41455 world development report 2008 Agriculture for Development world development report 2008 Agriculture for Development world development report 2008 Agriculture for Development THE WORLD BANK Washington, DC © 2007 The International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org All rights reserved 1 2 3 4 5 10 09 08 07 This volume is a product of the staff of the International Bank for Reconstruction and Devel- opment / The World Bank. The findings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgement on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. Cover design by Chris Lester of Rock Creek Strategic Marketing and Bill Pragluski of Critical Stages. Typesetting by Precision Graphics. Printed in the United States by Quebecor World. Cover photos by World Bank staff members, clockwise from top left: milk thermometer, Lillian Foo; wheat threshing, Alexander Rowland; Holstein cow, Lillian Foo; supermarket beans, Lillian Foo; Andean woman and baby at market, Curt Carnemark/World Bank Photo Library; cotton plant, Arne Hoel. Softcover Hardcover ISBN-13: 978-0-8213-6807-7 ISBN-13: 978-0-8213-6808-4 ISSN: 0163-5085 ISSN: 0163-5085 eISBN-13: 978-0-8213-6809-1 DOI: 10.1596/978-0-8213-7235-7 DOI: 10.1596/978-0-8213-7233-3 Contents Foreword xiii Acknowledgments xv Abbreviations and Data Notes xvii Overview 1 What can agriculture do for development? 2 What are effective instruments in using agriculture for development? 8 How can agriculture-for-development agendas best be implemented? 18 Part I What can agriculture do for development? 26 1 Growth and poverty reduction in agriculture's three worlds 26 The structural transformation 27 The three worlds of agriculture for development 29 Agriculture's development potential shortchanged 38 The political economy of agricultural policy 42 A new role for agriculture in development 44 focus A: Declining rural poverty has been a key factor in aggregate poverty reduction 45 2 Agriculture's performance, diversity, and uncertainties 50 Productivity growth in developing countries drove agriculture's global success 50 Growth across regions and countries has been uneven 53 Differences in performance reflect different underlying conditions 54 Opportunities for a new agriculture through diversification 58 Future perspectives: confronting challenges and rising uncertainties 61 Conclusion--a continuing production challenge 68 focus B: Biofuels: the promise and the risks 70 v vi CONTENTS 3 Rural households and their pathways out of poverty 72 Three complementary pathways out of rural poverty: farming, labor, and migration 73 The variation in rural households' income strategies 74 Rural occupations and income sources 77 Household behavior when markets and governments fail: rational, despite appearances 82 Rural household asset positions: often low and unequal 84 Pervasive risks and costly responses 89 Smallholder challenges to compete 90 Conclusions 92 focus C: What are the links between agricultural production and food security? 94 Part II What are effective instruments for using agriculture for development? 96 4 Reforming trade, price, and subsidy policies 96 Agricultural protection and subsidies in developed countries 96 Agricultural taxation in developing countries 98 Simulated gains from trade liberalization 103 Scope for achieving potential gains 110 Transitional support 112 Public investment for long-term development 114 Conclusions 116 5 Bringing agriculture to the market 118 Food staples: improving commodity trading and risk management 118 Traditional bulk export commodities: maintaining international competitiveness 122 Higher-value urban markets: linking producers to modern supply chains 124 Higher-value exports: meeting product standards 128 Conclusion 133 focus D: Agribusiness for development 135 6 Supporting smallholder competitiveness through institutional innovations 138 Land policies for secure rights and reallocating resources 138 Financial services for smallholders 143 Insurance to manage risk 147 Contents vii Developing efficient input markets 150 Producer organizations in a context of value chains and globalization 153 Institutional innovations--still a work in progress 157 7 Innovating through science and technology 158 Genetic improvement has been enormously successful, but not everywhere 159 Management and systems technologies need to complement genetic improvement 163 Investing more in R&D 165 Institutional arrangements to increase the efficiency and effectiveness of R&D systems 169 Using available technology better: extension and ICT innovations 172 Moving forward 176 focus E: Capturing the benefits of genetically modified organisms for the poor 177 8 Making agricultural systems more environmentally sustainable 180 Drivers of resource degradation 181 Improving agricultural water management 182 Greening the green revolution 188 Managing intensive livestock systems 189 Reversing degradation in less-favored areas 190 Payment for environmental services 197 Conclusions 199 focus F: Adaptation to and mitigation of climate change in agriculture 200 9 Moving beyond the farm 202 Rural employment: a daunting challenge 202 Agricultural wage employment 205 Rising rural nonfarm employment 209 Wages and earnings in the rural labor market 212 Labor supply: migration and the urban economy 214 Schooling, training, and transition to the labor market 216 Providing safety nets to reduce vulnerability 219 A final word on rural labor markets and migration: the need for policy attention 221 focus G: Education and skills for rural development 222 focus H: The two-way links between agriculture and health 224 viii CONTENTS Part III How can agriculture-for-development agendas best be implemented? 226 10 Emerging national agendas for agriculture's three worlds 226 New opportunities and challenges 226 The proposed approach 227 Agriculture-based countries--accelerating growth, poverty reduction, and food security 229 Transforming countries--reducing rural-urban income gaps and rural poverty 234 Urbanized countries--linking smallholders to the new food markets and providing good jobs 238 Political, administrative, and financial feasibility 242 Recognizing the policy dilemmas 243 11 Strengthening governance, from local to global 245 Changing roles: the state, the private sector, and civil society 246 Agricultural policy processes 248 Governance reforms for better policy implementation 251 Decentralization and local governance 254 Community-driven development 256 Aid effectiveness for agricultural programs 257 Progress on the global agenda 258 Moving forward on better governance for agriculture 265 Bibliographical note 266 Endnotes 269 References 285 Selected indicators 321 Selected agricultural and rural indicators 322 Selected world development indicators 333 Index 355 Contents ix Boxes 1 What is the future for the global food supply? 8 5.2 Innovative uses of information technology to link farmers 1.1 Rural population dynamics 29 to markets in India and West Africa 121 1.2 Cross-country evidence on the effect of agricultural growth 5.3 Price stabilization through international trade: saving $200 on poverty reduction 30 million in Bangladesh 123 1.3 Large countries have regional heterogeneity that replicates 5.4 Zambia and Burkina Faso: contrasting experiences in the three worlds of agriculture 33 liberalizing domestic cotton markets 124 1.4 Agriculture's comparative advantage in Sub-Saharan 5.5 Linking small farmers to high-value chains: Three Africa 34 approaches 129 1.5 A role for agriculture in Africa's mineral-rich 5.6 Employment gains and reduced poverty in rural countries 35 Senegal 131 1.6 Supporting farmers without a strong fiscal base: lessons 6.1 Benefits from community-driven land certification in from Thailand 36 Ethiopia 140 2.1 The green revolution in food staples that didn't happen: 6.2 Improving the efficiency of land administration services in Sub-Saharan Africa's variegated palette 55 Georgia 141 2.2 Population density and the definitions of "rural" 58 6.3 How land rentals can increase productivity and equity in China 142 2.3 Why are there lagging regions in countries with high agricultural growth? 59 6.4 Rental markets and the impact of restrictions in India 143 2.4 The global environmental footprint of expanding livestock 61 6.5 Banrural SA: from ill-performing agrarian bank to profitable public-private financial institution 146 2.5 Substantial potential for expanding irrigation in Sub- Saharan Africa--in the right way 65 6.6 Mongolia's index-based livestock insurance 149 3.1 Establishing the relative importance of the different 6.7 Is there a rationale for fertilizer subsidies? 