Community, Market and State in Development
9780230_274587_01_prexvi.indd i
6/23/2010 5:04:45 PM
9780230_274587_01_prexvi.indd ii
6/23/2010 5:04:46 PM
Community, Market and State in Development Edited by
Keijiro Otsuka and
Kaliappa Kalirajan
9780230_274587_01_prexvi.indd iii
6/23/2010 5:04:46 PM
Editorial and selection matter © Keijiro Otsuka and Kaliappa Kalirajan 2010 Individual chapters © Contributors 2010 All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission. No portion of this publication may be reproduced, copied or transmitted save with written permission or in accordance with the provisions of the Copyright, Designs and Patents Act 1988, or under the terms of any licence permitting limited copying issued by the Copyright Licensing Agency, Saffron House, 6–10 Kirby Street, London EC1N 8TS. Any person who does any unauthorized act in relation to this publication may be liable to criminal prosecution and civil claims for damages. The authors have asserted their rights to be identified as the authors of this work in accordance with the Copyright, Designs and Patents Act 1988. First published 2010 by PALGRAVE MACMILLAN Palgrave Macmillan in the UK is an imprint of Macmillan Publishers Limited, registered in England, company number 785998, of Houndmills, Basingstoke, Hampshire RG21 6XS. Palgrave Macmillan in the US is a division of St Martin’s Press LLC, 175 Fifth Avenue, New York, NY 10010. Palgrave Macmillan is the global academic imprint of the above companies and has companies and representatives throughout the world. Palgrave® and Macmillan® are registered trademarks in the United States, the United Kingdom, Europe and other countries. ISBN: 978–0–230–27458–7 hardback This book is printed on paper suitable for recycling and made from fully managed and sustained forest sources. Logging, pulping and manufacturing processes are expected to conform to the environmental regulations of the country of origin. A catalogue record for this book is available from the British Library. A catalog record for this book is available from the Library of Congress. 10 9 8 7 6 5 4 3 2 1 19 18 17 16 15 14 13 12 11 10 Printed and bound in Great Britain by CPI Antony Rowe, Chippenham and Eastbourne
9780230_274587_01_prexvi.indd iv
6/23/2010 5:04:46 PM
A Festschrift volume for Professor Yujiro Hayami
9780230_274587_01_prexvi.indd v
6/23/2010 5:04:46 PM
9780230_274587_01_prexvi.indd vi
6/23/2010 5:04:46 PM
Contents List of Figures
ix
List of Tables
xi
Preface
xiv
Part I An Overview 1 Community, Market and State in Development: An Introduction Keijiro Otsuka and Kaliappa Kalirajan 2
Crossing Borders: A Tribute to Yujiro Hayami Gustav Ranis
3 Pushing Frontiers of Development Thought Cristina C. David
3 11 14
Part II Role of State 4
5
6
7
8
The Role of the Government in Facilitating TFP Growth during Japan’s Rapid-growth Era Shuhei Aoki, Julen Esteban-Pretel, Tetsuji Okazaki, and Yasuyuki Sawada Development Performance across Indian States and the Role of Governments Kaliappa Kalirajan, Shashanka Bhide, and Kanhaiya Singh
45
The Role of Traditional Chinese State and the Origin of Modern East Asia Debin Ma
64
Government Distortions of Agricultural Prices: Lessons from Rich and Emerging Economies Kym Anderson
80
The Human Capital Basis of the Japanese Miracle: A Historical Perspective Yoshihisa Godo
Part III 9
21
103
State to Community
Community Governance and Public-Goods Investments in China’s Villages Renfu Luo, Linxiu Zhang, Jikun Huang, and Scott Rozelle
123
vii
9780230_274587_01_prexvi.indd vii
6/23/2010 5:04:46 PM
viii
Contents
10
Poverty, Politics and Projects under Community Participation in Zambia Alain de Janvry, Hideyuki Nakagawa, and Elisabeth Sadoul
11 Political Competition and the Impact of Local Governance on Pro-Poor Programs Pranab Bardhan and Dilip Mookherjee 12
Promises and Realities of Community-Based Agricultural Extension Gershon Feder, Jock R. Anderson, Regina Birner, and Klaus Deininger
148
172
187
Part IV Community and Market 13 The Community Mechanism of Contract Enforcement: What are the Differences between Rural Communities and Industrial Clusters? Keijiro Otsuka and Tetsushi Sonobe 14
The Transformation of Hayami’s Village Jonna P. Estudillo, Yasuyuki Sawada, Kei Kajisa, Nobuhiko Fuwa, and Masao Kikuchi
15 Recent Developments of Agricultural Markets in East Africa Takashi Yamano, Yoko Kijima, Tomoya Matsumoto, and Megumi Muto 16 Private-Sector Industrialization in China: Evidence from Wenzhou John Strauss, Edward Y. Qian, Minggao Shen, Dong Liu, Mehdi Majbouri, Qi Sun, Qianfan Ying, and Yi Zhu 17
The Role of Markets for Managing Agricultural Risks in Developing Countries Peter Hazell
211 224
245
262
291
Notes
311
List of Contributors
317
Index
319
9780230_274587_01_prexvi.indd viii
6/23/2010 5:04:46 PM
Figures 4.1 4.2 4.3 4.4 4.5 4.6 4.7 5.1 7.1
7.2
7.3
7.4 7.5 7.6 7.7 8.1 8.2 10.1 10.2
Changes in Japan’s gross domestic product per capita in logarithm of real GDP expressed in 1990 US dollars Changes in productivity of Japanese economy relative to the US Changes in sectoral total factor productivity indexes during the postwar period in Japan Changes in net technology import and R and D expenditures relative to GNP in Japan Changes in number of Type A technological assistance contracts Ratio of net payment for technology import to GDP and productivity growth by industry Comparison of average years of schooling in Japan and the US Pattern of India’s economic growth from 1951 to 2008 Nominal rates of assistance to agriculture in High-Income Countries (HIC) and European transition economies and in developing countries, 1955 to 2004 (percent, weighted averages, with “decoupled” payments included in the dashed HIC line) Nominal rates of assistance to exportable, import-competing and all covered agricultural products, high-income and developing countries, 1955 to 2004 (percent) Nominal rates of assistance to agricultural and nonagricultural tradable sectors and relative rate of assistance, developing and high-income countries, 1955 to 2004 (percent, production-weighted averages across countries) Trade reduction and welfare reduction indexes for tradable farm products, by region, 1960 to 2007 (percent) Nominal and relative rates of assistance, China and India, 1965 to 2005 (percent) RRAs and log of real per capita GDP, India and Northeast Asian focus economies, 1955 to 2005 NRAs for Japan, Korea and China and date of accession to GATT or WTO, 1955 to 2005 (percent) The Japan/US ratios in average schooling, GDP per capita, and physical capital per worker The percentage of graduate courses in the total enrollment of tertiary education Operational rules under SIF, CDD and decentralization Timeline for Zambia’s social programs and elections
22 24 25 27 27 29 38 48
84
85
87 88 94 98 99 110 116 150 152
ix
9780230_274587_01_prexvi.indd ix
6/23/2010 5:04:46 PM
x
Figures
10.3 15.1 15.2 15.3
Targeting of projects by ward characteristics across programs Farm-gate prices of milk and bananas Simplified milk-market structure in Kenya Banana market before and after mobile phone coverage in Uganda
9780230_274587_01_prexvi.indd x
159 247 250 255
6/23/2010 5:04:47 PM
Tables 4.1 TFP decomposition 26 4.2 Number and composition of licenses by industry 28 4.3 Decomposition of R and D expenditure in terms of share in total R and D expenditure in selected countries 33 5.1 Size and income of India’s state and union territories (2005–6) 49 5A.1 Explaining the variations in real per-capita growth in the post-reform periods 51 5.2 Explaining the variations in real per capita growth in the post-reform periods 52 5.3 The pace of development and spatial balance in development: Selected human development indicators 56 5.4 The pace of development and spatial balance in development: Selected infrastructural development indicators 56 5.5 Trends in state’s own revenue receipts and revenue expenditures 57 5.6 Finance Commission (FC), Planning Commission and nonstatutory transfer of resources from the centre to states 58 5.7 Results of regression analysis of interstate disparity in state government expenditures 61 7.1 Nominal rates of assistance to agricultural and nonagricultural tradables and relative rate of assistance, Asian developing economies and Japan, 1960 to 2004 (percent) 90 7.2 NRAs for Asian developing-country farmers, by product, 1960 to 2004 (percent, at primary product level) 92 7.3 Trade reduction indexes, Asian countries and other regions, all covered tradable farm products, 1960 to 2004 (percent) 93 7.4 Welfare reduction indexes, Asian countries and other regions, all covered tradable farm products, 1960 to 2004 (percent) 94 8.1 Human-capital accumulation through school education in Japan and the United States 109 8.2 Comparison in average schooling by levels of education between Japan and the United States 112 9.1 The way in which village leaders acceded to their office in 131 rural China, 1998–2003 (%) 9.2 The differences in the number (amount) of annual public-goods projects (investments) and the way in which village leaders acceded to their offices over time in rural China, 1998 and 2003 133 9.3 Baseline analysis (OLS) of the impact of directly electing the village leader on public investment in rural China 134 xi
9780230_274587_01_prexvi.indd xi
6/23/2010 5:04:47 PM
xii Tables
9.4 9.5 9.6
10.1 10.2 10.3 10.4 10.5 10.6 10.7 10.8 11.1 11.2 11.3 11.4 11.5 11.6 11.7 11.8 11.9 11.10 13.1 13.2 13.3 13.4
14.1 14.2 14.3 14.4
Fixed-effect analysis of the impact of directly electing village leader on rural public investment in rural China (1998–2003) Relationship between public-goods investment, size of the project and reelection in rural China, 1998–2003 Fixed-effects estimation of the effect of the annual number of public-goods projects (amount of investments) in latest term and reelection (reappointment) of village leader in rural China Number of projects implemented under SRP I, SRP II and ZAMSIF Distribution of wards by number of projects Targeting of projects by ward characteristics across programs Poverty targeting: Choice of wards within districts based on relative welfare From votes to projects in presidential elections From projects to votes in the 2006 presidential elections From votes in local elections to projects From projects to votes in local elections District-wise allocation of sample villages Village characteristics in sample villages, 1978 and 1998 Land reforms Trends in public supplies of agricultural inputs Trends in farm productivity, incomes and wages Left Front share regressions Land-reform panel regressions Targeting of IRDP credit subsidies Targeting of mini-kits Targeting of fiscal grants Proportions of enterprise managers by former occupation and formal schooling in Bingo, 1968–98 Previous occupation and years of schooling of enterprise managers by timing of new entry in Jili, China Characteristics of enterprise founders in the electric appliance industry by timing of new entry in Wenzhou, China Characteristics of enterprise managers by period of establishment in the garment industry in the suburbs of Hanoi Number of households in Hayami’s village in the Philippines, 1966–2006 Adoption of modern rice varieties and yield in Hayami’s village, 1966–2006 Tenure distribution of plots in Hayami’s village in the Philippines, 1966–2006 Sources of household income in Hayami’s village in the Philippines, 1966–2006
9780230_274587_01_prexvi.indd xii
138 141
142 157 157 158 160 162 165 167 168 176 176 178 178 179 180 182 183 183 184 216 217 218
218 225 227 228 229
6/23/2010 5:04:47 PM
Tables xiii
15.1 Milk marketing of dairy farmers in Kenya 15.2 Average milk price by sales outlet in Kenya 15.3 Determinants of milk price at the household level (Sublocation fixed-effects model) 15.4 Mobile phone network coverage and ownership in Uganda 15.5 Ratio of banana sellers and sales by mobile phone coverage in Uganda 15.6 Market participation and mobile phone coverage/possession: Banana production 16.1 Distribution of sampled firms by industries and regions 16.2 Distribution of type of ownership across period of establishment 16A.1 Logit regression of whether the firm was surveyed 16.3 How many founders did the enterprise have? 16.4 The shareholder’s relationship with the chief leader at the time of establishment 16.5 Characteristics of founders at the time of establishment 16.6 Sources of initial funding 16.7 Sources of funding during firm growth (% of funds) 16.8 What was the distance between your company and the nearest company making the same product? 16.9 Which firms export? 16.10 Channels of export (sales percentages) 16.11 Characteristics of interviewed workers 16.12 How did you find the job in this enterprise? 16.13 Migration 16.14 Income (yuan) 2005 and holding shares 16.15 Fringe benefits 17.1 Types of risk and losses, and the effectiveness of local capacity to manage 17.2 Financial performance: BASIX weather-insurance portfolio
9780230_274587_01_prexvi.indd xiii
252 253 254 256 257 258 267 268 269 270 271 272 274 275 279 279 280 282 283 283 284 285 293 304
6/23/2010 5:04:47 PM
Preface A person with an untiring quest for frontier discovery and an unshakeable commitment to his teaching and research pursuits in development economics – that is the quintessential Professor Yujiro Hayami. He is one of the most versatile development economists of our time. The breadth and depth of his academic achievements and contributions stand testimony to this fact. One of the major contributions of Professor Hayami to development economics is highlighting the importance of “community” as an important economic system. Conventionally, development economists have classified economic systems based on the combinations of the involvement of state and market in economic activities. Thus, there is the “free market” system at the one end, the “central planning” system at the other extreme, or some mixture of the two. By his logical argument, Professor Hayami has introduced “community” as the third pillar of development along with the conventional pillars, market and state. His novel proposition of including “community” as one of the important pillars of development has been widely accepted by leading development economists, and also appreciated by policy-makers around the world. But the role of community in economic development is the critical issue in development economics that needs further exploration. This volume, which is entitled Community, Market, and State in Development, is a token of appreciation for the invaluable contributions he has made as an academician. It is not an ordinary, but a special festschrift volume in the sense that all the authors are internationally recognized academicians, and they have extended Professor Hayami’s development model of “community, market, and state” by making new contributions with a clear focus on the role of community in economic development, and its interactions with the state and the market. This volume is also unique because various studies were well coordinated by prior discussions between the editors and authors. Therefore the editors are thankful to the authors who have cordially accepted our request. It is the hope of both editors and authors that this volume will command wide readership and contribute to policy discourse, now and in the near future. These papers were presented at the Festschrift Conference held on 27 and 28 February 2009 at the National Graduate Institute for Policy Studies in Tokyo. Each presentation was followed by discussion with a specialist in the concerned area, and general discussion by participants at the conference. We are grateful to Professor Hayami, who was present at the conference, and distinguished personalities from around the world, for their valuable contributions to this volume. We would like to acknowledge the contribution of Prabhu Pingali, who kindly participated in the discussion during the xiv
9780230_274587_01_prexvi.indd xiv
6/23/2010 5:04:47 PM
Preface
xv
conference. We are thankful to the FASID staff and the GRIPS staff, particularly Ms Mayuko Tanaka and Ms Yoko Saigo, for their kind support in organizing the conference successfully. We would also like to thank Ms Tanaka, who tirelessly edited the manuscripts following the guidelines of Palgrave Macmillan, and Paul Kandasamy, who edited the language of the manuscripts. We are grateful for the financial support of the Global Centers of Excellence Program of the Ministry of Education, Culture, Sports, Science, and Technology of the Japanese government. We appreciate Palgrave Macmillan, especially Ms Taiba Batool, and Ms Gemma Papageorgiou, for their understanding and cooperation in bringing out this volume with flying colours. K EIJIRO O TSUKA AND K ALIAPPA K ALIRAJAN February 2010
9780230_274587_01_prexvi.indd xv
6/23/2010 5:04:47 PM
9780230_274587_01_prexvi.indd xvi
6/23/2010 5:04:47 PM
Part I An Overview
9780230_274587_02_cha01.indd 1
6/23/2010 12:15:50 PM
9780230_274587_02_cha01.indd 2
6/23/2010 12:15:50 PM
1 Community, Market and State in Development: An Introduction Keijiro Otsuka and Kaliappa Kalirajan
1.1
Aim of the book
The classical economists rejected the mercantilist doctrine of state intervention and regulation, and called for a retrenchment of state activity in order to allow for the operation of free-market forces. However, a market cannot exist in a vacuum. It is artificially created, and can function only if the players in the market, such as producers, wholesale traders, retailers, and consumers, follow the rules of the free market. The implication here is that the government has to play a very positive role to encourage the smooth functioning of the market by enforcing rules and regulations to eliminate constraints to overall economic growth and development. Thus, the successful operation of a free-market economy is greatly dependent on the effective functioning of the competent government. However, there can be market failure and government failure, either individually or together. Market failure occurs in the supply of public goods and can occur in the case of pure private goods too. The latter can happen due to information asymmetry and also due to the presence of externalities. The consequence of information asymmetry and externalities is that transactions in the market will be different from the socially optimum level. Then the market does not guarantee economic efficiency, and possibly worsens equity in income distribution. Thus, government intervention may be needed whenever the market fails or market-induced income distribution is not socially desirable. Hirschman (1958) earlier strongly argued for governmental intervention to counteract the “polarization effects” of freemarket forces in order to mitigate the misfortune of the “backward regions.” The state has a monopoly on legal coercion within the jurisdiction recognized by other states, and it may use this for good or ill, for the interests of the governed or the interests of those doing the governing. Drawing on Locke’s theory of the social contract, if governments were for the interests of the governed, this means that government is created by individuals to solve problems that they alone cannot solve. Governments are supposed 3
9780230_274587_02_cha01.indd 3
6/23/2010 12:15:50 PM
4
Keijiro Otsuka and Kaliappa Kalirajan
to mitigate market failure to improve the overall welfare of the society. The governments’ failure, however, occurs from (a) misuse of budgets, (b) diversion of principal-agent interests, and (c) excessive regulation due to seeking “institutional rents” (Hayami, 2001). Moreover, government cannot solve market failure arising from the problem of asymmetric information, as it does not have access to unobservable information. Herein lies the important role to be played by the community, whose members know one another and can often detect cheating, shirking, and other dishonest behaviors of its fellow members. In other words, as forcefully argued by Hayami (1998, 2001, and 2009), the community can support market transactions by reducing information asymmetry. The case of the jeepney in the Philippines discussed in Otsuka et al. (1986) provides an example, which vividly demonstrates how asymmetric information is dealt with by the community. Furthermore, as argued by Otsuka and Kalirajan (2006), the contribution of markets towards improving economic welfare will gradually be increased, in response to the new opportunities opened up by the introduction of new technologies. So, what are the sources of technological inventions and innovations? Who would discourage or promote technological change? The key process in the economic force of technological change is described by Schumpeter as the introduction of innovation, and the cultural innovator is the entrepreneur. Schumpeterian literature argues that technological innovations emanate from a class of entrepreneurial community members, who are constantly trying to be on the frontier of technology in terms of production and management. What drives entrepreneurs to undertake innovation is nothing but profit motives shaped by free-market forces. In this context, the role of “community” as an important economic system introduced by Hayami (1998) assumes added importance. A prominent question in this context is whether there will be any community failure, as in the case of market failure and state failure. Hayami (2001) has not overlooked this possibility. While admitting the possibility of community failure in the sense that it would take a long time for communities to adjust their norms and traditions to their changing resource endowments and technologies, he has pointed out that it would not be universally true all the time. He has cited the Japanese experience, which shows that communities’ adjustment capabilities cannot be underestimated. Japan developed its “sub-contracting” system, which is still being followed in many developing countries, through community-type cooperation. Subcontracting is particularly effective when new parts and components which embody new technological ideas are introduced. Needless to say, the introduction and development of the subcontracting system is an institutional innovation. The implication is that any economic system built with inappropriate reliance on state, community, and market should lead to inefficiency and inequality. In these circumstances, it may appear that market competition
9780230_274587_02_cha01.indd 4
6/23/2010 12:15:51 PM
Community, Market and State in Development
5
would be useful to minimize or eradicate the failures of state and community, but it may well be that a competitive market works if and only if it is supported by state or community, or both. Then, an important task of policy-makers is to promote the effective working of “induced innovation” mechanisms to adjust cultural and institutional environments to be consistent with competitive market systems. Though it may look like a difficult task for governments to perform, historical evidence, particularly based on East Asian performance, shows that it is possible even with those countries’ different traditions. Given the fact that state, market, and community are strongly interrelated with each other, and also that there can be government failure, market failure, and community failure, the most important question, drawing on Hayami (2009), is how to arrive at an optimal combination of the roles of community, market, and state in promoting a country’s welfare. As Karl Marx said, the institution that exploited productive potential at a certain stage of development would become a restraint on further development. The implication is that appropriate institutional innovations and the effective functioning of such institutions are required, not based on abstract theory alone, but more importantly on dynamically changing empirical realities to provide sustained economic growth to developing and developed economies. As aptly pointed out by Hayami (2009, p. 118), “how to combine the community, the market and the state in the total economic system is probably the most important agenda for economists geared toward the reduction of poverty in developing economies, as well as the maintenance of economic vitality and social harmony in developed economies.” Accordingly, this volume, which is entitled as Community, Market, and State in Development, explores the different modes of interactions of government, market, and community followed in different countries to promote the welfare of their people, drawing from both theories and realities reflected in empirical data. This volume is a token of appreciation for the invaluable contributions Hayami has made as an academician. We, the authors of this volume, are proud and extremely privileged to have had the opportunity to be among his colleagues, to have worked under his research guidance, to have been able to derive wisdom from his rich experience, and most importantly to have had a long-term association with such a notable and noble person. It is not an ordinary, but a special festschrift volume in the sense that the authors, all of them internationally recognized academicians, have extended Professor Hayami’s development model of “community, market, and state” by making new contributions with a clear focus on the role of community in economic development, and its interactions with the state and market. To our knowledge, there are no other books available with similar objectives, or which contain consistent empirical findings from different countries around the world.
9780230_274587_02_cha01.indd 5
6/23/2010 12:15:51 PM
6
Keijiro Otsuka and Kaliappa Kalirajan
1.2
Scope of the book
Following this chapter, there are two special tributes to Hayami’s research contributions on community, market, and state discussed by Ranis and David. Ranis has summarized Hayami’s contributions to the literature as follows: “His work has been both Main Street and Side Street, broad as well as deep, concerned with such well-travelled issues as the government and the market, education and rural poverty, as well as more esoteric subjects including the economics of the commons, the economic consequences of Japanese colonialism, the economics of agricultural contract choice (with our workshop organizer) and his innovative proposal (of course, not implemented) of an alternative land reform paradigm for the Philippines.” David has pointed out that Hayami, through his intensive and convincing research, has drawn the attention of researchers and policy-makers to the role of communities in economic development. She concludes by saying, “Surely, Yujiro Hayami is not just ‘anyone’ of Japan, but one of the great economists concerned with development in this generation.” The research findings presented in this book are divided into three parts under the following headings: (i) role of state; (ii) state to community; and (iii) community and market. Hayami has cautioned that any economic system built with inappropriate reliance on state and community would lead to inefficiency and increase in inequality. In this context, the first set of five papers concern discussions on state success, state failure, and community failure in improving the overall socioeconomic welfare of people. Aoki, Esteban-Pretel, Okazaki, and Sawada have documented the role of different government policies that had contributed to high growth in Japan after World War Two. They have found that in nonagricultural sectors, total factor productivity (TFP) growth occurred at first through the import of foreign technologies via licensing, and subsequently through the innovation of its own technologies. Unique to the case of Japan is the establishment of R and D consortia of enterprises by the government, which reduces unnecessary duplication of basic research activities and promotes research cooperation based on “community-type” inter-enterprise relationships. Following the successful intervention by the state in Japan in promoting economic growth, Kalirajan, Bhide, and Singh discuss the importance of the community–state nexus. They show that government in India within a federal framework has mechanisms that foster development equitably across its states, particularly through health and education expenditures aimed at promoting humancapital development. However, they point out that the scheduled tribe communities, which are economically backward, are not taking part effectively in the human-capital development programs implemented by states due to their social and cultural backgrounds. Thus, in this context of community failure in India, as Hayami (2001) has argued, the authors conclude that it is imperative for the states to intervene by approaching the communities layer
9780230_274587_02_cha01.indd 6
6/23/2010 12:15:51 PM
Community, Market and State in Development
7
by layer to win their confidence, and to educate them in the benefits of participating in welfare programs, something not found in tribal thinking. While traditional beliefs and norms have been the major reason for community failure in certain states in India, Ma discusses how traditional norms and cultures have laid the foundation for modern economic growth in China. The traditional Chinese state, due to its “dynasty” characteristic, simulates a community with ties that go beyond blood relations. Particularly, within this political structure, the dominance of an autocracy in a unified empire gave birth to an enduring bureaucratic tradition of technocracy and impersonal rule, and a tradition of community mechanism of contract enforcement and public-good provision at the local level. These institutional endowments paved the way for the rise of a modern nationalist ideology and a bureaucratic developmental state for modern East Asia. After examining the country-specific experiences of the role of state in promoting economic welfare and the interaction of communities with state in nurturing growth, Anderson examines the impact of growth on implementing macroeconomic policies such as agricultural trade policies. Drawing on results from a new multi-country World Bank research project, he summarizes evidence on the changing extent of distortions to prices of farm products nationally, regionally, and globally by examining the extent to which later-emerging economies of Asia and elsewhere are following the policy trajectory of a gradual reduction in anti-agricultural policies and eventual transition to pro-agricultural pricing. Behind such changes are the changing politics of interest groups formed by communities of farmers. Godo examines the education policy since the Meiji Restoration to the present. He argues that the Japanese government as well as business communities failed to foresee the emerging need for building a system of strong tertiary education, especially research-oriented graduate schools at a world-class level, before catch-up growth occurred in the 1980s. He concludes that government failure in tertiary education policy is one of the major factors underlying the poor performance of the Japanese economy since the beginning of the 1990s. One of the major contributions of Hayami to understanding the effectiveness of local governance in economic development concerns the role of communities. He has argued elegantly that communities play a central role in making local governments accountable and effectively implementing development programs. Expanding Hayami’s work on the community–state nexus, the next set of four papers discuss how communities influence state administration, and state dissemination of information such as extension advice in developing countries, by analyzing community-level data collected particularly from China, India, Zambia, Kenya, and Uganda. Luo, Zhang, Huang, and Rozelle, using data from nearly 2450 randomly selected villages, examine the link between village governance reforms and the provision of public goods in rural China between 1998 and 2004. They
9780230_274587_02_cha01.indd 7
6/23/2010 12:15:51 PM
8
Keijiro Otsuka and Kaliappa Kalirajan
conclude that democratization has increased the quantity of public-goods investment at the village level because those elected village leaders who are able to implement more public projects during their terms of office are more likely to be reelected. Thus, the rural community in China is able to influence the state administration in providing pubic good through their election process. Similarly, de Janvry, Nakagawa, and Sadoulet argue that Zambia offers a unique laboratory to analyze the role of community participation in the provision of local public goods. They have shown that poverty correlates positively with politics, reinforcing the value of local politicization of projects when there is stronger community participation. Their result vindicates the detailed work of Professor Hayami on the rural community’s role in complementing the functioning of state by effectively participating in the delivery of local public-goods. Their analysis also shows greater state involvement in reinforcing the capacity of the community to participate in Zambia, which is consistent with the findings in China. Bardhan and Mookherjee, using nearly 90 village surveys, analyze the role of community in making local governments accountable by examining the link between the distribution of development program benefits and voting behavior in the state of West Bengal in India. The results of the analysis show that the intravillage distribution of benefits has been pro-poor, reflecting effective functioning of communities to promote equity. On the other hand, the intervillage allocation of benefits does not seem to be pro-poor. This suggests some evidence for political clientelism in voter behavior. Feder, Anderson, Birner, and Deininger examine the conceptual underpinnings of community-based extension approaches to influence what is attempted and how it is achieved in such service systems. The organizational structures represent vehicles for bringing what Hayami called “the social capital” and engaged linkage to help in the delivery of information and other services that can be provided through advisory endeavor. They indicate that the community-based extension service in Uganda was a community failure, while that in India was successful. They conclude by saying that Yujiro Hayami’s work on social capital and communities can guide this field of research as it provides an insightful perspective for analyzing the potentials and limitations of the community mechanism in agricultural development. Hayami has forcefully argued that the community’s role in development is complementary to the state, and to the market too. Each pillar of development is prone to failure due to their functioning mechanisms; nevertheless, the failure of one can always be mitigated by the functioning of the other. The final set of five papers provide strong evidence from developing countries of Asia and Africa to support Hayami’s thesis that the community’s role is complementary to the market. Otsuka and Sonobe consider the fact that transactions among manufacturers, and between them and traders, are highly active in industrial clusters in a large number of countries.
9780230_274587_02_cha01.indd 8
6/23/2010 12:15:51 PM
Community, Market and State in Development
9
They argue that industrial clusters are equipped with an excellent mechanism of contract enforcement which reduces transaction costs and facilitates the division of labor. Thus, in the case of industrial clusters, drawing on Hayami’s forceful arguments, the community and the market are complementary, as the community mitigates market failures. Therefore, in this respect, industrial clusters are no different from rural communities. In order to sustain the development of industrial clusters, government must invest in infrastructure, such as the establishment of industrial zones, and provide credit. Hence, they conclude that community, market, and state, which are the three pillars of economic organizations ought to play complementary roles in the development of industrial clusters. Estudillo, Sawada, Kajisa, Fuwa, and Kikuchi jointly discuss how community has been playing a complementary role to that of state and market, to facilitate growth over four decades in the well-researched “Hayami’s village” in the Philippines. Through their primary long-term panel data on historical changes in the village, they have disproved the common belief that the community is the yoke of underdevelopment because it strictly obeys traditional institutions and norms that are slow to respond to changes in economic opportunities. In particular, this study reviews the multifaceted roles of rural communities in irrigation management, land and labor relations, risks coping, social capital, and the development of the rural nonfarm sector. They conclude that rural communities are far from static, but respond systematically and dynamically to changes in technology, resource endowments, and market conditions. Although the generality of the findings in one village would need to be interpreted with caution, they believe that the finding that community principles of mutual help and income-sharing tend to be consistent with the market principle of profit-seeking is valid in many rural areas in Asia. Then the question immediately arises as to how communities in rural Africa behave in relation to state and market. Yamano, Kijima, Matsumoto, and Muto, using panel data from Kenya and Uganda, have examined the role of community-based mechanisms on market transactions. They have shown how the technology-induced market expansion for banana farmers in Uganda has increased their income by reducing market transaction costs through community participation and cooperation among farmers. For example, as long as one representative among farmers has access to a mobile phone, this person can make arrangements with traders on behalf of fellow farmers. This would encourage farmers to form farmer groups, which may help farmers not only in terms of marketing but also in terms of technology diffusions or other knowledge sharing. Strauss, Qian, Shen, Liu, Majbouri, Sun, Ying, and Zhu compare the process of the Wenzhou industrialization to date with the East Asian model of industrialization, and observe certain commonalities, particularly the role of communities. They argue that just as in the case of Taiwan and Japan before
9780230_274587_02_cha01.indd 9
6/23/2010 12:15:51 PM
10 Keijiro Otsuka and Kaliappa Kalirajan
World War Two, it was small-scale family firms that initiated new industrial production in Wenzhou, with finance coming initially from family members and friends. Thus, the community mechanism of contract enforcement based on social and family ties was certainly critical in the early growth stages of these firms, as is emphasized by Godo and Hayami (2002). Hazell discusses the role of markets for managing agricultural risks in developing countries in the context of the withdrawal of many public risk-management schemes. He argues that due to the inability of farmers and rural communities to handle covariate yield and price risks over wide areas effectively, market-mediated solutions to risk management would be the preferred solution. He cautions that it is becoming clear that market-mediated approaches to risk management are not achieving on anything like the scale needed. Thus, he recommends the intervention of governments or international agencies like the World Bank to initiate and strengthen such activities. We believe that the diverse empirical studies included in this volume are complementary and contribute significantly to a deeper understanding of the roles of community, market, and state in economic development. The volume also indicates that much more research needs to be done in this area to shed clearer light on the proper division of labor among the three pillars of economic organizations.
References Godo, Y. and Hayami, Y. (2002) “Catching Up in Education in the Economic Catch-up of Japan with the United States, 1890–1990,” Economic Development and Cultural Change, 50 (4), 961–78. Hayami, Y. (1998) “Community, Market and State,” in C. K. Eicher and J. M. Staatz (eds.), International Agricultural Development, 3rd edition (Baltimore: Johns Hopkins University Press), 90–102. Hayami, Y. (2001) Development Economics: From the Poverty to the Wealth of Nations, 2nd edition (Oxford: Oxford University Press). Hayami, Y. (2009) “Social Capital, Human Capital and the Community Mechanism: Toward a Conceptual Framework for Economists,” Journal of Development Studies, 45 (1), 96–123. Hirschman, A. (1958) The Strategy of Economic Development (New Haven, CT: Yale University Press). Otsuka, K. and Kalirajan, K. (2006) “Rice Green Revolution in Asia and Its Transferability to Africa: An Introduction,” Developing Economies, 44 (2), 1–10. Otsuka, K., Kikuchi, M. and Hayami, Y. (1986) “Community and Market in Contract Choice: The Jeepney in the Philippines,” Economic Development and Cultural Change, 34 (2), 279–98.
9780230_274587_02_cha01.indd 10
6/23/2010 12:15:51 PM
2 Crossing Borders: A Tribute to Yujiro Hayami Gustav Ranis
Yujiro Hayami has always been ahead of the curve; he has always had what might be called a nose for the jugular. He has been an intellectual pioneer but never an intellectual ideologue. He has consistently crossed borders, across disciplines and within disciplines – even when it wasn’t fashionable. Let me count the ways in which he has reached across the various intellectual boundaries which many of us, ensconced in our various fiefdoms, have been all too prone to erect.
2.1 Between agricultural economics and development economics We all know this is where he started and where he never really left. His interest increasingly moved from Japanese agricultural development, historically and in the contemporary context, toward its role in overall economic development. His contrasting of the Japanese intensive with the Mexican extensive styles of agricultural development, and their differential impact on their respective historical development patterns, has been a truly seminal contribution.
2.2
Between agricultural and rural nonagricultural activities
Yujiro is one of the few economists who early on recognized the importance for overall economic success, in terms of growth as well as distribution and poverty alleviation objectives, of encouraging the blossoming of rural industrial and service activities in balanced and mutual enhancement with agricultural productivity change.
2.3
Between macro and micro analysis
Yujiro is one of the few researchers I know who is comfortable in both paradigms, as his prodigious bibliography indicates, ranging from broad-gauged 11
9780230_274587_03_cha02.indd 11
6/23/2010 12:15:57 PM
12
Gustav Ranis
treatises like Development Economics: From the Poverty to the Wealth of Nations to Anatomy of a Peasant Economy, in which he analyzes household behavior in “his” adopted village in great detail. What distinguishes his microanalysis is that, in contrast to many others laboring in this field, he has always been interested in mobilizing such research to gain a better understanding of the macro economy and macro policies, and not simply for its own sake or for impressing other, microoriented, development economists.
2.4
Between classical and neoclassical analytical tools
Yujiro was never doctrinaire in these matters. He is well-versed in the history of thought, again not a common phenomenon these days. Most of his contributions are quite orthodox. Yet in his work with Kikuchi on Indonesian and Philippine agriculture, to cite but one example, he was ready to deviate from the conventional wisdom which, of course, warmed the cockles of my partisan heart: ... the wage adjustment process toward equilibrium seems to involve a mechanism very different from the market mechanism assumed in neo-classical economics. The fact that labor’s share in the traditional output-sharing contract is institutionally fixed implies that wage rates cannot adjust directly to change in labor’s marginal productivity ... only through modification of institutional arrangements. Arthur Lewis and I agree. Even Ken Arrow, a high priest of neoclassical economics, argues that equilibrium is achieved only with time – in the case of developing countries it may be anywhere from 10 to 30 years – before the “turning point” is reached.
2.5
Between economics and neighboring disciplines
Increasingly, Yujiro ventured into the no man’s land of political economy and institutional economics. His work with Vernon Ruttan on induced institutional innovations was an early indication of this, and his collaboration with Masahiko Aoki, on the Institutional Foundations of East Asian Economic Development, a more recent one.
2.6
Between theory and policy
Yujiro was not a reclusive ivory-tower economist, but deeply concerned with policy issues. Domestically he labored long and hard over a lifetime – if not always successfully – to alter the Japanese government’s rice policy, for example via his volume on Japanese Agriculture Under Siege: The Political
9780230_274587_03_cha02.indd 12
6/23/2010 12:15:57 PM
Crossing Borders: A Tribute to Yujiro Hayami
13
Economy of Agricultural Policies, as well as via his shorter public policyoriented pieces and his participation in government rice price policy committees, which I well remember his complaining about. Abroad, he once again made major academic contributions, comparing the political economy of agricultural protection in various countries of East Asia as well as making several hands-on contributions as the agricultural economist working with the International Rice Research Institute in the Philippines; as a valued member of my Comprehensive Employment Strategy Mission for the Philippines, where he, inter alia, developed the idea that irrigation, the application of fertilizer, and resorting to intensive agriculture, are generally likely to eventuate once the land frontier is reached; as a member of the Board of Trustees at IFPRI; as a participant in the World Bank’s CGIAR review; to cite but a few examples.
2.7 Between being a prolific and an innovative contributor Yujiro Hayami has 18 books and roughly 200 journal articles to his credit. Yet, he seldom repeated himself – a very distinctive accomplishment in our profession – and managed to be consistently path-breaking and innovative. His work has been both Main Street and Side Street, broad as well as deep, concerned with such well-travelled issues as the government and the market, and education and rural poverty, as well as more esoteric subjects including the economics of the commons, the economic consequences of Japanese colonialism, the economics of agricultural contract choice (with our workshop organizer), and his innovative proposal (of course, not implemented) of an alternative land-reform paradigm for the Philippines. If there is one area where there is no “trespassing” or exploring boundaries, but just a walk – or rather, a dash – down Main Street, it is his love of tennis and the high quantity and quality of it.
2.8
Finally, between research, teaching, and service
He has done it all and done it all exceedingly well. I have always thought of Kazushi Ohkawa as the Simon Kuznets of Japan, and I think of Yujiro Hayami as the Ted Schultz of Japan, distinguished as a scholar, as teacher, and a kind caretaker of others. The Emperor of Japan has already recognized this remarkable individual’s contributions. All of us, and many others who are not here today, count themselves lucky to have had lots of intellectual and personal contacts with him. Here’s hoping that these will continue for many more years to come.
9780230_274587_03_cha02.indd 13
6/23/2010 12:15:57 PM
3 Pushing Frontiers of Development Thought Cristina C. David
The theme of this festschrift, “Markets, State, and Communities,” in honor of Yujiro Hayami, is quite appropriate. It represents, thus far, the culmination of the evolution of his economic thought and contributions. As noted in his latest writings, the most important agenda for economists currently concerned with the reduction of poverty in developing economies, the maintenance of economic vitality and social harmony in developed countries, and the preservation of the environment globally, is how to combine the roles of markets, the state, and the community in the total economic system suited to each country’s culture and social values to achieve these objectives; as by nature, each is subject to failures – market failure, government failure, and community failure. I am honored to express this tribute, on behalf of the authors of this festschrift, to one of the most deserving economists in this generation. Yujiro Hayami devoted his entire professional life to the relentless pursuit of understanding the process of economic change, not just as an academic interest but with the clear goal of influencing the course of economic development through appropriate policy and institutional reforms. The quality of his research is evident from its strong theoretical underpinnings: macro, micro, and down to grassroots perspectives, and with interdisciplinary appreciation. And he dealt with big rather than trivial questions. The breadth and depth of his research contributions are undeniably exceptional, and many are path-breaking. The findings of his numerous researches on wide-ranging issues, reported in more than 200 journal articles and book chapters, are integrated into a series of books in English, Japanese, and even Chinese editions. The analytical syntheses of literature contained in his books alone are invaluable to students and professionals. Let me highlight a number of his most significant contributions:
3.1 Agriculture development and induced innovation Hayami-Ruttan’s book entitled Agricultural Development: An International Perspective (1971 and 1985) won the Enduring Quality Publication Award of 14
9780230_274587_04_cha03.indd 14
6/23/2010 12:16:04 PM
Pushing Frontiers of Development Thought
15
the American Agricultural Economics Association for its definitive analysis of the agricultural development process. It is predictable that technological change embodied in modern inputs played the key role in explaining differences in agricultural productivity observed across countries, and this was found as expected; however, it was novel to empirically prove that technical change is endogenous, induced by changing factor endowments – as reflected in relative factory prices, changing output prices, and research production function. Moreover, rapid population growth on limited land supply and technical change were also found to induce institutional innovations to overcome market failures, such as property rights, financial markets, public investments in agricultural research and extension, irrigation, and other factors of land and market infrastructure. These factors facilitated the adaptation of new and borrowed technologies for different production environments and effective use by farmers.
3.2 Green Revolution, Institutional changes, and communities Yujiro Hayami, working with many others, wrote on a variety of issues related to the three topics above. But the unique and important body of work that profoundly influenced his thoughts on the process of economic development, and sharpened his grassroots perspective, is the set of intensive and extensive village studies conducted in the Philippines and Indonesia over three decades with Masao Kikuchi, as well as Toshihikoe Kawagoe. These village studies provided opportunities to observe and collect detailed data for learning about the complexity of village economic activities and analyzing agrarian change – how product and factor markets adjust in response to population pressure, technological change, land redistribution and tenancy regulations, commercialization, and rural industrialization; and their impacts on production, incomes and their distribution, class structure, and traditional institutions in a different agro-ecological, socioeconomic, and cultural environment than Japan’s. While recognizing the value of an interdisciplinary approach in fully understanding these processes, he consistently applied neoclassical microeconomic reasoning throughout the analysis, including those for understanding contractual arrangements, institutional changes, and even social norms. These village studies clarified the confusion about the effects of the Green Revolution and rapid growth in population, and land-reform regulations. Because modern rice technology was essentially scale-neutral, the adoption rate of modern varieties did not vary by farm size or by tenure. Indeed, their analysis indicates that it is the inadequate spread of the new technology in the face of rapid population growth and limited land supply that have caused the polarization in the villages. On the other hand, rapid population growth and land market regulations have deepened stratification of
9780230_274587_04_cha03.indd 15
6/23/2010 12:16:04 PM
16
Cristina C. David
village population, increasing the landless households and creating a class of subtenants. Factor market adjustments to changes in technology and population growth have not always been explicitly through adjustments in wages and rental rates, but were caused by changes in the traditional institutions for paying labor and land inputs, such as the shift from “hunusan” to “gama” in the Philippine case, or from the “bawon” to the “tebasan” system in the Indonesian case. The choice of labor contracts, and hence institutional rather than wage adjustments, can be explained by the efforts to minimize risks and transaction costs and, at least in the Philippine case, by the social norm of mutual help and income-sharing within the village community. By quantifying implicit wages under different institutional arrangements and alternative land rentals and market and technological conditions, the achievement of a reasonably efficient resource allocation in a neoclassical sense through these institutional adjustments was demonstrated. Indeed, it was further suggested that the villages’ social norm of mutual help and income-sharing within the community itself may also be explained by economic rationality. They do acknowledge that social norms or cultural factors may constrain factor market adjustments. For example, the caste system in India, which strictly establishes the division of labor among different castes, segmented the labor market, preventing landless households from meeting higher labor demand for tasks traditionally reserved for cultivator households, and thus may have promoted mechanization prematurely, and promoted social conflicts. Social norms and cultural factors may also determine how well communities are able to organize and perform economic functions, in which they perform better than the state or the market because of transaction costs, moral hazard, asymmetric information, or public-good characteristics. The deterioration of the irrigation system in the village after the state drastically reduced support for operation and maintenance was not arrested by efforts to shift responsibility to irrigators’ associations or the village community. Unlike in East Asia where communities successfully manage irrigation systems because they are tightly knit, as agro-ecological conditions historically required collective action for certain economic activities, the efforts to organize irrigators’ associations in the Philippine village were less successful, in part due to the loosely structured nature of the village community that had little experience of collective action.
3.3 Political economy of agricultural protection Yujiro Hayami’s versatility is again demonstrated in his series of papers quantifying the welfare effects of agricultural price and trade policies in Japan, and explaining their causes in political economy terms. With Anderson and others (1986), they analyzed how political economy factors explain the
9780230_274587_04_cha03.indd 16
6/23/2010 12:16:04 PM
Pushing Frontiers of Development Thought
17
rising trend of agricultural protection in the course of economic development in East Asia. But most illuminating is his 1988 book entitled Japanese Agriculture Under Siege: The Political Economy of Agricultural Policies, which analyzed the cause of Japan’s sharp increases in agricultural protection and the difficulty of reversing that policy, particularly for rice, drawing on his broad and deep understanding of his country’s agriculture, institutional structure, and political economy. Agricultural protectionism was initially a response to the problem of agricultural adjustment in the era of rapid economic development, based on technology-borrowing as the gap between farm and nonfarm income widened sharply. Efficient response to the changing comparative advantage through factor and product market adjustments, however, was severely constrained by land market regulations, rural industrialization, and the phenomenon of part-time farming. Increasing farm size to facilitate larger-scale mechanization lowers costs, raising labor productivity through increasing human capital and specialization of labor, and the shift to higher-valued products per unit of scarce land resources has been too slow. But even more insidious is the political power exerted by vested interests involved in the marketing of agricultural outputs and inputs to maintain high rates of agricultural protection. In fact, under the pressure to conform with international trade agreements, Japan managed to increase market access, instead of lowering tariffs and allowing private-sector imports of rice; which caused little reduction of domestic rice production as much of the increased rice imports ended up being used as livestock feed.
3.4 Development economics and the roles of the market, state, and communities By the late 1990s, Hayami again drew on his growing economic thought to undertake a more comprehensive analysis of the process of economic development and identify the appropriate strategies, policy, and institutional frameworks for the poor countries to catch up with the level of development of wealthier nations. The nature and processes of industrial development were integrated throughout his writing. Again, his conclusion is that technological change, specifically technology-borrowing would be the driving force of economic development. And the state is able to ensure that the incentive structure reflects the scarcity value of inputs and outputs, public goods are provided, and other market and community failures are addressed through appropriate policy and institutional frameworks. This is critical in ensuring markets and communities function efficiently in harnessing modern technologies borrowed from more advanced countries. The set of papers presented in this workshop to honor Yujiro Hayami is our response to his challenge of designing the combination of markets, state, and communities to alleviate poverty in developing economies, maintain
9780230_274587_04_cha03.indd 17
6/23/2010 12:16:04 PM
18 Cristina C. David
economic vitality and social harmony in developed countries, and preserve the environment globally. From the intensive use of official statistics, to field surveys and recordkeeping of farm-households’ activities, his “random walks” across villages, and “pedestrian” interviews of traders, Hayami stayed focused on the use of economic rationality principles in explaining the processes of economic, technical, and institutional change. Yet he was strongly convinced from the start that social norms and other cultural factors are important in understanding why people from different backgrounds express their economic rationality in different forms, so that there is some degree of path dependency, as well as multiple equilibrium paths to economic development. By the late 1990s, his exploration of those phenomena, especially from the village studies, led him to highlight the important role of communities, and not just markets and the state, in the process of economic development. However, we do hope that in the years ahead he will be able to push further the frontiers of understanding cultural endowments, their effects and determinants, to improve the design of policies and programs that will enhance the functioning of markets, the state, and communities. Many individuals who had the opportunity to work with Yujiro Hayami in his fascinating and highly productive journey: his former students and now colleagues in Japan – Otsuka, Kikuchi, Kawagoe, Godo, Honma; international colleagues – Ruttan, Barker, Herdt, Anderson; and Filipinos – the IRRI research staff, Adriano, Quisumbing, Crisostomo-David, among others too many to mention, all join us in honoring him today. Surely, Yujiro Hayami is not just “anyone” of Japan, but one of the great economists concerned with development in this generation.
9780230_274587_04_cha03.indd 18
6/23/2010 12:16:04 PM
Part II Role of State
9780230_274587_05_cha04.indd 19
6/23/2010 5:04:53 PM
9780230_274587_05_cha04.indd 20
6/23/2010 5:04:53 PM
4 The Role of the Government in Facilitating TFP Growth during Japan’s Rapid-growth Era*1 Shuhei Aoki, Julen Esteban-Pretel, Tetsuji Okazaki, and Yasuyuki Sawada
4.1
Introduction
It is widely known that Japan experienced rapid economic growth in the late 1950s and 1960s, when per capita gross domestic product grew at a remarkable rate of over 10 percent (Figure 4.1). There are numerous books and academic papers written on the reasons behind this success of the postwar Japanese economy. Some studies rely on descriptive macro-level statistics (Inada et al., 1993; Kosai and Kaminski, 1986; Minami, 1994; Nakamura, 1995; Ohkawa and Rosovsky, 1973) and others employ the growth-accounting framework to decompose the high-growth rate into different factors (Denison and Chung, 1976; Hayami and Ogasawara, 1999, 2002; Yasuba, 2002; Young, 1995). More recent works on the Japanese economy rely on modern calibration techniques to replicate the postwar rapid growth (Braun et al., 2006; Braun et al., 2009; Chen et al., 2006; Otsu, 2007; Parente and Prescott, 2004). There is an emerging consensus, both in the growth-accounting studies, such as Young (1995) and Hayami and Ogasawara (1999), and in the calibration works, such as Otsu (2007) and Esteban-Pretel and Sawada (2009), that Japan’s postwar rapid growth was driven by a high-growth rate of total factor productivity (TFP). These studies, while careful in their analysis and accurate in their results, assume that the evolution of TFP is exogenous to their models, and, thus, do not address questions related to the sources of the high rates of TFP growth. However, uncovering the determinants of TFP growth is important, especially when studying government policies, * This paper has been prepared for the Festschrift Conference for Yujiro Hayam on State, Community, and Market in Development, February 27–28, 2009. We would like to thank Tina David, Yoshihisa Godo, Jikun Huang, Kaliappa Kalirajan, Keijiro Otsuka, Gustav Ranis, Scott Rozelle, Tetsushi Sonobe, John Strauss, and other participants for helpful comments.
21
9780230_274587_05_cha04.indd 21
6/23/2010 5:04:53 PM
22
Shuhei Aoki et al. 11 10 9 8 7
1870 1875 1880 1885 1890 1895 1900 1905 1910 1915 1920 1925 1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005
6
Figure 4.1 Changes in Japan’s gross domestic product per capita in logarithm of real GDP expressed in 1990 US dollars Source: Maddison’s Database
.
since some of these policies could have affected the economy by influencing TFP. This chapter studies the main forces behind the high growth of Japanese TFP in the postwar period, and the role that the government played in the evolution of Japanese technology. We argue that it is necessary to analyze the evolution of TFP for the agricultural and nonagricultural sectors separately. This is due to the changing level of economic development of the Japanese economy in the postwar period, and the initial importance of the agricultural sector, both in share of employment and total output, right after the war. We show that, for the nonagricultural sector, technology improvement in the early years following World War Two started with the import and licensing of specific foreign technologies. With time, the adoption of foreign technologies was gradually replaced by the development of Japan’s own domestic technologies. We argue that the government played a role during both stages. In the foreign technology adoption phase, the main impact of the government was through heavily regulating the number and type of technologies which companies were allowed to license and import from foreign firms. As for the phase of the development of their own new technologies, the stage in which the Japanese economy is still immersed, the government has not contributed as much as governments in other developed countries in terms of expenditures. However, its main influence has been through the establishment of R and D consortia, but not by affecting the R and D activities with distortionary subsidy systems, and finally by not strengthening patent laws. In particular, it is argued that the positive
9780230_274587_05_cha04.indd 22
6/23/2010 5:04:53 PM
Role of the Government in Facilitating TFP Growth
23
outcomes on the R and D consortia can be attributed to the community mechanism of enforcing agreement between government and industrial leaders in the sense of Hayami and Godo (2005). The remainder of this chapter is organized as follows. In Section 4.2 we decompose aggregate Japanese TFP into agricultural and nonagricultural sectors, and explain their main characteristics. In Section 4.3, we present evidence of foreign technology adoption and innovation in postwar Japan in the nonagricultural sector. Section 4.4 discusses the determinants of agricultural TFP. In Section 4.5 we focus on the domestic diffusion of technology. Finally, Section 4.6 summarizes and concludes.
4.2 Evidence on overall Japanese TFP and its sectoral decomposition It has been well established in the economic literature that the innovation and adoption of new technologies are the two key driving forces of the TFP growth. Models of endogenous technological innovation can be broadly split into three categories: first, new variety models (Romer, 1990); second, quality ladder models (Aghion and Howitt, 1992; Grossman and Helpman, 1991); and, third, the induced technological innovation model (Hayami and Ruttan, 1971 and 1985). Moreover, technologies can be adopted rather than innovated through three broad channels: First, imported intermediate inputs directly improve technology level (Coe and Helpman, 1995; Eaton and Kortum, 2002). Second, foreign direct investment (FDI) facilitates technology transfers (Markusen, 2002). Finally and more generally, sufficient stock of domestic knowledge and technology is a necessary condition for a country to be able to import, adapt, and absorb new technologies from overseas (Barro and Sala-i-Martin, 1997; Benhabib and Spiegel, 2005). The following sections review how new technologies were created and diffused, and how the government facilitated these processes in the postwar Japanese economy. When the Second World War ended in 1945, the Japanese economy found itself far behind the technology frontier of the world. In 1964, the Agency of Industrial Science and Technology, which was in charge of industrial R and D under the Ministry of International Trade and Industry (MITI), looked back at the 1940s and 1950s and wrote: “The technology gap between Japan and the advanced countries in the prewar period was preserved, and it further expanded due to the vacuum of technology adoption during the war” (Agency of Industrial Science and Technology, 1964). Indeed, in 1952 the productivity gap between Japan and the US was substantial. According to the estimate by Christensen et al. (1995), Japanese TFP was as low as 43 percent of that of the US. However, the productivity of the Japanese economy increased rapidly after that, and it became 80 percent of that of the US in the early 1970s, when the high growth came to an end (Figure 4.2).
9780230_274587_05_cha04.indd 23
6/23/2010 5:04:54 PM
24 Shuhei Aoki et al. 0.8
0.6
0.4
0.2
1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973
0
TFP Figure 4.2
Output per capita
Changes in productivity of Japanese economy relative to the US
Source: Christensen et al. (1995), pp. 316–20.
In the following sections, we divide the economy into the agricultural and nonagricultural sectors. The reason behind this division is twofold. First, right after World War Two, the share of output and employment in the agricultural sector was more than 50 percent. As is shown in Esteban-Pretel and Sawada (2009), accounting for the structural change that took place in Japan in the postwar period is essential to understanding the rapid growth. During a structural change process, TFP growth, or more precisely, labor productivity growth in the agricultural sector, is an important determinant of economic growth. In models with non-homothetic preferences (e.g., Eswaran and Kotwal, 1993; Esteban-Pretel and Sawada, 2009; Gollin et al., 2007), households first consume the subsistence level of food that is essential for their survival, and then consume other goods such as manufacturing goods or services. In such models, if agricultural TFP level is low, even if the TFP level in the nonagricultural sector is high, the economy needs to allocate a large fraction of resources to unproductive food production, which reduces economic growth. Second, while the adoption of foreign technologies in the manufacturing sector is widespread and has been documented for many developing economies, the international diffusion of agricultural technologies is often very difficult due to differences in climatic conditions. As we discuss later, given these difficulties, the Japanese agricultural sector was forced to innovate independently, while in the other economic sectors the adoption of foreign technologies was more common, at least right after World War Two.
9780230_274587_05_cha04.indd 24
6/23/2010 5:04:54 PM
Role of the Government in Facilitating TFP Growth
25
300 250 200 150 100 50
1956 1958 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000
0
Non-agriculture
Agriculture
Figure 4.3 Changes in sectoral total factor productivity indexes during the postwar period in Japan Note: Labor input is total hours. Source: Esteban-Pretel and Sawada (2009) dataset.
Figure 4.3 shows the evolution of TFP in the agricultural and nonagricultural sectors in postwar Japan. We can observe that both TFP series increase significantly until the first oil crisis, although the growth rate of nonagricultural TFP was significantly higher than that of agriculture. In order to further understand the relative contribution of the sector-specific TFP to the overall TFP level, we decompose aggregate TFP, A, into four components: Agricultural TFP, Aa, nonagricultural TFP, Am, and the contribution of the reallocation of capital and labor. We follow Syrquin (1984) and Basu and Fernald (2002) and decompose aggregate TFP as follows: A A r ⫺ r K i wi ⫺ w L i A ⫽ sVa a ⫹ sVm m ⫹ ∑ sVi sKi i ⫹ ∑ sVi sLi , A Aa Am i ∈{a, m} ri Ki i ∈{a, m} wi Li where K is capital; L is labor; r and w are capital and labor returns, respectively; sVi, sKi, and sLi are the nominal value-added share, capital share, and labor share of sector i, respectively. Note that variables without subscripts denote aggregate levels, and those with subscripts are those in the agriculture (a) or nonagriculture (m) sectors. The final two terms on the right-hand side of the previous equation are the “growth bonus” effect arising from reallocating capital and labor from a low to a high marginal productivity sector. When capital or labor are reallocated to a sector with higher marginal product, the third and fourth terms are positive. We refer to these last two terms as the capital and labor reallocation effects.
9780230_274587_05_cha04.indd 25
6/23/2010 5:04:54 PM
26 Shuhei Aoki et al.
Table 4.1 TFP decomposition
1956–73 1973–83 1983–91 1991–2000
Aggregate TFP (%)
Agr. TFP (%)
Non-agr. TFP (%)
4.78 0.50 2.26 0.17
0.11 0.00 –0.01 –0.01
3.96 0.25 2.13 0.09
Capital reallocation (%)
Labor reallocation (%)
0.05 0.04 0.01 0.01
0.66 0.21 0.13 0.08
Note: Labor input is total hours. Source: Esteban-Pretel and Sawada (2009) dataset.
Table 4.1 shows the decomposition of aggregate TFP for various subperiods starting in 1956. We can see that the main contributors to aggregate TFP growth are nonagricultural TFP and the reallocation of labor. Both terms are high during the rapid-growth era, from 1956 to 1973. The importance of the labor reallocation effect is consistent with the findings of Hayashi and Prescott (2008) and Esteban-Pretel and Sawada (2009), which show that the elimination of migration barriers from rural to urban areas can be seen as one of the important determinants for Japan’s postwar economic miracle.1 The contribution of agricultural TFP is low, which is partly due to the declining share of agricultural production in total output over the period of study.
4.3 Foreign technology adoption and innovation in the nonagricultural sector We now move on to study the way in which the remarkable improvement of nonagricultural technology occurred in Japan. Initially Japan relied on imported foreign technologies, and later started the development of its own technologies. This sequence, which is also observed in other countries that are not technology leaders, can be clearly seen in Figure 4.4. The ratio of net technology imports to GNP, which was initially high, has declined gradually over time. Simultaneous with this decline, Japan experienced a rise in the ratio of R and D expenditures to GNP. This indicates that over time, Japan substituted the import and adoption of foreign technologies for the development of their own. We now study these two processes in more detail. Foreign technology adoption As can be seen in Figure 4.2, the technology gap between Japan and the US was greatly reduced during the 1950s and 1960s. A major driving force of this rapid technological progress in the decades following World War Two was the adoption of foreign technologies. Figure 4.5 shows the trend of technology adoptions measured by the number of “Type A” technological
9780230_274587_05_cha04.indd 26
6/23/2010 5:04:55 PM
0.20
3.0
0.15
2.3
0.10
1.5
0.05
0.8
27
R&D/GNP (%)
Net technology import/GNP (%)
Role of the Government in Facilitating TFP Growth
0
Net technology import/GNP
1988
1986
1984
1982
1980
1978
1976
1974
1972
1970
1968
1966
1964
1962
1960
1958
1956
0
R&D/GNP
Figure 4.4 Changes in net technology import and R and D expenditures relative to GNP in Japan Note: The left-hand vertical axis is for “net technology import/GNP,” while the right-hand is for “R&D/GNP.” Source: Science and Technology Agency (1972, 1974, and 1991) and SNA.
2000
1500
1000
500
Figure 4.5
1978
1976
1974
1972
1970
1968
1966
1964
1962
1960
1958
1956
1954
1952
1950
0
Changes in number of Type A technological assistance contracts
Source: Science and Technology Agency (1980).
9780230_274587_05_cha04.indd 27
6/23/2010 5:04:55 PM
28
Shuhei Aoki et al.
Table 4.2 Number and composition of licenses by industry 1950–4 1955–9 1960–4 1965–9 1970–4 1974–9 Number Total Chemical products Petrochemical plant engineering Petroleum and coal products Iron, steel, and nonferrous metals Fabricated metal products General machinery Transportation equipment Precision instruments and machinery Electric machinery Food and tobacco Textile products Ceramic, stone, and clay products Plastic products Others Percentage Total Chemical products Petrochemical plant engineering Petroleum and coal products Iron, steel, and nonferrous metals Fabricated metal products General machinery Transportation equipment Precision instruments and machinery Electric machinery Food and tobacco Textile products Ceramic, stone, and clay products Plastic products Others
454 82 0
575 116 19
2039 280 38
3926 678 73
8295 1048 82
7846 808 92
15
17
24
73
217
84
38
56
96
181
201
147
6 98 42 0
20 120 34 8
74 589 94 48
124 1021 209 125
225 1973 438 261
217 1759 391 304
108 0 24 10
109 1 33 8
469 8 83 45
664 58 172 81
1251 260 794 206
1451 151 1060 134
1 30
5 29
94 97
209 258
403 936
187 1061
100.0 18.1 0.0
100.0 20.2 3.3
100.0 13.7 1.9
100.0 17.3 1.9
100.0 12.6 1.0
100.0 10.3 1.2
3.3
3.0
1.2
1.9
2.6
1.1
8.4
9.7
4.7
4.6
2.4
1.9
1.3 21.6 9.3 0.0
3.5 20.9 5.9 1.4
3.6 28.9 4.6 2.4
3.2 26.0 5.3 3.2
2.7 23.8 5.3 3.1
2.8 22.4 5.0 3.9
23.8 0.0 5.3 2.2
19.0 0.2 5.7 1.4
23.0 0.4 4.1 2.2
16.9 1.5 4.4 2.1
15.1 3.1 9.6 2.5
18.5 1.9 13.5 1.7
0.2 6.6
0.9 5.0
4.6 4.8
5.3 6.6
4.9 11.3
2.4 13.5
Source: Science and Technology Agency (1980).
9780230_274587_05_cha04.indd 28
6/23/2010 5:04:56 PM
Role of the Government in Facilitating TFP Growth
29
assistance contracts, that is, contracts whose terms were longer than one year, from 1950 to 1979. While the number of foreign technology adoptions per year was around 100 until 1959, it increased to around 2000 by the late 1960s. Breaking down these numbers by industry, we find that they were concentrated in a few industries, that is, chemical products, general machinery, and electrical machinery, from the 1950s to the 1970s (Table 4.2). The license share of these three industries over the total number of licenses was on average about 60 percent in the 1950s and 1960s. This large contribution of imported technologies to postwar Japanese economic development is widely accepted in the preceding studies. For example, Peck and Tamura (1976) characterize postwar Japanese technological development as one with “high returns from importing technology.” Goto and Odagiri (1996) point out that Japan efficiently utilized the channels of technology transfer such as import of machinery and equipment, and purchase of technology and technological services. Figure 4.6 shows a scatter diagram showing the ratio of net payment for technology import to GDP and productivity growth by industry. The horizontal axis denotes the number of foreign technological adoptions from 1955 to 1969, normalized by the average real value added in 1955 and 1970, while the vertical axis denotes the productivity growth from 1955 to 1970 (TFP in panel [a] and labor productivity in panel [b]). We can clearly observe a positive correlation between technology adoption and growth, both for TFP and labor productivity. The impact of technology adoptions on the different industries, suggested in Figure 4.6, has been well documented in the literature. Previous work on the industrial development of postwar Japan is full of anecdotes about the positive effects of technology imports (e.g., Agency of Industrial
200
(a) TFP growth (%)
(b) Labor Productivity 300
Electrical Machinery p=0.702 Transportation equipment
150
Machinery
p=0.702 Chemicals
Chemicals
Transportation equipment
225
Machinery
100
Fabricated metal Textile mill Primary metal
50
150
Others
75
0
Electric Machinery
Fabricated metal Primary metal Textile mill Others Food
Food
−50
0 0
0.75
1.5
2.25
3
0
0.75
1.5
2.25
3
Figure 4.6 Ratio of net payment for technology import to GDP and productivity growth by industry Sources: Economic and Social Research Institute of Cabinet Office (2001) and Economic Research Institute of Economic Planning Agency (1998).
9780230_274587_05_cha04.indd 29
6/23/2010 5:04:56 PM
30 Shuhei Aoki et al.
Science and Technology, 1964; Kohama, 2001; Society for Industrial Studies, 1995; Yonekawa et al., 1990–1). In particular, the Committee for Foreign Technology Survey (1961) comprehensively describes technology imports by individual industries and their impacts in the early stage of postwar growth. Let us briefly look at the cases of electric machinery, automobiles, and iron and steel, which became the major export industries of Japan. Electric machinery is usually subdivided into “heavy electric machinery” and “light electric machinery.” The former is composed of investment goods for electricity companies, while the latter encompasses consumption goods for households. With respect to heavy electric machinery, three of the major companies, Toshiba, Mitsubishi Electric, and Fuji Electric, entered into comprehensive technology adoption contracts with GE, Westinghouse, and Siemens, respectively, in the early 1950s. One of the most remarkable results of these contracts was that the Japanese electric machinery companies became able to produce thermal generators of large capacity. In the early 1950s, they started to shift from hydroelectric generation to thermal generation. It was for that purpose that they imported advanced thermal generators larger than 200,000 KW each, which they could not produce. Owing to foreign technology adoptions, they acquired the ability to produce these advanced machines. Concerning light electric machinery, foreign technology adoptions created a new industry, that is, the television set industry. TV broadcasting started in Japan in 1953. Since R and D for producing TV sets was suspended during the war, and the start of TV broadcasting occurred at a later time than in advanced countries, Japanese technology for TV production was substantially behind. Many basic patents for TV sets were owned by foreign companies, in particular RCA, Westinghouse, and EMI. Hence, more than 30 Japanese electric machinery companies entered into technology assistance contracts with these foreign companies in the early 1950s. While the cumulative number of TV sets produced in Japan until 1954 was 75,000, in 1959 alone 2.9 million TV sets were produced, and out of those, 27,000 sets were exported. The occupation authority (SCAP) allowed the production of automobiles in 1949, although at this time automobiles were just remodeled small trucks whose performance was 10 to 20 years behind the world standard. In the early 1950s, three major automobile companies, Nissan, Hino, and Isuzu, entered into technology adoption contracts with Austin Motors, Renault Corporation, and Roots Motors, respectively. According to the contracts, information on design, specifications, materials, processing, and other aspects, was provided to Japanese companies, which enabled them to produce world-class automobiles. Furthermore, accompanying these technologies, the mass-production system based on transfer-machines was introduced to the Japanese automobile industry.
9780230_274587_05_cha04.indd 30
6/23/2010 5:04:56 PM
Role of the Government in Facilitating TFP Growth
31
The iron and steel industry developed fairly well in Japan, but the technology for producing strip metal, whose demand was increasing as the durable consumption goods industry developed, lagged behind. In order to catch up with the advanced countries, MITI and iron and steel companies drew up the “First Iron and Steel Rationalization Plan” in 1951, the focus of which was the construction of strip mills. Based on the plan, Fuji Iron Works and Yawata Iron Works, the two largest iron and steel companies in Japan, entered into technological assistance contracts with Armco Steel to obtain know-how on the design, construction, and operation of strip mill factories. A similar contract was entered into between Kawasaki Iron Works and Republic Steel in 1958. The “First Iron and Steel Rationalization Plan” was followed by the “Second Iron and Steel Rationalization Plan” in 1956. The focus of the second plan was the introduction of the basic oxygen furnace (BOF), which was a substitute for the open hearth. In 1956, Nippon Kokan Corporation (NKK) acquired a general license in Japan with Alpine Co., and six other major Japanese steel companies acquired the sublicense of BOF from NKK in the late 1950s. There are few studies that quantitatively identify the effect of foreign technology adoption in 1950s and 1960s Japan. Kiyota and Okazaki (2005) examined the effect of technology adoption on the firm’s performance using firm-level data from 1957 to 1970. They matched the complete list of individual technology adoptions by the firms listed on the Tokyo Stock Exchange in that period with the financial data from the JDB database. They found that in the 1950s, technology adoption had a positive impact on labor productivity, value-added, the capital-labor ratio, and the R and D-sales ratio, although not on TFP. In a different study, Nakamura and Ohashi (2008) reveal that the adoption of BOF raised the annual TFP growth rate of the steel industry from 7 percent (counterfactual) to 17 percent (actual in 1957–68). At the same time, they find that just after a plant introduced BOF, its TFP dropped 9 percent and it took two years before TFP reached the level achieved by the old open-hearth technology. This is likely to be due to the learning effect, which seems to explain the result of Kiyota and Okazaki (2005) that technology adoption did not have a positive impact on TFP. Policy framework for technology adoption The Japanese government played an important role in the import of the technologies described above, by imposing a strict framework for such adoptions.2 This policy framework, which started in 1950 with the enactment of the Foreign Investment Law and had mostly disappeared by 1968, was intended for the selective adoption of what were thought of as key technologies. The 1950 law required government approval of all “Type A” technological assistance contracts. By this approval system, the government aimed at selectively importing technologies that would contribute to Japan’s balance of payments and the development of “important” industries.
9780230_274587_05_cha04.indd 31
6/23/2010 5:04:56 PM
32 Shuhei Aoki et al.
Meanwhile, the government guaranteed and protected the investments and contracts that were approved, and thereby intended to promote the adoption of desirable technology. In the late 1950s, requests for relaxing regulations on technology adoption were made by domestic industries as well as those from foreign countries. The domestic industries were keen to adopt more foreign technology to cope with the trade liberalization expected in the near future. Responding to these requests, the government simplified the examination procedure of technology adoption in 1959, which was called the “new method for capital import.” This measure was meant to be tentative until a more substantial deregulation, which was conducted in 1961. After the deregulation in 1961, foreign technology adoptions were approved automatically, except for cases which were thought to be detrimental to the Japanese economy, while the government continued to examine each application of technology adoption. In 1968, the government finally ceased the individual examination process, except for certain cases, such as aircraft, weapons, gun powder, and nuclear power. In other words, government intervention in technology adoption was basically abolished. We can observe in Figure 4.5 several qualitative jumps in the range of imported technologies. These jumps divide the sample period into the four phases generated by the deregulations of 1959, 1961, and 1968. Figure 4.5 suggests that the government regulation indeed affected technology adoptions in Japan. Studying the influence of such regulation of the number of imported technologies, Kiyota and Okazaki (2005) use firm-level data and find a negative impact of the regulation on the amount of technology adoptions. They also find that the government regulation affected the composition of the adopted technologies. Peck and Tamura (1976) point out that MITI gave priority to adopting technologies for intermediate and capital goods for heavy industries in the 1950s. The relatively high percentage of chemical products and “iron, steel and non-ferrous metals” in the 1950s seems to reflect the priorities of MITI. The advantages and disadvantages of government intervention in foreign technology adoptions in Japan have been controversial. Johnson (1982) and Lynn (1982 and 1998) evaluate the contribution of MITI’s control over technology adoptions, arguing that thanks to MITI’s intervention, Japanese industries were able to obtain cutting-edge technologies at lower prices. Their rationale is that MITI’s intervention and controlled competition among Japanese firms helped reduce the concerns of foreign technology suppliers, who otherwise may not have dealt with Japanese firms that did not have international reputations. Nakamura and Ohashi (2008) argue that, given the high estimated learning costs of new technologies in the steel industry, MITI’s intervention encouraged Japanese steel companies to adopt the more productive BOF technology through lowering royalties.
9780230_274587_05_cha04.indd 32
6/23/2010 5:04:56 PM
Role of the Government in Facilitating TFP Growth
33
In contrast, others have stressed that MITI’s intervention hindered or delayed competent Japanese firms from accessing foreign technologies. A case cited often is that of Sony in the early 1950s (Goto and Odagiri, 1996; Okimoto, 1989; Peck and Tamura, 1976). Sony had to wait for two years until the adoption of the transistor technology from Western Electric was approved by MITI. This delay was due to MITI’s skepticism about the abilities of Sony, a small start-up firm at that time. This anecdote is consistent with MITI’s general policy of screening applications for technology adoption. Kiyota and Okazaki (2005) reduced the number of technology adoptions by a firm based on its attributes including the cumulative number of technologies it had adopted, and found that its coefficient was significantly positive in the 1950s, whereas it was around zero after deregulation in the 1960s. This implies that MITI preferred to approve technology adoption by experienced firms, and this had an unfavorable effect on competent start-up firms like Sony in the 1950s. Innovation of new technologies As the Japanese economy caught up with the developed countries, it gradually shifted from the foreign technology adoption process to the innovation process, where Japanese firms engaged in R and D investments more intensively. The R and D expenditure to GNP ratio in Japan increased from about 50 percent of the US level in 1961 to about the same level in 1989.3 As in the case of the US, around 40 percent of R and D expenditure in Japan is used for basic and applied research, and the rest is used for the development of new products.4 Similar to the US, firms’ R and D activities also contributed to the development of science. Two Japanese Nobel laureates in the natural sciences were corporate researchers when they achieved their scientific breakthroughs. Table 4.3 Decomposition of R and D expenditure in terms of share in total R and D expenditure in selected countries
Government
Industry
University
Private laboratory
Provider User Provider User Provider User Provider (%) (%) (%) (%) (%) (%) (%) Japan (1988) US (1988) West Germany (1987) France (1983) UK (1987)
User (%)
18.4 48.0
9.3 11.5
76.3 47.9
73.9 71.8
4.5 2.8
12.6 13.8
0.7 1.4
4.3 2.9
36.6 53.8 38.7
3.4 26.4 15.1
62.3 42.0 49.7
73.1 56.8 67.0
– 0.2 0.6
12.9 15.9 14.2
– 0.4 1.9
10.6 0.9 3.7
Source: Westney (1994) (Originally taken from Science and Technology Agency, 1990).
9780230_274587_05_cha04.indd 33
6/23/2010 5:04:56 PM
34
Shuhei Aoki et al.
One characteristic of R and D in Japan is that the government’s share of R and D expenditure is the lowest among the developed countries. Table 4.3 shows that while the US government provided around 50 percent of R and D expenditure in 1988, the Japanese government only provided around 20 percent. Japanese firms accounted for almost all the remaining expenditure, 74 percent. At the user level, however, Japan’s R and D expenditure is similar to those of other developed countries. Policies on technology innovation One implication of the endogenous growth theory is that R and D investment in laissez-faire economies is lower than the social optimum. This is due to the existence of externalities and imperfect competition arising naturally from increasing-returns-to-scale technologies. Such externalities justify the intervention of the government through policies such as the provision of financial incentives via subsidies and tax benefits to R and D investment, or the strengthening of patent rights. However, evidence seems to suggest that the Japanese government did not pursue the pro-incentive policies to correct market failures, as is suggested by the endogenous growth theory. As we see in Table 4.3, the Japanese government’s subsidies for R and D investments were much smaller than those in other developed countries. Moreover, according to Okimoto (1989), specific tax incentives for R and D investment were also small compared with the US. The avoidance of such policies, which can easily generate an excuse for pork-barrel spending, might be one of the reasons why R and D policies in Japan were relatively free from distortions arising from government failures. According to Okimoto (1989), MITI has traditionally refused to protect depressed industries, while the declining sectors tend to be the biggest beneficiaries of tax exceptions in the US. These relatively nondistortionary tax policies of the Japanese government towards the R and D sector may have been beneficial for competition and ultimately innovation, making the R and D sector one of the highest-growth industries in Japan. On the other hand, Beason and Weinstein (1996) argue that the Japanese government, including MITI, allocated a disproportionate amount of resources to lowgrowth industries.5 They also find that the financial support by the government did not positively affect TFP. MITI also avoided other types of policy distortions. Okimoto (1989) argues that after World War Two, Japan discarded superpower aspirations. As a result, the Japanese government did not promote defense-related industries, which meant that, unlike other countries, the government did not spend substantial quantities on R and D investment in the defense industry. He also argues that the Japanese government de-emphasized the importance of national prestige. This implied that, unlike France and other European countries, MITI has not followed a strategy of cultivating one or two “national
9780230_274587_05_cha04.indd 34
6/23/2010 5:04:56 PM
Role of the Government in Facilitating TFP Growth
35
champions.” For example, Japan has had nothing like the US’s Strategic Defense Initiative, and no supersonic jet or Airbus-type project. The Japanese patent system had been traditionally beneficial to the user. While a user-favorable patent system helps and promotes the import of technologies in the adoption process, it is harmful to the innovation process. This is one of the reasons why patent rights in Japan were strengthened in 1988 through a substantial patent reform. However, the benefits of such a reform are debatable, since Sakakibara and Branstetter (2001) show that the 1988 Japanese patent reform does not seem to have increased either R and D expenditures or product innovation. The endogenous growth theory emphasizes the importance of knowledge spillover in the process of technological change. Since this kind of spillover is a type of externality, policies aimed at internalizing the benefits of such knowledge spillover can be justified. One such policy is the establishment of R and D consortia, where potentially competiting firms cooperate on the R and D of targeted technologies, and exchange and share their knowledge.6 Probably, the most famous R and D consortium is the project on very largescale integrated circuits. However, cooperation among the participants, which is indispensable for the success of an R and D consortium, is not easy to achieve. For instance, if participant firms compete in the product market, they are less likely to cooperate in R and D, because their efforts for innovation in the consortium help the competitors. Similarly, if differences in technology level among firms are large, leading firms are not willing to participate in R and D consortia, because they will gain little by participating. What, then, was the performance of R and D consortia in Japan? Findings in a series of papers by Branstetter and Sakakibara suggest that the Japanese R and D consortia have had positive effects on the innovation of participating firms. They find that even after controlling for fixed effects, a firm’s participation in R and D consortia increased the firm’s R and D expenditure and patent production (Branstetter and Sakakibara, 1998, 2002). They also note that the design of an R and D consortium was more important to its success than the level of resources expended on it (Branstetter and Sakakibara, 2002). These results indicate that although government financial support for R and D consortia was modest in Japan (Sakakibara, 1997), the role of such consortia may have been substantial. Okimoto (1989) attributes the success of R and D consortia to the relationship of trust between MITI and the private firms. He also notes MITI’s ability to consult with them. In this sense, the R and D consortia utilize what Hayami (2006) refers to as “the community enforcement mechanism in modern business societies in Japan.” On the other hand, Okimoto and Sakakibara point out that the success of the Japanese R and D consortia was owed to the special circumstances of the catch-up period. During that period, the Japanese firms evenly fell behind
9780230_274587_05_cha04.indd 35
6/23/2010 5:04:56 PM
36 Shuhei Aoki et al.
their Western counterparts in terms of the technology level and felt the need to catch up in order to prepare for the competition with Western firms after direct foreign trade liberalization (Okimoto, 1989; Sakakibara, 1997). In addition, a firm’s participation in R and D consortia was a sign of the firm’s ability to outsiders, such as buyer industries and banks (Sakakibara, 1997). However, the circumstances changed after the catch-up period ended. Differences in the technology capabilities of Japanese firms have become larger. The signaling effect was likely to have decreased too, because the government did not have superior information to the private firms. Another policy that promotes the development and sharing of knowledge is the promotion of institutions such as research institutes and universities. According to Okimoto (1989), the laboratories of Nippon Telegraph and Telephone Public Corporation, and its successor Nippon Telegraph and Telephone Corporation (NTT), have played an important role in R and D activities in Japan, as AT and T has done in the US. On the other hand, while the university share of R and D expenditure in Japan is similar to that of other developed countries (Table 4.3), Okimoto and Nishi (1994) report that the relations between firms and universities have been traditionally inactive, which has limited the knowledge-sharing between these two types of institution. Furthermore, Aldrich and Sasaki (1995) report that university-based R and D is not allowed in Japanese R and D consortia. All this suggests that university R and D in Japan has been conducted independently from that of industry.
4.4 Foreign technology adoption and innovation in the agricultural sector Unlike in the nonagricultural sector, foreign technology adoption is not easy in agriculture, primarily due to the fact that agriculture is strongly constrained by environmental conditions (Hayami and Godo, 2005). The induced innovation theory by Hicks (1932) and Hayami and Ruttan (1971 and 1985) formalizes the idea that technological innovations are directed toward technologies that use smaller amounts of relatively scarce factors. However, such innovations are not always possible with the effort of private farm producers because of the public-good nature of many biological and chemical technologies, such as new rice seeds and production practices. Hence, to facilitate TFP growth in agriculture, it is essential to develop the appropriate public-sector research, and extension systems. For this reason, the government must potentially play a very important role in the agricultural sector (Rustichini and Schmitz, 1991). In Japan, modern technological progress in the agricultural sector has taken place in two phases, the prewar and postwar eras. As Hayami (1975) notes, the rapid TFP growth in the prewar period was supported by a backlog of technological potential accumulated before the Meiji era, when superior
9780230_274587_05_cha04.indd 36
6/23/2010 5:04:57 PM
Role of the Government in Facilitating TFP Growth
37
methods and advanced knowledge were developed by the wealthy veteran farmers (“Rounou”). After the Meiji Restoration, these veteran farm leaders started identifying and disseminating their knowledge across regions. The government helped the exploitation and diffusion of the backlog of indigenous technologies by sponsoring “seed-exchange meetings” and “agricultural discussion meetings,” and by promoting scientific agricultural research and extension activities (Minami, 1994). These activities were complemented by irrigation investments by the government, and their maintenance performed by the neighboring communities in mobilizing the community mechanism of collective work is emphasized by Hayami (1975). Agricultural TFP improvement during the early postwar era shown in Table 4.1 was the consequence of many technological advancements that had been accumulated during the prewar period (Hayami and Ruttan, 1985). TFP growth in the postwar years played an essential role in growth of agricultural output (Hayami, 1975; Kuroda, 1995, 1997). It is argued that the postwar TFP growth in the agricultural sector has come from two main sources: the mechanization of agriculture, and the introduction of high-yielding varieties.7 As to the first source, mini-tractorization based on the adoption of small-scale machinery became an important factor in agricultural growth during the rapid-growth era. In this process, the research conducted during the war for nonagricultural purposes formed a backlog used in the advancement of agricultural techniques (Hayami, 1975). Lau and Yotopoulos (1989) show that there is a positive association between scale economies and tractorization, and Hayami and Kawagoe (1989) find that increasing returns to agricultural production emerged in Japan as agriculture developed and wages increased during the mid-1960s. These increasing returns and rises in wages occurred because of the substitution of large-scale machinery, such as riding tractors, for small hand tractors. As to the second source, the Japanese government played an important role in this process of agricultural technological improvement. In particular, the R and E efforts of the Ministry of Agriculture, Forestry and Fishery (MAFF) resulted in the development of new rice varieties, from both the pure-line and crossbred “Norin” selections.8 These R and E activities were crucial in the early postwar period because they allowed for the reduction of costs in agricultural production. However, the effect of R and E on TFP growth dramatically decreased after the late 1960s, despite the fact that the level of R and E investment did not decrease. The diminished impact of R and E on TFP growth is what Kuroda (1997) identifies as the main source of the agricultural growth slowdown after 1969.
4.5
Domestic diffusion of technology
The preceding sections have focused on the role of the Japanese government in the development and dissermination of foreign technologies. In
9780230_274587_05_cha04.indd 37
6/23/2010 5:04:57 PM
38
Shuhei Aoki et al.
the theoretical model of economic growth it is commonly assumed that once a new technology is developed, or imported from overseas, it is readily available for use by any firm willing to pay its price. While this perfect and instantaneous diffusion of technology may be a reasonable simplification for these models, in reality, technologies and know-how do not immediately diffuse within a country. Indeed, using a numerical dynamic generalequilibrium model with endogenous TFP, Oshima (2009) shows that a major part of TFP growth in Japan’s postwar rapid-growth era can be explained by domestic knowledge diffusion. In this section we discuss the factors that affect the domestic diffusion of existing technologies, including the role of the Japanese government. The level of human capital is an important lubricant for the diffusion of technologies. Once a new technology is imported or invented, for it to be widely used, workers in firms need to be able to learn how to use it. The ability of the labor force to learn and implement new techniques is strongly linked to the educational level of a country. Godo in Chapter 8 compares the progress of education between Japan and the US in the period from 1890 to 1990, finding that while Japan greatly caught up with the US in terms of education in the prewar period, there was a missing link between improvements in education and the macro performance of Japan. This missing link can be clearly observed in the postwar period, when Japan rapidly caught up with the US in terms of per capita GDP, although the catching-up process in terms of schooling was much less dramatic (Figure 4.7). Godo and Hayami (2002) argue that the weak contemporaneous correlation between education and TFP growth in the postwar period is due to the fact that what 15 12 9 6 3
1888 1893 1898 1903 1908 1913 1918 1923 1928 1933 1938 1943 1948 1953 1958 1963 1968 1973 1978 1983 1988 1993
0
Japan Figure 4.7
U.S.
Comparison of average years of schooling in Japan and the US
Source: Godo and Hayami (2002).
9780230_274587_05_cha04.indd 38
6/23/2010 5:04:57 PM
Role of the Government in Facilitating TFP Growth
39
was important in the diffusion of technologies after the war was the available educational level, rather than its current improvements. In particular, Godo and Hayami (2002) state that it was the sharp increase in Japanese vocational training prior to 1940 that was critical to the diffusion of technologies in the 1950s and 1960s. The government has implemented several policies that have been important in increasing the human-capital level of Japan. It introduced mandatory elementary education in the late-nineteenth century, made tuition-free in 1900, and started to implement vocational training programs as early as 1893. It also provided certificates for technicians through exams, such as the National Trade Skill Testing and Certification System (“Ginou Kentei Seido”) of the Ministry of Health, Welfare, and Labor, introduced in 1959, which helped workers move across jobs and diffuse the existing technologies. It is widely accepted that physical infrastructure plays an important role in the process of economic development. One way it helps is by facilitating the diffusion and adoption of technological advancements. Since the classical works of development theory, such as Hirschman (1958) and RosensteinRodan (1943), development economists have considered infrastructure as an indispensable precondition of industrialization. A government’s industrialization program would be successful when physical infrastructure is shared among the firms and investors, since the coordination of various investments generates strong pecuniary externalities, which is identified as the key to successful industrialization (Ciccone and Matsuyama, 1996; Murphy et al., 1989). In facilitating the diffusion of technologies, the Japanese government played an important role by providing physical infrastructures such as the postal system, telecommunications, railways, and paved roads and highways (Kohsaka et al., 2007). For instance, the government built the crucial transportation infrastructure that was needed to implement the BOF technology explained in section 3 (Okazaki, 2001). In diffusing borrowed technologies imported from advanced foreign countries, or those developed domestically, imitation through industrial clusters may have also been important. As an example of technology spillovers in clusters, Yamamura et al. (2005) examine the evolution of the motorcycle industry in Japan from 1948 to 1964. Using individual firm data, they show that the industry’s rapid growth in the early phase is explained by the massive entry and imitation of simple technologies. On the other hand, the sustained growth in the later phases is explained by domestic innovations and subsequent imitations by other local firms, as well as the exit of inefficient firms. Sonobe and Otsuka (2006) found similar patterns in different industries in Japan and other East Asian countries. The motorcycle industry example of Yamamura et al. (2005) also highlights the role of the government as the enactor and enforcer of patent laws, since most of the industry’s growth is explained by imitation of the
9780230_274587_05_cha04.indd 39
6/23/2010 5:04:57 PM
40 Shuhei Aoki et al.
competitor’s technology. Japan’s patent system somehow allows these types of imitation patterns. It is a system that promotes easy opposition to patents before they are even issued. As pointed out by Flath (2005), this provides incentives to firms to license their new inventions on terms favorable to users, and encourages the early revelation of new technologies.
4.6
Concluding remarks
In this chapter we have shown that the growth of aggregate TFP in postwar Japan was mainly driven by that of nonagricultural TFP, although agricultural TFP also played an important role in the structural transformation of the economy. The Japanese government played a role in the technological growth of both sectors, and its contributions comprise both financial support and the provision of the necessary organizational infrastructure. The high growth of nonagricultural TFP in the postwar period was driven initially by the adoption of foreign technologies and later by domestic innovation. Most of the initial adoption of technologies was done through licensing by Japanese firms under the strict supervision of the government. The government influenced the types and quantities of technologies imported by setting a very restrictive process for the approval of the import of foreign technologies. The government’s share of R and D in Japan is smaller than that in other developed countries, but it has been highly involved in the establishment of R and D consortia. This suggests that the role of the Japanese government in the innovation phase has been primarily to avoid distorting the incentives of innovating firms through subsidies, and to enhance coordination through R and D consortia by, for example, employing the community enforcement mechanism between government and industry in the sense of Hayami (Hayami and Godo, 2005). In the agricultural sector, TFP growth was driven mostly by innovations of new technologies. The mechanization of agriculture through locally invented mini-tractors, as well as the development of high-yielding crop varieties, has been the source of growth in agricultural TFP. The government’s participation in this process was crucial since many of these technologies were developed directly by the Ministry of Agriculture, Forestry and Fishery. The Japanese government also helped nonfarm sectors by investing in the human-capital level, building much of the necessary physical infrastructure, and by establishing a patent system that is fairly favorable to imitators. In summary, the Japanese government played a part in the growth of TFP in the rapid-growth era by directing the adoption of foreign technologies, promoting the coordination of R and D activities, and setting up ways in which available technology could be diffused across the country. Yet, we should also note that the Japanese government was not immune from failure. Indeed, the government failed to foresee the importance of higher
9780230_274587_05_cha04.indd 40
6/23/2010 5:04:57 PM
Role of the Government in Facilitating TFP Growth
41
education in the post-catch-up growth era. Japan was not successful in accumulating a high level of knowledge due to insufficient investments in tertiary or postgraduate education. Therefore, intangible capital-intensive technological progress was not fully induced (see Chapter 8). This government failure is reflected in the low TFP growth of the 1990s.
