WORLD DEVELOPMENT REPORT 2002 22825 September 2001 =~~'-a Building Institutions for Markets THE WORLD BANK WORLD DEVELOPMENT REPORT 2002 Bul in.lg nstitutions for Markets Published for the World Bank Oxford University Press Oxford University Press OXFORD NEW YORK ATHENS AUCKLAND BANGKOK BOGOTA BUENOS AIRES CALCUTTA CAPE TOWN CHENNAI DAR ES SALAAM DELHI FLORENCE HONG KONG ISTANBUL KARACHI KUALA LUMPUR MADRID MELBOURNE MEXICO CITY MUMBAI NAIROBI PARIS SAO PAULO SINGAPORE TAIPEI TOKYO TORONTO WARSAW and associated companies in BERLIN IBADAN C) 2002 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, N.W., Washington, D.C. 20433, U.S.A. Published by Oxford University Press, Inc. 200 Madison Avenue, New York, N.Y. 10016 Oxford is a registered trademark of Oxford University Press. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior permission of Oxford University Press. Cover design by Debra Naylor, Naylor Design, Inc. Cover photographs: top, © 2001 Stephen Simpson/FPG; middle, C 2001 Stone/Paul Chesley; bottom, Yosef Hadar, World Bank. Inside design and typesetting by Barton Matheson Willse & Worthington, Baltimore. Manufactured in the United States of America First printing September 2001 This volume is a product of the staff of the World Bank, and the judgments made herein do not necessarily reflect the views of its Board of Executive Directors or the countries they represent. The World Bank does not guarantee the accuracy of the data included in this publication and accepts no responsibility whatsoever for any consequence of their use. The boundaries, colors, denominations, and other information shown on any map in this volume do not imply on the part of the World Bank any judgment on the legal status of any territory or the endorsement or acceptance of such boundaries. ISBN 0-19-521607-5 clothbound ISBN 0-19-521606-7 paperback ISSN 0163-5085 Text printed on recycled paper. r o revvo r T his World Development Report is about building We recognize the central importance of institutions in market institutions that promote growth and the development process through the Comprehensive I reduce poverty, addressing how institutions Development Framework, which stresses the interde- support markets, what makes institutions work, and pendence of institutions with the human, physical, and how to build them. macroeconomic sides of development. This theme is a natural continuation of last year's The Report emphasizes the importance of histori- Report, which demonstrated that markets are central cal context: where countries are today affects where to the lives of poor people, that institutions play an im- they can go. It also takes a pragmatic approach to in- portant role in how markets affect people's standards of stitution building, focusing on what can be done prac- living and help protect their rights. This Report identi- tically rather than on what should be done in an ideal fies how institutions can promote inclusive and inte- world. Social and political factors affect the pace of grated markets, and ensure stable growth and thus dra- change, and sweeping reforms are not always possible. matically improve people's incomes and reduce poverty. It is important to work on the areas where opportuni- It is about equal opportunity and empowerment for ties present themselves; each step can take countries people, especially the poor. forward-if correctly designed. And smaller reforms Some countries have successfully harnessed market- can build constituencies for larger ones. oriented reforms to improve the welfare of all their peo- This Report recognizes that one size does not fit all ple. But in other countries, markets have not given peo- in institution building and provides policy guidance on ple as much incentive to engage in wider trade, the how to develop appropriate institutions. Building on ability to use fully their skills and resources, and oppor- the successes of countries, and learning from the fail- tunities to increase their income. ures, the Report provides a deeper understanding of Effective institutions can make the difference in the market-supporting institutions and a better apprecia- success of market reforms. Without land-titling insti- tion of how people may build such institutions. In iden- tutions that ensure property rights, poor people are un- tifying how to promote institutional change, it looks at able to use valuable assets for investment and income the roles of private and public, and national, local, and growth. Without strong judicial institutions that en- international, actors. It draws on a wealth of research force contracts, entrepreneurs find many business ac- and practical experience from inside and outside the tivities too risky. Without effective corporate gover- Bank, as well as on insights from many disciplines, pre- nance institutions that check managers' behavior, firms senting new research and data on institutions. waste the resources of stakeholders. And weak institu- From these diverse sources, the Report distills four tions hurt the poor especially. For example, estimates lessons on building effective institutions: show that corruption can cost the poor three times as much as it does the wealthy. a Design them to complement what exists-in terms Addressing the challenge of building effective institu- of other supporting institutions, human capabilities, tions is critical to the Bank's mission of fighting poverty. and available technologies. The availability and costs III IV WORLD DEVELOPMENT REPORT 2002 of supporting institutions and capacity determine * Promote competition among jurisdictions, firms, the impact of any particular institution. By under- and individuals. Developing country market actors standing how institutions interact, we can identify often face too little competition, and changing this priorities, will significantly improve institutional quality. * Innovate to identify institutions that work-and Greater competition modifies the effectiveness of ex- those that do not. Sometimes this requires experi- isting institutions, creates demand for new ones, and mentation. Even in countries with similar incomes increases choice for consumers. Competition among and capacities, innovation can create stronger insti- jurisdictions highlights successful institutions and tutions because of differences in local conditions, promotes demand for them. Competition among differences that range from social norms to geogra- firms and individuals does the same. phy. Countries can gain from expanding successful public innovations and adopting private innova- These broad lessons, as well as the detailed analysis tions. But they must also have the courage to drop and many examples throughout this Report, will help failing experiments. us and policymakers build institutions that ensure sta- * Connect communities of market players through ble and inclusive growth and thus improve people's in- open information flows and open trade. Exchanging comes and reduce poverty. information changes behavior. It creates demand for institutional change by holding people to account and by supplying ideas for change from outside the community. Linking communities of people in net- works of information and trade is thus a priority for those building market-supporting institutions. James D. Wolfensohn FOREWORD V This Report has been prepared by a team led by Roumeen Islam and comprising Arup Banerji, Robert Cull, Asli Demirgiic-Kunt, Simeon Djankov, Alexander Dyck, Aart Kraay, Caralee McLiesh, Russell Pittman; and Helena Tang, Nazmul Chaudhury, Jeffrey Hammer, Richard Messick, and Tatiana Nenova made additional contributions. The team was assisted by Theodora Galabova, Paramjit Gill, Yifan Hu, Olga loffe, Claudio Montenegro, Stefka Slavova, Mahesh Surendran, and LiHong Wang. Andrei Shleifer and Joseph E. Stiglitz provided valuable suggestions during the writing of the Report. Bruce Ross-Larson is the editor of the overview, chapter 1, and chapter 10. Andrew Balls provided editorial assistance. The work was carried out under the general direction of Nicholas Stern. The team was advised by a panel of experts comprising Carl Anduri, Abhijit Banerjee, Kaushik Basu, Tim- othy Besley, Francois Bourguignon, Antonio Estache, Cheryl Gray, Avner Greif, Nurul Islam, Emmanuel Jimenez, Daniel Kaufmann, Michael Klein, Yingyi Qian, and Kenneth Sokoloff. Many others inside and outside the World Bank provided helpful comments and wrote background papers and other contributions, and their names are listed in the bibliographical note. Much of the background re- search was supported by a generous grant from the Dutch government. Research was also supported by the Swiss Trust Fund. The World Bank Development Data Group contributed to the data appendix and was re- sponsible for the Selected World Development Indicators. The team undertook a wide range of consultations for this Report, from the initial outline to the final draft. During the Report's planning stage in 2000, a February workshop in Berlin and a July workshop in Washington, D.C. provided an exchange of ideas among academics and policymakers from around the world. During the drafting stage in 2001, a consultative meeting on the media was held in April, and a consultative meeting on judicial systems was held in May. In addition, the authors held consultations with a wider com- munity that included nongovernmental organizations, holding meetings in Paris (with representatives from the French Conseil d' Etat, the Organisation for Economic Co-operation and Development, the French De- velopment Agency, and the Competition Council); in London (Department for International Development, Overseas Development Institute, and nongovernmental organizations); and in Amsterdam (Amsterdam Insti- tute for International Development). The team also conducted a series of videoconferences with audiences in Bangladesh, Brazil, Egypt, Japan, Mexico, Morocco, South Africa, and Thailand. A consultation with the In- ternational Confederation of Free Trade Unions was also held. Rebecca Sugui served as executive assistant to the team; Leila Search as program assistant and technical support; and Shannon Hendrickson, Joanna Kata-Blackman, Mei-Ling Lavecchia, and Rudeewan Laohakit- tikul as team assistants. Maria Ameal and later Eva Santo Domingo served as resource management assistant. Book design, editing, and production were coordinated by the Production Services Unit of the World Bank's Office of the Publisher. PART q N7iROdUCTiON I Buildinr lnstitutions: Complement lrtrovzi ' C or nect, and Compete 3 How do institutions support markets?. .. 5 How do institutions support growth and poverty reduction?. . . 9 How do you build effective institutions? ...10 Organization and scope of the Report . . .25 Conclusions. . . 26 PART Xi F I R ! S L Farmers Building more secure and transferable rural land institutions .32 Building effective and accessible rural financial institutions .39 Building effective institutions for agricultural technology and innovation .43 Conclusions .52 3 Governance of Firms 55 What firms around the world look like .57 Private governance institutions for firms .58 Laws and formal intermediaries .63 Conclusions .73 4 Financial Systenmc 75 Should policymakers promote bank-based or market-based financial systems? .77 What form should financial regulation take? .79 Enhancing efficiency in the financial sector: the role of ownership and competition .84 How foreign entry and e-finance can change the nature of financial markets .88 How to enhance access to financial services .91 Conclusions .96 VII VIII WORLD DEVELOPMENT REPORT 2002 PART III GOVERNMENT 5 Political Institutions and Governance 99 Political institutions and policy choices ................................................. 101 Corruption ..................................................................... 105 Politics, institutions, and taxation . .................................................... 110 Conclusions ................................................................... 115 6 The Judicial System 117 Comparison of legal and judicial systems ............................................... 120 New evidence on two aspects of judicial efficiency: speed and cost ............................ 121 Judicial reform efforts .............................................................. 124 Fairness .................................................................... 129 Conclusions ................................................................... 131 7 Competition 133 Domestic competition . ............................................................ 135 International competition ........................................................... 142 Conclusions ................................................................... 149 8 Regulation of Infrastructure 151 Competition in infrastructure sectors .................................................. 154 Structure of the regulatory system ..................................................... 158 Designing infrastructure regulation to deliver services to poor people .......................... 161 Conclusions ................................................................... 166 PART IV SOCIETY 9 Norms and Networks 171 Informal institutions in markets: their utility and shortcomings .......... .................... 172 Building and adapting formal institutions ............................................ 176 Integrating informal and formal institutions ............................................ 178 Conclusions ............................................ 179 10 The Media 181 Independence ............................................ 183 Quality ............................................. 188 Broadening the media's reach ............................................ 190 Institutions to complement the media ............................................. 192 Conclusions ............................................ 192 Bibliographic Note 195 Selected World Development Indicators 229 Boxes 1.1 A poem on the problems of trade .................................................... 5 1.2 What are institutions? ............................................................. 6 1.3 Institutional evolution and economic development: private traders and public rulers .............. 7 1.4 Courts and the expansion of trade ................................................... 9 1.5 Weak institutions hurt poor people .................................................. 9 CONTENTS IX 1.6 Who builds institutions? .11 1.7 Human capital and institutional design .13 1.8 Computerization and land registration in Andhra Pradesh, India .14 1.9 Private innovation supported by formal institutional change .15 1.10 Distributional effects of innovation depend on who innovates: bankruptcy law in two countries .16 1.11 Experimentation and adaptation: bankruptcy institutions in Hungary .17 1.12 Trade and institutional change in Thailand .17 1.13 Institutional evolution of rice markets and standardization in Japan, 1600-1920s .18 1.14 Role of the news media in fighting corruption in Kenya .19 1.15 Competition and the evolution of corporate law .20 1.16 The interplay of social, political, and economic forces in the reform of land institutions in China .. 22 1.17 Applying the lessons to the social sectors .............................................. 24 1.18 Crises and institutional change in Malaysia .25 2.1 Agricultural marketing institutions .33 2.2 Early institutions of land ownership in Mesopotamia and Egypt .34 2.3 Quilombos in Brazil: infrastructure, social change, and a new demand for land registration .34 2.4 Examples of policy barriers to the operation of land markets .35 2.5 Informal collateral transactions using land titles in Thailand .37 2.6 Do indigenous land rights constrain agricultural investment and productivity in Africa? .37 2.7 A transparent and accessible institutional framework for granting land rights in Mexico .38 2.8 The intersection of formal and informal lending: marketing agents in the Philippines .41 2.9 Livestock as savings: contrasting evidence from India and Burkina Faso .43 2.10 Creating an information-sharing network for the poor: SRISTI in India .47 2.11 Private sector extension services in Argentina .47 2.12 Increasing information flows between farmers and researchers in Ghana ....................... 51 2.13 International spillovers and the CGIAR .51 3.1 Business groups and restrictions on competition in Kazakhstan .62 3.2 Business associations and trade credit .63 3.3 The need for formal laws: the development of corporate law .65 3.4 Measuring the strength of legal protections for shareholders .65 3.5 Legal and regulatory change in Brazil .66 3.6 Limitations to private governance in accounting .71 3.7 Evolution of international accounting standards .72 4.1 The financial revolution versus the industrial revolution .76 4.2 Financial structure varies across countries: better information and legal systems that protect property rights play a role .78 4.3 Promoting stock markets in developing countries .78 4.4 Designing a bank safety net: the role of deposit insurance .81 4.5 The role of subordinated debt in establishing credibility: the case of Argentina .82 4.6 Institutional design for bank supervisors .83 4.7 The political economy of banking reform .87 4.8 Strengthening bank exit mechanisms: lessons from Latin America .89 4.9 The effects of foreign ownership of banks in Hungary .91 4.10 Technology and provision of financial services .92 4.11 Credit registries .95 5.1 Political institutions, property rights, and fiscal outcomes in 17th century England .100 5.2 Packaging trade reforms ........................................................... 105 X WORLD DEVELOPMENT REPORT 2002 5.3 Political connections and firm value in Indonesia ........................................ 106 5.4 Discretion and truck inspection in Gujarat, India ........................................ 107 5.5 Constitutional reform for anticorruption in Thailand ..................................... 108 5.6 Business taxation in Uganda . ...................................................... 111 5.7 Market discipline versus state discipline: municipal bankruptcy in Hungary .................... 113 5.8 Tax sharing with weak tax administration: the case of Russia ............................... 114 6.1 How mediation resolves disputes . ................................................... 118 6.2 Who benefits from better courts? ..................................................... 119 6.3 Surveys on judicial performance ..................................................... 120 6.4 Comparing judicial efficiency ....................................................... 121 6.5 Index of the complexity of litigation . ................................................ 122 6.6 Debt recovery in Tunisia .......................................................... 123 6.7 The creation of a specialized commercial court in Tanzania ................................ 126 6.8 Alternative dispute resolution in Bangladesh ........................................... 127 7.1 Adam Smith on competition, 1776 . .................................................. 134 7.2 Measuring competition ........................................................... 134 7.3 Labor regulations and rigidities in the labor market: the example of India ...................... 136 7.4 Increased product market competition and increased labor market flexibility in India ..... ........ 137 7.5 Differences between the United States and the EU on competition law and its enforcement .... .... 141 7.6 Open trade and institutional change: product markets in India .............................. 143 7.7 Benefits of liberalization of industrial country markets for agriculture and textiles ................ 144 7.8 Lack of competition in services restricts gains from merchandise trade liberalization ..... ......... 145 7.9 Weak IPR systems promoted access to technology and growth in East Asia ..................... 147 8.1 Private provision: recent evidence from concession arrangements ............................ 153 8.2 Vertical integration and discrimination in the provision of Internet services .................... 157 8.3 Water concession in Tucuman, Argentina .............................................. 163 8.4 Targeting subsidies: Chile's approach . ................................................ 165 9.1 Exclusion in trading in African history ................. ............................... 175 9.2 Tiripur in Tamil Nadu (India): insiders and outsiders in the use of informal institutions ..... ...... 175 9.3 The influence of formal institutions on norms: colonial Uganda ............................. 177 9.4 Islamic banking: informal and formal approaches ........................................ 178 9.5 Education among the Orma in Kenya: adapting well-established norms ....................... 178 10.1 The media's role in reducing corruption in Peru .182 10.2 Improving education through the media in Panama .182 10.3 Measuring media independence through data on media ownership .183 10.4 Media ownership influences content: Ukraine .185 10.5 Controlling news on the Internet .187 10.6 Improving access to information in Thailand .190 10.7 Increasing access to the media: wall newspapers in Nepal .................................. 192 Figures 1.1 The concentration of ownership varies tremendously across countries ......................... 8 1.2 Financial depth generates growth .................................................... 10 1.3 Complexity of procedures in debt collection ............................................ 12 1.4 Cost of business registration across countries ........................................... 13 CONTENTS Xi 1.5 Greater openness and quality of institutions .18 1.6 Diversity of information providers and quality of institutions .19 2.1 Poverty headcounts, urban versus rural, selected countries, 1990s .32 2.2 Evolution of formal institutional borrowing of farm households, selected Asian countries .39 2.3 Financial performance of generalized agricultural insurance programs .44 2.4 Median rates of return on agriculture research and extension by region .45 2.5 Agricultural research intensity, public and private, 1993 .48 3.1 Flows of new investment are insensitive to value added in developing countries .56 3.2 Proportion of assets in publicly traded firms accounted for by group-affiliated firms .61 3.3 Shareholder rights and stock market development .66 3.4 Accounting standards across countries .70 3.5 Capital raised through new depository receipt programs .73 4.1 Financial system development across income groups .77 4.2 State ownership in banking, 1998-99 .85 4.3 Evolution of the Hungarian banking sector .86 4.4 Increase in the market share of majority foreign-owned banks, selected countries, 1994 and 1999... 89 5.1 Variation in the quality of policies around the world ..................................... 101 5.2 Divided governments have difficulty with fiscal adjustments ................................ 103 5.3 Governments with fewer checks and balances than others are less likely to enforce banking regulations during crises ................................................................... 104 5.4 Tax collection around the world ..................................................... 110 6.1 (a) Procedural complexity reduces efficiency ............................................ 122 (b) Rich countries also have complex regulations, but ..................................... 122 (c) They have more efficient systems because of complementary institutions and capacity .... ...... 122 6.2 Excessive written procedures limit access to justice ....................................... 124 6.3 The independence of the judiciary enhances property rights ................................ 129 7.1 The size of the unofficial economy rises with the number of procedures required to start up new business .................................................... 138 7.2 Corruption rises with the number of procedures required to start a new business ..... ........... 139 7.3 Effectiveness of competition law increases with per capita income ............................ 141 7.4 Openness reduces price-cost margins ................................................. 143 7.5 Patent strength rises with per capita income ............................................ 147 10.1 Who owns the media? ..................................................... 184 10.2 State media ownership and low competition are associated with poor outcomes ..... ............ 185 10.3 Media penetration rates by region and by OECD compared with non-OECD countries ........... 191 Tables 2.1 Types of rural lending institutions ............. ...................................... 40 2.2 Where will the private sector invest in agricultural technologies? ............................. 46 3.1 Share of small formal sector firms in selected economies, selected years ........................ 57 3.2 Financial statements do not disclose useful information for resource providers .................. 70 6.1 Inputs into the judicial system for selected countries, 1995 ................................. 120 XII WORLD DEVELOPMENT REPORT 2002 7.1 Benchmarks of product market dominance in competition laws around the world ..... .......... 140 8.1 Investment in infrastructure projects with private participation in developing countries by sector and region, 1990-99 ...................................................... 152 8.2 Strategy for vertical separation or integration ........................................... 156 8.3 Access to electricity, water, sanitation, and telephone services in 22 Latin American countries, 1986-96 ................................................................... 162 9.1 Types of informal sanctions in contract-enforcement mechanisms ............................ 173 Definitions and data notes The countries included in regional and income The following abbreviations are used: groupings in this report are listed in the Classification of Economies table at the end of the Selected World EU European Union Development Indicators. Income classifications are FDI Foreign direct investment based on GNP per capita; thresholds for income clas- GATT General Agreement on Tariffs and Trade sifications in this edition may be found in the Intro- GDP Gross domestic product duction to Selected World Development Indicators. GNP Gross national product Group averages reported in the figures and tables are HIWVAIDS Human immunodeficiency virus/ unweighted averages of the countries in the group un- acquired immune deficiency syndrome less noted to the contrary. IPR Intellectual property rights The use of the word countries to refer to economies . N o n n implies no judgment by the World Bank about the NOE Nrgaoisanion tforoEconomiciCo legal or other status of a territory. The term developing OECD Organisation for Economic Co- countries includes low- and middle-income economies operation and Development and thus may include economies in trasition from cen- PPP Purchasing power parity tral planning, as a matter of convenience. The term de- R&D Research and development velopedor industrial countries may be used as matter of SMEs Small and medium-size enterprises convenience to denote the high-income economies. TRIPS Trade-Related Aspects of Intellectual Dollar figures are current U.S. dollars, unless other- Property Rights wise specified. Billion means 1,000 million; trillion VTO World Trade Organization means 1,000 billion. PART I Intro uction CHAPTER 1 15Ua ilng nstitutLions: f . 