152 pathways 75 6.8 Thriving rural input supply retailers as agrodealers in 3.2 Constructing comparable measures of income across Africa 153 countries 76 6.9 Producer organizations with international 3.3 The challenge of drastic demographic changes from memberships 155 selective migration 81 7.1 When zero means plenty: the benefits of zero tillage in 3.4 Returning to the farm in Zambia--subsistence agriculture, South Asia's rice-wheat systems 164 AIDS, and economic crisis 86 7.2 Using legumes to improve soil fertility 165 3.5 New technologies and positive discrimination policies 7.3 Stronger IPRs in developing countries: effect on small reduce social inequalities in India 88 farmers 167 3.6 Pastoralists' precarious livelihoods 89 7.4 Sub-Saharan Africa's agricultural R&D challenge 168 3.7 Are farms becoming too small? 92 7.5 Long-term capacity development in Ghana 169 4.1 Types of instruments that distort trade 97 7.6 IPR options to give the poor access to modern 4.2 The political economy of agricultural reforms in developed science 171 countries 99 7.7 Mexican farmers lead research through PRODUCE 4.3 Nominal rates of assistance 100 foundations 172 4.4 Significant progress in reducing the antiagricultural bias in 7.8 Adding value to a poor farmer's crop: cassava in Colombia China and India 102 and Ghana 173 4.5 The political economy of agricultural reforms in developing 7.9 Private agribusiness and NGOs: leading ICT provision to countries 104 farmers in India 176 4.6 Simulating the effects of trade liberalization with global 8.1 Restoring the northern Aral Sea--by doubling the Syr models 105 Darya's flow 185 4.7 Net buyers and net sellers of food staples within a 8.2 Resource degradation in rice-wheat systems of South country 109 Asia 188 4.8 Examples of subsidies in India and Zambia 116 8.3 Integrated pest management to control the Andean potato weevil in Peru 189 5.1 Impacts of road infrastructure on markets and productivity 120 8.4 Managing poultry intensification in Thailand 190 x CONTENTS 8.5 Four trajectories: disappearing or rebounding forests, 10.4 Special features of agriculture in Europe and Central misery or growth 193 Asia 240 8.6 Agroforestry parklands in Niger turn back the desert and 11.1 Conflicts over land displace millions in Colombia 246 restore livelihoods 194 11.2 Translating vision into practice: a former minister's view of 8.7 Two tales of community-driven management, watersheds, Uganda's Plan for Modernizing Agriculture 249 and pastures 196 11.3 Empowering producer organizations and developing a 8.8 Managing drought and livestock in pastoral areas of the vision for agriculture in Senegal 250 Middle East and North Africa 197 11.4 Vietnam's progress in aligning budgets with sector 9.1 Horticulture development in Maharashtra 209 priorities 251 9.2 A women's cooperative in India 211 11.5 Regional integration: opportunities and challenges in West 9.3 Child labor: pervasive in agriculture 219 Africa 251 9.4 The gradual but incomplete move toward cash-based food 11.6 Making a green revolution through vision and aid 220 leadership 253 10.1 Four policy objectives of the agriculture-for-development 11.7 Delivering international public goods 262 agenda form a policy diamond 228 11.8 Global financing for climate change adaptation and 10.2 Comprehensive Africa Agricultural Development mitigation--the urgency of addressing the needs of Program 230 vulnerable countries and small-scale farmers 263 10.3 Middle East and North Africa--agriculture for jobs and as a safety net 235 Figures 1 The number of poor rose in South Asia and Sub-Saharan 1.1 Cereal yields are up and poverty is down in South Asia, but Africa from 1993 to 2002 ($1-a-day poverty line) 3 cereal yields and poverty were unchanged in Sub-Saharan 2 Agriculture's contribution to growth and the rural share in Africa 27 poverty distinguishes three types of countries: agriculture- 1.2 As countries develop, the shares of GDP and labor based, transforming, and urbanized 5 in agriculture tend to decline, but with many 3 GDP growth originating in agriculture benefits the poorest idiosyncrasies 28 half of the population substantially more 6 1.3 Agriculture-based, transforming, and urbanized countries 4 Public spending on agriculture is lowest in the agriculture- constitute agriculture's three worlds 31 based countries, while their share of agriculture in GDP is 1.4 Price and wage effects dominated the long-run elasticity of highest 7 rural poverty to cereal yields in India, 1958­94 33 5 Developing countries are taxing agricultural exportables 1.5 The ratio of food processing to agricultural value added less 11 rises with incomes 37 6 Domestic consumption and exports of high-value products 1.6 Labor productivity has been a more important source in developing countries are growing rapidly 13 of growth in agriculture than in nonagriculture, 7 The yield gap for cereals between Sub-Saharan Africa and 1993­2005 39 other regions has widened 15 1.7 Macroeconomic policy and agricultural growth have 8 Agriculture and deforestation are heavy contributors to improved in Sub-Saharan Africa 40 greenhouse gas emissions 17 1.8 Official development assistance to agriculture declined 9 The four policy objectives of the agriculture-for- sharply between 1975 and 2004 41 development agenda form a policy diamond 19 2.1 Cereal yields rose, except in Sub-Saharan Africa 51 10 Agricultural growth in Sub-Saharan Africa has increased as 2.2 Modern inputs have expanded rapidly but have lagged in macroeconomic conditions improved 20 Sub-Saharan Africa 52 11 The urban-rural income disparity has increased in most of 2.3 Growth in agricultural GDP per agricultural population is the transforming countries 21 lowest in Sub-Saharan Africa 53 12 Agriculture-based and transforming countries get low 2.4 Stagnation in Sub-Saharan African agriculture may be scores for governance 23 over 54 Contents xi 2.5 There are big differences across regions in agricultural 5.2 Wholesale prices in Ethiopia fluctuate within a wide potential and access to markets 57 import-export parity band 122 2.6 Per capita food consumption in developing countries is 5.3 World prices for traditional bulk exports continue to shifting to fruits and vegetables, meat, and oils 60 decline 123 2.7 High value exports are expanding rapidly in developing 5.4 Food consumption expenditures in Indonesia are shifting countries 61 from cereals to higher-value and prepared foods 125 2.8 Slower growth in cereal and meat consumption is projected 5.5 Rising per capita incomes drive supermarket for the next 30 years 62 growth 125 2.9 Developing countries will become even bigger markets for 5.6 Transport, handling, and packaging are major costs for cereals exported largely by developed countries 63 French bean exports in Bangladesh 129 2.10 Arable and permanent cropland per capita of the 6.1 Credit-constrained rural households use fewer inputs and agricultural population is falling in Sub-Saharan Africa and have lower incomes 144 South Asia 63 6.2 Transport costs make up about one-third of the farmgate 2.11 Water scarcity affects millions of people in Asia and the price of urea fertilizer in African countries, 2005 151 Middle East and North Africa 64 6.3 More than a third of Zambia's 2004/05 public budget for 2.12 Growth rates of yields for major cereals are slowing in agriculture went to fertilizer subsidies 151 developing countries 67 7.1 Improved varieties have been widely adopted, except in 2.13 Exploitable yield gaps are high for maize in Africa 67 Sub-Saharan Africa 159 3.1 Real per capita income varies widely for each livelihood 7.2 Estimated returns to agricultural R&D are high in all strategy 77 regions--averaging 43 percent 166 3.2 In most countries, the vast majority of rural households 7.3 Financing for extension services, the traditional and the participate in agriculture 78 new approach 175 3.3 Sources of income vary between poor and rich 79 8.1 Dependence on groundwater irrigation is highest in the 3.4 Women's reported participation in agricultural self- Middle East and South Asia 184 employment relative to men's varies by region 80 8.2 Groundwater aquifers in India are being depleted 186 3.5 Rural-urban gaps in educational attainment are 8.3 With PES, degraded pasture has been