References Agency of Industrial Science and Technology (1964) Gijutsu Kakushin to Nihon no Kogyo (Technological Innovations and the Japanese Industries: Fifteen Years History of the Agency of Industrial Science and Technology) (Tokyo: Nikkan Kogyo Shinbunsha). Aghion, P. and Howitt, P. (1992) “A Model of Growth Through Creative Destruction,” Econometrica, 60 (2), 323–51. Aldrich, H. E. and Sasaki, T. (1995) “R and D Consortia in the United States and Japan,” Research Policy, 24 (2), 301–16. Barro, R. J. and Sala-i-Martin, X. (1997) “Technological Diffusion, Convergence, and Growth,” Journal of Economic Growth, 2 (2), 1–26. Basu, S. and Fernald, J. G. (2002) “Aggregate Productivity and Aggregate Technology,” European Economic Review, 46 (6), 963–91. Beason, R. and Weinstein, D. E. (1996) “Growth, Economies of Scale, and Targeting in Japan (1955–1990),” Review of Economics and Statistics, 78 (2), 286–95. Benhabib, J. and Spiegel, M. M. (2005) “Human Capital and Technology Diffusion,” in Aghion and Durlauf (eds.), Handbook of Economic Growth, 1A, Chap. 13 (Amsterdam: North Holland), 935–66. Branstetter, L. and Sakakibara, M. (1998) “Japanese Research Consortia: A Microeconometric Analysis of Industrial Policy,” Journal of Industrial Economics, 46 (2), 207–33. Branstetter, L. and Sakakibara, M. (2002) “When Do Research Consortia Work Well and Why? Evidence from Japanese Panel Data,” American Economic Review, 92 (1), 143–59. Braun, A. R., Esteban-Pretel, J., Okada, T. and Sudou, N. (2006) “A Comparison of the Japanese and US Business Cycles,” Japan and the World Economy, 18 (4), 441–63. Braun, A. R., Ikeda, D. and Joines, D. H. (2009) “The Saving Rate in Japan: Why it Has Fallen and Why it Will Remain Low,” International Economic Review, 50 (1), 291–321.. . Chen, K., Imrohorog˘ lu, A. and Imrohorog˘ lu, S. (2006) “The Japanese Saving Rate,” American Economic Review, 96 (5), 1850–8. Christensen, L. R., Cummings, D. and Jorgenson, D. (1995) “Relative Productivity Levels, 1947–1973,” in D. Jorgenson (ed.), Productivity, Vol. 2 (Cambridge, MA: MIT Press). Ciccone, A. and Matsuyama, K. (1996) “Start-up Costs and Pecuniary Externalities as Barriers to Economic Development,” Journal of Development Economics, 49 (1), 33–59. Coe, D. T. and Helpman, E. (1995) “International R and D spillovers,” European Economic Review, 39 (5), 859–87. Committee for Foreign Technology Survey (1961) Gaikoku Gijutsu Donyu to Sangyo no Henbo (Foreign Technology Adoption and Transformation of Industries) (Tokyo: Industrial Science Association).
9780230_274587_05_cha04.indd 41
6/23/2010 5:04:57 PM
42
Shuhei Aoki et al.
Denison, E. and Chung, W. (1976) “Economic Growth and Its Sources,” in H. Patrick and H. Rosovsky (eds.), Asia’s Next Giant: How the Japanese Economy Works (Washington, DC: Brookings Institution), 63–151. Eaton, J. and Kortum, S. (2002) “Technology, Geography, and Trade,” Econometrica, 70 (5), 1741–79. Economic Research Institute of Economic Planning Agency (1998) Gross Capital Stock of Private Enterprises, 1955–1996 (Tokyo: Economic Planning Agency). Economic and Social Research Institute of Cabinet Office (2001) Report on National Accounts from 1955 to 1998 (Tokyo: Cabinet Office). Esteban-Pretel, J. and Sawada, Y. (2009) “On the Role of Policy Interventions in Structural Change and Economic Development: The Case of Japan’s Postwar Japan,” Research Institute of Economy, Trade and Industry (RIETI) Discussion Paper No. 09-E-001. Eswaran, M. and Kotwal, A. (1993) “A Theory of Real Wage Growth in LDCs,” Journal of Development Economics, 42 (2), 243–69. Flath, D. (2005) The Japanese Economy, 2nd edition (Oxford: Oxford University Press). Godo, Y. and Hayami, Y. (2002) “Catching Up in Education in the Economic Catch-up of Japan with the United States, 1890–1990,” Economic Development and Cultural Change, 50 (4), 961–78. Gollin, D., Parente, S. L. and Rogerson, R. (2007) “The Food Problem and the Evolution of the International Income Levels,” Journal of Monetary Economics, 54 (4), 1230–55. Goto, A. and Odagiri, H. (1996) Technology and Industrial Development in Japan: Building Capabilities by Learning, Innovation and Public Policy (Oxford: Clarendon Press). Grossman, G. M. and Helpman, E. (1991) “Quality Ladders in the Theory of Growth,” The Review of Economic Studies, 58 (1), 43–61. Hayami, Y. in association with Akino, M., Shintani, M. and Yamada, S. (1975) A Century of Agricultural Growth in Japan: Its relevance to Asian Development (Tokyo: University of Tokyo Press). Hayami, Y. (2006) “The Role of Community and Market in Economic Growth (in Japanese),” in Y. Sawada and T. Sonobe (eds.), Market and Economic Growth (Tokyo: Toyo Keizai Shinposha). Hayami, Y. and Godo, Y. (2005) Development Economics, 3rd edition (Oxford: Oxford University Press). Hayami, Y. and Kawagoe, T. (1989) “Farm Mechanization, Scale Economies and Polarization: The Japanese Experience,” Journal of Development Economics, 31 (2), 221–39. Hayami, Y. and Ogasawara, J. (1999) “Changes in the Sources of Modern Economic Growth: Japan Compared with the United States,” Journal of the Japanese and International Economies, 13 (1), 1–21. Hayami, Y. and Ogasawara, J. (2002) “Reply to Yasuba’s Rejoinder,” Journal of the Japanese and International Economies, 16 (2), 286–7. Hayami, Y. and Ruttan, V. W. (1971) Agricultural Development: An International Perspective (Baltimore: Johns Hopkins University Press). Hayami, Y. and Ruttan, V. W. (1985) Agricultural Development: An International Perspective (Baltimore: Johns Hopkins University Press). Hayashi, F. and Prescott, E. C. (2008) “The Depressing Effect of Agricultural Institutions on the Prewar Japanese Economy,” Journal of Political Economy, 116 (4), 573–632.
9780230_274587_05_cha04.indd 42
6/23/2010 5:04:57 PM
Role of the Government in Facilitating TFP Growth
43
Hicks, J. (1932) Theory of Wages (London: Macmillan). Hirschman, A. O. (1958) The Strategy of Economic Development (New Haven, CT: Yale University Press). Inada, K., Sekiguchi, S. and Shoda, Y. (1993) The Mechanism of Economic Development: Growth in the Japanese and East Asian Economies (Clarendon Press). Johnson, C. (1982) MITI and the Japanese Miracle: The Growth of Industrial Policy, 1925– 1975 (Stanford, CA: Stanford University Press). Kiyota, K. and Okazaki, T. (2005) “Foreign Technology Acquisition Policy and Firm Performance in Japan: 1957–1970: Micro-Aspects of Industrial Policy,” International Journal of Industrial Organization, 23, 563–86. Kohama, H. (2001) Sengo Nihon no Sangyo Hatten (Industrial Development in Postwar Japan) (Tokyo: Nihon Hyoronsha). Kohsaka, A., Yoshino, N. and Nakahigashi, M. (2007) “Japan,” in A. Kohsaka (ed.), Infrastructure Development in the Pacific Region, Routledge Studies in International Business and the World Economy (London: Routledge). Kosai, Y. and Kaminski, J. (1986) The Era of High-Speed Growth: Notes on the Postwar Japanese Economy (Tokyo: University of Tokyo Press). Kuroda, Y. (1995) “Labor Productivity Measurement in Japanese Agriculture,” Agricultural Economics, 12 (1), 55–68. Kuroda, Y. (1997) “Research and Extension Expenditure and Productivity in Japanese Agriculture, 1960–90,” Agricultural Economics, 16 (XXX), 111–24. Lau, L. and Yotopoulos, P. A. (1989) “The Meta-Production Function Approach to Technological Change in World Agriculture,” Journal of Development Economics, 31 (2), 241–69. Lynn, L. H. (1982) How Japan Innovates, A Comparison with the US In the Case of the Oxygen Steelmaking (Boulder, CO: Westview Press). Lynn, L. H. (1998) “Japan’s Technology-Import Policies in the 1950s and 1960s: Did They Increase Industrial Competitiveness?” International Journal of Technology Management, 15 (6/7), 556–67. Markusen, J. (2002) Multinational Firms and the Theory of International Trade (Cambridge, MA: MIT Press). Minami, R. (1994) Economic Development of Japan (New York: St Martin’s Press). Mundlak, Y. and Strauss, J. (1978) “Occupational Migration Out of Agriculture in Japan,” Journal of Development Economics, 5, 55–77. Murphy, K. M., Shleifer, A., and Vishny, R. W. (1989) “Industrialization and the Big Push,” Journal of Political Economy, 97 (5), 1003–26. Nakamura, T. (1995) The Postwar Japanese Economy: Its Development and Structure, 1937–1994 (Tokyo: University of Tokyo Press). Nakamura, T. and Ohashi, H. (2008) “Effects of Technology Adoption on Productivity and Industry Growth: A Study of Steel Refining Furnaces,” Journal of Industrial Economics, 56 (3), 470–99. Ohkawa, K. and Rosovsky, H. (1973) Japanese Economic Growth: Trend Acceleration in the Twenty Century (Stanford, CA: Stanford University Press). Okazaki, T. (2001) “The Government-Firm Relationship in Postwar Japan,” in J. E. Stiglitz and S. Yusuf (eds.), Rethinking the East Asian Miracle (Washington, DC: The World Bank). Okimoto, D. I. (1989) Between MITI and the Market: Japanese Industrial Policy for High Technology (Stanford, CA: Stanford University Press). Okimoto, D. I. and Nishi, Y. (1994) “R and D Organization in Japanese and American Semiconductor Firms,” in M. Aoki and R. Dore (eds.), The Japanese Firm: Sources of Competitive Strength (Oxford: Oxford University Press).
9780230_274587_05_cha04.indd 43
6/23/2010 5:04:57 PM
44
Shuhei Aoki et al.
Oshima, A. (2009) “Knowledge Diffusion and the Japanese Growth Miracle,” mimeographed, Department of Economics, University of Iowa. Otsu, K. (2007) “A Neoclassical Analysis of the Postwar Japanese Economy,” IMES Discussion Paper Series 07-E-01, Institute for Monetary and Economic Studies, Bank of Japan. Ozaki, R. S. (1972) The Control of Imports and Foreign Capital in Japan (New York: Praeger Publishers). Ozawa, T. (1974) Japan’s Technological Challenge to the West, 1950–1974: Motivation and Accomplishment (Cambridge, MA: MIT Press). Parente, S. and Prescott, E. (2004) Barrier to Riches (Cambridge, MA: MIT Press). Peck, M. J. and Tamura, S. (1976) “Technology,” in H. Patrick and H. Rosovsky (eds.), Asia’s New Giant (Washington, DC: Brookings Institution). Romer, P. M. (1990) “Endogenous Technological Change,” Journal of Political Economy, 98 (5, part 2), S71-S102. Rosenstein-Rodan, P. N. (1943) “Problems of Industrialization of South-Eastern Europe,” Economic Journal, 53 (210–11). Rustichini, A. and Schmitz, J. A., Jr. (1991) “Research and Imitation in Long-run growth,” Journal of Monetary Economics, 27 (1), 271–92. Sakakibara, M. (1997) “Evaluating Government-Sponsored R and D Consortia in Japan,” Research Policy, 26 (4–5), 447–73. Sakakibara, M. and Branstetter, L. (2001) “Do Stronger Patents Induce More Innovation? Evidence from the 1988 Japanese Patent Law Reforms,” RAND Journal of Economics, 31 (1), 77–100. Science and Technology Agency (1972, 1974, and 1991) White Paper on Science and Technology, available from http://www.mext.go.jp/b_menu/hakusho/hakusho.htm. Science and Technology Agency (1974, 1990, and 1991) Indicators of Science and Technology (Tokyo: Ministry of Finance Printing Bureau). Science and Technology Agency (ed.) (1980) Gaikoku Gijutsu Donyu Nenji Hokokusho (Annual Report of Foreign Technology Adoptions) (Tokyo: Printing Office of the Ministry of Finance). Society for Industrial Studies (1995) Sengo Nihon Sangyo Shi (Industrial History in Postwar Japan) (Tokyo: Toyo Keizai Shinposha). Sonobe, T. and Otsuka, K. (2006) Cluster-Based Industrial Development: An East Asian Model (London: Palgrave Macmillan). Syrquin, M. (1984) “Resource Reallocation and Productivity Growth,” in M. Syrquin, L. Taylor, and L. E. Westphal (eds.), Economic Structure and Performance: Essays in Honor of Harris B. Chenery (Orlando, FL: Academic Press). Westney, D. E. (1994) “The Evolution of Japan’s Industrial Research and Development,” in M. Aoki and R. Dore (eds.), The Japanese Firm: Sources of Competitive Strength (Oxford: Oxford University Press). Yamamura, E., Sonobe, T. and Otsuka, K. (2005) “Time Path in Innovation, Imitation, and Growth: The Case of the Motorcycle Industry in Postwar Japan,” Journal of Evolutionary Economics, 15 (2), 169–86. Yasuba, Y. (2002) “Rejoinder to Hayami and Ogasawara,” Journal of the Japanese and International Economies, 16 (2), 284–5. Yonekawa, S., Shimokawa, K. and Yamazaki, H. (eds) (1990–1) Sengo Nihon Keieishi (Business History in Postwar Japan) Vols. 1–3 (Tokyo: Toyo Keizai Shinposha). Young, A. (1995) “The Tyranny of Numbers: Confronting the Statistical Realities of the East Asian Growth Experience,” Quarterly Journal of Economics, 110 (3), 641–80.
9780230_274587_05_cha04.indd 44
6/23/2010 5:04:57 PM
5 Development Performance across Indian States and the Role of Governments Kaliappa Kalirajan,* Shashanka Bhide, and Kanhaiya Singh
5.1
Background
In the context of balanced regional development, drawing on the East Asian model of industrial development recently discussed in Sonobe and Otsuka (2006), a relevant point that needs to be made at the outset is that less equitable spatial development is not necessarily an indicator of constraints to overall growth of the economy. Yet, if large and populous spatial segments of the economy remain backward while the other regions move ahead for long periods of time, the overall national development strategy becomes unsustainable (World Bank, 2009). The crucial questions in this context are about how to bring spatially equitable development in a country, and which institution has the primary responsibility for bringing such equitable development. Conventionally, an economic system is comprised of state and market, which means the primary responsibility for spatially equitable development rests with either state or market, or both. While a federal framework provides for allocation of state resources in an equitable manner, the market operates based on business perspectives that may not aim for spatially equitable development. In contrast to the conventional method of assigning factors contributing to socioeconomic development either to the state or to the market, Hayami (2004) introduced another relevant entity: “community.” Communities are expected to maintain benefits for their members in a sustainable manner at the local level, a strategy that may again not have spatial equity as an objective. The task of achieving balanced development has, therefore, been primarily assigned to “state,” and neither to “market” nor to “community.” Even Hirschman (1958), who was a proponent * The authors are grateful to Keijiro Otsuka, Tetsushi Sonobe, and participants at the 2009 FASID Hakone Conference for their valuable comments and suggestions, which materially improved this paper.
45
9780230_274587_06_cha05.indd 45
6/23/2010 12:16:26 PM
46
Kaliappa Kalirajan et al.
of focused or “unbalanced” industrialization, earlier strongly argued for governmental intervention to counteract the “polarization effects” of freemarket forces in order to mitigate the misfortune of the “backward regions.” Though state has to take a leading role towards achieving balanced regional development, as Hayami (2001) has argued, the final outcome with respect to development, including equitable regional development, is a result of the interaction between the state, market, and community. India has a federal system of governance with both the state, or provincial, and the central governments responsible for the development of the nation as a whole. Policies at the central as well as state-level influence state-level variations in economic conditions, in turn. In this context, this chapter examines whether governments in India within a federal framework have been able to foster development equitably across its states, and whether there is any “community failure” in complementing the role of the state. The following section describes the approach followed in this chapter to examining equitable development across states in India. Spatial diversity of the Indian economy concerning its overall growth is shown in the next section, which is followed by a discussion of the existing spatial diversity in terms of basic development indicators. The next section attempts to explain the probable causes for variation in development across states in India. Empirical identification of factors influencing variations in development across states is presented in the following section. A final section brings out the overall conclusions of this chapter.
5.2
The approach
To provide a bird’s-eye view of the development of a country, per capita gross domestic product (GDP) is generally used as a measure. It is understood that GDP alone says nothing about distribution of income, nor about how it is used. If development is about people and how they live, then one needs a measure of the level of living of the population, into which is built a distributive element of income. In this context, it is worth citing the statement from the Indian prime minister’s foreword to the Planning Commission’s Eleventh Five-Year Plan: “The transition to high growth is an impressive achievement, but we must not forget that growth is not the only measure of development. Our ultimate objective is to achieve broad-based improvement in the living standards of all our people” (Planning Commission, 2008, p. iii). Drawing on the United Nations Research Institute for Social Development (UNRISD, 1968), and based on the uniform availability of data, the following variables, which represent a wide range of basic developmental aspects of an economy, all of which tend to change more or less simultaneously as the society develops, are considered in this study: infant mortality rate (IMR), life expectancy at birth (LE), literacy rate (LR), telecom density per
9780230_274587_06_cha05.indd 46
6/23/2010 12:16:26 PM
Development Performance across Indian States 47
thousand population (TD), and per capita electricity consumption in kwh (EC).9 These indicators refer to the very basic needs of life, and it is expected that the state and central governments at different levels would strive to bridge the spatial gap in these dimensions of development within India. Thus, first, spatial distribution of the above variables, along with gross state domestic product (GSDP), is used over the years to gauge the effectiveness of region-specific constraints or accelerators of development. One of the region-specific factors considered in this study is “community.”
5.3
Spatial diversity in overall growth of the Indian economy
India’s overall growth performance from 1951–2 to 2007–8 is shown in Figure 5.1, with probable reasons for its fluctuations. India’s past record, even before the launching of economic reforms in the early 1990s, points to wide variations across its subnational units, including states.10 The federal framework is expected to achieve more balanced development of the diverse country than centralized rule, as decentralized governance can be expected to utlilize the available resources more efficiently to meet the aspirations of the local population (Hayami, 2001). At a general level, the extent of variation in the Indian economy can be illustrated by broad measures of the size of state economies. There are 29 states in India with their own democratically elected assemblies. Among these, 10 states have populations of more than 50 million. There are 8 states with populations below 5 million. Six states have a gross state domestic product (GSDP) of more than $50 billion. Table 5.1 provides profiles of the states in terms of size of populations and economies. Some insights into the trends concerning divergence in economic growth across states can be obtained through the coefficient of variation in per-capita income across the states over time. Following Kalirajan and Akita (2002), the weighted coefficient of variation of per-capita gross state domestic product (GSDP) in constant 1993–4 prices for the period 1980–1 to 2004–5 is calculated.11 The trend indicates that there has not been a significant increase in disparity concerning per-capita gross state domestic product as captured by this measure. The weighted coefficient of variation for 1980–1 was 0.2297, and it increased to 0.2307 in 2004–5, which is not significant. This period had been one of relative stability in spatial disparity in India. What are the factors that contribute to such relative stability in development?
5.4
Explaining the growth differential across states
Following the seminal work of Barro (1991), based on theory and on the availability of uniform data across states, the following reduced-form equation was formulated to examine the conditional convergence of per-capita
9780230_274587_06_cha05.indd 47
6/23/2010 12:16:26 PM
1963–4
1961–2
1959–0
1957–8
1953–4
1971–2
1967–8
1965–6
3.5 3.75 100
4.57
1965–6
3.8 3.94 67
4.51
1957–8
1961–70
1973–4
1972–3 and 1979–80
4.64
3.6 3.16 105
1971–80
1951–2
5.40
5.4 5.40 70
–
TG = 7.9
Drought East Asian Crisis (97), Drought Gulf War, oil prices, USA economic reforms, fiscal deficit, Crises economic sanctions (98), Kargil war, 9/11/01 attack and global slowdown, swelling fiscal deficit
1981–90
Pattern of India’s economic growth from 1951 to 2008
Trend growth Average growth Coefficient of variation (percent) Average growth without drought and crisis years Years of the drought and crises
1969–70
1951–60
1975–6
Percentage growth
Figure 5.1
1955–6
Data
1977–8
−6.0 1981–2
−4.0
1979–80
War, oil prices, drought, current Drought A/C deficit, monetary Capital-intensive First 5tightening year plan, Stringent fiscal Drought, Nehru and Shatri and F Exchange control died, war, high fiscal deficit, Drought, oil crisis, current A/C deficit US aid suspended
1983–4
0.0
1987–8
−2.0
1991–2 and 1997–8
6.27
6.2 5.59 43
1991–2000
1991–2
2.0
1989–90
4.0
1993–4
6.0
1985–6
TG = 6.2
1995–6
8.0
TG = 5.4
8.34
7.9 7.58 22
2001–9
2002–3, 2008–9
1997–8
TG = 3.6
2001–2
TG = 3.5
1999–2000
TG = 3.8
2003–4
10.0
2005–6
12.0
2007–8
9780230_274587_06_cha05.indd 48
6/23/2010 12:16:26 PM
Development Performance across Indian States 49
Table 5.1
Size and income of India’s state and union territories (2005–6) GSDP
Sl. No. State/UT 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
Andhra Arunachal Pradesh Assam Bihar Jharkhand Goa Gujarat Haryana Himachal Jammu and Kashmir Karnataka Kerala Madhya Pradesh Chattisgarh Maharashtra Manipur Meghalaya Mizoram Nagaland Orissa Punjab Rajasthan Sikkim Tamil Nadu Tripura Uttar Pradesh Uttaranchal West Bengal Pondicherry All India
Population million Rs billion
Per capita GSDP Billion US$
Rs
US$
80.4 1.2 28.5 90.2 29.1 1.6 54.6 23.1 6.6 10.9
2360 29 575 802 622 124 2198 1064 255 265
53.32 0.66 13.00 18.11 14.06 2.80 49.65 24.03 5.75 5.99
29,369 25,086 20,186 8891 21,377 79,389 40,221 45,974 38,457 24,397
663 567 456 201 483 1793 909 1038 869 551
56.0 33.4 65.9 22.7 104.2 2.5 2.5 1.0 2.5 38.8 26.5 61.8 0.6 64.9 3.4 181.9 9.2 84.8 2.6 1116.1
1680 1190 1163 519 4381 57 63 27 57 785 1097 1242 18 2235 94 2798 262 2347 68 32,757
37.94 26.88 26.28 11.73 98.95 1.29 1.43 0.61 1.28 17.74 24.79 28.06 0.41 50.49 2.12 63.19 5.91 53.02 1.45 739.93
29,999 35,601 17,649 22,873 42,056 22,684 25,699 27,027 22,736 20,251 41,420 20,095 31,186 34,424 27,694 15,382 28,572 27,668 25,712 29,350
678 804 399 517 950 512 581 610 514 457 936 454 704 778 626 347 645 625 588 663
real gross domestic product across states, after controlling for important human capital, physical capital, and decentralization indicators: SYPC1 a1 a2 SYPC8081 a3 LIT8081 a4 STPOP a5 DCI a6 Cst u1
(1)
With the expected results of either convergence or divergence, the next interesting question is whether the initial conditions of the structure of the economy (agriculture share and manufacturing share) have been important variables explaining differences in growth of per-capita real gross domestic product.
9780230_274587_06_cha05.indd 49
6/23/2010 12:16:27 PM
50 Kaliappa Kalirajan et al.
Here, two periods of initial conditions were considered: 1993–4 was the period immediately after the institution of economic reforms in India, and 1980–1 was the period of recovery, with a 7-per-cent growth rate, from a severe drought and current-account deficit crisis. The following equations were formulated: SYPC2 b0 b1SYPC9394 b2 AGR9394 b3MFG9394 b4 LIT9394 b5STPOP b6DCI b7Cst u2
(2)
SYPC2 b0 b1SYPC8081 b2 AGR8081 b3MFG8081 b4 LIT8081 b5STPOP b6DCI b7Cst u2
(3)
The further interesting issues to be explored in this study are: (i) whether a productive agricultural sector is a prerequisite for the growth of the manufacturing sector, or whether a comparative disadvantage in agriculture stimulates the growth of manufacturing sector for survival; and (ii) whether richer states achieve higher growth of both manufacturing and agricultural sectors. The potential variables in testing the above hypothesis include percapita growth in agricultural real gross state domestic product, per capita growth in manufacturing real gross state domestic product, selected structural variables of initial conditions of income, agricultural share and manufacturing share, literacy rate, demographic composition, a decentralization index, and the coastal region. AGRI_GSDP g0 g1YPC8081 g2 AGR8081 g3MFG8081 g4 LIT8081 g5STPOP g6DCI g7Cst u3
(4)
AGRI_GSDP g0 g1YPC9394 g2 AGR9394 g3MFG9394 g LIT9394 g STPOP g DCI g Cst u3
(5)
MFG_GSDP d0 d1YPC8081 d2AGR8081 d3MFG8081 d4LIT8081 d5STPOP d6DCI d7Cst u4
(6)
4
5
6
7
MFG_GSDP d0 d1YPC9394 d2AGR9394 d3MFG9394 d4LIT9394 d5STPOP d6DCI d7Cst u4
(7)
The symbols in the estimated equations are as follows: SYPC = per capita growth of real gross state domestic product, YPC8081 and YPC9394 = per-capita gross state domestic product in 1980–1 and 1993–4 respectively; AGRI_GSDP = per-capita growth in agricultural real gross state domestic product; MFG_GSDP = per-capita growth in
9780230_274587_06_cha05.indd 50
6/23/2010 12:16:27 PM
Development Performance across Indian States 51
manufacturing real gross state domestic product; AGR8081 and AGR9394 = initial agricultural conditions, which are 1980–1 and 1993–4 respectively of shares of the agriculture sector in GSDP; MFG8081 and MFG9394 = initial manufacturing conditions, which are 1980–1 and 1993–4 shares of manufacturing sector in GSDP (all in fractions); LIT8081 and LIT9394 = literacy rates in 1980–1 and 1993–4 respectively as initial conditions variables; DCI = fiscal decentralization index, which is the share of state revenue including statutory transfers to denote fiscal decentralization in its expenditures; Cst = dummy variable, taking the value of 1 for the presence of coastal area and zero otherwise; STPOP = Schedule Tribe community, which is shown as a fraction of the population; and ST components are from the 1991 census data. Each of the above growth equations was estimated using the data for 29 Indian states for the post-reform period of 1993–4 to 1999–2000 following a general-to-specific approach, and the results are given in Table 5A.1 using 1993–4 initial conditions. Relevant variables are presented at 1993–4 constant prices.
Table 5A.1 Explaining the variations in real per-capita growth in the post-reform periods Variables
SYPC 1
SYPC 2
AGRI_GSDP
MFG_GSDP
Constant
AGR9394
–0.289** (0.140) 0.202 (0.198) –
MFG9394
–
–0.155* (0.052) 0.138 (0.146) 0.121 (0.098) 0.482** (0.228) 0.115** (0.050) 0.025 (0.019) 0.048** (0.022) 0.048** (0.023) 0.78 0.36 [0.54] 0.16 [0.23]
0.072** (0.034) 0.132 (0.148) 0.148 (0.104) –0.149 (0.172) –0.255 (0.364) 0.018** (0.008) 0.035 (0.038) –0.008 (0.007) 0.74 0.37 [0.58] 0.18 [0.28]
0.292** (0.136) 0.128 (0.146) –0.119 (0.109) 0.822** (0.409) 0.680** (0.338) -0.015 (0.025) 0.032** (0.015) 0.042** (0.020) 0.72 0.28 [0.52] 0.14 [ 0.25]
YPC9394
LIT9394 STPOP DCI Cst R-bar square Functional form CHSQ(1) Heteroskedasticity CHQ(1)
0.114** (0.052) 0.028 (0.026) 0.044** (0.020) 0.045** (0.022) 0.80 0.38 [0.57] 0.21 [0.28]
Notes: 1. Variables have been defined in the text. 2. Figures reported below each coefficient estimate are its standard errors. 3. * means significant at the 1 percent level. 4. ** means significant at the 5 percent level. 5. Figures in square brackets are critical values.
9780230_274587_06_cha05.indd 51
6/23/2010 12:16:28 PM
52
Kaliappa Kalirajan et al.
Table 5.2 periods
Explaining the variations in real per capita growth in the post-reform
Variables
SYPC 1
SYPC 2
AGRI_GSDP
MFG_GSDP
Constant
AGR8081
–0.277** (0.132) 0.252** (0.122) –
MFG8081
– 0.125** (0.052) 0.033** (0.016) 0.048** (0.022) 0.048** (0.024)
–0.164* (0.048) 0.138 (0.146) 0.142** (0.069) 0.458** (0.219) 0.116** (0.051) 0.026** (0.012) 0.058** (0.027) 0.044** (0.021)
0.065** (0.031) 0.128 (0.143) 0.162* (0.047) –0.145** (0.072) –0.262 (0.388) 0.023** (0.011) 0.032 (0.037) –0.012** (0.005)
0.266** (0.126) 0.131 (0.142) –0.122* (0.036) 0.768** (0.376) 0.782** (0.388) –0.022 (0.028) 0.046** (0.022) 0.062** (0.030)
0.89 0.42 [0.54] 0.18 [ 0.23]
0.81 0.36 [0.54] 0.16 [0.23]
0.72 0.38 [0.54] 0.12 [0.23]
0.76 0.29 [0.54] 0.10 [ 0.23]
YPC8081
LIT8081 STPOP DCI Cst R-bar square Functional form CHSQ(1) Heteroskedasticity CHQ(1)
Notes: 1. Variables have been defined in the text. 2. Figures reported below each coefficient estimate are its standard errors. 3. * means significant at the 1 percent level. 4. ** means significant at the 5 percent level. 5. Figures in square brackets are critical values.
Results using 1980–1 initial conditions are better in terms of number of significant variables and R bar squares, and hence these results are presented in Table 5.2, which are used for further analysis. This finding suggests that the influence of the selected macroeconomic variables on agriculture, manufacturing, and overall growth in the post-reform period is clearly different from growth in the pre-reform period. The selected model with 1980–1 initial conditions significantly captures some of the features of social and economic diversity across states. The results presented below discuss only those variables that are found to be important in terms of statistical significance for Indian states. The important results are that there is divergence with respect to per-capita growth in real gross state domestic product across states, and that unlike in the East Asian model of economic growth, having a productive agricultural sector as the initial condition for the growth of the manufacturing sector is not valid in the Indian context. The inference is that having comparative disadvantage in agriculture stimulates the growth of the manufacturing sector for
9780230_274587_06_cha05.indd 52
6/23/2010 12:16:28 PM
Development Performance across Indian States 53
survival.12 Another interesting result is that decentralization and initial conditions of income level have a significant effect on manufacturing growth. The results of the SYPC1 equation clearly show that there is divergence in the growth of real gross state domestic product across states (see Rao et al., 1999). The principal force driving convergence in the neoclassical growth model is diminishing returns to reproducible capital. Thus, states with lower initial values of capital–labor ratios will have high marginal products of capital and, therefore, will tend to grow at higher rates. However, as Hayami (2001) has argued, inefficient and poor-quality institutions and organizations could lead to violation of the critical assumption of diminishing returns on reproducible capital. This means divergence of income for a considerable period of time in the development process. Thus, it is logical to argue that the convergence hypothesis will hold only when country-specific institutions and organizations do not intervene in the process negatively to delay or constrain the convergence process. The distinction between growth that translates into “a rising tide lifts all the boats,” and growth that is disproportionately favorable to some states, has been recognized by the central government in its Eleventh Five-Year Plan document. The implication here is that lack of effective functioning of proper institutions at the central and state levels has been a problem. In the case of manufacturing, the inference is that integration of rural industries with urban sectors through the subcontracting system, which is built on community trust and cooperation in East Asia à la Hayami, is missing in the Indian growth process (see Sonobe and Otsuka, 2006). In the context of agriculture, the dynamism that was generated by the Green Revolution had worked its way fully into production in the 1980s, and there was no alternative source of strong productivity growth (Kalirajan et al., 2001). A lack of infrastructure and various policy constraints affecting agricultural productivity and trade have been major constraints on technological breakthrough in agriculture, as discussed by Vaidyanathan (1995). Results of SYPC2, which show a smaller coefficient for agriculture relative to manufacturing, appear to support these inferences, particularly with respect to an urgent need for improvement in agricultural productivity. Decentralization has a positive effect on overall gross state domestic product growth and on manufacturing growth. The coefficient implies that India’s overall growth rate can still be substantially stepped up should the central government decentralize economic policy, further allowing states to make important economic decisions based on their state-specific production environments. Significant and crucial fiscal, infrastructure, and regulatory decisions on economic management still remain at the central government level. The inference from the coefficient is that the decentralized system of governance would help the states to be more attractive to prospective domestic and foreign direct investors, which has a bearing on manufacturing growth. The second set of variables that are found to be important in the growth process across states is the structure of the economy. For the 1993–2000
9780230_274587_06_cha05.indd 53
6/23/2010 12:16:28 PM
54
Kaliappa Kalirajan et al.
data sets, the structure of the economy during 1980–1 is considered as one of the initial conditions in this chapter, as the 1993–4 initial conditions did not provide better results. The coefficient of initial income level, YPC80-81 in the SYPC2, and MFG_GSDP equations are positive, and statistically significant. The significant negative coefficient of initial manufacturing conditions in the AGRI_GSDP equation implies that states with comparative disadvantage in manufacturing appear to grow faster in agriculture for their survival. Similarly, the negative coefficient of initial agricultural condition in the MFG_GSDP equation means that states with comparative disadvantage in agriculture appear to grow faster in manufacturing for survival. The third set of variables found to be significant in explaining variations in per-capita overall growth, and growth in agriculture and manufacturing, are related to the social fabric of the Indian states.13 Drawing on Hayami (2001), they relate to to certain patterns of behavior of state governments, welfare organizations, and general culture. The variables falling into this category are initial literacy rate across states, and share of population of the scheduled tribal communities. Motivation for considering the latter community variable à la Hayami is rather complex, and needs some explanation. The effect of share of ST population is positive and statistically significant on AGRI_GSDP, SYPC1, and SYPC2, while it is negative but statistically not significant in the case of MFG_GSDP. It is important to understand the genesis of these two different results. The people in tribal regions are mostly isolated from the cosmopolitan culture and bound tightly by their local culture and traditional way of life. The presence of the tribal community across states varies significantly. For example, based on 1991 census data, 95%, 88%, 86%, and 64% of the population respectively in the states of Mizoram, Nagaland, Meghalaya, and Arunachal Pradesh are tribal communities. The desire to change is lacking in tribal regions, where the majority of members work as agricultural laborers in places near their habitations (see XaXa, 2001). It is true that states having larger ST population enjoy more welfare programs compared to other states, sponsored by both state and central governments. Such welfare programs, which are mainly focused on rural areas, have spillover effects and are likely to benefit all segments of the rural population. Thus, the welfare programs provide a link between the presence of ST community and agricultural growth, though these programs have not made significant impacts on the welfare conditions of the ST communities. The important question here is whether such a situation is due to state failure in not having proper institutions to implement the welfare programs effectively, or due to community failure in lack of enthusiasm for participating in the welfare programs. In this context, it is worth mentioning the study by Kijima (2006), which argues that despite policies targeting scheduled tribes (ST), there remain large disparities between living standards of ST and non-ST households in India. Much of the disparity between ST and non-ST comes from the fact that areas where the ST live are
9780230_274587_06_cha05.indd 54
6/23/2010 12:16:29 PM
Development Performance across Indian States 55
different from those where the non-ST live, and the education level of the ST is remarkably low compared to the non-ST population. In this context, it is worth citing the Indian prime minister’s recent statement: “The rapid growth achieved in the past several years demonstrates that we have learnt how to bring about growth, but we have yet to achieve comparable success in inclusiveness” (Planning Commission, 2008, p.iii). Quantitatively, each percentage-point difference in the ST community leads to an increase in per-capita agricultural growth of 0.02 percent. The presence of the ST community has no significant growth impact on manufacturing, which could be due to the poor literacy rate and unwillingness of the ST population to travel to urban areas. These results have significant implications for states like Mizoram, Nagaland, Meghalaya, and Arunachal Pradesh, where tribal communities are predominant. There is a need to end the region’s insulation from the dynamism of the rest of India. The average performance of these communities need to be improved, and lack of enthusiasm among the community to change its lifestyle is a kind of “community failure” discussed by Hayami (2009). As Hayami (2001) has argued, it is imperative for states to approach communities layer by layer to win their confidence and to educate them about the benefits of integrating with the rest of the society, which is not found in tribal thinking (see for example, XaXa, 2001). Another important result is that states with high initial levels of literacy appear to have higher per-capita growth in gross state domestic product, and also higher per-capita manufacturing growth, while such a relationship cannot be established in the case of agricultural growth. The coastal area can stimulate industrial growth, and overall growth is due to its capability of serving both domestic and international markets: hence its coefficient is positive in all equations except the AGR_GSDP equation. The much faster growth of coastal China provides solid evidence for the influence of coastal areas on industrial and overall economic growth.
5.5
Measures of spatial disparity in development
We present disparity patterns in other selected developmental indicators across states. We adopt the measure of weighted coefficient of variation to examine the trends in spatial balance in the basic development indicators: infant mortality rate (IMR), life expectancy at birth (LE), literacy rate (LR), telecom density per thousand population (TD), and per capita electricity consumption in kwh (EC), and the results are given in Tables 5.3 and 5.4. The trends clearly show that there is a reduction in disparity in literacy rate, and life expectancy at birth. In the case of infant mortality rate, the trends are less clear (Table 5.3). The trends in infrastructure development, which include telephone density and per capita electricity consumption, also show a general improvement, particularly in very recent times (see Table 5.4).
9780230_274587_06_cha05.indd 55
6/23/2010 12:16:29 PM
56 Kaliappa Kalirajan et al.
Table 5.3 The pace of development and spatial balance in development: Selected human development indicators Life expectancy at birth
Infant mortality rate (per thousand)
Year
Average
CV
Year
Average
1983 1988 1990 1991 1993 1995 1999 2000 2001 2003
60.2 61.1 59.2 59.7 60.5 61.2 62.1 62.3 62.5 63.0
0.7398 0.7299 0.7541 0.7472 0.7376 0.7298 0.7197 0.7175 0.7157 0.7107
1971 1976 1981 1986 1991 1994 1998 2005 – –
121 123 107 95 77 53 66 54 – –
Literacy rate
CV
Year
Average
CV
0.0565 0.0650 0.0681 0.0685 0.0708 0.1237 0.0670 0.0680 – –
1971 1981 1991 2001 – – – – – –
32.8 42.2 51.2 63.9 – – – – – –
0.0830 0.0655 0.0575 0.0370 – – – – – –
Note: The coefficient of variation is calculated over data for 15 major states of India. The CV is estimated as a weighted measure using population shares of the states as weights.
Table 5.4 The pace of development and spatial balance in development: Selected infrastructural development indicators Telecom density (% population)
Year
Average
1980 1985 1987 1988 1989 1990 1991 1992 1993 1994 1995 1998 1999 2000 2003 2004 2006 2007
0.38 0.40 0.47 0.48 0.52 0.55 0.62 0.71 0.83 1.00 1.20 1.68 2.37 2.68 4.24 5.69 10.74 15.85
Electricity consumption (kwh)
CV
Year
0.3350 0.1661 0.1717 0.1720 0.1726 0.1670 0.1654 0.1603 0.1592 0.1606 0.1629 0.1275 0.1499 0.1498 0.1555 0.1618 0.1250 0.1139
1975 1980 1983 1985 1987 1990 1991 1992 1993 1994 1996 2002 2004 – – – – –
Average (per capita) 97 121 150 168 196 232 247 263 288 306 336 359 387 – – – – –
CV 0.1157 0.1322 0.1335 0.1298 0.1352 0.1285 0.1244 0.1268 0.1197 0.1234 0.1235 0.1597 0.1523 – – – – –
See Note to Table 5.3.
9780230_274587_06_cha05.indd 56
6/23/2010 12:16:29 PM
Development Performance across Indian States 57
Thus, there has been an improvement in recent times in four out of the five development indicators considered, and progress appears to be more spatially balanced than before. The movement towards more equitable levels of performance along with rising development indicates that there are factors facilitating spatially equitable development. It is important to identify such factors, which is attempted in the following pages. Factors facilitating spatially equitable development The central government aims to ensure equal access to basic services such as primary education, primary health care, infrastructure such as roads, drinking water, electricity, and communication services across states. However, Table 5.5 shows that states could finance their current expenditures or revenue expenditures by their own revenues only to about 56 percent. Therefore, the centre shares with states some of its revenues from sources such as income tax. The norms for allocation of resources from the centre to states are mostly “formula”-based, which depends on factors such as population without any distinction between ST and non-ST populations. There are two main institutions that are constitutionally created every five years by the president of India to transfer resources from the centre to states: the Finance Commission and the Planning Commission. Both institutions first fix the amount of tax revenue that needs to be transferred from the centre to states, which is called “vertical transfer,” and then distributes the allocated amount to states, which is called “horizontal transfer,” using different formulas. The Planning Commission also makes specific-purpose transfers for various central schemes implemented by different ministries of the central government without using a formula, which are called “non-statutory” transfers (Rao and Chelliah, 1995). Table 5.6 shows the vertical transfer from the centre to states over
Table 5.5
Trends in state’s own revenue receipts and revenue expenditures
Year
State’s own receipts State’s rev. expend. State’s own rev. receipts to total receipts (%) to total expend. (%) to state’s rev. expend. (%)
1990–1 1995–6 1999–2000 2000–1 2001–2 2002–3 2003–4 2004–5 2005–6
35.2 39.2 38.6 37.8 40.2 38.28 37.54 38.10 37.07
54.6 57.0 56.4 56.0 56.9 54.31 56.27 56.28 56.59
53.1 58.6 49.8 48.6 50.0 56.01 53.83 60.05 56.07
Source: Table 7 in Rao et al. (2008).
9780230_274587_06_cha05.indd 57
6/23/2010 12:16:29 PM
9780230_274587_06_cha05.indd 58
6/23/2010 12:16:30 PM
12.11 13.56 14.48 10.57 10.39
5
25 50 10 7.5 7.5
Population Per-capita income Tax effort Fiscal management
60 25 7.5 7.5
Weight (%)
1.66 1.54 1.06 0.63 0.82
6
Source: Authors’ compilation from various Planning Commission plan reports and Finance Commission reports.
Population Income Area Tax effort Fiscal discipline
Weight (%)
Criteria
24.35 22.77 24.79 23.75 26.13
4
Criteria
1.96 2.52 3.42 2.34 5.2
3
Gadgil Formula used in X Plan (2002–7)
22.39 20.25 21.37 21.4 20.93
2
B. Horizontal Transfer Formula of 12th Finance Commission (2005–10)
VII FC VIII FC IX FC X FC XI FC (first two years)
1
Period
13.77 15.1 15.54 11.19 11.21
7
38.11 37.86 40.33 35.79 37.2
8
Total transfers Total other Total through Finance Grants through Share in central Share in Commission (2+3) Planning Non-plan grants transfers transfer taxes (%) grants (%) Commission (%) (nonstatutory) (%) (5+6) (%) (4+7) (%) (%)
Finance Commission transfers
Other transfers
Finance Commission (FC), Planning Commission and nonstatutory transfer of resources from the centre to states
A. Vertical transfer Transfers from centre to states as percentage of gross revenue receipts of the centre
Table 5.6
Development Performance across Indian States 59
the years, which indicates the overall transfer of about 38 percent of the centre’s revenues to states. The formula used for horizontal transfer by the XII Finance Commission covering the period 2005–10 is given in Table 5.6 along with the Planning Commission’s Gadgil formula used in the X Plan (2002–7).14 The Gadgil formula is binding on the Planning Commission, but not for the Finance Commission (Rao, 2005). How have the instruments of state and central governments worked in achieving spatial equitable distribution in development, in the Indian context? We examine this dimension of the process in the rest of this chapter.
5.6
Disparities in state government expenditures
What has been the pattern of expenditures across states since the beginning of a policy regime which has been more reliant on markets to deliver growth than in the past? We have used revenue (or, broadly, current expenditures) of the state governments for analysis, as they make up about 77 percent of total expenditures today. Within revenue expenditures, “development expenditures” make up to 80 percent of spending. “Development expenditures” refers to expenditures on various socioeconomic development programs in the social sectors and economic sectors. The “development expenditures” of the state governments are grouped into two broad categories: social services and economic services. Social services includes health and education. Economic services includes development programs in different sectors of the economy, particularly in infrastructure. For example, revenue expenditures on programs relating to agriculture, roads, electricity, and industries are grouped under economic services. These are not capital expenditures but relate to expenditures on operation and maintenance of ongoing programs. Data are obtained from the Handbook of Statistics on State Government Finances (RBI, 2004) and previous publications on state finances by the Reserve Bank of India (RBI). The state-level population estimates are obtained by interpolating decadal census estimates to obtain per-capita expenditures for the states. The values in current figures are deflated by the wholesale price index to obtain expenditures and GSDP in real or constant figures. The trends in government development expenditures are examined in a regression model in the framework of the “convergence analysis.” Although the underlying theoretical arguments are quite different, the methodology of income convergence analysis of Barro and Sala-i-Martin (1995) provides a useful tool to examine whether the disparity across states in state government expenditures is increasing. The conditional convergence regression model we used is: 1 PCREt 1 ln ln PCREt0 + a2 ln PCYt0 a0 a1 t − t PCRE 0 t0 t − t0 a3 STPOP + a4 DCI u
9780230_274587_06_cha05.indd 59
(8)
6/23/2010 12:16:30 PM
60
Kaliappa Kalirajan et al.
where PCRE = Per-capita real state government expenditures; PCY = Per-capita real gross state domestic product (average of three years ending in year t or t0 in all cases except for 1980–1 where the average is for three years beginning in 1980–81; the averages are taken to remove abnormal data points); t0 = First year of the time period; t = Final year of the time period; and the rest of the variables are as defined earlier. Given our interest in welfare indicators, we have extended the analysis to cover four types of development expenditures of the state governments: (1) expenditure on economic services within development expenditures; (2) expenditure on social services within development expenditures; (3) expenditure on medical and public health services within social expenditures; and (4) expenditure on education services within expenditures on social services. We have examined the pattern of expenditure for two periods: Pre-reform periods 1980–1 to 1985–6 Post-reform periods 1993–4 to 1999–2000 We are not particularly interested in the steady-state levels of expenditures, but merely wish to examine whether the expenditures across states are likely to be “converging” or “diverging” over time. Convergence would imply that the disparity is likely to be declining, and divergence would imply the opposite. In the above equation, if the coefficient “a1“ is positive, the percapita expenditures of the different states would be moving at different rates, with the “higher expenditure states” increasing expenditures faster than the “lower expenditure states.” Therefore, there would be no convergence of percapita expenditures across states and the disparity would increase over time. If the coefficient is negative, then the per-capita expenditures would be converging or disparity would decrease over time. If the coefficient is zero, then again disparity would not be rising. The role of per capita real GSDP in the equation is to control for overall “initial conditions” of the state economy. If the average income (per-capita GSDP) is larger in one state as compared to the other, the expenditures may also be higher because of the availability of larger resources (through their state taxes) to that state. Once we control for this variable, the pattern that remains should reflect the influence of the other factors, including the devolution of resources from the central government to the states, on the tendencies of the states to spend. In order to control for this latter aspect of central government transfer, and also the
9780230_274587_06_cha05.indd 60
6/23/2010 12:16:30 PM
Development Performance across Indian States 61
Table 5.7 Results of regression analysis of interstate disparity in state government expenditures Expenditures
PCREt 0
PCYt 0
STPOP
DCI
1980–5 Econ. services 0.3268**(0.1552) Social services –0.2283(0.1992) Pub. health 0.1165(0.1237) Education –0.0987(0.1013)
0.2572(0.2116) 0.1976(0.1882) 0.2001(0.1980) 0.1879(0.2011)
0.1872**(0.0884) 0.1743**(0.0820) 0.1552(0.1821) 0.1390(0.1542)
0.1932(0.1878) 0.1667(0.1740) 0.1223(0.1452) 0.1086(0.1277)
1999–2000 Econ. services 0.3129(0.2952) Social services –0.2627**(0.1292) Pub. health –0.2772**(0.1355) Education –0.2028**(0.1004)
0.2882(0.2416) 0.1955**(0.0899) 0.2174**(0.1080) 0.2072**(0.1033)
0.1762**(0.0807) 0.1881**(0.0920) 0.1825**(0.0875) 0.1954**(0.0961)
0.1774**(0.0878) 0.1375**(0.0640) 0.1448**(0.0723) 0.1220**(0.0589)
Note: ** Means significant at the 5 percent level.
capacity of the state to collect revenue, the decentralization index has been used as another controlling variable. As the earlier results show that the presence of the ST population has positive impact on overall economic growth, it is also used as a conditional variable in equation (8). The results obtained from applying the ordinary least-squares method of estimation to equation (8) with different dependent variables, with concerned independent variables explained above, are summarized in Table 5.7. Given our interest in finding out whether there is convergence or divergence in per-capita real state government expenditure and its components across states, it may be observed that the expenditures on “social services” show a statistically significant convergence pattern more consistently than the expenditures on “economic services.” In fact, the expenditures on “economic services” show no convergence for any subperiods considered, while the “social expenditures” show convergence in the post-reform periods. Thus, state government expenditures on social sectors appear to be more pro-spatially equitable than expenditures on “economic services.” Within the “social services,” expenditures on both “health” and “education” services show statistically significant convergence in the post-reform periods, indicating that the mechanisms driving these expenditures are based more uniformly on the needs of population across all regions of the country. The coefficients of decentralization index and the ST population are positive and significant at the 5 percent level for social services, medical and public health, and education expenditures. These results corroborate earlier findings that greater decentralization improves overall economic growth, which in turn facilitates higher expenditure on social services. Further, the significant and positive coefficient of the ST population implies that state
9780230_274587_06_cha05.indd 61
6/23/2010 12:16:30 PM
62 Kaliappa Kalirajan et al.
governments do pay special attention to the improvement of welfare of the ST population, though the central transfer to states does not make any allowance for the presence of ST population in states. However, these coefficients are not significant for the economic services expenditures. The analysis presented here does not clarify whether the patterns of divergence or convergence are a result of higher growth of expenditures by the low expenditure states, or declining rates of growth of expenditures in the high expenditure states. In other words, the fiscal imbalances that result from increased expenditures may also act as a moderator of spending behavior at the state level. The initial condition of average per-capita GSDP does not influence the dynamics of expenditures.
5.7
Conclusions
Though the Indian government’s specific objective of keeping the disparities in development indicators from widening has been achieved with respect to some development indicators, there has been a growing concern that the less developed states continue to lag behind the developed states and that this cleavage is increasing. The policy concern is explicit in the latest Eleventh Five-Year Plan (2007–8 to 2011–12), where inclusive growth has been a watchword. In this context, this chapter refers to the sources of disparity and the role that may be played by the three pillars of development, the state, market, and community, articulated by Hayami (2004). The economic reforms of the 1990s have given more space to the markets in the allocation of resources, as compared to the state relative to the pre-reform days. What implication does this have for spatial equity in development? The state continues to be responsible for the supply of public goods including basic human capital and infrastructural development services across the country. Analyses presented in this chapter suggest that the mechanisms by which state governments provide resources for such services do not continuously lead to higher interstate disparity. If this pattern is a result of equitable sharing of central resources by the states through decentralization, this element of state behavior is important in keeping the interstate disparities from widening. Further, the results point out that the expenditures on basic services such as health and education are pro-spatially more equitable than in the case of economic services. Overall, the results in this chapter indicate that government in India within a federal framework has mechanisms that foster development equitably across its states, particularly through health and education expenditures aimed at improving human-capital development. However, the slowly rising disparities in economic services across states, and community failures in the sense of not participating in government welfare programs that exist in certain states, which make some communities less educated than others, warrant the attention of the central and state governments. Drawing on
9780230_274587_06_cha05.indd 62
6/23/2010 12:16:31 PM
Development Performance across Indian States 63
Hayami (2001), it is conjectured that such tendencies arise mainly due to lack of appropriate and efficient institutions at state and central government levels in India, which indicates the need for further institutional reforms.
References Barro, R. J. (1991) “Economic Growth in a Cross Section of Countries,” Quarterly Journal of Economics, 106, 407–33. Barro, R. J. and Sala-i-Martin, X. (1995) Economic Growth (New York: McGraw-Hill). Hayami, Y. (1998, 2001) Development Economics (Oxford: Oxford University Press). Hayami, Y. (2004) “The Community, the Market, and the State in Economic Development,” Paper presented at the Beijing Forum. Hayami, Y. (2009) “Social Capital, Human Capital and the Community Mechanism: Toward a Conceptual Framework for Economists,” Journal of Development Studies, 45 (1), 96–123. Hirschman, A. (1958) The Strategy of Economic Development (New Haven, CT: Yale University Press). Kalirajan, K. P. and Akita, T. (2002) “Institutions and Interregional Inequalities in India: Finding a Link Using Hayami’s Thesis and Convergence Hypothesis,” Indian Economic Journal, 50 (2), 47–59. Kalirajan, K. P., Mythili, G. and Sankar, U. (2001) Accelerating Growth Through Globalization of Indian Agriculture (New Delhi: Macmillan). Kijima, Y. (2006) “Caste and Tribe Inequality: Evidence from India, 1983–1999,” Economic Development and Cultural Change, 54 (2), 369–404. Planning Commission (2008) Eleventh Five Year Plan 2007–2012, Volume 1 – Inclusive Growth (New Delhi: Oxford University Press). Rao, C. H. H. (2005) Essays on Development Strategy, Regional Disparities and CentreState Financial Relations in India (New Delhi: Academic Foundation). Rao, M. G. and Chelliah, R. J. (1995) A Survey of Research in Fiscal Federalism in India (New Delhi: Indian Council of Social Science Research). Rao, M. G., Shand, R. T. and Kalirajan, K. P. (1999) “Convergence of Incomes across Indian States: A Divergent View,” Economic and Political Weekly, XXXIV (13), 769–78. Rao, M. G., Sen, T. K. and Jena, P. R. (2008) “Issues before the Thirteenth Finance Commission,” Working Paper 2008–55, National Institute of Public Finance and Policy, New Delhi. Reserve Bank of India (2004 and selected previous issues) State Government Finances (Mumbai: RBI). Sonobe, T. and Otsuka, K. (2006) Cluster-Based Industrial Development: An East Asian Model (New York: Palgrave Macmillan). UNRISD (1968) Research Notes – A Review of Recent and Current Studies Conducted at the Institute, No. 1 (Geneva: United Nations Research Institute for Social Development). Vaidyanathan, A. (1995) The Indian Economy: Crisis, Response, and Prospects (New Delhi: Orient Longman). World Bank (2009) World Development Report 2009: Reshaping Economic Geography (The World Bank Group). XaXa, V. (2001) “Protective Discrimination: Why Scheduled Tribes Lag Behind Scheduled Castes,” Economic and Political weekly, July 21, 2765–72.
9780230_274587_06_cha05.indd 63
6/23/2010 12:16:31 PM
6 The Role of Traditional Chinese State and the Origin of Modern East Asia Debin Ma
6.1
Introduction
Part of Professor Yujiro Hayami’s lifelong research agenda has been to engage the large question of the growth and causal mechanism of modern economic growth in East and Southeast Asia. In a series of highly influential publications, he succinctly outlined the sequence of the East Asian miracle in dimensions of economic theory, empirical analysis, and historical implication (Hayami, 1998, 2001). He characterized the East Asian miracle initiated by Japan since the Meiji as a process of rapid catch-up with Western advanced economies based on massive technology and institutional borrowing. However this capacity to borrow did not come automatically, but was built on a creative adaptive learning process that was in turn preceded and supported by rapid improvements in human-capital formation (Godo and Hayami, 2002). Underlying this process is a more fundamental mechanism as explicated in Professor Hayami’s tripartite framework of market, community, and state. The rise of what he termed agricultural fundamentalism in East Asia underpinned by the prevalence of small and independent (relatively homogeneous) peasant holdings throughout East Asia is the historical linchpin to this framework. As “... rice farming demands care and judgment under diverse and variable natural environments, small farms dependent on family labor requiring little supervision and work enforcement tend to dominate” (Hayami, 1998, p. 27). Professor Hayami’s illuminating reinterpretation turned the old Wittfogel thesis of “oriental despotism” on its head. He pointed out that “in densely populated, highly stable settlements these small ‘peasants’ had to collaborate closely for the control of water ...”, which led to the rise of the irrigation community (Wittfogel, 1976, p. 23). This sense of community prevalent in East Asian agriculture laid the social capital foundation for industrialization, as the internal organization of modern firms also simulates a community to enhance trust and coordination. This community-based social capital coupled with a relatively homogeneous 64
9780230_274587_07_cha06.indd 64
6/23/2010 12:16:37 PM
The Role of Traditional Chinese State 65
population also supplied the ideological base for nationalism in the modern era that in turn led to the rise of a bureaucratic developmental state. Modern states in East Asia managed to maintain a sustained, strong role in the domestic economy with a relatively small government (as measured by its share of total GDP). This is possible precisely because the government operated through implicit agreements with firms through the community nexus such as trade associations (Hayami, 1998, p. 24). Meanwhile, nationalist ideology may have acted as a brake on governmental failures and corruption, which are ever-present. Furthermore, long-term stable employment of bureaucrats, with a structure relatively independent of politics, may be resistant to rent-seeking pressures from vested interest groups (p. 19). A notable contribution in Professor Hayami’s analysis is the utilization of a broad framework that encompasses culture, history, and ideology. My chapter is part of a large project to show the relevance of the political economy of the traditional Chinese state to its long-term growth trajectory. Acknowledging the importance of Professor Hayami’s tripartite framework, my article will focus mostly on the role of the early modern Chinese state and how it could shed light on many of the fundamental features, as identified by Professor Hayami, of modern East Asia. Drawing on the insights of several important recent studies related to institutions and economic growth as championed by Douglas North, Avner Greif, Daron Acemouglou, and others, this article delineates the political logic of the Chinese empire in the framework of three major actors: the emperor, the bureaucracy, and the people, each with their own objective functions and incentive schemes. In the framework of principal-agent and asymmetric information, the historical interplay of these three agents with their respective incentive schemes and structures of private information shape the internal logic of the political structure and the rationale of Confucian ideology. I argue that in the absolutist regime of traditional China, the absence of credible government commitment can be partially alleviated through the rulers’ monopoly of power and a long time-horizon, which led to a virtuous equilibrium of low extraction and the operation of a relatively free private economy. With a long time-horizon of monopoly rule, the rulers’ objective function switched from short-run revenue maximization to the long-term defence of monopoly rents. This traditional political structure generates important historical implications for the patterns of property rights and an incentive structure that not only sheds light on both the longterm growth as well as stagnation of the Chinese economy, but is also relevant for explaining the rapid catch-up of China and East Asia in general in the modern era. I divide the rest of the chapter into three sections. The next section provides a narrative model of the political structure and its historical evolution. The third section applies this model to Qing China (1644–1911), with a comparative perspective on English and Western European states in the
9780230_274587_07_cha06.indd 65
6/23/2010 12:16:37 PM
66
Debin Ma
early modern era, a perspective of relevance for explaining the economic divergence. The concluding section discusses how the legacy of this political regime is helpful for understanding some of the fundamental features of modern East Asia depicted by Professor Hayami as crucial for the rise of the East Asian miracle.
6.2 The historical origin of the Chinese state From the founding of the Chinese empire in Qin (221–206 BC) up until the fall of the last imperial Qing dynasty in 1911, a centralized rule with political power monopolized by a single imperial dynasty supported by a civil bureaucracy, has been indisputably one of her most distinctive characteristics. We start with a description of the theoretical model or, to borrow a terminology from Max Weber, the “ideal type,” of the Chinese empire as an absolutist regime before we proceed to historical reality. In this model of the absolutist regime, ultimate power was vested in the emperor who commanded property rights over all factors of production including land and labor. At the other or lower end of the spectrum are the people or masses (farmers or peasants in an agrarian regime) who are nominally the tenants and cultivators of land and resources owned by the emperor. The imperial household was entitled to rents from agricultural output, the bulk of which went into the supply of external defence and internal security. In this model, the dominance of a single imperial household over all social or political groups is essential. At the founding of the Qin empire, China’s First Emperor (⾺ྟⱛ) followed the advice of his legalist (⊩ᆊ) prime minister, Lishi (ᴢᮃ), and opted against a feudal (ᇕᓎ) type of political arrangement where the imperial power would coexist with various regional elites or aristocrats, often with hereditary status. Instead, the rulers implemented a regime of empire-wide administrative units and household registration. In this new regime, only the status of the imperial throne was hereditary. With the elimination of the aristocracy and self-contained political units, the administration of the empire – tax collection and the coordination and provision of basic public-goods – was governed by direct imperial rules and orders executed by a professional and impersonal bureaucracy. The subsequent rise of a Chinese bureaucracy as the administrative organ of the emperor, or the state, became a distinctively Chinese institutional innovation.15 This concept of the state, as pointed out by China historian Qian Mu, is in many ways an extension of the Chinese concept of a patriarchal household. With the elimination of the hereditary aristocracy, the transition from feudalism to central rule turned the stand-alone imperial household (ᆊ) into the national sovereign (). The literal translation of the Chinese character for nation-state (ᆊ) is really “state-family,” or what Max Weber termed a “familistic state” (Creel, 1964, p. 168). Etymology used by Qian Mu reveals
9780230_274587_07_cha06.indd 66
6/23/2010 12:16:37 PM
The Role of Traditional Chinese State 67
that the equivalent Chinese term of “prime minister” (ᆄⳌ) for the empire derived from titles that denoted managers of private royal households in the pre-Qin period. Thus, for Qian Mu, the rise of central rule also marks the beginning of a separation between ownership (the imperial ruler) and management (the bureaucracy).16 This political structure, to borrow present-day industrial organization terminology, is not dissimilar to large private (family) ownership managed by outsiders. This analogy will be a useful guide to examining the economic efficiency of this political organization in light of the principal-agent theory. We thus have a political model with three actors: the emperor, the bureaucrat, and the masses or peasant farmers. Each of these actors has its own distinctive objective functions, incentive schemes, and information structures. With the emperor claiming a share of peasant farm output collected through the bureaucrats, incentive misalignment among or between the actors would potentially induce the emergence of double principal-agent problems: between the emperor and the bureaucrat, and between the bureaucrat and agent. To understand both the logic and, to certain degree, the remarkable longevity of this model of political autocracy throughout Chinese history, it may be useful to start with Mancur Olson’s benchmark model based on the analogy of stationary and roving banditry. The crux of the argument is that monopoly political rule given a long time-horizon (especially with the throne being hereditary across generations, as in dynasties) could, contrary to what is often believed, lead to a “virtuous” equilibrium of a relatively low level of predation or extraction and a high level of provision of public goods by the rulers (Olson, 1993). The key assumption is that the ruler’s self-interest in maximizing tax revenue provides incentives to support economic growth, leading to a larger tax-base. The longer the time-horizon and the more stable the rule, the more likely it was that the ruler’s interest could align with the interest of the sovereign. Thus under conditions of monopoly rule, a long time-horizon and a low discount rate, the ruler’s high valuation of the stream of future tax income with onetime or short-term extraction constitutes a self-enforcing constraint on the grabbing hands of the autocratic rulers, in the absence of formal constitutional constraint.17 This Chinese model of autocracy differs quite significantly from that of Western Europe. There, even the most absolute and centralized rulers had to contend with local power magnates whether autonomous city states, landed aristocrats, or some type of representative body, not to mention the Church. In this sense, rulers in Western Europe could be defined as a coalition or a class consisting of landed and commercial elites, some of which were organized into corporate bodies or representative assemblies. The stand-alone nature of Chinese rulers was consistent with numerous historical examples of the rulers turning against the landed or commercial elites, as well as against bureaucrats.18
9780230_274587_07_cha06.indd 67
6/23/2010 12:16:37 PM
68
Debin Ma
It was precisely this extremely narrow interest of the Chinese rulers that, as predicted by the Olsonian model, allowed the imperial rulers to transcend the factional interests of elites or bureaucrats and to accommodate the moral claim of the “mandate of heaven,” the much larger encompassing interest of the sovereign. Without having to contend with independent and autonomous powers, the Chinese imperial ruler stood beyond the interests of any ruling class and in fact, if strictly defined, only represented the interests of himself and his lineage. Given the relatively small and unchanging size of the imperial household, the presence of asymmetric information in the bureaucratic structure and the possibility of insurrection, fiscal predation, and revenue maximization were of secondary relevance to the imperial objective: to survive and preserve its monopoly position. There were fundamental difficulties – both theoretical and technical – with a monolithic empire in the premodern era. Externally, it needed to deal with the periodic threat emanating largely from China’s unruly northern frontiers. Internally, the legitimacy of one ruler above all was itself highly contentious, and his rule was susceptible to dynastic struggles and internal rebellion. Chinese history is replete with violent rebellions and dynastic breakdowns led by regional magnates, frontier generals, or entrenched bureaucrats. The formidable information problem posed a fundamental dilemma for a unitary empire with premodern monitoring technology: the larger and more complex an empire grew, the more dependent a ruler would be on local, well-informed agents under conditions of asymmetric information. This dependence would in turn render the ruler vulnerable to the risk of an alternative power base built up locally. While in Europe this dilemma simply created conditions for permanent and entrenched political fragmentation, in China two complementary institutional changes in fiscal regime and civil bureaucracy gradually steered the country onto a path of state centralization. Both institutional changes took shape from the second millennium onward, forming what Wang Yanan claimed were the dual pillars of traditional Chinese polities. They were to have lasting consequences, producing long-term economic change in China. A key advantage of a centralized state is its capacity to control and mobilize all factors of production to achieve state objectives. In this case, China accomplished a precocious transition into a traditional fiscal state which rested on revenue obtained from taxing the economy in its entire territory, as distinguished from the domain states (or demesne states) in Western Europe, which drew income primarily from crown-owned property such as estates, forests, mines, and so on (see He, 2008, chapter 6). To ensure state revenue, Chinese imperial rulers throughout dynasties had actively engaged in the allocation of land to peasants who could in therefore cultivate and contribute taxes. The well-known equal-field system as practiced in Tang (618–907 AD) allocated land to a male adult according to his productive capacity, upon which the state levied the so-called triple tax (⾳ᒌ䇗).
9780230_274587_07_cha06.indd 68
6/23/2010 12:16:37 PM
The Role of Traditional Chinese State 69
Depending on the category of land title, some of the allocated land could be returned back to the state once the cultivator left or died. The equal-field system was aimed at directly taxing individual producers, and preventing the rise of large-scale land-holding and a regional power base. It required comprehensive household and land registration and a strong administrative and coercive governmental power to tie labor down to an immobile factor, the land. In reality, state extractions or bureaucratic lapses often drove the farmers into the tax shelter of powerful regional lords, who, in turn, built up the productive and fighting capacities of their dependents, posing a constant threat to the throne. In this light, the change from the triple-tax system to the dual-tax system (ϸࠊ) around 780 was a landmark transition. The dual-tax system proposed by the then prime minister, Yang Yan (ᴼ♢), shifted the taxation from labor to land irrespective of its ownership. This marked the beginning of a land-based taxation system which was to become the hallmark of Chinese fiscal regime all the way down to the twentieth century. A taxation system targeting the immobile factor of land also enabled, or led, the state to eventually relinquish control and regulation of property rights vested in land, opening the way to recognizing de facto private property rights and the development of a vibrant private land market. Imperial property rights in land and people began to transform into rights to taxation (Qian, 1966, chapter 2; Twitchett et al., 1986). This institutional change in the fiscal regime went hand in hand with another transformation in Chinese bureaucracy. As indicated earlier, Qin’s military success partly depended on his ability to draw loyal talents from below. However, before the Tang dynasty, the system of directly recruiting bureaucrats from the populace remained ad hoc, relying on a combination of informal recommendations and official examinations. By the Tang era, as the civil service examination system developed and became increasingly important, it had gradually fallen under the control of official schools monopolized by elite lineages who were turned into a social class of ruling bureaucrats with hereditary status. The rise of an entrenched and closed bureaucratic class became one of the fundamental destabilizing forces undermining central imperial rule. It was around the eighth century that the system of official schools managed by the closed elite lineages began to decline, gradually opening up the civil service examination system to the general populace beyond the pupils of the official schools. This marked the rise of the second pillar of China’s political structure that survived into the twentieth century: a highly competitive and open civil service examination and an impersonal bureaucracy. This opened up the prospect of upward mobility to ordinary people, and more or less allowed a Chinese style of political participation from the bottom up. From the ruler’s perspective, the creation of a body of career officials having no autonomous territorial or functional power base was
9780230_274587_07_cha06.indd 69
6/23/2010 12:16:37 PM
70 Debin Ma
indispensable to the long-term stability of the empire. With the installation of a system of rotating bureaucrats empire-wide in a period of three to five years, and the rule of avoidance which precluded bureaucrats from serving in their home county, the empire designed an institution that would ensure the bureaucrats’ loyalty to the emperor rather than their constituents. Their promotions and punishments were imposed through an administrative hierarchy guided by detailed administrative codes of conduct. A civil bureaucracy inculcated in Confucian ideology contributed to civilian rule and demilitarization of the empire, and was essential to its sustainability and longevity. Meanwhile the bureaucrats as a class, and the powerful agents of the rulers, were incentivized with special lifelong privileges such as exemption from taxation and favorable status in legal standing over ordinary people.19 In sum, these two pillars of institutional change which originated in the late Tang era and flourished in the Song era, constituted what historians of China have hailed as the Tang-Song Transformation (ᅟব䴽), which has had long-lasting historical consequences. The dual-tax policy allowed the government to retreat from direct management or regulation of property rights in land, and led to the transition from estate farming to a freestanding, family-based, owner-cum-tenant system of agricultural cultivation which was far more incentive-compatible. Another important political consequence is that state-formation in China had decidedly tilted towards stable and centralized rule under unitary imperial rule, with absolute imperial rule reigning supreme from the beginning of Ming in the fourteenth century. The rise of Chinese absolutist central rule, as predicted by this narrative model, also led a slow evolution towards fixing the target of the annual taxation base (ᅮ乱ЏН) (Huang, 1974; Iwai, 2004). The adoption of the so-called “single whip tax” around the late-sixteenth century, which denominated most land taxes in silver terms, further led to the monetization of an imperial tax-base fixed to a silver standard. Technically, setting a target of tax revenue based on a preset household and land registry record greatly simplifies the administrative procedures, requiring less information and monitoring, and thus reducing the potential for abuse. Low tax and light corvée had become the hallmark of benevolent rule in China (ҕᬓ). In fact, historically, any deviation from the policy of fixed revenue was the bad omen of a dynasty in decline or crisis. The policy of fixed revenue allowed the private sector rather than the state to capture or claim all the residuals of economic expansion brought about by rising productivity, growing territory, and increasing population.
6.3
Qing China (1644–1911) in comparative perspective
The last and possibly the most powerful and centralized Chinese dynasty, as I want to argue here, epitomizes the most prominent features of this
9780230_274587_07_cha06.indd 70
6/23/2010 12:16:37 PM
The Role of Traditional Chinese State 71
political economy model in action. The Qing imperial monarchy under the rule of Manchus, from a non-Han Chinese minority in China’s Northeast frontier, became a great defender of orthodox Confucius ideology and a champion of a centralized political system with a civil bureaucracy. The more than two-and-a-half century rule of the Qing saw roughly a tripling of China’s population and a doubling of her territory, and ushered in China’s prosperous eighteenth century, the so-called “Glorious World of Kanxi and Qianlong” (ᒋђⲯϪ) (Maddison, 2007, p. 43; Spence, 1999). The road to the heyday of eighteenth-century prosperity started in 1644, the year of the Qing’s official inauguration. As with all previous rulers who managed to conquer a territory as vast as China, it took about a further two decades for the Qing army to achieve the complete suppression of the former Ming loyalists, mostly based south of the Yangzi river. It was only by 1683, under the reign of emperor Kangxi (1661–1722), that Qing managed to quash the rebellion of these so-called “three feudatories” and annex their territories into his centralized administration. Two years later, after a longterm blockade and military campaign, Kanxi finally broke the resistance of the rebellious naval kingdom of Zheng Chenggong and officially integrated the island of Taiwan as an administrative unit of China. In the final decades of the seventeenth century, Qing’s attention turned to China’s northwestern territory, an area of recurrent threat to previous Chinese dynasties. Qing rulers managed to contain the threat from an expansionary Russia by signing the Treaty of Nerchinsk in 1689, and to conquer China’s Northwestern territory in 1696. From 1720, Qing achieved the control of Tibet with the installation of a new Dalai Lama (Spence, 1999). Clearly, by the early eighteenth century, the Qing had succeeded in the consolidation of power and the establishment of monopoly rule over a territory that had become the largest ever in Chinese history. More importantly, the Qing’s sphere of influence extended beyond its vast territory to other parts of Asia through the socalled tributary order, with states such as Korea, Vietnam, Burma, and other Southeast Asian states recognizing various degrees of Chinese suzerainty. With the closing of Japan under the Tokugawa era (1680–1866), the dominance of Qing throughout Asia had been placed beyond challenge.20 Consolidation of power and monopoly of rule had set out the kind of long time-horizon needed for Qing to move towards the low-taxation equilibrium of the eighteenth century. Our own Western model of political economy lends insight to the confident declaration by Kangxi emperor in 1712 that there would be no additional taxes on a newly added taxable population. Ten years later, the policy was firmly set in place by the fiscal reform measure of consolidating head tax into land tax (᨞ϕܹഄ), the total amount of which would be fixed to the level recorded in the national record of 1710 (Shen, 1996, p. 184). This was clearly the most concrete and dramatic step towards the principle of fixed fiscal revenue. Empirical evidence confirms the largely stationary character of fiscal revenue from the early eighteenth
9780230_274587_07_cha06.indd 71
6/23/2010 12:16:37 PM
72 Debin Ma
century until right before the mid-nineteenth century, which saw the breakout of the devastating Taiping Rebellion. If deflated by the rice price, real expenditure (or taxation) actually declined throughout the seventeenth and eighteenth centuries, and stabilized until the mid-nineteenth century. If we factor in population, real per-capita expenditure and tax burden declined even more sharply from the late-seventeenth century. Converted into kilograms of rice, per-capita Chinese taxation was remarkably modest, amounting to something like two or three days of the wage of an unskilled laborer.21 This seems consistent with Wang Yeh-chien’s more comprehensive calculation of tax revenue (which includes guesstimates for costs of tax collection as well as various extralegal local surcharges) as amounting to merely 2.4 percent of NNP even in the 1910s (p. 133). If Wang’s figure can be relied on for the 1910s, we have good reason to believe that government revenue as a percentage of GDP would have been even lower in the eighteenth century. It is possible that Qing tax rates were the lowest across the dynasties, especially in per-capita terms.22 Overall, land tax predominated in total government revenue, amounting to as much as 74 percent of the total (Wang, 1973, p. 80). Official government data on the expenditure side of the imperial government reported for 1776 reveals that the upkeep of armies took up 78 percent of the total budget, followed by 11.3 percent that went into the payment of bureaucrats. The imperial household or Nei-wu-fu (ࡵݙᑰ) consumption took up a mere 1 percent of the total budget.23 Of course, the major issue with taxation and expenditure in traditional China occurred outside the imperial center. As the traditional Chinese taxation system was highly centralized, most of the collected revenue was accrued or allocated by the imperial fiscal administration: the Board of Revenue, local finance, or fiscal needs at the lowest level of official bureaucracy, figured little in this system. It has long been recognized that the official tax revenue allocated to the local government fell far short of the requirements of normal administration, often being insufficient to cover the salaries of official bureaucrats, let alone their expenses and support staffs. Over time, it had become a tolerated practice for almost all levels of bureaucrats to find informal earnings, in particular the infamous extralegal surcharges (㢯ᤤᴖ), beyond the official level. At the local level, administration needs led the county magistrates to employ a large number of secretaries, clerks, runners, and personal servants. Yet none of these staffs were on the official rolls, being employed personally by the magistrates out of their extralegal extractions. Zelin’s study documents in detail the sources of these additional revenues, including the levying of various surcharges, manipulation of weights and measures, currency conversion in tax collection, falsifying reports, shifting funds across fiscal season years, retaining commercial tax revenue, hoarding tax revenue from newly claimed land, and exacting contributions and
9780230_274587_07_cha06.indd 72
6/23/2010 12:16:38 PM
The Role of Traditional Chinese State 73
donations from local farmers or merchants. Provincial-level officials and their “unofficial” staffs relied on the extraction of gifts and contributions from lower-level officials, and engaged in practises such as skimming funds off at purchase and allocation – buying at a low price but reporting at a high price (Zelin, 1985, p. 46–71). This kind of corruption at the local level is a pale reflection of the much larger networks of collusion at different levels of the state machinery. Because of the prevalence of informal revenue and often imperial acquiescence towards corruption, it is difficult to gauge how these methods of informal taxation could change the picture of fiscal revenue or extraction at the national level. It is important to note that Wang Yieh-chien’s calculation of total tax revenue included his guesstimates of various extralegal surcharges, which amounted to about 24 percent of the final total for 1753.24 Clearly more research is needed in this field, but given Wang’s inclusive figure which represents such a low share of the national income, I don’t believe any reasonable upper estimate would change the conclusion that the overall level of tax extraction remained modest. The real problem of informal taxation may not lie so much in its level of extraction, but rather in its distortionary effects on economic incentives. On the surface, there seemed to be a fundamental contradiction between the Qing imperial ideology of benevolence based on a fixed target of taxation, and their tolerance of, and to certain degree reliance on, informal fiscal extraction at the lower level of society. However, the structure of political economy in the historical context reconciles this seeming contradiction. With tax collection being one of the most important functions of bureaucracy, the imperial policy of fixed revenue meant a fixed number of bureaucratic and administrative units from the imperial perspective. Second, like other agents of the economy, bureaucrats were their own revenue maximizers, who were prone to fiscal expansion if unchecked. Imperial attempts to expand tax bases rarely curbed unofficial extractions on the part of bureaucrats, but rather ended up legitimizing a higher tax target at the expense of the imperial reputation for benevolence. Third, with the hierarchical command structure any “visible” bases of tax revenue at the lower level incurred the risk of extraction from the top. In this sense informal taxation, being outside the official purview, ironically became the most secure source of local finance. In fact, throughout the dynasties, corruption itself was connived at within the system as long as it was not excessive enough to disturb social stability and induce insurrection. Fiscal comparison with Western Europe, in particular with England, contributes a unique perspective to the question of economic divergence between China and Northwestern Europe, which had emerged at least by the early modern era. While revenue accrued to the Board of Revenue (᠌䚼) of Qing China remained largely stagnant or declined slightly in real terms, tax receipts at the Exchequer of Britain rose by a stunning 17-fold
9780230_274587_07_cha06.indd 73
6/23/2010 12:16:38 PM
74 Debin Ma
from 1665 to 1815. The total British revenue as a share of national income, before China’s Glorious Revolution of 1688, surged from slightly more than 3 percent – a rate not hugely different from that of the Qing in the eighteenth to nineteenth centuries – to about 18 percent by 1810 (O’Brien, 1988, p. 3). By the eighteenth century, Britain was the second most heavily taxed nation in the world after Holland (Prak and van Zanden, 2006, p. 130). Again, as distinct from Chinese revenue which had been firmly locked into direct taxes on land, the surge in British tax receipts came disproportionately from indirect taxes such as customs and excise duties, which accounted for nearly 80 percent of total revenue towards the end of the eighteenth century (O’Brien, 1988, p. 9–10). More importantly, as noted by various scholars, fiscal expansion transformed Britain into a modern fiscal state marked by the rise of a Weberian type of professional bureaucracy, which more than tripled in size between the 1690s and the 1780s (Brewer, 1989, p. 66–7). The British development of a civil and fiscal administration that evolved towards professionalism, political neutrality, transparency, and accountability presents a sharp contrast to the stagnation of Chinese bureaucracy, which continued to rely on informal staffs and extralegal revenue sources. The administrative capacity to generate greater revenue from the growing sectors of the economy enabled the British state to capitalize (or mortgage) its future revenue through the financial sector. Thus the growth of a modern fiscal state laid the foundation for the growth of a modern financial market and financial institutions (North and Weingast, 1989). Enhanced state capacity not only strengthened British warfare capabilities and financed overseas explorations in Europe and beyond, but also extended formal state power to the economic sphere.25 By contrast, in China there emerged a reverse incentive problem once the Chinese imperial state switching to a fixed target of tax revenue. The central government no longer had a stake in the growth of the economy, which led to the rise of “laissez-faire” Chinese-style, or what Avner Greif (2005) described as the absent government in some of the private sectors, where growth and expansion in scale increasingly required the involvement of formal state power. It is clear that by the eighteenth century, the bulk of Qing involvement in the private sector was aimed at risk reduction and social stability rather than economic expansion: encouraging the opening up of new land to accommodate a growing population, tax exemption, and famine relief in hard times. In areas not directly related to social stability or government tax revenue, formal imperial power rarely reached: the sectors of commerce, trade, finance and monetary remained largely unregulated and unprotected. Clearly, it was the fear of insurrection rather than the interests of the stakeholders that was the driving force behind the Qing policy. In the absence or weakness of formal state power, semi-official or informal local community organizations in the form of lineage organization,
9780230_274587_07_cha06.indd 74
6/23/2010 12:16:38 PM
The Role of Traditional Chinese State 75
village societies, native place merchant groups, and so on, often took on the role of economic regulation and dispute resolution, as well as the provision of a local granary system, charities, and schooling.26 While these informal organizations helped to provide some basic local public-goods for a largely agrarian society, where production units were small and properly incentivized, they became a problem when the scale of production and marketing grew to require a level of collective action beyond the organizational capacities of private agents or local communities. In short, there was a missing or weak link between the private agents and state power embedded in the incentive structure of the absolute regime of early modern China.
6.4 Concluding discussion: The traditional Chinese state and the origin of modern east Asia The narrative model of the traditional Chinese state in this article argues that institution, particularly the political structure of states, may be a far more important and relevant explanatory factor to account for both China’s long-term economic growth and the great divergence between China and Western Europe in the early modern era. I argue that the fundamental problem of incentive misalignment and asymmetric information in the traditional Chinese state institutions may have hindered the development of institutions which are essential for modern economic growth. Furthermore, with the achievement of monopoly rule in the absence of internal or external political competition, the incentives of the rulers and ruling elites (such as the bureaucratic and gentry class) led them to prioritize the defense of monopoly rents, and become resistant to social and institutional change that could disturb the traditional equilibrium. In this context, political events such as the arrival of Commodore Perry at the Japanese seashore and the outbreak of the Opium War in South China around the mid-nineteenth century, were watershed events that eventually signaled the end of China’s political isolation and exposed the monopoly rulers in East Asia to external military threat and political competition in a new global setting. Meiji Japan first recognized the possibility of complete erosion of monopoly rents previously retained through exclusive political rule, and rose to meet the challenge. With a fundamental restructuring of rulers’ incentives, economic growth, rather than the defence of monopoly rents, then emerged as the new state objective. The bureaucratic developmental state, with political legitimacy based on the growth of per-capita income, became by and large replicated throughout the East Asia in the post-World War Two era, and more recently in China since the late 1970s. It was in the context of the massive diffusion of Western technology and institutions, following the opening-up of these countries, that we rediscovered the powerful role of the institutional legacy associated with the traditional Chinese political structure.
9780230_274587_07_cha06.indd 75
6/23/2010 12:16:38 PM
76 Debin Ma
What is the institutional legacy of the traditional state? How has it contributed to the rapid economic catch-up of East Asia since the mid-nineteenth century? For that, we turn to Professor Hayami’s framework of market, community, and state. Starting with market, this article traces the historical origin of the small and independent peasantry to the institutional transformation which occurred in China’s medieval period. The momentous switch in tax regime from poll tax to land tax around the eighth century gradually but decisively moved the state away from direct management or regulation of property rights vested in land, and allowed the development of de facto property rights to land.27 The resultant policy of fixed revenues also converted the imperial rulers into the equivalent of fixed-rent landlords, in the contractual setting of a landlord-tenant relationship, leaving the private sector rather than the state to capture or claim all the residuals of economic expansion and rising productivity.28 This policy turned each individual peasant household into a highly independent and incentivized producing and market unit, in direct contact with the market, and in the process nurtured a long tradition of commercialization in China and East Asia. Turning to community, we can note that the retreat of the state, as symbolized by the limited reach of the state bureaucratic apparatus at the societal level, also left open space for the growth of some form of bottom-up community-based organization. These ranged from merchant-based lineage association in Southern China and native placed based associations to farming villages. For merchants, these organizations were critical for supplying rules and regulations, standardizing weights and measurements, resolving commercial and legal disputes, enforcing contracts, providing local charities and schooling, and so on. For villages, the farming communities became important in coordinating local irrigation and supplying local public-goods such as bridges and roads. Indeed, one could even argue that the functioning of a centralized state in many ways depended on the working of various communities at the local level to supply local public-goods and provide social and economic regulation. It is important to distinguish these social groups from the formally recognized corporate bodies in Western Europe in that their legal status was often vague and weak, and often lacked the European kind of countervailing power against the state. Nonetheless, they formed the very foundation of a community-based society in East Asia.29 Finally, turning to the state, the “familistic” origin of the traditional Chinese state engendered a powerful sense of a community with shared interest and common destiny. As an extension of the family, the traditional Chinese state also simulated a community, with ties that went beyond blood relations. More importantly, within this political structure the dominance of an autocracy in a unified empire gave birth to an enduring bureaucratic tradition of technocracy and impersonal rule. These institutional endowments paved the way for the rise of a modern nationalist ideology and a bureaucratic developmental state for modern East Asia.
9780230_274587_07_cha06.indd 76
6/23/2010 12:16:38 PM
The Role of Traditional Chinese State 77
Another possibly more important though unintended consequence of the traditional bureaucracy is its impact on average levels of human capital. Access to the privileged bureaucratic ruling class, or upward social mobility, was acquired through the passing of a relatively objective and standard examination system, which generated an incentive system peculiar to a traditional society such as China: a very high (often irrationally so) esteem was now attached to education and literacy in the wider society. Despite the possible welfare loss from the often wasteful efforts thrown into examination preparation, the civil service examination system bestowed a historical tradition highly conducive to human-capital accumulation in the modern era. Recent works by Baten et al. (2009) demonstrate quantitatively that, despite substantially lower per-capita income, level of literacy and numeracy in China (possibly in East Asia in general) may have come close to that attained by advanced Northwestern European countries during the early modern era before the onset of the Industrial Revolution. This finding offers a historical antecedent to the rapid build-up of human capital in modern East Asia in the nineteenth and twentieth century, as captured by the findings of Godo and Hayami (2002).
References In English Baten, J., Ma, D., Morgan, S. and Wang, Q. (2009) “Evolution of Living Standards and Human Capital in China in the Eighteenth to the Twentieth Century: Evidences from Real Wage and Anthropometrics,” Working Paper No.122/09, Department of Economic History, London School of Economics, London. Bogart, D. and Richardson, G. (2008) “Parliament, Property Rights and Public Goods in England: 1600 to 1815,” Working paper, Department of Economics, University of California, Irvine. Brewer, J. (1989) The Sinews of Power: War, Money and the English State, 1688–1783 (London: Unwin Hyman). Chang, C. L. (1955) The Chinese Gentry: Studies on their Role in Nineteenth-century Chinese Society (Seattle, WA: University of Washington Press). Chang, T. A. (1972) “The Economic Role of the Imperial Household in the Ch’ing Dynasty,” Journal of Asian Studies, 31 (2), February, 243–73. Ch’u, T. T. (1962) Local Government in China under the Ching (Cambridge: Harvard University Press). Creel, H. G. (1964) “The Beginnings of Bureaucracy in China: The Origin of the Hsien,” The Journal of Asian Studies, 23 (2), February, 155–84. Godo, Y. and Hayami, Y. (2002) “Catching Up in Education in the Economic Catch-up of Japan with the United States, 1890–1990,” Economic Development and Cultural Change, 50 (4), 961–78. Greif, A. (2005) “Commitment, Coercion, and Markets: The Nature and Dynamics of Institutions Supporting Exchange,” in C. Menard and M. M. Shirley (eds.), Handbook for New Institutional Economics, Chapter 28 (Norwell, MA: Kluwer Academic Publishers).
