The Forniationi of Toyota's Relationiship wvith Suippliers 81 Kyoho-kai. In this sense, the wartime experience was an important factor in the creation of today's relationships between Toyota and its suppliers. Reorganization of the Kyoho-kai The end of the war and the postwar upheaval disrupted the activities of the Kyoho-kai. In a meeting with suppliers in 1946, however, Kiichiro Toyoda expressed his keen desire to establish an automobile industry in Japan. He stressed that this would require the estab- lishment of a parts industry, and that Toyota and its suppliers had a common destiny. Soon after the meeting, Toyota reorganized the Kyoho-kai into three regional districts: Tokyo, Osaka, and the Tokai district, including Nagoya (Toyota Motor Company History Edito- rial Committee 1967: 253-54). Suppliers in the Tokyo and Osaka districts were larger en- terprises that were rather independent of Toyota, while the Tokai district suppliers were comparatively small-scale and sold their products almost exclusively to Toyota (Toyota Motor Company History Editorial Committee 1978: 150-51). For Toyota, therefore, estab- lishing a parts industry meant ensuring the stability and growth of suppliers in the Tokai district, so that in order to improve the quality of parts and establish a reliable parts indus- try, Toyota's first managerial efforts had to be directed at helping the Tokai suppliers. This reorganization of the Kyoho-kai symbolized a new era in the relationships between Toyota and its-suppliers. The strong personal ties, both among suppliers and between Toyota and the suppliers, that were nurtured during the war remained firm, and they facilitated the task of building a coherent organization. For some years after the war, Toyota's annual production remained low: 6,703 vehicles in 1948 and 10,824 vehicles in 1949. To make matters worse, massive labor problems erupted at Toyota in 1950, and the company stood on the brink of bankruptcy. The crisis led to the resignation of Kiichiro Toyoda and to the separation of Toyota Motor Sales Co., Ltd., from Toyota Motor Co., Ltd."' After Kiichiro Toyoda's resignation, Taizo Ishida was appointed president in July 1950. At the end of the same month, Toyota received its first special procurement order-for 1,000 vehicles-from the U.S. military, with several rep- etitions of the order to follow in the next few months (Toyota Motor Corporation 1988: 104-11)." The orders contributed a great deal toward putting the Toyota enterprise back on a sound footing, and the revenue derived from the procurement orders helped to re- build and modernize production facilities. In 1951 Toyota formulated its first long-term, five-year scheme for modernizing produc- tion facilities. At this stage, the company still could not afford to provide direct managerial guidance or technical instruction to suppliers. Its purchasing department, with a staff of only forty, was fully engaged in ordering and receiving parts, and it could not respond to requests from suppliers for technical support (Aichi Prefecture, Industry and Commerce Department 1953: 3). The first attempt to modernize management practices among suppliers came not from Toyota but from the Small and Medium Enterprises Agency. After its establishment in 82 Tozward thte Rtral-Based Development of Commiierce aind I;tditsfy 1948, the agency tried to modernize small and medium-size enterprises and undertook a variety of examinations of their business practices for their benefit (Ministry of Interna- tional Trade and Industry 1963:646). The agency extended these measures to include groups affiliated with larger firms and their suppliers. Its program was called 'keiretsui diagno- sis." Toyota and its twenty-one suppliers in the Tokai district were beneficiaries of the agency's program in 1952 and 1953. A few officials of Aichi Prefecture carried out this keiretsut diagnosis; they visited each plant and recommended improvements in management capabilities. They looked through the plant, collected data on management practices, and interviewed employees as well as employers. They devoted a few days to observing and advising each supplier. Most of the suppliers they visited were small-scale, employing under 100 people. The suppliers' man- agement practices were haphazard rather than systematic: some did not keep accounting records properly and did not know (at least not exactly) whether their companies were making a profit or posting a loss until the accountants visited them once a year to produce a balance sheet. Keiretsu diagnosis was a good opportunity for suppliers to learn basic management techniques and acquire other useful knowledge. One official recalled that suppliers were very skeptical, and sometimes fearful; about the diagnosis at first, but they became helpful once they realized how useful it was in improving productivity (Kyoho- kai History Editorial Committee 1967: 24). The suppliers eventually were very apprecia- tive of the diagnosis, and to express their gratitude they donated a bus to Aichi Prefecture (Kyoho-kai History Editorial Committee 1967: 25). At the end of each investigation, the officials produced a report based on their observa- tions, which they then gave to the supplier. The report would describe the weaknesses and strengths of management and suggest ways to improve the productivity of the plant. The report would also grade each supplier on management, production, labor, marketing and purchasing, finance, accounting, and research. This grading system was rather simple, but apparently Toyota had not undertaken such a grading procedure before the keiretsut diag- nosis was carried out. After the suppliers were visited, the officials produced a final report on the relationship between Toyota and its suppliers. This final report was read at a meeting at Toyota head- quarters in the spring of 1953. Senior government officials as well as suppliers were present. The officials disclosed their findings and recommended a number of ways the relations between Toyota and its suppliers could be improved. They also produced a table in which the suppliers were given point grades."2 Although the suppliers' names were not men- tioned, each supplier received a separate report on its plant and knew its own grading points. Individual suppliers were thus easily able to see how they ranked. Suppliers with lower grades received a strong stimulus to improve their management skills and raise their performance-the fear that they might start getting fewer orders from Toyota. The keiretstu diagnosis also taught Toyota a lesson in how to advise, monitor, and grade its suppliers (Kanto Kyoho-kai History Editorial Committee 1976: 32). Up to that time, Toyota apparently had no monitoring system at all. After the diagnosis, Toyota's staff be- gan to visit suppliers often and to check their manufacturing processes closely. In October Thze Formation of Toyotas Relationship with Suppliers 83 1953 Toyota decided to introduce quality control into every manufacturing process, estab- lishing a committee in charge of quality control. Toyota also asked suppliers at the annual meeting of the Kyoho-kai to introduce quality controls in their plants. In response to this request, the Kyoho-kai organized seminars on quality control. Toyota sent its own person- nel to the seminars to act as instructors and, following the seminars, Toyota's staff would visit supplier plants to observe how the suppliers implemented quality control and to give advice if needed (Wada 1991). In view of Toyota's apparent lack of advice and monitoring for suppliers before the keiretsut diagnosis experience, this course of action showed a re- markable change in attitude. After the keiretsut diagnosis, the character of the Kyoho-kai changed greatly. The Kyoho- kai had been a convivial gathering, a time for making or renewing contracts and explain- ing news and views on an informal basis. After the diagnosis, however, the suppliers started visiting one another's factories and studying how to rationalize procedures. The suppliers also established several study groups within the Kyoho-kai. As someone remarked, "The activity of the Kyoho-kai really began only at this time." Suppliers became serious about striving to rationalize and improve their management (Kyoho News, January 5, 1967). It was the grading of the suppliers in the final report that drove them, especially those rated lower in the table, to work harder. They realized that Toyota would be carefully monitor- ing their efforts to rationalize plant procedures. Thus, a sense of competitiveness emerged among the suppliers. In 1955 Toyota marketed the Crown, "Japan's first completely domestically produced passenger car" (Kato 1981: 62). In 1956 it also introduced a "supermarket system" in its plants, the forerunner of the "just-in-time method," and suppliers brought some of their parts directly to the assembly line (Ohno 1978:49-51; Shingo 1981: 136-37,256-57). Toyota could not have introduced such practices if its suppliers had not been striving so hard to rationalize after the keiretsu diagnosis experience. Conclusions The interfirm relationships between Toyota and its suppliers contrast greatly with those of American automobile manufacturers and their supply sources. Managerial efforts by both Toyota and its suppliers, and also the interactions between them, are what have shaped interfirm relations at Toyota over a long period. When the keiretsut diagnosis was carried out, Toyota seized the opportunity to create an atmosphere of competition among its suppliers and to monitor them. Personal ties created before and during the war helped smooth the way for the factory tours conducted by the suppliers themselves after the diagnosis and encouraged them to compete with one an- other in a friendly rivalry. Without a mutual trust based on personal ties among suppliers, as well as between the suppliers and Toyota, the diagnosis could not have produced such good results. If Toyota had tried to stimulate competition among the suppliers by hinting at their competitors' ratings, and if the suppliers could not count on Toyota to provide 84 Toward the Ruiral-Based Developmient of Commiierce and Indiistry good business opportunities for them, the suppliers would have stopped dealing with Toyota. And if the suppliers had felt that Toyota was merely abusing its authority, they would have looked for more attractive business opportunities with other companies rather than engaging in cutthroat competition among themselves to obtain more orders from Toyota.' 4 Trust was created between Toyota and its suppliers as a result of personal ties, but the long-standing business relationship that survived wartime government controls, a steel shortage, and the grading points of the company's venture into a new business provided both sides with the opportunity to test the strength of that trust. In a sense, Toyota's suc- cess was based on the creation of a community of suppliers according to the image of the traditional village communities pervasive in Japan.' While entry into this community was not easy, competition among the small number of suppliers was extremely intense because it was driven not only by short-run profit, but also by the praise or disgrace each member would face depending on its performance. Thus, this community mechanism of "cooperation and competition among the few" has proved to be instrumental in the devel- opment of the highly efficient production system that enabled Toyota to achieve supremacy in the world automobile industry. It is considered a pfime example of how an "apparently modern" agrarian organization can be transformed into the basis of a modern, high- technology production system. Notes 1. On the earliest attempts to assemble automobiles in Japan, see Japan Automobile Association (1965). 2. The patent rights covered all countries except Japan, China, and the United States. On this deal, see Taniguchi (1992). 3. In January 1934 a special meeting of shareholders formally approved the company's entry into the business of manufacturing automobiles. 4. This department became an independent firm in 1940, the Aichi Steel Company. 5. On the development of the relationship between Toyota and its suppliers, see Wada (1991). 6. The historical outline of Kyoho-kai is based on Kyoho-kai History Editorial Committee (1967). 7. Professor Yochio Ohba recently suggested that there were thirty-one original member compa- nies in the Kyoryoku-kai. According to his list, just seven of the thirty-one suppliers were located in the Tokyo and Osaka areas; the rest, the large majority, were found in Aichi Prefecture. The list produced by Ohba, however, does not include all the suppliers listed in the official history. See Ohba (1990). 8. It must be acknowledged that the greater distance between the Koromo Plant and Tokyo explains some of the differences in average purchase prices between the Tokyo and Nagoya areas, yet there was not a wide difference in the average purchase prices between Nagoya suppliers and Osaka. This suggests that the distance factor alone cannot explain the differences in the average prices among the three areas' suppliers. 9. On the annual production of Toyota, see the Appendixes of Toyota Motor Corporation (1988). The Formation of Toyota's Relationship woith Suippliers 85 10. Both Toyota Motor Co., Ltd., and Toyota Motor Sales Co., Ltd., remained independent until they merged again to become Toyota Motor Corporation in 1982. 11. The special procurement orders ended in March 1951. 12. The individual suppliers' grading points are reproduced in Wada (1991: 32). 13. The expanding volume of sales of Toyota cars was helpful in this regard. After 1955, and especially during the late 1960s and 1970s, Toyota and other Japanese car assemblers enjoyed a rising demand for passenger cars. Under this expanding market, most suppliers could increase their sales in absolute terms, although they might reduce their own share of the total parts-purchas- ing of Toyota. In order to increase their own share in the long run, suppliers competed with each other. Because Toyota did not greatly change its purchasing volume for a specific supplier annually, suppliers could collaborate too, especially in such a growing market. On the observation of suppli- ers' relationships, see Womack, Jones, and Roos 1990: 154-55. 14. John Kay identifies the importance of the "need to establish trust and penalize opportunism in a network of relational contracts." He defines the "relational contract" as follows: "An exchange relationship between two parties which is not fully articulated. The rules of behavior are implicit and the enforcement mechanism is the value of the continuing relationship between two parties." This is the very case to Toyota (see Kay 1995: 84). References Aichi Prefecture, Department of Statistical Survey. 1937. Sh6tva 11 nenz 12 gatsur niatsuijitsui genzai Aichi-ken koj6 soran (Directory of Factories in Aichi Prefectutre as of tlhe End of 1936). Nagoya: Aichi Prefecture. Aichi Prefecture, Industry and Commerce Department. 1953. "Keiretsu shindan-sho sokatsu: Toyota Jidosha Kogyo Kabushiki-gaisha" ("A Summary of kei,etsut Consultation: Toyota Motor Com- pany"). Nagoya. Photocopy. Amaya, Shogo. 1982. Nihon jid6sha k6gyO no shiiteki teniikai (Tlie Historical Developmenit of the Japanese Auttomiobile Indistrh). Tokyo: Nihon Hyoron-sha. Hayami, Yujiro, and Toshihiko Kawagoe. 1993. The Agrarian Origins of Commnerce antd Induistry. Lon- don: Macmillan. Japan Automobile Association, ed. 1965. Nihon jid6sha k6gy6shi-k6 (Draft History of thte Japanese Auito- mobile Indutstry). Vol. 1. Tokyo. Kanto Kyoho-kai History Editorial Committee. 1976. 30 nent n1o ayutmi (Thte History of thte First 30 Years). Tokyo. Kato, Seisi. 1981. My Years wuith Toyota. Tokyo: Toyota Motor Sales Company. Kay, John. 1995. Fouindationis of Corporate Suiccess. Oxford, U.K.: Oxford University Press. Kojima Stamping Company, Company History Editing Committee. 1988. Okagesanta-de 50 nen ninina genki-de (Th7antks to Yout We Have Prosperedfor 50 Years). Toyota City. Kyoho-kai History Editorial Committee. 1967. Kyoho-kai no ayudmi (The History of Kyoho-kai). Toyota City. Mason, Mark. 1992. Americani Mutltinational and Japan: The Political Economny of Japanese Capital Con- trols, 1899-1980. Cambridge, Mass.: Harvard University Press. Meido Iron Company. 1974. Meido 50 n1et7 n1o ayin7i (Meido's 50-Year History). Nagoya. 86 Toward the Rutral-Based Developmitent of Corninierce and Iniduistry Ministry of International Trade and Industry. 1963. Shoko seisakni-shi (The History of Policies for Conm- merce and Industry). Vol. 12. Tokyo: Shoko Seisaku-shi Kanko-kai. Ohba, Yochio. 1990. "'Jidosha seizo jigyo-ho' to Toyoda, Nissan, Isuzu no Nihon-teki seisan shisutemu" ("The Automobile Manufacturing Act and the Japanese Production System at Toyota, Nissan, and Isuzu"). Keizai ronshtu (Hokkai Gakuen Daigaku) 38(2). Ohno, Taiichi. 1978. Toyota seisan hoshiki (Toyota Produtction System). Tokyo: Daiamondo. Shingo, Shigeo. 1981. Study of "Toyota" Produiction Systemnfrom anz Indutstrial Engineerinig Vievpoint. Tokyo: Japan Management Association. Taniguchi, Yutaka. 1992. "1930 nen zengo no b6shoku kikai k6gy6 ni okeru nichiei kankei no ichidanmen" ("One Aspect of the Relationship Between Japan and Britain in the Textile-Machine Industry Around the 1930s"). In Kaichiro Oishi, eds., Senkaniki Nihon no taigai keizai kankei Japan's Econromic Relationship wvith Foreign Countries Betzveen the Wars). Tokyo: University of Tokyo Press. Toyoda Automatic Loom Works, Ltd. 1967. 40 nen-shzi (40th Anniversany History). Kariya. Toyoda, Kiichiro. 1937. Toyota Jidosha Kogyo Kabuishikigaisha no soritsut to soshiki (The Fountdation of the Toyota Automobile Conmpany and Its Organization). Toyota City. Toyota Motor Company History Editorial Committee, ed. 1958. Toyota Jidosla 20-nenshi (The First 20 Years of Toyota Motor Conmpany). Toyota City. 1967. Toyota fidosha 30-nenshi (The First 30 Years of Toyota Motor Company). Toyota City. 1978. Toyota t7o aylumi (Th1e History of Toyota). Toyota City. Toyota Motor Corporation. 1988. Toyota: A History of the First 50 Years. Toyota City. Wada, Kazuo. 1991. "The Development of Tiered Inter-firm Relationships in the Automobile Indus- try: A Case Study of the Toyota Motor Corporation." Japaniese Yearbook on Butsiness History 8. Watanabe, Shinichi, and Yasuharu Kiode, eds. 1953. Koromoshi keizzai hattatsui no sho-joken1 (Condi- tionsfor Koromo City's Economic Development). Koromo Town Office. Womak, James P., Daniel T. Jones, and Daniel Roos. 1990. The Macline That Changed the World. New York: Rawson. Part II. Current Developments in East Asia 5 Export-Oriented Garment Industries in the Rural Philippines Masao Kikuchi A lthough political and social unrest rendered the 1980s a lost decade for economic development, the Philippines is now vigorously pursuing the development path for industrialization taken by its ASEAN (Association of Southeast Asian Na- tions) counterparts. In government development strategies, industrialization is being pur- sued by the promotion of export-oriented, modem manufacturing industries, mainly through efforts to attract foreign investment and by the promotion of small- and medium- scale industries, especially in rural areas (National Economic and Development Authority 1992). The latter is of particular importance because it is imperative that the country eradi- cate rural poverty by providing productive employment opportunities for the overflow- ing rural labor force, while avoiding overcongestion in urban slums. Past industrialization efforts in the Philippines were biased toward urban centers, while rural industry was left in a rudimentary state. A major reason for underdeveloped rural industry can be found in the dual structure of the Philippine society. Dualism was, and still is, a common characteristic of all Asian countries in their early stages of development, but it is especially distinct in the Philippines because of the highly skewed asset distribu- tion, which has its origins in Spanish colonial rule. The majority of agricultural land, the most important traditional production asset, has been owned by a landed oligarchy in the peasant as well as plantation sectors, with many tenants and agricultural laborers on one I am deeply grateful for the invaluable assistance rendered by Nelly Fortuna and Wolly Tan during the field surveys of this research. I also thank all the garment factory owners, managers, and sewers who were interviewed during the surveys. My appreciation also goes to the govemment officials whose assistance was instrumental in the collection of information for this study. 89 90 Tohward the Rural-Based Development of Conunierce and Indulstry side and a small number of landlords on the other, the former residing in rural villages and the latter in cities. Resnick (1970) has clearly shown that the opening of the Philippines to the world economy around the turn of the century under the American colonial regime led to rapid economic development that was based mainly on the expansion of such export- oriented plantation crops as sugarcane and coconut, at the expense of household indus- tries in rural villages. Since the colonial period, the rural areas of the country have been predominantly agri- cultural; rarely were they the site of industrial activities.' Exceptions to this pattern have been sporadic agro-based industries, such as sugar-processing factories and coconut and rice mills, generally established by the landed elite or urban-based Chinese (including Chinese-Filipino ntestizo) merchants who emerged as entrepreneurs under American rule. In rural areas, this dual structure of the society, or "agrarian specialization and agrarian division of labor" (Resnick 1970: 65), has remained essentially unchanged since indepen- dence in 1946. This dual social structure may not be congruent with the promotion of rural industrial- ization. When agricultural surpluses are extracted by landlords, few funds are left in rural areas for investment in rural industries. More important, entrepreneurship, which is cru- cial for the evolution of rural industries, but has long been monopolized by a "landed- oligarchy-converted urban elite," may be nearly absent in the rural Philippines.2 Although the degree and mode of absentee landlordism differ across regions of the country, the Philippines bear a greater handicap in the indigenous evolution of rural industries than countries such as Japan, where rural industries were initiated in the early stages of mod- em economic growth by a large number of entrepreneurial peasants, mostly small to me- dium-size landowners living in rural villages. Nevertheless, the struggle to promote rural industries and other nonagricultural eco- nomic activities, if not overly successful, has always been a fixture of the rural Philippines (see, for example, Hayami and others 1990b). Furthermore, past land reform programs, coupled with the diffusion of green revolution technology, have contributed to keeping a significant portion of agricultural surpluses in rural villages. Such changes should have provided more favorable conditions for the emergence of rural industries. Recognizing the importance of promoting rural industries, the major questions to be asked in this chap- ter are: * Is the rural entrepreneurship needed to promote rural industries present in the Phil- ippines? If it is, who are the entrepreneurs? * How can the entrepreneurs mobilize and organize resources, capital and labor in par- ticular? * How are rural-based industries economically and institutionally viable? * What are their future prospects to transform themselves into modem industries? These questions are approached through a case study of the garment sector in Laguna Province. This industry was chosen because, as in other developing countries in Asia, it is one of the most important, rapidly growing industries in the Philippines in number of Export-Oriented Garnient Induistries in the Rutral Philippines 91 establishments, employment, and export earnings. More important, it is an industry found in some rural areas that employ the putting-out system. It is often said that rural-based industries enjoy advantages in mobilizing labor with low opportunity costs when adopt- ing the putting-out system, which carries minimal labor-monitoring costs because of close community ties (see the chapter by Hayami in this volume). The rural garment industry thus provides an opportunity to examine whether the advantages postulated in the devel- opment literature exist and, if they do, how rural entrepreneurs exploit them. Laguna, one of the provinces in Southern Luzon where many garment contractors and subcontractors are concentrated, was selected for this study. Situated southeast of Metro Manila, the province has been predominantly agricultural except in its northwest corner, which forms a corridor between Manila and all provinces in Southern Luzon and Bicol. It has rapidly been transformed into an urban/industrial area. Unlike such provinces as Batangas, where garment production was a major industry long before the garment ex- port boom began in the mid-1980s, Laguna is a relative newcomer in this industry, so that the diffusion of the industry into rural areas and the emergence of rural entrepreneurs, if any, can be observed more clearly than in other provinces. It is also expected that the broad representation of economic activity in Laguna, from peri-urban to purely agricultural ar- eas, makes it possible to observe different production and contract patterns in the garment sector under varying urban influences. The Garment Industry in the Philippines Struictural Changes in the Philippine Economy and the Garment Indutstry It is well-known that political instability has seriously hampered the economic develop- ment of the Philippines. During the 1980s in particular, political disturbance led to large negative growth rates. As a result of this economic regression, the level of per capita GNP (gross national product) in 1993, although significantly recovered from the low levels of the mid-1980s, was still at about the same level attained in the late 1970s. This does not, however, mean that the Philippine economic structure has been stable. The most drastic changes can be observed in the pattern of exports (table 5-1). Until the early 1970s, Philippine exports had been dominated by primary commodities, such as sugar, copra, and copper concentrates. Since the mid-1970s, this pattern, inherited from the colonial period before World War II, has been rapidly and radically replaced by one dominated by industrial commodities. (It should be noted that this change continued even during the economic turmoil of the 1980s.) Growth rates for industrial exports in two com- modity groups, electronics and garments, have been quite high. By 1990, the garment sec- tor had become the single largest export sector of the country, surpassing electronics.3 Export-oriented industrialization proceeded as in other developing countries in Asia. Partly because of its strong urban bias, industry was hit hardest during the political- economic upheaval of the mid-1980s. Although the sector has recovered since hitting bot- tom in 1985, some manufacturing subsectors have not come back to their 1980 production levels (table 5-2). It should be noted that this recovery process was disrupted again by 92 Toward the Rural-Based Development of Comnmerce and Industry Table 5-1. Changes in Export Shares of Major Product Groups in the Philippines (percent) Export 1971 1975 1980 1985 1990 1993 Ten traditional exports, 77 71 46 24 14 13 Other exports 23 29 54 76 86 87 Industrial products 58 71 n.a. Electronics 22 19 16 Garment 13 22 20 Total 100 100 100 100 100 100 n.a. Not available. a. Copra, sugar, bananas, logs and lumber, desiccated coconut, coconut oil, pineapples, gold, abaca, and copper concentrates. Soutrce: NSCB (various issues) and NSO (various issues). Table 5-2. GrossValue Added in Manufacturing by Industry Group (billion pesos) 1993/1980 Induistry 1980 1985 1990 1991 1993 (percent) Food manufactures 75.3 64.4 69.6 68.6 65.6 -13.0 Beverage 4.0 6.5 7.4 7.5 6.8 70.0 Tobacco 4.9 5.0 5.1 5.0 4.9 0.0 Textile 7.9 5.3 6.7 6.5 5.7 -28.0 Garments, footwear 7.2 5.7 10.0 10.7 11.6 61.0 Wood and cork products 6.4 2.8 4.1 4.0 3.5 -45.0 Furniture and fixtures 2.5 1.7 2.6 2.4 2.1 -16.0 Paper and paper products 1.7 1.5 2.2 2.2 1.8 6.0 Publishing and printing 2.3 1.4 3.0 3.0 2.8 22.0 Leather . 0.2 0.1 0.1 0.1 0.1 -50.0 Rubber products 2.8 1.9 2.8 2.8 2.3 -18.0 Chemical products 12.4 10.6 11.5 12.1 11.5 -7.0 Petroleum and coal products 16.4 18.1 28.1 26.3 32.0 95.0 Nonmetallic minerals 4.6 2.7 4.6 5.0 5.2 13.0 Basic metal industries 2.7 4.4 4.9 5.0 4.2 56.0 Metal industries 4.0 2.9 4.5 4.1 4.4 10.0 Machinery 2.3 1.3 2.1 2.1 2.3 0.0 Electrical machinery 3.7 4.7 7.1 7.3 8.7 135.0 Transport equipment 5.2 0.9 1.9 1.7 2.4 -54.0 Miscellaneous manufactures 2.0 2.0 3.7 4.0 3.5 75.0 Total 168.3 143.9 182.0 180.5 181.3 8.0 Note: Figures represent 1985 prices. Souirce: NSCB (various issues). Export-Oriented Garmienit Iizditstries in the Rutral Philippines 93 political instability during the early 1990s, when Ramo's administration succeeded Aquino's. The electric power crisis was the most serious, visible sign of adverse conditions for industry. The garment industry was one of the subsectors that recorded quick recovery by 1990. Further development-in spite of the difficulties in the early 1990s-has made this sector the third largest among the industry groups in gross value added, up from fifth largest in 1980. Characteristics of the Garment Indutstry The garment sector is the second-largest industry in numbers employed, after food manu- facturing (table 5-3). Rough estimates of labor's factor share in gross value added (the ratio of wages/salaries per paid employee to average labor productivity, expressed in gross value added) indicate that the garment industry is one of the most labor-intensive manufacturing sectors. Its share is 51 percent in the case of large firms (with 10 workers or more) and 69 percent in the case of small firms, compared with an average of 24 percent for the manufacturing industry as a whole. Development of the garment industry, with it many capital-intensive subsectors, should have contributed significantly to the provision of productive employment opportunities for the growing labor force. Its relatively low capital-labor ratio (the other side of the coin from the high share of labor in gross value added), however, leads to relatively low labor productivity in the garment sector in com- parison with capital-intensive industry groups. The level of wages/salaries paid to em- ployees of large garment firms is also less than the average for large firms as a whole. Table 5-3 also shows that the garment industry is characterized by a distinct dual struc- ture. In 1988 there were about 1,500 large garment firms (ninety-one employees on aver- age), while the number of small garment firms (three workers per firm on average) was more than 12,000. The difference in firm size corresponds to differences in labor produc- tivity and wage/salary levels, which in large firms are three and two times, respectively, their levels in small firms. It should be noted, however, that the duality in productivity and wages is much less pronounced in the garment industry than in other major manufac- turing sectors. For example, labor productivity in large firms in the food manufacturing industry is about ten times that of small firms, and it is nearly forty times greater in the beverage sector. Wage/salary levels differ by a factor of four in these two sectors. Subsectors of the Plhilippine Garmlent Industry At least three subsectors can be distinguished in the garment industry: the export, ready- to-wear (RTW), and order-made tailor sectors (figure 5-1).4 Exporters (also called contractors) are key figures in the export sector. They receive bulk orders directly from foreign buyers or through merchandisers (or brokers), together with design and materials. The materials-cloth, thread, and the like-are all imported, tax- and duty-free, through the Bonded Manufacturing Warehouse (BMW) of the Garment and Textile Export Board (GTEB), a government body set up in the Department of Trade 94 Torvard the Rural-Based Developmnent of Comnerce and lndistry Table 5-3. Profile of Philippine Manufacturing Firms by Major Industry Group, 1988 Wagesl Emfploynment Labor salaries Percent Value of Value produc- per paid of un?paid ouitput added tizvity employee Numreber Per enmploy- (million ratio (1,000 (1,000 Grouip of firmis Total firml ees pesos) (percent) pesos) pesos) Large firms: Total 11,488 856,951 75 1 385,010 35 156 37 Food manufacturing 3,070 161,127 52 3 83,010 31 158 37 Beverage 103 30,249 294 0 25,189 62 516 58 Tobacco 32 14,357 449 0 16,194 53 593 36 Textile 547 89,499 164 0 19,556 31 68 28 Garment 1,556 142,160 91 1 15,210 51 55 28 Leather 120 5,167 43 2 531 39 40 22 Footwear 441 9,589 22 9 948 34 34 14 Wood and cork 682 56,433 83 1 10,936 34 67 28 Furniture and fixtures 678 41,276 61 2 3,867 43 40 20 Paper and paper products 178 15,433 88 0 10,247 35 234 44 Printing and publishing 636 20,311 32 2 4,166 40 88 34 Industrial chemicals 115 10,879 95 0 14,433 38 499 64 Other chemicals 301 29,700 99 0 30,560 40 416 88 Petroleum refineries 4 2,431 608 0 44,482 18 3,322 189 Rubber 174 26,181 150 0 8,428 37 120 43 Plastic 300 17,616 59 0 6,698 34 128 31 Pottery and earthenwear 59 5,066 86 1 786 59 91 38 Glass 35 5,776 165 0 3,313 59 309 76 Cement 17 5,893 347 0 5,582 34 326 53 Other nonmetallic minerals 353 11,219 32 3 1,990 49 86 27 Iron and steel 127 15,507 122 0 19,762 26 330 48 Nonferrous metal 35 2,618 75 1 11,937 25 1,157 55 Fabricated metal 483 21,733 45 1 5,965 32 88 29 Machinery 552 18,576 34 2 3,318 31 55 31 Electrical machinery 214 54,374 254 0 25,628 30 132 53 Transport equipment 230 14,053 61 1 8,895 28 179 42 and Industry to promote garment and textile exports. In garment production, exporters depend heavily on subcontractors working through putting-out contracts. Subcontractors often put out parts of sewing operations to additional sub-subcontractors. Finished gar- ments are exported to foreign countries through buyers. The most popular foreign mar- kets for Philippine garments are supermarkets and department stores that make volume sales of relatively low-quality clothes for the middle and,lower classes in countries such as the United States and nations of the European Union.5 The numbers of these exporters and subcontractors have mushroomed since the late 1980s, supporting rapid increases in garment exports (table 5-4). There are signs that de- Export-Oriented Garmenet IndiLstries in the Rutral Philippines 95 Wages/ Employmnient Labor salaries Percent Value of Value produc- per paid of unpaid output added tizvity employee Numnber Per ermploy- (million ratio (1,000 (1,000 Groutp offirms Total firm ees pesos) (percent) pesos) pesos) Small firms: Total 67,147 234,428 3 47 11,294 36 17 11 Food manufacturing 30,359 109,229 4 48 5,443 31 16 10 Beverage 884 2,309 3 70 31 61 14 13 Textile 1,316 3,613 3 60 98 45 12 10 Garment 12,482 31,404 3 56 886 55 16 11 Leather 209 909 4 40 122 33 44 22 Footwear 1,703 11,024 6 39 368 47 16 11 Wood and cork 2,043 8,472 4 44 391 38 18 13 Furniture and fixtures 4,077 15,519 4 44 559 46 17 11 Paper and paper products 145 760 5 46 53 35 25 12 Printing and publishing 1,522 8,136 5 27 1,455 19 33 20 Industrial chemicals 37 180 5 15 62 43 150 27 Other chemicals 115 506 4 38 48 31 29 16 Rubber 254 1,185 5 40 52 52 23 15 Plastic 83 402 5 20 39 59 56 17 Pottery and earthenwear 712 2,009 3 75 29 61 9 8 Glass 24 126 5 27 7 69 18 16 Other nonmetallic minerals 1,830 6,714 4 42 276 41 17 10 Fabricated metal 6,035 17,406 3 53 580 49 17 14 Machinery 1,277 6,388 5 25 443 51 36 20 Iron and steel 60 285 5 27 21 35 26 18 Nonferrous metal 59 238 4 56 6 62 16 14 Electrical machinery 63 232 4 32 14 55 32 19 Transport equipment 283 971 3 38 57 49 28 14 Grand total 78,635 1,091,379 14 11 396,304 35 126 30 Note: Based on the 1988 Census of Establislh"ments. "Large" firms are those with ten or more workers; "small" firms have fewer than ten workers. Labor productivity is expressed in gross value added. Souirce: NSO (1990). velopment of the export subsector of the garment industry has accelerated since 1987. There was a clear increase in both the number of garment exporters registered with the GTEB and the total employment in the garment sector as a whole between 1986 and 1987. Paralleling the increase in exporters, the number of garment subcontractors for exports, also registered with the GTEB, increased rapidly from the late 1980s to the early 1990s (table 5-5). Exporters/contractors are located either in large cities or in export-processing zones (EPZs), where privileges are given to firms. This is clearly demonstrated by the regional distribution of subcontractors for export garments shown in table 5-5. As shown below, 96 Toward the Rutral-Based Devielopmenit of Comnterce anid Indistry Figure 5-1. Garment Industry in the Philippines: Subsectors Export Sector Domestic Market Sector Order-Made Final consumers RTW Subsector Tailor Subsector in foreign countries (e.g., U.S.) Final consumers in the Philippines Local markets in Department stores Luzon (Baguio), Supermarkets Visayas (Cebu), (e.g., in U.S.) Mindanao (Davao), and the like Buyers Department stores (outside the country) Supermarkets (Design, materials) in Metro Manila (High-end market) Merchandisers: Baclaran Market Brokers in Manila (outside the country) Shop = stallkeepers (Design, materials) (Wholesaling) (Low-end market) Exporters = Contractors S (Production) (di RTW garmnent firmsa _ _ ~~~~~~~(Design, material selection, ) c \ < ~~~~~~~production)v 4 ucntractors _ < > < <\ t ~~~~~~~~~~Tailors in towns/cities K Su-subotatrs > ucnrcors (Production upon order) <= ~~~~~~~~Podcin Export-Orienied Garment Iniduistries in the Rutral Philippinies 97 Table 5-4. Number of Garment Exporters, Total Garment Exports (FOB), and Employment Index of the Garment Sector Nulmber of Valute of garnmenit exports (US$ million) Employment in the gar?11en1t To quota To noinqutota garment sector Year exporters couintries conltries Total (Index: 1985 = 100) 1984 706 568 79 647 109 1985 747 577 82 659 100 1986 754 672 131 803 105 1987 919 993 155 1,148 127 1988 1,061 1,092 183 1,275 126 1989 1,086 1,352 238 1,590 148 1990 1,207 1,574 247 1,821 147 1991 1,069 1,588 320 1,907 147 1992 1,295 1,762 430 2,192 163 1993 1,428 1,869 473 2,343 163 Note: The employment index is for the garment sector as a whole. Source: Garment and Textile Export Board and Department of Trade and Industry. the primary reason exporters put out garment production to subcontractors is to cut labor costs. Subcontractors should therefore be located in regions where wage rates are lower and government labor regulations weaker. At the same time, the transaction costs involved in managing subcontractors, particularly for tight quality control of their products, make it more convenient-even necessary-to use subcontractors located as close as possible to Table 5-5. Garment Subcontractors in the Philippines Total )Distribution of garment subconztractors by region, 1993 number Nuimber of Total number of workers of garment garm7ent Nlinber subconi- subco7t- Admi77i- of Year tractors Region tractors Direct Indirect strative Total macliines 1988 1,669 Ilocos Region 0 0 0 0 0 0 1989 2,780 Cagayan Valley 1 8 2 5 15 8 1990 2,923 Central Luzon 470 10,568 2,524 1,146 14,238 12,119 1991 3,125 Metropolitan Manila 1,172 49,906 17,367 7,023 74,296 57,612 1992 2,613 Southern Luzon 715 22,250 6,245 2,752 31,547 24,517 1993 2,396 Bicol Region 1 36 0 6 42 31 Western Visayas 2 469 20 8 497 50 Central Visayas 35 818 175 62 1,055 764 Eastern Visayas 0 0 0 0 0 0 Mindanao 0 0 0 0 0 0 Total 2,396 84,355 26,333 11,002 121,690 95,101 Source: Division of Bonded Manufacturing Warehouse, Garment and Textile Export Board. 98 Towvard the Riral-Based Dev7elopment of Commerce and indistry the parent company. As a result, many subcontractors are located within Metro Manila, where most garment exporters gather, and in Southern and Central Luzon, adjacent to Metro Manila. Aside from Metro Manila and its surroundings, another small cluster of subcontractors is found in Central Visayas; Metro Cebu is the central city with an EPZ nearby. The RTW sector produces ready-to-wear clothes for domestic markets. Unlike export- ers/contractors, who generally produce garments according to designs made, and using materials given, by the merchandisers/buyers, either in-house or through subcontracting, RTW firms usually cover the entire production process for garments. They handle every- thing from the selection and purchase of materials to cutting and sewing, and often to marketing. Most RTW firms have a similar number of workers and machines to the gar- ment subcontractors for exports. It is not rare for them to depend on subcontractors, but RTW subcontractors (typically housewives) are more often not organized. They do con- tract jobs in individual houses rather than in organized factories or workshops. Two subsectors can be discerned within the RTW sector: products for low-end mar- kets and those for high-end markets. In the former, garments are typically sold by RTW firms to wholesalers at Baclaran Market in Manila; they are then sold to retailers and wholesalers in such remote regions as Visayas and Mindanao. RTW firm owners some- times keep stalls in Baclaran, engaging in marketing on their own and using the stalls as sources of market information and knowledge about consumers' preferences. In the high- end RTW subsector, garments are made specifically for select supermarkets and depart- ment stores in Metro Manila that have customers from rich families and the burgeoning middle class. Good design and neat clothes are more critical for this high-end market than for the low-end market.6 RTW firms and subcontractors for export are interrelated- many of the firms have converted themselves into subcontractors as the demand for subcontractors has increased. There are also some RTW-converted subcontractors who have switched back to the RTW sector. The technology and skills adopted by RTW firms and subcontractors for export are essentially the same. Garment entrepreneurs, and their associated labor and capital, are therefore largely interchangeable at this level between the two sectors. The flows of materials and finished garments, however, are entirely separate and do not cross between these sectors.7 The third garment subsector may be called the order-made tailor sector. In any town or city in the Philippines, there are many tailor shops that produce clothes upon order from customers. They produce a surpris- ingly wide range of clothes, from casual, everyday clothes to formal suits and dresses. Like RTW firms, these independent tailors do everything in garment production. How- ever, they differ from RTW firms on one important point: because they sell their own skills to their customers, they rarely depend on subcontractors. This would also explain why few large tailors hire a number of sewers. Many of the small garment firms shown in table 5-3 are these tailors. There are cases, although infrequent, in which subcontrac- tors for export or RTW firms accept jobs for order-made clothes from their neighbors, but they consider order-made jobs as their side business and clearly distinguish them from their main business. Export-Orienited Garmen1t Indiustries in the Ruiral Philippines 99 Garment Subcontractors for Exports in Laguna Province Laguna Province includes the southern coast of Laguna de Bay, the Philippines' largest lake (figure 5-2). A strip of lowland along the lake rises to upland areas that continue to the surrounding mountainous provinces of Cavite, Batangas, and Quezon. Until recently, the whole of Laguna had been agricultural; rice, coconut, and sugarcane were the major prod- ucts. The lowland strip forms one of the most productive rice-growing areas of the coun- try. Developed since the early Spanish period (sixteenth to seventeenth centuries), it was a major supplier of rice to Manila, using boat transport across the lake. Sugarcane has been planted in some upland areas between the lowland strip and the foothills or more moun- tainous parts of the province, particularly in Calamba and surrounding municipalities. Partly because this section of the province has been rapidly urbanized in the last few de- Figure 5-2. Laguna Province, Philippines ETRO. " RAL PROVINCE 7" - g' ' .. f/X8 - T ,R PROVINCE Z ~ ~ ~ ~~~x / Santa QEO 7 I ~~~~~~~~~~~~Maria iMabitac Siam \ Sinilioan' / > l - ' 4 : - y 8 . A Pangil,' \ Jr , ^ I I Paete /'Santa Rosa 7 Kalayaan \ CAVITE / / ox ' de 9y j PROVINCEv , Laguna, di Bay Grin' " Lumicin- 1~~~~ , , Ct l4rba P15-: s> y ,f\1____ Rants' gx,,.7 ~~~~~ Luisiana'-/ 'K / 'V '/B o) Magdalena /('. - / < , LAGUNA PROVINCE V %yvLjayjayX \ N X s,/'Calauan 0 - : t-s rv X \, o' ,,-~~~~ t ~ SN,,arlan ,.~~/ 74~~ RizaI\\ A fofmnp: m \ San Pablo"' , , . 3 '>> - \ \Alaminos, City ,t' . City P1LJP1~4S/ ,;3i - ¾dr-tX:- -: l \, -5-. (1 i) Province capital r v: `B-ATANAS 0 Municipalities - - f,"7."' - BATANGAS l_**_ ------- Municipal boundaries , 5 10 National road At7 K, 1D8KO ES rhPROVINCE 1/' -- laRD 29580 ~ KILOMETERS - Super highway 100 Towvard thte Rural-Based Developmenit of Co0mmnerce an1d Iniditstry cades, and partly because the price of sugar has been unfavorable, sugarcane has been slipping from its status as a "major" agricultural product of the province. Hilly terrain, which is covered nearly 100 percent by coconut trees, constitutes about three-fourths of the province. Among these agricultural areas are local towns (rnunicipio), where the municipal office, town church, market, and small stores are located. These towns are small in population, with the exception of several municipalities in the northwest corner of the province. Apart from their functions as administrative units and minor collection and distribution points for locally produced agricultural commodities, these towns do not perform the functions of regional centers. Their quiet but dormnant atmosphere shows only rare signs of rural industry, with the exception of a few rice mills, sugar centrals, and small shops for car repair and vulcanizing.8 The rural villages in Laguna are predominantly agricultural, ex- cept for some coconut-related industries mentioned below. The rice and coconut areas in the province, although intermingled at their margins, form not only different agricultural zones, but also involve farmers of different natures. Rice is basically a subsistence crop, and rice farming in this region is practiced as monoculture. As in other parts of the Philip- pines, absentee landlordism is pervasive in the area. Most rice farmers cultivate tiny plots of paddy fields as tenants of some sort,9 and rice villages are largely inhabited by tenant farmers and landless laborers, who are even more numerous than the farmers. In contrast, coconut is a multipurpose commercial crop. In addition to the main product of coconut oil (copra), it produces desiccated coconut, juice, wine, nata-de-coco, lumber, charcoal, brooms and mats (from leaves), and many other goods. Furthermore, many other commercial tree crops, such as bananas, papayas, lanzones, and coffee, are intercropped under coconut trees. Unlike rice farmers, coconut farmers in this area are mostly small to medium-size owner-operators (5 to 20 hectares), and they reside in coconut villages together with coco- nut tenants and coconut laborers. These differences between rice and coconut lead to a difference in the degree of agriculture-based rural industry. Few rice-based rural indus- tries are found in the rice-growing areas (with the exception of rice milling), but numerous industries, including copra making, wine (lambanog) distillery, food processing, and handi- crafts, have been developed in the coconut-growing areas. For convenience in this analysis of garment subcontractors in Laguna, the province is divided into four sections (see figure 5-2 and table 5-6). Laguna I, covering the municipali- ties from San Pedro to Calamba, is the corridor mentioned earlier. Rice and sugarcane were once grown in this area, but urban subdivisions and industrial factories have spread during the past two decades. Because of the superhighway running up to Calamba, it is a one-hour drive from the center of Manila to the end of this section, if traffic is not heavy. Laguna II consists of municipalities that are less than two hours' travel from Manila. La- guna III and IV are also plotted according to travel time from Manila, less than three hours in the case of Laguna III and more for Laguna IV. Unlike Laguna I, which is now virtually a part of Metro Manila, Laguna II to IV are still predominantly agricultural. The municipalities along the coast of Laguna de Bay from Calamba to Sta Cruz (part of Laguna II) form a major rice-growing area, and farmers there Export-Oriented Garment Indutstries in the Ruiral Phtilippines 101 Table 5-6. Number of Garment Subcontractors and Their Workers in Laguna, 1993-94 Denisity Total Numniber of (per 1,000 poptulationt) Nuember of iiniber zvorkers per Number of Nutmber Area contractors of zvorkers subcon tractor sutbconitractors of wvorkers Laguna, total 189 6,528 35 14 476 San Pedro 36 1,587 44 23 1,014 Binan 12 320 27 9 238 Sta Rosa 32 814 25 34 859 Cabuyao 10 298 30 15 445 Calamba 37 2,402 65 21 1,385 Laguna I 127 5,421 43 20 865 Alarninos 10 258 26 36 941 San Pablo 18 281 16 11 174 Los Banos 5 82 16 8 124 Bay 2 26 13 6 80 Calauan 3 43 14 9 131 Pila 1 11 11 4 40 Sta Cruz 7 149 21 9 195 LagunalI 46 850 18 11 200 Nagcarlan 3 66 22 8 175 Liliw 5 75 15 23 342 Laguna III 8 141 18 13 237 Magdalena 2 37 19 15 275 Luisiana 1 16 16 7 112 Pagsanjan 1 10 10 4 40 Lumban 1 18 18 5 91 Kalayaan 1 14 14 8 107 Pakil 1 9 9 7 67 Sta Maria 1 12 12 5 58 LagunaIV 8 116 15 7 97 Note: One subcontractor for weaving with 1,050 workers is excluded from the list. Source: Garment and Textile Export Board. are typically tenants of rice monoculture. Coconut is grown mainly in the hilly parts of the province, stretching from Alaminos/San Pablo (another part of Laguna II), through Nagcalran/Liliw (Laguna III), to Magdalena/Luisiana (part of Laguna IV). The north- eastern corner of the province (the rest of Laguna IV) is not only far from the center, but it also has relatively meager land resources because of the topography: a steep mountain range closely parallels the shore of the lake, leaving little room for either paddy fields or coconut land. To recapitulate, the degree of urban influence decreases rapidly as one goes further from Laguna I to Laguna IV, and it is almost entirely absent in the last two sections. 102 Toward the Rtural-Based Development of Commerce anid Indurstry Garment Stubcontractors Table 5-6 lists 189 export subcontractors who were registered with, and accredited by, the GTEB in 1993-94, according to their locations. In a clear trend, the intensity as well as the number of subcontractors decreases as we go farther from Metro Manila, where most of the garment exporters (contractors) are concentrated. Such a trend seems to indicate that the cost of managing subcontractors increases for the principal contractors as the distance between them increases. The low density of subcontractors at a location in Laguna would make the cost of using subcontractors directly proportional to the distance. In areas where many subcontractors are concentrated in one neighborhood, as in some traditional gar- ment-producing towns in Batangas, the cost per subcontractor could be low, even if the traveling distance from the parent contractor were greater." It should be noted, however (table 5-6), that although the density of garment subcontractors decreases according to the distance from Metro Manila, some irregularities in this trend are found in selected towns, such as Alaminos, Liliw, and Magdalena. These towns, with higher densities of subcon- tractors, are all located in the coconut area, and the density is clearly low in the Laguna II towns (from Los Bafios to Pila) situated in the best rice-growing area of the province. Most of the subcontractors in Laguna are small, employing thirty-five workers on the average. Eighty percent employ fewer than thirty, and more than half employ ten to twenty workers (table 5-7). However, not all are small: some are very large, and the largest em- ploy more than 500. There is a clear tendency for firms to become smaller as the distance from Metro Manila increases (table 5-6); all the large subcontractors with more than 100 workers are in Laguna I."1 The services these subcontractors render are numerous, ranging from cutting cloth through embroidery/smocking/appliqu6 and sewing (table 5-8). There are also some special services, including beading and injection molding (for hangers). Knitting services are for such products as sweaters and gloves. It is customary for each Table 5-7. Size Distribution of Garment Subcontractors in Laguna, 1993-94 Size offirnm in nutmber of workers Nuember offirns Fewer than 10 29 (15) 10-20 97 (51) 20-30 26 (14) 30-40 11 (6) 40-50 7 (4) 50-100 10 (5) 100-200 5 (3) 200-500 2 (1) 500-1,000 2 (1) Total 189 (100) Note: Figures in parentheses are percentages. Source: Garment and Textile Export Board. Export-Oriented Garmenit Iindutstries in the Rural Philippinies 103 Table 5-8. Number of Garment Subcontractors and Number of Workers per Subcontractor by Service and Major Products, Laguna Province, 1993-94 Region of Lagutna Number of subcontractors Numniber of 7workers per subcon tractor Task I II III IV Total I II III IV Total Service Sewing 94 42 7 6 149 47 18 17 14 36 Embroidery 14 1 0 2 17 32 15 0 17 29 Knitting 9 0 1 0 10 44 0 22 0 42 Beading 4 1 0 0 5 17 36 0 0 21 Smocking 1 0 0 0 1 26 0 0 0 26 Applique 1 1 0 0 2 10 24 0 0 17 Injection molding 1 1 0 0 2 18 23 0 0 21 Cutting 1 0 0 0 1 19 0 0 0 19 Other 2 0 0 0 2 15 0 0 0 15 Total 127 46 8 8 189 43 18 18 15 35 Product Baby/children's wear 40 32 4 4 80 27 18 20 13 22 Coat/jacket 20 4 2 1 27 81 15 13 21 64 Sweater 6 0 0 0 6 44 0 0 0 44 Polo/blouse 4 2 0 0 6 33 19 0 0 28 Athletic suit 12 0 0 0 12 71 0 0 0 71 Shorts 7 1 0 0 8 17 19 0 0 17 Gloves 1 1 1 0 3 26 18 22 0 22 Accessories 7 1 0 1 9 14 36 0 16 16 Trousers 6 3 1 0 10 31 24 15 0 27 Underwear 2 0 0 0 2 11 0 0 0 11 Dress 6 0 0 0 6 15 0 0 0 15 Pajamas 4 1 0 0 5 13 27 0 0 15 Other 12 1 0 2 15 75 23 0 14 63 Total 127 46 8 8 189 43 18 18 15 35 Note: For the regions in Laguna, see Table 5-6. Source: Garment and Textile Export Board. subcontractor to specialize in a selected service. Some perform more than one service, such as sewing and embroidery, but they are the exception. The overwhelming majority of Laguna subcontractors are specialized in sewing, the service in greatest demand in gar- ment production, which at the same time requires less skill than other specialties, and few machines except sewing machines. The list of garment products that the subcontractors make is also long. Here, too, there is specialization among subcontractors: some produce mainly baby and children's wear, while 104 Towvard the Rmral-Based Dezvelopnient of Comm1ierce anid Iniduistry others create coats and jackets. Specialization in products, however, is much looser than in services. Many subcontractors that specialize in baby and children's wear, for example, accept sewing jobs for coats and jackets, ladies' dresses, shorts, and other items. The most popular product among Laguna subcontractors is baby and children's wear, followed by coats and jackets and trousers and shorts. The specialized services supplied by subcontractors suggest that garment exporters depend heavily on subcontractors for selected steps in garment production. As will be revealed by the sample survey in the next section, some exporters put alnost all produc- tion processes out to subcontractors;"2 they merely perform the final checking and revis- ing of finished garments and their packing and shipping. Profile of Firms and Entrepreneurs Sample Suirvey In order to obtain detailed information on the nature and characteristics of garment firms in Laguna, a sample survey was conducted of the owners or managers of garment firms in August of 1994. The main target of the survey was subcontractors in the export garment sector of Laguna II-IV. In addition, a small number of exporters and RTW garment firms were interviewed to obtain general information about the garment sector as a whole. The subcontractor sample was first drawn randomly, based on a list of subcontractors registered in 1990-91 that had been prepared by the provincial office of the Department of Trade and Industry. Soon after commencing the survey, however, it was found that the turnover or closure rate of subcontractors was extremely high; in many municipalities, about 50 percent of the subcontractors listed were out of operation by the time of the survey. Many unregistered sub-subcontractors and subcontractors were also found. There- fore, many unlisted firms were added to the sample "randomly" whenever they were found. The sample exporters were selected partly from among those in contractual rela- tionships with the sample subcontractors, and partly from the subcontractor list-that is, exporters that had once been subcontractors. Altogether, individuals at forty-eight firms were interviewed; forty were subcontractors and the rest were divided equally between exporters and RTW firms (table 5-9). All the subcontractor samples but one were from Laguna II-IV."3 It should be noted that three subcontractors that have since closed their garment firms are included in the subcontrac- tor sample. Two of the exporters in the sample are located in Metro Manila, and two of the RTW examples are in Taytay, Rizal, a town that has long been famous for producing RTW garments.'4 Sample Garnment Firms Many sample subcontractors are recent establishments, with an average of fewer than eight years of operation (table 5-9). Some began their operations before 1970, which sug- gests that the garment subcontracting business for export is not entirely new. The starting Export-Oriented Garnmenit Itididstries in the Rutral Philippines 105 Table 5-9. Sample Garment Firms in Laguna: Number by Type, Location, Year Opened, Service, and Product Nuimber offirnis Firnisfor Suibconitractor domestic for export Exporter market Total Total 40 4 4 48 Location Lagunal 1 1 - 2 Laguna II 27 1 2 30 Laguna III 5 - - 5 Laguna V 7 - - 7 Metro Manila - 2 - 2 Rizal (Taytay) - - 2 2 Year opened Before 1970 1 1 - 2 1970s 2 1 3 6 1980-84 5 - 1 6 1985-86 4 - - 4 1987-88 7 - - 7 1989-90 6 2 - 8 1991-92 10 - - 10 1993-94 5 - - 5 Average years of operation 7.3 20.8 20.3 9.4 Type of service Sewing 38 - - 38 Sewing and embroidery 1 - 1 2 Knitting 1 - - 1 General - 4 3 7 Major product Baby dress (BD) 26 1 27 Jacket (J) 4 1 - 5 BD+J 2 1 - 3 Trousers 4 1 - 5 Shorts 2 - - 2 Girl/lady dress - - 3 3 Gloves 1 - - 1 Brassier 1 - 1 Blouse and T-shirt - - 1 1 Soutrce: Author's data. years for the subcontractors, however, clearly indicate that the garment export sector has grown dramatically since the late 1980s. This trend is consistent with the inclusion of two new exporters in the sample (both started their operations in the late 1980s), and one is quite old, established in the late 1940s. 106 Tozvard the Rutral-Based Developmenit of Conimerce and Inidustry The major service rendered by the sample subcontractors is sewing, as is the case for the subcontractor population in Laguna (table 5-9). As mentioned earlier, of the various ser- vices, sewing requires the least skill and capital. For instance, cutting requires more skill than simple sewing, and buttonholing requires both skill and special machines that are more expensive than sewing machines. Without exception, sewing subcontractors per- form only the sewing parts of garment production. A subcontractor receives a job order from a contractor (exporter) together with the designed pattern (sample clothes), cut ma- terials, and thread, and sews the ordered garment, usually finishing work before the final stage. The contractor completes the nearly finished garments, completing such difficult tasks as buttonholing and fixing the crotch/zipper."5 This incompleteness of the produc- tion process-or the separation of garment production into many different processes-in the export sector is in sharp contrast to the RTW sector. RTW firms, more or less equivalent in size of operation to export subcontractors, but with much smaller lot sizes for each garment, do everything from designing to selection and purchase of materials, sewing the final product, packing, and marketing. Subcontractors accept jobs outside their specialized products. What they will accept depends on the pace of demand (job orders) for their service and the price of contracts. Baby clothes, jackets, and shorts require only a sewing machine, which enables any sew- ing subcontractor to accept orders for these products, although their sewers sometimes seek higher productivity by trying to specialize in a given product. Subcontractors of prod- ucts that require special machines, such as brassieres and gloves, are generally specialized by product. Two kinds of subcontractors are distinguished by their degree of dependency on con- tractors/exporters. Freelancers contract with more than one contractor, not only over time but also at the same time. "Parent-child" subcontractors work exclusively for one contrac- tor who provides not only all job orders, but also the necessary machines, a quality con- troller, a machine mechanic, and even a supervisor; the subcontractor provides only sewers and a workshop. In an extreme form of freelance, subcontractors select contractors accord- ing to their own criteria. Although actual subcontractors lie between these two poles, freelancers usually produce multiple products, and subcontractors that specialize in a spe- cific product are generally of the "parent-child" kind."6 Most of the sample subcontractors employ ten to twenty workers (table 5-10). Just over 10 percent are "large," with more than thirty workers. The optimum size for subcontrac- tors in order to minimize transaction costs appears to be about ten to thirty workers; it is more difficult to get good job orders with fewer than ten workers,17 and the cost of labor management in single owner-manager-operated firms increases disproportionately when the number of workers exceeds thirty. The most serious problems subcontractors face are a lack of working capital and late payments by contractors, as well as difficulties in managing sewers. The labor-manage- ment problem is the most serious source of headaches for RTW firms (their size is compa- rable to that of export subcontractors). Two of the subcontractors who had gone out of the garment business mentioned this problem first as their reason for shutting down. Many Export-Oriented Garment Industries in the Rutral Philippines 107 Table 5-10. Size Distribution and Sewers' Wage, Sample Garment Firms Number offirnms Firmsfor Subcontractor domestic for export Exporter nmarket Total Total 40 4 4 48 Size (number of workers) Below 10 9 - 1 10 10-19 19 - - 19 20-29 7 - 2 9 30-49 3 - 1 4 50-99 2 - - 2 100 and more - 4 - 4 Average number of workers 17.0 272.3 24.0 38.9 Size (number of machines) Below 10 6 - - 6 10-19 20 - 1 21 20-29 8 - 2 10 30-49 4 2 1 7 50-99 2 - - 2 100 and more - 2 - 2 Average number of machines 20.7 121.5 24.5 29.4 Average weekly sewers' wage P395 P823 P988 P499 Source: Author's data. subcontractors within the "optimum" size group also expressed a desire to maintain their present size in the future. There are a few exceptions-subcontractors far beyond the opti- mum size that still want to expand. All of them have aspirations to climb the ladder to- ward principal contractor/exporter. The size distribution of sample garment firms in the number of machines in place has a pattern similar to that of their distribution by number of workers (table 5-10). For subcon- tractors and RTW firms, the number of workers is determined by the number of sewing machines, particularly high-speed sewing machines. On average, a subcontractor has twenty machines: fifteen high-speed sewing machines, two ordinary sewing machines with motors,"8 two over-edging machines, and one other special machine (table 5-11). This machine list corresponds to an average employment of seventeen workers: fourteen sew- ers, two revisers, and one supervisor (table 5-11). It should be noted that the subcontrac- tors with ordinary sewing machines are generally found in remote parts of Laguna; it is almost impossible to recruit sewers without high-speed machines. Exceptions to this sewer- machine matching rule are found among exporters (table 5-10). All four sample exporters hire more than 100 employees, but two of them have fewer than fifty machines. This anomaly 108 Towvard the Ruiral-Based Development of Conmmerce and Industry Table 5-11. Average Number of Machines and Workers for Thirty-nine Sample Subcontractors for Sewing Average niumber Average niumber of 7workers Category of machinzes per sulbcontractor Machines Total 20.8 Sewing High-speed single 14.9 High-speed double 0.2 Ordinary 2.3 Edging machine 5-thread 1.2 3-thread 1.0 Other special machines 1.2 Workers Total 16.8 Sewer 14.2 Reviser/trimmer 1.6 Supervisor 1.0 Source: Author's data. is explained by the practice of some exporters of putting-out all production operations to subcontractors except final checking, packing, and shipping. A major reason for exporters to subcontract is to exploit cheap labor, because small sub- contractors in rural areas are not subject to the minimum wage and other tight labor regu- lations. Although there would be a tradeoff between the transaction costs associated with utilizing subcontractors and labor cost differences, the disparity in the effective wage rates between subcontractors and exporters seems to be large enough to make subcontracting attractive to exporters (table 5-10). The minimum wage rate of the region is currently P145 daily; among the exporters, the average sewers' wages are P823 weekly, roughly equal to this daily rate multiplied by six working days. In contrast, the average sewer's wage among the sample subcontractors, all of whom are paid piece-rate wages, is as low as P395 a week. Considering that subcontractors' employees usually work about ten to fifteen hours each day, the wage gap between exporters and subcontractors could be even larger."9 Entrepreneturs What kind of people own and manage the sample garment firms? In all the sample sub- contractors and RTW firms that are unincorporated family businesses, the owners them- selves manage their firms, although in some cases spouses participate in the management and assist in firm operations. In the case of the sample exporters, all of which are incorpo- Export-Oriented Garmient Industries in the Rutral Philippines 109 rated, the two smaller examples are managed by the owners, but the two large ones are operated by hired managers. The owner/managers are relatively young, mostly in their thirties and forties, and pre- dominantly female (though there are male owner/managers). An overwhelming majority are college graduates (table 5-12). Nearly 60 percent have backgrounds, judged by their parents' occupations, on farms or other rural origins. This percentage increases to 70 per- cent among subcontractors. They can be legitimately called rtural entrepreneurs.20 Among those with rural origins, more are from coconut areas than from the rice areas. This is consistent with the findings in the previous section about the subcontractor population in Laguna, and it may indicate that the commercial nature of coconut farming fosters entre- preneurship better than subsistence-oriented rice farming. Of those from the rice areas, the majority are from the owner-farmer class, but a few owner/managers are from poor tenant families. Many accumulated their capital for their initial investment in their gar- ment firms from the rice trade (buying and selling). Coconut laborers are usually as poor as, or even poorer than, rice tenants, but some have become subcontractors, investing their savings from the coconut trade. All this suggests that rural commerce activities are impor- tant, not only for agriculture itself, but also in fostering entrepreneurship amenable to industrialization. Of the thirty-nine sample subcontractors for sewing, eleven have been, or still are, em- ployees of exporters in Metro Manila. Including two subcontractors whose families had been engaged in tailoring and three subcontractors who started their garment businesses as RTW firms, sixteen subcontractors had some experience in the garment sector. They may have some advantage either in sewing skill and knowledge or in getting contacts with exporters. It should be noted, however, that neither experience in the garment sector nor sewing skill and knowledge is essential to start as a subcontractor. The rest of the subcontractors had little knowledge or experience of the industry before they started.2" The entry barrier to becoming an export subcontractor is extremely low. Aside from the funds needed for initial investment in sewing machines and a workshop, one needs con- nections with exporters. In this respect, the employees and former employees of exporters have a clear advantage. Six subcontractors started as subcontractors. Their first contracts were with subcontrac- tors operating nearby, and they became subcontractors as soon as they could establish direct contact with exporters. Many subcontractors with no prior knowledge or experi- ence in garment work studied the garment business by visiting existing, reputable sub- contractors before they started their own businesses. Although difficult to quantify, there seems to be a clear correlation between the intensity of this effort and the performance of the subcontractors. The majority of sewing subcontractors obtained funds for initial investments from their own savings (table 5-13). The second important source of funds was the provision of credit, mainly by parent exporters, and generally in the form of machines. Only a limited number of subcontractors obtained bank loans for their investments. It is difficult for many of them to borrow from banks, not only for initial investments, but also for investments after 110 To7vard lie Rutral-Based Developmient of Comml111erce an?d lndutstry Table 5-12. Profile of Owner/Managers of the Sample Garment Firms Num1tber offirms Firmisfor Suibcontractor domestic for export Exporter niarket Total Total 40 4 4 48 Age of owner/manager 20s 2 - - 2 30s 13 2 - 15 40s 18 - 4 22 50s 3 - 3 60s and above 4 - - 4 Unknown - 2 - 2 Sex of owner/manager Male 11 2 - 13 Female 29 2 4 35 Educational level of owner/manager Primary 1 - - 1 Some high school 2 - - 2 High school graduate 6 - 2 8 Vocational 1 - - I Some college 5 - - 5 College graduate 25 2 2 29 Unknown - 2 - 2 Social class of owner/manager (parents' occupation) Farm Rice-owner 5 - - 5 Rice- and sugar-owner 1 - - I Rice-tenant 3 - - 3 Rice- and coconut-owner 5 - - 5 Coconut-owner 10 - - 10 Coconut/copra trade 2 - - 2 Coconut-laborer 2 - - 2 Farm total 28 - - 28 Nonfarm Lawyer - 1 - I Government employee 2 - - 2 Teacher 2 - - 2 Laborer/handicraft 2 - 1 3 Tailor 2 - - 2 Merchant/storeowner 3 - - 3 Garment firm owner - - 2 2 Footwear firm owner 1 - - 1 Construction contractor - 1 - 1 Unknown - 2 1 3 Nonfarm total 12 4 4 20 Export-Oriented Garment Indutstries in the Rutral Philippinies 111 Table 5-13. Sources of Initial Investment for the Sample Sewing Subcontractors Nutm1ber of Souirce sutbcoitractors Parent company + own savings 7 (18) Parent company + bank loan + own savings 1 ( 3) Parent company + loan from relatives + own savings 1 ( 3) Loan from relatives/friends + own savings 6 (15) Bank loan + own savings 1 ( 3) Middlewoman + own savings 1 ( 3) Own savings 19 (49) Donation 2 (5) Unknown 1 (3) Total 39 (100) Number of subcontractors who have ever received bank loans 8 (21) Note: Figures in parentheses are percentages. Source: Author's data. the business has been established, in spite of the program of government-subsidized loans intended to foster rural industries. Because of the relatively low capital requirement for subcontractors, which is one of the conditions that makes the garment industry easy to enter, the unavailability of bank loans does not seem to be a serious entry barrier. 