1etet Innovate, k_..otnlect, anl (0otnpete How do we accountfor the persistence ofpoverty in the ties. And the rules of the organization, although not midst ofplenty? If we knew the sources ofplenty, why written, were self-enforcing. Remaining in the coali- don't poor countries simply adopt policies that make tion of traders best served each member's interests. So- forplenty?... We must create incentivesforpeople to cial ties cemented mutually beneficial business rela- invest in more efficient technology, increase their skills, tionships, and cross-border trade flourished. and organize efficient markets. Such incentives are Today, a millennium later, people everywhere face embodied in institutions, similar problems in striving to improve their well- -Douglass C. North, 2000 being through market activity. African entrepreneurs lack information about potential business partners. Poor farmers in Latin America lacking formal title to - n the 11th century the Maghribi traders of North their land cannot use it as collateral to secure access to Africa wanted to expand business across borders, credit. Budding entrepreneurs in Central Asia, trying all around the Mediterranean. Trade in each center to start new businesses, run into political obstacles was free of formal regulations and restrictions, and from established firms and the state. competitive, with many buyers and sellers negotiating Despite the problems, many people in rich countries prices through brokers, open-bid auctions, and direct and poor are engaged in productive-and rewarding- dealings. Cross-border trade also was generally free of market activity. As World Development Report 2000/ formal regulations and restrictions. But it was fraught 2001 argued, income from participating in the market with uncertainty about selling prices, the quality on is the key to boosting economic growth for nations and arrival, and the possibility of theft. Only if merchants to reducing poverty for individuals. This Report is traveled with their goods to distant markets could they about enhancing opportunities for poor people in mar- ensure the safe arrival and sale of their merchandise. kets, and about empowering them. What makes mar- Such risks and costs naturally limited trade. ket activity rewarding and possible for some, and not So in all major trading centers around the Mediter- others? Why are some market systems inclusive and in- ranean, the Maghribis set up overseas agents to repre- tegrated, allowing benefits to flow to the poor as well sent their interests and exchange information about as the rich, the rural people as well as the urban? And markets. Being from the same community, these agents why are other markets localized and segmented? were seen as trustworthy. And with fewer contractual The Maghribi example illustrates some of the rea- problems, Maghribi merchants no longer needed to sons. Markets allow people to use their skills and re- travel to ensure that they would not be cheated. Infor- sources and to engage in higher-productivity activities mation flowed freely in this network bound by social if there are institutions to support those markets. What 3 4 WORLD DEVELOPMENT REPORT 2002 are these institutions? Rules, enforcement mechanisms, particular countries. Institution building is generally a and organizations supporting market transactions. Ex- cumulative process, with several changes in different tremely diverse across rich and poor communities and areas building up to complement and support each nations, they help transmit information, enforce prop- other. This Report identifies elements of such a strat- erty rights and contracts, and manage competition in egy. Even small changes can build momentum for fu- markets. All market-supporting institutions do one or ture changes. The whole is greater than the parts, and more of these things. And in so doing, they give peo- even moderate progress in the parts can contribute to a ple opportunity and incentives to engage in fruitful better system to promote growth and reduce poverty. market activity. Four main lessons emerge for institution building. This Report is about people building institutions that The first two are about supplying effective market- support the development of markets. The 2000/2001 supporting institutions. But supplying institutions is Report underscores the importance of institutions in af- not enough. People must want to use them too. Thus, fecting poor people's participation in markets. This Re- the second two lessons are also about creating the de- port discusses both institutions that support growth and mand for such institutions, and about the forces for those that directly affect access of people left out of change within countries. many market activities. It considers those institutions To ensure effective institutions: that provide opportunities for people and that empower them. It goes beyond the 2000/2001 Report by analyz- * Design them to complement what exists-in terms of ing what institutions do to promote growth and facili- other supporting institutions, human capabilities, and tate access and by suggesting how to build effective in- available technologies. The reason? The availability and stitutions. And it emphasizes how institutions can help cost of supporting institutions, existing levels of cor- people make better use of the assets they own and how ruption, human capacity and technology determine to accumulate more. In focusing on institution build- the impact of a particular institution. That is why in- ing, it does not devalue the importance of policy. But stitutions that achieve their goals in industrial coun- good policies are not enough. The details of institution tries may not do so in developing ones. Much of the building matter for growth and poverty reduction. important work in building institutions lies in modi- The Report contributes to existing work on institu- fying those that already exist to complement better tions and markets in several novel ways. It provides a other institutions and in recognizing what not to build diagnostic framework for understanding how institu- in a particular context, as much as what to build. "Best tions support market activity. Bridging the gap between practice" in institutional design is a flawed concept. theory and evidence across disciplines, it also builds on * Innovate to design institutions that work-and drop existing evidence on the role of institutions and insti- those initiatives that do not. Even in countries with tutional change. It extends previous empirical work on similar incomes and capacities, innovation can create institutional change to developing countries and pre- stronger institutions because of differences in local sents a framework for institutional change. It confirms conditions-differences ranging from social norms that one size does not fit all in considering institutional to geography. Experimentation, which has some costs design. But it does more than that. It illustrates how to that must be recognized, can nevertheless help iden- proceed in building more effective institutions. It pro- tify new and more effective structures. Countries can vides policy guidance by taking a pragmatic approach. gain from expanding successful public innovations The aim is not to define what should be done in an and adopting private innovations. But they must also ideal world, but what can be done in today's world. have the courage to drop failing experiments. In understanding what drives institutional change, * Connect communities of market players through open the Report emphasizes the importance of history. Many information flows and open trade. Exchanging goods developing countries have been nation-states for a short and services outside existing networks and commu- time compared with industrial countries. The evolution nities creates demand for market-supporting institu- of nations teaches that building institutions takes time tions. Exchanging information through open debate and that the process within each country may stall or creates demand for institutional change by holding reverse because of political conflicts or economic and people to account, by changing behavior, and by sup- social conditions. It offers lessons about the process of plying ideas for change from outside the community. change and the importance of norms and culture in Linking communities of people in networks of infor- BUILDING INSTITUTIONS: COMPLEMENT, INNOVATE, CONNECT, AND COMPETE 5 mation and trade is thus a priority for policymakers Box 1.1 building market-supporting institutions. A poem on the problems of trade * Promote competition among jurisdictions, firms, and individuals. Greater competition modifies the effec- If I knew you and you knew me tiveness of existing institutions, changes people's in- 'Tis seldom we would disagree; But never having yet clasped hands centives and behavior, and creates demand for new Both often fail to understand institutions. Developing country actors may face too That each intends to do what's right little competition, often because of current institu- And treat each other "honor bright" tional structures. Changing this will improve the qual- How little to complain there'd be ity of other institutions. Competition among jurisdic- If I knew you and you knew me. tions-for example, among different states within a When'er we ship you by mistake, country or between countries-highlights successful Or in your bill some error make institutions and promotes demand for them. Compe- From irritation you'd be free If I knew you and you knew me. tition among firms and individuals does the same. Or when the checks don't come on time And customers send nary a line, This chapter first provides a framework for evaluat- We'd wait without anxiety, ing the role of institutions in supporting market trans- actions, growth, and poverty reduction. It then focuses If knew you and you knew me. on the four main lessons on institution building, fol- Source: Who's Who in the Grain Trade 35 (June 20, lowed by a discussion of the impact of political and so- 1922-23); cited in Bernstein 2001, World Development cial forces on institutional evolution. Report 2002 background paper. How do institutions support markets? Small vendors engage in simple spot-market transac- Yet not all institutions promote inclusive markets. tions, with buyers and sellers dealing face to face in The Maghribis lowered transaction costs among them- fairly standard products whose quality is easy to verify. selves, but in so doing excluded other communities. In- A rural vegetable market in a poor country is such a stitutional designs that evolve through either historical market. Large multinational firms exchange more dif- circumstances or directed action by policymakers are ferentiated products, facing greater difficulties in veri- not necessarily the best institutions for all society-or fying quality and bigger separations in time and space for economic growth and poverty reduction. Moreover, between the quidand the quo. International exchange institutions that once supported market transactions of food products is an example of such a market. Most can outlive their usefulness-for example, privatization economies have both types of markets-the first more agencies and bank restructuring agencies. The challenge common in developing countries, the second in indus- for policymakers is to shape policies and institutional trial economies. development in ways that enhance economic develop- Developed markets, more global, inclusive, and inte- ment. The Maghribis operated under a policy of free grated, offer more opportunity choice. Underdeveloped trade that enhanced their opportunities. It was to take markets, more likely in poor countries, are more likely advantage of these opportunities that they developed to be local and segmented. So, compared with farmers their institutions. in Canada, poor farmers in Bangladesh have fewer op- Clearly there is no unique institutional structure portunities-and far fewer formal institutions (such as guaranteed to lead to economic growth and poverty banks and formal courts) to reduce their risks and in- reduction. Large firms in the United States and the crease their opportunities. United Kingdom are often publicly held, with dis- What limits market opportunities? Transaction costs persed ownership, and are widely traded. But that is not from inadequate information, incomplete definition the case in other high-income countries such as France and enforcement of property rights, and barriers to or Canada, where ownership structures are highly con- entry for new participants.' What increases them? In- centrated (figure 1.1). And to promote competition, stitutions that help manage risks from market ex- policymakers can use quite different guidelines. In East change, increase efficiency, and raise returns (boxes 1. 1, Asia competition authorities consider a market share of 1.2, and 1.3). 50 to 75 percent to be evidence of possible monopoly 6 WORLD DEVELOPMENT REPORT 2002 Box 1.2 What are institutions? Institutions are rules, enforcement mechanisms, and orga- and public and private organizations operating under public nizations. This Report considers those institutions that sup- law. For example, organizations include firms operating port market transactions.2 Distinct from policies, which are under corporate law. Informal institutions, often operating the goals and desired outcomes, institutions are the rules, outside the formal legal system, reflect unwritten codes of including behavioral norms, by which agents interact-and social conduct. Examples include land inheritance norms the organizations that implement rules and codes of conduct and moneylenders using social networks to determine cred- to achieve desired outcomes. Policies affect which institu- itworthiness based on the reputation of the agents involved. tions evolve-but institutions too affect which policies are People in both rich and poor countries rely on informal adopted. Institutional structure affects behavior. But behav- institutions to facilitate transactions, but these institutions ior may also change within existing institutional structures. are relatively more important in poor countries where for- Institution builders can be diverse-such as policymak- mal institutions are less developed. Moreover, poor people ers, businesspeople, or community members. Corporate, in poor countries are often ill served by the limited formal collateral, and bankruptcy laws are public institutions, as are institutions available. In poor countries, and poor regions in the judiciary, tax collection agencies, and regulatory agen- particular, informal institutions substitute for formal institu- cies. Banks, reciprocity between community members, and tions (box figure). Countries and communities can go a long land inheritance norms are private institutions. Many private way toward resolving information and enforcement prob- institutions exist under the aegis of public institutions. Pri- lems without using their formal public legal systems. vate banks, for example, operate within the framework of public law. Social norms exist within (or without) formal laws. The enforcement of rules can be internal, implemented Informal rural credit in selected developing by the parties affected by the rules, or external, imple- countries, 1980s and 1990s mented by a third party. Informal institutions and private for- Percentage of total credit mal mechanisms generally rely on their own members for 100 enforcement. Individual agents organize themselves into in- formal groups, such as business associations (chapter 3) or mutual insurance systems (chapter 9) when the cost of col- 80 - lective action is low and the rules can be easily monitored. In these groups, expulsion from the community is a form of punishment. 60 * External enforcement mechanisms, such as judicial sys- tems or third-party arbitration, are critical mechanisms for the development of integrated markets. They allow access 40 - to market opportunities for a broader group of market par- ticipants. For external enforcement mechanisms to be ef- 20 fective, the legitimacy of the enforcer is vital. When the 20 state acts as an agent that shares the objectives and beliefs of its citizens-and implements rules consistent with 0 them-it is more likely to build effective formal institutions Nigeria India Nepal Sri Lanka Pakistan Thailand to support market development.3 11980s) (1981) (1995-96) (1981) (1991) (1980sa Effective institutions are those that are incentive-compati- Source: Kochar 1997; Besley and others 2001; Ijere 1986, cited in ble. Institutions with internal enforcement mechanisms are ef- Adegbite 1997; Mansuri 1998; Desai and Mellor 1993. fective because there is a mutually recognized system of re- wards and penalties. An important issue in the design of public institutions is ensuring that the incentives that are created ac- Networks, such as those of the Maghribis, that are tually lead to desired behavior. Take the example of deposit in- based on common ethnic, religious, and other common ties, surance, which is designed to protect depositors from the risks are closed groups; that is, entry into the group is limited. In inherent in financial institutions (chapter 3). Experience has such groups, costs of information processing and definition shown that deposit insurance can weaken the incentives of fi- and enforcement of property rights are lowered by mutual nancial managers to lend depositors' funds prudently and can ties or trust. Although these transaction costs are lower in lead to excessive risk-taking. In circumstances like this, com- closed groups, the informal and norm-based institutions that plementary regulations are required to realign incentives, such such groups rely on tend to support a less diverse set of ac- as regulations to ensure that bank managers have a significant tivities than do formal legal institutions. As countries de- financial stake in bank performance. velop, the number and range of partners that market partic- ipants deal with increases and market transactions become Informal andformal institutions more complicated, demanding more formal institutions. Formal institutions include rules written into the law by gov- Conversely, public or private agents may build formal insti- ernment, rules codified and adopted by private institutions, tutions to enable a more diverse set of activities. BUILDING INSTITUTIONS: COMPLEMENT, INNOVATE, CONNECT, AND COMPETE 7 Box 1.2 (coritinueci] Legislators may purposefully base formal law and judi- state power, overtaxation, corruption, short time horizons, cial practice on social norms. In some cases this may con- cronyism, and the inability to uphold public order. For exam- sist of simply codifying and modifying existing practices and ple, governments may establish restrictive trading rules in re- writing them into law (Bernstein 1999). But this is not sim- sponse to lobbying by business monopolies intent on safe- ple, particularly in heterogeneous societies. Choosing how guarding their monopoly interests. The balance between to weigh each group's norms and standards is critical in de- markets and state power, and between business and social termining not just the efficiency impacts, but also legitimacy interests, is a delicate one in the course of institutional devel- and distributional implications.4 For example, in multiethnic opment. Historically, the government's role in the protection Uganda, English was adopted as a neutral common lan- of property rights and the provision of other public goods has guage for the formal functions of the state. Such concerns been closely linked to its role in ensuring peace or law and extend to standards or rules in international markets as well. order. Conflicts over property between private agents, and Ideally, informal and formal institutions should com- between the state and private agents, are some of the most plement each other. Together, they can reduce transaction important issues that governments have had to deal with, be- costs more than either can alone. Formal courts, for example, cause they often lead to a breakdown of law and order. deter litigation and facilitate informal settlement simply by Market development and private business flourish when providing the threat of enforcement (chapter 6). Far more dis- the behavior of those who govern is not arbitrary (see box putes arise in business transactions than go through a formal 1.3 ). For example, detailed analyses of the evolution of cor- dispute resolution process (Bernstein 1999). porate law in several countries show that in the early stages of development, private business was typically subject to Putblic versus pritwte roles the arbitrary whims of those in power. The state, with pri- Governments have an important role in providing public mary control rights, granted the permission to incorporate goods, such as laws that delineate property rights and the ju- case by case (Pistor and others 2000). At later stages, the dicial institutions that enforce these rights and establish the right to incorporate was no longer a personal favor but was rule of law. But governments have been known to impede granted to any entrepreneur that met a set of predeter- the development of markets through arbitrary exercise of mined conditions. Box 1I3 Institutional evolution and economic development: private traders and public rulers In medieval Europe, the political power of local rulers was ex- among traders, and increased problems of information and en- tensive. Local rulers could confiscate the property of individual forcement. Growth meant trading with members from other traders from other regions without incurring penalties. In re- social and ethnic backgrounds, which meant that social con- sponse, private mercantile guilds evolved to promote trade and nections could not easily be used as a basis for information or to guard against the arbitrary action of local rulers. These guilds enforcement. established agreements with merchants in foreign cities and Members no longer wanted to be collectively responsible with local authorities themselves. Arbitrary confiscation was for individual breaches of contract. So leaders pushed for an punished by the withdrawal of large amounts of business by enforcement and sanctioning system based on individual re- the guild, and so local rulers were forced to respect the rights sponsibility rather than community responsibility. To the ex- of its members. This change in the balance of power helped tent that community growth implied more intracommunity so- to promote the security of foreign traders. cial and economic diversity, it also reduced the political viability In the 12th century, traders in Europe established commu- of the community. But the extent to which communities could nity-based mechanisms to facilitate the exchange of credit and abolish community-based mechanisms depended on a reliable trade across borders. These mechanisms were based on the third party to enforce contracts. In England, the monarch per- community accepting responsibility for the performance of its formed this role, and in 1275 King Edward I issued a statute members vis-a-vis other communities. For example, when a outlawing community responsibility for debts. Genoese merchant defaulted on a loan from a merchant in The example illustrates a general principle: as economies London, community leaders in Genoa were responsible for en- grow and develop, different types of institutions are needed forcing the contract by imposing sanctions on the defaulter. to facilitate transactions. Many different actors can push for Community origin was easily established, meaning that repu- new institutions. But the role that the state plays depends tation within the community was important, and agents could on its capacity and political viability: a strong state that re- be trusted not to renege on their contracts. spects the law itself and refrains from arbitrary action is a criti- As cities grew in size and number, so did the communities cal factor. of merchants and traders, making collective action more diffi- cult. Unrestricted entry into trading led to more competition Source: Greif 1997a. 8 WORLD DEVELOPMENT REPORT 2002 Figure 1.1 rights to assets and income and being able to protect The concentration of ownership varies those rights are critical for market development. tremendously across countries These include the rights of the private sector in rela- tion to the state. Institutions can reduce the poten- Ownership concentration tial for disputes and help enforce contracts. Examples 0.58 - , include a country's constitution, its judicial system, 0.58 - * ++ + , and the full array of social networks. * * + + + * They increase competition in markets-or decrease it. 0.48 - Competition gives people incentives to do better and * ' + promotes equal opportunity. In competitive markets 0.38 -0 resources are more likely to follow the merits of a project than the social or political connections of an 0.28 - entrepreneur. The degree of competition also affects *.18 ,1 I I t I# innovation and economic growth (chapters 2 and 7). 0.18 u l , , + I I a But while some institutions facilitate competition, 5.5 6.5 7.5 8.5 9.5 10.5 Income others impede It. For example, by overregulatmg the (logarithm of per capita income} entry of new business, governments can constrain Note: Ownership concentration is measured by the combined stakes competition. And by organizing market activities of the three largest shareholders in the 10 largest privately controlled around a closed group of participants-recall the firms. Source: La Porta, Lopez-de-Silanes, and Shleifer 1999. Maghribis-outsiders will find it harder to compete _____________ _ even while opportunities for those in the group may increase (chapters 3 and 9). power, whereas in Africa the range is 20 to 45 percent. Within South Asia some farmers rely on cooperatives The transaction costs of acquiring information, en- to market their goods; others use informal contracts forcing property rights, and restraining competition with private traders. can prevent the emergence of inclusive markets. But ef- This Report provides a framework that applies across fective institutions can reduce those costs. Consider the the range of market-supporting institutions. It cuts following example. If the quality and value of the grain through the complexity and diversity of institutional that traders buy from a farmer cannot be easily deter- structures by focusing on what institutions do. Under- mined, and if traders have little information about a standing what they do is the first step in building effec- farmer, they have to inspect each bag of grain to assess tive institutions. Institutions do three main things:5 quality. Traders also provide credit to farmers. But if traders have little information on the ability of farmers * They channel information about market conditions, to repay the debt-or if farmers cannot use the assets goods, andparticipants. Good information flows help they own as security-providing credit is risky. These businesses identify partners and high-return activi- problems are magnified for smaller and poorer farmers. ties-and assess their creditworthiness. Information The trader may impose higher interest rates on poorer about businesses helps governments regulate effec- farmers, and the farmers may be more likely to default tively. Institutions can affect the production, collec- than if they were exposed to competition.7 tion, analysis, verification, and dissemination-or the Through these three functions, all institutional struc- withholding-of information and knowledge. They tures affect the distribution of assets, incomes, and costs do this for participants in, and between, communi- as well as the incentives of market participants and the ties and markets. Examples include accounting firms efficiency of market transactions. By distributing rights and credit registries, which facilitate information pro- to the most efficient agent, institutions can enhance pro- cessing, or government regulations on the media, ductivity and growth. By affecting the incentives to which restrict the dissemination of information.6 invest-for example, through strengthening property * They define and enforce property rights and contracts, rights-they can affect investment levels and adoption determining who gets what and when. Knowing one's of new technology. By delineating market rights, such BUILDING INSTITUTIONS: COMPLEMENT, INNOVATE, CONNECT, AND COMPETE 9 Box 1.4 Box 1.5 Courts and the expansion of trade Weak institutions hurt poor people Studies of manufacturing firms in eight African countries Mounting evidence shows that the poor bear the greatest demonstrate the supporting role institutions play in mar- burden of institutional failure. Take corruption, a highly re- ket development. These country studies show that the ab- gressive tax. Demands for bribes and unofficial fees for sence of effective public dispute resolution mechanisms services hit poor people hardest. In far too many cases in cases of breach of contract has limited the expansion legal systems and the judiciary fail to serve poor people. of trade and market development. Courts tend to be slow Their illiteracy and inability to pay for legal representation and inefficient. The absence of formal contract enforce- put formal legal institutions beyond reach. The failure of ment mechanisms has limited the growth of firms and the the state to protect property also hurts the poorest dispro- development of financial institutions. The small scale of portionately, because they cannot afford to protect them- the formal productive sector has, in turn, prevented the selves from crime. And badly designed regulatory institu- development of complementary institutions. tions reduce the provision of infrastructure to the poorest Another study analyzing six countries in Africa (Bu- in society. rundi, Cameroon, C6te d'lvoire, Kenya, Zambia, and Zim- World Development Report 2000/2001 stressed that babwe) shows that among these countries, the presence poor people are often more vulnerable than others to of a more developed legal system encouraged firms to macroeconomic crises and natural disasters. Market insti- undertake riskier activities because well-functioning legal tutions that support growth of overall incomes can reduce systems helped to adjudicate and settle disputes that their vulnerabilities to shocks and help them insure against arose from such market activities, bad times. Some of the institutions discussed in this Report have an important and direct role in this. For example, finan- Source: Bigsten and others 2000; Collier and Gunning cial institutions help mitigate their risks, allowing individuals Source: Bigsten and others 2000; Collier and Gunning to diversify their savings and risks and allowing them to smooth their consumption over good times and bad. Source: World Bank 2000d. as through competition law, they limit producer rents and protect consumers from high prices. And by clarify- overlapping sets of institutions. For example, the success ing rights for the disadvantaged in markets, institutions of the state in providing laws and the performance of the can directly affect the lives of poor people. For example, judiciary and police reflect whether citizens and in- giving formal titles to poor people whose occupancy vestors perceive the state as respecting property rights. rights were not recognized by lenders allows them to Access to financial services and the sophistication of fi- borrow and invest. nancial markets reflect how successfully institutions pro- tect the property rights of borrowers and lenders. High How do Institutions support growth Hnw dovinrtitrductions supportgrowtlevels of public corruption reflect how the behavior of public agents in state institutions responds to the types Institutions that support market transactions can thus of incentives that exist for politicians and civil servants affect poor farmers in Latin America as much as they af- to pursue the public good over their self-interest. fect wealthy businessmen in Canada. Country case stud- Positive relationships between economic develop- ies, as well as cross-country empirical work, provide im- ment and these indicators of institutional success have portant insights into institutional development and been widely documented. But most studies do not es- market development (box 1.4). They confirm how mar- tablish links between specific institutions and specific ket-supporting institutions affect people's lives by influ- outcomes. Instead, they highlight the wide variety of encing growth, determining people's access to markets, institutions that support markets. For example, income and enabling poor and rich people to make the best use and the rule of law-encompassing the collective im- of their assets. Moreover, weak market-supporting insti- portance of property rights, respect for legal institu- tutions can hurt the poor disproportionately (box 1.5). tions, and the judiciary-are highly correlated. For an- A growing body of research links institutional success other example, the development of financial institutions (and failure) to economic growth and market develop- predicts growth (figure 1.2). ment over time and across countries. A wide range of in- On institutional development and economic growth, dicators captures the performance of different, often important differences have been found between coun- 10 WORLD DEVELOPMENT REPORT 2002 Figure 1.2 institutions for conflict management.9 In divided soci- Financial depth generates growth eties, such as those with ethnic fragmentation or high inequality, low-quality institutions for managing con- Ratio of liquid liabilities to GDP, 1960 flict-including low-quality government institutions Deep (greater and inadequate social safety nets-magnify external than 0.5) shocks, triggering distributional conflicts and delaying policy responses. Prolonged uncertainty in the eco- 0.25toO.5 nomic environment and delayed policy adjustments curtail subsequent economic growth. How do you build effective institutions? 