9780230_274587_07_cha06.indd 77
6/23/2010 12:16:38 PM
78
Debin Ma
Hayami, Y. (1998) “Towards an East Asian Model of Economic Development,” in Yujiro Hayami and Masahiko Aoki (eds.), The Institutional Foundations of East Asian Economic Development (Basingstoke: Macmillan), 3–36. Hayami, Y. (2001) Development Economics, 2nd edition (Oxford: Oxford University Press). Hayami, Y., Quisumbing, M.A.R. and Adriano, L. S. (1990) Towards an Alternative Land Reform Paradigm: A Philippines Perspective (Manila: Ateneo de Manila University Press). He, W. (2008) Paths Toward a Modern Fiscal State: England (1642–1750), Japan (1868–1895), and China (1850–1911) Ph.d Thesis (Cambridge: MIT Press). Huang, R. (1974) Taxation and Governmental Finance in Sixteenth-Century Ming China (Cambridge Studies in Chinese History, Literature and Institutions) (London: Cambridge University Press). Liu, G. (2005) Wrestling for Power: The State and the Economy in Later Imperial China, 1000–1770, Phd. Thesis (Harvard University Press). Ma, D. (2004) “Growth, Institutions and Knowledge: A Review and Reflection on the Historiography of Eighteenth- to Twentieth-century China,” Australian Economic History Review, 44 (3) (Special Issue on the Economic History of Asia), 259–77. Ma, D. (2009a) “Incentives and Information: An Institutional Perspective on the Traditional Chinese State and Great Divergence in the Early Modern Era,” Unpublished working paper, LSE Economic History Dept. Ma, D. (2009b) “Law and Economic Change in Traditional China: A Comparative Perspective,” LSE working paper No. 124, 2009 at http://www.lse.ac.uk/collections/ economicHistory/pdf/WP124.pdf. Maddison, A. (2007) Chinese Economic Performance in the Long Run, 2nd edition, revised and updated: 960–2030 AD, ed. Development Centre Studies (Paris: Development Centre of the Organisation for Economic Cooperation and Development). North, D. C. (1981) Structure and Change in Economic History (New York and London: Norton). North, D. C. and Weingast, B. R. (1989) “Constitutions and Commitment: The Evolution of Institutions Governing Public Choice in Seventeenth-Century England,” Journal of Economic History, 49, December, 803–32. O’Brien, K. J. (1988) “China’s National People’s Congress: Reform and Its Limits,” Legislative Studies Quarterly, 13 (3), 343–74. Olson, M. (1993) “Dictatorship, Democracy, and Development,” American Political Science Review, 87 (3), September, 567–76. Otsuka, K., Chuma, H. and Hayami, Y. (1992) “Land and Labor Contracts in Agrarian Economies: Theories and Facts,” Journal of Economic Literature, XXX, December, 1965–2018. Prak, M. and van Zanden, J. L. (2006) “Towards an Economic Interpretation of Citizenship: The Dutch Republic between Medieval Communes and Modern Nation-States,” European Review of Economic History, 10, 111–45. Spence, J. D.(1999) The Search for Modern China, 2nd edition (New York and London: Norton). Twitchett, D. C., Loewe, M. and Fairbank, J. K. (1986) The Cambridge History of China, Volume1, The Ch’in and Han Empires, 221 BC–AD 220 (Cambridge: Cambridge University Press). Wang, Y.-C. (1973) Land Taxation in Imperial China, 1750–1911, Harvard East Asian Series 73 (Cambridge, MA: Harvard University Press). Wittfogel, K. A. (1976) Oriental Despotism: A Comparative Study of Total Power (New York: Yale University Press).
9780230_274587_07_cha06.indd 78
6/23/2010 12:16:38 PM
The Role of Traditional Chinese State 79 Zelin, M. (1985) The Magistrate’s Tael: Rationalizing Fiscal Reform in Eighteenth-Century Ch’ing China (Berkeley, CA and London: University of California Press).
In Chinese Qian Mu (䪅〚) (1966) Ёग़ҷᬓ⊏ᕫ༅ (The Gains and Losses of Chinese Dynastic Politics) (佭␃: Ёॄࠋॖ). Shi, Z. (ᖫᅣ) (2008) ⏙ҷ᠌䚼䫊ᑧᬊᬃᑧᄬ㒳䅵 (Statistics on the Stocks, Receipts, and Expenditures of the Vaults of the Qing Board of Revenue) (⽣ᓎҎ⇥ߎ⠜⼒). Shen, X. (ᄿ㖞߮) (1996) Џ㓪 Ё䊺 (A History of Chinese Taxation) (Ё䋶ᬓ㒣⌢ ߎ⠜⼒). Wang, Y. (⥟Ѯफ) (2005) Ёᅬڮᬓ⊏ⷨお (The Politics of Chinese Bureaucracy) (Ё⼒ Ӯ⾥ᄺߎ⠜⼒). Xu, M. M. (ᕤ㣖ᯢ) (2004) ∳फ㒙Ϣ∳फ⼒Ӯ: 1368–1911 ᑈ(Jiangnan Gentry and Jiangnan Society: 1368–1911) (࣫Ҁ: ଚࡵॄк佚).
In Japanese Iwai, S. (ችѩ㣖) (2004) Ё䖥Ϫ䉵ᬓȃⷨお (A Study on the Early Modern Fiscal History of China) (Kyoto: Kyoto University Press).
9780230_274587_07_cha06.indd 79
6/23/2010 12:16:38 PM
7 Government Distortions of Agricultural Prices: Lessons from Rich and Emerging Economies* Kym Anderson
7.1
Introduction
For many decades agricultural protection and subsidies in high-income (and some middle-income) countries have been depressing international prices of farm products, which lowers the earnings of farmers and associated rural businesses in developing countries. That worsened between the 1950s and the early 1980s (Anderson et al., 1986), thereby adding to global inequality and poverty because three-quarters of the world’s poorest people depend directly or indirectly on agriculture for their main income (World Bank, 2007). In addition to that external policy influence on rural poverty, however, the governments of many developing countries have directly taxed their farmers over the past half-century. A well-known example is the taxing of exports of plantation crops in postcolonial Africa (Bates, 1981). At the same time, many developing countries chose also to pursue an import-substituting industrialization strategy, predominantly by restricting imports of manufactures, and to overvalue their currency. Together those measures indirectly taxed producers of other tradable products in developing economies, by far the majority of them being farmers (Krueger et al., 1988, 1991).
* Financial assistance from World Bank Trust Funds (particularly those provided by the governments of Japan, the Netherlands and the United Kingdom), GRIPS and the Australian Research Council is gratefully acknowledged, as are the contributions of the country case study contributors to the Agricultural Distortions project, computational assistance by a team of research assistants led by Ernesto Valenzuela, and helpful comments from festschrift discussants and also commentators on numerous related conference and seminar papers over the past year. Views expressed are the author’s alone and not necessarily those of the World Bank or its Executive Directors. 80
9780230_274587_08_cha07.indd 80
6/23/2010 12:16:47 PM
Government Distortions of Agricultural Prices
81
This disarray in world agriculture, as Johnson (1973) described it in the title of his seminal book, means there has been overproduction of farm products in high-income countries and underproduction in more-needy developing countries. It also means there has been less international trade in farm products than would be the case under free trade, thereby thinning markets for these weather-dependent products and thus making these markets more volatile. Using a stochastic model of world food markets, Tyers and Anderson (1992, Table 6.14) found that instability of international food prices in the early 1980s was three times greater than it would have been under free trade in those products. During the past quarter-century, however, numerous countries have begun to reform their agricultural price and trade policies. To get a sense of how much that has reduced the distortions to global markets for farm products, the present chapter draws on the results of a recent World Bank multicountry study of policies affecting agricultural price incentives over the past 5 decades. That study includes 75 countries that together account for 92 percent of the world’s population and agricultural GDP, and 95 percent of total GDP. Specifically, this chapter summarizes estimates of the nominal rates of assistance (NRAs) for more than 70 different farm products, with an average of almost a dozen per country. Not all countries had data for the entire 1955–2007 period, but the average number of years covered is 41 per country. Having such a comprehensive coverage of countries, products and years offers the prospect of obtaining a reliable picture of long-term trends in price-distorting policies (as well as annual fluctuations around those trends, not reported here) for country groups, regions, and the world as a whole. The chapter begins with a summary of the methodology used to generate annual indicators of the extent of government interventions in markets, details of which are provided in Anderson et al. (2008). The NRA estimates are then summarized across regions and over the half-decades since the mid-1950s. A summary is also provided of an additional set of indicators of agricultural price distortions that are based on the trade restrictiveness index first developed by Anderson and Neary (2005) and modified for the Bank’s research project by Lloyd et al. (2009). Then a new set of results from a global economy-wide model provides quantification of the impacts on global agricultural trade of the reforms since the early 1980s and of the policies still in place as of 2004. The chapter concludes by drawing on the lessons learned to speculate on the prospects for further reform, while considering the political economy of agricultural protection backed by agricultural communities, in the sense of Hayami and Godo (2005) as well as the earlier study by Anderson et al. (1986).
9780230_274587_08_cha07.indd 81
6/23/2010 12:16:47 PM
82 Kym Anderson
7.2
Methodology
Government-imposed distortions can create a gap between domestic prices and what they would be under free markets. The Nominal Rate of Assistance (NRA) for each farm product is computed as the percentage by which government policies have raised gross returns to farmers above what they would be without the government’s intervention (or lowered them, if NRA <0). A weighted average NRA for all covered products is derived using the value of production at undistorted prices as weights. To that NRA for covered products is added a “guesstimate” of the NRA for noncovered products (on average around 30 percent of the total value of farm production) and an estimate of the NRA from nonproduct-specific forms of assistance or taxation. Since the 1980s governments of some high-income countries have also provided so-called “decoupled” assistance to farmers but, because that support in principle does not distort resource allocation, its NRA has been computed separately and is not included for direct comparison with the NRAs for other sectors or for developing countries. Each farm industry is classified either as importcompeting, or as a producer of exportables, or as producing a nontradable (with its status sometimes changing over the years), so as to generate for each year the weighted average NRAs for the two different groups of covered tradable farm products. Also computed is a production-weighted average NRA for nonagricultural tradables, for comparison with that for agricultural tradables via the calculation of a percentage relative rate of assistance (RRA), defined as: RRA 100 * [ (100 NRAag t) / (100 NRAnon-ag t) 1] where NRAag t and NRAnon-ag t are the percentage NRAs for the tradables parts of the agricultural (including noncovered) and nonagricultural sectors, respectively.30 Since the NRA cannot be less than –100 percent if producers are to earn anything, neither can the RRA (since the weighted average NRAnon-ag t is nonnegative in all our country case studies). And if both of those sectors are equally assisted, the RRA is zero. This measure is useful in that if it is below (above) zero, it provides an internationally comparable indication of the extent to which a country’s sectoral policy regime has an anti-(pro-)agricultural bias. The cost of government policy distortions to incentives in terms of resource misallocation tends to be greater, the greater the degree of substitution in production. In the case of agriculture which involves the use of farm land that is sector-specific but transferable among farm activities, the greater the variation of NRAs across industries within the sector then the higher will be the welfare cost of those market interventions. Thus, it is helpful to have a single indicator of the overall welfare effect of each
9780230_274587_08_cha07.indd 82
6/23/2010 12:16:47 PM
Government Distortions of Agricultural Prices
83
country’s regime of agricultural price distortions in place at any time, and to trace its path over time and make cross-country comparisons. To that end, the family of indexes first developed by Anderson and Neary (2005), under the catch-all name of trade restrictiveness indexes, are drawn upon. To generate indicators of distortions imposed by each country’s border and domestic agricultural policies on its economic welfare and its agricultural trade volume, Lloyd et al. (2009) define a Welfare Reduction Index (WRI) and a Trade Reduction Index (TRI) and estimate them for the same focus countries. As their names suggest, these two new indexes respectively each capture in a single indicator the direct welfare- or trade-reducing effects of distortions to consumer and producer prices of covered farm products from all agricultural and food price and trade policy measures in place (while ignoring noncovered farm products and indirect effects of sectoral and trade policy measures directed at nonagricultural sectors). Specifically, the TRI (or WRI) is that ad valorem trade tax rate which, if applied uniformly to all farm commodities in a country that year would generate the same reduction in trade (or economic welfare) as the actual cross-commodity structure of agricultural NRAs for that country, other factors being equal. The WRI measure reflects the partial equilibrium welfare cost of agricultural price-distorting policies better than the NRA because it recognizes that the welfare cost of a government-imposed price distortion is related to the square of the price wedge. It thus captures the disproportionately higher welfare costs of peak levels of assistance or taxation, and is larger than the mean NRA and is positive regardless of whether the government’s agricultural policy is favoring or hurting farmers. In this way the WRI and TRI go somewhat closer to what a computable general equilibrium (CGE) can provide in the way of estimates of the trade and welfare (and other) effects of the price distortions captured by the product NRA estimates: while not capturing the indirect distortions from other sectors as the RRA does, these indexes have the advantage over a CGE model of providing an annual time series and not requiring a formal model.
7.3 Estimates of the changing extent of agricultural price distortions This section first presents aggregate results for the world as a whole, and then provides more details of the results for Asia in particular, where the evolution of price distortions has been more dramatic than in any other region. Global findings The global summary of the new results from the World Bank project is provided in Figure 7.1. It reveals that the nominal rate of assistance to farmers in
9780230_274587_08_cha07.indd 83
6/23/2010 12:16:47 PM
84 Kym Anderson 70 60 50
Percent
40 30 20 10 2000–04
1995–99
1990–94
1985–89
1980–84
1975–79
1970–74
1965–69
−20
1960–64
−10
1955–59
0
−30 HIC & ECA
HIC and ECA, incl. decoupled payments
Developing countries Figure 7.1 Nominal rates of assistance to agriculture in High-Income Countries (HIC) and European transition economiesa and in developing countries, 1955 to 2004 (percent, weighted averages, with “decoupled” payments included in the dashed HIC line) Note: a Denoted by the World Bank as ECA, for (Central and Eastern) Europe and Central Asia. Source: Anderson (2009, Ch. 1), based on estimates in Anderson and Valenzuela (2008).
high-income countries rose steadily from the mid-1950s until the end of the 1980s, apart from a small dip when international food prices spiked around 1973–4. After peaking at more than 50 percent in the mid-1980s, that average NRA for high-income countries has fallen a little, depending on the extent to which one believes that some new farm programs are “decoupled” in the sense of no longer influencing production decisions (see dashed line in Figure 7.1). For developing countries, too, the average NRA for agriculture has been rising, but from a level of around –25 percent during the period from the mid-1950s to the early 1980s, to nearly 10 percent in the first half of the present decade. The average NRA for developing countries conceals the fact that the exporting and import-competing subsectors of agriculture have very different NRAs. Figure 7.2 reveals that while the average NRA for exporters has been negative throughout (going from –20 percent to –30 percent before coming back up to almost zero in 2000–4), the NRA for import-competing farmers in developing countries has fluctuated between 20 and 30 percent (and even reached 40 percent in the years of low prices in the mid1980s). Having increased in the 1960s and 1970s, the antitrade bias within
9780230_274587_08_cha07.indd 84
6/23/2010 12:16:47 PM
Government Distortions of Agricultural Prices
85
agriculture (the taxing of both exports and imports of farm products) for developing countries has diminished since the mid-1980s, but the NRA gap between the import-competing and export subsectors still averages around 20 percentage points. Figure 7.2 also reveals that the NRA for import-competing farmers in developing countries has increased at virtually the same pace as that in high-income countries. This suggests that growth in agricultural protection
(a) Developing countries 90 70 50 30
2000–4
1995–9
1990–4
1985–9
1980–4
1975–9
−30
1970–4
−10
1965–9
10
−50 Import-competing
Exportables
Total
(b) High-income countries plus Europe’s transition economies 90 70 50 30
2000–4
1995–9
1990–4
1985–9
1980–4
1975–9
1970–4
1965–9
−30
1960–4
−10
1955–9
10
−50 Import-competing
Exportables
Total
Figure 7.2 Nominal rates of assistance to exportable, import-competing and all covered agricultural products,a high-income and developing countries, 1955 to 2004 (percent) Note: a Covered products only. The total also includes nontradables. Source: Anderson (2009, Ch. 1), based on estimates in Anderson and Valenzuela (2008).
9780230_274587_08_cha07.indd 85
6/23/2010 12:16:47 PM
86
Kym Anderson
from import competition is something that begins at low levels of per-capita income rather than being a phenomenon exclusive to high-income countries. The improvement in farmers’ incentives in developing countries is understated by the above NRA estimates, because those countries have also reduced their assistance to producers of nonagricultural tradable goods, most notably manufactures. The decline in the weighted average NRA for the latter, depicted in Figure 7.3, was clearly much greater than the increase in the average NRA for tradable agricultural sectors for the period to the mid-1980s, consistent with the finding two decades ago of Krueger et al. (1988, 1991). For the period since the mid-1980s, changes in the NRAs of both sectors have contributed almost equally to the improvement in incentives to farmers. The RRA, defined in the previous section, provides a useful indicator of relative price change: the RRA for developing countries as a group went from –46 percent in the second half of the 1970s to 1 percent in the first half of the present decade. This increase (from a coefficient of 0.54 to 1.01) is equivalent to an almost doubling in the relative price of farm products, which is a huge change in the fortunes of developing-country farmers in just a generation. This is mostly because of the changes in Asia, but even for Latin America this relative price hike is one-half, while for Africa this indicator improves by only oneeighth. As for high-income countries, assistance to manufacturing was on average much less than assistance to farmers, even in the 1950s, and its decline since then has had only a minor impact on that group’s average RRA (Figure 7.3). Turning to the single indicators of the impact of agricultural distortions on national economic welfare and trade volume, Lloyd et al. (2009) estimate their TRI and WRI for the 75 countries in the above-mentioned World Bank study. The TRI estimates indicate that the trade-reducing impact of agricultural policies for developing countries as a group was roughly constant until the early 1990s and thereafter it declined, while for high-income countries the decline in TRI began a few years later (Figure 7.4(a)). The TRI for developing countries is driven by the exportables subsector which was being taxed until recently and the import-competing subsector which was, and is increasingly, being protected (albeit less than in high-income countries – see Figure 7.2 above). For high-income countries, policies have supported both exporting and import-competing agricultural products and, even though they strongly favor the latter, the assistance to exporters has offset somewhat the antitrade bias from the protection of import-competing producers. The WRI estimates, shown in Figure 7.4(b), indicate a steady rise from the 1970s to the 1980s for agricultural policies, but some decline in the 1990s. This reflects the fact that NRAs for high-income and developing countries diverged (in opposite directions) away from zero in the first half of the period under study and then converged toward zero in the most recent
9780230_274587_08_cha07.indd 86
6/23/2010 12:16:48 PM
Government Distortions of Agricultural Prices
87
(a) Developing countries 80 60
20
2000–4
1995–9
1990–4
1985–9
1980–4
−40
1975–9
−20
1970–4
0 1965–9
Percent
40
−60 RRA
NRA agric. tradables
NRA non-ag. tradables
(b) High-income countries 100 80
40 20 2000–4
1995–9
1990–4
1985–9
1980–4
1975–9
1970–4
−40
1965–9
−20
1960–4
0 1955–9
Percent
60
−60 RRA
NRA agric. tradables
NRA non-ag. tradables
Figure 7.3 Nominal rates of assistance to agricultural and nonagricultural tradable sectors and relative rate of assistance,a developing and high-income countries, 1955 to 2004 (percent, production-weighted averages across countries) Note: a The RRA is defined as 100 * [(100 NRAag t) / (100 NRAnon-ag t) 1], where NRAag t and NRAnon-ag t are the percentage NRAs for the tradables parts of the agricultural and nonagricultural sectors, respectively. Source: Anderson (2009, Ch. 1), based on estimates in Anderson and Valenzuela (2008).
quarter-century. That meant that their weighted average NRA traces out a fairly flat trend, whereas the WRI traces out a hill-shaped path and thus provides a less misleading indicator of the trend in resource misallocation in world agricultural markets.
9780230_274587_08_cha07.indd 87
6/23/2010 12:16:48 PM
88 Kym Anderson (a) Trade reduction index 60
40
20
Developing countries
Europe’s transition econs.
2005–7
2000–4
1995–9
1990–4
1985–9
1980–4
1975–9
1970–4
−20
1965–9
1960–4
0
High-income countries
(b) Welfare reduction index 80
60
40
20
Developing countries
Europe’s transition econs.
2005–7
2000–4
1995–9
1990–4
1985–9
1980–4
1975–9
1970–4
1965–9
1960–4
0
High-income countries
Figure 7.4 Trade reduction and welfare reduction indexes for tradable farm products, by region, 1960 to 2007 (percent) Source: Lloyd et al. (2009), based on NRAs and CTEs in Anderson and Valenzuela (2008).
Findings for Asia32 From the mid-1950s to the early 1980s, agricultural price and trade policies reduced earnings of farmers in developing Asia on average by more than 20 percent; but that implicit taxation declined from the early 1980s and, from the mid-1990s, the NRA switched sign and became increasingly
9780230_274587_08_cha07.indd 88
6/23/2010 12:16:49 PM
Government Distortions of Agricultural Prices
89
positive. That average hides considerable diversity within the region, however. Nominal assistance to farmers in Korea and Taiwan was positive from the early 1960s (although very small initially when compared with the 40+ percent in Japan), Indonesia had some years in the 1970s and 1980s when its NRA was a little above zero (as did Pakistan prior to Bangladesh becoming an independent country in 1971), and India’s and the Philippines’ average NRAs became positive from the 1980s (Table 7.1).33 This trend is present for the vast majority of the individual commodity NRAs for the region too, with meat and milk the only products to have seen their assistance rates cut over that period. As is true for other regions of the world, assistance is among the highest for the “rice pudding” products of sugar, milk and rice (Table 7.2). But even for those three products there is a great diversity across countries in their NRAs, with 5-year averages ranging from almost zero to as much as 400 percent for rice and 140 percent for milk in Korea, and to 230 percent for sugar in Bangladesh. There is a great deal of NRA diversity also across commodities within each Asian economy’s farm sector, and the extent (as measured by the standard deviation) has grown rather that diminished over the past 5 decades, from a regional average of less than 40 percent in the early years under study to more than 55 percent in recent years. This suggests there is still much that could be gained from improved resource reallocation both between Asian economies and within the agricultural sector of individual Asian economies, were differences in rates of assistance to be reduced. That possibility of trade and welfare gains from further reform is underscored by the estimates of WRIs and TRIs for Asian countries, which are reported in Tables 7.3 and 7.4. For Japan and Korea, their TRIs are very similar to their high NRAs (cf. Table 7.1), since all major farm products are importable and most are highly protected. Taiwan’s TRI until recently was negative, reflecting the fact that its rice producers were assisted even when the island was a significant exporter. China’s TRI was always positive and quite high in the 1980s and 1990s, because of the strong implicit taxes on both exports and imports of farm products. India’s TRI was even higher, peaking in the latter 1980s and still high compared with China’s and those for other South Asian countries. The TRIs for Southeast Asia are generally smaller, but vary considerably across countries and over time. The WRIs are necessarily positive and generally much higher than the NRAs. For China and India they have become considerably smaller over the past two decades, but they have declined little in such countries as Indonesia, the Philippines, and Sri Lanka (Table 7.4), reflecting the fact that a wide range of NRAs still prevail in those countries. The anti-agricultural policy biases of the past were due not just to agricultural policies. Also important to changes in incentives affecting intersectorally mobile resources have been the significant reductions in border protection to the manufacturing sector (which has been the dominant
9780230_274587_08_cha07.indd 89
6/23/2010 12:16:50 PM
90 Kym Anderson
Table 7.1 Nominal rates of assistance to agricultural and nonagricultural tradables and relative rate of assistance,a Asian developing economies and Japan, 1960 to 2004 (percent) 1960–4 1970–4 1980–4 1985–9 1990–4 1995–9 2000–4 Japan NRA Ag. NRA Non-Ag. RRA
44.5 3.9 39.1
47.3 2.8 43.3
67.0 1.1 65.2
127.7 1.3 124.8
129.7 1.1 127.1
133.4 0.8 131.4
133.6 0.7 132.1
Korea NRA Ag. NRA Non-Ag. RRA
4.9 37.1 –21.4
46.1 11.4 30.5
118.6 6.8 104.8
159.8 5.7 145.9
197.6 3.3 188.2
164.8 2.3 158.8
171.9 1.7 167.3
Taiwanb NRA Ag. NRA Non-Ag. RRA
4.7 9.3 –4.2
12.0 7.5 4.2
18.7 5.2 12.9
33.8 4.5 28.0
46.3 2.6 42.5
54.9 1.8 52.2
70.9 1.0 69.0
Chinab NRA Ag. NRA Non-Ag. RRA
–45.2 41.6 –60.5
–45.2 41.6 –60.5
–45.2 41.6 –60.5
–35.5 28.3 –49.9
–14.3 24.9 –31.1
6.6 9.9 –3.0
5.9 5.0 0.9
Indonesia NRA Ag. NRA Non-Ag. RRA
na na na
–3.8 27.7 –24.7
10.5 27.7 –13.5
–1.9 26.5 –22.5
–7.5 17.6 –21.3
–9.7 10.6 –18.3
13.9 8.1 5.4
Malaysia NRA Ag. NRA Non-Ag. RRA
–7.6 7.4 –14.0
–9.4 7.1 –15.5
–4.9 5.2 –9.6
1.4 3.9 –2.4
2.6 2.8 –0.3
–0.2 2.0 –2.2
1.5 0.9 0.6
Philippines NRA Ag. NRA Non-Ag. RRA
–1.7 19.0 –17.4
–6.0 16.3 –19.8
–4.0 12.9 –14.9
15.8 11.0 4.3
16.7 9.9 6.1
35.7 8.6 24.9
23.5 6.4 15.9
Thailand NRA Ag. NRA Non-Ag. RRA
na na na
–23.1 16.1 –33.7
–2.3 14.2 –14.4
–6.9 11.1 –16.3
–6.4 10.0 –14.9
1.8 8.9 –6.5
–0.2 7.8 –7.4
Vietnamb NRA Ag. NRA Non-Ag. RRA
na na na
na na na
na na na
–15.9 4.3 –19.2
–26.4 –11.2 –17.4
0.0 1.5 –1.3
20.7 20.8 0.0
Bangladesh NRA Ag. NRA Non-Ag. RRA
na na na
na na na
–3.9 22.4 –21.5
17.5 28.5 –8.6
–2.4 33.3 –26.7
–8.0 29.0 –28.6
4.0 23.4 –15.8
Continued
9780230_274587_08_cha07.indd 90
6/23/2010 12:16:50 PM
Government Distortions of Agricultural Prices
Table 7.1
91
Continued 1960–4 1970–4 1980–4 1985–9 1990–4 1995–9 2000–4
Indiab NRA Ag. NRA Non-Ag. RRA
5.2 113.0 –56.3
12.6 83.1 –38.3
4.1 59.3 –33.5
67.5 48.6 11.7
2.0 15.9 –12.1
–2.3 12.6 –12.9
15.4 5.2 12.5
Pakistanb NRA Ag. NRA Non-Ag. RRA
–1.0 169.7 –63.8
9.3 146.7 –55.9
–9.3 48.3 –38.6
–5.9 45.1 –35.1
–10.2 39.3 –35.2
–2.6 27.0 –23.0
1.5 14.6 –11.5
Sri Lanka NRA Ag. NRA Non-Ag. RRA
–25.7 124.6 –66.6
–18.5 70.7 –51.6
–15.4 57.1 –46.2
–11.2 59.0 –44.3
–1.3 47.1 –32.9
14.0 36.4 –16.3
10.8 22.9 –9.8
Asian developing economiesc NRA Ag. –27.7 –24.3 NRA Non-Ag. 67.1 50.3 RRA –56.4 –47.9 Dispersion of 30.7 37.6 national RRAsd
–18.8 38.3 –40.8 51.9
–11.2 15.4 –22.8 56.0
–2.6 14.9 –15.2 65.1
7.5 9.6 –1.9 50.5
11.7 4.3 7.1 50.8
Notes: a The RRA is defined as 100* [(100 NRAag t) / (100 NRAnon-ag t) 1], where NRAag t and NRAnon-ag t are the percentage NRAs for the tradables parts of the agricultural and nonagricultural sectors, respectively. b Estimates for China pre-1981 and India pre-1965 are based on the assumption that the nominal rates of assistance to agriculture in those years was the same as the average NRA estimates for those economies for 1981–4 and 1965–9, respectively, and that the gross value of production in those missing years is that which gives the same average share of value of production in total world production in 1981–4 and 1965–9, respectively. This NRA assumption is conservative in the sense that for both countries the average NRA was probably even lower in earlier years, according to the authors of those country case studies. c Weighted averages of the above national averages, using weights based on gross value of national agricultural production at undistorted prices. d Simple average of the standard deviation around a weighted mean of the national RRAs for the region each year. Source: Calculated from Anderson and Valenzuela (2008), which draws on national estimates reported in Anderson and Martin (2009).
intervention in the tradables part of nonagricultural sectors). That reduction in assistance to producers of nonfarm tradables has been even more responsible for the improvement in farmer incentives than the reduction in direct taxation of agricultural industries. For Asia as a whole, the average NRA estimates for nonfarm tradables declined steadily throughout the past four or five decades as policy reforms spread. This contributed to a decline in the estimated negative relative rate of assistance for farmers: the weighted average RRA was worse than –50 percent up to the early 1970s,
9780230_274587_08_cha07.indd 91
6/23/2010 12:16:51 PM
92
Kym Anderson
Table 7.2 NRAs for Asian developing-country farmers, by product, 1960 to 2004 (percent, at primary product level) 1960–4 1970–4 1980–4 1985–9 1990–4 1995–9 2000–4 Barley Beef Cassava Chickpea Cocoa Coconut Coffee Cotton Egg Fruits and vegetables Jute Maize Milk Oilseeds Palm oil Pig meat Poultry Rice Rubber Sorghum Sugar Tea Wheat Asian dev. economies:a
84 25 na 50 na –29 na –19 –21
120 44 –23 1 –3 –8 –7 63 0
166 101 –9 8 –2 –11 –9 –12 10
357 94 –17 12 –1 –19 –5 –2 22
524 145 –11 9 –2 –34 –5 –3 27
543 106 –14 15 –2 –22 –1 0 23
563 85 –10 19 0 –8 –2 5 51
0 na –10 na 24 –11 16 0 –6 –16 82 96 –39 –12
0 –30 19 122 11 –15 51 18 –17 –8 55 13 –28 15
–8 –29 –20 108 22 –1 –41 48 –27 –19 7 37 –18 –3
–3 –35 –6 124 35 –2 –39 –2 –6 –14 36 39 –19 12
–11 –38 –15 40 21 2 –3 20 –9 –16 7 13 –10 4
–6 –6 8 23 22 –9 7 17 2 5 21 20 –8 18
–4 –39 13 32 22 –3 4 12 18 4 16 43 –7 11
–21.6
–15.2
–4.8
6.0
10.2
–2.7
0.3
Note: a Weighted using as weights production valued at undistorted prices. Source: Calculated from Anderson and Valenzuela (2008), which draws on national estimates reported in Anderson and Martin (2009).
but it improved to an average of -32 percent in the 1980s, -9 percent in the 1990s, and is now positive, averaging 7 percent in 2000–4 (Table 7.1). Has the international location of production of farm products within Asia become more or less efficient as a result of policy changes over the past five decades? A global computable general-equilibrium model with a time series of databases is needed to answer that question well, but one very crude way of addressing the question is to examine the standard deviation of RRAs across the economies of the region over time. That indicator suggests intersectoral distortions have become more dispersed across countries over time: it averaged 35 percent during 1960–74, 50 percent during 1975–89, and 55 percent during 1990–2004 (final row of Table 7.1).
9780230_274587_08_cha07.indd 92
6/23/2010 12:16:51 PM
Government Distortions of Agricultural Prices
93
Table 7.3 Trade reduction indexes, Asian countries and other regions,a all covered tradable farm products, 1960 to 2004 (percent) 1960–4 1970–4 1980–4 1985–9 1990–4 1995–9 2000–4 Bangladesh China India Indonesia Korea Malaysia Pakistan Philippines Sri Lanka Taiwan Thailand Vietnam Asian DCs, total Africa Latin America All developing countries Japan High-income countries, all
na na na na 5 12 7 –4 26 –6 na na 15 32 22 26
–13 na 42 1 44 8 19 1 20 –16 25 na 23 33 19 27
–1 44 38 14 119 18 4 3 13 –19 13 na 34 18 19 28
24 44 70 5 158 21 12 16 5 –25 11 12 28 54 13 29
1 19 26 2 189 14 –3 18 23 37 9 28 18 17 23 21
–8 4 18 –1 164 5 –2 39 17 67 6 6 8 16 7 9
6 1 22 19 184 5 4 27 4 96 1 –11 6 22 8 10
64 19
73 16
105 27
144 28
134 28
132 18
127 18
Note: a Regional aggregates are weighted using the absolute value of net imports (computed as the difference between the value of consumption and the value of production) at undistorted prices. Source: Lloyd et al. (2009), based on product NRAs and CTEs in Anderson and Valenzuela (2008).
Of the striking changes in RRAs shown for individual economies over the past two decades, it is the move from negative to positive RRAs for China and India that matter most for the region – and indeed for the world. The extent of the decline in the nonagricultural NRA since the early 1980s is very similar for those two key countries, but the agricultural NRA has differed: in China the 5-year averages have risen steadily from –45 percent to 6 percent, whereas in India it has been close to zero except for a spike upward when international food prices collapsed in the mid-1980s, and for a rise in the present decade (Figure 7.5). This dramatic rise in the RRA for the world’s two most populous countries is of great significance to the current analyses of the causes of the international food price rises of the present decade. One of the contributors is said to be the growing appetite for food imports by these two countries as they industrialize and their per-capita incomes rise. Yet both countries have remained very close to self-sufficient in agricultural products over the past four decades. Undoubtedly the steady rise in their RRAs has contributed to
9780230_274587_08_cha07.indd 93
6/23/2010 12:16:51 PM
94 Kym Anderson Table 7.4 Welfare reduction indexes, Asian countries and other regions,a all covered tradable farm products, 1960 to 2004 (percent) 1960–4 1970–4 1980–4 1985–9 1990–4 1995–9 2000–4 Bangladesh China India Indonesia Korea Malaysia Pakistan Philippines Sri Lanka Taiwan Thailand Vietnam Asian DCs, total Africa Latin America All developing countries Japan High-income countries, all
na na na na 45 14 44 18 32 30 na na 27 52 42 44
30 na 49 18 69 10 75 30 29 52 30 na 39 52 38 42
29 55 54 31 130 57 39 33 26 43 22 na 48 51 44 48
49 48 87 21 176 95 46 46 29 85 18 22 46 81 39 48
29 25 31 24 211 71 31 32 39 124 16 30 28 52 42 32
25 12 22 28 194 31 24 51 35 155 19 24 19 37 20 19
31 8 27 27 228 34 29 42 30 190 12 37 16 36 23 18
74 49
106 46
150 69
248 71
240 52
210 38
213 38
Note: a Regional aggregates are weighted using the average of the value of production and the value of consumption at undistorted prices. Source: Lloyd et al. (2009), based on product NRAs and CTEs in Anderson and Valenzuela (2008).
CHINA 120
INDIA
2000–4
1995–9
1990–4
1985–9
1980–4
1975–9
1970–4
−40 −80
−80
NRA agriculture
Figure 7.5 (percent)
2000–4
1995–9
−40
1990–4
0 1985–9
0 1980–4
40
1975–9
40
1970–4
80
1965–9
80
1965–9
120
NRA non-agriculture
RRA
Nominal and relative rates of assistance, China and India, 1965 to 2005
Source: Calculated from Anderson and Valenzuela (2008), which draws on national estimates reported in Anderson and Martin (2009).
9780230_274587_08_cha07.indd 94
6/23/2010 12:16:52 PM
Government Distortions of Agricultural Prices
95
that outcome (Anderson and Martin, 2009, Tables 1.9 and 1.10). It may also have helped ensure that the trend in China’s ratio of urban to rural mean incomes (adjusted for cost-of-living differences) has been flat since 1980 (Ravallion and Chen, 2007, Figure 3), and that the Gini coefficient for India has hardly changed between 1984 and 2004 (World Bank, 2008). A major question, to which we return at the end of the chapter, is: Will their RRAs remain at their current neutral level of close to zero, or will they continue to rise in the same way, as observed in Korea and Taiwan and, before them, in Japan and Western Europe? To summarize, one of the most salient features of price and trade policies in the Asian region since the 1960s is the spate of major economic reforms, including significant trade liberalization. Overall levels of nonagricultural protection have declined considerably, which has improved the competitiveness of the agricultural sector in many Asian economies but especially in China and India. Two other salient features have been the gradual policy movement away from taxing agricultural exportables, but at the same time – and in contrast to nonagriculture – a rise in agricultural import protection. The latter means there is still scope for reducing distortions in resource use within agriculture even in countries with an average NRA for agriculture, and an RRA, close to zero. In particular, an antitrade bias in assistance rates within the farm sector remains in place. This may be understandable from a political economy viewpoint, but it nonetheless means that resources continue to be allocated inefficiently within the farm sector and, since openness tends to promote economic growth, that total-factor productivity growth in agriculture is slower than it would be if remaining interventions were removed.
7.4 Effects of past reforms and of remaining policies: Results from economy-wide modelling It is clear from the above that there has been considerable reform over the past quarter of a century of policy distortions to agricultural incentives throughout the world: the anti-agricultural and antitrade biases of the policies of many developing countries have been reduced, and the export subsidies of high-income countries have been cut. As well, there has been some re-instrumentation toward less inefficient and less trade-distorting forms of support, particularly in Western Europe (see the dashed line in Figure 7.1). However, protection from agricultural import competition has continued to show an upward trend in both rich and poor countries (Figure 7.2), notwithstanding the Uruguay Round Agreement on Agriculture that aimed to bind and reduce farm tariffs. What have been the net economic effects of agricultural price and trade policy changes around the world since the early 1980s? And how do those effects on global markets, farm incomes, and economic welfare compare
9780230_274587_08_cha07.indd 95
6/23/2010 12:16:52 PM
96
Kym Anderson
with the effects of policy distortions that were still in place as of 2004? Valenzuela et al. (2009) used a global economy-wide model known as Linkage (van der Mensbrugghe, 2005) to provide a combined retrospective and prospective analysis that sought to assess how far the world had come, and how far it still has to go, in rectifying the disarray in world agriculture. It quantifies the impacts both of past reforms and current policies by comparing the effects of the recent World Bank project’s distortion estimates for the period 1980–4 with those of 2004.34 Several key findings from that economy-wide modelling study are worth emphasizing. First, the policy reforms from the early 1980s to the mid-2000s improved global economic welfare by US$233 billion per year, and removing the distortions that remained in 2004 would add another US$168 billion per year (in 2004 US dollars). This suggests that in terms of global welfare the world moved three-fifths of the way towards global free trade in goods over that quarter-century. Second, developing economies benefited proportionately more than high-income economies (1.0 percent compared with 0.7 percent of national income) from those past policy reforms, and would gain nearly twice as much as high-income countries if all countries were to complete that reform process (an average increase of 0.9 percent compared with 0.5 percent for high-income countries). Of those prospective welfare gains from global liberalization, 70 percent would come from agriculture and food policy reform. This is a striking result given that the shares of agriculture and food in global GDP and global trade are only 3 and 6 percent, respectively. The contribution of farm and food policy reform to the prospective welfare gain for developing countries alone is slightly greater, at 72 percent. Third, the share of global farm products (excluding intra-European Union trade) exported in 2004 has been slightly smaller as a result of those reforms since 1980–4, because of less farm export subsidies. The 8 percent share for agriculture in 2004 contrasts with the 31 percent share for other primary products and the 25 percent for all other goods – a “thinness” that is an important contributor to the volatility of international prices for weatherdependent farm products. If the policies distorting goods trade in 2004 were removed, the share of global production of farm products that is exported would rise from 8 to 13 percent, thereby reducing instability of prices and reducing the quantities of those products traded. Fourth, the developing countries’ share of the world’s primary agricultural exports rose from 43 to 55 percent, and its share of global farm output from 58 to 62 percent, because of the reforms since the early 1980s, with rises in output of nearly all agricultural industries except rice and sugar. Removing the remaining goods market distortions would boost their export and output shares even further, to 64 and 65 percent respectively. Fifth, the average real price for agricultural and food products in international markets would have been 13 percent lower had policies not
9780230_274587_08_cha07.indd 96
6/23/2010 12:16:53 PM
Government Distortions of Agricultural Prices
97
changed over the past quarter-century. Evidently the impact of the fall in RRA in high-income countries (including the cuts in farm export subsidies) in raising international food prices more than offset the opposite impact of the RRA rise (including the cuts in agricultural export taxes) in developing countries over that period. By contrast, removing the remaining distortions as of 2004 is projected to raise the international price of agricultural and food products by less than 1 percent on average. This is in contrast to earlier modeling results based on the Global Trade Analysis Project (GTAP) protection database. For example, Anderson et al. (2006) estimated that they would rise by 3.1 percent or, for primary agriculture alone, by 5.5 percent. The smaller impact seen in these new results is because export taxes in developing countries, based on the above NRA estimates for 2004, are included in the new database (most notably for Argentina), and their removal would heavily offset the international price-raising effect of eliminating import protection and farm subsidies elsewhere. Sixth, for developing countries as a group, net farm income (value added in agriculture) is estimated to be 4.9 percent higher than it would have been without the reforms of the past quarter-century, which is more than ten times the proportional gain in nonagricultural value added. If the price and trade policies remaining in 2004 were removed, net farm incomes in developing countries would rise a further 5.6 percent, compared with just 1.9 percent for nonagricultural value added. In addition, unskilled workers in developing countries – the majority of whom work on farms – would see their returns rise more than returns on other productive factors from that liberalization. Together, these findings suggest that both inequality and poverty could be alleviated by such reform, given that three-quarters of the world’s poor are farmers in developing countries (Chen and Ravallion, 2008).
7.5
Prospects for further agricultural reform
The expectation is that, provided they remain open and continue to free up domestic markets and practice good macroeconomic governance, developing economies will keep growing rapidly in the foreseeable future once the current global recession passes. The growth in Asia will be more rapid in manufacturing and service activities than in agriculture, and in the more densely populated economies of the region that growth will be accompanied by rapid increases in per-capita incomes of low-skilled workers where labor-intensive exports boom. Agricultural comparative advantage is thus likely to decline in such economies. Whether these economies become more dependent on imports of farm products depends, however, on what happens to the RRA. The first wave of Asian industrializers (Japan, and then Korea and Taiwan) chose to slow the growth of food import dependence by raising their NRA for agriculture even as they were bringing down their NRA for nonfarm tradables, such that their RRA became increasingly above
9780230_274587_08_cha07.indd 97
6/23/2010 12:16:53 PM
98 Kym Anderson
the neutral zero level. A key question, foreshadowed above, is: Will later industrializers follow suit, given the past close association of RRAs with rising per-capita income and falling agricultural comparative advantage? When the RRAs for Japan, Korea, and Taiwan are mapped against real percapita income, it is possible to superimpose on that same graph the RRAs for lower-income economies to see how they are tracking relative to the first industrializers. Figure 7.6 does that for China and India, and shows that their RRA trends of the past three decades are on the same trajectory as the richer Northeast Asians. That provides reason to expect the governments of later industrializing economies to follow suit if other things are equal. However, might one expect different government behavior now, given that the earlier industrializers were not bound under GATT to keep down their agricultural protection? Had there been strict discipline on farm trade measures at the time Japan and Korea joined GATT in 1955 and 1967, respectively, their NRAs may have been halted at less than 20 percent (Figure 7.7). At the time of China’s accession to WTO in December 2001, its NRA was less than 5 percent according to this present study, or 7.3 percent for just import-competing agriculture. Its average bound import tariff commitment was about twice that (16 percent in 2005), but what matters most is China’s out-of-quota bindings on the items whose imports are restricted by tariff rate quotas. The latter tariff bindings as of 2005 were 65 percent for grains, 50 percent for sugar, and
Relative rate of assistance, %
200 Korea
150
Japan
100 Taiwan 50 0 7 −50
8
9
10
11
India China
−100 ln real GDP per capita Figure 7.6 RRAs and log of real per capita GDP, India and Northeast Asian focus economies, 1955 to 2005 Source: Calculated from Anderson and Valenzuela (2008), which draws on national estimates reported in Anderson and Martin (2009).
9780230_274587_08_cha07.indd 98
6/23/2010 12:16:53 PM
Government Distortions of Agricultural Prices 200
99
China (2001 = 4.5%)
150 Japan Korea (1955 = 16.6%) (1967 = 7.4%) NRA (%)
100
50
2000–5
1995–9
1990–4
1985–9
1980–4
1975–9
1970–4
1965–9
1960–4
−50
1955–9
0
−100 Japan
Korea
China
Figure 7.7 NRAs for Japan, Korea and China and date of accession to GATT or WTO, 1955 to 2005 (percent) Source: Calculated from Anderson and Valenzuela (2008), which draws on national estimates reported in Anderson and Martin (2009).
40 percent for cotton (WTO et al., 2007, p. 60). China also has bindings on farm product-specific domestic supports of 8.5 percent, and can provide another 8.5 percent as nonproduct specific assistance if it so wishes – a total of 17 percent NRA from domestic support measures alone, in addition to what is available through out-of-quota tariff protection. Clearly the legal commitments China made on acceding to WTO are a long way from current levels of domestic and border support for its farmers, and so are unlikely to constrain the government very much in the next decade or so (Anderson et al., 2009). The legal constraints on Asia’s developing countries that joined the WTO earlier (except for Korea) are even less restrictive. For India, Pakistan, and Bangladesh, for example, their estimated NRAs for agricultural importables in 2000–4 are 34, 4, and 6 percent, respectively, whereas the average bound tariffs on their agricultural imports are 114, 96, and 189 percent, respectively (WTO et al., 2007). Also, like other developing countries, they have high bindings on product-specific domestic supports of 10 percent and another 10 percent for nonproduct specific assistance, a total of 20 more percentage points of NRA that legally could come from domestic support measures – compared with currently 10 percent in India and less than 3 percent in the rest of South Asia. Hopefully China and South and Southeast Asia will not make use of the legal wiggle room they have allowed themselves in their WTO bindings and thereby follow Japan, Korea, and Taiwan into high agricultural protection. A
9780230_274587_08_cha07.indd 99
6/23/2010 12:16:53 PM
100 Kym Anderson
much more efficient and equitable strategy would be to instead treat agriculture in the same way they have been treating nonfarm tradable sectors. That would involve opening the sector to international competition, and relying on more efficient domestic policy measures for raising government revenue (e.g., income and consumption or value-added taxes) rather than trade taxes. It might be argued that such a laissez-faire strategy could increase ruralurban inequality and poverty and thereby generate social unrest. On the other hand, policies that lead to high prices for staple foods, in particular, involve potentially serious risks for the urban and rural poor who are net buyers of food in developing countries, as has been demonstrated by concerns about the recent increases in prices of these goods (Ivanic and Martin, 2008). Available evidence suggests that problems of rural-urban poverty gaps have been alleviated in parts of Asia and elsewhere by some of the more-mobile members of farm households finding full- or part-time work off the farm and repatriating part of their higher earnings back to those remaining in farm households (Otsuka and Yamano, 2006, Otsuka et al., 2009). Concerted government intervention through social policy measures can be important both in reducing the gaps between rural and urban incomes, identified by Hayami (2007) as a concern, and in raising national incomes overall. Efficient ways of assisting any left-behind groups of poor (nonfarm as well as farm) households include reducing any underinvestment in rural public-goods that have high social payoffs such as basic education and health and rural infrastructure, as well as agricultural research. The reasons why some countries have reformed their price-distorting agricultural and trade policies more than others in recent decades are varied. Some have reformed unilaterally, apparently having become convinced that it is in their own national interest to do so. China is the most dramatic and significant example of the past three decades among developing countries, and Australia and New Zealand among the high-income countries (Anderson et al., 2007; Huang et al., 2009). Also, some have reduced their agricultural subsidies and import barriers at least partly in response to the GATT’s multilateral Uruguay Round Agreement on Agriculture, the European Union (EU) being the most important example. There are numerous signs that the governments of developing countries want to keep open their options to raise agricultural NRAs in the future, particularly via import restrictions. One indicator is the high tariff bindings to which developing countries committed themselves following the Uruguay Round: as of 2001, actual applied tariffs on agricultural products averaged less than half the corresponding bound tariffs for developing countries of 48 percent, and less than one-sixth in the case of least-developed countries (Anderson and Martin, 2006, Table 1.2). Another indicator of reluctance to undertake agricultural trade reform is the demand by many developing countries to be allowed to maintain their rates of agricultural protection for reasons of food security, livelihood security, and rural development. In
9780230_274587_08_cha07.indd 100
6/23/2010 12:16:53 PM
Government Distortions of Agricultural Prices
101
reality, these are often simply attempts to protect agriculture in middle- and high-income countries, whose farm lobbies are particularly strong politically. To borrow the expression of Hayami and Godo (2005), such agricultural protection is the result of “community failure”: by taking advantage of reduced transaction costs and the increasing value of rural votes due to declines in the farm population, farmers succeed in gaining income support at the cost of domestic consumers’ welfare and that of poor farmers in lower-income countries. In principle the WTO’s Doha Development Agenda is capable of delivering a new trade-liberalizing outcome that would prevent such outcomes, but in practice WTO members are struggling to find common ground and may have to compromise greatly before reaching an agreement. The more wiggle room that is provided for in the eventual agreement, the more likely it is that developing countries will follow the same agricultural protection path this century as that taken last century by high-income countries.
References Anderson, J. E. and Neary, J. P. (2005) Measuring the Restrictiveness of International Trade Policy (Cambridge, MA: MIT Press). Anderson, K. (2008) “Distorted Agricultural Incentives and Economic Development: Asia’s Experience,” the inaugural The World Economy Annual Asia Lecture, University of Nottingham, Kuala Lumpur campus, January 16 (since published in The World Economy, 32 (3), March 2009). Anderson, K. (ed.) (2009) Distortions to Agricultural Incentives: A Global Perspective, 1955–2005 (London: Palgrave Macmillan and Washington, DC: World Bank). Anderson, K., Hayami, Y. and others (1986) The Political Economy of Agricultural Protection: East Asia in International Perspective (London: Allen and Unwin). Anderson, K., Kurzweil, M., Martin, W., Sandri, D. and Valenzuela, E. (2008) “Measuring Distortions to Agricultural Incentives, Revisited,” World Trade Review, 7 (4), October, 675–704. Anderson, K., Lloyd, P. J. and MacLaren, D. (2007) “Distortions to Agricultural Incentives in Australia Since World War Two,” The Economic Record, 83 (263), December, 461–82. Anderson, K. and Martin, W. (eds) (2006) Agricultural Trade Reform and the Doha Development Agenda (London: Palgrave Macmillan and Washington, DC: World Bank). Anderson, K. and Martin, W. (eds) (2009) Distortions to Agricultural Incentives in Asia (Washington, DC: World Bank). Anderson, K., Martin, W. and Valenzuela, E. (2009) “Long Run Implications of WTO Accession for Agriculture in China,” in C. Carter and I. Sheldon (eds.), China’s Agricultural Trade: Issues and Prospects (London: CABI). Anderson, K., Martin, W. and van der Mensbrugghe, D. (2006) “Distortions to World Trade: Impacts on Agricultural Markets and Farm Incomes,” Review of Agricultural Economics, 28 (2), Summer, 168–94. Anderson, K. and Valenzuela, E. (2008) Global Estimates of Distortions to Agricultural Incentives, 1955 to 2005, data spreadsheets available from October at www.worldbank.org/agdistortions. Bates, R. H. (1981) Markets and States in Tropical Africa: The Political Basis of Agricultural Policies (Berkeley, CA: University of California Press).
9780230_274587_08_cha07.indd 101
6/23/2010 12:16:53 PM
102
Kym Anderson
Chen, S. and Ravallion, M. (2008) “The Developing World is Poorer Than We Thought, But No Less Successful in the Fight Against Poverty,” Policy Research Working Paper 4703 (Washington, DC: World Bank). Hayami, Y. (2007) “An Emerging Agricultural Problem in High-Performing Asian Economies,” Policy Research Working Paper 4312, World Bank (Washington, DC), August (revision of Presidential Lecture, 5th Conference of the Asian Society of Agricultural Economists, Zahedan, Iran, August 29–31, 2005). Hayami Y. and Godo, Y. (2005) Development Economics: From the Poverty to the Wealth of Nations, 3rd edition (Oxford:Oxford University Press). Huang, J., Rozelle, S., Martin, W. and Liu, Y. (2009) “China,” in K. Anderson and W. Martin (eds.), Distortions to Agricultural Incentives in Asia, Chapter 3 (Washington, DC: World Bank). Ivanic, M. and Martin, W. (2008) “Implications of Higher Global Food Prices for Poverty in Low-Income Countries,” Policy Research Working Paper 4594, World Bank (Washington, DC), April. Johnson, D. G. (1973) World Agriculture in Disarray, revised 1991 (London: St Martin’s Press). Krueger, A. O., Schiff, M. and Valdés, A. (1988) “Measuring the Impact of Sectorspecific and Economy-Wide Policies on Agricultural Incentives in LDCs,” World Bank Economic Review, 2 (3), 255–72. Krueger, A., Schiff, M. and Valdés, A. (1991) The Political Economy of Agricultural Pricing Policy, Volume 2: Asia (Baltimore: Johns Hopkins University Press for the World Bank). Lerner, A. (1936) “The Symmetry between Import and Export Taxes,” Economica, 3 (11), August, 306–13. Lloyd, P., Croser, V. and Anderson, K. (2009) “Global Distortions to Agricultural Markets: New Indicators of Trade and Welfare Impacts, 1960 to 2007,” CEPR Discussion Paper No. 7160 (London), February. Otsuka, K., Estudillo, J. P. and Sawada, Y. (eds.) (2009) Rural Poverty and Income Dynamics in Asia and Africa (London and New York: Routledge). Otsuka, K. and Yamano, T. (2006) “Introduction to the Special Issue on the Role of Non-Farm Income in Poverty Reduction: Evidence from Asia and East Africa,” Agricultural Economics, 35 (supplement), November, 373–97. Ravallion, M. and Chen, S. (2007) “China’s (Uneven) Progress Against Poverty,” Journal of Development Economics, 82, 1–42. Tyers, R. and Anderson, K. (1992) Disarray in World Food Markets: A Quantitative Assessment (Cambridge and New York: Cambridge University Press). Valenzuela, E., van der Mensbrugghe, D. and Anderson, K. (2009) “General Equilibrium Effects of Price Distortions on Global Markets, Farm Incomes and Welfare,” in K. Anderson (ed.), Distortions to Agricultural Incentives: A Global Perspective, 1955 to 2007, Chapter 13 (London: Palgrave Macmillan and Washington, DC: World Bank). van der Mensbrugghe, D. (2005) “Linkage Technical Reference Document: Version 6.0,” Unpublished (Washington, DC: World Bank), at www.worldbank.org/prospects/linkagemodel. World Bank (2007) World Development Report 2008: Agriculture for Development (Washington, DC: World Bank). World Bank (2008) World Development Indicators (Washington, DC: World Bank). WTO, ITC and UNCTAD (2007) World Tariff Profiles 2006 (Geneva: World Trade Organization).
9780230_274587_08_cha07.indd 102
6/23/2010 12:16:54 PM
8 The Human Capital Basis of the Japanese Miracle: A Historical Perspective Yoshihisa Godo
8.1
Introduction
One of the most conspicuous events that has occurred in the Japanese economy since the Second World War has been the sharp downward kink in its growth trend, beginning in the 1990s. For nearly three decades, from the mid-1950s to the oil crisis of 1973–4, the Japanese economy had recorded an exceptionally high rate of growth of about 10 percent per year, during which its real GDP per capita nearly trebled. As a result, Japan advanced from being a lower-middle-income economy to one of the high-income economies. This high economic growth, which Japanese economists commonly call kodo keizai seichou (literally, “high-speed economic growth”) was unprecedented in history, although soon replicated by other high-performing economies in East Asia. Thus, this phenomenon can be seen as a forerunner of the East Asian Miracle (World Bank, 1993). After a relatively quick recovery from stagflation caused by the oil crisis, the Japanese economy continued to display such an outstanding performance throughout the 1980s that the expression “Japan as Number One” became the vogue during this period. As such, it may not be an exaggeration to call Japan’s four decades from the 1950s to 1980s the “Japanese Miracle.” However, the economy plunged at the beginning of the 1990s, with the country experiencing zero economic growth throughout the period commonly known as the “lost decade.” The Japanese economy has not yet been able to regain its previous vigor; until recently, its annual growth rate was below 3 percent, the lowest of the major industrialized countries. This chapter aims to shed light on the mechanism underlying the rise and fall of the Japanese Miracle from the approach of human-capital formation through school education. Similar to the case of the East Asian Miracle, the Japanese Miracle represented the success of underdeveloped economies in “catching up” with advanced economies through technology borrowing. Rapid economic growth based on technology borrowing depends on (1) the existence of a large backlog of borrowable technologies 103
9780230_274587_09_cha08.indd 103
6/23/2010 5:05:03 PM
104
Yoshihisa Godo
retained by advanced economies, and (2) the development of the capability in developing economies to borrow advanced technologies. Given the very large technology gap existing between developed underdeveloped economies and Japan in the 1950s–80s, the reason why Japan alone was able to fully exploit the advantage of backwardness in technology borrowing, as described by Gerschenkron (1962), cannot be explained without understanding the process by which this technology-borrowing capability was nurtured in Japan. In this chapter, we first attempt to clarify this process based on our new estimates of “the average number of years of schooling in the working-age population, defined as persons between 15 and 64 years of ages inclusive” (henceforth referred to as “average schooling”) over the period of modern economic growth from 1890 to 2000. By comparing the changes in average schooling in Japan with those of the United States, we investigate the process by which Japan accumulated its human capital through formal education to the point of becoming able to borrow worldfrontier technology. Further, we try to demonstrate, from an examination of average schooling for different levels of education, that the Japanese school education system, which was highly effective in supplying the human capital required for technology borrowing in the catch-up stage, failed to meet the demand for higher-level human capital after the catch-up growth was completed. The failure of the school education system to meet this newly emergent demand is considered to be one of the major factors underlying the descent of the Japanese economy into stagnation, beginning with the “lost decade” of the 1990s. We argue that this reflects the “community failure” in the sense of Hayami and Godo (2005), which facilitates erroneous agreements among business communities that undergraduate training at universities is sufficient to maintain high growth in the Japanese economy even after it has caught up with the US and other advanced economies.
8.2
Role of school education in technology borrowing
How does school education contribute to an economy’s capability to absorb foreign technologies? Ohkawa, who was the first to emphasize the importance of the absorptive capacity with reference to Japan’s industrialization, proposed the term “social capability,” which included not only the technical competence of workers but also commercial, industrial, and financial institutions, such as long-term employment contracts in large-scale firms to increase incentives for workers to acquire firm-specific skills, and subcontracting systems for large-scale principal firms to enable the saving of scarce capital (Ohkawa and Kohama, 1989; Ohkawa and Rosovsky, 1973). Further, Abramovitz (1986) extended the concept of social capability to include the political structure that prevents vested interests from blocking innovations.
9780230_274587_09_cha08.indd 104
6/23/2010 5:05:03 PM
The Human Capital Basis of the Japanese Miracle
105
As such, social capability is multifaceted, encompassing nearly all aspects of a social system. Yet among these, school education plays a dominant role. Thus, it is reasonable to assume that the technical competence of workers to produce goods and services, as well as managers’ competence to organize labor and capital for production, is roughly proportional to the years of school education, at least in modern societies (Abramovitz, 1986, p. 388). The competence of workers can also be increased by nonschool training, such as on-the-job training (OJT); however, the effectiveness of OJT and other training programs with practical orientation are critically dependent on the level of formal education among the trainees. How efficiently collective actions can be organized for the supply of appropriate institutions and policies depends on the level of intelligence among citizens and the associated advancement of mass media, which is also ultimately dependent on the level of education. Thus, it should be reasonable to assume that human capital formed by school education is the major determinant of the social capability of a country to borrow advanced technologies. Japan launched its catch-up to Western industrialized nations with the “Meiji Restoration” of 1868, a series of events that resulted in a shift from the feudal system under the hegemony of the Tokugawa Shogunate to the structure of a centralized nation-state under the rule of the emperor. As this reform was induced by the threat of colonization by Western powers, backed by their industrial supremacy, the Meiji government considered it a national priority to promote industrialization. To this end, an engineering college was established in 1886 within the newly-founded Tokyo Imperial University, along with a network of national technical high schools established in subsequent decades along the German model. These tertiary-level schools were critically important for technology borrowing, as they produced the high-caliber scientists and engineers able to understand the scientific principles underlying the mechanisms of foreign technology. These human resources were thus capable of designing appropriate production layouts using technology in a way consistent with the local conditions of the borrowing countries, as well as preparing production manuals for workers on the operation of new production systems. However, such high-level human resources alone are not sufficient for the borrowing countries to establish modern industrial systems. For a developing economy traditionally dependent on agriculture and small-scale cottage industries, the borrowing of modern industrial technologies usually involves the importation of the factory system that had been established in the Western world since the Industrial Revolution. The factory system is characterized by a division of labor, whereby a large number of workers are assembled in a single workshop for cooperative production. Each worker is responsible for performing one of the tasks along a production line designed for the effective operation of a single large machine according to a predetermined work schedule, under the command of a centralized management hierarchy.
9780230_274587_09_cha08.indd 105
6/23/2010 5:05:03 PM
106
Yoshihisa Godo
The high efficiency of this system for the mass production of standardized industrial commodities had been well known since the famous example of a pin factory described by Adam Smith (1776). Thus, for the effective operation of the factory system, a large supply of trained workers is also imperative. The discipline to work in close coordination with many fellow workers in a manner consistent with the rhythm of machine operation, as dictated by the inorganic time of a clock and under the command of supervisors, is not easy to acquire for a worker who is used to farming on his own land, in accordance with the rhythms of nature, and in implicit collaboration with family members and close neighbors (Thompson, 1967). For laborers who used to work according to “peasant time” (Smith, 1988), the development of their aptitude to work in the factory system involves a change in their value system or culture, as broadly defined. If workers who cannot change their culture, namely, those who continue to stick to peasant time, are employed in a factory, their frequent absenteeism would damage the working of the factory system, resulting in a serious loss in production efficiency (Kiyokawa, 2003). School education at the primary and secondary levels represents a highly effective apparatus for the government to embed such an aptitude for the factory system among the youth in traditional agrarian societies. By being forced to come to school at a prescribed time and to study for the duration of class hours together with classmates under the instruction of teachers, children are regimented to follow the disciplines of collective actions. School-wide activities, such as morning assemblies, entrance and graduation ceremonies, and sports days, provide opportunities for them to get accustomed to larger collective actions under hierarchical commands. In addition, basic literacy and arithmetic, as well as the elementary knowledge of sciences acquired by students at primary- and secondary-level schools, add much to the efficiency of their labor through a better understanding of production manuals and instructions given to them when they are employed at factories. The demand for schools that can mass-produce laborers with an aptitude for the factory system is especially large in economies that have newly introduced the factory system as an inseparable part of the process of technology borrowing. The Meiji government launched an ambitious project of propagating the compulsory elementary education system over the entire nation with the School System Rule (Gakusei) of 1872. This proclamation, passed only four years after the Meiji Restoration, had a famous preface announcing its goal that “there shall not be a single household in any village, which does not let children study in school.” This target was virtually achieved in the following four decades, with the school enrollment ratios of elementary school-aged children rising from 28 percent in 1873, to 81 percent in 1900, and up to 98 percent in 1910 (Japan Ministry of Education, 1967, p. 132).
9780230_274587_09_cha08.indd 106
6/23/2010 5:05:03 PM
The Human Capital Basis of the Japanese Miracle
107
It is thus important to recognize that the school education system represents a critically important institution for underdeveloped economies to catch up with advanced economies through the supply of the two types of human capital needed to achieve effective borrowing of advanced industrial technologies: (1) high-level scientists and engineers who can decode the scientific principles underlying the machines and equipment embodying advanced industrial technologies and design appropriate layouts and manuals for the use of foreign technologies in local conditions – that is, human capital produced by tertiary school education; and (2) laborers who have the aptitude to work under the factory system, in terms of conformity with the disciplines of collective work in the factory, and compliance with the instructions from employers conveyed through a hierarchy of management – that is, human capital produced by primary and secondary schools. In the rest of the chapter, we attempt to describe the evolution of the Japanese school system in response to the demand for these two types of human capital, and discuss how particular sequences in their accumulation gave rise to the Japanese Miracle and subsequent economic stagnation.
8.3
The Japanese miracle from a century’s perspective
In this section, we attempt to provide a quantitative overview of the process of Japan’s economic catching up with the United States in terms of real GDP per capita over 100 years since the late-nineteenth century in comparison with its educational catching-up in terms of average schooling. Here, it is assumed that the United States represents the frontier of the world economy with which Japan aimed to catch up during the period of analysis. Through this overview, the mechanism that gave rise to the Japanese Miracle after the Second World War is clarified. This exercise is based on Godo and Hayami (1999, 2002), using data that we consider to be more reliable, while focusing more precisely on the theme of this chapter. Data Average schooling is calculated by accumulating the “total school enrollment” (number of persons attending schools) at corresponding years and ages after adjusting for changes in the population due to immigration and mortality. For the sake of simplicity, we assume no difference in education level between immigrants and domestic citizens, and no correlation between school career and mortality. The total number of years of schooling of persons of age u years in year t, Eu,t, is calculated using the following equation: u⫺1 Gu,t Eu,t ⫽ ∑ N w,t + w⫺u , w ⫽0 Gw,t + w⫺u
9780230_274587_09_cha08.indd 107
(1)
6/23/2010 5:05:03 PM
108 Yoshihisa Godo
where: Nw,t = Total school enrollment of persons at the age of w years in year t, and Gw,t = Total number of persons at the age of w years in year t. By accumulating Eu,t for all ages between 15 to 64 years old and dividing it by the total working-age population, we can find the average schooling in year t, ASt, as follows: 64
∑E
u,t
⫽15 ASt ⫽ u64
∑
(2)
Gu,t
u⫽15
Equations (1) and (2) account for all the persons enrolled in schools regardless of education quality (such as teacher qualifications, student-teacher ratio, and number of schooling days per year), level, and type; a repeated schooling year is also counted as a year. Since there are unmeasurable quality differences in schooling between Japan and the United States, a direct comparison of schooling levels between the two countries is misleading. We believe, however, that changes in the relative ratio will provide useful information on the divergence and convergence of schooling levels between the two countries. The US Census Bureau and the Japanese Ministry of Education provided the official statistics for the school enrollment and population data. We utilized them with the appropriate adjustment and extrapolation based on various historical studies such as Folger and Nam (1967), Cremin (1980), Umihara (1988), and Umemura et al. (1988). Further details on the data sources and estimation procedures used are given in Godo and Hayami (1999) and Godo (2001, 2007). In order to compare economic development with educational development, we used two economic indicators: per-capita GDP and capitallabor ratio (in 1990 PPP US dollars). Capital was measured by the gross nonresidential, nonmilitary fixed stock at the beginning of the year and was estimated by the perpetual inventory method. Labor was measured by total employment (number of persons gainfully employed). The major data sources were OECD (1999), Maddison (1995), the Japan Economic and Social Research Institute (2002), the Japan Cabinet Office (2002), Umemura et al. (1988), the Japan Ministry of Health, Labor and Welfare (various issues), and The Economic Report of the President 1991 and 2002, the US Bureau of Economic Analysis (1993). Further details on the data sources and estimation procedures used are described in Godo (2007).
9780230_274587_09_cha08.indd 108
6/23/2010 5:05:04 PM
The Human Capital Basis of the Japanese Miracle
109
Table 8.1 Human-capital accumulation through school education in Japan and the United States School enrollment ratioa (%) Japan 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000
26 38 43 51 58 62 70 75 78 87 85 87
Average schoolingb (years per person)
US 40 40 62 68 73 74 78 85 87 85 87 89
Japan 1.3 2.0 3.3 4.3 5.6 6.5 7.6 8.7 9.8 10.7 11.5 12.3
US 6.5 7.2 7.7 8.3 9.1 9.8 10.5 11.3 12.0 12.8 13.5 14.0
Notes: a For persons aged 6–20 years. b Average number of years of schooling per person in the working-age population (persons aged 15–64 years). Source: Godo (2007).
Educational catch-up versus economic catch-up Table 8.1 summarizes the estimation results of the average schooling and enrollment ratio for the school-aged (ages 6–20) population. In both the average schooling and the school enrollment ratio, Japan lagged far behind the United States in 1890. However, Japan was able to catch up to the United States almost completely in the enrollment ratio by 1980, and in average schooling by 2000. Figure 8.1 compares the process of Japan’s economic catching-up with the United States with that of its catching-up in school education. The educational catch-up was measured by comparing the changes in the ratio of average schooling in Japan to that of the United States, while the economic catch-up was measured by comparing the changes in the ratio of real GDP per capita in Japan to that in the United States. In addition, the changes in the ratio of physical capital stock per worker in Japan to that of the United States are shown to clarify the mechanism of the economic catch-up. The average schooling in Japan rose from 1.3 years in 1890, which was only 20 percent of that of the United States, to 5.6 years, or 62 percent of the US level, in 1930. Such rapid catch-up in school education was based on the Japanese government’s determination to shoulder the extraordinarily heavy cost of education due to the strong belief that catch-up in education was the most efficient means to catch up with Western economic
9780230_274587_09_cha08.indd 109
6/23/2010 5:05:04 PM
110
Yoshihisa Godo 110 100 90
Average schoolinga
% (US=100)
80 70 60 50
GDP per capitab
40 30 20 10
Physical capital per workerc
0 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 Figure 8.1 The Japan/US ratios in average schooling,a GDP per capita, and physical capital per worker Notes: a. Average number of years of schooling per person in the working-age population (persons age 15–64 years). b. GDP is measured in PPP 1990 US dollars. c. Physical capital stock divided by the total number of workers employed. Physical capital stock is measured by the gross nonresidential nonmilitary capital stock at the beginning of the year. Source: Godo (2007).
and military power. However, this rapid catching-up in education of Japan with the United States before the Second World War was not paralleled by a catching-up in income level. Indeed, the GDP per capita in Japan remained at about 20 percent of the United States level throughout the period 1890 to 1930. Meanwhile, the physical capital stock in Japan increased significantly faster than in the United States, but Japan’s capital-labor ratio remained very low, at only 13 percent of the US level, as late as 1930. However, Japan’s economic catch-up sharply accelerated after the Second World War. Japan’s capital-labor ratio relative to the United States rose rapidly from 10 percent in 1955 to 80 percent in 1990. This was paralleled by an increase in the percentage of Japan’s per-capita GDP from 20 percent to about 80 percent over the same period. Meanwhile, the level of education in Japan continued to approach that of the US, but this rate became much slower than the rate before the Second World War, with the Japan-US gap in education closed by only about 10 percentage points after the war. Why was the gap in GDP per capita between Japan and the United States not reduced significantly before the war, especially before 1930, despite the very rapid closing of the educational gap? And why did the income gap begin to close rapidly in the postwar period, the major investment era of the Japanese Miracle, despite relatively slow growth in average schooling?
9780230_274587_09_cha08.indd 110
6/23/2010 5:05:04 PM
The Human Capital Basis of the Japanese Miracle
111
A decomposition of the average schooling by levels of education will provide a clue to answering those questions. Let: P w,t = Total primary school enrollment of persons at the age of w years old in year t, Sw,t = Total secondary school enrollment of persons at the age of w years in year t, and Tw,t = Total tertiary school enrollment of persons at the age of w years in year t. By definition, Nw,t in equation (1) is equivalent to the summation of P w,t, Sw,t, and Tw,t. By substituting Nw,t by P w,t, Sw,t, and Tw,t, the average schooling of three different levels of education can be obtained respectively from equations (1) and (2). One problem here is that there is no universal agreement among researchers about how to classify the levels of education. In this chapter, following the classification of the school statistics from the US Office of Education, grades 1–8, grades 9–12, and beyond grade 12 are assumed to be primary, secondary, and tertiary, respectively. Table 8.2 shows the summary of the average schooling by levels of education. For 2000, the average schooling of primary education in Japan is almost 8. This means that primary education is universal. The average schooling of primary education in the United States slightly exceeds 8, because repeating a year at the primary level is not rare in the United States. In Table 8.2, as in the case of Figure 8.1, the catch-ups at the three different levels of education (primary, secondary, and tertiary) are measured separately by the average schooling at these levels in Japan relative to those of the United States. This comparison shows that Japan’s catching-up in school education progressed rapidly before the Second World War, first at the primary level, followed by a catch-up at the secondary level. From 1890 to 1930, the average schooling at the primary level in Japan as a percentage of the United States level increased from 20 to 64 percent, and that of the secondary level from 5 to 50 percent. Meanwhile, the catch-up in tertiary education was slow, especially before the mid-1920s, when Japan’s level remained at only about 10 percent of the United States’ level. It appears that the stock of high-level human capital accumulated through tertiary education in Japan had not reached the threshold at which frontier technology practiced in the United States could be effectively borrowed. However, this does not mean that there was an absence of technology borrowing by Japan from the United States and other Western nations during the first half-century of Japan’s modern economic growth. As mentioned in the previous section, significant efforts had been made to selectively borrow appropriate foreign technologies suited to Japan’s economic environment during this period, including the importation of secondhand machinery. The success of such efforts was manifested in the growth of GDP per capita at a rate equivalent to that of the United States, which was then the leader of the world economy. This success is considered to have been based on the rapid accumulation of mid-level human capital through primary and
9780230_274587_09_cha08.indd 111
6/23/2010 5:05:05 PM
9780230_274587_09_cha08.indd 112
6/23/2010 5:05:05 PM
1.27 1.97 2.98 4.08 5.14 5.73 6.33 6.89 7.31 7.55 7.78 7.99
Primaryb
0.01 0.02 0.06 0.17 0.40 0.69 1.12 1.59 2.11 2.56 2.96 3.33
Secondaryc 0.002 0.004 0.012 0.022 0.07 0.12 0.17 0.24 0.37 0.58 0.75 1.03
Tertiaryd
Japan
6.31 6.94 7.36 7.74 8.01 8.25 8.43 8.48 8.40 8.29 8.18 8.13
Primaryb 0.12 0.15 0.25 0.42 0.80 1.23 1.67 2.17 2.70 3.11 3.40 3.60
Secondaryc
Source: Godo (2007).
0.07 0.09 0.12 0.18 0.26 0.33 0.45 0.61 0.89 1.35 1.85 2.24
Tertiaryd
United States
20 28 40 53 64 69 75 81 87 91 95 98
Primaryb
Comparison in average schoolinga by levels of education between Japan and the United States
Notes: a Average number of years of schooling per person in the working-age population. b Schooling of 1st to 8th grades. c Schooling of 9th to 12th grades. d Schooling beyond 12th grade.
1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000
Table 8.2
5 12 25 40 50 56 67 73 78 82 87 93
Secondaryc
3 5 9 12 27 38 38 39 42 43 41 46
Tertiaryd
(United States =100)
The Human Capital Basis of the Japanese Miracle
113
secondary school education, with which the old-vintage technologies in the West were effectively operated in Japan. However, the gap in the GDP per capita between Japan and the United States did not narrow, presumably because Japan’s human-capital stock did not reach the threshold at which world-frontier technologies could be effectively implemented. It appears, however, that Japan reached such a threshold by the 1930s. By this period, the average schooling at the primary level in Japan had reached over 60 percent of that of the United States, and that of secondary education had reached 50 percent, which may be sufficient as a source of human capital in the form of labor equipped with the aptitude to work under the factory system using American technologies for large-scale industrial production. Further, after the promulgation of the University Rule (Daigaku Rei) in 1918, universities and vocational colleges under private management were allowed to be established. The number of public-supported tertiary schools also increased in accordance with the increased demand for highlevel scientists and engineers, as well as modern corporate managers, as the Japanese economy rapidly shifted its focus from labor-intensive to capitalintensive industries after the First World War. Correspondingly, the average schooling for tertiary education in Japan relative to that of the United States began to rise sharply from the 1930s (Table 8.2). The increased supply of high-level human capital produced from these new tertiary education institutions is thought to have contributed greatly to upgrading Japan’s capacity to borrow world-frontier technologies effectively. Indeed, while it had previously imported sophisticated large weapons from abroad, Japan began to produce domestic weapons in preparation for the Second World War. The quality of its products, such as the fighter Kamikaze and the battleship Yamato, soon rivaled those of the United States at the beginning of the hostilities. Mechanism of the Japanese miracle If Japan had made an immediate and effective investment in physical capital for economic production to complement its increases in human capital, instead of squandering it in the pursuit of war, the rapid economic catch-up that occurred after the Second World War would have progressed from the 1930s, even if the catch-up rate might have been slower than that from the mid-1950s to the mid-1970s. In fact, a significant surge in the percentage of Japan’s per-capita GDP relative to that of the United States during the 1930s can be seen in Figure 8.1, a surge which was unfortunately disrupted by the war. With much of its physical capital having been destroyed, the GDP per capita in Japan relative to that of the United States immediately after the war dropped to less than half of the prewar level. However, Japan’s capability to profitably employ premier frontier technology with sufficiently elevated human-capital survived. In fact, many scientists and engineers who played
9780230_274587_09_cha08.indd 113
6/23/2010 5:05:05 PM
114
Yoshihisa Godo
pivotal roles in the promotion of major innovations in the Miracle era, such as National Rail bullet trains, Honda cars, and Sony TVs, had previously been employed in military laboratories during the war. Meanwhile, opportunities to borrow foreign technologies in Japan were particularly numerous in the period immediately after the war, as a large backlog of technologies ripe for borrowing had been accumulated during the country’s complete wartime isolation from the international business, science, and engineering communities. The high-profit opportunity for physical capital investments based on previously accumulated human capital, and the increased backlog of borrowable technologies, are thought to have sharply raised the rate of return to business investments geared for borrowing advanced foreign technologies. A major investment spurt did indeed occur, resulting in the rapid economic catch-up of Japan with the United States during the first three decades after the Second World War. This process was the essence of the Japanese Miracle.
8.4
From the miracle to long-term stagnation
If the Japanese Miracle was based on the borrowing of advanced technology from abroad at the end of the Second World War, the speed of its economic growth would have inevitably decreased as the technology gap between Japan and advanced economies (typically the United States) closed as a direct result of successful technology borrowing. Thus, it is easy to understand why the period of miracle growth did not last very long. However, one question remains: Why, in the 1990s, did the Japanese economy slide into the worst recession among the major industrialized countries? A clue to answering this question can be found in Table 8.2, which shows that, during the Miracle period, educational catching-up continued to progress mainly at the primary and the secondary levels, while no significant catch-up occurred at the tertiary level. The percentage of Japan’s average tertiary schooling relative to that of the United States remained virtually constant. This observation implies that Japan was able to borrow world-frontier technology with no concurrent investment for upgrading its human-capital base relative to the United States in terms of the stock of high-level scientists and engineers. This may not be an anomaly, however, considering the conditions surrounding Japan at that time. Since advanced technologies were abundantly available abroad, Japan did not need to spare its own resources to develop them. A high rate of economic growth could thus be readily achieved with a relatively small cadre of high-level scientists and engineers who could adapt foreign technology for domestic use. A more critical question would have been how to secure a supply of laborers who could understand the basics of the translated technology and master its use in accordance with the instructions of a hierarchical management within a large-scale factory system.
9780230_274587_09_cha08.indd 114
6/23/2010 5:05:05 PM
The Human Capital Basis of the Japanese Miracle
115
As explained previously, a considerable number of scientists and engineers who had been trained to develop military technology were able to survive the Second World War. They played a pivotal role in implementing world-frontier technology for domestic industrial use in Japan after the war. It appears that the business entrepreneurs in Japan at that time did not feel much need for scientific and engineering talents beyond the legacy of the war. Their strategy was to increase the productivity of new enterprises based on borrowed technology by promoting its mastery by workers. Indeed, the continuous improvement (kaizen) of production processes by workers and foremen at the shop-floor level to achieve zero-defect production under the “Japanese system of management,” which aims to forge the loyalty of workers through such means as lifetime employment and seniority-based promotion, was the major source of international competitiveness for Japanese industries (Abegglen and Stalk, 1985; Aoki and Dore, 1994). Thus, the Japanese business leaders in the Miracle period voiced little need for the strengthening of tertiary education. Instead, they strongly demanded the increase of middle-level schools oriented towards vocational education (Ohta, 1978). The allocation of a large budget for middle-level education at the expense of higher-level education was probably efficient for maximizing the economic growth of Japan in the postwar catch-up process. However, the highgrowth performance of the Japanese economy at that time depended on the availability of advanced technology abroad, which could readily be borrowed by a relatively small cadre of high-level scientists and engineers. This backlog was certain to be exhausted as the foreign advanced technology was successfully borrowed. When Japan closed its technology gap visà-vis advanced industrial economies at the end of its successful economic catch-up in the 1980s, it required new and original technologies to compete with other industrial economies in the world market. Much larger and higher-quality human capital is necessary for producing innovative ideas and designs at home than for borrowing them from abroad. The Japanese government failed to prepare such a human-capital base during the Miracle growth period. The very success of the Miracle growth mechanism, which relied on the adaptation of foreign technology for domestic use by a small number of scientists and engineers who had survived the Second World War, then effectively utilized by high-quality labor with middle-level education at relatively low cost, had blinded the eyes of communities of Japanese entrepreneurs and policy-makers to the need for higher-level human capital at the end of the catch-up process. This lag resulted not only from the inability of the government to foresee the emerging need for Japan to build world-class, research-oriented graduate school systems, but also from private entrepreneurs’ preoccupation with preserving the Japanese system of management. As this system was geared to promote cooperation among employees for the prosperity of the company,
9780230_274587_09_cha08.indd 115
6/23/2010 5:05:05 PM
116
Yoshihisa Godo
management was very careful to maintain harmony among the employees. For this purpose, employees were treated as equally as possible, for example, in principle, they were promoted at the same speed according to seniority. Bound by tradition like the traditional village community, the management of Japanese companies was reluctant to employ the holders of advanced degrees such as MBAs and PhDs for placement in positions above the workers who had been promoted through OJT after being employed with mid-level education. This management practice undoubtedly reduced private returns to investment in advanced education, particularly graduate-level education (Clark, 1995). Such a corporate culture is difficult to change in response to changes in economic environments. Additionally, a long time- lag is usually involved for current investment in schooling to materialize into a stock of human capital useful for economic production. To that extent, the speed with which high-level human capital has accumulated in Japan just prior the end of the economic Miracle has not been able to compensate for diminishing opportunities for borrowing technology from abroad. The percentage of graduate students in the total enrollment of tertiary education is shown in Figure 8.2 (the total enrollment of tertiary education = graduate course enrollment + undergraduate course enrollment). As can be seen, graduate programs were accorded low priority in Japan. Up until the
20 United States
(%)
15
10
5 Japan 0 1940
1950
1960
1970
1980
1990
2000
Figure 8.2 The percentage of graduate courses in the total enrollment of tertiary education Note: a. The total enrollment of tertiary education = the total enrollment of undergraduate courses + the total enrollment of graduate courses. Source: Japan Ministry of Education, School Statistics, various issues, US Bureau of the Census, Historical Statistics of the United States, 1973, US Bureau of the Cenus, Statistical Abstract of the United States, various issues.
9780230_274587_09_cha08.indd 116
6/23/2010 5:05:05 PM
The Human Capital Basis of the Japanese Miracle
117
middle of the 1980s, the Japanese government had not promoted graduatelevel social science education for business people. The lack of high-caliber social science education has been a serious disadvantage for Japanese business people wanting to work in the international arena (Inoki, 2009). While the Japanese government has turned to the promotion of graduate programs since the middle of the 1980s, Japan’s percentage of graduate students in the total tertiary education is still significantly lower than the US level even now. Since there are no visible signs thus far that Japan will begin catching up with advanced economies such as the United States in terms of high-level education, as shown in Figure 8.2, the prospect for Japan to remain competitive in the global market appears bleak, since the competitiveness of an advanced economy vis-à-vis other advanced economies of equal rank on the world frontier of technology is critically dependent on its ability to develop innovative ideas by using its own resources to push technological and scientific frontiers, instead of borrowing the ideas of others. Although the Japanese government began to place more emphasis on tertiary education, especially graduate programs, from the latter half of the 1980s, it was too late to achieve the catch-up in the stock of advanced education by the time the economic catch-up was completed around the end of the 1980s. As a result, the shift from being the borrower to the originator of new and innovative ideas progressed much more slowly than was needed. This lag in the shift of priority in educational investment from middle to higher levels of schooling appears to be one of the major factors underlying the poor performance of the Japanese economy following the “lost decade” of the 1990s.
8.5 Summary and conclusion In this chapter, the processes by which the so-called Japanese Miracle, characterized by exceptionally high economic growth, emerged during the first four decades after the Second World War, were explored. Furthermore, the process by which this Miracle growth was succeeded by prolonged recession and poor growth performance in Japan among the major industrialized countries was examined. For these purposes, comparisons were made between Japan’s catching up with the United States in the stock of school education, and its economic catching up in terms of real GDP per capita and physical capital stock per worker after over a century of modernization since the Meiji Restoration. Our observations are consistent with the hypothesis that the Japanese Miracle represents a case of rapid economic catch-up by a relatively underdeveloped economy through the effective borrowing of advanced technology from abroad. In the aftermath of the Second World War, Japan was forced to search hard for ways to survive after the severe damage inflicted
9780230_274587_09_cha08.indd 117
6/23/2010 5:05:06 PM
118
Yoshihisa Godo
by the conflict, with much of its physical capital having been destroyed. However, at that time, the Japanese economy was presented with an extraordinary opportunity to achieve rapid recovery and growth. The technology gap between Japan and other developed economies had widened due to its wartime isolation from the rest of the world. As a result, a large backlog of borrowable technology was suddenly made available. A small cadre of scientists and engineers who had received a high level of education before the war and whose research and development capability had been enhanced through their participation in military technology development projects was able to survive the war. Their capability was sufficient to adapt world-frontier technology from abroad for domestic use, which could be effectively utilized by workers who had acquired the necessary knowledge and skills, as well as the aptitude to work in large-scale factories through education in well-developed primary and secondary school systems since the Meiji Restoration. These conditions sharply raised the returns to investment in physical capital for new businesses based on borrowed technology. Correspondingly, an extraordinary investment boom occurred from the 1950s, resulting in the rapid economic catching up of Japan with the United States in GDP per capita and in physical capital stock per worker over the next three decades. This was the basic mechanism behind the Japanese Miracle. During this economic miracle period, however, Japan made no significant catch-up with the United States at the level of tertiary education, though educational catching up continued to take place at the primary and secondary levels. At that time, it is likely that the community of Japanese entrepreneurs felt no urgent need to strengthen high-level educational institutions, since the human-capital base inherited from the wartime period was sufficient to adapt advanced foreign technology for domestic use. They could be competitive in the international market for high-quality, mass-produced industrial commodities by increasing the capability of their employees with middle-level education through OJT geared to promoting their skills and motivation, in order to use the borrowed technology more efficiently by eliciting a cooperative effort under the Japanese system of management. Such a strategy worked very efficiently in supporting the catch-up growth in the Miracle era, but it turned out to be a major constraint on sustaining economic growth when such a process was completed in the 1980s. At that stage, Japan found that the stock of high-level human capital as the basis of producing innovative ideas within Japan was much less than sufficient. The inability of the government to foresee the emerging need for building the system of tertiary education, especially research-oriented graduate schools at a world-class level, before catch-up growth was completed, is one of the major factors underlying the poor performance of the Japanese economy since the beginning of the 1990s. Furthermore, largely similar and inappropriate expectations about the future role of tertiary education held by a group of private entrepreneurs was a fundamental cause for the failure to strengthen
9780230_274587_09_cha08.indd 118
6/23/2010 5:05:06 PM
The Human Capital Basis of the Japanese Miracle
119
such education. The absence of competition among Japan’s giant enterprises is likely to have led to the lack of innovative policies to hire highly educated scientists by such enterprises, which may have been a consequence of excessive coordination among business leaders in Japan, similar to that of village communities which attempt to conserve traditional systems. In other words, the failure to strengthen graduate-level education is an example of the “community failure” discussed by Hayami and Godo (2005). The current stagnation of the Japanese economy is but one of the many examples in which a system that has proved to be very effective in promoting economic growth at a certain stage of development has turned out to be an obstacle to further growth in the next stage. Since previous cases of economies able to escape from these constraints are quite rare in history, the prospect of the Japanese economy being able to regain its vigor in the competitive global arena appears to be rather bleak. However, the high-performing economies currently comprising the East Asian Miracle may have a better chance to escape from being trapped is this cycle if they can learn from Japan’s failure. In this regard, the current strategy of Singapore to build world-class education and research systems to support the transition of its economic center of gravity from labor-intensive to knowledgeintensive industries before the catch-up growth process is completed might represent a promising model for developing economies to sustain economic growth beyond the stage of catching up. Indeed, real GDP per capita in Singapore is already higher than that in Japan.