converted to large 84 sustainable land use in Nicaragua 198 3.6 Farm size distributions are often bimodal 87 9.1 Agriculture is not enough to absorb new rural 3.7 Yields on small farms lag behind large farms in staples in workers 203 Brazil and Chile 91 9.2 The share of wage workers in agricultural employment rises 4.1 Progress has been slow in reducing overall support to with per capita income 205 agricultural producers in the OECD, but there has been 9.3 The share of wage labor in agricultural employment is some move to less-distorting "decoupled" payments 97 rising in many countries 206 4.2 For agriculture-based countries, net agricultural taxation 9.4 Formal employment in Brazilian agriculture has become fell in 9 of 11 countries 100 more cyclical 207 4.3 Developing countries are taxing exportables less 101 9.5 Labor requirements are considerably higher for vegetables 4.4 For transforming countries, 9 of 10 either increased than for cereals 209 protection or reduced taxation 102 9.6 Retail trade and services dominate nonfarm wage 4.5 For urbanized countries, 6 of 7 either increased protection employment 210 or reduced taxation 103 9.7 Most rural nonfarm enterprises have only one or two 4.6 Estimated real international commodity price increases workers, mostly self-employed, Indonesia, 2005 210 following complete trade liberalization 106 9.8 Wages are much higher in rural nonfarm employment 4.7 The corresponding gain in the estimated trade shares of than in agricultural employment in India, Mexico, and developing countries 106 Uganda 212 4.8 Latin America and Sub-Saharan Africa are expected to 9.9 For workers with no education, wages in agricultural and have higher agricultural output growth under global trade rural nonfarm employment are not so different across reforms 107 sectors 212 4.9 Subsidies have risen while public goods investments have 9.10 Agricultural wages have been declining in most Latin declined in India 115 American countries 213 5.1 Layers of intermediaries characterize Ghana's maize 9.11 Agricultural wages have been rising in most Asian and markets 119 African countries 214 xii CONTENTS 9.12 Labor productivity in rural nonfarm self-employment is 10.2 Policy diamond for transforming countries 237 heterogeneous in Indonesia 215 10.3 Policy diamond for urbanized countries 241 9.13 Growth of manufacturing and service employment in 11.1 Agriculture-based and transforming countries get low Mexico is a function of distance to an urban center with scores for governance 245 more than 250,000 inhabitants 215 11.2 Good fits to country-specific conditions for demand- 9.14 Average years of education in rural areas, by age 216 side and supply-side approaches are needed to improve 10.1 Policy diamond for agriculture-based countries 231 agricultural sector governance 252 Tables 1 Characteristics of three country types, 2005 5 5.2 Public and private sector roles to enhance trade-related SPS 1.1 Demographic and economic characteristics of three country compliance and quality management capacity 132 types, 2005 31 7.1 Total public agricultural R&D expenditures by region, 1981 1.2 Poverty in three country types, 2002 32 and 2000 167 1.3 Public spending in agriculture-based countries is 7.2 Assets of public and private sectors in agribiotechnology low 41 research 171 3.1 Changing market participation among farming households 7.3 Ways of providing and financing agricultural advisory in Vietnam 74 services 174 3.2 Typology of rural households by livelihood strategies in 8.1 Agriculture's environmental problems onsite and three country types 76 offsite 181 3.3 Changes in farm size and land distribution 87 9.1 Rural households' diverse sources of income 204 4.1 Nominal rates of assistance by commodity in developing 9.2 Rural employment by sector of activity, selected countries 101 countries 205 4.2 Estimated cost distribution of current trade 9.3 Average years of education of rural 18­25 year olds, selected policies 105 countries 217 4.3 Illustrative poverty effects from agricultural trade reform in 11.1 Types of global organizations and networks relevant for developed and developing countries 108 agriculture 261 5.1 Public and private options for strengthening farmer links to the market 128 Maps 2.1 Agroecological zones in agricultural areas 56 8.1 Overexploitation has caused severe water stress in many 2.2 Market access in agricultural areas of Africa, Asia, and river basins 183 Latin America 56 8.2 Many deforestation hotspots are in tropical areas 192 Foreword Agriculture is a vital development tool for achieving the Millennium Development Goal that calls for halving by 2015 the share of people suffering from extreme poverty and hun- ger. That is the overall message of this year's World Development Report (WDR), the 30th in the series. Three out of every four poor people in developing countries live in rural areas, and most of them depend directly or indirectly on agriculture for their livelihoods. This Report provides guidance to governments and the international community on designing and implementing agriculture-for-development agendas that can make a difference in the lives of hundreds of millions of rural poor. The Report highlights two major regional challenges. In much of Sub-Saharan Africa, agriculture is a strong option for spurring growth, overcoming poverty, and enhancing food security. Agricultural productivity growth is vital for stimulating growth in other parts of the economy. But accelerated growth requires a sharp productivity increase in smallholder farming combined with more effective support to the millions coping as subsistence farm- ers, many of them in remote areas. Recent improved performance holds promise, and this Report identifies many emerging successes that can be scaled up. In Asia, overcoming widespread poverty requires confronting widening rural-urban income disparities. Asia's fast-growing economies remain home to over 600 million rural people living in extreme poverty, and despite massive rural-urban migration, rural poverty will remain dominant for several more decades. For this reason, the WDR focuses on ways to generate rural jobs by diversifying into labor-intensive, high-value agriculture linked to a dynamic rural, nonfarm sector. In all regions, with rising land and water scarcity and the added pressures of a globalizing world, the future of agriculture is intrinsically tied to better stewardship of natural resources. With the right incentives and investments, agriculture's environmental footprint can be lightened, and environmental services harnessed to protect watersheds and biodiversity. Today, rapidly expanding domestic and global markets; institutional innovations in markets, finance, and collective action; and revolutions in biotechnology and information technology all offer exciting opportunities to use agriculture to promote development. But seizing these opportunities will require the political will to move forward with reforms that improve the governance of agriculture. Ultimately, success will also depend on concerted action by the international develop- ment community to confront the challenges ahead. We must level the playing field in inter- national trade; provide global public goods, such as technologies for tropical food staples; help developing countries address climate change; and overcome looming health pandem- ics for plants, animals, and humans. At stake are the livelihoods of 900 million rural poor, who also deserve to share the benefits of a sustainable and inclusive globalization. Robert B. Zoellick President World Bank Group xiii Acknowledgments This Report has been prepared by a core team led by Derek Byerlee and Alain de Janvry and comprising Elisabeth Sadoulet, Robert Townsend, and Irina Klytchnikova. The team was assisted by Harold Alderman, Beatriz Avalos-Sartorio, Julio Berdegué, Regina Birner, Lynn Brown, Michael Carter, Luc Christiaensen, Marie-Helene Collion, Klaus Deininger, Peter Hazell, Karen Macours, Michael Morris, Paula Savanti, and Dina Umali-Deininger, all of whom drafted parts of the Report. The team was assisted as well by Noora Aberman, Jorge Aguero, Shahrooz Badkoubei, Sarah Baird, Leandre Bassole, Benjamin Davis, Nango Dem- bele, Ashok Gulati, Corinna Hawkes, Tidiane Kinda, Melissa Klink, Alex McCalla, Claudio Montenegro, Stefano Pagiola, Eija Pehu, Catherine Ragasa, Antti Seelaff, and John Staatz. The work was conducted under the general guidance of François