0.15 to 0.25 Recalling the framework of information, enforcement, and competition, policymakers building institutions first need to assess what is inhibiting market develop- Shallow (less ment or leading to certain market outcomes (box 1.6). than 0.15) Rather than focusing first on specific structures, they l l l l need to focus on the functions that are missing and de- 0 1 2 3 termine why. Policymakers need to ask: Average GDP growth 1960-95 Note: Figure based on partial scatter from the instrumented cross- * Who needs information on what? For example, do sectional regressions in Beck, Levine and Loayza 2000. bankers lack information on the creditworthiness of potential borrowers? * Are everybody's property rights and contracts clearly de- tries that once were colonies and are now industrialized fined and enforced? For example, do farmers have en- and former colonies that are still developing. Both groups forceable rights to land they use? trace key features of their institutions to former settlers. * Is there too little competition-or too much? For exam- A big part of the difference in later institutional develop- ple, is an infrastructure monopoly inhibiting entry ment-and its impact on growth-is the effort of set- or are firms not undertaking high-return research be- tlers in establishing well-functioning legal institutions.8 cause they lack safeguards on intellectual property? In the United States and New Zealand, colonizers settled in large numbers and transplanted institutions Once the institutional gap is identified, the next step common to, and understood by, the general populace, is to design the appropriate institution. Both supply and mostly new immigrants. In such countries the trans- demand factors are important. Moreover, as countries planted legal institutions were widely used, adapted to change and develop, so will the appropriate institution. local circumstances, and changed with economic de- To be effective, such an institution must be designed so velopment. Developing countries on every continent that the incentives of market actors are aligned to also received formal legal systems, transplanted by col- achieve the desired outcome. Four key approaches to- onizers. But their indigenous populations had little ac- ward institution building hold across all sectors and cess to or understanding of these legal systems. So the countries: complement what exists, innovate to identify institutions were not adapted to local circumstances. institutions that work, connect communities through Cross-country evidence suggests that the quality of in- information flows and trade, and promote competition. stitutions that support growth and poverty reduction through market development is lower in these coun- Completnent what exists tries than in the former group and has therefore not Developed market economies have institutional struc- supported economic growth and poverty reduction to tures that depend heavily on a capable state-a pro- the same extent. vider of public goods, a regulator, and an adjudicator. Institutions also affect how countries deal with con- But the involvement of the state in markets must be flict. A recent study found that growth and poverty out- consistent with its capacity. World Development Report comes in Asia, Latin America, and Sub-Saharan Africa 1997 emphasized matching the capability of the state since the mid-1970s have depended on the quality of with the tasks that government organizations take on. BUILDING INSTITUTIONS: COMPLEMENT, INNOVATE, CONNECT, AND COMPETE II Box 1.6 Who builds institutions? Institutional reform is not just the preserve of national govern- mestic reform in Mexico. Some countries in Eastern Europe ments. Individuals and communities, local entrepreneurs, multi- are implementing wide-ranging institutional reforms as they national companies, and multilateral organizations can build strive to become members of the European Union. institutions, often in partnership with each other. National gov- Recent developments surrounding the AIDS crisis illustrate ernments may initiate reform or may simply respond to pres- how different groups may affect the process of institutional sures from the private sector or from external actors. change. In some cases of systemic institution building, govern- ments have been effective in successfully transplanting laws, Many agenits of ehange: healtlh crises andpatents organizations, and agencies. In other cases systemic reforms More than 95 percent of HIV/AIDS cases are in developing did not have the desired outcomes. The contrast between countries. But the average cost of the antiretroviral treatments, Poland and Russia is instructive in this regard. Poland had a which have reduced AIDS mortality by 70 percent in industrial more recent history of a market system, and Polish policymak- countries, is still more than $10,000 a year, far beyond the ers and business people had a better understanding of the req- reach of most people in poor countries. uisite institutional framework. Polish reforms focused on clari- Some developing countries-Brazil, India, South Africa, and fying property rights between the state and private actors-for Thailand-have taken steps to reduce the cost of AIDS treat- example, by imposing hard budget constraints on public firms. ment through the design and application of their intellectual Russia did not have a recent history of market development, property rights laws-an international institution-to allow com- and reforms did not initially have the desired effects, partly be- pulsory licenses permitting the production of generic drugs and cause there was no clear delineation between private and pub- the import of cheaper generic drugs. In Thailand generic drugs lic institutions. Firms were not immediately exposed to hard became available at just 10 percent of the price of the patented budget constraints, as shown by widespread arrears in taxes product. and other payments (Recanatini and Ryterman 2000). These measures led to threats of trade sanctions and law Institution building at the sectoral level has also met with suits from the drug manufacturers. But collective action, initi- varying success. In Tanzania and Zambia the public sector in- ated by international agencies and NGOs, helped increase tervened in agricultural marketing with the stated aim of stabi- access to AIDS drugs by enforcing existing public health safe- lizing farmer incomes. In most cases these reforms failed- guards, permitted under the Trade-Related Aspects of Intellec- leading to lower marketable output and often corruption. tual Property Rights agreement but not previously imple- Worse, the experiences affected perceptions of the overall in- mented. The news media were instrumental in publicizing the tegrity of public institutions. Successes include the reform of disparities in the availability of AIDS treatment and promoting business registration in Bulgaria, now conducted online and public debate on the issue. As a result, the U.S. government taking around two days, not three weeks as in the past. retracted its trade sanction threats. And pharmaceutical com- Local business interests, the foreign business community, panies agreed to reduce prices-and more recently to drop a nonprofit organizations, the media, and international organiza- lawsuit on intellectual property rights against the South African tions have all been involved in direct institution-building efforts government. in developing countries. For example, membership in the North American Free Trade Association has hastened the pace of do- Source: Perez-Casas and others 2000. This Report builds on that analysis by examining how a Costs, relative to per capita income, of establishing existing information, enforcement costs, and the cost and maintaining institutions of building and maintaining institutions affect the way * Administrative capacity, including human capabilities governments support private transactions in markets. a Technology. It also examines how market development is affected by the extent to which government actors themselves Both existing and newly transplanted institutions respond to the institutions they build. As countries de- can be more effective in poor countries if they are sys- velop, the types of institutions they need and demand tematically modified to take these differences into ac- also change. count."0 This may sometimes mean changing priorities One of this Report's messages is that institutions in terms of which types of institutions to build first, that work in industrial countries may not produce sim- and whether to build at all at a given time. ilar results in poorer countries because of differences in: Complementary institutions. Government interven- tions can reduce many market failures, but governments * Complementary institutions, such as those promot- may also fail in trying to support market transactions. ing transparency and the enforcement of laws For example, governments may impose regulations to * Existing levels and perceptions of corruption try to compensate for market failures or as a way of re- 12 WORLD DEVELOPMENT REPORT 2002 stricting private activity. Choosing between market fail- Figure 1.3 ures and potential government failures is not easy, but Complexity of procedures in debt collection measures can be taken to limit both. However, the lim- ited capacity of developing country governments to im- Economy plement regulations means that many activities in Philippines poorer countries are overregulated. Peru For regulatory systems in developing countries to - have a realistic chance of success, they need to be sim- Senegal pler, often less information-intensive, and less burden- Mozambique some on the courts. Many developing countries, how- Bulgaria ever, despite their weaker judicial systems, tend to have Ecuador very complex debt collection procedures (figure 1.3). Venezuela, Regulations in industrial countries can also be very R.B. de complex, but they do not impose as many additional Argentina costs as they do in poorer countries-for several reasons. El Salvador Enforcement capacity in richer countries is stronger, and Ghana judges may face other incentives that affect their perfor- Kazakhstan mance and judicial efficiency (chapter 6). Regulators are more accountable, and complementary institutions (such Sample average as those affecting judges' wages or careers, or those which Japan promote transparency) provide checks and balances to Canada protect market participants. In developing countries, Sweden where there are fewer supporting institutions (for exam- ple, where courts are weak or lack credibility), one solu- United States tion is to write simple rules and have fewer of them. Switzerland Where informal institutions operate effectively, and Germany when formal institutions require supporting institu- Australia tions, building new formal institutions may not be a Hong Kong, priority for policymakers. China Finland * Studies of land titling in various countries show that New Zealand formal titles may not have the desired effects when United Kingdom input, output, and credit markets and institutions are underdeveloped and the demand for agricultural Increasing complexity of litigation goods is low (chapter 2). In such cases traditional community-based mechanisms are more effective in Note: For the definition of complexity see chapter 6. The sample average is based on 96 countries. delineating property rights. Source: Survey done for World Development Report 2002 in * Corporate governance is difficult in poorer countries conjunction with Lex Mundi, an international association of law firms. because of weak legal systems and the lack of private information intermediaries. In this situation concen- trated ownership structures-and business groups countries, the financial cost of complying with regula- and associations-may provide more effective cor- tions for registering a business is very high relative to porate oversight than dispersed ownership structures. per capita gross national product (GNP) (figure 1.4a) and higher than industrial country averages."t Surpris- Costs, capacity, and corruption. The cost of govern- ingly, developing countries that have less administra- ment regulation, whether in financial or other terms, tive capacity also require more procedures to register a needs to be consistent with a country's per capita in- business (figure 1.4b). The high cost, whether in com- come to be effective. For example, a recent study cov- plexity or resources, deters entry into the formal sector, ering 85 countries found that in many developing potentially reducing competition and incurring ad- BUILDING INSTITUTIONS: COMPLEMENT, INNOVATE, CONNECT, AND COMPETE 13 Figure 1.4a Box 1.7 Cost of business registration (as percentage of Human capital and institutional design GNP per capita) is higher for lower-income countries Human capital affects the quality of the rules that govern market transactions and the enforcement of these rules. Cost of registration (NO Literacy levels and technical skills vary greatly across and 140 - within countries. The poorest economies of the former So- 120 - viet Union have income levels lower than many countries in Asia and Africa but nearly universal primary education. 100 _ So literacy is less of a barrier for Armenians using formal in- 80 - stitutions than it may be for some Angolans-and it is less 60 - of a problem for today's Malaysians than it was for those of a generation ago. The rules and organizations that gov- 40 - ern markets have to allow relevant market actors to use 20 - them easily. This argument holds within countries as well- for example, across poorer rural and richer urban areas. ° ' S 2nd 3rd 4th l The usefulness of institutions also depends on the ca- pability of their administrators. Judges untrained in corpo- Income quartiles rate law and accountancy, for instance, may not be the best arbiters of bankruptcy cases. Successful institution Figure 1.4b builders have had either to tailor institutions to prevailing Lower-income countries have more procedures administrative capacity (using, for example, simpler bank- ruptcy rules) or to complement institution building with a Average per capita GDP (thousands of U.S. dollars) strong focus on concurrently developing technical exper- 25 - tise for administrators (from accountancy skills to regula- tory economics). 20 - 15 - ditional costs in the form of increasing corruption 10 - (figure 1.4c).2 A World Bank study also finds that in 5 - many African countries, restrictive regulations and _I I_ - practices are often aimed at generating rents for offi- o 0 _J I _ S _ | - cials and favored private agents or groups, constraining 1-5 6-10 11 15 16 and more business activity in both agriculture and industry."3 Number of procedures Since building institutions is costly, requiring a min- Figure 1.4c imum threshold demand before they can operate effi- More procedures are associated with higher ciently, small countries can face problems. Small coun- corruption tries and those countries wishing to expedite access to Index of corruption (higher = worse} institutions may wish to rely on foreign institutions- 90 - such as foreign banks or foreign stock market listings- 80 - rather than build supervisory and regulatory capacity 60 - at home (chapter 4). Hungary and Estonia, for exam- 50 - ple, encouraged the entry of foreign banks, supervised 40 - and regulated in their country of domicile. 30 - Human capability. More human capital may be 20 - needed to use some market institutions-such as for- o u- * | | | * S *mal judicial methods to resolve disputes-and to ad- 1-5 6-10 11-15 16 and more minister regulations or develop standards (box 1.7). For Number of procedures example, competition authorities need people who un- derstand the complex details of competition cases. As Note: Costs are defined as official fees as a percentage of 1999 GNP countries build human capabilities, they need to con- per capita. sider where to focus their attention. Human capital and Source: Dlankov and others 2001, World Development Report 2002 background paper. the array of market institutions In an economy have a dynamic relationship. Agents need human capital to 14 WORLD DEVELOPMENT REPORT 2002 benefit from certain institutions. And over time, as Box 1.8 agents learn, institutions need to be adapted. As can be Computerization and land registration in seen from the experience of East Asia, actively promot- Andhra Pradesh, India ing literacy and primary education can have a big pay- off in the eventual quality and success of formal insti- Buying property in Andhra Pradesh used to be complex and take a long time. After the purchase the buyer visited tutions, as both users and administrators are more able the local office of the Sub-Registrar of Assurances in per- to work with market institutions. son, had the property valued and stamp duty calculated, Technology. Infrastructure regulation shows that purchased stamp paper, and had a writer draft the deed in technical standards used in industrial countries may be the requisite legal language. The purchaser also had to pro- vide additional documents related to income and other inappropriate for developing countries (chapter 8). In properties owned. All these documents were then scruti- poor countries service providers using low-cost tech- nized by the registrar, and recorded, before an exact copy nology often operate in the informal sector for parts of of the final deed was copied by hand and certified. society not reached by formal operators. Regulators are In Andhra Pradesh, 387 subregistrar offices registered about 1.2 million documents a year, 60 percent of them typicalyhostletonforma providers.Busomedevelfor agricultural land. A yearly manual update of property in- oping countries recognize the benefits of allowing those formation was carried out, since hundreds of thousands providers to operate. In Paraguay about 400 private of property files were updated with the new sales from water suppliers operate their own wells and provide the year. Land registration offices throughout the state are now piped water to households unserved by the public sec-eqipdwtcouerzdonesudrthCi- tor. Imposng strictstandardson providrsuigim equipped with computerized counters under the Com- ror. Imposing strict standards on providers using sim- puter-aided Administration of Registration Department pie technology would immediately drive these private (CARDi project, initiated and financed by the state govern- suppliers out of business. A more gradual evolution in ment to improve efficiency and increase duty collections. regulation is needed.14 Starting with a pilot project in 214 locations over 15 months, Countries do not have to go through a long learning- the entire database was transferred to computers, the Countres do not have to go through a long learnng- copying and filing system was replaced with imaging, and by-doing process in all aspects of institutional develop- all back-office functions were automated. Standardization ment. They can transplant and modify some insti- and greater transparency in property valuation procedures tu tional forms from other countries and shorten the boosted stamp duty revenues. Registration processing time was cut from 10 days to 1 hour. development process by learning from other countries. They can also use Internet technology to reduce insti- Source: Case study by Dr. Subhash Chandra Bhatnagar, tutional constraints and improve the effectiveness of in- University of Delhi. World Bank 2000, as part of the stitutions. In many developing countries the Internet E-Government Focus Group, available at http://wwwl. is already providing the means for accelerated learning, worldbank.org/publicsector/egov. improved information flows, reduced enforcement costs, and enhanced competition in markets (box 1.8). But to leapfrog stages in development through technol- distributional effects, can discriminate systematically ogy, policymakers need to increase access to technology against poor countries. For example, the Trade-Related for market agents. Market rules affect access. Aspects of Intellectual Property Rights Agreement International rules and standards. Standardizing laws (TRIPS) can impose significant costs on poor countries, and regulations generally reduces the information and because strong patent protection is not as appropriate enforcement costs of transactions across borders and can for them as it is for rich countries. Many industrial enhance trade efficiency. International standards also countries themselves only recently adopted laws safe- have the potential to provide benefits much larger than guarding intellectual property, and the nature of these those under bilaterally agreed standards between coun- laws has evolved over time in response to changing do- tries for both poor and rich countries. But depending mestic economic and political factors. Developing coun- on which standards are chosen, international standards tries also lack the supporting institutions to implement can also be costly for poor countries and can have sig- TRIPS effectively-these will take time and resources nificant distributional consequences between countries. to build. International trading rules and principles, enshrined Another example is the adoption of international ac- in the World Trade Organization (WTO), promote trade counting rules by companies in many developing coun- (chapters 5 and 7). But some standards, through their tries. This has enhanced their access to credit in inter- BUILDING INSTITUTIONS: COMPLEMENT, INNOVATE, CONNECT, AND COMPETE I5 national markets. Voluntary adoption of standards by Sicw- E %. firms wanting to obtain credit in international markets Private innovation supported by formal is likely to be beneficial. But these standards are not ap- -nstitutioal change_ propriate for smaller firms (chapter 3). And forcing In Bangladesh an economics professor had an idea-to small firms in developing countries to adopt them help poor people help themselves by giving them small would raise their costs and possibly push them into the loans to start businesses despite their lack of collateral or informal sector. credit histories. He started the Grameen Bank in 1976 For international standards to truly benefit all coun- using his social connections in government to manage a tries by facilitating trade-and to avoid systematic bi- village branch of a government bank. The success of this endeavor, followed by expansion to other bank branches, ases against developing countries-the standards need led the government to eventually change the laws govern- to reflect realities in developing countries. These include ing the Grameen Bank. It was established first as an inde- the costs of adhering to standards as well as the benefits, pendent entity with government control, then as an effec- and particularly important are the costs imposed on the tively private bank run by a public official, and finally as an effectively private bank run by a private individual and an poor. Important questions are: W hose standards should independent board of directors. Today, Grameen Bank has be adopted and why, and what is the process under branches in more than half the villages in Bangladesh and which these standards are negotiated? The process of re- more than 2 million borrowers. forming international rules needs to be transparent, and In Peru another innovative individual began with an ex- be active participants to periment. He found that in Lima it took 728 bureaucratic developing countries need to be acuve partlclpants to steps for a person with an informal right to housing to get influence outcomes in their favor. But human capital legal title. He followed up with a 10-year public informa- constraints may prevent developing countries from rep- tion campaign, proving to politicians that there was a "hid- resenting their interests. In such circumstances interna- den consensus for reform" for simplifying the procedures tional donors could help enhance their representation, for formalization. Faced with overwhelming public support for simplification, the Peruvian congress unanimously or developing countries could pool their scarce technical passed legislation to formalize titles. Today, a simple legal skills and have common representation at international procedure for establishing land titles for poorer people negotiations or hire private specialists to represent them. works in parallel with the formal system. Variation within countriers. Some variation in insti- These two stories show how the state can work with Variationsmaybe w sitn coutri Somhefvaiaincy in ist.i- private actors to promote institutional innovation by directly tutions may be desirable for both efficiency and distri- supporting experiments-or at least by allowing them to butional reasons, even between regions within coun- proceed and be tested and then, if they are successful, tries. Even industrial countries do not standardize all by encouraging their growth. The stories also show the im- laws and regulations within the country. For example, portance of other factors in promoting innovation. Social connections and networks can reduce barriers to experi- Australia and Canada have different laws in different mentation. Openness in information sharing provides the states for secured transactions. Different states in the impetus to adopt and expand successful experiments. United States have different corporate laws. The differ- ences exist because of variations in economic and social source: De Soto 2000; Yunus 1997. structures-variations that can be particularly instruc- tive for large countries such as Brazil, China, India, and the Russian Federation. Of course the costs of standard- however, and these need to be balanced against poten- ization versus diversity will vary depending on the in- tial benefits. stitution and the relative distribution of gains and Policymakers can replicate successful local innova- losses. Where spillover effects across jurisdictions are tions. But they also need to be flexible enough to drop large and not sustainable at the macroeconomic level, unsuccessful experiments. Because innovation can come variation may be less desirable. from many sources, collaboration by the different ac- tors in society is vital, as shown by the development of Innovate to identify5 itnst.ittion7s that work microfinance institutions in Bangladesh, where the Even at similar levels of development, countries differ government adapted its formal legal structure to ac- in many ways-in their norms, geography, and endow- commodate private innovation, and the process of land ments. Innovation, often through experimentation, can titling followed in Peru (box 1.9). help accommodate those differences and produce more In some cases, greater local autonomy and participa- effective institutions. Experimentation also has costs, tion may foster institutional experiments that lead to in- i6 WORLD DEVELOPMENT REPORT 2002 novation. For example, Aguas Argentina, a privatized Box 1.10 monopoly that provides water and sanitation services Distributional effects of innovation depend on in Buenos Aires, worked under a novel institutional who innovates: bankruptcy law in two countries arrangement to design new ways to organize service de- livery. The monopoly worked with local government, a The United Kingdom created its bankruptcy regime livery. T mn lwthrough explicit legislation that recognized the importance low-income community, and a nongovernmental orga- of decentralized contracting: legislation stated that corpo- nization (NGO) to create a new organizational form to rations were free to make the rules under which they improve service delivery. The community was experi- would transact. Lenders and borrowers in the United King- menting with two systems: a low-cost sewerage system dom had the power to innovate through contracts, and aenting a . , over time commercial practice was incorporated into law. and a double water system (with one connection to the In the United States judges and legislators held that power. network for small volumes of potable water and another The U.K. system, designed by private agents engaged drawing on groundwater sources too salty for drinking in borrowing and lending, is today characterized by a great but good for washing and bathing). The double water concentration of rights in favor of the principal lender. The principal claimant appoints a receiver who uses his powers system was dropped at the experimental stage because it for the sole purpose of repaying this principal's debt. The was too expensive to develop, while the sewerage system court's role is much less significant than it is in the United was maintained. To expand its water network, Aguas States, and the judgment is not subject to court review. Argentinas took over those systems built at lower cost In the United States, Chapter 11 bankruptcy law is the community, giving customers a discount on the characterized by a partial dispersion of rights away from by tne communlty, glvmg customers a alscount on tne secured claims (priority lenders). U.S. legislation was price in exchange. In effect, it had contracted out con- amended several times at moments of economic crisis at struction to the community.15 the instigation of the judiciary. At these times preservation Innovation through experimentation can happen at of companies rather than their dissolution was uppermost different,levels..Experimentation and innovation occur in the minds of legislators and judges-leading to a debtor- friendly bankruptcy law. Upon default a company in the on at least three levels: national public policymaking, United States may seek protection from its creditors, usu- private commercial practices, and local action by com- ally retaining control over the business. munities and civil society leaders (see box 1.16). Local experimentation has the advantage of allowing many innovations to be tried simultaneously-with the suc- cessful ones replicated and the failures contained. But not all innovations can be left to local or decentralized novation.16 The tension between experimenting and communities-since local actions may have conse- standardizing public institutions within countries will quences across communities and too much experimen- be settled in favor of the latter when effective institu- tation can lead to each community having different tional forms are found. Policymakers have to ensure that rules. Local innovation can also open institutions to successful local innovations can be scaled up. They must capture by local elites, inviting corruption. When ef- also be willing to drop outdated institutional forms. fective innovations are identified, policymakers can help Hungary, in the early years of its transition, for exam- expand such institutions by replicating them in other ple, experimented with a particular form of bankruptcy areas (for example, through adopting a law) or by shar- law, which was later dropped when conditions changed ing information on the innovation. and a more effective alternative emerged (box 1.11). Who innovates determines institutional evolution. De- pending on who innovates, institutions can evolve in Connect communities through information flows quite different ways (and with quite different distribu- and trade tional consequences), as shown by the evolution of Open information exchange and open trade promote bankruptcy law in the United Kingdom and the United institution building by creating demand for market- States (box 1. 