References Abegglen, J. C. and Stalk, G., Jr. (1985) Kaisha: The Japanese Corporation (New York: Basic Books). Abramovitz, M. (1986) “Catching Up, Forging Ahead, and Falling Behind,” Journal of Economic History, 46: 385–406. Aoki, M. and Dore, R. (eds.) (1994) The Japanese Firm: Sources of Comparative Strength (Oxford: Oxford University Press). Clark, B. R. (1995) Places of Inquiry: Research and Advanced Education in Modern Universities (Berkley, Los Angeles, London: University of California Press). Cremin, L. A. (1980) American Education: The National Experience, 1783–1876 (New York: Harper and Row). Economic Report of the President: Transmitted to the Congress February 1991 (1991) (Washington, DC: US Government Printing Office). Economic Report of the President: Transmitted to the Congress February 2002 (2002) (Washington, DC: US Government Printing Office). Folger, J. K. and Nam, C. B. (1967) Education of the American Population (Washington, DC: US Government Printing Office). Gerschenkron, A. (1962) Economic Backwardness in Historical Perceptive (Cambridge, MA: Harvard University Press). Godo, Y. (2001) Estimation of Average Years of Schooling by Levels of Education for Japan and the United States, 1890–1990, FASID Development Database 2000–01 (Tokyo: Foundation for Advanced Studies on International Development).
9780230_274587_09_cha08.indd 119
6/23/2010 5:05:06 PM
120 Yoshihisa Godo Godo, Y. (2007) “Revised Annual Estimates of Average Years of Schooling for Japan, Korea and the United States,” mimeo. Godo, Y. and Hayami, Y. (1999) “Accumulation of Education in Modern Economic Growth: Comparison of Japan with the United States,” ADBI Working Paper 4 (Tokyo: Asian Development Bank Institute). Godo, Y. and Hayami, Y. (2002) “Catching-Up in Education in the Economic Catch-Up of Japan with the United States, 1890–1990,” Economic Development and Cultural Change, 50 (4), July, 961–78. Hayami, Y. and Godo, Y. (2005) Development Economics: From the Poverty to the Wealth of Nations, 3rd edition (Oxford: Oxford University Press). Japan Cabinet Office, Director-General for Policies (2002) Social Capital of Japan (Tokyo: Toyo Keizai Shimpo-sha). Japan Economic and Social Research Institute (2002) Gross Capital Stock of Private Enterprises (At Market Prices in Calendar Year of 1995) – 1990–2000 (Tokyo: Printing Bureau of the Ministry of Finance). Japan Ministry of Education (1967) Wagakuni Kyoiku no Ayumi (History of Japanese Educational Development) (Tokyo: Meiji Tosho Shuppan). Japan Ministry of Health, Labor and Welfare, Monthly Labour Survey, various issues. Kiyokawa, Y. (2003) Ajia ni Okeru Kindai teki Rodoryoku no Keisei (The Formation of Modern Labor in Asia) (Tokyo: Iwanami-shoten). Maddison, A. (1995) Monitoring the World Economy 1820–1992 (Paris: OECD). Ohkawa, K. and H. Kohama (1989) Lectures on Developing Economies: Japan’s Experience and Its Relevance (Tokyo: University of Tokyo Press). Ohkawa, K. and H. Rosovsky (1973) Japanese Economic growth: Trend Acceleration in the Twentieth Century (Stanford, CA: Stanford University Press). Ohta, T. (1978) Sengo Nihon Kyoikushi (The History of Education in Japan in the Postwar Period) (Tokyo: Iwanami-shoten). Organization of Economic Development and Cooperation (OECD) (1999) National Accounts: Main Aggregates 1960–97 (Paris: OECD). Smith, A. (1937) (original publication, 1776) An Inquiry into the Nature and Causes of the Wealth of Nations, in E. Cannnan (ed.), 6th edition (New York: Modern Library). Smith, T. C. (1988) Native Sources of Japanese Industrialization, 1750–1920 (Berkeley and Los Angeles, CA: University of California Press). Thompson, E. P. (1967) “Time, Work-Discipline, and Industrial Capitalism,” Past and Present, 38, December, 56–97. Umemura, M., Akasaka, K., Minami, R., Takamatsu, N., Arai, K., and Itoh, S. (1988) Manpower (Estimates of Long-term Economic Statistics, Volume 2) (Tokyo: Toyo Keizai Shinposha). Umihara, T. (1988) Kinsei no Gakko to Kyoiku (Education and Schools in Japanese Modern Times) (Kyoto, Japan: Shibunkaku Shuppan). US Bureau of Economic Analysis (1993) Fixed Reproducible Tangible Wealth in the United States, 1925–89 (Washington, DC: US Government Printing Office). World Bank (1993) The East Asian Miracle (New York: Oxford University Press).
9780230_274587_09_cha08.indd 120
6/23/2010 5:05:06 PM
Part III State to Community
9780230_274587_10_cha09.indd 121
6/23/2010 12:17:13 PM
9780230_274587_10_cha09.indd 122
6/23/2010 12:17:13 PM
9 Community Governance and Public-Goods Investments in China’s Villages* Renfu Luo, Linxiu Zhang, Jikun Huang, and Scott Rozelle
9.1
Introduction
Yujiro Hayami and his coauthors have studied one of the most fundamental questions of development economics: How does democracy (or any form of community governance) affect the development process (Aoki and Hayami, 2001)? The unique approach of Hayami is that he has built his foundation for examining these questions on the study of rural communities themselves. In his books on the Philippines (Hayami and Kikuchi, 2000) and Indonesia (Hayami and Kawagoe, 1993), Hayami has studied the roles of rural communities in irrigation management, land and labor relations, risk coping, social capital, and development of the rural nonfarm sector. The major finding in many pieces of his research is that rural communities help allocate resources that increase both efficiency and equity. A branch of the literature concerned with understanding the mechanism that underlies the relationship between rural governance and growth has more recently emerged, centering on works that track the way governance institutions affect public resources allocation and public-goods investment. A number of papers, many of which focus on developed countries, seek to answer questions such as why elected officials respond to the demands of citizens, positing that democratic governments are responsive to public demands because citizens use periodic elections to sanction bad politicians and reward good ones (Besley and Coate, 2003a; Fiorina, 1981; Key, 1966). Empirically, the linkage between elections and public goods in developed countries has been examined (Deacon, 2003). *We want to acknowledge the financial assistance of National Natural Science Foundation of China (70225003), Chinese Academy of Sciences (KSCX2-YW-N-039), the World Bank, the Ford Foundation, and the Canadian International Development Agency.
123
9780230_274587_10_cha09.indd 123
6/23/2010 12:17:13 PM
124 Renfu Luo et al.
In recent years a number of theoretical and empirical papers have addressed these issues in the developing world. Research on these issues in developing countries may be equally or even more important, given the early stages of democratic transition in which many countries find themselves. In addition, public goods in developing countries are often severely underfunded. Theoretically, some studies analyze the question of how local governance affects the provision of public goods (Bardhan and Mookherjee, 2000a, 2000b, 2001; Besley and Coate, 2003b). Unfortunately, according to much of the theoretical literature, the exact nature of the linkage between democracy and public-goods investment is ambiguous. One exception is a paper by Rosenzweig and Foster (2003 – henceforth RF). In their paper, RF derive the unambiguous result that links the election process to public-goods investments by embedding a two-party voting model of democracy in a general-equilibrium framework (which seeks to capture the activities in the rural economy). The RF paper finds that increasing the population weight of the poor segment of a village’s population in a democracy induces public resource allocations in a direction that increases the welfare of the poor. Although there is a desperate need to explore this issue empirically, applied researchers were precluded from studying the linkages between democracy and public goods empirically until recently because of the absence of detailed information (Dethier, 1999). However, the logjam has been broken. For example, Besley and Burgess (2002) use data from India to show that the election of local leaders has a positive effect on public food distribution and disaster relief. Other research teams have shown that when there is a fair election process, increasing the number of representatives of a specific subgroup (or interest group) on the village council (through some mechanism such as affirmative action or the reservation system) will increase the allocation of public resources to favor the group with the largest voting block (Chattopadhyay and Duflo, 2004). Finally, Rosenzweig and Foster (2003) demonstrate that local democratization is positively correlated with the provision of local public-goods. In their work they demonstrate that when the constituency of the leaders is composed mostly of farmers (in the case of India, large landowners in a governance system that is based on hereditary leadership), local leaders tend to allocate more investment funding to irrigation. In contrast, when the local voting population is made up mostly of landless laborers and there is a democracy, local leaders tend to invest more in roads. Social scientists have examined similar sets of issues in China, a nation that is now into its fifth or sixth round of local elections in many rural villages. However, despite the intense attention from political scientists (e.g., Kelliher, 1997; O’Brien, 1994), there has been relatively little effort by economists to study the linkages between elections and public-goods investment in rural China. There does appear to be considerable scope
9780230_274587_10_cha09.indd 124
6/23/2010 12:17:14 PM
Community Governance and Public-Goods Investments
125
for studying election–investment links. Although China began having village-level elections in the 1980s, the pace of the emergence of elections and their quality vary both over time and across space (Zhang et al., 2004). Moreover, while there seems to be a high return to investment in public goods (Fan et al., 2004), village-to-village differences in investment are great (Zhang et al., 2006). Somewhat surprisingly, it has only been in recent years that economists have begun to examine these issues. For example, Brandt and Turner (2003) find that village elections provide a strong disincentive to rent-seeking. While this may mean that leaders might use less of a village’s fiscal revenues on activities that improve their own personal welfare, they do not attempt to measure the effect. The only published paper that we know of to address the issue of the linkage between elections and public service provision in rural China (a concept somewhat broader than public-goods investment) is Zhang et al. (2004). Using survey data from 60 villages in Eastern China, it is found that when villages hold elections (regardless of the nature of the election), the village’s leadership tended to tax constituents less and allocate more local resources to public service provision. The Zhang et al. (2004) paper, however, is based on data from only one region of China; it does not differentiate whether the village leader was directly elected; and it does not seek to quantify the mechanism that is driving the election–investment relationship. In short, in China, a country that has held more than four million local elections during the past two decades, there are a number of fundamental questions that have not been answered. When rigorous statistical sampling is used at the national level, can we find any relationship between the emergence of elections and public-goods investment? Holding all other factors constant, when villagers directly elect their leaders, does the leader increase investment in public goods during his/her term of office? More fundamentally, why? When leaders are in office, do they increase investment in public goods in a way that aids their efforts to extend their term in office? The overall goal of this chapter is to answer some of these questions as a way to better understand the relationship between rural governance reforms in China and investment in public goods. To meet this overall goal, we have three specific objectives. First, using survey data we provide a profile of village elections and public-goods investments, and chart the way they move together. Second, we seek to understand if elections, all other things being constant, lead to higher levels of public-goods investment. Finally, we hope to push the empirical literature further than has been done in the past, and seek to explain empirically the mechanism through which elections operate by testing whether or not leaders who are in office are able to enhance their chances for reelection by investing in public goods.
9780230_274587_10_cha09.indd 125
6/23/2010 12:17:14 PM
126
9.2
Renfu Luo et al.
Adapting the framework to rural China
In adapting the RF conceptual model to China, we need to do two things. First, we need to show that there are two distinct classes in China’s villages. This is needed primarily because the distinction between landed and landless does not pertain to China, where virtually 100 percent of households in every village have access to almost equal quantities of land within the villages (Brandt et al., 2002). While land is allocated equally, this does not mean that all households are created equal. According to Yan (1992), there are several types of households in rural villages, which can be separated into two distinct groups. On the one hand there is a category that we call the elites. The elites include cadres (both former and current ones that work in the village and in upper levels of government) and their family members, relatives, and close friends. In addition, there are the ordinary villagers (henceforth, simply called villagers), which constitute the rest of the people in the village. In villages that do not have elections, the types of households are characterized mainly on the basis of the access that each group – the elites and villagers – has to public resources. Elites, who are mostly appointed to their positions (or who have “inherited” their positions by virtue of family ties), control the community’s budget and prefer to spend the budget on those things that benefit the elites (e.g., on banquets, office supplies, travel, and office buildings). According to a World Bank report (Fork and Wong, 2005), even during the 2000s village leaders spent about half of village fiscal expenditures (44 percent) on official salaries, entertainment, administrative expenses, and other related expenses (henceforth called entertainment expenditures for short). Clearly, these kinds of expenditure have little if anything to do with villagers. In our chapter we assume that these types of expenditures are the counterpart to expenditures on irrigation in the RF paper. The second assumption that we need to make to establish the validity of the conceptual framework of our chapter is about the spending preferences of villagers who gain a voice in village affairs through elections. The first question to answer is whether there are any “third categories of spending” besides entertainment and public infrastructure. According to Fork and Wong (2005), public-good infrastructure spending constitutes the other main spending category of villages. In fact, as a share of total spending, allocations to public-goods investment accounted for 43 percent in the early 2000s – which is just a bit less than the share spent on entertainment. There are only minor amounts of spending on other categories, such as social security (4 percent); militia training (2 percent); and environmental protection (1 percent). Since there is almost no change in total fiscal revenues or expenditures during our study period (Fork and Wong, 2005), if we find that elections lead to greater spending on public goods, this means that there is less spending on entertainment.
9780230_274587_10_cha09.indd 126
6/23/2010 12:17:14 PM
Community Governance and Public-Goods Investments
127
With such little spending on the other categories, the remaining question is whether or not villagers have a strong preference for spending on public infrastructure. In fact, our survey (which is described in detail below) contains strong evidence that villagers have a strong demand for spending on village public-goods infrastructure. According to the data 70 percent of our sampled households responded that they believed there had been an improvement in the level of public-goods infrastructure between 1998 and 2004. Despite the fact that there had been improvements in rural publicgoods infrastructure over the years previous to the survey period, nearly 80 percent of households stated that they were still dissatisfied with the current level of infrastructure. This finding, based on our data, is consistent with the findings of the World Bank (2004) that public-goods provision in rural areas of developing countries are far from sufficient and are often in high demand by villagers. Therefore, for these reasons we believe that there is support for our assumption that rural residents in rural China have high demand for infrastructure investments. In fact, our data set also contains more direct evidence that villagers have a strong demand for infrastructure. In our sample villagers can also be shown to be willing for resources to be directed towards infrastructure projects. In our survey we asked villagers: If the upper-level government gave 50,000 yuan to the village, how would they want the funds to be spent. Villagers were given a choice of answers (including infrastructure investments; investment into village office space; environmental projects; and social welfare-oriented projects). In response, by far the majority of them (90 percent) stated that they would prefer that the funds were spent on public-goods investments. In addition, when we questioned the willingness of individuals to contribute to the public-goods investment efforts with their own money (20 yuan per family member – or about 100 yuan per family), most (nearly 70 percent) said they would be willing to contribute. Based on our assumption and the logic of the RF model, we can state our own version of the election–investment hypothesis. When there are no elections, elites have control over the village budget and spend more on nonpublic goods infrastructure expenditures. When elections emerge, villagers gain more of a say over village affairs and village leaders begin to allocate more spending towards the things villagers demand: public-goods infrastructure. In short, the hypothesis to be tested is: Elections in rural China stimulate more spending on public-goods infrastructure.
9.3
Data sources and empirical model specification
To test the effect of village governance on public-goods infrastructure we conducted a survey in rural China at village level in 2003. Although China initiated reforms aimed at transforming village governance in villages towards self-governance over the past two decades, the promulgation of the
9780230_274587_10_cha09.indd 127
6/23/2010 12:17:14 PM
128 Renfu Luo et al.
PRC Organic Law for Villager Committees formally took place in 1998 (Shi, 2004). One of the most salient features of the Organic Law is that there is a clear message that village leaders are supposed to be elected. Because of this timing, we collected village information (based on both accountant records and village official recall) in 1997 as a baseline. The survey itself was conducted in late 2003, so the survey period covers six years – from 1998 to 2003. To ensure the national representative nature of our sample, a standard procedure was used to choose the sample villages. In each of China’s major agro-ecological zones, we randomly selected a sample province. Sample counties and sample townships were also selected randomly. Within each township, we included all villages in the survey (unless there were more than 20 villages in the township, where we randomly chose 20 villages). In total, we surveyed 2448 villages in 6 provinces, 36 counties, and 216 townships. In the survey form we mainly asked the respondents three sets of questions about sample villages. First, the survey included a section that elicited information about the general characteristics of the village, such as its resource base (human and land), the economic structure of the village, income, geography, location of the village, the timing of other major reforms, ties to government officials in the township and county government, and so on. In the second part of the survey, there were a number of questions about public-goods investment in the village. For each year of the survey, the size, timing, and source of funding of each public-goods investment was enumerated. Finally, the survey had a section that examined the system of governance that prevailed in each year in each sample villages. For example, we recorded the name, level of education, and tenure of each village leader. Most importantly, we also asked how each village leader gained office – by direct election or by appointment. In this chapter, a direct election is defined as one in which villagers vote for two or more candidates and the winner becomes the village leader. All other village leaders are considered to have been appointed. Among other things, this information allowed us to “count” how many different people served as village leader between 1998 and 2003, and we created a list of all of the “terms.”35
9.4 Empirical model, variable definition, and estimation approach In the analysis we will have two broad sets of equations. The first, which we call the baseline analysis, will be estimated by OLS and will include a number of time-invariant (as well as time-varying) variables. We also will take advantage of the panel nature of our data set and estimate a fixed-effects model, which will entail dropping the time-invariant variables since they will be subsumed into the 2448 village dummy variables.
9780230_274587_10_cha09.indd 128
6/23/2010 12:17:14 PM
Community Governance and Public-Goods Investments
129
In order to empirically test the hypothesis of the China-version of the RF model, we assume that public-goods investments are a linear function of village governance and other factors. Using the assumptions that will be used to conduct the baseline analysis, the relationship can be written as: yijt b0 b1Dit b2Vi b3Vit mi it for each j,
(1)
where thes b s are parameters to be estimated and yijt is a measure of the level village public-goods investment (measured either as the number of projects per year or investment level per year) in village i for investments from source j during the village leader’s term t. Because not all investments are from the same source, we need to be careful to understand the effect of governance on investments from different sources (identified by j). Specifically, according to our survey, individuals in Chinese villages contribute a large share of the funding of public-goods investment (Zhang et al., 2006). Therefore, in equation 1, we specify the dependent variable in one of four ways. First, we estimate equation (1) for total investment which we identify as the sum of all sources of funds (j = 1). We also run separate regressions for projects that are funded in one of three ways – solely funded by the village itself (j = 2); projects that are jointly funded by both the village and the upper-level government (j = 3); and projects that are solely funded from above (j = 4). On the right-hand side of the estimated equation the main variable of interest is, Dit the variable that represents the governance mode of the village. As discussed above, we use the variable “Was the village leader elected directly (one if the village leader was elected directly)?” We use this in part because there is little error in this formulation of the governance variable and in part because it is argued in the political science literature that the direct election of the village leader is the most salient feature in China’s village governance (Louie, 2001). Though our measure of governance reforms is fairly rudimentary, what we give up in richness, we believe we gain in coverage. In our data set we observe the mode of governance in 7041 terms (in the 2448 villages and 6-year time-period). We also have at least two observations for each village. To obtain more consistent estimates of the coefficient Dit, in the baseline analysis we also add a number of control variables (Vi and Vit) to the right-hand side of equation (1). These variables come from our data set and they (or similar variables) have been used by others who empirically estimate similar equations (Banerjee et al., 2005; Besley and Case, 1995b; Chattopadhyay and Duflo, 2004; Miguel and Gugerty, 2005; Rosenzweig and Foster, 2003). In our equation the vector of time-invariant variables (Vi) includes socioeconomic as well as locational, geographical, and other factors. We also include a number of time-variant variables (Vit), including the level of education of the village leader (measured in years of attainment); the age of the village leader (measured in years); the occupation of the village
9780230_274587_10_cha09.indd 129
6/23/2010 12:17:14 PM
130 Renfu Luo et al.
leader (where the variable is equal to one if the village leader has never held an off-farm job or is a “pure farmer,” and zero otherwise); and a policy variable that is equal to one if the election term occurred after China’s Tax for Fee reform (a policy that has had a major effect on local fiscal management: Fork and Wong, 2005, and zero otherwise). While the baseline analysis provides us with important information (it at least shows us correlations), it is necessary to use caution in our interpretations of the results. The coefficient of interest (Dit) may still be affected by unobserved heterogeneity. For example, the unobserved heterogeneity may be caused by some unobserved village time-invariant factors (e.g., it could be that a community has a long history of collective action and there is trust within the community). Although these factors may be difficult (or impossible) to measure, they still may affect both whether or not there is a direct election, and the level of investment. To get a consistent estimation of the coefficient of variable Dit, we add a set of village dummy variables mi on the right side of the estimated equation to control for all village fixed effects. Fixed-effects estimation can account for a large part of the unobserved heterogeneity because we also add a time-trend variable in our analysis (which is measured as the first year of the village leader’s term).
9.5 Empirical results: Descriptive and multivariate Despite the fact that villages are supposed to have been electing the village leader, there are still a significant number of villages that do not elect the village leaders (Table 9.1). Over the entire study period (1998 to 2003), although 79 percent of villages elected their leaders, 21 percent still did not. In some of these villages township and county leaders still appointed leaders by policy as a matter of course. Other villages just did not allocate the time to run an election (Table 9.1, row 1).36 In other villages villagers nominated or elected a slate of representatives and the village leader was chosen from among the group by the group members themselves. In all of these cases, we count such villages as being run by a village leader who was not directly elected or (for simplicity) appointed. Our data also show that there are notable differences in the propensity of villages to directly elect their village leaders across space and time. During the survey period nearly all (more than 90 percent) of villages in some provinces, such as Jilin, directly elected the village leader (Table 9.1, row 6). The record in villages in other provinces, for example Hebei province (only a little more than 65 percent), was less complete (row 7). Despite the incomplete record on directly electing the village leader, the share of villages that do elect their leaders directly over time has risen (Table 9.1, rows 8 to 16). Because the village leaders are, by policy, supposed to serve three-year terms, to look at the rise in the propensity of villages to
9780230_274587_10_cha09.indd 130
6/23/2010 12:17:15 PM
Community Governance and Public-Goods Investments
131
Table 9.1 The way in which village leaders acceded to their office in rural China, 1998–2003 (%)a Elected directly Total Jiangsu Gansu Sichuan Shaanxi Jilin Hebei
79 72 77 89 81 93 68
Appointed 21 28 23 11 19 7 32
The difference in the ways village leaders acceded to their office by election cycles (over time) Starting year, election term cycle Term cycle 1 1995 77 23 1998 84 16 2001 86 14 Term cycle 2 1996 1999 2002
73 79 83
27 21 17
Term cycle 3 1997 2000 2003
62 74 79
38 26 21
Note: a Based on 7041 village leaders’ terms in 2448 villages. In our survey, some villages belong to the 1995–1998–2001 term cycle, and the other villages belong to the other two term cycles. Data source: Authors’ survey.
choose their leaders over time, we divide our sample into three distinct sets of villages, based on the starting year of the normal term.37 According to our data, no matter whether we look at villages that run their regulation elections during election cycle 1 (or those that should hold their direct elections in 1995, 1998, and 2001) or election cycle 2, or election cycle 3, the share of villages that directly elect their leaders rises: from 77 to 86 percent; from 73 to 83 percent; and from 62 to 79 percent.
9.6 Public-goods investments On the basis of our data, during the time that elections were beginning to spread across China villages invested in a wide variety of different types of public-goods projects and there was also a great deal of heterogeneity in the number of, and investment levels into, projects – over time, and across
9780230_274587_10_cha09.indd 131
6/23/2010 12:17:15 PM
132 Renfu Luo et al.
villages. Public-goods investment in China’s villages can be categorized into more than 20 different types of projects (e.g., roads and bridges, school construction, irrigation and drainage – henceforth called total investment). By far the majority of these projects (60 percent of them) can be considered core public-goods projects (roads and bridges; schools; irrigation and drainage systems; drinking water; and clinics). Public-goods investment into these activities has also been rising over time. According to our data, the average village invested 60,000 yuan into 0.63 projects per year between 1998 and 2000. Between 2001 and 2003, investment in the average village rose to 71 thousand yuan into 0.67 projects per year. While public goods in many countries are almost entirely the responsibility of upper-level governments, it is not difficult to see that households in Chinese villages also contribute a large share of funding in public-goods investment. While 36 percent of projects are fully funded from above (as is the norm in most countries), nearly half (46 percent) are funded with matching funds from the villages and upper-level government. Eighteen percent of all public-goods projects were funded solely by the village itself. In terms of total investment levels (denominated in real yuan), villages in China were funding 47 percent of their public-goods investments; only a little less than the contribution of funds coming from above (53 percent). Moreover, the level of total investment in our study does not count the investment by China’s villagers in in-kind labor contributions. If the labor days that villagers invested into projects were monetized (at 10 yuan per day, less than half of the going daily, unskilled wage rate), the overall contribution of the local village into their community’s public-goods total investment would far exceed 50 percent. The number of projects and the level of investments (and the split between the sources) also differs by province.
9.7 Elections and investments While there are other factors that explain differences in the level of investment across villages (Zhang et al., 2006), our data suggest in a number of ways that investment is somewhat higher in villages that directly elect their leaders. When looking at average numbers of projects per year, villages that directly elect their village leaders have more projects overall (0.68/year) than those with appointed leaders (0.59/year). The differences appear for all types of investments. The same pattern also appears when we look at the level of investment. Villages that directly elected their leaders invested more in yuan – in total, and in all different types of investments. In all of these cases, the differences in investment are significant, in a statistical sense at least, at the 10 percent level. The relationship between elections and public-goods investment is even clearer when we look at differences over time. When a village went from an elected leader to an elected leader (or from an appointed leader to an
9780230_274587_10_cha09.indd 132
6/23/2010 12:17:15 PM
Community Governance and Public-Goods Investments
133
Table 9.2 The differences in the number (amount) of annual public-goods projects (investments) and the way in which village leaders acceded to their offices over time in rural China, 1998 and 2003 Total projects
Village-funded Govt-funded Co-funded projects projects projects
The difference in the number of annual public-goods projects between two terms No differencea From appointed to directly elected village leader From directly elected to appointed village leader
0.15 0.31
0.03 0.07
0.07 0.12
0.06 0.12
–0.07
–0.07
0.03
–0.03
3.0 9.3
1.4 29.8
1.2
–23.1
The difference in the amount of public-goods investments between two terms No differencea From appointed to directly elected village leader From direct elected to appointed village leader
4.5 42.4
–24.3
0.04 3.2
–2.4
Note: “No difference” means that in between the two election terms there was no difference in the way that village leaders acceded to office. This includes when a directly elected leader followed a directly elected leader, and when an appointed leader followed an appointed leader. Data source: Authors’ survey.
appointed leader) there was little change in total investment (although still relatively small, the change is larger in terms of number of projects than in terms of level of investment – Table 9.2, rows 1 and 4). Likewise, there was little variation between village-funded-only projects, government-fundedonly projects, or co-funded projects.38 However, when a village went from an appointed leader to a directly elected leader the number of projects per year and the level of investment per year both rose relatively more (Table 9.2, rows 2 and 5). In contrast, when a village went from a having a directly elected leader to an appointed leader, the number of projects and level of investment fell (for both total investment and village-funded-only projects).
9.8 Multivariate analysis Because descriptive statistics do not account for the impact of other observed and unobserved factors, we first use equation 1 to assess the effect of direct elections on public-goods investment. In the first set of results, using
9780230_274587_10_cha09.indd 133
6/23/2010 12:17:15 PM
9780230_274587_10_cha09.indd 134
6/23/2010 12:17:15 PM
(5.18)***
(4.22)***
–0.0003 (0.90)
–0.01 (2.27)**
–0.03 (0.49)
Percentage of minority population in 1997 (%)
Per capita land in 1997 (mu)
The illiteracy rate of village labor force in 1997 (%)
–0.19 (2.67)***
0.04 (3.57)***
Total population in 1997 (1000 people)
The migrant ratio of village labor in 1997 (%)
–0.012 (0.99)
Net per-capita income in 1997 (1000 yuan)
–0.14 (3.66)***
–0.04 (1.34)
–0.01 (2.49)**
–0.0001 (0.39)
0.01 (0.99)
0.049 (7.31)***
0.06
Villagefunded
0.09
Time-invariant social-economic factors
The way village leader acceded to office (0 = appointed, 1 = elected directly)
All sources
–0.02 (0.42)
–0.01 (0.18)
–0.003 (1.01)
–0.0003 (1.13)
0.02 (3.00)***
–0.039 (4.68)***
(0.78)
0.01
Co-funded
–0.03 (0.74)
0.02 (0.71)
–0.001 (0.44)
0.00003 (0.15)
0.01 (1.91)*
–0.022 (3.50)***
(1.66)*
0.02
Above funded
Annual public-goods projects in village leader’s term
0.7 (0.03)
31.1 (1.68)*
2.9 (1.94)*
0.32 (2.73)***
14 (4.19)***
7.917 (2.05)**
(1.53)
10.1
All sources
–9.0 (2.55)**
–1.4 (0.47)
0.1 (0.29)
0.03 (1.40)
2.4 (4.58)***
5.314 (8.71)***
(2.85)***
3.0
Villagefunded
18.5 (0.89)
19.1 (1.11)
2.1 (1.52)
0.23 (2.12)**
7.1 (2.28)**
5.159 (1.44)
(0.63)
3.9
Co-funded
–8.8 (1.46)
13.4 (2.68)***
0.7 (1.77)*
0.06 (2.00)**
4.5 (4.99)***
–2.556 (2.44)**
(1.81)*
3.2
Above funded
Annual public goods investment in village leader’s term
Table 9.3 Baseline analysis (OLS) of the impact of directly electing the village leader on public investment in rural China
9780230_274587_10_cha09.indd 135
6/23/2010 12:17:16 PM
0.005 (4.22)***
Number of fellow villagers working in township or county government (each)
0.0003 (0.45)
0.02 (4.98)***
0.0006 (1.54)
–0.001 (4.04)***
–0.002 (1.97)**
0.004 (1.44)
–0.001 (0.75)
Hilly land over 25 degrees in total land area in the village in 1997 (%)
Percentage of effectively irrigated land in 1997 (%)
The distance of the nearest road to the village seat in 1997 (kms)
The farthest distance between two small groups within this village in 1997 (kms)
The distance between village committee and township seat in 1997 (kms)
0.001 (0.60)
–0.001 (0.65)
–0.002 (3.59)***
–0.0001 (0.56)
0.00002 (0.08)
Time-invariant locational-geographical factors
0.02 (2.70)***
The number of village/ group Enterprises in 1997 (each)
–0.0004 (0.38)
0.002 (1.13)
0.0001 (0.16)
–0.0002 (0.74)
0.0002 (0.80)
0.004 (4.88)***
0.0003 (0.06)
–0.001 (1.57)
0.003 (1.96)**
–0.0001 (0.23)
–0.001 (6.23)***
0.0004 (1.83)*
0.001 (1.30)
–0.0005 (0.12)
–0.20 (0.36)
–1.1 (1.12)
–0.40 (1.72)*
–0.04 (0.38)
0.15 (1.30)
1.3 (3.34)***
50.7 (21.26)***
–0.03 (0.32)
–0.1 (0.94)
–0.08 (2.04)**
0.04 (2.61)***
0.00003 0.00
0.1 (1.33)
2.7 (7.12)***
0.12 (0.24)
–1.1 (1.25)
–0.36 (1.67)*
–0.04 (0.47)
0.04 (0.34)
0.9 (2.42)**
48.3 (21.78)***
Continued
–0.29 (1.98)**
0.2 (0.70)
0.04 (0.57)
–0.04 (1.32)
0.11 (3.63)***
0.3 (3.23)***
–0.3 (0.50)
9780230_274587_10_cha09.indd 136
6/23/2010 12:17:16 PM
(3.49)*** 7041
7041
(10.02)***
–56
Yes
Yes
–0.02 (1.81)*
–0.01 (1.64)
–0.001 (1.15)
–0.17 (8.70)***
Co-funded
7041
(13.65)***
–59
yes
yes
0.01 (0.94)
0.01 (1.30)
0.001 (0.81)
–0.11 (7.50)***
Above funded
7041
(2.38)**
–6180
Yes
Yes
–7.6 (1.33)
–1.9 (0.48)
0.41 (1.09)
–35.5 (3.94)***
All sources
Data source: Authors survey.
7041
(0.38)
157
Yes
Yes
–2.6 (2.85)***
0.9 (1.41)
0.12 (1.98)**
–3.7 (2.58)**
Villagefunded
7041
(1.59)
–3844
yes
yes
–3.9 (0.73)
–5.1 (1.37)
–0.05 (0.14)
–25.3 (3.01)***
Co-funded
7041
(3.55)***
–2494
Yes
Yes
–1.1 (0.73)
2.3 (2.12)**
0.34 (3.35)***
–6.6 (2.70)***
Above funded
Annual public goods investment in village leader’s term
Notes: Absolute value of t statistics in parentheses, * significant at 10%; ** significant at 5%; *** significant at 1%.
7041
(15.78)***
Observations
–16
–130
Constant
Yes
Yes
–0.03 (2.58)***
Yes
–0.04 (2.15)**
Prior occupation of village leader (pure farmer = 1, others = 0)
0.02 (3.51)***
Yes
0.02 (1.48)
Education of village leader
0.002 (2.86)***
Time trend
0.001 (1.21)
Age of village leader
–0.06 (3.90)***
Villagefunded
Province dummy
–0.34 (11.92)***
All sources
Annual public-goods projects in village leader’s term
Tax-for-fee reform (after = 1, before = 0)
Time-variant factors
Table 9.3 Continued
Community Governance and Public-Goods Investments
137
the baseline model, we produce an initial set of estimates of whether or not directly elected village leaders have an effect on village public-goods investment, while holding a number of other observables (Vi) constant. The sample for use with the baseline model includes more than 7000 pairs of election/appointment-investment observations from 2448 villages. There are eight different versions of the baseline model equation, which use eight different measures of the dependent variables: four using the number of public goods (total number; from village-funded-only projects; jointly fund projects and government-funded-only projects) and four using the level of investment. These results are reported in Table 9.3. In running the model in equation 1 (in its eight different versions), the regression equations appear to perform well. In particular, a number of the control variables behave as expected. First, as also found in Zhang et al. (2006), poorer villages in China are shown to receive more projects that are funded from government only; richer villages invest more of their own funds into village-funded-only public goods (Table 9.3, row 2, columns 2, 4, 6, and 8). Villages with higher levels of out-migration, as expected, have lower levels of public-goods investment because there is little incentive to invest when a large number of the villagers are living outside the village (row 7, columns 2 and 6). The results also imply that political connections are important ways to raise investment; villages with more fellow villagers working in town and county governments have more public-goods investments (Table 9.3, row 9). Most importantly, the baseline results show the positive relationship between directly electing a village’s leader and the number of projects and level of investment into public goods (Table 9.3, row 1). All of the signs are positive and most of them are significant at the 10 percent level. Above all, when village leaders are elected directly, there are more village-funded-only projects and a higher level of investment: and the t-ratios of the coefficients are high (row 1, columns 2 and 6). Importantly, although there is a positive coefficient, the average magnitude of the effects are relatively small (about 0.1 project more per year and around 10,000 yuan more investment – a level equal to about 10 per cent of the mean number of projects and level of investment.
9.9
Fixed-effects model
Although the baseline analysis controls for a number of village-specific characteristics, it is still difficult, if not impossible, to control for all of the effects. If the unobserved factors also help to explain public-goods investment, and if they are in some way correlated with the propensity of a village to directly elect their leader, the baseline results could be subject to endogeneity-induced bias. Because our sample contains multiple observations on elections and investment over time, we can use a fixed-effects model
9780230_274587_10_cha09.indd 137
6/23/2010 12:17:16 PM
9780230_274587_10_cha09.indd 138
6/23/2010 12:17:17 PM
(2.55)**
0.001 (1.59) –0.01 (0.46) –0.0005 (0.03)
–0.05 (3.36)*** yes yes –18 (4.27)*** 7,041
(4.67)***
–0.001 (0.44) 0.005 (0.20) –0.03
(0.77)
–0.30 (9.99)*** yes yes –129 (15.3)*** 7,041
–0.14 (7.12)*** yes yes –52 (9.36)*** 7,041
(1.40)
–0.003 (2.23)** –0.01 (0.36) –0.03
(3.38)***
0.08
Co-funded
–0.11 (7.31)*** yes yes –59 (14.1)*** 7,041
(0.34)
0.0005 (0.51) 0.02 (1.34) 0.01
(2.40)**
0.04
Govt-funded
–24.7 (3.09)*** yes yes –4,805 (2.17)** 7,041
(0.68)
0.10 (0.19) –1.1 (0.18) –6.4
(3.83)***
35.5
All sources
–3.2 (2.55)** yes yes –245 (0.71) 7,041
(2.22)**
0.25 (3.26)*** –0.3 (0.36) –3.2
(1.55)
2.2
Village funded
Data source: Authors’ survey.
Notes: Absolute value of t statistics in parentheses, * significant at 10%; ** significant at 5%; *** significant at 1%.
Observations
Education of village leader Prior occupation of village leader (pure farmer = 1, others = 0) Tax-for-fee reform (after = 1, before = 0) Village fixed effect Time trend Constant
0.04
Villagefunded
0.16
All sources
–15.4 (2.08)** yes yes –2,144 (1.05) 7,041
(0.33)
–0.31 (0.67) –1.3 (0.23) 2.9
(3.39)***
29.0
Co-funded
–6.1 (2.55)** yes yes –2,416 (3.62)*** 7,041
(2.15)**
0.15 (1.02) 0.5 (0.29) –6.0
(1.52)
4.2
Above funded
Yearly public-goods projects in village leader’s term Yearly public-goods investment in village leader’s term
Fixed-effect analysis of the impact of directly electing village leader on rural public investment in rural China (1998–2003)
The way village leader acceded to office (0 = appointed,1 = elected directly) Age of village leader
Table 9.4
Community Governance and Public-Goods Investments
139
by including a set of 2448 village dummy variables on the right-hand side. The village fixed-effects control for village-specific factors that are fixed over time. We also include a time trend to control common time-varying factors. Despite the potential problem of endogeneity, in fact, when comparing the estimated coefficients from the baseline model with those of the fixedeffects models, there is little change in sign or level of significance (and the magnitudes are about the same, at most, only a bit lower – Table 9.4, row 1). In other words, we find strong evidence that in villages that have directly elected leaders the number of public-goods projects and the level of investment are higher. Moreover, the effect appears in all eight versions of the equations (although the t-ratio is low when looking at the level of investment into village-funded-only and government-funded-only public goods). Like the case of the baseline model (though most of the coefficients are larger in fixed-effect estimates), the magnitude of the effect is still relatively reasonable. In this way, however, our results are somewhat like those of Rosenzweig and Foster (2003), Pande (2003) and Chattopadhyay and Duflo (2004) in that they find positive effects of changes in elections and political democracy (the effect are significant but the magnitude of the effect is not very large).
9.10 Pork-barrel politics, Chinese style While there are a number of reasons why village leaders who are directly elected might invest more in public goods, we examine if village leaders are doing so as part of their strategy to get reelected. The assumption behind our test of motivating behavior is that village leaders gain utility (in some way) from holding their position. If, as shown above, villagers who vote demand public-goods, then leaders who are able to influence the direction of spending and increase allocations to public-goods investment should be expected to enhance their reelection hopes. While most of the literature focuses on the actions that leaders take to improve their chances for reelection, it is still not clear whether these effects exist in countries that are in the early stages of democracy. China provides an especially interesting case for several reasons. First, there are formal legal provisions in China that are pushing villages to hold direct elections for their leaders. Second, there are sharp differences among communities in the way that villages select their leaders (Kelliher, 1997; Kennedy, 2002; Morduch and Sicular, 2000; Oi, 1989; Oi et al., 2000; Paster and Tan, 2000; Tan, 2004; Tsai, 2002; Shi, 2004). This diversity of governance environment is actually a benefit to econometricians who are interested in exploring these questions empirically. Third, since there are no term limitations on village leaders in rural China, all incumbents will always face incentives to deliver those things demanded by their constituents.
9780230_274587_10_cha09.indd 139
6/23/2010 12:17:17 PM
140
Renfu Luo et al.
To test whether or not an incumbent’s reelection depends on his/her ability to deliver public-goods, the first step is to create a subsample of election cycle terms in which the village leader who was elected runs for office again and faces an election. Of the 5553 election terms in which the village leader was directly elected (which as shown above account for 79 percent of election terms), we need to drop a number of them. First, nearly 12 percent of the election terms were prematurely interrupted because the leader resigned or was asked to step down. Second, around 7 percent of village leaders were promoted to party secretary so they did not run in the next direct election. After dropping these election terms, the election terms that remain in our sample include village leader incumbents who ran for office and won, and incumbents who ran for office and lost. In addition, we also look for comparison purposes at a sample of village leaders who were appointed and were subsequently either reappointed, or not. In looking at the empirical relationship public-goods investment behavior and the propensity of a leader to be reelected, we also will look at two time periods:the first three years of our data (1998–2000) and the last three years (2001–3). This dimension of the analysis is needed because it is possible that before 2000 villagers were more concerned about issues such as tax reform and family planning policy implementation (Li, 2006; Yep, 2004). Villagers started to become highly interested in public-goods investment after 2000, when many of the tax and fee burdens began to be reduced. In looking at the raw descriptive statistics, there is little evidence that is consistent with an argument that those village leaders who invest more in public goods are rewarded with reelection. In fact, before 2001 we see that those village leaders who invested in fewer projects with less money were elected. The same results hold for those leaders who were reappointed. Even after 2001 there is little evidence that reelection was related to public-goods investment (although the direction of the shift between 1998–2000 and 2001–3 supports the hypothesis). However, if we divide the sample into villages that have relatively few (Table 9.5, panel A) and relatively more projects (panel B) and those that have relatively small (Table 9.5, rows 1 and 3) and relatively large (rows 2 and 4) projects, we find that the descriptive statistics, in fact, do suggest that there is a relationship between public-goods investment and reelection. In villages with few projects, when the average size of project was small, leaders were reelected 62 percent of the time (Table 9.5, row 1). In contrast, when there were more projects, and the average size of project was small, the reelection rate of leaders rose to 73 percent (and was statistically significant). When the average size project was large, however, the share of terms in which leaders were reelected actually fell (from 66 to 61) as the number of projects increased (rows 2 and 4). One interpretation of this finding is
9780230_274587_10_cha09.indd 140
6/23/2010 12:17:17 PM
Community Governance and Public-Goods Investments
141
Table 9.5 Relationship between public-goods investment, size of the project and reelection in rural China, 1998–2003 Average size (level) of the average public-goods investment (per project)
Reelected (%)
Not reelected (%)
Frequency
Panel A. In villages that implemented 0.5–1 public investment projects per year Lowa Higha
62 66
38 34
224 223
Panel B. In villages that implemented more than 1 public investment project per year Lowa Higha
73 61
27 39
200 200
Note: a A project’s average-sized investment was considered “low” if it fell beneath the median (about 50,000 yuan per project); it was considered “high” if the average size of the project fell above the median. Data source: Authors’ survey.
that while villagers like more projects (and reward leaders with reelection if there are more investment projects) they do not like it when the size of the project is too big (since perhaps there is a perception that money and labor were being wasted). Hence, in those villages where projects were kept relatively small, villagers reelected those leaders who allocated more funding to public-goods investment.
9.11
Multivariate analysis
To understand whether of not investment into public goods helps village leaders become reelected, we use a similar empirical strategy as Besley and Case (1995a) to test for the nature of the reelection incentives. The empirical function is as follows: Pit g0 g1Iit1 g2Vit mi mit it
(2)
where Pit is an indicator variable measuring if the incumbent leaders was reelected/reappointed (equaling one if the incumbent in village i during term (t 1) was reelected as village leader in term t, and otherwise was zero). On the right-hand side of the equation (2), Iit1 is a variable that is a measure of the number of public-goods projects (or level of investment) that the incumbents in village i during term (t 1) implemented. In this section 1 is
9780230_274587_10_cha09.indd 141
6/23/2010 12:17:17 PM
9780230_274587_10_cha09.indd 142
6/23/2010 12:17:18 PM
–0.0001 (0.70)
Average investment level of govt-funded public projects
Average investment level per project
–0.0002 (1.75)*
(0.91)
0.00002 (0.17)
0.03 (0.55)
–0.17 (2.20)**
0.08 (0.75)
(0.70)
0.00000 (0.00)
–0.02 (0.56)
0.0004
–0.02 (0.54)
Reappointed (1 = yes, 0 = no)
–0.0001
0.01 (0.37)
Co-funded public projects, number
Public investment level
0.08 (1.84)*
Village-funded public projects, number
0.04 (1.96)** 0.09 (2.05)**
0.04 (1.72)*
Govt-funded public projects, number
Public projects number
Reelected (1 = yes, 0 = no)
Table 9.6 Fixed-effects estimation of the effect of the annual number of public-goods projects (amount of investments) in latest term and reelection (reappointment) of village leader in rural China
9780230_274587_10_cha09.indd 143
6/23/2010 12:17:18 PM
2686
2686
6.83 (10.55)***
yes
yes
0.25 (1.19)
–1.67 (12.75)***
–0.75 (4.35)***
–0.74 (4.32)***
–0.75 (4.34)***
2686
6.90 (10.64)***
Yes
Yes
0.24 (1.17)
875
3.03 (4.15)***
yes
yes
0.45 (1.34)
875
3.00 (4.11)***
yes
yes
0.46 (1.36)
875
3.03 (4.14)***
yes
yes
0.45 (1.34)
875
2.91 (4.00)***
Yes
Yes
0.47 (1.39)
–0.72 (4.20)***
0.00001 (0.12)
–0.0001 (1.69)* –1.68 (12.80)***
–0.0003 (0.40)
–0.0004 (1.07)
Notes: Absolute value of t statistics in parentheses, * significant at 10%; ** significant at 5%; *** significant at 1%. a This is unbalanced panel data, with some villages having one observation and the others having two observations.
Observationsa
2686
6.89 (10.63)***
6.82 (10.53)***
Constant
yes yes
yes
yes
0.25 (1.19)
0.25 (1.20)
Village fixed effect
–1.68 (12.83)***
–1.66 (12.70)***
Year dummy
Education of incumbent
Age of incumbent
Average investment level of village-funded public projects Average investment level of co-funded public projects
144
Renfu Luo et al.
the coefficient of interest. We also include a number of control variables, Vit, such as the age and education of the village incumbent. We also add village fixed-effects and year effects (trend variables) in order to hold constant all non-time-varying unobservables. According to our results of the estimation of the effect of public-goods investment on the possibility of being reelected, there is a positive relationship. Although the magnitude of the coefficient is small, it is statistically significant in the reelection equations (Table 9.6, row 1, columns 1 and 3). In other words there is evidence that villagers rewarded incumbents with reelection if they implemented more public-goods projects. And, while the level of investment is not statistically significant (column 2, row 5), when we hold the average size of the project constant (rows 6 to 9), the number of government-funded-only projects and village-funded-only projects become a significant determinant of reelection. In the same way as seen in the descriptive statistics, villages apparently do not like village leaders to invest in frivolously large projects, but when accounting for the size of the project, the higher the number of projects, the higher the probability that incumbents are reelected. In other words, our results are consistent with the story that China’s rural leaders also are engaged in “Pork-barrel politics.” Interestingly, when we look for a relationship between public-goods investment and the propensity to be reappointed, it is absent (Table 9.6, columns 5–8). Unlike the case of direct election where leaders are trying to please villagers who demand public-goods projects, it seems that public-goods projects/investments are not a major concern in the decision to reappoint or not. When village leaders are appointed, there appears less of an imperative to deliver “pork” to their villagers – at least in the form of public-goods investments.
9.12
Conclusion
The purpose of this chapter was to provide empirical evidence pertinent to theories of rural governance, using survey data from rural China. Based on a theoretical framework developed by Rosenzweig and Foster (2003), we created a conceptual model for China and produced the key hypothesis that was tested in the chapter: Village elections lead to greater investment in rural villages. We also sought to push the work of RF further by testing whether or not elections have been able to discipline the behavior of village leaders. There are two major empirical findings from our analysis. First, compared to the traditional modes of governance in rural China, the shift towards elections has resulted in more public-goods projects/investments. In particular, we found that when village leaders were elected directly, they implemented more public-goods projects/investments. In contrast, when village leaders were appointed investments into public goods was lower. In other words,
9780230_274587_10_cha09.indd 144
6/23/2010 12:17:18 PM
Community Governance and Public-Goods Investments
145
according to our analysis, we cannot reject the hypothesis that village elections lead to greater investment into roads, irrigation, and drinking water. Second, this chapter also found that villagers can indeed use elections to discipline their incumbent leaders. Specifically, our results indicate that since village leaders face reelection, they have an incentive to act in a way consistent with their constituents’ demands. In other words, when there are two leaders – both facing reelection – and one has invested more in public goods during his term than the other, the village leader with the record of greater public investment was shown to have a greater propensity to be reelected. At the same time, we do not find any significant evidence that those village leaders who were appointed had any incentive to promote infrastructure investment. These results, perhaps more than anything, also can be seen to build on the work of Yujiro Hayami and his studies of rural communities (e.g., Hayami and Kikuchi, 2000). In much of the research by Hayami, it was found that institutions in rural communities played a key role in producing more effective irrigation systems and forming types of land–labor relations that enhance the outcomes of the poor. In this chapter we find that community governance reform in China is also associated with rising investments into public goods, such as roads, irrigation works, and drinking-water systems. This is another example of how communities’ demands influence the efficiency and equity of those living and working in developing countries.
References Aoki, M. and Hayami, Y. (2001) Communities and Markets in Economic Development (Oxford: Oxford University Press). Bardhan, P. and Mookherjee, D. (2000a) “Decentralizing Anti-Poverty Program Delivery in Developing Countries,” mimeo, Institute for Economic Development (Boston). Bardhan, P. and Mookherjee, D. (2000b) “Capture and Governance at Local and National Levels,” American Economic Review, 90 (2), 135–9. Bardhan, P. and Mookherjee, D. (2001) “Corruption and Decentralization of Infrastructure Delivery in Developing Countries,” mimeo, Institute for Economic Development (Boston University). Banerjee, A., Somanathan, R. and Iyer, L. (2005) “History, Social Divisions, and Public Goods in Rural India,” Journal of the European Economic Association, 3 (2–3), 639–47. Besley, T. and Burgess, R. (2002) “The Political Economy of Government Responsiveness: Theory and Evidence From India,” The Quarterly Journal of Economics, 117 (4), 1415–51. Besley, T. and Case, A. (1995a) “Incumbent Behavior: Vote-Seeking, Tax-Setting, and Yardstick Competition,” The American Economic Review, 85 (1), 25–45. Besley, T. and Case, A. (1995b) “Does Electoral Accountability Affect Economic Policy Choices? Evidence from Gubernatorial Term,” Quarterly Journal of Economics, 110 (3), 769–98. Besley, T. and Coate, S. (2003a) “Elected versus Appointed Regulators: Theory and Evidence,” Journal of the European Economic Association, 1 (5), 1176–206.
9780230_274587_10_cha09.indd 145
6/23/2010 12:17:18 PM
146
Renfu Luo et al.
Besley, T. and Coate, S. (2003b) “Centralized versus Decentralized Provision of Local Public Goods: A Political Economy Approach,” Journal of Public Economics, 87 (12), 2611–37. Brandt, L., Huang, J., Li, G. and Rozelle, S. (2002) “Land Rights Rural China: Facts, Fictions and Issues,” The China Quarterly, 47, 67–97. Brandt, L. and Turner, M. (2003) “The Usefulness of Corruptible Elections,” William Davidson Institute Working Papers Series, No. 2003–602, William Davidson Institute at the University of Michigan Stephen M. Ross Business School. Chattopadhyay, R. and Duflo, E. C. (2004) “The Impact of Reservation in the Panchayati Raj: Evidence from a Nationwide Randomized Experiment,” Economic and Political Weekly, 39 (9), 979–86. Deacon, R. T. (2003) “Dictatorship, Democracy, and the Provision of Public Goods. University of California at Santa Barbara,” Economics Working Paper Series, No. 1172. Dethier, J. J. (1999) “Governance and Economic Performance: A Survey,” ZEF discussion paper on development policy, No. 5. Fan, S., Zhang, L., and Zhang,X. (2004) “Reforms, Investment, and Poverty in Rural China,” Economic Development and Cultural Change, 52 (2), 395–421. Fiorina, M. P. (1981) Retrospective Voting in American National Elections (New Haven, CT: Yale University Press). Fork, A. and Wong, C. (2005) “Extending Public Finance to Rural China,” background for presentation at the MOF-World Bank International Seminar on Public Finance for Rural Areas. Hayami, Y. and Kawagoe, T. (1993) The Agrarian Origins of Commerce and Industry: A Study of Peasant Marketing in Indonesia (London: Macmillan). Hayami, Y. and Kikuchi, M. (2000) A Rice Village Saga: Three Decade of Green Revolution in the Philippines (London: Macmillan). Kelliher, D. (1997) “The Chinese Debate over Village Self-government,” China Journal, 37, 63–86. Kennedy, J. J. (2002) “The Face of ‘Grassroots Democracy’ in Rural China: Real versus Cosmetic Elections,” Asian Survey, 42 (3), 456–82. Key, V. O., Jr. (1966) The Responsible Electorate: Rationality in Presidential Voting, 1936–1960 (Cambridge, MA: The Belknap Press of Harvard University Press). Li, L. C. (2006) “Embedded Institutionalization: Sustaining Rural Tax Reform in China,” The Pacific Review, 19 (1), 63–84. Louie, K. S. (2001) “Village Self-Governance and Democracy in China: An Evaluation,” Democracy, 8 (4), 134–54. Miguel, E. and Gugerty, M. K. (2005) “Ethnic Diversity, Social Sanctions, and Public Goods in Kenya,” Journal of Public Economics, 89 (11–12), 2325–68. Morduch, J. and Sicular, T. (2000) “Politics, Growth, and Inequality in Rural China: Does it Pay to Join the Party?” Journal of Public Economics, 77, 331–56. O’Brien, K. J. (1994) “Implementing Political Reform in China’s Villages,” The Australian Journal of Chinese Affairs, 32 (9), 33–59. Oi, J. C. (1989) State and Peasant in Contemporary China (Berkeley, CA: University of California Press). Oi, J. C. and Rozelle, S. (2000) “Elections and Power: The Locus of Decision-Making in Chinese Villages,” China Quarterly, 162, 465–89. Pande, R. (2003) “Can Mandated Political Representation Increase Policy Influence for Disadvantaged Minorities? Theory and Evidence from India,” American Economic Review, 93 (4), 1132–51.
9780230_274587_10_cha09.indd 146
6/23/2010 12:17:18 PM
Community Governance and Public-Goods Investments
147
Rosenzweig, M. R. and Foster, A. D. (2003) “Democratization, Decentralization, and the Distribution of Local Public Goods in a Poor Rural Economy,” BREAD Working Paper, No. 010. Shi, W. (2004) “The Development of Grassroots Democratic Elections in China,” Social Science in China, Special Issues: Peasant Worker Migration to China’s Cities, 1, 113–25. Tan, Q. (2004) “Building Institutional Rules and Procedures: Village Election in China,” Policy Sciences, 37, 1–22. Tsai, L. L. (2002) “Cadres, Temple and Lineage Institutions, and Governance in Rural China,” China Journal, 48, 1–27. World Bank (2004) World Development Report: Making Services Work for Poor People (Washington, DC: World Bank). Yan, Y. (1992) “The Impact of Rural Reform on Economic and Social Stratification in a Chinese Village,” The Australian Journal of Chinese Affairs, 27, 1–23. Yep, R. (2004) “Can ‘Tax-for-Fee’ Reform Reduce Rural Tension in China? The Process, Progress, and Limitations,” The China Quarterly, 177, 42–70. Zhang L., Luo, R., Liu, C. and Rozelle, S. (2006) “Investing in Rural China: Tracking China’s Commitment to Modernization,” Chinese Economy, 39 (4), 57–84. Zhang, X., Fan, S., Zhang, L. and Huang, J. (2004) “Local Governance and Public Goods Provision in Rural China,” Journal of Public Economics, 88 (12), 2857–71.
9780230_274587_10_cha09.indd 147
6/23/2010 12:17:18 PM
10 Poverty, Politics and Projects under Community Participation in Zambia* Alain de Janvry, Hideyuki Nakagawa, and Elisabeth Sadoul
10.1 Understanding the role of the rural community in development Among the many seminal contributions made by Yujiro Hayami to our understanding of international agricultural development, his research on the rural community has been hailed by the profession for its unique insights (Hayami, 1988). This is in part because of the distinct approach he followed for this research, analyzing in great detail village microcosms over long periods of time to extract generic propositions about the role of the community in development. He labeled it as the approach of a “bonsai economist.” According to him, “the bonsai lover makes a lifelong effort to perfect the miniature shape of a tree to reflect the spirit of the universe” (Hayami and Kikuchi, 2000). As with the effort of patience and abstraction of the bonsai lover, his work has been grounded in a lifelong effort to observe and document the transformation of particular rural communities, while projecting his observations into broad conceptual frameworks capturing the role of the community in development. Following the bonsai principle, reality is simplified to the maximum. No unnecessary embellishments: it is a parsimonious model, capturing the essence of what matters in order to understand the whole. In Yujiro Hayami’s work, the model used to conceptualize the role of the rural community in development was triangular: it specified the mutually reinforcing roles of the state, the market, and the community. Much of his work showed how the community could play an important role in making * This paper was prepared for presentation at the Festschrift Workshop in honor of Professor Yujiro Hayami held at the National Graduate Institute for Policy Studies, Tokyo, February 27–28, 2009. Authors’ email addresses are:
[email protected],
[email protected],
[email protected] 148
9780230_274587_11_cha10.indd 148
6/23/2010 12:17:27 PM
Poverty, Politics and Projects
149
markets work better. This is because distinctive features of the community such as local information, trust and norms, social capital, and interlinked transactions can help reduce transactions costs, adverse selection, and moral hazard, overcoming many market failures. But the community can also play an important role in making the state work better. This is because it can assume functions on behalf of or instead of the state. This includes the delivery of local public-goods, securing a minimum subsistence for all, and regulating externalities across community members. The state can in turn work with the community in helping it to more effectively assume these functions. This includes reinforcing local administrative capacities, promoting intercommunity cooperation when there are economies of scale beyond the community, and changing community institutions, for instance property rights, or facilitating more participatory decision-making processes. In other chapterss, we made use of two fundamental propositions derived from Yujiro Hayami’s approach to the role of the community in development. One is that the African rural community tends to be driven by strong egalitarian norms in ensuring subsistence for all (Hayami, 2006; Platteau and Hayami, 1998). While effective in securing survival in a context of broadly-shared poverty and high risks, egalitarian norms can also become an obstacle to growth. Community conservatism in proactively preserving the sharing norm can oppose initiatives that would induce growth at the cost of economic differentiation. We used this to explain why it is difficult for market-oriented producer organizations, the potential source of net social gains, to emerge in community contexts with high social homogeneity (hence a strong capacity to enforce sharing norms) and high exposure to natural risks (hence the need for sharing mechanisms) (Bernard et al., 2009). The other proposition is that strong egalitarian norms can defeat elite and leader capture in community organizations, making these organizations effective for the delivery of pro-poor local public-goods. However, lack of financial resources and weak management capacity tend to limit their economic effectiveness to “penny capitalism” (Tax, 1963). As a consequence, institutional richness under the form of widely prevalent and formalized community organizations does not always translate into significant economic benefits, and the rural community remains an underused opportunity for rural development (Bernard et al., 2008). In this chapter, we address the community–state linkage through the experience of devolving to communities greater roles in delivering local public-goods. This is done by analyzing the introduction of a sequence of institutional arrangements giving increasing roles to the community in time and across space. We use this process to reveal econometrically two functions of the community in development documented by Yujiro Hayami at the bonsai level: to help achieve progressive targeting of public expenditures across communities and households; and to place pro-poor checks on
9780230_274587_11_cha10.indd 149
6/23/2010 12:17:27 PM
150 Alain de Janvry et al.
electoral clientelism in using public resources to reward votes or to mobilize electoral support. In line with Yujiro Hayami’s results, both conclusions are optimistic about the potential for complementary roles between state and community in development.
10.2 Giving greater roles to the community in delivering local public-goods Many developing countries that implemented structural adjustment programs in the last two decades established Social Investment Funds (SIF) to deliver small-scale social and economic investment projects to communities in order to mitigate the negative impacts of adjustment on the poor (Rawlings et al., 2001). Under this approach, community-based organizations (CBOs) were invited to submit project proposals to an ad hoc central agency that had the responsibility of selecting from among proposals received, providing budget support, and monitoring the implementation of projects (Figure 10.1). This approach to the delivery of local public-goods was based on the presumption that delegating to CBOs the responsibility of identifying projects for investment in local public-goods would improve their poverty-reduction value, compared to a centralized approach. Which communities were selected for support, and which projects within these SIF (SRP I and SRP II, ZAMSIF district level 1)
CDD (ZAMSIF district level 2)
Decentralization (ZAMSIF district level 5)
Central agency
Central agency
Central agency
Project appraisals
Project proposals
Budget support
Community Investment Fund allocation
Budget support
District (Appraise and monitor projects)
Project proposals
District (Select projects, disburse, and monitor projects) Project proposals
Community-based organizations
Community-based organizations
Level 1: 22
Number of districts by level under ZAMSIF Level 2: 24
Figure 10.1
Selection, budget support Community-based organizations
Level 3: 19 Level 4: 3 Level 5: 4
Operational rules under SIF, CDD and decentralization
9780230_274587_11_cha10.indd 150
6/23/2010 12:17:27 PM
Poverty, Politics and Projects
151
communities, thus depended on the composition and capacity of CBOs in formulating demands and getting them approved by the central agency. Because community capacity is quite uneven, and citizen interests very unevenly represented by CBOs, the question arose whether this approach would effectively serve the poor across and within communities. For this reason, the centralized SIF approach gradually evolved towards a community-driven-development (CDD) approach giving greater roles to a structured representation of local interests. Local representation could be through the district, often with the assistance of an appointed development council in charge of representing the interests of the communities (e.g., wards, the lowest formal administrative unit in Zambia) in district-level decision-making. The district was then in charge of receiving proposals from CBOs, appraising projects, transmitting project appraisals to the central agency, and monitoring implementation of projects that had received budgetary support from the central agency (Figure 10.1). The CDD approach met with considerable support among international development agencies as a pro-poor instrument to allocate funds to the provision of local public-goods and to support local productive projects. There were high expectations that the community could assume functions that markets and states could not perform so well (Platteau and Abraham, 2002). It was estimated that, in 2003, up to one-fourth ($7 billion) of the World Bank’s annual disbursements were occurring through this modality (Mansuri and Rao, 2004). As community capacity and social capital were gradually strengthened by purposeful interventions of the central agency, community roles could be correspondingly increased. From playing an intermediary role between CBOs and the central SIF agency, districts were gradually entrusted with greater roles in resource disbursement. The ultimate step was the decentralization of fiscal resources, eventually transferring to districts financial resources to be competitively allocated to the projects submitted by CBOs. Yet, in spite of large-scale implementation of the modality, there are still few evaluations of the SIF-CDD-decentralization approach to the delivery of local public-goods, leaving strong reservations about the actual effectiveness of the approach and conditions for success (Mansuri and Rao, 2004), and inviting calls for much greater efforts at conducting rigorous impact evaluations of CDD-based projects (World Bank, 2006). Lack of rigorous evaluations largely derives from the great degree of difficulty in establishing identification of causalities in such complex processes. The main issue that requires evaluation is the trade-off that may exist between taking advantage of local information and local social capital to better target resources on the poorest communities; and the risks of capture of benefits by richer communities and local elites, and use of project resources for clientelistic gains by local politicians (Platteau and Gaspard, 2003). While asymmetrical information is expected to make centralized approaches less effective, asymmetrical
9780230_274587_11_cha10.indd 151
6/23/2010 12:17:27 PM
9780230_274587_11_cha10.indd 152
6/23/2010 12:17:27 PM
125
1997 SRP II 86
1998
0
1999
27
2000
58
2001
Timeline for Zambia’s social programs and elections
November Results not available
December 2795 899/1287 MMD, UNIP
December 6921 572/1206 MMD, FDD UNIP, UPND
116
263
2002 2003 ZAMSIF 119
2004
90
December 72/148**
December MMD
2006
1426
Total
December 4091 727/1394 MMD,PF, UDA
2005
Notes: * The second largest party (UNIP) boycotted the 1996 election. ** Elections in 2 constituencies were postponed due to the death of candidates.
Figure 10.2
Local elections (councilors) Number of candidates MMD victories Main parties
December 69/150
96
1996
November 131/150*
104
1995
October 125/150
36
1994
Parliamentary elections (MP) MMD victories
104
1993
December MMD
199
1992 SRP I
November MMD*
1
1991
October MMD
2
1990
Predidential elections Winning party % votes
Number of projects
Program
Poverty, Politics and Projects
153
power relations at the local level and weak accountability mechanisms in the use of public funds can open the door to elite and politician capture. Zambia offers a unique laboratory to analyze the targeting effectiveness and political economy of the SIF-CDD-decentralization approach. It has had a sustained 16 years of experience with community projects administered through three programs: two centrally managed SIFs, SRP I (Social Recovery Project I, 1990–4) and SRP II (Social Recovery Project II, 1995–9), and a CDD program, ZAMSIF (Zambia Social Investment Fund, 2000–5), with increasing devolution of control over resources to districts, ultimately converging into full decentralization (Figure 10.2). To analyze program effectiveness, we use administrative data on the projects supported by the programs and match them with information from the 1990 and 2000 population censuses. To analyze clientelism and political rewards, we match four national election results, in 1991, 1996, 2001, and 2006, with project placement data under the three programs to examine whether there is an association between electoral outcomes and project allocation, or between project placement and electoral rewards. ZAMSIF introduced a systematic effort at enhancing the administrative capacity of districts in managing participatory projects and in tracking progress. The degree of devolution of control over resources was increased with district capacity. This allows us to assess whether greater district capacity and the associated greater control over resources is associated with greater program effectiveness in reaching the poor and with different electoral rewards. To examine the overall targeting performance of the centrally administered SRPs and increasingly decentralized ZAMSIF, we analyze the relation between poverty and projects across programs at different levels of community participation. We then analyze the relation between politics and projects to quantify both patronage as votes are rewarded by projects, and voter support as projects are rewarded by votes. We finally ask whether greater politicization of projects in local elections as community participation rises leads to politics improving or worsening the poverty-reduction value of projects.
10.3 What we know about poverty, politics, and projects in the SIF-CDD approach Previous research gives mixed evidence on the effectiveness of community participation in the allocation of local public-goods. Faguet (2004) suggested that municipal decentralization in Bolivia allowed better adjustment of public expenditures to the specificity of community needs in human-capital formation and social services, particularly in the poorest municipalities. This result is important as it shows that distributional gains can indeed be achieved for the poor through decentralization, as opposed to inviting capture by local
9780230_274587_11_cha10.indd 153
6/23/2010 12:17:27 PM
154
Alain de Janvry et al.
elites as had been broadly presumed. Paxon and Schady (2004), using data for Peru, found that SIFs were better at reaching the poorest communities but not the poorest within communities. In this case, capture was an intracommunity phenomenon as project benefits did not reach the neediest. Galasso and Ravallion (2005) for Bangladesh, Alderman (2002) for Albania, Arcand and Bassole (2006) for Senegal, and Bardhan and Mookherjee (2006) for West Bengal all found the opposite, namely that inequality in the appropriation of benefits of decentralized programs was more an intercommunity than an intracommunity problem. In Senegal, it is local politics that is important in influencing the allocation of projects across communities in a CDD approach to local public-goods, leading to eventual “community capture” (Arcand and Bassole, 2006). In Thailand, it is the differential social capital endowment of a community, in particular the strength of its organizations and networks, and not necessarily its wealth, that determines its success in attracting CDD projects (Chase et al., 2006). Combining inter- and intracommunity distribution of Zambia’s SIF benefits across households, Chase and Sherburne-Benz (2001) observed that the program reached poor households in rural areas, but not in urban areas where benefited households were better off than urban households overall. Van Domelen (2002) analyzed six social funds in Armenia, Bolivia, Honduras, Nicaragua, Peru, and Zambia and concluded favorably that investments were generally targeted at the poorer districts and benefited more the poorer households, and that they largely responded to stated community needs. But Rao and Ibáñez (2003) found that, in the Jamaica social fund, gains were extensively captured by the elites within the community. Hence, evidence about the needs-capture trade-off is mixed at both the community and the individual level, and outcomes depend on local specificities and program implementation. Several studies addressed the political economy of public investment when local participation is involved. One class of theories – the core-supporter model – suggests that a politician allocates investments to the communities where he has received the strongest electoral support as rewards for their loyalty (Cox and McCubbins, 1986; Dixit and Londregan, 1996; Verdier and Snyder, 2002). In this case, projects follow votes as clientelistic rewards, redistributing income toward proven supporters. This is the politics of patronage that seems to apply well to the African context, where use of state resources to maintain governmental power by rewarding core supporters is widespread (Bratton and van de Walle, 1997). The other class of theories – the swing-voter model – predicts that politicians target communities with more swing voters whose political behavior could be influenced by the public-goods provision (Dixit and Londregan, 1996). In this case, votes are expected to follow projects, and projects to be targeted where they can have the greatest impact on swaying votes. Empirical results testing these theories show that both can apply. Finan (2005) finds that Brazilian federal deputies allocated public works in 1996
9780230_274587_11_cha10.indd 154
6/23/2010 12:17:27 PM
Poverty, Politics and Projects
155
and 1999 to the municipalities where they had received more electoral support in 1994, supporting the core-supporter model. By contrast, Dahlberg and Johansson (1999) found evidence that Swedish incumbent governments distributed temporary grants for ecologically sustainable development programs to regions where there were more swing voters. Manacorda et al., (2009) find that beneficiaries of a cash transfer program in Uruguay are more likely than nonbeneficiaries to favor the current government relative to the previous one. There is also evidence that government spending increases the incumbent’s vote share in US congressional elections (Levitt and Snyder, 1997) and likewise for political incumbents in Spain (Sole-Olle and Sorribas-Navarro, 2008). For Mexico, Rodriguez-Chamussy (2009) finds that Oportunidades expenditures at the municipal level not only increased the share of votes for the incumbent presidential party in municipal elections (a legitimate reward since the program is fully under the authority of the federal government), but also for the local incumbent party even if in opposition to the presidential party, indicating the capacity of local mayors to successfully engage in credit-claiming for benefits delivered by others in their municipality. Whether decentralization reinforces the two-way link between projects and politics is an important question as it may suggest a trade-off between more efficient delivery of local public-goods (based on local information and local social capital) and greater use of public budgets to reward or mobilize votes.
10.4
Zambia’s social programs and their political context
The SIFs in Zambia had the objective of funding small, simple, and locally initiated projects to mitigate the hardships that poor communities were facing under structural adjustment following the debt crisis. A detailed explanation of how it functioned is provided by Chase and Sherburne-Benz (2001). Under SRPI and SRPII, the Micro Project Unit (MPU) of the Ministry of Finance was the central agency in charge of project selection and resource allocation. Staff of the provincial offices of MPU provided information on potential funding opportunities to the communities and local authorities. To enhance the likelihood that projects submitted for funding dealt with the perceived needs of those in the poor communities, a participative approach was adopted through CBOs. Under this scheme, potential beneficiary organizations were asked to express and prioritize their needs, and were encouraged to formulate and submit project proposals for potential funding. SIFs in Zambia were transformed into CDDs with the initiation of ZAMSIF in 2000. Compared to the former SRPs, the ZAMSIF program allowed more decentralization in the decision-making process with regard to the initiation and selection of projects by interested communities. Efforts were made
9780230_274587_11_cha10.indd 155
6/23/2010 12:17:27 PM
156
Alain de Janvry et al.
to train the districts and the communities to enable them to increase their roles in program appraisal and monitoring. The administrative capacity of a district was rated from level 1 to 5 based on a set of indicators updated on an annual basis through the district assessments conducted by the Provincial Assessment Committees. Depending on these levels, a district could enjoy an increasing degree of autonomy in the management of funds originating in the central agency.39 For districts with administrative capacity level 1, ZAMSIF supported all community project activities in the same fashion as under SRP II. Districts with level 2 were given the responsibility to field appraise, desk appraise, and monitor projects. Districts with level 3 had the role of costing and budgeting projects, and of supporting implementation with monitoring and technical advice. Districts with level 4 could approve projects up to US$50,000, and communities were accountable for these expenditures to the district. Finally, districts with level 5 received an annual Community Investment Fund allocation from the central agency, and could disburse and monitor all community projects. As discussed later, district councils mainly consist of elected representatives including the district’s Member of Parliament (MP) and the locally elected councilors, giving politicians a high level of control over project allocation. We have an exhaustive list of the 1426 projects approved under SRP I, SRP II, and ZAMSIF.40 It gives information on the location, type, and budget of each project. Table 10.1 reports the number of projects by sector and program period. It shows that the vast majority of projects under SRP I and SRP II concerned education, health, and water/sanitation. Projects became more diversified under ZAMSIF, including projects in support of income generation. Table 10.2 shows the distribution of projects received by wards. In all three program periods, there were still many wards which had never received a project: 81 percent, 80 percent, and 68 percent of all wards under SRP I, SRP II, and ZAMSIF, respectively. After 16 years, 50 percent of the wards had had at least one project. Among those that participated, most of them did so only once or twice, with very few receiving three or more projects. We combine the project data with the 1990 and 2000 population censuses. The censuses provide detailed ward-level information on individual characteristics such as education, fertility, ethnicity, religion, migration status, and housing conditions, as well as asset-holding status, source of access to water, and availability of latrine facilities. Since the three most frequent types of projects are in education, health, and water/sanitation, we focus on the ward-level characteristics related to these sectors. Data show that the level of satisfaction of needs in these three sectors is low. For example, school enrollment for ages 7 to 15 is 59 percent, access to latrines 66 percent, and access to water 71 percent.
9780230_274587_11_cha10.indd 156
6/23/2010 12:17:28 PM
Poverty, Politics and Projects
Table 10.1
157
Number of projects implemented under SRP I, SRP II and ZAMSIF Program SRP I
Sector
Number
Education Health Water supply/ Sanitation Community welfare Environment/ Income Food security/ Market HIV/AIDS Infrastructure Roads Others Total
SRP II
ZAMSIF
Total
Share Share Share Share (%) Number (%) Number (%) Number (%)
266 69 31
70 18 8
266 38 16
82 12 5
255 134 62
44 23 11
787 241 109
61 19 9
–
–
–
–
22
4
22
2
–
–
–
–
10
2
10
1
–
–
–
–
16
3
16
1
– 12 – 2 380
– 3 – 1 100
– 4 – 2 326
– 1 – 1 100
24 11 39 – 576
4 2 7 – 100
24 27 39 4 1282
2 2 3 0 100
Table 10.2 Distribution of wards by number of projects SRP I Number of Projects
Number
SRP II
ZAMSIF
All programs
Share Share Share Share (%) Number (%) Number (%) Number (%)
0 1 2 3 4 or more
1042 170 52 15 9
81 13 4 1 1
1029 215 34 7 3
80 17 3 1 0
875 305 86 15 7
68 24 7 1 1
640 331 177 81 59
50 26 14 6 5
Number of wards Wards with projects Total projects
1288
100
1288
100
1288
100
1288
100
246
259
413
648
361
317
553
1177
Note: * This number excludes supplementary projects.
9780230_274587_11_cha10.indd 157
6/23/2010 12:17:28 PM
158 Alain de Janvry et al.
10.5 How effective are SIF and CDD in reaching poor communities? We use descriptive statistics in Table 10.3 to characterize the degree of progressiveness in allocating projects across wards under the three programs. Ward characteristics are for those which had projects in that particular year. Progressiveness is measured by the wards’ position on two welfare indicators – school enrollment ages 7 to 12 and household assets (ownership of a radio and/or a television) – relative to the district average. District averages are from the 1990 census for the years 1990 to 1999, and from the 2000 census data for the years 2000 onwards. Positive signs thus indicate regressive allocations, and negative signs progressive allocations. The data show a
Table 10.3
Targeting of projects by ward characteristics across programs Characteristics of wards with projects in deviation from the corresponding district mean (%)
Year 1990 census average SRP I
SRP II
2000 census average ZAMSIF
1990 1991 1992 1993 1994 1990–4 average 1995 1996 1997 1998 1999 1995–9 average 2000 2001 2002 2003 2004 2005 2000–5 average
Number of School enrollment Household projects rate ages 7–12 assets
2 1 199 104 36 342 104 96 125 86 0 411 27 58 116 263 119 90 673
50 1.1 11.1 3.7 3.9 5.5 4.0 3.1 5.2 4.0 2.6
15 –5.6 2.7 1.0 0.9 3.2 1.1 1.0 2.1 1.4 0.6
3.7
1.3
56 5.5 5.1 2.2 0.5 –1.2 0.5 1.1
25 3.6 1.5 0.3 –0.7 –2.3 0.3 –0.3
Note: Deviations from district mean use the 1990 census data for years 1990 to 1999 and the 2000 census data for years 2000 and onward. Household assets are the percentage of households with radio and television ownership.
9780230_274587_11_cha10.indd 158
6/23/2010 12:17:28 PM
Poverty, Politics and Projects
159
Characteristics of wards with projects in deviation from the corresponding district mean (%) (Weighted averages for enrollment and assets in parentheses) 6 5 4 3 2 1 0 1990 −1
1992
1994
1996
1998
−2 −3
SRP I (4.0, 1.1)
SRP II (3.7, 1.3) School enrollment
Figure 10.3
2000
2002
2004
2006
ZAMSIF (1.1, −0.3) Asset ownership
Targeting of projects by ward characteristics across programs
clear evolution from regressive targeting under the SRP toward progressive targeting under ZAMSIF. For example, average deviation in household assets was 1.1 percent under SRPI, 1.3 percent under SRPII, and –0.3 percent under ZAMSIF. For school enrollment, average deviation was 4.0 under SRPI and 3.7 percent under SRPII, falling to 1.1 percent under ZAMSIF. While ZAMSIF also started as regressive, most of its progressiveness occurred in the last three years. As can be visualized in Figure 10.3, we do not see any trend toward more progressive targeting within the SRPI and II periods. The equation we estimate to identify the role of district capacity level is: p Pwdp ⫽ b1p Wwd ⫹ b2p Wwd L2d ⫹ b3p Wwd L3d ⫹ υdp ⫹ « wd
p
where Pwd is a binary indicator of project placement in ward w in district d under program p (SRPI, SRPII, or ZAMSIF). Wwd is the relative welfare of ward w in district d, L2d is administrative level 2 in district d (with level 1 as the p base), L3d are administrative levels 3 to 5 in district d, d a district fixed effect p under program p, and « wd a random term. Administrative levels are only effective under ZAMSIF. If greater community participation with district levels 2 and above influences the placement of projects in wards with lower relative welfare in the district, b2p and b3p (for p = ZAMSIF) will be different from zero. By contrast, levels should not matter for project placement in a ward under SRP I and II, in which case b2p and b3p ⫽ 0. This counterfactual on the role of administrative levels under SRP I and II provides strong identification of the role of community responsibility under ZAMSIF.
9780230_274587_11_cha10.indd 159
6/23/2010 12:17:29 PM
160
Alain de Janvry et al.
Table 10.4 welfare
Poverty targeting: Choice of wards within districts based on relative
Dependent variable: Dummy of project placement in ward SRP I
SRP I&II
ZAMSIF
School School School enrollment enrollment enrollment rate ages Household rate ages Household rate ages Household assets assets assets 7–12 7–12 7–12 Ward relative welfare in district Interacted with level 2 Interacted with levels 3 to 5 Ln (population in ward) Percent rural population Constant Observations R-squared Fixed effect Tests Effect: Base ⫹ level 2 F test: Base ⫹ level 2 (p-value) Effect: Base ⫹ levels 3 to 5 F test: Base ⫹ level 3 to 5 (p-value) SRP I&II and Zamsif level 1 Zamsif level 2 Zamsif levels 3 to 5
0.347** (0.171)
⫺0.0974 (0.390)
0.327*** (0.125)
⫺0.0221 (0.284)
0.414** (0.209)
0.642** (0.262)
0.0185 (0.223) ⫺0.0602 (0.215)
0.483 (0.448) 0.323 (0.459)
⫺0.0137 (0.163) ⫺0.0496 (0.157)
0.275 (0.326) 0.228 (0.335)
⫺0.365 (0.278) ⫺0.260 (0.266)
⫺0.977*** (0.295) ⫺1.038*** (0.313)
0.0337** (0.0155)
0.0403*** (0.0155)
0.0345*** (0.0113)
0.0402*** (0.0113)
0.119*** (0.0229)
0.121*** (0.0227)
⫺0.0634 (0.0605) ⫺0.0834 (0.176) 1234 0.148 District
⫺0.117* (0.0637) ⫺0.0898 (0.179) 1234 0.141 District
⫺0.0753* (0.0441) ⫺0.129 (0.128) 2468 0.084 District
⫺0.126*** (0.0464) ⫺0.131 (0.130) 2468 0.078 District
0.0761 (0.0643) ⫺0.683*** (0.254) 1288 0.101 District
⫺0.00957 (0.0705) ⫺0.612** (0.257) 1288 0.108 District
0.3655
0.3856
0.3133
2468
0.049
⫺0.335
0.0160
0.115
0.00472
0.155
0.806
0.0487
0.2868
0.2256
0.2774
0.2059
0.154
⫺0.396
0.0468
0.447
0.00855
0.342
0.402
0.0652
Regressive
Neutral
Regressive
Neutral
Regressive
Regressive
Regressive
Neutral
Regressive
Neutral
Neutral
Progressive
Regressive
Neutral
Regressive
Neutral
Neutral
Progressive
Notes: Standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1.
9780230_274587_11_cha10.indd 160
6/23/2010 12:17:30 PM
Poverty, Politics and Projects
161
Results in Table 10.4 show that SRP I and II were regressive in terms of relative welfare of wards measured by school enrollment and household assets. There is no difference across wards even in the districts that are later to become levels 2, and 3 to 5, under ZAMSIF. In districts with administrative capacity level 1, ZAMSIF operates just like the SRPs, and it is as regressive as the SRPs in placing projects across wards. Shift to progressiveness occurs with the responsibility of wards in appraising and monitoring projects (level 2) and continues to increase with greater decentralized selection, disbursement, and monitoring (levels 3 to 5). F-tests of significance of the effect of welfare on the probability of receiving a project under the different programs show that regressiveness in districts level 1 was transformed under ZAMSIF into neutrality using the school enrollment criterion, and into progressiveness using the household asset criterion. This change occurs with level 2 and is reinforced with greater district roles at levels 3 to 5. The fact that there was no effect of future administrative levels on project placement under SRP, and that it did not occur for level 1 under ZAMSIF, gives us a good counterfactual indicating that it is devolution of roles to the community that affected the shift out of regressiveness, rather than either a simple crosssection difference or a time trend. These results give strong support to the proposition that greater community participation, here at the district level, leads to better targeting of projects on relatively poorer wards within the district.
10.6 The political economy of SIF and CDD Are projects rewarded by granting votes (the swing-voter model), and are votes received rewarded by projects (the core-supporter model)? Does greater community participation, in making the link between projects and politics more precise, increase the political value of projects for elected officials? And does this potentially greater politicization associated with greater community participation favor or hurt the poverty value of projects? The information we have on voting patterns in national and local elections allows us to answer these questions. After 20 years of being a one-party state, Zambia in 1991 held its first multi-party elections for both the presidency and parliament. These twolevel national elections were held again in 1996, 2001, and 2006. The same party, the Movement for Multiparty Democracy (MMD), won all four elections, although with a declining percentage of votes as the number of candidate parties increased.41 As seen in Figure 10.2, the MMD won 125, 131, 69, and 72 of the 150 seats in parliament in 1991, 1996, 2001, and 2006, respectively. In addition to the four national elections, local elections were held in 1992 (results not available), 1998, 2001, and 2006 to elect one councilor
9780230_274587_11_cha10.indd 161
6/23/2010 12:17:30 PM
162
Alain de Janvry et al.
Table 10.5
From votes to projects in presidential elections
Dependent variable: Project budget per capita (log) 1991 election, 1992–4 projects District (1)
(2)
Constituency (3)
(4)
1996 election, 1997–8 projects District (5)
(6)
Constituency (7)
(8)
Number of 0.481*** 0.541** 0.549*** 0.647*** 0.0257 –0.226 0.560** votes MMD (0.171) (0.209) (0.210) (0.228) (0.308) (0.334) (0.244) received (log) Population –0.726*** –0.705*** –0.532*** –0.447** –0.184 –0.257 –0.521*** (log) (0.146) (0.156) (0.193) (0.202) (0.232) (0.235) (0.194) Number of votes MMD received in 1991 (log) Number of votes MMD received in 1996 (log) Number of votes MMD received in 2001 (log) Population in 1991 (log) Population in 1996 (log) Population in 2001 (log) Ethnic –0.854 –0.908 0.523 fragmentation (0.671) (0.809) (1.053) Percent rural –0.362 –0.180 3.270*** population (0.826) (0.736) (1.201) Child school –1.430 2.026 0.600 enrolment (1.181) (1.327) (1.878) Asset –0.827 –5.626 16.94** (4.489) (3.421) (6.498) Observations 49 49 124 124 57 57 146 R-squared 0.617 0.668 0.159 0.193 0.169 0.330 0.139 Fixed effects Province Province Province Province Province Province Province
0.942*** (0.268) –0.420** (0.192)
–1.766** (0.818) 1.531** (0.659) 3.201** (1.262) –0.478 (2.946) 146 0.225 Province
Notes: Standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1.
9780230_274587_11_cha10.indd 162
6/23/2010 12:17:30 PM
Poverty, Politics and Projects
2001 election, 2002–4 projects District
Constituency
(9)
(10)
(11)
(12)
0.278* (0.149)
0.322** (0.156)
0.347** (0.169)
0.398** (0.175)
–0.334*** –0.265** (0.122) (0.122)
–0.222 (0.185)
–0.0710 (0.197)
Pooled election and projects District
Constituency
(13)
(14)
(15)
(16)
0.481** (0.198)
0.493** (0.207)
0.549*** (0.194)
0.652*** (0.195)
0.0257 (0.212)
0.0438 (0.221)
0.560** (0.217)
0.753*** (0.224)
0.278 (0.263)
0.328 (0.275)
0.347 (0.215)
0.460** (0.215)
–0.726*** –0.705*** –0.532*** (0.170) (0.174) (0.179) –0.184 –0.167 –0.521*** (0.159) (0.163) (0.173) –0.334 –0.349 –0.222 (0.215) (0.221) (0.235) –0.507** –0.449 –0.342 (0.244) (0.400) (0.349) –0.426 0.136 0.514 (0.470) (0.523) (0.481) 1.020 1.026 0.432 (0.668) (1.034) (0.730) –2.449* –1.473 0.979 (1.376) (1.447) (1.709) 72 72 150 150 178 178 420 0.176 0.294 0.050 0.083 0.472 0.481 0.274 Province Province Province Province Province- Province- ProvinceElection Election Election
9780230_274587_11_cha10.indd 163
163
–0.464*** (0.178) –0.468*** (0.171) 0.0311 (0.241) –0.856** (0.362) 0.621* (0.362) 1.888*** (0.672) –1.400 (1.195) 420 0.306 ProvinceElection
6/23/2010 12:17:31 PM
164
Alain de Janvry et al.
from each ward for the district council. As required by law, a government council is formed in each district to supervise the provision of services such as infrastructure to community members. A district government council is composed of (1) the members of parliament in the district, (2) two representatives of the traditional chiefs, appointed by the chiefs in the district, and (3) all the elected councilors in the district representing their respective wards. The government council is thus a highly political institution, opening the door for greater decentralization in the allocation of local public-goods to be accompanied by greater use of projects for clientelistic or electoral gains. A district consists of one to seven constituencies – 2.7 on average – each of which elects one member of parliament. A district also consists of 11 to 30 wards – 19 on average – each of which elects one councilor. For each candidate, the election data provide the name, party membership, and number of votes received.42 The national majority-party (MMD) won 70 percent of the ward elections in 1998, 47 percent in 2001, and 52 percent in 2006. So, political competition exists at the local level, and can be influenced by project placement in wards. From votes to projects: The politics of patronage at the national level We analyze in Table 10.5 the relation between the number of votes received in presidential elections by the party in power, the MMD, and project budgets per capita at the level of a district or a constituency in the three years following the election. This is done for the presidential elections held in 1991, 1996 (with distorted results due to a boycott of the election by the UNIP, the second-largest party), and 2001. The estimated equation for a particular election is: Pip ⫽ aVip ⫹ Xip g⫹m p⫹« ip , where Pip is project budget per capita in log during the three years after the election in district or constituency i in province p, Vip is the number of votes received by the MMD in logs, Xip is a set of characteristics of the district or constituency, mp is a province fixed effect, and «ip a random term. Results show that votes are strongly rewarded by projects, suggesting that projects are used for political patronage, as predicted by Bratton and van de Walle (1997). These effects are robust to the introduction of controls. The relation tends to be stronger at the constituency than at the district level, which makes sense: at the constituency level, which are subsets of districts, projects can be more precisely targeted to votes received. The effects are large. In two constituencies with the same population size, a 10 percent increase in votes received is associated with a 4 to 9 percent increase in budgets per capita. When introducing district levels
9780230_274587_11_cha10.indd 164
6/23/2010 12:17:31 PM
Poverty, Politics and Projects
Table 10.6 elections
165
From projects to votes in the 2006 presidential
Dependent variable: Percent of votes received by the MMD Constituency
Project budget per capita (log) in 2003–5 Ethnic fragmentation
(1)
(2)
1.591 (1.306)
2.248** (1.116) 17.17*** (6.145) 16.57** (7.770) 8.342 (16.15) -53.21** (22.62) 41.95*** (14.06) 150 0.720 Province
Percent rural population Child school enrollment Household assets Constant Observations R-squared Fixed effects
58.63*** (4.990) 150 0.592 Province
Notes: ***p<0.01, **p<0.05, *p<0.1. Standard errors in parentheses.
of administrative capacity (not reported here), we see no effect of greater community participation on project rewards. This also makes sense as decentralization should affect local politics, not the way presidential votes are rewarded. We can thus conclude that the core supporter/patronage model seems to apply as votes are rewarded by projects, and that this effect persists across programs from SIF to CDD and also as community participation rises. From projects to votes: The politics of voter support at the national level We analyze in Table 10.6 the relation between project budget per capita at the constituency level over the two years before the election and the percentage of votes received in that constituency by the MMD in the presidential election. We can only do this for the 2006 election and projects in 2003–5, as previous elections have too few prior projects. The estimated equation is: Vip ⫽ b Pip ⫹ Xip g ⫹ m p⫹« ip
9780230_274587_11_cha10.indd 165
6/23/2010 12:17:32 PM
166
Alain de Janvry et al.
with the same notations as before. Results show that projects do buy votes for the MMD, but the relation is not as strong as that going from votes to projects. The relation is, however, meaningful. Doubling the project budget at the level of the constituency is associated with an increase of 2.2 percent in votes received. Decentralization does not affect this result (not reported here). We thus conclude that projects are meaningful to obtain more electoral support in presidential elections for the ruling party. Yet, the politics of patronage seems to be a more prevailing practice than using projects to acquire the support of swing voters.