Bourguignon in collab- oration with the Sustainable Development Network. Bruce Ross-Larson was the principal editor. Extensive and excellent advice was received from Kym Anderson, Hans Binswanger, Karen McConnell Brooks, Mark Cackler, Manuel Chiriboga, Kevin Cleaver, Christopher Delgado, Shantayanan Devarajan, Josue Dione, Gershon Feder, Alan Harold Gelb, Ravi Kanbur, Jeffrey Lewis, Were Omamo, Keijiro Otsuka, Rajul Pandya-Lorch, Prabhu Pingali, Pierre Rondot, Kostas Stamoulis, Erik Thorbecke, C. Peter Timmer, Joachim von Braun, staff of the Agriculture and Rural Development Department and of the Sustainable Devel- opment Network of the World Bank, staff of RIMISP (Latin American Center for Rural Development), and many others to whom the team is grateful without implication (see page 266). Numerous others inside and outside the World Bank provided helpful com- ments and inputs. The Development Data Group contributed to the data appendix and was responsible for the Selected World Development Indicators. The team also acknowledges the generous support of the multidonor programmatic trust fund (Knowledge for Change Program), the Canadian International Development Agency, Ford Foundation, France's Ministry of Foreign Affairs, Global Donor Platform for Rural Development, International Development Research Centre, International Fund for Agricultural Development, InWEnt (Capacity Building International), Japan's Ministry of Finance, Science Council of the Consultative Group on International Agricultural Research, Swedish International Development Cooperation Agency, Swiss Agency for Development and Cooperation, UK Department for International Development, United States Agency for International Development, and The William and Flora Hewlett Foundation. The team benefited greatly from a wide range of consultations. Meetings and regional workshops were held locally and in Australia, Canada, France, Germany, India, Italy, Japan, Kenya, Mali, Norway, Sweden, and the United Kingdom; and discussions of the draft Report were conducted online. The team wishes to thank the participants in these workshops, videoconferences, and discussions, which included academics, researchers, government officials, and staff of nongovernmental, civil society, and private sector organizations. Rebecca Sugui served as senior executive assistant to the team, Ofelia Valladolid as pro- gram assistant, and Jason Victor and Maria Hazel Macadangdang as team assistants. Evan- geline Santo Domingo served as resource management assistant. xv Abbreviations and Data Notes Abbreviations AATF African Agricultural Technology Foundation IMF International Monetary Fund ADB Asian Development Bank IPCC Intergovernmental Panel on Climate Change AfDB African Development Bank IPR Intellectual property rights CAADP Comprehensive Africa Agriculture IRI International Research Institute for Climate Development Programme and Society CDD Community-driven development IRRI International Rice Research Institute CGIAR Consultative Group on International IWMI International Water Management Institute Agricultural Research MDG Millennium Development Goal CIAT International Center for Tropical Agriculture MFI Microfinance institution CIMMYT International Maize and Wheat Improvement NEPAD New Partnership for Africa's Development Center NERICA New rice for Africa CIRAD Agricultural Research for Developing Countries NGO Nongovernmental organization DAC Development Assistance Committee NRA Nominal rate of assistance EU European Union ODA Official development assistance FAO Food and Agriculture Organization OECD Organization for Economic Cooperation GAEZ Global agroecological zones and Development GDP Gross Domestic Product PPP Public-private partnerships GHG Greenhouse gas PES Payment for environmental services GMO Genetically modified organism PSE Producer support estimate GPS Global Positioning System R&D Research and development HIV/AIDS Human immunodeficiency virus/ SAFEX South Africa Futures Exchange acquired immune deficiency syndrome SPS Sanitary and phytosanitary standards ICARDA International Center for Agricultural SWAps Sector-wide approaches Research in the Dry Areas TFP Total factor productivity ICRAF International Center for Research UN United Nations in Agroforestry UNDP United Nations Development Programme ICTs Information and communication technologies UNCTAD United Nations Conference on Trade IDA International Development Association and Development IDB Inter-American Development Bank UNFCCC United Nations Framework Convention IEA International Energy Agency on Climate Change IFAD International Fund for Agricultural USAID United States Agency for International Development Development IFAP International Federation of Agricultural USDA United States Department of Agriculture Producers WDR World Development Report IFDC International Center for Soil Fertility and WHO World Health Organization Agricultural Development WTO World Trade Organization IFPRI International Food Policy Research Institute WWF World Wide Fund for Nature ILO International Labor Organization xvii xviii ABBREVIATIONS AND DATA NOTES Country Code Country Name Country Code Country Name AGO Angola MLI Mali ARG Argentina MOZ Mozambique AZE Azerbaijan MWI Malawi BDI Burundi MYS Malaysia BEN Benin NER Niger BFA Burkina Faso NGA Nigeria BGD Bangladesh NPL Nepal BGR Bulgaria PAK Pakistan BLR Belarus PER Peru BOL Bolivia PHL Philippines BRA Brazil PNG Papua New Guinea CHL Chile POL Poland CHN China PRY Paraguay CIV Côte d'Ivoire ROM Romania CMR Cameroon RUS Russian Federation COL Colombia RWA Rwanda CZE Czech Republic SDN Sudan DOM Dominican Republic SEN Senegal DZA Algeria SLV El Salvador ECU Ecuador SVK Slovak Republic EGY Egypt, Arab Rep. of SYR Syrian Arab Rep. ETH Ethiopia TCD Chad GHA Ghana TGO Togo GIN Guinea THA Thailand GTM Guatemala TJK Tajikistan HND Honduras TUN Tunisia HUN Hungary TUR Turkey IDN Indonesia TZA Tanzania IND India UGA Uganda IRN Iran, Islamic Rep. of UKR Ukraine KEN Kenya VEN Venezuela, R. B. de KHM Cambodia VNM Vietnam LAO Lao PDR YEM Yemen, Republic LKA Sri Lanka ZAF South Africa MAR Morocco ZAR Congo, Dem. Rep. of MDG Madagascar ZMB Zambia MEX Mexico ZWE Zimbabwe Data notes The countries included in regional and income groupings in low- and middle-income economies and thus may include this Report are listed in the Classification of Economies table economies in transition from central planning, as a matter at the end of the Selected World Development Indicators. of convenience. The terms advanced countries or developed Income classifications are based on gross national income countries may be used as a matter of convenience to denote (GNP) per capita; thresholds for income classifications in this high-income economies. edition may be found in the Introduction to Selected World Dollar figures are current U.S. dollars, unless otherwise spec- Development Indicators. Group averages reported in the fig- ified. Billion means 1,000 million; trillion means 1,000 billion. ures and tables are unweighted averages of the countries in the Serbia and Montenegro is used in this Report either because group, unless noted to the contrary. the event being discussed occurred prior to the independence The use of the word countries to refer to economies implies of the Republic of Montenegro in June 2006 or because sepa- no judgment by the World Bank about the legal or other rate data for the Republic of Serbia and the Republic of Mon- status of a territory. The term developing countries includes tenegro are not available. Overview An African woman bent under the sun, continuing to reduce poverty and hunger weeding sorghum in an arid field with a for several decades thereafter. Agricul- hoe, a child strapped on her back--a vivid ture alone will not be enough to massively image of rural poverty. For her large fam- reduce poverty, but it has proven to be ily and millions like her, the meager bounty uniquely powerful for that task. With the of subsistence farming is the only chance last World Development Report on agri- to survive. But others, women and men, culture completed 25 years ago, it is time have pursued different options to escape to place agriculture afresh at the center of poverty. Some smallholders join producer the development agenda, taking account of organizations and contract with export- the vastly different context of opportunities ers and supermarkets to sell the vegetables and challenges that has emerged.2 they produce under