2 Many of the subcontractors have sideline businesses and occupations (table 5-14), which often allow them to secure working capital for running their garment businesses. The owners or spouses of thirty-four of the thirty-nine subcontractors held, or still hold, salaried jobs.23 In addition to working for exporters, popular positions include working as government employees and as teachers. Some have accepted important positions in their towns, such as mayor or vice mayor. Many, including government employees and teachers, express the opinion that they set up their garment workshops not only for private profit but also for the employment benefits to their localities.24 Another source of funds is overseas remit- tances. In spite of the widely held belief that remittances from these workers are wasted- mainly to increase consumption-our study revealed that they are an important source of investment for the development of rural industry. What about the possibility of subcontractors expanding their businesses to the level of contractors/exporters? Table 5-15 lists the contractors/exporters with whom the sample subcontractors are affiliated. Many are joint ventures owned by Americans; Chinese nies- tizo, most probably from the "traditional," urban-based Chinese merchant class, are also represented among owners of export garment houses, as well as exporters who are ethnic Filipinos. Of the twenty-three parent garment companies, 35 percent are owned by Filipi- nos, although most of them are relatively small, with fewer than 500 employees. The family backgrounds of these Filipino owners are unknown, except for two export- ers who had previously been garment subcontractors. One of the firms was started in the 1970s by the mother of the present owner as a tiny RTW garment shop in a town in 112 Towvard the Rutral-Based Developmiienit of C&iniiiierce an7d Indstrwy Table 5-14. Major Sideline Jobs/Sources of Funds: Thirty-nine Sample Sewing Subcontractors NIumnber of subconitractors Farm-related 11 Coconut farming/trade 5 Rice farming/trade 4 Livestock farming (duck, pig) 2 Nonfarm, self-employed business 6 Rice mill 1 Footwear factory I Garment trade 1 Construction 1 Jeepny owner 1 Farm machinery store 1 Nonfarm, salaried job 34 Teacher/ex-teacher 4 Government employee 8 Garment exporter 9 Other large company 4 Overseas employment 6 Mayor/ex-mayor 2 Vice mayor 1 Note: Major sideline jobs/sources of funds refer to those of owners/spouses, allowing double counting. Em- ployees in government, garment exporters, and employees of other companies/institutions, as well as workers for overseas employment, include those who no longer hold these positions. Source: Author's data. Laguna, which began operation with three ordinary sewing machines. The mother later converted the shop into an export subcontractor, and finally, in the late 1980s, the cur- rent owner made the transition to becoming an exporter by securing export quotas and foreign buyers. Another subcontractor who also became an exporter is of urban origin. After working as an employee of a company for more than ten years, he engaged in the food trading business, then used his savings to set up a garment workshop for export subcontracting. Very few can climb the ladder, but at least the upward route is open.' One exporter who had previously been a subcontractor listed the necessary conditions for becoming an exporter: * Funds for investment * Good connections with trading companies and key persons in the garment export business * The ability to speak English-important in establishing those connections * A good sense of management, especially cost accounting. Many sample subcontractors who aspire to climb the ladder are well aware of these conditions, particularly the first and second. Export-Oriented Garmenit Induistries in the Rural Philippines 113 Table 5-15. Garment Exporters Affiliated with Sample Subcontractors * Owner offitrm Size offirmn Americani Indian Chinese Filipino Filipino Total Large 7 0 3 1 11 (64) (0) (27) (9) (100) Small 1 2 2 7 12 (8) (17) (17) (58) (100) Total 8 2 5 8 23 (38) (9) (22) (35) (100) Note: About ten exporters are not induded here, because of nonavailability of information. Figures in paren- theses are percentages. Source: Author's data. Garment Production by Subcontractors The Process and Structure of Garment Produlction by Suibcontractors The operations of export subcontractors for sewing are rather simple: they receive sample clothes and cut materials, or fabric, from an exporter; sew the clothes following the sample pattern by dividing the sewing process into simple operations; and deliver the finished garments to the exporter within a specified period. A subcontractor usually provides a workshop and sewing machines and hires sewers. Rough estimates of revenue, costs, and returns of garment operations are listed in table 5-16. Following this table, the production process of garment subcontractors in Laguna can be observed. A contract between a subcontractor and an exporter consists of three essential clauses: the contract price for each piece of the product, the number of pieces contracted, and the date for delivery of the finished clothes to the exporter. The annual revenue of a subcon- tractor depends on the first two clauses, as well as the number of orders from the exporter(s). The contract price of a garment contracted by subcontractors varies widely-from as little as P3 a piece to P80 a piece-depending on the quality and sophistication of the job. The number of pieces in each order ranges from several hundred to several hundred-thousand. The contract period also varies, from two weeks to a few months. Contract prices are high for urgent jobs. It is critical for subcontractors to secure a continuous flow of contract jobs from export- ers throughout the year. They make every effort to this end, but most subcontractors, par- ticularly freelancers, are subject to fluctuations and lulls between job orders. One source of instability, as in agricultural production, is the seasonality in the garment industry. The demand for garments is stronger during July to December (before Christmas), and it de- clines markedly for the February-May period.26 The fixed capital assets needed for sub- contractors are machines and a workshop. There are various ways for subcontractors to acquire sewing machines, the most important means of production for them, and other 114 Towvard the Rutral-Based Development of Commerce and Induistry machinery. Some buy new machines from machine shops, but a more popular method is to purchase machines in the secondhand markets, often with arrangements to pay on an installment plan. The machines of subcontractors who close their shops and machines used as collateral by subcontractors who fail to make repayments are also occasionally available. There is a machine rental market, but it is narrow. Machines are sometimes provided, usually rent-free, by an exporter if a subcontractor wants them. Many subcon- tractors prefer to own the machines themselves, because they do not want to be restricted by a given exporter. All the sample subcontractors have at least one workshop; two of them have two. The size and quality of these workshops vary enormously. Some use a part of their houses as their workshop, with or without renovation, and others convert warehouses for farm equip- ment to this purpose. Some have built or leased an independent house. With few excep- tions, their workshops are humble and small, as judged from the estimated rental value listed in table 5-16. It is commonly observed that too many sewers are working with ma- chines in too confined a space. For machines, capital services are estimated by applying an annual interest rate of 20 percent to the stock values. This interest rate is selected for conversion because the sample Table 5-16. Estimates of Revenue, Costs, and Returns of Average Garment Subcontractors for Exports, for OneYear of Operation, 1994 Prices Item Pesos (1,000) Percent Revenue 630 100 Capital services Machines 66 10 Workshop (central value) 30 5 Total 96 15 Labor Sewers 325 52 Revisers/trimmers 26 4 Supervisors 23 4 Total 374 59 Current inputs Needles 2 0 Thread 23 4 Electricity 21 3 Machine maintenance 6 1 Total 52 8 Operator's surplus 108 17 Note: For the average subcontractor with 16.8 workers and 20.8 machines (table 5-11). The following assump- tions are taken from the sample averages: contract price per contracted garment = P17.24; sewers' share = 51.6 percent; labor productivity = 208 pieces/sewer/month; net working months = 11 months; rejection rate = 5 percent. Capital service for a workshop is obtained from owner's self-evaluation; for machines, it is taken from the asset values by applying 20 percent/year interest rate. Source: Author's data. Export-Oriented Garment Indutstries in the Ruiral Philippinies 115 subcontractors unanimously point out that they would never borrow money from banks if the interest rate were greater than 20 percent. The actual interest rates for those who have received bank loans range from 12 percent (institutional loans) to 19 percent (com- mercial loans). The rental value of workshops is estimated by the sample subcontractors in capital service. The share of capital services thus estimated absorbs 15 percent of total revenue. The owner/managers of subcontracting firms determine the piece rate their sewers will receive for each job order. The share of the sewer's wage in the total price varies greatly, not only across subcontractors but also for the different orders received by a single sub- contractor, from 33 to 75 percent, depending on the difficulty of the job. A 50:50 sharing, however, is most popular, and the average sewer's share is 51.6 percent. Revisers/trim- mers and supervisors, whose main function is in-house quality control, are usually paid a fixed daily rate. In some cases, they work as sewers when free from their own duties and receive piece-rate wages in addition to their fixed wages. In about half of the sample sub- contractors, the owner works as reviser and supervisor. In these cases, not all of the labor earnings are paid out. For labor as a whole, the factor share in total revenue is estimated to be nearly 60 percent. In addition to labor and capital, the subcontractors require current input, including needles, thread, electricity, and machine maintenance. Thread is provided by exporters for some high-quality job orders, but usually is not provided for ordinary work. Because sewing machines run on electricity, this item is the second-largest current input, fol- lowed by machine maintenance. The needles for the sewing machines are provided by the subcontractors, but this expense is negligible. The combined cost of these current inputs is estimated to be 8 percent of the total revenue of the average subcontractor. In other words, the gross value-added ratio of the garment subcontracting business can be as high as 92 percent. As a result, an average P108,000 of surplus, or 17 percent of total revenue, is left to the operators. This means a return of P9,000 monthly for their management, excluding the returns to fixed capital. If compared with the monthly incomes of P2,000 to P4,000 of government employees and teachers, the customary jobs available to people with higher educations in rural areas, this level of return is quite high. It should be noted that this is an average figure; some subcontractors earn much more than this, and others much less.27We can also estimate a rough order of the rate of return to capital investment. The average annual rate of return, as measured by the ratio of the operator's surplus to the value of fixed assets, is estimated to be 23 percent. If payments for wages and current inputs for one month are added to the capital requirement as working capital, the rate of return becomes 21 percent, roughly comparable to the interest rate used to convert asset values into flows. This rate of return should be much lower than the prevailing level of interest rates in the informal money markets in rural areas, but it is high compared with the rate on deposits in commercial banks in rural areas, currently about 5 percent annu- ally. The garment subcontracting business is lucrative enough to attract investment by rural entrepreneurs. 116 TowTard the Riural-Based Development of Commerce and indutstny Sub-Subcontracting Arrangements It is not uncommon for subcontractors to put out part of a contracted job to sub-subcon- tractors, and a number of sub-subcontracting arrangements are used. Some subcontrac- tors have relationships with regular sub-subcontractors that are just like those between exporters and subcontractors. In these, subcontractors add the sewing capacity of their sub-subcontractors into their total capacity. They often provide sub-subcontracors with machines. The jobs given to the sub-subcontractors are frequently as complete as those the subcontractors receive from exporters, with contracted prices that are usually the same as the rates the subcontractors pay to their in-house sewers. In an extreme form of this ar- rangement, a subcontractor will become an intermediary between an exporter and sub- subcontractors; the subcontractor will have no in-house production, but depend entirely on the sub-subcontractors for garment production.28 The second kind of sub-subcontract- ing arrangement is to put out specific operations, such as trimming, to sub-subcontractors at a piece rate. Sub-subcontractors for this kind of work are usually housewives, who do the jobs at home. The third kind of contract is formed among subcontractors at the same level. When subcontractors have difficulty finishing jobs within the promised period, they ask their fellow subcontractors for help. Sometimes the subcontractors with a plethora of jobs give part of them to those who have difficulty finding orders. In this case, the con- tracted price between the first and second subcontractors is lower than that between the exporter and the first subcontractor. All three sub-subcontracting practices are fairly com- mon, although some exporters prohibit sub-subcontracting. Subcontractors, particularly freelancers, interact with each other rather intensively. When sub-subcontracting is practiced, it is common for sewing machines to be rented among the firms. Subcontractors exchange information about exporters, sewers, and the like through the Laguna Garment Subcontractor Association (founded in 1992, but no longer in exist- ence), on exporters' premises, and at other occasions, such as seminars held by exporters for garment subcontractors. Within such an environment, it is intriguing that some sub- contractors have large and long-outstanding payments due from other subcontractors, with whom they have had sub-subcontracting arrangements. In addition to the problem of debt insolvency, there are cases of swindlers-usually from outside the locality-who disappear after collecting the finished products.29 Sewers and Labor Management Sewers working in garment subcontracting firms are predominantly female, but about 10 percent of the total sewer population is male.A' Most sewers are from the bottom strata of rural villages-agricultural laborers, small tenants, and fishermen's families-and gener- ally have lower-level educations, such as primary school only, or perhaps some high school. Many sewers come from remote areas, and nearly 90 percent of subcontractors have dor- mitories for live-in sewers; the accommodations are very humble, with free room and board. Export-Oriented Garment Indiustries in the Rural Philippines 117 Sewers typically work from 7 o'clock in the morning to 5 o'clock in the evening; over- time work extends the day up to 10 o'clock at night, and they work at least six days each week. It is not unusual for them to work on Sundays to complete rush jobs. The average sewer receives a weekly wage of about P400 (table 5-10), which yields a daily wage of about P65, more than 50 percent lower than the minimum wage rate of this region.3' This daily wage is comparable to, or a bit lower than, the daily wage rate for rice transplanting in Laguna, which is P70 to P80, although transplanting takes only four to five hours each day. Nevertheless, few sewers take leave during the transplanting season, but many do so during the rice-harvesting season, because harvesting work gives them an in-kind daily wage (paddy) worth more than P100. This seems to suggest the value sewers place on the job security found in the garment sector in relation to the higher-paying, but less certain, job opportunities in farming, as well as the clean work of sewers compared with the drudg- ery of working in muddy paddy fields. The labor supply is abundant, and thus it is not difficult for subcontractors to hire as many local sewers as they need.32 Nevertheless, it is difficult to find and keep good, de- pendable sewers, and the turnover rate among sewers is very high for many subcontrac- tors. After acquiring the skill of operating high-speed sewing machines,33 young sewers seek jobs with big exporters in Metro Manila, which often results in turnover rates of more than 50 percent within a six-month period.34 Absenteeism and a "get-a-cash-advance-and- run-away" attitude among sewers are the major management problems for subcontrac- tors. Their concerns about labor management may be seen in the "Rules and Regulations" posters often found on the walls in garment workshops (an example is shown in box 5-1 in English translation). The nature of the labor market for the garment industry can be understood from an examination of newspapers. For example, the August 20, 1994, issue of the People's Joturnal, a popular English daily tabloid, published 140 recruitment advertisements; thirty-five (25 percent) were for sewers and revisers in the garment sector. The following day, in the same newspaper, the ratio was 35/129 (27 percent). About 50 percent of the advertisements were large, and had been placed by big garment exporters; the other 50 percent, placed by small garment subcontractors, were only a few lines long. The two subsectors compete with each other to recruit sewers. From a different perspective, the August 19, 1994, issue of the Manila Builletin, an English daily read by the upper classes, carried only nine gar- ment-related advertisements of the total of 220 recruitment advertisements (4 percent). This seems to show that highly educated manpower is not sought by the export garment sector.35 One way for subcontractors to get good sewers is to recruit them through relatives or friends from such remote areas as Quezon Province and Visayas through relatives or friends. Because of substantial wage differentials between Laguna and these remote areas, it is easy to recruit sewers, as long as good connections are available. About one-fifth of the sample subcontractors use this recruitment method. Another source of reliable sewers is the population of former sewers of exporters, usually older persons, who returned to their home villages after marriage. Some of them are employed as supervisors and trimmers. 118 Toward thte Rutral-Based Deevelopmenit of Comnmerce and Inditstry Box 5-1. Sample"Rules and Regulations"Posted at a Garment Subcontractor with Twenty Sewers Company's Rules and Regulations Time: Start 7:00 - 7:15 7:16 Totally late 10:00 - 10:30 Break time 10:31 Late 12:00 - 1:00 Noon break 1:01 Late 3:00 - 3:30 Break time 3:31 Late 5:00 - 7:00 Dinner 7:01 Late 7:00 - 10:00 Overtime 1. A penalty of P5.00 will be charged each time an employee comes late for prescribed work- ing hours. 2. Snacks will only be permitted during the break time. 3. Unnecessary standing, doing nothing, and chatting are strictly prohibited during work- ing hours. 4. Smoking is not allowed during working hours, particularly inside the room. 5. Place cloth and thread at their designated areas. Do not scatter them. 6. Turn off the sewing machines, electric fans, and lights before leaving. 7. All employees are enjoined to accomplish their respective daily reports. No report, no pay. 8. Any worker found guilty of cooking foods during the working hours will not be granted an allowance privilege. As a matter of compensation, those employees who strictly observe the above rules and regulations, without committing any violation in a given period, will be entitled to an incen- tive allowance of P50.00 per week. Sewers working for subcontractors are, without exception, paid piece-rate wages. There are, however, at least two different systems of piece-rate wages; one is by operation and the other is by completed garment. In the former, the sewing process is divided into indi- vidual operations, and sewers are paid the piece-rate wage that has been set for their task, according to the number of pieces they finish. An example of such piece-rate wages by operation is shown in table 5-17. In the latter, a piece rate is set for each completed gar- ment, and sewers are paid according to the number of garments they make from start to finish. The system of piece-rate wage by operation is commonly adopted by subcontrac- tors for export. The system of piece-rate wage by completed garment is often applied in the RTW sector. Among the sample subcontractors, the latter system is used by only one operator, the owner of a premanufacture operation with a very small workshop in the most remote town in Laguna, but it is used by many home-sewers.36 Export-Orien?ted Garm11enit Inidutstries in the Rural Phtilippines 119 Table 5-17 Sample Sewing Operations and Piece-Rate Wages, Baby Dress Item Rate Item Rate Shoulder 0.10 Crotch-back 0.15 Arm hole 0.35 L-A-pocket 0.15 Sleeve garter 0.40 Side 0.30 A-collar 0.40 Crotch 0.30 J-collar 0.30 Lace-leg 0.20 Sleeve lace 0.20 Bib 0.20 Waist-front 0.20 ShirTing 0.20 Waist-back 0.20 Back placket 0.20 Pocket 0.40 Edging-over all 0.35 Crotch-front 0.15 Edging-lace 0.15 Total P4.90 Note: Contacted price P10.00/piece; payments to sewers = P4.90 (49 percent of the contracted price). Sotirce: Author's data. Exporter-Subcontractor Relations and Quality Control How can freelance factory operators become regular subcontractors to exporters? They first show the exporter their GTEB accreditation and the mayor's permit for operating a factory in their town. Then, a test is made to identify whether the subcontractor's sewing skill meets the company's standard. In addition to these checks and tests, some exporters ask their subcontractors to pay surety bonds every year. As long as a subcontractor's (or his/her sewers') skill is above a certain standard, it is not difficult to become a subcontrac- tor for exporters. The demand by exporters for good-quality subcontractors seems to be quite strong. Many exporters search for good ones, for instance, by sending letters of invi- tation to the subcontractors on the GTEB subcontractor list and by visiting subcontractors in weak financial condition to persuade them to convert to a parent-child subcontractor relationship. Subcontractors usually pay weekly wages to their sewers, while exporters pay subcon- tractors when the contracted garments are delivered, a few weeks to more than a few months after the contract work begins. The gap between the payments to sewers and the payments from exporters creates a serious problem for subcontractors who lack working capital. This problem is aggravated by a mode of payment that many exporters adopt: 50 percent is customarily paid upon delivery (after spot checks of the delivered garments), 40 percent a few weeks later (after finishing the final processes of garment production), and 10 percent is retained by the exporter. This 10 percent, which is paid to the subcontractor at the end of the year, is said to be insurance against nondelivery for the exporter. Exporters' payments are made by postdated checks, so that even the first payment car- ries a built-in lag: subcontractors have to wait a few weeks to cash the checks. Subcontrac- tors with financial difficulties sometimes have to take the checks to a rediscounter-who cashes them, taking a 5 to 20 percent fee, before their due dates. For subcontractors, par- 120 Tohvard thie Rural-Based Developmenit of Conmmerce and lindlstry ticularly freelancers, good exporters are those who pay good prices and do not delay their payments. It is quite important for a subcontractor to have a continuous flow of job orders from exporter(s). A long period without contracts not only makes firm operations difficult, but it also results in the dispersion of the sewers who are mobile, even when work is steady. One means exporters often use to exercise their control over subcontractors is to give spe- cial favors or priority to certain subcontractors in allocating orders. Exporters also make quick, on-time payments to favored subcontractors with whom they want to maintain contracts. The seasonality in garment demand affects the employment pattern of exporters, as well as their relations with subcontractors. For instance, an exporter that does not depend heavily on subcontractors hires about 25 percent of their total labor force as casual work- ers; these individuals work during the busy season but are laid off during the slack season. For exporters who rely on subcontractors, the ability to grant priority in allocating orders during the slack season can be a strong means of control over subcontractors. Subcontracting arrangements thus offer advantages to exporters, not only in exploiting cheap labor in rural areas, but also in dealing with the seasonality of the garment markets. At the same time, controlling the quality of garments made by subcontractors and assur- ing timely delivery remain the two major difficulties exporters face. It is of great impor- tance that exporters who sell their garments to foreign buyers in bulk-tens of thousands of dozens of the same garments; for example, baby clothes or jeans-maintain a standard- ized quality. To deal with these problems, in addition to the control and incentive systems described so far, exporters utilize a variety of quality-control measures. First, it should be mentioned that, in subcontracting between exporters and subcontrac- tors, there is less room for cheating (such as stealing materials) than there was in the textile industries in England and Japan during the time the putting-out system was practiced widely (see Chapter 3 by Itoh and Tanimoto). Because the materials for producing gar- ments-imported cloth and yam-are given to subcontractors in the form of cut pieces, which can be counted exactly,37 and because the work subcontractors perform on con- tracted garments usually ends just before the final products are completed, there is little room for the subcontractors to cheat. Unlike sewing, the cutting of cloth, while requiring more skill, could present more opportunities to cheat, and this is why the process is per- formed by the exporters themselves or by subcontractors that specialize in cutting. The most direct measure of quality control for exporters is the check or quality inspec- tion (called a quality audit) conducted at the exporter's factory when finished garments are delivered by subcontractors. In the case of rush jobs, exporters make this quality in- spection at the subcontractor's work site when they collect the finished garments. If sub- standard garment products are found, they are rejected, and revising is needed. Rejections also occur during the final production processes in the exporter's factory. If there are many rejections, the rejected garments are generally returned to the subcontractors for revision; if there are only a few, the revisions are made by the exporters, and the rejected pieces are deducted from the payments. Export-Oriented Garment Indutstries in the Rural Philippines 121 Some exporters send their quality controllers (QC) to make final inspections at the sub- contractors' factories. The QCs make sample checks of 4 to 5 percent of the finished gar- ments; if major defects are found in a given percentage-10 percent of the sample, for example-the entire batch is rejected for revision. If defects are found in a smaller percent- age, the entire batch is accepted and packed immediately for export. This method is ap- parently used for subcontractors whose skill level is relatively high. Another measure of quality control, which is more process oriented, is direct visits by exporters' employees to subcontractors' factories. Although the frequency and intensity of visits differ among exporters, or for an exporter among its subcontractors, this measure is almost universal. Foreign merchandisers sometimes visit subcontractors, without notice to the exporter involved.3 There are two forms of visits by exporters: one is mainly to inspect the progress of subcontractors' work and is usually carried out by an industrial engineer from the production coordination department, and the other is purely for quality control and is the responsibility of employees dispatched from the quality control depart- ment. Both visits involve quality control, but the latter is more of a final check. Subcontrac- tors do not usually distinguish between these two kinds of visits and call them both QCs. For parent-child subcontractors, QCs stay at the subcontractor and usually report to their company only once a week. In such an extreme case, these subcontractors are a de facto branch of the exporter. For freelance subcontractors, the frequency of QC visits var- ies, from daily to weekly for each job contract; the frequency probably depends on the skill level of the subcontractor and the distance to the exporter's headquarters. In any case, a large number of QCs from garment exporters travel around rural areas every day to visit the subcontractors assigned to them. Yet another measure to control quality and to encourage subcontractors' work is to set incentives and penalties. Many exporters adopt incentive payment systems that provide additional payments, ranging from 5 to 25 percent, to subcontractors who deliver their con- tracted garments before the specified due date. The system is often constructed so that the earlier the delivery, the higher the additional payment. The system is frequently balanced by a penalty system when subcontractors fail to meet the due date. Another penalty system, keyed to the quality inspection upon delivery, is adopted by many exporters: if the rate of rejection of a batch of delivered garments exceeds a selected percentage, money is deducted from the payment at a rate that increases as the percentage of rejected garments rises. The quality inspection of delivered garments at the exporter's factory is a popular source of complaints from subcontractors. When exporters pick up the finished garments, make the inspection at their factories, and reject garments during the final processes there, the checking is done in the absence of the subcontractors, and the reduction of payment is calculated by the exporters without the consent of the subcontractors. The criteria for re- jections are thus not clear to the subcontractors. In some cases, it appears that exporters make sporadic rejections just to demonstrate their strict quality controls. Some quite ca- pable subcontractors, when told that a quantity of their garments had been rejected, were able to prove that the rejected items were not theirs; these subcontractors had put some identification on their garments that was not easily visible. 122 Towvard the Rutral-Based Developnmenit of commerce and iltdtustry The Role of Government The government has played an important role in the development of the garment industry in rural areas. The new drive for garment exports, begun during the 1970s through tax privileges for export industries and export processing zones (EPZs), as well as other means, has strengthened since the late 1980s as the political turmoil associated with the downfall of the Marcos regime has receded. The vigorous efforts by the government in this area are readily apparent if one visits the Garment and Textile Export Board (GTEB), a government organization established under the auspices of the Department of Trade and Industry (DTI) to promote garment and tex- tile exports. A visit to the office shows it to be bubbling with enthusiasm for promoting the garment export business: swarms of applicants for accreditation certificates for garment subcontracting are rushing to the counter; many exporters are busily asking the officers in charge for the list of subcontractors, applying for export quotas, or checking the status of their imported materials; and foreigners are inquiring about possible partners for joint ventures. The government has clearly played important roles by coordinating and provid- ing information. The same enthusiasm prevails in the DTI offices at the regional level.39 These offices try to promote the garment industry in rural areas through a variety of programs, including the provision of consultation services and seminars for potential subcontractors. Tax ex- emptions for subcontractors for a set period after their establishment also encourage po- tential entrepreneurs to enter the industry. Many municipal governments in rural areas hold training courses for rural youths and housewives to help them develop sewing skills using high-speed and other special machines. These programs, however, seem to be the total contribution of the government to the garment sector in rural areas. Many policies have been implemented repeatedly to pro- mote rural industries, such as subsidized credit and input subsidies. But the impact of these policies is barely visible in the rural garment sector. Very few sample subcontractors obtained institutional loans. This means that the transaction costs to obtain institutional credit are too high for small rural entrepreneurs. A major reason that exporters put out their garment production to subcontractors is the wage differential between the urban labor market, governed by minimum wages and other labor regulations, and the rural labor market, in which such regulations are absent. If the government had tried to implement these regulations in rural areas, the development of the rural garment industry would have been much less significant. All this suggests that the government should concentrate its efforts on supporting the markets, rather than trying to intervene in the market mechanism. The provision of public infrastructure, a responsibility that has traditionally rested with the goverrnent, is impor- tant to such efforts. For example, the development of efficient road and highway networks connecting urban centers with outlying regions would help rural industries develop in remote rural areas by reducing travel time between the rural areas and the centers, which would reduce the transaction costs involved in subcontracting arrangements. The devel- Export-Orientted Garment Iiididstries in the Rural Philippin2es 123 opment of industrial centers in regions other than Metro Manila through the provision of infrastructure (such as ports and EPZs) is also important if rural industrialization based on subcontracting is to succeed. What rural entrepreneurs in the garment sector need most is information on subcon- tractors, exporters, foreign buyers, available technologies, and so forth; the provision of adequate and efficient information is essential. The government can also play critical roles in research, training, and technology extension, including research on trends in foreign demand and labor and industrial management technology, training for skill development among sewers/special machine operators, and seminars for owners/supervisors of gar- ment subcontracting firms to enhance knowledge/technology on labor management/ methods of quality control, and the like. In these respects, the GTEB's performance to date may be adequate in its direction, but there are many areas that can be improved and strengthened. Conclusions The rapidly growing export garment sector in the Philippines has given rise to garment firns in rural areas adjacent to Metro Manila. These firms engage in the production of garments under subcontracting arrangements with exporters/contractors who contract garment production with foreign buyers, who in turn act as the principals of these con- tracts. The major reason for exporters to put out garment production to subcontractors is to utilize cheap labor in rural areas-wage levels are 50 percent lower than in the urban sector-by avoiding the minimum wage and other labor regulations enforced in the met- ropolitan area. Subcontracting in the export garment sector is not limited to sewing. Subcontracting firms for other services, such as embroidery, knitting, beading, and the like, have also burgeoned. Because the technology used in the garment industry is labor intensive, the spread of garment subcontracting in rural areas has created employment opportunities with low opportunity costs. More fundamentally, the rise of the export garment sector gives an opportunity for "new" rural entrepreneurs to create a new rural industry. Be- cause of its colonial heritage, there have been few rural industries in the Philippines, ex- cept some agro-based industries run by "traditional" entrepreneurs. The impact of such recent changes can be observed clearly in Laguna, where the gar- ment industry has been insignificant, not only in export but also in ready-to-wear gar- ments for the domestic market. Since the early 1980s, a number of garment subcontracting firms have emerged in the rural areas of the province. Seventy percent of these new entre- preneurs are from rural farming communities. Unlike traditional entrepreneurs, who are either from the landed oligarchy or from the Chinese merchant class, many of the new entrepreneurs are from the peasant class. Both the rice- and coconut-growing areas in the province produce rural entrepreneurs, but it appears that the coconut area yields a greater number. Because coconut is a commercial crop, peasants engaged in this area would be 124 Tozvard the Ruiral-Based Development of Co0mmerce and Indnstry more readily trained as rural entrepreneurs. Furthermore, many subcontractors of rural origin have engaged in trading (buying and selling) coconut or rice, which supports the contention that rural commerce is a cradle for fostering rural entrepreneurs (Hayami and Kawagoe 1993).4°Many subcontractors started their garment businesses as sub-subcon- tractors and then became subcontractors. Opportunities are also open for subcontractors to climb to the exporter/contractor level. If they can obtain export quotas and find foreign investors, it is possible for subcontractors to become exporters/contractors. Thus, the lad- der from peasant to rural entrepreneur-and further, to modem industrialist-does exist, however narrow it may be. A necessary condition for being a successful subcontractor or to succeed to exporter is the ability to develop good personal relationships with people, including sewers; export- ers' employees, such as managers, merchandisers, finance clerks, and quality controllers; foreign buyers; and government officials. This is the skill fostered in commercial agricul- ture and trade activities in rural areas. The college-level educations achieved by most sub- contractors may also be useful in fostering such skills. If so, however, what is useful in entrepreneurship would not be specific knowledge or skill, but the wider perspective and higher aspirations obtained in college life. A critical factor underlying the rapid diffusion of the garment business in rural areas is the ease in entering the industry through subcontracting. Special skill or experience in the industry is not a prerequisite. The technology used in the industry is neither very sophis- ticated nor capital-intensive. The initial investment requirement is not large. Many sub- contractors began their businesses with a few ordinary sewing machines, easily affordable for rural people from their savings or by borrowing from relatives or friends. Further- more, the active market for secondhand machines with the common arrangement of pur- chase on credit is instrumental in keeping the initial investment needs low.41 As in other industries that adopt the putting-out system, quality control for garments made by sub- contractors is important to exporters /contractors. A number of measures are adopted for quality control, such as sending quality controllers to subcontractors' factories, quality inspection upon delivery of finished garments, checking and rejection during the final process of garment production at exporters' factories, and retaining a part of the payment to subcontractors. There are also many measures to avoid opportunistic behavior on the part of subcontractors, which can include stealing supplied materials or not meeting the delivery date. These measures include preferential allocation of job orders, prompt pay- ments, retention of a part of payments, incentives and penalties for beating or missing delivery dates, and so forth. Because cloth is given only after it has been cut into pieces and parts and counted exactly, there is not much room for cheating by subcontractors. Quality control within subcontractor workshops is carried out by the owners themselves and by hired supervisors/revisers. Their garment production process is divided into as many pieces as possible. The piece-rate wage payments for the countable work divisions seem to function well as a built-in measure of work monitoring and enforcement. The government has played important roles in the development of the garment indus- try in rural areas, but its contribution is mainly the promotion of garment exports through Export-Oriented Ganrentt Industries in tihe Rural Philippines 125 tax privileges for exporters and rendering such services as coordination and information provision for exporters and subcontractors. In the rural garmnent industry, it is hard to find a positive impact of the policies that intervene directly into markets. The case study in this chapter suggests that the government should concentrate its efforts in supporting the markets, rather than trying to intervene in the market mechanism. Further improvements in public infrastructure, information provision, and research and technology extension in the garment industry are the fields in which the government should-and can-play criti- cal roles. Notes 1. As the rural household industry declined, the percentage of the labor force engaged in agri- culture increased from 51 percent in 1902 to 61 percent in 1918, and further, to 71 percent, in 1938 (Resnick 1970: 64). 2. The lack of knowledge and ability among tenant-farmers and agricultural laborers-even as agriculturists-was the first pretext given by the landlords who were against the programs when land reform programs were envisioned and implemented (Hayami and others 1990a). 3. Among the sectors earning foreign exchange in 1993, the garment sector was the second- largest earner (P62 billion), exceeded only by remittances from overseas Filipino workers (P71 bil- lion), and followed by the tourism sector (P57 billion) and electronics exports (P50 billion). 4. Another garment subsector not shown in this figure, although related only to marketing and not to production, is secondhand garment transactions. This subsector could still be of some impor- tance in poorer sections of society, urban as well as rural, but is not discussed in this chapter because of the lack of information. 5. Philippine garments are exported to two groups of countries, quota and nonquota countries. Quota countries are the United States, EU countries, Canada, and Norway (total export value of US $1,869 million in 1993); all other countries are nonquota countries (total export value of US$473 million in 1993), of which Japan; Hong Kong, China; and the United Arab Emirates are major im- porters. This means that for a garment firm to be an exporter, it is a near necessity to have export quotas (EQs), which are allocated by the GTEB according to its rules and regulations. 6. RTW firms for high-end markets usually get contracts through the merchandisers of super- markets and department stores; garment contractors for export get them from foreign merchandis- ers. In domestic markets, the kind of garments/apparel to be made depends heavily on the garment firms: unless they can produce apparel fashionable enough to attract customers, they cannot sur- vive. Therefore, they make the utmost effort to collect information and ideas on what is popular. It should be noted, however, that RTW firms for low-end markets also make an effort to update their fashion sense, although to a much lesser extent. In contrast, contractors and subcontractors for export garments, particularly the latter, do not need to gather such information. 7. One may wonder if imported materials with tax-free privileges, to be used for garment firms for exports such as textile/cloth and thread/yam, might be diverted to those for the domestic mar- ket. Such moral hazards are prevented, at least between contractors and subcontractors, as explained in this chapter. 126 ToTward the Rural-Based Development of Comminierce antd Industry 8. There are certainly other exceptions to this statement. In addition to the towns in the urban- ized northwest corner of the province, one finds some traditional rural industries in towns in La- guna: jeepny building in San Pablo City, footwear factories in Liliw, and woodcarving in Paete. These industries, however, are by no means widespread, and it can be said that until quite recently the "agrarian separation and agrarian division of labor" described by Resnick (1970) were appli- cable to Laguna. 9. In spite of past attempts to implement land reform programs, few rice owner-operators have been created in Laguna (see, for example, Otsuka 1992). 10. Distance is a cost to subcontractors too. Information on the subcontracting market becomes more difficult to obtain as the distance from Metro Manila increases. For instance, the subcontractor in the most distant town in Laguna, when she set up her garment firm, obtained information about the industry from an RTW firm operating in a nearby town in Rizal Province, on the other side of the lake, not from someone in Laguna. 11. Among the Laguna subcontractors, eight maintain liaison offices in Metro Manila in addition to factory/workshops in Laguna. All of them are also from Laguna I. 12. In the case of baby clothes with smocking on the front breast and embroidery on the pockets, for example, imported cloth is first put out to cutting subcontractors who cut pieces, including the front breast and pocket parts; the former parts are put out to smocking subcontractors and the latter to embroidery subcontractors. After getting these smocked and pocket parts back, all parts are brought to sewing subcontractors for assembly. 13. This sample subcontractor in Laguna I was interviewed when he was visiting one of his fellow subcontractors, whose workshop is located in a Laguna II town. 14. It is said that there are now as many as 300 garment factories in Tatay. All of them, except a few that have recently converted to subcontractors for export, are RTW firms. 15. According to subcontractors, these parts are difficult and need special skills and machines. For example, the machine to make a buttonhole for jeans, called an "eyelet," is as expensive as a brand new Mercedes Benz. It seems, however, that exporters make these parts in the factories partly for the purpose of quality control. 16. Of the forty subcontractors in the sample, sixteen have been working for only one contractor. Not all of these sixteen are of the parent-child variety; some are looking for better contractors. 17. This is because transaction costs for contractors become too high to have such small subcon- tractors with tiny lot sizes for contracted garments. 18. Few ordinary manual sewing machines without motors are used in the garment industry, except by some tailors and unorganized home-sewer subcontractors for RTW. 19. It should be noted that the average sewer's wage in the sample RTW firms is higher than that of exporters. This difference is probably because the RTW firms, particularly those aimed at high- end markets, usually employ highly skilled sewers. 20. Among subcontractors, there is no significant difference between those of rural origins and those of nonfarm origins in the owner's age, years of operation, and firm size. 21. As stated earlier, three-fourths of the subcontractors have received higher education. At col- lege, most enrolled in the school of commerce, majoring in banking and finance, accounting, or business administration. A few studied education, and one chemical engineering. It is interesting to note that the majority of newly emerging rural entrepreneurs in the cotton textile industry during the late Meiji period (around 1900) did not have any experience in the industry when they started textile factories (Amano and Abe 1989). Export-Oriented Garment Inditstries in the Rueral Philippines 127 22. This, of course, does not mean that there is no need to provide bank loans for entrepreneurs in the garment industry. Many sample subcontractors expressed a need for bank loans, particularly for working capital, for which commercial banks do not lend without collateral. In one reported case, a subcontractor had to borrow working capital from a fime-six, a usurious moneylender who lends 5 pesos on Monday and recovers 6 pesos on the following Monday, a weekly interest rate of 20 percent, or a monthly rate of about 100 percent. Although not listed in the table, many subcon- tractors obtained loans in the form of machines purchased on credit. Markets are well developed, not only for brand new machines but also for secondhand machines, and purchase on credit is widely practiced. 23. Some of them resigned from office when they started their garment businesses, but others continue to work while managing their garment firms. 24. An extreme case of this public consideration for employment creation among concerned people in rural areas is the establishment of garment factories by agricultural cooperatives or religious missionaries. There are at least three garment subcontractors like this in Laguna, two of which are in our sample: one was founded by an agricultural cooperative and the other is an association of sewers formed by a Roman Catholic church. 25. The ability to be a good entrepreneur may be clearly revealed when management environ- ments become harsh. Many subcontractors with upward aspirations gained large benefits from their business when electricity was not provided steadily in the early 1990s, an event that hit the garment industry hard. The subcontractors acquired their own generators, at a price roughly ten times that of the most expensive high-speed machine. 26. This seasonality problem is similar, or even greater, in the RTW sector for domestic markets. In order to deal with the slack season, for example, one of the sample RTW firms aiming at high-end markets operates a side business of renting student caps and gowns for graduation ceremonies held during this lean season. 27. As mentioned earlier, garment subcontracting is easy to enter because it doesn't require much skill or capital. But this does not mean that subcontractors perform the same way. Performance varies enormously, depending on the owner or manager's ability in management, particularly la- bor and financial management. 28. In the sample in this study, there is a sub-subcontractor who gets 100 percent of his job orders through this kind of subcontractor. This sub-subcontractor provides a workshop and hires sewers, while the subcontractor provides all the machines. The subcontractor takes 35 percent of the con- tracted price as a return for her services and machines. 29. They approach less-experienced subcontractors who have difficulty securing sufficient job contracts, give them a job, and, after receiving the finished garments, disappear without paying for the job. 30. Although younger generations dominate, female sewers are from all age groups. It is rare to see older male sewers. 31. You can buy 4 to 5 kilograms of white rice daily out of this wage. 32. The simplest method of recruiting sewers is to hang a "Sewers Wanted" sign at the gate of the garment workshop. 33. Having sewing skills helps a sewer find an employer, but it is not necessary. Many subcon- tractors say that one to two weeks of training is sufficient even for novices, and some prefer recruit- ing and training fresh workers over hiring "secondhand" sewers. 34. In contrast, the turnover rate of workers is very low in exporters' factories (see Ohno 1994). 128 Tozward the Rtral-Based Dezvelopmenit of Conmnerce and Industry 35. The advertisements that took the largest space in these newspapers, but were not the largest in number, were for overseas employment. Another noticeable feature is that there are few recruit- ment advertisements from the electronics industry, which has developed as rapidly as the garment sector in the value of exports. This may partly be the result of the low labor absorption power of this industry, and partly a result of the different kind and class of worker required by the industry. Whereas the garment sector recruits those who have some skill, almost regardless of age, the elec- tronics industry wants trainable, not trained, persons who have graduated from high schools, voca- tional schools, and colleges. 36. Although this study does not explore how large exporters manage their labor process, of the four sample exporters, the two that used to be subcontractors still adopt piece-rate wages for their in-house sewers; also, the manager of another exporter complained repeatedly that because of the tight government labor regulations, he could not adopt the system of piece-rate wages, so it is difficult to motivate sewers to work hard and to monitor workers' efforts. On labor management and quality control in a large garment exporter in Manila, see Ohno (1994). 37. Checking the number of pieces for all the parts upon delivery by an exporter is an important and laborious task for a subcontractor. If the number of some parts is less than specified, it must be reported to the exporter within a specified period, usually a few days. Otherwise, the exporter deducts the cost for the deficit in parts from the payment to be made to the subcontractor, assuming that the subcontractor lost them. 38. The necessity for this monitoring of exporters/contractors by foreign merchandisers or buy- ers may indicate that moral hazard could arise on the part of exporters/ contractors. This could indeed be the case, because room for maneuvering is large for exporters who handle raw materials imported without tax or duty. 39. When I visited a DTI regional office, I was enthusiastically welcomed by the officer in charge, who mistook me for a Japanese businessman seeking investment opportunities in the garment sec- tor in the region. 40. The positive correlation between the commercial nature of the agricultural crops and the burgeoning of rural entrepreneurs is found widely in the early stage of modern economic growth in Japan. See, for example, Amano and Abe (1989) on how indigo production fostered rural entrepre- neurs in Tokushima Prefecture in Japan during the late nineteenth century. 41. One may recall that, in the interwar period in Japan, easy entry, backed by development of the market for secondhand power looms, was an important condition for Japanese rural entrepreneurs to join the cotton textile industry, the most popular rural industry of the time (Saito and Abe 1987). References Amano, Masatoshi, and Takeshi Abe. 1989. "Zairaisangy6 ni okeru shihon to keiei" ("Capital and Management in Traditional Industries"). In Mataji Umemura and Yuzo Yamamoto, eds., Kaiko to ishin (T7he Port Opening anid the Meiji Restoration). Tokyo: Iwanami. Hayami, Yujiro, Ma Quisumbing, R. Agnes, and Lourdes S. Adriano. 1990a. Tozward an Alterntative Land Refornm Paradigm. Manila: Ateneo de Manilo University Press. Hayami, Yujiro, Masao Kikuchi, Luisa M. Bambo, and Esther B. Marciano. 1990b. "Transformation of a Laguna Village in the Two Decades of Green Revolution." IRRI Research Paper Series No. 142. Los Bafios: IRRI. Export-Oriented Garnment Itndutstries in1 the Rural Philippines 129 Hayami, Yujiro, and Toshihiko Kawagoe. 1993. Thre Agrariani Origins of Conrmerce and Indutstry. Lon- don: Macmillan. NEDA (National Economic and Development Authority). 1992. Phiilippinie Developmienzt Report 1991. Manila: NEDA. NSCB (National Statistical Coordination Board). Various issues. Philippinie Statistical Yearbook. Makati, Metro Manila: NSCB. NSO (National Statistics Office). 1990. 1988 Cenisufs of Establishnments. Manila: NSO. . (various years). Philippine Yearbook. Manila: NSO. Ohno, Akihiko. 1994. "Kindai-teki seiz6-gy6 bumon ni okeru gin6 ikusei" ("Skill Development in Modem Manufacturing Firms"). In Ajia shokokn no ningenshigen kailatsiu to rodOryokut ni kansuirut ch6sa kenrkyti1 h6kokusho: Philippines (A Report on Humnan Resouirces Development and Labor Force in Asia: The P1rilippines). Tokyo: Association of Population and Development in Asia. Otsuka, Keijiro. 1992. "Green Revolution, Land Reform, and Household Income Distribution in the Philippines." EDCC 40(4): 719-41. Resnick, Stephen A. 1970. "The Decline of Rural Industry under Export Expansion: A Comparison among Burma, Philippines, and Thailand, 1970-1938." Jouirnial of Economic History 30(1): 51-73. Saito, Osamu, and Abe Takeshi. 1987. "Chinbata kara rikishokki k6j6 e: Meiji k6ki ni okeru menorimonogy6 no baai" ("From Fee-Contract Looms to Power Loom Factories: The Case of the Cotton Textile Industry in the Late Meiji Period"). In Ryoshin Minami and Yukihiko Kiyokawa, eds., Nihoni n1o k6gyoka to gijntsui kailhatsiu (Inzdustrialization and Technological Development in Japan). Tokyo: T6yo Keizai Shinp6sha. 6 The Rural Garment and Weaving Industries in Northern Thailand Akihiko Ohno and Benja Jirapatpimol r t Ihe recent, rapid industrialization of Thailand has been realized by an enormous T influx of foreign direct investment in the manufacturing industries of Bangkok and neighboring provinces. The resulting concentration of factories caused hyperurbanization in the metropolitan area, which has produced skewed distribution of income and employment opportunities. Rural Thailand is struggling with poverty, unem- ployment, and underemployment, and urban-based industrialization cannot be an effec- tive remedy for these problems. Great expectations are placed on the development of rural household nonfarm industries, which are more labor intensive and make fewer demands on scarce capital resources than do their modem, large-scale counterparts. These charac- teristics of rural household industries have often been put forward as a strong argument for government policies in their support. The purpose of this chapter is to identify the socioeconomic factors that facilitate the market competitiveness of the rural household textile industries in the process of eco- nomic development. Our inquiry is focused on rural household enterprises in the ready- made garment and hand-weaving industries in northern Thailand-Chiang Mai and its environs. Both are labor-intensive sectors of the textile industry. This region has recently witnessed a mushrooming of rural household enterprises in these sectors (Mead 1981; Amara and others 1989; Angkarb 1990; Sungidh and Poonpanich 1994), followed by the establishment of modem garment factories (table 6-1). The most crucial condition for an industry's survival is its market competitiveness. Any support to an industry that is insufficiently competitive will result in wasted resources and, consequently, will have an adverse effect on the national economy in the long run. It is inter- esting to note that the growth of these two rural household industries gained momentum 131 132 Tonard the Riural-Based Developmentt of Conmmrerce anld Indutstry Table 6-1. Modern Garment Factories in Chiang Mai Province Year Emnployees establislhed Male Femnale Location 1978 23 435 C 1979 17 390 S 1980 10 70 S 1984 4 160 S 22 170 1987 5 147 C 1988 12 100 C 1989 8 130 - 120 S 20 282 S 1990 5 75 C 25 173 26 165 S 1991 2 100 S 15 84 S 1992 3 64 C 1993 30 50 12 - S - 204 1994 - 90 60 S Note: C and S represent Chiang Mai Town and Sankampeang, respectively; blanks represent other locations. Source: Thailand, Ministry of Industry 1995. after the areas began to industrialize at the beginning of the 1990s; until recently there were few large-scale modem textile factories in the region. In addition, the growth of the rural household industries was achieved without explicit government intervention. Accordingly, we can safely state that these rural household industries are sufficiently competitive with their modem counterparts. The region is, therefore, an appropriate area for the study of issues related to rural household industries during the early stages of industrialization. The textile industry in northern Thailand is composed of a modem, large-scale sector and a rural household sector. Although large-scale handloom weaving factories are not as prevalent as large-scale garment factories, in Sankampaeng, for example, a large weaving factory (Workshop G) employs more than 200 workers and has 90 handlooms. Technologi- cal differences are generally used to demarcate the modem and rural household sectors. Technology, however, does not appear to be a significant factor distinguishing the two sectors in either of these two industries, because they use the same production tools- sewing machines and wooden-frame handlooms. This allows us to compare the organiza- tional efficiency of the modem and rural household sectors without having to consider the effects of technological differences. Tlte Rutral Garnentef and Weav)ing Indutstries in Northernt Thailanzd 133 As we illustrate below, the rural sectors of the garment and weaving industries show distinctive characteristics in their origins, modes of contractual relationships with urban businesses, and types of entrepreneurship. An investigation of the rural household indus- tries from a two-dimensional perspective, comparing the rural and the modem sectors and the garment and the weaving industries, offers a key to identifying the economic and social factors that promote or impede the development or rural household industries. The villages surveyed are mainly located in the Sankampaeng, Chom Thong, and San Patong districts of Chiang Mai Province and in the Basang district of Lamphung Province. They are all upland areas situated around the periphery of the Chiang Mai valley and are agriculturally backward areas because of inadequate water supply. It is often pointed out that agricultural sterility provides fertile ground for rural nonfarm industries.' All the rural informants are female (most of them married), since the rural textile industries are traditionally the exclusive domain of women.2 Our research was conducted in March 1994. History of the Thai Textile Industry Evidence from archeological excavation has confirmed that the Thai people of prehistory learned to use natural fiber for cloth weaving in order to make clothing and other useful materials, and this ability continued, through the Sukothai and Ayudhaya periods to the modem time. Cloth was woven in the past for household use and for sale to local people, as well as to neighboring countries. Several historical studies reveal that products from the Thai textile industry were traded with China, Laos, Burma, and England during the nineteenth century, with Bangkok and Chiang Mai acting as the centers for the Central and the Northern regions respectively (Bowie 1992). Thai cloth weaving in the early period employed simple throw-shuttle handlooms. A Silk Craft Department was set up in 1909 in the reign of King Rama V, and a Silk Craft School was established. This activity was extended to other provinces-Nakon Rachasima and Buri Rum. Since then, cloth weaving has changed from a mere household activity to production outside the household. In addition, fly-shuttle handlooms were introduced in the middle of King Rama VI's reign, although they were not yet popular among Thai people. The picture of the Thai weaving industry drastically changed with the use of the power loom and the development of the modem weaving and the initiation of the ready-made garment industry. It is said that in the 1960s modern textile production not only had re- placed import needs, but had also nearly superseded the old method of production. Al- though the household weaving industry declined, it picked up in some areas when domestic and foreign markets for artisanal goods were created through development of the rela- tional contract. This is where our story on the hand-weaving industry starts. The ready-made garment industry expanded in the 1980s, and in some years it had the highest export value among all export industries (Supat 1992). Up until 1991, there were 2,029 factories producing garments, ranging from those with fewer than 10 to those with 134 Towoard the Rutral-Based Developmenit of Conmmerce anid Inidutstry more than 1,000 sewing machines. Nevertheless, the stated number could be lower than the actual number of factories because small factories, with fewer than thirty (now ten) sewing machines, were not required to register as factories. Most small factories tended not to register because they wished to evade the regulations involved. Since most of gar- ment factories were initially located in Bangkok and its metropolitan region, the govern- ment had a policy to distribute siting of industries into the provinces by giving privileges to factories scheduled to locate outside Bangkok, in such locations as Chiang Mai. This is where our story on the garment industry starts. Scope of Analysis This chapter is mainly concerned with entrepreneurship and the mode of transactions between rural and urban businesses, and it attempts to analyze the emergence of rural household industries in northern Thailand. Entrepreneuirship The entrepreneurial or managerial functions necessary for the complex task of operating an enterprise are mainly in the areas of marketing management, including procurement of raw materials; labor management; and financial management. All these functions are usu- ally performed by the same person, the entrepreneur-manager, in household enterprises. In a rural setting, people seldom have formal training in functional areas, and therefore the entrepreneurship necessary for success in enterprises is often thought to be absent in a rural society. If this is true, what explains the present prosperity of rural household enter- prises in northern Thailand in tandem with industrialization? We approach this subject by investigating the areas in which rural entrepreneurs have advantages or disadvantages compared with their urban counterparts. Rural household enterprises, because of their small size and the difficulty in communi- cating with the outside world, are said to suffer from managerial disabilities, with the most critical in the area of marketing. Rural manufacturers, in general, struggle with diffi- culties in knowing and exploiting the needs of urban and overseas consumers. This sug- gests the importance of accessibility not only to marketing channels, but also to information concerning product details, which change according to the needs of consumers. Also, the procurement of raw materials suitable for marketable products is a crucial, but difficult, aspect of marketing. Insufficiency of funds biought about by a weak collateral position and poor access to credit are often suggested as other obstacles to the development of rural household indus- tries. This argument is unconvincing, however, because the capital requirements for rural household enterprises are not always prohibitively large. This capital bottleneck thesis thus must be reconsidered with regard to both fixed and working capital. We assume that rural household enterprises have an entrepreneurial advantage over their modern counterparts in the area of labor management. Two points may clarify this The Riral Garnientt and Weaz7ing Indiustries in Northernt Thailand 135 perspective. First, the enforceability of labor contracts, which are often implicit, is one of the major problems associated with employment relations. In the absence of an enforce- ment mechanism, no labor contract would ever be completed. In the early stages of indus- trialization, an insufficient formal enforcement mechanism, in the absence of a trained middle management, often casts doubt on the superiority of the factory system over rural household enterprises. In addition, although the creation of a disciplined workforce is imperative for successful factory management, newly recruited factory workers from tra- ditional societies offer tenacious resistance to factory organizations (Thompson 1967; Smith 1988; Clark 1994).3 This often leads to counterproductive behavior among workers, such as high turnover, absenteeism, lack of work discipline, and low morale (Ohno 1994). One can assume the possible predominance of informal over formal labor control in the early stages of industrialization, and this may offer a key to understanding the advantages of rural entrepreneurs over their urban counterparts. Second, the degree of compliance with labor regulations in Thailand is not high, par- ticularly among small enterprises (Naruemol 1993). Such enterprises can avoid the costs imposed by labor regulations, such as those concerning minimum wages, overtimne, paid leave, maternity leave, and welfare facilities for workers. Accordingly, this dualism en- dorses comparative advantages for rural household enterprises in the labor market. It follows that the more rigid the regulations on the modern sector, the more advantageous and attractive the rural household sector becomes. Mode of Transactions A distinctive feature of rural household enterprises in northern Thailand is their heavy reliance on urban businesses in the relational contracting system. However, while the put- ting-out system is widely used in the ready-made garment industry, this practice is rarely observed in the weaving industry. Instead, advance-order contracts (or guaranteed-pur- chase contracts) on a continuing basis are prevalent in the rural weaving industry.4 This contrast demonstrates how the rise of rural household enterprises occurred. Putting out is often seen as a transition phase from the craft shop or the guild system to the factory system (Wardell 1992). Landes (1969) finds the reason for the transition in "the usual difficulty of compelling performance by cottage workers" (work control) and "the poor quality of hand work" (quality control), which he refers to as "the frictions inherent in putting out." This shift can be achieved, as mentioned above, only when a system of formalized supervision is smoothly institutionalized in hierarchical factory organizations. When no such system exists, difficulties in work and quality control will cause managerial problems, even in the factory system. This is where we assume that informal labor control predominates over formal control in the early stages of industrialization. The shift of pro- duction from the cottage system to the factory through the putting-out system is not a universal historical sequence (see Chapter 1 by Hayami). The predominance of putting- out arrangements in northern Thailand since industrialization requires a different per- spective on the new role of the putting-out system. 136 Towvard t1te Rtiral-Based Development of Comnierce an7d Industry Hypotheses On the basis of the above discussion, the following hypotheses were framed to explain the dynamic growth of rural household enterprises after industrialization began in northern Thailand. HYPOTHESIS 1: Rural household enterprises have a comparative advantage in labor man- agement compared with modem enterprises, whereas rural enterprises suffer from a disadvantage in marketing, especially in the early stages of economic development. HYPOTHESIS 2: As a corollary of hypothesis 1, the relational contracting system-in the form of either a putting-out contract or an advance-order contract-is a device that allows rural household enterprises to overcome disadvantages in marketing. Con- versely, it is a device for their large-scale counterparts to overcome difficulties in la- bor management. Needless to say, the relational contracting system solves the problems of adverse selec- tion and moral hazard, which are not only the frictions inherent in the putting-out system as Landes pointed out, but also are frequently observed in the imperfect market. The Garment Sector The large garment factories of Thailand have recently begun to move gradually toward reliance on putting out part or all of the production process to small workshops in villages. Putting out in the ready-made garment industry involves about 280,000 workers in Bangkok and 500,000 workers in other provinces (table 6-2).5 The marketing process for ready-made garments in the putting-out system in northern Thailand is depicted in figure 6-1. Although its contractual arrangements take a variety of forms, the system consists of three distinctive tiers: urban businesses (the principals), group leaders (the subcontractors or agents), and the sewers. Although there are group leaders Table 6-2. Pattern of Employment and Wages in the Ready-Made Garment Industry Num1lber of Wages per day Prodtcer Toorkers intvoltved (baltt) Modem factory 17,000 80-100 Putting-out Bangkok 280,000 70-80 Other provinces 500,000 50-70 Total 797,000 Source: Sungsidh and Poonpanich (1994). Thze Ruiral Garnment and Weaving Inzduistries in Northerm Thailand 137 Figure 6-1. The Marketing Process in the Rural Garment Industry Overseas market Domestic market Export agencies 'A 61 Bangkok ........... ...... ................................ .. . . . . . . . _________ '4' _________ _________ Urban businesses Merchant-manufacturers Modem factories (Production) [GH] (Production) Chiang Mai, Sankampeang ~~~...... .......... .... . .. . .....,.. ................. Group leaders (Production) . ~~~[GC] [GD] [GE] [GF] Home sewers [GA] [GBI Note: Solid and dotted lines denote the flows of finished products and raw materials (fabrics), respectively. in the weaving sector as well, they do not rely on the putting-out system. Therefore, in this chapter the term group leaders is used interchangeably with subcontractors or agents in the garment industry. Urban businesses are either modem garment factories or wholesalers; wholesalers are middlemen bound by contract to work for a particular exporter in Bangkok. The group leaders are supplied with raw materials, especially fabrics that generally have been precut. The leaders bring the sewers together at their premises (hereafter, rvorkshop sewers), distribute the fabrics supplied among the sewers working at home (hereafter, honme 138 Towvard the Rlral-Based Developnmenit of Conmmerce anid Iznduistry Table 6-3. Workshops Studied Workshop Hon2e Year of Workshop wuorker wvorker opeatioyn Note Garments GC 0 8 3 Orders from a trader in Sankampeang GD 6 2 7 Orders from a modern factory, ex-factory worker GE 5 0 3Ms Subcontractor of GD, used to work for GD GF 0 12 2 The industrial estate, rather impersonal labor relations GG 2 7 7 Orders from a modem factor, ex-factory worker GH 20 Va 10 Merchant-manufacturer, difficulties in labor management Weaving WA* 5 2 3 A food-processing zone, custom order for local consumption WB* 3 12 20 Outskirts of Sankampeang town, sample orders, low quality WC+ 9 46 20 Remote upland village, medium quality, loose supervision WD* 21 40 10 Distance of 60 km from town, time-rate wages for preloom activities WE+ 35 105 20 Traditional homeweaving area, workers engaging in QC WF* 30 0 0 Superior varieties, no piece-rate wages, promotion system Note: Ms = months; Va = variable; widow; and + = village chief's wife. Sou(rce: Authors' data. sewers), or use a combination of the two systems. Individual cases of these entrepreneurs (shown in table 6-3) from our field survey data are presented in Appendix 6-1. The Weaving Sector The marketing process in the weaving sector is simpler than that in the garment sector. The weaving leaders in the villages have links with particular wholesalers or export agencies on advance-order arrangements. That is, the urban traders do not supply the leaders with raw materials. As in the garment workshops, the weaving leaders bring weavers together at their premises (hereafter, wtiorkshop wveavers), subcontract work out to home weavers, or em- ploy a combination of the two techniques. The leaders supply their weavers with cotton yam dyed at their premises and arrange for the sale of the finished products. Although their marketing system is simple, the weaving workshops have a wider vari- ety of organizational styles than do those of the garment sector. The overall picture of the sample workshops surveyed (from WA to WG) is indicated in figure 6-2, along the axes of The Rutral Garment and Weazlinig Industries in Northern Thailanid 139 Figure 6-2. Cases of the Weaving Industry A s High 30 + 0 + 0 quality F 90+0+0 21+0+40 9+46+0 35+0+105 G D C E Factory system Traditional management 3+0+12 5+5+0 A B Poor quality Note: Number of looms = leaders' premises + lending out + home-workers homes. "product quality" and "mode of labor management." The horizontal axis corresponds to the continuum of the factory system based on formal labor control compared with the traditional system based on informal labor control. The weaving process can be divided into two stages: preloom activities (dyeing, loosen- ing and unwinding, pirn winding, warping, and sizing) and loom-stage activities. In this chapter, the latter is emphasized. Individual cases of rural entrepreneurs (table 6-3) in the weaving industry are presented in Appendix 6-2. Characteristics of Rural Entrepreneurs and Workers Although there is a variety of leaders and workers in both the garment and weaving sec- tors, some common features can be identified. 