10.7
Decentralization and local politics
Decentralizing to communities the role of project approval and budgeting should shift the political economy of projects from presidential elections under SIF and centralized CDD to local elections under a decentralized CDD, with district administrative levels 3 to 5. Local elections allow wards to elect one councilor who will represent the ward in the district council. With greater decentralization, the district council acquires an increasing role in the placement and budgeting of projects across wards, enhancing the importance of local elections in attracting projects or in rewarding local politicians for placing projects in the ward. From votes to projects: The politics of patronage at the local level The equation we estimate is similar to that for votes to projects in national elections, with fixed effects now at the district level. In this case, we consider the results of local elections and the allocation of projects across wards within a district. Results in Table 10.7 show that it is in wards located in urban districts that projects follow local votes when there is greater community participation (district levels 3 to 5). Votes received by the local MMD candidate in wards located in urban districts (there were no urban nonMMD candidates in the 2001 elections) are rewarded by larger project budgets per capita for the ward. Also important to note for the interpretation of the results is that wards with greater poverty, as measured by household assets, are rewarded by more project expenditures, as observed in Table 10.4, even after controlling for political support to the local MMD candidate. From projects to votes: The politics of voter support at the local level We estimate again an equation where decentralization (district levels 3 to 5) should matter in mobilizing votes under ZAMSIF (2006 elections), but not under SRPII (1998 elections). Results in Table 10.8 show that project budgets received by urban wards with more community participation (levels 3 to 5) are rewarded by votes for the MMD candidate for councilor representing
9780230_274587_11_cha10.indd 166
6/23/2010 12:17:33 PM
Poverty, Politics and Projects
167
Table 10.7 From votes in local elections to projects Dependent variable: Project budget per capita (log)
Rural districts
2001 local elections
(1)
Percentage of votes –0.00269 received by the (0.00420) MMD candidates Percentage of votes received by the MMD candidates in district levels 1 and 2 Percentage of votes received by the candidates from district majority party in district levels 1 and 2 Percentage of votes received by the MMD candidates in district levels 3 to 5 Percentage of votes received by the candidates from district majority party in district levels 3 to 5° Ethnic fragmentation Child enrollment
(2)
0.181* (0.097) 968 0.045 District
(3)
(4)
(5)
(6)
0.0198 (0.0122) –0.00169 (0.00574)
–0.00198 (0.00576)
0.00644 (0.0182)
0.00173 (0.0185)
0.00388 (0.00513)
0.00420 (0.00517)
–0.0159 (0.0195)
–0.0144 (0.0195)
–0.00855 (0.00690)
–0.00886 (0.00695)
0.00871 (0.00632)
0.00878 (0.00636)
0.179* (0.097) 968 0.048 District
0.182 (0.381) 0.130 (0.596) –0.740 (1.047) 0.185* (0.101) 968 0.049 District
Household assets Population in ward (log) Observations R-squared Fixed effects
Urban districts
0.0647*** 0.0566** (0.0209) (0.0221)
–0.0543 (0.198) 230 0.073 District
–0.0114 (0.197) 230 0.105 District
0.296 (3.121) 1.434 (1.715) –1.808* (1.018) –0.0418 (0.196) 230 0.128 District
Notes: ***p<0.01, **p<0.05, *p<0.1. Standard errors in parentheses. °There are no non-MMD candidates in urban wards.
9780230_274587_11_cha10.indd 167
6/23/2010 12:17:33 PM
9780230_274587_11_cha10.indd 168
6/23/2010 12:17:33 PM
From projects to votes in local elections
(6)
–47.94* (24.45) 11.51 (14.71) 8.256 (8.965) 231 0.301 District
–1.084 –1.349* (0.837) (0.803)
(5)
–4.436 (4.318) –3.691 (7.686) –18.14 (11.15) 897 897 897 231 231 0.280 0.280 0.292 0.223 0.223 District District District District District
–0.235 (0.828)
–0.598 (0.820)
–1.120 (0.709)
(4)
–1.694 (1.298)
–0.194 (0.820)
–0.607 (0.818)
(3)
–1.212 (1.342)
–0.603 (0.579)
(2)
Urban districts
945 0.553 District
0.400 (0.293)
(1)
0.657 (0.449)
0.270 (0.383)
(3)
3.878 (3.381) –12.05** (5.857) –15.68* (8.600) 945 945 0.553 0.563 District District
0.569 (0.453)
0.278 (0.385)
(2)
Rural districts
226 0.320 District
1.154** (0.521)
(4)
0.128 (0.538)
(6)
226 0.334 District
57.43** (24.75) –23.67 (14.60) –14.69 (8.979) 226 0.508 District
2.840*** 1.832** (0.944) (0.831)
0.434 (0.617)
(5)
Urban districts
2003–5 projects, 2006 local elections
Notes: Weighted regressions using ward population. *** p<0.01, ** p<0.05, * p<0.1. Standard errors in parentheses.
Project budget per capita (log) Project budget per capita (log) in district levels 1 and 2 Project budget per capita (log) in district levels 3 to 5 Ethnic fragmentation Child school enrollment Household assets Observations R-squared Fixed effects
(1)
Rural districts
1996–8 projects, 1998 local elections
Dependent variable: Percentage of votes received by the MMD candidate for councilor
Table 10.8
Poverty, Politics and Projects
169
the ward. This did not happen under SRPII, nor under ZAMSIF in the districts with low community participation (levels 1 and 2). Decentralization thus makes local politics more relevant. We also see (columns 3 and 6 for the 2006 elections) that there is a positive correlation between votes for the MMD and poverty, particularly in wards located in rural districts: less child enrollment and lower household assets correlate with more votes for the MMD in local elections. We can thus conclude that greater community participation in the provision of local public-goods makes local politics more relevant. Votes follow projects, and projects follow votes in local elections when communities have a greater role to play. And votes for the MMD correlate to poverty. Hence, in this context, there is no contradiction between the politicization of local public-goods that follows greater community roles, and the pro-poor value of these public goods.
10.8 Conclusion: Community participation for poverty and politics Zambia offers a unique laboratory to analyze the role of community participation in the provision of local public-goods. Over a 16-year period, provision was increasingly shifted to community roles through the local administrative process and with the assistance of the state in strengthening local administrative capacity. Due to the administrative structure put into place to allocate funds at the district level, community participation is very closely linked to political outcomes. We use this experience to ask whether greater community responsibility leads to better poverty targeting across communities, both directly through the links between poverty and projects and indirectly through the link between poverty and politics. We can now assemble into a consistent story the evidence that we obtained on the role that greater community participation can have on the targeting of projects and the politicization of local votes. We saw in Table 10.4 that community participation helps poverty to bring projects. There is better targeting on poor wards with greater community participation. We saw in Table 10.7 that, with community participation, votes received by the local MMD candidate bring projects, particularly in the more urbanized wards. The question is whether the greater targeting on poverty and the greater politicization of local public-goods, both associated with greater community participation, are in conflict or complementary to pro-poor targeting. The answer is that politicization of local public-goods is here complementary to reaching the poorer wards. We saw in Table 10.8 that votes for the MMD and poverty correlate positively under community participation. Poverty brings projects, politics brings projects, and poverty correlates positively with politics, reinforcing the value of local politicization of projects when there is stronger community participation. This result vindicates the
9780230_274587_11_cha10.indd 169
6/23/2010 12:17:34 PM
170 Alain de Janvry et al.
detailed work of Yujiro Hayami on the rural community: greater community participation in complementing the role of the state – by participating to effect the delivery of local public-goods for poverty reduction – and greater state roles in reinforcing the capacity of the community to participate – by raising administrative capacity to levels 3 to 5 – lead to a situation where participation helps achieve greater pro-poorness in the delivery of public goods, both directly through targeting and patronage, and indirectly through poverty influencing local politics.
References Alderman, H. (2002) “Do Local Officials Know Something We Don’t? Decentralization of Targeted Transfers in Albania,” Journal of Public Economics, 83, 375–404. Arcand, J. L. and Bassole, L. (2006) “Does Community-Driven Development Work? Evidence from Senegal,” CERDI-Université d’Auvergne. Bardhan, P. and Mookherjee, D. (2006) “Pro-Poor Targeting and Accountability of Local Governments in West Bengal,” Journal of Development Economics, 79, 303–27. Bernard, T., de Janvry, A. and Sadoulet, E. (2009, forthcoming) “When Does Community Conservatism Constrain Village Organizations?” Economic Development and Cultural Change, 58 (4). Bernard, T., Collion, M. H., de Janvry, A., Rondot, P. and Sadoulet, E. (2008) “Do Village Organizations Make a Difference in African Rural Development? A Study for Senegal and Burkina Faso,” World Development, 36 (11), 2188–204. Bratton, M. and Van de Walle, N. (1997) Democratic Experiments in Africa (New York: Cambridge University Press). Chase, R., Christensen, R. and Thongyou, M. (2006) “Picking Winners or Making them? Evaluating the Social Capital Impact of CDD in Thailand,” The World Bank. Chase, R. and Sherburne-Benz, L. (2001) “Household Effects of African Community Initiatives: Evaluating the Impact of the Zambia Social Fund,” The World Bank. Cox, G. and McCubbins, M. (1986) “Electoral Politics as a Redistributive Game,” Journal of Politics, 48, May, 370–89. Dahlberg, M. and Johansson, E. (1999) “On the Vote-Purchasing Behavior of Incumbent Governments,” Department of Economics, Working Paper No. 24, University of Uppsala. Dixit, A. and Londregan, J. (1996), “The Determinants of Success of Special Interests in Redistributive Politics,” Journal of Politics, 58 (4), 1132–55. Faguet, J. P. (2004) “Does Decentralization Increase Government Responsiveness to Local Needs? Decentralization and Public Investment in Bolivia,” Journal of Public Economics, 88 (3–4), 867–93. Finan, F. (2005) “Political Patronage and Local Development: Evidence from Brazil,” Working Paper (Department of Economics, University of California at Los Angeles). Galasso, E. and Ravallion, M. (2005) “Decentralized Targeting of an Anti-Poverty Program,” Journal of Public Economics, 89 (4), 705–27. Hayami, Y. (1988) “Elmhirst Memorial Lecture: Community, Market, and State,” in A. Maunder and A. Valdés (eds.), Agriculture and Governments in an Interdependent World (UK: Dartmouth Publishing Company). Hayami, Y. (2006) “Communities and States in Rural Development: East Asia Compared With Africa” (Tokyo: Foundation for Advanced Studies on International Development).
9780230_274587_11_cha10.indd 170
6/23/2010 12:17:34 PM
Poverty, Politics and Projects
171
Hayami, Y. and Kikuchi, M. (2000) A Rice Village Saga: Three Decades of Green Revolution in the Philippines (USA: Barnes and Noble). Levitt, S. and Snyder, J. (1997) “The Impact of Federal Spending on House Election Outcomes,” Journal of Political Economy, 105 (1), 30–53. Mansuri, G. and Rao, V. (2004) “Community-Based and -Driven Development: A Critical Review,” World Bank Research Observer, 19 (1), 1–39. Paxon, C. and Schady, N. (2004) “Child Health and the 1988–92 Economic Crisis in Peru,” Policy Research Working Paper, WPS 3260 (World Bank). Platteau, J. P. and Abraham, A. (2002) “Participatory Development in the Presence of Endogenous Community Imperfections,” Journal of Development Studies, 32 (2), 104–36. Platteau, J. P. and Gaspard, F. (2003) “The Risk of Resource Misappropriation in Community-Driven Development,” World Development, 31 (10), 1687–703. Platteau, J. P. and Hayami, Y. (1998) “Resource Endowments and Agricultural Development: Africa vs. Asia,” in M. Aoki and Y. Hayami (eds.), The Institutional Foundation of Economic Development in East Asia (London: Macmillan). Rao, V. and Ibáñez, A. M. (2003) “The Social Impact of Social Funds in Jamaica: A MixedMethod Analysis of Participation, Targeting, and Collective Action in CommunityDriven Development,” World Bank Policy Research Paper 2970 (World Bank). Rawlings, L., Sherburne-Benz, L. and van Domelen, J. (2001) “Letting Communities Take the Lead: A Cross-Country Evaluation of Social Fund Performance,” World Bank, PREM Network. Rodriguez-Chamussy, L. (2009) “Local Electoral Rewards from Centralized Social Programs: Are Mayors Successful at Credit Claiming?” Department of Agricultural and Resource Economics, University of California at Berkeley. Sole-Olle, A. and Sorribas-Navarro (2008) “The Effects of Partisan Alignment on the Allocation of Intergovernmental Transfers, Differences-in-Differences Estimates for Spain”, Journal of Public Economics, 92 (12), 2302–19. Tax, S. (1963) Penny Capitalism: A Guatemalan Indian Economy (Chicago, IL: The University of Chicago Press). Van Domelen, J. (2002) “Social Funds: Evidence on Targeting, Impacts, and Sustainability,” Journal of International Development, 14, 627–42. Verdier, T. and Snyder, J. M. (2002) “The Political Economy of Clientelism,” CEPR Discussion Paper 3205. World Bank (2006) The Effectiveness of World Bank Support for Community-based and -driven Development: An OED Evaluation, The World Bank, Operations Evaluation Department.
9780230_274587_11_cha10.indd 171
6/23/2010 12:17:34 PM
11 Political Competition and the Impact of Local Governance on Pro-Poor Programs Pranab Bardhan and Dilip Mookherjee
11.1
Introduction
Yujiro Hayami has been a leader in village studies of institutions and community governance in Asia (Hayami, 1999 and 2001). In the context of community governmence, he has argued elegantly that communities play a central role in making local governments accountable and effectively implementing development programs, such as the distribution of agricultural inputs. One of the means for communities to play an important role in local governance in a democratic country is through voting in local government elections. The interesting question in this context is about the prospect of capture of local governments by rural elites. This chapter is in the Hayami tradition, and focuses on how political competition influences the effectiveness of pro-poor programs carried out by local village governments. Our data are primarily from village surveys in nearly 90 villages in West Bengal, India. Since the late 1970s, West Bengal has pursued an approach to rural development that involved reforms in agrarian institutions, particularly land reform and decentralized governance. These reforms were initiated by a Left Front government which has been in power in the state continuously since 1977. A functioning three-tier system – at the district, block, and village levels – of local governments (panchayats) elected every five years was instituted since 1978, well in advance of the 73rd and 74th Constitutional Amendments in the early 1990s which mandated such a system throughout India. In contrast to most other states (Karnataka and Kerala being notable exceptions), significant responsibilities were devolved to the West Bengal panchayats, including implementation of land reforms, selection of beneficiaries of various development schemes funded by upper-level governments, and administration of local infrastructure building programs. A significant land-reform program was administered by panchayats at the block and village levels, involving implementation of land ceilings, distribution of titles (patta) to surplus land to the landless, and regulation of tenancy 172
9780230_274587_12_cha11.indd 172
6/23/2010 5:05:10 PM
Political Competition & Impact of Local Governance
173
contracts. The magnitude of the land-reform effort exceeded those of other Indian states by a considerable margin: for example, Appu (1996, Appendix IV.3) estimates over 6.7 percent of operational land had been redistributed in West Bengal by the early 1990s, in contrast to less than 1 percent in most other states. Only one other state, Jammu and Kashmir, had redistributed more land than West Bengal; no other state has embarked on a tenancy registration program – called operation barga – on a similar scale. Over a million sharecroppers were registered by 1981, up from 242,000 in 1978, and increasing to almost one and a half million by 1990 (Lieten, 1992, Table 5.1). Registered sharecroppers were protected from eviction and from receiving crop shares below mandated minimum levels. The panchayats played a key role in identifying suitable beneficiaries of both kinds of land-reform program. By most accounts, these institutional reforms have transformed the balance of economic and political power away from big landowners and state bureaucrats, towards local governments more responsive to ordinary farmers and the poorer sections of rural society. Apart from land reform, responsibilities entrusted to the panchayats included selection of beneficiaries of two principal poverty alleviation schemes: the Integrated Rural Development Program (IRDP) which distributed subsidized credit to the poor to enable them to acquire incomegenerating assets, and employment programs: Food for Work (FFW), the National Rural Employment Program (NREP), and the Rural Labor Employment Guarantee Program (RLEGP) were merged in 1989 into the Jawahar Rozgar Yojana (JRY). The panchayats also played a leading role in choosing beneficiaries of other development programs such as farm extension programs, the distribution of subsidized agricultural inputs (in the form of mini-kits containing seeds, fertilizers, and pesticides), and miscellaneous welfare schemes (old-age assistance, disaster relief, housing programs for the poor, and so on). The panchayats administered construction of local infrastructure (roads, irrigation, public buildings), allocating funds across different projects and subsequently implementing these projects. The bulk of the funds for these programs were devolved to the panchayats under various schemes sponsored by the central and state governments. Despite the financing role played by upper-level governments, we shall henceforth refer to these programs as provided or administered by the panchayats, to recognize their key role in the delivery process (in contrast to most other Indian states, where these programs are implemented almost entirely by state government-appointed bureaucrats). Our focus is on the political economy of local governance in West Bengal: That is, how accountable were local governments? And what were the major determinants of land-reform implementation and agricultural input deliveries at the local level? The overriding concern commonly expressed about a strategy of rural development based on local democracy is the prospect of capture of local governments by rural elites, or their proneness
9780230_274587_12_cha11.indd 173
6/23/2010 5:05:10 PM
174
Pranab Bardhan and Dilip Mookherjee
to corruption. In our earlier theoretical work – Bardhan and Mookherjee (2000, 2005, 2006) – we have emphasized that local governments may be more or less prone to corruption than bureaucrats appointed by central or state governments, depending on how well institutions of local democracy function. Hence the relative effectiveness of a strategy of rural development centered on local democracy can be context-specific, and for any given setting can only be determined empirically. Since direct evidence about capture and corruption is difficult to obtain, we can infer these indirectly by studying how land-reform implementation and other developmental efforts of panchayats in different West Bengal villages have varied according to socioeconomic inequality and political competition. Concerning the political will of local governments to implement land reforms and target development programs on the poor, our empirical results lead us to reject polar hypotheses that local government officials were motivated by ideology or electoral opportunism alone. The ideology hypothesis predicts policy outcomes to be the result of intrinsic policy preferences of elected officials. On the basis of the notion that Left Front candidates are intrinsically more motivated to redistribute than those from its principal rival, the Indian National Congress, such an approach implies that panchayats will implement antipoverty programs more intensively if Left Front candidates win a larger fraction of panchayat seats. In contrast, the Downsian approach – see Downs (1957) – assumes all officials act out of opportunistic motives, to maximize their chances of being elected (or reelected, once in office). This leads to the convergence of policies adopted by officials of rival parties to that which is consistent with the preferences of the median voter. Hence the Downsian prediction is that actual policies will be independent of the party composition of panchayats. Both theories predict there will be greater antipoverty effort when there is greater inequality and poverty to start with (in the ideology-based theory, the redistribution desired by the Left intensifies; in the Downsian theory the median voter tends to be poorer). We find evidence in our statistical analysis below that party composition of panchayats mattered in a number of instances, but not in the way that the ideology hypothesis predicts. Antipoverty efforts often slackened when the Left Front became more entrenched, and when there was greater local poverty or inequality. These results are consistent with the hypothesis that there was some capture of the panchayats by local elites (e.g., medium-level landowners) or abuse of power among elected officials. There is also some evidence of strategic manipulation of resource flows to different village panchayats by upper-level government bodies. Nevertheless, the extent of political distortions owing to capture varied across different contexts. For instance, within villages, “private good” programs such as the IRDP credit program or agricultural mini-kits were well-targeted: more than 95 percent of these were allocated to small, marginal, and landless households. On the
9780230_274587_12_cha11.indd 174
6/23/2010 5:05:11 PM
Political Competition & Impact of Local Governance
175
other hand, employment-generating programs (such as JRY) were administered in a way that less employment was generated out of allotted funds when poverty within the village rose. Higher-level panchayats were subject to noticeably greater distortions: a jurisdiction at the village or Gram Panchayat (GP) level experiencing higher poverty and higher demographic share of scheduled castes (SCs) and tribes (STs) received significantly less resources as a whole from upper-level bodies.
11.2 A description of the sample village characteristics All the results reported in this chapter pertain to a sample of 89 villages. This sample originated in a stratified random sample of villages drawn for the purpose of cost of cultivation surveys by the West Bengal Department of Agriculture, covering all the districts of the state (excepting Calcutta and Darjeeling, in which there is minimal production of foodgrains). For these sample villages, detailed farm-level records of inputs and outputs were procured for three successive five-year panels, spanning 1981 to 1996. The farm data is complemented by village data collected from a variety of sources. We carried out household surveys in these villages, and collected data relating to composition and activities of local governments, spanning the period 1978 to 2004. These enabled us to assess changes in the distribution of land, literacy, caste, and occupations in the village; composition of elected GPs; details of infrastructure programs; and yearly budgets of the GPs. We visited local land-reform offices to obtain data on yearly land-reform implementation (land titles distributed and tenants registered, including names of beneficiaries and cultivation areas involved). Visits to local lead banks and block development offices generated yearly data on distribution of IRDP credit and agricultural mini-kits in each village. GP records yielded the yearly allocation of spending and the scale of various infrastructure projects. We also collected data on rainfall from local recording centers of the state meteorological department, leading economic indicators at the district or regional level from published statistics of the state government, and outcomes of elections to the state and national legislatures in each constituency spanning the sample areas. Summary statistics concerning the villages in our sample are provided in Tables 11.1 and 11.2. The 89 villages are located in 57 village government (Gram Panchayat [GP]) jurisdictions. Each GP consists of 10 to 20 elected members of a council governing administration of the jurisdiction of the GP, which usually consists of 8 to 15 villages or mouzas. On average each district comprises 20 blocks and 200 GPs. Each district (or zilla) has a single zilla parishad (ZP), the top tier of the panchayat system, and each block has a panchayat samiti (PS), the middle tier. The top official at each level is an ex-officio member of the next higher level; other officials at each tier are elected directly by voters. For the most part, we focus on the composition
9780230_274587_12_cha11.indd 175
6/23/2010 5:05:11 PM
176
Pranab Bardhan and Dilip Mookherjee
Table 11.1 District-wise allocation of sample villages
District 24 Parganas (N) 24 Parganas (S) Bankura Birbhum Bardhaman Cooch-Behar Hooghly Howrah Jalpaiguri Malda Midnapur Murshidabad Nadia Dinajpur Purulia West Bengal
Number of villages in sample
Left Front percentage of seats in GP (average 1978–98)
6 8 5 6 8 8 6 4 5 2 8 6 5 4 8 89
56 54 87 56 84 85 70 79 74 60 78 46 79 51 62 69
Table 11.2 Village characteristics in sample villages, 1978 and 1998
Number of households Operational land-household ratio (acre/hh) % households landless % households marginal (0–2.5 acres) % households small (2.5–5 acres) % households medium (5–12.5 acres) % households big (12.5– acres) % land small % land medium % land big % poor households low-caste % small households illiterate % big households illiterate % households in nonagricultural occupation Population-bank ratio
1978
1998
228 1.54 47.3 35.2 11.2 4.7 1.6 56.7 23.9 19.5 38.3 44.1 4.4 41.1 41.6
398 0.87 52.3 39.1 6.4 2.0 0.3 73.9 18.5 7.6 39.8 31.9 3.2 51.4 23.1
Notes: “Poor” household means landless or marginal landowner. “Up to small” household means landless, marginal, or small landowner. All land information pertains to distribution of cultivable non-patta land that is owned. Source: Indirect household survey. Population-bank ratio from West Bengal Economic Review, various years.
9780230_274587_12_cha11.indd 176
6/23/2010 5:05:11 PM
Political Competition & Impact of Local Governance
177
of seats in GPs as they are the main implementing agencies at the ground level (e.g., with respect to the selection of beneficiaries of various developmental schemes and infrastructure projects within villages). Moreover, party composition of GPs and higher tiers are highly correlated with one another. The 20-year period 1978 to 1998 witnessed four successive elected bodies in each GP, each with a five-year term (which we sometimes refer to as a timeblock). The Left Front coalition won an absolute majority in approximately three-quarters of the elected GPs, with a mean seat proportion of 69 percent. The main opposition party was the Indian National Congress and its various offshoots (such as the Trinamul Congress, which broke away for the 1998 elections). Most electoral constituencies witnessed a contest between the Left and the Congress (or the Trinamul Congress): there were hardly any three-way contests. In most cases, these two parties collectively garnered more than 90 percent of all elected positions. The dominance of the Left Front was greater at higher tiers; for example, the mean Left Front share in ZP positions during the period was 86 percent. Table 11.2 shows the principal demographic and asset distribution changes in the sample villages between 1978 and 1998. The number of households almost doubled, the result of population growth, household subdivision, and in-migration. Illiteracy rates fell, especially among the poor (landless or marginal landowners). The incidence of nonagricultural occupations among household heads rose from 41 percent to 51 percent. The distribution of cultivable non-patta land (i.e., excluding land distributed through the land reforms) changed in a variety of ways: landlessness increased, while the distribution of land among landowners became more equal. The proportion of households without any such land or with marginal holdings below a hectare (2.5 acres) increased by almost 10 percent. In this sense, poverty increased. On the other hand, the proportion of land in small holdings (below 2 hectares) rose by 17 percent, indicating a reduction of land inequality among those holding land. These occurred as a result of market sales of land, and fragmentation of land holdings accompanying household subdivision. Table 11.3 indicates the extent of land-reform implemented. In our sample, approximately 6.1 percent of cultivable land area had been registered under Operation Barga, and 5.4 percent in the patta distribution program, by 1998. Approximately 15 percent of households in 1998 had received pattas, and 4.4 percent had been registered under Operation Barga. The former program benefited three times as many households as the latter, essentially because the average size of a patta plot was one-third that of an average plot registered under the Barga program. Undoubtedly this was a large program, one of the largest land-reform initiatives in India in recent memory. Also distinctive was the involvement of panchayats in this process, who were instrumental in mobilizing mass participation in village meetings to
9780230_274587_12_cha11.indd 177
6/23/2010 5:05:11 PM
178 Pranab Bardhan and Dilip Mookherjee
Table 11.3 Land reforms 1978 Average
1998 Average
16.4* 1.4 4.9 2.7 2.4 3.1 43.4
15.3 5.4 14.9 4.2 6.1 4.4 51.2
% operational land vested % operational land distributed % households receiving pattas % operational land leased % operational land with registered barga % households with registered bargadar % tenants registered
Notes: Average across sample villages, weighted by operational land areas. * Only available for 34 villages. Source: Block land records offices (land reforms); indirect household survey (operational land and tenancy).
Table 11.4 Trends in public supplies of agricultural inputs
Mini-kits per household IRDPa per household GP irrigation expenditureb GP road expenditurec GP employment man-days per household Area irrigated by state canals (hectares) State road length (kms) % operational land registered by patta % operational land registered by barga
1982
1985
1990
1995
0.11 36 5233 6470 d 3.9
0.11 29 4265 4501 3.0
0.08 25 1485 2501 2.2
0.06 18 2627 4572 1.9
72,793
72,168
1271
1282
1309
1320
4.5
5.8
6.2
6.3
1.9
4.2
4.9
5.1
79,774 84,672
Notes: Average across sample villages, weighted by operational land areas. a IRDP credit subsidy, 1980 prices. b, c Expenditure out of employment program funds, 1980 prices. d For year 1983. Source: Block agricultural development offices, lead banks, GP budgets, West Bengal Economic Review, various years.
identify ownership of land among households in each village, selecting suitable beneficiaries, and pursuing contested cases through the courts. Table 11.4 depicts trends in agricultural inputs provided by the government, as guided by the panchayats between 1982 and 1995. The table shows a weighted average across the sample villages, with operational land areas used as weights (to render these comparable to the land-reform estimates).
9780230_274587_12_cha11.indd 178
6/23/2010 5:05:11 PM
Political Competition & Impact of Local Governance
179
Table 11.5 Trends in farm productivity, incomes and wages
Rice yield, kg/hectare HYV rice area/rice area HYV rice area/total cropped area Value-added/acre Wage rate Hired labor annual hrs/acre
1982
1985
1990
1995
1414 .08 .05 723 0.62 156
1932 .15 – 725 0.69 176
2758 .39 – 1303 0.88 249
3647 .66 .44 1401 1.01 369
Note: Simple average across sample villages. All rupee figures deflated by costof-living index, 1974 = 100. Source: Cost of cultivation surveys.
The 1980s witnessed larger supplies of IRDP credit and mini-kits compared with the 1990s. One out of every nine households received mini-kits, containing seeds, fertilizers, and pesticides, in any given year during the 1980s. This implies that the scale of the mini-kit program was substantially greater than the land reforms. The bulk of employment funds were spent by GPs on building and maintenance of local roads; these employment programs created 2–4 man-days of employment per household every year. Spending on local roads and irrigation was highest in the first half of the 1980s, fell sharply in the second half, and rose again in the 1990s. The downward trend was less sharp compared with the land reforms, which also decreased throughout the 1980s and virtually ceased thereafter. There was some expansion of areas irrigated by state canals; greater expansions were witnessed in medium and small irrigation projects, many of which were managed by panchayat officials. Table 11.5 shows changes over time in cropping patterns and incomes, based on the farm panel in the sample villages. The most spectacular change was in rice yields, which increased two-and-a-half times between 1982 and 1995. This is partly explained by widespread diffusion of high-yielding varieties (HYV) of rice, with acreage devoted to such varieties expanding from less than 10 percent of total rice acreage in 1982 to 39 percent by 1990, and 66 percent in 1995. In real terms, farm value-added per acre more than doubled. Wage rates for agricultural workers rose by 66 percent, and employment more than doubled. Since the poorest sections of the rural population are landless and rely mainly on agricultural labor, the incomes of the poor rose significantly during this period.
11.3
Results of the statistical analysis
In this section we review our findings concerning the political economy of the programs implemented by the West Bengal panchayats. We start by explaining determinants of electoral success of the Left Front in panchayat
9780230_274587_12_cha11.indd 179
6/23/2010 5:05:12 PM
180
Pranab Bardhan and Dilip Mookherjee
Table 11.6 Left Front share regressions
No. observed (GPs) F-st (d.f.) p-value, 2nd order Ser. corr. diff. res. Assembly vote share Difference, district % Congress seats in Parliament % Congress seats in Parliament Lagged GP Left Front share Inflation rate Inflation rate lagged GP Left Front share
CrossSection (OLS)
Panel (Ar-Bond)
Panel (Ar-Bond)
Panel (Ar-Bond)
57 3.11 (9,47)
221 (56) 28.00 (6,214)
221 (56) 17,90 (14,206)
221 (56) 18.41 (15,205)
.14
.04
.06
–0.45* (0.24) 0.72*** (0.23) 3.77*** (0.85) –7.34*** (1.05)
0.85*** (0.28) –0.64*** (0.24) 1.01*** (0.25) 3.43*** (0.85) –6.60*** (1.04)
1.32*** (0.40) –0.51** (0.24) 0.80*** (0.23) 3.48*** (0.91) –6.97*** (1.13)
Notes: Panel regressions: Arellano-Bond GMM estimator. Controls include distribution of land, literacy, caste; small factory employment growth robust standard errors in parentheses; *** significant at 1%, ** at 5%, * at 10%.
elections. We then turn to the determinants of the extent of land-reform implemented and farm extension services delivered in villages, drawing on the village panel. Table 11.6 provides panel regression results for the fraction of seats won by the Left Front in GPs, in terms of a variety of state and national factors likely to affect relative voter loyalties to the Left Front vis-a-vis the Congress party, after controlling for distribution of land, literacy, and caste within the villages. The nationwide popularity of the Congress party is measured by the proportion of seats secured by it in the national parliament. This also captures the ability of the Congress party to influence policies of the central government towards West Bengal, including construction of infrastructure and other central government projects in select constituencies. Another measure of the relative popularity of the Left Front and the Congress party within the district concerned is the average vote-share difference in preceding state assembly elections between the two parties, averaged across all polling constituencies in the same district. As a measure of economic performance of the incumbent that most concerns voters, we include the rate of inflation of a cost-of-living index in the nearest of four centers of the state where this is computed by the state government (Asansol, Ranigunj, Jalpaiguri, and Kolkata). The first and last columns of Table 11.6 show that election outcomes in GP elections closely mirrored district-level vote-share differences in preceding
9780230_274587_12_cha11.indd 180
6/23/2010 5:05:12 PM
Political Competition & Impact of Local Governance
181
state assembly elections. Other columns show that the presence of the Congress party in the national parliament and the inflation rate were significantly correlated with panchayat election outcomes. The effects of these state and national factors differed according to historical incumbency: in traditionally Congress party-dominated areas, a rise in Congress presence in the national parliament or a fall in the inflation rate caused the Congress party to become more entrenched, with the opposite effect in traditionally Left Front areas. In contrast to the effect of these broader factors influencing voter loyalties, changes in local land, literacy, or caste distributions had no effect on panchayat election outcomes. These results form the basis of our identification strategy for various analyses reported below, for example, the effect of varying political competition on policies implemented by GPs, and also subsequently for effects of these policies on farm outcomes. Fluctuations in political events at the state or national level (interacting with local incumbency patterns) provide exogenous sources of temporal variation in Left Front share at the local GP. Twelve fluctuations in Left Front share can be explained by fluctuations in general voter loyalties based on larger state or national issues, which are likely to correlate with time-varying village-specific fluctuations in voter preferences for redistribution, or unobserved determinants of farm productivity. Table 11.7 provides regression estimates for different measures of land-reform implemented (pattaland, bargaland refer to the proportion of cultivable land distributed as land titles or registered under Operation Barga, while pattadar, bargadar refer to the proportion of households receiving land titles or registered as sharecroppers). Since in many years there were no land-reforms implemented at all in any given village, the regression has to incorporate endogenous censoring; accordingly we use a TLAD (trimmed least absolute deviation) estimator with village fixed-effects (Honore, 1992). Table 11.7 shows no evidence of any tendency for land reforms to increase monotonically with the Left Front share of local GP seats. With the exception of the pattaland regression, there was an inverted-U relation instead, statistically significant in the bargadar regression. In the latter regression, the second-last column shows that the top turning point of the U appeared at approximately 50 percent of the Left Front share. In other words, once the Left Front commanded an absolute majority in the local GP, further increases in its seat share reduced rather than increased the proportion of households registered under Operation Barga. This runs counter to the predictions of the ideology hypothesis (which states that the Left Front implements more land reform than the Congress party owing to its ideological commitments), in favor of the quasi-Downsian hypothesis that electoral competition is subject to interest-group pressure (Grossman and Helpman, 1996) or political moral hazard (Bardhan and Mookherjee, 2004), and that political contestability increases redistributive effort.
9780230_274587_12_cha11.indd 181
6/23/2010 5:05:12 PM
182
Pranab Bardhan and Dilip Mookherjee
Table 11.7 Land-reform panel regressions Pattaland % Left % Left Sq. % Left* % 1978 households Landless % Left* % 1978 land Big % Election year Dummy % Preelection year Dummy Total obs. Censored obs. No. groups
–0.66 (0.62) 0.23 (0.33)
–0.08 (0.09) 0.13 (0.18) 1755 1570 89
–0.35 (0.63) 0.12 (0.55) 0.91 (0.82) –1.25*** (0.40) 0.14 (0.15) –0.07 (0.08) 1755 1570 89
Pattadar –0.21 (0.40) –0.11 (0.33)
–0.08 (0.06) –0.01 (0.06) 1755 1570 89
0.64 (0.50) –0.15 (0.30) –1.12** (0.47) –1.02** (0.50) –0.01 (0.06) –0.08 (0.06) 1755 1570 89
Bargaland 0.43 (0.94) –1.03 (0.97)
0.35 (0.26) –0.04 (0.04) 1755 1588 89
Bargadar
1.10 0.51** 0.90*** (0.71) (0.30) (0.23) –0.66 –0.51*** –0.59*** (0.56) (0.19) (0.22) –0.82 –0.38 (0.88) (0.47) –2.88*** –0.47* (1.08) (0.25) –.04 0.16*** 0.00 (.03) (0.05) (0.01) 0.36 0.001 0.13*** (0.31) (0.01) (0.05) 1755 1755 1755 1588 1588 1588 89 89 89
Notes: Trimmed least absolute deviation regressions, yearly data, 1978–98. * Village controls include land distribution, illiteracy rates, proportion of low caste; also included: timeblock dummies, village fixed-effects. Standard errors in parentheses, *** significant at 1%, ** at 5%, * at 10%.
This Table does not report the instrumental variable estimates, for which the reader is referred to Bardhan and Mookherjee (2004). Additional confirming evidence against either a pure ideology or Downsian hypothesis is a significant negative interaction between Left Front share and (either of two measures of) 1978 land inequality. In villages with a more unequal land distribution to start with, increases in Left Front share significantly reduced every measure of land-reform implementation. The quasi-Downsian hypothesis provides a natural interpretation of this finding: higher land inequality implies greater proneness of political parties to capture by landed elites. Given the general dominance of the Left Front, an increase in Left front share represents reduced political competition between the two parties, which permits greater capture of elected officials, resulting in less land reform. The role of electoral competition is further indicated by significant election- or pre-election-year spikes in Barga registration rates. Tables 11.8, 11.9, and 11.10 display implications of village panel regressions for targeting of subsidized credit under the IRDP program, agricultural mini-kits, and fiscal grants to GPs respectively. For the explanation of the estimation procedures the reader is referred to Bardhan and Mookherjee (2004). The first three columns of Table 11.8 show intravillage shares of intended beneficiaries of IRDP loans: landless, small landowning households
9780230_274587_12_cha11.indd 182
6/23/2010 5:05:12 PM
Political Competition & Impact of Local Governance
183
Table 11.8 Targeting of IRDP credit subsidies (Intravillage) (Intravillage) (Intravillage) (Inter-GP) GP Landless % Up to small % SC/ST % share average per share share household Mean (sd) at 1980 prices Effect of following hypothetical changes: 2.5% households switch: medium to landless 2% households switch: big to medium 2% households switch: big to marginal 10% cult. land shifts: small to big 5% poor households switch: non-SC/ST to SC/ST Left Front share of ZP seats rises: 86% to 96% Left Front share of GP seats rises: 66% to 76%
45.0 (39.7)
96.5 (13.8)
45.4 (42.0)
23.56 (66.24)
7.8
3.3
–11.1
–11.5
–6.6
–2.1
–0.8
13.9
–3.7
0.7
–13.1**
3.8
–6.6
–7.1***
–1.1
–10.4
3.7
–1.0
0.4
3.9
10.9
–4.1*
–0.7
–8.5
–7.5
–0.8
0.7
–0.3
Notes: ***, **, * denote significant at 1%, 5%, 10% respectively.
Table 11.9 Targeting of mini-kits (Intravillage) Up to (Inter-GP) GP average small % share per household Mean (sd) Effect of following hypothetical changes: 2.5% households switch: medium to landless 2% households switch: big to medium 2% households switch: big to marginal 10% cult. land shifts: small to big 5% poor households switch: non-SC/ST to SC/ST Left Front share of ZP seats rises: 86% to 96% Left Front share of GP seats rises: 66% to 76%
97.7 (8)
0.085 (0.114)
0.3
–0.053
–0.2
0.017
–0.3
–0.012
0.6
–0.062*
–0.1
–0.078***
1.2
–0.029
–0.3
–0.007
Notes: ***, **, * denote significant at 1, 5, 10% respectively.
9780230_274587_12_cha11.indd 183
6/23/2010 5:05:12 PM
184
Pranab Bardhan and Dilip Mookherjee
Table 11.10 Targeting of fiscal grants
Mean (sd) Mean (sd) Effect of following hypothetical changes: 2.5% households switch: medium to landless 2% households switch: big to medium 2% households switch: big to marginal 10% cult. land shifts: small to big 5% poor households switch: non-SC/ST to SC/ST Left Front share of ZP seats rises: 86% to 96% Left Front share of GP seats rises: 66% to 76%
(Intravillage) Empl. (Inter-GP) GP (Inter-GP) GP days generated average per average per per rupee household household all employment grant employment grant grants 0.024 (0.095)
195 (366)
464 (1247)
–17
–84*
2
60
0.001
–17
–18
–0.008*
–66*
–146**
0.001
–39
–125***
–0.005
–0.31
–16
0.000
–46**
–97***
–0.010** 0.009**
Notes: ***, **, * denote significant at 1, 5, 10% respectively.
(owning less than 5 acres of cultivable land), and scheduled castes and tribes (SC/ST). The mean shares of these groups were close to their demographic weights, indicating that on average these schemes were targeted well. The leakage of IRDP credit to medium and large landowners was small, though roughly in proportion to their demographic weights. Examining the effect of varying land inequality, significant effects are observed for the share of the “up to small” group with a rise in share of land in big holdings, and a particularly strong effect of a rise in the proportion of marginal landowning households of the SC/ST group share. These findings are consistent with the notion that these villages are characterized by political and economic conflict between three broad classes: big landowners, small and marginal landowners, and agricultural workers (of whom low-caste members form a large fraction). The effects on the inter-GP allocation of credit were generally not statistically significant. An increase in Left Front share at the district level (by 10 percent, starting from the mean of 86 percent) tended to reduce credit allocation to a village, which was statistically significant at 15 percent. It also resulted in a significant drop in intravillage targeting towards the poor. These results are broadly consistent with the quasi-Downsian theory: as the Left Front became entrenched
9780230_274587_12_cha11.indd 184
6/23/2010 5:05:13 PM
Political Competition & Impact of Local Governance
185
in the district, less effort was made to direct IRDP credit to the village, and also within the village, to the poor. However, the effects of altered political competition at the GP level appeared to be insignificant. Table 11.9 shows corresponding implications for targeting of mini-kits. The first column shows they were almost exclusively delivered within villages to small and marginal landowning households, rather than to medium or big landowners. Moreover, there was almost no effect of change in land inequality, caste composition, or political competition on intravillage targeting. On the other hand, the intervillage allocation responded significantly to land inequality and caste composition, favoring small non-SC/ST households relative to big landowners and SC/ST households. Table 11.10 shows results for targeting of grants received by GPs from higher levels of government. The first column shows that increased landlessness and the land share of big landowners lowered employment generated from allotted funds by about a third, another indication of local elite capture. The inter-GP allocation of these grants discriminated against the SC/ST households and big landowners, just as in the case of the mini-kits. The same is true for fiscal grants, where an additional negative effect is observed with respect to the extent of landlessness. The quantitative magnitudes of these biases are much stronger than in the case of kits or IRDP credit allocation. Moreover, political competition at the GP level also mattered: entrenchment of the Left Front (i.e., increases in the Left Front share beyond the mean) lowered provision of employment grants to the GP significantly. In general, the results are broadly consistent with the quasi-Downsian model, and run counter to the ideology hypothesis. This indicates that electoral competition promoted accountability of local governments (with respect to sharecropper registration and employment programs). These results are consistent with views expressed by many political commentators and critics that the Left Front has pursued the “politics of middleness,” being more responsive to the needs of middle-level farmers than to those of the poor.
11.4
Conclusions
In this chapter, using nearly 90 village surveys, the role of community in making local governments accountable is analyzed by examining the link between the distribution of development program benefits and the voting behavior in the state of West Bengal in India, where the Left Front, comprising mainly the Communist Party, has been in power for the last three decades. Two important questions examined in this context are: Do program benefits go to the poor households, and are there signs of political clientelism in West Bengal? The results of the analysis show that the intravillage distribution of benefits has been pro-poor. On the other hand, the intervillage allocation of
9780230_274587_12_cha11.indd 185
6/23/2010 5:05:13 PM
186
Pranab Bardhan and Dilip Mookherjee
benefits does not seem to be pro-poor. This latter result corroborates our earlier findings using village-level surveys. This chapter thus contributes to the small literature on the micro-studies of accountability of local governments, particularly with respect to pro-poor programs. It also shows that micro village studies, of the kind encouraged by Yujiro Hayami, can bring out nuances and complexities that are not usually captured in models of macro political theory.
References Appu, P. S. (1996) Land Reforms in India (Delhi: Vikas Publishing House). Bardhan, P. and Mookherjee, D. (2000) “Capture and Governance at Local and National Levels,” American Economic Review, 90, 135–9. Bardhan, P. and Mookherjee, D. (2004) “Political Economy of Land Reform in West Bengal,” Discussion Paper, Institute for Economic Development, Boston University. Bardhan, P. and Mookherjee, D. (2005) “Decentralizing Anti-Poverty Program Delivery in Developing Countries,” Journal of Public Economics, 89, 675–704. Bardhan, P. and Mookherjee, D. (2006) “Corruption and Decentralization of Infrastructure Delivery in Developing Countries,” Economic Journal, 116, 101–27. Downs, A. (1957) An Economic Theory of Democracy (New York: Harper Collins). Grossman, G. and Helpman, E. (1996) “Electoral Competition and Special Interest Politics,” Review of Economic Studies, 63, 265–86. Hayami, Y. (1999) “Community Mechanism of Employment and Wage Determination: Classical or Neoclassical,” in G. R. Saxonhouse and T. N. Srinivasan (eds.), Development, Duality and the International Economic Regime (Ann Arbor, MI: University of Michigan Press), 85–106. Hayami, Y. (2001) “Ecology, History and Development: A Perspective from Rural South-east Asia,” World Bank Research Observer, 16, 169–98. Honore, B. (1992) “Trimmed LAD and Least Squares Estimation of Truncated and Censored Regression Models with Fixed Effects,” Econometrica, 60, 533–66. Lieten, G. (1992) Continuity and Change in Rural West Bengal (New Delhi and London: Sage).
9780230_274587_12_cha11.indd 186
6/23/2010 5:05:13 PM
12 Promises and Realities of Community-Based Agricultural Extension Gershon Feder, Jock R. Anderson, Regina Birner, and Klaus Deininger
12.1
Introduction
Economists have often neglected the role of the community as a third pillar in the economic system next to the state and the market. Yet in agricultural development, which is often thwarted by both market and state failures, the “community mechanism” has a particular and promising role to play. It is one of Yujiro Hayami’s unique contributions to the economics literature that he developed “a conceptual framework for economists” that emphasizes the role of communities and shows how the contested concept of social capital can be applied to analyze them from an economic perspective (Hayami, 2009). At the same time, Hayami is not a “community romanticist” – he is well aware that, just like states and markets, communities can also fail. Hayami’s insights on the community mechanism are particularly important for agricultural extension, which plays a leading role among the public services through which governments have traditionally sought to promote agricultural performance. Hundreds of thousands of public (or publicly-paid) extension agents are employed by governments in developed and developing countries. International development agencies have provided in the past five decades several billions of dollars in programs to support and upgrade extension services in developing countries. Yet, development scholars and practitioners have generally concluded that the performance of extension services in developing countries has been disappointing. A 2001 review sponsored by the Food and Agriculture Organization of the United Nations (FAO), for example, characterized extension services across the developing world as “failing” and “moribund,” in “disarray or barely functioning at all” (Rivera et al., 2001, p. 15). Many agricultural development scholars attribute the inadequate performance of public extension to an incentive failure on the part of extension agents. The failure derives from the fact that, in most public systems, 187
9780230_274587_13_cha12.indd 187
6/23/2010 12:17:51 PM
188
Gershon Feder et al.
agents are nominally accountable to their superiors (who may often not provide effective supervision), and are only indirectly (if at all) accountable to their farmer-clients. Moreover, the lack of information and feedback on different farmer groups’ needs and priorities hinders the design of relevant and effective extension programs (Anderson and Feder, 2007). In view of these realities, many extension scholars and development practitioners have focused on assessing the merits and feasibility of demanddriven approaches to extension (e.g., Anderson, 2007; Birner and Anderson, 2007; Birner et al., 2006; Chapman and Tripp, 2003; Chipeta, 2006; Kidd et al., 2000; Rivera and Alex, 2005; Rivera and Zijp, 2002). In their purest form, in such extension approaches the providers of service are contracted directly by farmers’ groups or communities to deliver information and related services that are specified by farmers. The funds to finance the contracts may be provided in part or in full by governments. But the contracts are issued and monitored by the community of clients. Decisions about contract renewal are made by the community as well, based on its satisfaction with the quality of services rendered in the past. Such a modus operandi is expected to give agents stronger incentives to provide quality advice. Furthermore, the design ensures that information and other services are in line with farmers’ needs and priorities. Several variants of such communitybased and demand-driven extension approaches have been implemented in the past 15 years, mostly with public funding of the extension providers’ costs (e.g., Anderson and Feder, 2007; Birner and Anderson, 2007; Chapman and Tripp, 2003; Dinar and Keynan, 2001; Hanson et al., 2006; Keynan et al., 1977; Kidd et al., 2000). In this chapter, we apply Hayami’s insights on the community mechanism to discuss in some detail the conceptual underpinnings of communitybased extension approaches, highlight theoretical and practical challenges inherent in their design, and assess the evidence available so far on their actual performance. Accordingly, the next section describes features of agricultural extension relevant to understanding systems’ performance. The subsequent section outlines approaches to meeting the challenges through community-based demand-driven systems. We then review actual experiences with such systems, and conclude with a summary of the insights gained and their implications for development policy.
12.2 Market and state failure in agricultural extension The literature on agricultural extension highlights a number of market failures that provide justification for some form of collective action to ameliorate the failure. Birner and Anderson (2007) review several characteristics of extension that lead to market failures, including the public-good nature of information, the merit good nature of some types of extension, externalities, and the transaction costs involved in reaching smallholder farmers.
9780230_274587_13_cha12.indd 188
6/23/2010 12:17:51 PM
Community-Based Agricultural Extension
189
These failures explain the practically universal involvement of national or lower-level governments with extension services, although their involvement does not necessarily require the provision of service through the staff of a public-sector organization. The disenchantment with direct service provision by the public sector, to which we alluded in the introduction, stems from a number of attributes that are frequently inherent in the way public extension systems are organized and managed. Feder et al. (2001), Anderson and Feder (2007), and Birner and Anderson (2007) provide an extensive discussion of these (often interrelated) attributes leading to state failure, which include the following: i. Scale and complexity of agricultural extension services; ii. dependence on broader policy environment; iii. interaction with knowledge generation, which requires functioning two-way feedback flows; iv. public duties other than knowledge transfer that are often imposed on extension agents; v. difficulty in attributing the impact of agricultural extension; vi. weak incentives of extension agents to perform and weak accountability to farmers; vii. cumbersome bureaucratic procedures; and viii. Weak political commitment and support. The weaknesses and issues outlined above have been recognized over the years by extension scholars and practitioners, and various reforming fixes and modifications have been proposed and introduced to address them (e.g., Feder et al., 2001). Among these “fixes,” community-based extension (CBE) is perceived by many observers as a particularly important strategy since it promises to overcome both the state failures and the market failures inherent in extension (see, for example, the World Bank in its Participation Sourcebook http://www.worldbank.org/wbi/sourcebook/sba203.htm; Rivera and Zijp, 2002; and the World Bank in its web-based Agriculture Investment Sourcebook).43
12.3
Community-based extension in principle
One of the advantages of community-based extension systems is that they can benefit from the social capital of rural communities and farmers’ organizations. Hayami (2009, p. 98) defined social capital as the “structure of the informal social relationships conducive to developing cooperation among economic actors aimed at increasing social product, which is expected to accrue to the group of people embedded in those social relationships.” In CBE systems, in their purest form, the extension service is contracted by the community, or is part of the staff of a farmers’ association. In a more
9780230_274587_13_cha12.indd 189
6/23/2010 12:17:51 PM
190
Gershon Feder et al.
diluted format of community extension, farmers’ organizations, whether representing single communities or wider constituencies, have a say in the design and execution of extension programs. Such an approach can provide improvement in the extension service received by farmers, and may provide advantages to smallholders. In most developing countries, the farmers-toagent ratio is more than 1000:1 (e.g., Anderson and Feder, 2004). Hence, it is difficult for farmers to exercise demand and hold service providers accountable without some form of organization. Farmers’ associations can play an important role in aggregating farmers’ demands for extension and in representing farmers in participatory models of extension management so as to make extension more demand-driven. Moreover, farmers’ organizations can help reduce the transaction costs of providing extension, as in groupbased extension approaches. An important advantage of CBE is the fact that the providers of extension advice are, in principle, directly accountable to the farmers, who are members of the community or the organization. An effective role for farmers in the contracting of service providers and in the assessment of their performance is crucial in creating such accountability. Thus, when designed properly, CBE can overcome the critical issue afflicting traditional top-down extension systems, namely, weak incentive to perform to the satisfaction of farmers. The difficulties of monitoring and attributing impact, and assessing relevance, are also less acute, as the services are focused on issues reflecting farmers’ demand, and farmers are involved in providing feedback or even in assessing the service. There are also better prospects for garnering political support (and the consequent budgetary support and financial sustainability); a more effective service with attributable impacts can produce political payoffs for local and national politicians. Moreover, in CBE the public-good aspect of extension is defined at a local or community level, and it may enable some cost recovery, as the “free-rider” problem is easier to resolve. However, there are several issues, some specific to CBE and some typical of extension systems, which may hamper the effectiveness of CBE. We turn now to consider these issues. A major challenge for CBE is the classic problem of collective action. If the benefits of extension advice are “non-excludable,” farmers have limited incentives to incur the transaction costs of participating in the organizational activities that are related to the establishment and management of the organization (the “free-rider” problem). The incentives to join local farmer groups for the purpose of group-based extension may be rather high, because the participants expect to benefit directly from their participation, and the already high social capital is reinforced. However, to participate in extension planning and management beyond the local level, farmers need to become organized at a more aggregate level, which poses its own challenges. The literature on the role of group size and heterogeneity in collective action is extensive, and these relations continue to be debated (e.g., Poteete and Ostrom, 2004). Often, farmers may have social
9780230_274587_13_cha12.indd 190
6/23/2010 12:17:51 PM
Community-Based Agricultural Extension
191
capital that has been generated in the context of their small community, and that would facilitate group activities and induce leaders and members to exert unpaid efforts on behalf of the group. But that capital may not exist across communities (e.g., Hayami, 2009). In such a case, to the extent that CBE requires organizations and collective action across communities, the incentives may not suffice to facilitate that. Donor-sponsored CBE programs often tackle this problem by financing the organization of cross- community farmer entities, and compensating farmers’ leaders and representatives for the time and effort they invest in these collective activities. However, it is a common phenomenon that organizations formed for donor-funded projects collapse once project funding ends (e.g., Ameur, 1994; Purcell and Anderson, 1997 for the case of extension). As pointed out by Hayami (2009, p. 115): “The hasty expansion of community-driven participatory development projects by external aid agencies, along the current vogue of social capital and community participation in the absence of real understanding of local economic and social systems, has resulted in a serious waste of development resources.” Another problem encountered in many participatory development programs arises from the entrenched top-down and patronizing attitudes that often characterize all levels of governments, leading to resistance against CBE approaches, even if they are championed by influential policy-makers (and possibly encouraged and fortified by the financial and intellectual support of external donors). Resistance may reflect itself in attempts to dilute participatory bodies’ powers, foot-dragging on budget transfers, and co-option of farmer leaders to support the interests of the bureaucracy rather than those of the community. Training at various levels of government in participatory concepts is a solution often adopted by champions of CBE, but it may not suffice to resolve the problem, as pointed out by Braun et al. (2006, p. 35): “... given that farmers, and especially women, have very little political influence in most developing countries, such efforts to move towards participation [in extension systems] often flounder on bureaucratic procedures, hierarchical structures, political exigencies, budget constraints, etc.” Consistent with Hayami’s (2009) analysis of the community mechanism, CBE systems are not equally suited for all kinds of extension. They have particular comparative advantages to facilitate extension for activities that require collective action, such as many types of natural resource management and pest management. Where farms are diverse in enterprise mix, and where more farm-specific advice is required, the comparative advantage of CBE approaches is lower. Thus, different farming systems require different types of extension systems. Another major challenge of CBE is avoiding social exclusion and elite capture – a common problem in rural development programs, and in extension specifically. Rural communities and farmers’ organizations are often dominated by middle-class and relatively wealthy farmers. Poor farmers
9780230_274587_13_cha12.indd 191
6/23/2010 12:17:51 PM
192
Gershon Feder et al.
and socially marginalized groups typically play a limited role in the leadership of communities and rural organizations, even if they are members. In particular, the representation of women in them is often low, a problem linked to the sociocultural role of women in many societies as well as to the time constraints faced by women (see, for example, Meinzen-Dick and Zwarteveen, 1998; Quisumbing, 2003). One strategy to deal with the elite capture and the social exclusion problems is the formation of specialized organizations, such as a group exclusively for women farmers, or the allocation of reserved seats for women and disadvantaged groups in participatory planning and management boards. The concept of CBE presumes that sufficient numbers of would-be qualified advisory professionals (or organizations containing such professionals) are available to do the needful (Chapman and Tripp, 2003, p. 7). Often, the expectation is that the existing public extension service cadres would form the core of such an industry, either as part of a semi-autonomous public (or privatized) entity that can be contracted, or as individual professionals following a privatization of the extension service. The NGOs dealing with rural development services in some countries are also viewed as potential extension service providers. The designers of CBE projects commonly pay attention to the issue of accrediting and regulating service providers. However, in countries with endemic governance deficiencies, the process of accreditation may be corrupted. The background and training of the pool of potential service providers (including the existing public extension personnel) may not suffice to address the more specific and localized issues that are likely to be brought up in a demand-driven system. Large-scale training (at public expense, at least initially) may be required (Chapman and Tripp, 2003). Community organizations are not immune to mismanaging funds. When an organization deals with various farmers’ agendas, rather than exclusively with extension, agricultural extension may not necessarily be a priority for the leadership of the organization. Lobbying for state support in the form of subsidies is often a higher priority than helping the members become more productively competitive. On the other hand, integrating extension into an existing entity that was founded for a broader set of goals is an important strategy to reduce the transaction costs of collective action. Elite capture of a different type occurs when leaders of large organizations may have incentives to run for political office, which in turn could lead to conflicts of interest. As noted, politicians have also been known to misuse public extension systems for their own purposes. Private provision of extension avoids most such political-capture problems but fails to serve many impoverished groups adequately. So, CBE approaches, although not risk-free, may present the most attractive options for extension provision in many situations. Accordingly, the important matter of performance in practice should be considered, as in the section that follows.
9780230_274587_13_cha12.indd 192
6/23/2010 12:17:51 PM
Community-Based Agricultural Extension
193
12.4 Evidence on the performance of community-based extension The literature provides extensive coverage of experiences with communitybased extension in the past two decades. However, much of the material is descriptive, and even within the empirical literature there is a paucity of rigorous econometric studies. Yet there are ample indications that the difficulties envisaged theoretically are indeed encountered in practice, and are amplified by practical implementation hurdles that are specific to community-based extension initiatives. The discussion focuses on three extension approaches that have important elements of community-based extension: the National Agricultural Advisory Services (NAADS) program of Uganda; the Farmer Field School (FFS) approach; and India’s Agricultural Technology Agency (ATMA) model. NAADS is a decentralized long-term program that started in 2001 in 24 subcounties of several districts of Uganda. Since then, with support from several donors, the program has been expanded to dozens of additional districts, aiming to reach national coverage in the near future (Government of Uganda, 2007). The program is centrally guided, but at least partially implemented within the country’s existing local government structures. A national board formulates strategic objectives, selects districts and subcounties to be included in the program, defines key implementation procedures (e.g., procurement procedures, and monitoring and reporting requirements), and provides technical services to local government levels of the program on issues that are beyond their capacity. The board is also responsible for certification, regulation, and overall quality verification of extension service providers (who may come from the private sector, NGOs, or semi-public organizations). In participating districts, the implementation of the program is overseen by a chief administrative officer supported by elected district councils, which are expected to oversee procurement boards and guide the process of converting local public extension workers into private advisory service providers (relatively little progress was achieved on this). Within NAADS subcounties, farmer-participants (who comprise only a fraction of the farming population) organize themselves in farmers’ groups based on members’ shared priorities and needs. Each group elects two representatives to the subcounty farmers’ forum, and it is at that level that priority technology needs are identified. The program manual requires that only three priorities be selected in each subcounty (NAADS, 2004). Members of the farmers’ forum are involved in various committees set up under NAADS (executive committee, technical committee, procurement committee, and monitoring and evaluation committee), which, together with local officials, are responsible for local-level strategy, planning and setting guidelines, managing program funds, and liaising with districts (Government of Uganda, 2001; NAADS, 2004). While the budget for the program at the local level
9780230_274587_13_cha12.indd 193
6/23/2010 12:17:51 PM
194
Gershon Feder et al.
is provided mostly through earmarked transfers from the national government (largely using targeted donor-supplied financing), these resources are to be complemented by an allocation from subcounty and district governments’ regular budget (5 percent) as well as a nominal contribution (2 percent) from farmer-participants (Government of Uganda, 2001). In 2009, the government decided to stop outsourcing of extension services and rather to hire extension staff through local governments, thus removing an important element in the community-based nature of the approach.44 This review focuses on the NAADS model that was in place prior to this change. The farmer field school (FFS) approach to delivering information and educational services was designed originally as a means to introduce knowledge of integrated pest management (IPM) to irrigated rice farmers in Asia, and has since been expanded to numerous countries, covering various agricultural themes (van den Berg and Jiggins, 2007). A typical FFS educates farmer participants, typically groups comprising 20–25 farmers, on agro-ecosystems analysis as well as specific technological features of their crops and the field environment. The FFS approach relies on participatory training methods to convey knowledge to field school participants, with the extension agenttrainer expected to act not just as a transmitter of information, but mainly as a facilitator encouraging farmers’ own discovery and discussion of their experiences and observations. The selection of participants for the training is done with strong community involvement through its established leadership and existing social structures. The participants are expected to contribute to the wider community through dissemination of knowledge, and follow-up activities such as field experiments and collective action. Through group interactions, attendees sharpen their decision-making abilities and are empowered by learning leadership, communication, and management skills (e.g., van de Fliert, 1993). The Agricultural Technology Management Agency (ATMA) model of participatory decentralized extension was implemented originally as a pilot program in a World Bank-funded agricultural development project in India in the period 1998–2003. It has since been expanded to a national program funded by the Indian union government. The program operates at a district level, and entails within each district a semi-autonomous agency dealing with extension matters. The agency can receive both public and private funds, as well as charge fees to clients. Each ATMA is directed and overseen by a governing board that includes representatives of all farmer classes in the district, as well as other stakeholders (private sector, rural banks, NGOs, and official agencies dealing with agricultural development). The governing boards define strategic priorities reflecting the interests of the constituencies represented on the board. Within each village, famers are also organized in self-help groups and other farmer-interest groups, and these groups elect representatives into the block (subdistrict) farmer advisory committee (FAC), the chairs of which serve on the ATMA governing board. The design
9780230_274587_13_cha12.indd 194
6/23/2010 12:17:51 PM
Community-Based Agricultural Extension
195
of the system provides, at least nominally, for ample farmer influence on extension activities, since every village is represented in the FAC (Swanson, 2008). With this background on NAADS, FFS and ATMA we can turn to a review of the limited evidence on performance of CBE approaches. Exclusion and elite capture Insights on the risk of capture (which, as we noted earlier, is not unique to CBE) are derived from the experience of the NAADS program in Uganda. Several observers whom we cite below commented on a bias in implementation, leading to the favoring of wealthier farmers. Thus, one early study concluded that the mobilization of groups through local government leaders appealed to the progressive wealthier elite, while “the poorer sections of the population ... were perceived to be excluded,” and that NAADS had “a strong bias towards the better-off” (Boesen et al., 2004, p. 66). A study of the Mukono district reports that farmers perceive that “poor farmers were left out,” and suggests that the “required relatively high levels of literacy and the lengthy debates (on prioritization) precluded women and the poor” (Obaa et al., 2005, p. 8, 9). An OXFAM/FODOWE study cited by Kibwika (2006, p. 101) reports on the basis of a 2004 survey that “the only people who benefit from NAADS are those with convertible assets ... or those with access to external financing such as remittances.” Similarly, Parkinson (2009) observes that, by 2006, the program had “introduced a number of short-term approaches that systematically rewarded wealthier and more connected farmers” (p. 123). She provides data showing that NAADS group leadership comprises farmers of higher education and wealth status than that of regular members, concluding that “poorer farmers ... were less able to benefit from the types of technology NAADS provided” (p. 164–5). Bukenya (2008) arrives at a similar conclusion. Farmer field-school training programs have also been reported to be vulnerable to elite capture. In her study of the early phases of FFS in Indonesia, van de Fliert (1993, p. 157) commented that “The composition of the field school groups observed was not representative of the farmer population in the villages. The groups contained many village officials, and farmers with relatively high education, large (owned) fields, and off-farms jobs, and no women at all.” A later report on the Indonesian experience cites observations of a field operative complaining that the relegation of participant selection to the established village leadership resulted in preference for members of the village elite (Fakih et al., 2003, p. 36). More generally, the report concluded that FFS effort in Indonesia has not been characterized by equity in its coverage (p. 64–5). Similarly, Feder and Savastano (2006) demonstrated econometrically that wealthier farmers with higher educational attainment had a higher probability of being selected for participation in the Indonesian FFS training than others.
9780230_274587_13_cha12.indd 195
6/23/2010 12:17:51 PM
196
Gershon Feder et al.
A study of livestock field schools in Vietnam claims a necessity (in the context of that particular training theme) to skip the poorest segment of the community because of a perception that they could not marshal the resources to raise even small livestock (Dalsgaard et al., 2005, p. 10). The study cites, however, evidence from livestock field schools in Bangladesh, Benin, and Senegal, where adherence to poverty-focused selection criteria yielded satisfactory levels of participation by the poor. Difficulties in implementing farmers’ priorities The NAADS program in Uganda provides an illustration of the practical difficulties faced in attempting to reflect farmers’ priorities in the advice delivered by service providers, even within a nominally demand-driven, community-based, extension system. Several observers of the program cited below suggested that farmers’ actual control of priority-setting for their groups’ training is limited in reality (especially with the rather arbitrary restriction to only three commodities), undermining one of the key expected advantages of community-based extension systems. Obaa et al. (2005) provide a detailed description of the various elements of the decision chain that leads to the design of a service contract by a provider within the NAADS program. Ultimately the priorities in the service contract are not defined for each group separately, but rather are determined at the subcounty level, and consequently the content of the training reflects an aggregate demand of many groups as well as top-down priorities of the national and local government (emphasis on commercial crops). The training received by any group may therefore contain only some or none of the priorities expressed by that specific group. Indeed, Obaa et al. (2005) report cases where farmers were actually getting advice on enterprises that were not their priority. Similarly, a study in one NAADS district (Friis-Hansen, 2004) reports participants’ frustration with the length of time consumed by various layers of the enterprise prioritization process, noting that “the enterprises on which they receive advisory services only rarely (are) similar to those for which they articulated needs” (p. 9), partly because of NAADS’ emphasis on commercial crops that are often not farmers’ first priority. Frustration with the priority-setting process is reported to have led to high dropout rates by participants from the program (Parkinson, 2009). The practical difficulty seems to stem from the high administrative transaction cost of establishing a financing and procurement system that can transfer funds down to a community (single farmer group) level and enable direct contracting of services by a single farmer group. Such a system also faces a challenge in monitoring the uses of funds and has vulnerabilities to capture and corruption even at the community level, as discussed by Platteau and Abrams (2002) and Deininger and Liu (2008). In Chile and Costa Rica, the government attempted to reduce the transaction cost of making funds for contracting extension available directly to farmers through the distribution
9780230_274587_13_cha12.indd 196
6/23/2010 12:17:51 PM
Community-Based Agricultural Extension
197
of vouchers. The farmer-holders of vouchers could use them to pay qualified extension service providers. As the payment was made directly by farmers to the service provider, it was expected that farmers’ priorities would be fully addressed by the extension agent. However, the institutional arrangements required to monitor and control the use of the vouchers so as to prevent fraud were inadequate, leading to major abuses and reduced effectiveness (Bebbington and Sotomayor, 2005; Berdegué and Marchant, 2002; Cox and Ortega, 2004; Dinar, 1996). Evidently, an appropriate control system needs to be significantly more extensive and hence more expensive. Difficulties in implementing farmers’ control of service providers’ contracts As argued in Section 12.3, effective control of service providers’ contracts by the community is critical to generate incentives for high-quality service delivery. Again the NAADS system in Uganda illustrates the practical difficulties in translating this principle into reality. Contracts within the NAADS system are awarded by the subcounty administration because there is no mechanism to transfer funds down to a village community level. The performance of service providers is not formally monitored by members of each group but, rather, it is the formal responsibility of the farmers’ forum that represents farmer groups from the entire subcounty. The existence of a wedge between the actual service recipients and the issuers and enforcers of the contracts weakens providers’ accountability to the farmers who receive advice, and thus diminishes the incentives to provide high-quality service. Even the monitoring by representatives of the farmers’ forum is deficient; farmers who are entrusted with this task are supposed to receive compensation for the time and effort exerted in monitoring on behalf of the larger community. However, funds to pay them are often not available, reducing the incentives to perform a role that only marginally serves their own farming group (Muwonge, 2007). Evidently, the farming population of a subcounty cannot be perceived as a community in a social sense, and thus it does not possess the social capital that would facilitate and induce voluntary (unpaid) actions on behalf of the group. As argued by Hayami (2009, p. 111): “For the community to have sufficient social relation capital in organizing cooperation among its members, its membership must be small enough to ensure intense social interactions.” Limited availability of competent service providers The advantages of CBE (and, more generally, all demand-driven extension systems) are predicated on the availability of a cadre of skilled service providers who can compete for the extension contracts issued by communities. Those providers who do not perform satisfactorily would then be weeded out by the competitive market process. Centralized screening and accreditation mechanisms can be introduced to limit the ability of unqualified
9780230_274587_13_cha12.indd 197
6/23/2010 12:17:51 PM
198
Gershon Feder et al.
service providers to take advantage of farmers’ lack of familiarity with the true capacity of a bidder in the initial phase of a CBE program. However, in many developing countries there is a rather thin market of qualified service providers, and the situation is exacerbated by the vulnerability of the accreditation mechanism to corruption and political manipulation. In some more-developed countries (e.g., the UK and New Zealand), the privatization and dismantling of public extension systems released into the market for extension service provision large numbers of competent would-be advisors. In contrast, the skills of extension staff in many developing countries were built on a slender educational preparation and were geared towards generic technology messages, rather than the more specific and localized issues that farmers tend to identify as their priorities. A transitional period whereby potential service providers equip themselves (through training and recruitment) with the skills that are likely to be demanded would need to be taken in account. However, this generates an incentive problem, as it is unlikely that potential providers would heavily invest their own resources in training staff prior to actual identification of demand and assurance of landing contracts. This induced Chapman and Tripp (2003, p. 7) to perceive a public-sector role in organizing and funding training in the transitional phase. The issue is even more acute when the dismantling of the public extension system is held up by political or administrative obstacles. This is illustrated by the experience in Uganda, where the retrenchment of the public extension system that was planned to run in parallel with the implementation of the CBE seemingly did not materialize due to legal issues. Cadres of the public extension service were therefore mostly not available to compete for the provision of services, and various NGOs, private groups, and semi-public entities, of varying backgrounds and records, gained contracts. Procedures for providers’ qualification and accreditation could be implemented only with delay and were fraught with irregularities, and the training and skills updating (funded in part by the NAADS program) were not systematically carried out. The inadequacy of service providers and the resultant low quality of service were highlighted in the NAADS midterm review (Kazigati, 2005, p. 45; Nyanzi, 2005, p. 46). Similarly, Ekwamu and Brown (2005, p. 28) reported that “the quality of service provision emerged as a major issue in personal interviews with farmer groups.” Deficiencies and irregularities in the procedures for procuring service provision are claimed to have led to the award of contracts to providers who lacked qualifications and who consequently delivered low-quality service (Parkinson, 2008, p. 139–40). Based on field interviews, Muwonge (2007) raised concerns regarding service quality, which he attributed to deficient monitoring of provider performance. In recent months, the Uganda government announced the reintroduction of the public extension service as the main extension provider within the NAADS program. Problems due to inadequate availability of qualified service providers also afflicted the voucher-based extension programs in Chile
9780230_274587_13_cha12.indd 198
6/23/2010 12:17:52 PM
Community-Based Agricultural Extension
199
and Costa Rica in the early 1990s (Ameur, 1994; Bebbington and Sotomayor, 1995). China’s program whereby farmer associations contract technical services from public entities such as research institutes, universities, and individual scientists was also assessed to have suffered due to the limited access to subject-matter specialists (Kidd et al., 2000). The current scaling-up phase of the ATMA program in India is claimed to be adversely affected by limited capacity of existing personnel and the inadequate skills of extension agents inherited from the long-defunct “T and V” extension system (Sulaiman and Hall, 2008; Swanson, 2009). Change in attitudes and top-down orientation of extension organizations is difficult Most CBE programs maintain a dependence on public-sector organizations (whether at the local or national level, or both), and they are introduced with the encouragement and initial funding of external donors. The longterm effectiveness and sustainability of such reforms depend crucially on the extent to which government bureaucracies and field-level workers can adopt the participatory mode of operation that underlies CBE. The pilot phase of the Indian ATMA scheme illustrated that a formal structure of governance that deliberately engages beneficiaries, and appropriate training of extension personnel, can indeed flourish in what is traditionally a top-down field of extension endeavor in this nation. According to Swanson (2008), the farmer orientation of the ATMA model in India positively affected the motivation and morale of the field extension staff: “For the first time, they could see the direct impact of their work on the lives of farmers, farm women and rural young people within their block and district. This new arrangement had a direct and positive impact on their performance. In the process, they were transformed from merely transferring technology (i.e., delivering information) to becoming problem solvers in working with farmer groups to identify and help solve specific problems or needs in pursuing different enterprises” (p. 32). However, Swanson (2009) points out that when the program moved to a national scale, funds for appropriate training of extension staff in the concepts and methods of participatory processes were not provided, resulting in disappointing performance. Similarly, Sulaiman and Hall (2008, p. 3) refer to “attitudinal barriers at all levels” and “lack of local ownership” as two of the factors underlying their concern that ATMA will suffer the same fate as the failed T and V extension system (p. 32). Given that most extension officials have been brought up in a top-down organizational culture, where farmers are viewed as wards of the state, it is not surprising that change is hard and reversals occur. In the Ugandan NAADS system, local governments do not fully “own” the communitybased approach, even though it has some strong champions at the national level (Muwonge, 2007). This is reflected in foot-dragging on local complementary budget allocations, and in the assertion of de facto control of
9780230_274587_13_cha12.indd 199
6/23/2010 12:17:52 PM
200 Gershon Feder et al.
contracting by officials at the subcounty level, even though nominally this was a responsibility for which the farmers’ forum was to have much influence. The emphasis on commercial crops that reflects the agricultural development strategy of the national government often takes precedence in selecting training programs over expressed priority needs of farmers’ groups (Bahiigwa et al., 2005). In Indonesia, a review of two decades of FFSs concluded that while many field workers subscribe to the participatory principles of this community-based approach “... there is little evidence that the culture and practice of participation in Community IPM has fundamentally influenced government bureaucracies” (Fakih et al., 2003, p. 45). Observers of the experience of the decentralized extension project in Indonesia that was completed in 2005 highlighted the limited comprehension by local-level staff of the CBE concept (Bourgeois and Kusumaningrum, 2006, p. 77). This is reflected in the fact that a key local institution (the district extension committee – DEC) that was created by the project as a stakeholders’ forum (including, in particular, farmer representatives) to provide strategic guidance and define extension priorities has not functioned effectively. The main reasons were that “most DECs remain dominated by government through an over-representation (usually 50 percent or more) or perceived authority of government members” (World Bank, 2005, p. 8). In Vietnam, the introduction of participatory extension through a livestock field school program faced significant difficulties because of “a strong inclination towards applying the familiar (and thus comfortable) style of top-down instruction” (Dalsgaard et al., 2005, p. 5). The external donors (who were promoting a CBE concept) had to compromise and accept a system that retained a distinct flavor of traditional top-down, technocratic extension development, because “too much ‘participatory arm-twisting’ by external advisors was not welcome” (Dalsgaard et al., 2005, p. 6). More generally, after almost two decades of experimentation with FFS approaches in various crops in Vietnam (invariably with external donor funding), the field school model has not yet been widely mainstreamed into the agricultural extension system of Vietnam (van de Fliert et al., 2007).
12.5 Quantitative studies of the impact of community-based extension There have been several quantitative analyses of the impacts of CBE programs. A number of these studies deal with small-scale pilot programs, or with a limited geographical area within a national program. It is therefore difficult to ascertain to what extent their results would hold for a largescale program, or whether they are representative of a national program. Studies of CBE are also subject to the typical methodological and data challenges that make the evaluation of extension impact difficult (Birkhaeuser et al., 1991).
9780230_274587_13_cha12.indd 200
6/23/2010 12:17:52 PM
Community-Based Agricultural Extension
201
A study of two pilot programs in Nicaragua (Dinar and Keynan, 2001) used secondary data to calculate incremental farm-level net revenue margins for one program and benefit–cost ratios for the other program. Both programs arranged for service contracts between public (in the case of the first program) or private (in the case of the second program) service providers and farmers, with a substantial public subsidy. Gross margins varied among the four regions in the first program and the aggregate gain, while positive, was not considered robust. Moreover, the calculation did not account for government expenditures and did not allow statistical testing. The analysis of the second program suggested a high benefit-cost ratio (1.77), but it did not account for all the public costs, and was not subjected to statistical testing. A Honduran small-scale public program to provide privately-contracted extension services to groups of farmers has been evaluated by Hanson et al., (2006). As with the NAADS program, the service provider contracts are not drawn up directly between the farmer group and the provider, but in the Honduran case representatives of each group are interviewed at the end of the season by program officials to assess client satisfaction. The overall rates of return on operations in two geographical areas were 8 percent and 10 percent, with a number of arbitrary assumptions on the profile of benefits over time. It is not clear whether the overhead costs of the program administration were included in the calculation. Farmer field school programs have been the focus of quite a few empirical studies (e.g., Anderson and Feder, 2004; Feder et al., 2004a, 2004b; RickerGilbert et al., 2008; Tripp et al., 2005) and extensive literature reviews citing such studies are available in Davis (2006) and van den Berg and Jiggins (2007). The studies are too numerous for us to review here, but our conclusion is that the results are mixed. Some of the positive results claimed are based on inadequate econometric analysis that does not properly account for selection biases. Insufficient attention is given to the full cost of the training, which tends to be higher (per farmer) than less intensive extension methods. It has also been noted that diffusion of information between trained farmers and other members of the community is problematic due to the detailed and experiential nature of the training, as highlighted by Braun et al. (2006, p. 39): “This [limited diffusion] leaves open the question how the beneficial impact of FFS on participating farmers can be scaled up beyond the relatively small numbers that can be reached directly through FFS ... The experience so far is that too many key characteristics of the FFS erode during mass replication for the benefits to be sustained.” The NAADS program in Uganda, being one of the largest national CBE programs, has been studied by many, although only two studies undertook a rigorous econometric analysis based on a sufficiently large sample. A study by Benin et al. (2007) relying on a 2004 survey of 894 households suggests that the program had a significant and positive impact on the adoption of new crop and livestock enterprises (vanilla, groundnut, goats,
9780230_274587_13_cha12.indd 201
6/23/2010 12:17:52 PM
202 Gershon Feder et al.
bees), modern technology, and postharvest technologies. However, the results do not indicate significant differences between NAADS and nonNAADS subcounties in crop yields or total (crop and livestock) income, although it suggests that farmers in NAADS subcounties avoided some of the large drops in crop income experienced by those in non-NAADS subcounties. A serious limitation on the robustness of the result is due to the fact that the study does not systematically account for the possible selection bias whereby the subcounties covered by NAADS are different from those not selected for inclusion in the program. Another drawback is the estimation of income and consumption changes on the basis of respondents’ recall over a five-year horizon. Another study of NAADS impact, based on a more limited sample of households interviewed in 2005–6, attempts to tackle selection bias through an instrumental variable procedure (Muwonge, 2007). In one of the specifications (that actually relies on ordinary least squares regression) the study estimates that NAADS participation increased the value of farm production by 18 percent, but the result may not be very robust as it is statistically significant only at the 90 percent confidence level. It is noteworthy that, once instrumental variables are used, extension impact is not significant. The author rejects the instrumental variables specification on the grounds that there is no indication of an endogeneity problem, but this conclusion is most likely due to the choice of weak instrumental variables (none of which is at the household level). A study conducted at the pilot phase of the ATMA program in 28 project districts suggested that the reformed extension system contributed to increasing farm income and rural employment through agricultural diversification. Thus, during a four-year period (1999−2003), the horticultural cropping area increased from 12 to 16 percent; oilseed crop area increased from 3 to 11 percent; and the crop area for herbs, medicinal, and aromatic crops increased from 1 to 5 percent. During this period, the area planted to cereal crops (primarily wheat and rice) declined from 55 to 47 percent, but yields increased 14 percent, resulting in no appreciable loss in the production of staple food crops. During this period, average farm income across these 28 pilot project districts increased 24 percent, in contrast with only 5 percent in nonproject districts (Tyagi and Verma, 2004). The analysis may have inadequately tackled selection biases and baseline advantages of some of the program districts, thus the attributions may entail an overestimate. The scaled-up program that expanded the ATMA model to the national level has not been evaluated rigorously, but it is perceived to be floundering: Rashid Sulaiman and Andy Hall (2008) attribute the difficulties to numerous implementation challenges, including: “Insufficient support; Mismatch with diversity of application contexts; Lack of local ownership; Capacity and institutional constraints.” Swanson (2009) agrees with the assessment of the poor status of the national program, but argues that the key problem
9780230_274587_13_cha12.indd 202
6/23/2010 12:17:52 PM
Community-Based Agricultural Extension
203
is insufficient resources to invest in training extension staff in the Indian states in participatory methods. A study of the impact of the Indonesia Decentralized Agricultural and Forestry Extension uses a panel sample of program participants and nonparticipants to infer impact on income and other performance criteria (Bourgeois and Kusumaningrum, 2006). The study, which is afflicted by a number of data and methodological issues, produced numerous mixed results, and curiously it indicated no project impact on agricultural incomes but some impact on nonfarm income. The results are summarized by one of the authors as follows (Kusumaningrum, 2007, p. 1): “There is no clear indication that the Participatory Action and People Participation method directly ... improved the welfare of rural people. Field research indicated that provision of inputs and services to target groups was not constant over the project’s time frame (five years). But still the project provided some benefits for the farmers, including better knowledge in how to acquire resources; learning how to make proposals and how to discuss and decide together about priorities.” The few cases reviewed above serve to highlight the need for cogent quantitative analysis of extension schemes of different types, but especially CBE variants.
12.6 Conclusion The belief that agricultural producers can productively use improved information to advance their productivity and profitability as well as contribute to higher agricultural and economic growth and poverty reduction is widely held and has been a key rationale for agricultural extension being an important element of agricultural development strategy for decades. But the validity of this belief has come under increasing scrutiny in recent years as evidence of less than satisfactory experience has been accumulating in the developing world, as well as among the agencies that are engaged in assisting national efforts. We began our review by examining the conceptual underpinnings of community-based extension approaches, whereby users of agricultural extension services are empowered, usually through their community structures such as farmers’ associations (of varying degrees of formality of constitution) or other associations that bring people together over shared interests, to influence what is attempted and how it is achieved in such service systems. The organizational structures represent vehicles for bringing social capital and engaged linkage to help in the delivery of information and other services that can be provided through advisory endeavor. The key rationale in pursuing such approaches is to overcome the theoretical and practical challenges inherent in the design of traditional public extension models, typically conducted by agencies located within a ministry of agriculture. The essence of the community-based, or “demand-driven,” approaches is
9780230_274587_13_cha12.indd 203
6/23/2010 12:17:52 PM
204
Gershon Feder et al.
to make extension workers accountable to the users of the services, and to enable beneficiaries to articulate well their needs and get them attended to by extension providers, and more directly to be aware of and reactive to the effectiveness of delivery. Since user beneficiaries may not be fully conscious of the opportunities available, there is an element of faith or hope (bolstered by appropriate institutional design and implementation) that the providers will indeed be able to deliver cogent information in timely and effective ways. In light of Hayami’s (2009) insights on community failures, it should come as no surprise that the reality of recent experience with CBE approaches to delivery of agricultural advisory services, as reviewed in the final section, finds them to be somewhat lacking in terms of having successfully overcome all the problematic features of extension delivery in the past. In the unfortunately rather few cases where performance has been relatively carefully studied, we find that elite capture constitutes a major constraint. The experience has not been much better in having the empowered farmer beneficiaries be successful in setting the extension agenda and “supervising” delivery of desired services. For instance, articulating farmer priorities and getting these into extension work-programs was a central design feature of NAADS in Uganda, but this was seriously compromised in implementation both because of influences that might be called political, and because of bureaucratic constraints on devolving authority to the village-based farmer group. The experience with the pilot phase of the ATMA model in India in this respect may have been better, perhaps because of the greater number of checks and balances incorporated into the arrangements for ATMA governance combined with proper training of staff involved. The disappointing experience of the same model at the scaled-up phase points out that the conclusions derived from a relatively small (in the Indian context) pilot cannot be simply extrapolated. A related aspect of farmer empowerment is in implementing farmers’ control of service providers’ contracts. This is evidently an aspect that requires considerable administrative imagination and novel mechanisms that have seemingly yet to be forged effectively, given the experience overviewed in Section 12.4. Part of the challenge for CBE derives from the problem, often encountered in developing countries, of a chronic shortage of competent potential service providers. The important role for public investment here is largely one of providing resources and mechanisms for training cadres of extension workers with relevant skill mixes, not an easy thing to do with the educational facilities available in many developing countries. Top-down thinking is hard to shake off among those accustomed to its lures, or so it seems in the cases studied. Deep-seated cultural attitudes are surely important features of the challenge of effectively empowering farmers, and changing mindsets of extension workers ostensibly working on their behalf.
9780230_274587_13_cha12.indd 204
6/23/2010 12:17:52 PM
Community-Based Agricultural Extension
205
To summarize these various challenges reviewed in this chapter, CBE, in spite of its promise, is no panacea! There are important knowledge gaps yet to be filled, as transpires from our review of the limited econometric evidence, to inform the public debate on investment in agricultural extension. Thus, we conclude our review with the inevitable academic call for further research in this surely important area of investment for agricultural and economic development. Yujiro Hayami’s work on social capital and communities can guide this field of research, as it provides an insightful perspective for analyzing the potential and the limitations of the community-mechanism in agricultural development – limitations that must be recognized more realistically, and potentials that must be exploited more creatively.