irrigation. Some work Agriculture operates in three distinct as laborers for larger farmers who meet the worlds--one agriculture-based, one trans- scale economies required to supply mod- forming, one urbanized. And in each the ern food markets. Still others, move into agriculture-for-development agenda differs the rural nonfarm economy, starting small in pursuing sustainable growth and reduc- enterprises selling processed foods. ing poverty. While the worlds of agriculture are vast, In the agriculture-based countries, varied, and rapidly changing, with the right which include most of Sub-Saharan Africa, policies and supportive investments at local, agriculture and its associated industries are national, and global levels, today's agricul- essential to growth and to reducing mass ture offers new opportunities to hundreds poverty and food insecurity. Using agricul- of millions of rural poor to move out of ture as the basis for economic growth in the poverty. Pathways out of poverty open to agriculture-based countries requires a pro- them by agriculture include smallholder ductivity revolution in smallholder farming. farming and animal husbandry, employ- Given Sub-Saharan Africa's unique agricul- ment in the "new agriculture" of high-value ture and institutions, that revolution will products, and entrepreneurship and jobs in have to be different from the Asian green the emerging rural, nonfarm economy. revolution. How to implement it after many In the 21st century, agriculture continues years of limited success remains a difficult to be a fundamental instrument for sustain- challenge. But conditions have changed, able development and poverty reduction. and there are many local successes and new Three of every four poor people in develop- opportunities on which to build. ing countries live in rural areas--2.1 billion In transforming countries, which living on less than $2 a day and 880 million include most of South and East Asia and on less than $1 a day--and most depend the Middle East and North Africa, rapidly on agriculture for their livelihoods.1 Given rising rural-urban income disparities and where they are and what they do best, pro- continuing extreme rural poverty are major moting agriculture is imperative for meet- sources of social and political tensions. The ing the Millennium Development Goal of problem cannot be sustainably addressed halving poverty and hunger by 2015 and through agricultural protection that raises 1 2 WORLD DEVELOPMENT REPORT 2008 the price of food (because a large number of on a complex agenda of interrelated agree- poor people are net food buyers) or through ments and international public goods. Civil subsidies. Addressing income disparities in society empowerment, particularly of pro- transforming countries requires a compre- ducer organizations, is essential to improv- hensive approach that pursues multiple ing governance at all levels. pathways out of poverty--shifting to high- This Report addresses three main value agriculture, decentralizing nonfarm questions: economic activity to rural areas, and pro- · What can agriculture do for develop- viding assistance to help move people out of ment? Agriculture has served as a basis agriculture. Doing this calls for innovative for growth and reduced poverty in many policy initiatives and strong political com- countries, but more countries could ben- mitment. But it can benefit 600 million of efit if governments and donors were to the world's rural poor. reverse years of policy neglect and rem- In urbanized countries, which include edy their underinvestment and misin- most of Latin America and much of Europe vestment in agriculture. and Central Asia, agriculture can help reduce the remaining rural poverty if small- · What are effective instruments in using holders become direct suppliers in modern agriculture for development? Top pri- food markets, good jobs are created in agri- orities are to increase the assets of poor culture and agroindustry, and markets for households, make smallholders--and environmental services are introduced. agriculture in general--more produc- With rising resource scarcity and mount- tive, and create opportunities in the ing externalities, agricultural development rural nonfarm economy that the rural and environmental protection have become poor can seize. closely intertwined. Agriculture's large envi- · How can agriculture-for-development ronmental footprint can be reduced, farm- agendas best be implemented? By design- ing systems made less vulnerable to climate ing policies and decision processes most change, and agriculture harnessed to deliver suited to each country's economic and more environmental services. The solution social conditions, by mobilizing politi- is not to slow agricultural development--it cal support, and by improving the gov- is to seek more sustainable production sys- ernance of agriculture. tems. The first step in this is to get the incen- tives right by strengthening property rights What can agriculture do and removing subsidies that encourage for development? the degradation of natural resources. Also Agriculture has features that imperative is adapting to climate change, make it a unique instrument which will hit poor farmers the hardest-- for development and hit them unfairly because they have Agriculture can work in concert with other contributed little to its causes. sectors to produce faster growth, reduce Agriculture thus offers great promise for poverty, and sustain the environment. In growth, poverty reduction, and environ- this Report, agriculture consists of crops, mental services, but realizing this promise livestock, agroforestry, and aquaculture. It also requires the visible hand of the state-- does not include forestry and commercial providing core public goods, improving capture fisheries because they require vastly the investment climate, regulating natural different analyses. But interactions between resource management, and securing desir- agriculture and forestry are considered in able social outcomes. To pursue agriculture- the discussions of deforestation, climate for-development agendas, local, national, change, and environmental services. and global governance for agriculture need to be improved. The state will need greater Agriculture contributes to development in capacity to coordinate across sectors and many ways. Agriculture contributes to to form partnerships with private and civil development as an economic activity, as a society actors. Global actors need to deliver livelihood, and as a provider of environ- Overview 3 mental services, making the sector a unique · As a livelihood. Agriculture is a source instrument for development. of livelihoods for an estimated 86 per- · As an economic activity. Agriculture cent of rural people. It provides jobs for can be a source of growth for the national 1.3 billion smallholders and landless economy, a provider of investment oppor- workers, "farm-financed social welfare" tunities for the private sector, and a prime when there are urban shocks, and a foun- driver of agriculture-related industries dation for viable rural communities. Of and the rural nonfarm economy. Two- the developing world's 5.5 billion people, thirds of the world's agricultural value 3 billion live in rural areas, nearly half of added is created in developing countries. humanity. Of these rural inhabitants an In agriculture-based countries, it gener- estimated 2.5 billion are in households ates on average 29 percent of the gross involved in agriculture, and 1.5 billion domestic product (GDP) and employs are in smallholder households.3 65 percent of the labor force. The indus- The recent decline in the $1-a-day tries and services linked to agriculture in poverty rate in developing countries-- value chains often account for more than from 28 percent in 1993 to 22 percent 30 percent of GDP in transforming and in 2002--has been mainly the result of urbanized countries. falling rural poverty (from 37 percent to Agricultural production is important 29 percent) while the urban poverty rate for food security because it is a source remained nearly constant (at 13 percent). of income for the majority of the rural More than 80 percent of the decline in poor. It is particularly critical in a dozen rural poverty is attributable to better countries of Sub-Saharan Africa, with a conditions in rural areas rather than to combined population of about 200 mil- out-migration of the poor. So, contrary lion and