140 ToWard the Rlural-Based Developmitenit of Commnerce and Inidustry Grouip Leaders as Rural Entrepreneurs It is a relatively recent phenomenon for leaders to set up their workshops in rural areas, employing neighboring villagers. These leaders can legitimatelybe called "rural entrepre- neurs" because they introduced new ventures into agrarian societies. The leaders in the garment sector are often women who worked for modem garment factories as line supervisors. Through this experience, they acquired the entrepreneurial skills of labor management in addition to sewing skills. Furthermore, they gained access to orders from modem garment factories; factory managers regard these leaders as quali- fied superintendents in the villages because of their previous experience as supervisors in the garment factories. The garment leaders are from a poor social stratum, a landless class of the agricultural society that is the source of sewers in the modern industry. The weaving leaders, however, are from diverse strata of rural society. Some are from poor families, often widows, while others are the wives of village chiefs. It should be noted that four of the six weaving lead- ers studied had started their businesses in order to earn a living after their husbands passed away. Weaving entrepreneurship is, therefore, not an ascribed status, but it can be as- sumed that there are many potential entrepreneurs in the rural weaving sector. Their no- table characteristic is that they themselves are skilled weavers who have inherited the traditional weaving culture. Both categories of leaders employ neighboring villagers on their premises, subcontract work out to home workers, or use a combination of both techniques. The relationship between the leaders and the home workers is akin to a putting-out contract: the former advances raw materials to the latter on the condition that the home workers supply fin- ished products.6 As we have found, however, there are no underlying differences between the home workers and the workshop workers. In our cases, the demarcation between the putting-out and employment relationships is blurred. For convenience, we regard home workers as wage earners employed by the leaders. Sezvers and Weavers The greater part of the sewers have prior experience in modem garment enterprises. After getting married, they returned to their villages and became sewers. In the early stages of industrialization, most of the workforce comes from agrarian communities. Female workers from such communities usually leave the factories in towns after marriage and return to their villages. The garment leaders, who are often returnees of the same kind, can thus hire an experienced workforce in rural areas. This life pattern of village women in northern Thai- land can be accounted for by matrilateral bias in the kinship system and preference for matrilo- cal residence (Karim 1995).7 It is safe to say that the modern garment factories create external economies by providing rural entrepreneurs (group leaders) and a skilled rural workforce (sewers).8 Therefore, as more modern garment factories are established, the number of ex- perienced returnees in rural areas will increase. A trait of the returnees is that their propen- T1re Rutral Garnmenit anid WeazJing Inidutstries in Nortternt Thtailanid 141 sity to migrate or to quit is very low. As a result, this labor market is highly segregated from the urban market and preserves the rural-urban wage differential. Weavers, in contrast, are local; weaving is a skill traditionally acquired by village women in northern Thailand. Weaving is now on the decline because of the penetration of mass- produced, ready-made garments into rural areas and the emergence of alternative job opportunities in modem factories. The weaving leaders, especially near towns (for ex- ample, the case WA), are suffering from a shortage of weavers. Most weavers are in their forties to sixties, while most of the sewers are in their twenties and thirties. It appears that recruitment of weavers is becoming difficult. Can Workers Be Entrepreneuirs? The economic and social differences between leaders and workers are generally small. But while being an entrepreneur is not an achieved status, in the case surveyed we could not find workers who had turned into leaders.9 In other words, a master-apprentice relationship of the guild style does not exist in the garment and weaving workshops. This implies that leaders are distinguished from workers by something more than the level of their production skills. What the leaders have-and the rural workers do not have-is access to an assured market for their products. In developing economies, where traditional values and prac- tices coexist with modernity, transactions between the two different worlds need inter- preters who are familiar with both worlds. The rural entrepreneurs can access and understand the information that links the villagers with the urban market, and they can interpret it for the villagers.'" In contrast, the urban entrepreneurs, who are not familiar with the rules and norms of village communities, are faced with difficulties in communi- cating directly with individual village producers directly. This is where the rural entrepre- neurs find their raisoti d'etre. As we have observed, the leaders are individuals who have links with both urban contractors and village workers. With these links, a group leader can claim the title of rural entrepreneur. Although most sewers and a number of weavers possess their own means of produc- tion, they are not independent, self-employed entrepreneurs; they are heavily dependent on the group leaders for procurement of raw materials and marketing of the products. As small producers living in remote areas, they are incapable of handling marketing issues at an individual level. To become entrepreneurs, individual rural workers must form a link with an urban business and develop an ability to manage villagers based on informal labor control. The ability to achieve these goals is not necessarily present in every villager. Contractual Arrangements Associations Between Leaders and Urban Butsinesses The relational contracting system characterizes the association between leaders and urban businesses. We must note the sharp contrast between the garment and weaving sectors in 142 Toward the Rtral-Based Developniort of Conmmiierce and Inidutstry their contractual arrangements. The garment leaders rely on urban businesses for the pro- curement of fabrics, as well as the sale of the products produced through putting-out ar- rangements. The weaving leaders, however, rely on the wholesalers only for the sale of their products on advance-order arrangements. The weaving leaders usually purchase cotton yam themselves in the market. Neither advances nor requests on the particulars of fabrics are common in the industry. Weaving wholesalers advance money only in the case of a bulk order."1 The reason for the heavy dependence of the garment leaders on urban businesses is that the leaders are not in a position to get information on the product market, such as designs, size, and kinds of fabrics, which largely determine product marketability. Advancing raw materials is indispensable in the garment sector because the materials provided suit the market demands urban businesses wish to satisfy. The weavers, in contrast, have a mo- nopoly on the skill and local artistic designs that capture overseas markets and satisfy grow- ing domestic demand for high-quality indigenous fabrics. This contrast makes clear the different functions of the relational contracting system in the two Thai textile industries. The above discussion, however, does not refute the capital bottleneck thesis altogether as the explanation of the putting-out system introduced in the garment industry. As far as working capital is concemed, the garment leaders are confronted with difficulties. While material costs in the garment industry account for approximately 90 to 95 percent of the delivered price, in the weaving industry they account for 50 percent of the price for supe- rior fabrics to 70 percent for inferior fabrics. In addition, garment leaders are relatively more impoverished than weaving leaders. It seems reasonable to suppose that a high con- straint on working capital for the garment leaders forces them to rely on the putting-out system. When the putting-out system is adopted, working capital is not urgently required. The question of a fixed capital constraint will be taken up later. All these things make it clear that rural household enterprises suffer from a disadvan- tage in marketing (see HYPOTHESIS 1, above), and that the relational contracting system is an effective device for overcoming the difficulties associated with marketing (see HYPOTHESES 2). The putting-out system is thus a modification of relational transactions. The system is useful in industries where rural workshops have difficulty obtaining market information on their products. The next question is why urban businesses decide to "buy" in the market, and not to "make" the products themselves. Before going into detail, we must clarify some differ- ences between the urban garment and weaving businesses. The urban garment businesses are modern garment factories or merchant-manufacturers (for example, the case of GH) who also "make" their products. The urban businesses in the weaving industry, however, are wholesale merchants. Therefore, the make-or-buy issue in the weaving industry is paraphrased into the question of why the decentralized system is more prevalent than centralized handloom factories.'2 The reason for decentralized production in the weaving industry is simple. Skilled weavers are village women, mostly married, whose skills are region-specific. Little needs to be said about the difficulties of shepherding such weavers into centralized factories. The Rutral Garment and Weaviing Induestries ini Northern Thlailand 143 There remains the second question of why the modem garment factories in northern Thailand rely on the putting-out system. Although we did not collect much information on putting out by urban businesses, the manager of a large modem garment factory in Chiang Mai indicated that by putting out, they can reduce the risks associated with invest- ment, including the enlargement of a factory, and they can reduce the cost of labor man- agement. In modern factories, labor protection laws make dismissals difficult. The factory can achieve greater flexibility by dividing its workforce into core and peripheral groups. The rural industries are useful as buffers that absorb fluctuations. Therefore, the relation- ship between the urban principals and the rural workshops is symbiotic. Nevertheless, the system has its drawbacks. According to one manager, it is difficult to complete production on schedule and to keep a close check on quality because most group leaders do not properly supervise their sewers. This is "the frictions inherent in putting out" observed by Landes in the phrase quoted earlier. To cope with these problems, the factory managers subcontract work out to villagers who previously worked in modem garment factories as supervisors. They transact in a recurrent fashion instead of in the spot market."3 The stable and recurrent transactions encourage mutual trust between urban businesses and leaders (HYPOTHESIS 2). The trust avoids adverse selection and moral hazard problems to a considerable extent (Akerlof 1970), because trust works as "an important lubricant of a social system" (Arrow 1974). Support for Rural Entrepreneurs Rural workshops have spawned spontaneously without any exogenous support. At the same time, support from govemmental and nongovemmental organizations has played an important role in bringing forth the rural entrepreneurs. As discussed above, a working capital constraint for the rural textile industries is not critical as long as they rely on the relational contracting system, especially on putting-out arrangements. However, a constraint on fixed capital reveals differences. Garnent leaders employ sewers who have their own sewing machines. In contrast, the weaving leaders have to lend looms for fabrics of double width without charge, since traditional handlooms are generally for fabrics of a single width. Therefore, the weaving leaders are faced with a fixed capital constraint. At the time they set up their businesses, weaving leaders often seek financial assistance from the Thai government to purchase the needed handlooms. Access to credit does not seem to be a problem-some leaders from a relatively poor stra- tum were provided credit without difficulty. Further expansion of a rural crediting ser- vice, sewing machines for the sewers, and handlooms for the leaders will be effective steps toward the formation of rural household enterprises. Institutions stimulatirig latent rural entrepreneurial talent play another vital role. Sev- eral government agencies are in charge of craft support, including tax privileges from the Board of Investment, crediting service by the Small Industries Finance Office, and market- ing promotion by the Thai Trade Center (Amara and others 1989). The Community Devel- 144 Toward ftie Ruiral-Based Dezvelopmentt of Comm1Z11erce anid lnditstry opment Department of the Ministry of the Interior assists the launching of group activities in rural areas (Thailand National Commission on Women's Affairs 1980). Nongovernmental organizations also engage in the formation of rural enterprises. The efforts of the YMCA in the promotion of rural entrepreneurs afford a good example of this. The following information is based on an interview with a staff member of the YMCA in Sankampaeng. The YMCA started a Chiang Mai Homenet Project, a five-year experimen- tal action research project, in 1990. The project assisted rural women in organizing and managing group activities. Major services and assistance provided by the project included group management, product quality improvement, product design development, market- ing information, and working capital. A major problem for such groups is marketing, and the project has undertaken several activities to promote marketing: * It organizes product fairs at least once a year. It also notifies its members about other trade fairs and advises them on the products they should prepare. * The project coordinates overseas orders coming through the YMCA. * The project publishes newsletters to inform its members about popular products and possible places to send their product samples. As a consequence, many groups have received orders directly from customers, and their products fetched favorable prices because the members do not have to pay a broker's fee. This assistance seems effective as a pump-priming policy in stimulating small rural entrepreneurs. Associations Between Leaders and Workers A salient feature of rural workshops is that the management hierarchy is shallow, even nonexistent. Instead, the mechanisms of community-based informal labor control are at work. Although direct evidence concerning the function of community labor control is, and is likely to remain, limited, its role should receive more attention in developing coun- tries, especially when the modern factory system has not been effectively established (see HYPOTHESiS 2). Informal labor control takes the form of either a peer group of socially homogenous members or a patron-client relationship in a simple hierarchy.'4 The two forms, however, are not mutually exclusive. They concurrently regulate the behavior of village workers in the rural workshops we surveyed. While in the garment workshops the control of the peer group predominates, in the weaving workshops that of a patron-client relationship pre- vails. This contrast can be explained by the different patterns of decisionmaking in the two industries. Compared with the simple hierarchy, the peer group is thought to be inefficient in both making and implementing decisions regarding work plans. In the garment work- shops, however, decisionmaking is not crucial because the necessary information on pro- duction is provided by parent factories through putting-out arrangements. Therefore, the drawbacks of a peer group would not become an issue for the garment workshops. In the weaving workshops, however, the leaders have a monopoly on the skill and designs of The Rtral Garment and Weaviing Indutstries in Northern Thailand 145 Figure 6-3. The Marketing Process in the Weaving Industry Urban businesses Urban areas (Chiang Mai, Sankampeang, Bangkok) . . . . . . . .. . . .. .. . . . .. .. . .. . . .. .. .. . . .. .. .. . .. .. .. . . .. .. . .. . . . . .. . .. .. . . .. . . . . . ........... Group leaders with weavers -- Market Home weavers Note: Solid and dotted lines denote the flows of finished products and raw materials (cotton yarn), respectively. traditional fabrics and play a vital role in making decisions on the products to be woven. A boss-free organization, therefore, is not practical for weaving workshops. Accordingly, the relative predominance of the controlling modes is largely determined by the differences in production techniques. It should be reemphasized that in the majority of cases studied, the social differences between leaders and workers are small. Strong ties of kinship and affinity-or at least friendship-exist among the members of the workshops, and they share common values and beliefs in village communities.Y' These relationships promote the formation of peer groups, especially in rural Thailand where the socially smoothing and conflict-avoiding value of kreng jai (care and consideration, as described further in Appendix 6-1) influences the villagers' attitudes and behavior.'6 Since the peer group works as a mutual monitoring system, effort-aversion leading to poor-quality output is largely prevented without fur- ther expense. Another point to note is that the association between the leaders and the workers in the workshops is not simply a pecuniary nexus, but is one of the many associations linking them in the village community. Under such tightly woven village relationships, a breach of any association, such as ill behavior in the workshop, would easily lead to village ostra- cism. Fear of such censure strengthens informal control (Hayami and Kikuchi 1980). 146 Tohvard thie Rutral-Based Development of Coninmerce and Inditstry It follows that in the rural workshops, extra-economic forces motivate villagers to work intensively and deters adverse selection and moral hazard. In large modem factories, in contrast, the moral reasons for hard work must be articulated.'7 Enforcing such imperson- alized norms involves high transaction costs in an economy where trained manpower for supervision is deficient, and people have not become familiar with factory organizations. This tends to cause organizational dysfunction in modern factories in developing coun- tries, which reveals a major advantage of the rural household enterprises over their large- scale counterparts (see HYPOTHESIS 1). It can thus be assumed that the prosperity of the rural industries in the early stages of industrialization partly reflects the limitations of early factory management. The principle of "work and you will be rewarded" is strictly applied in the rural work- shops, although in a mode different from that employed in modem factory organizations. In conclusion, the rural household industries have the advantage of low labor costs, a result of the dualism created by labor regulations, the cost of migration to the urban labor market, and the low monitoring costs of household work. Conclusions and Policy Recommendations It should be concluded that the rural household industries would have a great chance of development if their comparative advantage in labor management were fully utilized and their comparative disadvantage in marketing were covered by appropriate relations with urban businesses to assure recurrent transactions, including putting-out arrangements. The informal labor control in the village communities should be recognized as one of the most important resources that rural entrepreneurs can mobilize. Equally important is that the expansion of the relational contracting system into rural areas may enhance the mar- ket competitiveness of the modem garment sector.'8 Based on the above discussion, the following steps may be recommended in a policy program designed to create and foster rural entrepreneurs. 1. Establishment of modern garmentefactories in rural areas as a training groundfor both gar- ment leaders (rural entrepreneurs) and sewers. Modem garment factories create external economies by providing rural entrepreneurs and a skilled workforce, and thus tax incentives may be offered to promote their introduction in rural areas. 2. Fturnishing information on produict market to weaving entrepreneurs. Although there are numerous latent weaving entrepreneurs in remote villages, their lack of marketing ability has forced them to remain small producers for local consumption. As our cases indicate, once they enter contractual relationships with urban businesses, they are able to expand their operations. Steps to assist this linkage should be taken. One pos- sibility is to hold trade fairs to assist local producers in forming connections with the urban and overseas markets. Some of the weaving leaders we interviewed had linked with urban businesses through trade fairs sponsored by the government. Institutions, including nongovernmental organizations, that link latent rural entrepreneurs with the urban and overseas markets are effective in spawning rural entrepreneurs. The Rural Garment and Weavinig Inidustries in Northern Thiailand 147 3. Preservation of dial labor standards. For a developing country, Thailand's labor laws are rather strict. The rural household enterprises, however, either stay outside the frame- work of the law or simply do not comply with it. This dualism enhances the market competitiveness of rural household enterprises. Therefore, this dualism should be preserved for the time being. Many of the rules in the labor law are inappropriate for workers in rural household industries. Loose application of labor laws will benefit the rural labor class through increased employment opportunities in expanded rural industries. Some developed countries are insisting that such dualism violates ILO (International Labor Organization) conventions and that implementation of core la- bor standards should be linked to access to world markets. It is mainly in the small- scale enterprises that there have been violations, and a convention on home-based workers was proposed by the ILO. Reactions of developing countries toward strict imposition of the ILO conventions have been strong. It should be noted that as far as the Thai garment industry is concerned, the dualism in legislation is not a direct cause of the trade friction. Garments destined for export are produced by large factories without the involvement of the putting-out system because of the necessity for qual- ity control. Garments that are produced through the putting-out system are for do- mestic consumption and export to neighboring countries that do not require the high standards that other exports demand. Accordingly, our standpoint is that the dualism in legislation should be preserved for a time so that the diffusion of industrial and cornmercial activities over rural areas can be achieved. 4. Suipply of chieap production tools. Sewing machines and handlooms are the main fixed capital of the two household industries examined in this study. Roughly speaking, two types of sewing machines were in use: manual machines and high-performance electric machines. The difference between them accounts for the levels of labor pro- ductivity, and hence for the wages the sewers receive. In addition, the possession of a sewing machine is a prerequisite to becoming a member of a sewing group. One can say that an ample supply of cheap, high-performance sewing machines and the for- mation of a secondhand market for these machines will facilitate the emergence of garment groups. A credit service for individual sewers, not for the garment leaders, to enable them to purchase sewing machines is another recommended policy. Our study demonstrates that the role of the Thai government in providing rural credit for the weaving leaders, not for the individual leaders, in installing handlooms assisted the growth of the weaving workshops. Expansion of a rural credit service will be an effec- tive step in fostering the formation of weaving groups. 5. Development of infrastructutre and preservation of traditional wveaving skills. Develop- ment of transportation and communication networks is imperative for efficient marketing of village products. A majority of the leaders surveyed take orders by telephone, and some weaving leaders have created new product designs by adopt- ing patterns in vogue that they had seen in fashion journals or on television. Be- cause home weaving is on the decline in rural areas as a result of the penetration of ready-made garments, the weaving craft must be encouraged and taught to the younger generation. 148 Toward the Rutral-Based Developmenit of Comnmerce and Inidutstry Appendix 6-1. Workshops and Sewers in the Garment Industry GA,. home sewer 1 in a village near Sankampaeng. Mrs. GA, in her late twenties, with two children, migrated from Bangkok to the village of her husband after marriage. In Bangkok she worked for a modern garment factory (more than 200 workers) for six years. Last year she bought a used electric sewing machine (9,000 baht) and a used hem-stitch sewing machine (2,500 baht) and entered into a contractual relationship with a leader from an adjoining village. Mrs. GA gets batches of 200 pieces of precut fabrics for shirts and sewing thread on a five- day cycle. Although she sometimes fails to meet deadlines, no explicit penalties are im- posed. After handing over finished garments to her leader, she gets another batch of materials. She sews only a part of the garment and is paid 2.5 baht per unit. She works an average of twenty days a month, and her monthly income is 2,000 baht, or 100 baht each working day. In 1993 the legal daily minimum wage in Chiang Mai Province was set at 101 baht, and 93 baht in Lamphun Province. Home sewing thus brings in a sizable income for a housewife. Mrs. GA's leader employs fifteen workshop sewers at her premises and subcontracts work out to home sewers. The leader, in turn, acts as a subcontractor for a modern gar- ment factory in Chiang Mai. When the leader cannot complete contracted work at her workshop before a deadline, she asks Mrs. GA to do a different type of sewing. Although this pays 5 baht for each piece, only ten pieces can be finished in a day (50 baht per day). When asked why she does not turn down this work, her answer was "krenlgjai."'9 Krengjai denotes caring and consideration. It is often regarded as a Thai value essential for smooth- ing and maintaining good social relations (Suntree 1990). It can also be viewed as a social obligation toward others. GB, home sewer 2 in a village near Sankamiipaeng. After marriage, Mrs. GB quit a garment factory where she had worked for ten years. She is in her early thirties, with two children. She bought a manual sewing machine and started to receive putting-out work some years ago. Her leader subcontracts work out to twenty home sewers and employs five workshop sewers on her premises. A month before the interview, Mrs. GB bought a used electric sewing machine for 8,000 baht. She estimates that with the new machine she can sew about 20 percent more than with the- manual machine. Her leader supplies precut fabrics, but Mrs. GB has to buy the sewing thread herself in the market. Her contracted work has to be finished on a two- or three-day cycle. No bonus is given for keeping the appointed date of delivery. She makes complete shirts and receives 2.5 to 4 baht for each unit. On average, she earns about 100 baht a day, although she has to work long hours. Her products are for domestic consumption only because of their poor quality. GC, leader 1 in a village near the town of Sankampaeng. Mrs. GC, in her early thirties, is the wife of a carpenter. She receives orders from a trader, Mrs. Z, in the town of Sankampaeng, who is supplied with fabrics and designs by a whole- The Rurml Garnment anzd Weaving Indzustries int Northerm Thailand 149 saler in Bangkok. Mrs. Z cuts the fabrics and distributes them to her subcontractors. The products are exported mainly to the Middle East. According to Mrs. GC, she happened to meet Mrs. Z and became her subcontractor, bringing together twelve neighbors with manual sewing machines (3,000 baht each) to work for her. Mrs. GC herself does not sew but engages in quality control and brokerage. The sewers are all married, in their thirties or forties, and from a poor social stratum of the village. Their husbands are either carpenters or construction workers. In the beginning the sewers installed their own sewing machines at Mrs. GC's premises, but now all are home sewers because working at home is more convenient for allocating time between sewing and household chores. Mrs. GC now controls eight sewers, since four left her to work at a new ceramic factory near the village. The factory pays 50 to 60 baht a day, but provides guaranteed work throughout the year (1,250-1,500 baht monthly). Mrs. Z assigns a shirt to Mrs. GC to sew for 4 baht, and Mrs. GC pays her home sewers 2.5 baht per shirt. Mrs. GC supplies her home sewers with sewing thread at no cost and allows them to use her hem-stitch sewing machine (6,500 baht) without charge. The pro- portion of defective products sent back from Mrs. Z is about 2 percent. Mrs. GC asks her home sewers to resew them. When resewing is impossible, although this is quite rare, the sewers have to buy the shirt back for 60 baht, which is the cost of the fabric. On average, leader GC receives orders for 2,000 to 3,000 shirts monthly and distributes work to her home sewers according to their requests. One of her most important roles is filling orders and handling the sewers' requests. To this end, she must stay on good terms with her home sewers; in other words, she must evoke krenig jai to coordinate the distribu- tion of work smoothly. An average sewer completes twenty to thirty shirts a day. This brings in 50 to 75 baht daily, or 625 to 920 baht a month (2,000-3,000 shirts/eight sewers). The wage rate is sig- nificantly lower than those in the cases of GA and GB, mainly because the sewers of GC use manual sewing machines in contrast to the electric machines used in the cases of GA and GB. The net earnings of leader GC amount to 3,000 to 4,000 baht a month, which is almost equal to the salary of blue-collar workers in modem garment factories in the area. The orders from the urban trader, Mrs. Z, involve only four types of designs. Although Mrs. GC wants more orders, she is unwilling to deal with other urban traders. This is because her sewers are disinclined to master different sewing patterns. During the periods of peak agricultural activity (May to July), the daily wage rate for agricultural laborers rises to 80 baht. Because many sewers want to work as agricultural laborers during this season, Mrs. GC has to take smaller orders during this period. GD, leader 2 in a village near Sankamnpaeng. Mrs. GD, in her late thirties, is the wife of a driver and worked as a supervisor in a modern garment factory in Chiang Mai for sixteen years. Mrs. GD takes putting-out orders from large garment factories. Fabrics supplied by the factories are cut by her elder sister accord- ing to designs specified by the factories. The sewers are paid two baht per product, and the leader receives four baht when each is delivered to the factory. The costs of the sewing 150 Towvard the Rutral-Based Development of Conimerce and Indutstry thread, a hem-stitch sewing machine, and electricity are borne by the leader. An average sewer completes thirty pieces daily. This brings her daily wage to 60 baht, which is almost equal to the agricultural wage rate in the surrounding villages. Mrs. GD employs six workshop sewers and three home sewers. According to her, there are no differences between the two groups. Some sewers gather at the leader's premises partly because they can use Mrs. GD's hem-stitch machine and partly because they just enjoy working with friends. The sewers have their own manual sewing machines (3,000 baht). All had worked for large garment factories until marriage. Mrs. GD engages in quality checks, sewing, and buttonholing. The work hours in the workshop are from 8 a.m. to 5 p.m., although sometimes they are extended to 9 p.m. in order to meet a delivery date set by her urban business. Mrs. GD stated that it was critical for this kind of business to obtain stable orders, or sewers would leave for higher wages. She subcontracts work out to some subleaders on a short-order basis-for example, 200 pieces in two days. The subleaders are former em- ployees of the workshop. When the subleader fails to finish the assigned work, Mrs. GD withdraws the work. She tries to allocate constant work to subleaders who do a neat job and meet deadlines. This policy acts as an enforcement mechanism that deters the oppor- tunistic behavior of the subleaders, which is shown in the next case. GE, leader 3, a subleader of GD. Mrs. GE in her late twenties, the wife of a carpenter, and used to work for leader 2 (GD). She quit the work when her child was born and became a subcontractor of Mrs. GD. She employs five neighbors, who have their own manual sewing machines (3,000 baht each). They are all married and have working experience as sewers in modem garment factories. All the sewers work together at the leader's premises, bringing their machines. Mrs. GE has a hem-stitch sewing machine that all the sewers can use without charge. Although precut fabrics are supplied by Mrs. GD, the cost of the sewing thread is borne by Mrs. GE. The sewers are paid three baht per product. An average sewer completes ten sets a day. This activity provides a significantly lower wage than those paid in the other cases de- scribed here. The leader, Mrs. GD, sometimes speeds up the work, threatening to supply fewer orders if the sewers do not do satisfactory work. When many orders come in, all the sewers have to work until 9 or 10 p.m. to meet the delivery date. Peer pressure seems to force them to work intensively. GF, leader 4, a married couple in a village near the Industrial Estate in Lamphun. Mr. GF, age thirty-one, is from Bangkok and was a dealer of machine parts. His wife, age twenty-four, worked for a modern garment factory in the industrial estate as a line super- visor for four years. She quit after having a child. They have high-performance electric sewing machines (13,000 baht each) and a hem-stitch machine (20,000 baht). Thte Rutral Garnienit anid Weazving Inidutstries in Northern Thtailand 151 Mrs. GF sews and checks the quality of the products, and Mr. GF cuts fabrics with his cutting machine (2,000 baht). They take orders for factory uniforms exclusively from the factories in the industrial estates. At present they employ twelve home sewers, all married, in their thirties, and with experience as sewers in modem garment factories. Three of them once served under Mrs. GF in the garment factory. The electric sewing machines are lent out to the home sewers without charge. When asked why the sewers do not work at the leader's premises, Mr. GF replied that if he employed more than ten workers at the workshop, he would have to register his business and would be charged corporate taxes.20 An average sewer can complete ten uniforms a day with a high-performance sewing machine, but the resewing rate because of poor quality is about 20 percent. With an ordi- nary electric sewing machine, only two or three pieces can be finished a day, Mr. GF stated. A sewer receives 20 to 25 baht per piece, a daily wage of 160 to 200 baht, taking into account the resewing rate. Although this wage rate is quite high, Mr. GF stated that it was necessary to hire skilled sewers for the high-performance sewing machines. Sewers who produce poor work are subject to dismissal. By the time of our interview, three sewers had been replaced. Mr. and Mrs. GF have no intention of expanding their operations because it would be- come difficult to check quality and to manage sewers. When the number of orders is be- yond their capacity, they sell orders to other garment workshops. GG, leader 4 in a village near Sankampaeng. Mrs. GG, age fifty, worked for a modem garment factory in Sankampaeng for twenty years. She left the factory seven years ago and started her own business with her twenty- seven-year-old daughter. They installed three high-performance electric sewing machines (16,000 baht each). In addition, nine sewing machines have been borrowed from a garment factory in Chiang Mai for which they do subcontract work. They have several sources of orders: two modern garment factories in Chiang Mai and Sankampaeng, a five-star hotel in Chiang Mai that requires evening gowns, and a school that needs uniforms. Mrs. GG secured those con- tracts through her personal network. There are two workshop sewers and seven home sewers, all of whom have work expe- rience in modem garment factories. They are all married and in their thirties. Although there are no differences in wages between the two groups, Mrs. GG tries to assign constant work to the sewers working at her workshop as much as possible. The leader organizes the sewers along an assembly-line process. Each worker sews a small component of the whole. Mrs. GG engages in quality checks and sewing as well. When asked whether it was difficult to regulate the flow of components from each sewer so that they would reach the final assembly point smoothly, Mrs. GG stated that she found little difficulty, partly because she had experience as a line supervisor in a large factory and partly because there was a small number of workers. 152 Tow ard the Rutral-Based Devtelopmentt of Comminierce and tindutstry GH, an export-oriented merchant-manutfactutrer in the town of Sankampaeng. A married couple in their sixties, who used to run a small retail shop, started their gar- ment-manufacturing business after they were connected in business ten years ago with a Bangkok-based export agency. Fabric and designs are supplied by the agency. Their products are exclusively for export, mainly to the Japanese market, which requires high-quality products. In the beginning, the products were often rejected when the work was not satisfactory. The rejected garments were sold in Chiang Mai and fetched a higher price than when exported. Exports, however, promise a more constant return than domes- tic sales because the local market is too small for continuous sale of the products. Since they tightened quality controls, few products have been rejected. They have a workshop and employ twenty sewers there. Most of the workers are local married women. The owners now have to rely on seasonal workers from remote areas, because increased job opportunities with higher wages in modem factories make it diffi- cult to recruit sewers in and around Sankampaeng. Five single women are seasonal work- ers from remote villages who stay in rooms provided by the owner at no cost. During the peak agricultural season, they return to their home villages to assist with agricultural work. Large orders are subcontracted out to home sewers nearby. These sewers are former employees, and thus are skilled and trustworthy. According to the owners, the sewers used to rely on the owners for work, and the threat of being laid off could be used to enforce higher work performance. Nowadays the owners have to resort to other means of persuasion because of the increasing alternative job opportunities in and around the town. The home sewers sometimes do not comply with the requests of the owners, who found managing sewers a source of anxiety and deplored the lack of loyalty of the sewers. "It's a headache," Mrs. GH said twice. The assembly process is split into several parts, and each sewer engages in only one part so that the owners can monitor the quality of the work. For workers who do not have sewing machines, the owners either lend a machine without charge or assign them em- broidery tasks. The owners do not contemplate expanding their business because of a lack of capital and the possible difficulties in managing a larger number of workers. Appendix 6-2. Workshops and W\rkers in the Weaving Industry WA, Workshop A, in a village located in afood-processing zone. Mrs. WA, in her late fifties, started her weaving business three years ago after her husband passed away. She borrowed seven handlooms through a government agency that assists rural women. She employs five workshop weavers and two home weavers. Fabrics are made based on the orders of neighbors. The quality is relatively poor. Mrs. WA is a skilled weaver from Lamphun Province, where traditional weaving has flourished. In her present area, however, weaving has not traditionally been practiced Tile Riral Garment and Weavinig 11tditstries in Northernt Thailanid 153 because the people enjoy good job opportunities in the food-processing industry. There- fore, Mrs. WA has to teach her workers how to weave. Wages are paid on a piece rate, 50 to 55 baht per meter. On average, workers weave two to four meters a day, which brings the weaver 100 to 200 baht daily. Their wages are significantly higher than those for agricul- tural laborers and factory workers because of the scarcity of weavers in the area. Mrs. WA is finding it difficult to run her business because the flow of orders is not secure. WB, Workshop B near Sankampaeng, producing low-qtualityfabrics. Mrs. WB, in her early fifties, started her business twenty years ago after her husband died. Since then, she has formed links with three wholesalers, two in Bangkok and one in Chiang Mai. The workshop produces fabrics of inferior quality for curtains, bedspreads, and lun- cheon mats. Mrs. WB accepts orders only for selected fabrics of varying designs and tex- tures traditional in the area. There are three workshop weavers and twelve home weavers who have their own looms. In addition, ten workers are engaged in preloom activities. All are older women, of whom the youngest is forty years old. In the case of a bulk order, Mrs. WB will receive a cash advance equivalent to around 40 percent of the delivery price. When working capital is in short supply, she buys the cotton yam on credit from a yarn merchant. The cost of materials accounts for approximately 70 percent of the wholesale price, and wages account for 25 percent (15 percent for weavers and 10 percent for workers engaged in preloom activities). The remaining 5 percent is the profit margin for the leader. Mrs. WB's profit margin sometimes disappears when the price of cotton thread rises. Wages are paid on a piece rate. Although different fabrics have different rates, the aver- age weaver earns 80 to 90 baht a day. Mrs. WB offers some incentive wages, such as a bonus, to enhance weavers' work performance, but her workers have not responded posi- tively. The availability of weavers is another problem in the area, because the village is on the outskirts of Sankampaeng, where younger people have job opportunities with better working conditions. As a consequence, the workforce is aging, and with the passing of the present generation, this kind of workshop, as well as workshop A, will certainly decline, if not die out altogether. WC, Workshop C in a remote village. Mrs. WC, sixty-five years old, the wife of a village chief, lives in a village about 20 kilome- ters from Chiang Mai. She began her weaving business ten years ago with five handlooms. Her products are average in quality. At the outset, the products were sold in town markets. For the past five years, she has had constant orders from a wholesaler/exporter in Sankampaeng. This led to the expan- sion of her business; she now has fifty-five handlooms. The purchase of these handlooms (3,000 baht each) was financed by government assistance. MIrs. WC installed nine looms in her house. The remaining looms are loaned to the home weavers in the village. Although 154 Towvard the Rutral-Based Detvelopment of Comnnerce and Industry some villagers have their own looms, they are indigenous looms for fabrics of single width, which do not have a large market. Handlooms last about thirty years, and depreciation is thus negligible. Therefore, as a rule, no rental charge is levied. Whether a weaver works at the workshop makes no difference to the owner, but the space she can provide for install- ing handlooms at her premises is limited. The wholesaler sometimes offers suggestions on color and quality of cotton yam for the products, but no advances are given. Although the workshop is open from 8 a.m. into the evening, there is no strict rule about working hours. When we visited, there were five weavers at work, and some were taking naps on a bench. Because the owner usually stays in her room, no strict supervision is exercised. Stopping and chatting seem to be common practices in this workshop. The work discipline required in modem factories is irrelevant in rural workshops. Wages are paid on a piece rate. In other cases, except workshop F, a piece rate is usually based on the length of the fabrics woven. In this workshop, the workload is measured in the gross of thread bobbins used in the production process. This is because a piece rate based on the length of fabrics woven with loose supervision offers opportunities for the weavers to indulge in reducing the number of picks of weft yam, which results in inferior production quality. Workers are paid 600 baht when they have woven 6 kilograms of woof and 4.5 kilograms of weft into fabrics, which usually takes ten days. The daily wage rate is thus about 60 baht, which is almost equal to the agricultural wage rate in the area. Workshop C can be characterized as having loose supervision, since the owner is too old to supervise intensively. Nevertheless, the problems of adverse selection and moral haz- ard seldom occur. There may be two main reasons for this: (1) there are few alternative job opportunities in this remote area for women in their forties and fifties; therefore, losing a job in the workshop immediately leads to joblessness for life; and (2) as discussed in the text, community relationships are likely to function as an implicit supervisor. These condi- tions deter opportunistic behavior by workers. In addition, because product quality need not be superior, strict supervision is not necessarily required. WD, Workshop D in a remote village. Mrs. WD lives in a village 60 kilometers from Chiang Mai. She is sixty-six years old and still active. After her husband passed away ten years ago, she started the weaving busi- ness with two handlooms. She established business relationships with an export agency in Bangkok after she sub- mitted her products to an exhibition sponsored by the government. Thereafter, her opera- tion expanded, and now she has twenty-one handlooms at her workshop. She employs twenty-one workshop weavers and seventeen workers for preloom activities. In addition, some forty home weavers used to work in her shop. Only in the case of a bulk order will the wholesaler advance money, up to 30 to 50 percent of the delivery price. Mrs. WD would like to have business relationships with other traders, but she thinks it will be impossible for her to contact them by herself. She feels that all she can do is wait until an export agency approaches her. At present, overseas The Rutral Garmiient and Weazving Indutstries in Northern Thailapid 155 sales account for half the gross returns of the workshop. High-quality fabrics with intri- cate designs are for export; plain, coarse fabrics are for the domestic market, One of the distinct characteristics of Workshop D is that the leader makes frequent in- spection visits to the floor. While the weavers are paid on a piece-rate basis, the workers for the preloom activities are paid on a time-rate basis, 45 to 50 baht a day. It must be noted that a time-rate wage system is made possible because of the intensive supervision of the owner. Working hours are fixed, from 8 a.m. to 5 p.m. for preloom workers, while the hours of the weavers are rather flexible. Although the designs of the fabrics are generally from the traditional patterns of the area, the owner creates new designs for fabrics by adapting patterns from a fashion jour- nal and television programs. She teaches her weavers how to weave the new patterns. Her daughter, forty-eight years old, takes care of a retail shop that Mrs. WD set up in the town of Chiang Mai. The most crucial problem this workshop faces, according to Mrs. WD, is the decline of home weaving prompted by the penetration of mass-produced, ready-made garments into rural areas. Since the younger generation has not woven at home, weavers need to be trained to acquire the necessary skill before employment. WE, Workshop E in an areafamtouesfor traditiotial honme weaving. Mrs. WE, the wife of a village chief, lives in a village 30 kilometers from Chiang Mai. She started her weaving business twenty years ago with fifteen handlooms. At the outset, she sold her products to a wholesaler in Chiang Mai. Her sales picked up considerably when she formed a link with a Japanese export agency five years ago. This connection was estab- lished when she submitted her products to an exhibition sponsored by the govemment. Now 60 percent of her production is sold to the export agency. She currently employs 35 workshop weavers and controls 105 home weavers. At the beginning of business dealings with the agency, the proportion of the products rejected because of deficiencies was 20 percent. This problem was overcome by hiring three workers, relatives of Mrs. WE, to check the quality of the products carefully. Their wages, in addition to those of the workers for the preweaving activities, are paid on a time rate, 100 baht daily. Working hours are fixed, from 8:30 a.m. to 5 p.m. One of the notable features of this workshop lies in the distinction between workshop weavers and home weavers. While the home weavers are assigned fabrics with general designs, workshop weavers are assigned fabrics with designs that are specific to the work- shop. Because of this, the wage rate of the workshop weavers is higher by about 15 percent than that of the home weavers. In order to keep the weaving technique secret, strangers are not permitted on the work floor. Mrs. WE finds it difficult to meet the delivery dates of the wholesaler. She gives bonuses to home weavers who meet deadlines, and to strengthen the commitment of the weavers to her workshop, she reiterates "if you are good enough to work for my workshop when it is busy, you will have work when it goes slack," or "this work is important not only to the 156 Toward the Rutral-Based Developnmenit of Commterce anid Industty workshop, but also to you." She recognizes that she must stay on good terms with her workers to obtain kreng jai. WF, Workshop F produtcing superior varieties offabrics. Mrs. WF set up a small workshop thirty years ago, employing several weavers in a village 50 kilometers southwest of Chiang Mai, after her husband passed away. At the start there were handlooms for fabrics of a single width, and the products were rarely marketed. After ten years, Mrs. WF installed ten handlooms that she improved for fabrics of double width, 36 inches. In addition, she opened a shop in Chiang Mai for the sale of her products. The workshop continues to use traditional methods, such as using cotton of traditional varieties, hand-spun yarn, and natural dyes made from plants instead of chemical dyes, to produce soft fabrics with delicate colors. Hand-spinning has practically disappeared in Thailand because of growing imports and the production of mill-spun yarn. In 1985 she was awarded the coveted National Artist Award. When Mrs. WF died a few years ago, her daughter took over the business. There are now thirty handlooms, ranging from those with a single treadle to those with four treadles. The latter are for weaving quality fabrics with intricate designs. No home weavers are employed by the workshop because it would be difficult to keep a close check on quality if work were subcontracted out. Time-rate wages are paid both to weavers and to workers performing preloom activities. Newly hired weavers have to serve a probation period of six months and are then promoted according to their weaving skills. Earnings from weaving activities among the individual weavers ranges from 55 to 150 baht a day, according to their abilities. The weavers are ranked in several classes of differ- ent abilities. There are twelve weavers ranked as superior, of which four enjoy the highest wages of 150 baht a day. This is significantly higher than the agricultural wage rates, 60 to 70 baht, in the area. Weavers who produce fabrics of inferior quality often are subject to demotion. Mrs. WF's daughter is the only supervisor in the workshop. She stated that a supervisory system relying on foremen, the practice in large factories, would not work. because the workers were not accustomed to supervising acquaintances, nor to being su- pervised by acquaintances from the same social group. Working hours are from 8 a.m. to 5 p.m. Mrs. WF's daughter picks up the workers living in two villages by car. There are no holidays except for festival days, because the workers do not take days off. During the interview we often encountered stories that revealed a strong preference for work rather than leisure because of the low income levels in the rural areas. Workshop F has features distinct from the other cases in that its products are renowned as high-quality, hand-woven fabrics of the traditional style. To keep the quality of its fab- rics, the workshop has institutionalized systems that characterize the modem factory or- ganization, such as time-rate wages and promotion. Nevertheless, labor management is still based on personalized control. It must be noted that the indigenous labor manage- ment of personalized control is not incompatible with the features of the modem factory system, such as a promotion and a time-rate wage system. The Rutral Garnnent and Weaving Industries in Northern Thlailand 157 Notes 1. This is often asserted in discussion of proto-industrialization. See Mendels (1972). 2. On workers in rural textile industries, see Benja (forthcoming). 3. This is expressed bestby Landes (1969). As far as "factory equals discipline cum supervision," for the first generation of factory workers from a traditional working milieu, "the factory is a new kind of prison; the clock of a new kind of jailer." 4. For definitions of types of contracts, see the chapter by Hayami in this volume. 5. On putting out in the garment industry in Bangkok, see Voravidh (1993). 6. Chapter 3 by Itoh and Tanimoto examines the relationship between leaders and home weav- ers in putting-out contracts in the Japanese weaving industry. 7. Matrilocal postnuptial residence among rural Thai women is common. See Apichat, Morgan, and Rindfuss (1986). 8. Note that this statement is only possible in economies where female factory workers return to their home villages after marriage. See Chapter 10 by Ohno and Kikuchi. 9. This shows a sharp contrast with the garment entrepreneurs in the Philippines (see Chapter 5 by Kikuchi). 10. For this statement, we rely heavily on Walter (1984). 11. Note that an advance in the weaving industry is made in cash, not raw materials. 12. Although a powered-loom weaving factory on a large scale is a plausible alternative, it is of no immediate relevance to discuss this industry because of product differentiation. Rural weavers pro- duce either coarse or high-quality varieties such as twilled fabrics, while the powered-loom indus- try produces plain fabrics of medium quality. This habitat segregation was observed in India as well. See Yanagisawa (1993). 13. This phenomenon was analyzed in terms of clientelization by Geertz (1978). 14. On the mechanisms of the peer group, see Feldman (1984), Leibenstein (1987), Baron (1988), and Casson (1991). 15. This produces the high goal congruence among group members that is a crucial characteristic of clan organization. In clan organizations, the employment relationship becomes relatively effi- cient because monitoring of work performance is unnecessary (see Ouchi 1980). 16. It is pointed out that with the introduction of capitalism in Thailand, kin-based relations have been replaced by patron-client relationships (see Akin 1975). As far as our cases are concerned, however, kin-based relations are still predominant. 17. A list of factory regulations was posted on the wall of the plant of a large-scale weaving factory in Sankampaeng (Workshop B in Appendix 6-2). They said, "In the workshop, i) do not bring in drinks, ii) do not play games, iii) do not steal, iv) do not bring strangers into the workshop, v) follow the supervisors' instructions, vi) do not quarrel with your fellow workers, and vii) if you violate more than two of the regulations, you will be fired on the spot." 18. Landes insists that "it was this expansion into rural areas that made European and especially English industrial products competitive worldwide." He stresses the availability of the cheapest labor in rural areas as a main advantage of the putting-out system (Landes 1986). Note that our study insists that comparative advantage in the area of labor management based on informal con- trol is another advantage of rural household enterprises. 19. This does not deny the possibility that Mrs. GA accepts the request to obtain more and con- tinuous orders in the future. 158 Toward the Riural-Based Development of Commerce and Industry 20. His statement referred to the Social Security Act of July 1980. Establishments with more than twenty employees are now required to contribute to the Workers' Compensation Fund. In mid-1993 the act was extended to firms with ten or more employees. References Apichat, Chamaratrithirong, Philip Morgan, and Ronald R. Rindfuss. 1986. When to Marry and Where to Live? A Sttudy of Postnuptial Residence and Age of Marriage among Thai Women. Nakorn Pathom (Thailand): Mahidol University, Institute for Population and Social Research. Akerlof, George. 1970. "The Market for 'Lemons': Quality Uncertainty and the Market Mechanism." Quarterly Joutrnial of Econiomiiics 84 (August): 488-500. Akin, Rabibhadana. 1975. "Clientship and Class Structure in the Early Bangkok Period." In G. Wil- liam Skinner and A. Thomas Kirsch, eds., Chaniges and Persistence in Thai Society. Ithaca, N.Y.: Cornell University Press. Amara, Pongsapich, and others. 1989. Women Honmeworkers in Thailand. Bangkok: Chulalongkom University, Social Research Institute. Angkarb, P. Korsiepom. 1990. "Rural Women and Industrial Homeworking in Thailand." Paper presented to the SID East Asia and Pacific Regional Conference, Jakarta, 18-20 July 1990. Arrow, Kenneth J. 1974. Limits of Organization. New York, N.Y.: Norton. Baron, James N. 1988. "The Employment Relation as a Social Relation." Joturnal of Japanese an d In ter- national Econiomics 2 (4): 492-525. Benja, Jirapatpimol. Forthcoming. "Women and Industrial Home-based Work." Jotintal of Social Re- search. Bowie, Katherine A. 1992 "Unraveling the Myth of the Subsistence Economy: Textile Production in the Nineteenth-Century Northern Thailand." Jornial of Asian Sttudies 51 (4): 797-823. Casson, Mark. 1991. The Economics of Butsiness Culture: Game Theory, Transactiont Costs, and Econonmic Performiance. Oxford, U.K.: Clarendon. Clark, Gregory. 1994. "Factory Discipline." Joutrnal of Economic History 54 (1): 128-63. Feldman, Daniel C. 1984. "The Development and Enforcement of Group Norms." Acadenmy of Man- agemnenit Reviewv 9 (1): 47-53. Geertz, Clifford. 1978. "The Bazaar Economy: Information and Search in Peasant Marketing." Ameri- can7 Economic Review 68 (2): 280-320. Hayami, Yujiro, and Masao Kikuchi. 1980. Asian Village Economy at the Crossroads. Tokyo: University of Tokyo Press. Karim, Jahan Wazir, ed. 1995. "Male" and "Female" in Developinig Souttheast Asia. Oxford, U.K.: Berg. Landes, David S. 1969. The Unbotunid Prometheuts: Techlmological Chanige and indtustrial Developnment in Western Eutropefrom 1750 to the Present. Cambridge, U.K.: Cambridge University Press. . 1986. "What Do Bosses Really Do?" Journal of Econonmic History 46 (3): 586-623. Leibenstein, Harvey. 1987. Inside the Firm: The Inefficiencies of Hierarchiy. Cambridge, Mass.: Harvard University Press. Mead, Donald C. 1981. Suibcontractinig in Rural Areas of Thailanid. Bangkok: Kasetsart University, Center for Applied Economic Research. Mendels, Franklin. 1972. "Proto-industrialization: The First Phase of the Industrialization Process." Jouirnal of Economic History 32 (March): 241-61. The Rlral Garnment and Weaving Induistries in Northern Thailand 159 Naruemol, Bunjongit. 1993. "Small Entrepreneurs and Legal Framework in Thailand." Paper pre- sented at the Seminar on Entrepreneurship and Socioeconomic Transformation in Thailand and Southeast Asia, Chulalongkorn University, Bangkok, 2-4 February 1993. Ohno, Akihiko. 1994. "Organizational Maladaptation of the First Generation of Factory Workers- A Case Study in Northern Thailand." Osaka City Univzersity Economic Review 33 (October): 57-76. Ouchi, William G. 1980. "Markets, Bureaucracies, and Clans." Adminiistrative Science Quiarterly 25 (March): 127-41. Smith, Thomas C. 1988. Native Souirces of Japanese Indutstrializationi, 1850-1920. Berkeley: University of California Press. Sungidh, Piriyarangsan, and Kanchada Poonpanich. 1994. "Labour Institutions in an Export-Ori- ented Country: A Case Study of Thailand." In Gerry Rodgers, ed., Workers, Institutions and Eco- nomic Growthl in Asia. Geneva: International Institute for Labour Studies. Suntree, Komin. 1990. Psychology of the Thai People: Valtes and Behavioral Patterns. Bangkok: National Institute of Development Administration. Supat, Suphachalasai. 1994. Tlhailanid's Clothing and Textile Exports. Bangkok: Institute of Southeast Asian Studies. Thailand, Ministry of Industry, Provincial Industry Office. 1995. "List of Factories in Chiang Mai." Bangkok. Thailand National Commission on Women's Affairs, Office of the Prime Minister. 1980. "Aspects of Thai Women Today." Paper presented to World Conference of the United Nations Decade for Women, Copenhagen, 14-30 July 1980. Thompson, E. P. 1967. "Time, Work-Discipline, and Industrial Capitalism." Past and Present 38: 56- 96. Voravidh, Charoenlost. 1993. Infornmal Sector: The Econonmy of the Poor Entrepreneur. Bangkok: Chulalongkorn University. Walter, C. Neale. 1984. "The Role of the Broker in Rural India." In Peter Robb, ed., Rural Souith Asia: Linkages, Change and Development. London: Curzon. Wardell, Mark. 1992. "Changing Organizational Forms: From the Bottom Up." In M. Reed and M. Hughes, eds., Rethinking Organization: New Directions in Organization Theory and Analysis. London: Sage. Yanagisawa, Haruka. 1993. "The Handloom Industry and Its Market Structure: The Case of the Madras Presidency in the First Half of the Twentieth Century." Indian Economic and Social History Review 30 (1): 1-27. I 7 Township-Village Enterprises in the Garment Sector of China Deqiang Liu and Keijiro Otsuka S-1 ince the inception of economic reform in 1978, the Chinese economy has grown rapidly, with an average annual growth rate of about 9 percent. While the efficiency of the state sector as a whole improved during the reform process, a far more impor- tant contribution to industrial development came from the development of rural industry, widely known as township-village enterprises (TVEs). In their recent literature survey, Chen, Jefferson, and Singh (1992) conclude that TVEs have been a leading sector of the Chinese economy.' The production share of TVEs in the manufacturing sector increased from 9 percent in 1979 to 33 percent in 1993. Naughton (1992) argues that the entry of TVEs into industrial sectors has destroyed the monopolistic industrial structure domi- nated by state enterprises (SEs), thereby contributing to the overall efficiency of the economy. TVEs are not necessarily new; some were established during the prereform period by people's communes or brigades to serve agricultural production and peasant life in rural areas. They were subject to regulation, including restrictions on the production of com- modities produced by SEs. Those regulations, however, were relaxed over time, particu- larly after 1984 when the government announced policies to abolish unfair regulations against TVEs. Thus, TVEs have gradually entered into new markets to capture profitable business opportunities (Byrd and Lin 1990; Findlay and Watson 1992). TVEs tend to spe- cialize in the production of labor-intensive commodities such as textiles and garments, where the performance of TVEs has far outweighed that of SEs because of the ample sup- ply of labor in rural areas created by the restrictions on rural-to-urban migration (Otsuka, Liu, and Murakami 1998). These industries are labor intensive, for which China ought to have comparative advantages, but they were suppressed by the heavy-industry-oriented policies of the socialist regime. These observations suggest that one of the factors underly- 161 162 To,tard the Rutral-Based Developmenit of Commtnterce anid Ilndutstry ing the success of rural industrialization in China has been the effective mobilization of the cheap labor force in rural areas. TVEs, however, did not have good access to technological knowledge, management know-how, markets, and financial resources compared with SEs, particularly in the early stage of their development. Nonetheless, according to a production function analysis of the garment industry in China by Murakami, Liu, and Otsuka (1994), TVEs are more effi- cient than SEs, both technically and allocatively. How have TVEs acquired technical skill and management know-how and developed marketing channels? What sorts of individu- als have played the role of entrepreneurs, introducing new knowledge, organizing the new industrial activities in rural areas, and coordinating marketing activities? This study extends the analysis of Murakami, Lui, and Otsuka by focusing clearly on the mechanisms by which TVEs have achieved rapid growth despite obvious handicaps. Several different kinds of enterprises coexist in this industry, including TVEs, SEs, and joint ventures (JVs); JVs have been established mostly with firms from Hong Kong, China.2 Furthermore, TVEs can be subclassified into ordinary or independent enterprises (ITVEs) and those managed in cooperation with SEs or state trading enterprises (STEs), which may be designated as CTVEs. ITVEs actively engage in subcontracts with SEs under piece- rate arrangements, in which the latter assign raw materials and sell the final products through their own marketing channels.3 CTVEs and SEs or STEs maintain subcontracts through domestic cooperation, an institutional arrangement in which parent enterprises send managers and engineers to TVEs and share profits in accordance with their respec- tive investment shares. CTVEs are also called "branch factories" of the parent enterprises.4 JVs are similar to CTVEs in the manner in which profits and investment costs are shared with parent enterprises. Compared with CTVEs, however, the management of JVs is some- what more independent from parent enterprises. Although both TVEs and SEs have estab- lished JVs, the former case is the focus in this study.5 An important question is why TVEs engage in various contractual and institutional arrangements with SEs, STEs, and foreign enterprises. In practice, the latter group pro- vides technology and management know-how to the TVEs. The central process of gar- ment production is sewing, which is technically simple and requires no complicated machine process.6 The major management problem is the supervision of female workers, who usually work in groups of ten to twenty under a relay system of sewing operations. It is therefore doubtful that the provision of technology and know-how constitutes the major reason for the prevalence of subcontracts and the emergence of interenterprise coopera- tion. It seems more reasonable to postulate that it is essentially the development of stable markets for TVEs that has prompted the emergence of the contractual and institutional relationships. As Hayami argues in Chapter 1 of this volume, the development of rural industries must be supported by the network of commerce and other related services. This must be particularly true in China, where state sectors (state manufacturing enterprises, state trading enterprises, and state-run retailers) still dominate commerce and trading, . and the private marketing sector is grossly underdeveloped, except in small retailing busi- nesses in urban areas. We hypothesize that subcontracts and domestic cooperation are designed to utilize the traditional marketing channels controlled by the state sectors. We Towvniship-Village Enterprises in the Garmnit Sector of Chin a 163 also hypothesize that the establishment of JVs represents an attempt to circumvent the traditional marketing systems and to exploit new trading opportunities in export markets. While there is a possibility that the emergence of the potential threat represented by the entry of JVs is the reason that the monopolistic SEs sought subcontracts, the introduction of profit motives to the management of SEs seems to have been a more decisive factor in the establishment of CTVEs. Indeed, as will be shown later, CTVEs were generally estab- lished earlier than JVs. Development of Rural Enterprises and the Garment Sector in China Before proceeding to an analysis of garment enterprises based on our own survey data, it will be useful to characterize the importance of the garment sector in China's overall economy in the production and export shares it represents, as well as the importance of TVEs compared with SEs, by using official statistics. The official statistics on TVEs by industry group, however, are limited. According to the China Statistical Yearbook (State Sta- tistical Bureau 1987, 1994), the total number of employees in the TVE sector as a whole increased from 28 million in 1978 to 51 million in 1992, while the real value of output increased twelvefold during the same period.7 Disaggregated data by industry are avail- able only after the mid-1980s. Table 7-1 shows the nominal value of production by major industry, the production share in total production of the manufacturing sector, and the production share of TVEs in Table 7-1. Value of Production, Production Share in Total Manufacturing Sector, and Share of Production of TVEs by Major Industry, 1986 and 1993 1986 1993 Valiie of Prod uction Slhare Value of Prodzuction Shlare produiction share of TVEs production shtare of TVEs Induistry (billion yen) (percent) (percent) (billion yen) (percent) (percent) Metal products 33 3.5 36.1 191 4.3 56.8 Building materials 67 7.1 47.3 337 7.6 55.5 Garment 21 2.2 28.0 133 3.0 51.8 Textiles 125 13.1 18.5 416 9.4 41.9 Ordinary machinery 110 11.6 17.9 428 9.6 36.7 Food 77 8.0 15.4 299 6.7 35.2 Electric machinery 43 4.5 15.1 212 4.8 32.3 Chemical 67 7.0 10.1 274 6.2 28.4 Ferrous metals 67 7.1 5.5 433 9.7 17.7 Electronics and 25 2.6 6.8 139 3.1 13.5 telecommunications Transport equipment 36 3.7 6.8 273 6.2 13.1 Petroleum 30 3.1 0.0 150 3.4 6.3 Total manufacturing 953 100.0 17.6 4,441 100.0 32.8 Source: State Statistical Bureau (1987, 1994). 