10). As history shows, during the devel- supporting institutions. opment process the institutions adopted favor those Open trade. Going beyond allocative efficiency, open who control the process. trade does more. Debates among people who formulate policy, those who implement it, and those outside government can * It exposes market participants to a larger, more di- help in disseminating information on institutional in- verse, group of trading partners, increasing the de- BUILDING INSTITUTIONS: COMPLEMENT, INNOVATE, CONNECT, AND COMPETE 17 Box 1.11 Box 1.12 Experimentation and adaptation: Trade and institutional change in Thailand bankruptcy institutions in Hungary In the early 19th century, with labor scarce and land abun- In 1992 the Hungarian government adopted a bankruptcy dant, land had little value in Thailand. Slaves rather than code giving creditors very strong rights to file for bank- land were taken as collateral in financial markets. Corre- ruptcy. The intention was to impose a hard budget con- spondingly, land markets were underdeveloped. There straint on firms, particularly on large enterprises. The law was little demand or need for the development of formal therefore stated that the creditor could file for bankruptcy institutions. But there was a well-developed legal system if a company was three months or more overdue on any to govern transactions in labor commitments. In theory, all debt (known as an automatic trigger). Since accounting land belonged to the king. In practice, individuals could use systems were underdeveloped, information on the true and sell the land, as long as they paid taxes and did not let performance of firms was not readily available, and the it lie fallow for more than three consecutive years. available information was not always reliable. The solvency In the latter part of the century, international trade or insolvency of a firm was therefore hard to measure. opened up, and transport costs declined. A rice export The short time frame and the establishment of such a boom led to a rapid expansion of production and use of strong trigger for bankruptcy proceedings led more than land. Land became more valuable, land disputes more 5,000 firms to file for bankruptcy. The government had not common. The demand for formal institutions, such as reg- expected such a large number of bankruptcies, particularly istries, to convey information and enforce property rights of small firms. The automatic trigger allowed the govern- increased. ment to assess quickly the true condition of firms. But be- The government responded by implementing a series cause the courts dealt with so many cases, they quickly of procedural and administrative changes, beginning in developed experience in handling bankruptcies. The re- 1892. The first initiative, to document land rights, was sult: the authorities abolished the trigger in 1998. Not only modified and improved several times; the final legislation were courts better able to adjudicate bankruptcies, but was passed in 1954. The current legislation is a compro- better information systems had developed to allow credi- mise between traditional practice, which allowed citizens tors to monitor companies. Market dynamics and support- to bring unoccupied forestland under cultivation as private ing institutions had evolved enough so that the law was property, and the more formal requirement of land titling no longer needed. based on detailed land surveys. Such institutional evolution is not unique to Thailand, Source: Gray and others 1996. for industrial countries have also shown that trade, by changing the terms of trade, gives rise to the demand for clear property rights and a need for the state to define them. mand for formal institutions to provide information and enforce contracts.17 Source: Siamwalla and others 1993; Stifel 1976. * It helps firms learn about technology and about or- ganizational and managerial forms. * It exposes markets to greater competition and land titles in many countries (chapter 2) developed changes in relative returns, which induce institu- once markets for goods produced on the land were ac- tional change (see below). cessible or when new members entered the community. * It exposes countries to a different set of risks, possi- Empirical work spanning over 110 countries shows bly supporting the creation of additional institutions that measures of institutional effectiveness (such as the to manage the new risks. quality of institutions for public service delivery, or per- * It brings new market participants from other coun- ceptions of the rule of law) are significantly related to tries or regions who also demand more effective in- openness in international trade. This is so even after ac- stitutions to support market transactions. counting for differences in size, per capita income, legal heritage, years the country has been independent, and The case of Thailand illustrates how liberalization of other factors (figure 1.5).i9 trading rules led to a shift in agricultural returns-and Greater openness in trade and capital markets has to institutional change in the market for land (box been associated with the development of financial sys- 1.12).i8 Similar patterns are observed in other coun- tems, as historical and cross-country analyses clearly tries and sectors. The development of standards for rice show. Large incumbent firms that have access to fi- within Japan was spurred after markets within Japan nance-through either retained earnings or established were connected (box 1.13). The demand for formal links with financial institutions-do not always have an 18 WORLD DEVELOPMENT REPORT 2002 Box 1.13 Figure 1.5 Institutional evolution of rice markets and Greater openness and quality of institutions standardization in Japan, 1600-1920s Rule of law index In Japan's Tokugawa period (1600-1868), local private 0.6 - traders collected and marketed the rice shares of both the (a3 daimyo (feudal lord) and the peasant. The traders had to be big, since poor inland transport meant that rice was 0.4 - shipped in large sailing vessels and later steamships-a costly and risky venture. When the network of railroads was extended to local areas, locally segmented markets 0.2 - began to form a nationwide market. And with the econo- mies of scale in transport and related risks, small traders _ I could market their rice, using small shipments from many local centers. ° - The competition among small traders from different 1st 2nd 3rd rice-producing regions increased the pressure to standard- Openness (terciles) ize rice grades. Better and more stable quality and stan- dards ensured higher prices in urban markets. Groups of Government effectiveness index farmers and traders began taking the initiative by labeling the quality of rice in various regions. By 1900 these volun- (b) tary efforts were transformed into official regulations by 0.4 - local government agencies, which began to set standards for the packaging of rice shipped to other regions. By 1910 there were 33 rice-grading warehouses (beiken soko), managed by private companies or cooperatives, serving 0.2 - several purposes-inspecting, grading, repackaging, and storing. Innovations in finance followed. As farmers and traders I - brought ungraded rice to the warehouse, it issued a "rice 0 - exchange note." The precursor to today's inventory credit, 1st 2nd 3rd these notes were also used as collateral for loans from Openness (terciles) banks and pawnshops, easing capital constraints for farm- Note: The figures show the partial relationship (after accounting for the ers and traders. effect of differences in the legal systems, ethnic diversity, GNP per More trade among different communities led to the de- capita, years that the country in question has been independent, velopments of standards, first adopted by private traders country size, and inequality of income) between an indicator of rule of and later by government. These early institutional changes law/ government effectiveness and openness for over 100 countries promoted new institutions to support market exchange. in 1997-98. The countries have been divided into three groups of equal size. Source: Islam and Montenegro forthcoming, World Development Source: Kawagoe 1998. Report 2002 background paper. incentive to promote financial systems that would fa- with underdeveloped supervisory and regulatory sys- cilitate new entry into their markets. Opening the econ- tems. To get around weak judicial systems, poor coun- omy to trade and financial flows can automatically re- tries can export the enforcement of contracts. For infra- duce rents that incumbents receive from preferential structure deals in which private investors from rich access to financial institutions. And over time the lower countries invest in poor countries, for example, interna- rents can reduce opposition to financial sector reform.20 tional arbitration clauses can be used in cases of dispute. Rather than improve their own systems, policymak- Open information flows. Open information exchange, ers in open economies can import whole aspects of the a driver of institutional development, can both improve institutional system: laws, regulations, and enforcement the quality of other existing institutions and create a de- systems. Because of the political problems and costs of mand for new ones. Better information makes monitor- importing foreign agencies, including foreign human ing peoples' behavior easier. This ability to monitor be- capital, there are not many examples. Many countries havior changes behavior and institutional quality even have allowed foreign banks to operate in the domestic when institutional structure does not change. Better in- financial sector, helping financial services grow even formation can also change social norms and so change BUILDING INSTITUTIONS: COMPLEMENT, INNOVATE, CONNECT, AND COMPETE 19 people's incentives to participate in different institu- ing-both in time spent and in bribes paid-led the tions. And it can inform policymakers and other mar- government to undertake reforms that reduced licens- ket participants about the benefits of institutional re- ing time from years to days.22 With poor information form and about the constraints on institutional reform. flow in an economy, regulatory rules and policies are Information from the media and low-cost informa- unclear. So regulated firms and customers do not know, tion on the Internet can enhance the functioning of or cannot find out, what regulations apply to them or public institutions. Evidence indicates that corruption, how to comply with them.23 for example, is lower in countries with a free press (box 1.14). There is also evidence that free media, by pro- Promlote cuiiipetition--amAxaongj,,m?isdictiotas,firins, viding a check on political actions, can raise policymak- anid individuals ers' awareness of the social effects of policies, improv- Competition among jurisdictions, among firms in ing the provision of social services. A study in India product markets, and among individuals does much found that the media affected how the government re- for institutional change.24 Often, current institutional sponded to floods and famines: the distribution of re- lief was greater in states with higher newspaper circula- Figure 1.6 tions. The more information the local media provided, Diversity of information providers and quality of the more effectively citizens could develop a collective institutions voice and put pressure on the government.21 Recent research for this Report shows that competi- Press freedom index tion in the provision of information can significantly 0.8 - increase the impact of the media on the quality of in- 0-7 - stitutions. For example, where the state does not con- 0.6 - trol information through monopoly or concentrated 0.5 - ownership of the media industry, the media can do _ much in checking corruption (figure 1.6). The effect of 0.4 private monopolies on information flow can be ex- 0.3 - pected to be similar. 0.2 - Information about the potential benefits and costs 0.1 - of particular institutional arrangements can change the 0 L I I I incentives for those who engage in market transactions 0-25 25-50 50-75 75-100 and the demand for institutions. In Nepal the publica- State ownership of press tion of simple facts about the costs of business licens- (percent) Effectiveness of government Box 1.14 0.6 - Role of the news media in fighting corruption 0.4 - in Kenya 0.2 - 0- In 1996 investigative lournalists from a privately owned newspaper uncovered evidence of corruption in Kenya's -0.2 - ministry of health. A purchase of unapproved malaria -0.4 - chemicals was planned through a foreign firm at a substan- 0.6 - tially higher price than local firms were charging. It was____ also reported that the health minister paid the foreign firm -0.8 - I 400 million Kenya shillings, even though no goods were re- -1 L ceived in return for this payment. The press revealed these 0-25 25-50 50-75 75-100 findings and made daily reports on the scandal. Eventually, State ownership of press under unrelenting pressure from the media, the minister (percent) was dismissed. Source: D)ankov and olheTs 2001, Wordd Deve$opment Reporl 2002 Source: Githongo 1997; Stapenhurst 2000. background paper 20 WORLD DEVELOPMENT REPORT 2002 structures may inhibit competition. Competition makes Box 1.15 institutions more or less effective by affecting relative Competition and the evolution of corporate law returns and changing the incentives of agents. For ex- ample, as competition in markets increases, traditional A study investigating legal change in 10 jurisdictions, in- norm-based institutions may become inadequae or .cluding both industrial and developing countries over more nombse ntiuinsmy e in othan 100 years, found competition among firms operating obsolete.'5 Competition can reduce the effectiveness within countries and across borders to be important in pro- of closed groups, such as guilds or business networks, moting changes in the corporate law. The changes were whose existence and effectiveness depend on superior often enacted in response to crises, owing to competitive access to such inputs as information. This can create the pressures, or as a conscious effort to standardize corpo- accessto suc mputsas mtomatlon Thls an crete therate law across countries. demand for new institutions or improve the quality of Studies of Europe in the late 19th century highlight ju- existing institutions by changing behavior. In places as risdictional competition in the development of corporate varied as Thailand and Uganda, greater competition for law. There was a shift from concession systems, in which rulers granted the right to incorporate case by case and land increased land disputes and created the demand for often as a special favor, to a system of registration in more formal procedures for recording transactions. which any company meeting certain minimum require- Competition in product markets has led to institutional ments could incorporate. For example, in France in 1867, change in labor markets (chapter 7). And there is some the shift was induced by the expansion of English compa- nies on the continent. Once France allowed companies in- evidence that competition between firms can be a par- corporated in England to operate as a corporation in France, tial substitute for strong shareholder rights in inducing without special approval by parliament, it faced pressure managers to act in the interest of owners. from domestic companies to drop the concession require- Firms competing in product markets, forced to in- ment at home. Israel in 1999, Japan in the 1990s, Chile in 1981, and crease efficiency, have the incentive to lobby policymak- Delaware in the United States (where there have been ers to implement institutional changes that lower their continuous changes) provide examples of jurisdictions that costs. Competition also affects the distribution of gains changed their corporate laws in response to competitive among market players, and so increases the demand for pressures in the 20th century. institutional change among those who want to maintain Source: Pistor and others 2000, World Development Re- their gains in the light of changing economic factors. port 2002 background paper. But sometimes institutions, such as rules governing in- tellectual property, may be needed to limit the degree of competition in markets and to foster innovation. and corporate laws vary across the states. Education sys- For firms in international capital markets, competi- tems vary across districts. tion can produce demand for better institutions, such Markets with more competition may require fewer as accounting standards (chapter 5). In turn, domestic formal institutions, since competition can substitute banks, to compete with foreign banks outside their for regulation. Take infrastructure: greater competition, home markets, may pressure their regulators to improve possible with technological changes, has allowed regu- prudential regulations. This happened in Mexico after lators to lower the frequency of price reviews (chapter it signed the North American Free Trade Agreement. A 8). Sectors previously considered natural monopolies World Bank study looking at institutional performance became potentially competitive, so governments now cites competition as a key factor affecting institutional rely more on competition to deliver desired outcomes, performance, since it changes the incentives for indi- such as affordable prices for consumers. viduals to succeed.26 But the competition from new infrastructure pro- Jurisdictional competition also fosters institutional viders can also complicate regulation. Before the priva- evolution. A study of corporate law evolution shows tization of state monopolies in many countries, state that competition between countries-and between for- infrastructure monopolies cross-subsidized some cus- eign firms operating in a country-has created pressure tomers-in many cases lowering costs for poorer house- for change in corporate laws (box 1. 1 5). In the United holds by charging higher prices to business users. After States competition between states to attract business privatization, governments aiming to protect poorer has led to institutional evolution of different forms in customers have found it difficult to regulate the priva- the various states. For example, personal bankruptcy tized firms in a way that provided adequate profits for BUILDING INSTITUTIONS: COMPLEMENT, INNOVATE, CONNECT, AND COMPETE 2I firms while providing adequate services to the poor since they could keep any collections above a certain (chapter 8). level. Local governments were also able to raise addi- There are times when institutions restricting com- tional taxes, not shared with the national government. petition are desirable. Some market rents may need to But all reforms are not equally difficult politically. be tolerated to fund the adoption of new technology, Some ineffective institutions may exist in part because and institutions restricting competition may be needed there are no interest groups pressing for change-not to promote market development. And regulating the because some interest groups oppose change. Or it may degree of competition among banks can enhance finan- be that those who would oppose change do not have cial stability by reducing the incentives for risk taking. much political sway. Whatever the reason, reforms in these areas could be accelerated. And as these reforms How1 do poditicaiftrwces. s;ia1lpr pssllres, breed new constituencies and forces, they can lead to a H7mdl shocks affect thc pace of change?J demand for greater change. The key is to find the op- Political forces and social pressures can either acceler- portunities and to work in these areas. ate or retard the development of new institutions. Shift- Although indigenous institutional development re- ing social, political, and economic balances are in turn sponds to changing economic and social conditions, a affected by a government's institutional reform efforts. central issue for transplants is managing distributional In industrial markets, however, the state is constrained conflicts. Institutional change creates winners and losers. from arbitrarily changing rules and laws, and there tend For example, bankruptcy law designates the rights to in- to be more checks and balances on various actors, pub- come and assets for creditors. Corporate law distributes lic and private. rights among owners, managers, and the government. Politicalforces. An institution exists in part because Regulation covering service provision to the poor trans- some constituencies gain from its existence and so have fers economic gains from producers to poor consumers the incentives and influence to support it. This distri- and between levels of government (chapter 8). butional aspect is particularly important when institu- The distribution of power among different levels of tions benefit a small group or minority in society for government largely determines the type of regulatory whom the costs of collective action are low and bene- structure likely to be effective. A study of the evolution fits are large. Checks and balances on political power, of regulation in infrastructure conducted for this Re- from firms and interest groups, can support the inter- port argues that the allocation of regulatory authority ests of the majority. But minority interests may in some in the now industrial countries closely followed the po- cases oppose the modification of institutions. litical structures of those countries.27 The degree of po- So policymakers wishing to embark on reforms may litical and administrative centralization in a country sig- have to create new institutions rather than modify ex- nificantly affects the intervention by the upper level of isting ones. According to some, this was important in government in regulation. In the United States-where the recent establishment of a regulatory authority for the states are large and there is a great deal of auton- telecommunications in Morocco. But even though omy-local regulation of concessions for water and building new institutions may be desirable, the costs of electricity was gradually overtaken by state-level regula- collective action-including those of information col- tion. The greater centralization was hastened by corrupt lection, enforcement, and competition-may be so municipalities or complex regulatory issues between great relative to perceived benefits that they would frus- local jurisdictions. Traditionally, regulation of natural trate the formation of a new political coalition that monopoly infrastructure firms evolved in response to would push for institutional change. political pressure from firms or communities. In re- Institutions often change when the power of those sponse to high prices and high profits, the public de- who directly benefit from the existing structures is un- manded government intervention. By contrast, France dermined or when they no longer reap any benefits so has a very centralized political system and has generally that they no longer have the incentive to oppose change. adopted a much more centralized regulatory structure. One way to accelerate institutional change is to co-opt When transplanting regulatory agencies from indus- the opponents of reform. In China after 1978, local trial countries, the domestic political structure and bal- governments were encouraged to collect federal taxes ance of power must be considered along with the qual- 22 WORLD DEVELOPMENT REPORT 2002 Box 1.16 The interplay of social, political, and economic forces in the reform of land institutions in China Before the 1960s rural land in China was the responsibility of the institutional structures to protect farmer welfare probably communes. In the early 1960s farmers in Anhui Province began slowed the effective implementation of land contracts. calling for a restructuring of communes, so that earnings could In some cases land readjustments also reflected the desire be linked to work. Local leaders began experimenting, allow- of local cadres to maintain influence. Control over land remains ing some households to contract for individual plots. The one of the main sources of economic and political power for demonstration led others to push for plots, and the resulting local officials. Perhaps as a result, land system rules regarding productivity increases led to formal sanctioning by local lead- tenure-both formal and informal-and practices have varied ers of this system. At that time the central government was widely around the country. Although the central government not involved. Later, the system was partially reversed because approved 15- and then 30-year tenures, this was not imple- of central government disapproval. Then in 1978 a severe mented. Field research indicates that county-and provincial- drought in Anhui led to a food crisis, and provincial leaders al- level officials from jurisdictions that rely heavily on agriculture lowed households to cultivate any land the collective farms are more likely to have interests that are similar to those of were unable to work. Nearby villages emulated the practice. farmers than local officials. The central government began accepting local institutional China's story reveals some important lessons for institu- innovation after almost 20 years, when faced with an economic tional reform. crisis. Central government officials formally adopted the House- hold Responsibility System, under which households could con- * Experimentation has been key for institutional reform, in this tract with local leaders to produce on their "own" land. Initial case at local levels. The central government was important distributions, although different from village to village, were es- in validating a successful experiment and thus in accelerat- sentially the same within the village. In other words, both so- ing its acceptance around the country. cial and productivity considerations determined land allocations * Institutional reform takes time. Chinese land policies will in the transition to individual and more formal rights. But the continue to be modified as several important issues are re- contracts with households did not assure them of very stable solved and as other supporting institutions evolve. land use rights. Although such rights were supposed to be allo- a At different stages in the process of institutional reform, the cated for a period of years, most villages in China adopted the role of local versus other leaders varied significantly. practice of periodically readjusting landholdings in accord with U When changing established norms, governments need to changes in household makeup. Chinese farmers and officials be aware of dual role played by institutions-in this case for- have not been of one mind regarding social versus efficiency mal, but in many others informal-in affecting both effi- considerations. Surveys of farmers in the 1990s indicated that ciency and equity. Social concerns affect the pace of reform. they wanted more secure land rights, but many also favored Explicit considerations of these issues can help policymak- readjustments. Chinese farmers indicated that they would over- ers undertake institutional reform. whelmingly support a no-readjustment policy if their welfare concerns could be addressed by other means (such as prefer- Source: Prosterman, Schwarzwalder, and Hanstad 2001; World ential allocation of wasteland or taxes). Lack of consensus on Development Report 2002 background paper. ity of information that is available to different levels of Many developing countries have recently tried to estab- government. Such issues are particularly important in lish autonomous revenue agencies to free tax collection large countries such as Brazil, India, and Russia. Infor- from political influence. What determines the success mation problems at the national level tend to be more of these reforms? The authority granted to these insti- severe, but so could be the risk of regulatory capture at tutions, and the political commitment to support their local levels. While economic analysis may argue for a greater autonomy (chapter 5). certain design, for effective institution building, politi- Political instability also affects investment within cal and social realities and their dynamics will need to countries, as cross-country empirical studies show. In be considered. countries more polarized and less politically stable, pol- History shows that politics influence the develop- icymakers are less committed to strengthening the legal ment of financial systems.28 Financial