with highly variable domestic to common perceptions, migration to production, limited tradability of food cities has not been the main instrument staples, and foreign exchange constraints for rural (and world) poverty reduction. in meeting their food needs through But the large decline in the number of imports. These countries are exposed rural poor (from 1,036 million in 1993 to to recurrent food emergencies and the 883 million in 2003) has been confined uncertainties of food aid, and for them, to East Asia and the Pacific (figure 1). In increasing and stabilizing domestic pro- South Asia and Sub-Saharan Africa, the duction is essential for food security. number of rural poor has continued to Figure 1 The number of poor rose in South Asia and Sub-Saharan Africa from 1993 to 2002 ($1-a-day poverty line) Rural poverty Urban poverty Millions of people Millions of people below $1-a-day poverty line below $1-a-day poverty line 500 500 400 1993 2002 400 1993 2002 300 300 200 200 100 100 0 0 Sub- South East Asia Middle East Europe & Latin Sub- South East Asia Middle East Europe & Latin Saharan Asia & Pacific & North Central America & Saharan Asia & Pacific & North Central America & Africa Africa Asia Caribbean Africa Africa Asia Caribbean Source: Ravallion, Chen, and Sangraula 2007. 4 WORLD DEVELOPMENT REPORT 2008 rise and will likely exceed the number of This group of countries has 417 million urban poor until 2040. In these regions, a rural inhabitants, mainly in Sub-Saharan high priority is to mobilize agriculture for countries. Eighty-two percent of the rural poverty reduction. Sub-Saharan population lives in agricul- · As a provider of environmental services. ture-based countries. In using (and frequently misusing) nat- · Transforming countries--Agriculture ural resources, agriculture can create is no longer a major source of economic good and bad environmental outcomes. growth, contributing on average only It is by far the largest user of water, con- 7 percent to GDP growth, but poverty tributing to water scarcity. It is a major remains overwhelmingly rural (82 per- player in underground water depletion, cent of all poor). This group, typified agrochemical pollution, soil exhaustion, by China, India, Indonesia, Morocco, and global climate change, accounting and Romania, has more than 2.2 billion for up to 30 percent of greenhouse gas rural inhabitants. Ninety-eight percent emissions. But it is also a major pro- of the rural population in South Asia, 96 vider of environmental services, gener- percent in East Asia and the Pacific, and ally unrecognized and unremunerated, 92 percent in the Middle East and North sequestering carbon, managing water- Africa are in transforming countries. sheds, and preserving biodiversity. With · Urbanized countries--Agriculture con- rising resource scarcity, climate change, tributes directly even less to economic and concern about environmental costs, growth, 5 percent on average, and pov- business as usual in the way agriculture erty is mostly urban. Even so, rural areas uses natural resources is not an option. still have 45 percent of the poor, and agri- Making the farming systems of the rural business and the food industry and ser- poor less vulnerable to climate change is vices account for as much as one third of imperative. Managing the connections GDP. Included in this group of 255 mil- among agriculture, natural resource lion rural inhabitants are most countries conservation, and the environment must in Latin America and the Caribbean and be an integral part of using agriculture many in Europe and Central Asia. Eighty- for development. eight percent of the rural populations in both regions are in urbanized countries. Agriculture's contributions differ in the Countries follow evolutionary paths three rural worlds. The way agricul- that can move them from one country type ture works for development varies across to another. China and India moved from countries depending on how they rely on the agriculture-based to the transforming agriculture as a source of growth and an group over the past 20 years, while Indo- instrument for poverty reduction. The nesia gravitated toward the urbanized (fig- contribution of agriculture to growth and ure 2). In addition, countries have sharp poverty reduction can be seen by categoriz- subnational geographical disparities--for ing countries according to the share of agri- example, many transforming and urban- culture in aggregate growth over the past 15 ized countries have agriculture-based years, and the current share of total poverty regions (such as Bihar in India and Chiapas in rural areas, using the $2-a-day poverty in Mexico). line (figure 2). This perspective produces Classifying regions within countries three types of countries--three distinct according to their agricultural potential rural worlds (table 1): and access to markets shows that 61 per- · Agriculture-based countries--Agricul- cent of the rural population in developing ture is a major source of growth, account- countries lives in favored areas--irrigated, ing for 32 percent of GDP growth on humid, and semihumid areas with little average--mainly because agriculture moisture stress, and with medium to good is a large share of GDP--and most of market access (less than five hours from a the poor are in rural areas (70 percent). market town of 5,000 or more). But two- Overview 5 Figure 2 Agriculture's contribution to growth and the rural share in poverty distinguish three types of countries: agriculture based, transforming, and urbanized Agriculture's contribution to growth, 1990­2005, % 80 Actual poverty data Agriculture-based BDI Predicted poverty data countries Poverty data over time RWA 60 CMR MWI SDN PRY BEN NER INDIA NGA PNG 40 CMR (1965­94) BGR TGO LAO GHA AZE CIV SYR MDG KEN BFA ETH NPL ZMB MOZ GIN MLI TCD UGA 20 HND PAK IRN YEM KHM SEN DZA GTM CHINA BRA PHL BOL AGO IDN BGD PER EGY VNM (1981­2001) ARG UKR RUS DOM IND TUN VEN SVK TUR MAR THA LKA CHN BRAZIL CHL CZE SLV ECU INDONESIA 0 (1970­96) MEX ZAR POL ROM ZAF MYS (1970­96) TJK HUN COL Transforming countries BLR Urbanized countries ZWE ­20 0.0 0.2 0.4 0.6 0.8 1.0 Rural poor/total poor, 2002 Source: WDR 2008 team. Note: Arrows show paths for Brazil, China, India, and Indonesia. The list of 3-letter codes and the countries they represent can be found on page xviii. Table 1 Characteristics of three country types, 2005 Agriculture- Transforming Urbanized based countries countries countries Rural population (millions), 2005 417 2,220 255 Share of population rural (%), 2005 68 63 26 GDP per capita (2000 US$), 2005 379 1,068 3,489 Share of agriculture in GDP (%), 2005 29 13 6 Annual agricultural GDP growth, 1993­2005 (%) 4.0 2.9 2.2 Annual nonagricultural GDP growth, 1993­2005 (%) 3.5 7.0 2.7 Number of rural poor (millions), 2002 170 583 32 Rural poverty rate, 2002 (%) 51 28 13 Source: Ravallion, Chen, and Sangraula 2007; World Bank 2006y. Note: Poverty line is $1.08 a day, in 1993 purchasing power parity dollars. thirds of the rural population in Sub-Saha- Heterogeneitydefinestheruralworld. Eco- ran Africa lives in less-favored areas defined nomic and social heterogeneity is a defining as arid and semiarid or with poor market characteristic of rural areas. Large commer- access. In five countries with detailed pov- cial farmers coexist with smallholders. This erty maps, the poverty rate is higher in less- diversity permeates the smallholder popu- favored areas, but most of the poor live in lation as well. Commercial smallholders favored areas. So using agriculture to reduce deliver surpluses to food markets and share poverty requires not only investing in less- in the benefits of expanding markets for the favored areas to combat extreme poverty, new agriculture of high-value activities. But but also targeting the large number of poor many others are in subsistence farming, in favored areas. mainly due to low asset endowments and 6 WORLD DEVELOPMENT REPORT 2008 unfavorable contexts. Consuming most poverty as GDP growth originating outside of the food they produce, they participate agriculture (figure 3). For China, aggregate in markets as buyers of food and as sellers growth originating in agriculture is esti- of labor. Membership in these categories mated to have been 3.5 times more effective is affected not only by asset positions, but in reducing poverty than growth outside also by gender, ethnicity, and social status, agriculture--and for Latin America 2.7 as they