164 Towtiard the Ruiral-Based Development of Commerce and Iidinstril each industry in 1986 and 1993. Industries are arranged in descending order by the pro- duction share of TVEs in 1993. The production shares of TVEs were particularly large in the metal product, building material, garment, and textile industries. Note that metal prod- ucts include machinery parts supplied from TVEs to SEs under subcontracts. It is clear that industries in which TVEs have large production shares are those that produce simple products with a large application of labor inputs. In contrast, the production shares of TVEs were low in ferrous metal, electronics and telecomnmunication, transport equipment, and petroleum. This group requires large capital investments and modem technologies. In other words, TVEs tend to specialize in the production of commodities that require inten- sive use of labor and extensive use of modern technologies. These are the industries that were given low priorities in the heavy-industry-oriented development strategies of the socialist regime, despite the comparative advantages China ought to possess in these in- dustries given the abundance of unskilled labor. As Naughton (1992) argues, TVEs have entered underdeveloped industries where profitabilities were high. Overall, the production share of TVEs has increased, and growth in the production share of TVEs was particularly impressive in the garment industry. It seems clear that the gar- ment sector is one of the major industries in which TVEs have comparative advantage over SEs. The comparative advantage of TVEs over SEs in the production of garment products can be more clearly seen in the export share (see table 7-2). As may be expected, major export items are products of light industries, including garments. Table 7-2 demonstrates that TVEs are highly export-oriented. The export shares for garment products are larger than the production share in garment industries, which ranged from 2.2 percent in 1986 to Table 7-2. Value of Export, Share in Total Exports, and Share of Exports by TVEs by Major Commodity Group, 1988 and 1991 1988 1991 Value of Export Sliare Valute of Export Share export share of TVEs export share of TVEs Export (billion yen) (percent) (percent) (billion yen) (percent) (percent) Garment 6.1 4.3 49.6 12.7 5.6 77.5 Food 21.0 14.9 12.9 n.a. n.a. n.a. Textile 26.2 18.6 16.4 42.8 19.0 25.0 Chemical 6.4 4.5 23.6 10.3 4.6 31.5 Machinery 6.9 4.9 16.2 15.9 7.0 23.8 Crafts and arts 9.7 6.9 39.0 15.7 7.0 52.0 Silk fabrics 6.1 4.3 21.4 10.1 4.5 33.7 Total export 140.7 100.0 19.1 225.6 100.0 29.7 n.a. Not available. Source: Township and Village Enterprise Yearbook Editorial Board (1989,1992). To7vnship-Village Enterprises in the Garnment Sector of China 165 3.0 percent in 1993. The ratio of exports to domestic production of garment products was about 20-25 percent in 1991-92. Thus, we may characterize the garment industry in China as a fast-growing industry because the development of TVEs is oriented toward exports. Changing Organizational Structures of TVEs in the Garment Industry TVEs have developed with particular vigor in the suburban areas of major cities in coastal areas. Although these areas are often within the jurisdiction of large cities, they used to be, and many still are, agriculturally based. It is therefore legitimate to call many of these areas rural. Furthermore, we will argue that such a concentration of TVEs is a consequence of the underdeveloped marketing sectors in China. While private enterprises have flourished in some areas, such as Zhejiang Province, they are generally very small, with an average of four to five employees. They seem to engage primarily in small-scale service sectors, such as local commerce, transportation, and construction activities. In the manufacturing sectors, it is TVEs run by townships and village governments that have led to rapid rural industrialization. In this study, we focus on TVEs located in the suburbs of Shanghai and Beijing where rural industrialization has successfully taken place. Samplintg Framework Resurvey is an effective method to obtain accurate information and to correct erroneous responses, and we originally intended to resurvey the TVEs surveyed by Murakami, Liu, and Otsuka (1994) in 1990. They sampled thirty TVEs in Shanghai and twenty in Beijing. In order to compare the productivity performance of different kinds of enterprises, they randomly selected enterprises that produced, at least in part, men's shirts and women's blouses.' These two products are relatively easy to manufacture and are homogeneous from the viewpoint of the production process. Neither JVs nor CTVEs were intentionally selected. Therefore, their sampling is unbiased with respect to the selection of ITVEs, CTVEs, and JVs. In this study, we had to drop nine of the originally sampled enterprises in Beijing and five in Shanghai because they did not reply to our inquiries. In order to restore a reason- able sample size in Beijing, we randomly added ten samples. Thus, our present study is based on sample data collected from twenty-five TVEs in Shanghai and twenty-one in Beijing. Since our sampling is not random in the rigorous sense of the word, the sample size is small, and only two locations were selected, the possibility of selection and other biases cannot be ruled out. At the same time, implementing field research in China is difficult, and thus very rare. Our research is expected to make a unique contribution to the understanding of the mechanisms of growth in TVEs. We conducted a very informal survey in March and April 1994, which was followed by a formal survey carried out by local government authorities through a questionnaire we 166 Tow7ard the Rutral-Based Development of Contnterce and Indulstry prepared based on the informal survey. Questions addressed the years in which the enter- prises, domestic cooperation, and joint ventures were established, and, most recently, the years 1990-93. The number of sample enterprises in 1993 is shown by type in the last column of table 7- 3. We did not distinguish between township- and village-run enterprises, because we did not observe any significant structural differences between them. Sixteen, or about one- third, of the sample enterprises were ITVEs; only nine were CTVEs. It must be noted, however, that seven CTVEs had established JVs with foreign enterprises between 1990 and 1993. Eight out of nine CTVEs were located in Shanghai. This large concentration of CTVEs reflects Shanghai's status as China's major traditional industrial center, and there are many SEs seeking cooperative management with TVEs. In contrast, Beijing is a newly industrializing area with relatively few existing SEs. The largest number of TVEs were found to be JVs in 1993. With respect to the relative distribution of JVs and ITVEs, there was no marked difference between Shanghai and Beijing. Strtuctutral Evolution of Township-Village Enterprises All kinds of TVEs were widely engaged in subcontracts with SEs and STEs under piece- rate arrangements. SEs are essentially integrated firms that carry out not only manufac- turing but also marketing, which reflects the legacy of the socialist planning system in China. Under the socialist system, STEs were monopolistic traders that regulated imports of high-quality raw materials and exports of garment products. SEs and STEs have been close partners in international trade. SEs as well as STEs allocate raw materials to TVEs and sell finished garment products through their own marketing channels. In contrast, we did not observe any subcontracting from TVEs to smaller production units, such as small workshops and rural households, which is often observed in other countries. Unlike many other countries, women in rural areas of China regularly work throughout the year, so that the utilization of women's time during agricultural off-sea- sons will not provide strong motivation for establishing subcontracts between TVEs and homeworkers. According to several TVE managers we interviewed, some TVEs had made subcontracts in the past for such activities as sewing on buttons and brushing off dirt, but stopped such subcontracting because of the difficulty of controlling the quality of work and enforcing the delivery date. We conjecture, however, that a more fundamental prob- lem is the high marketing costs of securing raw materials and selling final products. In the absence of an efficient private trading network, TVEs must contact and negotiate directly with material suppliers and retailers, which is costly. Before proceeding to an examination of the nature of the interenterprise contractual relationships and their economic significance, let us clarify two important characteristics of garment enterprises in rural China: the changing composition of enterprise categories over time and the large size of enterprises. As the first column of table 7-3 shows, among the forty-six sample enterprises, only nine existed in 1975, and all of them were ITVEs. The number of TVEs increased gradually with Towntship-Village Enterprises in the Garnment Sector of China 167 Table 7-3. Changes in the Number of Sample Enterprises by Type, 1975-93 Enterprise 1975 1980 1985 1990 1993 Independent (ITVEs) 9 24 30 27 16 Cooperation (CTVEs) 0 2 7 16 9 Joint ventures (JVs)' 0 0 1 3 21 (1) (3) (7) Total 9 26 38 46 46 a. Numbers in parentheses refer to the joint ventures between foreign enterprises and CTVEs, which are regarded as JVs. Source: Authors' data. the reform of the economy. It is important to observe that the number of ITVEs increased until 1985, but then declined in the late 1980s and early 1990s. In contrast, the number of CTVEs increased markedly in the late 1980s but declined from sixteen to nine for the three- year period from 1990 to 1993, because seven CTVEs had initiated joint ventures with TVEs, SEs, and foreign enterprises. Why has the number of ITVEs and CTVEs declined since the mid-1980s and the early 1990s, respectively? According to table 7-4, the ITVEs in place in 1993 started operation, on average, in 1979, with seventy-two workers. Average size in the number of employees increased threefold, to 221, in 1993, but it was less than one-half that of CTVEs and JVs in the same year. They were also much smaller than others when the enterprises were estab- lished. These observations seem to indicate that enterprises that remained ITVEs tend to be less efficient-in production, in marketing, or in both-than the TVEs that have been reorganized as CTVEs and JVs. We will examine this point further when we analyze the production data. Table 7-4. AverageYear of Establishment of Enterprise, Domestic Cooperation, and Joint Ventures, and Changes in the Average Number of Employees by Type of Enterprise Average nml7ber of employees Average year established wlhen established Type Coopera- joint Coopera- Joint (as of 1993) Enterprise' tion venture Enterprise' tion venture 1993 ITVEs 1979 n.a. n.a. 72 n.a. n.a. 221 CTVEs 1976 1986 n.a. 182 556 n.a. 534 JVs 1980 n.a. 1992 138 n.a. 402 463 n.a. Not applicable. a. TVEs were all independent when established, except for five enterprises that launched operations as CTVEs. Source: Authors' data. 168 Towvard tlhe Ruiral-Based Development of Commerce anid Iniduistry The average CTVE in operation as of 1993 was established in 1976 as an ITVE; it initi- ated cooperation with SEs or STEs around 1986. During the period from 1976 to 1993, the number of employees increased threefold. The typical JV of 1993 was established around 1980 and became a JV very recently The average size of the enterprises in this category also increased by a factor of three over the thirteen-year period. The size of not only ITVEs but also CTVEs and JVs roughly tripled from the year of establishment to 1993, which suggests that TVEs commonly seek to expand enterprise size. According to our informal survey, a major advantage of JVs over CTVEs lies in the ease of obtaining export licenses under the preferential treatment JVs receive from the central government. Without such a license, CTVEs and ITVEs have to export products through STEs and pay commissions. It is thus distortionary government policies that provide a decisive advantage for JVs over other enterprises. There are several possible advantages of CTVEs and JVs over ITVEs in the transfer of technology and management know-how, as well as the acquisition of financial and mar- keting assistance from urban enterprises. Transfer of technology and management know- how, however, do not seem to offer a strong justification for domestic and international cooperation, because subcontracts are also designed to transfer technology and to sim- plify management on the part of the subcontractor, or TVEs in our context. Furthermore, it is possible for TVEs to recruit engineers and managers from SEs, even though labor mar- kets are imperfect. In principle, marketing constraints can also be also reduced by subcon- tracts. Yet subcontracts are essentially short-term contracts, subject to frequent changes. Marketing channels in China are institutionally limited because marketing is still domi- nated by monopolistic state manufacturing and trading enterprises, and it may well be that such long-term institutional arrangements as domestic and international cooperation have been devised in order to secure stable marketing channels. In the absence of an effi- cient marketing sector, an industrial structure that allows a small number of large enter- prises to cooperate may be more efficient than an alternative structure in which a large number of small enterprises operate, even if the organizational costs of an enterprise in- crease more rapidly than its size. We are assuming here that given the total volume of market transactions, the total cost of transactions is lower when there are fewer enter- prises, while the total cost of production tends to be lower when there is a larger number of enterprises. The optimum industrial structure clearly depends on how the costs of mar- ket transactions and production change with changes in the size and the number of enter- prises. Without efficient market transactions, the interenterprise division of labor is difficult, so that the vertically integrated enterprise operation becomes advantageous, as argued by Hayami (see Chapter 1 of this volume). In this connection, it is important to realize that the average size of garment enterprises is distinctly larger in China than in other countries. In our sample, the number of employ- ees in each enterprise varies, generally between 200 and 600, but there are TVEs that em- ploy more than 1,000. According to the chapters in this volume by Kikuchi (Chapter 5) and by Ohno and Jirapatpimol (Chapter 6), a size of fewer than ten employees is not uncom- mon in rural areas of the Philippines and Thailand, even though there are much larger Township-Village Enterprises in the Garment Sector of China 169 enterprises in large cities. These small enterprises are engaged in subcontracts with urban enterprises. In the case of China, subcontractors are large and do not create another layer of subcontracts with smaller enterprises. Furthermore, subcontractors are often subsidiar- ies of urban and foreign enterprises in China. These observations imply that the large size of enterprises and the prevalence of interenterprise cooperation are unique organizational features of rural industrialization in China. As such, they must be explained by factors unique to the Chinese economy. There are several possible explanations for the prevalence of large enterprises among our samples. First, men's shirts and women's blouses are simple, standardized products, so that large-scale production has a certain advantage. Second, entrepreneurs in rural ar- eas are scarce, so that the number of enterprises tends to be small. Third, the larger size of enterprises strengthens the bargaining position of TVEs compared with SEs and STEs. Fourth, given the preference for large enterprises of township and village governments, TVEs attempt to grow to gain favorable policy treatments. While we do not deny the im- portance of these factors, we believe that a more important reason for the large size of enterprises in China is the lack of an efficient marketing sector that can effectively orga- nize small production units through market transactions. Contractual and Institutional Arrangements Between Enterprises All of our sample TVEs are heavily dependent on SEs, STEs, or foreign enterprises in acquiring raw materials and selling garment products. Traditionally, SEs purchased in- puts, such as cloth, directly from domestic weaving enterprises, and sold products to do- mestic markets through state-run retailers, including the SEs' own shops, and to foreign markets through STEs (see figure 7-la). New marketing channels have been developed and grafted onto this traditional mar- keting structure as TVEs 'have developed. Table 7-5 presents the relative importance of state garment, state trading, and foreign enterprises as partners in subcontracts, both at the time the TVEs were established and in 1993. Note that when established, most TVEs were ITVEs, but five TVEs started as CTVEs, so that the numbers in the first line under the heading "year of establishment" show average data of forty-one ITVEs and five CTVEs. It must be also pointed out that CTVEs and JVs were engaged in subcontracts with both parent and nonparent enterprises. Major partners of ITVEs, both in the year of enterprise establishment and in 1993, were SEs, followed by STEs. In terms of figure 7-la, ITVEs acquired raw materials from SEs and delivered products to them under subcontracts, while relying on the traditional marketing channels of SEs. The whole range of transactions by ITVEs, however, was not covered by subcontracts. ITVEs also acquired raw materials directly from both SEs and TVEs and engaged in self- marketing final products to newly emerging private retailers, as the dotted lines in figure 7-la indicate. Given the lack of well-organized markets, however, such market transac- tions are bound to be costly. Furthermore, market transactions require large sums of work- 170 Towvard the Ruiral-Based Developnment of Comnmerce antd Indutstry Figure 7-1. Major Marketing Channels of TVEs (a) ITVEs - Flow of raw materials State-run b Flow of garment products j- Domestic State State Foreign material - garment - trading 4 enterprises suppliers enterprises enterprises and traders | Private 4-I > retailers … - - - - > ITVEs - - - - - - - - - - - - - - - - - - - - - _ (b) CTVEs and JVs State < Foreign trading enterprises enterprises and traders It Its CTVEs JVs ing capital because the purchase of raw materials occurs long before sale revenues are received. The requirement for working capital of TVEs can be lessened by subcontracts, because contractors purchase, or assist in the purchase of, raw materials. CTVEs have subcontracts primarily with STEs, directly as well as indirectly through the mediation of SEs, for the purchase of high-quality foreign raw materials and the export of Townshlip-Village Enterprises in the Garmenzt Sector of China 171 Table 7-5. Major Partners in Subcontracts,Year of Enterprise Establishment and in 1993 (percent) State State Foreign garment trading garment and (SEs) (STEs) trading enterprises Other Year established 63 35 0 2 1993 ITVEs 75 25 0 0 CTVEs 11 78 11 0 JVs 30 22 42 6 Source: Authors' data. their products. Parent companies of CTVEs are SEs in four out of nine cases, and STEs in the other five. That STEs are also parent enterprises of CTVEs indicates that TVEs do not necessarily seek the acquisition of technology and management capacity through domes- tic cooperation and that the establishment of marketing channels is an important motive behind subcontracting as well as domestic cooperation. JVs carry out direct transactions with foreign enterprises and traders using export licenses (see figure 7-lb for their major marketing channels). Note that both CTVEs and JVs maintain subcontracts with SEs to some extent (see table 7-5). Stubcontracts The subcontract is a piece-rate contract in which contractors (urban and foreign enter- prises) either commit to purchase or to assign specific raw materials to subcontractors (TVEs); specify designs, and usually provide technical assistance; set delivery dates; and make piece-rate payments in exchange for products finished by a specified date. In 1993 all types of TVEs were engaged in subcontracts with multiple enterprises (table 7-6); ITVEs appear to have relied greatly on subcontracts. The subcontract is essentially a short-term contract, which is often renewed after its successful completion. The average length of the most important contracts for ITVEs, however, was as long as 5.9 years, which was longer than the subcontracts made with other kinds of enterprises. It appears that Table 7-6. Average Number of Subcontracts per TVE and Average Duration of the Most Important Subcontract, by Type of Enterprise in 1993 Average dutrationi of the most Average nzumber importanzt subcontract Enterprise of sibcon tracts (years) ITVEs 8.3 5.9 CTVEs 6.6 3.3 JVs 7.4 2.7 Average 7.5 3.9 Source: Authors' data. 172 To7vard the Rutral-Based Developnment of Comnmerce and Indutstry ITVEs relied on subcontracts more heavily than CTVEs and JVs, both of which had secure, long-term trading partners under interenterprise cooperation. Table 7-7 shows the mode of raw material transactions between ITVEs and contractors and between TVEs and their parent enterprises under domestic cooperation and joint ven- tures. In China, subcontracts with "free" provision of raw materials or putting-out con- tracts were not common, particularly for ITVEs; more common was the purchase of specified raw materials by TVEs from designated sources, with the purchase of products guaran- teed in advance by the contractor, which may be called an advance-order contract. The main reason for the low incidence of putting-out contracts seems to be the weak financial position of SEs, which often run deficits. The free provision of raw materials in advance may also be risky for the principal if the agent is not trustworthy. In subcontracts between parent enterprise and TVEs, putting-out contracts were somewhat more common.9 It may well be that cooperation and joint ventures were ways to save in the transaction costs of marketing by establishing reliable, long-term trade partnerships. Piece rates are determined in advance, and the fees are usually paid with a single pay- ment within a specified period after the delivery of finished products. In this arrange- ment, contractors attempt to make payments after receiving the final revenue from sales. This was the kind of arrangement made in more than 70 percent of cases; in the rest of the cases, small advance payments were made or bonus payments were added upon success- ful fulfillment of all contractual terms. Technical and/or financial assistance was provided in about 60 percent of cases when the initial subcontracts were made, but the importance of such assistance has declined over time. The quality and quantity of final products are inspected by the contractors and by the parent enterprises. During the informal survey it was commonly reported by a large num- ber of TVE managers and engineers that quality controls had been effectively enforced. According to them, the rate of returned products, which fail to meet quality standards, has been almost zero. Thus, quality shirking, which is considered a major problem of piece- rate contracts, seems to have been effectively prevented."' Other opportunistic behavior, Table 7-7. Mode of Raw Material Transactions under Subcontracts at theYear of Enterprise Establishment and in 1993 (percent) Advance order Advance order contract Pultting-oult contract' uitlh partial paymentb contractc Year established 56 15 28 1990-93 ITVEs 68 11 20 CTVEs with parent enterprise 55 5 40 JVs with parent enterprise 53 10 38 a. Purchase of designated or assigned materials by TVEs. b. Purchase of assigned material with partial payment by TVEs. c. Free provisions by parent enterprises and/or contractors. Souerce: Authors' data. Tozvnship-Vi1lage Enterprises in the Garmnient Sector of China 173 Table 7-8. Family and Friend Ties and Other Prior Relations Leading to Establishment of Subcontracts, Domestic Cooperation, and JointVentures (percent) Other Internmediation Family anrd Prior suib- business by local Relationslhip frienid ties conitracts relations government Subcontracts, year established 55 n.a. 30 15 Subcontracts, ITVEs, 1990-93 36 n.a. 61 2 Domestic cooperation 33 21 14 32 joint venture 29 14 38 19 n.a. Not applicable. Source: Authors' data. such as concealed transfer of designated and assigned raw materials for other purposes, also seems uncommon. A critical element in the effective enforcement of subcontracts is the prior experience of contractors in manufacturing the ordered products, so that experienced engineers have accurate knowledge of the appropriate positioning of cloth for cutting-and, hence, exact cloth requirements-and the expected quality of final products if the assigned tasks have been accomplished properly. The threat of nonrenewal of current subcontracts and loss of social reputation seem to work against shirking and cheating, even though such activities were rarely reported. Competition among TVEs for subcontracts also seems to play a role in preventing oppor- tunistic behavior. Some TVE managers informed us that if urgent jobs were requested, they would operate their factories day and night to obtain more jobs in the future. A widely observed means of reducing the transaction costs associated with contractual relations and the internal management of firms in the less-developed countries is to use "F-connections," meaning ties with family and friends (Ben-Porath 1980; Hayami and Otsuka 1993). For example, managers of SEs may prefer to offer subcontracts to TVEs that employ relatives and close friends. Such personal or community ties were observed in 55 percent of subcontracts in the year of enterprise establishment (table 7-8). Over time, how- ever, a variety of business relations become increasingly important as reasons to enter into contractual relations. As in the case of putting-out subcontracts in weaving in prewar Ja- pan (Itoh and Tanimoto, Chapter 3 in this volume), community ties seem to have played a role in enforcing the terms of subcontracts in contemporary China. Donmestic Cooperation and Joint Ventutres I As in the case of ITVEs, CTVEs rely on subcontracts. The major differences between the two lie in the large commitment of parent enterprises (SEs or STEs) to the management of TVEs. A written contract is exchanged that specifies the period of cooperation, appoint- ment of managers and engineers, and investment and profit shares, among other things." 174 To7ard the Ruiral-Based Development of Comnierce and Indnstry The contract period in our sample is usually ten years. SEs often provide brand names and technology as a part of their investment requirements, in addition to funding and used machinery. In order to enforce the terms of contracts effectively, a parent enterprise sends a man- ager as well as engineers to the subcontractor. On average, three engineers were sent in 1993. The manager is usually provided with a high salary, although he will be fired if malfeasance is found to be practiced or enterprise performance is unexpectedly low. By and large, the parent enterprise takes charge of production management and marketing, while leaving the responsibility of labor management to the CTVE. Therefore, it is perti- nent to characterize the CTVE as a branch factory or a subsidiary of an SE or STE. While SEs, STEs, and TVEs seek domestic cooperation, final decisions are often made by SEs and STEs because there are fewer SEs and STEs than TVEs, which gives greater bar- gaining power to the former. SEs and STEs look for partners in domestic cooperation to obtain profits, because their management is subject to a variety of regulations, such as life- long employment, and thus their profit-seeking behavior is necessarily restricted. Success- ful subcontracts in the past, which indicate a high quality of work, sometimes led to domestic cooperation (see table 7-8). Local governments also assisted in the formation of CTVEs through intermediation. Since TVEs are ultimately owned by local governments and the policies of local governments influence the fate of TVEs, it must be that the support of CTVEs by local government provides incentives to establish new forms of production or- ganizations in rural areas of China. By far the most important reason for TVEs to establish domestic cooperation has been to improve market access (table 7-9). In contrast, technological, managerial, and financial reasons are of secondary importance. Since parent enterprises share profits, they have incentives to provide a stable flow of sewing jobs to CTVEs directly, as well as through the mediation of subcontracts between STEs and CTVEs (see table 7-5). In this way, the estab- lishment of CTVEs is expected to contribute to stable market transactions. Qualitatively, the contractual terms of JVs do not seem substantially different from those of CTVEs in investment and profit sharing, marketing arrangements, and production management. While there are cases in which parent enterprises send managers, it is more common for JVs to recruit managers locally. Like the parent enterprises of CTVEs, these Table 7-9. Major Reasons for Establishing Domestic Cooperation and International Joint Ventures (percent) Access Preferential Improz?e- Improve- Access to Access to export tax ment of ment of finanzcial Item to markets licenise treatmenit techinology maniagenienit resources Domestic cooperation 68 n.a. 0 11 5 16 Joint venture 27 33 16 4 2 19 n.a. Not applicable. Soturce: Authors' data. Township-Village Enterprises in the Garment Sector of Ciina 175 Table 7-10. Distribution of After-Tax Profits and Investment Shares among Local Government, TVEs, and Parent Enterprises, 1993 (percent) Recipients of profits Inivestmzent slares Local Local government TVEs Total Parent gozverinmelnt Parent Enterprise (A) (B) (A) + (B) enterprises and TVEs entterprises ITVEs 41 59 100 n.a. 100 n.a. CTVEs 36 35 71 29 57 43 JVs 30 39 69 31 60 40 n.a. Not applicable. Source: Authors' data. parent enterprises sent engineers to JVs (an average of about three in 1993) to ensure the quality of garment products. While ties to family and friends, prior putting-out contracts, and intermediation by local governments seem to play some role in the establishment of JVs, other business relations, such as open market transactions of raw materials and gar- ment products, were more important reasons for the initiation of JVs (table 7-8). According to table 7-9, the most important direct reason for the establishment of JVs was to obtain an export license; the second most important was improvement of market access. Export li- censes are readily granted to JVs as a part of the central government's policy to promote JVs. Thus, the establishment of JVs leads to the development of efficient marketing chan- nels for exports. To put it differently, the major objective of establishing JVs was to obtain export channels through direct partnerships with foreign enterprises, which was aided by preferential issuance of export licenses to JVs.'2 The distribution of after-tax profit and investment shares among the local government, TVE, and parent enterprise is shown in table 7-10. Profits were distributed in roughly equal portions to local governments and TVEs in the cases of CTVEs and JVs. Parent enter- prises received profit shares of about 30 percent in both cases. These profit shares were somewhat smaller than their investment shares, although we were told that profit sharing was proportional to the sharing of investment costs."3 The discrepancy between the two may be explained by the larger contributions by TVEs to management than originally planned. Marketing Channels and the Performance of Enterprises Based on the preceding discussion, it seems reasonable to postulate that while both sub- contracts and the two forms of interenterprise cooperation are designed to provide mar- keting channels for newly emerging TVEs, the cooperation provides more stable marketing channels. In the cooperation, the establishment of JVs seems to offer an advantage over CTVEs in access to export markets. If these hypotheses are valid, we expect to observe that 176 Towvard the Rriml-Based Development of Comm1erce and Induistry (1) reliance on self-marketing is highest for ITVEs; (2) greater resources are expended for marketinig by ITVEs; and (3) the observed production performance (for example, value added per worker) tends to be lower for ITVEs and higher for JVs. In this section we will examine these points. Clhanging Marketing Chainnels In the survey we failed to obtain comprehensive data on the quantitative importance of various marketing channels in the acquisition of raw materials and sales of final prod- ucts."4 Table 7-11 summarizes the frequency of the most important marketing channels reported by our sample enterprises in the year of enterprise establishment and in 1993 by type of enterprise. For this reason, the reliance on SEs and STEs may be underreported to the extent that they did not provide the major marketing channels for TVEs. When TVEs were established, they relied on self-purchase of raw materials in 40 per- cent of cases and on self-marketing output in 33 percent. Thus, subcontracts did not al- ways cover a major part of TVEs' transactions. The importance of subcontracts is reflected in the frequency of transactions with state garment and trading enterprises. An important observation here is that ITVEs continued to depend considerably on the self-acquisition of raw materials and self-marketing of final products in 1993. If input and product markets are underdeveloped and imperfect in China, as they certainly have been, ITVEs must have faced serious marketing problems. CTVEs acquired raw materials and sold final products almost exclusively through sub- contracts with state garment and trading enterprises. This indicates that domestic coopera- tion reduced the need to rely on self-acquisition and self-marketing. JVs depended greatly on foreign enterprises in marketing activities, yet they also depended on SEs and STEs, partly because some SEs and STEs also joined the establishment of JVs. More fundamentally, JVs seem to seek diverse marketing channels to stabilize the flow of market transactions. Larger resource expenditures by ITVEs for marketing activities are clearly indicated by table 7-12, which shows the number of specialized marketing personnel for each million Table 7-11. Major Marketing Channels at theYear of Enterprise Establishment (YOE) and in 1993 by Type of Enterprise (percent) Acqutired raw0 materials Produtcts marketinzg Foreign Foreign TVEs SEs STEs eniterprises TVEs SEs STEs eniterprises YOE 40 30 28 2 33 30 33 4 1993 ITVEs 47 20 33 0 31 25 38 0 CTVEs 11 40 44 4 0 44 56 0 JVs 19 21 22 38 5 19 24 52 Source: Authors' data. Townsihip-Village Enterprises in the Garmient Sector of China 177 Table 7-12. Marketing Personnel per MillionYuan of Gross Revenue and Working Capital, 1993 Marketing Workinrg capital (ratio to gross revenue) personn?el! Borrowed Sufppliedfrom Enterprise million yen Total Self-suipplied fronm banks parent enterprisesa ITVEs 1.04 52 23 28 0 CTVEs 0.31 49 9 18 22 JVs 0.34 37 12 18 7 a. Estimated as residual. Source: Authors' data. yuan of sales revenue and the ratio of working capital to sales revenue. The following observations can be made. First, the ratio of marketing personnel is significantly higher for ITVEs than CTVEs and JVs. Second, although the total working capital requirement for each unit of sales revenue is roughly similar across the three kinds of enterprises, the pro- portion of self-financed working capital is much larger for ITVEs than for others. The difference between the total working capital and self-financed capital essentially corre- sponds to the portion of working capital financed by parent enterprises and bank loans. Third, consistent with the above interpretation, the ratio of working capital borrowed from banks to sales revenue is higher for ITVEs. These observations are consistent with our hypothesis that domestic cooperation and joint ventures provide stable and efficient mar- ket channels for TVEs, over and above those made possible by reliance on subcontracts alone. Marketinig and Prodluctivity The overall efficiency of enterprise management will depend not only on technological and managerial efficiency, but also on efficiency in marketing, insofar as the latter affects purchase of raw materials, outlets for final products, and the timeliness of these transac- tions. Table 7-13 compares basic enterprise performance in gross value of production, la- Table 7-13. Comparison of Basic Performance by Type of Enterprise, 1993 Gross valtue Labor Capital-labor Export of produiction prodiuctivity' ratio' Wage rate ratio, Enterprise (million yen) (yen) (yen) (yen/year) (percent) ITVEs 11.5 6,919 4,968 3,019 42 CTVEs 21.8 7,831 3,651 3,705 88 JVs 31.9 10,365 4,776 3,671 75 a. Average value added per employee. b. Average value of machinery per employee. c. Value of export divided by gross value of production. Source: Authors' data. 178 Towvard the Riral-Based Deelopnment of Comnierce and Indistry bor productivity, average yearly wage, capital-labor ratio, and export ratio. Wage pay- ments are commonly made on a piece-rate basis. The labor productivity of ITVEs was found to be the lowest, despite the highest capital- labor ratio, in 1993, indicating that the overall production efficiency was also lowest for ITVEs. Presumably reflecting such inefficiency, the wage rate was also lowest among ITVEs. That the ratio of exported products was also lowest among ITVEs indicates the low quality of their products, because export markets usually require high-quality products. Thus, it is clear that the overall efficiency of enterprise management was lowest for ITVEs. This re- sult was anticipated, because ITVEs faced greater difficulties in marketing and received less technical and management assistance from SEs. It is difficult to compare the relative efficiency of CTVEs and JVs. First, since both labor productivity and capital-labor ratios were lower for CTVEs, overall production efficiency cannot be compared without knowledge of production function parameters. Second, the seemingly good performance of JVs may be apparent more. than real, because their access to export licenses may confer price advantages to JVs over CTVEs. According to the estimation results of the production function by Murakami, Liu, and Otsuka (1994), the total factor productivity is 30 to 4V percent less for ITVEs and 25 to 8 percent less for CTVEs compared with JVs. The difference in productivity between CTVEs and JVs, however, is statistically insignificant. Thus, the statistical analysis does not indi- cate the critical importance of export licenses dictated by the restrictive export policy in China. Characteristics of Rural Entrepreneurs The next question we would like to explore is who rural entrepreneurs have been in the context of rural industrialization in China. Rural entrepreneurs facilitate the introduction of new technologies, new products, and new market opportunities to rural areas, and they exploit the availability of the cheap rural labor force. The use of the cheap rural labor force is of particular significance in China, because labor migration to large cities has been re- stricted, which has artificially accentuated the dual structure of labor markets. Another important question is the mobility of factory managers and engineers from urban to rural enterprises. If rural enterprises can recruit managers and engineers from urban enterprises, they do not have to develop interenterprise cooperation and subcon- tracts to gain management know-how and technological knowledge. In that case, the es- tablishment of marketing channels is likely to be a major reason for establishing domestic cooperation and joint ventures. Characteristics of Managers The major rural entrepreneurs in the development of TVEs can be identified as enterprise managers, who have clear and strong decisionmaking authority. Table 7-14 demonstrates Township-Village Enterprises ini thze Garmtenit Sector of Chinia 179 Table 7-14. Personal Characteristics and Mode of Appointment of Enterprise Managers,Year of Establishment (YOE) and 1993 Previous occitpationl/enmployers (percent)l Appointed by (percent)< Local Parenrt Other Proportiont of Local Parent gozerni- eniter- eniter- Average higlh school govern- eniter- Farmier uient prises prisesb age graduiates nwenit prises YOE 22 41 11 24 41 22 87 11 1993 ITVEs 19 38 n.a. 44 44 50 94 n.a. CTVEs 22 0 44 33 49 33 22 78 JVs 5 29 29 38 41 67 33 57 n.a. Not applicable. a. There are a few cases classified as other categories. b. Includes TVEs for which managers worked before. c. There are a few cases classified as other methods of appointment. Souirce: Authors' data. the changing personal characteristics of TVE managers-from the years of enterprise estab- lishment (YOE) to the present, together with the changing mode of their appointment. At the time TVEs were established, their managers were mostly farmers or from the local government cadres. With few exceptions, they have been males. In cases where TVEs were established as CTVEs, managers were sent from parent enterprises in the year of establishment. In the majority of cases, however, managers were rural-based and had little experience in factory management during the early stage of development. The decisionmaking authority on enterprise management often rested in local governments, particularly in negotiations with external agencies. Many, however, were replaced by indi- viduals who had experience working in garment factories; that is, those who were re- cruited from parent and other enterprises. While the average age of managers did not change much between establishment and 1993, the ratio of those with high school educa- tions or above significantly increased. Thus, it is clear that entrepreneurs in the garment industry now are more experienced in garment production and more educated than be- fore. Over time they were given greater decisionmaking authority over internal manage- ment and larger claims on residual profits, whereas local governments gradually withdrew from internal management and assisted external management only occasionally (Chen and Rozelle 1995; Otsuka, Liu, and Murakami 1998). A number of China experts argue that TVEs are unlikely to pursue profits rationally because of excessive intervention in enterprise management by local governments (see, for example, Byrd and Lin 1990). For example, it is said that TVEs are required to protect and promote the welfare of resident workers even at the sacrifice of profits. To support such a view, these experts often allude to the appointment of TVE managers by local gov- ernments without due consideration of their qualifications. The appointment process, how- ever, has been changed significantly. It is true that 87 percent of the managers were appointed by local governments and 11 percent by parent enterprises when TVEs were 180 Towvard thte Rutral-Based Develo0pment of Comminierce anid Industrty established. But the cases of appointment by local governments declined to 50 percent, on average, in 1993. It must also be pointed out that even though local governments have appointed managers in the 1990s, in practice, many of them simply approved managers nominated by TVEs. It is also interesting to note that most managers were appointed by parent enterprises under domestic cooperation and joint ventures in 1993. This implies that many of the current managers of CTVEs and JVs have the management abilities re- quired by parent enterprises. In short, given the increasing pressure of market competition among TVEs, highly quali- fied managers seem to have been selected in recent years. Characteristics of Engineers The characteristics of not only enterprise managers but also of engineers have changed over time. Engineers may not be major entrepreneurs, but they often assist managers in the introduction of new machinery, new designs, and new materials. Table 7-15 compares the job experience and training of engineers between YOE and 1993. Note that we used multiple-choice questions among a group of engineers; thus, if even a single engineer had worked as a rural artisan, for example, the answer to the ques- tion of rural artisan as a previous occupation was yes for the group. According to this table, 31 percent of enterprises recruited rural artisans as engineers in YOE, whereas such cases became rare by 1993, except for ITVEs. The reliance on retired workers and the re- cruitment of engineers from other enterprises has been high for ITVEs, which indicates that ITVEs have acquired technical know-how through employment of experienced engi- neers. In contrast, engineers from the parent enterprises constituted the majority of engineers Table 7-15. Major Characteristics of Engineers in Sample Enterprises, Year of Establishment (YOE) and 1993 (percent) Previouts occupatioll/employersa Experienice Proportion7 Parenit Other of highi Rural Retired entter- enter- Vocatio7nal school Average Entterprise artisan wvorker prises prises schiooling OJT' graduiate age YOE 31 52 11 46 21 93 36 31 1993 ITVEs 25 56 n.a. 50 31 100 36 25 CTVEs 0 22 78 11 33 78 39 25 JVs 7 14 67 12 52 76 33 50 n.a. Not applicable. Note: This table is based on the responses to questions to a group of engineers. Thus, for example, the answer is yes if there is even one engineer who has been rural artisan before. a. There are a few cases classified as other categories. b. On-the-job training. Source: Authors' data. Townlship-Village Enterprises in the Garment Sector of China 181 in CTVEs and JVs. This last observation indicates that domestic cooperation and joint ventures facilitate technology transfer from parent enterprises to their subsidiaries. Engineers now have received more formal training in vocational schools. In the case of ITVEs, they have learned through on-the-job-training within TVEs. There is no question that the technological base of TVEs has been strengthened through the employment of more qualified engineers. Support Policies of Local Government In addition to intermediation in business partnerships, local governments actively pro- mote the development of TVEs through a variety of actions, such as the provision of infra- structure and guarantees for bank loans. While there is no denying that such policies contributed to the development of TVEs, it is difficult to establish causation because it is likely that local policies are endogenously determined. For example, local governments may want to promote more efficient enterprises because they are interested in their own revenues as owners of TVEs. Or they may want to assist TVEs in financial difficulty. In this section, we briefly examine the extent of implementation of various promotion policies for rural industrialization by enterprise category, while keeping in mind the ana- lytical difficulties mentioned above." Indutstrial Promiotioni Policies Table 7-16 exhibits the proportions of enterprises that were supported by industrial pro- motion policies from 1990 to 1993 by kinds of enterprises. Direct loans and subsidies have been infrequent in the 1990s. More important were tax exemptions, both full and partial. According to the Income Tax Law for Joint Ventures, which was issued in 1980 and amended in 1983, JVs are wholly exempt from income taxes for the first two years and are given 50 percent reductions in the subsequent three years. Similarly, CTVEs were given various tax reductions. But favorable treatment was much less frequently given to ITVEs. Thus, tax policies clearly favor the development of CTVEs and JVs. Table 7-16. Proportions of Enterprises Receiving Support from Industry Promotion Programs of Local Government, by Type of Enterprise, 1990-93 (percent) Suibsidized Direct Tax Guiaranteed Enterprise loan sutbsidies exemption banik loans ITVEs 0 0 50 63 CTVEs 2 13 88 88 JVs 5 19 76 67 Soutrce: Authors' data. 182 Towvard the Ruiral-Based Development of Commerce and Industry Also important were the guarantees of local governments for loans to TVEs from state banks. It is well known that local governments are influential in the allocation of bank loans, which are often determined by administrative considerations. It is a common belief that while budget constraints have become "hard" as govermnent subsidies have declined, the constraints on bank loans have become "soft," with loans easily available if local gov- ernments provide endorsements. In the frequency of local government provision of guar- antees for bank loans, CTVEs were favored.16 One possible interpretation is that local governments responded to the demand for investment funds by CTVEs because they played an important part in the establishment of CTVEs and expect high revenue from CTVEs in the future. By and large, however, local governments took some responsibility for the fi- nancial management of TVEs regardless of their relative efficiencies. Provision of lItfrastrtctutre In most of the areas where our sample enterprises are located, suburban industrial towns have developed through the rapid development of TVEs. In general, however, social in- frastructure, such as roads, telephones, and electricity, was inadequately supplied. Table 7-17 shows the proportions of enterprises that received improved infrastructure services from 1990 to 1993 by type of enterprise. This analysis is merely indicative; we did not collect detailed information about the intensity of the use of infrastructure services, nor did we know the quality of infrastructure services before 1990. For roads, telephones, and the supply of electricity, there were some tendencies for CTVEs and JVs to be favored over ITVEs. Local governments sometimes offer short training courses for engineers and workers in the garment industry and provide market information, particularly about popular com- modities and designs in foreign markets such as Japan. According to Table 7-17, CTVEs most often participated in training and information dissemination programs. Whether this is the choice of the governments or the enterprises, however, is difficult to ascertain. In sum, there is some indication that local governments attempted to promote CTVEs and JVs more than ITVEs, particularly in tax policy and infrastructure development. Be- cause CTVEs and JVs are more efficient than ITVEs, this selective promotion policy seems Table 7-17. Proportions of Enterprises Receiving Improved Infrastructure Services Provided by Local Government, by Type of Enterprise, 1990-93 (percent) Training and En7terprise Road Telephone Electricity informizationt ITVEs 31 50 19 19 CTVEs 63 63 13 50 JVs 57 90 38 29 Souirce: Authors' data. Tovwnship-Village Enterprises in the Garment Sector of China 183 consistent with the maximization of future revenue for local governments from the opera- tion of TVEs. Conclusions The major advantages of TVEs over SEs are management and market environments free from government regulation on employment and wage policies and the availability of cheap, unskilled labor. Budget constraints are much harder for TVEs than SEs, because local governments do not support unprofitable TVEs as much as the central government supports such SEs. What is lacking for TVEs is technical and management know-how, marketing capacity, and financial resources. We therefore hypothesize that the key to the development of TVEs is entrepreneurship leading to successful innovation in the sense of Schumpeter-that is, introduction of new products, new technologies, and better manage- ment methods to rural industries and creation of new marketing channels. Within the context of the institutional setting in China, it is SEs that possess the technical skills, man- agement know-how, and marketing channels. SEs are subject to regulation, including regu- lations on wage determination and dismissal of employees, that constrains them from seeking profits. Therefore, SEs have both the capacity and the incentive to play the role of entrepreneur in rural areas. Taking advantage of favorable policy environments for joint ventures, foreign enterprises also have interests in promoting rural industrialization. In this study, we obtained clear evidence that TVEs have acquired technology, manage- ment know-how, and marketing channels from urban enterprises through subcontracts, cooperation with SEs, and the establishment of joint ventures with foreign enterprises. In other words, the traditional socialist enterprises evade the regulations by engaging in vari- ous contractual arrangements with TVEs. In the process, entrepreneurs have changed from rural-based to urban-based, indicating that the entrepreneurship required for the manage- ment of TVEs is scarce in rural areas. We also found that the establishment of CTVEs and JVs was a more comprehensive and thorough method of transferring technology and man- agement skills, as well as strengthening marketing capacity, than subcontracts, which are essentially short-term contracts. ITVEs, which rely on short-term subcontracts, are much less productive than CTVEs and JVs. As a result, the number of ITVEs declined over time, as CTVEs and JVs rose. Among the different motives behind the prevalence of subcontracts and interenterprise cooperation, improved access to markets was identified as the most compelling. Kikuchi (Chapter 5 in this volume ) and Ohno and Jirapatpimol (Chapter 6 in this volume) found in the Philippines and Thailand, respectively, that in order to overcome marketing con- straints in rural areas, putting-out contracts were adopted. In these countries, however, interenterprise cooperation, as observed in China, was not developed. In China, market- ing channels are extremely limited, because marketing is still controlled by a relatively small number of state enterprises. Under such imperfect market conditions, reliance on short-term subcontracts did not assure stable transactions in inputs and outputs for TVEs. 184 Towvard thte Rutral-Based Development of Contmerce and Inditstry ITVEs had to engage in costly transactions in newly emerging, but still underdeveloped, free markets. In our view, the essence of both domestic cooperation and joint ventures is to assure long-term, stable contractual relationships between enterprises, which include put- ting-out and other forms of subcontracts. Another main feature of garment enterprises in China is their large size. One reason may be their proximity to major cities, which tend to be characterized by relatively high wages. Garment factories in large cities in other countries are also relatively large, even though they are usually smaller than TVEs in China. Of course, there is a possibility that TVEs do not seek profits because they are essentially "public enterprises." As we have seen, however, managers of TVEs have increasingly been appointed by the enterprises rather than by local governments. We also found that about 20 percent of the workers employed in our sample enterprises were from other areas, despite the common allegation that TVEs are constrained to employ only resident workers. In our view, the most plau- sible characterization of TVEs is that they are the "spin-off" of urban enterprises, which rely on urban-based managers, use high-speed sewing machines, and produce standard- ized products. The puzzle is why TVEs did not develop another layer of subcontracts with smaller rural enterprises and workshops, as is widely practiced in other countries. As Itoh and Tanimoto (Chapter 3 in this volume) demonstrate, the operation of small, family-operated weaving workshops in a rural area of prewar Japan was supported by a large number of local merchants, local wholesalers, and national-level wholesalers who were connected by community ties and other long-standing relationships. Similarly, in Indonesia, small rural enterprises that process soybeans are supported by the finely structured network of vil- lage-based, town-based, and urban-based merchants (Hayami and Kawagoe 1993). Fur- thermore, Levy (1990) argues that the generally smaller size of enterprises in Taiwan, China, than in the Republic of Korea in the footwear industry can be explained largely by the active participation of traders in a large number of small transactions in Taiwan. In rural China, private traders play no role. Trading is still dominated by large state enterprises, state trading enterprises, and state-run retail shops. This is a legacy of the socialist system of commodity transactions based on the material balance in which no role is given to free market transactions. It will probably take some time to develop efficient marketing net- works based on enduring personal relationships. Another constraint is the difficulty in establishing small private enterprises in rural areas because of the unclear protection of private properties in China (Chen and Rozelle 1995). If our view is correct, unless and until an efficient private marketing sector develops, including a competitive network of local traders, wholesalers, and urban merchants, China's rural industrialization will continue to be led by large TVEs. Such an industrial organization, however, will not be truly efficient if a major advantage of rural industri- alization lies in the low cost of labor management associated with the small size of an operation. As an efficient marketing sector develops, we expect to observe the develop- ment of subcontracts between large and small rural enterprises and between traders and small rural enterprises. Torwntshrip-Village Entterprises in tlte Garnient Sector of Chinza 185 Notes 1. See an earlier literature survey by Perkins (1988), which also ascribes the success of economic growth in China to the growth of TVEs. 2. Strictly speaking, there are also urban collective enterprises, which are another form of tradi- tional socialist enterprise. They are smaller than SEs, but similar to SEs in many respects. 3. In the garment industry in China, putting-out contracts in which the agent processes the material provided by the principal are uncommon. More common are cases in which the agent purchases the material and processes it for sale to the principal in accordance with contractual arrangements made in advance. We call the latter contract an "advance-order contract." Both of these contracts can be considered variants of the subcontract. 4. One may wonder if the relationship between CTVEs and their parent companies is similar to the relationship between part suppliers and the parent company of Toyoda discussed by Wada in Chapter 4. Although parent companies commonly provided technical assistance, there are more dissimilarities than similarities in the Toyoda case. While Toyoda attempted to develop long-stand- ing, tight personal relationships with inexperienced part suppliers, SEs looked for promising TVEs in the garment business as partners for a specified period, typically ten years. In our observation, many CTVEs wish to become independent of the SEs after the contractual period rather than to continue the business relationship for longer periods, as in the case of Toyoda. 5. It is found that the size of enterprises and the type of technologies employed are generally similar among JVs, regardless of whether they are joint ventures with SEs or with TVEs. 6. This is particularly true for such products as men's shirts, women's blouses, and uniforms, but not necessarily so for men's suits, women's dresses, and baby clothes. 7. Note that the numbers here refer strictly to township- and village-run enterprises, with the exclusion of private enterprises. 8. They attempted to select thirty TVEs in each location, but the sample size was actually smaller in Beijing, partly because of the limited number of TVEs producing men's shirts and women's blouses, and partly because of poor responses to their questionnaire. A total of fifty-two TVEs were engaged in the garment industry in Shanghai and thirty-six in Beijing. 9. It may well be that the relatively low incidence of "free" provision of raw materials reflected the effect of the stringent monetary policies of the early 1990s, which resulted in a shortage of working capital for SEs. 10. Workers are motivated to produce larger quantities of output without regard to quality under piece-rate contracts, as is demonstrated by Stiglitz (1975). 11. Written contracts were prepared for 57 percent of subcontracts in the year of enterprise estab- lishment. 12. Preferential tax treatment was also cited as another important reason for establishing JVs. In China, JVs are fully exempt from income tax payment for the first two years, and partly exempt (that is, 50 percent) for the subsequent three years. 13. Caution must be exercised in interpreting data in table 7-10, because the numbers were based on survey results not reported by all of our sample enterprises. 14. This is partially because TVEs do not have accurate information on the market values of raw materials and products, which are transacted under subcontracts. 15. In qualitative terms, the industry promotion policies of local governments in China are quite similar to the promotion policies for small- to medium-scale enterprises implemented by prefec- 186 To7vard the Ruiral-Based Development of Conmmerce and Inidutstry tural governments in Japan. It is also interesting to note that parent enterprises usually provide technical and financial assistance to subcontractors in Japan (see Itoh and Urata 1994). 16. In our survey we had little success obtaining information on the average amount of bank loans, which is an important limitation of the analysis. References Ben-Porath, Yoram. 1980. "The F-Connection: Families, Friends, Firms, and the Organization of Exchange." Popiulation and Devielopment Reviewv 6 (1): 1-30. Byrd, William A., and Qinsong Lin. 1990. ChiZna's Rural Indutstry: Struicture, Development, and Reform. Oxford, U.K.: Oxford University Press. Chen, H., and Scott Rozelle. 1995. "Local Leaders, Managers, and the Organization of Township and Village Enterprises in China." Food Research Institute, Stanford University, Palo Alto, Calif. Photocopy. Chen, Kang, Gary H. Jefferson, and Inderjit Singh. 1992. "Lesson's from China's Economic Re- form." Joiurnal of Com11parative Economics 16 (June): 201-25. Findlay, Christopher, and Andrew Watson. 1992. "Surrounding the Cities from Countryside: China's Rural Enterprises and Their Implications for Growth, Trade, and Economic Reform." In Ross Garnaut and Guogang Liu, eds., Ecoionomic Reforn and Implemiientationi: The China and Pacific Region. Sydney: Allen and Unwin. Hayami, Yujiro, and Toshihiko Kawagoe. 1993. The Agrarian Origins of Comm1lerce and Industry: A Stuidy of Peasant Marketinig in Indonesia. London: Macmillan. Hayami, Yujiro, and Keijiro Otsuka. 1993. The Economiics of Conitract Clhoice: An Agrariani Perspective. Oxford, U.K.: Clarendon. Itoh, Motoshige, and Shujiro Urata. 1994. "Small and Medium Enterprise Support Policies in Ja- pan." Faculty of Economics, University of Tokyo. Photocopy. Levy, Brian. 1990. "Transaction Costs, the Size of Firms, and Industrial Policy: Lessons from a Com- parative Case Study of the Footwear Industry in Korea and Taiwan." Jollrnlal of Developmenlt Eco- nomics 34 (November): 151-78. Murakarni, Naoki, Deqiang Liu, and Keijiro Otsuka. 1994. "Technical and Allocative Efficiency among Socialist Enterprises: The Case of the Garment Industry in China." Jornial of Comparative Econom- ics 19 (December): 410-33. Naughton, Barry. 1992. "Implications of the State Monopoly over Industry and its Relaxation." Modern Chlinla 14 (January): 14-41. Otsuka, Keijiro, Deqiang Liu, and Naoki Murakami. 1998. Inidustrial Reformn in Chinia: Past Perfor- mance and Futture Prospects. Oxford, U.K.: Oxford University Press. Perkins, Dwight H. 1988. "Reforming China's Economic System." louirnial of Econo1mic Literatutre 26 Uune): 601-45. State Statistical Bureau. 1987, 1994. China Statistical Yearbook. Beijing: China Statistics Publisher. Stiglitz, Joseph. 1975. "Incentive, Risk Sharing, and Information: Notes Towards a Theory of Hier- archy." Bell Journial of Econiomizics 6 (2): 552-79. Township and Village Enterprise Editorial Board. 1989, 1992. Chinese Yearbook of Township-Village Enterprises. Beijing: China Agricultural Publisher. 8 Rural Entrepreneurship and Industrial Development in Korea Jung-Hwan Lee and Chong-Hyuk Suh Until the early 1960s, the Republic of Korea had remained a typical preindustrial country; half its gross national product (GNP) was generated by the agricultural sector. A vigorous export-oriented industrialization policy begun in 1962, however, transformed the low-income agrarian economy into a middle-income industrialized economy. GNP has expanded at an average annual rate of 7 percent during the last thirty years; per capita GNP was more than US$7,000 in 1993 and was expected to reach US$10,000 in 1995. Although Korea has accomplished successful industrial development, the nation is far behind in rural industrial development compared with Japan and Taiwan, China. As a result, a large number of rural people have migrated to a few large cities; population has been declining rapidly in many rural areas while explosive population growth has oc- curred in a few cities. The growth of population was so fast in a few of the largest cities that urban infrastructure could not keep up with the expansion of the urban areas. Korean society could not help developing a number of urban problems, including housing, traffic, and environmental turmoil. The income disparity between farm and nonfarm households has been persistent. In 1990 the average farm household income was about 84 percent of its urban counterpart, while it was 112 percent in Japan (see table 8-1). Moreover, there is no prospect for improv- ing this gap with agricultural income in the near future. The growth rate of the farm sector is not anticipated to surpass that of the nonagricultural sector, and farm household in- come will be adversely affected by trade liberalization. Most agricultural products, in- cluding rice, cannot compete with imported foreign products. Income parity between farm and nonfarm households in Japan is partly attributed to the very large portion of farm household income that comes from off-farm activities. When 187 188 Towvard the Ruiral-Based Developmentt of Comnterce antd Indulstry Table 8-1. Income Disparities Between Farm and Nonfarm Households in Korea, Japan, and Taiwan (China), SelectedYears Houisehold income in Korea (1,000 wvon) Farm-nonofarnm disparity (percent) Year Urban (A) Farm (B) Korea (B/A) Japani Taiwan, Chinia 1965 113 112 99.1 99.2 94.8a 1970 381 256 67.2 110.1 72.2 1975 859 873 102.0 126.1 79.6 1980 3,205 2,693 84.0 118.0 74.2 1985 6,044 5,736 94.9 110.5 71.3 1990 13,184 11,026 83.6 111.6 66.4b 1992 19,416 14,505 75.8 n.a. n.a. n.a. Not available. a. 1966 data. b. 1989 data. Source: Korea: Economic Planning Board, Anntal Report of Urban Housel7old Survey; Ministry of Agriculture and Fishery, Report of Farm Household Econonoy Survey. Japan: Ministry of Agriculture and Fishery, Report of Fartt Houselhold Economy Survey, Office of Economic Planning, Annual Report of Household Survey. Taiwan, China: Taiwan Provincial Government, Survey Report of Houtselhold's Income and Expenditure, selected years. per capita GNP was about US$5,000 in current prices, the share of off-farm income out of total farm household income was about 38 percent in Korea, while it was about 60 percent and 58 percent in Japan and Taiwan, China, respectively (figure 8-1). This consideration leads us to the notion that off-farm income is very important for small Korean farmers and that the development of rural industry is critically important for their well-being in the future. This information raises a basic question: why has rural industry been so underdevel- oped in Korea, while Japan and Taiwan, China, have achieved successful rural industrial- ization in the process of industrial development? This issue has important policy implications for many developing countries as well as for Korea. The prevailing answers to this basic question in Korea can be summarized in two general assertions. The first is the notion that government support to rural enterprises was inadequate to promote rural in- dustrialization, and government support should have been substantially increased to pro- mote rural industry (Choe and Lee 1984; Kim 1987). The second notion is that the industrial policy of the Korean government was heavily biased toward large, vertically integrated enterprises and that small enterprises, including rural industry, could barely develop, in spite of government support (Kim and Whang 1987; Lee and others 1995). These assertions may partly answer the question, but they cannot be reconciled with the substantial effort of the Korean government to promote rural industrialization and the success of rural industrialization in some areas. It remains to be explained why govern- ment efforts did not yield good results and why industrialization was successful in some rural areas in spite of its overall failure. In order to find a coherent answer to the question raised above, macro and micro ap- proaches are taken in this study. In the macro analysis, the initial conditions for rural in- Ru.ral Eniterprenieurslhip and Inidistrial Developmient in Korea 189 Figure 8-1. Share of Off-Farm Income in Farm Household Income in Korea, Japan, and Taiwan (China), Selected Years Share of off-farm income (percent) 100 90 18 85 Japan 880 0-*..... '80 '89 70--- 60 7 7 8 o- - - * -I ~~~~~~~~~~~~'93 50 - '65 ~~~~~~~~~~~Korea 30- 20- '75 10 100 400 800 1,200 1,600 2,000 2,400 5,000 9,000 13,000 25,000 Per capita GNP (US$ current) Source: Suh and others (1991: 191). dustrialization, the government industrial development policies, and the condition of ru- ral areas for industrial development will be investigated, based mainly on secondary data and literature surveys. The micro analysis is based on a case study conducted in an area of successful rural industrialization. This analysis will focus on the factors that contributed to the success of rural industrialization in the study area. Historical Sketch of Rural Industrialization Policy in Korea The Korean government has made efforts to promote rural industry for decades. At the end of the 1960s, the government began to adopt rural industrial development policies to alleviate the regional disparity between urban and rural areas. In 1968 the Farm House- hold Side-Business Program was introduced to promote traditional rural manufacturing activities. It is also to be noted that the Agriculture and Fishery Development Corporation 190 Towuard the Rufral-Based Developnmenit of Conmnmerce and Itndutstry was founded in 1967 by the government to increase the processing of agricultural prod- ucts in rural areas. And in 1973 the Saemaul Factory Program (a program promoting small village factories) was launched to establish independent factories in rural areas. The farm households that participated in the Side-Business Program were encouraged to produce folk crafts, processed agricultural and livestock products, and unsophisticated rural household items such as brooms made of bush clover. About one-third were small- scale factories, while the rest were cottage handicraft operations. Almost all of the farm households participating in this program, however, suffered from poor sales promotion and lack of operating funds. The Agriculture and Fishery Development Corporation established twenty-three compa- nies for processing agricultural products in rural areas. Most of them suffered from poor management and marketing and were transferred to private companies by the mid-1970s. The Saemaul Factory Program was designed to stimulate the rural economy and to pro- vide more nonfarm job opportunities to farmers through the establishment of independent factories in rural areas. Preferential credits were extended to the Saemaul factories for in- vestment in plant and equipment, as well as to provide initial operating funds. Tax reduc- tions or exemptions were also offered to the factories. Six-hundred and eighty-six designated factories were active in 1984, when this program was terminated by the government. The program originally planned to establish more than one factory in each township, but less than half the townships received a Saemaul factory. Moreover, more than 50 percent of the factories were located in the vicinity of large cities. More important, many of the Saemaul factories were eventually shut down. In the end, the Saemaul Factory Program did not con- tribute much to creating nonfarm employment in rural areas (Suh and others 1991: 6). In the 1980s, rural industrialization policy shifted to emphasize the construction of rural industrial estates in order to provide needed infrastructure for rural enterprises. Rural industrialization was conceived as the most important farm-income policy under the on- going liberalization of the agricultural market, so that rural industrial estates were strongly emphasized by the government. In order to accelerate rural industrialization, the government enacted the Farm House- hold Income Source Development Act in 1983. Under the provisions of this act, industrial estates were to be built in the center of rural areas. The size of these estates was to range from 33,000 to 99,000 square meters. Government supports were provided to the firms or farm households participating in this program. Firms participating in the Rural Industrial Estate Program received direct subsidies for land acquisition, subsidized loans for factory construction and operation, and exemption from income and property taxes. During 1984-90, a total of 201 billion won (US$258 million) in direct subsidies were provided to the rural enterprises participating in the government programs. During the same period, about 837 billion won (US$1,074 million) in concessional policy loans was provided to participating firms or farm households. A large portion of these direct subsi- dies and concessional policy loans was allocated to the Rural Industrial Estates Program. About 94 percent of direct subsidies and 86 percent of concessional government loans were allocated to firms or entrepreneurs participating in this program. Rttral Entterprenetursitip and Indutstrial Developmentt in Korea 191 Table 8-2. The Number of Firms and Employees in Rural Industry Total Numntber of Nutnmber emtployees Slare of rural area Year offirmns (1,OOOs) Firms Employees 1955 8,810 n.a. 44.0 n.a. 1958 12,971 260 38.2 24.2 1970 24,114 861 32.5 17.2 1975 22,787 1,420 26.7 16.1 1980 30,823 2,015 28.7 20.8 1985 44,047 2,438 21.2 20.1 1990 68,872 3,020 23.3 21.0 1991 72,213 2,918 24.0 28.2 n.a. Not available. Souirce: Survey of Mining and Manufacturing Firms, Ministry of Commerce and Industry (1993). Although the Korean government has operated these rural industrial development pro- grams over the last twenty-five years, the results have been far short of what was in- tended. First, many rural industrial estates have suffered from underutilization of their capacity because few enterprises actually moved into the estates, and many of the enter- prises that did move in have gone bankrupt. Second, the number of rural firms has doubled during the past twenty years, but their share in the total number of enterprises has sub- stantially decreased. The percentage of rural firms out of the total number of firms was 33 percent in 1970, but about 24 percent in 1991, meaning that urban concentration has in- creased (table 8-2). Third, even in rural areas, manufacturing activities are not evenly dis- tributed. Most rural enterprises are located in the vicinity of large cities and industrial zones, mainly Seoul, Pusan, Taegu, and southeastern coastal areas. Macro Perspective on the Reasons for Unsuccessful Rural Industrialization There was little basis for entrepreneurship in rural areas in the initial stage of industrial- ization in Korea. Throughout its long history, Korea had been a centralized society in which local administrators appointed by the central government were not very interested in the development of infrastructure and industries in the areas of their mandate because their job security was not guaranteed. Historical records indicate that, on average, local admin- istrators served in a position for less than two years. Similarly, local entrepreneurs were not interested in developing specialized manufac- turing products because a special commodity tax was charged whenever a new commod- ity was developed (Kang 1984). At the same time, private entrepreneurs were not regarded as a respectable class in traditional Korean society. Such social customs continued until the 192 Tozvard thte Rutral-Based Developnment of Comnmerce anid Inidutstry early twentieth century. Thus, most educated and capable persons were not willing to engage in commercial and industrial activities. Moreover, during the thirty-six years of the Japanese colonial period (1910-45), most industrial facilities were developed in the northern part of Korea because of favorable endowments of mineral and hydraulic resources. Agriculture was developed in the south because of its favorable agricultural land endowment. Furthermore, the Korean War of 1950-53 thoroughly destroyed the industrial basis; few industrial facilities and little infra- structure remained after the war. For example, about 40 percent of paved roads were de- stroyed between 1947 and 1953 (as shown below). In addition, social as well as human relationships in rural communities, which are very important factors for the development of rural entrepreneurship, were seriously disrupted during the war. Because the initial conditions for industrialization were so poor, the Korean govern- ment took strong initiatives in industrial development by introducing a series of ambi- tious five-year economic development plans. To accelerate industrial development, from the beginning the industrial policy concentrated on strategic sectors and relatively advan- tageous areas-large enterprises, export-oriented and heavy industries, and a few urban areas (Choe and Kim 1986). For example, from the mid-1960s to the early 1980s, the gov- ernment developed industrial estates in a few local growth areas and strongly promoted new modern enterprises in big cities. Large amounts of concessional loans were directed to these enterprises. As a result, very few funds and little effort were available from the government to develop small rural enterprises. The strong bias of government's industrial policy toward new, modem enterprises in large cities triggered a massive out-migration from rural areas beginning in the mid-1960s. Until that time, the annual net off-farm migration rate was only 1.22 percent, as shown in table 8-3. The rate increased to 4.05 percent in the late 1960s, however, and to more than 5 percent in the late 1970s. Most migrants were young and educated. For example, the mi- gration rate was around 8-11 percent for those younger than thirty (table 8-4). It is also to be noted that the migration rate of young women was much higher than for their male counterparts. As a result, for industries located in rural areas, acquiring young and/or female labor became more difficult, and the advantage of cheap labor faded in rural areas. This change is partially revealed by the wage ratios between the rural and urban sectors in the 1970s and 1980s, as shown in table 8-5. In this table, urban wages are represented by the wages Table 8-3. Net Off-Farm Migration and Migration Rates 1960-66 1966-70 1970-75 1975-80 1980-94 Net off-farm migration (1,OOOs) 1,133 2,516 2,561 3,350 2,449 Migration rate (annual percentage) 1.22 4.05 3.60 5.39 5.95 Sozurce: Park (1989). Rural Enterpreneurship anid Indutstrial Developmenit in Korea 193 Table 8-4. Net Off-Farm Migration Rates by Age and Gender (annual percentage) 1966-70 1975-80 Age Male Fenmale Male Female 13-17 7.8 9.9 11.0 11.7 18-22 6.5 10.2 7.2 10.4 23-27 8.9 9.4 8.3 9.9 28-32 7.1 2.9 3.3 2.0 33-37 1.7 1.1 1.8 1.5 Source: Sloboda (1982). for workers in production and related sectors, the workers most in demand in small in- dustry. According to this table, in the 1970s and 1980s, male wages in rural areas were approximately 80 percent of urban male wages, and female wages were higher in rural areas than in the urban sector. The wage comparison in table 8-5, however, may overstate the actual rural-urban wage parity because rural wages are represented by the wages for daily workers, who are gen- erally hired during the peak seasons-that is, rice transplanting and harvesting periods. In real wages, the figures in table 8-5 may understate the rural-urban wage ratios because the cost of living in rural areas is likely to be lower than in urban areas. In spite of these reservations, we can say that very tight labor supply conditions have prevailed in rural areas. A survey revealed that 42 percent of rural entrepreneurs pointed to their difficulty in acquiring qualified labor as the most difficult problem they faced (Suh and others 1991:229-31). Table 8-5. Wage Ratio Between Rural and Urban Sector by Gender Year Male Female Average 1972 0.78 1.17 0.97 1975 0.79 1.16 0.98 1978 0.77 1.15 0.96 1981 0.90 1.37 1.13 1984 0.82 1.19 1.01 1987 0.74 0.99 0.86 1990 0.76 0.96 0.86 Note: Rural wage includes all payment of cash and in kind for farm workers. Urban wage denotes all pay- ment, including overtime and bonus for production and related workers. Source: Ministry of Agriculture and Fishery, Statistical Yearbook (annual). Ministry of Labor, Yearbook of Labor Statistics (annual). 