References Ameur, C. (1994) “Agricultural Extension: A Step beyond the Next Step,” World Bank Technical Paper No. 247, World Bank, Washington, DC. Anderson, J. R. (2007) “Agricultural Advisory Services,” Background Paper for 2008 World Development Report, Agriculture for Development. World Bank, Washington, DC. Anderson, J. R. and Feder, G. (2004) “Agricultural Extension: Good Intentions and Hard Realities,” World Bank Research Observer, 19 (1), 41–60. Anderson, J. R. and Feder, G. (2007) “Agricultural Extension,” in R. E. Evenson and P. Pingali (eds.), Agricultural Development: Farmers, Farm Production and Farm Markets. Handbook of Agricultural Economics, Vol. 3 (Amsterdam: Elsevier), 2343–78. Bahiigwa, G., Rigby, D. and Woodhouse, P. (2005) “Right Target, Wrong Mechanism? Agricultural Modernization and Poverty Reduction in Uganda,” World Development, 33 (3), 481–96. Bebbington, A. J. and Sotomayor, O. (1995) Demand-Led and Poverty-Oriented – or Just Sub-Contracted and Efficient? Learning from (Semi)-Privatised Technology Transfer Programmes in Chile (Oxford: FSG). Benin, S., Nkonya, E., Pender, J., Okecho, G. Mugarura, S. and Kato, E. (2007) Quantifying the Impact of the National Agricultural Advisory Services in the Uganda Rural Livelihoods (Washington, DC: IFPRI). Berdegué, J. and Marchant, C. (2002) “Chile: The Evolution of the Agricultural Advisory Service for Small Farmers: 1978–2000,” in W. M. Rivera and W. Zijp (eds.), Contracting for Agricultural Extension. International Case Studies and Emerging Practices (Wallingford: CABI Publishing), 21–8. van den Berg, H. and Jiggins, J. (2007) “The Impacts of Farmer Field schools in Relation to Integrated Pest Management,” World Development, 35 (4), 663–86. Birkhaeuser, D., Evenson, R. E. and Feder, G. (1991) “The Economic Impact of Agricultural Extension: A Review,” Economic Development and Cultural Change, 39 (3), 607–50. Birner, R. and Anderson, J. R. (2007) “How to Make Agricultural Extension DemandDriven? The Case of India’s Agricultural Extension,” Discussion Paper No. 729, IFPRI (Washington, DC). Birner, R., Davis, K., Pender, J., Nkonya, E., Anandajayasekeram, P., Ekboir, J., Mbabu, A., Spielman, D., Horna, D. and Benin, S. (2006) “From Best Practice to Best Fit: A Framework for Analyzing Agricultural Advisory Services Worldwide,” Development
9780230_274587_13_cha12.indd 205
6/23/2010 12:17:52 PM
206 Gershon Feder et al. Strategy and Governance Division, Discussion Paper No. 39, IFPRI (Washington, DC). Boesen, J., Miiro, R. and Kasozi, S. (2004) “Basis for Poverty Reduction? A Rich Civil Society, Farmer Innovation and Agricultural Service Provision in Kabale, Uganda,” Working Paper Series, No. 2/2004, Danish Institute for International Studies (Copenhagen). Bourgeois, R. and Kusumaningrum, D. (2006) Impact of the Decentralised Agricultural and Forestry Extension Project (Jakarta: CIRAD), December. Braun, A., Jiggins, J., Röling, N., van den Berg, H. and Snijders, P. (2006) A Global Survey and Review of Farmer Field School experiences (Nairobi: ILRI). Bukenya, C. (2008) “Stakeholder Participation in the National Agricultural Advisory Service in Uganda: Analysis of Practice and Contextual Factors,” PhD dissertation (Wageningen: Wageningen University). Chapman, R. and Tripp, R. (2003) “Changing Incentives for Agricultural Extension – A Review of Privatized Extension in Practice,” Working Paper Series, No. 123. Agricultural Research and Extension Network, Overseas Development Institute (London). Chipeta, S. (2006) “Demand-Driven Agricultural Advisory Services,” Green Booklet Series, Neuchatel Initiative. Cox, M. and Ortega, H. (2004) “Chile: Origin and Evolution of a Privatized Extension System,” in W. M. Rivera and G. Alex (eds.), Extension Reform for Rural Development, Vol. 2: Privatization of Extension Systems (Washington, DC: World Bank and USAID), 9–15. Dalsgaard, J. P. T., Minh, T. T., Giang, V. N. and Riise, J. C. (2005) “Introducing a Farmers’ Livestock Training Approach into the National Extension System in Vietnam,” Agricultural Research and Extension Network (AGREN), Network Paper No. 144, January. Davis, K. (2006) “Farmer Field Schools: A Boon or Bust for Extension in Africa?” Journal of International Agricultural and Extension Education, 13 (1), 91–7. Deininger K. and Liu, Y. (2008) “Economic and Social Impacts of Self-help Groups in India,” Paper presented at AAEA Conference (Orlando, Florida), July. Dinar, A. (1996) “Extension Commercialization: How Much to Charge for Extension Services,” American Journal of Agricultural Economics, 78 (1), 1–12. Dinar, A. and Keynan, G. (2001) “Economics of Paid Extension: Lessons from Experience in Nicaragua,” American Journal of Agricultural Economics, 83 (3), 769–76. Ekwamu, A. and Brown, M. (2005) “Four Years of NAADS Implementation: Programme Outcomes and Impact,” in Proceedings of the Mid-term Review of the National Agricultural Advisory Services (Kampala: Ministry of Agriculture, Animal Industry and Fisheries), 25–39. Fakih, M., Rahardjo, T. and Pimbert, M. (2003) “Community Integrated Pest Management in Indonesia: Institutionalising Participation and People-centred Approaches,” Institutionalising Participation Series No. 05. IIED (London, UK). Feder, G., Murgai, R. and Quizon, J. B. (2004a) “Sending Farmers Back to School: The Impact of Farmer Field Schools in Indonesia,” Review of Agricultural Economics, 26 (1), 45–62. Feder, G., Murgai, R. and Quizon J. (2004b) “The Acquisition and Diffusion of Knowledge: The Case of Pest Management Training in Farmer Field Schools, Indonesia,” Journal of Agricultural Economics, 55 (2), 221–43.
9780230_274587_13_cha12.indd 206
6/23/2010 12:17:52 PM
Community-Based Agricultural Extension
207
Feder, G. and Savastano, S. (2006) “The Role of Opinion Leaders in the Diffusion of New Knowledge: The Case of Integrated Pest Management,” World Development, 34 (7), 1287–300. Feder, G., Willett, A. and Zijp, W. (2001) “Agricultural Extension: Generic Challenges and the Ingredients for Solutions,” in S. Wolf and D. Zilberman (eds.), Knowledge Generation and Technical Change: Institutional Innovation in Agriculture (Boston, MA: Kluwer), 313–56. van de Fliert, E. (1993) “Integrated Pest Management: Farmer Field Schools Generate Sustainable Practices,” A Case Study in Central Java Evaluating IPM Training, Paper No. 93–3, Wageningen Agricultural University (Netherlands). van de Fliert, E., Dung, N. T., Henriksen, O. and Dalsgaard, P. T. J. (2007) “From Collectives to Collective Decision-Making and Action: Farmer Field Schools in Vietnam,” Journal of Agricultural Education and Extension, 13 (3), 245–56. Friis-Hansen, E. (2004) “Smallholder Agricultural Technology Development in Soroti District: Synergy between NAADS and Farmer Field Schools,” Ugandan Journal of Agricultural Sciences, 9 (1), 250–7. Government of Uganda (2001) “Plan for Modernization of Agriculture: Eradicating Poverty in Uganda, Ministry of Agriculture, Animal Industry and Fisheries,” Ministry of Finance, Planning and Economic Development (Kampala). Government of Uganda (2007) National Agricultural Advisory Services: Annual Report 2005/6. Ministry of Agriculture, Animal Industry and Fisheries (Kampala). Hanson, J. C., Just, R. E. and Lainez, J. R. (2006) “Evaluating a Public-Funded, PrivatelyDelivered Agricultural Extension System in Honduras,” Journal of Extension Systems, 22 (2), 14–27. Hayami, Y. (2009) “Social Capital, Human Capital and the Community Mechanism: Toward a Conceptual Framework for Economists,” Journal of Development Studies, 45 (1), 96–123. Kazigati, G. (2005) “Performance of NAADS Districts: Kabarole,” in Proceedings of the Mid-term Review of National Agricultural Advisory Services (Kampala: Ministry of Agriculture, Animal Industry and Fisheries), 44–6. Keynan, G., Olin, M. and Dinar, A. (1977) “Co-financed Public Extension in Nicaragua,” World Bank Research Observer, 12 (2), 225–47. Kibwika, P. (2006) Learning to Make Change: Developing Innovation Competence for Recreating the African University of the 21st Century (Wageningen: Wageningen Academic Publishers). Kidd, A., Lamers, J., Picarelli, P. and Hoffmann, P. (2000) “Privatizing Agricultural Extension: Caveat Emptor,” Journal of Rural Studies, 16 (1), 95–102. Kusumaningrum, D. (2007) “Repackaging Participatory Approach as an Option for Rural Development?” United Nations Economic and Social Commission for Asia and the Pacific, CASPA Flash, 5 (2), February. Meinzen-Dick, R. and Zwarteveen. M. (1998) “Gendered Participation in Water Management: Issues and Illustrations from Water Users’ Associations in South Asia,” Agriculture and Human Values, 15, 337–45. Muwonge, A. (2007) “Local Government Financing and Provision in an Institutionally Constrained Decentralized System: The Case of Agricultural Extension in Uganda,” Doctoral dissertation, Andrew Young School of Policy Studies, Georgia State University (Athens). NAADS (2004) National Agricultural Advisory Services, Revised Implementation Guidelines 2004/05, NAADS Secretariat (Kampala).
9780230_274587_13_cha12.indd 207
6/23/2010 12:17:52 PM
208 Gershon Feder et al. Nyanzi, A. (2005) “Performance of NAADS Districts: Mukono,” in Proceedings of the Mid-term Review of National Agricultural Advisory Services (Kampala: Ministry of Agriculture, Animal Industry and Fisheries, Kampala), 46. Obaa, B., Mutimba, J. and Semana, A. R. (2005) “Prioritizing Farmers’ Extension Needs in a Publicly-Funded Contract System of Extension: A Case Study from Mukono District, Uganda,” Working Paper Series, No. 147. Agricultural Research and Extension Network, Overseas Development Institute (London). Parkinson, S. (2008) Learning Participation in Rural Development: A Study of Uganda’s National Agricultural Advisory Services, Unpublished PhD thesis (Guelph, Ontario: University of Guelph). Platteau, J. P. and Abraham, A. (2002) “Participatory Development in the Presence of Endogenous Community Imperfections,” Journal of Development Studies, 39 (2), 104–36. Poteete, A. R. and Ostrom, E. (2004) “Heterogeneity, Group Size and Collective Action: The Role of Institutions in Forest Management,” Development and Change, 35 (3), 435–61. Purcell, D. L. and Anderson, J. R. (1997) Agricultural Extension and Research: Achievements and Problems in National Systems (Washington, DC: World Bank). Quisumbing, A. R. (ed.) (2003) Household Decisions, Gender, and Development: A Synthesis of Recent Research (Washington, DC: International Food Policy Research Institute and Johns Hopkins University Press). Ricker-Gilbert, J., Norton, G. W., Alwang, J., Miah, M. and Feder, G. (2008) “CostEffectiveness of Alternative Integrated Pest Management Extension Methods: An Example from Bangladesh,” Review of Agricultural Economics, 30 (2), 252–69. Rivera, W. M. and Alex, G. (eds.) (2005) Extension Reform for Rural Development: Case Studies of International Initiatives, Vols. 1–5 (Washington, DC: World Bank and United States Agency for International Development). Rivera, W. M., Qamar, K. M. and Crowder V. L. (2001) Agricultural and Rural Extension Worldwide: Options for Institutional Reform in Developing Countries (Rome: FAO). Rivera, W. M. and Zijp, W. (eds.) (2002) Contracting for Agricultural Extension: International Case Studies and Emerging Practices (Wallingford: CABI Publishing). Sulaiman, R. and Hall, A. (2008) “The Fallacy of Universal Solutions in Extension: Is ATMA the New T and V?” LINK News Bulletin, United Nations University, September, 1–4, at http://innovationstudies.org/index.php?option=com_docman and task=cat_view and gid=45 and Itemid=99999999. Swanson, B. E. (2008) Global Review of Good Agricultural Extension and Advisory Service Practices (Rome: FAO). Swanson, B. (2009) “The Fallacy of Universal Solutions in Extension: Comment,” Blog entry in LINK Blog, at http://innovationstudies.org/index.php?option=com_ myblog and show=The-Fallacy-of-Universal-Solutions-in-Extension-139.html and Itemid. Tripp, R., Wijeratne, M. and Piyadasa, V. H. (2005) “What Should We Expect from Farmer Field Schools? A Sri Lanka Case Study,” World Development, 33 (10), 1705–20. Tyagi, Y. and Verma, S. (2004) “Economic Rate of Return of the Innovations in Technology Dissemination Component of National Agricultural Technology Project,” Report Submitted to National Institute for Agricultural Extension Management (MANAGE) (Hyderabad, India). World Bank (2005) “Implementation Completion Report for the Indonesia Decentralized Agricultural and Forestry Extension Project 2,” Washington, DC, November 25.
9780230_274587_13_cha12.indd 208
6/23/2010 12:17:53 PM
Part IV Community and Market
9780230_274587_14_cha13.indd 209
6/23/2010 12:17:59 PM
9780230_274587_14_cha13.indd 210
6/23/2010 12:17:59 PM
13 The Community Mechanism of Contract Enforcement: What are the Differences between Rural Communities and Industrial Clusters? Keijiro Otsuka and Tetsushi Sonobe
13.1
Introduction
Because of asymmetric information, markets tend to fail due to moral hazard, adverse selection, and holdup. Such problems tend to be serious in the context of the dynamic development of an economy, where new technologies are continuously adopted, new products are widely introduced, and new markets are successively developed. But if markets fail to function, an economy is unlikely to develop. Thus, the question is how such market failures can be reduced so as to stimulate market-based development. In this chapter we would like to explore whether effective mechanisms exist that reduce the extent of market failures by inducing the cooperative behaviors of transacting parties in developing countries. Hayami (2009) argues that a rural community has an inherent informal mechanism to enforce contracts effectively among community members, partly because the members know each other very well and, hence, the information asymmetry is not so serious, and partly because the penalty for dishonest behavior is high in communities characterized by multifaceted and enduring personal relationships.45 In such communities, social capital based on mutual trust will be developed to induce cooperative behavior, so that markets tend to work well. Hayami developed this idea based on his study of rural villages in the Philippines in collaboration with Kikuchi (Hayami and Kikuchi, 1982), and elaborated the idea further by applying it to the case of tenancy contracts (Hayami and Otsuka, 1993), and also to the case of the marketing of agricultural products in Indonesia (Hayami and Kawagoe, 1993). 211
9780230_274587_14_cha13.indd 211
6/23/2010 12:17:59 PM
212
Keijiro Otsuka and Tetsushi Sonobe
The question arises as to whether similar mechanisms exist outside rural communities. In this connection, Sonobe and Otsuka (2006) argue that the industrial cluster, which may be defined as a geographical concentration of enterprises producing similar and related products, is a voluntarily formed “community” by entrepreneurs, not necessarily from the same community, because of the benefits arising from an environment in which everybody knows everybody else and in which rumors can become quickly widespread, as in rural communities. Thus, the cost of contract enforcement is expected to be low within an industrial cluster, which explains, at least partly, why indigenously developed industries in developing countries are mostly cluster-based.46 The first purpose of this chapter is to examine the hypothesis that the development of both rural communities and indigenously established industrial clusters is based commonly on the community mechanism of contract enforcement. Because of the low transaction costs, the imitation of new ideas is widespread in industrial clusters. Thus, Sonobe and Otsuka (2006) argue that private efforts to innovate fall short of the social optimum because of information spillovers, or imitation. So long as innovation is an engine of industrial development, the development of industrial clusters is hampered by imitation. There are several ways to overcome this market failure, such as the use of patent systems and the provision of government support for innovations. The same issue should potentially arise in rural communities where imitation, or learning from others, is a common practice. The second purpose of this article is to consider how to promote innovations in rural industries and the production of high-value crops in rural communities, where information on new technology is widely shared. The organization of this article is as follows. We review, first, Hayami’s arguments on contract enforcement in rural communities in Section 13.2 and, second, the empirical literature on industrial clusters in Section 13.3. We discuss how the issue of weak incentives to innovate can be overcome in both industrial clusters and rural communities in Section 13.4. Finally we conclude this chapter by providing policy implications for the development of both rural communities and industrial clusters.
13.2 Contract enforcement in rural communities According to Hayami and Godo (2005, p. 311), the community is defined as “the organization that guides community members to voluntary cooperation based upon close personal ties and mutual trust ...” Relationships of mutual trust are created and maintained through long-term, continuous, and multiple transactions among a group of people in the community. Since people know one another and meet regularly, and since they are engaged in similar and related activities (e.g., farming and processing of farm products), information asymmetry is not as serious as in outside societies. Moreover,
9780230_274587_14_cha13.indd 212
6/23/2010 12:17:59 PM
The Community Mechanism of Contract Enforcement
213
the circulation of malicious gossip by word of mouth will lead to heavy penalties for those who commit moral hazard. Thus, if a tenant shirks or fails to maintain soil fertility, their conduct is likely to not only be detected by his present landlord but also to be known by other community members including potential landlords, so that he will lose the opportunity to work as a tenant in his community. Indeed, share tenants, who are considered to shirk because of the disincentive effect of output sharing on a tenant’s work effort (i.e., the so-called Marshallian inefficiency of share tenancy), are found to work as hard as owner-cultivators and leasehold tenants according to the extensive review of empirical literature (Otsuka, 2007; Otsuka et al., 1992). This is likely due to the fact that landlords select trustworthy community members as share tenants, for example relatives and sons of close friends, so that the community mechanism of contract enforcement works. Such results are widely reported by empirical studies on share tenancy in Asia (e.g., Sadoulet et al., 1994, 1997).47 Hayami (e.g., 1996, p. 1162) emphasizes the fundamental contribution of trading activities to the development of rural communities: “... profitable opportunities created by new technology and improved infrastructure such as roads cannot be exploited without the activities of middlemen.” According to Hayami and Kawagoe (1993), in both long-settled Java and newly opened Sumatra, indigenous entrepreneurs belonging to the upper peasantry actively engaged in the collection of agricultural products such as fresh vegetables, corn, and soybeans, their processing, such as the production of tempe and tofu, and the shipment to large cities, such as Jakarta. These village-based traders out-competed town-based ethnic Chinese. The strength of these village-based traders lay in their ability to enforce contracts with farmers and low-rank collectors by means of the community mechanism of contract enforcement. Small farmers benefit from the sale of commercial crops to village-based traders, as the trading sector is highly competitive. In this way, the community mechanism of contract enforcement contributes to the development of rural communities. As Asian economies have grown rapidly, the importance of nonfarm income in rural households has grown enormously, which directly contributes to the significant reduction in rural poverty (Hayami and Kikuchi, 2000; Otsuka et al., 2009). There is a tendency that rural nonfarm jobs do not require as much education as urban jobs and, hence, even uneducated rural workers can find reasonably remunerative nonfarm jobs. Recognizing the importance of rural industrialization, Hayami (1998) observes the critical role played by relational contracting based on long-term and continuous contract relations between small, rural-based manufacturers and traders or agents of large, urban-based enterprises. Traders bring marketing information regarding urban and foreign markets and new production methods to the rural communities, and often provide credit to local producers as well. Otsuka (1998) observes that many of the owners/managers of rural
9780230_274587_14_cha13.indd 213
6/23/2010 12:17:59 PM
214
Keijiro Otsuka and Tetsushi Sonobe
enterprises have had work experience at the urban enterprises with which they have subcontract relationships. Thus, the trust forged through longterm employer-employee and contractor-subcontractor relationships contributes to the development of rural industries. Needless to say, such rural producers are born in the locality where they operate factories, so that they know many of their employees personally.
13.3 Contract enforcement in industrial clusters Advantages of industrial clusters Manufacturing industries in developing countries, which have been initiated by native entrepreneurs, are mostly cluster-based and led by SMEs (small and medium-sized enterprises), particularly in low-income countries. Good examples are the garment and shoe industries. Machinery industries are also often clustered, as the inter-enterprise transactions of a large number of parts and components are required for the production of final products. In our observation, the formation of industrial clusters has been booming in many areas of China. At present, industrial clusters are ubiquitous not only in East Asia, including Japan and Taiwan, but also in Southeast and South Asia and, to a lesser extent, sub-Saharan Africa. According to Marshall (1920), the advantages of industrial clusters or agglomeration economies are: (1) information spillovers or imitation, (2) the division and specialization of labor among enterprises producing parts, components, and final products, and (3) the development of skilled labor markets. While we do not have any objections to these advantages associated with industrial clusters, we would like to point out that the major sources of information spillovers are the leakage of technological information by partsuppliers and skilled workers recruited from innovative enterprises. Thus, information spillovers, the division of labor, and the development of skilled labor markets are intimately related. While Marshall recognizes the role of industrial clusters in facilitating the division of labor among manufacturing enterprises, he did not explicitly consider the role of industrial clusters in facilitating the transactions between manufacturers and traders. Because of the geographical proximity, an enterprise manager knows many of the other enterprise managers, partsuppliers, and traders. Indeed they often exchange information about the personality, performance, and conduct of manufacturers and traders. Since most indigenous industries are clustered, it is difficult to provide quantitative evidence for the role of the “community mechanism of contract enforcement” in industrial clusters by comparing clustered areas with nonclustered areas. However, the fact that the transactions among manufacturing enterprises and between manufacturing enterprises and traders are active in industrial clusters strongly suggests that it works well, not only in rural communities but also in industrial clusters.
9780230_274587_14_cha13.indd 214
6/23/2010 12:17:59 PM
The Community Mechanism of Contract Enforcement
215
Two types of industrial clusters There are rural and urban industrial clusters, which differ in the type of products produced, the type of entrepreneurs who lead the cluster development, and the way in which the industries are initiated. In urban areas, a variety of goods and human resources are available. Such diversity of urban economic activities turns out to be an advantage for the production of skillintensive products and products that requires a variety of intermediate inputs. A typical product of urban industrial clusters is machinery, whose production requires substantial skills and mechanical knowledge as well as a large quantity of parts and materials. Since engineers are often the initiators, we call this type of clusters “engineer-led” (Sonobe and Otsuka, 2006). As the cluster grows large and gets congested, it often sprawls out or is relocated to the suburbs. Urban industrial clusters can be viewed as quasi-communities, in which the community mechanism of contract enforcement functions even though the constituents are not necessarily from the same community. We observed such quasi-communities in the machine-tool cluster in Taichung, Taiwan, the motorcycle clusters in Hamamatsu, Japan and Chongqing, China, and the metalworking and car-repair clusters in Kumasi, Ghana, and Nairobi, Kenya (Sonobe and Otsuka, 2006; Sonobe et al., 2009). On the other hand, the advantage of rural areas in manufacturing lies in the abundance of low-wage labor. However, this advantage will be realized only if the production using low-wage labor is connected to large demand in urban or foreign markets. Thus, the central players in the development of rural industrial clusters tend to be traders. Because of the necessity of the link to urban and foreign markets, the location of rural clusters cannot be too far away from major cities. In addition, their location is often characterized by weak comparative advantage in agriculture due to the shortage of extensive fertile land. Because skilled workers, designers, and engineers are unavailable in rural areas, technically simple products, such as garments, shoes, woodwork products, and simple metal products, are produced.48 Roles of traders in industrial clusters In what follows, we would like to show that traders play the role of entrepreneurs in rural industrial clusters by reviewing the development of the following four rural clusters: (1) the garment industry in Bingo in Japan, (2) the garment industry in Jili in China, (3) the electric appliance industry in Wenzhou in China, and (4) the garment industry in the suburbs of Hanoi in Vietnam (Nam et al., 2010; Sonobe and Otsuka, 2006).49 In each of these four industrial clusters, there are 100 to 1,000 manufacturing enterprises producing final products, aside from parts-suppliers and traders. There is remarkable resemblance between the garment clusters in Bingo in Hiroshima Prefecture in Japan, and in Jili (which can also be spelt Zhili) in Zhejiang Province in China. First, they are located in rural areas but not
9780230_274587_14_cha13.indd 215
6/23/2010 12:18:00 PM
216
Keijiro Otsuka and Tetsushi Sonobe
very far away from large cities (e.g., 200 kms), which are Osaka in the case of Bingo and Shanghai in the case of Jili.50 Second, because of the poor soil and limited availability of farmland, the wives and other members of farm households had to supplement their meager farm incomes by weaving traditional cotton cloth (called kasuri) in the case of Bingo, and producing pillowcases and other miscellaneous merchandise in the case of Jili. Third, to sell these products, local traders traveled around various parts of the country, and in this way the tradition of commerce was established in these rural townships. Fourth, it was these local traders who introduced the current major products, that is, working clothes in Bingo and infant clothing in Jili. In Bingo, the development of the garment industry was led by former local traders who used to deal in traditional cotton cloth. Their presence remained strong for many years after the industry was initiated before the war. Even as of 1968, the enterprises owned and operated by the former local traders accounted for 41 percent (see Table 13.1). The spin-offs, that is, imitators who had worked for other garment enterprises before, have accounted for 42 to 56 percent for several decades since the late 1960s. Their entry was prompted by the high entrepreneurial profits earned by their employers, that is, the former local traders. As such new entries expanded the production of the cluster, the division of labor involving a large number of small subcontractors specializing in particular parts of the entire production process was highly developed. This development rested on the community mechanism of contract enforcement, like rural industrialization in other places in Asia (Hayami, 1998; Otsuka, 1998, 2007). It is also interesting to observe that the former local traders were not only the first to introduce the factory production system to the Bingo area but also the first to deal directly with large urban retailers, such as department stores, when indirect transaction with small retailers through local traders was the common practice. Direct transactions with urban retailers became profitable for the former local traders because they began producing high-quality, differentiated products. In general, increases in direct transactions are associated with quality improvement and attempts to establish brand names.
Table 13.1 Proportions of enterprise managers by former occupation and formal schooling in Bingo, 1968–98 Former occupation (%)
Average schooling (years)
Year
Local traders
Spin-offs
Others
No. of sample enterprises
1968 1977 1986 1998
41 33 31 35
42 53 56 52
17 14 13 13
66 86 89 75
9780230_274587_14_cha13.indd 216
Local traders 10.2 11.5 13.1 14.1
Spin-offs 11.9 12.0 12.1 13.1
Others 9.7 9.9 12.0 13.6
6/23/2010 12:18:00 PM
The Community Mechanism of Contract Enforcement
217
Table 13.2 Previous occupation and years of schooling of enterprise managers by timing of new entry in Jili, China Year Occupation (%) Farmer Factory worker Marketing specialist Other manager and specialist Total No. of sample enterprises Years of schooling
1980–90
1991–5
1996–9
52 19 4 26 100
33 33 13 23 100
32 23 23 23 100
27 7.1
66 7.4
98 7.5
In Jili, the proportion of managers who were formerly marketing specialists (i.e., local traders and those who had worked in the marketing division of the garment enterprises) gradually increased, as marketing assumed increasing importance in the development process of the garment town (see Table 13.2). Like their counterparts in Bingo, they produced higher-quality products and engaged more actively in direct transactions with urban traders, and achieved higher labor productivity. According to Fleisher et al. (2008), the last decade has witnessed a great development of the subcontracting system in the Jili cluster, which now ranks as one of the largest production centers of children’s clothing in the world. The production of electric fittings, such as power outlets and switches, in Wenzhou in Zhejiang Province was initiated by a farmer in his teens, but the development of this industry was led by local traders. Like the garment town of Jili, Wenzhou used to be a poor rural area and had a strong tradition of commerce, in which low-quality merchandise produced by farm households were sold by local traders in large cities across the country.51 Indeed, “made in Wenzhou” used to be synonymous with inferior products for Chinese consumers. The electric fitting enterprises in Wenzhou, therefore, had to overcome this bad reputation in order to market their improved products when consumers with rising income levels became increasingly fastidious about product quality from the late 1980s. This challenge was first met by a former trader in the late 1980s. Moreover, the massive new entry of former traders began in this period (Table 13.3). Clearly, the development of this cluster is trader-led. After improving the quality of their products, the entrepreneurs employed a number of sales agents selling their products in major cities across the country. They prohibited the sales agents from dealing in products of other producers in order to differentiate their improved products from the lowquality products of other enterprises, and to establish their brand names. While the use of sales agents might invite the agency problem, each of the
9780230_274587_14_cha13.indd 217
6/23/2010 12:18:00 PM
218 Keijiro Otsuka and Tetsushi Sonobe
Table 13.3 Characteristics of enterprise founders in the electric appliance industry by timing of new entry in Wenzhou, China
No. of enterprises Years of schooling
Before 1980
1981–5
1986–90
1991–5
1995–2000
4 8.0
32 9.8
30 10.0
36 10.5
10 10.9
50 0 25 0 0 25
13 25 25 3 9 25
7 17 43 13 10 10
3 17 64 6 6 6
0 10 50 0 10 30
Occupation (%) Farmers Factory workers Traders Engineers Managers Others
Table 13.4 Characteristics of enterprise managers by period of establishment in the garment industry in the suburbs of Hanoi
Years of prior experience in garment marketing Proportion of enterprise managers who have work experience in garment marketing (%) Years of prior experience of working at garment factory Number of Vietkieu relatives per enterprise manager Proportion of enterprise managers who have Vietkieu relatives (%) Average years of schooling Number of enterprises
Before and in 2000
After 2000
0.4
1.4
15
57
0.3
5.7
0.6
2.3
19
50
10.0 124
11.3 14
leading enterprises employed hundreds of agents. The reason why they did so was that there were a large number of traders who had migrated from Wenzhou and settled in major cities. We found that more than 95 percent of the sales agents were such Wenzhou traders. Thus, the electric fitting enterprises used the community mechanism of contract enforcement in long-distance transactions between the major urban markets and the industrial cluster. As is shown in Table 13.4, similar observations can be made of the garment cluster in the suburbs of Hanoi, which has had a tradition of producing garment products for the last several decades under the socialist system. There
9780230_274587_14_cha13.indd 218
6/23/2010 12:18:00 PM
The Community Mechanism of Contract Enforcement
219
are thousands of household workshops producing parts for large registered enterprises under subcontracts, and the owners of the small workshops visit large enterprises almost every day to receive materials and deliver finished parts. Obviously, the community mechanism of contract enforcement is at work in coordinating the dense subcontracting relationships with a large number of household producers, all of whom are native residents in the community.52 Reflecting the increasing importance of improving product quality and marketing such products, the number of proprietors who were former local traders increased appreciably after 2000. Also the fact that the number of years of prior experience working in a garment factory has increased indicates the increasing importance of spin-offs in this cluster. This garment cluster is the home village of a large number of Vietkieu, overseas Vietnamese traders who went to Russia and Eastern Europe in the late 1980s and started small businesses. Once they settled there, they began going back and forth between their home village and Russia or Eastern Europe to engage in exporting the garment products. They tend to choose manufacturers who are relatives in order to take advantage of the community mechanism of contract enforcement. Surprisingly, a common practice is that these Vietkieu pay the local producers only once a year after Christmas, that is, after a major business season. Such contracts can hardly be agreed upon by local producers without strong trust. Interestingly, both prior marketing experience and the number of Vietkieu relatives are the critical variables that significantly affect the value-added and export ratio of the enterprises, according to regression analyses.53
13.4
Community as a potential obstacle to innovations
As in other sectors of the economy, successful innovation is the key to the development of industrial clusters. More often than not new industries in developing countries are born based on the imitation of imported commodities from developed countries. Needless to say, the initiation of new production is an innovation, as it is risky and the establishment of a production system using poor substitutes for unavailable parts, materials, and machines is not an easy task. Indeed, the initiator receives Schumpeterian entrepreneurial profit. Such profit attracts a swarm of imitators, which tends to result in the “excess supply” of low-quality products, reduced product prices, and lower profitability. Sonobe and Otsuka (2006) found that in order for an industrial cluster to keep growing, it has to have another set of innovations for upgrading product quality, production organization, and marketing methods. The production of an improved product may be achieved simply by employing competent designers, engineers, and skilled workers, and by using modern machines and expensive materials. While a higher-quality product is more costly to produce, it cannot immediately command a higher price in the
9780230_274587_14_cha13.indd 219
6/23/2010 12:18:00 PM
220
Keijiro Otsuka and Tetsushi Sonobe
market until consumers are convinced that the product quality has been improved. Therefore, brand names must be established and products must be sold directly to large retailers or through exclusive sales agents. Since the improved products are differentiated products, their production requires specific new parts, which embody new ideas. Thus, the entrepreneur must build a long-term relationship with trustworthy parts-suppliers or produce key parts internally. In order to assure the high quality of products, a quality control system and a labor management system are required. If they are successful in such product and management innovations, the entrepreneurs need to look for export possibilities and expansion of production, by means of merger or loans from the formal financial sector. Although each “innovation” may be better termed an “improvement,” carrying out such multifaceted “improvements” is a formidable task. Thus, the entrepreneurs who introduce innovations to the industrial clusters are highly educated. Tables 13.1 to 13.4 show that the schooling levels of enterprise managers were high, and increased over time faster than the national averages in the respective countries. Furthermore, the schooling of enterprise managers is a highly significant variable in explaining the frequent use of direct transactions, the high prices of products, and other indicators of product quality, not only in our East Asian studies (Sonobe and Otsuka, 2006) but also in all the other studies we have conducted. It must also be emphasized that the availability of useful human resources for innovations, such as skilled workers, engineers, and parts-suppliers, is a distinct advantage of industrial clusters in carrying out multifaceted innovations. Another factor facilitating successful innovations is learning from abroad. Although we have only anecdotal evidence, entrepreneurs in Taiwan learned much from the Japanese experience, those in China learned from Taiwan and Japan, and those in Vietnam are learning from China. In Addis Ababa, some of the most successful shoemakers travel abroad, mostly to Italy, several times each year to obtain new knowledge. In other words, the international transfer of technology and management know-how contributes to innovations in industrial clusters in developing countries.54 A major obstacle to the development of industrial clusters is the endemic imitation of innovations, which reduces the private returns to innovation below the social returns. Thus, multifaceted innovations for quality improvement do not take place as often as they should from the viewpoint of the society or the cluster. Patent protection is unfeasible, as many of the new ideas are improvement types. The insufficient incentive to innovate is also a major obstacle to the development of rural communities. Recognizing this issue, Hayami (2009, p. 115) points out: “Our knowledge is grossly insufficient on the ways and means to induce the community mechanism to act toward promoting innovation in response to changes in economic environment.”
9780230_274587_14_cha13.indd 220
6/23/2010 12:18:00 PM
The Community Mechanism of Contract Enforcement
221
There are a few countermeasures to the insufficient innovation problem. One is to organize “community” actions. In order to supply a large volume of improved and standardized farm products, small farmers form production and marketing groups in some cases, while others participate in collective production schemes organized by cooperatives, and in other places larger farmers subcontract production to smallholders (McCullough et al., 2008). Schmitz (1995) attaches importance to the role of voluntarily organized production associations in upgrading the quality of products in the industrial clusters. Altenburg and Meyer-Stamer (1999), however, question the effectiveness of such systems from the viewpoint of transaction costs. It is not clear how effective organized community action is. Since its success depends greatly on the ability of its leaders, it is important to nurture social entrepreneurs. Questions, however, arise as to who the social entrepreneurs are, what roles they are expected to play, and what changes are needed in the traditional community norms. Hayami (2009) argues that social entrepreneurs are those who transform the traditional norms and conventions into new forms of cooperation toward innovations. A possible interpretation of his argument is that social entrepreneurs may be able to create an incentive system to appreciate inventions and improvements made by other members of a rural community or an industrial cluster, and to recognize them as contributions to the welfare of the community. Such recognition may encourage farmers or enterprise managers to attempt innovations even if imitations are expected to be common. According to Hayami (1996, 1998, 2009), social entrepreneurs can be efficient providers of new business information as a local public-good. For example, a village association can function as a community technology center which introduces advanced knowledge and collects market information for the association members. How far such organized community actions led by social entrepreneurs can contribute to the improvement of farm or manufactured products remains to be seen. In our view, the government or international organizations should take the initiative in providing technological and management knowledge conducive to innovations for promising social entrepreneurs. Sonobe and Otsuka (2006) advocate training programs in which enterprise managers learn the skills of marketing, quality improvement, and production management. Such entrepreneurial training will have greater effects if it is provided for enterprise managers in stagnant clusters who are thirsty for new knowledge.55
13.5
Conclusion
It is difficult to assess how the development of industrial clusters benefits from the community mechanism of contract enforcement because industrial production outside clusters is so uncommon that the comparison of clusters
9780230_274587_14_cha13.indd 221
6/23/2010 12:18:00 PM
222
Keijiro Otsuka and Tetsushi Sonobe
and nonclusters is unfeasible in many industries. Moreover, punishment of dishonest behaviors by the community will be seldom required if it is, in fact, effective. Nonetheless, the fact that transactions among manufacturers and between them and traders are highly active in industrial clusters in a large number of countries, indicates that industrial clusters are equipped with an excellent mechanism of contract enforcement which reduces transaction costs and facilitates the division of labor. As Hayami forcefully argues, the community and the market are complementary, as the community mitigates market failures. In this respect, industrial clusters are no different from rural communities. Hence, assisting the formation of industrial clusters by investing public funds in infrastructure in general, and in industrial zones in particular, makes sense. Then, all three pillars of economic organizations, that is, community, market, and state, can play complementary roles in the development of industrial clusters. From the viewpoint of sustainable economic development, the critical question is how to create social and economic environments conducive to innovations in rural communities and industrial clusters. Since the transaction of information, which cannot be patented, is known to be costly, markets for innovative ideas do not exist or fail to function. This market failure can be resolved by the community mechanism, if the community has the capacity to organize community actions and supply “local public goods” efficiently. It is, however, possible that the government, international organizations, or aid agencies in developed countries may be in a better position to fill the gap between the social and private net benefits from innovations. An exploration of how to design efficient systems facilitating innovations for rural communities and industrial clusters is a major remaining issue in development economics.
References Akoten, J. and Otsuka, K. (2007) “From Tailors to Mini-Manufacturers: The Role of Traders in the Transformation of Garment Enterprises in Kenya,” Journal of African Economies, 16 (4), 564–95. Altenburg, T. and Meyer-Stamer, J. (1999) “How to Promote Clusters: Policy Experience from Latin America,” World Development, 27 (9), 1693–713. Fleisher, B., Hu, D., McGuire, W. and Zhang, X. (2008) “The Evolution of an Industrial Cluster in China,” mimeo, Department of Economics, Ohio State University. Hayami, Y. (1996) “The Peasant in Economic Modernization,” American Journal of Agricultural Economics, 78 (5), 36–53. Hayami, Y. (ed.) (1998) Toward the Rural-Based Development of Commerce and Industry: Selected Experiences from East Asia (Washington, DC: World Bank Economic Development Institute). Hayami, Y. (2009) “Social Capital, Human Capital and the Community Mechnism: Toward a Conceptual Framework for Economists,” Journal of Development Studies, 45 (1), 96–123.
9780230_274587_14_cha13.indd 222
6/23/2010 12:18:00 PM
The Community Mechanism of Contract Enforcement
223
Hayami, Y., Dikishit, A. K. and Mishra, S. N. (2006) “Waste Pickers and Collectors in Delhi: Poverty and Environment in an Urban Informal Sector,” Journal of Development Studies, 42 (2), 41–69. Hayami, Y. and Godo, Y. (2005) Development Economics: From the Poverty to the Wealth of Nations (Oxford: Oxford University Press). Hayami, Y. and Kawagoe, T. (1993) The Agrarian Origin of Commerce and Industry: A Study of Peasant Marketing in Indonesia (New York: St Martin’s Press). Hayami, Y. and Kikuchi, M. (1982) Asian Village Economy at the Crossroads: An Economic Approach to Institutional Change (Baltimore: Johns Hopkins University Press). Hayami, Y. and Kikuchi, M. (2000) A Rice Village Saga: Three Decades of Green Revolution in the Philippines (London: Macmillan). Hayami, Y. and Otsuka, K. (1993) The Economics of Contract Choice: An Agrarian Perspective (Oxford: Clarendon Press). Holden, S., Otsuka, K. and Place, F. (2009) The Emergence of Land Markets in Africa: Assessing the Impacts on Poverty, Equity, and Efficiency (Baltimore, MD: Resources for the Future). McCullough, E. B., Pingali, P. and Stamoulis, K. G. (2008) “Small Farms and the Transformation of Food Systems: An Overview,” in E. B. McCullough, P. Pingali and K. G. Satmoulis (eds.), The Transformation of Agri-Food Systems: Globalization, Supply Chains, and Smallholder Farmers (Rome: Food and Agriculture Organization). Marshall, A. (1920) Principles of Economics (London: Macmillan). Nam, V. H., Sonobe, T. and Otsuka, K. (2009) “An Inquiry into the Transformation Process of Village-based Industrial Clusters: The Case of an Iron and Steel Cluster in Northern Vietnam,” Journal of Comparative Economics, 37 (4), 568–81. Nam, V. H., Sonobe, T. and Otsuka, K. (2010) “An Inquiry into the Development Process of Village Industries: The Case of a Knitwear Cluster in Northern Vietnam,” Journal of Development Studies, 46 (2), 312–30. Otsuka, K. (1998) “Rural Industrialization in East Asia,” in Y. Hayami and M. Aoki (eds.), The Institutional Foundations of East Asian Economic Development (London: Macmillan). Otsuka, K. (2007) “Efficiency and Equity Effects of Land Markets,” in R. E. Evenson and P. Pingali (eds.), Handbook of Agricultural Economics, Volume III (Amsterdam: Elsevier). Otsuka, K., Chuma, H. and Hayami, Y. (1992) “Land and Labor Contracts in Agrarian Economies: Theories and Facts,” Journal of Economic Literature, 30 (4), 1965–2018. Otsuka, K., Estudillo, J. P. and Sawada, Y. (eds.) (2009) Rural Poverty and Income Dynamics in Asia and Africa (London: Routledge). Sadoulet, E., Fukui, S. and de Janvry, A. (1997) “The Meaning of Kinship in Sharecropping Contracts,” American Journal of Agricultural Economics, 79 (2), 394– 407. Sadoulet, E., de Janvry, A. and Fukui, S. (1994) “Efficient Share Tenancy Contracts under Risk: The Case of Two Rice-Growing Villages in Thailand,” Journal of Development Economics, 45 (2), 225–43. Schmitz, H. (1995) “Collective Efficiency: Growth Path for Small-Scale Industry,” Journal of Development Studies, 31 (4), 529–66. Sonobe, T., Akoten, J. and Otsuka, K. (2009) “An Exploration into the Successful Development of the Leather-shoe Industry in Ethiopia,” Review of Development Economics, 13 (4), 719–36. Sonobe, T. and Otsuka, K. (2006) Cluster-Based Industrial Development: An East Asian Model (Basingstoke, UK: Palgrave Macmillan).
9780230_274587_14_cha13.indd 223
6/23/2010 12:18:01 PM
14 The Transformation of Hayami’s Village* Jonna P. Estudillo, Yasuyuki Sawada, Kei Kajisa, Nobuhiko Fuwa, and Masao Kikuchi
14.1
Introduction
One of the most important areas of inquiries in development economics is the choice of the most appropriate economic system that promotes economic development. In reality, an economic system is a combination of various interactions of three important organizations, that is, the community, market, and state. A community is defined as a small group of people characterized by intensive social interactions that tend to minimize information asymmetry between transacting parties, thereby lowering transaction costs of economic activities (Aoki and Hayami, 2001). A market is an organization that coordinates the activities of profit-maximizing firms and utility-maximizing individuals under the guidance of prices, while a state specializes in the supply of public goods by means of legitimate coercive power (Hayami and Godo, 2005). Community, market, and state can be potentially combined to lower the costs of economic transactions, leading to an allocation of resources that is both efficient and equitable. Traditional customs and moral principles that are slow to change could be corrected by the promulgation of dynamic legal laws. Rule enforcement and intensive social interactions that are limited to small communities only can be corrected by common rules set by the state. Bilateral monopolistic markets in land and labor can be abolished through the penetration of the market system and the development of the nonfarm sector. The community has the potential of correcting market failures. Externalities of both biological and physical types can be corrected by the community through customary rules and principles, and the high transaction costs due to quality uncertainty in both the product and labor markets could be mitigated in a patron-client relationship involving interlinked * The authors benefited greatly from the comments given by Elisabeth Sadoulet, Cristina C. David and Keijiro Otsuka. Vivencio Marciano, Ester Marciano, and Jan Irish Villegas provided excellent research assistance. The usual caveat applies.
224
9780230_274587_15_cha14.indd 224
6/23/2010 5:05:19 PM
The Transformation of Hayami’s Village
225
transactions in land and labor markets and input and output markets (Hayami and Otsuka, 1993). This chapter reviews the literature on the role of the community in the management of irrigation system, land and labor relations, risk coping, social capital, and the development of the rural nonfarm sector. As a typical example of a rural community, we explore community mechanisms in the so-called “Hayami’s village” in the Philippines. The major finding of this piece of work is that rural communities are far from static, but respond systematically and dynamically to changes in economic conditions that often result in an efficient and equitable allocation of resources.
14.2
Description of Hayami’s village
Prof. Hayami calls “his village” East Laguna Village, which is located in the town of Pila in the province of Laguna. In 1966, there were 66 households in the village with a population of 399 (Table 14.1). The largest majority of them were engaged in rice production, either as farmers or agricultural laborers. He conducted 11 rounds of survey in the village between 1966 and 1997, and published three books and 14 journal articles using data from these surveys.56 His scholarly work gave a microscopic picture of the process by which agricultural production, community institutions, and people’s economic well-being evolved in the course of the development of a village economy. Household description Households in the village consist of three major classes: (1) farmers, who operate their own farms, either as tenants or as owners; (2) agricultural
Table 14.1 Number of households in Hayami’s village in the Philippines, 1966–2006 Year
Farmers
Agricultural workers
Nonfarm workers
Total
1966
46 (70)a 54 (50) 53 (34) 51 (21) 36 (9)
20 (30) 55 (50) 98 (62) 150 (62) 138 (32)
0 (0) 0 (0) 7 (4) 41 (17) 254 (59)
66 (100) 109 (100) 158 (100) 242 (100) 428 (100)
1976 1987 1995 2006
Note: a Numbers in parentheses are percentages. Data in 1966, 1976, 1987, and 1995 were drawn from Hayami and Kikuchi (2000, Table 3.3, p. 54), while data in 2006 were drawn from the database of Y. Sawada.
9780230_274587_15_cha14.indd 225
6/23/2010 5:05:19 PM
226
Jonna P. Estudilo et al.
workers, who have no farm to operate but eke out a living from casual farm work; and (3) nonagricultural workers, who are employed solely in the nonagricultural sector either in salaried work or self-employed activities. The number of agricultural-worker households rose more rapidly than did the farmer households from 1966 to 1995 (Table 14.1). The sharp rise in the number of these households can be attributed to population pressure, landreform regulations on tenancy contracts, and increased demand for hired labor associated with the diffusion of MVs and the substitution of hired for family labor. The land-reform laws prohibit the transfer to other parties, except legitimate heirs, of cultivation rights acquired through land reform, thereby limiting the opportunities of landless households to become tenants. A spectacular growth in the number of nonagricultural households from 1995 to 2006 is evident, due to the opening of a relatively low-priced residential subdivision in 2000 that attracted nonagricultural households from nearby areas. The International Rice Research Institute (IRRI) released the first modern variety (MV) of rice, IR8, on November 28, 1966. Adoption of MVs in the village was rapid because of the presence of a well-developed irrigation infrastructure, and access to information regarding new rice technology owing to the village’s proximity to the IRRI and the University of the Philippines at Los Banos. According to the 1976 survey, almost all the farmers adopted MVs (Table 14.2). Yield rose from about 2 tons to about 4–5 tons per ha because of the diffusion of MVs, higher application of fertilizer, and adoption of better cultural practices. The IRRI, the governor of Laguna, the local branch of the Bureau of Agricultural Extension, and relatives and neighbors were all instrumental in the diffusion of MVs in the early period of their release. Landholdings in Hayami’s village are generally small and scattered, and characterized by pervasive landlordism. Land accumulation in the village proceeded slowly, mainly through the pacto de retrovenda (land-mortgaging operations). The relationship between the landlord and the tenant can be considered paternalistic because a large majority of landlords live in nearby places such as the municipality of Pila, local towns in Laguna, and neighboring provinces, and only a few reside in Manila. Five major forms of land tenure exist in the village: (1) ownership, (2) leasehold tenancy, (3) share tenancy, (4) pawning of owned land, and (5) subleasing of leasehold land under three arrangements (Table 14.3). Share tenancy was common in 1966, but it has been increasingly replaced by leasehold tenancy after the implementation of land reform in 1974. Discussion of the land-reform program and its impact on the land tenure system can be found in the section on landlord-tenant relations below. A larger proportion of household income came from farm sources in 1974–6 and 1980–3 (Table 14.4) including rice farming, livestock and poultry raising, propagation of high-value tree crops, backyard vegetable farming,
9780230_274587_15_cha14.indd 226
6/23/2010 5:05:19 PM
The Transformation of Hayami’s Village
227
Table 14.2 Adoption of modern rice varieties and yield in Hayami’s village, 1966–2006 MV adoption (% of farmer adopters)
Yield (tons per ha)
1966 Wet TVa MVb
100 0
1.9 nac
1976 Wet TV MV
3 97
1.1 2.8
1976 Dry TV MV
2 98
2.0 3.2
1987 Wet TV MV
0 100
na 4.0
1987 Dry TV MV
2 98
4.6 4.7
1996 Wet TV MV
0 100
na 4.0
1996 Dry TV MV
0 100
na 5.3
2006 Wet TV MV
0 100
na 4.4
2007 Dry TV MV
0 100
na 4.9
Year and season
Notes: a Traditional variety of rice. b Modern variety of rice. c Means not available. Data Source: Data in 1966, 1976, 1987, and 1995 were drawn from Hayami and Kikuchi (2000, Table 5.1, p. 107), while data in 2006 were calculated from the database of Y. Sawada.
and farm wages. Nonfarm sources included formal salary work, informal self-employed enterprises in commerce, manufacturing, and transport, and remittances. In 1995–6, income from nonfarm activities had become the dominant source of income because of the increase in nonfarm wage earnings and remittances from members working outside the village, including overseas contract workers. The proportion of nonfarm income became
9780230_274587_15_cha14.indd 227
6/23/2010 5:05:19 PM
228 Jonna P. Estudilo et al.
Table 14.3 Tenure distribution of plots in Hayami’s village in the Philippines, 1966–2006 Plot Year
Tenure a
Area
No.
%
Ha
3 19 70 8
1.3 29.9 66.1 6.9 0.5 6.4 104.2
1966
Owned Leasehold Share Sublet Pawning arrangement Sharecrop arrangement Total
2 12 44 5 1 4 63
1976
Owned Leasehold Share Sublet Pawning arrangement Leasehold arrangement Sharecrop arrangement Total
3 44 30 16 5 2 9 93
1987
Owned Leasehold Share Sublet Pawning arrangement Leasehold arrangement Sharecrop arrangement Total
14 39 21 10 3 3 4 84
1995
Owned Pawning of ownership Leasehold Share Sublet Pawning arrangement Sharecrop arrangement Total
25 2 37 13 4 2 2 91
31 2 46 16 5
28.0 0.8 47.7 14.6 5.9 3.5 2.4 97.0
2006
Owned Leasehold Share Sublet Pawning arrangement Sharecrop arrangement Total
15 25 5 4 4 0 49
31 51 10 8
14.53 36.93 3.80 4.24 4.24 0.00 59.50
100 3 48 32 17
100 17 46 25 12
100
100
1.7 67.7 29.7 9.1 1.5 0.8 6.8 108.2 11.6 46.1 21.9 9.0 1.4 1.8 5.8 88.6
% 1 29 63 7
100 2 63 27 8
100 13 52 25 10
100 29 1 49 15 6
100 25 62 6 7
100
Note: aIncludes land under certificate of land transfer and emancipation patent. Data Source: Data in 1966, 1976, 1987, and 1995 were drawn from Hayami and Kikuchi (2000, Table 4.4, p. 87); data in 2006 were calculated from the database of Y. Sawada.
9780230_274587_15_cha14.indd 228
6/23/2010 5:05:19 PM
The Transformation of Hayami’s Village
229
Table 14.4 Sources of household income in Hayami’s village in the Philippines, 1966–2006 Year 1974–6 1980–3 1995–6 2006–7
Totala 58(100) 53(100) 56(100) 76(100)
b
Farm origin
Nonfarm origin
50(87) 33(62) 20(36) 24(32)
8(13) 20(38) 36(64) 52(68)
Note: a Deflated by the CPI outside Metro Manila (1995 = 100). b Numbers in parentheses are percentages. Data Source: Data in 1966, 1976, 1987, and 1995 were drawn from Hayami and Kikuchi (2000, Table 10.3, p. 235); data in 2006 were calculated from the database of Y. Sawada.
particularly large in 2006 because of the increase in the number of nonagricultural households.
14.3 Community roles in Hayami’s village This section presents a review of literature on the different aspects of community mechanisms in irrigation management, land and labor relations, risks coping, social capital, and the development of the rural nonfarm sector.57 Irrigation management The role of irrigation in MV adoption and subsequent yield increase has been well documented in Asian countries (David and Otsuka, 1994). Irrigation water has the unique characteristic of being a common-pool or commonproperty resource, subject to the danger of overexploitation (the so-called “tragedy of the commons” (Hardin, 1968)). The experience in Hayami’s village shows that a hasty turnover of the irrigation system to the community led to the system’s unsatisfactory management because the local community had no strong organizational capacity to control free-riding problems. Thus, a major role for the government is to facilitate the development of the communities’ own capabilities first to manage their local affairs before turning over the system’s management to them. Hayami’s village had a long tradition of single-cropped rice production until 1958 with the opening of gravity irrigation systems built and maintained by the National Irrigation Administration (NIA). The NIA system is widespread and complex, designed to serve 4000 ha cultivated by about 2500 farmers. The system consists of a main canal, five laterals (A to E), and 16 sublaterals, with Hayami’s village located at lateral C. There is no metering device to measure the water consumption of each farmer so there is a tendency for farmers located upstream to overuse the water and thereby
9780230_274587_15_cha14.indd 229
6/23/2010 5:05:19 PM
230 Jonna P. Estudilo et al.
induce water shortages in the lower streams. Thus, the NIA irrigation system has the typical problem of communal irrigation management. Prof. Hayami noticed a serious deterioration in the village’s irrigation during a visit in July 1995, when the NIA’s irrigation had reached only about 20 percent of the paddy fields. This situation is not unique to his village, and was common in the lower reaches of the major laterals. The deterioration was brought about mainly by a decline in the operation and maintenance (O and M) activities of the NIA. The budget of NIA for O and M activities was reduced to about one-half from the early to the late 1980s. This situation came about because of the decline in rice prices and the decline in the so-called “project fund,” a portion of which is intended to finance NIA’s O and M activities. Irrigation fees were paid in paddy, so the real revenues of the NIA decreased sharply, corresponding to the decline in rice prices during the 1980s. There was also a decline in the contribution of international aid agencies and the national government to “project funds” for the construction of new systems and funds designated for small repairs, on the premise of an abundant supply of rice. With the onset of budgetary pressure, the NIA tried to mobilize farmers’ participation in its O and M activities by organizing farmer-beneficiaries into irrigator associations (IAs) and giving the IAs various cash incentives, depending on the extent of the responsibilities of IAs for canal cleaning, collecting irrigation fees, and other O and M activities. The NIA was able to establish four IAs for the five laterals covering about 70 percent of its total service area. Two IAs were successful in suppressing free-riding problems, by strictly monitoring water rotation schedules, while the other two IAs failed to reduce illegal water take-outs of upstream farmers. The successful IAs had smaller and simpler systems involving fewer farmers and villages, whereas the IAs that failed had larger laterals with more technical designs and social structures.58 Hayami’s village, unfortunately, belonged to a failed IA. To cope with the declining water supply, some farmers decided to exit the system by purchasing water pumps and not paying irrigation fees, which naturally led to the decline in revenue of NIA and fewer O and M activities, thus further lowering water supply. The NIA failed to invest adequately by adding more personnel responsible for institutional development of the IAs that involved larger systems, despite the high social returns associated with such an endeavor. This was because the private benefit to the NIA in terms of increases in irrigation fees associated with the effective management of IAs was much less than the corresponding cost of pooling more NIA personnel into the system. Yet, it is necessary for the NIA to play a facilitating role in the management of local affairs in the face of difficulties in organizing community efforts through the initiatives of the local people alone. Overall, the experience in Hayami’s village shows that community management is far from a panacea, and it underlines the importance of the potentially complementary role of the
9780230_274587_15_cha14.indd 230
6/23/2010 5:05:20 PM
The Transformation of Hayami’s Village
231
government in effectively managing local matters, which sharply contrasts with the widely held preference for leaving the management of such matters entirely in the hands of the community. Landlord-tenant relations Prolific literature on landlord-tenant relations has focused on the choice between shared and fixed-tenancy contracts (Binswanger and Rosenzweig, 1984; Otsuka and Hayami, 1988) by exploring the role of risks and transaction costs in the choice between the two contracts (Coase, 1937; Williamson, 1985). Traditional criticism of share tenancy rests on the Marshallian thesis that share tenancy results in inefficient resource allocation because the share tenant receives only a fraction of the marginal product of his labor, and this lessens her/his work incentives (Sen, 1966). While share tenancy may be an inferior form of land contract because of the high cost of monitoring the tenant’s work effort, it can be rationalized on the assumption that share tenancy has the benefit of risk-sharing (Stiglitz, 1974). Moreover, the costs of the supervision of share tenants can be substantially reduced in a small agrarian community through an interlinked contract involving consumption and production credits, as well as through emergency insurance (Hayami and Otsuka, 1993). The interlinked contract is effective in suppressing dishonest behavior in a small community, where personal ties among family members, relatives, and friends are valuable for efficient contract enforcement (Ben-Porah, 1980; Pollak, 1985). Yet, despite the relative merits of share tenancy, the land-reform programs in developing countries aim to abolish share tenancy on the premise that it is an inefficient and exploitative form of tenancy contract (Binswanger and Elgin, 1998). In Hayami’s village, the abolition of share tenancy led to the rise of subtenancy, which was effectively enforced in the presence of strong community ties between contracting parties. The village had experienced major changes in tenurial relations when the Marcos land-reform program was implemented in 1972. The program had two major components: (1) tenancy reform (Operation Leasehold), and (2) land redistribution (Operation Land Transfer). Tenancy reform converted share tenancy to leasehold tenancy and applied to tenants whose landlords owned less than the maximum ceiling of seven ha. The land redistribution program converted share tenants to amortizing owners (holders of a certificate of land transfer) and applied to tenants whose landlords owned more than seven ha. Leasehold rent and amortization fees were fixed at 25 percent of the average rice yield for three normal crop years preceding 1972. Rice yields rose in the village because of the diffusion of MVs so that a divergence between the returns to land and fixed leasehold rents and amortization fees prescribed by law was created. This divergence conferred an economic value on the possession of cultivation rights by the land-reform beneficiaries.
9780230_274587_15_cha14.indd 231
6/23/2010 5:05:20 PM
232 Jonna P. Estudilo et al.
Before the land reform in 1966, the most common form of land tenure was share tenancy (Table 14.3). Seventy percent of farms and 63 percent of paddy areas were under a share tenancy contract. There was a decrease in the incidence of share tenancy and a corresponding increase in the incidence of leasehold tenancy and ownership because of the progress of Operation Leasehold and Operation Land Transfer. It is surprising that 27 percent of the rice area remained under share tenancy in 1976 and 6 percent remained so in 2006, despite the prohibition of the share tenancy contract. Some share tenants simply refused to go to the local agrarian reform office to change their status because of their good relationships with their landlords, who were either relatives or friends, an attitude that reflected the social norm of utang na loob (debt of gratitude). The presence of economic rent led to the rise in the incidence of subtenancy. Subtenancy was already observed in the village even before the advent of the land-reform program but its incidence increased with the implementation of land reform (Hayami and Kikuchi, 2000, p. 91). Under the subtenancy contract, a leasehold tenant rents out part (or the whole) of her/his operational holdings to other farmers or agricultural laborers. The leasehold tenant charges the market rent, which is over and above the leasehold payment to his landlord that is fixed by the land-reform program. Since the leasehold tenants can pocket the difference between the market rent and the actual rent prescribed by law, we can reasonably assume that the “excess” profit is the main factor behind the rise in subtenancy. In brief, this section has demonstrated how subtenancy had risen as an alternative form of land tenure contract to avoid the rigidities of the landreform laws. While deemed illegal by the land-reform laws, subtenancy is consistent with community norms and tradition, as the terms of agreement are acceptable to stakeholders, who are both members of a close-knit community. The question is whether the subtenancy is consistent with efficiency and equity. Subtenancy is consistent with efficiency because land is transferred from the land-rich sublessor to the land-scarce sublessee household. In the case of pawning, the sublessor is able to ease his or her credit constraint and explore profitable opportunities in the nonfarm sector by using the pawning revenues. Subtenancy is arguably consistent with equity because it opens up opportunities for the noncultivating households to have access to land in the midst of institutional constraints imposed by the land-reform laws. The remarkable fact is that subleasing is practiced, even if the sublessor loses the cultivation right, if the sublessee reports the illegal subleasing arrangement to the agrarian reform office. Labor relations Studies on labor contracts in agriculture have focused on the choice between casual- and permanent-labor contracts (Bell and Srinivasan, 1985; Eswaran and Kotwal, 1985) or the choice between piece-rate and time-rate
9780230_274587_15_cha14.indd 232
6/23/2010 5:05:20 PM
The Transformation of Hayami’s Village
233
compensation schemes within the casual labor contracts (Roumasset and Uy, 1980). A piece-rate contract in harvesting manifests the principle of mutual help and income sharing to secure a subsistence wage for the casual labor force. We discuss the changes in rice harvesting contracts in Hayami’s village to illustrate how the community is able to adjust labor contracts to accommodate changes in labor market conditions brought about by the increase in labor supply and technological change. Hunusan (a Tagalog word for “sharing”) is the traditional labor contract in rice harvesting in the village, whereby anybody in the village can participate in harvesting without restriction. Harvesters commonly receive a one-sixth share of the gross harvest as their compensation. Family members seldom come to monitor the work until the harvest is piled on the ground ready for sharing. Such an attitude is consistent with the social norm of income- and work-sharing because retreating from farm work means giving up income-earning activities to relatively lower-income members of the community. When rice yield was low, the one-sixth share of harvest may well represent the harvesting labor’s contribution to output (labor-marginal productivity). When rice yield rose because of the diffusion of MVs and higher use of fertilizer, the harvester’s income rose in proportion to rice yield. Meanwhile, market wage rates remained relatively stable because of the increase in labor supply. The harvester’s compensation under the traditional hunusan contract exceeded the prevailing market wage rates, making this contract disadvantageous to the farmer-employer. A new system of labor contract called gama (a Tagalog word for “weeding”) emerged and rapidly replaced hunusan in the late 1960s and early 1970s. Under gama, the harvesters receive the traditional one-sixth share of harvest, but employment on harvesting is limited to workers who perform weeding services without receiving compensation. Under gama, weeding is free so that the one-sixth share represents compensation for both harvesting and weeding. The shift from hunusan to gama effectively adjusts labor compensation per hour of work downward. Yet, this is acceptable to the community because it maintains the long-established traditional sharing arrangement, which is consistent with the community norm. Gama declined while a new form of hunusan with reduced sharing rates of up to one-tenth came forth in the 1980s and 1990s. In the new hunusan, only the laborers who receive a specific invitation from the farmer-employer are allowed to participate in harvesting. In this way, the new hunusan resembles the piece-rate contract that is prevalent in market economies. The shift from gama to the new hunusan with a reduced sharing rate was socially acceptable as the availability of job opportunities with higher remuneration rates in the nonfarm sector broke down the traditional patron-client relationship between the farmer-employer and gama workers. Farmer-employers were not satisfied with the weeding performance of gama workers, which
9780230_274587_15_cha14.indd 233
6/23/2010 5:05:20 PM
234 Jonna P. Estudilo et al.
pushed them to contract out weeding services to daily-wage workers. Thus, the reduced sharing rate under the new hunusan was accepted by gama workers because of the reduction in their work assignment under the new contract in the face of increasing work opportunities in the nonfarm sector. We found that the shift from hunusan to gama to new hunusan with reduced sharing rates was consistent with the community principle of mutual help and income-sharing, and with market efficiency which dictates that wages are equal to the value of the workers’ contribution to output. Despite the dynamic changes in the labor market brought about by new rice technology and population pressure, the change in the labor contract did not produce excess “profits” that give undue advantage to one party. The change in the labor contract may well be consistent with equity, as such a change could be considered a “win-win” scenario for all stakeholders. Coping with risks In the face of negative shocks to income and assets, households, especially the poor, had to adopt a wide variety of formal and informal forms of self and mutual insurance mechanisms in order to maintain the same level of consumption (Alderman and Paxson, 1992; Besley, 1995a; Dercon, 2005; Fafchamps, 2003; Morduch, 1995; Walker and Jodha, 1986). The risk-coping strategies may be specific to the characteristics of the household and the nature of the losses caused by income and/or asset shocks. Here we discuss how the community mechanism in Hayami’s village was able to help households to cope with the aftermath of super-typhoon Milenyo (with the international code name of Xangsane) that hit the Philippines (directly hitting Laguna, where Hayami’s village is located) on September 28, 2006. Sawada et al. (2009) conducted a survey barely three months after the occurrence of Milenyo, from 20 January to 15 February 2007, to assess the typhoon damage and identify the risk-coping strategies of households and the various ways by which the community and state responded to the calamity. There was a wide degree of heterogeneity in the damage caused by Milenyo, even within the same village, so we can investigate the effectiveness of formal and informal insurance mechanisms. As to human losses, there were no dead or seriously injured household members reported. A total of 108 households (27 percent) encountered serious damage to their house: typically the roofing was completely or partially destroyed. Including the cases of multiple damage, 54 percent of the farmers and 27 percent of the landless reported having experienced a decline in income, especially because of damage to the standing rice crop. The paddy harvest in the village as a whole declined by about 72 tons (i.e., a 32 percent decline from the normal harvest) and the paddy price declined from the expected price of PHP8.89 per kg to PHP7.44 per kg (i.e., a 16 percent reduction from the normal price). The loss in paddy production was, on average, PHP260 per household, which
9780230_274587_15_cha14.indd 234
6/23/2010 5:05:20 PM
The Transformation of Hayami’s Village
235
is almost equivalent to the minimum wage rate of PHP277 per day (equivalent to US$5.29 at US$1 = PHP52.35). In contrast, the total loss of standing mango trees was PHP680 per household, which is 2.72 times the minimum wage rate per day. Clearly, there was heterogeneity in Milenyo damage, relating to ownership of fruit trees and paddy fields. Households in the village were able to cope with Milenyo by adopting five important strategies, that is, by (1) switching consumption from purchased food and food taken outside to own produce; (2) obtaining emergency loans from relatives and village moneylenders; (3) receiving remittances from overseas; (4) receiving aid from local government and private individuals; and (5) engaging in nonfarm employment. Households change the quality and composition of food expenditure in response to a natural or man-made disaster, as reported in the literature (Frankenberg et al., 2003; Kang and Sawada, 2008; McKenzie, 2006; Strauss et al., 2004). A larger proportion (76 percent) of the landless households compared with farmers (27 percent) and nonagricultural households (47 percent) reported having to decrease their food purchases, while maintaining their total nutritional intake by shifting to consumption of own fish catch (or consuming food given by relatives). Borrowing from close relatives is one of the important coping strategies during times of disaster (Glewwe and Hall, 1998; Sawada and Shimizutani, 2005) because rural households lack consumption insurance (Besley, 1995b; Eswaran and Kotwal, 1989; Glewwe and Hall, 1998). Also, rural households are constrained from borrowing from the formal credit market because of high information costs and lack of assets for collateral or policy-induced financial repression (Fafchamps, 2003). In Hayami’s village the village moneylenders, who were community members, were important sources of emergency funds. For the landless, the nearby sari-sari (village variety) stores provided either on credit or by cash purchase the most basic needs such as rice, canned goods, candles, and kerosene, evidence that village stores do not hoard basic goods during calamity, perhaps because of intricate and close associations among the village people that prohibit such opportunistic behavior. As reported in earlier studies (Otsuka et al., 2009), remittances have become an important source of income in rural Philippines because of the rise in the number of overseas Filipino workers (OFWs). There were 31 OFWs in Hayami’s village as of the 2007 survey. Twenty-five percent of the farmer households, 16 percent of the landless households, and 21 percent of the nonagricultural households reported having received remittances after Milenyo. In brief, the availability of emergency borrowing and remittances indicates the importance of personal networks in surviving such a crisis. Local government, through village officials, played a particularly important and effective role during the disaster by (1) immediately opening the village meeting-hall as a temporary shelter to households that lost their roofs
9780230_274587_15_cha14.indd 235
6/23/2010 5:05:20 PM
236
Jonna P. Estudilo et al.
or were affected by flash floods, (2) distributing grocery bags containing the most basic food items, valued at about US$2 per household, (3) providing galvanized iron sheets to 16 households whose houses lost their roofs, and (4) giving cash gifts amounting to about US$50 to eight households. There was also strict food price and supply monitoring undertaken by the local government in the local public market, in order to avoid unnecessary price hikes and disruption of local food supply maneuvered by opportunistic businessmen. Overall, the disaster management was effective because there was no substantial information asymmetry between the donor and the recipient of aid owing to the village officials’ many years of association with the community. Nonfarm employment serves an important role in consumptionsmoothing in the face of a disaster, as shown by studies in India (Kochar, 1999; Walker and Ryan, 1990) and Indonesia (Frankenberg et al., 2003).59 We cannot identify major changes in primary occupation after Milenyo, which indicates that it was long-term employment in the nonfarm sector that enabled households to insure against the disaster. The number of unemployed, that is, those workers who reported “none” for their primary occupation, rose from 57 to 67 people, indicating an increase in unemployment of 18 percent. While households utilize various forms of precautionary savings such as grain storage (Park, 2006; Townsend, 1995), cash holdings (Townsend, 1995), and liquidation of livestock (Fafchamps et al., 1997; Rosenzweig and Wolpin, 1993) in response to smaller typhoons that occurred from 1994 to 2003, and demographic shocks such as death or illness of household members (Fuwa et al., 2006), cash assistance and remittances from relatives and friends in and outside the village were also evident as major coping strategies during Milenyo.60 This is consistent with the hypothesis that idiosyncratic shocks created by Milenyo were buffered by private transfers. The community can play the role of an effective insurance mechanism against various forms and levels of intensity of idiosyncratic shocks. Against covariant shocks, the role of government was also highlighted as a complementary role, given the lack of effective community mechanisms during a disaster when only the state can effectively undertake a large-scale operation providing relief from correlated shocks. Social capital61 It is widely recognized that social relationships, which are shaped through trust, norms, and networks, play important roles in economic transactions. Such social relationships have been called the “social capital” (Coleman, 1990; Putnam, 1993) which is conceptually akin to physical and human capital. Hayami (2009) proposed the following narrow definition: “Social capital is defined as the structure of informal social relationships conducive to developing cooperation among economic actors aimed at increasing social
9780230_274587_15_cha14.indd 236
6/23/2010 5:05:20 PM
The Transformation of Hayami’s Village
237
product, which is expected to accrue to the group of people embedded in those social relationships.” Words in italics are clarified as follows. First, the term “informal” excludes the relations to be enforced by the state’s apparatus or to be realized as the result of market transactions. Second, the function of social capital is limited in facilitating cooperation or collective action as Coleman (1990) and Putnam (1993) originally intended. Third, since the benefits of social capital accrue mainly to the group of people embedded in the specific social relationship that carries out the collective action, the positive externality is largely limited within a specific group. Therefore, social capital has the attribute of local public-goods, and a community is the central organization to produce it. There are two kinds of social capital classified by function: bonding and bridging social capital (Gittell and Vidal, 1998; Woolcock and Narayan, 2000). Bonding capital strengthens the sense of loyalty, solidarity, and trust within the community, while bridging capital enhances collaboration with organizations outside the community. The community has the advantage of providing and maintaining bonding social capital because there is little asymmetric information among its members due to repeated and intense social interactions. The presence of bonding social capital is evident in the efficient management of common-pool property, such as irrigation water (Fujiie et al., 2005; Ostrom et al., 1990; Kajisa et al., 2007) and forest resources (Tachibana et al., 2001). Social capital is also effective in preventing labor-shirking in share tenancy contracts (Hayami and Otsuka, 1993; de Janvry et al., 1997), and in smoothing consumption patterns in the face of shocks (Fafchamps and Lund, 2003; Sawada et al., 2009). The presence of strong bonding capital, however, may limit the scope of cooperation between communities, resulting in a less desirable resourceallocation outcome. According to Hayami (2009), such situations can be overcome by exposing individuals in the community to new ideas and opportunities, by expanding the scope of intercommunity interactions through bridging. To move in this direction, the community, which is the central organization for producing social capital, must be complemented by a competitive market, which in turn is supported by the state’s judicial apparatus. Putnam’s (1993) comparative study between north and south Italy revealed that the exposure to competitive markets enables the community to seriously recognize the opportunity cost of maintaining the status quo, and this motivates it to increase social capability in bridging between communities. Needless to say, the role of the state is also important because market competition must be facilitated by formal institutions. Therefore, Hayami (2009) emphasizes the importance of seeking the best mix of these three entities. Using a sample set of residents and emigrants from Hayami’s village, Kajisa (2007) finds that bonding social capital has been instrumental for the
9780230_274587_15_cha14.indd 237
6/23/2010 5:05:20 PM
238 Jonna P. Estudilo et al.
efficient functioning of the labor market for small businesses located within and around the community. At the same time, villagers are found to be paid less in the nonfarm small businesses distant from the community, unless they acquired the jobs through family or relative networks; and not many villagers are part of such nascent networks. Hayami’s village has therefore shown the limitation of bonding social capital, which suggests the importance of bridging social capital in the course of economic development. Development of the rural nonfarm sector Household survey data reveal that nonfarm incomes typically account for 30–50 percent of rural household incomes in developing countries (Davis et al., 2007; Lanjouw and Lanjouw, 2001; Otsuka et al., 2009). Within the rural nonfarm sectors, manufacturing sectors appear to account for relatively small portions, and service sectors for larger income shares. According to Haggblade et al. (2009), manufacturing typically accounts for 20–25 percent of rural nonfarm employment, whereas services (such as trade, transport, and construction) account for 75–80 percent. Studies show that the main driver of rural poverty reduction has increasingly shifted from agricultural growth to expansion of nonagricultural income opportunities in many parts of the developing world (Lanjouw and Lanjouw, 2001; de Janvry and Sadoulet; 2001; de Janvry et al., 2005). Correspondingly, the relative importance of access to human capital has increased vis-à-vis that of land in rural areas (Estudillo and Otsuka, 1998; Fuwa, 2007). A series of studies by Hayami and his collaborators focus on the roles played by rural communities and local institutions in the development of the rural nonfarm sector. Hayami (1998b) advocates rural-based industrialization that hinges on the development of rural nonfarm sectors through “relational contracting,” which is a substitute for vertical market integration. He argues that local community and informal institutions can be key elements for the success of rural industrialization. The informational advantages of local entrepreneurs provide them with competitive advantages (visà-vis their urban-based counterparts) in effectively utilizing a low-cost rural labor force while reducing the potential for moral hazard, thereby ensuring required product quality and timely delivery. In the system of vertically integrated “relational contracting,” the rural-based traders and urban manufacturers can function as effective intermediaries between the rural labor force and urban-based traders, retailers, or exporters, who in turn can bring in new market information (e.g., required product quality, shifting consumer preferences, and so on), technology, capital, and brand recognition. Local communities with well-established informal local institutions for risk-sharing, credit transactions, and community enforcement can facilitate the development of rural nonfarm sectors. Examples of “relational contracting” can be found in Japan’s modern industrial sector and the contemporary developing world (Hayami and Kawagoe, 1993), as well as in Japan’s
9780230_274587_15_cha14.indd 238
6/23/2010 5:05:20 PM
The Transformation of Hayami’s Village
239
historical processes (Hayami, 1998a). A prime example of the former is the subcontracting arrangement practiced by the Toyota motor company (Wada, 1997). Subcontracting in Hayami’s village came in with the establishment of seven metalcraft factories during the period 1991–5 (Hayami et al., 1998).62 During this period, there was a diversion of foreign demand for laborintensive products away from Taiwan, Korea, and Hong Kong to Southeast Asian countries, including the Philippines. These rural factories produced Christmas ornaments and other crafted items from tinplate, based on orders from exporters operating in Metro Manila. Subcontracting connects Hayami’s village with the global market through a system of customary trade practices and informal contracts between the village entrepreneurs and the exporters. Subcontracting enhances production efficiency as it is able to exploit the capabilities of rural self-employed informal entrepreneurs who are efficient in utilizing scarce capital and management input while employing cheap labor. Subcontracting is also consistent with equity because the production process in metalcraft is extremely labor-intensive, utilizing the labor of landless workers with low opportunity cost.
14.4 Summary and conclusions There is a common belief that the community is the yoke of underdevelopment because it strictly obeys traditional institutions and norms that are slow to respond to changes in economic opportunities. While reviewing the literature on the multifaceted roles of rural communities in irrigation management, land and labor relations, risks coping, social capital, and development of the rural nonfarm sector, this chapter explores the community mechanisms in Hayami’s village in the Philippines. Irrigation water has the attributes of a common-pool resource, which can be used by farmers in the locality but is exhaustible if overexploited. In Hayami’s village, there was a premature handover of the irrigation system’s management to the community at a time when the community had not yet fully developed a strong organizational capacity to control free-riding problems. As a result the farmers were dissatisfied with the system’s performance, inducing them to adopt the more expensive pump irrigation system. Clearly, the government needs to assume a facilitating role by helping communities to develop their own capability to manage their local affairs, rather than turning over the system’s management to the government. Subtenancy emerged as a common form of land tenure contract that is able to mitigate the institutional rigidity of the land-reform laws. Subtenancy is consistent with efficiency in land allocation, enabling the beneficiary of land reform (the former share tenants) to exploit the “excess profits” that were created when the land-reform law set the land rent over and above the normal market rate. Subtenancy is arguably consistent with equity as well
9780230_274587_15_cha14.indd 239
6/23/2010 5:05:20 PM
240 Jonna P. Estudilo et al.
because it opens up opportunities for the noncultivating households to gain access to land despite institutional constraints imposed by the land-reform laws. While the land-reform laws proscribe subtenancy as illegal, it works smoothly because it is transacted within close circles of friends and relatives belonging to the same community. Hunusan is a labor contract in rice harvesting where every villager can participate in harvesting and receive one-sixth of the gross harvest. Gama restricts employment in harvesting to workers who participate in weeding without receiving compensation. The shift from hunusan to gama was an institutional innovation in labor relations that is consistent with market efficiency in which adjustments in labor market conditions took place so as to bring down wages in response to an increase in labor supply. The shift is also consistent with equity because it does not violate the traditional community norm of mutual help and income-sharing. The community can play the role of an effective insurance mechanism against various forms and levels of intensity of idiosyncratic shock. When super-typhoon Milenyo struck Hayami’s village, the affected community constituents identified state-contingent transfers and emergency loans from relatives, friends, neighbors, and moneylenders as one of their most important coping mechanisms. Against covariant shocks, government money transfers played an important role in enabling people to weather the disaster. Joint and mutually complementary effort undertaken by the local government and private individuals was the key in effectively managing a large-scale relief operation in the aftermath of a disastrous typhoon. Social capital in the form of personal networks is instrumental for the efficient functioning of the labor market in the early stages of its development. Intense and repeated social interaction within a narrow local community decreases the search costs for both workers and employers, and lessens the asymmetric information between the two parties. A case study indicates that small businesses and industries located within or around Hayami’s village take advantage of this function of social capital. It also shows that villagers working distant from the village cannot take such an advantage unless they have personal networks there. Professor Hayami argues that the key to development is the self-employed informal agents who devise efficient ways to utilize scarce capital and managerial inputs, and to mobilize cheap labor. A metalcraft industry was established in Hayami’s village through a subcontracting arrangement between the villagers and exporters operating in Metro Manila. Subcontracting is made possible when both parties are bound by community rules to obey the customary trade practices and informal trade institutions. Overall, the historical changes in Hayami’s village show that rural communities are far from static but respond systematically and dynamically to changes in technology, resource endowments, and market conditions. Although the generality of the findings in one village would need to be
9780230_274587_15_cha14.indd 240
6/23/2010 5:05:20 PM
The Transformation of Hayami’s Village
241
interpreted with caution, we believe that the finding that the community principle of mutual help and income-sharing tend to be consistent with the market principle of profit-seeking is valid in many rural areas in Asia. The state can work effectively if it is in close collaboration with community members, as in the case of the efficient management of a relief operation in the aftermath of Milenyo. The most appropriate economic system for promoting economic development seems to be the right combination of community, market, and state. Our knowledge on this issue is far from adequate, but a number of studies of Hayami’s village provide important clues.
References Alderman, H. and Paxson, C. H. (1992) “Do the Poor Insure? A Synthesis of the Literature on Risk and Consumption in Developing Countries,” in Policy Research Working Paper 1008 (Washington, DC: World Bank). Aoki, M. and Hayami, Y. (2001) Communities and Markets in Economic Development (Oxford: Oxford University Press). Bell, C. and Srinivasan, T. N. (1985) The Demand for Attached Farm Servants in Andra Pradesh, Bihar, and Punjab (Washington, DC: World Bank, mimeo). Ben-Porah, Y. (1980) “The F-Connection: Families, Friends, and Firms and the Organization of Exchange,” Population Development Review, 6 (1), 1–30. Besley, T. (1995a) “Nonmarket Institutions for Credit and Risk Sharing in Low-income Countries,” Journal of Economic Perspectives, 9, 115–27. Besley, T. (1995b) “Savings, Credit and Insurance,” in J. Behrman and T. N. Srinivasan (eds.), Handbook of Development Economics, 3A (North-Holland, Sections 4.5, 4.6, and 4.7). Binswanger, H. P. and Elgin, M. (1998) “Reflections on Land Reform and Farm Size,” in C. K. Eicher and J. M. Staatz (eds.), International Agricultural Development, 3rd edition (Baltimore: Johns Hopkins University Press), 316–28. Binswanger, H. P. and Rosenzweig, M. (1984) “Contractual Arrangements, Employments, and Wages in Rural Labor Markets: A Critical Review,” in H. P. Binswanger and M. Rosenzweig (eds.), Contractual Arrangements, Employments, and Wages in Rural Labor Markets in Asia (New Haven, CT: Yale University Press). Coase, R. H. (1937) “The Nature of the Firm,” Economica, 4 (16), 386–405. Coleman, J. S. (1990) Foundations of Social Theory (Cambridge, MA: Harvard University Press). Cox, D., Zekeriya, E., and Jimenez, E. (1998) “Motives for Private Transfers Over the Life Cycle: An Analytical Framework and Evidence for Peru,” Journal of Development Economics, 55 (1), 57–80. Davis, B., Winters, P., Carletto, C., Covarrubias, K., Quinones, E., Zezza, A., Stamoulis, K., Bonomi, G. and DeGiuseppe, S. (2007) “Rural Income-generating Activities: A Cross-country Comparison,” Background paper for the World Development Report 2008 (Washington, DC: World Bank). David, C. C. and Otsuka, K. (1994) Modern Rice Technology and Income Distribution in Asia (Boulder, CO: Lynne Riener). de Janvry, A. and Sadoulet, E. (2001) “Income Strategies Among Rural Households in Mexico: The Role of Off-farm Activities,” World Development, 29, 467–80. de Janvry, A., Sadoulet, E. and Fukui, S. (1997) “The Meaning of Kinship in ShareCropping Contracts,” American Journal of Agricultural Economics, 79, 394–406.
9780230_274587_15_cha14.indd 241
6/23/2010 5:05:20 PM
242 Jonna P. Estudilo et al. de Janvry, A., Sadoulet, E. and Zhu, N. (2005) “The Role of Non-farm Incomes in Reducing Rural Poverty and Inequality in China,” mimeo. Dercon, S. (2005) Insurance Against Poverty (Oxford: Oxford University Press). Dercon, S. and Krishnan, P. (2000) “In Sickness and in Health: Risk-Sharing within Households in Rural Ethiopia,” Journal of Political Economy, 108 (4), 688–727. Dubois, P., Jullien, B. and Magnac, T. (2008) “Formal and Informal Risk-sharing in LDCs: Theory and Empirical Evidence,” Econometrica, 76 (4), 679–725. Estudillo, J. P. and Otsuka, K. (1998) “Green Revolution, Human Capital, Off-Farm Employment: Changing Sources of Income among Farm Households in Central Luzon, 1966–94,” Economic Development and Cultural Change, 47 (3), 497–523. Estudillo, J. P., Sonobe, T. and Otsuka, K. (2007) “Development of the Rural Non-Farm Sector in the Philippines and Lessons from the East Asian Experience,” in A. M. Balisacan and H. Hill (eds.), The Dynamics of Regional Development: The Philippines in East Asia (Cheltenham, UK: Edward Elgar). Eswaran, M. and Kotwal, A. (1985) “A Theory of Two-Tier Labor Markets in Agrarian Economies,” American Economic Review, 75 (1), 162–77. Fafchamps, M. (2003) Rural Poverty, Risk and Development (Edward Elgar). Fafchamps, M. and Lund, S (2003) “Risk-Sharing Networks in Rural Philippines,” Journal of Development Economics, 71, 261–87. Fafchamps, M., Udry, C. and Czukas, K. (1997) “Drought and Saving in West Africa: Are Livestock a Buffer Stock?” Journal of Development Economics, 55, 273–305. Frankenberg, E., Smith, J. P. and Thomas, D. (2003) “Economic Shocks, Wealth, and Welfare,” Journal of Human Resources, 38 (2), 280–321. Fujiie, M., Hayami, Y. and Kikuchi, M. (2005) “The Conditions of Collective Action for Local Commons Management: The Case of Irrigation in the Philippines,” Agricultural Economics, 33 (2), 179–89. Fuwa, N. (2007) “Pathways Out of Rural Poverty: A Case Study in Socioeconomic Mobility in the Rural Philippines,” Cambridge Journal of Economics, 31, 123–44. Fuwa, N., Marciano, E. and Reaño, J. (2006) Report on the 2003 Livelihood Systems of Rural Households Survey in the Philippines (Los Banos, Laguna, Philippines: International Rice Research Institute). Gittell, R. and Vidal, A. (1998) Community Organizing: Building Social Capital as a Development Strategy (Newbury, CA: Sage). Glewwe, P. and Hall, G (1998) “Are Some Groups More Vulnerable to Macroeconomic Shocks Than Others? Hypothesis Tests Based on Panel Data from Peru,” Journal of Development Economics, 56, 181–206. Haggblade, S., Hazell, P. B. R. and Reardon, T. (2009) “Transforming the Rural NonFarm Economy: Opportunities and Threats in the Developing World,” IFPRI Issue Brief 58, International Food Policy Research Institute, Washington, DC. Hardin, G. (1968) “The Tragedy of the Commons,” Science, 162, 1243–8. Hayami, Y. (1998a) “Community, Market, and State,” in C. K. Eicher and J. M. Staatz (eds.), International Agricultural Development, 3 (Baltimore: Johns Hopkins University Press), 90–102. Hayami, Y. (ed.) (1998b) Toward the Rural-Based Development of Commerce and Industry: Selected Experiences from East Asia (Washington, DC: World Bank). Hayami, Y. (2009) “Social Capital, Human Capital, and Community Mechanism: Toward a Consensus among Economists,” Journal of Development Studies, 45, 96–123. Hayami, Y. and Godo, Y. (2005) Development Economics: From the Poverty to the Wealth of Nations (Oxford: Oxford University Press).