institutions, par- system and protecting private property rights.30 Weak ticularly banks, provide an easy way for governments property rights in politically unstable countries lead to to channel the economy's resources in directions they lower investment. deem politically desirable.29 The effective functioning Socialpressures. Social structures such as inequalities of government agencies, such as tax collection agencies in income distribution and in the influence of different and financial supervisory authorities, depend critically ethnic groups also affect the demand for institutional re- on politics and checks and balances on political power. forms and their sustainability (box 1.16). BUILDING INSTITUTIONS: COMPLEMENT, INNOVATE, CONNECT, AND COMPETE 23 More inequality sometimes means lower institu- The recent experience of the East Asian economies tional quality. Empirical work across countries-using suggests that policies to promote equality, through in- indicators of institutional development that measure vestments in education, can yield high returns. People the rule of law, corruption, enforcement of property who are literate and educated are more likely to par- rights, and an overall index of these indicators-sug- ticipate in and demand formal market-supporting in- gests that there is some association between the distri- stitutions. This Report provides some guidance for bution of income and institutional quality, with very institution building in the social sectors (box 1 .17). Pro- unequal distributions of income being associated with moting opportunity in this way can promote social co- a lower quality of institutional development. hesion, important for consensus-building on reforms. Why might this be? Perhaps more unequal societies Large initial inequalities in wealth in closed markets are more polarized or less likely to engage in social or can also engender situations where strong economic in- economic transactions with each other. More polarized terests may "capture" the state, leading to regulatory societies also may be less likely to agree on institutional structures that favor their narrow interests and prevent reform, much as they may find it more difficult to agree broad-based markets. Market participants can have a on policy reforms.31 Or perhaps when a few players, key role in the design of institutions that affect their such as large business groups, dominate economic transactions. Creating inclusive institutions with more transactions, they have little incentive to support for- social legitimacy-systems in which business interests mal institutions that would enhance competition in and government can work together in an open and their activities. Those players, often part of well-knit transparent fashion in establishing institutions-can networks, can conduct most of their business through lead to faster progress than in closed systems. reputational mechanisms. Shocks. Large shocks to economic and political sys- The differences in the development paths of North tems change the balance of economic, social, and po- and South America are often cited as examples of how litical power-and thus the effectiveness of institutions. social factors-such as equality in the distribution of Sometimes shocks forestall reform and at other times human capital and other resources, differences in ethnic they can accelerate it. During economic depressions, diversity, and the economic power of the dominant for example, business and financial groups often come group in these economies-can affect institutional de- under greater scrutiny. It is claimed that periods of eco- velopment and growth.32 Countries in both regions im- nomic depression in Europe reduced political and so- ported institutions from Europe. A more equal initial cial support for financial development, particularly for distribution of income and a less polarized society in the the development of equity markets.33 But country ex- United States is cited as an important factor promoting perience also shows that since market-supporting insti- institutional reform. There was more participation by tutions need some stability to be effective, large eco- broad segments of the population in a competitive mar- nomic or political shocks may be needed for all but ket economy. More egalitarian societies may also be less gradual change. And sometimes several large shocks are polarized. This factor is probably more important for needed.34 For policymakers and politicians, periods ethnically diverse countries, particularly during eco- of crisis can sometimes provide opportunities, at least nomic downturns when conflicts tend to be magnified. in some sectors, to undertake bolder institutional re- Other forces may be at work. True, the history of the forms-and these are opportunities to be seized. industrial countries is full of examples of periods and A detailed analysis of the evolution of corporate law countries of high inequality. Consider the prevalence in industrial and developing countries shows that eco- of sweatshops, unhealthy working conditions, and ex- nomic crises create demand for reform. For example, tensive child labor in much of the United Kingdom the recent financial crises affected reform in Malaysia during industrialization. But this inequality did not (box 1.18). 3 keep the United Kingdom from the forefront of indus- Shocks in technology also create demand for new in- trial development. So it is not clear that initially high stitutions. Regulators need to develop new institutions income inequalities will always prevent later broad- to deal with such technological breakthroughs as the based market development. There may be countervail- Internet. For example, the spread of e-banking and the ing forces at work, such as the open exchange of infor- provision of financial and other information over the mation, open trade and competition, and innovation, Internet offers lower transaction costs-and new op- all promoting institutional development. portunities for fraud. 24 WORLD DEVELOPMENT REPORT 2002 Box i'7 Applying the lessons to the social sectors Some of the key problems institutions face in the deliv- servants-can be a solution for both service delivery and ery of social services are information, enforcement, and the provision of social assistance. What many NGOs bring competition. to the table is a credible promise not to exploit weaknesses Limited information about the beneficiary is available to in the monitoring systems of government. the provider of services, complicating the targeting of in- In many sparsely populated and poor areas, such as one come transfers to those truly in need. In Moldova, for exam- might find in rural Africa, it is unusual to find private, mod- ple, before recent changes in transfer systems, a 1997 sur- ern medical facilities unless provided by NGOs, particularly vey found that the richest 10 percent of the population faith-based NGOs. In delivering social assistance, NGOs enjoyed almost a fifth of all social assistance payments, based in the community may be better able than formal while 38 percent of poor households got no form of social agencies to discover who is most in need of aid and may be assistance at all. organizationally more flexible in delivering appropriate assis- Enforcement by public officials of good quality is difficult. tance to the neediest. The community may also serve as the In relatively poor areas of most countries, there are difficul- arbiter of who has the most needs-as in the mahalla sys- ties in maintaining staff and providing services, especially tem in Uzbekistan. for public providers. Evidence from Canada (Anderson and For many of the poorest countries, the best option for tar- Rosenberg 1990) to Indonesia (World Bank 1 994b), and from geting social transfers effectively may be to experiment with India (The Probe Team 1999, p. 44) to Zambia (World Bank different self-targeting mechanisms to find the system that 2001 f) shows substantial differences in vacancy rates in best ensures that few other than the poor use the transfers. health posts between urban and rural areas. Innovative approaches using less desirable consumption Then there are the issues of competition. For example, goods for the poorest (as in Bangladesh in the 1970s and competition by the government in providing social transfers Tunisia in the early 1 990s) have proved useful (for Tunisia, see may drive out private institutional arrangements-such as Tuck and Lindert 1996). Well-designed public works programs family networks, which can be targeted more effectively to that pay below-market wages are also a good self-targeting the poor than more arm's length (public) social assistance.36 way to ensure that resources get to those who need them. A study from the Philippines, simulating the results of intro- Connecting communities. Promoting open information ducing an unemployment insurance scheme, found that net exchange has been very important in building successful private transfers to the unemployed would fall by 92 pesos for service delivery institutions. In the state of Ceara in Brazil, every 1 00 given by the government (Cox and Jiminez 1995). one factor in the dramatic improvement of health service de- Complementing what exists. The demand for modern livery was an innovative monitoring approach. But also criti- public institutions to deliver universal social services and cal was a substantial public relations campaign that pre- widespread social assistance is fairly recent. In health the ceded the program, increasing its visibility, enthusiasm, and large national systems of the United Kingdom and Canada prestige. In this way, the program recruited a cadre of inter- date from 1948 and the 1970s, respectively. and in education, ested local monitors (Tendler and Freedheim 1994). the achievement of universal primary education, requiring Sometimes, simply providing information to local com- public funding, occurred late in the 19th century. The origins munities is enough to stimulate improvements in quality in of national social assistance schemes are also fairly recent. service delivery. Recent technological advances, including In each case, extensive reliance on the private sector the Internet, allow government and private agents to pro- preceded the participation of governments. In fact, as to- vide information cheaply. The rate of sharing information day's richer countries grew more advanced, they could pro- is dramatically enhanced. Take Uganda. In 1995 a study to vide more forma/ social services at a price-quality mix de- track expenditure from the central government to individual manded by the population. They could ensure adequate schools found that as little as 30 percent of nonsalary recur- training of public providers. And they had the complemen- rent budget allocations meant to reach schools actually did. tary institutions (such as more reliable income and asset The results of this study were publicized in newspapers and ownership records) to better target social assistance to the posted at local facilities. A follow-up survey in 1999 showed neediest and that were free from corruption. an increase in actual disbursements, averaging very close to For developing countries, public involvement in these 100 percent (Ablo and Reinikka 1998). Another variation that areas has accelerated. With poor complementary institu- does not depend on technological improvements is sharing tions-inadequate monitoring capacity, poor communication information through the use of traveling teachers. networks-providing universal coverage immediately may be Promoting competition. Competition between public and too ambitious. Public financial constraints, including low fiscal private providers improves institutional quality. In Malaysia a resources, may also worsen quality of services. So recogniz- reliable system of public clinics has maintained pressure on ing the need to be flexible in price-quality goals is important. the private sector to keep prices reasonable (van de Walle Innovating to identify what works. Despite the lack and Nead 1995; World Bank 1992). But competition is pos- of complementary institutions, developing countries can sible only in areas densely populated enough to support mul- use innovative methods to ease many of the information tiple providers. This leaves unaddressed the problem of re- and enforcement problems in these areas. The use of pro- mote areas with many poor people. In the United States, for viders closer to the community-such as NGOs, whose example, voucher systems are almost always advocated motivations are different from both private sectors and civil only for urban areas. BUILDING INSTITUTIONS: COMPLEMENT, INNOVATE, CONNECT, AND COMPETE 25 Box 1.13 Crises and institutional change in Malaysia -. . ,,%$ t4% 4. .. flv,IW,,A* .f };:r .5 i. _ t. ,. 2t'J .g12 .53.._iS Malaysia had one of the most developed capital markets in crisis and with the aim of reducing capital outflows, policymak- East Asia in the early 1 970s. At first, securities market regu- ers adopted selective capital controls. lation followed mostly the English system of market self- After an evaluation of the crisis, a series of more substan- regulation. Although a comprehensive securities act was en- tial institutional changes were introduced in preparation for acted in 1973, jurisdiction over market supervision was divided continued liberalization. Focused on transparency and gover- among several state agencies-including the ministry of fi- nance, these changes included new accounting standards, nance, the registrar of the companies, and the capital issues merger and acquisition rules, capital adequacy rules for stock committee. brokering companies, and broker commission liberalization. In 1993, after a decade of rapid market development, con- The implication for policymakers is clear: if crises expose trols were unified in a new securities commission. Before the real vulnerabilities in markets, policymakers should take advan- financial crisis in 1997, the commission had determined to re- tage of these times to fix the vulnerabilities. place the detailed merit regulations system with a liberalized system based primarily on disclosure. But in the wake of the Source: World Bank staff. Organization and scope of the Report around the world and on competition authorities and The second part of this year's Report concentrates on legislation. Last, it assesses the regulation of monopo- firms. It addresses institutional issues that affect pro- lies in developing countries and the consequences for ductivity and risk management in agriculture: the service delivery to poor people. rights to land, the credit in rural areas, and the institu- Parr IV of the Report concentrates on society. It dis- tions that support innovation and dissemination of cusses how norms and codes of conduct in societies in- ideas in agriculture. It also concentrates on the prob- fluence markets and public institutions and in turn are lems of governance for firms, looking at institutions, influenced by market developments. It also explores the internal and external to the firm, that enhance invest- role of the media in expressing and disseminating the ment in firms and ensure good management-espe- concerns and values of society-and the effects such in- cially the interaction between ownership structures and formation flows have on institutional quality and thus legal frameworks and between private institutions (such on economic and social outcomes. It draws on a new as business associations) and public ones. And it ex- study of media ownership around the world written for plores the critical role of financial institutions, the nec- the Report. essary supporting institutions for their development, Market-supporting institutions are a big topic, for and the role of the supervisory and regulatory system these institutions are everywhere and varied. So much in ensuring a healthy financial system. It draws on new remains to be learned about them. This Report offers research done for the Report on the role of politics in policymakers some guidance that has been distilled financial development, institutions to secure access for both from the history of institutional evolution and new borrowers, and the effects of foreign bank entry from the lessons of recent experience-the varied expe- and privatization. riences of the transition economies in the 1990s, the Part III of the Report concentrates on government. continuing struggles in many poor countries around It examines how political institutions support good the world, and the successes of some of the emerging governance, focusing on the policymaking process, the economies in the past decades. incentives for corruption, and the institutions of taxa- At the same time, the Report does not address all tion. It next explores issues of judicial efficiency, and possible institutional problems in all possible fields. the experience with reforms aimed at improving effi- Rather, it focuses on a subset of these institutions from ciency, and examines the causes and consequences of many fields to illustrate that the framework (inform, cross-country differences in judicial procedures from a enforce, compete) and messages (complement, inno- new survey covering over 100 countries. It then dis- vate, connect, and-again-compete) can be applied cusses the main impediments to competition in mar- regardless of the specific sector studied. It does not kets, gathering new data on business entry regulations cover in detail institutions that previous World Devel- 26 WORLD DEVELOPMENT REPORT 2002 opment Reports have covered. This Report, one in a se- are not adequately organized to do so. Reforms to such ries looking at critical development issues, is a natural institutions are not as difficult to implement politically continuation of World Development Report 2000/2001, and, once implemented, could not only improve the which discusses the central role of markets in the lives way markets work but can help build momentum for of poor people. It leaves some important issues for further change. Both the supply of institutions and the WorldDevelopment Report 2003, which will focus on is- demand for them matter. Development experience sues related to the environment as well as on social co- does not provide a universal guide as to which partic- hesion and stability,. ular institutions should always be created first. How- ever, within each sector, the Report does identify areas where the introduction of a particular institutional Development experience shows that markets can pro- structure may need to wait for the development of vide the means to attain sustained increases in living other supporting or complementary institutions or standards for people around the world. World Develop- conditions. In other words, some priorities can be ment Report 2000/2001 argued that markets are central identified. to the lives of poor people. By providing opportunities This Report also considers the interaction between to engage in productive activities, and by empowering informal or norm-based institutions and formal insti- citizens, they can promote growth and reduce poverty. tutions. Many poor people, particularly those in poor But for markets to provide widespread benefits, they countries, do not have access to formal institutions. need to be inclusive and integrated. Policies that pro- Innovative designs may help bridge the gap between in- mote growth and reduce poverty are important, but the formal and formal institutions and gradually increase details of institutional design matter as well. the access of those left out. Simplifying formal institu- Improvements in living standards, and overall im- tions, providing more information about them to users, provements in the lives of poor people, depend on in- strengthening human capital, and accepting informal stitutions that support growth as well as those that di- institutions when formal institutions would not have rectly enhance the access of poor people to markets. their desired impact are some of the ways in which in- That is, poor people are affected by what other market stitutional designs can be modified to suit the needs of actors do. poorer countries and of poor people. Building effective institutions is a complex task. Ex- Local, national, and international actors, public or perience indicates that one size does not fit all. But not- private, affect how institutions evolve over time. The withstanding the uniqueness of countries, analysis of balance of power between private and public actors, country experience does hold important lessons for in- and the state's recognition of both its strengths and lim- stitutional development. itations, is an important factor in market development. This Report provides a framework for institutional A strong and capable state is necessary to support mar- development. It builds on the work of several disci- kets, and an arbitrary and corrupt state can impede plines, combining theory and evidence. It extends em- their development. But it is not only the balance be- pirical evidence on the details of institutional design tween private and public actors that matters. The de- across a wide range of countries, and within countries sign of institutions and the pace of reform are affected over time, to understand the process of institutional by how local and national leaders and national and in- change. And it provides guidance on how to build new ternational leaders interact. All of these interactions are institutions, modify existing ones, and create the forces affected by the nature of information flows and the ca- for change. pabilities of the various parties. Most times institutional change is a step-by-step The four main lessons of this chapter are that for ef- process. The Report acknowledges as well that many fective institution building policymakers need to com- reforms are difficult because there are constituencies plement what exists, innovate to suit local conditions, which benefit from existing institutions and often in- foster open trade and open information exchange, and terest groups which would promote change do not do foster competition among regions, firms, and individ- so. But it is also true that some institutions continue uals. The incentives provided to people depend on the to exist not because there is concerted support for whole set of institutions and affect their performance. them, but because forces that would press for change So when building an institution or modifying one, the BUILDING INSTITUTIONS: COMPLEMENT, INNOVATE, CONNECT, AND COMPETE 27 key thing to consider is whether supporting institu- tated by complexity of regulation in nontransparent tions-without which the institution would not be ef- markets and where other incentives for bureaucratic fective-exist. If not, perhaps it would be better to efficiency (such as wages or promotion) are weak. In work on the supporting institutions first or to modify these countries, to complement existing conditions, design so that the planned institution can work with- regulation needs to be streamlined. Technological dif- out the supporting institution. Also important are ferences are also relevant. To accommodate country- the levels of human capital needed, the extent of cor- specific differences in culture and endowments, in- ruption, and costs relative to per capita income. With novation should be encouraged and accepted. Finally, scarce human capital, complex regulations cannot be providing opportunities for trade will develop markets enforced as they are in countries with highly skilled per- and the demand for institutions that support transac- sonnel. These factors argue for simplification of insti- tions in markets. Open information sharing will do the tutional design. Higher costs relative to per capita in- same. Competition among regions and among firms, come of accessing formal institutions will mean that the often limited by current institutional structures, will disadvantaged and poorer members of society will be help identify new institutional forms and create the de- unable to access these institutions. Corruption is facili- mand for new institutions. PART 11 irms FIRMS ARE KEY BUILDING BLOCKS OF MARKETS, PRODUCING GOODS and providing services that form the basis of market exchange. This part of the Report considers institutional issues for three groups of firms. Farmers are the focus of chapter 2, which looks at ways for agricultural producers in developing countries to increase their productivity and reduce their risks through improving the institutions that gov- ern the markets for land, credit, technology, and agricultural output. For firms, a key institutional question is how to enhance investment and ensure good management; chapter 3 covers the problems of Governance ofFirms (focusing on those outside the financial sector), through an examination of interactions between ownership structures and legal frameworks, and public and private institutional players. In chapter 4, Finan- cial Systems, which perform critical functions for market systems, are considered; the chapter discusses the necessary institutional conditions for their development and the role of the supervisory and regulatory system in ensuring a healthy financial system. CHAPTER 2 .Farmners Most of the world's poor people earn their livingfrom effectively in some environments, but as incomplete agriculture, so if we knew the economics of agriculture surrogates in others. we would know much of the economics ofbeingpoor. Second, farming in many countries has historically -Theodore W. Schultz, 1980 suffered from urban bias in public policy. For example, state marketing boards in several African countries re- sembled the systems used by the colonists to gather armers operate in the market, like other entrepre- food during the Second World War.3 Both these neurs. But markets in rural areas, and particularly systems subsidized urban consumers of food by re- agricultural markets, suffer especially from prob- quiring farmers to sell their output at less than the lems of information, inadequate competition, and market price. Other examples of urban bias include weak enforcement of contracts. Building institutions overvalued exchange rates to make imports cheaper for that reduce transaction costs for farmers, therefore, can urban consumers, excessive agricultural export taxes, greatly improve the way agricultural markets operate. and high effective rates of protection for domestic in- This is especially important for poverty reduction, be- dustries that provide agricultural inputs. cause poor people are more likely to live in rural areas Public investment in infrastructure, education, and and make their living from agriculture-related activities other services in rural areas also tends to be lower than (figure 2.1). Well-functioning agricultural markets also in urban areas. Lower investment increases transaction have important benefits for the rest of the economy. As costs in marketing, which can be a major institutional agricultural productivity improves, farmers leave agri- constraint to developing agricultural productivity. culture for more productive employment in industry Public or private efforts to build specific institutions and services, promoting overall growth.' that ease information costs, such as grades and stan- Three particular challenges face policymakers build- dards or market information systems, can help to boost ing institutions for agricultural markets. First, agricul- agricultural development (box 2.1). Beyond physical tural activity is usually geographically dispersed and dis- access to markets, large segments of the rural popula- tant from major urban centers. A problem in providing tion, and the rural poor in particular, often face con- rural credit, for example, is that formal providers of siderable obstacles in accessing agricultural markets. credit, such as banks, may find it costly to obtain infor- This is because their relative lack of education can mation on geographically scattered small farmers. So make some useful formal institutions, such as institu- interest rates on formal loans to small farmers, if loans tions for disseminating technological information, are available at all, tend to be prohibitively high. Simi- harder to access. larly, costs for judicial services and the marketing of Third, agriculture is heavily dependent on the va- produce can be high because of the distance between garies of climate. Poor farmers often rely on their own farms and major towns.2 In all such cases, informal in- savings and the help of family and friends when floods stitutions serve as substitutes for formal institutions- or droughts strike. But these insurance mechanisms are 3I 32 WORLD DEVELOPMENT REPORT 2002 Figure 2.1 in turn, generates demand for inputs. The more open Poverty head counts, urban versus rural, the market, the greater is the demand for effective for- selected countries, 1990s mal institutions for farmers-from documented prop- erty rights in land to better access to credit. Zambia Informal institutions and simplified procedures may Madagascar I be appropriate in situations where complementary for- Guatemala r m mal institutions are absent or where the overall demand Ecuador for agricultural output is low. In such cases the costs of NepalU Urban complex formal institutions may be high compared Bangladesh Rural with their benefits. For example, in areas where there is little competitive pressure on land, communal rights Romania to land can be sufficient for tenure security. Formal ti- Indonesia tles may be more appropriate in situations where high China demand for land gives rise to disputes over land and in- 0 10 20 30 40 50 60 70 80 90 100 formal institutions can no longer resolve these disputes National poverty rates satisfactorily. (as percentage of population below poverty line) Innovation, often through experimentation, can Note: National poverty lines are used, so results are not comparable identify techniques that overcome the inherent high across countries. transaction costs in rural areas. These can range from Source: World Bank data. simple databases that permit technological information sharing among small farmers to improved enforcement mechanisms inherent in group-based lending. of little use when savings are meager or when the entire This chapter concentrates on specific agricultural in- circle of family and friends suffers from the same disas- stitutions, including land rights, rural financial