194 Totvard the Rutral-Based Developmiient of Conmmerce anid Industry It should also to be noted that rural-urban wage parity does not necessarily mean in- come parity between farm and nonfarm households, because the working hours of family farm workers are limited by the seasonality of labor demand in farming, particularly rice farming. As mentioned earlier, the income difference between farm and nonfarm house- holds has been persistent. Another cause of unsuccessful rural industrialization can be seen in misdirected gov- ernment efforts to promote rural industrialization. First, government efforts were directed to the provision of direct incentives to enterprises, as mentioned in the previous section, and the importance of preconditions for industrial development in rural areas was ne- glected (see Chapter 9 in this volume by Lane). As a result, transportation facilities, electri- fication, and credit markets were barely developed in rural areas. For example, almost all provincial roads were unpaved until the late 1970s, as shown in table 8-6 and figure 8-2. The density of paved highway in Taiwan, China, was 76.4 kilome- ters per thousand square kilometers in 1962 and 214.5 in 1972, while in Korea it was only 10 kilometers in 1960 fewer than 50 kilometers in 1975. At the same time, only 12.6 percent of Korean farm households had electricity in 1964 and 61.1 percent by 1974, while in Tai- wan, China, electrification of farm households had reached 70 percent by 1960 (Ho 1982; Keidel 1982). Because of the poor infrastructure and unfavorable labor supply, many enterprises pre- ferred to locate in urban areas. Many kinds of pecuniary and nonpecuniary incentives provided by the government to newly established enterprises in rural areas had only mar- ginal influence on attracting entrepreneurs to rural areas. It was still risky, in spite of the incentives, to start a business in a rural area; the only advantage in rural areas was cheap land prices. Second, the rural industrialization policy emphasized industrialization in rural areas rather than industrialization of rural areas (Suh 1994). In other words, central as well as local governments have neglected the importance of rural entrepreneurship rooted in the community, instead emphasizing the relocation of transplanting urban entrepreneurs or urban enterprises to rural areas by providing direct incentives. All of these direct incen- Table 8-6. Length and Share of Paved Highway Lenzgthz of highwzvay (k.n) Share of paved wvay (percent) Year Total Paved Total Nationial way Provinicial 1936 24,283 538 2.2 4.8 0.4 1947 24,433 1,061 4.3 14.1 0.7 1953 26,033 611 2.3 5.1 0.3 1961 27,169 1,122 4.1 12.6 0.5 1967 34,799 2,092 6.0 17.6 0.5 1975 44,905 10,000 22.3 44.0 0.5 1980 46,951 15,599 33.2 67.4 12.6 1990 56,715 40,545 71.5 89.1 63.6 Source: Ministry of Construction, Statistical Yearbook, 1960 and thereafter. Rural Enterprenieturship and Indutstrial Development in Korea 195 Figure 8-2. Paved National Highway in the Republic of Korea Source: Keidel (1982) 1967 1975 , -' -n -, tI, AC 7 * Paved Highway _. Paved Highway IB123S e :'- Paved Highway,- > ._ hipeway tives were provided only to enterprises involved in the government programs. In order to be involved in the government programs, enterprises must pass the selection procedure of the government-nominated institute. It is likely that the selection procedure was biased toward external conditions, including funding capacity, because the government was un- aware of the importance of rural entrepreneurship rooted in the community. Most rural entrepreneurs were likely to be too small and too new to meet the criteria. As a result, the suffocation of rural entreprenfeurship was even greater under the policy to move urban enterprises into rural areas than it had been before government intervention. Meanwhile, transplanted enterprises faced serious difficulties in adapting to new and unfamiliar busi- ness environments and failed to take root. A Case of Successful Rural Industrialization in Korea To confirm the broad perspective presented above, we conducted a case study that fo- cused on the role of rural entrepreneurs and their relationship to industrial development in a rural area. The Punggi area, very remote but very successful in rural industrialization, was selected as the case study site. 196 Towvard the Rutral-Based Dezvelopment of Commerce and Industry An intensive method was adopted to collect data and information relating to historical changes in the production and marketing systems of Punggi's weaving industry. The re- searchers had many interviews with weavers, wholesalers, and officers working in the Taegu-Kyungbook Weavers' Cooperative and county offices from September 1994 to May 1995. Data on the rural industrial development policy program implemented at the local level were collected from YoungPoong, Bongwha County, and YoungJoo City offices. Punggi is a typical small, rural township located in YoungPoong County, in the northern part of Kyongsang Book-Do Province. This area does not have a locational advantage for industrial activities because it is mountainous and far from large cities (figure 8-3). Never- theless, this area is a very successful case of rural industrial development. As shown in table 8-7, there were 170 manufacturing enterprises in this county in 1994, which contrasts greatly with the neighboring areas of Bonwha County and YoungJoo City. YoungPoong County's success in rural industrial development is mainly attributed to the growth of the weaving industry in the Punggi area. Of the 170 small manufacturing factories, 145 are engaged in weaving synthetic cloth using rayon, nylon, and polyester yarn. The product is mainly used for lining cloth, and 80 percent of the total national production of lining cloth is produced in this area. In 1993 the total value of cloth production was about 53 billion won (US$66 million), which was equivalent to one-third of the total gross regional production of the county Figure 8-3. Map of Study Area KWANGWON DO EK= Il, O S i S ' X AS02 iS. 0S': 0' ------i- Raiilroads ': t::i: f: S Wx:f Bngwha O YourngJoo n taltEi parksV * Uisong "Study Area KYONGSANGBOOK - DO Seoul ,.ci VREPUBLIC Xu q < Koomi {i *;S; 0303 >t 7% | OF Pungg: Kim c S on Pohwang ChungJoo Youngc . KOREA Taegun Kyongju OK Area of map ,, * fTaegu 7 .-guszvKaT Ulsan . KYONGSANG NAM DO E3 Rural Eniterpreneuirslhip and Indutstrial Developnmenit in Korea 197 Table 8-7. Number of Manufacturing Firms in the Study Area, End of 1994 Couinty or city Clothl Food Other Total YoungPoong 145 15 10 170 Bongwha 1 1 21 23 YongJoo 5 6 31 47 Note: The numbers include only firms employing more than five workers. Source: Statistical Yearbook of each county. (table 8-8). About 1,000 local employees work in the weaving industry and receive about 600 million won (US$750,000) monthly in wages. The weaving industry contributes sub- stantially to the employment and nonfarm income of the rural households in this area. Socioecononzic Change and Developmient of the Produiction Systemn in the Punggi Weaving Indutstny The weaving industry in the Punggi area has a long history. About sixty years ago, a silk weaving factory was established by an immigrant from Pyungannam-Do Province, in the northeastern part of northern Korea. Before he came to Punggi, he had operated a weav- ing factory in his home town. In the late 1940s the weaving industry in the Punggi area began to produce rayon cloth instead of silk, and the number of weaving factories in- creased dramatically from 5 in 1945 to 120 in 1951, as shown in table 8-9. This expansion was attributed in part to the many immigrants from northern Korean who moved into the Punggi area and started weaving businesses. The expansion, however, was accelerated by the Korean War. Most weaving factories, as well as infrastructure in other regions, were destroyed during the war, while the Punggi area was untouched. Punggi rushed to fill domestic demand for rayon cloth. Punggi had good initial conditions compared with other areas because it received the input of new entrepreneurs and its industrial basis was preserved. Table 8-8. Quantity andValue of Weaving Cloth Produced in the Punggi Area, 1992-93 Qutanitity (1,000 yards) ValuIe (millionl tuon1) 1992 1993 B/A 1992 1993 B/A Cloft1h (A) (B) (percent) (A) (B) (percen1t) Total 95,700 105,300 110.0 49,764 52,650 105.8 Rayon 11,797 21,060 178.5 7,343 12,630 172.0 Acetate 31,128 21,060 67.6 16,361 9,900 60.5 Nylon 25,222 29,480 116.9 10,868 13,270 122.1 Polyester 27,553 33,700 122.3 15,192 16,850 110.9 Source: Taegu-Kyungbook Weaver's Cooperative (1994). 198 Tozvard the Rural-Based Developniet of Conmmerce and Indutstry Table 8-9. Number of Weaving Factories in Punggi Area, SelectedYears 1938 1945 1948 1951 1968 1985 1988 1994' Number 2 5 31 120 48 76 100 160 a. Includes firms employing fewer than five workers. Source: Taegu-Kyungbook Weaver's Industrial Cooperative (1994). Punggi's weaving industry faced its first depression at the end of the 1950s. By this time, most industrial facilities-including weaving facilities-in other areas had recovered from the war, and industrial production regained its prewar level. Moreover, rayon cloth pro- duced in the Punggi area was inferior in quality because weaving enterprises in urban areas had adopted electric weaving machinery, while those of Punggi still used handlooms. Therefore, Punggi weavers lost in the competition with urban weavers because of the low quality of their products and their high production costs. To make matters worse, nylon cloth began to be produced by the urban weaving industry. Demand for the rayon cloth produced in the Punggi area decreased rapidly because most customers substituted nylon for rayon cloth in garment production. Yet Punggi producers could not adopt elec- tric weaving machinery because there was no electricity in their area. They tried using gas engines as a source of power for weaving machinery, but these could not provide suffi- cient power to produce high-quality rayon as well as synthetic cloth. As a result, Punggi's weaving industry faced a serious depression. The situation changed at the end of the 1960s, when the government constructed an electric railroad through the Punggi area. With this electricity, electric looms could be in- troduced, and many factories began to use nylon and polyester yarn. To reduce the large amount of investment needed for machinery, they bought secondhand looms from urban weaving enterprises in Taegu, Taegeon, and Mokpo. Labor productivity increased as much as two to six times. Along with adopting electric looms, weavers also began to change their final products. Instead of rayon cloth for garments, they began to produce nylon and polyester cloth for lining garments and bags. It must be noted that electrification played a critical role in Punggi's weaving industry. Fortunately, weavers in Punggi could use bank loans instead of private credit to finance this new investment. This was made possible when a formal financial institution, the Citi- zens Bank, was established in YoungJoo City in 1963, and the government urged the bank to provide loans to weaving enterprises. This was a great opportunity for the enterprises because the annual interest rate for these loans was about 25 percent, while the rate for private loans was about 60 or 70 percent. The introduction of this credit institution was another contributing factor in the second expansion of Punggi's weaving industry. In 1986, many enterprises began to adopt labor-saving technologies because of a labor shortage. They tried to replace labor with new machinery. Air-jet and water-jet weaving machinery and a shuttle-change system were adopted to increase labor productivity. Also, Ruiral Enterpreneuirslhip and Induistrial Development in Korea 199 Table 8-10. Historical Changes of the Punggi Weaving Industry Machines Stage of induistrial operated Social development Period Produtction Technology by a wvorker situtation Starting stage 1934-47 Cottage Hand loom 1 Japanese colonial time First expansion 1948-59 Cottage Hand loom 1 Before and after stage Korean War First depression 1960-67 Cottage Hand loom 1 Introduction of stage nylon yam Second expansion 1968-85 Cottage and Electrical weaving 2 Adoption of stageq factory loom electric power (motor system) (4-6) Third expansion 1986-present Cottage and Air and water-jet 14-20 Spread of labor stage factory weaving loom union (shuttle change system) Source: Authors' data. many skilled laborers began to establish their own weaving enterprises with the aid of their previous employers. As a result, the number of weaving factories increased rapidly after 1986, although most have remained small-scale, cottage-style factories. Transformation from cottage to modem factories was made possible by the construction of rural industrial estates in this area. Based on the Farm Household Income Source Devel- opment Act, a rural industrial estate of 105,000 square meters was constructed in the Punggi area by the government in 1989. Twenty-four weaving enterprises moved into this indus- trial estate and were able to make the transition from cottage to modem factory operations. Development of Marketing Systems The development of a marketing system is essential to promote rural industrialization, especially in remote places such as the Punggi area. The weaving industry of the Punggi area has developed a unique marketing system, as shown in figure 8-4, which has contrib- uted substantially to the success of the area's weaving industry. The marketing system has developed in concert with changes in technology and market demand. As mentioned earlier, the weaving industry in other areas was destroyed by the Korean War, while Punggi's weaving factories were left intact. As a result, many local collectors and peddlers rushed to this area to purchase rayon cloth with cash. In other words, the market for rayon cloth was a seller's market and transactions were made in a kind of village spot market between the weavers and many buyers. At that time, rayon weavers had their own dyeing facilities and dyed plain cloth them- selves. Meanwhile, rayon yam was supplied to weavers by yarn dealers, and payment to 200 Torward the Rural-Based Development of Coninerce anid ilnditstry Figure 8-4. Marketing System of Punggi's Woven Cloth During the 1950s period yarn dyed cloth Rayon yarn Weaver > Local dealer < collector cash or cash in-kind cash cloth dyed cash ~~f ,j,jloth Tcashot I Peddler l Wholesaler/ retailer in large cities 1960-present (System for domestic market) Synthetic draft producer nylon , ~~verbal dyed Rayon yam Weaver order coWholesaler oth Garment importer < (principal) < (Seoul, Taegu) < maker draft I . draft draft order and yam drafJ Tcash order dyed I /oth >1 1 anddraftl cloth Weaver Bank plain Dyeing (subcontractor) cloth factory (System for export) | Dyeing | factory PAw?nC, dyed order cloth and draft yarn Garment maker > Weaver (exporter) ) (subcontractor) draft draft 4, cash Bank Rutral En terpreneurship anud Indutstrial Development in Korea 201 the yarn dealers was made in cash or in kind, mostly in woven cloth. In-kind payments were very common at that time because of the scarcity of cloth and high inflation. Nevertheless, this system could not continue indefinitely, because the market for rayon cloth became competitive following the development of many new enterprises in other areas, and the marketing system was changed in the 1960s. Two marketing systems have developed in the Punggi area: the "putting-out contract," in which the principal advances materials for processing to his agent and later collects the finished products at a piece-rate payment, and the "advance-order contract," in which the principal guarantees the pur- chase of the finished products but does not advance materials (see Chapter 1 by Hayami in this volume). About 48 percent of weaving manufacturers trade in advance-order contracts with whole- salers in Seoul, as shown in table 8-11. They have no written contracts. Instead, verbal orders by telephone are very common. Both parties-weaver and wholesaler-expect the transaction to continue to be repeated as long as contractual agreements are observed. A wholesaler places an order with a weaver for the delivery of plain cloth to a dyeing factory. At the same time, the wholesaler specifies a color and pattern to the dyeing factory for dyeing the plain cloth delivered by the weaver. Most dyeing factories are located in Ansan Industrial Estate near Seoul. The weaver delivers his product to the dyeing factory selected by the wholesaler and receives a certificate of delivery. Dyed cloth is delivered to the wholesaler by the dyer, and most of the dyed cloth is sold for lining to garment mak- ers. It is worthwhile to note that the dimension and quality of the cloth are rarely checked when the cloth is delivered to the dyeing factory and then to wholesaler. This reflects the close, long-term relationship and credibility established between them. Even without in- spection, very few claims are put forward by garment makers, and they don't think the time and effort involved in inspecting the cloth are worthwhile. In other words, wholesalers organize two stages of subcontracts to acquire the final product of lining cloth: the first is the advance-order contract with weavers, and the other is the putting-out contract with dyeing factories. It is to be noted that while weaving manu- facturers do not have dyeing facilities now, they did until the 1950s. There were two rea- sons for this change: (1) in the Punggi area, it is difficult to operate a dyeing factory because of government regulations on pollution in rural areas; and (2) wholesalers prefer to send the cloth to specialized dying factories because weavers do not have much information or Table 8-11. Number of Weaving Enterprises by Form of Contract, 1994 Formi Numnlber offirms Advance-order contract (A) 75 (47.8) Putting-out contract (B) 53 (33.8) Mixed (A + B) 29 (18.4) Total 157a (100) Note: Parentheses represent percentages. a. This figure excludes three enterprises that were not operated at the time of survey. Source: Authors' data. 202 Tozward the Rutral-Based Development of Conmnerce aid Indtstry sense of fashion in the important areas of color, technology, and quality, and thus it is not worthwhile for weavers to maintain dyeing facilities. As payment, weavers receive drafts payable after three months, and they usually ask for cash from the local financial institution in exchange for the draft. The financial institu- tion pays cash to the weaver in exchange for the draft after taking collateral equivalent to the face value of the draft. The weaver pays an annual interest of 12 to 17 percent to the financial institution. Meanwhile, weaving manufacturers buy synthetic or rayon yam di- rectly from yam producers or importers. The weaver pays a three-month draft, instead of cash, to the yam suppliers. Thirty-four percent of weaving manufacturers produce woven cloth on the basis of the putting-out contract system. The principals are usually garment exporters in urban areas, but they are sometimes weavers in the Punggi area. The garment exporter gives an order to the weaver for processing the yam provided. Time of delivery and quality standards are strictly enforced by the garment exporter. If the weaving manufacturer violates the terms of delivery, the exporter puts in a claim for the violation. The manufacturer delivers plain cloth to the dyeing factory, as in the case of the advance-order contract system. The garmentmaker pays the weaving enterprises at a piece rate, usually based on the length of the woven cloth. Exporters prefer the putting-out system to the advance-order system because, as pointed out by Hayami in Chapter 1, the principal can use the provision of raw material as credit paid in advance to the subcontractor and use it as a strong instrument to enforce commod- ity delivery and quality standards among weavers. Another type of putting-out contract is made between a small cottage weaver and a large factory weaver. The principal weaver gives an order to a subcontracting weaver for processing the yam he or she provides, and later collects the woven cloth. In this case, the larger firm is usually the previous employer of the owner of the smaller weaving firm. The principal weaver uses putting-out contracts as a short-cut to expand his sales without expanding the size of the factory. This relationship is described in detail in the following section. Because the terms of contracts are different in these two kinds of contract systems, a weaver must choose one system. For a weaver who has enough operational funds and secures a wholesaler to buy his products, an advance-order contract is preferable because he usually earns more profit per unit of weaving machinery with this system. At the same time, the weaver will bear the risks, because the draft issued by the wholesaler would become worthless if the wholesaler were to go bankrupt. To a weaver who has the ability to operate, but not enough operational funds and insuf- ficient reliable wholesalers to buy his products, the putting-out contract system is prefer- able. In this system, he produces woven cloth on the basis of a subcontract that usually specifies quality standards, a loss rate, and the time of delivery. Because the yarn is pro- vided by the principal, the weaver does not need much in the way of operational funds, except funds for wages and miscellaneous expenses. Rutral Enterprenieturship and Inidtustrial Developmenit in Korea 203 Evolultion of Entrepreneuirship Development of the weaving industry in the Punggi area might not have been possible without the emergence of capable entrepreneurs. The managerial capacity of current en- trepreneurs, as well as the generation of new entrepreneurs, is essential for the sustainable development of Punggi's weaving industry. Entrepreneurs can be categorized into new entrants of local origin and immigrants or children of imnmigrants. In 1994 about 72 per- cent of weaving factories were owned and operated by entrepreneurs of local origin (table 8-12). At the same time, 28 percent of the factories were operated by immigrants from northern Korea or their children. Before the 1970s, the share of immigrants or their chil- dren was dominant, and they represented about 70 percent in 1970. The emergence of new entrepreneurs of local origin is very similar to the apprenticeship system in Western society. Drawing on interviews with many weavers, we can tentatively say that a potential entrepreneur of local origin usually follows a course of learning to become a weaving manufacturer before embarking on independent work. A potential entrepreneur usually joins a weaving factory as an unskilled worker. He spends ten to fifteen years learning skills related to weaving production; he receives nor- mal wages. After this on-the-job training, he may want to start his own business, but he needs a certain amount of capital for the new business and must find a wholesaler to buy his product. It is difficult for him to mobilize sufficient funds to start a new business and to find a wholesaler to buy his product. In this circumstance, the enterprise where he served as a skilled worker can play the role of patron. The patron enterprise sells two or three used weaving machines to the entrepreneur on credit. The payment for the machines can usually be postponed for up to twelve months. At the same time, the patron enterprise offers to make the individual a subcontractor if he has been a diligent worker and proved to be a reliable and credible person. In this subcon- tract system, the patron enterprise provides his subcontractor with yarn and prescribes the level of quality and the loss rate for the production of cloth. Payment for the order is made on a piece-rate basis after delivery. The patron-client relationship usually continues for a few years. During this period, the new entrepreneur not only learns management skills but also finds wholesalers to buy his product. If he performs this process well, then he can be an independent entrepreneur by the age of thirty-five to forty. Table 8-12. Origin of Entrepreneurs in the Weaving Industries in Punggi Area, 1994 Origin Nitumber Percent Immigrants frorn northern Korea 15 9.4 Second generation of immigrants 30 18.8 Local origin 115 71.8 Total 160 100.0 Soturce: Authors' data. 204 Towvard the Rutral-Based Dezelopmnent of Cormnerce and Inditstry In the case of a second-generation entrepreneur, the successor to the entrepreneur usu- ally joins the weaving factory in his twenties to learn business management. After five to ten years of management training, he becomes the executive vice president of his parents' enterprise. At this stage, he is in charge of managing the enterprise, under the supervision of his parents. Several years later, usually in his late thirties or early forties, he becomes the president of the enterprise. With these patterns of the emergence of new entrepreneurs, the weaving industry has been prosperous in the Punggi area. The existence of many small enterprises and the sub- contract system have played the role of a seed-bed for the emergence of new entrepre- neurs, as asserted by previous studies (Staley and Morse 1965: 232-35; O'Farrell and Crouchley 1985: 117; Rao and Rani 1994: 123-27). Because there are many small enter- prises, workers can learn many techniques of production, marketing, and procurement of raw materials. Hence, they may acquire the general knowledge and information that is essential to start their own businesses. Let us trace the history of some typical weaving enterprises to show the emergence process of new weaving entrepreneurs. Three typical cases are presented below. CASE 1: HYUNG-DoNG KIM, OWNER OF THE DONG HOONG WEAVING FACTORY. Kim has oper- ated his own weaving factory since 1969, and he is now considered one of the most suc- cessful weaving businessmen in the Punggi area. He operates a well-structured weaving factory with twenty-three employees in the Bonghyun Industrial Estate. He is a high school graduate and is fifty-six years old. He served as president of the Punggi Weavers' Coop- erative, which is a member of the Taegu-Kyungbook Weavers' Cooperative. Until 1969, Kim had worked in a weaving factory as a spinner. He made up his mind to start his own business and bought eight secondhand weaving machines. He installed them in the kitchen of his house and began to produce woven cloth for the patron factory on a putting-out contract. He did not have to worry about sales and could concentrate on production while he relied on the putting-out contract system. After seven years, he was able to afford his own factory and decided to find a wholesaler who would buy his woven cloth on an advance- order contract. Kim walked through the wholesale markets for woven cloth in Taegu and Seoul several times. Whenever he met a wholesaler, he guaranteed the dimensions of his products be- cause many weavers cheat on the dimensions of theirs. He was able to find two interested wholesalers in Seoul and began to sell his woven cloth on advance-order contracts. The number of wholesalers that buy his products has increased because he has maintained strong credibility. Since 1987, Kim has offered to start some of his employees in their own weaving busi- nesses under putting-out contracts because he could not fill all his orders from wholesal- ers. At present, he has eight subcontractors. CASE 2: JAE-Doo LEE, OWNER OF THE DOO SAN WEAVING FACTORY. Lee is forty-one years old and operates a cottage-style workshop with his wife. He was born in a village of Rutral Entterpreneutrship anid Iindutstrial Developmenit in Korea 205 YoungPoong County. When he was seventeen years old, he started to work in a weaving factory as an unskilled employee. From 1972 to 1989 he worked in four weaving factories in the Punggi area. In 1990 he started his own weaving business with the aid of his previ- ous employer, Hyung-Dong Kim, owner of the Dong Hoong Weaving Factory mentioned in Case 1. Kim believed in Lee's ability to operate a weaving factory because he had been a dili- gent and credible worker while he was Kim's employee. Kim helped him in two ways: he gave him the chance of a putting-out contract and gave advice whenever Lee had a prob- lem operating his factory. Lee currently has sixteen electric weaving looms; he and his wife operate all of them. He now has two subcontracts, one with Kim, and another with a garment exporter in Taegu City. CASE 3: JI-YONG SONG, OWNER OF THE POONG YONG WEAVING FACTORY. Song is thirty-six years old and has been operating his weaving factory since 1993. He succeeded his father, who was seventy-three years old that year and had operated the factory since 1950. He em- ploys twenty-two workers to produce polyester cloth. Song joined the factory in 1983 when he finished three years of military service. Before his 'military service, he studied textile design for two years in a community college. Because his father operated the weaving factory for more than forty years, Song had begun learning the weaving business in child- hood. He did not need much time to learn the management of the weaving business. Nev- ertheless, his father required him to work in the factory as an employee for ten years to learn techniques and skills related to weaving, as well as skills related to labor manage- ment, procurement of raw materials, and marketing his products. His father had produced polyester cloth on advance-order contracts with a textile whole- saler in Seoul, but the wholesaler went bankrupt in 1980, and his father had to switch from advance-order contracts to the putting-out system. He now operates his factory on put- ting-out contracts from a textile exporter in Taegu. The Role of Governimenit in the Developnment of Pninggi's Weaving Indtustry In the Punggi study area, there was little government support aimed at promoting rural industry before 1989. By the end of the 1980s, the only support the central government provided to the weaving factories was in the form of institutional loans. These govern- ment supports were intended to assist the weaving enterprises in replacing their old-style, inefficient weaving machines with new ones, but fewer than 20 percent of the weaving enterprises received these loans, because most weaving enterprises did not have the col- lateral required for loans provided by local banks. The land zoning policy of the government was too conservative and has worked as a substantial barrier to the development of the weaving industry. Until the mid-1980s, all the weaving enterprises had located in residential areas, where manufacturing activities were prohibited by law. In other words, all the weaving enterprises in the Punggi area had been violating the law. As a result, they were always uncertain about their future because they were afraid the government might stop their manufacturing activities. 206 Tozvard the Rrural-Based Developmiert of Commt)iterce anid hndtistrq Local weaving manufacturers repeatedly asked the government to change the resi- dential area into an industrial area or to provide a new industrial area. It was hard for the government to change the whole township into an industrial area because many residents who did not operate weaving enterprises did not want their residential area to become industrial. At the same time, designating a new industrial site and developing an industrial estate would have been difficult for the government because most of the land suitable for industrial purposes in the vicinity of Punggi township was prime farm- land. It was not an accepted practice in Korea until the late 1980s to change prime farm land into industrial sites. When the government enacted the Farm Household Income Source Development Act in 1983, the local weaving manufacturers decided to ask the central government to construct an industrial estate in the Punggi area. Their proposal was accepted by the government, and 1,267 million won (US$1.6 million) was subsidized by the central government to con- struct the industrial estate. Government support provided about 34 percent of total invest- ment, which included land procurement, construction, and infrastructure development. Following the government decision, the Bonghyun Rural Industrial Estate was constructed at the end of 1989, and twenty-four weaving enterprises moved into this estate together with another two enterprises from an urban area. Special loans were provided to the enterprises for their investment. Each enterprise could receive a maximum of 500 million won (US$632,000) for its physical investment. The interest rate for these loans was 7.5 percent yearly, with the condition of a ten-year repayment period and a grace period of five years. In addition, each enterprise was eli- gible to borrow a maximum of 200 million won (US$253,000) in operating funds. The interest rate for these loans was also 7.5 percent yearly, with a three-year repayment pe- riod, including a one-year grace period. The policy of establishing rural industrial estates played a critical role in transforming Punggi's weaving industry from a cottage industry to modern factories. It should be noted, however, that the privileges provided to the Punggi estate were available for all enter- prises established in rural industrial estates in many other places, but they were not al- ways as successful as they were in Punggi's case. For example, two industrial estates were designated by the government in the neighboring county and city: Gahung Industrial Es- tate in YoungJoo City, and Bonghwa Industrial Estate in Bonghwa County (table 8-13). Nevertheless, 35 and 40 percent of the total enterprises in the Gahung and Bonghwa es- tates, respectively, went bankrupt, while only 8 percent (just two enterprises) went bank- rupt in the Bonghyun estate. Moreover, it did not take much time-eighteen months-for all factories to begin opera- tion in the Bonghyun Industrial Estate. Two firms in the Gahung estate and five firms in the Bonghwa estate did not start construction until February 1995. There may be several reasons that the Bonghyun Rural Industrial Estate has been so successful compared with the other two rural industrial estates. The most important factor was that most enterprises established in the Bonghyun estate were owned by local entre- preneurs, while the enterprises of the Gahung and Bongwha estates were mostly trans- planted enterprises of urban origin. Rutral Enterp7reneurship and industrial Developmen2t in Korea 207 Table 8-13. Number of Firms Operating in the Three Rural Industrial Estates, end of 1994 Name of indiustrial estate Bongltyvu Galhiug Bonigwvha Item11 (Youn11gpoon1g) (You11ngjoo) (Bongzvh1a) Size (pymig)o 32,000 33,000 45,000 Construction year 1989 1992 1993 Price of industrial estate (won/pyuing) 90,000 78,000 66,000 Number of firms initially accepted to locate (A) 26 17 15 Number of firms bankrupted (B) 2 6 6 B/A (percent) 7.7 35.3 40.0 a. One pyuwg equals to 3.3 m2. Souirce: Authors' data. According to Lloyd and Mason (1985: 71), local entrepreneurs have substantial advan- tages. Entrepreneurs who start their own businesses usually encounter a lot of risk and uncertainty, and they are likely to face serious difficulty in adapting to a new environment if they establish their enterprises in an unfamiliar place (Kim and Cha 1994). The enter- prises that moved to the Gahung and Bonghwa estates from urban areas faced many un- anticipated problems, including difficulty in acquiring qualified and reliable workers, problems in marketing their products, and difficulty in adapting to the rural area. In con- trast, the local entrepreneurs of Bonghyun estate did not have to make a new effort to acquire reliable workers and marketing channels, and they could gain support and coop- eration from the local community because they had long, personal relationships with the community. It is worthwhile to note that even in the Bonghyun estate, the two firms moved from urban areas went bankrupt. In labor management, the enterprises of local origin in the Bonghyun estate have a sig- nificant advantage over those of the other two rural industrial estates. In Bonghyun estate, a weaving factory usually has a residential quarter within the compound for the owner's family. Because the factory owner is a local person, he can live within the factory and manage the business very efficiently. In the two other rural industrial estates, most of the factory owners live in urban areas and only occasionally visit the factories. Labor manage- ment cannot be performed as efficiently as in the case of Bonghyun estate under these circumstances. Conclusions Although Korea has accomplished successful industrial development, it is far behind in rural industrial development compared with Japan and Taiwan, China. As a result, explo- sive population growth has taken place in a few large city areas, while the population has fallen drastically in many rural areas. A basic question is raised: why has rural industry 208 Tozvard the Rutral-Based Developmenit of Conmerce and Industsy been so underdeveloped in Korea, while Japan and Taiwan, China, have achieved success- ful rural industrialization in the process of economic development? As implied by Punggi's success in the 1950s, the initial conditions should be considered a basic factor. For the initial stages of industrialization, Korea inherited a very poor entre- preneurial base in rural areas because of its long history of a strong central political system and Confucianism. Moreover, the poor industrial base was almost completely destroyed by the Korean War. Because the initial conditions of industrialization were so poor, the Korean government adopted a strong industrial policy biased toward the strategic urban sector. As a result, very little room was left in the government program for the development of small rural enterprises. It should be noted, however, that the Korean government has made efforts to develop rural industry since the end of the 1960s through the Farm Household Side- Business Program, the Agricultural and Fishery Development Corporation Project, the Saemaul Factory Program, and the Rural Industrial Estate Program. These efforts, however, were misdirected, and made things worse. First, government efforts were made to provide direct incentives to rural enterprises to create individual successes, and the preconditions for industrial development in rural areas such as trans- portation facilities, electrification, and a credit market were neglected. As a result, poor infrastructural conditions prevailed until the end of the 1970s, which created difficulties for rural enterprises, as illustrated in the case of Punggi. That is, Punggi's contraction in the 1960s and its recovery in the 1970s were mostly the result of an infrastructural condi- tion-electrification. Second, government support was directed to transplanting urban enterprises to rural areas and neglected the importance of rural entrepreneurs rooted in rural communities. Government provided subsidies, generous credit, and tax exemptions or reductions in order to attract urban enterprises to rural areas. Transplanted enterprises, however, failed to adapt to the new environment in rural areas, as shown in the case of Punggi and neigh- boring rural industrial estates. Third, government concern was only directed at the physical dimension of rural indus- try and ignored the importance of institutional and entrepreneurial development. As shown in the case of Punggi's success, a marketing system, organized by various forms of con- tracts to establish links in the marketing chain, is essential for the success of rural small industry. The marketing system organized by subcontractors not only links the marketing chain but also generates new entrepreneurs. The subcontracting system is made workable by personal ties and mutual credibility, and, as shown in Punggi's case, by the existence of many small enterprises and a subcontracting system to provide rural employees with the chance to learn business management and to become new entrepreneurs through the pro- cess of patron-client relationships with previous employers. The policy implications for rural industrialization derived from this analysis are as fol- lows. First, government efforts should be made to improve infrastructure, and not to pro- vide direct incentives to rural enterprises to create individual successes. Second, government policy should be directed to rural entrepreneurs rooted in rural communities, and not to RIsral Enterprenewrship and indutstrial Development inl Korea 209 transplanting urban entrepreneurs to rural sites. Third, government should focus on the development of institutions and entrepreneurship, not only on the physical side of build- ing rural industry. How to promote the development of institutions and entrepreneurship in rural areas remains a critical question. Nevertheless, here is a caveat: any intention of the government and/ or an organization supported by the government to try to act as a marketing agent by itself will be risky and is likely to end in failure, as shown in the Agriculture and Fishery Development Cooperation Project. References Choe, Sang-Chul, and Chong-Kee Kim. 1986. "Rural Industrial Policy in Korea: Past Experiences and New Approach." In Yang-Boo Choe and Fu-Cohen Lo, eds., Rtral Indutstrialization annd Non- Farm Activities of Asian Farmlers. Seoul: Korea Rural Economic Institute and Asia and Pacific De- velopment Center. Choe, Yang-Boo, and Dong-Phil Lee. 1984. "Development of Rural Industrial Estate: Directions and Issues." Korea Rural Economic Institute Report, Seoul. Ho, Samuel P. S. 1982. "Economic Development and Rural Industry in South Korea and Taiwan." World Devielopmiienit 10: 973-90. Kang, Man Kil. 1984. Hankook Ken ndaesa (Modern Korean History). Seoul: Changjakgoe Bepyung. Keidel, Albert. 1982. "Regional Agricultural Production and Income." In Sung-Hwan Ban, Pal-Yong Moon, and Dwight H. Perkins, eds., Rural Development. Cambridge, Mass.: Harvard University Press. Kim, Chong-Kee. 1987. "Sahwoe Ganjupjabon ie Jiyuk Kyungje Sungjang ei Michin Hyokwa" ("Im- pact of Infrastructure on Regional Economic Growth"). Jot rntal of Korean Econom1tic Studies 9: 3-25. Kim, Chong-Kee, and Il-Chung Whang. 1987. "Nongchon Gonjup eui Teuksung gwa Baljun Jungchaeck" ("Characteristics of Rural Industries and Their Development Policies"). Korea De- velopment Institute Report, Seoul. Kim, Yong-Woong, and Mee-Sook Cha. 1994. "Regional Difference in New Firm Formation and Policy for Promoting Rural Enterprises." Instituttefor Humnian Settlemlent Planninlg Inforlmationi Bulle- tint 154: 30-40. Lee, Dong-Pil, and others. 1995. "Strategies for Rural Nonfarm Industries." Korea Rural Economic Institute Report, Seoul. Lloyd, F. E., and C. M. Mason. 1985. "Spatial Variations in New Firm Formation in the United Kingdom: Comparative Evidence from Merseyard, Greater Manchester and South Hampshire." In D. J. Storey, ed., Small Firms in Regiontal Econonmic Devielopmtent. Cambridge, U.K.: Cambridge University Press. O'Farrell, F. N., and R. Crouchley. 1985. "An Industrial and Spatial Analysis of New Firms Forma- tion in Ireland." In D. J. Storey, ed., Small Firms in Regional Economic Developmtent. Cambridge, U.K.: Cambridge University Press. Park, Jin-Do. 1989. "Nongga Nodongryuk eui Yuchool gwa geu Teucksung" ("Migration of the Farm Labor Force and its Characteristics"). In Jong-Hwan Joo, ed., Hankook Jabonjufelnronl (Capital- ism in Korea). Seoul: Hanul. 210 Toward the Rmral-Based Development of Commerce and Inidustry Rao, D. Chitra, and C. Rani. 1994. "Entrepreneruship Development, Status, and Trends." In Ruiral Induistrializationi. New Delhi: AFRO-ASIAN Rural Reconstruction Organization. Suh, Chong-Hyuk. 1994. "Problems and Strategy for the Adjustment of Rural Industrial Structure." Paper presented at the Symposium on Changes of Agricultural Environment and Adjustment of Rural Industrial Structure, Seoul. Suh, Chong-Hyuk, and others. 1991. "Evaluation of Off-Farm Income Policy and Its Long-term Development Strategy in Korea." Korea Rural Economic Institute Report C 91-12, Seoul. Sloboda, John E. 1982. "Off-Farm Migration." In Sung-Hwan Ban, Pal-Yong Moon, and Dwight H. Perkins, eds., Rutral DevelopmeWt. Cambridge, Mass.: Harvard University Press. Staley, E., and R. Morse. 1965. Modern Smnall Indutstryfor Developing Countries. New York,: McGraw- Hill. 9 Political Bases of Rural Industrialization: Korea and Taiwan, China David W. Lane T* his chapter illustrates the importance of the political system in shaping economic T organization and development by offering a political explanation for the divergent patterns of rural industrialization in Taiwan, China, and the Republic of Korea. Just as relational contracting between economic actors spreads risk and reduces transac- tion costs in imperfect markets (see Hayami, in this volume), so too can analogous ex- change between political and economic actors reduce investment costs and promote rural entrepreneurship. The extent to which power is decentralized influences the emergence of rural entrepreneurship by offsetting pressures for concentrated urban industry and verti- cal integration. This occurs when politics is structured to allow local politicians discretion- ary influence over policy implementation and government resources. Incentives for politicians to seek local-especially electoral-support prompt efforts to extract resources for their constituents, easing the pressures of a discriminatory policy environment for small rural industrialists whose access to, and influence on, central policy formulation is other- wise limited. As a result, for small industrialists seeking working and investment capital under conditions of credit rationing and capital scarcity, cognizance of political factors helps predict business behavior. I thank Yujiro Hayami, James Alt, Peter Hall, Dwight Perkins, two anonymous reviewers, and semi- nar members at Hitotsubashi and Harvard Universities for their helpful comments on previous versions of this paper. 211 212 Toward thle Rlural-Based Developmnenit of Coimierce antd Industry As will be shown below, aside from rural policy and at cross-purposes to financial policy, state, and particularly electoral, structure can influence the survival strategies of small businesses, tying clientelist social norms or authority patterns to political resources for economic ends and reinforcing patterns of business organization and the political status quo. Entrepreneurs seeking the cheapest effective self-help strategy3-in this case, to ob- tain financing-can choose between maximizing under existing constraints (economic strat- egies), trying to alter those constraints (political strategies), or market exit. Subcontracting and borrowing on informal credit markets are examples of economic self-help strategies; voting, bribery, lobbying, and clientelism are examples of political strategies; and bank- ruptcy and offshore production are examples of market exit. Because government regula- tion and tolerance for these survival strategies affects their cost to the small firm, the political structures institutionalizing these costs in Taiwan, China, and Korea will prompt cross- country variance in the strategies entrepreneurs choose to adopt. The patterns of individual or collective self-help selected by small firms on the basis of these politically imposed costs help to explain the varying patterns of business organiza- tion. The evidence shows that in Taiwan, China, state and electoral structure leaves room for political agents to capture state resources for local'economic benefit, but that in Korea, the centralization of economic and political power works to deny opportunities for politi- cal and economic entrepreneurship to local businessowners. The lesson is that political structure can create economic externalities. To the extent that political structure is a func- tion of choice rather than necessity, therefore, the evidence here suggests that political organization and political rules can be shaped consciously to affect economic outcomes. The Puzzle of Rural Industrialization in Taiwan, China, and Korea Divergence in the industrial structures of Korea and Taiwan, China, is anomalous in light of the countries' many commonalities. Despite precolonial differences, at approximately the same time over most of the twentieth century, Korea and Taiwan, China, have shared similar resource endowments, colonial experiences, external security threats and trading opportunities, initial economic conditions, development paths, and even development poli- cies. Even so, each has followed a different path to industrial growth. Korea's industry is concentrated in urban areas and is dominated by diversified, verti- cally integrated conglomerates. Table 9-1 shows that just over half of Korea's small and medium-size manufacturers are located in the six largest cities. If the small industrial es- tablishments of Kyonggi Province are included in this group-Kyonggi being the prov- ince in which the capital, Seoul, and its port, Inchon, are located-then a whopping 78 percent of Korea's small industry can be classified as urban. Indeed, Jong-Gie Kim (1991: 3) states that "between 1971 and 1988, the proportion of manufacturing firms located in urban areas rose from 54.4 to 75.6 percent," suggesting that the inclusion of provincial territory within urban boundaries is entirely appropriate.4 Kim (1991: 3-4) continues: Political Bases of Rutral Inidutstrializationz 213 In 1970, 17.8% of rural manufacturing firms were located in South Cholla province, 15.5% in Kyonggi province, and 15.1% in South Chungchong province. However, by 1980, while almost all other provinces experienced a decline in their share of manu- facturing firms ... Kyonggi. . . increased its share to more than 30%. During the same period, rural manufacturing employment in ... Kyonggi province increased more than four times and now accounts for 40.9% of the nation's manufacturing employ- ment. Other provinces experienced lower growth rates in manufacturing or no growth at all both in terms of employment and number of firms. Table 9-1. Small and Medium-Size Manufacturers by Location: Korea and Taiwan, China, 1990 Korea Taizvan, Clinla Numnlber Percenit Numniber Percenit Total 67,679 100 Total 155,263 100 Cities 35,095 51.9 Cities 28,611 18.4 Seoul 17,379 25.6 Taibei 9,591 6.1 Pusan 6,909 10.2 Gaoxiong 3,678 2.3 Taegu 4,778 7.0 Jilong 832 0.5 Inchon 4,270 6.4 Xinzhu 4,308 2.7 Kwangju 787 1.1 Taizhong 8,392 5.4 Taejon 972 1.4 Jiayi 1,810 1.1 Tainan 7,582 4.8 Provinces 32,584 48.1 Counties 119,126 76.7 Kyonggi 17,853 26.3 Taibei 33,393 21.5 Kangwon 1,041 1.5 Taizhong 23,414 15.0 N. Chungchong 1,043 1.5 Zhanghua 17,369 11.1 S. Chungchong 1,600 2.3 Taoyuan 9,134 5.8 N. Cholla 1,876 2.7 Tainan 8,635 5.5 S. Cholla 1,832 2.7 Gaoxiong 5,322 3.4 N. Kyungsang 3,073 4.5 Miaoli 4,070 2.6 S. Kyungsang 4,041 5.9 Ilan 2,908 1.8 Cheju 225 0.3 Jiayi 2,715 1.7 Nantou 2,647 1.7 Pingdong 2,539 1.6 Yunlin 2,284 1.4 Xinzhu 2,258 1.4 Hualian 1,351 0.8 Taidong 562 0.3 Penghu 525 0.3 Sour ce: Korea Small and Medium Industry Promotion Corporation, Major Statistics of Smniall and Medium11 Indus- tries, 1994. Taiwan Ministry of Economic Affairs, Small and Medium-sized Enterprise Agency, Wlhite Paper on Smiall and Medium-sized Enterprise (Zhongxiao qiye baipishin), 1991. 214 Toward the Rutral-Based Developmenit of Conmierce and Indlstlry Ho (1982: 980) extends the argument further: Between 1958 and 1975, there was a more than three-fold increase (from 109,000 to 369,000) in rural factory employment, but the relative share of rural factory employ- ment in total factory employment declined from 42 to 26%. It should be noted that the increase in rural factory employment is probably biased upward because included in "rural" is Taegu, one of Korea's most rapidly expanding cities in the 1960s and 1970s. In contrast, the industry of Taiwan, China, is smaller in scale, and its firms are far more numerous and more evenly distributed between rural and urban locations. Table 9-1 shows that less than one-fifth of Taiwan's small and medium-size manufacturers are located within the largest cities. Once again, however, counties adjacent to Taiwan's major cities may be considered urban when their enterprises gain from proximity to large contractors and the urban marketplace. If Taibei and Taizhong Counties are tallied with the cities, about 55 percent of Taiwan's small industry can be considered urban-based. Lack of data classified both by firm size and precise location makes accurate appraisal difficult, but Amsden (1991: 1130, 1132) shows that industrial employment rose most between 1966 and 1986 in the rural regions abutting Taiwan's metropolitan areas. Over these two decades, this "sub- urban" employment rose by a factor of nearly seven, compared with that in rural areas of almost five, and urban growth of only three. Evidence from farm household income surveys buttresses the view that Korea's rural industry is unusually quiescent. Data in table 9-2 show that the overwhelming majority of Korean farm incomes remain farm-based despite the structural transformation of the economy as a whole into industry and services. Taiwanese farmers, however, to say noth- ing of their Japanese counterparts, obtain most of their income from wages, salaries, and other nonfarm business. Many Koreans have quit the countryside altogether for urban employment: table 9-3 shows that population in the six largest cities alone has doubled every two decades, while that in the rest of the country stayed flat.' Figure 9-1 shows that despite the industrial deepening of the 1970s, Taiwan, China, remains less urbanized than Korea. Taipei, like Seoul in Korea, is the seat of central government and the center of Taiwan's business, finance, medicine, publishing, and education activities. Yet only about one-tenth of Taiwan's 20 million people live in the capital, compared with the one-quarter of Korea's 40 million people who live in Seoul. Two hypotheses have been advanced to explain the spatially dispersed industry of Tai- wan, China, relative to Korea. The first argues that Taiwan's agricultural sector, both un- der Japanese occupation and since, was healthier, more advanced, and more diverse than Korea's, and therefore produced more surplus to invest in alternative undertakings. Tai- wan processed sugar, fruit, and vegetables in the countryside from an early date, thereby providing impetus for rural commercial and support activities (Ho 1982: 982-83). Con- versely, Korea was basically a rice monoculture economy until the state's attempts in the mid-1960s to diversify crops and stabilize commodity prices through the Agriculture and Political Bases of Rutral Indutstrializationt 215 Table 9-2. Composition of Farm Household Income: Korea, Taiwan (China), and Japan (percent) Agricultutiral Nonagrictultutral incomne income Total Wages Side business Other 1962 Korea 79.6 20.4 9.4 3.8 7.2 Japan 51.4 48.6 35.7 9.4 3.5 1965 Korea 79.2 20.8 8.9 3.5 8.4 Taiwan, China 66.0 34.0 20.1 3.8 11.1 Japan 48.0 52.0 41.4 7.0 3.6 1970 Korea 75.9 24.1 10.4 3.8 9.9 Taiwan, China 48.7 51.3 36.0 2.7 12.6 Japan 36.5 63.5 52.8 6.9 3.8 1975 Korea 81.9 18.1 9.4 2.5 6.2 Taiwan, China 48.1 51.9 37.3 3.5 11.1 Japan 31.3 68.7 - - - 1980 Korea 65.2 34.8 14.5 2.5 17.8 Taiwan, China 26.4 73.6 52.2 6.2 15.2 Japan 20.0 80.0 - - - 1985 Korea 64.5 35.5 13.8 3.7 18.0 Taiwan, China 24.8 75.2 - - - Japan 19.4 80.6 69.5 8.3 2.8 1987 Korea 61.4 38.5 14.7 4.3 19.5 Taiwan, China 22.0 78.0 - - - Japan 16.5 83.5 71.9 9.8 1.8 Note: Data for Taiwan, China, and Japan are from 1966 and 1974 rather than 1965 and 1975; 1987 data for Taiwan, China, and Japan are from 1989 and 1988 respectively. Source: Korea: Ministry of Agriculture and Fisheries, various years. Report on the Results of the Farm Household Survey; Taiwan, China: Council for Agricultural Planning and Development. 1982. Basic Agricultuiral Statistics; Japan: Ministry of Agriculture, Forestry, and Fisheries. nd. Pocket Agriculture Statistics, cited in Jong-Gie Kim (1991: 8-9). Fisheries Development Corporation (AFDC). The AFDC financed and oversaw construc- tion of cold-storage facilities and processing plants in the face of incentives making moneylending more profitable to rural entrepreneurs than direct investment in produc- tive capacity.6 Little systematic evidence has been gathered to test this proposition, but its positive implications for Taiwan, China, may be tempered by the evidence noted below that agri- 216 Toward the Rutral-Based Development of Coninmerce and Inditstry Table 9-3. Population by Location, Korea and Taiwan, China, 1960-92 Popuilation by area, TathiLan, Cliina 1992 1990 1985 1980 1975 1970 1965 1958 Total 20,752,494 20,352,966 19,258,053 17,805,069 16,149,702 14,675,964 12,483,193 9,691,378 Cities 6,542,459 6,486,377 6,034,400 4,944,645 4,454,026 3,844,774 2,773,994 1,842,164 Taibei 2,696,073 2,719,659 2,507,620 2,220,429 2,043,318 1,769,568 1,135,500 718,237 Gaoxiong 1,405,909 1,386,723 1,302,849 1,202,123 998,919 828,191 596,092 377,446 jilong 359,482 352,919 351,524 344,867 341,383 324,040 278,320 190,810 Xinzhu 332,707 324,426 304,010 Taizhong 794,960 761,802 674,936 593,427 546,838 448,140 364,262 256,088 Jiayi 258,698 257,597 253,573 Tainan 694,630 683,251 639,888 583,799 523,568 474,835 399,820 299,583 Counties 14,210,035 13,866,589 13,223,653 12,860,424 11,695,676 10,831,190 9,709,199 7,849,214 Taibei 3,162,346 3,048,034 2,663,683 2,258,757 1,629,105 1,240,576 1,061,552 725,105 Taizhong 1,317,505 1,258,157 1,142,189 1,013,176 861,606 785,903 688,092 572,050 Zhanghua 1,264,955 1,245,288 1,223,209 1,166,352 1,102,835 1,050,246 978,636 858,985 Taoyuan 1,415,546 1,355,175 1,211,249 1,052,800 861,792 726,750 588,786 442,319 Tainan 1,046,659 1,026,983 1,000,781 962,827 943,075 934,865 873,174 759,016 Gaoxiong 1,146,578 1,119,263 1,076,761 1,000,645 943,650 830,661 712,201 565,075 Miaoli 553,557 547,609 550,343 542,745 538,820 524,744 479,685 406,285 Ilan 456,857 450,943 449,981 442,988 427,655 412,787 378,881 325,447 Jiayi 556,580 552,277 569,932 825,967 840,664 849,914 793,110 672,715 Nantou 542,396 536,479 535,572 524,245 517,927 511,040 467,526 384,346 Pingdong 901,491 893,282 901,981 888,270 857,089 828,761 744,111 559,010 Yunlin 753,841 753,639 791,186 796,276 801,422 800,578 749,416 644,979 Xinzhu 385,668 374,492 366,566 641,937 617,311 587,652 525,008 431,218 Hualian 355,609 352,233 361,549 355,178 346,914 335,799 298,663 229,440 Taidong 255,362 256,803 276,389 281,218 291,111 291,761 260,322 182,379 Penghu 95,085 95,932 102,282 107,043 114,700 119,153 110,036 90,845 cultural surplus was often extracted from the countryside, first by the Japanese during the colonial period (Gold 1988: 104-109), and later by the Nationalist government. Further- more, the value of Taiwan's food and processed agricultural exports fell sharply through- out the 1960s, from 67 percent of total exports in 1960 to barely 20 percent by 1970 and to under 10 percent in 1980. Agriculture's share of the net national product also traces a downward slope, from 33 percent in 1960 to 18 percent in 1970 and 9 percent in 1980 (CEPD 1988:41,213). Thus, while Taiwan's agricultural surplus may have been an impor- tant source of investable funds for rural industry in the 1960s, and was certainly more important than Korean agriculture in this respect at that time, the role of agricultural sur- plus inevitably waned thereafter with its ebbing absolute volume, and with it went most of its impact on the emergence of rural entrepreneurship. The second and more important hypothesis is that the high quality and early installa- tion of physical infrastructure made rural industrialization possible in Taiwan, China, but Political Bases of Ruiral Indutstrializationi 217 Population by area, Korea 1992 1990 1985 1980 1975 1970 1965 1960 Total 44,563,318 43,471,628 42,220,980 41,503,762 37,931,617 33,926,536 30,588,956 27,727,606 Cities 21,573,682 20,664,428 18,343,496 15,592,254 12,567,186 9,902,374 6,860,628 5,235,634 Seoul 10,969,862 10,612,577 9,639,110 8,364,379 6,889,502 5,433,198 3,470,880 2,445,402 Pusan 3,887,278 3,798,113 3,514,798 3,159,766 2,453,173 1,842,259 1,419,808 1,163,671 Taegu 2,286,305 2,229,040 2,029,853 1,604,934 1,310,768 1,063,553 811,406 676,692 Inchon 2,070,616 1,817,919 1,386,911 1,083,906 799,982 646,013 485,511 402,009 Kwangju 1,134,843 1,062,084 866,695 651,642 506,703 414,598 307,412 229,393 Taejon 1,224,778 1,144,695 906,129 727,627 607,058 502,753 365,611 318,467 Provinces 22,989,636 22,807,200 23,877,484 25,911,508 25,364,431 24,024,162 23,728,328 22,491,972 Kyonggi 6,619,629 6,155,632 4,794,135 4,933,862 4,039,885 3,358,105 2,984,374 2,721,630 Kangwon 1,555,082 1,580,430 1,724,809 1,790,954 1,862,107 1,866,928 1,771,035 1,636,726 N. Chung- chong 1,403,633 1,414,295 1,391,004 1,424,083 1,522,203 1,453,899 1,550,475 1,369,313 S. Chung- chong 1,872,095 2,027,766 3,001,572 2,955,999 2,948,649 2,860,690 2,899,838 2,525,646 N. Kyong- sang 2,873,336 2,865,676 3,010,001 4,954,559 4,858,551 4,559,092 4,479,614 3,848,424 S. Kyong- sang 3,846,825 3,671,509 3,516,660 3,322,132 3,280,052 3,119,669 3,176,401 4,182,042 N. Cholla 2,028,956 2,069,848 2,202,243 2,287,689 2,456,403 2,434,522 2,516,528 2,373,487 S. Cholla 2,283,858 2,507,439 3,748,484 3,779,475 3,984,849 4,005,735 4,013,011 3,553,041 Cheju 506,222 514,605 488,576 462,755 411,732 365,522 337,052 281,663 Source: 1958: Taiwan Provincial Government, Department of Civil Affairs (Taiwan shengzhengfu minzhengting): Hoousehiold Registrationi Statistics of TaiW'an (Tafintnsheng h1wji ton gji ynolan), 1965: Monthily Bulletini of Popiulation Reg- istration Statistics of Tainanii, ROC (Z1honghu7a milignto Taiaiushieng l1uiji tongji yuiebao), 1970: Tain'an Demograply Monthly (Zhonghlua iniiiguio Tai7an renkou tongji yuekan), 1975-92: Taiwan, Ministry of the Interior (Zhonghua minguo neizhengbu). Taiwntn Demographiy Quarterly, ROC (Zhlonghlua minguio Taiu'an renikou tongji jikan). Korea Institute of Urban Administration, National Statistical Yearbook (Hanguk toshi hengchongyunguso, Chonguk tongkye yeonkam), 1994. its absence in Korea afforded fewer opportunities for rural entrepreneurs to prosper. The data on paved roads and electrification offer powerful evidence for this contention. According to Ho (1982: 983): The density of paved highway and feeder roads in Taiwan was 76.4 kilometers per thousand square kilometers in 1962 and 214.5 kilometers per thousand square kilo- meters in 1972.... In 1960, the density of paved road in Korea was below 10 kilome- ters per thousand square kilometers and only 6.4 kilometers per thousand square kilometers if city roads were excluded, and much of this roadway was in the Seoul area where the density of paved road was about three times the national average. 218 Toward the Rutral-Based Development of Commerce and Induistry Figure 9-1. Urban to Total Population, Korea and Taiwan, China Percent 50 48.4 Taiwan, China 43.4 EKorea ? 40 376 33.1 33 31.9 31.5 30 2229.2 27 28 190 189 222224 20 10 0 1960 1965 1970 1975 1980 1985 1990 1992 Taiwan, China, cities: Taipei; Gaoxiong; Jilong; Taizhong; Tainan; and, for 1985-92, Xinzhu and Jiayi. Korea cities: Seoul, Pusan, Taegu, Inchon, Kwangju, and Taejon. Even in 1975, after Korea had built a national highway grid, its paved road density (excluding city roads) was still below 50 kilometers per thousand square kilometers, a level reached in Taiwan in the late 1950s. The absence of paved rural roads in Korea offered few advantages to rural entrepreneurs. Lacking significant wage or productivity differentials to drive urban agents to seek rural labor and subcontracting, poor Korean infrastructure encouraged entrepreneurs and sub- contractors to set up shop in close proximity to their urban-based customers rather than assume the heavy inventory, transportation, and information costs necessitated by a rural location. Conversely, physical infrastructure in Taiwan, China, created trading links to urban markets, facilitating rural entrepreneurship and rural-urban interaction. The case of rural electrification presents a similar contrast. By 1960, Ho states, 70 per- cent of farm households in Taiwan, China, received electricity, compared with a mere 13 percent of Korean households in 1964, most of these in Kyonggi Province surrounding Seoul (1982: 983-84). Early rural electrification in Taiwan, by providing an important re- source to industry, surely favored local development over urban industrial concentration. Wade (1990: 268), among others, claims that this was the state's intent from the beginning. Overcapacity was built into the system in the expectation that rural industry would arise Political Bases of Rutral In7dustrialization 219 to exploit it. Ranis (1979: 215) notes that government attempted to charge rural and urban users equal rates for electricity, or at least tried to avoid subsidizing either group. Provision of physical infrastructure to rural areas is a necessary but not sufficient condi- tion for rural entrepreneurship. Local, independent entrepreneurship will not bring about rural industrialization absent the physical infrastructure making cost-effective production and urban linkage possible. Rural producers require urban markets or their agents for information-gathering, production inputs, and sales, as well as urban services for interna- tional trade. Hence, Taiwan, China, was advantaged over Korea by its early superiority in rural infrastructure, and perhaps by its diversified rural economy as well. Yet the presence of a favorable physical environment does not, in itself, ensure sustained entrepreneurship and industrialization. Moreover, while Korea's infrastructural development has since been converging with that of Taiwan, its rural industrialization has not. Continued divergence between Korea and Taiwan in rural industrial vitality suggests either that physical infra- structure plays a more limited role than their comparison otherwise implies, or that Taiwan's gains from infrastructure came early as a once-and-for-all boost. Both conjectures point to the need for a better account of the incentives affecting the mobilization and persistence of rural entrepreneurship. Policy Incentives and Politics Aside from the advantages of Taiwan, China, in infrastructure and agricultural diversity, political variables also work to favor Taiwan over Korea in rural entrepreneurship and industry. Political variables take many forms, with diverse effects. Political decisions alter factor prices and create rents. Political criteria also prompt the provision of infrastructure, and even migratory pressures, and thus merit examination for their systemic economic impact.7 This section discusses some economic effects of political variables in Korea and Taiwan, ranging from direct policy intervention to less obvious, but no less meaningful, political influences on entrepreneurial decisionmaking. Rutral Policy Rural policy is the most straightforward political link to rural industrial development. Yet although we intuitively expect new resources to spur growth, several state efforts to create rural industry in Korea have produced few of the hoped-for results.' One such program, begun in 1967, tried to combine readily available local labor and raw materials with loans and tax incentives to create a rural handicrafts sector. Partly because handicrafts offer few linkages to other industries and are neither an intermediate input nor an essential con- sumer item, the program added little significant value to local labor. The Saemaul Factory Program was a more successful attempt to encourage local in- dustrialization, this time through generous incentives to relocate or expand factories in small townships to process local raw materials.9 Although received well in local areas, 220 Towvard the Rural-Based Developmeit of Comminerce and inditstry the unEavailability of skilled labor, ancillary services, and adequate physical infrastruc- ture hobbled many of the factories that moved from urban areas to the countryside. Profitable Saemaul factories remained so partly because of their proximity to urban markets and industry (Kim 1991:14): Almost 75% of the Saenratul factories in operation are concentrated in the provinces of Kyonggi, and North and South Kyongsang, and most of them [are located] near the major metropolitan areas of Seoul, Pusan, and Taegu. Indeed, one source argued forcefully that much money was wasted on the Saemaul program precisely because political concerns prompted a balanced distribution of funds by region rather than according to the economic potential of site-specific investment."0 Current rural industrialization efforts build on the lessons learned from previous mis- takes, but rather than improve rural conditions, they perpetuate the urbanization of in- dustry and the migration of the past by establishing industrial estates in existing rural centers deemed capable of providing the requisite labor and infrastructure. Korea's rural development policy limits its own effectiveness by continuing to seek and create indi- vidual successes directly and from scratch, rather than indirectly through broad incen- tives that appeal to the self-interest of the whole stratum of potential rural entrepreneurs. The effects of Taiwan, China's, rural policy on local industrialization are also limited. Apart from constructing physical infrastructure and conducting land reform-which gen- erally failed to transform dispossessed landlords into industrialists" -government did little to encourage rural industry directly. Although Kuo (1983: 23-29) suggests that land reform aided industrialization by promoting the cultivation of cash crops for potential processing, the contributions of agriculture to industry should not be overstated. Capital outflow from agriculture ended by 1970 (Yager 1988: 254-55), while Taiwan has continued its rapid growth apace. Moreover, Taiwan's reliance on large-scale public enterprise and the use of the postal savings system and grain procurement policy to withdraw agricul- tural surplus from the countryside for long-term industrial and infrastructure investment suggests that Taiwan's authorities placed a low priority on balanced rural development through local entrepreneurship."2 Small-industry policy in Taiwan took on distinct shape only with the creation of targeted financial programs and technical assistance centers in the early 1970s, but these were not specifically rural-based. The problems of rural policy in Korea and its low profile in Taiwan, China, suggest that this most direct of policy measures has been largely irrelevant to local industrialization. Other political variables, however, have played significant roles. Policy Incentives Intentionally or not, both targeted and macro-level policies often produce incentives that affect a wide range of other actors and outcomes. Tax regulations, for example, although aimed at raising revenue without hindering growth, regularly create perverse economic Political Bases of Rutral Inidustrialization 221 incentives instead. Thus Taiwanese (China) businessowners form broad networks of small companies, ultimately controlled by an individual or a very small coterie of intimates, which as a group successfully avoid taxation by virtue of their inconspicuous size. The same corporate resources, integrated as a single conglomerate, would quickly draw un- welcome official attention."3 Patterns of policy enforcement also produce economic incen- tives: lax enforcement of tax laws in Taiwan and Japan is a noted tacit form of government support for small business."4 Regulatory policy also affects the transaction costs of economic exchange. Given con- straints on the availability of working capital in Taiwan, China, for example, there is a broad-based market in postdated checks as a substitute for formal credit facilities and commercial paper. While the Taiwanese are not alone in using the postdated check as an informal credit instrument, government policy toward curb market financing there pre- sents a notable contrast to repeated Korean efforts to stamp out informal lending. In a tacit admission of the inadequacies of the state-operated banking system, until 1974 Taiwan's government permitted postdated checks to be cashed prior to maturity (Caldwell 1976: 740), not only spurring their use but also improving efficiency in the otherwise fragmented curb market by creating a secondary market in discounted checks. More important, until its repeal in mid-1987, was a statute that made passing bad checks a criminal offense and allowed the attachment of a borrower's assets to recover losses. By contrast, dishonoring a promissory note or trade bill carried only civil penalties (Caldwell 1976: 739). Although conclusive studies are lacking, one may reasonably infer that in the absence of legal enforcement mechanisms, the cost of discounting and transacting with postdated checks rose significantly. Indeed, the government hoped the increased risk of transacting after the statute's repeal would force these unregulated money flows into the formal sector (Winn 1986). The Korean government, in contrast, has attempted on several occasions to eliminate the curb market by discouraging potential creditors rather than by satisfying unmet de- mand. A 1972 presidential decree dissolved the curb market by fiat, but its effects were only temporary (Cole and Park 1983: 158-68; Clifford 1994: 102-04). Vested interests re- pelled attempts to eliminate false-name bank accounts (one means for curb market credi- tors to stockpile cash) until President Kim Young Sam's successful initiative in 1993. State and Electoral Structture Beyond the indirect effects of individual policies, however, broad political institutions can influence economic entrepreneurship. The extent of decentralization of political influence has a particularly strong affect on the prosperity of small entrepreneurs. Embedded in state organization are rules and norms governing, among other things, the extent to which lobbying is tolerated and the channels for its expression. These institutions are located in disparate areas, ranging from legislative procedure to electoral rules, from the power of the executive branch to factionalism in the governing political party. The argument pre- 222 Towoard the Rural-Based Developnwet of Conimerce anid Indulstry sented (but not tested) here is that these institutional differences in state structure affect the costs of self-help strategies to economic entrepreneurs, and therefore the choice of political and economic strategies that firms are likely to pursue. As applied to the question of rural industrialization in Korea and Taiwan, China, the argument focuses on the varying extent of decentralization of political influence in the two otherwise similarly authoritarian political regimes. Despite the shared inability of rural entrepreneurs in each system to directly influence policy formulation at the central government level, and in the face of financial policies exacerbating the normally discrimi- natory borrowing environment, small firms in Taiwan, especially rural-based firms, are able to use the electoral dynamic locally in a compensatory way to ease the constraints on investment and operating capital. In the absence of similar mechanisms in Korea, rural entrepreneurs there are forced to rely on costly economic survival strategies such as infor- mal borrowing or, when available, subcontracting. The following section illustrates the argument for Korea and Taiwan, with brief comparative mention of Japan as well. Financing Small Industry Among the most powerful policy tools available to officials in postwar Korea; Taiwan, China; and Japan has been influence over the financial system and credit allocation (Cole and Park 1983; Johnson 1982; Wade 1985, 1990; Woo 1991). Until recently, interest rates were fixed and banks and credit markets were highly segmented in all three systems. Indirect financing (bank loans) predominated and capital markets were left weak. Government policy was to aggregate funds from small savers and invest them in building large-scale industry. Post offices offered savings accounts bearing above-market effective interest rates, and the deposits were reinvested by government. According to one observer, Korea's central bank requires rural bank branches to transfer 70 percent of new deposits out of the coun- tryside to Seoul for reallocation (Jung-Hwan Lee and Chong-Hyuk Suh, in this volume). For small business, it was critical that banks in Taiwan, China, and Korea were virtually all state-owned, and banking-market entry almost nonexistent, resulting in credit ration- ing to large and politically well-connected firms and the rise of significant curb markets to serve unmet demand. Even now, for loans to small firms, Taiwan's banks demand collat- eral of 120 percent or more; Korean banks demand 150 percent."5 State ownership of Taiwan's banks ensures conservative lending patterns by their civil servant employees, whose prospects depend on minimizing error. Although uncommon in practice, Taiwan's bankers can still be required to recoup loan losses out of personal funds. Small firms are ordinarily disadvantaged relative to large firms in the freest of markets: their usually vague track records, poor bookkeeping, limited collateral, and relatively small borrowings all make them less reliable and more costly borrowers for the banks than larger firms. The point of the above summary, however, is that development policy in Korea and Taiwan, China, said to favor producers over consumers, increased the costs of investment for smaller producers. Political Bases of Rutral Iiidustrialization 223 Predicting Snmall Firm Choice Despite such policy constraints, small entrepreneurs do not lack adaptive financing op- tions, at least in theory. Individually, they may seek out subcontracting relationships with firms that benefit from government largesse, thereby ensuring occasional access to credit, technology, and a source of stable demand for their products (Yokokura 1984). Some may pursue clientelist strategies with local political representatives or seek favors as part of political support associations. They may express their interests at the ballot box, solicit investment funds from friends and relatives, or go bankrupt. Collectively, firms may orga- nize, locally or nationally, to press for wholesale change in government policy or for com- pensation for policy-driven losses. Further, entrepreneurs may move outside the bounds of state regulation to form unofficial credit markets or try to extract state resources with the aid of local officials (Oi 1989). Aside from the organizational costs of collective action,'6 state and electoral structure can influence the calculus of entrepreneurial decision by defining the limits of permis- sible political participation, the effectiveness of participation, and the cost of ignoring state proscription of political action. For example, two revealing indicators of openness to political participation are the presence and role accorded to voluntary associations and the existence and impact of elections for political representatives. Of no less import are the incentives to politicians to assist small entrepreneurs and their control over the resources to do so. Focus on the role of local politicians and the electoral process is primary in this in- stance because of the limits set on collective lobbying by the state. In both postwar Korea and Taiwan, China, for example, state controls over collective political opposition made lobbying of the central government costly and difficult. In Taiwan, associations were illegal without government approval and registration with the appropriate ministry. By law, every economic sector of more than five firms formed an industrial association; the more important associations received both a state-appointed full-time secretary to run them and the oversight of the Nationalist Party's social affairs department (Wade 1990: 271). Korea's attitude toward business associations and voluntary association has dif- fered little, permitting no formal means to articulate, let alone organize behind and act upon, positions conflicting with government policy. The only voluntary associations to survive locally in Taiwan were the traditional clan, fraternal, and temple societies (Rig- ger 1993: 21), which by default tended to become repositories of communal and political activity (Feuchtwang 1981: 53). Where political controls on free expression make collective lobbying impossible, ratio- nal entrepreneurs may resort to individual political action or to economic alternatives. Yet economic strategies are also costly, and they put rural entrepreneurs at a disadvantage to firms with better access to government, even in cases when that access comes only at a price. At the same time, where political access is readily available and political representa- tion cheap, rational entrepreneurs forgo costly economic coping mechanisms to press their interests in the political arena. 