imply differing abilities to use the times more. Rapid agricultural growth-- same assets and resources in responding to in India following technological innova- opportunities. tions (the diffusion of high yielding vari- Heterogeneity is found in the rural labor eties) and in China following institutional market where there are many low-skill, innovations (the household responsibility poorly remunerated agricultural jobs and a system and market liberalization)--was small number of high-skill jobs that offer accompanied by major declines in rural workers pathways out of poverty. It is found poverty. More recently, in Ghana, rural in the rural nonfarm economy where low- households accounted for a large share of a productivity self- and wage-employment steep decline in poverty induced in part by coexists with employment in dynamic agricultural growth. enterprises. And it is found in the outcomes of migration, which lifts some of the rural Agriculture can be the lead sector for poor out of poverty but takes others to overall growth in the agriculture-based urban slums and continued poverty. countries. Agriculture has a well-estab- This pervasive heterogeneity in agricul- lished record as an instrument for poverty ture and rural society has deep implications reduction. But can it also be the leading for public policy in using agriculture for sector of a growth strategy for the agricul- development. A particular policy reform is ture-based countries? Besides the sheer size likely to have gainers and losers. Trade lib- of the sector, two arguments, applied to the eralization that raises the price of food hurts agriculture-based countries of Sub-Saharan net buyers (the largest group of rural poor Africa, support the view that it can. in countries like Bolivia and Bangladesh) The first is that in many of these coun- and benefits net sellers (the largest group tries, food remains imperfectly tradable of rural poor in Cambodia and Vietnam). because of high transaction costs and the Policies have to be differentiated according prevalence of staple foods that are only to the status and context of households, tak- lightly traded, such as roots and tubers and ing particular account of prevailing gender local cereals. So, many of these countries norms. Differentiated policies are designed not necessarily to favor one group over the other but to serve all households more cost- Figure 3 GDP growth originating in agriculture benefits the poorest half of the population effectively, tailoring policies to their condi- substantially more tions and needs, particularly to the poorest. Balancing attention to the favored and less- Expenditure gains induced by 1% GDP growth, % 8 favored subsectors, regions, and households Agriculture is one of the toughest policy dilemmas fac- 6 Nonagriculture ing poor countries with severe resource constraints. 4 2 Agriculture has a strong record in development 0 Agriculture has special powers in reducing ­2 poverty. Agricultural growth has special Lowest 2 3 4 5 6 7 8 9 Highest powers in reducing poverty across all coun- Expenditure deciles try types. Cross-country estimates show Source: Ligon and Sadoulet 2007. that GDP growth originating in agricul- Note: Based on data from 42 countries during the period 1981­2003. Gains are significantly different for the lower half of ture is at least twice as effective in reducing expenditure deciles. Overview 7 must largely feed themselves. Agricultural Yet agriculture has been vastly underused productivity determines the price of food, for development. Parallel to these suc- which in turn determines wage costs and cesses are numerous failures to use agricul- competitiveness of the tradable sectors. ture for development. Many agriculture- Productivity of food staples is thus key to based countries still display anemic per growth. capita agricultural growth and little struc- The second is that comparative advan- tural transformation (a declining share of tage in the tradable subsectors will still lie agriculture in GDP and a rising share of in primary activities (agriculture and min- industry and services as GDP per capita ing) and agroprocessing for many years, rises). The same applies to vast areas within because of resource endowments and the countries of all types. Rapid population difficult investment climate for manufac- growth, declining farm size, falling soil fer- tures. Most economies depend on a diverse tility, and missed opportunities for income portfolio of unprocessed and processed pri- diversification and migration create distress mary-based exports (including tourism) to as the powers of agriculture for development generate foreign exchange. Growth in both remain fallow. Policies that excessively tax the nontradable and tradable sectors of agriculture and underinvest in agriculture agriculture also induces strong growth in are to blame, reflecting a political economy other sectors of the economy through mul- in which urban interests have the upper tiplier effects. hand. Compared with successful transform- That is why, for many years to come, the ing countries when they still had a high growth strategy for most agriculture-based share of agriculture in GDP, the agriculture- economies has to be anchored on getting based countries have very low public spend- agriculture moving. Success stories of agri- ing in agriculture as a share of their agricul- culture as the basis for growth at the begin- tural GDP (4 percent in the agriculture-based ning of the development process abound. countries in 2004 compared with 10 percent Agricultural growth was the precursor to in 1980 in the transforming countries, fig- the industrial revolutions that spread across ure 4). The pressures of recurrent food cri- the temperate world from England in the ses also tilt public budgets and donor priori- mid-18th century to Japan in the late-19th ties toward direct provision of food rather century. More recently, rapid agricultural than investments in growth and achieving growth in China, India, and Vietnam was food security through rising incomes. the precursor to the rise of industry. Just as Where women are the majority of small- for poverty, the special powers of agricul- holder farmers, failure to release their full ture as the basis for early growth are well potential in agriculture is a contributing established. factor to low growth and food insecurity. Figure 4 Public spending on agriculture is lowest in the agriculture-based countries, while their share of agriculture in GDP is highest Agricultural GDP/GDP Public spending on agriculture/agricultural GDP Percent Percent 35 35 30 29 29 1980 2000 30 1980 2000 25 24 25 20 20 16 17 15 14 15 12 10 11 10 10 10 5 5 4 4 0 0 Agriculture-based Transforming Urbanized Agriculture-based Transforming Urbanized Source : Fan, forthcoming. 8 WORLD DEVELOPMENT REPORT 2008 Underuse of agriculture for development agriculture of staple crops and traditional is not confined to the agriculture-based export commodities also finds new markets countries. In transforming countries with as it becomes more differentiated to meet rapid growth in nonagricultural sectors, changing consumer demands and new uses the reallocation of labor out of agriculture is (for example, biofuels) and benefits from typically lagging, leaving large numbers of regional market integration. However, agri- poor people in rural areas and widening the culture faces large uncertainties that are rural-urban income gap. The farm popula- difficult to predict and call for caution in tion demands subsidies and protection. But managing the global food supply (box 1). weak fiscal capacity to sustain transfers large An emerging vision of agriculture for enough to reduce the income gap and con- development redefines the roles of produc- tinuing urban demands for low food prices ers, the private sector, and the state. Produc- create a policy dilemma.4 The opportunity tion is mainly by smallholders, who often cost of subsidies (which are three times pub- remain the most efficient producers, in par- lic investments in agriculture in India) is ticular when supported by their organiza- reduced public goods for growth and social tions. But when these organizations cannot services in rural areas. Raising incomes in capture economies of scale in production agriculture and the rural nonfarm economy and marketing, labor-intensive commercial must be part of the solution. farming can be a better form of production, and efficient and