institu- ter. Wealthier farmers and those in richer countries can tions, and institutions that create and deliver agricul- purchase forms of disaster insurance and benefit from tural technology. Many other critical issues that relate public subsidies when struck by adversities. These sub- to agriculture are dealt with in other chapters of the sidies are costly for poorer economies, and markets for Report-for example, openness to international trade disaster insurance require an array of complementary in- in chapter 7 and water and electricity pricing in chap- stitutions unavailable in most developing countries. ter 8-and in other Bank reports.5 How can governments or communities build effec- tive institutions to raise farmers' returns and lower their Building more secure and risk?4 This chapter addresses this question by drawing transferable rural land institutions on evidence from successes and failures of institutional Historically land was abundant and was held commu- arrangements across countries. The interlinked institu- nally or could be obtained by any who laid first claim tions governing farmers' physical and financial assets- to it. But as population grew, land in many parts of the those for land and for finance-are particularly impor- world-specifically, agricultural land-became more tant. Secure and transferable rights to land stimulate scarce, until its relative scarcity raised its value. As land income-generating investment and reduce uncertain- became more important as a productive asset, it moved ties about future incomes. So do well-functioning rural into individual or family ownership. With private prop- financial institutions, which provide credit for both erty came the need to prove ownership. Even in ancient income-enhancing and risk-reducing investments and Egypt and Mesopotamia, titles for land were important insurance. Institutions for generating and disseminat- for land transactions (box 2.2). ing agricultural technology directly affect the yields and Today the nature of property rights in agricultural risk inherent in agricultural production. land varies widely across countries. Both governments In many countries, marketing problems are the and communities have built institutions to define these biggest institutional constraints to increasing agricul- rights. In some countries, such as Uzbekistan, the state tural productivity. Connecting small, isolated commu- owns all land. In China private ownership of land is nities into larger markets, and particularly into global also prohibited; government regulations allow private markets, stimulates demand for farmers' output. This, citizens to lease land legally for 15 years, although prac- FARMERS 33 Box 2.1 Agricultural marketing institutions When complex agricultural marketing arrangements in develop- kets. Throughout the world the expansion of trade between com- ing countries fail, it is usually because of the lack of effective sup- munities has created a similar demand for such standards and porting institutions. The state has a role to play in building better grades. Private merchants usually initiate standards. But as the marketing institutions, but not through state marketing bodies, volume of exchange increases, the importance of public interven- which have clearly been unable to balance conflicting state objec- tion to promote the use and adaptation of standards increases. tives, including credit provision, tax collection, and food security International standards are often sponsored by larger farmers and price stability for urban consumers. Instead, the state can fa- and firms in developing countries. These standards may promote cilitate private marketing institutions, such as contract farming and overall exports. Yet smaller farmers who are currently involved in cooperatives. export markets may be left out of the process. The setting of high- Contract farming arrangements, or "out-grower schemes," level standards may raise their costs. These farmers have two op- are attractive for farmers in developing countries because they tions. First, they may reap part of the benefits of standardization, help small farmers access modern inputs, such as credit and such as lower information costs, through the use of informal insti- seeds, and market their produce to domestic and international tutions that have evolved to mitigate informational problems (as markets. These schemes range from agreements between indi- with informal brokering arrangements in Ethiopia). Second, policy- vidual traders and farmers, as in many Asian countries, to more makers can reach out to enroll poor farmers in certification pro- formal systems in countries of Latin America, Central Europe, and grams to integrate them into the wider agricultural markets, as is East Asia. Even though the institution is private, governments can being done by Mayacert, a nongovernmental organization (NGO) help build it in two ways: as information facilitator, helping to operating in Guatemala. match small farmers with domestic traders and agrobusiness Market information systems (MIS) generically describe dis- firms; and through complementary institutions, especially contract semination networks of public data that provide information on enforcement mechanisms (such as courts to resolve contractual agricultural markets. For farmers, knowledge of market informa- disputes, or grades and standards). tion (such as the prevailing price of a commodity in key wholesale Farmer cooperatives are more common in industrial countries markets) can help them to plan their production, harvesting, and than in the developing world, dominating the dairy sector in Fin- sales according to market demand. For traders, better information land, wheat in Canada, rice in Japan, and grain in Argentina. By improves their ability to decide whether to hold products in stor- tackling the problems caused by the relative smallness of indi- age or ship them to the most lucrative markets. In both cases MIS vidual farmers, cooperatives can be very successful in dealing are of special use to smaller farmers or traders, who lack the scale with both information asymmetries and competitive power ver- economies to gather such information on their own account. sus purchasers. They do this through collective action, pooling re- In most industrial countries, private agencies provide agricul- sources and lowering the unit costs of transactions. For market- tural market information for a fee, while public agencies collect ing cooperatives in developing countries, the record has a clear market data and make the information available free of charge. lesson: governments should stop trying to impose "top-down" Given the high cost of collecting and disseminating such informa- cooperative structures on farmers. Cooperatives such as Anand tion in areas lacking standardization of quality and weights and ad- in Gujarat, India, or UGC in Mozambique have seen success be- equate communication infrastructure, any user fees charged by cause they are voluntary in nature, which helps mitigate some col- private agencies are likely to be high. So the public sector has an lective action problems, such as low effort by participants (a prob- important role in poorer countes. lem that has plagued state-led agricultural cooperatives), Also, Public sector MIS systems are not widespread-a survey of they have experimented with context-specific institutional design, 120 developing countries identified only 53 such systems (Shep- which has improved trust, transparency, and innovation. herd 1997). But they are usually barely functional. This is due to Whatever the organization of marketing, purchasers can still inadequate financing, inability of bureaucrats to collect reliable incur high costs to verify the quality of goods they buy. Two insti- market information, and reluctance of traders to divulge informa- tutions have evolved to meet these needs: grades and standards tion for fear of being taxed. provide a greater level of certainty about the quality of produce, Nevertheless, several innovative strategies for effective dis- and market information systems provide information to farmers. semination of marke information are being explored. For exam- Because agricultural products have a vast array of characteris- ple, the government of Andhra Pradesh, India, makes prices of tics, grades (classifications based upon quantifiable attributes) and produce in different regional markets available on a website that standards (rules of measurement) are used to separate similar is updated daily. Again, a major role for the public sector may be products into categories and describe them with consistent ter- to help market participants improve their own information flows minology. This evaluation system can significantly reduce infor- by expanding the availability of low-cost communication technol- mation costs by allowing traders to contract "remotely" through ogy For example the exchange of market infoTmation in Ghana, commodity specification rather than through on-site visual inspec- the Philippines, and Bangladesh was boosted when governments tion. But the benefits go beyond this. Because grades and stan- made rural access a condition for granting licenses to mobile tele- dards can be independently certified, they facilitate access to phone companies. Consequently, market traders increasingly credit, through the use of warehouse receipt schemes, inventory gather and convey information among themselves through the credit, and commodity exchanges. They can also expand the mar- ket by allowing price and quantity comparisons, and thus trade, use of their own cellular telephones. across markets with common standards. In the rice market in Japan, standards and grades were cre- Source: Chaudhury and Baneri 2001, World Development Report ated when the spread of railroads began to link once-isolated mar- 2002 background paper. 34 WORLD DEVELOPMENT REPORT 2002 Box 2.2 Box 2.3 Early institutions of land ownership in Quilombos in Brazil: infrastructure, social change, Mesopotamia and Egypt and a new demand for land registration From the dawn of agriculture around 10,000 years ago Until recently settlements known as quilombos have been until a couple of centuries ago, land has been abundant rel- hidden away in Brazil. Tucked away in geographically re- ative to population in much of the world. Land in early mote settings, these communities are inhabited by the de- times was usually owned, if at all, by the king or the tem- scendants of runaway slaves. Their remoteness was origi- ple. In Egypt the Pharaoh Menes (c. 3100 B.C.) carried nally an effort to avoid discovery and scrutiny by slave around deeds certifying his ownership of all land, granted owners. Today the settlements are connected to the rest to him by the king of the gods. But private land ownership of Brazil and the world by new and improved road, river, and land sales were also recorded. In Uruk (in southern and rail links. As a result the value of this land has grown Mesopotamia, c. 3000 B.C.), there are records of individu- for ranchers, mining companies, and land speculators, als who "owned" land, although titles did not exist-and who have been attempting to take over some of the tablets give information on the sale of this land. quilombo lands for development. Initially these efforts Some of the earliest existing physical records of pri- were relatively successful because quilombo inhabitants vate landholdings date from the reign of Hammurabi (1 6th did not have formal titles. Since their ancestors had been century B.C.), whose famous code also laid down specific illiterate, no documents testified to the existence of their circumstances under which the king would step in to re- communities, and all official records of slavery were offi- solve land disputes. In Hammurabi's time, land assign- cially destroyed in 1890. ments were delineated by pegs around the boundary, and Yet a government-sponsored effort is now under way a record of all landholdings was kept in the palace. By the to give quilombo dwellers legal title to ancestral lands This third century B.C., it had become common for Egyptian process was eased by the adoption of a new constitution landholders to keep a document of possession with them- in Brazil in 1988, 100 years after slavery ended, which fi- selves. By the time of the Sassanian era (A.D. 224 to 651), nally recognized the rights and status of descendants of property ownership in Mesopotamia required a written runaway slaves. By 2000, 743 quilombos, some dating deedr witnessed and then registered with the state back to the 17th century, had been identified across Brazil and were seeking formal recognition of their status. There Source: Powelson 1998. have been some attempts to expedite the titling process, such as accepting the oral testimony of the oldest resi- dents as proof of settlement claims. Still, the process of regularization has not gone as fast as could be hoped, and the government is trying to accelerate it. tices differ across different localities. Similarly, although Source: Rohter 2001; Funda,co Palmares 2000. private ownership is forbidden in most Central Asian countries, some governments have developed well- defined and often-codified use rights to state property In many cases establishing formal titling is an un- and have built organizations to administer them.6 Pri- necessary cost in the medium term. But formal prop- vately determined and ill-defined "squatters' rights" con- erty rights systems enforced by the state are needed to stitute ownership and transferability of cultivated land reduce land disputes where population growth or de- for many small farmers in Latin America. Communities mand for agricultural produce leads to competitive and tribes in many African countries have informal, in- pressures on land or where transactions with those out- dividual use rights to communally owned property.7 side the community are common. For example, the de- Land tenure is transferable in most of South Asia, but mand for formal individual property rights was stimu- uncertain institutional arrangements have resulted in lated in England by the demand for wool and thus for clashes about ownership and the potential for govern- sheep. In Kenya the demand for formal land rights was ment seizure of land, leading to insecurity in some areas. triggered by the emerging global market for plantation Secure and transferable land rights can be provided crops, while in Thailand the cause was the internation- by both informal and formal institutions. Such systems alization of its market for rice following the 1826 Bow- must provide information on who owns the land, who ering Treaty. Better infrastructural services that con- has a secured interest in the land, where land transac- nect remote lands to the market can also stimulate the tions are registered, and how to access this information. demand for formal institutions to delineate and en- The community-defined ownership or use rights in force property rights, as has been the case in Brazil parts of Africa, for example, perform these functions. (box 2.3). FARMERS 35 throlfgh) better 1 i[ Examples of policy barriers to the operation of better laud ~~~~~ land markets Improved security of tenure can raise the expected re- -_______ turns from investment and ease credit constraints. This in turn can raise investment levels and productivit,v. Se- Even when land transfer is allowed by law, extensive reg- ulation of transactions can frustrate the operation of the cure tenure to land helps assure investors that the re- land market. The following barriers can be reduced by gov- turns to their investment will not be expropriated by ernment actions. government or private agents. Better land tenure also Restrictions on land sales. Some countries prohibit land sales outright. In many transition countries land priva- increases access to credit, since land can be used as col- tization has been accompanied by ceilings on sale prices lateral (discussed in chapter 4). and moratoriums on resales. Governments have indirectly If land tenure is secure, a functioning land market restricted land sales by mandating that any land transac- tion has to be approved by a higher authority. that allows transfer of property from one owner (otra High sales costs. High transaction taxes or high fees possessor of user rights) to another can help raise pro- can discourage land sales or drive them into the informal ductivity by transferring land from less efficient culti- sector. In the Philippines and Vietnam the tax on land transactions is almost 20 percent of land value. Costs can vators to more efficient ones. This overall productivity also be high when lack of competition results in high fees gain, of course, is greater if there are functioning credit for services associated with land sales. markets-otherwise the more efficient farmers would Restrictions on land subdivision. Such restrictions have not be able to raise the capital needed for the purchase. been established in former colonial environments to Productivity increases also depend on sellers being able prevent the disintegration of large farms often formerly Productivity increases also depend on sellers being able owned by colonialists, without any economic justification. to engage in other income-generating activity. Several For example, Zimbabwe continues to have these restric- studies of China, one of the few countries that has ex- tions, while South Africa has just recently begun the pro- perimented with allowing different systems of transfer cess of repealing its regulation. Restrictions on the use of land for collateral. Examples rights across different provinces, have confirmed that include Vietnam, where the value of land as collateral is higher levels of transferability were positively correlated limited by law and where foreign banks are not allowed to with higher levels of farm investment. 8 take land as security for credit. Also, creditors cannot own or exchange land use rights, and any land that is repos- In many developing countries extensive regulation sessed is auctioned off by the state. In Mexico banks can of land market transactions has meant that land mar- obtain the use right but not the ownership of land. kets seldom operate freely. Since transparency tends to Lengthy land registration processes. In Mozambique there is a backlog of about 10,000 applications for land be low and administrative capacity limited, these regu- rights, which means long delays between receipt of an in- lations also encourage corruption. Lowering these costs vestment plan and eventual granting of the land right. In of land transactions may be of particular importance in Cameroon the minimum amount of time it takes to regis- ter a plot is 15 months, and registration commonly takes parts of Asia, especially South Asia, where a flourishing between 2 and 7 years. In Peru the official adjudication land market could improve productivity by avoiding process takes 43 months and 207 steps in 48 offices, al- excessive fragmentation and subdivision of landhold- though an expedited process is now being implemented ings (box 2.4). in selected areas. Source: Deininger 2001, World Development Report 2002 When are JUr mal tithnia i'?sditzf fin .z,;e±'d,-' background paper; de Soto 2000. Formal land titles can increase access to credit and raise ..u.- E. investment in land. But these functions greatly depend on the broader institutional environment. Govern- ple in box 2.3, to occupancy rights by squatters on ments should embark on large-scale titling programs land.9 Formal land titles have also been established dur- only where competitive pressures and potential disputes ing land privatization processes (as in the transition mean that community land tenure arrangements are countries). As discussed in World Development Report ineffective. 2000/2001, clearly defining land rights during land re- Efforts to issue documented and registered land titles forms is key to improving the lives of poor people- have gained prominence in recent years. In many cases, farmers and nonfarmers alike.10 titles have formalized undocumented tenancy rights, But the financial and administrative costs of a for- which can range from long-established community- or mal land titling program are high. This raises the ques- tribe-based systems, as in the Brazilian quilombos exam- tion of what types of institutions are needed, and when. 36 WORLD DEVELOPMENT REPORT 2002 Formal land titles create secure and transferable systems administered by communities or their leaders. property rights by providing better information. Infor- The great majority of these holdings are held under mal land right systems are based on the knowledge of rights of inheritance derived ultimately from commu- community members and neighbors. These individuals nity membership-rights that are defensible in the may know the quality of a piece of land, who truly local courts.14 owns it, and its precise physical demarcation. But out- The third situation arises where administrative and siders to the community who want to buy land have lit- institutional shortcomings mean that formal titling tle access to this information, and no way to ensure the does not result in more secure tenure than informal al- reliability of the information they obtain. Formal land ternatives. The effectiveness of formal titles depends titles can help to remove this source of uncertainty. At on the quality of the title-such as clarity-and respect the same time, by resolving ownership disputes, they for the law. National legislation for tenure reform has can thwart arbitrary seizure. They also ensure that the limited capacity to change behavior when indigenous price of the land more closely reflects its value rather arrangements on land persist.15 than the added costs associated with verifying its own- The fourth situation occurs where the benefit of for- ership status and physical location. mal titles is low because of failures in other agricultural Property owners may clearly demand the establish- institutions. That has been the case in Kenya, for ex- ment of formal titling systems when informal systems ample.16 If complementary markets for credit and for become less effective. This usually occurs when land be- marketing of inputs do not work, then the first policy comes relatively scarce and in dispute. Increased open- responses, given limited institutional or organizational ness to other communities and competition in product resources, should be in those areas. markets-reflecting strong market demand for agricul- Available empirical evidence from studies on Asia tural output-has often increased the demand for for- and Latin America suggests a positive relationship be- mal titles. In areas of new settlement or frontiers (as in tween tenure security and investment. For example, Brazil), formal titles can enhance the security of agri- studies for the Brazilian frontier found formal titles in- cultural ownership.11 For 35,000 squatter families liv- creased productivity by providing clear information ing on encroached forest reserve land in Thailand, a about ownership rights in undeveloped areas.17 land reform project in the 1980s provided occupancy Land titles can also improve access to credit. But ti- certificates that could be upgraded to full land titles. tles alone are not sufficient-cross-country experience Since the 1980s the World Bank has been supporting suggests that the difference has been the existence of land titling projects in Thailand. Studies have found complementary institutions. In Thailand the existence that these efforts have encouraged significant produc- of formal land titles has facilitated the flow of both for- tivity-increasing investments and greater access to mal and informal credit (box 2.5). Moreover, increased credit. The Bank is continuing to work with the Thai investments in titled land raised its value and improved government on a 20-year program to improve the land access to credit.18 titling and administration system.12 At the same time, studies have generally found that In other situations community-based approaches formal titles have little effect on access to and use of offer a cheaper and effective alternative to formal insti- credit in very poor regions in Africa, India, and some tutions.13 The first situation occurs where buyers and parts of Latin America (box 2.6). Two related factors sellers know each other at the local level and where explain this finding. First, complementary formal credit there is strong peer pressure to avoid socially disruptive institutions may not be widely available. A study of two property disputes. In these cases the main source of de- villages in southern India found that transferable land mand for land is often from within the community; the rights had little effect on credit, probably for this rea- community is strong and close-knit; there is consis- son.19 Land titles alone may not lower the high costs of tency and continuity of community leadership; and any enforcement and of managing very small loans that for- certificates of ownership issued by those in the commu- mal lenders deal with in lending to small farmers. For nity are accepted by others in the community. The sec- instance, a study for Paraguay found that the effect of ond situation occurs where community arrangements formal titles on credit varied strongly with size. Smaller are also legally valid and enforceable. Under Nigerian producers holding fewer than 20 hectares were ex- law, for instance, "customary tenure" is defined as those cluded from the credit market.20 Second, the lack of FARMERS 37 Informal collateral transactions using land titles Do indigenous land rights constrain agricultural in Thailand investment and productivity in Africa? While formal title documents for rural land can facilitate Most African farmers still hold their land under indigenous, credit transactions, the costs of registering liens can be customary, or communal land tenure systems. In the tra- high, and the process can be time-consuming. Therefore, ditional African society, the household, the village, and the lien registration may not be compatible with loan transac- kin group provided insurance against risks, access to in- tions of relatively smaller amounts or short duration. Alter- formal credit, and security. Lineage rules of inheritance native arrangements have emerged, however, that take helped to enforce intergenerational transfers. The threat advantage of the value-enhancing effects of titles on col- of sanctions, which included exclusion from the social lateral while avoiding the high transaction costs of formal structure and its benefits, was the major instrument of en- lien registration. In Thailand in the 1980s a study found forcement of the rules. Even where households have be- that borrowers sign a power-of-attorney authorization to a come geographically dispersed, the common inheritance lawyer representing the lender (typically a trust of local of land in the village and the social support system of the businessmen) and leave the title document with the attor- traditional society continue to bind them together. ney. The cost of foreclosure in such a case is low, and the In the past such land tenure systems were thought to risk to the lender is reduced considerably. This procedure provide insufficient tenure security to induce farmers to screens out borrowers with overly risky projects at a very make necessary investments In land (World Bank 1974; low transaction cost compared with a formal registration Harrison 1987). But research has shown that such sys- of lien. While these arrangements have been documented tems can be effective. The evidence from rain-fed crop- for Thailand, they are likely to exist in many other regions ping areas suggests that indigenous tenure systems have of the world where the transaction costs of registering for- been flexible and responsive to changing economic cir- mal liens are high. cumstances (Place and Hazell 1993; Bruce and Migot- Adholla 1994). Harrison (1990) found that smaliholders in Zimbabwe, despite not having private title to their land, have achieved rapidly increasing maize yields and that their productive performance is not inferior to that of the biggest commercial farmers in the country. Mighot- other complementary formal institutions, specifically Adholla and others (1 994a) found similar results for Ghana. enforcement mechanisms, makes a difference. In Kenya, Source. Bruce and Mighot-Adholla 1994; Collier and Gun- for instance, where banks were prevented from foreclos- ning 1999; Soludo 2001, World Development Report 2002 ing on property used as collateral, a study found that background paper. banks did not make loans to farmers despite the exis- tence of formal titles.21 rules for original adjudication of registered title, estab- Al.; fi;#;: ef7 + I '.K.'' .g' fl ½,i,'. n/5s"ei lish if and how provisional rights can be registered, and Formal land market institutions include land registries, stipulate how these rights subsequently mature. The titling services, and land mapping. In building these in- government also needs to establish an authority (which stitutions, three characteristics should be kept in mind: can be public or private) to ensure the impartial main- clear definition and sound administration of property tenance of land registers, to determine the nature of rights; simple mechanisms for identifying and transfer- these registries, and to delineate the method by which ring property rights; and thorough compilation of land a register for the whole jurisdiction is to be compiled titles and free access to this information.22 Although and subsequent transactions are to be recorded. this discussion focuses on rural land markets, most of Clearly defined land parcels need to be based on cred- the lessons hold for urban markets as well.23 ible land surveys. Otherwise, increasing land disputes- Clear definition and sound administration. A land the resolution of which, given the overworked judicial registry, where titling information is filed, helps to solve systems in many developing countries, usually takes a the central problem of information on property rights. long time-can undermine the fundamental aim of Many of the functions of a land registry can be per- land registries. In Indonesia, for example, land disputes formed by the private sector. But the government has a account for 65 percent of all court matters. role in ensuring that the registry provides