224 Towarmd tlie Rutral-Based Developnment of Conmmerce anid Ilnduistry In sum, state structure both constrains and facilitates political action. By placing costs on the self-help alternatives available to businesses, such as collective lobbying, state struc- ture constrains choice. Yet state structure can also provide entrepreneurial opportunity by binding with social structures and institutions to produce (often unintended) positive eco- nomic externalities, in this case at the level of local electoral politics. The evidence that follows shows that as a constraint on entrepreneurship, state struc- ture in both Korea and Taiwan, China, since the 1960s similarly raised the cost of political action and forced local industrialists into costly economic adjustment strategies. By con- trast, during the capital-scarce early postwar years in democratic Japan, the constraints of state structure were minimal and small entrepreneurs used the ready access to the politi- cal system to press their interests. In their facilitating effects on entrepreneurship, however, the state structure of Taiwan, China, has offered more to rural entrepreneurs than that of Korea. By centralizing political authority and economic resources in a strong executive branch that monopolizes local political and economic life, the Korean state has nearly eliminated the opportunity-based effects of state structure on local entrepreneurship. In Taiwan, a strong executive branch has left room for electoral competition for local and provincial posts to provide preexist- ing local political factions and economic interests informal access to state resources for local development. Although not benefiting all rural entrepreneurs, this dynamic has none- theless spurred local activity, particularly by easing investment and operating costs. La- tent entrepreneurship can thus be roused by the opportunities that the specifics of government organization create. Eviden1ce In Taiwan, China, small industrialists use the electoral dynamic at the local level to sub- vert constraints imposed by a centralized state and a powerful ruling political party. Poli- ticians link existing clientelist relations among political factions and economic networks at the local level with opportunities offered by local elections and electoral rules in order to secure a power base for themselves and to serve their constituencies (Wang 1994). Local entrepreneurial impulses are thereby given material support by political agents on the basis of patron-client ties. This outcome is at odds with the conventional portrait of the distribution of power in Taiwan, China. Most observers rightly conclude that overall political control and the power to formulate and implement policy rests with the ruling Nationalist Party (KMT), whose Leninist organization has thoroughly penetrated government. In practice, power is fur- ther concentrated among several dozen top KMT members who influence government operations through their personal as well as institutional power bases. In organizational terms, however, the government of Taiwan, China, is divided into both a national legislature, whose membership was wholly controlled by the KMT, and a "powerless" provincial legislature, whose membership is chosen by vote. The KMT made this decision partly to maintain the pretense of democracy while under martial law, and Political Bases of Rural Indutstrialization 225 partly to coopt and control local political actors after the KMT's retreat to Taiwan. The creation of an elected provincial assembly provided a platform for the expression and aggregation of local interests. While the KMT long made clear that basic political ques- tions regarding its rule were not open for debate, provincial assemblymen have felt little compunction in seeking personal and constituent benefits from their office (Lerman 1978). As a result, the electoral dynamic at the provincial level has had at least two effects helpful to local businessmen at the expense of KMT policy and economic controls. First, small industrialists in the early 1970s successfully lobbied members of the provin- cial assembly to press the party for relief from the debilitating effects of credit rationing. The provincial assembly members prevailed upon the provincial finance commissioner, a high-ranking KMT official, to act. New emphasis was placed on policy for small industry, and several state corporations were established as a result: the Small Business Integrated Assistance Center (1973), the Small and Medium Business Credit Guarantee Fund (1974), and the network of small and medium-size enterprise banks (1975). These either lend directly or reduce the cost to banks of small business loans by lowering transaction costs and informational asymmetries. Second, and more important, the sole administrative authority of the provincial assem- bly over Taiwan, China's, three largest commercial banks has resulted in the election to the assembly of local faction bosses who press the banks to lend to local clients on other than strictly economic grounds. Bad loans are lost in a maze of provincial- and central-govern- ment-level accounting.'7 While data are scarce for this informal and illicit, but common, activity, it is possible to depict the financial constraints facing a typical entrepreneur and circumstantial evidence for political exchange. Taking 1986, a year for which a variety of financial data are avail- able, as an example, Taiwan, China's, private firms as a group could borrow only 55 per- cent of the funds they actually invested from formal financial institutions: 45 percent of their borrowing came from informal sources, including interfirm lending.'8 Small firms received just over one-third of the funds loaned by Taiwan's banks in 1986, even though they accounted for 98 percent of all firms at the time.'9 Moreover, 92 percent of the funds made available to small firms were lent on a short-term basis: less than 3 percent of bank loans to Taiwan's small firms had terms of over one year.20 The cost of money was an additional constraint: in 1986, interest on long-term loans from the commercial banks ranged between a floor of 5.50 percent and a ceiling of 10.25 percent; interest on short-term secured loans ranged between 5.00 and 9.50 percent. This published range is somewhat misleading, however. Given the absence of competition from either the bond market (which is minuscule) or informal credit markets (whose rates are much higher; see table 9-4), interest rates charged borrowers by Taiwan, China's, banks are always at the maximum permissible rate, whatever the quality of the credit or the maturity of the loan (Caldwell 1976:738). If anything, given the normal risks and transac- tion costs involved, the interest rates charged to small firms in practice are raised further by informal demands for compensating balances. The 1986 interest rate on loans for ex- ports was a uniform 5.50 percent to all, several points lower than short-term secured loans. 226 Towvard thze Riral-Based Detvelopment of Commerce atd Indtstry Table 9-4. Margins Between Curb Loan Rates and the Secured One-Year Loan Rate, Korea and Taiwan, China Korea Taiwvan, China Year (A) (B) (A) - (B) 1963 38.86 1964 45.90 1965 40.42 9.72 30.70 1966 32.68 9.78 22.90 1967 30.52 10.53 19.99 1968 30.24 10.57 19.67 1969 26.86 12.10 14.76 1970 26.16 8.65 17.51 1971 23.44 9.49 13.95 1972 21.30 10.60 10.70 1973 17.74 9.20 8.54 1974 25.06 14.40 10.66 1975 32.38 14.01 18.37 1976 24.37 15.53 8.84 1977 23.07 15.86 7.21 1978 24.60 15.24 9.36 1979 14.46 1980 14.43 1981 15.71 1982 17.53 1983 16.44 1984 15.97 1985 15.46 1986 13.51 1987 13.90 1988 11.87 1989 11.36 Soutrce: Korea Small and Medium Industry Promotion Corporation 1994; Taiwan, Ministry of Economic Af- fairs 1991. Yet in aggregate, these concessionary loans amounted to only 3.1 percent of all exports in 1986, or less than 1 percent of all bank lending in 1987 and 1.6 percent of total lending in 1985.21 Small firms accounted for 70.05 percent of Taiwan's 1986 manufactured export sales.22 Assuming export loans for small firms proportionate to their share of exports, con- cessionary export loans to small firms amounted to 5.74 percent of all bank lending to small firms and 2.97 percent of total small firm borrowing from all sources. The large amount of money being supplied even today by informal sources is clear evidence of the limited scope of political exchange in financing rural entrepreneurs; it is wrong to conclude that informal political access is available to all. Even so, the high cost of informal borrowing (see table 9-4)23 makes borrowing from the banking system the most Political Bases of Rtural Industrialization 227 attractive option for small business. But bank credit can satisfy only a small slice of overall demand. Recourse to the political process can provide entrepreneurs in Taiwan, China, with preferential compensatory treatment. Whether political exchange results in the ac- quisition of a stable source of cheap funds, or simply a loan guarantor in the form of a local or provincial politician to ensure future borrowing, the political process has value to any Taiwanese entrepreneur, urban or rural. Particularly for rural businessowners lacking the variety of financing options available to their urban counterparts,24 politics is a significant opportunity for economic gain. The correlation between lending to small rural firms and the level of bank politicization pro- vides circumstantial evidence for this suggestion. Table 9-5 shows that the three provincial conmnercial banks-First, Zhanghua, and Hua Nan-make nearly 70 percent of their loans to small firms. They are the only banks that make a majority of their loans to small enter- Table 9-5. Bank Lending to Small Firms, Taiwan, China, 1986 (thousands of $NT) SMEI total lendinig Bank namiie Total l,endintg SME lendinig (percent) Commercial banks Bank of Taiwan 212,143,564 27,034,063 12.7 *Hua Nan Commercial Bank 125,181,340 83,871,271 67.0 *First Commercial Bank 120,368,761 83,290,037 69.2 *Zhanghua Commercial Bank 113,293,682 77,039,704 68.0 Taipei Bank 72,696,093 5,294,412 7.3 Shehua United Commercial Bank 21,805,048 718,057 3.3 Huaqiao Commercial Bank 12,633,003 5,429,271 43.0 Shanghai Commercial Savings Bank 6,023,586 2,383,234 39.6 Gaoxiong Bank 5,961,497 1,273,371 21.4 Agricultural banks Cooperative Bank of Taiwan 199,111,294 97,404,562 48.9 Taiwan Land Bank 188,795,042 9,544,362 5.1 Farmers Bank of China 64,505,032 6,168,565 9.6 Trade banks International Commercial Bank of China 35,089,234 2,923,066 8.3 China Export-Import Bank 10,728,776 173,556 1.6 Development banks Communications Bank 95,439,098 3,185,070 3.3 Central Trust Bureau 21,814,570 215,764 1.0 Total 1,305,589,620 405,948,365 31.1 Note: Data as of June 30, 1986. Source: Small Business Integrated Assistance Center (1986: 37). 228 Totward the Rtural-Based Development of Cotmmerce and Inidustry prises and the only banks that are administered by local rather than central government. Although data do not permit an accurate breakdown of commercial bank lending by loca- tion, that the three commercial banks have around 300 branches each throughout the is- land suggests more than an urban business focus. Crudely assuming that the three commercial banks lend equally to rural and urban businesses, loans to rural firms by these three institutions alone account for 46.14 percent, or almost half, of all lending to all small firms, urban and rural. Conversely, loans to small urban manufacturers from all commer- cial banks amount only to 12.93 percent of total small firm lending.25 Political pressure on the officers of the three provincially administered banks carries more than nuisance value. As Sender (1994: 60) observed: Provincial assemblymen have a say in everything from staff appointments to budgets to buying cars for overseas offices. In return for cooperative hearings on operating budgets and a recommendation for promotion, bank executives are expected to make the politicians' lives easier. This may mean arranging loans for their acquaintances, buying land at exorbitant prices, or providing campaign funds at low interest rates. Budget hearings take up two months of every year and require the presence of the senior bank management in the provincial capital for extensive lobbying. And despite excellent credentials, to gain appointment as president of the Hua Nan Bank, Edward Chien found it necessary to visit each of the eighty or so provincial assembly members to secure suffi- cient support (Sender 1994: 61). The reach of county-level factional elites into local finance is deep: Chu (1989: 158; cited in Wang 1994: 185) claims that nearly 70 percent have KMT-sanctioned interests in coop- erative banks and the few private banks, and 60 percent gain from similar ties to credit associations. Chen (1989: 17-18; cited in Huang 1990: 736) goes further, finding that al- most all local factions (eighty-one out of eighty-nine) monopolize some form of public resource, most commonly cooperative banks, credit cooperatives, farming and other co- operatives, and passenger transport companies. Moreover, 442 of 518 candidates for the provincial assembly affiliated with local factions controlled at least one such monopoly. Only 72 of 774 candidates unaffiliated with a local faction had similar direct interests. Factional control of economic resources has electoral implications: between 1954 and 1985, only 38.2 percent of candidates for the provincial assembly were affiliated with local po- litical factions, yet they won 61 percent of the contested seats (Huang 1990: 731-32). Inter- ests in these economic institutions serve a rural constituency as well as the politicians themselves; lineage and clientelist norms require that their fruits be shared. Anthropolo- gists find credit relations are particularly affected by such personal ties in Taiwan, China (Deglopper 1972; Gallin and Gallin 1974); in effect, local politicians act as gatekeepers to capital for entrepreneurs. In other areas of economic interest to rural entrepreneurs as well, Taiwan, China's, local politicians manipulate the rules to benefit themselves and their constituents. With account- ability and often multiple ties to individual constituents, and with institutional access to political resources through their official positions, local politicians are in a position to in- Political Bases of Ruiral Inidutstrialization 229 fluence zoning and licensing and to confer other advantages (for example, in locating bus stops) that target business opportunities at particular constituents and locations.26 At the central government level, as much as an upsurge might be expected, the legaliza- tion of opposition parties and the strengthening of the legislature since the mid-1980s has not resulted in wholesale realignment of small industrialist behavior away from local clientelism toward the lobbying of the central government. While there are a number of plausible explanations for this non-event, including the collective action problem at the central level and the gradual elimination of credit scarcity conditions, only one follows directly from my cost-based political argument: that the unknown cost of lobbying the central government for indeterminate benefits outweighs the known cost of obtaining the tangible benefits secured by the provincial assembly member's capture of government resources. Small industry's lobbying costs at the central level are indicated both by the challenge of matching the favorable lobbying position of big business in general and by the particular conflicts of interest posed by members of large industrial families holding seats in the national legislature. While legislators at the national level must now seek popular support partly in the form of small-business votes, their constant search for funds also results in consideration for the often antithetical positions of potential big-business backers. Lim- ited national-level organization by small business, on the one hand, in part the product of central government's past licensing requirements and agenda control over interest asso- ciations, and the substantial investment required to learn to lobby the central government, on the other, effectively suggest the continued importance of local rather than central small- business political activity to sustain entrepreneurial vitality at the local level. The case of Taiwan, China, shows that the constraints on political action imposed by the central government's domination of policymaking and its restrictions on political lobby- ing can be mitigated by the synergies generated through linkage between local govern- ment structure and underlying social organization. From the perspective of principal-agent theory, the dynamic of electoral accountability established by the provincial assembly and Taiwan's other local elected offices forces the KMT to accept significant agency costs. From the perspective of rural entrepreneurial vitality, however, the space for political exchange permitted by these agency losses reduces policy and economic constraints on local indus- trial viability. As such, the design of political institutions contributes to the success of small industry. In Korea, however, tight control over not just political access but also over the power of local politicians to distribute state resources and implement state rules raises the cost to small business of political action, closes off opportunities for state structure to inter- act with local social structure, and thereby limits would-be entrepreneurs either to expen- sive economic self-help strategies, such as curb market borrowing, or to market exit and urban migration. The absence of elections for local officials and the restrictions placed on interest associa- tions in Korea have effectively precluded the political representation of small industry. Central control is accomplished by several means. The local bureaucracy acts without leg- islative review, but its authority is superseded by all central government decisions. The 230 Tovward the Rutral-Based Developmenit of Conmmerce and Indtustry Ministry of Home Affairs affects local governance by reviewing local budgets, supervis- ing local officials, recommending new appointments, and controlling the police. Promi- nent local officials (about 17 percent of the total) are members of the central civil service, and the most important of these are central higher civil service members. Influential fig- ures such as the provincial governor are rotated rapidly to avoid local entrenchment (Jacobs 1985: 33-34).7 Appointed by the central government, Korea's local officials are not moti- vated by local conditions, an interest in local development, or by any need to cultivate authority over the local populace, but rather by the hope of acquiring a good reputation in Seoul's halls of power. Gradual change in this situation with limited local elections in 1991 culminated in June 1995 elections for provincial governors and urban mayors. Yet while these elections imply greater local government accountability, the continued retention by the central government of political control, state resources, and regulatory power means only marginal change in the constraints on political action small entrepreneurs face. Entre- preneurship in rural Korea has been largely stifled by state organization for centralized control, leaving little space for local political exchange. Collective political action also has only limited effect in Korea. Although the Chambers of Commerce have demonstrated their independence as interest groups in recent years, their policy goals do not necessarily overlap those of rural entrepreneurs. Other interest groups embracing small-firm interests are creations of the state; they receive government funding and staff and implement state policy. They collect data on small business and consult with officials in the economic bureaucracies in an ad hoc manner, but the bureau- cracy relies upon the president and his staff rather than the private sector to steer policy change. An annual Small Business Day allows small-firm representatives to meet with the Korean president and articulate their views, but few other institutionalized opportunities exist for small firms to broach their concerns to Korean policymakers. Faced with limited access to those with the power to lessen the economic and political constraints belaboring rural entrepreneurs, and plagued by weak demand for subcontracting from the vertically integrated chiaebol,28 only informal borrowing and other informal, individual activities re- main as costly but necessary self-help strategies for Korea's small entrepreneurs." The democratization process in Korea, like that in Taiwan, China, has meant little to Korea's rural entrepreneurs in practice; presidential elections and a freer press have nei- ther significantly eased their access to the policy process nor sanctioned new forms of political participation. Gains will come to Korea's rural entrepreneurs only when specific institutional changes are made that induce local government to focus on local constituents and permit local officials the discretion to implement policy and utilize state resources. Japan Postwar Japan presents a very different situation. Bureaucratic records indicate that the Ministry of International Trade and Industry's (MITI) predominant small-business policy goal in the early postwar years was continued industrial rationalization,m even at a time of high unemployment, social stress, and Occupation policies aimed at the dissolution of Political Bases of Rutral I,idutstrialization 231 concentrated economnic power. Such a policy meant continued hardships and tight finances for all small entrepreneurs. Yet numerous policy measures were enacted by the Diet to aid small-business financing: in 1951 bank reform laws were passed creating new classes of financial institutions specifically for small business (the sOgo gink6 and shin'y6-kink6); state corporations were established in 1949 and 1953, again to lend specifically to small enter- prises;3" and in 1956 a statute was enacted to protect the usually small-scale subcontrac- tors from late payments for services rendered. In this case, a political dynamic operated to overcome the designs of the powerful economic bureaucracy. The rapid postwar lifting of restrictions on independent interest associations produced an explosion of well-organized vocational associations.32 At the same time that postwar democratization legitimized interest articulation and lobbying, it also strengthened the powers of the Diet relative to other government organs (cf. the prewar and wartime peri- ods), incited competition for supporters among the emergent political parties, and, finally, given the absence of strong central party organization, encouraged individual national- level legislators to develop their own local networks of supporters.33 Taken together, the extensive organization of small industrialists and the incentives to politicians to seek out and cultivate new constituencies eased access for small entrepreneurs to the political pro- cess and encouraged political action as a way of adjusting to constraints on the availability of investment capital. Thus, under a particular set of institutional and organizational cir- cumstances, small entrepreneurs could both articulate their interests and push politicians for policies that addressed their concerns over the stated interests of the bureaucracy.34 Political Institutions and Social Organization Efficiency and Local Politics Political institutions facilitate economic outcomes in other ways as well. For example, with knowledge of the consequences to economic efficiency of the legislative influence of Japa- nese and Korean farmers, one might justifiably question how access to the political sys- tem, as suggested by the case of Taiwan, China, has possibly managed to redound to the benefit rather than to the detriment of local development. In certain cases, however, social ties and political institutions can create the conditions that allow political action and lob- bying to result in social gains rather than social costs. First, local politicians may define their personal interests in terms of local development if they believe that gains from local development will benefit them. Self-interest and com- munity interest will overlap when there is an incentive to value absolute over relative gains. This situation is most likely to obtain when, as is the case for Taiwan, China, eco- nomic and political exchange between political agents and their constituencies is person- alized and linked by family, clan, or other social ties.35 Second, political institutions can lengthen time horizons. Although not necessarily mov- ing politicians to concern for long-term constituent welfare, a long-term view can at least 232 To7ward thte Rulral-Based Development of Commerce an1d Indutstry discourage the political predation on the private sector commonlv visible in the develop- ing world by reducing the pressure on politicians to value power maintenance and oppor- tunity for personal gain over public service. Long time horizons can obtain under conditions like those found in Taiwan, China, where the electoral rules, including those rewarding vote- and office-buying, produce a lock on electoral districts and offices for politicians (Rigger 1993). Contrast this situation with that in Korea, where short time horizons and two-year rotations are the norm for local political appointees, or to that elsewhere in the developing world, where weak institutions produce precarious tenures for politicians and prompt short-term rapacity. When the political rules of the game do not provide politicians with the confidence to ignore considerations of political survival, politicians became obsessed with the mainte- nance of their political authority and legitimacy and use all available means to perpetuate it. This also may result in benefits for constituents in the form of policies and perks aimed at buying support, although such gains are likely to be transitory in the absence of rules that help to secure long incumbency. Appointed officials in Korea, however, almost cer- tainly will not care to seek out local political support, for they govem at the pleasure of their superiors in Seoul. Political authority and legitimacy may facilitate local acceptance, but they are not required for appointed politicians to maintain their positions. Finally, the existence of clientelist social norms suggests that all political support comes to a politician with clear reciprocal obligations and that elected politicians will direct re- sources and opportunities toward their supporters accordingly. Indeed, to a greater de- gree than complex and atomistic urban society can, rural patron-client ties and factional organization enmesh politicians and constituents in obligatory yet instrumental roles- ensuring credible commitments. Absent clientelist obligations, little pressure exists for appointed politicians in rural Korea, even if of local origin, to share personal or institu- tional access to government resources with their constituencies. Politics and Patron-Client Relations As is obvious with a moment's thought, patron-client ties exist everywhere, but produce varying effects. As an explanatory variable, therefore, patron-client relations cannot stand on its own, but must be tied to a theory that explains the way in which clientelism matters. An understanding of patron-client relations and their role in encouraging entrepreneur- ship benefits, therefore, from an awareness of political conditions and institutions. Clientelism in Taiwan, China, enables rural entrepreneurship by joining with a particular political structure to direct otherwise unavailable state resources toward local purposes. Clientelism and the social organizations that distribute its spoils do not create entrepre- neurship; they simply facilitate its emergence. Because state structure affects two sets of actors-government actors and societal actors-clientelism can act to exploit the interests of both for mutual benefit. Korea's state structure allows little place for the political assertion of local economic interests, mainly because the authority of local political appointees derives from the center Political Bases of Rutral Inidutstrialization 233 rather than the locality, and because local officials have less flexibility on policy than elected officials might. Together with the high economic cost of entrepreneurship, finance, and subcontracting, the inaccessibility and high cost of political recourse precipitates rural market exit among potential industrialists and contributes to Korea's flaccid rural indus- trial sector. Altered state structure should lower the cost of political action, however, and the local elections of June 1995 should produce new coalitions of local entrepreneurs, both economic and political. Policy and Enitrepreneuirship In addition to the effects of interacting political rules and social organizations, policy may also create openings for entrepreneurship. Contrasting Korean and Taiwanese stances to- ward foreign direct investment are a case in point. Korea prefers foreign lending to foreign direct investment: for most years since 1965, foreign direct investment has amounted to less than 5 percent of total foreign capital in- flow (Amsden 1989: 76). Taiwan, China, in contrast, has tried to attract multinational cor- porations for both economic and national security reasons since the early 1960s, first in labor-intensive and then in technology-intensive industries. In Korea in 1975, only 17.8 percent of foreign-firm subsidiaries were 95 to 100 percent parent-owned. In Taiwan it was 45 percent, compared with a figure for the rest of the developed and developing world of 67.4 percent.36 T'he Korean state acts both as gatekeeper to foreign investment and as the agent of tech- nological transfer, approving for direct investment only projects that promote sanctioned production goals, sectors, and technologies, and for which licensing arrangements are unavailable. Foreign firms normally are obliged to take a Korean joint-venture partner and surrender equity after transferring effective know-how (Mardon 1990). But when that know-how is monopolized by the Korean joint-venture partner, few market mechanisms exist to ensure the diffusion of the transferred technology to other Korean firms. The mo- nopolist has no incentive to share knowledge and, in the absence of competitors, no reason to improve its efficiency. In Taiwan, China, too, the state encouraged foreign firms to transfer new technology, but Taiwan allows foreign-owned factories to hire local labor and produce for both export and domestic use; Taiwan has not succeeded in barring entry to all firms that compete with domestic producers. Technology transfer is accomplished primarily through work- ers in the multinational's factories who are trained but quickly leave their employers to start their own firms, often as subcontractors to the foreign producer.37 How can such a dynamic aid rural industrialization? Multinational spin-offs have little reason to leave the urban marketplace, particularly if the parent that provides work orders is based there. General Instruments, one of the earliest foreign firms to set up in Taipei, is now surrounded by smaller subcontractors, founded and run by its former employees. Taiwan, China's, policies on foreign direct investment benefit rural industry only insofar as social norms are given their head to locate subcontracting and supply sources in the 234 Toward the Ruiral-Based Developnmentt of Conmnerce anid hidistry countryside. Extended families and lineage ties reduce the costs of investment and the risks of failure. The political exchange produced by the interaction of state structure and clientelism can also reduce capital and other business costs. Nevertheless, the required social networks and clientelist organization are entrenched only in the countryside; urban associations are more varied but less robust. Conclusions Broad political factors have as much power to shape markets as the specifics of policymaking. Politics cannot be defined out of markets, but this does not mean that the market mechanism cannot operate freely within the constraints that government, policy, and other variables impose. Analysis of the bases of rural industrialization and the mani- festation of entrepreneurship is improved, therefore, wherever it is possible to explain the existence and implications of these political factors. This chapter suggests that only a po- litical-economy approach to economic problems makes such an explanation possible. Through politics, interests compete for relative advantage and distributional gains. Those who control resources often distort economic rationality and cannot always be persuaded to surrender their own stakes for disinterested administration. Accordingly, one of the main themes here is the need to direct attention toward the political context in which economic development occurs so as to note and harness political interests for develop- mental ends rather than to dissipate developmental resources in ignorance of political constraints on the market. The contrast between Korea, where development policy measures have failed partly because of the obstacles to entrepreneurship posed by state structure, and Taiwan, China, where state structure assists local development by rewarding political and social interests, is a case in point. The case of Taiwan shows first that political structure has consequences for economic as well as political outcomes, and second that institutional variation can significantly alter the opportunity set for small entrepreneurs. Although Taiwan's vital rural industrial sector may be the product of an unintended synergy among political struc- ture, actor interests, and underlying social norms of interaction, the understanding that political structure is a matter of political choice, and that actor interests are often shaped decisively by the incentives organizational and institutional structure provide, suggests that regardless of the historical path that led to positive outcomes for small entrepreneurs in Taiwan in the face of central government control over access to policymaking and in- vestment capital, similarly beneficial conditions can be created as a matter of political choice elsewhere as well. The general lesson is that although cross-country differences suggest that the origins of variation in state structure are historically and socially based, this should not block policymakers from realizing that changes in both political organization and institutions can auger not only a new style of politics with new political actors, but also previously unobtainable consequences for the economy and economic actors. In contrast to the differ- Political Bases of Rural Industrialization 235 ences between Taiwan, China, and Korea in infrastructure and economic diversity, the elimination of which required years of effort, the differences in political structure between Taiwan and Korea that block rural entrepreneurship are products of conscious decision and can be altered much more quickly. Notes 1. Other assumptions include the following: (a) Entrepreneurs prefer to maximize their access to capital as a means of maximizing growth. Relational subcontracting here is thus either a source of capital or a cost-reducing strategy. (b) Entrepreneurs need significant amounts of stable capital for start-up or technical upgrading in such areas as machine tools, electronics, metalworking, or even bicycles and carbon-fiber tennis racquets. Weaving and garment sewing are not typically capital- or investment-intensive processes. (c) There are significant constraints on the availability of such capi- tal from the formal banking system or from family and friends, regardless of whether those con- straints are products of market or govemment failure. 2. The difficulty, of course, is in obtaining data with enough detail to make the actual rural/ urban boundaries clear. 3. The lack of separation between provincial cities and the less densely populated countryside in the data probably also masks a net migration to the smaller cities. 4. Author's interview of Mr. Ducksoo Lee, former AFDC vice-president, November 26, 1994. 5. Bates (1981) suggests how agricultural policies bom of a ruler's desire for political self-preser- vation can result in both urban and transnational migration and in the selective provision of agri- cultural inputs and infrastructure. 6. This description relies upon Kim (1991: 11-15). 7. The Saemaul ("New Community") movement was a 1970s attempt by Park Chung Hee to consolidate political support for his regime at the grassroots level. Saemaul's ideology called for self-sacrifice, hard work, and political loyalty. Its economic aspects included attempts to slow the pace of urbanization and to support the rural economy by encouraging firms to relocate to the countryside; hence the term Saemaul factories. 8. Author's interview with a senior analyst at the Korea Development Institute, Seoul, October 1995. Note that the Saemaul movement constituted a general effort to support agriculture and re- duce pressures for urbanization; it was not a developmental program for rural industry alone. 9. See Yager (1988: 122-23) and Simon (1988: 144-46). 10. On Taiwan, China's, public enterprises and the KMT-state enterprise nexus, see Wu (1992) and Chen and others (1991). One summary of grain procurement policy and the rice-fertilizer bar- ter system is Yager (1988: 258-60). 11. Author's interviews, October 1992-August 1993. This phenomenon creates a difficult mea- surement problem, for there are few ways to distinguish from the outside small companies spun off from a parent for tax avoidance purposes from similar entities born of less instrumental entrepre- neurial motivation. More to the point, it is difficult to know when economic factors alone justify or explain a particular form of business organization. 12. For Japan, see Patrick and Rohlen (1988: 366-68); for Taiwan, China, author's interviews, 1992-93. Observers of the Korean scene also frequently comment that much business takes place off 236 Tozvard the Rnral-Based Development of Commerce and Industry the books, partly because the Korean government takes audits seriously as a political tool. The most prominent case involves audits of Hyundai Corporation and charges against its chairman, Chung Ju Yung, in response to his political ambitions (for a summary see Koh 1992: 7-8). Clifford (1994: 308-09) notes actions by tax authorities to enforce policy objectives on a much smaller scale at the grassroots level. Note that these anecdotes illustrate the plain connection between regulation and behavior; they clearly do not define the behavioral effects of the tax code or imply that stringent or lax enforcement explains variance in the rural industrial vitality of Korea and Taiwan. 13. Author's interviews, autumn 1994 in Korea and spring 1993 in Taiwan, China; confirmation for the Taiwan case in remarks by Wing Thye Woo at the World Bank workshop on the "Role of Govemment in the Evolution of System Change, " Hakone, Japan, December 19, 1993. 14. Olson (1965) holds that for large numbers of individually rational actors, large-scale collec- tive action is discouraged in the absence of selective incentives preventing free-riding by some on the organizational efforts of others. Bates (1981) suggests that low rural population densities also discourage collective action by raising coordination costs. This claim is challenged in East Asia, however, by high population densities and the relatively short physical and psychological distance between city and countryside. 15. Author's interview with former Minister of Finance Shirley Kuo (Kuo Wanrong), July 31, 1993. 16. Interfirm lending allegedly accounts for just 3 percent of all informal lending, but more may be subsumed in the "nonprofit financial groups" (feiyingli tutanti) and "personal" funding catego- ries. Economic Research Department, Central Bank of China, Flowv of Flunds in the Taitvan Area, Re- putblic of Chin7a, cited in Huang (1988: 214). 17. For figures on the number of small firms, see Small and Medium Enterprise Agency (1991: 257). For lending figures (36.43 percent), see Small Business Integrated Assistance Center (1986:38). Note that the eight small and medium-size enterprise banks, chartered with the mission of lending to Taiwan's smaller firms, manage to divert one-quarter of their funds to large enterprises. Their impact is significant but small. Total lending in 1986 by the small enterprise banks amounted to only 13.2 percent of the ordinary banks' lending, and their lending to sniall firms amounted to 32.6 per- cent of the ordinary banks' lending. SMEI SME banks Ordinary banks ordinary banks SME lending (NT$) 132,307,934 405,948,365 32.59% Total lending (NT$) 172,019,709 1,305,589,620 13.18% SME/total lending 76.91% 31.09% 18. Office of Economic Research, Bank of Taiwan (1987: 9). 19. Economic Research Department, Central Bank of China, Zlhonghlua Minguto Taiwan Dichut Jinrong Tongji Yuiebao (Monthly financial statistics, Taiwan area, Republic of China), cited in Ya-Hwei Yang (1989: 48, 50). 20. Small and Medium Enterprise Agency (1991: 288). 21. Cases exist of interest rates exceeding 10 percent weekly. Author's interviews in Taipei, au- tumn 1993. 22. For a sense of the diversity of informal lending options, see Biggs (1988: 21-33). Political Bases of Rutral Induestrialization 237 23. The proportion of small rural and urban industrialists was taken from table 9-1; lending fig- ures are from table 9-5. 24. Author's interviews, spring 1993, especially with Professor Yun-Han Chu and Judge Zhiming Jiang, Taizhong Regional Court. This is not the place for a normative evaluation of informal politi- cal behavior, but it is worth noting that political corruption redounds to socially beneficial ends in Taiwan, China, by creating opportunities for entrepreneurial vitality in otherwise economically hostile environments. Without the possibility of political intervention or without a politician's guar- antee, for example, central government restrictions helped ensure that potential entrepreneurs had only limited access to formal credit markets until the mid-1980s. 25. For an incisive study of Korean authority relations, see Jacobs (1985). Another influential study is Henderson (1968). 26. Since the protracted labor disputes of 1986-87, however, the cliebol have increasingly sought out small firms as subcontractors in order to spin off the now more costly, but lower value-added, production tasks previously kept in-house. 27. As an example of the costs involved, informal rural borrowing, mostly for productive invest- ment, sent farmer debt levels skyrocketing over the course of the 1980s, to the point that the central government intervened with several bailout packages of a trillion won each. According to one re- port, average debt levels in 1985 alone exceeded farmer financial assets by 15 percent, making their liquidation insufficient to repay the loans. See Suh Myong-soo, "Farm Debt Relief Measures to Infuse Vigor to Rural Front," Korea Herald, March 17, 1987. 28. In particular, see Japan, Ministry of International Trade and Industry (1963: 365-70). While rationalization normally connotes productivity and efficiency gains, the Japanese term also bears the historical implication of consolidation or liquidation for inefficient firms through merger or other means. 29. The People's Finance Corporation (Kokumin Kin'yet Koko) was established in 1949 for small business in general. The Small Business Finance Corporation (Chuish6 Kigyo Kin'yu K6ko) was established in 1953 to serve small industrialists in particular, and was funded by MITI's budgetary requests under the new Fiscal Investment and Loan Program. 30. Some of these were the product of the state's wartime organization of production, but after the war they shifted from representing the state's interests toward representing those of their members. 31. See Masumi (1992) on the evolution of politician support groups. 32. Ramseyer and Rosenbluth (1993: 125-26) offer a summary of the case of Japan's credit unions (nnijink6), which also blocked bureaucratic pressures for the consolidation of small financial institu- tions by taking their concerns to the politicians. 33. In the absence of such ties, policy incentives can accomplish the same purpose. Compare China's tax structure, which increasingly allows local government to retain some portion of tax revenue: whether rural or urban, local officials have since become converts to the cause of local economic growth. See Oi (forthcoming). 34. The figure for Japan was 18.5 percent (Curhan, Davidson, and Suri 1977: 316-17). 35. Paper presented by Danny Kin-Kong Lam at Harvard University, August 26, 1992. The phe- nomenon is not restricted to MNC employees, but is widespread and is reflected in high urban employment turnover. 238 Toward the Rural-Based Developmentt of Comnterce and Industry References Amsden, Alice. 1989. Asia's NeTv Giant: Soufth Korea and Late Indutstrialization. New York: Oxford University Press. .1991. "Big Business and Urban Congestion in Taiwan: The Origins of Small Enterprise and Regionally Decentralized Industry (Respectively)." World Developnment 19(9) 1121-35. Bates, Robert. 1981. Markets and States in Tropical Africa. Berkeley: University of Califomia Press. Biggs, Tyler S. 1988. "Financing the Emergence of Small and Medium Enterprise in Taiwan: Hetero- geneous Firm Size and Efficient Intermediation." Employment and Enterprise Policy Analysis Discussion Paper 16. Washington D.C.: United States Agency for Intemational Development. Caldwell, J. Alexander. 1976. "The Financial System in Taiwan: Structure, Functions, and Lessons for the Future." Asian Suirvey 16(8): 729-51. Chen, Ming-tong. 1989. "Local Economies of Collective Monopoly, Local Factions, and Elections of Taiwan Provincial Assemblymen." Unpublished report to the National Science Council of Taiwan's Executive Yuan. Taipei: National Science Council. Chen, Shih-Meng S., Chung-Cheng Lin, C. Y. Cyrus Chu, Ching-hsi Chang, Jun-ji Shih, and Jin-Tan Liu. 1991. Jiegout Dang-Guo Zibenzhutyi: Lun Taizwan Guianyinigshtiye zhi Minzyinightua (Disintegrating KMT-State Capitalism: A Closer Look at Pritatizing TaiTvan's State- and Party-Owntved Enterprises). Taipei: Taipei Society. Chu, Yun-han. 1989/1992. "GuazhanJingji yu Weiquan Zhengzhi Tizhi" ("The Oligopolistic Economy and Authoritarian Political System"). In Taiwan Yanjiu Jijinhui, ed., liepon Taiwan lingji: Weiquanl Tizhixia de Longduan yut Boxute (Dissecting Taiwan's Economy: Monopoly and Expectation tender an Autthioritariani System). Taipei: Taiwan Research Fund. Clifford, Mark L. 1994. Troubled Tiger: Businessnen, Bureaucrats, and Genemls in South Korea. Armonk, N.Y: M. E. Sharpe. Cole, David C., and Yung Chul Park. 1983. Financial Developmienzt in Korea, 1945-1978. Cambridge, Mass.: Harvard University Press. CEPD (Council for Economic Planning and Development, Republic of China). 1988. Taiwan Statisti- cal Data Book, 1988. Taipei: Council for Economic Planning and Development. Curhan, Joan P., William H. Davidson, and Rajan Suri. 1977. Tracing the Muiltinationials. Cambridge. U.K.: Cambridge University Press. Deglopper, Donald R. 1972. "Doing Business in Lukang." In W. E. Wilmott, ed., Econonmic Organiza- tion in Chinese Society. Stanford, Calif.: Stanford University Press. Feuchtwang, Stephan. 1981. "Blending of Mainland Religious Influence in Taipei." In James C. Hsiung, ed., The Taiwvan Experience, 1950-1980. New York: Praeger. Gallin, Bernard, and Rita Gallin. 1974. " The Integration of Village Migrants in Taipei." In Mark Elvin and G. William Skinner, eds., Thze Chinese City Betveeni Twvo Worlds. Stanford, Calif.: Stanford University Press. Gold, Thomas B. 1988. "Colonial Origins of Taiwanese Capitalism." In Edwin A. Winckler and Susan Greenhaigh, eds., Contendinig Approachtes to the Political Economtiy of Taiwan. Armonk. N.Y.: M. E. Sharpe. Henderson, Gregory. 1968. Korea: The Politics of the Vortex. Cambridge, Mass.: Harvard University Press. Ho, Samuel P. S. 1982. "Economic Development and Rural Industry in South Korea and Taiwan." World Development 10(11): 973-90. Political Bases of Rural Indutstrialization 239 Huang, Teh-Fu. 1990. "Local Factions, Party Competition, and Political Democratization in Tai- wan." Joturnal of National Chengchli University 61: 723-45. Huang, Yongren. 1988. "Zhongxiao Qiye Jinrong Tixi de Jiantao yu Zhanwang" ("An Evaluation and Forecast of the Small and Medium-Sized Enterprise Financial System"). Jiceng Jingrong (Grassroots Finance) 17 (September): 201-26. Jacobs, Norman. 1985. Korean Road to Modernization and Development. Urbana and Chicago: Univer- sity of Illinois Press. Japan, Ministry of International Trade and Industry. 1963. Shoko seisaku shi (History of Indutstrial Policy). Vol. 12. Tokyo: Ministry of International Trade and Industry. Johnson, Chalmers. 1982. MITI and the Japanese Miracle: The Growth of Indtustrial Policy, 1925-1975. Stanford, Calif.: Stanford University Press. Kim, Jong-Gie. 1991. "Rural Industrialization Policy in Korea: Past Performance and Future Direc- tion." KDI Working Paper No. 9112. Seoul: Korea Development Institute. Koh, Byung Chul. 1992. "Domestic Politics and External Relations." In Donald N. Clark, ed., Korea Briefing. Boulder, Colo.: Westview. Kuo, Shirley W. Y. 1983. The Taiwan Economy in Transition. Boulder, Colo.: Westview. Lerman, Arthur J. 1978. Taiwan's Politics: The Provincial Assemblyman's World. Washington, D.C.: University Press of America. Mardon, Russell. 1990. "The State and Effective Control of Foreign Capital: The Case of South Ko- rea." World Politics 43(1): 111-38. Masumi, Junnosuke. 1992. "The 1955 System: Origin and Transformation." In Tetsuya Kataoka, ed., Creatinig Single-Party Democracy: Japan's Postwar Political Systemn. Stanford, Calif.: Hoover Institu- tion Press. North, Douglass. 1990. Institutions, Instittutional Chanige, and Economic Performance. New York: Cam- bridge University Press. Office of Economic Research, Bank of Taiwan. 1987. Zhonghuia Miugnio Taiwan Dichut Gongye Caiwn Zhuiangkutang Diaocha Baogao (Survey Report on7 the Finanicial Condition of Industry, Taiwan Area, Re- public of China) 28 (1985-86). Taipei. Oi, Jean C. 1989. State anid Peasant in Conitemporary China: The Political Econonmy of Village Government. Berkeley: University of California Press. . Forthcoming. Rural China Takes Off: Incentives for Induistrialization. Berkeley: University of California Press. Olson, Mancur. 1965. The Logic of Collective Action. Cambridge, Mass.: Harvard University Press. Patrick, Hugh T., and Thomas P. Rohlen. 1988. "Small-Scale Family Enterprises." In Kozo Yamamura and Yasukichi Yasuba, eds., The Political Economy of Japan. Vol. 1, The Domestic Transformation. Stanford, Calif.: Stanford University Press. Ramseyer, J. Mark, and Frances McCall Rosenbluth. 1993. Japan's Political Marketplace. Cambridge, Mass.: Harvard University Press. Ranis, Gustav. 1979. "Industrial Development." In Walter Galenson, ed., Economic Growth and Struc- tuiral Change in Taiwan: The Postwar Experience of the Repu7blic of China. Ithaca, N.Y., and London: Cornell University Press. Rigger, Shelley. 1993. "Electoral Strategies and Political Institutions in the Republic of China on Taiwan." Harvard University, Fairbank Center Working Papers No.1. Cambridge, Mass.: Fairbank Center for East Asian Research. Sender, Henry. 1994. "Borrowed Time." Far Eastern Economic Review 157: 60-63. 240 Towvard the Ruiral-Based Development of Comnmerce and Industry Shea, Jiadong. 1990. "Financial Development in Taiwan: A Macro Analysis." Paper presented at the Conference on Financial Development in Japan, Korea, and Taiwan, Academia Sinica, Taipei, Taiwan, August 27-28,1990. Simon, Denis Fred. 1988. "Extemal Incorporation and Internal Reform." In Edwin Winckler and Susan Greenhalgh, eds., Contending Approaches to the Political Economy of Taiwanl. Armonk, N.Y.: M. E. Sharpe. Small and Medium Enterprise Agency, Taiwan. 1991. Zhongxiao Qiye Baipishu (W1tite Paper on Small and Mediiumn-Size Enterprises). Taipei: Ministry of Economic Affairs. Small Business Integrated Assistance Center, Taiwan. 1986. Zhongxiao Qiye Jinrong Nianbao (Smiall and Meditmni-Size Enterprise Finanicial Yearbook, 1986). Taipei: Small Business Integrated Assistance Center. Wade, Robert. 1985. "East Asian Financial Systems as a Challenge to Economics: Lessons from Tai- wan." California Ma\nagement Reviezw 27(4): 106-27. . 1990. Governing the Market: Econonmic Theory an1d the Role of Government in East Asian Indutstrial- ization. Princeton, N.J.: Princeton University Press. Wang, Fang. 1994. "The Political Economy of Authoritarian Clientelism." In Luis Roniger and Ayse Gunes-Ayata, eds., Democracy, Clientelismii, and Civil Society. Boulder, Colo.: Lynne Reinner. Winn, Jane Kaufman. 1986. "Decriminalizing Bad Checks Should Help to Rationalize Taiwan's Fi- nancial System." East Asia Executtive Reports 8(8): 9-22. Woo, Jung-en. 1991. Race to the Swift: State and Finance in Korean Indutstrialization. New York: Colum- bia University Press. Wu, Ruoyu. 1992. ZhanlhouL Taiwan Gonigying Shiye zhi Zhengjinig Fenshi (An Analysis of the Political Econony of Postzvar Taiwvan's Putblic Enterprise). Taipei: Yeqiang Publishing. Yager, Joseph A. 1988. Transforminig Agricultutre in Taiwan: The Experience of the Joinzt Commnissioni on Rutral Reconstruictioni Ithaca, N.Y.: Cornell University Press. Yang, Ya-Hwei. 1989. "Taiwan zhi Zhengcexing Rongzi Tixi" ("Taiwan's Policy-Based Financial System"). Jiceng Jinronig (Grassroots Finance) 18 (March): 45-60. Yokokura, Takashi. 1984. "Chfish6 kigy6" ("Small and Medium-Sized Enterprises"). In Komiya Ryutaro, Masahiro Okuno, and Kotaro Suzumura, eds., Nihonz no sangyO seisakut (Induestrial Policy of Japan). Tokyo: University of Tokyo Press. Part III. Comparative Perspectives 10 Organizational Characteristics of Rural Textile Industries in East Asia Akihiko Ohno and Masao Kikuchi T he chapters in this volume investigate the rural garment and weaving industries of five Asian countries. The basic working hypothesis is that the wide diffusion of industrial and commercial activities in rural areas depends largely upon develop- ment of the relational contracting system rather than the factory system or discrete trans- actions in the spot market. The country case studies reveal that the relational contracting system has indeed been widely adopted in the textile industries, and that with this system, the rural textile industries maintain their competitive edge in urban markets by mobiliz- ing local resources, particularly entrepreneurship and labor of low opportunity cost. Despite a basic similarity, there is no single mode of the relational contracting system across the countries, between the two textile industries studied, or over time. Each mode of the contracting system has its own peculiarities. This diversity, however, does not imply that there is no rationale behind the contracting arrangements. Instead, these arrange- ments offer clues to a comprehensive understanding of how a particular mode of the rela- tional contracting system is chosen, in keeping with differences in socioeconomic and institutional environments. Our concern in this chapter is, while confining our main attention to the textile indus- tries examined in this volume, to illustrate a simple taxonomy of the relational contracting system. Analytical Framework Five chapters in this volume (Chapter 3 on Japan, Chapter 5 on the Philippines, Chapter 6 on Thailand, Chapter 7 on China, and Chapter 8 on Korea) depict the modes of the rela- 243 244 Toward the Rlral-Based Developnment of Comnierce and Induistry tional contracting system adopted under different conditions in the textile industries of East Asian countries. In this section, we set forth an analytical framework that allows us to place these studies in a comparative perspective. Relational Contracts as an Organizational Form In the framework of organizational forms, formal organizations (vertical integration or the factory system), markets, and relational arrangements are the major alternative medi- ating structures of transactions. In this view, the factors that determine the "make or buy" decision must be explored first. When the choice of urban businesses or middlemen is to buy, our second question is whether spot market transactions or relational transactions are preferred. When the choice is relational transactions, a third question concerns the reasons for the choice of a particular mode of that system. A taxonomy of the relational contracting system, therefore, needs to be explored within the framework of the choice among three organizational forms by specifying the local conditions that prompt the emergence of par- ticular forms. Positions of the Economic Actors in the Relational Contracting System The five country chapters focus on the aspects of the relational contracting system. Figure 10-1 illustrates the relative positions of the economic actors involved in the system in the rural-urban continuum and the main areas examined in each chapter. An important contrast is that in Chapter 3 on Iruma, the association between the rural merchants (rural entrepreneurs) and the independent weavers (agents) is viewed as a putting-out system. In Chapter 6 on Thailand, however, the association between the weav- ing/garment leaders (rural entrepreneurs) and the independent weavers/sewers is re- garded as an employment relationship, while that between rural entrepreneurs and urban businesses is considered a putting-out system. According to agency theory, rural entrepre- neurs are principals in the case of Iruma, while they are agents in the case studies of the Philippines, Thailand, Taiwan (China), and Korea. Accordingly, in finding common char- acteristics among the textile industries in the five countries, meticulous attention should be given to the different focal points of the papers. Modes of the Relational Contracting System The relational contracting system can be divided into the putting-out system and the ad- vance-order system. In the former, principals advance materials for processing to their agents, and later collect the finished products for a piece-rate payment; in the latter, prin- cipals only guarantee purchase of the finished products and there are no advances (see Chapter 1 of this volume by Hayami). The case studies in this volume, however, require finer distinctions in the relational contracting system from the viewpoint of decisionmaking on product specifications, which Organizational Characteristics of Rural Textile Induistries in East Asia 245 Figure 10-1. Relative Positions of the Economic Actors in the Relational Contracting System in East Asian Textile Industries Philippines Thailand Japan (Weaving) China Korea Garment Garment Weaving Iruma Banshu Garment Weaving since 1890 since 1960 .... . ....... Urban 0 D 0 0 areas : ' ...... * ......... : :* : . Sm-: *~ : : : : : : : : Semi- ' . a* *** urban :: : : * areas ~~~~~~~...... Rural * areas . .. . System adopted [PJ [P] [A] [P] [PI [A] [P&AI Note: * = merchants, * = workshops, and * = individual workers (including family-run workshops). Shaded and unshaded marks imply indigenous and modem respectively. Asterisk (C) indicates rural entrepreneurs. Dotted rectangles are the focal areas investigated in respective papers. P and A denote the putting-out and the advance-order systems, respectively, according to the classification of respective case studies. largely determine product marketability. Product specifications involve product design, product standardization, and quality control. In general, market information on product specifications is poorly communicated across markets because of asymmetric information and the untransferability of production skills between principals and agents. Asymmetri- cal distribution of market information and production skills between principals and agents among the textile industries in the five nations examined in this volume has led to the various modes of the relational contracting system. To achieve this greater refinement, we introduce a concept-"hold" over a product-to refer to obtaining market information on product specifications that is unobservable to others and/or to having distinct skills that are often untransferable to others. Because such "hold" is a technological factor, it should be distinguished from the concept of mer- its/demerits associated with a specific contractual mode. In addition, those who have "hold" over products do not necessarily possess marketing ability. By taking hold over the 246 Toward the Rutral-Based Development of Commerce and Ilnduistry Figure 10-2. Typology of the Relational Contracting System Product Specifications Precise Loose A d Yes Putting-out system (l) Putting-out system (ll) v a n c No Advance-order system (I) Advance-order system (II) e product into consideration, the following four modes of the relational contracting system can be identified (figure 10-2). - THE PrLJFNG-ouT SYSTEM (I): Principals advance materials to agents along with precise product specifications. Advances are generally made in the form of materials, be- cause market information on product specifications is usually embodied in those materials, such as precut cloth in the garment industry and dyed, warped, and sized yarn in the weaving industry.' e THE PUrrING-OUT SYSTEM (II): Although principals advance materials or cash, they des- ignate only general product specifications. * THE ADVANCE-ORDER SYSTEM (I): Although principals do not advance materials or cash, they designate precise product specifications, such as the kinds of materials to be used. * THE ADVANCE-ORDER SYSTEM (II): Principals do not advance anything, nor do they desig- nate precise product specifications; they merely guarantee the purchase of the prod- ucts from their agents. Among the four relational contracting modes, the putting-out system (I) and the ad- vance-order system (II) are polar cases, while the putting-out system (II) and the ad- vance-order system (I) are in-between. On the continuum of transactions from the spot market to vertical integration, or even the factory system, the putting-out system (I) is closest to vertical integration, while the advance-order system (II) is most akin to spot- market transactions in that neither advances nor product specifications are made in the transactions. Organizational Chiaracteristics of Rtral Textile Indulstries in East Asia 247 Stages of Market Development The stage of market development is another important determinant of contractual mode. By market development, we refer to how effectively the local market is integrated into urban markets or into overseas markets. It will be useful to distinguish three stages of market development: transactions within rural areas (local transactions); transactions with particular urban businesses (relational transactions); and transactions with numerous, anonymous urban businesses (ecumenical spot-market transactions). When transactions are mainly confined to rural areas, a spot market is apt to emerge. This is because product specifications are of less concern in the local market and, among other factors, rural producers have little trouble marketing their products in their areas. As rural societies are engulfed by urban markets, the localized spot market is subject to change be- cause many products of rural industries, although not all, are inferior goods, which dimin- ishes the competitive edge of rural industries. At this stage, rural industries must maintain a competitive edge against their modem counterparts. Our basic working hypothesis on the relational contracting system can thus be extended as follows: the relational contracting sys- tem is instrumental in enhancing the competitive edge of rural products by solving the prob- lems associated with rural industries in the process of market development. In economic development, marketing channels within the urban market are developed by urban businesses in particular; rural producers often lack marketing ability (a stage of relational transactions). At this stage, urban businesses become indispensable not only in their role as wholesalers, but also as conduits of market information and as quality con- trollers for product specifications. When the rural market develops further, necessary in- formation on product specifications would become directly available to rural producers and the marketing difficulties of rural producers will be largely solved. This is the stage of ecumenical spot-market transactions that none of the rural economies discussed in this volume has reached. The taxonomy of the relational contracting system is summarized in tables 10-1 and 10- 2. Table 10-2 presents the relative dominance of hold over the products among economic actors involved in the relational contracting system discussed in this volume, and the con- tracting systems adopted in the five countries. Table 10-2 shows the taxonomy of the sys- tem in accordance with the flows of materials and information on product specifications from principals to agents. In these tables the relational contracting systems discussed in this volume are renamed according to our classification. Propositions The above discussion can be summarized into the following two propositions on modes of the relational contracting system in association with hold over products. PROPOSmON 1: When principals (mostly urban businesses) have hold over products in relation to agents (mostly rural entrepreneurs), the putting-out system (I) or the advance- 248 Tozvard the Rutral-Based Development of Coninierce and Inidutstry Table 10-1. Relative Dominance of Hold over the Products Modes of Independent Riral Urban1 contracting workers entrepreneurs actors systent Garment industry Philippines and Thailand - - P (I) China NE _ A (I) Weaving industry Thailand ** - A ([) Korea: 1950s NE * _ Spot market 1960s to present NE - * P (II) & A (II) Iruma: before 1890 * - NE Spot market after 1890 ** NE P(l) Banshu Vertically striped * NE P(I) Cross striped ** ** NE P (I) Note: Asterisks denote hold. Their number irnplies the importance of hold. NE and - denote insignificant hold and nonexistence of economic actors involved in the transactions. P and A denote the putting-out system and the advance-order systems respectively, with their type in parentheses. Souirce: Authors' data. order system (I) will emerge. Principals are obliged to supply market information on product specifications because they can acquire it at relatively low cost. Because supplied materi- als often embody necessary informnation on product specifications, the putting-out system (1) is a usual consequence; a possible alternative is the advance-order system (I). It emerges when necessary materials and services designated by principals are readily available in the market and agents do not suffer from working-capital constraints. PRoPos1TIoN 2: When rural producers have a hold in relation to urban businesses, the putting-out system (II) or the advance-order system (II) may appear. It would be assumed that the former would be adopted when agents suffer from working-capital constraints, Table 10-2. Modes of the Relational Contracting System in East Asian Countries Sutpplyfrom principals Modes of Produect contractinig Material specificatiotts system Cases in this volurme 1 Yes Yes Putting-out system(l) Garment industries of the Philippines and Thailand, Iruma since 1890, Banshu 2 Yes No Putting-out system(ll) Korea since 1960 3 No Yes Advance-order system(l) China 4 No No Advance-order system(II) Weaving industry in Thailand, Korea since 1960 No No Spot market Iruma before 1890, Korea in the 1950s Sottrce: Authors' data. Organiizational Characteristics of Rutral Textile Induistries in East Asia 249 and the latter would be selected when they do not. These modes may also appear when product specifications are not enumerated in the contract between principals and agents- for example, when products are inferior or a third party is in charge of product specifica- tions, such as printing and dyeing. In any mode of the relational contracting system, rural actors are supposedly faced with difficulties in placing their products on the urban markets. The advances made in the relational contracting system have two distinct functions: (1) when made in the form of materials, they embody market information on product specifications, such as design, color, size, and the like, and (2) they allow agents to cope with their working capital constraints. The first form of advance is a pivotal feature of the putting-out system (I), although it entails the second function as well. The putting-out system (II), which by definition offers the second function, is akin to the advance-order system (II) in that the transfer of market information on product specifications from principals to agents is not as critical as it is in the putting-out system (1).2 As mentioned above, the ecumenical spot market works much like the advance-order system (II). This spot market will appear when the following condition is satisfied in addi- tion to those of the advance-order system (II): rural producers are not faced with critical difficulties in placing their products on the urban markets and can secure market informa- tion on product specifications. The basic working hypothesis of this volume, framed by Hayami (in Chapter 1), can be paraphrased to argue that the relational contracting system is instrumental in diffusing industrial and commercial activities over rural areas because the above conditions for the ecumenical spot market are seldom met in the rural areas of developing economies. Rural Textile Industries in an East Asian Perspective Insights from the Japanese Experience The emergence of the putting-out system in Iruma in about 1890 has important implica- tions for the evolution of rural textile industries in a historical context. Before 1890, imper- sonal spot market transactions were prevalent in Iruma for textile products; local merchants purchased cotton cloth at local markets, or directly from the weavers, and sold it to local wholesalers. The agrarian society of Iruma had been involved in the market economy before the emergence of the putting-out system. During this period, the weavers had a hold over their products. Consumers, however, were less fussy about the quality of prod- ucts, and the Iruma brand had not yet been established. Around 1890, the putting-out system emerged in Iruma as the local market became integrated in the national market. In this process, product specifications were imperative for the maintenance of the market competitiveness of Iruma-brand cloth. Accordingly, the role of local merchants as information conduits or quality controllers became crucial. For this purpose, they started to supply dyed, warped, and sized yam to village weavers. 