9780230_274587_15_cha14.indd 242
6/23/2010 5:05:21 PM
The Transformation of Hayami’s Village
243
Hayami, Y. and Kawagoe, T. (1993) The Agrarian Origins of Commerce and Industry: A Study of Peasant Marketing in Indonesia (London: Macmillan). Hayami, Y. and Kikuchi, M. (2000) A Rice Village Saga: Three Decade of Green Revolution in the Philippines (London: Macmillan). Hayami, Y., Kikuchi, M. and Marciano, E. (1998) “Structure of Rural-Based Industrialization: Metal Craft Manufacturing on the Outskirts of Greater Manila, the Philippines,” Developing Economies, 36 (2), 133–54. Hayami, Y. and Otsuka, K. (1993) The Economics of Contract Choice: An Agrarian Perspective (Oxford: Claredon Press). Jacoby, H. and Skoufias, E. (1997) “Risk, Financial Markets and Human Capital in Developing Countries,” Review of Economic Studies, 64, 311–35. Kajisa, K. (2007) “Personal Networks and Non-agricultural Employments: The Case of a Farming Village in the Philippines,” Economic Development and Cultural Change, 55 (4), 669–707. Kajisa, K., Palanisami, K. and Sakurai, T. (2007) “Effects on Poverty and Equity of the Decline in the Collective Tank Irrigation Management in Tamil Nadu, India,” Agricultural Economics, 36 (3), 347–62. Kang, S. J. and Sawada, Y. (2008) “A Credit Crunch and Household Welfare in Korea,” Japanese Economic Review, 59(4), 438–58. Kochar, A. (1999) “Smoothing Consumption by Smoothing Income: Hours of Work Responses to Idiosyncratic Agricultural Shocks in Rural India,” Review of Economics and Statistics, 81 (1), 50–61. Lanjouw, J. and Lanjouw, P. (2001) “The Rural Non-farm Sector: Issues and Evidence from Developing Countries,” Agricultural Economics, 26, 1–23. Ligon, E. (2008) “Risk Sharing,” in New Palgrave Dictionary of Economics, 2nd edition (London: Palgrave Macmillan). Lucas, R. E. B. and Stark, O. (1985) “Motivations to Remit: Evidence from Botswana,” Journal of Political Economy, 93, 901–18. McKenzie, D. J. (2006) “The Consumer Response to the Mexican Peso Crisis,” Economic Development and Cultural Change, 55 (1), 139–72. Morduch, J. (1995) “Income Smoothing and Consumption Smoothing,” Journal of Economic Perspectives, 9, 103–14. Ostrom, E. (1990) Governing the Commons (New York: Cambridge University Press). Otsuka, K., Estudillo, J. P. and Sawada, Y. (2009) Rural Poverty and Income Dynamics in Asia and Africa (Oxford: Routledge). Otsuka, K. and Hayami, Y. (1988) “Theories of Share Tenancy: A Critical Survey,” Economic Development and Cultural Change, 37 (1), 31–68. Park, A. (2006) “Risk and Household Grain Management in Developing Countries,” Economic Journal, 116, 1088–115. Pollak, R. (1985) “A Transactions Cost Approach to Families and Households,” Journal of Economic Literature, 23 (2), 581–608. Putnam, R. D. (1993) Bowling Alone (New York: Simon and Schuster). Rose, E. (2000) “Gender Bias, Credit Constraints, and Time Allocation in Rural India,” Economic Journal, 110 (465), 738–58. Rosenzweig, M. R. and Wolpin, K. (1993) “Specific Experience, Household Structure, and Intergenerational Transfers: Farm Family Land and Labor Arrangements in Developing Countries,” Quarterly Journal of Economics, Supplement 100, 961–88. Roumasset, J. and Uy, M. (1980) “Piece Rates, Time Rates, and Teams: Explaining Patterns in Employment Relation,” Journal of Economic Behavior and Organization, 1 (1), 343–60.
9780230_274587_15_cha14.indd 243
6/23/2010 5:05:21 PM
244 Jonna P. Estudilo et al. Sawada, Y. (2007) “The Impact of Natural and Man-Made Disasters on Household Welfare,” Agricultural Economics, 37 (s1), 59–73. Sawada, Y., Estudillo, J. P., Fuwa, N. and Kajisa, K. (2009) “How Do People Cope With a Natural Disaster? The Case of Super-Typhoon Milenyo in the Philippines,” in G. P. Carnaje and L. S. Cabanilla (eds.), Development, Natural Resources, and the Environment (Department of Economics, University of the Philippines at Los Banos, Laguna, Philippines), 99–129. Sawada, Y. and Lokshin, M. (2009) “Obstacles to School Progression in Rural Pakistan: An Analysis of Gender and Sibling Rivalry Using Field Survey Data,” Journal of Development Economics, 88, 335–47. Sawada, Y. and Shimizutani, S. (2005) “Are People Insured Against Natural Disasters? Evidence from the Great Hanshin-Awaji (Kobe) Earthquake in 1995,” CIRJE Discussion Paper Series F-314 (Tokyo: Faculty of Economics, University of Tokyo). Sen, A. K. (1966) “Peasants and Dualism With or Without Surplus Labor,” Journal of Political Economy, 74 (5), 425–50. Smith, J. P., Thomas, D., Frankenberg, E., Beegle, K. and Teruel, G. (2004) “Wages, Employment, and Economic Shocks: Evidence from Indonesia,” Journal of Population Economics, 15 (1), 161–93. Stiglitz, J. E. (1974) “Incentives and Risk-Sharing in Sharecropping,” Review of Economic Studies, 41 (2), 219–56. Strauss, J., Beegle, K., Dwiyanto, A., Herawati, Y., Pattinasarany, D., Satriawan, E., Sikoki, B., Sukamdi and Witoelar, F. (2004) Indonesian Living Standards Before and After the Financial Crisis: Evidence from the Indonesia Family Life Survey (Rand Corporation and ISEAS). Tachibana, T., Nguyen, T. M. and Otsuka, K. (2001) “Agricultural Intensification vs Extensification: A Case Study of Deforestation in the Northern Hill Region of Vietnam,” Journal of Environmental Economics and Management, 41 (1), 44–69. Townsend, R. M. (1994) “Risk and Insurance in Village India,” Econometrica, 62, 539–91. Townsend, R. M. (1995) “Consumption Insurance: An Evaluation of Risk-bearing Systems in Low-income Economies,” Journal of Economic Perspectives, 9, 83–102. Udry, C. (1994) “Risk and Insurance in a Rural Credit Market: An Empirical Investigation in Northern Nigeria,” Review of Economic Studies, 61, 495–526. Wada, K. (1997) “The Formation of Toyota’s Relationship with Suppliers: A Modern Application of the Community Mechanism,” in Y. Hayami (ed.), Toward the RuralBased Development of Commerce and Industry: Selected Experiences from East Asia (Washington, DC: World Bank). Walker, T. S. and Jodha, N. S. (1986) “How Small Farm Households Adapt to Risk,” in P. Hazell, C. Pomareda and A. Valdes (eds.), Crop Insurance for Agricultural Development: Issues and Experience (Johns Hopkins University Press for the International Food Policy Research Institute). Walker, T. S. and Ryan, J. G. (1990) Village and Household Economics in India’s Semi-Arid Tropics (Baltimore: Johns Hopkins University Press). Williamson, O. (1985) The Economic Institutions of Capitalism (New York: Free Press). Woolcock M. and Narayan, D. (2000) “Social Capital: Implications for Development Theory, Research, and Policy,” The World Bank Research Observer, 15 (2), 225–49.
9780230_274587_15_cha14.indd 244
6/23/2010 5:05:21 PM
15 Recent Developments of Agricultural Markets in East Africa* Takashi Yamano, Yoko Kijima, Tomoya Matsumoto, and Megumi Muto
15.1
Introduction
The importance of the working of community mechanisms in rural markets is recognized by many development economists. Professor Yujiro Hayami’s insightful field observations, based on the economic theories of games and transaction costs, have helped us understand the roles of community mechanisms in rural markets (Aoki and Hayami, 2001a; Hayami and Kikuchi, 2000; Hayami and Otsuka, 1993). He argues that the major barrier against the development of markets in developing economies is the absence of an effective mechanism to protect property rights and to enforce contracts (Aoki and Hayami, 2001b). This absence raises transaction costs, which may keep farm households out of markets by holding the farm-gate prices of crops below their reservation prices. Hayami (2009) points out, however, that the community mechanisms of contract enforcement based on mutual trust or social capital may significantly reduce transaction costs. Although his observations are mostly from Asian countries, his argument is also applicable to African rural markets. Agricultural markets in sub-Saharan Africa have experienced drastic changes in market institutions after the structural adjustment period from the late 1980s to the early 1990s, transforming agricultural markets from government-supported to liberalized markets. Although the market liberalization was expected to reduce the high transaction costs of governmentsupported agencies, such as cooperatives or marketing boards, the sudden withdrawal of governmental controls and support from market transactions
* Financial support for the data collection used in this chapter is provided by the 21st Century Center of Excellence project at National Graduate Institute for Policy Studies. The authors appreciate valuable comments from Keijiro Otsuka, Takashi Kurosaki, and participants at the CSAE Economic Development in Africa Conference in 2008.
245
9780230_274587_16_cha15.indd 245
6/23/2010 7:15:50 PM
246
Takashi Yamano et al.
has left some commodity markets without adequate institutional support, leaving evaluations of the market liberalization controversial (Badiane and Shively, 1998; Jayne et al., 2002; Negassa et al., 2004). Some evaluations of the liberalized markets, however, were premature because many markets were still in the process of transition. Separately, regarding market information technology, a rapid diffusion of mobile phone networks in the region is expected to reduce transaction costs in agricultural markets in developing countries, where the landline telecommunication infrastructure is usually poor. Recent studies suggest that traders and producers use mobile phones to reduce transaction costs, by reducing search costs and product spoilages, and to achieve smaller price variances (Aker, 2008; Goyal, 2008; Jensen, 2007; Overa, 2006). After about two decades since market liberalization and with the recent introduction of improved information technology in the agricultural markets, evidence on how these changes affect agricultural markets in sub-Saharan Africa is still thin. The purpose of this chapter, therefore, is to describe two examples of agricultural market developments in Kenya and Uganda by using panel data of 862 rural households in Kenya and 856 households in Uganda. In Kenya, we examine the impact of the liberalization of the milk market in 1992, where liberalization removed the ban on selling raw milk in urban areas, including Nairobi. The removal of the ban has encouraged many private traders to enter the milk market in Kenya. From Uganda we use panel data, collected in 2003 and 2005, that cover a period when the mobile phone network expanded rapidly. Between the two surveys, we find that the proportion of sample communities that were covered by the mobile phone network increased from 42 to 92 percent. This gives us an opportunity to estimate the impact of the mobile phone network expansion on banana marketing among our samples because we would expect that the improved information technology would help reduce transaction costs among traders of highly perishable crops, such as bananas. In the next section, we analyze the impact of market liberalization on the milk farm-gate price in Kenya, and the impact of information technology improvements on the farm-gate price of bananas in Uganda. Section 15.3 explains the panel data used in this chapter. The results of the dairy market liberalization in Kenya are discussed in Section 15.4, which is based on Kijima et al. (2009). Section 15.5 is based on Muto and Yamano (2008) and presents the results of the mobile phone network expansion on banana marketing in Uganda. We conclude this chapter in Section 15.6.
15.2
Conceptual framework
In this section, we present simple models on the farm-gate prices of agricultural commodities and apply the models to milk marketing in Kenya and banana marketing in Uganda. Suppose that the farm-gate price of a
9780230_274587_16_cha15.indd 246
6/23/2010 7:15:51 PM
Recent Developments of Agricultural Markets
247
commodity of farmer i at time t is defined as PitF Pt M g τ i , where PtjF is the M farm-gate price of the commodity, Pt is the price of the commodity at the nearest market from the household i at time t, g is the marketing cost per unit of distance, and τi is the distance between the market and farmer i. We assume that g represents the sensitivity of the output price of the commodity with regard to the distance to the market. The milk market in Kenya In the case of the milk market in Kenya, the liberalization of the market resulted in a shift from a dairy-cooperative-dominant market to a privatetrader-dominant market, as we explain later in Section 15.4. Dairy cooperatives pay a fixed and largely uniform milk price to their members, to be fair to their members. As a result, the farm-gate price of milk paid by the dairy cooperative is less sensitive to the distance from the retail market than the farm-gate price paid by private traders. In panel A of Figure 15.1, we see two lines showing the farm-gate price paid by dairy cooperatives (flat) and private traders (steep). As the figure shows, dairy farmers who live close to the retail market would gain by selling milk to private traders. Note, however, that before liberalization, selling raw milk in urban areas was prohibited by law. Thus, the dairy farmers who lived near the urban areas had to accept the price paid by dairy cooperatives. Moreover, as new private traders gain experience in milk trading, and as untrustworthy traders are weeded out through community mechanisms, we would expect the farm-gate price paid by private traders to increase. To investigate the change in the farm-gate price for dairy farmers, we compare retail and farm-gate prices at a given time and over time. We focus Panel A Dairy market in Kenya pM
p2
Panel B Banana market in Uganda
Farm-gate price with mobile phones
Farm-gate price paid by traders
p1
Farm-gate price paid by cooperatives
pF
Farm-gate price without mobile phones
Distance Figure 15.1
pF
Distance
Farm-gate prices of milk and bananas
9780230_274587_16_cha15.indd 247
6/23/2010 7:15:51 PM
248
Takashi Yamano et al.
on the price range because this can be seen as a proxy for marketing costs, which include transportation costs as well as transaction costs. Then, we hypothesize that the price range shrinks, especially in areas close to retail markets, as dairy farmers sell more to private traders than dairy cooperatives, and as private traders become more efficient in milk marketing over time after liberalization. Thus, our hypothesis is: H 0 : ( p2M p2F ) ( p1M p1F ) < 0,
(1)
where t = 1 and 2 indicate the first and second round of the panel data, respectively. In the empirical analyses, we test the hypothesis by estimating the community fixed-effects model with pooled cross-section data:
(
)
pit b0 b D DitR d DitR T bT T b g g (di ) v j eit ,
(2)
where pit is the milk price paid to dairy farmer i, either at the retail shops R or at the farm gate; Dit is a dummy variable for the retail price; T is a yeardummy for the second round; g(di) is a function of the traveling time from household i to the nearest urban town; and vj represents the community fixed-effects. The coefficient of the interaction term, d, captures the change in the marketing cost as explained in equation (1). Mobile phones and banana marketing in Uganda To investigate the impact of the mobile phone network on banana marketing in Uganda, we modify the above model by introducing information technology, It, and assume that the marketing cost of bananas per distance unit, g, is a function of the information technology, where the marketing g I cost decreases as the information technology improves: ∂ ∂ I( ) < 0. Thus, the mobile phone network coverage decreases the marketing cost of bananas by improving the information technology, leading to a lower level of g j (I t ). We think that the marginal change in g j (I t ) due to a one unit increase in the information technology, It, is expected to be large for perishable crops, such as bananas. The potential gain for farmer i when the information technology improves, while keeping the market price and distance constant, can be derived as: j
t
t
∆ pitF piF2 piF1 , p M g ( I 2 ) τ i p M g ( I 1 ) τ i , {g ( I 2 ) g ( I 1 ) }τ i ,
9780230_274587_16_cha15.indd 248
(3)
6/23/2010 7:15:52 PM
Recent Developments of Agricultural Markets
249
where I2 > I1 and {g(I 2) g(I1)}<0. Thus, the banana farmers should gain by the introduction of the mobile phone network, and the gain should be larger for farmers who are located farther away from the market. Note that the impact of the introduction of the mobile phone network is larger for more perishable crops than for less perishable crops because the marketing costs of more perishable crops would be larger than that for less perishable crops.63 Marketing of perishable crops can be quick if the marketing activities are organized. To organize the marketing activities between traders and producers, every single producer does not need to own mobile phones to benefit from such organized collections. One representative of the farmers can make arrangements, such as agreeing on a collection time and place with a buyer, via mobile phones. Thus, the new information technology can motivate farmers to organize themselves to make the marketing of perishable crops more efficient.
15.3 Data We use panel data from the Research on Poverty and Environment and Agricultural Technology (RePEAT) Project in Uganda and Kenya. The Kenya RePEAT data are based on three comparable cross-sectional data sets of 3300 randomly selected householders, interviewed in the late 1990s by the Smallholder Dairy Project (SDP), a collaborative team from the Ministry of Livestock Development and Fisheries, the Kenya Agricultural Research Institute, and the International Livestock Research Institute. For ease of presentation, we denote the SDP data as “the 1998 data” hereafter. In 2004, the RePEAT project randomly selected 100 sublocations (the smallest administrative unit in Kenya) from the sublocations where the SDP households resided.64 In each sublocation, ten SDP households were randomly selected for reinterviews. In this chapter, we use panel data from 862 households.65 The Uganda RePEAT data largely build upon and complement a completed research project on policies for improved land management in Uganda, conducted by the International Food Policy Research Institute (IFPRI) and Makerere University between 1999 and 2001 (Pender et al., 2004). The latter involved a survey of 107 communities selected from two-thirds of the regions in Uganda. Because of instabity in the north and northeastern parts of the country, these areas were excluded from the samples. The RePEAT project was conducted in 29 out of 32 districts and in 94 out of 107 Local Council 1s (LC1s) studied by IFPRI. From each LC1, ten households were selected from the 94 LC1s, and 940 households were interviewed in 2003.66 In 2005, a follow-up survey of the 940 households was conducted, and 856 households were interviewed successfully. In both surveys, community surveys were also conducted along with the household surveys. As we mentioned earlier, the proportion of sampled households covered by the mobile phone network increased from 42 to 92 percent between the surveys. As for
9780230_274587_16_cha15.indd 249
6/23/2010 7:15:54 PM
250 Takashi Yamano et al.
household mobile phone ownership, only 4.3 percent of the households possessed at least one unit in 2003, while in 2005 the percentage had increased to 11.5 percent.
15.4 Dairy market development in Kenya Dairy liberalization in Kenya After independence, to guarantee a stable milk price to all dairy farmers, the Kenyan government mandated that KCC accept all milk delivered to its plants and introduced a uniform pricing system (pan-seasonal and panterritorial) in which KCC buys milk at the same price no matter where the dairy farmers are located and no matter when they sell their milk, mainly for political reasons.67 In Figure 15.2, we describe the simplified milk-
Before liberalization
Rural consumers
Dairy farmers
Private traders
Dairy cooperatives
KCC
Urban consumers
After liberalization
Dairy farmers
Rural consumers Private traders
Cooperatives
Raw milk Figure 15.2
KCC/ Private processors
Urban consumers
Processed milk
Simplified milk-market structure in Kenya
9780230_274587_16_cha15.indd 250
6/23/2010 7:15:54 PM
Recent Developments of Agricultural Markets
251
marketing channels before and after liberalization in Kenya. Before liberalization (panel A), selling raw milk in urban areas was prohibited. Thus, private traders used to sell raw milk only to consumers in rural areas. Dairy farmers sold milk to dairy cooperatives, who sold it to KCC. By the late 1980s, KCC had accumulated deep debts because of poor management and costly operations, and the indebted Kenyan government could not support KCC financially. As a result, in 1992, the formal milk-market was liberalized. The liberalization removed the ban on selling raw milk in urban areas, including Nairobi, and allowed private milk-processing firms to buy milk from dairy farmers.68 Although the official intent of the 1992 dairy sector liberalization was to allow private processors to sell pasteurized milk in urban areas, the interpretation by market actors was that all actors, including raw milk traders, could operate in urban areas (Owango et al., 1998). When the liberalization effectively removed the ban on raw milk sales in urban areas, dairy farmers living close to urban areas started selling raw milk in urban areas (panel B, Figure 15.2). The transformation of the dairy milk-market, however, created some problems in the beginning. While the dairy cooperatives purchased raw milk from members throughout the year, traders tended not to buy milk from the dairy farmers during the rainy season due to the excess supply of milk. In addition, some traders disappeared without making payments to the dairy households, and some dairy farmers and traders adulterated the milk to increase the quantity. Because of these problems with private traders, the closures of the dairy cooperatives created a need for a new efficient marketing channel which could prevent such difficulties (Holloway et al., 2000; Staal et al., 1997). According to interviews conducted from January to March 2005 by some of the authors of this chapter, the frequency of problems related to milk marketing has declined as trust-based and long-term relationships have developed between the traders and the dairy farmers, and between the traders and the retailers. This is because transactions based on local connections and family ties have the effect of restraining business counterparts from breaking community rules, as argued by Hayami and Kawagoe (1993). In the early postliberalization period, there were few long-distance traders transporting raw milk to large cities. This is partly because trading raw milk in large quantities over a long distance requires capital for the purchase of pickup trucks. More importantly, this is because it was difficult for the traders to sell raw milk in large cities when there were no institutions and, hence, no practical ways to assure the quality of the raw milk. As traders established milk-marketing chains of kiosks, generally called “milk bars,” and brand names to gain a good reputation regarding the quality and safety of their product, the number of long-distance raw milk traders increased gradually to fill the gap between milk demand and supply (ArgwingsKodhek and Karin, 1999).
9780230_274587_16_cha15.indd 251
6/23/2010 7:15:55 PM
252
Takashi Yamano et al.
Descriptive statistics To examine the changes in milk marketing among dairy farmers during the transition period of the milk market after liberalization, we present the percentages of households who produced and sold milk in 1998 and 2004, sorted by the traveling time to urban towns, in Table 15.1. In this table, we find that there is a large increase in the proportion of households selling milk, from 58 percent to 80 percent, among milk producers between 1998 and 2004. Columns (D) to (F) show the percentages of milk producers who sold milk to the three different milk-buyer types,69 that is, local customers, cooperatives/KCC, and private traders/processors. We find that in 1998, half of the milk sellers sold milk to local customers and the share of KCC/cooperatives was about 30 percent, regardless of the distance to urban towns. In 2004, the share of KCC/cooperatives declined to 12 percent. It is interesting that the cooperatives could maintain relatively high shares in areas closer to towns, but reduced their shares significantly in areas farther away from towns. Private traders and processors, instead, gained shares in all areas. Note also that in areas far from towns (60 minutes and more), the share of local customers increased from 69 percent to 86 percent from 1998 to 2004.
Table 15.1
Traveling time to towna
Milk marketing of dairy farmers in Kenya Percentage of milk Number Percentage producers of houseof milk who sold holds producers milk
% of sellers sold milk to: Traders, Local Cooperative/ private customersb KCC processor
(A)
(B)
(C)
(D)
(E)
(F)
1998 0–15 mins 15–30 mins 30–60 mins 60 mins Total
127 280 361 94 862
% 55 61 62 62 60
% in (B) 69 67 51 45 58
% in (C) 58 41 55 69 51
% in (C) 28 38 20 31 29
% in (C) 14 21 25 0 20
2004 0–15 mins 15–30 mins 30–60 mins 60 mins Total
127 280 361 94 862
54 65 69 59 64
80 86 79 64 80
46 36 49 86 47
12 20 6 5 12
42 44 45 9 41
Notes: a “Traveling time to town” is the average traveling time to the nearest urban area from each household. This variable is measured by GPS information which identifies three types of road: tarmac, other all-weather road, and dry-weather road. To calculate traveling time, tarmac road can be driven at 80 km per hour, other all-weather road can be driven at 50 km per hour, and dryweather road can be driven at 30 km per hour. b “Local Customers” includes restaurants, hotels, kiosks, and individual customers.
9780230_274587_16_cha15.indd 252
6/23/2010 7:15:55 PM
Recent Developments of Agricultural Markets
253
Thus, in these remote areas, private traders and processors failed to replace dairy cooperatives as milk buyers. Retail and farm-gate price difference It is not clear from Table 15.1 if the shift from dairy cooperatives to private traders has benefited dairy farmers. To investigate this issue, we focus on the price difference between the retail and farm-gate prices shown in Table 15.2. If the price difference has declined over time, we take this as a sign of reduced transaction costs. In Table 15.2, we stratify the samples according to the distance between dairy households and urban areas because transportation costs account for a large share of the price difference. In all areas, the retail price is higher than the farm-gate price. Both the retail and farm-gate prices tend to decrease as the traveling time to urban towns increases up to 60 minutes. The difference between the retail and farm-gate prices is presented in column C. The change in the transaction costs over time is presented in column D. The results indicate that the price difference has declined in the areas closest to urban towns and the areas farthest from urban towns, while it stayed at the same level in the other areas. The result in the areas closest to urban towns is consistent with panel A of Figure 15.1. In the case of the areas farthest from urban towns, many dairy farmers exited from the market. Thus, the result could be driven by the exit of dairy farmers who received low milk prices. These results suggest that transaction costs have been reduced as private traders have replaced cooperatives as the largest milk-buyer group. Table 15.2
Average milk price by sales outlet in Kenyaa DID Estimator (C) in 2004– 1998
Retail price
Farm-gate price
Price differentialc
(A) Kshs/liter
(B) Kshs/liter
(C) Kshs/liter
(D)
1998 0–15 mins 15–30 mins 30–60 mins 60 mins
29.7 25.2 23.3 26.3
24.7 22.3 20.0 18.0
5.0 2.9 3.3 8.3
–2.2 0.1 0.0 –4.8
2004 0–15 mins 15–30 mins 30–60 mins 60 mins
21.5 18.3 17.2 18.9
18.7 15.3 13.9 15.4
2.8 3.0 3.3 3.5
Traveling time to townb
Notes: a Milk prices are adjusted for 2004 October Kenya shillings (Kshs) per liter of milk. b Deflators are from Statistical Bulletin, Central Bank of Kenya (June 2002, 2004, and 2005). c Price differential between retail and farm-gate price (p M – pF).
9780230_274587_16_cha15.indd 253
6/23/2010 7:15:55 PM
254 Takashi Yamano et al.
Table 15.3 Determinants of milk price at the household level (Sublocation fixedeffects model)
Retail price dummy
1998 (A)
2004 (B)
Pooled model (C)
0.940* (1.88)
0.825** (2.39)
1.516** (3.46) –0.908*
–10.59 (–0.88) 5.655 (0.65)
–20.91** (–2.85) 14.94* (2.55)
Retail price dummy x year 2004 dummy Travel time to town (minutes) x 0.01 Travel time squared x 0.01 Travel time x 2004 dummy x 0.01 Travel time squared x 2004 dummy x 0.01 Year 2004 dummy Constant Sublocation fixed effects Overall r-squared Number of sublocations Number of observations
27.17** (7.36) Included 0.11 75 280
22.39** (10.69) Included 0.14 92 479
(1.78) –10.08 (1.44) 7.647 (1.44) –7.055* (1.89) 4.809* (1.79) –4.985** (4.17) 26.18** (12.66) Included 0.381 93 759
Notes: Observations include only those householders who sold milk in the survey years. * and ** indicate 10 and 5 percent level of significance, respectively.
Since Table 15.2 only shows the average marketing cost by the broad category of traveling time to urban towns, we conduct regression analyses next. The results from the regression analyses in Table 15.3 indicate that the price difference has shrunk over time. In column C, we pool the two surveys and include the interaction terms. The coefficient of the retail price dummy measures the price difference between the retail and farm-gate prices in the first round, and the coefficient of the interaction term of the retail price dummy and the time dummy identifies the change in the price difference. The coefficient of the price is negative and significant at the 10 percent level, indicating that the marketing cost has been reduced between 1998 and 2004.
15.5 Banana marketing and mobile phones in Uganda Banana marketing in Uganda Before the introduction of mobile phones, town-based traders used to visit banana-producing areas without any prior announcement due to the lack
9780230_274587_16_cha15.indd 254
6/23/2010 7:15:56 PM
Recent Developments of Agricultural Markets
255
Panel A: Before the mobile phone coverage
Farmer Farmer
by bicycle
by truck Urban areas
Traders
Panel B: After the mobile phone coverage Farmer
by bicycle by truck
Farmer Figure 15.3
Traders
Urban areas
Banana market before and after mobile phone coverage in Uganda
Note: In panel A, town-based traders make unannounced visits to banana-producing communities and collect bananas by sending out village-based buyers on bicycles. This may take a few days before filling up a truck. In panel B, town-based traders negotiate banana-collection arrangements with a farmer representative via mobile phones. The representative asks fellow farmers to bring bananas to a collection point at an agreed time. It may take only a few hours to fill up a truck.
of any communication method, and might spend several days collecting banana bunches by sending out village-based buyers on bicycles until they collected enough bananas to fill up their trucks (panel A, Figure 15.3). For banana farmers, this was the only available option to sell their bananas to outsiders. Because it is costly to transport banana bunches, banana farmers often wait for traders to come by and, as a result, they do not have much bargaining power over the price the trader offers, particularly if they do not have information on the prevailing price in the market. Moreover, in Uganda it was uncommon to find farmer associations where farmers could potentially share market information (Smale and Tushemereirwe, 2007). The introduction of mobile phones has changed this situation in some areas (panel B, Figure 15.3).70 When areas are covered by the mobile phone network, individual banana farmers or their representatives can contact banana traders in advance, negotiate the price, and arrange a time and place to sell the bananas. Banana farmers can select a trader who offers the highest buying price by making phone calls to some potential buyers. Because one representative can negotiate with traders on behalf of fellow banana producers, who are relatives and friends in the same community, individual banana producers do not need to own mobile phone units. Futher, such arrangements for negotiation with the traders by a representative of the community facilitate business transaction without breaking community rules, as discussed by Hayami and Kawagoe (1993). For traders, prearrangements can reduce their
9780230_274587_16_cha15.indd 255
6/23/2010 7:15:56 PM
256 Takashi Yamano et al.
transaction costs significantly because they can fill up their trucks in a few hours and then transport fresh bananas to urban markets. Fresh bananas collected in such an efficient transaction can fetch higher prices at urban markets, and suffer less spoilage. Thus, the benefits from the reduced transaction costs can be shared by both banana producers and traders. Recent studies report such benefits of mobile phone use in agricultural markets in other developing countries as well. For instance, Jensen (2007) shows that fishermen in India have increased arbitrage among local fish markets after the introduction of mobile phones, leading to a decline in geographical variation in fish prices and a reduction in spoilage. On the marketing of crops, Overa (2006) finds from a trader survey that Ghanaian traders reduce food spoilage through marketing their crops efficiently by using mobile phones. Similarly, Goyal (2008) uses a market-level dataset to identify an increase in the market price of soybeans after internet kiosks were introduced in central India. Aker’s work (2008) is the most closely related to our study. She carefully compiles secondary market price data on grains across 42 domestic and cross-border markets in Niger, and merges that with mobile phone coverage. She uses price information from a panel survey of 395 traders and finds that mobile phones reduce grain-price dispersion across markets by a minimum of 6.4 percent, and reduce intra-annual price variation by 12 percent. Descriptive analysis By using the Ugandan panel data, described in Section 15.2, we stratify our sample of banana farmers into three groups in Table 15.4 according to the timing of the mobile phone coverage, and present the ratio of households who have sold bananas and the ratio of sales to production. The first group is the households who were covered by the network even at the time of the
Table 15.4
Mobile phone network coverage and ownership in Uganda
Proportion (%) of households by mobile phone network coverage
Region
Number of households (A) Number
2003 and 2005 (B) %
Eastern Central Western Total
383 250 223 856
34.7 49.2 47.9 42.4
9780230_274587_16_cha15.indd 256
Only in 2005 (C) % 57.7 50.8 34.5 49.6
Proportion (%) of households who own mobile phone units
Never covered (D) %
2003 (E) %
7.5 0 17.4 7.9
2.6 8.0 3.1 4.3
2005 (F) % 8.6 17.6 9.8 11.5
6/23/2010 7:15:56 PM
Recent Developments of Agricultural Markets
Table 15.5
257
Ratio of banana sellers and sales by mobile phone coverage in Uganda Ratio of sellers to producers (%)
Distance Mobile to district coverage center (A) All locations
N/Y Y/Y N/N N/Y More than Y/Y 20 miles N/N Less than N/Y 20 miles Y/Y N/N
Ratio of sales to production (%)
2003 (B)
2005 (C)
Dif. (D)
2003 (E)
2005 (F)
Dif. (G)
48.3 54.0 74.4 42.9 67.5 25.0 47.6 50.9 76.6
53.4 50.7 83.7 68.0 72.2 25.0 46.1 46.5 90.0
5.1 –3.3 9.3 25.1* 4.7 0.0 –1.5 –4.4 13.4
20.0 19.6 24.0 19.0 25.2 16.6 20.8 18.5 23.8
24.6 22.8 27.2 30.2 26.8 5.5 22.0 21.2 26.4
4.6** 3.2* 3.2 11.2** 1.6 –11.1 1.2 2.7* 2.6
Notes: N/Y indicates the group of households that were not covered by the mobile phone network in 2003 but were covered in 2005. Y/Y indicates the group of households that were covered by the mobile phone network both in 2003 and 2005. * and ** indicate significance at the 5 and 1 percent level, respectively.
2003 survey, and the second group is the households who attained access to the network coverage between the 2003 and 2005 surveys. The third group is the households who have never been covered by the network. We note that the number of sample households in this last group comprises only 7.9 percent of the total sample. Then, we further stratify our samples by their distance to the district centers, where banana markets are usually located, because we think that the impact of the mobile phone network on crop sales would depend on the distance to markets. In Table 15.5, we find a large increase in the proportion of households selling bananas in locations more than 20 miles away from the district centers. The proportion of households who sold bananas increased from 42.9 percent in 2003 to 68.0 percent in 2005 in locations where the mobile phone network became available between the 2003 and 2005 surveys (the second group). We also find that the proportion of banana sales compared to their production increased by 11.2 points among the same group. In contrast, the proportion of households who sold bananas and the proportion of banana sales compared to production did not increase significantly in locations where the mobile phone network was already available in 2003 (the first group). Thus, we conjecture that this increase in sales is due primarily to both the reduction in spoilage and the increase in production induced by the higher farmgate price. As for locations never covered by any mobile phone network, the ratio of sellers to producers increased in areas less than 20 miles from the district centers, but the extent was much less than the increase observed in areas newly covered between 2003 and 2005. The increase in sellers may be
9780230_274587_16_cha15.indd 257
6/23/2010 7:15:57 PM
258
Takashi Yamano et al.
Table 15.6 Market participation and mobile phone coverage/possession: Banana production Dummy variable for selling bananas Variables Household mobile phone possession dummya Community mobile phone coverage dummy Distance x mobile coverage Year 2005 (=1) F-stat on IVs Number of observations
Ratio of sales to production
FE (A)
FE (B)
FE-IV (C)
FE (D)
FE (E)
FE-IV (F)
0.203* (2.35)
0.209* (2.44)
0.151 (0.19)
0.118* (2.39)
0.121* (2.45)
–0.062 (0.14)
0.055 (1.57)
–0.095 (1.41)
–0.094 (1.11)
0.054** (2.68)
–0.011 (0.29)
0.000 (0.02)
0.481** (17.7)
0.008** (2.62) 0.502** (17.8)
0.174** (11.2)
0.003* (1.99) 0.183** (11.3)
1161
1161
0.007** (2.53) 0.505** (10.85) 2.58 1151
1161
1161
0.003* (1.81) 0.191** (7.04) 2.58 1151
Notes: Numbers in parentheses are absolute t-values. * and ** indicate significance at the 5 and 1 percent level, respectively. The distance to the district center is not included in the models because it is fixed over time. a Instrumental variables: household mobile possession (HHmobile) is instrumented by the four interaction terms between the mobile coverage dummy and the four household characteristics: log of farm equipments value, age of household head, education of male adult, and education of female adult. These IVs together pass the overidentification test at the 1% significance level.
a result of the diffusion of market price information through a mechanism other than mobile phones in areas close to the district center. Estimation results To estimate the impacts of the mobile phone network on banana marketing, while controlling for unobserved community and household characteristics, we estimate the fixed-effect model and present the results in Table 15.6. The results indicate that the mobile phone network increases the probability of banana sales in remote areas: the interaction term of the distance to the district center and the mobile phone coverage has a positive and significant coefficient. The mobile phone network coverage itself, however, has a negative coefficient and thus decreases the probability of banana market participation by 9.5 percentage points. As a result, the combined impact is negative near the district center and positive in remote areas. According to the results, the combined impact turns to positive at 12 miles away from the district center. The results, therefore, suggest that the farther away the farmer is located from the district center, the larger the impact on market participation caused by mobile phone coverage. Because banana farmers in remote areas have lower costs for
9780230_274587_16_cha15.indd 258
6/23/2010 7:15:57 PM
Recent Developments of Agricultural Markets
259
banana production in general, the mobile phone network may have favored the banana producers in remote areas over those near the district centers. The result does not change much even when we treat household-level mobile phone possession as an endogenous variable and apply the fixed-effects instrumental variable (FE-IV) model. The results for the ratio of banana sales quantity to production are very similar to the results for market participation. In summary, according to the results from the instrumental variables model in Table 15.6, household-level mobile phone possession does not have a significant impact on banana sales participation, but the community-level mobile phone network coverage has a significant impact in remote areas. The results, therefore, suggest that even those households who do not have mobile phone units benefit from the mobile phone network, possibly because traders use mobile phones to contact farmer representatives in banana-producing areas to make arrangements to buy bananas. As we mentioned earlier, as long as one representative farmer has a mobile phone unit in a village community where personal networks are well established, the representative can arrange a time and place to collect bananas, and even negotiate the selling price with potential buyers on behalf of fellow farmers. In all likelihood, the community-based personal network supports the expansion of the market when the mobile phone becomes a new marketing tool.
15.6 Conclusion Drawing on Hayami’s propositions about the important role playing by communities in the functioning of markets, this chapter highlights the importance of enhancing community participation in the market to reduce rural poverty in Africa. Panel data from Kenya and Uganda have been used to report recent developments in agricultural markets, following market liberalization and the introduction of information technology. By using panel data of 862 rural households in Kenya, we find that the proportion of dairy farmers who sold milk to dairy cooperatives drastically decreased from the late 1990s to 2004, while that of private traders increased greatly. Furthermore, our estimation results suggest that the marketing costs of milk have been reduced over time. In Uganda, we find that the mobile phone network expansion has had a large impact on the market participation of farmers in areas farther away from district centers. This implies that mobile phone coverage expansion in Uganda encourages the market participation of farmers who reside in remote areas. Although we are unable to quantify the roles of community mechanisms in assisting market developments in the data, our field observations suggest that problems related to private milk traders have ceased as trust-based and long-term relationships have developed between the traders and the dairy farmers, and between the traders and the retailers, in Kenya. This is
9780230_274587_16_cha15.indd 259
6/23/2010 7:15:57 PM
260 Takashi Yamano et al.
clear evidence supporting the proposition of Hayami (2009) that business contractor experiments based on mutual trust significantly reduces transaction costs. Our field observations also suggest that farmers do not need to own mobile phones individually because market information can be shared widely among farmers in the same community. As long as one representative among them has access to a mobile phone, this person can make arrangements with traders on behalf of fellow farmers. This may encourage farmers to form farmer groups, which may help farmers not only in terms of marketing but also in terms of technology diffusions or other knowledgesharing. The two examples in this chapter suggest the successful development of agricultural markers in East Africa, where more such developments are needed.
References Aker, J. C. (2008) “Does Digital Divide or Provide? The Impacts of Cell Phones on Grain Markets in Niger,” Working Paper 154 (New York: Center for Global Development). Aoki, M. and Hayami, Y. (2001a) Community and Markets in Economic Development (New York: Oxford University Press). Aoki, M. and Hayami, Y. (2001b) “Introduction: Community and Markets in Economic Development,” in M. Aoki and Y. Hayami (eds.), Community and Markets in Economic Development (New York: Oxford University Press). Argwings-Kodhek, G. and Karin, F. (1999) “Draft Report on Revitalizing the Dairy Sector in Kenya,” mimeo (Nairobi: Tegemeo Institute). Badiane, O. and Shively, G. (1998) “Spatial Integration, Transport Costs, and the Response of Local Prices to Policy Changes in Ghana,” Journal of Development Economics, 56, 411–31. Central Bank of Kenya (2002, 2004, 2005) Statistical Bulletin (Nairobi: Central Bank of Kenya). Goyal, A. (2008) Information Technology and Rural Markets: Theory and Evidence from a Unique Intervention in Central India’ processed (Washington, DC: the World Bank). Hayami, Y. (2009) “Social Capital, Human Capital and the Community Mechanism: Toward a Conceptual Framework for Economists,” Journal of Development Studies, 45 (1), 96–123. Hayami, Y. and Kikuchi, M. (2000) A Rice Village Saga: Three Decades of Green Revolution in the Philippines (London: Macmillan). Hayami, Y. and Kawagoe, T. (1993) “Middlemen in a Peasant Community: Vegetable Marketing,” in M. Aoki and Y. Hayami (eds.), Community and Markets in Economic Development (New York: Oxford University Press). Hayami, Y. and Otsuka, K. (1993) The Economics of Contract Choice (London: Oxford University Press). Holloway, G., Nicholson, C., Delgado, C., Staal, S. and Ehui, S. (2000) “Agroindustrialization Through Institutional Innovation: Transaction Costs, Cooperatives, and Milk-Market Development in the East-African Highlands,” Agricultural Economics, 23, 279–88. Jayne, T. S., Govereh, J., Mwanaumo, A., Nyoro, J. K. and Chapoto, A. (2002) “False Promise or False Premise? The Experience of Food and Input Market Reform in Eastern and Southern Africa,” World Development, 28, 293–316.
9780230_274587_16_cha15.indd 260
6/23/2010 7:15:57 PM
Recent Developments of Agricultural Markets
261
Jensen, R. (2007) “The Digital Provide: Information (Technology), Market Performance, and Welfare in the South Indian Fisheries Sector,” Quarterly Journal of Economics, 122 (3), 879–924. Kijima, Y., Yamano, T. and Baltenweck, I. (2010) “Emerging Markets in the PostLiberalization Period: Evidence from the Raw Milk Market in Rural Kenya,” Journal of African Economics, 19 (1), 88–110. Muto, M. and Yamano, T. (2008) “The Impact of Mobile Phone Coverage Expansion on Market Participation: Panel Data Evidence from Uganda,” World Development, 37 (12), 1887–96. Negassa, A., Myers, R. and Gabri-Madhin, E. Z. (2001) “Grain Marketing Policy Changes and Spatial Efficiency of Maize and Wheat Markets in Ethiopia,” MTID Discussion paper No. 66, IFPRI, Washington, DC. Ngigi, M. (2005) “The Case of Smallholder Dairying in Eastern Africa,” EPT Discussion Paper 131 (Washington, DC: International Food Policy Research Institute). Overa, R. (2006) “Networks, Distance and Trust: Telecommunications Development and Changing Trading Practices in Ghana,” World Development, 34 (7), 1301–15. Owango, M., Lukuyu, B., Staal, S. J., Kenyanjui, M., Njubi, D. and Thorpe, W. (1998) “Dairy Co-operatives and Policy Reform in Kenya: Effects of Livestock Service and Milk Market Liberalisation,” Food Policy, 23, 173–85. Pender, J., Nkonya, J., Jagger, E., Sserunkuuma, P. and Ssali, H. (2004) “Strategies to Increase Agricultural Productivity and Reduce Land Degradation: Evidence from Uganda,” Agricultural Economics, 31, 181–95. Smale, M. and Tushemereirwe, W. K. (2007) An Economic Assessment of Banana Genetic Improvement and Innovation in the Lake Victoria Region of Uganda and Tanzania, Research Report 155 (Washington, DC: International Food Policy Research Institute). Smallholder Dairy Project (SDP) (2004) “Public Health Issues in Kenyan Milk Markets,” SDP Policy Brief 4 (Nairobi: International Livestock Research Institute). Staal, S. J., Delgado, C. and Nicholson, C. (1997) “Smallholder Dairying under Transactions Costs in East Africa,” World Development, 25, 779–94. Yamano, T., Otsuka, K., Place, F., Kijima, Y. and Nyoro, J. (2005) “The 2004 REPEAT Survey in Kenya (First Wave): Results,” GRIPS Development Database 1 (Tokyo: National Graduate Institute of Policy Studies).
9780230_274587_16_cha15.indd 261
6/23/2010 7:15:58 PM
16 Private-Sector Industrialization in China: Evidence from Wenzhou* John Strauss, Edward Y. Qian, Minggao Shen, Dong Liu, Mehdi Majbouri, Qi Sun, Qianfan Ying, and Yi Zhu
16.1
Introduction
The rapid growth of the Wenzhou economy is very unusual in that it has been based on privatization without tumult within transitional China, even during the planned economy regime before China’s industrial economic reforms started in the early 1980s. Since then, the “Wenzhou Model” has become well-known in China. Wenzhou is a municipality in southeast Zhejiang province on the east coast of China, with a total population of seven million. During the communist-planned economic regime, there had been little building of state-owned enterprises (SOEs) or township-village enterprises (TVEs) in the area. Some scholars claim that local poverty had a part in this, since TVEs and sometimes SOEs were often financed out of local resources. Also Wenzhou is just across the straits from Taiwan and was always a stronghold of the nationalists, and strongly pro-capitalist. Infrastructure around Wenzhou has been very poor until recent years. The port is poor and not heavily used; the roads have long been poor: in the mid-1980s it apparently took 12 hours to drive to Hangzhou, the capital city of Zhejiang province; now it takes 4 hours or so. Rail transport was not built-up either, until the mid-1990s. So infrastructure development followed, and did not precede, industrialization in Wenzhou. And yet, beginning in the mid- to late-1980s, strong industrial growth occurred in the Wenzhou region in various manufacturing industries such
*
We thank participants from both Chu KeZheng College and School of Economics at Zhejiang University, for the survey and their work on the database set-up and data process. Special thanks are due for the excellent research assistance from Yuan Ma, Yang Liu, Meng Meng Ge, Qiqi Cheng, Jing Hao, Jia Li, Zhipeng Liao, Hongchun Zhao, Quan Li and Lina Kay for their excellent research assistance. Many thanks to the participants of the Workshop on State, Community and Market for their useful comments, particularly to Jikun Huang, Kei Otsuka, Gus Ranis, Scott Rozelle and Tetsushi Sonobe. All remaining errors are ours. 262
9780230_274587_17_cha16.indd 262
6/23/2010 5:05:39 PM
Private Sector Industrialization in China
263
as clothing, shoes, cigarette lighters, spectacles, umbrellas, and many others. Almost all the firms in these industries in Wenzhou were private, even though the private sector was quite small in the early and mid-1980s. This was not as true in other areas of China, where TVEs and SOEs were many times more important than private-sector firms. The local government in Wenzhou mostly left the private sector alone, although they helped in some ways, such as building some industrial parks and marketplace areas in the early 2000s. There was even a very brief time around 1981 during which the local Wenzhou government actually allowed private banks (unheard of in China then) to establish themselves. These banks were short-lived: once the central government knew of them, they were closed. Perhaps an equally important role of government was at the central and provincial levels, in spurring the economic reforms which led to rapid growth throughout China. This undoubtedly encouraged industrialization in Wenzhou through a general “equilibrium effect.” As a result of this rapid industrial expansion, per-capita incomes rose in Wenzhou far faster than in the rest of China, despite starting at lower levels in the late 1970s. According to data in Sonobe et al. (2004), today mean incomes in Wenzhou are more than double those of the rest of China, whereas the reverse was true in the 1970s. There are reports these days of Wenzhou entrepreneurs now going outside Wenzhou and starting buyouts of SOEs and TVEs in other provinces. Much of the early industrial products in Wenzhou got the reputation, within China, for being very low-quality, but sold at a very low price. Shoes apparently would fall apart after little wear, and so forth. However, as Sonobe et al. (2004) show, and we provide corroborating evidence, firms in Wenzhou successfully began to climb the quality ladder during the 1990s. Today Wenzhou is a center of private-sector activity in China, and the source of many exports to the US and Europe. About 10 percent of all clothing production in China (for export or not) occurs in the Wenzhou area: for shoes it is 25 percent, for spectacles and cigarette lighters over 50 percent. Hence Wenzhou is a very important area for the study of industrialization in China. We note, however, that industrialization in Wenzhou is not typical of other provinces in China. It has been based to date on low-technology, labor-intensive industries and the private sector, particularly family firms, many of which are densely clustered. As is discussed in Chapter 13, the industrial cluster can be regarded as a voluntarily formed “community” in which the community mechanism of contract enforcement is at work. It is the purpose of this study to help shed light on the entrepreneurship, entrepreneurs and enterprise growth in industrial firms in Wenzhou. We rely on a probabilistic firm survey that we carried out in Wenzhou in early 2006 for three industries: shoes, spectacles, and general equipment. We fielded a formal survey, but we also asked many informal questions, which helps to enrich our story. In our survey we focused on getting detailed firms’
9780230_274587_17_cha16.indd 263
6/23/2010 5:05:39 PM
264 John Strauss et al.
histories, enabling us to learn how the firms started and grew, and what roles personal networks played in the growth process. We focused on the origins of the firms, including previous firms that may have been antecedents. We collected information on who the founders were, how many, their relationships with each other, and their background in terms of experience and other human capital. We collected detailed information on how they financed their start, and how they financed their expansion. We also have detailed information on the sources of technology introduced into the firms, particularly whether it was Chinese or imported, and whether the firms were getting technical instructions from foreign firms they may have been exporting to. We also assembled detailed information on markets, especially how such markets, including export markets, were found. We collected detailed information about explicit assistance that came from governments, local and/or central. Finally, we interviewed workers at each firm: managers, skilled workers such as designers, and production workers. We collected information on the human-capital background of the workers; how workers got their jobs, distinguishing between market mechanisms and community mechanisms in the sense of Godo and Hayami (2002); whether they had formal contracts, and for how long; whether they continued to come back to the same firm year after year; and detailed information on wages, as well as fringe benefits such as free housing. Virtually all the initial Wenzhou entrepreneurs are family-based, but some have partnerships. They all started with only a handful of workers, less than ten in almost all cases, and often less. However, the successful ones have grown into very large-scale firms. All were initially funded by their own family resources or with money put up by shareholders, who were usually close family members or friends, which relates to the community mechanism of contract enforcement. Government banks simply did not lend to small-scale private enterprises during this period. During expansion, reinvested profits contributed almost half of the finance. The structure of the rest of the chapter is as follows: in Section 16.2 we review the literature on the Wenzhou model, in Section 16.3 the survey is discussed in detail, and the results are presented in Section 16.4. In Section 16.5 we discuss these results and how they fit into the literature. Finally, we draw conclusions in Section 16.6.
16.2 The “Wenzhou model” Wenzhou is very unusual within China. It is famous historically for its entrepreneurs. During the communist regime, there had been little building of state-owned enterprises (SOEs) or township-village enterprises (TVEs) in the area. Indeed, if anything, Wenzhou was discriminated against under Mao because Zhejiang province was the home province of Chiang Kai-shek (Liu, 1992). Infrastructure around Wenzhou has been poor until recent
9780230_274587_17_cha16.indd 264
6/23/2010 5:05:39 PM
Private Sector Industrialization in China
265
years, including in particular transport. And yet, beginning in the midto late-1980s, strong industrial growth occurred in the Wenzhou region. Almost all the firms in these industries were private-sector and most of those small and medium enterprises (SMEs), though some of these firms have grown into large ones. As a result of this rapid industrial expansion, per-capita incomes rose in Wenzhou far faster than in the rest of China, despite starting at lower levels in the late 1970s. Today Wenzhou is the center of private-sector activity in China, and the source of many of the exports to the US and Europe. Hence Wenzhou is a very important area for the study of industrialization in China. Products in Wenzhou in the early stages of industrialization were lower-end technology products, and of lower quality. Imitation was prevalent among small firms. Firms producing the same type of goods were clustered in a small areas. Lenders were always borrowers’ friends, relatives, neighbors, or other shareholders. Lenders had a good understanding of the borrowers, so monitoring costs were less and moral-hazard problems were reduced. Further, entrepreneurs in Wenzhou had an extraordinary motivation to work hard from an instinct of survival, given the large number of firms they were competing with. With firm growth, a number of characteristics began to change. First, entrepreneurial ability arguably may have increased with the accumulation of experience. After starting up the firms, entrepreneurs began to accumulate experiences of firm management and how to expand the product market. Also, as the central government adopted a more favorable attitude toward the private entrepreneurs, many entrepreneurs become more ambitious and wanted to enlarge their investment scale. The technological level adopted by enterprises began to rise. Although low-quality products have their advantages at the start-up stage of company growth, they also brought a bad reputation to Wenzhou enterprises. Thus, in the new stage of company growth, many firms began to climb the quality ladder, as documented by Sonobe et al. (2004) for one industry: low-voltage electronics. Learning by doing assisted managers, as well as new and better inputs (including labor) becoming available. Even at an early stage of company growth, employees in small firms were always entrepreneurs themselves: as the firms grew, more skilled workers became necessary. Firms had to hire skilled workers, which for shoe manufacturers included shoe designers, for general equipment firms possibly engineers, and so forth. The growth of these private small- and medium-scale enterprises (SMEs), and of the regional economy, has been called the “Wenzhou Model.” The Wenzhou Model attracted researchers’ attention, particularly in the early 1990s, and now again since 2000. Liu (1992) has an excllent review of the early history of Wenzhou industrialization. Historically, people around Wenzhou were farmers in specialty crops (tea, fruit, sugar cane) for commercial sale, which agents from Wenzhou sold
9780230_274587_17_cha16.indd 265
6/23/2010 5:05:40 PM
266
John Strauss et al.
throughout China. Collectivization went against the grain for Wenzhou area farmers, and they resisted it. Liu (1992), as well as many other observers, argue that the population in Wenzhou adapted their traditional institutions to modern conditions. What was especially important, it is claimed, was initiative (particularly entrepreneurship), mobility, and markets. The rapid growth of household industries is a manifestation of the first factor, initiative. The sales of new products relied heavily on Wenzhou migrants to other parts of China helping to discover markets: that is mobility. Within the Wenzhou area market towns developed, sometimes with each specializing in a different good. As early as 1986, according to Liu, Wenzhou was given special status by the state council in Beijing as an “experimental zone.” In part Wenzhou was like other low-income areas of China that were receiving similar permission to experiment in order to improve economic outcomes. Wenzhou also received support at the regional level from the Shanghai party establishment. Although the existing literature analyzes the role of different factors in the establishment and growth of Wenzhou enterprises, there are few studies based on reliable statistical data, from probabilistic surveys. Examples include Sonobe et al. (2004), Sonobe and Otsuka (2006), and Ruan and Zhang (2009). Sonobe et al. (2004) look at firms that manufacture lowvoltage electronic products in the Wenzhou area, emphasizing that these firms started by imitating easy-to-replicate technologies. Later marketing expertise became important to expand sales, and these firms learned and at the same time became more technologically sophisticated. Sonobe and Otsuka (2006) focus on the roles of firm clustering in firm growth. Ruan and Zhang (2009) examine firms in a cluster of cashmere sweater producers, focusing on the initial financing of the firms and how clustering apparently aided this because firms could specialize in producing only certain parts of the whole sweater, lowering initial capital requirements. However, in Puyuan township in northern Zhejiang province, the conditions are rather different from those in Wenzhou, and the “Wenzhou” model does not exactly apply. In addition, we study firms that are involved in the entire manufacturing process for their industries, not parts suppliers. Our study is also somewhat different in coverage than the Sonobe-Otsuka and RuanZhang studies, which are much more focused, particularly on firms climbing the technological ladder or the mechanisms through which clustering of firms plays an important role. Since the economic reforms began, economic growth has proceeded rapidly. It is arguably the economic reforms that have been the primary cause of Wenzhou’s industrialization and rapid growth, along with the liberalization of migration. The latter has allowed a constant and very large flow of labor into Wenzhou, which has kept the wages of production workers lower than they might have been.
9780230_274587_17_cha16.indd 266
6/23/2010 5:05:40 PM
Private Sector Industrialization in China
267
16.3 Data and sampling design We conducted a formal probabilistic survey in 2006. First, we choose 72 firms, 24 each from three popular manufacturing industries in Wenzhou: shoes, spectacles, and general equipment. We decided to do our fieldwork in three areas of Wenzhou: Lucheng, Ruian, and Ouhai; these represented two areas of urban Wenzhou (Lucheng being downtown), plus a rural area (Ouhai). For each industry we took samples in two of the three areas (the top two areas for firms in each industry), choosing 12 firms in each region for each industry. The sampling frame was obtained from a complete census of firms in each of these three areas, listed by industry, obtained from the county statistical bureau. Information was supplied on names of the firms, addresses, sales and revenue estimates, and numbers of employees, as of December 2004. For each of the three areas and three industry combinations, we ranked firms by sales and sampled a first firm randomly from the smallest x firms, where x was chosen in order for us to be able to sample 12 firms from each area. We then sampled every xth firm by size. In this way we were able to get a sample of firms across the firm-size distribution. Table 16.1 shows average sales and employment in December 2004 of the sampled 72 firms. Note that there is one firm that is much larger than the others, with over 3000 employees as of December 2004. This turns out to be one of the largest shoe-manufacturing firms in China. We discuss this firm in detail below, as it is an outlier in many dimensions of our results. Apart from this firm, only five others have more than 300 employees, while 56 firms have 100 employees or less. So the firm-size distribution in these three industries in Wenzhou is skewed towards small and medium firms, as is the case in much of the developing world. As is usual in surveys, we were not able to find and complete surveys for all of the 72 firms. Two firms outright refused to see us, 8 had gone out of business, which we found out when we went to the December 2004 address,
Table 16.1
Distribution of sampled firms by industries and regions Shoemaking
Spectacles
General equipment
Lucheng Ouhai Lucheng Ouhai Lucheng Ruian Total number of firms 469 459 167 292 233 600 Number of sample firms 12 12 12 12 12 12 Average sales (10,000) 5825.23 468.86 771.5 673.18 1204.43 735.03 (4936.51) (190.40) (418.69) (342.94) (880.52) (418.69) Average employment 359.00 82.08 74.5 128.41 81.16 55.08 (881.48) (107.56) (87.13) (224.49) (195.76) (111.19) Note: Standard errors are in parenthesis.
9780230_274587_17_cha16.indd 267
6/23/2010 5:05:40 PM
268 John Strauss et al.
Table 16.2
Distribution of type of ownership across period of establishment Pre1979–84 1985–94 1995–2004 Total Observation
Industry
Shoemaking Spectacles General equipment Total
Ownership Collective Type Private Partners (outside family) Share-owned (also private) Other
0.27 (0.21) 0.13 (0.15) 0.60 (0.26) 0.16 (0.06)
0.44 (0.29) 0.09 (0.11) 0.47 (0.31) 0.34 (0.13)
0.27 (0.22) 0.63 (0.25) 0.11 (0.10) 0.49 (0.17)
0.33 (0.22) 0.36 (0.23) 0.31 (0.23) 1.00 (0.00)
0.39 (0.17) 0.27 (0.21) 0.22 (0.10)
0.00 0.00 0.79 (0.06) 0.13 (0.09)
0.00 0.00 0.49 (0.09) 0.32 (0.14)
0.06 (0.04) 0.55 (0.10) 0.24 (0.09)
0.07 (0.07) 0.06 (0.06)
0.08 (0.06) 0.00 0.00
0.20 (0.10) 0.00 0.00
0.14 (0.07) 0.01 (0.01)
21 17 17 55 4 31 10
9 1
Notes: There are 55 observations. Standard errors are in parenthesis. The corrected Pearson chi-squared test of equality across groups of industries has value 4.58 and P-value 0.04, and across type of ownership has value 2.71 and P-value 0.11.
and 7 others had apparently moved, but we could not find out where. As can be seen from Table 16.2, we were able to contact 55 out of the 72 sampled firms. This of course raises a selection issue, but this is part of a deeper issue, that over the years many firms have no doubt entered and exited these three industries in these three areas, and so the sample of firms that existed in December of 2004 will not be representative of the firms that have existed over this period. The kind of sample that would be ideal for this study would include entrants for each year from the mid-1980s until now, and the sampled firms would be followed over time. That sampling design would have picked up both new entrants and failures. What we have is a random sample of firms that existed in December 2004, which will be biased towards successful firms, and particularly older firms. But our 55 firms do not even constitute a random sample of the existing firms from December 2004, because of the closures and moved firms from December 2004 until February/March 2006. We have attempted to address the latter source of selection bias, while recognizing that we have a broader selection of successful firms. What we do is to include all 72 firms in a simple logit regression where the dependent variable is a binary variable equal to one if the firm was contacted. The covariates are dummy variables for
9780230_274587_17_cha16.indd 268
6/23/2010 5:05:40 PM
Private Sector Industrialization in China
Table 16A.1 Logit regression whether the firm was surveyed
269
of
Dependent Variable = 1 if firm was surveyed Ruian Lucheng Spectacles General equipment Small Medium Constant Pseudo-R-squared
3.39 (1.51) 0.61 (0.79) –1.22 (0.82) –2.62 (1.06) –2.40 (0.93) –1.19 (0.94) 3.20 (1.05) 0.22
Notes: Number of observations is 72. Standard errors are in parenthesis. The dependent variable is a dummy which is 1 if the firm was surveyed and zero otherwise. The independent variables are all dummy variables for location, industry, and size.
tercile of firm size in December 2004, where the largest tercile is dropped, and dummies for industry and area are added. The results are reported in Table 16A.1. Being located in Ruian made it somewhat more likely that we could contact the firm than if it was in Ouhai. Being a general equipment manufacturer made it less likely to be contactable than if made shoes. Most important from the point of view of potential selection bias is the positive relationship between firm size and being contacted. This is not surprising, because larger firms are more likely to be successful and therefore to be still in business. Because of the selection results, all the statistics we report hereafter are weighted. In fact there are two steps for the weighting, first a simple sample weight that takes care of the fact that the sampling proportions are different for each industry/location stratum. For this we use the ratio of the population fraction to the sample fraction of firms by industry located in each of the three areas. We use our sample frame as our estimate of the population, since it is supposed to be a census of all manufacturing firms, no matter the size. The second step is to apply inverse probability weighting using the inverse of the predicted probabilities from the logit regression (see
9780230_274587_17_cha16.indd 269
6/23/2010 5:05:40 PM
270
John Strauss et al.
Wooldridge, 2002, for a justification). We truncate the inverse probabilities at the 90th percentile and multiply that by the sampling weight described above to arrive at the final sample weight. We use these for all of the firm analyses, but not for the labor figures.
16.4 Empirical results Characteristics of entrepreneurs Who are the entrepreneurs who founded these firms? What is their background? Table 16.2 shows the year the firm was founded. Some 85 percent of these firms were founded after 1990. When we examine founding dates by type of ownership of firms, we find that almost all started out as privatesector firms, and all are so today. Four firms started as collective or red-hat firms (former township village enterprises, TVEs): in fact these firms were really private but used as a cover that they had been TVEs. All of these four were founded before 1985. Over half of our firms are purely family businesses, while 38 percent had partners or shareholders outside the immediate family. Table 16.3 shows that Wenzhou entrepreneurs, generally speaking, started the firm by themselves, or with a very small number of co-owners. Some 85 percent of firms had three or fewer founders, with 43 percent having only one. Shoe firms are especially likely to have only one founder. The other founders are most likely to be close relatives of the chief founder (usually the spouse or brother), followed by other relatives and friends (Table 16.4). This shows the importance of family and informal networks of friends in the establishment of these small businesses. It fits well Table 16.3
How many founders did the enterprise have? Industry
1 2 3 4 or above Total
Shoemaking
Spectacles
General equipment
Total
0.75 (0.05) 0.25 (0.05) 0.00 (0.00) 0.00 (0.00) 1.00
0.31 (0.08) 0.22 (0.04) 0.35 (0.00) 0.12 (0.04) 1.00
0.18 (0.02) 0.34 (0.15) 0.11 (0.01) 0.38 (0.14) 1.00
0.42 (0.11) 0.26 (0.05) 0.16 (0.07) 0.16 (0.10) 1.00
Observations 23 15 8 8 54
Notes: Number of observations is 54. Standard errors are in parenthesis. The corrected Pearson chi-squared test of equality across groups of industries has value 5.18 and P-value 0.02.
9780230_274587_17_cha16.indd 270
6/23/2010 5:05:41 PM
Private Sector Industrialization in China
271
Table 16.4 The shareholder’s relationship with the chief leader at the time of establishment Industry Shoemaking
Spectacles
General equipment
Total
Observations
0.48 (0.04) 0.21 (0.15) 0.21 (0.18) 0.00 (0.00) 0.00 (0.00) 0.10 (0.07) 1.00
0.44 (0.04) 0.19 (0.00) 0.24 (0.02) 0.13 (0.06) 0.00 (0.00) 0.00 (0.00) 1.00
0.48 (0.10) 0.24 (0.09) 0.22 (0.01) 0.00 (0.00) 0.05 (0.01) 0.02 (0.03) 1.00
0.46 (0.05) 0.21 (0.04) 0.23 (0.03) 0.05 (0.04) 0.02 (0.02) 0.02 (0.02) 1.00
25
Close relatives Extended relatives Friends Former colleagues No relationship Others Total
14 11 2 1 2 55
Note: There are 55 observations. Standard errors are in parenthesis. The chi-squared test of equality across groups of industries has value 0.30 and P-value 0.76.
with Professor Hayami’s notion of the “community mechanism” of contract enforcement. Table 16.5 displays key background characteristics of the chief founders. The chief founders tend to be young at the time of founding, a mean of 35 years. Interestingly, in our sample the age of the chief founders is lower the larger the firm size at its founding. Unfortunately we did not collect data on founders’ education completed at the time of founding, but only at the time of our survey. This may overstate education at the time of founding because some entrepreneurs went back to their university and received honorary degrees. At the time of the survey, about equal numbers of chief founders had junior high school and senior high school schooling. A small number had only primary schooling, and an equally small number had had some college education. So clearly the level of schooling of these entrepreneurs was not especially large. Over 90 percent of founders were not Communist Party members at the time of founding. For large firms the percentage of party members is somewhat higher, at 17 percent.71 In terms of experience, 90 percent had some experience of farming, but 22 percent had some sales and marketing experience (possibly in addition to farming). The importance of sales as a former occupation is very similar to results found in other micro- and small-firm surveys (see Otsuka
9780230_274587_17_cha16.indd 271
6/23/2010 5:05:41 PM
272 John Strauss et al.
Table 16.5
Characteristics of founders at the time of establishment Industry Shoe General making Spectacles equipment
Age
Average age
Education
College or above Senior high school Junior high school Primary school Not educated
Total Party Yes membership Total
31.78 (3.20) 0.20 (0.06) 0.29 (0.01) 0.25 (0.07) 0.10 (0.07) 0.16 (0.06) 1.00 0.12 (0.01) 1.00
37.87 (1.93) 0.22 (0.14) 0.35 (0.03) 0.43 (0.12) 0.00 0.00 0.00 0.00 1.00 0.12 (0.13) 1.00
34.92 (0.93) 0.11 (0.01) 0.23 (0.17) 0.37 (0.08) 0.21 (0.07) 0.08 (0.03) 1.00 0.00 0.00 1.00
Total Observation 35.08 1.55 0.18 (0.06) 0.29 (0.07) 0.36 (0.07) 0.10 (0.06) 0.08 (0.04) 1.00 0.08 (0.05) 1.00
48 11 18 16 5 4 54 5 46
Notes: Number of observations for age is 48, for education is 54, for party membership is 46. Standard errors are in parenthesis. The corrected Pearson chi-squared test of equality across groups of industries has value 1.26 and P-value 0.38 for age, 1.29 and 0.30 for education, 0.39 and 0.59 for party membership.
and Sonobe, Chapter 13, this volume). Over half the chief founders had been apprentices, but virtually none had worked for the government, including SOEs. It is instructive to recount some actual firm histories at this point, to get a better sense of what happened over this period. We begin with the story of the largest firm in our sample, a very prominent shoemaking firm. The company was established in 1980. The founder was about 30 years old at the time and was having a lot of difficulty paying the basic living expenses of his family, especially because he had to take care of his brothers and sisters in addition to his own family. At the time he was working at a machinery factory and the company produced a range of machinery, some of which could also have been used in shoe factories. He had a friend at the time who suggested he make and sell shoes in order to earn more money. Motivated by this idea, he built a small, poorly-equipped workshop in the basement of his house and started making shoes, with his brothers and two hired workers. He borrowed 500 RMB from his neighbor to start the business. During those years, he encountered many difficulties in maintaining the business. He did everything he could to earn enough money to finance the investment and growth of the company, such as selling fruit on the streets.
9780230_274587_17_cha16.indd 272
6/23/2010 5:05:41 PM
Private Sector Industrialization in China
273
Despite this arduous process, he was capable of managing the business growth process. The small, under-equipped workshop became bigger and better every year. In 1985, when the government announced that private enterprise was considered legal (as a result of the state council ruling), he rushed that same day to the government registration office to register his business. By the late 1980s the small workshop had become medium-sized, with about 50 workers. The firm at this point revolutionized its product design and quality. At the time, Wenzhou shoes were regarded as very low-quality in China. Although the company’s products were not actually of low quality, the misperception of people was an obstacle. In 1987, the Zhejiang local government organized a large-scale burning of shoes made in Wenzhou in a major public square in the city of Hangzhou, the capital of Zhejiang province. In the wake of this disastrous event for the Wenzhou shoe industry our owner traveled to Guangzhou, Shanghai, and Italy to learn more about the shoemaking business. In 1989, the owner set up a new company aiming to produce shoes with “quality and credibility.” He set up the first mechanized stringing assembly line in Wenzhou, and invested 1.2 million RMB on technological improvements such as computer-aided design (CAD) software. He tried to find a brand name associated with “first-class quality, style, and first-class service.” This completely distinguished his company from his competitors in Wenzhou, who were still focused on low-quality, low-priced lines of shoes. In 1993, the company introduced a production line bought from Taiwan and became one of the top ten brands in China. In 1996, another production line was imported from Italy. In the same year, the company completed a project to computerize key parts of production lines, and also the shoe design processes. Currently, the firm dominates the high- and middle-end markets for shoes in China. It has more than 2500 chain stores in China and 120 stores in over 10 countries, including the US and Europe. It has won many prizes and awards. It is essential to understand the completely different direction that the firm took after 1989. This was an enormous risk that the owner undertook. A second example comes from the general equipment industry in our sample. The firm that we visited was a kind of general contractor. A customer would come to them and give them an order, that is, the blueprint of the product to be produced. In return the company agreed to supply the orders at low prices. If the customer agreed, the firm would subcontract the order to production shops. This firm was a family business. The founder, the father and head of the family, used to be a salesman of plastic machinery. Being acquainted with the industry, he saw the opportunity for an intermediary firm that could connect customers with firms that made quality products at low prices. In 1989 he borrowed money from his relatives and started his own company based on this idea. In the beginning the company specialized in plastic products, though later they moved into metal products. Their
9780230_274587_17_cha16.indd 273
6/23/2010 5:05:42 PM
274 John Strauss et al.
main competitive advantage was the network they built inside the industry, both with customers they were selling to and with product suppliers who could provide better price/quality combinations. This intermediary firm repeatedly built relationships, both with buyers and with supplier firms. Now, the firm is actually starting production. Financing the firm at its founding and during its growth Equity financing, that is, shareholders’ investment, is the major source of financing for start-up firms in our sample, with 74 percent saying they used that form of financing (Table 16.6). However, shareholders in our sample were largely close relatives and friends, as shown in Table 16.4. Finance from shareholders is followed by resources from the family, who are not strictly speaking shareholders (38 percent of firms), and banks (11 percent). Rural credit unions are not an important source of initial lending in our sample. Hence personal ties, or what Professor Hayami called the “community mechanism of contract enforcement,” are of critical importance in the financing of start-up firms in Wenzhou. During the firm’s early growth stage the financing sources are somewhat different than for start-up. In particular, reinvestment of the firm’s profits
Table 16.6
Sources of initial funding Industry
Firm size
ShoeGeneral making Spectacles equipment Small Medium Large Shareholders Yes 0.53 0.97 (0.06) (0.03) Family Yes 0.47 0.38 (0.10) (0.14) Other personal loans Yes 0.00 0.00 0.00 0.00 Informal loans Yes 0.05 0.00 (0.04) 0.00 Banks Yes 0.05 0.19 (0.03) (0.10) Rural credit unions Yes 0.00 0.03 0.00 (0.03)
Total Observations
0.66 (0.08)
0.77 (0.16)
0.71 (0.14)
0.70 (0.12)
0.74 (0.10)
35
0.30 (0.11)
0.41 (0.08)
0.37 (0.12)
0.36 (0.11)
0.38 (0.08)
18
0.08 (0.03)
0.00 0.00
0.07 (0.08)
0.00 0.00
0.02 (0.02)
1
0.00 0.00
0.00 0.00
0.00 0.00
0.05 (0.05)
0.02 (0.02)
1
0.08 (0.03)
0.22 (0.10)
0.00 0.00
0.11 (0.06)
0.11 (0.05)
4
0.08 (0.03)
0.07 (0.05)
0.00 0.00
0.04 (0.04)
0.04 (0.02)
2
Note: Number of observations is 50. Standard errors are in parenthesis.
9780230_274587_17_cha16.indd 274
6/23/2010 5:05:42 PM
Private Sector Industrialization in China
275
becomes important, as shown in Table 16.7. Funds from shareholders still dominate, providing 48 percent of new investment, and retained earnings supply about 36 percent. Non-shareholder family and friends, and banks, are much less important, but still provide something. There are differences between small and large firms, with shareholder financing being appreciably larger for small firms, and retained earnings being more important. Firms believe that formal-sector banks are quite difficult for them to borrow from, and that family and friends are a much easier source. This helps explain why shareholders are the major source of funds at start-up (note that it is virtually impossible in our data to distinguish between resources from individual shareholders and from their families). The major reason for shareholders, families, and friends being the major source of financing at the birth of these firms is that formal-sector banks were simply not available to these private-sector firms, at least at the start Table 16.7
Sources of funding during firm growth (% of funds) Industry Shoe Making
Retained from profit (%) 27.83 (24.23) Family and friends 10.25 (6.96) Shareholders 44.97 (19.69) Other personal loans 0.00 0.00 Informal lending 5.39 (4.69) Banks 11.56 (2.28) Rural credit unions 0.00 0.00 Firm payables 0.00 0.00 Total 100 Observations 9
9780230_274587_17_cha16.indd 275
Firm Size
General Spectacles equipment
Small
Medium
Large
Total
30.60 (21.69)
48.32 (14.26)
40.34 (14.84)
52.94 (19.37)
15.18 (12.10)
36.53 (11.69)
4.30 (3.05)
5.66 (5.66)
5.61 (2.07)
10.40 (8.27)
3.35 (3.85)
6.41 (2.48)
60.12 (28.27)
36.63 (17.55)
47.95 (18.00)
31.43 (13.89)
62.28 (17.13)
48.07 (13.95)
0.00 0.00
1.12 (0.53)
0.00 0.00
1.34 (1.48)
0.00 0.00
0.42 (0.37)
0.00 0.00
0.00 0.00
0.00 0.00
0.00 0.00
4.70 (4.27)
1.45 (1.54)
1.39 (0.99)
7.72 (3.22)
3.64 (2.25)
3.89 (2.35)
12.70 (5.34)
6.51 (2.23)
0.00 0.00
1.12 (0.53)
1.11 (0.75)
0.00 0.00
0.00 0.00
0.42 (0.37)
1.55 (1.10) 100 10
0.00 0.00 100 11
0.00 0.00 100 9
0.00 0.00 100 11
1.79 (2.05) 100 10
0.55 (0.54) 100 30
6/23/2010 5:05:42 PM
276 John Strauss et al.
of the Wenzhou industrialization process. But there is a second important reason as well, which is that firms can generally start extremely small, as our shoe and clothing firm examples show. Indeed, there is old literature on microenterprises that demonstrates that the assets used to start new microenterprises are very small, and well within the reach of many households (see, for example, Kilby, 1971; Kilby and Liedholm, 1990; Liedholm and Mead, 1987, 1999, plus many of their students). While there are studies demonstrating that initial wealth is a positive correlate of entering into a family business under these conditions with limited availability, or none, of formal-sector lending (see Paulson and Townsend, 2004 for a discussion of Thailand), and we cannot rule that out for our Wenzhou founders, the amounts needed to start a business are generally small enough that many people are evidently able to do so. This evidence in Wenzhou is in fact remarkably consistent with the historical evidence for post-World War Two Japan and Taiwan starting in the 1950s and 60s (see Ranis, 1995). The next question is whether micro and small firms can grow into medium and large firms. In the Wenzhou case the answer is obviously yes, at least in some cases. More generally there is a fairly large literature that measures firm growth as a function of initial size, usually, though not always, finding that smaller firms grow faster in developing countries.72 These case studies demonstrate that economic theories of low-level growth traps that rely on credit constraints to deter people from becoming entrepreneurs may have little basis in fact, certainly in east Asia. Technology selection and innovation There were many reasons why Wenzhou enterprises adopted low levels of technology during the early stages of enterprise growth. First, at the beginning of the economic reforms, the prices of many major production factors such as natural resources and high skilled workers were regulated by the government in China, and kept at a high level. There were also dual price systems: the government set a relatively low price for production factors for the in-system SOEs in order to protect their profits. For out-of-the-system firms such as private firms in Wenzhou, it was very difficult to gain the necessary production factors, and even if they could, the price was much higher than the in-system price. This made it difficult for Wenzhou enterprises to produce the same products as SOEs, or adopt advanced technologies like those of the SOEs. On the other hand, because the development of the old planned economy was so imbalanced, many consumer goods were in short supply. As a result, the entrepreneurs had easy access to unregulated but profitable small-commodity markets. Therefore goods produced by small private firms tended to be low-end products and the technologies were traditional and nonexclusive, meaning that they were obtained either by imitating advanced technologies or from self-innovation. Finally, the initial wealth of new entrepreneurs was insufficient to purchase large
9780230_274587_17_cha16.indd 276
6/23/2010 5:05:42 PM
Private Sector Industrialization in China
277
and expensive equipment, especially given the state of the credit markets. In the later stages of firm growth, many firms began to climb the quality ladder and therefore had to raise their technological level (Sonobe et al., 2004). We began by asking the firms’ leaders to characterize the sophistication of the technologies that they were using when the firm was established. We divided technologies into ordinary technologies, ordinary plus some selfinnovation, intermediate-level technologies, and advanced technologies. It was found that only 10 percent of firms said that they were using advanced technologies, and the others were split between intermediate and ordinary technologies. Of course the meaning of these terms is certainly questionable, so that one well might want to focus on other dimensions. For example, the spectacle firms in Wenzhou generally make the plastic frames. They do not in general make prescription lenses, which require a much more advanced technology. They do make sunglasses lenses, though, which require a much easier-to-master technology. Indeed much of the output of the spectacle industry in Wenzhou consists of sunglasses, though they do purchase prescription spectacle lenses from suppliers in northern China and insert them into the frames that they make. We asked if the equipment was domestically made, or imported. Most of the firms use domestic equipment (82 percent) and only an estimated 23 percent of machines are imported. In this regard the large shoemaking firm in our sample is quite different, not at the time of founding, but starting nine years later. About 1 in 3 of the firms said they made improvements to the technologies they used. In light of the literature on technology acquisition in East Asia (see, for example, Evenson and Ranis, 1990; Pack and Westphal, 1986; Westphal et al., 1985), it is instructive to examine whether these firms received instruction from abroad in the use of their machines. The answer is generally no, although again our large shoemaking firm is an exception. We found that learning from doing was the most important way in which these early technologies were mastered, without outside assistance. Some of our firms did receive instructions from the equipment providers, which were mostly Chinese, and some by borrowing technicians from state-owned enterprises. We also investigated who was responsible for quality control. Nearly 80 percent of the firms claimed that they learned by themselves. Generally the firms trained workers to engage in this task. About 20 percent said that clients had some control over the firms learning about quality control (our categories are not mutually exclusive). This latter is a model that one might have expected on the basis of technology acquisition in South Korea in the late 1960s, as described by Westphal et al. (1985), so it is interesting that in Wenzhou, it was left much more to local firms to fend for themselves.
9780230_274587_17_cha16.indd 277
6/23/2010 5:05:43 PM
278
John Strauss et al.
As an instructive example of technology acquisition, one of our spectacle firms was one of the first such firms in Wenzhou. It is now a large firm of over 1000 employees, and sells both domestically and internationally. Among its largest customers is Walmart, with whom it has a contract for sunglasses. When this firm started, the owners were looking for a machine that would make the plastic spectacle frames. They found such an item, an elderly Chinese machine that had been discarded by an older company. It was not very usable, but technicians at the new firm took it apart and reassembled it, and in doing so discovered how to improve its construction. The firm began making its own machines because there were no local suppliers and because, through “reverse engineering,” improved the quality of the original machine. So at this point the firm became “vertically integrated.” In later years many other firms entered the spectacle industry in Wenzhou, and a supply industry of machine-makers developed. Now this firm contracts out to another supplier firm in Wenzhou, to make spectacle-frame machines to its specifications. So here is a case in which the clustering of firms induced machine-supplier firms to develop. These machines were all locally made, and no import of machines or parts was undertaken. On the other hand there are some important exceptions, the principal one in our sample being the large shoe-manufacturing firm, which in a later phase of development began to import its machines from abroad. Indeed its owner chose to import machines that are very capital-intensive, unlike the rest of the industry. The market discovery and competition Competition There is fierce competition in these low-technology industries in China, and particularly in Wenzhou. Table 16.1 shows the number of firms in each industry in our sampled areas: there are many of them. Table 16.8 shows that the firms are close together, generally within one km of the nearest firm of the same type, so there is a large amount of clustering of firms in our sample. The major areas of competition are price and quality. In shoemaking and spectacles both factors are claimed to be important, while in general equipment quality is said to be the main area of competition. When measuring the quality of the goods they make against national and international standards, half the firms think that the quality of their goods ranks with the average national quality, but there is a distribution around that point. In shoe and spectacle manufacture roughly two-thirds of the firms believe their products are of lower quality, and lower price, compared to Chinese products nationally. However, more than a third believe that they make higher-quality and lower-priced products than the national average. Only about 15 percent of firms believe that they manufacture higher-quality,
9780230_274587_17_cha16.indd 278
6/23/2010 5:05:43 PM
Private Sector Industrialization in China
279
Table 16.8 What was the distance between your company and the nearest company making the same product? Industry Shoemaking Less than 1 km 1 to 5 km More than 5 km Total
Spectacles
0.80 (0.06) 0.13 (0.11) 0.07 (0.05) 1.00
0.77 (0.14) 0.13 (0.13) 0.10 (0.01) 1.00
General equipment 0.75 (0.07) 0.14 (0.05) 0.10 (0.02) 1.00
Total Observations 0.78 (0.06) 0.13 (0.06) 0.09 (0.02) 1.00
33 7 5 45
Notes: Number of observations is 45. Standard errors are in parenthesis. The corrected Pearson chi-squared test of equality across the groups of industries has value 0.05 and P-value 0.94.
Table 16.9
Which firms export? Industry
Firm size
ShoeGeneral making Spectacles equipment Small Medium Large Total Observations Yes Observations
0.50 (0.08) 20
0.83 (0.17) 17
0.11 (0.01) 16
0.35 (0.23) 14
0.51 (0.15) 18
0.65 0.50 (0.17) (0.18) 21 53
25
Note: Number of observations is 53. Standard errors are in parenthesis.
higher-priced goods, and those firms are almost all in the general equipment industry. Markets and market discovery Remember that 22 percent of founders said that they had had some sales experience prior to founding their firms. Previous analysts have suggested that the earlier migratory experiences of Wenzhou entrepreneurs before the 1990s helped create clusters of Wenzhou people all over China, and thus greatly helped in market discovery. This may well be true, though our survey does not confirm it. One part of market discovery that we can shed light on from our data concerns export markets. We see from Table 16.9 that about half of the surveyed firms export, but the figures are very different across industries: over 80 percent of spectacle firms and 50 percent of shoe firms export, but only 11 percent of general equipment firms. The latter is probably due to the fact that these firms are supplying parts to other firms in the region. Exporting is also monotonic, in terms of firm size, across
9780230_274587_17_cha16.indd 279
6/23/2010 5:05:43 PM
280 John Strauss et al.
these three industries. In terms of the estimated percentage of export sales, almost no general equipment, just under 30 percent of shoes, and almost 80 percent of spectacles are exported. As regards other sales destinations, other provinces within China are the next most important destination, followed by local sales. One aspect of government assistance that could have helped in exporting was the establishment of “economic free zones.” These are small tracts of land within Wenzhou where the government allows firms to locate, and they receive certain advantages if they do. Firms locating in the economic free zones are more likely to export, and to export a higher proportion of their sales. We need to be careful here not to attribute causality to the economic free zones; it could be that the more profitable firms are located there, and that they are more likely to export their products. However, note that firms outside these zones are also highly likely to export. The other important consideration is how these firms learn about exporting possibilities. Table 16.10 sheds light on this. Nearly 80 percent of exporting firms in our sample do so through the Foreign Trade Company, which is a national governmental organization that matches Chinese firms with exporters. This is especially likely to be important for new exporters. One pattern in our data, though rather weak, is that larger firms are more likely to have direct export connections, and so not to have to export through the Foreign Trade Company. Although local transport conditions were thought by us to have been very poor during the period of Wenzhou industrialization, when we asked about transportation difficulties, for example slow shipments from suppliers or to markets, we were universally told that this was not a problem. Table 16.10
Channels of export (sales percentages) Industry
Firm size
ShoeGeneral making Spectacles equipment Small Medium Through 70.07 foreign trade (21.41) company Direct export 29.93 (21.41) Total Observations
100 9
Large
Total Observations
88.65 (2.41)
30.42 (10.78)
86.86 (15.90)
100.00 0.00
61.11 (14.26)
79.08 (9.69)
19
11.35 (2.41)
69.58 (10.78)
13.14 (15.90)
0.00 0.00
38.89 (14.26)
20.92 (9.69)
5
100 13
100 2
100 3
100 14
100 24
24
100 7
Notes: Number of observations is 24. Standard errors are in parenthesis. The Adjusted Wald test of equality across groups of industries has value, 11.33 and P-value, 0.02 and across groups of employment size has value 2.98 and P-value 0.16.
9780230_274587_17_cha16.indd 280
6/23/2010 5:05:43 PM
Private Sector Industrialization in China
281
Labor inputs, human-capital accumulation, and incentives As part of the firm survey, we “sampled” three types of workers at the firms: managers, skilled workers such as designers or engineers, and production workers. The sample was purposive, since we were not allowed to access worker lists from the sample firms. At each firm we tried to interview two managers, three skilled workers, and three production workers. We went to cafeterias at each firm during lunchtime and spoke to workers there. While the firm’s representatives were sometimes present, we asked supervisors to let us interview the workers alone, and generally, once they had seen one or two interviews, they let us continue. In this section, we start by comparing simple human-capital background characteristics of the three types of workers we interviewed. We then ask how they found the job they have now, how many years they have been working at the firm, whether they are from the Wenzhou area, and if not where they are from, whether they brought their family members with them, whether they have formal contracts, for how long, and the terms of the contract, including current monthly salary, year-end bonuses, and fringe benefits. Table 16.11 shows background characteristics of the sampled workers. Our sample is largely male, though of the production workers we interviewed 30 percent are women. Note that we do not argue that these proportions hold for the entire workforce, indeed for production workers at least these proportions are likely to be low. The average age of our interviewees is about 30, and somewhat higher for the managers. Two-thirds of the production and skilled workers interviewed are educated to junior high school level, or lower. For managers this fraction is only one-third. Still, only 16 percent of managers, 13 percent of skilled workers, and 6 percent of production workers claim to have any college education. Hence these industries clearly do not rely on the number of highly schooled managers and workers that other higher-tech industries might. In later years this situation may well change: if China does not expand its college-educated workforce, at some point it may have difficulty competing in the world, as happened in Japan in the 1990s and later, emphasized by Godo and Hayami (2002). Fewer than 5 percent of production workers claim to be Communist Party members, compared to 10 percent of skilled workers and 20 percent of managers. In terms of finding their current job, our main question was how much they used market mechanisms such as newspaper ads, and how much they used informal networks such as information from friends. For managers, 31 percent used friends and about 20 percent used ads (Table 16.12). Another 10 percent say that the firm found them, through informal or formal contacts. Some 56 percent of the managers we interviewed said they had no relationship with the chief founder, and another 12 percent said they were friends with him. Fifteen percent said they were close relatives, and another
9780230_274587_17_cha16.indd 281
6/23/2010 5:05:43 PM
282
John Strauss et al.
Table 16.11
Characteristics of interviewed workers Work type
Gender
Male Female
Manager
Skilled worker
Production worker
0.80 (0.03) 0.20 (0.03)
0.90 (0.03) 0.10 (0.03)
0.69 (0.06) 0.31 (0.06)
Observations
Age
Average Age
37.97 (0.89)
29.62 (1.61)
33.43 (1.06)
316
Education
University or above Senior high school Junior high school Primary school Not educated
0.16 (0.04) 0.50 (0.07)
0.13 (0.07) 0.22 (0.06)
0.06 (0.02) 0.27 (0.05)
36
0.28 (0.06)
0.58 (0.10)
0.51 (0.02)
140
0.05 (0.03) 0.00 0.00
0.08 (0.02) 0.00 0.00
0.13 (0.04) 0.03 (0.01)
30
Yes
0.21 (0.04)
0.11 (0.05)
0.06 (0.04)
319
Party membership
109
3
Notes: Number of observations for gender is 318, for age is 316, for education is 318, for Party membership is 319. Standard errors are in parenthesis.