comprehen- Administration of the surveys has to address two sive ownership evidence to the public at low cost. For concerns. First, survey standards should be commensu- this, it has to enact land registration laws that define rate with the country's (and region's) level of economic 38 WORLD DEVELOPMENT REPORT 2002 development. In Zambia, for instance, standards re- Box 2.7 quire the same degree of survey precision for office A transparent and accessible institutional blocks in the capital as for 5,000-hectare farms in framework for granting land rights in Mexico sparsely populated areas.24 In poorer countries, more comprehensive survey coverage of land boundaries at a Mexico established a special institutional infrastructure for comprehensiv surve coverage of land boundaries at a granting land rights, which has three elements: (a) an om- lower level of precision and cost (using neighboring budsman's office to supervise the regularization of land parcels and landmarks) may be preferable to a low level ownership; (b) a system of special courts to attend to the of coverage at a high level of precision (say, satellite- large number of existing land conflicts (and provide an op- aided mapping). Second, there needs to be an ade- portunity for quick appeal of any irregularities occurring dur- ing the regularization process); and (c) a modern registry to quate supply of survey professionals, so that the land record land rights that had been established to ensure that registration process is not unnecessarily lengthy. In they could be used in commercial transactions. Zambia, for instance, supply restrictions by the tightly Beneficiaries agree that the program has increased knit association of surveyors meant that there were only tenure security, and evaluations have demonstrated the positive impact it has had on the functioning of land mar- seven qualified surveyors in the entire country in 1994. kets. This transparent and accessible institutional frame- Indonesia, Malaysia, and the Philippines have also re- work has also improved governance in areas that were ported similar restrictions and lack of surveying capac- hitherto dominated by local cliques and party bosses. ity.25 Pressure by media groups, civil society, and gov- ernment to eas suc aniopttv beavo coul Source: Deininger 2001, World Development Report 2002 eminment to ease such anticompetitive behavior could akrudppr yield results. _____________________________background paper. Simplicity of identifi cation and transfer. In practice, establishing formal land rights can be a lengthy and cumbersome process. There are simple ways to ease this. technical, administrative, and legal impediments may One is to convert occupancy rights into full title. In cause this problem. For example, in Indonesia and Mozambique, for instance, land rights are granted to Madagascar a lack of coordination between the legal cultivators based on actual occupation for the last 10 and the fiscal cadastre, or official register of land own- years. Oral testimony is sufficient to support land own- ership, prevents the government from knowing how ership claims, and communities can request formal ti- much land it owns. tles at any point. Similarly, oral testimony is being ac- Land registries that are not publicly accessible raise cepted in many cases of formalizing the quilombos in the cost of transactions. In Tajikistan and several other Brazil (see box 2.3). Suffering from a backlog of land countries of the former Soviet Union bureaucratic in- disputes in the regular court system, Mexico established termediaries are needed because the land registries are specialized agrarian courts (box 2.7). The admission of closed to the public. In other countries, such as Indone- oral evidence and a degree of decentralization have made sia, there is separate title recording, which requires ex- such courts accessible to the poor at reasonable cost. tensive cross-referencing between the legal and fiscal Computerization can also simplify the identification systems. This separation affects not only the speed of and transfer process (as in the example of Andhra access (which can be eased by computerization), but Pradesh, India, given in box 1.8). Although the initial also the integrity of the system. investments in technology can be large, they can have Experimentation through pilot projects helps iden- high payoffs in speeding up land transactions in densely tify institutions effective for a given context. An exam- populated areas. Tax payments can also be used as a ple comes from C6te d'Ivoire, where the World Bank proof of possession that can eventually be converted is working with the authorities on the Plan Foncier into ownership. This mechanism also provides an incen- Rural. A pilot project helped to develop the methodol- tive for landowners to pay taxes. Finally, transferabil- ogy for the systematic clarification and certification of ity of land rights can be eased by reducing regulation- rural land tenure. Mapping and documentation of land induced costs of transactions (see box 2.4). rights is carried out by the professional team, accom- Thorough compilation andfree access to titling infor- panied by the land user or owner, neighbors, and vil- mation. Incomplete land registries, where certain land lage chiefs, moving from field to field within a village plots are not part of the information base, are a com- area. Any disputes that arise are settled by the entire en- mon feature of developing countries. A combination of tourage on the spot. When mapping and documenta- FARMERS 39 tion of ownership and user rights are completed, the in- Figure 2.2 formation is made publicly available so that claimants Evolution of formal institutional borrowing of can openly register disagreements. If no conflicting farm households, selected Asian countries claims to a parcel have been made within three months, the tenure status is considered satisfied.26 percent Japan Building effective and accessible 80 - Korea rural financial institutions 8 A One study of the rural environment states that 60- Few banks would even consider making agricul- 40 - 4, tural loans, and those who did charged extremely high interest rates. Rural credit was fertile ground 20 - Philippines , for the loan sharks, and year after year, farmers , India turned over their crops to help pay exorbitant inter- oL L I I I est charges on loans made to keep theirfarms oper- 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 ating. Should a crop fail, the chances ofa farmer ex- tricating himself and his familyfrom a loan sharks Source Desai and Mellor 1993. clutches were virtually non-existent.27 This study was depicting the situation in rural North Carolina in the United States in the early 20th century. Indonesia's Bank Rakyat Indonesia, was improved by The description could apply just as accurately to many pooling agricultural risks with nonagricultural ones. developing countries today, where formal institutions The broader development of financial markets (chap- such as commercial banks have relatively little incen- ter 4) and the development of complementary institu- tive to offer services to rural clienteles. tions such as those for enforcing contracts are also im- Over time, formal credit provision has increased in portant for this process. rural areas of industrial and some developing countries But in many developing countries, effective and (figure 2.2 shows the evolution for some Asian coun- accessible formal rural financial institutions are still tries). Increasing prosperity among farmers; better rural rare-largely because of the lack of complementary in- infrastructure; integration of the urban and rural finan- stitutions. Accessing finance is particularly difficult for cial systems; and the development of complementary poor farmers. To address this problem, policymakers institutions such as formal credit histories or collateral initially created financial institutions that were specific systems for rural borrowers, which lower the costs of to the agricultural sector, without much success. lending, have all contributed to this increased access to Developing countries have often relied on trans- formal credit (chapter 4).28 In some countries, specific plants of rural and agricultural credit institutions that rural credit institutions such as cooperative banks and have been successful in the United States, Western Eu- credit unions have also been successfully developed. rope, and Japan. Among them are various forms of co- But the provision of crop insurance for farmers has had operatives (often adapted from the German Raiffeisen mixed results even in industrial countries. 29 model), post office deposit schemes, and state market- In industrial countries today, specialized commer- ing agencies that provided credit. French colonies in cial institutions for offering credit, such as microcredit Africa built specialized agricultural banks based on the organizations and agricultural development banks, are original design of the Credit Agricole. Latin American relatively rare. As markets became more integrated, non- countries have often used lessons from the historical ex- specialized commercial banks began to supply credit perience of the United States to set up public and co- to agriculture, supplanting both informal credit insti- operative farm credit systems.30 tutions and specialized agricultural banks. The finan- By the late 1980s, most of these institutions had cial viability of many of the formal specialized agri- clearly failed. Part of the problem was specific-trans- cultural banks, such as France's Credit Agricole and plants were not suited to country circumstances. But 40 WORLD DEVELOPMENT REPORT 2002 other issues also contributed. Urban bias policies, re- Table 2.1 pressive financial policies, and loan targeting to narrow Types of rural lending institutions interest groups reduced the financial viability of banks lending to rural sectors. Debt relief programs also raised Type Informal Formal effective costs for lenders. of loan institutions institutions As a result, informal financial institutions still dom- Secured Pawnbrokers Commercial banks inate in most poorer countries and for poorer farmers. Moneylenders who Rural development In the mid-1990s, 81 percent of rural borrowing in of tand titlession banks Nepal was from informal sources, while farmers in Labor-pawning programs Nigeria received 30 percent of their loans from infor- institutions Mortgaged credit mal moneylenders and 40 percent from esusu clubs (co- fro(l traders operative credit arrangements).31 Within countries, informal sources of credit are dis- Unsecured Moneylenders Some credit unions proportionately used by poorer farmers. Data from Credit from traders Credit cooperatives proportionately used by poorer ~~~~~~~(purchasers of Farmers' Nepal in the 1990s show that 97.5 percent of those output) associations with per capita consumption below 2,000 rupees (the Credit from traders Nonbank financial very poorest rural group) borrowed from informal (sellers of inputs) institutions Friends and family (e.g., insurance sources. No rural borrowers with per capita consump- Savings groups companies) tion above 50,000 rupees borrowed informally.32 Sur- (including Microcredit groups veys from India and Pakistan show similar results. 33 In ROSCAs, ASCRAs, and Thailand a study found that nonborrowers and those chit funds) who borrowed exclusively from the informal sector Labor-bonding tended to have fewer assets as well as lower incomes. institutions Only the largest farmers borrowed from commercial Note: ROSCAs are rotating savings and credit associations; ASCRAs are accumulated savings and credit associations. Labor pawning, a banks as well as some farmers associations, and infor- historical artifact, involved offering a family member's labor as se- mal loans were also smaller in size.34 curity. Labor pawning was prevalent in precolonial West Africa But the mix between different types of informal Austin and Sugihara 19931. lenders has changed with the increased commercializa- tion of agriculture.35 In India, Pakistan, and the Philip- pines studies report that crop loans from moneylenders weak formal institutions, these costs are still cheaper and landlords have been replaced by loans from "com- than the costs of gathering credit information about mission agents," or traders, who advance credit to buy many small and scattered borrowers or of attempting to inputs against a promise thar the farmer will sell the enforce a contract through an inadequate legal system. produce through them at harvest.36 These arrange- Building effective systems for secured transactions ments are usually better for the poorer farmers because can promote the development of formal rural financial they provide access to otherwise unavailable input and institutions (chapter 4). In some developing countries, marketing arrangements (see box 2.1). the use of formally collateralized loans is quite limited. Land cannot be used as collateral if legal claims, such I/mp rovinlg r,'ii-al leindinig institutions as laws limiting "ownership" of seized land to use rights Information and enforcement issues are fundamental only, are not clear. In some developing countries many problems faced by both informal and formal lenders in assets, such as warehouse receipts or inventory credit, agriculture. Lenders overcome these problems through are still not recognized as collateral. Indian law explic- a range of institutional mechanisms, which can be "se- itly recognizes warehouse receipts as title documents, cured" or "unsecured" (table 2.1). but there is no such recognition in Ghanaian law. One Secured loans, offered in exchange for some collat- way around this enforcement problem is to use mov- eral, are the preferred option for most formal lenders able property as collateral-it can be given to the cred- in rural areas because such loans automatically lower itor to hold. information and enforcement costs.37 Assessing the Collateralized loans tend to be given mostly to larger value of collateral and selling the collateral in case of farmers in developing countries. For the lender, such default can be costly. But in environments with overall farmers usually have two advantages over poorer and FARMERS 4I smaller counterparts. First, collateral is easily available- Box 2.8 their property rights over land are more likely to be The intersection of formal and informal lending: formally registered, and they are more likely to have marketing agents in the Philippines movable property of high value. Second, because loan Rice marketers in the Philippines-paddy traders, rice amounts are larger, the lender's unit cost of processing millers, wholesalers, and retailers-act as moneylenders the loan or using the formal judicial system for enforce- primarily to establish a claim over the farmers' produce and ment are lower. to ensure that they are part of the trading chain. A key to Poorer farmers often rely on unsecured loans from their success as credit intermediaries is that in the absence both formal and informal sources. Rural credit institu- of sufficient resources of their own, the rice marketers bor- tionsthn formaloping mfountrm es, whetherforal or infor- row much of their lendable capital (80 percent in one sur- tions in developing countries, whether formal or infor- vey) from formal financial institutions. The traders' intimate mal, gather information about the borrower's reputa- knowledge of and close contact with the farmers and sub- tion by word of mouth. These institutions are usually sidiary traders ensure that information and enforcement localized and'have easy access to information. A study costs are low and that repayment rates are high. More- for Thailand, for example, found that informal lend, over, the rice marketers are a good risk for formal institu- for Thnailand, for example, found that informal lend- tions because they have a good history of repayment and ers are often the shopkeepers in the village because the size of their loans is relatively large. These informal the store acts as a center for village gossip and thus in- rural lenders may also use banks as places to keep their formation.38 Some formal institutions, such as credit savings. unions and cooperatives, also tend to be localized and Source: Floro and Ray 1997. can use past credit history with the institution itself as the main guide for future lending. Formal and informal lenders, and large and small rural borrowers, interact in complex ways in financial to another. Experimentation around a basic institu- markets in developing countries. Formal lenders such tional form has been a good way to identify success- as commercial banks lend directly to the larger farmers ful institutions. So thriving microfinance institutions and traders for their personal use. But they may also around the world differ in operational details. Innova- make wholesale loans to traders, who then act as infor- tions on different aspects, such as the targeted group mal lenders, making retail loans in smaller amounts to and the repayment periods, have been altered to suit small farmers and middlemen (box 2.8). the characteristics of different countries. How can rural lending institutions be improved? Ex- Giving incentives to loan officers, and rebating a perimenting with context-specific institutional design small part of the loan for early repayment, the Unit and using new technology to reduce costs are two prom- Desas, part of Bank Rakyat Indonesia (BRI-UD), im- ising avenues. proved the repayment rate (to 92.5 percent in 1995) .41 The design of new rural lending institutions can be Successfully experimenting with its institutional form, improved by incorporating innovations based on the such as the size and composition of its "solidarity successful elements of informal institutions and formal groups" of borrowers, BancoSol in Bolivia has grown interventions. Social and peer pressures, which are par- from a subsidized lending program operated by an ticularly effective enforcement mechanisms in informal NGO to a self-sustaining commercial bank. Learning lending, can contribute to the success of formal rural from successive experiments about how to adapt credit credit programs in weak institutional environments.39 delivery to the local context and farmer needs, an Al- Newer microcredit institutions, including those backed banian rural credit program funded by the World Bank by the World Bank and other donors, have adapted grew from offering small-scale credit in seven village many of the same mechanisms for inducing repay- credit funds in 1992 to a full-scale rural development ment.40 Some nongovernmental organizations (NGOs) project supporting hundreds of village credit funds by are also trying to emulate informal lenders by serving 1995 and a follow-up microcredit project since 1999.42 as a bridge between banks and poor borrowing groups. Experimentation helps to understand the impact of MYRADA in southern India acts as such an interme- a rural finance institution. For example, an element of diary, aiming to help borrowing groups deal directly the Grameen Bank's programs is the requirement that with the banks after a few loan cycles. borrowers repay their loans in small installments ac- Successful rural finance institutions cannot always cording to a rigid (weekly) schedule.43 Imposing a reg- be transplanted from one socioeconomic environment ular repayment schedule can be costly, however. It re- 42 WORLD DEVELOPMENT REPORT 2002 duces the attractiveness of long-gestation projects, such transaction costs and risks by realizing economies of as those in agriculture, and helps to explain why infor- scale and diversifying their portfolios.52 mal lenders appear to thrive even in villages where mi- The most successful institutions began by financing crofinance programs are active.44 mainly nonfarm activities and started making agricul- Technological innovations can also help credit pro- tural loans only after they had grown into mature in- vision. First, the continuing extension of credit-rating stitutions. A striking aspect of successful rural financial services to rural areas brings the promise of eventual in- institutions is that they all operate in relatively densely tegration of urban and rural banking, as has occurred populated rural areas.53 Geographic density reduces in industrial countries. So far, such information inter- costs of transactions and makes it feasible, for example, mediaries are developing mostly in middle-income for SafeSave in Bangladesh to send out staff to collect countries, such as Argentina.45 "Meta-information in- savings from its members on a daily basis.54 Thus there termediaries" are also being developed. These rate fi- is a question whether these designs can exist in sparsely nancial intermediaries themselves rather than their populated countries. clients, the first step in the development of credit refer- ence bureaus. Further, they offer financial information Developing rural savings institutions in a standardized format. The Micro-Banking Stan- Besides access to credit, safe and liquid savings instru- dards Project, funded by the Consultative Group to As- ments are vital for farmers' well-being. In the absence sist the Poorest, has recently collected, analyzed, and of loans, savings are the only resource for investments. published data on the financial status of participating They also provide "self-insurance" against the periodic microfinance organizations.46 Micro-Rate, a private shocks to income common to agriculture, as farmers credit-rating agency that specializes in evaluating micro- add to savings in good times and draw on their savings lenders, offers a similar service.47 when times are difficult. Market women in rural parts Second, information technology can reduce transac- of western Africa often save their daily earnings by giv- tion costs for both state and private actors. For exam- ing them to susu men (itinerant savings collectors). The ple, Compartamos, a Mexican NGO, has started giv- fact that the depositors are willing to pay the deposit ing its field staff inexpensive handheld computers to taker a fee suggests that there is a demand for safekeep- record data, thereby reducing paperwork and speeding ing institutions. synchronization of data. This has allowed field staff to Savings institutions in rural societies are still infor- access and update records far more easily.48 mal, and savings are often not in financial assets. Rural Some institutional designs tailored to poor rural households in developing countries save in physical as- areas have been successful. Small minimum balance re- sets such as livestock and jewelry. This does not always quirements and liquid savings products are attractive to provide security because these assets may not hold their rural borrowers. In such schemes offered by some rural value in bad times. The success of such a strategy also banks in Asia and Latin America, lenders offset high depends on the level of development of the market for unit costs by having interest rates increase with the ac- that asset (box 2.9). count balance. Administrative costs can also be lowered The development of formal rural savings institutions, by maintaining lean field offices and offering efficiency as with rural credit institutions, is inhibited by high bonuses to motivate staff to be more productive.49 costs of operation. Governments have attempted to pro- Many of the elements explaining the successes of in- vide savings facilities in rural areas. But these efforts, on stitutions such as the BRI-UD in Indonesia and the average, have failed to cover their administrative costs. Bank for Agriculture and Agricultural Cooperatives in Even some of the fast-growing microfinance programs Thailand can be found in any successful institution.50 have relied on external agencies or governments for their These include simplicity in financial contracts, trans- sources of funds. The only major exception has been parency in operations, and integration across markets. BRI-UD in Indonesia, but here the flow of savings has Operational autonomy and freedom from political in- been from the rural sector to the urban sector.55 terference are critical for providing the institutions with the freedom to experiment with the terms and types of Insuring agaitnst risk in agriculture financial products offered.51 Moreover, successful rural Agricultural risk is considerable and covariant-usually financial institutions tend to be large, usually serving all borrowers in an area are affected similarly. These millions of households. This allows them to reduce problems are compounded by information problems, FARMERS 43 which are especially large in developing countries. For Box 2.9 crop insurers, specific events such as floods or a locust Livestock as savings: contrasting evidence from attack are verifiable, and thus these risks are insurable. India and Burkina Faso But when the yield on an insured crop is reported to be lower than expected, the reasons can be many and are Faced with risky environments, rural households often rely on the sale of assets to smooth consumption in the face difficult to untangle, and the true value of output is of income shocks. The main assets that farm households hard to verify. Insurance also provides incentives for possess are productive assets, in the form of land or live- fraud or "moral hazard" (low effort or investment by stock. Unlike land, livestock is portable and may offer a the farmer). useful way for households to buffer against production shocks. In an influential article, Rosenzweig and Wolpin As a result, formal insurance mechanisms for agri- (1993) presented evidence that the sale of livestock, no- cultural households are difficult to implement even in tably bullocks, is used as a consumption-smoothing device richer countries. Unsuccessful attempts to offer gener- by rural households in India. The market for bullock sales alized crop insurance in developing countries have con- and purchases is well integrated regionally, with 60 per- ibuted to the decline of agricultural banks.56 cent of bullock sales in the sample villages taking place tnbutr d to the dechne ot agncultural banks. with buyers outside the village. As a consequence, bullock In both industrial and developing countries, premi- prices do not seem to vary with village-specific production ums collected in general agricultural insurance schemes shocks, an important consideration for choosing an asset have never been enough to offset the indemnities paid that one might have to sell in bad times. An interesting contrast is provided by Fafchamps, out to farmers (figure 2.3). The situation is clearly Udry, and Czukas (1998), who examined livestock sales worse when high administrative costs are added to the and purchases in Burkina Faso and found very little evi- costs related to monitoring the insured. These schemes dence of a similar phenomenon. What explains the differ- have historically needed significant government subsi- ence in these two sets of findings? Livestock markets dies to stay operational, in rural Burkina Faso, which is much less densely popu- lated than India, are less integrated. Furthermore, the Narrowly focused "named-peril" schemes are the only more widespread the agricultural shock (in the case of agricultural insurance mechanisms that have functioned Burkina Faso, the study period included a drought that af- without large government subsidies; they have succeeded fected large parts of the country), the more contempora- p)recisely because they minimize the potential for decep- neous are household decisions to sell livestock, and the precisely becauseIthey minimize the potential for decep- lower the efficacy of sales in smoothing consumption. tion by farmers and do not depend on the farmers' ac- tions or investment. In industrial countries today, agri- cultural insurance is offered only as event insurance, for difficult (chapter 9). Village-based mutual insurance is example against hail or floods-risks whose occurrence also limited because the main sources of risk affect the is relatively easy to monitor. In the United States, entire community. Informal insurance arrangements named-peril plans are the only mechanisms offered by thus face a difficult tradeoff. The very factors that make private insurers without government subsidies.57 these informal risk-sharing mechanisms work-geo- If general crop yield insurance is to be provided in graphic proximity and social ties-also limit partici- developing countries, it is likely to require subsidies, pants' ability to diversify as a way to lower risk. even if administrative costs are kept at a minimum. A 1995 study of a general crop insurance scheme in India confirmed that it offered considerable subsidies.58 technology and innovation Given the limited avalability of formal insurance for Two centuries ago, Thomas Malthus argued that the farmers in developing countries, most insurance ar- world would exhaust its food supplies because popula- rangements are informal.59 A study of northern Niger- tion grew geometrically but agricultural production ian villages found that credit contracts were dependent grew arithmetically. Technological change has proved on the nature and amount of shocks affecting borrow- Malthus wrong. Agricultural innovations-such as ers, with lenders bundling credit and insurance.60 high-yielding seeds, herbicides, fertilizers, agricultural "Social insurance" agreements between members of machinery, and resource management techniques- a village stipulate that those who are better off once allowed food production growth to outpace population crops are harvested and sold are required to make trans- growth. One of the fastest ways to increase agricultural fers to the needy. Sometimes, as when individuals have productivity rapidly is the adoption of new agricultural an incentive to leave the community, enforcing this technologies. Rapid productivity growth boosts farmer agreement so that these transfers actually occur can be incomes and helps farmers manage risk. 44 WORLD DEVELOPMENT REPORT 2002 Figure 2.3 Financial performance of generalized agricultural insurance programs Total costs/premiums 7- 6 * Indemnities paid/premiums g Administrative costs/premiums 5- 4- 3- 2- Brazil Costa Rica India Japan Japan Mexico Philippines United States (1975-81) (1970-89) (1985-89) (1947-77) (1985-90) (1980-89) (1981-891 (1980-90) Country and period Note: The height of the bars indicates the sum