250 Tow-ard the Rutral-Based Developmenit of Commiiierce anid Inzdutstry Consequently, hold over the products was passed into the hands of local merchants, lead- ing to the adoption of the putting-out system (I), as predicted in Proposition 1, above. Under the growing need for product standardization, the weavers themselves were also obliged to adopt the putting-out system (I). This clearly shows that the putting-out system is complementary to the market mechanism. It is also interesting to note that along with the advent of the putting-out system, the merchants decreased the geographical range of their transactions. 'his does not imply that the ecumenical spot market preceded the putting-out system in Iruma;3 rather, the localized market preceded the relational contracting system. Despite the ostensible mar- ket segmentation, the emergence of the putting-out system is indubitably an institutional change brought about by change in market environments. The putting-out system in Iruma brought the rural economy from the stage of local transactions to the next stage of rela- tional transactions: the introduction of the putting-out system implies not the weakening, but the strengthening, of the market mechanism in rural areas. Itoh and Tanimoto (Chapter 3) examined the period before the introduction of mechani- cal looms. A useful, related example is the case of a rural weaving workshop in Banshu, now Hyogo Prefecture, in western Japan, in the 1910s after the diffusion of mechanical looms.4 This example provides a good standard for comparison with the weaving indus- tries of other East Asian countries where mechanical or power looms are in use. The workshop in Banshu was operated by a family that acted as a principal (hereafter, the principal-weaver). Banshu was known for producing indigenous striped cotton cloth. The cloth produced in Banshu was similar to Iruma's-a large variety of plain cloth in small quantities. Two styles of plain cloth were woven in Banshu-vertically striped and cross-striped (plaid) cloth. Before the 1910s, while the former could be woven on mechani- cal looms, the latter could be made only on handlooms, primarily fly-shuttle looms.5 This principal-weaver engaged in manufacturing vertically striped cloth with mechani- cal looms in his workshop; he employed ten to twenty weavers, mostly married women. In addition, he put work out to village producers (hereafter, agent-weavers) for both kinds of cloth. Vertically striped cloth was put out to rural workshops that had installed me- chanical looms, while cross-striped cloth was made by farm household workshops that owned handlooms. The former workshops employed relatively more weavers; for example, one workshop hired sixty-nine workers. The latter were either household workshops that employed about ten weavers or family-operated workshops. For cross-striped cloth, the agent-weavers themselves possessed weaving skills because the cloth could only be wo- ven on handlooms. Because of the need for product standardization, however, the princi- pal-weaver supplied dyed, warped, and sized yarn for the agent-weavers for both vertically and cross-striped cloth. Hold over the products was in the principal-weaver's hands in Banshu, because product standardization had already been required in the market. Under these circumstances, the contracting mode employed for both kinds of cloth was the put- ting-out system (I), as assumed in Proposition 1. There remains a second question: why wasn't the factory system selected instead of the putting-out system? In other words, why did the principal-weaver choose to "buy" rather Organizational Characteristics of Rutral Textile Indutstries in East Asia 251 than "make"? Many of the chapters in this volume point to flexibility in the size of produc- tion as a major factor that supports adoption of the relational contracting system rather than the factory system. With the relational contracting system, principals can achieve production flexibility without worrying about the utilization rate of fixed capital. The fol- lowing may lend further support to this argument. A technological constraint on the prin- cipal-weaver in Banshu was removed when mechanical looms were improved to weave cross-striped cloth. Even so, the principal-weaver kept putting work out to farm house- hold workshops for cross-striped cloth. By putting work out, the principal-weaver could achieve greater flexibility against demand fluctuations. The agent-weavers were gener- ally married village women who enjoyed flexible working hours and could allocate their time between weaving and household chores at will. This relates to the difficulties associated with labor management in early factory organi- zations, as pointed out in Chapter 6 on Thailand. For example, the workshop operated by the principal-weaver in Banshu to produce vertically striped cloth suffered from high ab- senteeism among its weavers.6 This was partly because the principal-weaver had not es- tablished a labor-management system that encouraged regular attendance by the weavers, and partly because village women engaged in weaving as a side job. The cheap labor available in rural areas, a feature related to the difficulties with labor management in early factory organizations, is, of course, an important reason for the choice of the relational contracting system. The heterogeneous nature of cloth products, which makes it difficult for a centralized factory to reach minimum economies of scale in production, is yet another reason for se- lecting the putting-out system over the factory system. In these circumstances, it could be more efficient to put work out to agents who specialize in certain types of products. This was actually the case in Banshu, as well as in the garment industries in the Philippines and Thailand. In general, a workshop deals with only certain kinds of products. Note that these are the merits of the relational contracting system when compared with centralized production, not those of the putting-out system alone. Current Developments in China and Southeast Asia The Japanese experience points out that hold over the products is a key determinant of the choice of contractual modes. Let us expand this argument into the textile industries in other East Asian countries (see tables 10-1 and 10-2). In the garment industries in the Philippines and Thailand, the urban principals (modem garment factories or wholesalers) supply precut cloth to rural agents (rural entrepreneurs). The urban businesses in the garment industry generally have better access to market in- formation on product specifications than the rural agents. Accordingly, when the urban businesses choose to "buy" rather than "make," the putting-out system (I) is a rational consequence for both principals and agents (Proposition 1). In the Chinese garment industry, in contrast, although the urban principals (state or foreign garment factories) have hold over the products, they do not advance materials or 252 Toward the Rural-Based Developnment of Conmmerce and Indutstry cash to the agents (township-village enterprises as the rural entrepreneurs), but let them purchase cloth from designated sources. The materials contain market information on prod- uct specifications. In China, however, it is the urban principals that suffer from working- capital constraints. To accommodate these constraints, the advance-order system (II) is adopted instead of the putting-out system (I), as predicted in Proposition 1. In the weaving industries investigated in this volume, diverse situations concerning hold over products, and hence different contractual patterns, are observed. While the ad- vance-order system (I) is widely practiced in northern Thailand, the putting-out system (I) was employed in Iruma and Banshu in Japan. In Punggi, Korea, the putting-out system (II) and the advance-order system (II) have coexisted since around 1960, but spot-market trans- actions were prevalent in the 1950s. In discussing this diversity, two attributes of woven cloth need to be considered: plain versus twilled cloth and yarn-dyeing versus piece-dye- ing methods.7 Although either dyeing method can be employed for plain cloth, twilled cloth is woven only with dyed yarn. Plain cloth requires less specialized weaving skills than twilled cloth. The Thai rural entrepreneurs/weavers are skilled at weaving twilled cloth with the traditional, intricate patterns of the region. Their skills for twilled cloth are hardly stan- dardized, and the weavers thus have strong hold over their products, which are suffi- ciently differentiated to maintain market competitiveness. Therefore, the Thai weavers do not have to rely on urban businesses to obtain market information on product speci- fications. In addition, the rural entrepreneurs do not suffer greatly from working-capital constraints, which is made clear by their habit of requesting cash advances from urban wholesalers only in the case of bulk orders. Nevertheless, they lack marketing skills, and the advance-order system (II) is consequently a rational choice in the Thai weaving in- dustry (Proposition 2). Other weaving industries examined in this volume produce plain cloth: Iruma and Banshu produced striped plain cloth, while Korea has produced white plain cloth since about 1960. A yarn-dyeing method and a piece-dyeing method, respectively, were thus employed. Under a yarn-dyeing method, dyeing, warping, and sizing are generally done by those who have hold over the products: local wholesalers in Iruma, and the principal- weaver in Banshu. Because principals supplied materials that embodied market informa- tion, the putting-out system (I) was adopted in Iruma and Banshu (Proposition 1). Among the three case studies, piece-dyeing is practiced only in the Korean weaving industry. In the 1950s, the rural weaving entrepreneurs dyed cloth in their own facilities; the rural producers had hold over the products. Transactions were carried out in the spot market, because the market economy in Punggi was developed in the 1950s. Product stan- dardization, however, was not crucial in the market, but it became imperative in the pro- cess of economic development. Product standardization reduces the transaction costs related to information asymmetry on product quality, and thus improves market efficiency. Nev- ertheless, it should be noted that a product standard is costly to develop, set, and enforce. Accordingly, standardization becomes economically possible only when market size ex- pands enough to compensate the associated costs. Since around 1960, in response to the Organizational Characteristics of Ruiral Textile Iindutstries in East Asia 253 growing need for product standardization, white cloth woven in rural workshops has been dyed in the dye-houses selected by the principals (urban wholesalers) to their specifi- cations. It is now the urban principals who have hold over the products, because they are in the position to secure market information about fashion in the urban market. In Korea, the dyeing process is carried out by a third party, and no precise product specifications are offered to the agents by the principals. Therefore, the contractual modes observed in Ko- rea since about 1960 are the advance-order system (II) and the putting-out system (II), depending on the financial positions of the agents (Proposition 2). According to our taxonomy, the contractual modes adopted in the Thai and Korean weaving industries can be classified as the putting-out system (UI). A sharp contrast, how- ever, exists between them. Hold over the products is in the agents' hands in Thailand, while it is in the principals' in Korea. In the latter, no precise product specifications are given to the agents, because a third party is in charge of dyeing. This requires a more detailed taxonomy than we can undertake here. In the context of agency theory, the Japanese rural entrepreneurs in Chapter 3 are regarded as principals whose agents are independent village weavers, while the Thai and Korean rural entrepreneurs in Chapters 6 and 8 are regarded as agents whose prin- cipals are urban wholesalers. Contractual relations in the former are described as the putting-out system, while in the latter they conform to the advance-order system. This contrast is primarily a product of the different stages of market development in the two countries. In industrializing Japan, urban businesses had not grown enough to expand their marketing channels into rural areas, and urban markets did not demand product standardization. Therefore, rural merchants played an important role in marketing rural products to urban consumers.' In Korea (since around 1960) and in Thailand, urban businesses had vigorously infiltrated the rural areas; the markets were further devel- oped and linked with the urban market, or even overseas markets, unlike the markets in Iruma and Banshu. Let us turn to the second question on the choice between spot-market transactions and relational transactions. As expressed in Proposition 2, hold by rural producers over prod- ucts is a necessary condition for the prevalence of spot-market transactions. This condi- tion was met in the three industries examined in this volume: Iruma before 1890, the Korean weaving industry in the 1950s, and the Thai weaving industry. In Iruma and Punggi, cloth was sold in the spot market. In Thailand, the advance-order system (II) is prevalent. This contrast can be explained by the quality of the products and the difference in geo- graphical distance to urban centers. Commercial activities had already spread in Iruma in the latter half of the nineteenth century, and in Punggi in the 1950s, although their prod- ucts were inferior in quality and sold mainly in local markets. The Thai rural weavers wove for home consumption and rarely marketed their products, although they would have had a strong competitive edge if they had been sold in urban markets. The market economy did not emerge in Thailand until the advent of the advance-order system. In addition, Thai cloth is now mostly consumed in Bangkok and in overseas markets far from the production areas. Therefore, the advance-order system (II) is beneficial for both parties 254 Towvard the Rtural-Based Development of Comnierce and Inditstry in northern Thailand. The weavers' marketing difficulties can be overcome by the system, while the urban principals can largely reduce the transaction costs associated with the procurement of products from weavers who live in remote villages. In summary, a comparison of these case studies reveals that the relational contracting system can appear in a variety of modes; the method selected depends on hold over prod- ucts, working capital constraints, and the stage of market development. Altogether, the investigation supports the working hypothesis of Hayami in Chapter 1: the relational con- tracting system, rather than spot-market transactions or vertical integration, facilitates the wide diffusion of industrial and commercial activities over rural areas. Dualism in Legislation and the Relational Contracting System Institutional arrangements, especially legislation, governing the labor market and factory organizations play a decisive role in shaping the modes of production organization. De- veloping countries today have much stricter legislation governing economic activities than developed countries experienced during their early stages of economic development. In Japan, for example, the first labor act (Factory Act) was enacted in 1911 and was in force in 1916. The act, however, regulated only minor areas of labor conditions. A full-scale Japa- nese labor law was not enacted until 1947. Note that Itoh and Tanimoto studied the Japa- nese weaving industry in the prefactory era. Laws protecting labor were enacted in developing countries at earlier stages of industri- alization than had been the case in the developed world. These legislative restrictions, however, are mainly confined to the formal sector, and the informal sector is often outside the legal framework or does not comply fully with the laws. This dualism in legislation makes the effective labor costs of the informal sector much lower than those of the formal sector, as discussed in the Philippine and Thai studies. Consequently, dualism in legisla- tion promotes efforts by large factories to contract work out to establishments smaller than those regulated by legislation.9 We need to mention here only the following five major elements of dualism in legislation. FLEXIBILITY. Enterprises in the formal sector have institutional buffers that absorb fluctua- tions in the size of the work force according to changes in business conditions. Temporary workers, including part-timers, dispatched workers, and apprentices, are often subject to employment adjustments during recessions. Subcontracting part of production is another means of achieving flexibility; the need to achieve such flexibility, of course, depends on the labor laws of the respective countries. The following serve as good examples. In Thailand, the difference between perma- nent and temporary employees was recently eliminated by Interior Decree No. 11. While many temporary workers were transferred to permanent status, large factories began to contract work out to smaller enterprises on a large scale (Sungsidh and Poonpanich 1994). In India, factories subject to laws regulating work conditions are defined as those Organiizational Chlaracteristics of Rural Textile Indutstries in East Asia 255 employing ten or more workers if using power, and twenty or more workers if not using electricity. As a result, an increasing proportion of establishments employ nine workers, and the practice of contracting work out to such establishments is increasing (Papola and Rodgers 1992). FEMALE LABOR. Most of the labor acts in Asian countries have special provisions regard- ing the protection and welfare of female workers, including maternity leave, separate welfare amenities such as child care, and prohibition of their employment during night hours and in dangerous and heavy operations. These provisions make the effective labor costs of female workers higher than those for male workers. This has prompted the prac- tice of contracting work out, especially in female-dominated industries. In India, for example, the existing factory act promulgated in 1948 stipulates compre- hensive provisions covering the health and welfare of female workers. Previously appli- cable to factories employing more than fifty workers, in 1976 the act was amended to cover factories employing more than thirty. In order to evade the factory act, large facto- ries began to rely increasingly on the relational contracting system (Malavika 1987). LABOR UNIONS. The capital-labor relationship is one of the core influences on the modes of production organization. In this sphere, developing Asian countries present diverse con- ditions. Labor unions in thfe Philippines provide a good example. Strict restraints on em- ployment retrenchment brought about by the strong presence of labor unions in the Philippines draw a different picture of the relational contracting system from that in coun- tries where labor unions are relatively modest or nonexistent. A collective bargaining agree- ment in a modem garment industry in the Philippines near Metro Manila, for example, has the following clause concerning lay-offs: "where both temporary, casual probationary and trainee and regular employees are present, such temporary, casual, probationary and trainees shall be laid off first and the regular employees shall be retained." In countries where the employment of core workers is guaranteed by strong labor unions, it is rational for the modem factories to maintain a buffer work force to cope with demand fluctua- tions. Consequently, the work force is divided into insiders (union members), whose em- ployment is protected by job-preserving measures, especially labor unions, and outsiders, who are not under such protection.'0 Workers involved in the relational contracting sys- tem are typical outsiders. The dualism created by labor unionism makes it costly to in- crease insiders in the modem factory; unionized factories are thus induced to contract work out to small enterprises that are not unionized."' TAXEs: Tax legislation varies significantly from one country to another. In Thailand, the business tax was recently replaced by a value added tax that exempts small enterprises.12 In the Philippines, manufacturing enterprises that employ fewer than twenty in rural ar- eas are exempted from major taxes for five years after registration (in accordance with the Magna Carta for Countryside and Barangay Business enacted in 1990). Such a differential tax system encourages large enterprises to contract work out to smaller enterprises. 256 Towcvard the Ruiral-Based Development of Conmmerce antd Industry LICENSING. The licensing or preferential allocation of quotas for exports to privileged establishments in the urban formal sector creates another form of dualism in legislation. Where such dualism exists, it certainly facilitates the relational contracting system, as is the case in China (see Chapter 7). A Long-Term View As pointed out by Hayami in Chapter 1, enterprises operated by rural entrepreneurs under the relational contracting system survive in Japan today; they have transformed their organizational features to adapt themselves to the modern industrial environment. The major textile zones of postwar Japan, for example, have developed in areas where the traditional textile industries once prospered (see Chapter 3). This historical continu- ity can also be observed in the Korean case in Chapter 8, where the handloom weaving industry evolved into the modem weaving industry with the installation of electric looms in the second half of the 1960s, and even air- and water-jet looms in the 1980s. More noteworthy is that in both countries these industries still rely greatly on the relational contracting system. The question we must consider here is whether there are indications that the rural textile industries of contemporary East Asia will evolve into modem enterprises in the course of economic development. We briefly examine this possible evolution from the viewpoint of the introduction of formal labor management into rural enterprises, be- cause one of the advantages of these enterprises is thought to rest on their informal labor management. Second, we touch upon the choice between the relational contracting sys- tem and the inside contract system that prevailed in the West during the early stages of economic development. Historical Continuity A predominant feature of the rural textile workshops in the Philippines and Thailand is the absence of any significant arrangement of the management hierarchy. Middle manag- ers for supervision are virtually absent in rural workshops, partly because the employees are suspicious about being supervised by fellow villagers, and partly because appointed supervisors are reluctant to supervise fellow villagers. Instead, the entrepreneur-manag- ers, who usually work side-by-side with their employees, exercise a substantial degree of control over labor. Informal labor management under a simple hierarchy is found to be an advantageous feature of the rural workshops. At the same time, however, it imposes restrictions on the number of employees, partly because an entrepreneur-manager can supervise only a small number of workers, and partly because a collective responsibility, such as a peer group sentiment, can be effectively fostered among a small number of homogenous workers. As shown in Chapters 5 and 6, the number of workers in the average workshop in the Philip- Orga7tizational Chiaracteristics of Rlral Textile Indutstries in East Asia 257 pines and Thailand is as low as ten to fifteen. Some of the rural entrepreneurs were found to have little intention of expanding their businesses because of possible difficulties in managing a larger number of workers. In the progress of organizational forms of production identified by mainstream histori- ans, the relational contracting system (especially the putting-out system) is often seen as a transition phase from the craft workshop or the guild system to the factory system, or a phase of proto-industrialization. Experiences in Japanese textile industries demonstrate historical continuity between rural industries and modem industries. From this perspec- tive, the immediate question becomes how rural enterprises involved with the relational contracting system evolve into modem factory organizations in the course of economic development. Can we expect similar continuity in the rural enterprises in developing coun- tries today? This question is deeply connected with a fairly general agreement that the socioeco- nomic-economic conditions that have sustained the management of rural workshops through community relationships will decline in the course of economic development. Our studies of the Philippines and Thailand (Chapters 5 and 6) serve as evidence of this. In these countries, the erosion of both social norms and the cohesion in rural society are observed. Such erosion can be explained primarily by differing levels of exposure to an urban economy. The rural workshops that are most exposed to an urban economy, such as the Philippine workshops and cases GF and GH of the Thai chapter, have transferred elements of the factory system into their workshops. Another example occurs in weaving workshop F of the Thai study, It institutionalized labor-management practices specific to the modem factory system, such as the time-rate wages and promotion practices, although not because of urban exposure. The point is that labor management in the factory system is not necessarily incompat- ible with that in the rural workshops, and rural entrepreneurs have eminent potential for incorporating the characteristics of modem factory organizations. The Relational Contract and the Inside Contract Systents The inside contract (or internal contracting) system was used extensively in Britain, Japan (Littler 1982), and the United States (Buttrick 1952; Chandler 1977) in the nineteenth cen- tury, before the transition to bureaucratic factory organizations. The two major forms of the contracting system, the putting-out and inside contract systems, bear a close resem- blance to one another. A comparison between them will help us to understand the func- tions of the relational contracting system in a new light. Although the inside contract system varied among industries, over time, and across countries, the primary form of the system was as follows. According to Buttrick (1952: 205), "under the system of inside contracting, the management of a firm provided floor space and machinery, supplied raw materials and working capital, and arranged for the sale of the finished product.... They [inside contractors] hired their own employees, supervised the work process, and received a piece rate from the company for completed 258 Towvard the Ruiral-Based Developnmenit of Commerce and Induistry goods." Inside contractors worked as agents of the manager-owners of the enterprises. When compared with the putting-out system, the only major difference is the working sites of the agents. Both systems are decentralized forms of production; the employment relationship is constructed around the agents, inside contractors, and rural entrepreneurs. Discussion of the inside contract system in comparison with the putting-out system may provide an answer to the question addressed in Chapter 3: why wasn't the putting-out system prevalent in Europe and North America, despite its several advantages? Chapter 6 on Thailand revealed that the absence of trained manpower for middle man- agement in large factories during the early stages of economic development is a cogent reason for decentralized responsibility in the form of the relational contracting system. In this context, the inside contract system can be an alternative to the relational contracting system. As many historical studies have proven, the inside contract system reflects the managerial difficulty associated with direct employment relations, as can be seen in the following quotation: under the inside contract system, "responsibility, risks and costs should be partly shifted on to contractors, thus creating greater flexibility in circumstances where managerial skills and knowledge of work operationDs were limited" (Thompson and McHugh 1990: 53). By relying on the inside contract system, the "capitalist ... was freed from most of the technical problems associated with production, improvement of the manu- facturing process and labor supervision" (Buttrick 1952: 207). Accordingly, it can be said that the inside contract system in Western countries worked as a substitute for the rela- tional contracting system. It can be assumed, therefore, that both the inside contract sys- tem and the relational contracting system are middle-management-saving devices. The question, therefore, is why the inside contract system was more prevalent than the relational contract system in the Western world. Population density, as suggested by Itoh and Tanimoto (Chapter 3), could be a main reason. Another reason could be the village community as it functions in developing Asian countries. This is, as pointed out in many chapters in this volume, an important economic resource for efficient production in rural industries. In order to utilize such a resource, relational contracting is instrumental. In the West, economic development started after the demise of the village community."3 There- fore, in the absence of a resource associated with the village community in the rural econo- mies of the West, the relational contracting system with rural producers is not advantageous for utilizing rural resources. The Rural Garment Industry in the Philippines and Thailand In this section, we compare the rural garment industry in the Philippines and Thailand, where the putting-out system is commonly practiced. These two case studies contain rela- tively more and similar sets of information on the industry today, enabling a comparison under a condition close to ceteris paribus. It is expected that such a comparison will give us further insights on the organizational characteristics of the rural textile industries. Organiizational Characteristics of Rutral Textile Itndutstries in East Asia 259 Observed Comnionalities The case studies in the Philippines (Chapter 5) and Thailand (Chapter 6) reveal common characteristics of the rural garment industry. First, the development of modem garment factories at urban centers has induced the evolution of many garment workshops in rural areas through the diffusion of putting-out arrangements. Second, this evolution of the rural garment industry has been pushed by rural entrepreneurs who, emerging mostly from the peasant class, can seize the opportunities created by economic development through mobilizing and organizing rural resources, capital and labor in particular. Third, an important condition for those who become entrepreneurs is that the entry barrier to the industry is low, not only because of the low fixed capital requirement to set up a rural workshop, but also because the putting-out system lowers the information requirements on marketable products and makes financial management a secondary consideration. Fourth, a major reason that garment firms in urban centers adopt the putting out system, despite the inherent disadvantage in the maintenance of quality standards, is that it can reduce labor costs, either by using rural labor of low opportunity cost or by avoiding the tight labor regulations enforced in the formal sector-or both. Fifth, the rural garment industry as it has evolved has contributed to creating productive employment opportuni- ties in rural areas for those who would otherwise work in less remunerative jobs, migrate to urban centers to seek jobs, or remain unemployed. In spite of these basic commonalities, however, there are also intriguing differences be- tween the garment industries of the two countries. We first enumerate the major differ- ences observed in the studies and then try to explain why such differences have emerged. Observed Differences Of the many observed differences, the most salient and interesting may be the following. 1. Worker experience in the modern garment indutstry. While most of the rural entrepreneurs (leaders) and sewers in Thailand have previous work experience in modem garment factories, many of those in the Philippines (although some had such experience) were novices when they entered the industry. 2. Edutcation and class backgrotnd. In the Philippines, a clear difference exists in educa- tional background between the rural entrepreneurs and the sewers: the former are mostly college graduates, while the latter are primary school graduates or high school drop-outs. In addition, while the sewers are from the bottom stratum of rural society, the rural entrepreneurs are either from the farmer class or from the landless worker class; the former source provides the largest number. In contrast, both Thai leaders and sewers are primary school graduates from the bottom stratum of rural society. 3. Contractual relationships. In the Philippines, there are rural entrepreneurs who free- lance and change exporters/contractors frequently. On the part of sewers, the behav- 260 Towvard thie Rural-Based Developnment of Commerce and Industry ior of getting a cash advance and running away was observed. In Thailand, no such behavior was found: the associations between the urban principals and the agents (rural entrepreneurs) and between the rural entrepreneurs and the sewers are both stable and self-perpetuating. 4. Tlhe mode of labor managemenit. The association between subcontractors and sewers, or the nature of labor management in rural garment workshops, demonstrates a sharp contrast between the two countries. Absenteeism and a high turnover among sewers caused by job-hopping are major complaints of the Philippine rural entrepreneurs, while these problems are not serious in the Thai garment workshops. In Thailand, informal labor control, based either on a peer group or on a patron-client relationship, is presumably effective. In contrast, more formal control seems necessary in the Fili- pino rural entrepreneurs' workshops. 5. Workshtop sewvers and home seuwers. In Thailand, the rural entrepreneurs (leaders) de- pend more on home sewers. They are considered employees of the rural entrepre- neurs (leaders) rather than independent workers under putting-out arrangements. In the Philippines, although there are rural entrepreneurs who "employ" home sewers, they are of less importance in garment production than they are in Thailand. 6. Aspiration of ruiral entrepreneutrsfor upward mobility. Some Filipino rural entrepreneurs have a strong desire to move up to the exporter/contractor level. Some started their careers in the industry as sub-subcontractors. In contrast, no sample rural entrepre- neurs in Thailand express such aspirations. Most hesitate even to expand their busi- nesses because of possible increasing difficulties in labor management in proportion to the number of workers. 7. Owvnership of prodtuction tools. The rural entrepreneurs in the garment industry in the Philippines must own high-speed sewing machines in order to employ sewers. In Thailand, the rural entrepreneurs can employ sewers who have their own sewing machines. The weaving leaders in Thailand, however, have to install handlooms to enter the weaving business. 8. Government interventiont. Both the Philippine and Thai governments have implemented policies intended to foster rural industries; the provision of institutional loans to rural entrepreneurs is the most typical. These policies appear to be more effective in Thai- land than in the Philippines. Many Thai rural entrepreneurs (leaders) in the weaving industry obtained such institutional loans to acquire handlooms, but the number of rural entrepreneurs in the Philippines who received such loans is quite limited. At the same time, few rural leaders in the garment industry in Thailand receive such loans. This is because they can employ sewers who already have their own sewing ma- chines. In contrast, the weaving leaders have to lend handlooms for fabrics of double width without charge, since traditional handlooms are generally for fabrics of a single width, which do not have a large market. As long as the garment leaders in the Phil- ippines have to own high-speed sewing machines in order to employ sewers, a credit service will be an effective step in the formation of garment groups. Organizational Charncteristics of Ruiral Textile Industries in East Asia 261 Factors Behind the Differences Many factors were involved in bringing about the differences mentioned above; they will not all be discussed in detail here. Nevertheless, a brief listing of the elements that gener- ated the differences, either independently or in concert, will include the following. GEOGRAPHICAL DISTANCE FROM AN URBAN CENTER. Laguna, the study area in the Philippines, is adjacent to Metro Manila; in travel time, the distance is only one hour (Laguna I) to three hours (Laguna IV). The northern region of Thailand, in contrast, is about 600 kilometers from Bangkok, and there are no large cities in between. Bangkok, like Metro Manila, is the center of consumption in the country, and most export agents have their offices there. This difference in geographical distance from the market center no doubt has profound impact on the organization of the garment industry. For instance, information on product markets is most important, but also most scarce, for the rural entrepreneurs in the garment industry. In Laguna, however, it is not entirely impossible for rural entrepreneurs to obtain the information on exporters by contacting them directly. In northern Thailand, the long distance from Bangkok makes it nearly impossible for rural entrepreneurs to gather such information. As a.consequence, the urban principals play the vital role of linking village producers with urban markets: only through the urban principals can the rural entrepre- neurs secure the information. Such a situation makes the rural entrepreneurs more subor- dinate to the urban principals and renders their relationship more stable and long-lasting. The different patterns of the ownership of production tools is another illustration of the difference in geographical distance from an urban center. The labor market in the Philip- pines case is stiffer than in the Thai case. Close proximity to Metro Manila provides alter- native job opportunities, and the Filipino garment subcontractors must have high-speed sewing machines to attract and employ competent sewers. The Thai garment leaders, in contrast, can employ sewers who have their own sewing machines, exploiting their rela- tively favorable bargaining position in a labor market where job opportunities are scarce. It is worth noting that the rural entrepreneurs in the most remote towns in Laguna have a mode of operation similar to that found in Thailand. These findings suggest that to understand what is required to promote rural industrial- ization in remote areas, it is necessary to study the roles played by individuals such as the urban principals in northern Thailand: they are the change agents who intermediate criti- cal market information between the centers/outside world and rural entrepreneurs. BEHAVIOR OF FEMALE WORKERS. Thailand is well known as a country with an exceptionally high rate of female labor participation. A considerable portion of the newly supplied work force for labor-intensive industries, particularly in unskilled occupations, is women from rural areas. There is, however, a clear tendency for these female workers to return to their home villages after marriage or after delivering a child."4 It appears that the notion that a married woman's place is at home remains the social norm. Therefore, in rural villages 262 Towuard the Rutral-Based Developnment of Comnierce anld ildlistry there are many returnees who were line supervisors or sewers in modern garment facto- ries. They create a large reservoir of sewers and rural entrepreneurs (leaders) for the rural garment workshops. As revealed in the Thai study, almost all leaders and sewers are mar- ried women. This naturally makes them immobile, both geographically (less job-hopping) and vertically (less aspiration to move up). The gender division of labor and limited mar- ket demand for female labor in rural areas make home manufacturing one of the few ways that rural women with family responsibilities can supplement family income. The labor participation rate of females is quite high in the Philippines as well, but the social norm of women staying at home after marriage is much weaker, even nonexistent. In other words, gender segregation in the Philippines is not as distinct as in Thailand.'" Therefore, modern garment factories in Metro Manila and the surrounding area are full of married sewers. This is a major reason that the turnover rate of sewers in modern garment factories is almost zero in the Philippines. WAGE DIFFERENTIAL BETWEEN PROTECTED AND UNPROTECTED SECTORS. The two case studies re- veal that lower labor costs in rural areas are behind the diffusion of the putting-out system in the garment industry. It must be noted, however, that there is a marked difference be- tween the two countries in the wage differential of modern factories and rural workshops. In the Philippines, the wage rate for sewers in modern factories is more than twice as high as that in rural workshops, while the wage gap is.not as distinct in Thailand-6 This does not mean that the labor cost advantage in garment production under the putting-out sys- tem has disappeared in Thailand. The advantage can still be utilized if the transaction costs associated with hiring-such as fringe benefits for sewers under well-enforced labor regulations in the modern sector and monitoring costs for labor efforts-are taken into consideration. This raises the interesting question of what can explain the difference in the nature of the dualism in the labor market of the two countries. But, setting aside this question, it would explain, at least in part, the differences in the association between the rural entre- preneurs and the sewers and in the mode of labor management of the rural entrepreneurs. With the large wage differential in the Philippines, it is reasonable to observe frequent job- hopping by sewers from rural workshops to modern garment factories. In contrast, with the relatively negligible wage differential in Thailand, there is little incentive for Thai sew- ers to switch jobs. This brings about more stable associations between the rural entrepre- neurs and the sewers. Note the opposite directions of labor mobility in the two countries. In the Philippines, the sewers working in rural garment workshops try to secure employ- ment in modern garment factories, while in Thailand, the village sewers are generally returnees from modern garment factories. It would be appropriate to mention that some Thai garment leaders, such as GH, face the same difficulties in managing sewers as Filipino subcontractors do. This is because workshop GH is located in a town and must recruit workers from remote villages. It re- sembles the Laguna rural entrepreneurs in many respects. Laguna is a province with rela- tively high wage rates, partly because of its proximity to Metro Manila and partly because Organiizatioinal Characteristics of Rutral Textile Industries in East Asia 263 of its high agricultural productivity. Some rural entrepreneurs there recruit sewers from remote regions where wage rates are lower, with the expectation that loyal and patient sewers can be obtained from this population. In many cases, however, these sewers begin job-hopping after working for a time at the rural entrepreneurs' workshops in Laguna. ROLE OF EDUCATION. In Thailand, the social differences between the rural entrepreneurs (leaders) and their sewers are insignificant. Both are from the bottom stratum of agrarian society and they have the same educational background (primary school graduates). The group cohesiveness among them is, therefore, relatively strong, and the strong sense of membership in a peer group is conducive to making informal labor control more relevant than the formal controls common in Thailand, and here lies the comparative advantage of rural household enterprises over large modern factories. Conversely, in the Philippines, there is a large difference between the rural entrepre- neurs and their sewers. Although some rural entrepreneurs come from well-to-do rural families with some land and assets, there are also some rural entrepreneurs whose family backgrounds are as humble as the sewers'. It cannot be said, therefore, that the rural entre- preneurs and the sewers are from different social classes in the village communities, but there are large differences in education between the two groups. Most of the rural entre- preneurs are college graduates, while the sewers are mostly primary school graduates or high school drop-outs. Because pseudo-social classes are created by differences in educa- tion in the Philippines, such a large difference in educational levels would result in a social cleavage wide enough to weaken group cohesiveness when compared to the Thai cases. This may be a factor behind the need for more formal labor controls in the rural work- shops, as indicated by the "rules and regulations" reproduced in the Philippines chapter. Such rigid factory regulations are found only in modem factories in Thailand, as shown in the Thai chapter. The differences in education of the rural entrepreneurs might be a significant factor behind the differences in the aspiration levels and upward mobility of the entrepreneurs in the two countries. The enrollment rate in the tertiary level of education is about 30 percent in the Philippines and about 15 percent in Thailand. With a smaller percentage of people going on to the tertiary level, college graduates in Thailand are the elites of society and, coupled with the rapid economic development in recent years, which increases the demand of the modem sector for a highly educated work force, it is improbable that a college graduate would take a job in an industry such as garment subcontracting in a rural area. College graduates in the Philippines are also members of an elite group, but in a greatly diluted sense, and the unemployment rate is higher for those of the age of college graduates.17 In such an environment, it is not at all unusual for college graduates to go into the garment subcontracting business. The presence of a ladder toward becoming export- ers should make the profession even more attractive to college graduates. To the extent that a college education fosters wider perspectives and higher aspirations among the stu- dents, as described in the chapter on the Philippines, it is reasonable to expect that the aspirations of the Filipino rural entrepreneurs are higher than those of Thai rural entrepre- neurs, who have only primary educations. 264 Towvard the Rural-Based Developnment of Commerce and Iindustry SOCIAL/CULTURAL NORMS AND COHESION OF RURAL SOCIETY. As mentioned above, the com- parative advantage in labor management of rural household enterprises over modem fac- tories in Thailand rests partially on the strong cohesiveness of rural societies. This cohesiveness is instrumental in forming either strong peer groups or patron-client rela- tionships among the people involved in the rural garment industries. In contrast, the na- ture of the association between the rural entrepreneurs and sewers in the Philippines makes it necessary to employ more formal labor controls. It may be adequate to point out here that such an advantage depends not only on the cohesiveness of the society but also on the factors explained so far in this section. In other words, it may be misleading to overemphasize the strong social cohesiveness in the Thai case relative to the social conditions in the Philippines. The Thai expression kreng jai, mean- ing "caring and consideration," connotes in Thai culture the spirit of getting along well with other people, or a kind of social norm among people to do so. Such a spirit in rural society makes informal control in rural workshops superior to the formal control in mod- em factories, especially in the early stages of industrialization, when the system of formal labor control in factory organizations has not been fully established. A Filipino expression, pakikisana, meaning to have friendship with others or to be good to other people, has a connotation similar to kreng jai. Those who violate this social norm overtly are considered zuarang hiya (shameless persons) by others. In this sense, the societies of both countries share similar "Asiatic" cultural values, in spite of their different histories. Another similarity between the two countries is that their societies are "loosely" struc- tured, as described by Embree (1950) for Thailand. Such looseness in the societies is largely a reflection of resource endowments. In Southeast Asia, both Thailand and the Philippines are rather exceptional countries in that there were still land frontiers until recently because of favorable endowments of land relative to population (Hayami and Kikuchi 1981). As the population grew, people migrated to the frontiers when the pres- sure overwhelmed the available stock of land. In these countries, people in rural societ- ies have thus been very mobile, which would have contributed to nourishing rural societies that are loosely structured. It seems, therefore, that the two countries have more similarities than differences in social norms and degree of cohesiveness in rural societies compared with other societies in Southeast Asia. If this is so, the differences in the mode of labor management between the two countries should stem primarily from the different levels of cohesiveness of rural societies that result from geographical distance from the urban center and the associated differences in their labor markets and educations. Notes 1. After dyed yarn is unwound from hanks, it has to be warped. Warping is a laborious yam preparation task. Sizing is necessary to strengthen yam. 2. Some economic historians who regard the putting-out system as a premodern, feudal ar- rangement may argue that the presence of a credit advance from the principal, or credit tying, is an essential feature of the putting-out system. Organizational Characteristics of Rural Textile Indutstries in East Asia 265 3. A similar situation was observed in the Korean weaving industry, as will be discussed below. The spot market in the 1950s was replaced by the relational contracting system in about 1960. 4. This section is based on Sasaki (1991). According to personal communication of the authors with Tanimoto, what took place in Iruma after the introduction of mechanical looms was quite similar to the events described by Sasaki in the weaving industry in Banshu. 5. In Iruma, only vertically striped cloth was woven. 6. On average, the weavers attended the workshop only about ten days each month. 7. In the yam-dyeing method, dyeing is a preloom activity, while in the piece-dyeing method, it is a postloom activity. 8. According to personal communication with Tanimoto, in the later days the urban wholesalers joined in decisionmaking on the kind of products to be made with the local wholesalers. This seems to have come about because the urban businesses were able to acquire market information on prod- uct specifications. Note that this situation closely parallels the garment industries of the Philippines and Thailand, where the urban businesses have strong "hold" over the products. 9. It should be noted that the developing countries in Asia have enacted diverse provisions pertaining to factory workers. This leads to different patterns of production organization, including the relational contracting system. To inquire further into the matter as a whole is beyond the scope of this brief chapter. 10. On the insider-outsider theory, see Lindbeck and Snower (1988). 11. For further details on the labor market in the Philippines, see Ofreneo (1994). 12. Manufacturers, sellers, or providers of a service, who have annual gross receipts of less than 600,000 baht, are not covered by the value added tax. 13. This part of the statement relies on Ishikawa (1988). 14. For example, the ratio of married female workers in the rural food industry (N = 150) is 70 percent, while the ratio in the garment industry (N = 150) located in urban areas is as low as 43 percent. See Varnee and Jirapatpimol (1994). 15. For example, while in Thailand no male sewers and leaders were found, in the Philippines there are some male sewers and subcontractors. 16. In the Philippines, the minimum wage rate, which modem factories have to observe, is P145/ day, whereas the wage rate paid by subcontractors is about P65/day on average, in spite of the longer daily working hours for the latter. In Thailand, the minimum wage rate is B106/day for factory workers, while the wage rate for sewers who have electric sewing machines is about B100/ day; for those who have manual sewing machines, it is about B60/day. The exchange rate with the U.S. dollar is almost the same for the baht and the peso. 17. The unemployment rate in the Philippines in 1993 from the cohort twenty to twenty-four years old was as high as 19 percent, compared with 10 percent for the work force as a whole. References Buttrick, John. 1952. "The Inside Contract System." Journal of Econonmic History 12 (3): 205-21. Chandler, Alfred D. 1977. The Visible Hand. Cambridge, Mass.: Harvard University Press. Embree, John F. 1950. "Thailand: A 'Loosely Structured' Social System." Anierican Anthropologist 52: 181-93. Hayami, Yujiro, and Masao Kikuchi. 1981. Asian Village Econonmy at the Crossroads. Tokyo: University of Tokyo Press. 266 Toward the Rutral-Based Developnment of Commerce and In7dustry Ishikawa, Shigeru. 1988. "Problems of Late Industrialization: An Asian Perspective." In Kenneth J. Arrow, ed., The Balance Betweeni Induistry and Agricultuire in Economic Development, Vol. 1: 85-104. Hong Kong: Macmillan. Lindbeck, Assar, and Dennis J. Snower. 1988. The Insider-Oiutsider Tlheory of Employment and Unem- ployment. Cambridge, Mass.: MIT Press. Littler, Craig R. 1982. The Development of the Labiour Process in Capitalist Societies. London: Heinemann. Malavika, Karlekar. 1987. Poverty and Women's Work. New Delhi: Ahakti. Ofreneo, Rene E. 1994. "The Labour Market, Protective Labour Institutions and Economic Growth in the Philippines." In Gerry Rodgers, ed., Workers, Instituttions and Econonmic Growth in Asia. Geneva: ILO. Papola, T. S., and Gerry Rodgers. 1992. Labour Institutionis and Econonmic Development in India. Geneva: ILO. Sasaki, Jun. 1991, 1992. "Sanchi-menorimonogyo ni okeru rikishokki donyugo no tonyaseodo" ("The Putting-Out System after the Introduction of the Power Loom into a Cotton-Weaving Dis- trict"). Keizaigakut zasshii 91 (5-6): 110-29; 92 (1): 53-65. Sungsidh, Piriyarangsan, and Kanchada Poonpanich. 1994. "Labour Institutions in an Export-ori- ented Country: A Case Study of Thailand." In Gerry Rodgers, ed., Workers, Instituttionls and Eco- ionomic Growthi in Asia. Geneva: ILO. Thompson, Paul, and David McHugh. 1990. Work Organizationi. Hong Kong: Macmillan. Varnee, Purisinsit, and Benja Jirapatpimol. 1995. "Women's Industrial Work Conditions and the Changing of Family Relations." Unpublished manuscript, Department of Sociology and Anthro- pology, Chiang Mai University. Photocopy. 11 On the Relevance of East Asian Experiences: A South Asian Perspective V V Bhatt T he East Asian experiences have been presented in great detail in this volume and summarized with elegant simplicity and lucidity by Hayami in Chapter 1. These varied experiences are interesting in themselves. Their significance, however, is not restricted to the countries concerned; they have relevance for other developing coun- tries that are experiencing acute problems of rural underemployment and poverty. They at least suggest the feasibility of nonfarm rural industrial and commercial development, if it is organically linked with urban industrial and commercial growth. In this context, these experiences illustrate the effectiveness of relational contracting as a coordinating mecha- nism that links rural industry and commerce to urban-metropolitan and export markets. The objective of this chapter is to indicate the relevance of these experiences to the South Asian countries, which face severe, inter-linked problems of poverty and underemploy- ment and to present some Indian case studies of relational contracting. The Relational Contracting Model, East Asian Experiences, and Economic Rationale of Coordinating Mechanisms The stylized characteristics of the relational contracting model, as presented in Chapter 1 by Hayami, seem to be the following: (a) There is a long-term contractual relationship-or the expectation of such a relation- ship-between the urban principal and the rural agents, based on economic ties and fortified by sociocultural or community norms and traditions. 267 268 Toward the Rutral-Based Detvelopment of Commerce and Industry (b) There is interlinking of contracts, as in the putting-out contract, which implies a con- tract for provision of materials and/or credit as well as guaranteed purchase of fin- ished goods, or as in the advance-order contract, which interlinks the contract for purchase of finished goods with the contract for the provision of technical assistance and/or credit facilities. (c) Transaction costs are reduced through mutual trust and confidence generated by (a) and (b). The enforcement of contracts generally does not require resort to legal mecha- nisms, although they are essential as a last resort. Weaving Industry How well do East Asian case studies reflect this model? The Iruma (after 1890) and Banshu (in 1910s) cases (Chapter 3) conform to the model very closely; in both cases, yam was supplied by the principal to the rural workshops with a contract for guaranteed purchase of finished cloth. There was thus an interlinking of contracts. The merchants in Iruma and the principal weaver in Banshu were both located close to the rural workshops, and thus it is quite likely that there were both economic and social bonds between the principal and the agents. In the weaving industry of Thailand, the weavers are highly skilled and their products already have a market; the principal (merchant) in this case places only an advance order. There is no interlinking of contracts; there are economic ties, but one wonders whether there are any community ties between the merchant and the weavers. In the Korean case of Punggi, there is interlinking of contracts because materials are supplied by the mer- chant with guaranteed purchase of the finished cloth, but social or community ties do not appear to be important. The second Korean case is similar to the Thai case: there are eco- nomic ties formed by the advance-order contract, but there is no linkage with either tech- nical assistance or credit facilities. Social or community bonds do not seem to be important. Garnment Industry In the case studies of the garment industry, the Thai and Philippines cases are similar in some respects. In both cases, materials are given by the principal to the agents with the contract for guaranteed purchase of the finished garments. The Thai case, however, corre- sponds closely to the relational contracting model-the rural entrepreneurs are trained in garment factories, and this seems to have built mutual trust, although there may not be any community bonds. Further, both rural entrepreneurs and sewers belong to the same social stratum and live close to one another in a rural area, and neither group has access to an alternative occupation. These characteristics may have created a social or communal bond, resulting in mutual trust. The Philippines case is quite different. Most of the agents or entrepreneurs are located near a metropolitan area; are highly educated, although without much training in the gar- ment industry; and aspire to become principals by directly obtaining export orders. The On the Relevance of East Asian Experiences 269 sewers are also not trained as well as those in Thailand, they have no close social links with the entrepreneurs (because of differences in educational and economic status), and they aspire to work directly in metropolitan garment industries as sewers. Both economic and social ties are thus much weaker among principals, entrepreneurs, and sewers in the Philippines than in Thailand. For these reasons, no long-term, durable relationship, or expectation of such a relationship-which is the main characteristic of the relational con- tracting model-has evolved. The Chinese case displays some characteristics of the relational contracting model. Ad- vance-order contracting takes place between state enterprises and rural enterprises, and it appears that the state enterprises provide technical guidance as well as trained manpower. There seem to be some personal ties between the rural enterprise managers and managers in the state enterprises, because some of the rural managers probably worked in state enterprises in some capacity. The evolution of this form of relational contracting does not seem to have relied on the economic logic of the practice. Instead, it is probably the result of the incentives for state enterprises to evade or avoid government regulations by engaging in contractual arrange- ments with rural enterprises. Once these regulations change and the state enterprise sec- tor is restructured to function effectively and efficiently, the incentive for this kind of relational contracting may disappear, which renders the contractual arrangement unstable. Further, most of the Chinese enterprises are located in or near two cities and are quite large compared with enterprises in the Philippines or Thailand. The organization of enter- prises is more like that of the large urban garment factories in Bangkok or Metro Manila than it is like rural enterprises elsewhere, and in the absence of restrictive government regulatory policies, the scale of these enterprises would have permitted them to secure export orders directly, following the pattern of the large Thai and Philippine garment fac- tories. Thus, the nature and characteristics of relational contracting vary in different contexts. With the exception of the historical cases of the Japanese weaving industries and the Thai garmnent industry, the contractual arrangements in the other cases do not show all the characteristics of the relational contracting model. After all, a model captures only the stylized characteristics, and there are bound to be deviations from the model in actual cases. Further, as is shown in Chapter 10 by Kikuchi and Ohno, the contractual modes also differ by context, depending on the hold over products, working capital constraints, and the nature and size of the market. Toyota Case The Toyota case is much closer to the relational contracting model. The long-term relation- ship between the principal and parts suppliers was deliberately nurtured by the principal through advance-order contracts that are interlinked with technical assistance to the agents. There was a deliberate effort by the principal to forge close economic ties with the parts suppliers by encouraging them to locate near the main factory. This encouraged the for- 270 Towvard tHe Rural-Based Dezvelopment of Coninierce and Iidutstry mation of close economic as well as community ties and bonds, which created an environ- ment of mutual trust, facilitating not only the enforcement of contractual obligations but also fostering learning by parts suppliers through interaction with the principal, as well as among themselves. The Toyota system was the outgrowth of the founder's experience as a manufacturer of automatic looms, which he supplied to the weavers. He had experienced the benefits of having close contractual relationships with the parts suppliers of automatic looms, and he had probably found that given training and technical guidance, even rural artisans can be trained in the manufacture of the sophisticated parts required by the car factory. This may have been the reason for his vision of "a pastoral factory" in Koromo, a rural town not very far from Toyoda's automatic loom enterprise. He was thus familiar with the region and the community. Rice-Marketinig Case The rice-marketing case illustrates not only the entrepreneurial qualities among rural farm- ers and traders that come to the fore when transport costs are reduced and markets are expanded, it also shows that relational contracting is feasible in the trading sector. Al- though Kawagoe does not explicitly mention it, there mttust have been relational contract- ing arrangements within the hierarchy of traders or merchants. For example, wholesale merchants in cities or urban areas must have been linked through several chains of inter- mediaries to the retail merchants or agents of wholesalers who carried out the grassroots task of purchasing rice from the farmers. In such a case, there must have been advance- order contracting among the hierarchy of merchants. Such contractual arrangements are generally durable and long-term because of the code of behavior among merchants: a mercantile economy cannot expand without the mutual trust and confidence that is essen- tial if merchants are to engage in credit transactions with one another. The bills of ex- change involved may not offer any legal protection (Hicks 1969: Chapter 3). In Japan, tradable rice drafts, which are similar to bills of exchange, were used, and there must have been a code of conduct for honoring such drafts (discounting them and the like). Without such a code, financial transactions among traders would not have been possible. These considerations seem to suggest that there were advance-order contracts among merchants, and this relationship must have been durable and built on a tradition of trust; otherwise, the credit transactions represented by rice exchange notes or rice drafts would not have been possible. A farmer would have been paid in cash, while credit trans- actions must have been used among the merchants. This inference is consistent with the historical experience of other countries (Hicks 1969). Thus, the rice-marketing case reflects the main characteristics of the relational contracting model. Econonmic Rationalefor Relationial Conttracting Subcontracting by the principal to the agents can evolve only in industries with distinct and technologically separable production processes. In such industries-including gar- Oni the Relevance of East Asia7t Experiences 271 ment making, weaving, automobiles, tractors, bicycles, and electronics-separate processes have different economies of scale. For example, some processes, such as market search for final products; designing, finishing, selling of finished products; and purchase of basic inputs, require specialized skills and / or specialized equipment, and hence have economies of scale. Such operations must be carried out on a large scale and located at urban-metro- politan centers. If the products have export markets and/or require imported materials, the obvious location would be a port, as in the Thai, Chinese, and Philippines garment cases. The principal thus has the advantage of large-scale operations. Other processes or operations may have different scale economies, or may even be scale- neutral. To reap the advantages of scale economies, it would be advantageous to special- ize. In order to gain the economies of specialization or division of labor, the principal thus has an incentive to enter into subcontracting relationships with the agents, who specialize in selected processes. Such specialization improves efficiency in several ways. There would be learning by doing, with resultant identification of technical innovations that may reduce cost. The entrepreneurial incentive would be greater in such decentralized operations than in enter- prises that centralize production under one roof, as in the vertical integration system. Some processes may not require specialized skills and may be highly labor-intensive. If the opportunity cost of labor is low-in rural areas, for example-the principal would gain a cost advantage by forming subcontracting relationships with rural entrepreneurs. This can be seen in the Thai garment case. In Thailand, women had the skills and preferred to stay in rural areas and engage in operations that gave them flexibility in timing and hours of work, and their opporttnity cost was thus lower than it would be in Bangkok. Further, the principal does not have to invest in premises and equipment and has lower labor management costs. In this fashion, the principal can enjoy economies that arise from choice of techniques or technology, depending on labor market conditions. Subcontracting relationships also offer economies of scope. Production is flexible and can be adapted to changing demand conditions. This is particularly so in industries that pro- duce not one product but several. The garment industry is known for products that are differentiated in design and does not produce one standardized product, but several prod- ucts of different styles. Similarly, in the car industry, there are several models. The pattern of demand may change, but subcontracting enables the principal to adapt product mix to such changes. The subcontracting system thus has economies of scale, specialization, and scope in addition to economies arisingfrom choice of techniques or technology. Nevertheless, there are transac- tion costs, including transport costs. Contracts can never be complete, and there is always scope for implicit contracts. The initial costs of transactions in choosing subcontractors may be high, but they may not eclipse the economies of subcontracting. The problem is, how can the transaction costs arising from moral hazard and adverse selection be reduced further? There may be opportunistic behavior, cheating, and fraud. If the principal has to change subcontractors often and the agents must frequently change the principal, the trans- action costs could be prohibitive. Both parties gain if the contractual arrangements are- or are expected to be-enduring. 