9 percent said they were distant relatives. So roughly 40 percent of the managers we interviewed were either relatives or friends of the chief founder (the remaining 5 percent are chief managers themselves, as respondents). Indeed, half of the managers we interviewed said that members of their close family or more distant relatives worked in the same firm. So these are family enterprises, presumably using family members as managers, which helps to solve the moral-hazard problem of managers shirking and not necessary representing the family interests – though this may be at the expense of using less qualified managers. Reflecting the proximity of family members, 24 percent of the managers are actually from Wenzhou, and only 30 percent come from outside Zhejiang province. However, of those managers who migrated to take their jobs, only half brought their families with them (Table 16.13). For skilled and production workers, the situation is a little different. A higher proportion of these workers used friends to find their jobs, roughly 50 percent in each case (Table 16.12). However, between 37 percent and 43
9780230_274587_17_cha16.indd 282
6/23/2010 5:05:43 PM
Private Sector Industrialization in China
Table 16.12
283
How did you find the job in this enterprise? Managers
Skilled workers
0.31 (0.07) 0.04 (0.02) 0.15 (0.07) 0.10 (0.05) 0.10 (0.03) 0.01 (0.01) 0.03 (0.03) 0.26 (0.09) 1.00
0.52 (0.05) 0.33 (0.10) 0.04 (0.02) na
0.47 (0.07) 0.39 (0.10) 0.04 (0.03) na
na
na
0.04 (0.03) 0.01 (0.01) 0.06 (0.03) 1.00
(0.02)
Help from friends Want ad or labor market AD from enterprise By chance The enterprise found me Headhunter/agency Government Other Total
Production workers
na 0.04 (0.01) 1.00
Notes: There are 73 observations. Standard errors are in parenthesis. N/A is not applicable. The corrected Pearson chi-squared test of equality across groups of industries has value 0.67 and P-value 0.58, and across groups of employment size has value 1.49 and P-value 0.26.
Table 16.13
Migration Worker type
Do you have a local Hukou? If you are a migrant, where do you come from?
Did your whole family migrate?
Manager
Skilled worker
Yes
na
Other county Other city
na
0.23 (0.09) na
Other province Yes
na na 0.55 (0.07)
0.05 (0.03) 0.95 (0.03) 0.31 (0.09)
Production worker 0.14 (0.03) 0.03 (0.02) 0.06 (0.02) 0.92 (0.04) 0.17 (0.02)
Notes: Standard errors are in parenthesis. Number of observations is 94, 71, 79 respectively.
percent used ads, a higher proportion than for managers. Skilled and production workers used both formal ads and informal networks, in roughly equal measure, to find their jobs. Some 23 percent of skilled workers have a local hukou, while only 14 percent of production workers do (Table 16.13). A hukou is a local residence
9780230_274587_17_cha16.indd 283
6/23/2010 5:05:44 PM
284
John Strauss et al.
card, which usually indicates the place the holder comes from. Generally, if workers do not have a hukou in the place where they work, there are social benefits that they miss out on, although this is beginning to change in some areas. Having a nonlocal hukou is generally a good indicator that the worker is a migrant, so many such workers are migrants. Of the migrant workers, nearly all (92–95 percent) come from outside Zhejiang Province (Table 16.13). Most come alone: of production workers, only 17 percent say they came with family, while 31 percent of skilled workers say they brought family members with them (Table 16.13). Just under half of managers have formal contracts, the average length being about 5 years. For skilled workers, half also have contracts, of about 1.66 years duration. For production workers, only 38 percent have contracts, for somewhat less than 1 year each. For a firm, issuing contracts is costly because there are contractual obligations set by the central government which they are required to undertake if they give a contract to a worker. For workers, some return to their firm year after year, and some of those get contracts. So, signing repeated contracts over a period of years is not unusual, especially for production workers. For skilled workers it is less likely, but presumably there are more incentives for firms to give longer contracts to skilled workers, particularly if some of their skills are firm-specific. Even for industry-specific skills such as those of a shoe designer, if the worker is good a firm may want to both pay a high salary and provide a multiyear contract, to avoid such a skilled worker being “poached” by another firm. Yet firm owners constantly complained about tight labor markets for skilled workers, so the low percentage that are in repeated contracts may indicate that others have been “bidded away” by other firms. Monthly salaries by worker type are given in Table 16.14. Managers received just over 2400 yuan a month in 2005 (the RMB–US$ exchange rate in early 2005 was between 8 and 7.75). Production workers on average received a little under half the monthly salary of a manager, 1156 yuan, while skilled workers’ salaries were in between, at 1767 a month. In Wenzhou firms year-end Table 16.14
Income (yuan) 2005 and holding shares Worker type
Salary per month Year-end bonus Do you hold shares in this enterprise?
Yes
Manager
Skilled worker
Production worker
2418.81 (241.71)
1767.22 (134.02)
1155.77 (53.65)
262
6,681.56 (1933.88) 0.18 (0.08)
1794.98 (357.19) 0.05 (0.02)
415.66 (172.56) 0.03 (0.02)
252
Observations
314
Note: Standard errors are in parenthesis.
9780230_274587_17_cha16.indd 284
6/23/2010 5:05:44 PM
Private Sector Industrialization in China
285
bonuses are an important part of annual income, so need to be examined. As shown in Table 16.14, year-end bonuses represent over 2.5 months’ pay for managers, about one month’s pay for skilled workers, and a little over one-third of a month’s pay for production workers. Less than 20 percent of managers own shares in the firm, and almost no workers do (Table 16.14). Fringe benefits are quite important, particularly housing (Table 16.15). Two-thirds of production workers get housing benefits from the firms, usually in terms of dormitories that they can live in. For 56 percent of production workers housing is free, and for another 13 percent it is subsidized. While conditions vary, some recently built or refinished dorms that we saw included air-conditioners. For skilled workers and even managers, some 55–60 percent have subsidized or free housing from their firms. About 40–50 percent of workers get meals supplied at work, and between roughly 30 percent and 50 percent of all workers get their food at subsidized prices, or free (Table 16.15). Health insurance is provided for half of the managers, a quarter of skilled workers, and 17 percent of production
Table 16.15
Fringe benefits Worker type
Does the enterprise provide housing support?
Yes, free Yes, subsidized No
Does the enterprise provide meals? Does the enterprise provide health insurance for you?
Yes, free Yes, subsidized Yes, not subsidized No Yes, for free Yes, subsidized Yes, not subsidized No
Manager
Skilled worker
Production worker
0.50 (0.12) 0.05 (0.03) 0.45 (0.12) 0.41 (0.07) 0.09 (0.02) 0.05 (0.03) 0.45 (0.09) 0.29 (0.04) 0.11 (0.04) 0.09 (0.04) 0.51 (0.05)
0.52 (0.11) 0.06 (0.03) 0.42 (0.10) 0.32 (0.02) 0.07 (0.03) 0.11 (0.06) 0.50 (0.05) 0.16 (0.07) 0.07 (0.05) 0.03 (0.02) 0.74 (0.11)
0.56 (0.10) 0.13 (0.04) 0.32 (0.12) 0.22 (0.05) 0.06 (0.03) 0.12 (0.04) 0.60 (0.08) 0.11 (0.03) 0.04 (0.02) 0.02 (0.01) 0.83 (0.04)
Note: Standard errors are in parenthesis.
9780230_274587_17_cha16.indd 285
6/23/2010 5:05:44 PM
286
John Strauss et al.
workers.73 When provided it is generally free, or with a subsidy. Other fringe benefits are very limited, including transport subsidies, provision of education for children, and pensions. Annual leave is a common benefit, however. Managers get more fringe benefits than skilled or production workers, as one would expect. A key issue for these workers, particularly the migrant production workers, is what their wages might have been had they stayed in their provinces of origin. They report that in rural areas of the main provinces of origin for Wenzhou workers (according to the 2005 Intercensus data), mean monthly salaries are about half to one-third the level of our Wenzhou firms’ salaries for production workers. This comparison is for workers in the areas of origin who have primary or junior high school education, which, we saw in Table 16.11, were the most common levels of schooling of our production workers. So rural migrant workers in Wenzhou are earning far larger salaries, double to triple what they could earn in their provinces of origin. On the one hand, these salary differences must keep new migrants coming to the extent they are allowed by the government, and on the other hand, such migration must put downward pressure on local wages for production workers. This looks remarkably like a surplus labor story (Fei and Ranis, 1964; Lewis, 1954; Ranis, 2003). Unfortunately we cannot test this assumption since we do not have access to a local Wenzhou data series on salaries over time, or on labor productivity, to see if wages have been rising more slowly than labor productivity over time in Wenzhou (see Ranis, 1995).74 Still, the large volume of rural migrant labor, with wages far higher in Wenzhou than in their originating areas, strongly suggests a surplus labor story. Government price subsidies and other support We divide government support into several parts in Wenzhou, including ease of getting business licenses; ease of getting land, and price; financial aids and loans; import and export rights; and other preferential treatment at the time of firms’ establishment. In general, the local government did not provide direct subsidies to the firms, such as financial support or tax preferential treatment. However, the Wenzhou government did not hold back the development of private firms; most firms in Wenzhou could obtain lands, business licenses, and import and export rights easily. In fact, the local government’s policy was “Just let them develop by themselves.” Only 12 percent of firms report getting any preferential tax incentives when they started up. Firms in shoemaking and general equipment were more likely to have gained some tax breaks, as were larger firms (at the time of the survey). Firms were more likely to have received favorable tax treatment at start-up if they were founded prior to 1985. Only 4 percent of firms report any direct financial assistance, or loans from the government.
9780230_274587_17_cha16.indd 286
6/23/2010 5:05:44 PM
Private Sector Industrialization in China
287
Business licenses and permissions to export and import were easy to get, according to our firms. Land has been the one factor that firms have at times worried about getting at reasonable prices. Land was owned by the government, and firms could purchase from the government. The price of land charged by the government had a great impact on the costs of establishing new firms. We found that 19 percent of firms could get land from the government free, and 9 percent for less than the market price, while most of firms (72 percent) needed to pay the market price. In different locations, the enterprises faced different price for lands. Thirty-one percent of firms in Ruian (an urban county), 24 percent in Ouhai (a rural town), and 14 percent in Lucheng (downtown Wenzhou) got free land from the government, and the percentages paying “less than market price” were 20 percent, 17 percent, and 3 percent respectively. Except for Ruian, more than three-quarters of firms had to pay the market price.
16.5
Implications and further discussion
At first glance, Wenzhou industrialization to date seems to have many commonalities with the East Asian model, although the details vary between countries (see Ranis, 1995; World Bank, 1993). In both cases it is lowtechnology, labor-intensive industries that served as the initial basis of industrialization. It was small-scale family firms that were the source of the initial firms in Wenzhou, Taiwan, and Japan before World War Two, with financing coming initially from shareholders, family, and friends, with reinvested profits being important during the growth stages. The size distribution of firms is tilted towards small- and medium-sized firms, although firm growth has been going on, certainly for successful firms. So the community mechanism of contract enforcement based on social and family ties were certainly critical in the early growth stages of these firms, as emphasized by Hayami and Godo (2002). Bank and government financial support was not needed in these cases for firms to be able to start and grow. In particular any credit constraints, if binding at all, did not prevent the very rapid growth of industries in Wenzhou. There is doubtless a lot of firm start-up and closure, although we could not establish that with our data. No doubt the dynamics of firm start-up and closure offer many lessons to be learned. That is another exercise. Technologies were being scaled up over time, although in Wenzhou it was through the use of local technologies (or those imported many years earlier), upgraded through “learning by doing” on the factory floor. Wenzhou may have benefited even more from being a late starter than did Taiwan and Korea, at least as far as technology acquisition is concerned, making it easier to use and upgrade low-quality local machines. Human capital was improved, especially with the acquisition of skilled labor for designing products, but also over time with the increasing education of the workforce.75 The roles
9780230_274587_17_cha16.indd 287
6/23/2010 5:05:44 PM
288
John Strauss et al.
of government were somewhat different in the different cases. Certainly government reforms that gave rise to overall economic growth and price stability were important to the expansion of firms in the Wenzhou and East Asian cases. Overall the Wenzhou experience seems quite consistent with that of Taiwan in its early period of industrialization, with some similarities with Japan during the prewar phase of its industrialization. One important difference must have been the much larger size and the rapid growth of the Chinese economy, and therefore its domestic market, compared to other East Asian countries, allowing in principal for firm growth over a much longer period without exports, although as we have seen exports have played a major role in Wenzhou. In addition, China’s size has undoubtedly made a difference in allowing a much larger volume of surplus labor migrating into Wenzhou than was available in Taiwan or Korea. Nonetheless, today in Wenzhou entrepreneurs are wringing their hands over rising wage costs and wondering how long they will be able to stay in Wenzhou rather than moving to a more remote, lower-wage area of China, or even outside China. On the other hand, one must be cautious in noting that Wenzhou might be a special case within China because of its initial conditions in the early 1980s, its remoteness, and its not having a history of SOEs or TVEs in the area, as well as its special treatment by both local and national governments over this period. Perhaps Wenzhou also benefited from its recent history of entrepreneurship and from migrant networks throughout China, which can be considered as a form of community relations, though this is much more speculative. However, a very interesting recent study by Zhang et al. (2006) suggests that Wenzhou is not a special case within China, at least not for private-sector firms. They survey 1199 households in 60 villages in 6 provinces in China, examine sources of income, and collect detailed information on the self-employed. There are remarkable similarities between what we find for Wenzhou and Zhang et al. find throughout China, in terms of privatesector, family-business start-ups. Start-ups in their study are very small family firms, based on family finances without explicit government help, just like those in Wenzhou. Their study suggests that a special-case argument is incorrect.76 However, one important way in which the Wenzhou experience may not be generalizable is because of the nature of the industries that have arisen there, in particular their low-technology nature. Scholars such as John Sutton argue that for high-technology industries such as auto manufacturers or auto parts suppliers, quality is much more important, and he further argues that quality thresholds must be attained by such firms in order for them to be internationally competitive. In this case larger-scale firms may be necessary, along with government subsidies to get started. For instance, Sutton notes that the car parts supply industry in China has now virtually reached international quality levels (Sutton, 2004; see
9780230_274587_17_cha16.indd 288
6/23/2010 5:05:44 PM
Private Sector Industrialization in China
289
also Brandt et al., 2008). He further argues that a period of domestic protection with import substitution policies protected these firms for a time (fortunately the government realized when to begin to reduce the import protection). To the extent that Sutton is correct, there may be limited applicability of the Wenzhou experience to future industrialization in China, to the extent that more sophisticated industries start to play a larger role. Indeed, based on the historical experiences of other countries, particularly in East Asia we should expect the types of industries in Wenzhou to change in future, becoming more high-tech (Nelson and Pack, 1999). This may not happen overnight, as long as wages in Wenzhou stay low because of largescale immigration. However, as wages in Wenzhou rise lower-tech industries may move westward, to poorer parts of China with lower-wage rates, in which case the Wenzhou experience may still have some relevance. Indeed, as long as rural “surplus labor” exists in China, labor-intensive, low-tech industries will still likely be important in Wenzhou.
References Bigsten, A. and Gebreeyesus, M. (2007) “The Small, the Young and the Productive: Determinants of Manufacturing-Firm Growth in Ethiopia,” Economic Development and Cultural Change, 55 (4): 813–40. Brandt, L., Rawski, T. and Sutton, J. (2008) “China’s Industrial Development,” in L. Brandt and T. Rawski (eds.), China’s Great Economic Transformation (Cambridge: Cambridge University Press). Evenson, R. and Ranis, G. (1990) Science and Technology: Lessons for Development Policy (Boulder, CO: Westview Press). Fei, J. and Ranis, G. (1964) Development of the Labor Surplus Economy: Theory and Evidence (Homewood, IL: Irwin Press). Godo, Y. and Hayami, Y. (2002) “Catching-Up in Education in the Economic Catch-Up of Japan with the United States,” Economic Development and Cultural Change, 50 (4): 961–78. Hayami, Y. and Godo, Y. (2005) Development Economics: From the Poverty to the Wealth of Nations, 3rd Edition (Oxford: Oxford University Press). Kilby, P. (1971) Entrepreneurship and Economic Development (New York: Free Press). Kilby, P. and Liedholm, C. (1990) “The Role of Non-farm Activities in the Rural Economy,” in J. Williamson and V. Panshamukhi (eds.), The Balance Between Industry and Agriculture in Economic Development, Vol. 2 (Macmillan). Lewis, W. A. (1954) “Economic Development with Unlimited Supplies of Labor,” Manchester School of Economic and Social Studies, 22: 139–91. Liedholm, C. and Mead, D. (1987) “Small-Scale Industries in Developing Countries: Empirical Evidence and Policy Implications,” International Development Papers No. 9, Department of Agricultural Economics, Michigan State University. Liedholm, C. and Mead, D. (1999) Small Enterprises and Economic Development: The Dynamics of Micro and Small Enterprises (London and New York: Routledge). Liu, A. (1992) “The ‘Wenzhou Model’ of Development and China’s Modernization,” Asian Survey, 32 (8): 696–711. Nelson, R. and Pack, H. (1999) “The Asian Miracle and Modern Growth Theory,” Economic Journal, 109: 416–36.
9780230_274587_17_cha16.indd 289
6/23/2010 5:05:44 PM
290 John Strauss et al. Pack, H. and Westphal, L. (1986) “Industrial Strategy and Technological Change: Theory versus Reality,” Journal of Development Economics, 22: 87–128. Paulson, A. and Townsend, R. (2004) “Entrepreneurship and Financial Constraints in Thailand,” Journal of Corporate Finance, 10: 229–62. Ranis, G. (1995) “Another Look at the East Asian Miracle,” World Bank Economic Review, 9 (3): 509–34. Ranis, G. (2003) “Is Dualism Worth Revisting?” Economic Growth Center Discussion Paper No. 870, Yale University. Ruan, J. and Zhang, X. (2009) “Finance and Cluster-based Industrial Development in China,” Economic Development and Cultural Change, 58 (1), 143–64. Sonobe, T., Hu, D. and Otsuka, K. (2004) “From Inferior to Superior Products: An Inquiry into the Wenzhou Model of Industrial Development in China,” Journal of Comparative Economics, 32: 542–63. Sonobe, T. and Otsuka, T. (2006) Cluster-Based Industrial Development: An East Asian Model (New York: Palgrave Macmillan). Sutton, J. (2004) “The Auto-Component Supply Chain in China and India – A Benchmarking Study,” mimeo (London: London School of Economics). Westphal, L., Kim, L. and Dahlman, C. (1985) “Reflections on Korea’s Acquisition of Technological Capability,” in N. Rosenberg and C. Frischtak (eds.), International Technology Transfer (New York: Praeger). Wooldridge, J. M. (2002) Analysis of Cross Section and Panel Data (Cambridge, MA: MIT Press). Zhang, J., Zhang, L., Rozelle, S. and Boucher, S. (2006) “Self-Employment with Chinese Characteristics: The Forgotten Engine of Rural China’s Growth,” Contemporary Economic Policy, 24 (3): 446–58.
9780230_274587_17_cha16.indd 290
6/23/2010 5:05:45 PM
17 The Role of Markets for Managing Agricultural Risks in Developing Countries Peter Hazell
17.1
Introduction
Agricultural production is typically a risky business. Farmers face a variety of price, yield, and resource risks that make their incomes unstable and unpredictable from year to year. In many cases, farmers are also confronted by the risk of catastrophe. For example, crops may be totally destroyed by hurricane, fire, drought, pests, or diseases, and product prices may plummet because of adjustments in local or world markets. The type and severity of the risks confronting farmers vary with the farming system and with the agroclimatic, policy, and institutional setting. Agricultural risks are particularly burdensome to small-scale farmers in the developing world. Farm households and rural communities have, over many generations, developed a plethora of traditional methods and social capital for managing risk (Hayami, 2009). These methods work best for managing independent rather than covariate risks, and for losses that are small relative to household and community assets. For larger and covariate losses, rural communities have had to look for outside help. Historically, feudal and tribal authorities helped fill that need, but recent decades have seen the introduction of government-sponsored interventions such as price-stabilization schemes, crop-insurance programs, and drought relief. With few exceptions, these public interventions have proven expensive and inefficient, and many were removed during the market-liberalization policies of the 1990s. The vacuum that has been left has contributed to an explosion in publicly-funded social-safety-net programs to provide relief in times of need. These expenditures are not only increasingly costly as growing population pressure exacerbates losses from natural disasters, but they contribute little to economic growth and long-term poverty reduction. New hopes lie in market-mediated approaches to risk management, and recent years have seen considerable innovation and experimentation in this area. However, market-mediated approaches have not yet achieved the scale that is needed. 291
9780230_274587_18_cha17.indd 291
6/23/2010 12:23:12 PM
292
Peter Hazell
This chapter reviews the context for market-mediated interventions for handling covariate risks, which cannot be dealt with effectively by local communities. The chapter also discusses the more important types of instruments that have been tried, and draws lessons that can help guide future attempts to scale up market-mediated approaches.
17.2 Existing risk-management approaches Types of risk in agriculture Agricultural risks can usefully be characterized by the degree to which they are correlated across households within a community or region (ranging from independent to highly covariate), their frequency of occurrence, and the types and severity of the losses incurred – Table 17.1. At one extreme are highly covariate risks such as floods, hurricanes, and severe droughts, which occur with low frequency but have catastrophic impact within affected regions. In addition to short-term humanitarian challenges arising from loss of lives, production, and incomes, the associated damage of key assets can make recovery slow and uncertain. Without help, many people can end up in long-term poverty. At the other extreme, there is a wide array of risks that occur much more frequently and which are weakly, if at all, covariate. These risks impact more randomly on individual households, and often the share of households affected each year is predictable (for example, deaths and illnesses of people or livestock). Many of the risks that affect seasonal yields and production are due to localized weather and pest problems that affect pockets of households, or even just particular fields in a farm. Between these two extremes lies a variety of risks that occur with moderate frequency and which are moderately covariate, such as losses in production and income, damage to assets due to less severe droughts or excess rainfall at harvest, or market and price risks. Covariate risks can have widespread ramifications throughout a regional economy. Agricultural losses affect farm incomes, agricultural wage earnings, and food supplies, and they spill over into the rural nonfarm economy. Widespread debt defaults undermine rural financial systems. The loss of productive assets can have important and adverse dynamic impacts, pushing households into poverty from which it may be difficult to recover in subsequent years. How farmers and rural communities manage risk Farm households and rural communities pursue a number of well-honed strategies for reducing their exposure to risk, and for coping with losses when they occur. The first line of defense is on-farm risk management. This has been well documented, and includes risk-reducing and risk-coping strategies (e.g., Walker and Jodha, 1986). Risk-reducing strategies include: (a) diversification
9780230_274587_18_cha17.indd 292
6/23/2010 12:23:12 PM
Role of Markets for Managing Agricultural Risks 293
Table 17.1
Types of risk and losses, and the effectiveness of local capacity to manage
Type risk
Type loss Seasonal production/ income
Frequency
Degree covariation Life
Low
High
Widespread loss of life and injuries from catastrophic weather events like hurricanes, floods, or severe droughts. Little or no capacity to manage locally, and recovery is difficult and slow.
Widespread loss of homes and productive assets from catastrophic weather events. Little or no capacity to manage locally, and recovery difficult and slow.
Regional production and income impacts can be large from catastrophic weather events with limited local capacity to manage. Recovery can be slow if lives and assets are also lost.
Medium
Medium
Some loss of life and widespread health problems can arise from seasonal malnutrition. Moderate capacity to manage locally, and recovery occurs.
Widespread loss of animals from drought or contagious diseases. Moderate capacity to manage locally, and recovery slow. Some people fall into poverty traps.
Loss of production from droughts, new pest outbreaks, and animal diseases Regional production and income impacts can be widespread, with induced shrinkage of the rural nonfarm economy. Moderate capacity to manage locally, and recovery can be rapid unless assets are also lost. Some people fall into poverty traps.
High to medium
Low
Deaths, accidents, and illnesses that affect a predictable share of the population each year. Some local capacity to pool these risks, but recovery from losses can be slow for the households involved.
Loss/damage/disease for a predictable share of the total stock of homes and productive assets each year. Good local capacity to pool these risk, but recovery from losses can be slow for the households involved.
Low yields for some farmers due to a variety of localized weather and pest problems. Good local capacity to manage these risks, and recovery is usually rapid.
9780230_274587_18_cha17.indd 293
Assets
6/23/2010 12:23:12 PM
294
Peter Hazell
of crops and livestock; (b) intercropping and staggered planting dates; (c) fragmentation of the farm to include plots that have different risk exposures in the landscape; (d) investing in irrigation; and (e) diversification into nonfarm sources of income. Risk-coping strategies include: carrying food stocks, extra animals, and other liquid assets (e.g., jewelry, savings) that can be easily sold in times of need; borrowing credit; and temporarily migrating for employment. The second line of defense is risk-pooling within rural communities. Accumulated social capital in the form of religious groups, credit groups, and kin-support networks enables risks to be pooled by assisting individuals in times of crisis on a reciprocal basis (e.g., Hayami, 2009; Sommerfeld et al., 2002). In Ghana, village chiefs have traditionally managed community grain stocks (Keyzer et al., 2007). Sharecropping contracts have also emerged in many societies as a way of sharing risks between landlords and tenants (Newbery and Stiglitz, 1979; Otsuka and Hayami, 1993). In pastoral areas, reciprocal grazing arrangements between spatially dispersed communities enable mobile or transhumant grazing practices that reduce the risk of having insufficient forage in any one location (McCarthy et al., 1999). Limitations of traditional risk management Traditional risk-management strategies have proven effective in helping rural communities to survive for countless generations, including in many drought-prone environments. But they have several limitations (Table 17.1). On-farm diversification strategies mean that farmers cannot fully exploit their comparative advantage by specializing in the most profitable enterprises. This leads to a reduction in expected farm income, though also a more stable one. Studies of drought-prone areas in India and Burkina Faso suggest that farmers sacrifice 12 to 15 percent of average income through risk-avoidance strategies (Gautam, Hazell, and Alderman, 1994; Sakurai and Reardon, 1997). Farm- and regional-level studies using mathematical programming models show that the trade-off in average income increases exponentially with the level of risk aversion (Hazell and Norton, 1986). Farmers may also be reluctant to invest in improved seeds and fertilizers or to make long- term investments in trees and land improvements, because they cannot recover their costs and repay any credit they may have borrowed in bad years. Financial institutions are also less willing to lend to farmers when the risk of default is exacerbated by significant weather and market hazards. The problem is compounded by lack of suitable collateral that can be easily liquidated to meet debt payments in crisis years. Traditional risk-management arrangements frequently fail to provide an adequate safety net for the poor. Poor people have fewer assets to fall back on in hard times, and are more exposed to food-price increases, contractions in local employment opportunities, and wage levels affected by local production or market shocks. There is a growing literature showing that
9780230_274587_18_cha17.indd 294
6/23/2010 12:23:12 PM
Role of Markets for Managing Agricultural Risks 295
repeated income shocks and asset losses can conspire to keep poor households trapped in poverty. Credit, which might offer a viable pathway out of poverty, is also much less likely to be available to small farmers caught in poverty traps (Carter and Barrett, 2006). Traditional risk-management arrangements are generally better for managing independent risks that affect a few farmers at a time (e.g., accidents, fires, theft) than large and covariate risks that affect many people simultaneously. This limitation is greatest for catastrophic losses associated with extreme weather events, new pest outbreaks, or a collapse in market prices. Some of these events may also involve loss or damage of assets and sometimes loss of lives, making recovery more uncertain and slow. They can also have knock-on effects for income earnings in the rural nonfarm economy. The aggregate losses associated with catastrophic events can easily overwhelm community risk-sharing arrangements, dry up informal credit, reduce asset prices and agricultural wages, and force many people into poverty. Some of the most dramatic evidence of the failure of traditional riskmanagement arrangements in handling covariate risk comes from studies of drought. For example, detailed studies of the impact of droughts in Ethiopia (Webb and von Braun, 1994), Eastern India (Pandey et al., 2007) and South India (Hazell and Ramasamy, 1991) all show that in percentage terms, income losses can far exceed initial production losses because of a collapse in local agricultural employment and wages, nonfarm income, and asset prices. What is missing in most developing countries is a third line of defense in risk management, namely mechanisms to enable farmers and rural communities to pool large and covariate risks at more aggregate levels (see the first row of Table 17.1).
17.3
Past experience with public interventions
Recognizing the limitations of traditional risk-management institutions, many governments have intervened with a range of risk-management programs for farmers, including crop insurance, price stabilization, livestockfeed subsidies, and emergency relief. Unfortunately, most have proved an expensive drain on the public purse with little evidence to show that they have generated any sizeable net social benefits. We review some of these experiences below. Crop insurance Publicly-provided crop-yield insurance has been implemented in many countries (Hazell et al., 1986). Most programs have been multiple-peril or all-risk programs. Indemnities are typically based on the difference between actual yield and a prespecified target yield, not on actual crop damage or input costs lost. Setting the target yield has proven difficult given poor or
9780230_274587_18_cha17.indd 295
6/23/2010 12:23:12 PM
296 Peter Hazell
nonexistent historical records for individual farms. In many countries the insurance has been tied to loans from a state-operated agricultural development bank, with the bank paying the premium and collecting the indemnities on behalf of its borrowing farmers. In some countries the insurance has been made compulsory for all farmers growing specific crops or borrowing from a state bank. The financial experience with publicly-provided, multiple-peril crop insurance has been ruinous. Programs have had to be heavily subsidized, with governments paying part of the farmers’ premium and a share of the administration costs. Often the government has also had to contribute to the indemnity payment in years when the insurer had insufficient reserves. To make a profit without government support, an insurer needs to price its contracts so that over time the total premiums collected from farmers will exceed the total indemnities paid plus all administrative costs. Hazell (1992) analyzed the experience with public crop-insurance programs in 7 countries with 5 or more years of available financial data, and found that the total payouts exceeded the premiums collected from farmers by a factor ranging from 2.4 to 5.7. This implies that governments had to subsidize 40 to 80 percent of the total cost. The total cost to governments ranged from $10 to $408 per insured hectare in 1987 prices. Despite these high costs, there is little evidence to show that crop insurance had any positive impacts on agricultural lending, agricultural production, or farm income. For example, social cost-benefit analyses of the Mexican and Japanese schemes show negligible social returns in relation to their high costs (Hazell et al., 1986; Tsujii, 1986). Pomareda (1986) found that a small increase in interest rates would have been just as beneficial to the agricultural development bank in Panama as the compulsory cropinsurance program for its borrowers. Crop insurance, when heavily subsidized, can even have important negative social impacts by encouraging farmers to grow unsuitable crops in risky environments. Why has public crop insurance failed so badly? There are several prominent reasons: ●
●
Many of the risks covered by multiple-risk insurance were open to moral-hazard problems. Since there was rarely any attempt to limit the kind of losses covered or to attribute cause, farmers could receive compensation even when part of the damage was due to their own negligence. Lack of data to quantify yield risks at farm level also led to unrealistic setting of yield targets and inflated estimates of crop losses. These conditions led to greater losses than necessary, increased the risk exposure of the insurer, and made “fair” actuarial calculations of those risks almost impossible. Premium rates were typically fixed by government and at the same rate for large groups of farmers, leading to selection problems.
9780230_274587_18_cha17.indd 296
6/23/2010 12:23:12 PM
Role of Markets for Managing Agricultural Risks 297 ●
●
●
●
Public insurers were often mandated to extend their insurance to small farms, and this added significantly to their administration costs. Inappropriate incentive problems arise within insurance institutions when the government underwrites most of their programs. When insurers know that the government will automatically cover most losses, they have little incentive to pursue sound insurance practices when assessing losses. In fact, they may find it profitable to collude with farmers in filing exaggerated or falsified claims. Hazell (1991) reports that in Mexico prior to the closing of the national agricultural insurance agency, it was not uncommon for inspectors to receive bribes of about 30 percent of the value of the indemnity payments made to farmers. When the insurer underwrites the loans of an agricultural development bank these incentive problems can easily infect the bank too, leading to a serious loss of discipline in banking practices. Why, for example, should bank staff try to collect loans from tardy borrowers if they can more easily obtain repayment from the state insurer? Governments have undermined public insurers for political reasons. Hazell (1991) gives examples where insurers have had to pay out against exaggerated losses in election years. Also, governments have sometimes undermined their crop-insurance programs by providing direct assistance to producers in disaster areas (Goodwin and Smith, 1995). Many crop-insurance programs also became too specialized, focusing on specific crops, regions, and types of farmers, particularly when the insurance was tied to the loans of an agricultural development bank that had a mandate to serve particular target groups identified by the government. Without a well-diversified insurance portfolio, crop insurers are susceptible to the covariate risk problem, and face the prospect of sizeable losses in some years.
Many of the public crop-insurance programs in developing countries were phased out in the 1990s as part of market-liberalization policies. Private crop insurance has since grown in a number of developing countries, but to avoid the pitfalls of the public programs, private insurers typically only offer contracts against specific perils (e.g., hail or frost damage), and sell mostly to commercial farmers growing higher-value crops (Gudger, 1991). Price stabilization Price stabilization affects food prices for consumers as well as farm-gate prices for producers, and hence has been widely implemented by governments as much to protect consumers from price spikes as to protect farmers from market-price collapses. Like public crop insurance, price-stabilization schemes managed by state agencies became very expensive. This was partly because of high administration and storage costs, but also because of built-in subsidies when
9780230_274587_18_cha17.indd 297
6/23/2010 12:23:12 PM
298
Peter Hazell
border prices were distorted to favor consumers or producers. Many pricestabilization schemes were phased out as part of the market-liberalization programs of the 1990s. The surviving schemes are nearly all in Asia, where governments face stiff political resistance to their removal from entrenched interest groups. Key problems that contributed to the demise of price-stabilization schemes were as follows: ●
●
●
It is difficult to set the price band within which prices are to be stabilized. Structural changes in production technology, consumer demand, and world markets need to be recognized or markets cease to efficiently guide resource allocation within the agricultural sector. Price bands are open to political manipulation. In Africa, many marketing boards used price-stabilization schemes to tax farmers (Newbery, 1989). In Asia, price-stabilization schemes were often used to improve the terms of trade for farmers or to offset macroeconomic biases against agriculture, at least during the Green Revolution era (Cummings et al., 2006). When stabilization programs distort average or trend prices for consumers, producers, or both, they soon lead to high subsidy costs and to economic inefficiencies by sending farmers and consumers the wrong market signals. Market-price interventions driven by social and political agendas undermine private storage and trade.
Feed subsides Feed subsidies have been an important public intervention for managing drought risks in countries with significant pastoral farming systems. In the west Asia and north Africa (WANA) region, the programs have been widely used to provide supplementary feed, mostly barley, to safeguard livestock in drought years (Hazell et al., 2001). These programs have been quite successful in protecting livestock numbers and production during droughts, but they have also encouraged unsustainable farming practices. In particular: ●
●
They have accelerated rangeland degradation in the long term by undermining the traditional process of adjusting flock size to interannual climatic variations. Herd sizes have increased sharply since the introduction of feed subsidies, and grazing practices have changed so that many of the animals no longer leave the rangeland areas during the dry season but have their feed and water trucked in. This practice leads to overgrazing during the dry season, reduces the natural seeding of annual pasture species, disturbs the soil, and contributes to wind erosion, particularly in areas near water and feed supply points. They have led to high government procurement prices for barley that has encouraged the mechanized encroachment of barley cultivation
9780230_274587_18_cha17.indd 298
6/23/2010 12:23:12 PM
Role of Markets for Managing Agricultural Risks 299
onto rangeland areas, where it causes serious soil erosion and cannot be sustained. Although typically introduced as a relief measure in severe droughts, once established feed subsidies have tended to become permanent and expensive to governments, and they have proved difficult to target, with the lion’s share of the subsidized concentrates going to large herders and to commercial farms. Total costs became high, and they were scaled back in most countries as part of market-liberalization policies. Relief programs Many governments have found it necessary to provide direct disaster assistance to relieve the problems of rural areas stricken with catastrophic losses caused by natural hazards like drought, flood, and hurricane. For many small, risk-prone countries such government assistance can be extremely costly, and may represent a high percentage of GNP when the disaster is large. These costs may escalate in the future as more people live in vulnerable areas, and as global climate change increases the frequency and severity of many natural disasters. Relief programs are driven by humanitarian rather than development agendas and their primary value is in saving lives and rebuilding assets. However, they have run into a number of problems: ●
● ●
●
It is difficult to target relief aid to the truly needy, and large leakages to others are common. Emergency relief often arrives too late. Relief can distort incentives for development; for example, food aid can depress local prices for farmers. Once disaster assistance has been institutionalized and people know they can count on it, it has many of the longer-term effects of an insurance subsidy that inadvertently worsens future problems by encouraging people to increase their exposure to potential losses. For example, compensation for flood or hurricane damage to homes can lead to the building of more houses in flood- and hurricane-prone areas.
General lessons from past public interventions A common problem with subsidized public-risk management interventions is that they distort economic incentives. Subsidies for risk management have similar effects to those of subsidies for any other input: they encourage overuse of that input. In this case the “overuse of the input” is the adoption of farming practices and livelihood strategies that lead to a growing dependence on government assistance. For example, compensation for crop losses in drought-prone areas encourages farmers to grow more of the compensated
9780230_274587_18_cha17.indd 299
6/23/2010 12:23:12 PM
300 Peter Hazell
crops. Feed subsidies have encouraged the encroachment of barley production onto the steppe in the WANA region. Governments have tried to manage too much risk. It is simply too costly to try and pool all the price, yield, or income fluctuations confronting farmers and rural communities. Many of these risks are too open to moral hazard and asymmetric information problems, and occur with too high a frequency to be insurable at affordable cost. Most households and rural communities have already proved that they are quite capable of managing independent risks and small covariate risks. It is the catastrophic risks in the lower tail of the loss distribution that are more problematic and need to be addressed. There has been inadequate differentiation of the needs of vulnerable people who cannot afford to insure from the more general population who can. This is equivalent to a failure to separate a social from a development (or growth) agenda. Many households are quite capable of bearing and managing most of the risks they face if suitable commercial risk-management instruments are available and, except in the event of major disasters, they do not need access to subsidized government interventions. But there are vulnerable households who do not have this capacity, and for whom all riskmanagement interventions serve primarily as social safety nets. The latter can only be provided on a heavily subsidized basis. Mixing these two types of needs leads to the design of public interventions that are heavily subsidized for all, and which end up being very costly both in terms of their direct government cost and the economic inefficiencies they induce through distorted incentives. If the needs of these two types of households can be differentiated, then more efficient and targeted instruments can be designed. Recent years have seen two important developments. The first has been the termination or downsizing of many heavily subsidized, public riskmanagement interventions. The second has been the rapid growth in the use of safety-net programs, with real advances in their design and implementation to improve targeting and reduce costs (Grosh et al., 2008). These developments have provided the opportunity for a more rational approach to risk management in agriculture, assigning the social agenda to the safetynet programs and allowing risk-management interventions to focus on the development agenda. This has spawned a great deal of interest in marketmediated approaches to risk management in agriculture.
17.4 A role for markets to manage risk In countries with well-developed financial systems, the banking system offers an efficient way for many farmers to cushion interseasonal shocks in income and consumption. But banking services are much less accessible to farmers in developing countries, and credit for consumption purposes is limited to traditional sources that are costly and suffer from the covariate risk problem.
9780230_274587_18_cha17.indd 300
6/23/2010 12:23:12 PM
Role of Markets for Managing Agricultural Risks 301
Alternative market-based instruments are therefore needed that deal more directly with important price and production risks. The most promising developments so far are area-based index insurance and futures markets. Area-based index insurance Area-based index insurance is an attractive instrument for managing the kinds of covariate and catastrophic risks that local communities cannot manage on their own. The essential principle of index insurance is that contracts are written against specific perils or events which are defined and recorded at regional levels (e.g., drought measured at a local weather station). To serve as agricultural insurance, the index should be defined against events that are highly correlated (on the downside) with regional agricultural production or income. For example, an insured event might be that rainfall during a critical period of the growing season falls 70 percent or more below normal. All buyers in the same region are offered the same contract terms per dollar of insurance coverage. That is, they pay the same premium rate and, once an event has triggered a payment, receive the same rate of payment, and their total payments and indemnities would be that rate multiplied by the value of the insurance coverage purchased. Payouts for index insurance can be structured in a variety of ways, ranging from a simple zero/one contract (once the threshold is crossed, the payment rate is 100 percent), through a layered payment schedule (e.g., a one-third payment rate as different thresholds are crossed), to a proportional payment schedule. Area-yield insurance was the first type of index-insurance product to be developed for agriculture. The insurance is written against the average yield for a county or district as measured through official crop cuts, and a payment is made whenever the officially reported yield for the region falls below some predefined yield (say, 90 percent of normal). Area-based yield insurance requires long and reliable time series of area-yield data, and this kind of data is not widely available in developing countries. The need to wait for official yield reports after the harvest also leads to delays in compensating farmers (in India the delays have become a serious problem: see Hazell and Skees, 2006). Hence alternative indices based on weather data have become more attractive, such as area rainfall or temperature. Weather data are much more widely available and are collected on a regular basis at large numbers of local weather stations. Advantages Area-based index insurance has a number of attractive features: ●
Because buyers in a region pay the same premium and receive the same indemnity per unit of insurance, it avoids all adverse-selection and moralhazard problems.
9780230_274587_18_cha17.indd 301
6/23/2010 12:23:12 PM
302 ●
●
●
●
Peter Hazell
It can be inexpensive to administer, since there are no individual contracts to write, no on-farm inspections, and no individual loss assessments. It is also relatively easy to market. At the extreme, contracts could even be sold rather like travelers’ checks or lottery tickets to any willing buyer. The insurance could in principle be sold to anyone. Purchasers need not be farmers, and the insurance could be attractive to anybody in the region whose income is correlated with the insured event, including agricultural traders and processors, input suppliers, banks, shopkeepers, and laborers. It is easy for the private sector to run, and might even provide an entry point for private insurers to develop other kinds of insurance products for rural people. As long as the insurance is voluntary and unsubsidized, it will only be purchased when it is a less expensive or more effective alternative to existing risk-management strategies. If the insurance is offered in small denominations it could also appeal to poor people.
Challenges Index-insurance faces a number of challenges: One challenge is generating sufficient demand for a sustainable insurance market to emerge. Clearly, the greatest potential will lie in regions where weather-related risks are the dominant risks confronting farm households. Index insurance also needs to be affordable compared to viable alternatives for managing risk. Studies in drought-prone areas have demonstrated that farmers are often willing to pay 12 to 20 percent above the pure-risk cost for drought insurance (Binswanger and Sillers, 1984; Gautam et al., 1994; McCarthy, 2003; Sakuri and Reardon, 1997). However, the total cost of insurance can become prohibitive if the probability of the insured event is too large. As a practical rule of thumb, events that occur more frequently than 1/7 may be too costly for most farmers to insure without a subsidy. Basis risk can also be a deterrent to demand. This is the problem that arises if an individual suffers a loss but is not paid because the major event trigging a payment for the region has not occurred. With index contracts it is also possible for an individual to be paid when they suffer no losses. Index contracts essentially trade off basis risk for transaction costs, and the insurance will not be attractive if the basis risk becomes too high. Basis risk can be reduced by limiting the insurance to severe or catastrophic weather events that affect all people in a region. Individual losses are then much more likely to coincide with events recorded at the local weather station. Another way to reduce basis risk is to increase the number and dispersion of weather stations, so as to better capture spatial variation in climatic conditions in written contracts. However, adding weather stations is expensive and they come without site-specific historical records. This has led to interest in new types of indices that can be assessed remotely with satellites, such as cloud
9780230_274587_18_cha17.indd 302
6/23/2010 12:23:13 PM
Role of Markets for Managing Agricultural Risks 303
cover or soil moisture content, for a chosen region during critical agricultural periods. This kind of data is becoming increasingly available and may prove the wave of the future. But its current use in writing contracts for individual households is constrained by a credibility problem; people may not trust financial decisions made by others on the basis of “unseen” data that may, because of basis-risk problems, fail to correlate highly with their own on-the-ground observations. It may be more relevant for the present to writing index contracts enabling lending institutions to cover their own portfolio risks. Other government interventions such as subsidized crop insurance, bank credit guarantees, or relief programs may crowd out demand for index insurance. While a high covariance risk reduces basis risk and hence increases the attractiveness of the insurance to buyers, it presents a serious problem for the insurer. This is because all those who have purchased insurance against the regional index must be paid at the same time. The insurer can hedge part of this risk by diversifying their portfolio to include indices and sites that are not highly and positively correlated, but it may also be necessary to sell part of the risk in the international financial or reinsurance markets. International reinsurance is already available for some kinds of natural disaster risk, though its high cost remains an impediment to its use for developing countries. This cost can be reduced by having government play a role in aggregating risk within a country, and insuring part of the aggregate risk itself before going to the global reinsurance market. The need to go to the global market could also be reduced if international development banks, such as the World Bank or others in the capital markets, provide governments with contingency loan arrangements so that the government would have sufficient capital to pay all losses in the bad year. This kind of arrangement could be especially valuable during the early years of an insurance program, before an adequate financial reserve and performance record has been built up. Experience India has had a government-backed area-yield insurance scheme since the early 1980s (Hazell and Skees, 2006). Although the program is generally well designed and managed by a public-insurance agency (the Agricultural Insurance Company of India, AICI), government intervention in premiumsetting has led to a situation in which the insurance is not financially sustainable without hefty government subsidies (the average loss ratio between 1985 and 2002 was 5.0, implying an 80 percent subsidy). Not surprisingly, the private sector does not offer area yield insurance in India. Weather-index insurance has been piloted in a number of developing countries, but the number of farmers that have been reached remains small, often less than 1 percent of the target populations. The following examples illustrate the experience so far.
9780230_274587_18_cha17.indd 303
6/23/2010 12:23:13 PM
304 Peter Hazell
Table 17.2 Financial performance: BASIX weather-insurance portfolio 2003–4 2004–5 Number of customers Claims settled % of farmers who received claims Premium collected (Rs) Claim amount paid (Rs) Loss ratio (%) Weather stations
2005–6
2006–7
230
402
6689
11,716
154 67
319 79
864 13
2379 2
2007–8 4545
2008–9
Cumulative
10,604
34,186
537 793(due) 12 7.5
5046 15
88,685
824,681 1,703,098 1,430,171 1,539,175 2,098,638 6,145,273
41,860
471,485
47 1
57 5
950,000 2,063,160 56 36
144 50
298,922 20 45
470,671 22 40
3,997,176 65
Source: Presentation by Vijay Mahajan (BASIX) at a WFP/IFAD/BASIX workshop on “The Current State of Practice in Index Insurance,” Hyderabad, India, November 18, 2008.
In India, a private insurer (ICICI Lombard) has partnered with a microfinance organization (BASIX) that serves as an intermediary for promoting and selling the insurance to its borrowers. The financial performance of the insurance has been favorable to date (Table 17.2), with an accumulated loss ratio since 2003/04 of 0.65. But BASIX has only sold 34,000 policies so far, and seems to be having trouble scaling the product up. Emergent problems from this experience include: the need to establish additional private weather stations to reduce basis risk; low numbers of contract sales (and hence high costs per sale) against each weather station; and the high cost of international reinsurance. Scaling up weather insurance in India is also challenged by competition from the heavily subsidized crop-yield insurance program, which is compulsory for all farmers borrowing from the state banks. In Malawi, the Insurance Association of Malawi offers weather-index insurance to small farmers to insure loans from local banks for buying improved groundnut seed (Skees and Collier, 2008). A crop-yield simulation model is used to structure the rainfall-insurance payouts to ensure a high correlation with yields. The product has been piloted in four areas, and in the 2005–6 growing season, 892 farmers purchased weather insurance for a total sum insured of $35,000. In the 2006–7 growing season farmer uptake increased to 1710 groundnut farmers and a rainfall-based insurance contract was also purchased by some 826 farmers for maize production. No indemnities have been required so far. One difficulty for the program is that some groundnut farmers with uninsured loans have been allowed to default on part of their loans when the proceeds from the sale of their crop were too low, and this has reduced incentives for farmers to buy the insurance (Giné and Yang, 2009).
9780230_274587_18_cha17.indd 304
6/23/2010 12:23:13 PM
Role of Markets for Managing Agricultural Risks 305
Mongolia has recently launched a pilot-index insurance program to protect herders against high livestock losses due to severe winters. The insurance is sold by private insurance companies, and was bought by 2400 herders (10 percent of the eligible herders) in 2006, the first pilot year. The index is based on county-level livestock mortality rates that are collected by the national statistics office. Though the index is based on livestock mortality and not on a specific weather event, the major underlying cause of large livestock losses is severe winter weather. Some more general lessons emerging from these and other pilots are as follows: ●
●
●
The insurance is often tied to farmers, individual crops, or farm credit. This greatly narrows the type of households that can purchase the insurance, and when tied to the yields of specific crops may even make the insurance unattractive to many farmers. Many small farmers are diversified, producing a mix of crop and livestock products and often having significant off-farm sources of income. In this context losses in household income that affect debt repayment capacity may not highly correlate with the yields of individual crops, and hence would not be improved much by insurance tied to specific crop-yield outcomes. Unfortunately, the insurance regulations prevailing in many countries require a proven business exposure in any insurance contract and therefore make it difficult to market insurance for more general purposes. Insurers can cut transaction costs by using an intermediary organization with an existing delivery infrastructure, such as a microfinance organization (e.g., BASIX in India). Linking to a finance organization can also enable the insurance to be linked directly to credit, and for credit to be provided to finance the insurance premium. There may be other types of useful intermediary in the market chain, such as fertilizer distributors and seed companies, that could sell insurance to cover the costs of their products. These kinds of options may be even greater in high-value market chains where processors or retailers are willing to contract with farmers and provide credit and other inputs. Farm households sometimes have access to season-specific information that enables them to forecast insured weather events with greater than normal probability. For example, in India the date of onset of a monsoon has been shown to be a reasonable predictor of the season’s rainfall pattern, and increasingly accurate seasonal-climate forecasts are available from meteorology services. In Mongolia, the condition of the summer pastures gives some indication of the vulnerability of livestock in the coming winter (Skees, 2000). It is therefore critical to either establish closing dates for selling insurance that are far enough in advance to preempt the value of such forecast information, or to adjust the premium rates on a dynamic basis. More difficult problems can arise in
9780230_274587_18_cha17.indd 305
6/23/2010 12:23:13 PM
306 Peter Hazell
areas that are impacted by cyclical weather patterns. For example, in Peru a rise in sea temperatures can be used to predict the onset of the El Nino Southern Oscillation some seven to eight months before it occurs. In extreme cases, insurance might have to be sold on a multiyear contract basis. Although the private sector is playing a key role in many of the new indexinsurance programs, still the schemes have been initiated by the public sector or international players such as the World Bank. It appears there are serious setup problems that need to be overcome, and private insurance companies are not able to do this on their own. If index insurance is to scale up, governments will need to intervene more actively by playing important facilitating roles. According to Skees and Collier (2008), these roles include: ●
●
●
●
●
Collecting, maintaining, and archiving the data needed. These data should be made available for public use and for use by those with commercial interests who wish to develop innovative weather-insurance products. Increasing the number of weather stations at which such data are collected is also an important need in many developing countries. Establishing a legal and regulatory environment for enforceable contracts that both buyer and seller can trust. Additionally, laws and regulations need to be consistent with international standards to improve the chances of insurers for gaining access to global markets for risk transfer. Unfortunately, in many countries, laws and regulations are simply not in place to accommodate the development and use of weather-insurance products. Educating farmers about the advantages and disadvantages of indexinsurance products. Supporting product innovation. One of the challenges associated with private-sector development of new financial products is the ease with which they can be copied and replicated by others. This “free-rider” problem discourages many companies from making initial investments in new product development, especially in underdeveloped markets. Thus, some level of government and/or donor support for product development can be justified. These investments should be targeted at feasibility studies, and at developing pilot tests of new products with the involvement of local private-sector partners. Underwriting part of the insurer’s exposure in the early years. Until a sufficient volume of business has been established to attract global reinsurers, extreme losses for the insurance pool may need to be underwritten by government and/or donors, perhaps through contingent loan arrangements. For example, the World Bank has a provided a contingent loan arrangement for the Mongolian livestock insurance scheme.
9780230_274587_18_cha17.indd 306
6/23/2010 12:23:13 PM
Role of Markets for Managing Agricultural Risks 307
Futures and options markets The removal of many public price-stabilization schemes as part of marketliberalization programs has increased exposure to price risk for many developing-country farmers. Although producer prices typically have high coefficients of variation (often 20 to 30 percent or more), it is the downside risks that are of greatest concern. Forward-pricing arrangements are primarily a way of hedging against downside risk. There are two market mechanisms for hedging price risks. One is contract farming, where the buyer provides the farmer with a forward price contract early in the growing season. Essentially the buyer assumes the price risk for the grower. Forward contracting is most likely to be found in high-value market chains where buyers have clear reasons for wanting to source their supplies: for example, high-value export crops like flowers and vegetables where the buyer needs assured quality and reliable supplies. These kinds of market chains have grown rapidly in recent years but there are not many studies of their effectiveness in sharing risk, or of the number of farms that have benefited. One exception is a study by Ramaswami et al. (2006) who find that contract farming for poultry meat in India does lead to significant reductions in the coefficient of variation of income for contract farmers. Futures and options markets are a less individualized way of enabling farmers to lock into forward prices. Since there are not many futures markets operating around the world, their use is constrained by basis risk. Hedging price risks for an export crop may be attractive because the basis risk can be low. This is because the farmer is selling the same product in parallel and highly linked world spot and futures markets and, since both transactions are undertaken in the same currency, there is no foreign-exchange-rate risk (Morgan et al., 1999). On the other hand, incentives for hedging the price of nonexport crops in a world futures market are much less favorable. In this case there may only be a low correlation between the domestic price of the commodity and the foreign futures price, and the two transactions are carried out in different currencies leading to an exchange-rate risk. These problems can only be adequately resolved by having a local futures market. Morgan et al., (1999) discuss the requirements for setting up local commodity and futures markets. Few farmers use futures and options markets, even in industrial countries. In the US, only about 10 percent of farmers participate directly in futures and options markets (Tomek and Peterson, 2001), and the level of participation is even lower in Australia (Simmons, 2002). Although the cost of futures trades are not particularly high (about 2 percent), access to those markets is a problem, especially if the trading is to be done overseas. Individual farmers need marketing intermediaries who can trade on their behalf. In Costa Rica, coffee exporters serve as intermediaries. They offer farmers forward contracts and then hedge the risk in the New York coffee futures market (Myers, 1993). This way the farmer may not even know that his/her price
9780230_274587_18_cha17.indd 307
6/23/2010 12:23:13 PM
308 Peter Hazell
risk has been transferred to the New York market; all he or she sees is a forward-contract price with the mill or the exporter. Similar arrangements are possible for other high-value export crops. It seems unlikely that marketing intermediaries can be found for commodities like food staples sold in domestic markets, until such time as national or regional futures/options markets for those commodities have been developed.
17.5 Conclusions This chapter analyzed the issue of managing covariate risk in agriculture, which cannot be dealt with by the community mechanism of contract enforcement, as all the community members suffer from the same risk so that risk cannot be shared effectively by them. Since the withdrawal of many public-risk management schemes, there has been much hope that market-mediated solutions to risk management would emerge to handle the kinds of covariate yield and price risks that farmers and rural communities cannot manage on their own. This has led to considerable innovation and the piloting of new approaches, particularly with index-insurance and forward-pricing arrangements. These are still early days, but it is becoming clear that market-mediated approaches to risk management are not achieving anything like the scale needed. There has also been limited spontaneous development by the private sector, and governments or international agencies like the World Bank have had to initiate activities. This reluctance by the private sector seems related to the high basis risks associated with too few weather stations or futures exchanges; setup problems that private firms cannot easily overcome on their own; the need for marketing intermediaries to link farmers with insurers and futures traders; and the fact that many risk-management products are simply too expensive for many small farmers to afford. There is need for further product and institutional innovation, and also for a stronger public-sector role in helping to launch new programs. In particular, governments need to create more enabling regulatory environments, set up more weather stations and futures market exchanges, and – in the case of index insurance – provide a first line of reinsurance. If marketmediated solutions cannot be made to work better, many governments and donors are likely to become even more entrapped into providing expensive safety-net programs for rural people.
References Binswanger, H. P. and Sillers, D. A. (1984) “Risk Aversion and Credit Constraints in Farmers’ Decision-Making: A Reinterpretation,” Journal of Development Studies, 21, 5–21. Carter, M. R. and Barrett, C. B. (2006) “The Economics of Poverty Traps and Persistent Poverty: An Asset-Based Approach,” Journal of Development Studies, 42 (2), 178–99.
9780230_274587_18_cha17.indd 308
6/23/2010 12:23:13 PM
Role of Markets for Managing Agricultural Risks 309 Cummings, R., Jr., Rashid, S. and Gulati, A. (2006) “Grain Price Stabilization Experiences in Asia: What Have We Learned?” Food Policy, 31, 302–12. Gautam, M., Hazell, P. and Alderman, H. (1994) “Rural Demand for Drought Insurance,” Policy Research Working Paper 1383 (Washington, DC: World Bank). Giné, X. and Yang, D. (2009) “Insurance, Credit, and Technology Adoption: Field Experimental Evidence from Malawi,” Journal of Development Economics, 89, 8–11. Goodwin, B. K. and Smith, V. H. (1995) The Economics of Crop Insurance and Disaster Aid (Washington, DC: The AEI Press). Grosh, M., del Ninno, C., Tesliuc, E. and Ouerghi, A. (2008) For Protection and Promotion: The Design and Implementation of Effective Safety Nets (Washington, DC: World Bank). Gudger, M. (1991) “Crop Insurance: Failure of the Public Sector and the Rise of the Private Sector,” in D. Holden, P. Hazell and A. Pritchard (eds.), Risk and Agriculture: Proceedings of the Tenth Agricultural Sector Symposium (Washington, DC: World Bank). Hayami, Y. (2009) “Social Capital, Human Capital and the Community Mechanism: Toward a Conceptual Framework for Economists,” Journal of Development Studies, 45 (1), 96–123. Hazell, P., Bassoco, L. M. and Arcia, G. (1986) “A Model for Evaluating Farmers’ Demand for Insurance: Applications in Mexico and Panama,” in P. B. R. Hazell, C. Pomareda and A. Valdés (eds.), Crop Insurance for Agricultural Developments: Issues and Experience (Baltimore: Johns Hopkins University Press). Hazell, P. B. R. and Ramasamy, C. (1991) The Green Revolution Reconsidered: The Impact of High-Yielding Rice Varieties in South India. (Baltimore: Johns Hopkins University Press). Hazell, P., Oram, P. and Chaherli, N. (2001) “Managing Livestock in Drought-Prone Areas of the Middle East and North Africa: Policy Issues,” in H. Löfgren (ed.), Food and Agriculture in the Middle East. Research in Middle East Economics, Vol. 5 (New York: Elsevier Science). Hazell, P. and Skees, J. (2006) “Insuring against Bad Weather: Recent Thinking,” in S. K. Rao, S. Mahendra Dev and K. Subbarao (eds.), India in a Globalizing World, Essays in Honour of C. H. Hanumantha Rao (New Delhi: Academic Foundation). Hazell, P. B. R. (1992) “The Appropriate Role of Agricultural Insurance in Developing Countries,” Journal of International Development, 4 (6), 567–81. Hazell, P. B. R. and Norton, R. D. (1986) Mathematical Programming for Economic Analysis in Agriculture (New York: Macmillan). Keyzer M. A., Molini, V. and van den Boom, G. J. M. (2007) “Risk Minimizing Index Functions for Price-Weather Insurance, with Application to Rural Ghana,” Center for World Food Studies, SOW-VU Working Paper 07–02. McCarthy, N. (2003) “Demand for Rainfall-index-based Insurance: A Case Study from Morocco,” Environment and Production Technology Division Discussion Paper No. 106 (Washington, DC: IFPRI). McCarthy, N., Swallow, B., Kirk, M. and Hazell, P. (eds.) (1999) Property Rights, Risk, and Livestock Development in Africa (Washington, DC: International Food Policy Research Institute). Meyers, R. J. (1993) “Strategies for Managing Coffee Price Risks in Costa Rica,” in S. Claessens and R. Duncan (eds.), Managing Commodity Price Risks in Developing Countries, Chapter 6 (Baltimore: Johns Hopkins University Press for the World Bank).
9780230_274587_18_cha17.indd 309
6/23/2010 12:23:13 PM
310
Peter Hazell
Morgan, C. W., Rayner, A. J. and Vaillant, C. (1999) “Agricultural Futures Markets in LDCs: A Policy Response to Price Volatility?” Journal of International Development, 11, 893–910. Newbery, D. M. (1989) “Agricultural Institutions for Insurance and Stabilization,” in P. Bardhan (ed.), The Economic Theory of Agrarian Institutions, Chapter 14 (Oxford: Clarendon Press). Newbery, D. M. G. and Stiglitz, J. E. (1981) The Theory of Commodity Price Stabilization (Oxford: Oxford University Press). Otsuka, K. and Hayami, Y. (1993) The Economics of Contract Choice: An Agrarian Perspective (Oxford: Clarendon Press). Pandey, S., Bhandari, H. and Hardy, B. (eds.) (2007) Economic Costs of Drought and Rice Farmers’ Coping Mechanisms (Los Banos, Philippines: International Rice Research Institute). Pomareda, C. (1986) “An Evaluation of the Impact of Credit Insurance on Bank Performance in Panama” in Hazell, P. B. R., Pomareda, C. and Valdés, A. (eds.), Crop Insurance for Agricultural Development: Issues and Experience (Baltimore: Johns Hopkins University Press), 101–16. Ramaswami, B., Birthal, P. S. and Joshi, P. K. (2006) “Efficiency and Distribution in Contract Farming: The Case of Indian Poultry Growers,” MTID Discussion Paper No. 91 (Washington, DC: International Food Policy Research Institute). Sakurai, T. and Reardon, T. (1997) “Potential Demand for Drought Insurance in Burkina Faso and Its Determinants,” American Journal of Agricultural Economics, 79 (November 4), 1193–207. Skees, J. and Collier, B. (2008) “The Potential of Weather Index Insurance for Spurring a Green Revolution in Africa,” paper presented at the AGRA policy convening, June 23–25, Nairobi. Simmons, P. (2002) “Why do Farmers Have so Little Interest in Futures Markets?” Agricultural Economics, 27, 1–6. Sommerfeld, J., Sanon, M., Kouyate, B. A. and Sauerborn, R. (2002) “Informal RiskSharing Arrangements (IRSAs) in Rural Burkina Faso: Lessons for the Development of Community-Based Insurance (CBI),” International Journal of Health Planning and Management, 17, 147–63. Tomek, W. G. and Peterson, H. H. (2001) “Risk Management in Agricultural Futures Markets: A Review,” The Journal of Futures Markets, 21 (10), 953–85. Tsujii, H. (1986) An Economic Analysis of Rice Insurance in Japan (Baltimore: Johns Hopkins Press). Walker, T. S. and Jodha, N. S. (1986) “How Small Households Adapt to Risk,” in P. B. R. Hazell, C. Pomareda and A. Valdés (eds.), Crop Insurance for Agricultural Development: Issues and Experience, Chapter 2 (Baltimore: Johns Hopkins University Press). Webb, P. and von Braun, J. (1994) Famine and Food Security in Ethiopia: Lessons for Africa (Chichester, New York: published on behalf of the International Food Policy Research Institute [IFPRI] by John Wiley).
9780230_274587_18_cha17.indd 310
6/23/2010 12:23:13 PM
Notes 1. Mundlak and Strauss (1978) show that income differentials between the agricultural and nonagricultural sectors were important in determining the rate of migration from the agricultural sector in Japan. 2. See Ozaki (1972), Ozawa (1974), and Peck and Tamura (1976) for details. 3. Science and Technology Agency (1974 and 1991). 4. Science and Technology Agency (1990). 5. The size of the distortion effect, compared with other countries such as the US, is an open question. 6. According to Sakakibara (1997), participant firms perceived that the single most important objective of the R and D consortia was knowledge-sharing. 7. In order to assess the validity of the traditional view, Kuroda (1995) decomposes TFP changes into the scale-economy factor and the technological-improvement factor. He finds that 90 per cent of the TFP growth rate for the period 1960–90 is explained by technological improvements, and 10 per cent by scale economies. This implies that the introduction of high-yielding varieties is more important than mechanization in explaining agricultural TFP growth in postwar Japan. 8. “Norin” is the Japanese abbreviation of MAFF. 9. Infant mortality rate, life expectancy at birth, and literacy rate can be considered as basic indicators of human capital development. Telephone density and percapita electricity consumption can be considered as indicators of infrastructural development of the economy. 10. India is a federation of states with responsibilities divided between the states and the centre for development efforts. The third tier of government, the local governments, are now gradually emerging as significant independent entities. 1 P 11. The weighted coefficient of variation is calculated as follows: cv ⫽ Y ∑ (Y ⫺Y ) P where Pi refers to population of the ith state; P refers to the population of the country; Yi refers to per-capita GSDP; Y* refers to per-capita national income; and n refers to the number of states. 12. The authors are grateful to Keijiro Otsuka for discussion on this point. 13. It must be made clear that such variables, while explaining differences in growth pattern across states, may not be construed as having a causal relationship with growth. 14. In the Indian case, decentralization has come about more from the center than the states. Many states did not find the need to decentralize below their administrative level until the constitution was amended in 1992. Decentralization was initiated to provide easy access to public goods locally. There were 247,033 rural bodies known as panchayats and 3682 urban bodies in 2005. Nevertheless, the constitution grants strong powers to the central government, including the control of the central executive over state legislation, and the right to take over state administration in a state of emergency (Rao, 2005). 15. See H. G. Greel, 1964 for an in-depth description of the origin of the direct administrative county system (䚵ওࠊ) in China. 16. For details and other references, see Ma (2009a). *
n i
*
2
i
i
311
9780230_274587_19_not.indd 311
6/23/2010 12:23:20 PM
312
Notes
17. The problem of the absence of formal constraints against the emperor is succinctly summarized by Ray Huang’s study of Ming despotism: “Final authority rested in the sovereign, bureaucratic action was limited to remonstrance, resignation, attempted impeachment of those who carried out the emperor’s orders, and exaggeration of portents as heaven-sent warnings to the wayward emperor. When all these failed, there was no recourse left.” See Huang (1974, p. 7). 18. For Ming emperors’ brutal treatment of unruly bureaucrats, see Huang (1974, pp. 13–14). 19. The special status accorded to the class of people who passed various levels of civil service examinations produced a class of nonhereditary elites, the so-called gentry, who were part of the local elite that served the function of managing local affairs, often in collaboration with the magistrates and governors: see Chang (1955). Also see Ma (2009a) for references to this part of the historical narrative. 20. This China-centered hierarchical world view was internalized by peripheral states under strong Confucian influence such as Korea, Vietnam, and to a certain extent Japan. 21. See Ma (2009a) for details and data on taxes and expenditure. 22. See Liu’s study (2005) on per-capita tax rates across several benchmark periods since the Song dynasty. For the 1930s, government consumption was only 3.6 per cent of NDE: Liu and Yeh (1965, p. 68). 23. See Shi (2008, p. 68) and Iwai (2004, p. 32). The so-called Nei-wu-fu (ࡵݙᑰ) also collected its own tax revenue, which occurred outside the official balance sheet of the Board of Revenue. See Chang (1972). 24. See Wang (2005, p. 72, Table 4.2). It is very likely that Wang’s estimate is on the lower side, as he himself pointed out that the mid-eighteenth century represented the heyday of Chinese bureaucracy in terms of relative efficiency and probity. 25. For the surge of parliamentary legislation from the late seventeenth century that supported changes in the property rights regime, and the construction of public infrastructure such as roads and canals conducive to economic growth, see Bogart and Richardson (2008). For the contrasting path of state formation in Western Europe and China, see Ma (2009a). 26. The literature on Chinese mercantile and village communities is huge: see Ch’u (1962) Chapters 9 and 10, Xu (2004), Chapter 2, for their role and interaction with local society and government. Also see Ma (2004) for a selective survey of merchant groups. 27. It is interesting to note that in a different context, Hayami et al. (1990) have pointed out the advantage of a progressive land tax system due to its simplicity of administrative procedure and minimum information requirement. 28. See Otsuka et al. (1992) for a review of incentive structures in various forms of land and labor contracts. 29. For a comparison of the different roles and status of communities in China and Europe, see Ma (2009b). 30. Farmers are affected not just by prices of their own products but also by the incentives nonagricultural producers face. That is, it is relative prices and hence relative rates of government assistance that affect producer incentives. More than 70 years ago Lerner (1936) provided his Symmetry Theorem that proved that in a two-sector economy, an import tax has the same effect as an export tax. This carries over to a model that also includes a third sector producing only nontradables.
9780230_274587_19_not.indd 312
6/23/2010 12:23:21 PM
Notes 313 31. Australia and New Zealand were clear exceptions, where manufacturing protection had been very high and its decline occurred several decades later than in other high-income countries (Anderson et al., 2007). 32. This section draws on and is further elaborated in Anderson (2008). 33. Note that it has been assumed that NRA estimates for China pre-1981 and India pre-1965 are the same as the average NRA estimates for those economies for 1981–4 and 1965–9, respectively, and that the gross value of production in those missing years is that which gives the same average share of value of production in total world production in 1981–4 and 1965–9, respectively. This NRA assumption is conservative in the sense that for both countries the average NRA was probably even lower (more negative) in earlier years. 34. While international food prices in mid-2008 were well above those of 2004, the slump in these prices over the second half of 2008 suggests that prices in 2009 may not be so different from those of 2004, and in any case the Doha Round negotiations have been using such a historical period against which to draw up reform proposals. 35. A term in this paper is defined as the interval of time during which a person serves as village leader. A full election term in most villages is three years. If a leader served for three years and then was reelected for three more years, he would be said to have served for two terms. If a leader served for one (or two) years of his full term and then left office (for whatever reason), the one year would be counted as one term. The time (one or two years) served by the person who replaced the former leader would also be counted as one term. 36. During our interviews and periods of survey we were told a number of stories by officials and villagers about why there was no election. In some villages that refused to elect their leaders it was because no one would run. In some villages officials and farmers told us that because there was so much control from the township (because the township controlled the nomination process), villagers decided that they did not want an election. 37. In counting the number of village leaders that acceded to their positions, if a village leader was elected in 1999 and reelected in 2002, we counted this as “two.” Though a large turnover of village leaders occurred in 1995, 1998, and 2001, there were still some village leaders who acceded to the office in other years, since the times of village-leader turnover differed between regions. In addition, there were still some (about 10 per cent) irregular terms for reasons such as village leaders resigning, dying, or being fired. In these cases, some villages replaced their village leaders by running a new election, or just appointed a new village leader. Thus, to account for these irregular terms, we use the average public-goods investment in the term (per year) instead of the total public-goods investment in one term. 38. For public-goods projects/investments from all sources, when we divide the “no difference across terms in the mode of governance” into two groups (elected-toelected and appointed-to-appointed), there is almost no difference. 39. The potential impact of district capacity on project choice by wards, in particular achieving a better fit between project type and community needs, can thus come through both greater administrative capacity and more autonomy in project management. 40. In the analysis, we excluded 144 “supplementary” projects, leaving 1282 projects, due to the fact that these projects were undertaken following criteria quite different from those of the other projects. 41. There is one province missing presidential election results for 1991.
9780230_274587_19_not.indd 313
6/23/2010 12:23:21 PM
314
Notes
42. We do not have information on the two representatives of the chiefs for each district, which implies that the majority party constructed for the two elections (members of parliament and district councilors) is not completely accurate. 43. http://web.worldbank.org/WBSITE/EXTERNAL/TOPICS/EXTARD/EXTAGISO U/0,,contentMDK:20930620~menuPK:2756949~pagePK:64168445~piPK:6416 8309~theSitePK:2502781,00.html. 44. See NAADS website http://www.naads.or.ug/news.php?id=90, accessed September 14, 2009. 45. Hayami (e.g., 2009) argues that the community mechanism is effective not only for the enforcement of contract-based transactions but also for the management of local affairs and the provision of local public goods. In this paper, however, we focus on the enforcement of contract-based transactions. 46. Here we distinguish between a cluster which has been established by local enterprises and one led by foreign direct investment. We believe that the community mechanism of contract enforcement plays an essential role in the former case, whereas the hierarchical relationships between large foreign enterprises and small local enterprises play a critical role in the latter case. 47. According to recent studies of tenancy and land market transactions in subSaharan Africa (Holden et al., 2009), share tenancy is less common in this region than in Asia. 48. So far we have conducted more than 20 case studies in Japan, Taiwan, China, Vietnam, the Philippines, Bangladesh, India, Sri Lanka, Pakistan, Kenya, Ethiopia, Tanzania, and Ghana. 49. To our knowledge, there are no large rural industrial clusters in sub-Saharan Africa (SSA), presumably because of the lack of well-developed infrastructure connecting rural producers to large markets in urban areas or foreign countries. This does not imply, however, that the community mechanism of contract enforcement does not work in SSA. See, for example, the case of the shoe-manufacturing cluster in Addis Ababa in Ethiopia (Sonobe et al., 2009) and the garment-manufacturing cluster in Nairobi in Kenya (Akoten and Otsuka, 2007). 50. Although Jili has become urbanized, it was a poor rural area when its successful industrialization began. This is also the case in Wenzhou. 51. According to our inquiries into the development of the garment- and shoemanufacturing clusters in Wenzhou, their development patterns are common and surprisingly similar to those in Jili, even though Jili’s development lagged behind Wenzhou’s. 52. Relatively simple parts are also produced in the neighboring communities. 53. Although unreported here, community, kinship, and family ties are strong in the garment-manufacturing cluster in Nairobi (Akoten and Otsuka, 2007) and in the iron- and steel-making cluster in Northern Vietnam (Nam et al., 2009b). Similar observations are made among waste pickers and collectors in Delhi (Hayami et al., 2006). 54. Such transfer of production and management knowledge is seldom found in other industrial clusters in SSA. 55. Actually we have already offered such training programs in Ghana, Kenya, and Ethiopia in collaboration with the World Bank and the Japan International Cooperation Agency. 56. Mahabub Hossain conducted surveys in Hayami’s village in 1992, 1997, and 2001; Nobuhiko Fuwa in 2003; and Yasuyuki Sawada in 2006. 57. Some parts of the next sections were drawn from chapters 4 to 7 and 9 of Hayami and Kikuchi (2000).
9780230_274587_19_not.indd 314
6/23/2010 12:23:21 PM
Notes 315 58. According to Fujiie et al. (2005), collective action in irrigation management in the Philippines is difficult to organize where (1) water supply is uniformly abundant; (2) water supply differs greatly between upper and lower streams in laterals; (3) the size of the association is large; (4) population density is low; (5) the share of nonfarm households is high; and (6) the history of irrigated farming is short. 59. Existing studies describe how households can use human assets to cope with negative shocks. This is done by increasing the number of workers, extending labor hours, or migrating to places where there are jobs (Kochar, 1999; Rose, 2000; Smith et al., 2004; Walker and Ryan, 1990, pp. 87–8). Even child labor income, which requires dropping out of education, is used as a coping device against parental income shortfalls (Jacoby and Skoufias, 1997; Sawada and Lokshin, 2009). 60. Informal assistance from relatives and friends is an informal risk-sharing scheme within a local community (Dercon and Krishnan, 2000; Dubois et al., 2008; Ligon, 2008; Townsend, 1994; Udry, 1994). Informal state-contingent transfers and gifts among relatives, friends, and neighbors can be based on altruism (Cox et al., 1998; Lucas and Stark, 1985) or seen as mutual-reciprocal insurance schemes for self-interested risk-coping purposes (Coate and Ravallion, 1993). 61. This section draws from Hayami (2009) and Kajisa (2007). 62. Industrial development in the country has continued to be concentrated in Luzon and increasingly in the southern Tagalog region where Hayami’s village is located (Estudillo et al., 2007). 63. In Muto and Yamano (2008), we compare the same model for bananas and maize, which is less perishable than bananas. We do not find any significant impact of the mobile-phone network on maize price and marketing. 64. The RePEAT Project is a collaborative research project of the Foundation for Advanced Studies on International Development, the National Graduate Institute for Policy Studies, the World Agro-forestry Center, and Tegemeo Institute in Kenya. More details on RePEAT are available in Yamano et al. (2005). 65. Out of the 1000 targeted households, 914 were successfully identified and 874 were interviewed in 2004. The rest (40 households) were successfully identified but not available for interview for several reasons: noncontact (27 cases), moved away (6 cases), refusal (5 cases), unavailable (1 case), and household dissolved (1 case). Among the 874 households, we do not use 12 since some of the variables used in the analyses are missing. 66. The Uganda RePEAT surveys were jointly conducted by Makerere University, the Foundation for Advanced Studies on International Development (FASID), and the National Graduate Institute for Policy Studies (GRIPS). 67. As with other developing countries, there is seasonality in milk production in Kenya. This is probably because the feed for cattle depends on grass which is grown under rain-fed conditions. Since milk production is seasonal, the price tends to fluctuate across seasons, if there is no price regulation. 68. Ngigi (2005) notes that even before liberalization, raw milk was sold directly to households in the neighborhood of the producing households in urban areas. However, raw-milk traders were harassed by the police in urban areas, and the quantity of raw milk traded was considered to be much lower before liberalization than after liberalization. 69. We only asked about the largest milk buyer. Thus, dairy farmers who identified local customers as the largest milk buyer could have sold some milk to traders. The data show, however, that about 92 per cent of the sample households sold the entire quantity of milk produced to just one buyer type.
9780230_274587_19_not.indd 315
6/23/2010 12:23:21 PM
316
Notes
70. The following observations are based on our own field observations and also from FASID (2008). 71. The size-of-firm gradient may be a result of founders joining the party later, after their success. Our question was intended to elicit party membership at the time of founding, but there may have been some deviations from that intent in answers. 72. See the review in Bigsten and Gebreeyesus (2007). 73. Note that since this survey was fielded, a major expansion of health insurance has taken place. 74. Census data in China do not have proper income data, and we would need a very large sample size, such as from a census, to obtain sufficient numbers of workers who live in Wenzhou. 75. Though we must be careful because, based on our study, we do not know if the average schooling of production workers is rising over time. We can guess that it may be, since overall schooling levels are rising, but that may not be true for the particular workers in these three industries, who even in 2005 had mainly only junior-high and primary-school education. 76. Note that very similar arguments were made in the late 1960s and 1970s about conditions in Taiwan and Korea being special, so that replication in other parts of the world might be difficult. This further suggests discounting these specialcase arguments.
9780230_274587_19_not.indd 316
6/23/2010 12:23:21 PM
Contributors Jock R. Anderson, University of New England, Armidale, NSW, Australia Kym Anderson, School of Economics, University of Adelaide Shuhei Aoki, Graduate School of Economics, University of Tokyo, and National Institute for Research Advancement Pranab Bardhan, Department of Economics, University of California at Berkeley Shashanka Bhide, National Council of Applied Economic Research, New Delhi Regina Birner, International Food Policy Research Institute, Washington DC Cristina C. David, Philippine Institute of Development Studies Klaus Deininger, International Food Policy Research Institute, Washington DC Alain de Janvry, Agricultural and Resource Economics, University of California at Berkeley Julen Esteban-Pretel, National Graduate Institute for Policy Studies Jonna P. Estudilo, Foundation for Advanced Studies on International Development and National Graduate Institute for Policy Studies Gershon Feder, International Food Policy Research Institute, Washington DC Nobuhiko Fuwa, Graduate School of Horticulture, Faculty of Horticulture, Chiba University Yoshihisa Godo, Department of Economics, Meiji Gakuin University Peter Hazell, School of Oriental and African Studies, London University Jikun Huang, Center for Chinese Agricultural Policy Kei Kajisa, Social Sciences Division, International Rice Research Institute, Philippines, and Foundation for Advanced Studies on International Development Kaliappa Kalirajan, The Australian National University, Foundation for Advanced Studies on International Development and National Graduate Institute for Policy Studies Masao Kikuchi, Graduate School of Horticulture, Faculty of Horticulture, Chiba University Yoko Kijima, Graduate School of Systems and Information Engineering, University of Tsukuba Dong Liu, Center for Research of the Private Economy, Zhejiang University Renfu Luo, Center for Chinese Agricultural Policy Debin Ma, Economic History Department, London School of Economics Tomoya Matsumoto, National Graduate Institute for Policy Studies 317
9780230_274587_20_con.indd 317
6/23/2010 12:23:27 PM
318 Contributors Mehdi Majbouri, Department of Economics, University of Southern California Dilip Mookherjee, Department of Economics, Boston University Megumi Muto, Japan International Cooperation Agency Hideyuki Nakagawa, Department of Agricultural and Resource Economics, University of California at Berkeley Tetsuji Okazaki, Graduate School of Economics, University of Tokyo Keijiro Otsuka, Foundation for Advanced Studies on International Development and National Graduate Institute for Policy Studies Edward Y. Qian, Center for Research of the Private Economy, Zhejiang University Gustav Ranis, Department of Economics, Yale University Scott Rozelle, Freeman Spogli Institute, Stanford University Elisabeth Sadoulet, Agricultural and Resource Economics, University of California at Berkeley Yasuyuki Sawada, Graduate School of Economics, University of Tokyo Minggao Shen, Caijing Group, Beijing Kanhaiya Singh, National Council of Applied Economic Research, New Delhi Tetsushi Sonobe, Foundation for Advanced Studies on International Development and National Graduate Institute for Policy Studies John Strauss, Department of Economics, University of Southern California Qi Sun, Department of Economics, University of Southern California Takashi Yamano, Foundation for Advanced Studies on International Development and National Graduate Institute for Policy Studies Qianfan Ying, Center for Research of the Private Economy, Zhejiang University Linxiu Zhang, Center for Chinese Agricultural Policy Yi Zhu, Department of Economics, University of Southern California
9780230_274587_20_con.indd 318
6/23/2010 12:23:27 PM
Index agricultural development, 15, 178, 187, 194, 200, 203, 205, 296, 297 agricultural extension, 188, 189, 192, 200, 203 agricultural policies, 7, 83, 86 agricultural productivity, 11, 15, 53 Agricultural protectionism, 17 agricultural risks, 10 agricultural trade policies, 7 balanced regional development, 45, 46 banana marketing, 246, 248, 258 catch-up growth, 7, 41, 104, 118, 119 Chinese empire, 65, 66 clientelism, 8, 150, 153, 185 community failure, 4, 6, 7, 54 community governance, 145, 172 community mechanism, 7, 8, 10, 23, 37, 187, 188, 212, 213, 214, 215, 216, 218, 219, 220, 221, 234, 263, 264, 274, 287, 308, 314 community participation, 8, 9, 153, 159, 161, 165, 166, 169, 170, 191, 259 community-based extension systems, 189 community-based organizations, 150 comparative advantage, 17, 97, 191, 215, 294 competitive market, 5, 197 conditional convergence, 47, 59 Confucian ideology, 70 contract choice, 6, 13 contract enforcement, 7, 9, 10, 212, 213, 215, 218, 219, 221, 222, 231, 245, 263, 287, 314 contract farming, 307 crop insurance, 296, 297 decentralization, 49, 50, 51, 53, 61, 150, 151, 153, 155, 164, 165, 166, 311 decentralization index, 50, 51, 61 democracy, 123, 124, 139, 173, 174 democratization, 8, 124
developing countries, 8, 10, 12, 80, 84, 85, 86, 93, 94, 95, 96, 97, 99, 100, 101, 124, 127, 150, 187, 198, 212, 219, 231, 238, 256, 295, 297 development expenditures, 59, 60 domestic technologies, 22 East Asia, 13, 16, 53, 64, 65, 66, 75, 76, 77, 101, 170, 171, 222, 242, 244, 277, 289 East Asian miracle, 64 economic catch-up, 76, 109, 110, 113, 114, 115, 117 economic change, 68 economic development, 7, 14, 17, 22, 29, 45, 59, 108, 241 economic services, 60, 62 economic welfare, 4, 6, 7, 83, 86, 95, 96 education expenditures, 6, 62 elementary education, 39, 106 elite capture, 191, 192, 195, 204 emerging economies, 7 endogeneity, 202 equal-field system, 68, 69 externalities, 3, 34, 149 factor endowments, 15 family firms, 10, 287 farm size, 15, 17 Finance Commission, 57, 59 foreign technologies, 6, 22, 24, 26, 40, 107, 111, 114 free trade, 81, 96 free-market, 3, 4 futures markets, 307 GATT, 98, 99 Gini coefficient, 95 governance reforms, 7, 125, 129 government failure, 7, 41 government intervention, 3, 32, 100, 303 Green Revolution, 10, 15, 53, 146, 171, 223, 243, 260, 298, 309, 310 growth-accounting, 21 319
9780230_274587_21_ind.indd 319
6/23/2010 12:23:34 PM
320 Index human capital, 17, 38, 62, 77, 104, 105, 107, 111, 113, 115, 116, 118, 238, 311 immigration, 107 income distribution, 3 index-insurance, 301, 308 Indian economy, 46, 47 induced innovation, 5, 14, 36 industrial clusters, 8, 9, 39, 212, 214, 215, 220, 221, 222, 314 Industrial Revolution, 77, 105 infant mortality rate, 46, 55 information asymmetry, 3, 211, 212, 224, 236 infrastructure, 39, 53, 55, 57, 126, 127, 145, 164, 172, 173, 175, 177, 180, 213, 222, 246, 262, 312 institutional economics, 12 institutional endowments, 7, 76 institutional innovation, 240 International Food Policy Research Institute, 208, 249, 310 IRRI, 18, 226 irrigation management, 315 irrigation systems, 16, 145, 229 Japan Economic and Social Research Institute, 108, 120 Japanese economy, 7, 21, 22, 23, 24, 33, 103, 104, 113, 114, 115, 117, 118, 119 Japanese miracle, 107, 113 Japanese system of management, 115, 118 labor contracts, 232, 233 land inequality, 177, 182, 185 land tax, 70, 71, 72, 76, 312 land-reform, 13, 15, 172, 173, 174, 175, 177, 178, 180, 181, 182, 226, 231, 232, 239, 240 landlord-tenant relations, 226, 231 Learning by doing, 265 life expectancy, 46, 55, 311 literacy rate, 46, 54, 55, 311 local governance, 7, 124, 172, 173 macroeconomic policies, 7 marginal product, 231 market access, 17 market competition, 4, 237 market discovery, 278, 279
9780230_274587_21_ind.indd 320
market failure, 3, 4, 222 mechanization, 16, 17, 37, 40, 311 Meiji Japan, 75 Meiji Restoration, 7 migration, 26, 107, 156, 286, 311 milk market, 246, 247, 252 mobile phone network, 246, 248, 249, 256, 257, 258, 259 moral hazard, 181, 300 neoclassical growth model, 53 nominal rates of assistance, 81, 91 OECD, 108 oil crisis, 103 operation barga, 173 panchayats, 172, 173, 174, 175, 177, 178, 311 panel data, 9, 246, 249, 259 Patent protection, 220 Planning Commission, 57, 63 political clientelism, 8, 185 political economy, 12, 13, 16, 65, 71, 73, 81, 95, 153, 154, 161, 166, 173, 179 population growth, 15, 16 poverty alleviation, 11, 173 price-stabilization, 297, 298, 307 private goods, 3 productivity, 6, 11, 12, 15, 17, 21, 23, 24, 25, 29, 53, 95, 115, 203, 286 property rights, 65, 66, 69, 70, 76, 245, 312 public goods, 3, 7, 17, 62, 67, 123, 124, 125, 127, 132, 137, 140, 141, 144, 145, 224, 311 Qin empire, 66 R and D, 6, 22, 23, 26, 27, 30, 33, 34, 35, 36, 40, 41, 43, 44, 311 regional development, 45, 46 relative rate of assistance, 82, 91 Reserve Bank of India, 59, 63 resource endowments, 4 rice price policy, 13 risk aversion, 294 risk-coping strategies, 234, 292 rural communities, 9, 10, 123, 145, 189, 212, 214, 222, 225, 238, 239, 240, 291, 292, 294, 295, 300, 308
6/23/2010 12:23:34 PM
Index 321 rural governance, 123, 125 rural industrialization, 216 rural poverty, 213, 238, 259 scheduled tribes, 54 school enrollment, 106, 108, 109, 111, 156, 158, 161, 165 small and medium enterprises, 265 social capability, 104, 105, 237 social capital, 8, 64, 151, 154, 187, 189, 190, 191, 197, 203, 205, 211, 237, 238, 245, 291, 294 social services, 59, 60 Southeast Asia, 89, 99 spatial equity, 45, 62 state-owned enterprises, 262, 264 technocracy, 7, 76 technological change, 4, 15 technological innovations, 4, 36 technology acquisition, 277, 287 technology adoption, 22, 23, 26, 29, 30, 31, 32, 33, 36 technology leaders, 26 tenancy contracts, 211, 231, 237
9780230_274587_21_ind.indd 321
tertiary education, 7, 111, 113, 116, 117, 118 TFP, 6, 21, 22, 23, 24, 25, 26, 27, 29, 31, 33, 34, 35, 36, 37, 38, 39, 40, 41, 43, 311 total factor productivity, 6, 21, 25 township-village enterprises, 262, 264 Trade Reduction Index, 83 trade restrictiveness index, 81 transaction costs, 9, 16, 101, 188, 190, 192, 222, 224, 231, 245, 246, 253, 256, 305 unobserved heterogeneity, 130 UNRISD, 63 voting behavior, 8, 185 Welfare Reduction Index, 83 Wenzhou model, 264 Western Europe, 67, 75, 76, 95, 312 World Bank, 7, 10, 63, 80, 81, 83, 84, 86, 96, 102, 120, 126, 127, 146, 147, 171, 186, 189, 205, 206, 207, 208, 222, 244, 290, 303, 306, 308, 314 WTO, 98, 99, 101
6/23/2010 12:23:34 PM
9780230_274587_21_ind.indd 322
6/23/2010 12:23:35 PM