of indemnities and administrative costs as a ratio to premiums collected. For premiums to fully cover costs, this ratio should not exceed one. The f gure for India does not include the 1989 rabi season, and data for administrative costs are not available. For Japan data are for paddy only, and administration cost data are based on 1989 on y. For Mexico figures are for crop insurance only. Source: Hazell 1992. The Green Revolution in South Asia during the nonrival (one person's use does not lower another per- 1960s and 1970s illustrates the benefits of agricultural son's benefit from it) and nonexcludable-a person technology. During the Green Revolution small farm- who does not pay for the product can still receive it ers dramatically increased their productivity by adopt- (table 2.2). Private firms will not supply goods and ser- ing high-yielding rice and wheat varieties and using vices based on these technologies because they cannot complementary inputs of irrigation and fertilizer. The restrict the benefits from the technologies to only those Green Revolution also generated secondary income ef- who paid for them. Farmers may not pay for market- fects for landless households.62 More generally, new ing information, for example, if they are able to receive technologies have more than doubled global crop yields it free from friends and peers. One study in the United over the last four decades.63 Between 1965 and 2000 States estimated that between 1975 and 1990 private productivity gains in output per hectare of cereal crops returns to seed companies were only 10 percent of so- averaged 71 percent globally.64 cial returns for nonhybrid seeds.66 Research to develop agricultural technologies, as well These problems are compounded by the large exter- as extension services to deliver them, generate high so- nalities associated with new agricultural technologies. cial rates of return across regions-usually more than 30 For example, a farmer may impose a negative external- percent (figure 2.4).65 Newer irrigation management ity on his neighbors by failing to vaccinate his livestock techniques, as well as seeds resistant to drought and to against a disease that then spreads to their herds. Con- pests, have helped to reduce risk. Finally, as discussed in versely, natural resource management techniques pro- World Development Report 2000/2001, numerous stud- duce positive externalities by protecting the quality of ies show that the poor benefit from advances in agricul- resources for future generations. As the Green Revolu- tural technologies, not only through reduced risk, but tion showed, adopting new agricultural technology has also through increased demand for their labor and lower significant positive externalities for the rural poor. Fi- food prices. nally, the lengthy time needed to develop new tech- Many agricultural technologies have characteristics nologies and the uncertain payoffs can lead to less pri- of public goods. That is, they may be at least partially vate research than would be socially desirable. FARMERS 45 F igur 2 4 International experience has shown that for private Median rates of return on agriculture research goods technologies such as machinery and biotechnol- and extension by region ogy for commercial farms, research and extension ser- vices can be left to the private sector. Public involve- 62% ment is required only for public goods, such as knowledge-based technologies, and where there are 50% strong market failures and externality effects (such as 47% | 43% 40% the secondary effects of technology for poor house- lll 34%o N N 36% holds). Yet even in these cases, public involvement does 29% I * not necessarily translate into public provision or mo- *l * 11 | 11 11 nopoly. Moreover, in research there are potentially large * * *3 * 11 | payoffs from regional collaborations among several fl* *fl fl countries-particularly for smaller and poorer coun- I *fl * *fl tries, where research capacity is low and markets are Number of ,N , Np r # small. studies 09 Q x *t *G ,c One of the primary barriers to the adoption of new agricultural technology is lack of knowledge.67 Exten- sion services are an institution aimed at filling this gap. 4 * 4 * Another major barrier is overregulation of agricultural Industrial Developing technology transfer. This section discusses these two countries countries factors. Note: Rates of return include private and social returns on Investment Extension services. The main functions of extension for both extension and research programs between 1953 and 1998. Source: Alston and others 2000 services are twofold: to inform farmers of new prod- ucts and techniques, and to gather and transfer infor- mation from farmers to other participants. This in- Taken together, these characteristics of agricultural cludes collecting feedback on farmer needs as input for technology suggest a need for at least some public in- research priorities, and learning techniques from one volvement in the development and delivery of agricul- farmer and sharing them with others, for example, ir- tural technologies. The question is, to what extent and rigation techniques. in what form? In the 1950s and 1960s large-scale extension sys- Public institutions that have provided agricultural tems emerged when governments throughout the world technology are under pressure to reform. Fiscal pressures invested heavily in services for delivering new agricul- and criticism for inefficiencies have slowed financial tural technologies.68 Estimates show that public sources support for public institutions in many countries. More- provide 81 percent of total extension services, with uni- over, technological shocks, such as the advent of new versities, parastatals, and NGOs accounting for 12 per- biotechnologies, and global movements to strengthen cent, and the private sector accounting for only 5 per- intellectual property rights have focused attention on cent. Most of the private sector extension services are the role of the private sector in developing agricultural provided in industrial countries.69 technologies. Liberalization of entry into agricultural re- The benefits of extension services are enormous- search and extension services, as well as increased com- more than 80 empirical studies have demonstrated that petition, have strengthened existing institutions and led extension services generate rates of return averaging to innovative institutional designs. Better information more than 60 percent.70 Numerous other studies show sharing between providers and users of technology- that extension services substantially improve technology often through decentralization and through interna- adoption rates, awareness, and productivity.71 These re- tional arrangements among technology providers-has turns are not only for dissemination of sophisticated also improved institutional quality. technologies. Information sharing on rural technology, 46 WORLD DEVELOPMENT REPORT 2002 Table 2.2 Where will the private sector invest in agricultural technologies? Type of good Public good Common pool Toll Private Features Nonexcludable and Rival but not excludable Excludable but nonrival Excludable and rival nonrival Examples * Weather forecasts * Self-pollinated seed * Soil analysis * Hybrid seeds * Market information varieties * Farm management * Biotechnology * Livestock manage- * Shared fishery, computer programs products (for example, ment techniques common pasture a Training courses and inputs and seeds) * Fertilizer application management private consultation in * Fertilizers, chemicals schedules techniques farm management and * Agricultural machinery * Natural resource production practices * Veterinary supplies management techniques Likelihood of Very low Low Higher High private provision Source: Umali-Deininger 1997; World Bank 2000a. including simple innovations for the poor and illiterate, tralized public extension institutions. This strategy in- can have a substantial impact on productivity (box 2.10). volves transferring responsibility for administrative, Despite these successes public extension services fiscal, and political decisions from central to local or have been criticized for being inefficient, ineffective, regional authorities-usually to government agents but and poorly targeted. In Kenya, for example, an evalua- potentially to community groups. By bringing decision- tion in 2000 found that government extension services making closer to clients, decentralization can increase supported by the World Bank did not meet farmer information flows, build local capacity, and improve ac- needs and were targeted toward groups that had a low countability. This in turn can improve efficiency, ser- marginal impact on overall productivity.72 A 1997 vice quality, and access. After the decentralization of ex- World Bank review of 31 extension projects revealed tension in Colombia, costs per farmer fell 10 percent, pervasive problems of inadequate client orientation, the area covered by extension services tripled, and the weak human resource capacity, and low levels of gov- number of beneficiaries more than doubled.76 ernment commitment.73 Despite its potential, decentralization of extension Some of the problems in public extension services services presents three major challenges. A national originate from external factors, such as lack of politi- framework is required to avoid confusion of responsi- cal commitment and dependence on complementary bilities between administrative levels and wide varia- policies. Another problem is that public extension pro- tions in quality. Also, decentralization should not apply viders are not always made accountable for their ac- uniformly to all extension functions. Some activities, tions, and the capacity to manage large and complex such as policy development, highly specialized techni- extension schemes is limited.74 Three main types of in- cal support, and market information services (in which stitutional reform for improving extension services are there are more significant economies of scale), can be discussed here-decentralization, privatization, and sep- conducted more efficiently by centralized authorities. aration offundingfrom execution. Each reform consists Finally, local governments may lack capacity to im- of a different combination of public and private in- plement these new institutional responsibilities. (For volvement. Other important techniques include more example, in the Philippines inadequate linkages be- participatory approaches and increased use of the media tween research and extension services were exacerbated (chapter 10).75 by decentralization.) 77 While keeping both service delivery and funding A second strategy is the privatization ofpublic exten- within the public ambit, several countries have decen- sion services. The private sector is likely to invest in dis- FARMERS 47 Box 2.10 Box 2.11 Creating an information-sharing network Private sector extension services in Argentina for the poor: SRISTI in India During the 1 970s the productivity of Argentine dairy farm- In Gujarat, India, a seven-year-old NGO, SRISTI, has pio- ing was seriously hampered by poor cattle nutrition and neered a pathbreaking way for poor farmers to tap into the poor farm hygiene. Faced with unstable supply and qual- innovations of their compatriots in the 5,500 villages scat- ity problems, the two largest dairy processors-Santa tered across the state, Volunteer workers armed with lap- Fe-Cordoba United Cooperatives (SANCOR) and La Sere- top computers travel from village to village searching for nisima-established extension services for their suppliers. low-cost innovations that can improve the earning power SANCOR's program included financing for agronomist or quality of life for poor villagers. Innovations covered in technical assistance, farm visits, artificial insemination ser- the SRISTI database include an eminently affordable (less vices, and accelerated heifer-rearing programs. By 1990, than $10) shoulder-carried pump that can be used to spray 120 farmers' groups were participating in the program, the small fields that most poor farmers cultivate and a and each group had assumed responsibility for the cost of small stopper that, when attached to rope-and-pulley sys- technical assistance. La Serenisima created 25 extension tems in wells, allows women to rest during the fatiguing branch offices, each of which provided technical assis- process of drawing water. These innovations are cata- tance to groups of up to 25 medium-to-large-scale farm- logued in a database and then circulated through a quar- ers. La Serenisima's program also made extensive use of terly newsletter. Work Is under way to disseminate the press and broadcasting media to inform farmers of live- database on-line, with villagers retrieving information stock management techniques. through kiosks. To improve access for the illiterate, the The results of these private extension efforts were ex- kiosks can provide data through a voice interface. By di- tremely positive. Although the number of dairy farms sup- rectly addressing the informational constraint faced by dis- plying SANCOR decreased by 24 percent, milk production persed rural communities, SRISTI has a tangible impact on increased by 15 percent between 1976 and 1985, Milk easing the burden of poverty for its constituents. production for La Serenisima jumped by almost 50 per- cent despite a 6 percent decrease in dairy farm areas of Source: Slater 2000. suppliers. Source: Umali-Deininger 1997; World Bank 1989a. semination of goods where knowledge is embodied in the technology itself-for example, in hybrid seeds. a portion of the cost of the extension service and has The private sector is also better able to extract a return the obvious benefit of cost recovery. This type of in- from extension services in commercial farming, even stitutional design may also increase competition by for technologies with public goods characteristics. As encouraging alternative providers to enter the exten- illustrated in Argentine dairy farming, private agropro- sion market. Moreover, the fee payment increases the cessing and marketing firms that contract with farmers accountability of service providers to farmers. An inno- may provide extension services for knowledge-based vative approach in Nicaragua that introduced paid ex- technologies as well as for private goods (box 2.1 1). tension services significantly improved cost-effective- Privatization of extension services has enhanced ness and led to a more responsive service. Even poor competition and helped develop more effective institu- farmers purchased extension services.81 tions. For example, partial privatization in the Nether- The separation of funding from execution can also lands reduced overhead expenditure by 50 percent and take the form of privateprovision with publicfunding. increased farmer satisfaction ratings by 40 percent.7 The main advantage of this separation is to stimulate Commercial providers are not the only solution. In- competition among private sector providers to improve stitutions for collective action such as farmers' associa- efficiency and service quality. Contracting private pro- tions have played a central role in delivering extension viders in specific functions of extension has proved a services-as in the Central African Republic in the successful reform strategy in countries ranging from Es- early 1970s.79 Complete privatization, however, can tonia to Madagascar. In Nicaragua the government has lead to underprovision of public goods or make exten- financed extension services by issuing to farmers vouch- sion services unaffordable for small and subsistence ers for extension services that could be redeemed with farmers, as was the case in Chile in the 1970s.80 either private or publicly provided extension.82 Separating public provision of extension with private Deregulation of input markets. Institutional obstacles finding usually involves charging farmers a fee to cover often restrict the delivery of new technology. Although 48 WORLD DEVELOPMENT REPORT 2002 most industrial countries have liberalized agricultural Developing new technologies for aggricultutre technology markets, governments in developing coun- in developing countries tries tend to overregulate the transfer of agricultural Research to develop agricultural products presents fun- technologies. This is particularly the case in seed mar- damental institutional challenges. First, it is often long- kets, but it also applies in markets for machinery, fer- term and risky and can require significant human re- tilizers, low-risk pesticides, and feed mix.83 Overregu- source capacity. Also, agricultural technologies often lation is of special concern in developing countries, have public goods characteristics and generate externali- since it creates opportunities for corruption in less ties. Third, there is a tension between economies of scale transparent environments and may hinder innovation. in research and development (R&D) and the need for Several types of barriers are applied. First, many de- location-specific technologies. The existence of econ- veloping countries restrict competition, by limiting chan- omies of scale in R&D suggests that research activities nels for the introduction of inputs to parastatal monopo- should be concentrated. But many agricultural products lies or by controlling market entry. Second, governments must be tailored to local conditions, such as climatic have introduced complex systems for testing, approval, and soil conditions-a fact that suggests a need for frag- and release of new varieties. In particular, compulsory mentation in research. For example, frost-resistant registration and certification of seed varieties, often de- wheat developed for Canadian farmers is of little value signed on the basis of public seed-breeding programs, are to farmers in Sudan. This effect is compounded by the unsuited to testing seeds from private plant-breeding pro- information gaps between researchers and users, which grams.84 Finally, key channels for technology transfer, suggests that research institutes need to have effective such as trade, technology licensing, and foreign direct in- communication with end users, often through physical vestment, are often restricted in developing countries. proximity. Removal of various regulatory barriers and introduc- In developing countries, the demand for location- tion of more flexible standards encourages greater pri- specific technologies may be too small to attract private vate sector participation in both research and distribu- sector investment, as evidenced by the lower levels of tion. A powerful illustration of these effects took place such investment (both absolute and relative to gross do- in Turkey during the 1980s, when deregulation of the mestic product) in developing countries (figure 2.5).87 government seed production and sales monopoly (sup- Similarly, patterns of research expenditure indicate that ported by the World Bank) significantly increased most private R&D on agricultural seed focuses on de- introduction of new seed technologies. As a result the velopment products with longer shelf life, herbicide returns to maize yields increased by 50 percent and resistance, and greater suitability for mass production income per hectare rose by $153-equivalent to an an- nual net economic gain of $79 million.85 Similar exam- Figure 2.5 ples exist in the deregulation of agricultural machinery Agricultural research intensity, public and markets in Bangladesh, seed markets in Peru, and agri- private, 1993 cultural input markets in Zimbabwe.86 Besides reducing import and entry barriers to agri- 30 Expenditure as a percentage of agricultural GDP cultural technologies, competition and information , Private flows are stimulated by (a) introducing voluntary seed 2-5 - U Public certification systems, supported by incentives for certi- 2.0- fication and enforcing strict disclosure laws for infor- 1.5 - mation on seeds and other agricultural inputs; (b) in- 1-0 troducing voluntary seed varietal registration or, as an interim measure, introducing automatic registra- 0.5 - tion for seeds approved in selected other countries; and o.o . (c) maintaining only those regulations that address gen- Developing Industrial countries countries uine public health and environmental externality con- cerns, but not on the grounds of protecting farmers Source. Pardey and Beintema 2001. from potential misinformation. FARMERS 49 techniques. In contrast, developing country priorities strong, they can reduce access to agricultural inventions are often greater nutritional content and robustness.88 by increasing prices, as documented by various studies, With some exceptions, notably in research for export and can potentially enable restrictive business prac- crops, the private sector invests little in adapting tech- tices.9'2 In this scenario, poorer farmers in developing nologies to local conditions or refining agricultural re- countries will not have access to wealth-enhancing source management techniques in developing countries. opportunities because they will be unable to afford These are areas of high social returns.89 new technologies. Furthermore, technology develop- Location-specific technologies for developing coun- ment could be hindered when new products are depen- tries are more likely to require public intervention and dent on many other IPRs, raising the costs of innova- local research or adaptation. Three main institutions af- tion. The genetically modified "golden rice," which has fect innovation in agriculture: intellectual property the potential to alleviate vitamin A deficiency, a major rights, agricultural research institutions, and competi- cause of blindness and immune dysfunction in poorer tive grants and negotiated contracts.90 countries, is reported to be developed with technol- Intellectualproperty rights andprivate sector research. ogies protected by up to 70 patents from 31 different Intellectual property rights (IPRs) protect the rents organizations. from innovation by regulating replication. R&D costs Two main strategies can help ensure that IPRs main- of developing some agricultural technologies are high tain incentives for innovation but do not restrict access compared with technologies in some other industries. to new agricultural technologies in developing coun- For example, it is estimated that new plant protection tries. The first strategy concerns the type of IPR system chemicals cost more than $150 million to develop.91 that is implemented. Under the 1994 Trade-Related Also, private firms are less able to appropriate the re- Aspects of Intellectual Property Rights Agreement turns from agricultural technologies than from those in (TRIPS), which sets minimum standards for IPRs in many other industries. In agriculture, products can be member countries of the World Trade Organization, readily imitated through reverse engineering, or seeds two types of plant variety protection systems are per- can be bred and resold. mitted-patent protection, or a sni generis system (mean- Yet much of the empirical evidence on the benefits of ing a design unique to the context).93 IPRs is inconclusive (see chapter 7 and World Develop- The difference between these two options is vast. ment Report 1998/1999). Critics note that numerous Under a sui generis system, farmers can replant seeds agricultural inventions were made without the protec- saved from a previous crop, but under a patent system tion of IPRs, even for private goods. Also, the protection they are generally prohibited from doing that. Simi- conferred by IPRs is highly limited for some technolo- larly, a sui generis system allows breeders to use seeds gies-it would be difficult for a technology producer to freely as they research new plant varieties; a patent usu- enforce IPRs against a heterogeneous group of small ally forbids such use. By choosing the option of a sui farmers who save and replant seeds for their own use. generis system over patents, therefore, countries can Because of this, private sector investment has concen- help to maximize farmer and breeder access to new trated more on seeds with built-in protection for intel- plant varieties. Argentina, Chile, and Uruguay are ex- lectual property. Examples include genetically trans- amples of countries that have already successfully im- formed seeds that will not germinate after the first crop plemented sui generis systems with clauses to help pro- and hybrid seed technology that increases yields and re- tect farmers' access to plant varieties. sistance by combining varieties so that the seeds do not The second strategy being followed is to build the breed true and subsequent crops do not perform as well. capacity to manage intellectual property. This approach The lack of empirical evidence on the benefits of is also relevant for public research institutes in indus- IPRs and the problems with their enforcement raise trial countries, where intellectual property has required questions about whether IPRs create value. A more se- these institutions to invest in resources, appropriate rious concern is that IPRs may impose substantial costs policies, and systems to manage it. In developing coun- on developing countries. IPRs balance the need to cre- tries, managing intellectual property includes seeking ate incentives for innovation with the need to ensure partnerships and development assistance from private fair access to new technologies. If IPRs are excessively technology producers. 50 WORLD DEVELOPMENT REPORT 2002 Some successful examples of capacity building exist. These pressures for institutional reform are rein- The Kenyan Agriculture Research Institute (KARI) and forced by a global slowdown in rates of public agricul- Monsanto established a partnership to develop virus- tural research investment over the last two decades.96 resistant sweet potatoes, with Monsanto providing Some NARIs have faced drastic cuts; in Russia, for royalty-free licensing of intellectual property, direct example, funding for some agricultural research insti- funding, basic research components, and technical as- tutes plummeted by more than 50 percent during the sistance for KARI to develop and test the product in 1990s.97 preparation for its release in 2002.94 In Mexico a multi- Two strategies have proved effective in addressing national corporation contracted to sell intellectual the competition and informational problems facing property to large-scale farmers in the lowlands but do- many public research institutions. These are to clarify nated the technology to small, poor farmers in the high- the public research mandate, and to introduce mecha- lands. In both cases, the private firms enhanced their nisms for information sharing among stakeholders in public relations image at little opportunity cost, since research. neither Kenyan nor highland Mexican farmers would Under the first strategy, specific priorities and re- have purchased the technologies without the donation. sponsibilities for the public sector, based upon public The potential of these arrangements is limited, how- goods and externalities issues, are identified. Key areas ever, because they apply almost exclusively to segments for public sector research include plant breeding for of markets, such as marginal farming areas or markets crops and environments that are overlooked by the pri- in small and poor countries, that would not support a vate sector but that will generate social returns, and pub- private sector return. The arrangements also require sig- lic research where the primary products are information nificant negotiating power and are highly unlikely to and advice, such as resource management techniques be sustained if farmers develop the capacity to pay for and prebreeding products.98 Malaysia, Zimbabwe, and technologies. the Maghreb countries are all examples where NARIs Public research institutions. Agricultural technology are refocusing on smallholders rather than commercial markets in developing countries often cannot support market needs for these reasons.99 By contrast, refocus- private sector returns, even with IPRs, and there are ing on commercial markets in China weakened public public goods and externality effects. So some level of agricultural research output and productivity.100 public sector involvement is required. One such form The second strategy is to build more effective insti- of involvement is public agricultural research institutes. tutions through open information sharing. Several in- Currently, national agricultural research institutes stitutional changes can help address information gaps (NARIs) account for a large share of agricultural re- between technology developers and farmers. For exam- search activity in almost every country and yield high ple, farmer representation on governance committees returns on investment (see figure 2.5). can help to ensure that information on farmer needs is Despite this record, public research institutes are incorporated in research. Successful examples exist in under pressure to reform. They have been criticized for Mali and Zimbabwe. Another approach involves farm- stifling competition by crowding out efforts of the com- ers in testing and adapting new products. This helps to mercial sector. Furthermore, two types of information refine technologies that meet user needs and can also asymmetries-those among public research institutes increase awareness and therefore dissemination of new themselves and between the institute and the farmer- technologies. For example, farmer testing was a factor have caused waste. One study revealed that 40 percent in the rapid adoption of the West African Rice Devel- of African wheat-breeding programs would have gener- opment Association's drought-resistant rice varieties in ated higher returns by screening and adapting foreign Guinea.10' wheat varieties rather than by locally breeding new va- Strengthening the links between extension and re- rieties. Open information sharing could help build bet- search services is another way to improve informa- ter institutions. Public research institutions have also tion flows to researchers about farmers' needs and to faced widespread criticism for lacking information on farmers about new technologies (box 2.12). Research- farmer needs and the incentives to respond to those extension links have, however, had a mixed record of needs.95 success. In China a pilot scheme to establish research- FARMERS 5I Box 2.12 o3