272 Toivard the Rural-Based Developmenit of Comnmzerce and Iindustry The risk of opportunistic behavior can be minimized if both the principal and the agents choose their partners after repeated transactions. This experience would provide each party with enough information to make an informed choice, and asymmetry of information would be reduced. Further, the principal would have an interest in establishing long-term, du- rable relations with subcontractors. Otherwise, his own operations would be adversely affected, and he has an incentive to take a long-term view. Similarly, the agents would be risk-averse and hence would take a long-term view; otherwise, they would have no incen- tive to undertake the risk of investment. With regard to the prisoners' dilemma game, it is shown that even if the game is played repetitively, treachery pays. Nevertheless, this result is valid only with the strong unreal- istic assumption of what Simon calls substantive ratioinality (Simon 1981: 32). As he demon- strates, this assumption is implausible in the face of bounded rationality. When this assumption is relinquished, and when players are striving for a satisfactory rather than optimal payoff, the cooperative solution is likely to be stable, even for finite repetitions of the game (Simon 1981: 46). This long-term relationship can be cemented by the interlinking of contracts. The prin- cipal can provide materials, technical assistance, or credit facilities along with the guar- anteed purchase of the product. It is easier to develop long-term relationships when both principal and agents share common sociocultural or community values and are linked by economic as well as sociocultural ties. Such long-term relationships create mutual trust and confidence, and thus reduce transaction costs (Arrow, 1974: 23-25). Relational contracting cannot evolve without such mutual trust and confidence in one another's integrity. Inside Contracting Inside contracting, which was prevalent in the United States in the late nineteenth century (Buttrick 1952), can also be considered as a form of relational contracting. The principal in this case provides not only materials, but also floor space and equipment to the agent; the agent is responsible for the technical aspects of production as well as recruitment and management of labor and receives a piece rate from the principal for completed goods. The relationship is based on mutual trust, because the principal chooses the agent on the basis of the agent's past performance, possibly as a worker. The agent has an incentive to be efficient and reliable because his continuing employment depends on his performance. This kind of relational contracting can evolve when there are economies of scale, not only in the design and marketing processes, but also in raising the necessary finance from banks or the capital market. Apart from these scale economies, the principal has an infor- mation advantage: since the agent is located close to the principal, the latter can easily detect any signs of opportunistic behavior. Thus, there is adequate assurance of the qual- ity of the product. The Indian garment industry case presented later in this chapter shows that insider contracting is relevant even now, when the associated additional economies are significant. On the Relevanice of East Asiani Experienices 273 Italiani District Model A variant of relational contracting has evolved to what has come to be known in the litera- ture as the Italian District Model. It has the following characteristics: * A cluster of primarily small and medium-size enterprises is spatially concentrated and sectorally specialized. * Relational contracting between the principal and agents is based on economic ties, buttressed by strong social or community ties or a sociocultural identity that facili- tates and promotes mutual trust. Such clusters, for example, have evolved in Emilia Romagna region of Northern Italy (Rabellotti 1995) and in the Sinos Valley in Brazil (Schmitz 1995). Both clusters have spe- cialized in the shoemaking industry, which is labor-intensive, like other traditional indus- tries such as weaving and garment making. The Italian District Model also has all the economies associated with relational contract- ing. In addition, such a cluster has external economies or economies of agglomeration, first discussed by Marshall in his Principles (Marshall 1920). These economies relate to the mar- ket for skilled workers, knowledge spillovers, and learning by interaction among firms. It is difficult to keep any market or technical secrets in a cluster like this, and the resulting external economies improve the collective efficiency (Schmitz 1995) of the whole cluster. Further, the cost of opportunistic behavior is heavy. Each firm's performance of their con- tractual obligations would be readily known, and those indulging in opportunistic behav- ior would suffer a loss of reputation and a consequent fall in orders, which could easily lead to the failure of the enterprise. The business code established in the cluster-an out- growth of social and cultural ties-prevents such behavior. Such clusters are likely to evolve in a marketing center located in rural areas in the rural-urban continuum, as the experi- ence of Italy and Brazil demonstrates. Of course, when the industries concerned lose their competitive advantage in the in- ternational market, the location of production could shift to other areas. For example, Italy and Brazil faced severe competition in the North American market in the 1980s from East Asian countries with relatively low labor cost. Further, Italy has specialized in classic leather shoes, while the demand pattern had changed in favor of synthetic tennis shoes made in Korea, Taiwan (China), and Hong Kong (China) (such shoes were also imported by Italy). Both Italy and Brazil had to shift to the production of higher-quality shoes for the elite market. Relational contracting modes had to change in order to adapt to this shift. It is useful in this context to refer to the New York experience. By the turn of the century, New York City was the center for garment making and distribution in the United States. After World War II, the industry was relocated to the southern states of the United States. Two decades later, the production centers shifted to East Asian areas such as Korea, Tai- wan (China), and Hong Kong (China), and later to China and South Asia as well. New York still has a competitive edge in garment marketing and design, which has consider- 274 Tohvard the Rutral-Based Development of Comnmerce and Industry able economies of scale. Indeed, much of the current production in Asia originates as an idea or an order in Manhattan (Hansen 1996). It appears from historical and current experience that with the expansion of the market, highly skilled activities such as weaving and shoemaking do tend to form clusters, in part because of economies of agglomeration and the consequent improved access to marketing channels. It is quite likely that such clusters in the weaving industry evolved in marketing centers in Iruma and Banshu. The Indian case of the garment industry presented later also shows the development of a cluster in a marketing center. The weaving industry of north- ern Thailand also appears to have evolved as an industrial cluster. In modem industries such as the production of automobiles, tractors, and bicycles, clus- ters of parts suppliers form around the principal. The Toyota case is an illustration of such a cluster, as are the Indian cases of tractor and bicycle production presented later in this chapter. Rationale for Nonfarm Development in South Asia With the current rate of increase in population and the extent of underemployment, the South Asian countries face a severe employment problem. The annual growth rate of the population in the region is more than 2 percent. In Pakistan, it exceeds 3 percent, while in Bangladesh, it is more than 3 percent, and in India it is 2.2 percent. The landless and small marginal farmers form more than 30 percent of the rural population in both India and Bangladesh, and 20 percent in Pakistan. As a result, rural unemployment and underem- ployment involves 35 percent of the rural work force in India, 60 percent in Bangladesh, and 15 percent in Pakistan. The proportion of rural population below the poverty line is quite high: 35 percent in India, 50-60 percent in Bangladesh, and 10-12 percent in Pakistan (table 11-1). The purely agricultural sector cannot absorb the surplus manpower. Instead, this sur- plus leads to fragmentation and subdivision of agricultural holdings and to an inciease in landless labor, which adversely affects agricultural productivity and intensifies poverty (Singh, 1990: Chapters 1-3). The modem industrial sector, even if it grew at an annual rate of 10-15 percent, cannot create employment opportunities at the rate required to absorb the burgeoning work force because it is small and the technology is capital-intensive (Bhatt 1973: Chapters 3-4; Singh 1990: Chapters 6-7). Further, because of large-scale rural underemployment, there is con- siderable migration from rural areas to the urban areas in search of employment, which aggravates the problems of urban congestion and leads to open urban unemployment and social, economic, and political tensions and conflicts. For example, twenty-three cities in India have a population exceeding one million and experience severe strain on the urban infrastructure-drinking water, sanitation and hygiene, housing, and intracity transport. To extend this infrastructure would probably be much more costly-both from the eco- nomic and social points of view-than to improve the infrastructure-roads, transport, On the Relevance of East Asian Experiences 275 Table 11-1. South Asia: Basic Economic Characteristics Item India Pakistan Bangladesh Sottth Asia A. Poputlation 1. Total population: (1989; millions) 832.5 109.9 110.7 1,156 2. Population, annual growth rate (percent), 1980-90 2.2 3.1 2.6 2.3 3. Urban population, proportion of total, 1989 (percent) 27 32 16 n.a. 4. Annual growth rate of urban population (percent), 1980-89 3.8 4.6 6.6 n.a. 5. Rural population: proportion of total (percent), 1989 73 68 84 n.a. 6. Annual growth rate of rural population (percent), 1970-79 1.88 3.0 2.40 n.a. 7. Rural population per square kilometer of agricultural area 345 280 990 n.a. B. Employment and potverty in rutral areas 1. Rural households: landless and small, marginal farmers, proportion to total (percent), 1985 33 20 32 n.a. 2. Open rural unemployment: proportion of total population of working age (percent) 6.7 2 10 n.a. 3. Rural unemployment and underemployment: proportion of total (percent) 35 15 60 n.a. 4. Rural population below poverty line: proportion of total (percent), 1980s 35 10-12 50-60 n.a. 5. Employment pattern: Proportion of rural work force (percent) (a) Agriculture 75 70 70 n.a. (b) Nonfarm 25 30 30 n.a. C. GNP per capita (US$, 1989) 340 360 180 320 Annual growth rate of GNP per capita: 1965-89 (percent) 1.8 2.0 0.4 2.9 (1981-90) Source: Singh (1990); World Bank (1991); Vaidynathan (1994); Visaria and Visaria (1994). communications, and power-in rural areas. As it is, the rural-urban continuum is already evolving: for example, in India, nonfarm employment is growing at a faster rate in towns with a population between 20,000 to 50,000 than in the urban areas (India, National Com- mission on Urbanization 1988; UNIDO 1995). There is no alternative to the promotion of highly labor-intensive commerce and indus- try if the related problems of underemployment and poverty are to be tackled effectively. 276 Towvard the Rriml-Based Development of Comnierce and Industry Such nonfarm rural development could provide employment opportunities in rural or semi-urban areas at much lower capital and social costs than similar employment devel- opment in large cities or metropolitan areas. At the same time, it would prevent migration from rural to urban areas, thus saving on costs of urban infrastructure as well as the indi- vidual costs, both social and economic, of migration (Bhatt 1960, 1980; Singh 1990). Such a rural-based development process, however, would require the institutional and organizational mechanisms that could link rural industry and commerce to urban-metro- politan and export markets. The relational contracting system is one such coordinating mechanism, as the East Asian experiences show, and it is quite relevant to the South Asian countries. Even the Indian cases, presented in the next section, illustrate the significance and effectiveness of relational contracting and other such models in ensuring a high rate of growth in the nonfarm sector in the South Asian countries. Appropriate systems have not yet been able to develop adequately because of the government's institutional and policy framework, which tended to support large-scale industrial enterprises. Licensing policies-industrial, import, and export-and the finan- cial system have been more geared toward the development of capital-intensive, large- scale enterprises than to labor-intensive, small-scale enterprises. Further, the policy of reserving selected areas for small enterprises in India became counterproductive. The small enterprises had no incentive to improve their technology by growing in size, and because large, urban-industrial enterprises could not enter these areas, they had no incentive to forge links with small enterprises through subcontracting mechanisms (UNIDO 1995). With deregulation and liberalization measures initiated since 1990 and the thrust to- ward rural, nonfarm development, the policy framework in South Asia is now more favor- able to the evolution of linkages between rural commerce and industry and urban commerce and industry. In the changed context, the relational contracting system or its variants may evolve, provided the constraints on infrastructure (roads, transport and communication, and power), technical and vocational education, and technological and marketing infor- mation are removed. Cases of Rural Industrialization in India Organizational and Technological Innovations in Marketing Milk The tradition of livestock rearing and dairying in the South Asian countries is centuries old, and one-fourth of the world's cattle reside in the area. (For source materials for this case study, see Thakur 1978; Jul 1979; Paul 1982; and Singh 1990.) Of the gross value of agricultural production, 15 percent in India and 25 percent in Pakistan is accounted for by this activity (Singh 1990). This is a household enterprise, managed by village women as their side occupation. These Indian household enterprises faced several constraints until the early 1950s. First, there was the marketing constraint. Although metropolitan and city demand for milk had On thle Relevance of East Asian Experiences 277 been increasing at more than 6 percent annually, the demand for milk has a high income elasticity. Milk, a perishable commodity, could not be transported over long distances be- cause the transport technology was inadequate. Private merchants purchased milk only in volumes they could sell in nearby towns or cities, and they did not provide stable and remunerative prices. Thus, milk producers had no incentive to increase the supply of milk by adopting innovative processes such as artificial insemination of cattle. In addition, vet- erinary care, extension, fodder, and other essential inputs needed to improve milk produc- tion were unavailable or inadequate in most villages. These marketing and production constraints limited the supply of milk, even when the urban-metropolitan demand was increasing and milk prices were rising. An attempt was made by the state (provincial) governments to meet the urban demand for milk by setting up modern dairy plants in metropolitan centers, close to consumers rather than close to rural areas, where milk was produced. These dairies promoted the export of cattle and cattle owners from rural areas to cities rather than the export of milk. The cattle colonies proved very costly, partly because of the high transport costs of cattle and feed and the managerial monitoring costs of cattle colonies. In addition, these cattle colonies could not supply adequate milk to the modem dairy plants, which were operat- ing much below their capacities and at a loss. This experiment in vertical integration proved very costly. The milk producers in Kaira District of Gujarat State in India developed a different approach. At the suggestion of their highly respected political-social leader, they created a cooperative structure for milk marketing in 1946. The Kaira Cooperative Union consists of a two-tier system-the District Cooperative Milk Producers' Union (milk union) and the Village Cooperative Milk Producers' Societies (milk societies). The village-level coopera- tive consists of rural members, each of whom buys a share in the society and also pays a membership fee. All members agree to sell their surplus milk to the society. The members elect a managing committee, which elects its own chairman. The society employs three to six persons for the day-to-day business of measuring and testing the milk and making payments to the milk producers. The milk is collected twice a day, morning and evening. The milk producers are paid in cash for the morning delivery in the evening, and for the evening delivery the next morning. The price depends on the fat content of the milk and is kept stable in real terms, while the society agrees to buy whatever volume of milk the sellers want to sell. Thus the marketing constraint is removed. The district union, formed by the village societies, collects the milk from the village societies and transports it to the union, which has its headquarters in a district (like a county in the United States) center (the town of Anand). There are currently more than 850 village societies, and each society has 300-400 members. At the union, milk is pasteurized and then sold as liquid rnilk. The remaining milk is converted into milk products at the plants of the union (established in 1955 and expanded in 1958 and 1960). Milk production has increased, but not only because of the removal of the market con- straint, although this provided an incentive for farmers to increase milk production. The capacity of milk producers to expand production has also been increased through the 278 Towoard thte Rural-Based Development of Conimerce an7d Indutstry introduction of technical inputs by the district union to the members of the village societ- ies, including balanced cattle feed concentrates, fodder seed, artificial insemination facili- ties, veterinary services, medicine, and the like as a part of their milk production enhancement and milch animal improvement programs. As a result of these activities and the provision of an assured daily income from the sale of milk, the milk cooperatives have increased the employment intensity of women and increased their incomes. Landless la- bor and small farmers account for more than 50 percent of the union's milk supply, and landless labor now derives 70 percent of its annual income from milk sales, while small farmers derive 30 percent. It has been estimated that one cross-bred cow does more to raise the standard of living in a landless household than giving the household two to four acres of irrigated land in most parts of India (Singh 1990). The technological development that made this possible involves the transport sector and the adoption of modern dairy technology under the leadership of a dedicated and able dairy engineer. Expansion of the railway system and the introduction of long-dis- tance hauling using refrigerated rail cars made it easier to transport milk to cities located as far as 300 miles from the district town. Milk powder production and chilling and feeder- balancing plants for milk made it easier to collect milk from rural areas and to store and transport it to the cities. Thus, a capital-intensive technology to process and market milk has been successfully combined with a relatively simple labor-intensive production tech- nology traditionally used by rural women. The union was very careful in selecting the chief executive of the dairy at Anand. It selected V. Kurien, a young man who had returned from the United States after receiving a degree in dairy engineering. He supplied the requisite managerial and technological expertise, without which the experiment would not have succeeded. The union managing committee had the good business sense to give him complete operational autonomy and to allow him to build his own team of executives and experts. The cooperative structure became feasible in Kaira because of two traits of the farmers: they have always been self-reliant and possess entrepreneurial traits. These qualities made them pioneers in the adoption of high-yielding seed and fertilizer technology. The farmers are also part of a community, which has strong cultural community links because the farmers belong to the same caste and have a highly evolved code of cooperative behavior, based on mutual trust and confidence (Pocok 1972). Their reputation in the community depends on how well they keep their word and how honest and truthful they are, not only in busi- ness but also in social dealings. This contracting pattern has been successfully replicated in the Kamatak State of India (Alderman 1987). Since 1970, attempts have been made to extend the Amul pattern in other districts of India. So far, 267 districts have been brought into the system, but it ap- pears that, with the exception of Gujarat and Karnatak States, the experiment is facing critical shortages of fodder, veterinary and breeding services, and refrigerated transport facilities. Unlike the pattern seen in Anand, the cooperative infrastructure is not evolving through the entrepreneurial spirit of the farmers, but as a result of the efforts of the Na- tional Dairy Development Corporation (NDDC), established in 1965 with the active assis- On the Relevance of East Asian Experiences 279 tance of the top managemnent of AMUL (UNIDO 1995). The net benefits of this NDDC experiment, Operation Flood, have not yet been objectively evaluated. Garmenit Indutstry in a Snmall Town Tiruppur, a small town 50 kilometers due east of Coimbatore (district town in Tamil Nadu State), is a hive of textile-related industrial activity (for source materials for this case study, see Cawthorne 1995). The town lies in the middle of the cotton belt and, as a result, gin- ning, spinning, and weaving mills are highly concentrated in this area. This local resource base has facilitated the development of the garment industry, which produces cotton knit- ted cloth and garments. It is one of the three main centers of this industry in India; the other two are Calcutta (in West Bengal State) and Ludhiana (in Punjab State). The industry currently has more than 1,500 small enterprises. More than 100 of the en- terprises are large (each employing 50-100 workers), 151 are medium-size (each employ- ing 20-49 workers), and the remainder are small workshops. Of the small workshops, some are family-operated enterprises (relying solely on family labor). The industry has grown very rapidly since 1960 to meet the growing demand for garments in the other states of India and to accommodate the growth in exports after 1975. In 1960 there were only 100 enterprises with total sales of about $3 million and total employment of 3,000 workers. There were few exports. By 1985 the number of enterprises was 1,500, with total sales of $80-90 million and exports of about $30 million; employment exceeded 40,000 workers. Sales have now grown to $1.8 billion, of which exports supply $600 million. The share of this town in total all-India output and exports exceeds 40 percent. The industry is highly segmented both in markets and in the kind of garments pro- duced. The small workshops use a very coarse knitted fabric and produce simple gar- ments such as underpants and undershirts, largely for the local market. The medium-size enterprises cater to the all-India market and produce a variety of garments with sophisti- cated designs. The large enterprises have contracts from abroad and produce garments to the specifications (design, fabric, dyeing, and finishing) of foreign buyers; the quality of garments for export is quite sophisticated. The profit margins for exports are much higher than for products in the local or all-India markets. All entrepreneurs are of rural origin, and some are from villages near the town. Those who own medium-size and large enterprises come from trading backgrounds. Owners of small enterprises belong to the class of landless farm labor. They originally migrated as workers for medium-size and large enterprises and then became entrepreneurs. The en- trepreneurs of medium-size and large enterprises are educated, and quite a number of them have high school and college degrees, while the owners of small shops generally have only rudimentary education in primary schools. The large enterprises receive orders from exporters and agents of foreign multinational firms, the medium-size firms from agents of wholesalers from other states of India, and the small shops from local wholesalers. The large enterprises contract out work to me- dium-size and small workshops that specialize in one of the production processes (all of 280 Towvard the Rural-Based Developntent of Comnmerce and Industry them contract out work for specific processes). Because the processes are distinct and sepa- rable, this extreme specialization has been possible. The processes involved are twisting and sizing of yarn, cloth making (knitting), dyeing, bleaching and finishing, and garment making. Large enterprises concentrate on receiving export orders and contracting out work for each process to medium-size and small enterprises; the remuneration is on a piece-rate basis. They merely supply the designs; the necessary materials are bought by the subcon- tractors. The large enterprises also have another form of subcontracting. Some of them own, but do not manage, several workshops, each specializing in a selected process. They engage a subcontractor, who is responsible for doing what is called the "job" work (work related to a process). The subcontractors are provided with the machines and raw materi- als necessary for a particular "job." These subcontractors are often individuals who have worked for the large entrepreneurs, and they were appointed as subcontractors because of the large entrepreneurs' familiarity with their expertise, skills, and management ability. The subcontractors hire labor as required and are responsible for managing it. Eventually some of these subcontractors start their own small workshops. Contracting out is common to all enterprises. It enables even a small enterprise to en- gage in the production of garments with little capital and a small number of workers. Because of the concentration of interconnected, extremely specialized enterprises in a sort of industrial cluster, there are economies of agglonmeration. Enterprises are able to obtain information on designs and processes from each other, through formal and informal shar- ing of information and close contacts. There is intense competition in the industry, and contracts can easily be made afresh; partnership relations can be changed easily both by the principal and the agents. Since familiarity and trust arise out of repeated transactions, however, the contractual relation- ships are generally long term and durable. Changing such a relationship involves transac- tion costs and risks, unless the terms are not met. One subcontractor works for several principals, depending on his capacity and skills, but his relationship with all of them is enduring. The workers include men and women, as well as children, and whole families are some- times also employed by large enterprises. The wages and other emoluments are lower than in large-scale factories, because small enterprises are exempt from all labor laws, including minimum wages. The hours of work are longer than in factory enterprises. Government help is of two kinds. In the field of exports, the Apparel Export Promotion Council conducts market research studies, identifies export potential, and explores new markets for the further development of exports. It sponsors trade delegations, study teams, and sales tours to markets abroad and it offers subsidies of approximately 50 percent, depending on the size of their units. The second form of assistance is credit on preferential terms from the nationalized banks, which also assist in the dissemination of market infor- mation by arranging meetings and seminars. Active associations in Tiruppur-the South Indian Hosiery Manufacturers' Association and South Indian Textile Research Associa- tion-also provide technical and market information. On the Relevance of East Asian Experiences 281 Because of relational contracting and rural-based industry development, this sector has proved to be superior in profitability to some vertically integrated, city-based enterprises. Some large enterprises in metropolitan centers such as Calcutta could not compete with small enterprises and were operating at a loss by 1960. All three centers-Calcutta, Ludhiana, and Tiruppur-currently have similar production and marketing patterns; Ludhiana, a small town, is more like Tiruppur than Calcutta. Modern Sutbcontracting in the Bicycle and Tractor Indutstries in Ptunjab BICYCLE INDUsTRY (for source material for this case study, see Singh 1994). By the mid- 1950s, there were seventeen bicycle producers, which were dominated by metropolitan- based, large-scale units with foreign collaboration. Set up by experienced businessmen in Madras, Calcutta, and Bombay, these large companies had established themselves in the growing market for bicycles in India. Within a decade, however, the up-country manufac- turers (such as Hero, Atlas, and Aron), located in small towns in rural areas, and set up with little funding and indigenous technology, emerged as the dominant actors in the Indian market. This was when the metropolitan units started incurring losses brought about by the organizational problems created by the integrated nature of their operations: high wage costs, strikes, other labor problems, and lack of subcontracting. While these older units gradually lost their position in the market, the three up-country units claimed more than 80 percent of the market by 1991. The share of Hero Cycles was 46 percent. Hero Cycles found a place in the Gutiness Book of World Records with their annual produc- tion of 2 million cycles. This enterprise was started by the Munjal family, which had mi- grated from Pakistan in the wake of the partition of old India. The family started with a small bicycle repair shop, then moved into components manufacturing, and converted this into a bicycle manufacturing unit by 1951. The firm had started manufacturing com- plete cycles by 1956. The family had previous experience in the bicycle industry when they lived in the former West Punjab (now in Pakistan). It belonged to the Khatri Caste, which is traditionally associated with business success. A local legend would have us be- lieve that even if a Khatri puts ashes on his head (that is, becomes a holy man), he will still make a profit. Hero is a family business, managed by the members of the extended family, and there is no separation from ownership and control, as happens in corporate enterprises. Further, the technology used is highly labor-intensive and based on indigenous machinery called "Punjab-type machines." These are assembled locally in a process learned through "re- verse engineering" of foreign machines, which makes replacement of machines easier and less costly than it is for metropolitan-based firms, which relied on the capital-intensive technology of imported machines. Further, Hero relies for 80 percent of its parts on a large number of subcontractors (more than 100). The enterprise itself has helped the establishment of small family workshops, which are the parts suppliers to the main enterprise. It helps these units in the acquisition of scarce raw materials and the selection of machinery and provides them with technical 282 Towvard the Ruiral-Based Development of Comnierce and Indutstry guidance. It also encourages its own employees to set up their own bicycle component manufacturing units and provides them with funds, market support, and other facilities. This durable and long-term contracting relationship with the parts manufacturers, who use highly labor-intensive technology as well as family labor to the extent possible, has improved the cost-effectiveness of Hero Cycles. The just-in-time inventory system, made famous by Japanese automobile companies, has been applied very effectively: suppliers are clustered around the main firm and orders can be placed and received in a matter of hours. The three-tiered system of large, small, and very small units in the bicycle industry ensures that waste is minimal; even the scrap material is used so efficiently that it reminds us of the waste recycling industry in Japan. The subcontractors are of rural origin and belong to farming families. Their educational levels are not very high-most have schooling up to the primary or secondary level, al- though a few may hold a high school degree. But this class has mechanical talents. Some of the members of the families have military experience, and some entrepreneurs are them- selves retired military personnel or rural artisans. Further, the Punjabees like to operate their own family enterprises rather than work as factory workers in a large enterprise. This is the reason for the large number of small, family-operated enterprises in Punjab. The durability of the contractual relationship is based on two factors. First, most of the subcontractors have been actively assisted by Hero Cycles as a part of its strategy, and thus have a strong sense of loyalty to the main enterprise. Second, Punjabees have a strong sense of ethical conduct and pride in achieving "excellence" in performance. Thus, the principal and the subcontractors form a cormnunity, and the community spirit is enhanced by the family, social, and cultural ties that bind them; such ties are stronger in Punjab than elsewhere. The government policy framework has been helpful. For refugees, the government pro- vided factory sites on a deferred-payment basis at a very low rate of interest; loans to build workshops and to purchase machinery at low rates of interest; and a special allotrnent of quotas for materials in short supply, such as steel and iron. Moreover, the Punjab govern- ment was among the first to set up a state financial corporation with the objective of pro- viding medium- and long-term financial assistance to industrial concerns. Government also exempted all small-scale units from paying excise duty on production and from all labor laws. TRACrOR INDUSTRY (this section is based on Bhatt 1978). The domestic production of trac- tors started in India in 1963-64 to meet the increasing demand from farmers, who had adopted the new hybrid seeds/fertilizer technology. Two firms were set up, one in col- laboration with Massey-Ferguson of the United Kingdom and the other with Eicher of the Federal Republic of Germany. The capacity of the two firms, however, was not adequate to meet the growing demand. By 1972, nine other companies were started, including Swaraj Tractor. Swaraj Tractor started with an annual capacity of 5,000 tractors in 1974 and is currently manufacturing 55,000 tractors each year. It is now the third-largest manufac- turer of tractors and is the only one to cultivate export markets. On the Relevance of East Asian Experientces 283 While all other units were based on collaboration with foreign firms and on imported inputs, Swaraj was the only firm based on Indian technological research and indigenous inputs. Its first product was a 20-horsepower tractor, which was suited to the relatively small holdings in Punjab-68 percent of holdings in Punjab were smaller than 9 hectares- and accounted for 20 percent of the stock of tractors in 1971. Swaraj was the only firm that provided a distribution network and after-sales service, and it was the only firm that de- veloped capability in the fields of R&D, tool design, and industrial engineering and tool making. These capabilities made it possible for Swaraj to introduce a 35-horsepower trac- tor as well as a small, economical tractor by 1980. The factors that made Swaraj Tractor an internationally competitive, viable enterprise also made it much less capital-intensive than the other tractor manufacturers in India. Its R&D capability has been used to change the design of components to cut down capital investments and to introduce innovations to conserve capital. The major reason, however, has been its departure from the vertical integration system and the adoption of the rela- tional-contracting system. It received 80 percent of its parts/components from subcon- tractors; that is, 780 kinds of components. To achieve this end, it assisted more than 100 small workshops around its location in Mohali, a small town near Chandigarh (the capital of Punjab), over a period of three years. These workshops were given technical assistance, including designs of components and technical training, and the expectation of a long- term relationship, provided they met quality standards (which most of them have). Sev- enty subcontractors are located within a radius of 10 kilometers from Swaraj, and the rest are within 20 kilometers. Obtaining parts in time has thus not been a problem. As with the bicycle industry in Punjab, all these subcontractors are of rural origin and have links with farm families. These are family workshops. The characteristics of the Punjab entrepreneurs outlined in the bicycle case also apply here. Swaraj Tractor also established a network of dealers in rural areas, providing them with technical training and informa- tion. As a result, it has promoted a large number of rural entrepreneurs in the commerce sector. As with the subcontractors, Swaraj also has a long-term and durable relationship with dealers. The government played a role in designing the 20-horsepower tractor-the Central Mechanical Engineering Research Institute did the research required. Because the tractor design was new, private entrepreneurs were not willing to undertake its production. The Punjab government, through its Punjab State Industrial Development Corporation, be- came the promoter, with financial assistance from the national development bank, the Industrial Development Bank of India. Comparative Evaluation of the East Asian and Indian Experiences The Indian experiences are comparable to the East Asian experiences, and it is interesting to note similarities and differences between the two. 284 Towvard the Rutral-Based Developnient of Commerce and Indlistry The Rice and Milk-Marketing Cases The rice-marketing case (see Chapter 2 by Kawagoe) of Japan from 1868 to 1930 and the Indian milk-marketing case have several common elements. In both cases, the develop- ment of transport and communications brought about the integration of local markets into a large, national market and created opportunities for rural entrepreneurs to emerge. The institutional innovation of rice-grading warehouses (beikeni s6ko) in Japan made it possible for a sophisticated market to evolve; it performed the function of inspection/grading, repackaging of rice into standardized qualities, and storage. These warehouses were started not only by private companies, but also by farmer cooperatives. A similar function was performed in India by the district union (a cooperative enterprise): it collected milk, graded it by fat content, and stored and processed it for marketing in a modem dairy plant. This institutional mechanism was created by rural entrepreneurs through the formation of a cooperative structure for milk marketing. Private merchants performed the marketing func- tion in Japan, while a cooperative structure formed by small milk producers looked after marketing in India. In the milk-marketing case, relational contracting was built into the cooperative mar- keting structure itself. The principal guaranteed the purchase of whatever milk the pro- ducers were willing to sell. There has also been interlinking of contracts: the principal supplies some inputs, technical assistance, and some other services. The long-term rela- tionship between the principal and agents is based on strong community links. Thus, this case has all the characteristics of the relational contracting model, discussed earlier. In Japan, relational contracting must have evolved among the hierarchy of merchants- rice collectors/retail traders, urban wholesalers, resalers, and retailers-although Kawagoe does not explicitly mention such evolution. There must have been a long-term relation- ship among these merchants; one can infer such a relationship on the basis of such trading systems elsewhere (Hicks 1969). The innovation of rice bills/drafts is similar to the bill of exchange elsewhere, and such a credit arrangement is not possible without a code of con- duct among the merchants. Without such a code of conduct, credit transactions cannot take place, for the transaction is based on mutual trust. Thus, as in the Indian case, the Japanese case has all the characteristics of the relational contracting model: a long-term relationship based on mutual trust and advance-order contracting interlinked with credit transactions. In both the cases, there is product specification by the principal: in the milk case, fat content of the milk, and in the rice case, the quality of rice. In the language of Chapter 10, the classification system of Kikuchi and Ohno, both the rice and the milk case are of the advance-order contracting I form. The Garment Indutstry The garment industry cases of Thailand (Chapter 6 by Ohno and Jirapatpimol) and the Philippines (Chapter 5 by Kikuchi) are also similar to the Indian case in some respects. On the Relezvance of East Asian Experiences 285 There is a stable, long-term relationship between the principal and the agents based on close contacts in both the Thai and the Indian cases. In Thailand, the rural entrepreneurs were known to the principal garment factories, and in India, the subcontractors lived in close proximity to the principals. In the Philippines, the economic and social bonds do not seem to be as strong as in Thailand or India because of the differences is social status arising from education and economic class, and the long-term relationships lack the stabil- ity observed in the Thai and Indian cases. In all three cases there is interlinking of con- tracts-a contract for guaranteed purchase linked with the provision of product specification. In Thailand and the Philippines, there is also provision of materials. In India, materials are supplied to the subcontractors for all processes except the twisting and siz- ing of yam. Thus, in Thailand and the Philippines, the putting-out contract I has been used, while in India, advance-order contracting I has been used in one process and a put- ting-out contract I has been utilized in all the other processes. The Thai and Indian cases have all the characteristics of the relational contracting model, while in the Philippines case, mutual trust is not as strong as in India or Thailand. The division of labor or specialization among subcontractors seems to be greater in the Indian case than in the Thai or Philippines cases. In Thailand and the Philippines, the subcontractors largely specialized in sewing operations. In the Philippines, for example, although Kikuchi mentions that some processes-such as cutting cloth, embroidery/smock- ing/appliqu6, sewing, and special processes including beading, injection molding (for haners), and knitting-are done by specialized contractors, the majority of his sample sub- contractors (149 out of 189) seem to specialize only in sewing, as in Thailand. In contrast, in India the division of Labor and specialization appear to be much greater than in Thai- land or the Philippines. The processes include twisting and sizing of yam, knitting, dye- ing, bleaching and finishing, and garment making; each subcontractor specializes in one of the distinct and separable processes. This has been possible because both the principals and subcontractors are located in one town, while in Thailand and the Philippines, the subcontractors are largely located in villages and small towns and the principals are lo- cated in metropolitan centers. In the Philippines, Laguna I and Laguna IV are in regions adjacent to Metro Manila, but there is no close proximity between the principals and the subcontractors as in India. The major difference between India and the other two cases is that in India, all princi- pals and subcontractors are located in what was earlier described as an industrial cluster. In this respect, the Indian case resembles not only the relational contracting model but also the Italian District Model. In addition to the economies associated with thte relational contract- ing model, the Indian case has the economies of agglomeration, discussed earlier. Further, insider contracting is also prevalent in the Indian case. The subcontractors in all three cases are of rural origin. In the Thai and Philippines cases, however, the subcontractors and their employees are largely women, while in the Indian case, the subcontractors are men, although some of their employees are women and children. In Thailand, the subcontractors had previous experience as line supervisors in large, urban-based garment factories, while in the Philippines they do not seem to have 286 Towvard the Rural-Based Development of Commerce and Indistry such experience and have other occupations as well. In India, as in Thailand, the owners of small workshops had previous experience as employees in large workshops, while the owners of medium-size workshops came from trading backgrounds, as is the case with some subcontractors in the Philippines. A lack of adequate previous training and strong personal ties between the principals and the subcontractors (because exporter-contractors in the Philippines are large, incorpo- rated firms, located in big cities) seems to be responsible for the need for close monitoring of subcontractors by principals. In India and Thailand, subcontractors have close personal ties with the principals that make complicated monitoring unnecessary. The education levels of subcontractors in India and the Philippines are higher than those of their counter- parts in Thailand. In all three cases, the market is segmented into local, national, and export sectors. The rapid growth of the garment industry in all three countries since the 1970s and 1980s has largely been the product of the opportunities available in the export sector. Government assistance in all three cases has been generally similar: tax incentives, provision of market information, provision of infrastructure, and so forth. In all these cases, the rural labor market is free from strict government regulations, and government policies encourage the growth of small enterprises. The Toyota and Tractor Cases With regard to forming close personal ties with suppliers and providing them with mana- gerial and technical assistance to improve their products, the Toyota case (Chapter 4 by Wada) is somewhat similar to the Swaraj Tractor case in India. In the early stages, how- ever, up to almost 1950, Toyota did not provide adequate managerial and technical assis- tance to its suppliers. This was the "keiretsiu diagnosis," promoted by the Small and Medium Enterprises Agency, which taught Toyota a lesson in how to advise, monitor, and grade its suppliers, as mentioned by Wada. Thus, it was only in the early 1950s that Toyota shaped an active policy of upgrading the managerial and technical competence of its suppliers, although the earlier close personal ties and the inculcation of strong community bonds certainly helped this process. This new, active strategy made it possible for Toyota to mar- ket its first completely domestically produced car, the Crown, in 1955 and to introduce into its plants what is called the "supermarket system"-a forerunner, as Wada mentions, of the later "just-in-time method." In the tractor case, the tractor was of indigenous design, as was Toyota's Crown. Even before producing tractors, however, Swaraj had decided to rely on subcontractors for 80 percent of its parts by establishing a relational contracting system. Over a period of three years, it actively assisted more than 100 small workshops around its location in the small town of Mohali in producing tractor parts/components by giving technical assistance, including designs of components, and technical training. These activities to foster close personal ties, and the extra-economic ties of sociocultural bonds, created a strong partner- ship relationship between Swaraj and its suppliers. Otn the Relevaance of East Asian Experiences 287 In the cases of both Toyota and Swaraj, all subcontractors were of rural origin and had links with farm families. Government policies were quite helpful to both. The bicycle case is similar to the Swaraj case and has certain common elements with the Toyota case. In the bicycle case, however, the entrepreneur has been a migrant, initially started a repair shop, and later became a bicycle manufacturer. He actively supported all his subcontractors and had close economic as well as sociocultural ties with them. This case shows how repair experience can generate skills, initially for making parts and later for the manufacture of the final product. Repair experience, like trading, can be a seed-bedfor entrepreneutrship. Thus, all three cases conform to the relational contracting model: there are long-term relationships based on economic as well as social ties and interlinking of contracts (advance-order contracting linked with technical guidance) that cements long- term relationships, as do the community bonds. Some product specifications are given by the principal, and all three cases represent advance-order contracting I. Because the parts suppliers and the principal are located not very far from one another, there are also econo- mies of agglonteration, or of an industrial cluster, discussed earlier. All three, although mod- ern enterprises, are located in small towns in rural areas, which allowed entrepreneurs of rural origin to emerge. Conclusions The East Asian and Indian case studies suggest thefeasibility of rural industrial and com- mercial development through the coordinating mechanism of the relational contracting system that links rural development with the urban industrial and commercial sector. Case studies can only be suggestive or illustrative, they cannot prove or disprove any hypoth- esis. Nevertheless, it is essential to inquire into the social and economic environment that can lead to such rural-urban linkages. As Schumpeter (1975) has emphasized, there must be opportunities for entrepreneurs to emnerge. In rural areas, such opportunities depend on the growth of the agricultural sector. As Kaldor (1967) has noted: The growth of the secondary and tertiary sectors is dependent on the growth of the 'agricultural surplus'; that is the excess of food production over the food consump- tion of the food producers themselves. This aspect of development was first empha- sized by Adam Smith . .. the emergence of a progressive agriculture was the key to the progress of industrialization in Europe. It is no accident that in England, as else- where in Europe, the so-called 'agricultural revolution' historically preceded the "in- dustrial revolution." . . . Land reform, with the consequent agricultural revolution, has also played a vital role in the development of Japan after the Meiji Restoration. For rural industrialization, however, one additional condition appears to be essential, and that relates to the agrarian economic and social structure. In countries where large landlords dominate the agrarian sector, rural nonfarm employment does not seem to in- 288 Toward the Rutral-Based DeVelopment of Conimerce and Induistry crease substantially. In England, the enclosure movement led to farms dominated by large landlords, and peasants were reduced to the status of wage earners or migrated to the urban areas. Rural industrialization and the emergence of rural entrepreneurs are possible only when farm size is relatively small and there is peasant farming. This condition pre- vailed in countries such as Italy, France, and Japan. Rapid agricultural growth under conditions of small-scale peasant agriculture creates opportunities for the emergence of rural entrepreneurs in commerce, industry, and ser- vices because of the backward and forward linkages of agriculture with industry and com- merce. Backward linkages comprise trade and industry relating to agricultural inputs such as fertilizer, agricultural implements, and high-quality seeds. Forward linkages relate to marketing and processing of agricultural produce. With rising incomes, opportunities arise in the manufacture and trade of basic consumption goods such as processed food, cloth- ing, shoes, milk, and milk products. Service sector opportunities relate to transport, re- pairs of agricultural implements, and consumer durables such as radios, bicycles, watches, and the like. With large-scale agriculture, these linkages promote urban industrialization and com- merce. Landlords, with their urban-metropolitan connections, generally prefer to satisfy their demand for products and services by importing them from urban areas, or even from abroad. Small-scale peasant farming promotes such linkages in the rural areas, and as the market expands with rising rural incomes, rural entrepreneurs tend to export their prod- ucts to urban areas, or even abroad. Rural entrepreneurs generally emerge from among the farmers, traders, artisans, and landless labor. For farmers, the inducement to start rural enterprises is to provide enter- prises and work for the members of growing families; all of them cannot be absorbed in agriculture because of the scarcity of land. Relatively rich farmers branch out, through family members, into small rural enterprises, largely related to agriculture, such as trad- ing in agricultural inputs and products, manufacture of inputs for agriculture, and con- sumption goods such as processed food and clothing for expanding rural and urban markets. There are retail traders in rural areas, and once rural production increases, their market expands and some of them form links with urban areas, while others tend to start small industrial enterprises. The traditional skills of the artisans prompt them to expand their production, first for the rural market, and later for the urban markets. Landless labor may initially find employment in all kinds of rural enterprises and, with improved skills and experience, may also start other enterprises. Because of capital constraints, the enterprises would be small, and with expanding mar- kets, there would be increasing specialization and division of labor, as Adam Smith em- phasized long ago (Smith 1970). Because of family connections and strong community links, these opportunities would be seized by establishing close linkages among the spe- cialized small enterprises. In time, with expanding markets, such linkages would develop with urban export markets. It is worth illustrating such evolution. In the Punjab State in India, food grain output increased by 9.5 percent annually during 1960-70, and 5.15 percent annually during 1970- On the Relevance of East Asian Experiences 289 79. Punjab has small farms. During this period, nonfarm employment rose significantly. The rural entrepreneurs emerged from among the farmers as well as from among the artisans. A large number of small enterprises emerged in products such as farm tools and equipment, (threshers, harrows, and tillers), woolen garments, bicycle and bicycle parts, sewing ma- chine parts, tractors, and machine tools (Pandit 1978; Little, Mazumdar, and Page 1987; Taub and Taub 1989; and Singh 1990). Because of linkages among families and strong community bonds, relational contracting evolved in sectors such as garments, farm equipment, bicycles, machine tools, and tractors. These.industries evolved in clusters in small towns; thus they have, in addition to eco1?omies of relational contractinig, economies of agglomeration as well. Their products are also exported to the urban areas and foreign countries. Similar development has taken place in the Kaira District of the Gujarat State in India. The rate of agricultural growth is similar to that of the Punjab State; like Punjab, there is small-scale peasant farming and strong community traditions (Pocok 1972). Out of fifty- nine small enterprises studied by Rutten (1992), thirty-nine were started by farmers; the rest were set up by artisans and traders. But all of them belonged to the rural area. In both the regions, Punjab as well as Kaira District, the population below the poverty line is much less than the all-India average; in Punjab, it is currently-less than 5 percent, as against the all-India average of 35 percent. Taiwan represents a similar case. It has the advantages of peasant farming; rapid agri- cultural development, and strong family, clan, and community bonds (Ho 1979). Historical and current experiences suggest that employment generation and industrial and commercial development in rural areas are highly correlated with the rate of growth of agricultural output. This evidence, however, comes mainly from areas that have experi- enced high agricultural growth rates over long periods, such as the Indian and Pakistan Punjabs, Kaira District in the Gujarat State of India, Malaysia, Taiwan (China), Japan, and some European countries, including Italy and France. That such an outcome is possible is shown by Mellor and Mudahar (1974) in their simulation model. They show that if food grain output grew at 4 percent annually over more than a decade in a country with an annual population growth rate of 2.2 percent and an initial underemployment rate of 30 percent of the labor force, the share of nonfarm employment would increase from its initial level to 55 percent, and underemployment would be reduced to 3 percent of the work force over a period of twelve years. The backward and forward linkages of agriculture with other sectors would bring about this result. Of course, no development, let alone rural development, can take place without ad- equate infrastructural development (roads, transport and communication, power, and the like) and institutional infrastructure in education, technological training and assistance, and the provision of information regarding existing and potential markets. Since the developing countries have a comparative advantage in labor-intensive indus- tries, they can seize this opportunity only if the trade policy is not only oriented toward exports, but also actively promotes them (World Bank 1993). These are some of the indus- tries where rural-urban linkages are feasible with a coordinating mechanism such as the relational contracting system or its variants. 290 Towvard the Rural-Based Developmnent of Commerce anid Indulstry References Alderman, H. 1987. Cooperative Dairy Developmenit in Karnatak, India: Ani Assessment. International Food Research Institute, Report No. 64, Washington, D.C. Arrow, Kenneth J. 1974. The Limnits of Organization. New York: W. W. Norton. Bhatt, V. V. 1960. Employmnent anid Capital Formiation in Underdeveloped Economies. Bombay: Orient Longman. .1973. Twvo Decades of Dezvelopmienit: The Indiani Experimenit. Bombay: Vora. .1978. Decision Making in the Puiblic Sector: A Case Stuidy of Swvaraj Tractor. World Bank Domes- tic Finance Studies, No. 48, Washington, D.C. .1980. Dezelopmenit Perspectives. Oxford: Pergamon. Buttrick, John. 1952. "The Inside Contract System." The Joutrnal of Economic History 12 (3). Cawthorne, P. M. 1995. "Of Networks and Markets: The Rise and Rise of South Indian Town: The Example of Tiruppur's Cotton Knitwear Industry." World Developmenlt 23 (1), January. Hansen, Gordon H. 1996. "Agglomeration, Dispersion, and the Pioneer Firm." Jornlal of Urban Econiomtics 39 (3), May. Hicks, John 1969. A Theory of Economiiic History. Oxford, U.K.: Oxford University Press. Ho, Samuel P S. 1979. "Decentralised Industrialization and Rural Development: Evidence from Taiwan." Economiiic Developmenit and Cultutral Chaange 28 (1), October. India, National Commission on Urbanization. 1988. Report or the National Commalission on7 Urban iza- tion. New Delhi. Jul, M. 1979. "Unexpected Benefits from a Dairy Project." Food and Nutrition Bulletin 1 (3), May Kaldor, Nicholas. 1969. Strategic Factors in Econtomnic Developnment. Ithaca: Cornell University. Little, Ian M. D., Dipak Mazumdar, and John M. Page, Jr. 1987. Small Manufacturing Enterprises: A Comparative Analysis of India and Other Couniitries. Washington, D.C.: World Bank. Marshall, A. 1920. Principles of Economics. London: Macmillan. Mellor, John W., and M. Mudahar. 1974. Simutlating a Developing Economy with a Modernizinzg Agricul- tutral Sector: Implicationisfor Employment and Economic Growth. International Food Research Insti- tute, Occasional Paper 76, Washington, D.C. Pandit, M. L. 1978 "Some Less Known Factors Behind Recent Industrial Changes in Punjab and Haryana." Econiomiiic and Political Weekly, November 25. Paul, Samuel. 1982. Manlaging Developmenit Programs. Boulder, Colo.: Westview. Pocok, D. 1972. Kanbi anzd Patidar: A Stuidy of Patidar Communniity in Glujarat. Oxford, U.K.: Clarendon. Putnam, Robert. 1993. Making Democracy Work: Civic Traditionis in Modernt Italy. Princeton, N.J.: Princeton University Press. Rabellotti, R. 1995. "Is there an 'Industrial District Model': Footwear Districts in Italy and Mexico Compared." World Developmeent 23 (1), January. Rutten, Mario. 1992. "Artisan or Merchant Industrialists? Small-Scale Entrepreneurs in the Coun- tryside of West India." The Jouirnal of Entrepreneutrship 1 (2), July-December. Schmitz, Hubert. 1995. "Small Shoemakers and Fordist Giants: Tale of a Supercluster." World Devel- opnment 23 (1), January. Schumpeter, Joseph A. 1975. Capitalism, Socialism, and Denmocracy. New York: Harper & Row, Harper Colophon. Simon, Herbert A. 1981. The Sciences of the Artificial. Cambridge, Mass: MIT Press. Oni the Relevance of East Asiani Experiences 291 Singh, Inderjit. 1990. Thte Great Ascent: Thte Rtral Poor in Sontth Asia. London: The Johns Hopkins University Press (for the World Bank). Singh, Sukhpal. 1994. "Refugees as Entrepreneurs: The Case of Indian Bicycle Industry." The Journtal of Entreprenteturshtip 3 (1), January-June. Smith, Adam. 1970. The Wealth of Nations. New York: Penguin. Taub, Richard P., and Doris L. Taub. 1989. Entreprenteuirshlip in India's Small Scale Inidutstries. New Delhi: Monohav. Thakur, D. S. 1978. "Impact of Dairy Development Through Milk Cooperatives: A Case Study of Gujarat." Indian Journral of Agricuiltntral Econzomiiics 33 (3), July. UNIDO (United Nations Industrial Development Organization). 1995. INDIA: Indlustrial Develop- mentt Reviewu. London: The Economist Intelligence Unit. Vaidyanathan, A. 1994. "The Employment Situation: Some Emerging Perspectives." Economic anzd Political Weekly 29 (50), December 10. Visaria, Pravin, and Leela Visaria. 1994. "Demographic Transition: Accelerating Fertility Decline in the 1980s." Ecoiionomic and Political Weekly 29 (51-52), December 17-24. World Bank 1991. World Developmenzt Report 1991: The Challenige of Developmenit. Washington, D.C. - .1993. East Asian Miracle. Washington, D.C. DI'RS TRIBUTORSOFGHANA Palestinian AuthontylMiddle East DISTRIBUTORS OF Center for Applied Research Epp Books Services Index Information Services W ORLD BANK Cyprus College P.O. Box44 OB. 19502 Jerusalem 6, Diogenes Street, Engomi TUC Tel: (972 2) 6271219 PU'BLICATIONS pO. Box 2006 Accra Fax: (972 2) 6271634 * ~~~~~~~~~~Nicosia ARGENTINA Tel: (3572) 59730 GREECE ITALY Oficina del Libro Intemacional Fax. (357 2) 66-2051 Papasotiriou S.A. Licosa Commissionaria Sansoni SPA Av. Cordoba 1877 35, Stoumara Str. Via Duca Di Calabda, 1/1 1120 Buenos Aires CZECH REPUBLIC 106 82 Athens Casella Postale 552 Tel: (54 1) 815-8354 USIS, NIS Prodejna Tel: (301) 364-1826 50125 Firenze Fax: (54 1) 815-8156 Havelkova 22 Fax: (301) 364-8254 Tel: (55) 645-415 E-mail: olilibro@satlink.com 130 00 Prague 3 Fax: (55) 641-257 Tel: (420 2) 242341486 HAITI E-mail: licosa§tlbcc.it AUSTRAUA, FlJI, PAPUA NEW GUINEA, SOLOMON Fax: (420 2) 24231114 Culture Diffusion URL: httlpAwwwftbcc.it/icosa ISLANDS, VANUATU, AND SAMOA 5, Rue Capois D.A. Intorrnaion ServicesURL: httpYA/www.nis.czl ,ReCpi D.A. nfDrmabion Services C.R 257 JAMAICA 648 Whrtehorse Road DENMARK Port-au-Prnce Ian Randle Publishers Ltd. Mitcharn 3132 SamfundsLifteratur Tel: (509) 23 9260 206 Old Hope Road, Kingston 6 Victora Rosenoems Alle 11 Fax: (509) 23 4858 Tel: 876-927-2085 Tel: (61) 392107777 DK-1970 Frededksberg C Fax: 876-977-0243 Fax: (61) 392107788 Tel: (45 31) 351942 HONG KONG, MACAO E-mail: irpl@colis.com E-mail: service@dadirect.com.au Fax: (45 31) 357822 Asia 2000 Ltd. URL: ht:pi/www .dadirectcom.au URL: htp1/www.sl.cbs.dk Sales & Circulation Department JAPAN 302 Seabird House Eastem Book Service AUSTRIA, ECUADOR 22-28 Wyndham Street, Central 3-13 Hongo 3-chome, Bunkyo-ku Geroid and Co. Libri Mundi Hong Kong, China Tokyo 113 Weihburggasse 26 Libreria Intemacional Tel: (852) 2530-1409 Tel: (81 3) 3818-0861 A-1011 'Men PO. Box 17-01-3029 Fax: (852) 2526-1107 Fax: (81 3)3818-0864 Tel: (43 1) 512-47-31-0 Juan Leon Mera 851 E-mail: sales@asia200O.com.hk E-mail: orders@svt-ebs.co.jp Fax: (431) 512-47-31-29 Ouito URL: httpilwww .asia2000.com.hk URL: htpJ/www .bekkoame.orjp/-svt-ebs URL: httpilwww.gerold.co/at.online Tel: (593 2) 521-606; (593 2) 544-185 Fax: (593 2) 504-209 HUNGARY KENYA BANGLADESH E-mail: librimu1 @librmundi.com.ec Euro Info Service Africa Book Service (E.A.) Ltd. Micro Industries Development E-mail: librimu2@librimundi.com.ec Margitszgeti Europa Haz Quaran House, Mtangano Street Assistance Society (MIDAS) H-1 138 Budapest PO. Box 45245 House 5, Road 16 CODEU Tel: (36 1) 350 80 24, 350 80 25 Nairobi Dhanmondi R1 Area Ruiz de Castilla 763, Edit. Expocolor Fax: (36 t) 350 90 32 Tel: (254 2) 223 641 Tel (880 2) 326427 Primer piso, O. #2 E-mail: euroinfo@mail.matav.hu Fax: (254 2) 330 272 Fax: (88112) 811188 Quito Fax: (881) 2) 8111S8 TelFax: (593 2) 507-383, 253-091 INDIA KOREA, REPUBUC OF E-mail: codeu@impsat.net.ec Allied Publishers Ltd. Dayang Books Trading Co. BELGIUM E 751 Mount Road Intematonal Division Jean Du Lannoy EGYPT, ARAB REPUBUC OF Madras - 600002 783-20, Pangba Bon-Dong, Socho-ku Av. du Rol 202 Al Ahram Distributon Agency Tel: (91 44) 852-3938 Seoul 1060 B(ussels Al Galaa Street Fax: (91 44) 852-0649 Tel: (82 2) 536-9555 Tel: (32 2) 538-5169 Cairo Fax: (82 2) 536-0025 Fax (32:2) 538-0841 Tel: (20 2) 578-6083 INDONEStA E-mail: seamap@chollian.net Fax: (20 2) 578-6833 ~~Pt. Indira Limited BRAZIL Fx(0)7-83Jalan) Borobudur 20 Eulyoo Publishing Co., Ltd. Publicactes Tecnicas Intemacionais Ltda. PBe Middl East Observer. Box 181 46-1, Susng-Dong Rue Peixoto Gomide, 209 ~~41, Sherif Street Jakarta 1 0320 Jongro-Gu 01409 Sac, Paulo, SRPar Tel: (62 21) 390-4290Seu Tel: (55 11) 259-6644ioSeu Fai: (55 11) 258-6990 Tel: (202) 393-9732 Fax: (62 21) 3904289 Tel: (82 2) 734-3515 E-mail: postmaster@pti.uol.br Fax: (202)393-9732 IRAN Fax (822)732-9154 URL: httpJ/www .uol.br FINLAND Ketab Sara Co. Publishers LEBANON Akateeminen Kirakauppa Khaled Eslamboli Ave., 6th Street Libraire du Liban CANADA PBo18Delafrooz Alley No. 8 PO. Box 11-9232 R,enouf Pjblishing Co. Ltd. FIN-00101 Helsinki PO. Box 15745-733 Beirut 5369 Canotek Road Tel: (358 0) 121 4418 Tehranl15117 Tel: (961 9) 217 944 Ottawa, Ontado K1JgJ Fax (358J0) 121443 Tel: (98 21) 8717819; 8716104 Fax: (961 9)217434 Tel: (613) 745-2665Fa.35 )1145 Fax: (613) 74S-7660 ~~~E-mail: akafilaus@stockmann.fi Fax: (98 21) 8712479 Fax: (613) 745-7660 ~~~~~~~~~~~~E-mail: ketab-sara@neda.net.ir MALAYSIA E-mail: order.dept@renoufbooks.com URL: htpiJvwwv akateeminen.comEl University of Malaya Cooperative URL: http:R/ www .renoufbooks.com FRANCE Kowkab Publishers Bookshop, Umited CHINA Editions Eska PO. Box 19575-511 PO. Box 1127 ChinaFinacial Ecoomic ublihing ouse5, avenue dv IlOp6ra Tehran Jalan Pantal Baru China Fancial & EconomicPublishingHouse a ris Tel: (98 21) 258-3723 59700 Kuala Lumpur 18 ) Tel: (331) 42-86-56-00 Fax: (98 21) 258-3723 Tel: (60 3) 756-5000 6333-8257 ~~~~~Fax: (33 1) 42-60-45-35 Fax: (60 3) 755-4424 Tel: (86 1 C) 6333-8257 Fax: (331 ) 424045-35 IRELAND E-mail: umkoopQtm.net.my Fax: (86 1) 6401-7365 (until July 1, 1998) Govemment Supplies Agency China Bor,k Import Centre W66, BavenkPublalions Olifg an tSolathair MEXICO Chin Boxk Impo5 Cet6W Bavnke Publinatin 4-5 Harcourt Road INFOTEC Boxe 225g 75116 Paris Dublin 2 Av. San Femando No. 37 BeXmng Tel: (331) 40-69-30-56/57 Tel: (3531) 661-3111 Col.TorielloGuerra COLOMBIA Fax: (331) 40-69-30-68 Fax: (3531) 475-2670 14050 Mexico, D.F Infoenlace Ltda. ISALTel: (52 5)624-2800 Carrera 6 Nqo. 51-21 GERMANY ISRAEtLitrtr t.Fax: (52 5) 624-2822 Apartado Aiereo 34270UNO-Veaag PO. Bot Liertue5 td E-mail: infctec@rtn.netmx Santa d BogoA D PoppelsdoerAllee 55 3 Yohanan Hasandlar Street U htl nn22 Tel: (57 1) 285-2798 Tel: (49 228) 949020 Tel Aviv 61560 Mundi-Prensa Mexico S.A. de C.V. Fax: (57 1) 285-2798 Fax: (49 228) 217492 Tel: (972 3) 5285-397 c/Rio Panuco, 141-Colonia Cuauhtemoc COTE D'IVOIRE URL: http:IJww.uno-vedag.de Fax: (972 3) 5285-397 06500 Mexico, D.E Center dEdifon et de Diffusion Africaines E-mail: unovedag@aol.com ROY. Aineatol Fax (52 5) 514-6799 04EDA) 541 Tel Aviv 61130 Abidjan 04 ~~~~~~~~~~~~~~~~~Tel: (972 3) 5461423 Tel: (225)2.46510;246511 Tel (9723)54142 Fax: (225) 25 0567 E-mail: royil@netvision.net.il NEPAL PORTUGAL SWITZERLAND 1784 Sofia Everest Media Iniemational Services (P) Ltd. Livraria Portugal Libraide Payot Service Institutionnel Tel: (359 2) 76 81 57 GPO Box 5443 Apartado 2681, Rua Do Carmo 70-74 Cotes-de-Montbenon 30 Fax: (359 2) 76 35 34; 76 27 84 Kathmandu 1200 Lisbon 1002 Lausanne E-mail: chr@mgu.bg Tel: (977 1) 472152 Tel: (1) 347-4982 Tel: (41 21) 341-3229 Fax: (9771) 224 431 Fax: (1) 347-0264 Fax: (41 21) 341-3235 HUNGARY Foundation for Market Economy NETHERLANDS ROMANIA ADECO Van Diermen EditionsTechniques 112 Pf 249 De LindebontrlnOr-PublikaKes Compani De Librarii BucuresKt S.A. Ch. de Lacuez 41 1519 Budapest PO. Box 202, 7480 AE Haaksbergen Str. Lipscani no. 26, sector 3 CH1807 Blonay Tel: (361) 204 2951, 204 2948 Tel: (31 53)574-0004 Bucharest Tel:(4121)9432673 Fax: (361)2042953 Fax: (31 53) 572-9296 Tel: (40 1) 613 9645 Fax: (41 21) 943 3605 E-mail: ipargazd(hungarynet E-mail: JindeboD@worldonline.nI Fax: (40 1) 312 4000 URL: http:l/www .woridonline.nV-lindeboo THAILAND JORDAN RUSSIAN FEDERATION Central Books Distribution Global Development FoAum NEW ZEALAND Isdatelstvo 306 Silom Road RO. Box 925273 EBSCO NZ Ltd. 9a, Kolpachniy Pereulok Bangkok 10500Ama110 Private Mail Bag 99914 Moscow 101831 Tel: (662) 235-5400 Amman11110 New Market Tel: (7 095) 917 87 49 Far. (66 2) 237-8321 KENYA Auckland Fax: (7 095) 917 92 59 Legacy Books Tel: (64 9) 524-8119 TRINIDAD & TOBAGOoos Fax: (64 9) 524-8067 SINGAPORE; TAIWAN, CHINA; AND THE CARRIBBEAN Loita House, Loita Street, Me67 1 MYANMAR; BRUNEI Systemabcs Studies Ltd. . Box 68077 Oasis Official Hemisphere Publication Services St. Augustine Shopping Center Nairobi PO. Box 3627 41 Kallang Pudding Road #04-03 Eastem Main Road, St. Augustine Tel: (254 2) 330853/221426 Wellington Golden Wheel Building Trnidad & Tobago, West Indies Fax: (254 2) 330854/561654 Tel: (64 4) 499 1551 Singapore 349316 Tel: (868) 645-8466 E-mail: Legacy0torm-net.com Fax: (64 4) 4991972 Tel: (65) 741-5166 Fax: (868) 645-8467 E-mail: oasis@actrixgen.nz Fax: (65) 742-9356 E-mail: tobe@trinidad.net KOREA, Republic of URL: httpJ/www.oasisbooks.co.nz/ E-mail: ashgate@asianconnect.com Seiong Books, Inc. UGANDA 81-4 Neung-dong NIGERIA SLOVENIA Gustro Ltd. Kwangjin-ku University Press Limited Gospodarski Vestnik Publishing Grnoup PO Box 9997, Madhvani Building Seoul 143-180 Three Crowns Building Jericho Dunaska cesta 5 Plot 614 Jinja Rd. Tel: (822) 498-0300 Private Mail Bag 5095 1000 Ljubljana Kampala Fax: (822) 3409-0321 lbadan Tel: (38661)1338347; 132 1230 Tel: (256 41)251 467 E-mail: sjbk@mail.nuri.net Tel: (234 22) 41-1356 Fax: (386 61)133 80 30 Fax: (256 41) 251 468 URL: http:/1203.248.78.1/sejong/ Fax: (234 22) 41-2056 E-mail: repansekj@gvestnik.si E-mail: gus@switluganda.com NEPAL NORWAY SOUTH AFRICA, BOTSWANA UNITED KINGDOM Bazaar Intemational NIC Info A/S Forsingle tiKes: Microinfo Ltd. GPO Box 2480 Book Department, Postboks 6512 Etterstad Oxford University Press Southem Aftca PO. Box 3, Alton, Hampshire GU34 2PG Kathmandu N-0606 Oslo Vasco Boulevard, Goodwood England Tel: (9771) 22-29-83 Tel: (4722)97-4500 RO. Box 12119, NI City 7463 Tel: (441420) 86848 Fax: (977 1) 22-94-37 Fax: (47 22) 97-4545 Cape Town Fax: (441420) 89889 Tel: (27 21) 595 4400 E-mail: wbank@ukminfodemon.co.uk SLOVAK REPUBLIC PAKISTAN Fax: (27 21) 595 4430 URL: httpilwww .microinfo.co.uk Slovart G.TG. Ltd. Mirza Book Agency E-mail: oxford@oup.co.za Krupinska 4 65, Shahrah-e-Quaid-e-Axam The Stationery Office PO. Box 152 Lahore 54000 For subscription orders: 51 Nine Elms Lane 852 99 Bratislava 5 Tel: (92 42) 735 3601 IntemaKonal Subscription Service London SW8 5DR Tel: (42 7) 839471; 839472; 839473 Fax (92 42) 576 3714 PO. Box 41095 Tel: (44171)873-8400 Fax: (427) 839485 Craighall Fax: (44171)873-8242 E-mail: gtg@intemetsk Oxford University Press Johannesburg 2024 URL: http:/www.theso.co.ukl 5 Bangalore Town Tel: (27 11) 680-1446 THAILAND Sharae Faisal Fax: (27 11) 880-6248 VENEZUELA Chulalongkom University Book Center PO Box 13033 E-mail: iss@is.co,za Tecni-Ciencia Libros, S.A. Phyathai Road Karachi-75350 Centro Cuidad Comercial Tamanco Bangkok 10330 Tel: (9221)446307 SPAIN Nivel C2, Caracas Tel: (662)2187292 Fax (9221)4547640 Mundi-Prensa Libros, S.A. Tel: (582)9595547; 5035; 0016 Fax: (662)255 4441 E-mail: ouppak@TheOffice.net Castelto 37 Fax: (58 2) 959 5636 28001 Madrid TURKEY Pak Book Corporation Tel: (34) 914 36 37 00 ZAMBIA Diinya Intotel, A.S. Aziz Chambers 21, Oueen's Road Fax: (34) 915 75 39 98 University Bookshop, University of Zambia 100 Yil Mahallesi Lahore E-mail: libreriaOmundiorensaes Great East Road Campus 34440 Bagcilar-lstanbul Tel: (92 42) 636 3222; 636 0885 URL: httpJiwww .mundiprensa.coml PO. Box 32379 Tel: (90 212) 629 0808 Fax: (92 42) 636 2328 Lusaka Fax: (90 212) 629 4689: 629 4627 E-mail: pbc@brain.net.pk Mundi-Prensa Barcelona Tel: (260 1) 252 576 E-mail: dunya@dunya-gazete.com.tr Conseli de Cent, 391 Fax: (260 1) 253 952 URL: htpJ/www.dunya.conV PERU 08009 Barcelona Editorial Desarrollo SA Tel: (34 3) 488-3492 ZIMBABWE UNITED ARAB EMIRATES Apartado 3824, Lima 1 Fax (34 3) 487-7659 Academic and Baobab Books (Pvt.) Ltd. Al Hamim Stationary & Bookshop Tel: (51 14) 285380 E-mail: barcelona@mundiprensa.es 4Conald Road, Graniteside PO. Box 5027 Fax: (51 14) 286628 RO. Box 567 Sharjah PHILIPPzNES SRI LANKA, THE MALDIVES Harare Tel: (9716)734687 IntemaPonaEBosouc etrIc Lake House Bookshop Tel: 263 4 755035 Tel: (971 6)7384473 International Booksource Center nc. 1 00, Sir Chittampalam Gardiner Mawatha Fax: 263 4 781913 Pager (971 6) 9760976 1127-A Antipolo St, Barangay, Venezuela Colombo 2 Pgr(7 )707 Makat City Tel: (941)32105 URUGUAY Tel: (63 2) 896 6501; 6505; 6607 Fax: (94 1) 432104 Fax: (632) 896 1741 E-mail: LHL@sri.lanka.net BOOKSELLERS OF Libreria43cnica Uruguaya E-mail: LHL~~~~~~~~sd.lanka.net ~~~~~~Colonia 1543, Piso 7, Of. 702 POLAtDjIP bli hi Si SWEDEN WORLD BANK Casilla de Correo 1518 POLAND SWEDEN ~~~~~~~~~~~~~~~~~~~~~~~~~Montevideo 11000O Intemational Publishing SevceV Wennergren-Williams AB PUBLICATIONS Tel: (5982)490072 Ul. Piekroa 31/37 2 0, Box 1305 P B I A I N e:(9 )407 00-677 Warzawa S-17125 So)na Prices varyfrom country to country. Fax: (5982)41 3448 Tel: (48 2) 628-6089 Tel (46 8) 705-97-50 Consult your local booksellerfor Fax: (48 2) 621-7255 Fax: (46 8) 27-00-71 E-mail: books%ips@ikp.atm.com.pl E-mail: mailOwwi.se prices and availability. URL: http/Aww. ipscg.waw.pljps/expor BULGARIA Humanites Research Center PO. Box 1784 Ij E 1 t'alr11i11 R[-'Su irtCt-S I-rt' Lit'Siilu'L for 11Yt' in FU m()com it'- tZiti st'iii1ni7rs. Tliw'yx u (1~~~~~~~~~~~~~~~~~~iscL1lss iSS1 ll 1 't't)11oll111ikC ,1 pilwil(t2z t poii 7tt 1/t1(? 11 17'11t1'1Sti t) 1XP(T t7'1Sct, 11 ( *~~~~~~~~~~Ti'l I/fit )1 t C'le'i N ' u)Id clag-; v 11to ( 1d bY 'wl \?'l 9e TI8H S eE NWRLD B ANK K- tt'chiiical t7x'1 tiS;c. T116, /will N7t' 181 8 H Street. N.W. pir?tiCu11ar i11to'c-'St toIa() 1 '1? 1t'1S Ct)1lcti1& Washington. D.C. 20433. U.S.A. wit/i public p/liCl/. Telephone: (202) 477-1234 Facsimile: (202) 477-6391 Telex: MCI 64145 WORLDBANK MCI 248423 WORLDBANK World Wide Web: http://www.worldbank.or-g E-mail: booksa worldbank.org U3