fair labor markets are the New opportunities are emerging. The key instrument to reducing rural poverty. world of agriculture has changed dramati- The private sector drives the organization cally since the 1982 World Development of value chains that bring the market to Report on agriculture. Dynamic new mar- smallholders and commercial farms. The kets, far-reaching technological and insti- state--through enhanced capacity and new tutional innovations, and new roles for the forms of governance--corrects market fail- state, the private sector, and civil society all ures, regulates competition, and engages characterize the new context for agricul- strategically in public-private partnerships ture. The emerging new agriculture is led to promote competitiveness in the agribusi- by private entrepreneurs in extensive value ness sector and support the greater inclu- chains linking producers to consumers and sion of smallholders and rural workers. In including many entrepreneurial smallhold- this emerging vision, agriculture assumes a ers supported by their organizations. The prominent role in the development agenda. What are effective instruments in using agriculture B O X 1 What is the future for the global food supply? for development? Agriculture has been largely successful in percent from 2000 to 2030. Added to this Agriculture can be the main source of meeting the world's effective demand for is the burgeoning demand for agricultural food. Yet more than 800 million people feedstocks for biofuels, which have already growth for the agriculture-based coun- remain food insecure, and agriculture has pushed up world food prices. tries and can reduce poverty and improve left a huge environmental footprint. And Managing the aggregate response of the environment in all three country types, the future is increasingly uncertain. agriculture to rising demand will require albeit in different ways. This requires Models predict that food prices in good policy and sustained investments, global markets may reverse their long- not business as usual. Sharply increased improving the asset position of the rural term downward trend, creating rising investment is especially urgent in Sub- poor, making smallholder farming more uncertainties about global food security. Saharan Africa, where food imports are competitive and sustainable, diversifying Climate change, environmental degrada- predicted to more than double by 2030 income sources toward the labor market and tion, rising competition for land and water, under a business-as-usual scenario, the higher energy prices, and doubts about impact of climate change is expected to the rural nonfarm economy, and facilitating future adoption rates for new technologies be large with little capacity to cope, and successful migration out of agriculture. all present huge challenges and risks that progress continues to be slow in raising make predictions difficult. per capita food availability. Increase access to assets To meet projected demand, cereal Household assets are major determinants production will have to increase by nearly 50 percent and meat production by 85 Source: Rosegrant and others 2007. of the ability to participate in agricultural markets, secure livelihoods in subsistence Overview 9 farming, compete as entrepreneurs in the women's rights. Redistributing underuti- rural nonfarm economy, and find employ- lized large estates to settle smallholders can ment in skilled occupations. Three core work if complemented by reforms to secure assets are land, water, and human capital. the competitiveness of beneficiaries-- Yet the assets of the rural poor are often something that has been difficult to achieve. squeezed by population growth, environ- Targeted subsidies to facilitate market- mental degradation, expropriation by dom- based land reform are used in Brazil and inant interests, and social biases in policies South Africa, and lessons must be derived and in the allocation of public goods. from these pioneering experiences for Nowhere is the lack of assets greater than potential wider application. in Sub-Saharan Africa, where farm sizes in many of the more densely populated areas Water. Access to water and irrigation is are unsustainably small and falling, land is a major determinant of land productivity severely degraded, investment in irrigation and the stability of yields. Irrigated land is negligible, and poor health and educa- productivity is more than double that tion limit productivity and access to better of rainfed land. In Sub-Saharan Africa, options. Population pressure together with only 4 percent of the area in production is declining farm size and water scarcity are under irrigation, compared with 39 per- also major challenges in many parts of Asia. cent in South Asia and 29 percent in East Enhancing assets requires significant public Asia. With climate change leading to rising investments in irrigation, health, and edu- uncertainties in rainfed agriculture and cation. In others cases, it is more a matter of reduced glacial runoff, investment in water institutional development, such as enhanc- storage will be increasingly critical. Even ing the security of property rights and the with growing water scarcity and rising costs quality of land administration. Increasing of large-scale irrigation schemes, there are assets may also call for affirmative action many opportunities to enhance produc- to equalize chances for disadvantaged or tivity by revamping existing schemes and excluded groups, such as women and eth- expanding small-scale schemes and water nic minorities. harvesting. Land. Land markets, particularly rental Education. While land and water are crit- markets, can raise productivity, help house- ical assets in rural areas, education is often holds diversify their incomes, and facilitate the most valuable asset for rural people to exit from agriculture. As farmers age, as pursue opportunities in the new agriculture, rural economies diversify, and as migration obtain skilled jobs, start businesses in the accelerates, well-functioning land markets rural nonfarm economy, and migrate suc- are needed to transfer land to the most pro- cessfully. Yet education levels in rural areas ductive users and to facilitate participation tend to be dismally low worldwide: an aver- in the rural nonfarm sector and migration age of four years for rural adult males and out of agriculture. But in many countries, less than three years for rural adult females insecure property rights, poor contract in Sub-Saharan Africa, South Asia, and the enforcement, and stringent legal restric- Middle East and North Africa. Improving tions limit the performance of land mar- basic rural education has been slower than kets, creating large inefficiencies in both in urban areas. Where demand for educa- land and labor reallocation and reinforc- tion is lagging among rural households, it ing existing inequalities in access to land. can be enhanced through cash transfers (as Safety nets and access to credit are needed in Bangladesh, Brazil, and Mexico) con- to minimize distress land sales when farm- ditional on school attendance. However, ers are exposed to shocks. increasingly it is the quality of rural educa- Land reform can promote smallholder tion that requires the most improvement, entry into the market, reduce inequalities with education conceived broadly to include in land distribution, increase efficiency, vocational training that can provide tech- and be organized in ways that recognize nical and business skills that are useful in 10 WORLD DEVELOPMENT REPORT 2008 the new agriculture and the rural nonfarm · Make agriculture more sustainable and economy. a provider of environmental services (chapter 8) Health. Widespread illness and death from HIV/AIDS and malaria can greatly Improve price incentives and increase reduce agricultural productivity and dev- the quality and quantity of pub- astate livelihoods. The majority of people lic investment. Recent reforms have affected by HIV work in farming, and there improved price incentives for agricultural is tremendous scope for agricultural policy producers in developing countries, reduc- to be more HIV-responsive in supporting ing but not eliminating historical policy adjustments to labor shocks and the trans- biases against agriculture. Between 1980­ mission of knowledge to orphans. In rural 84 and 2000­04 net agricultural taxation Zambia, population declines have been declined on average from 28 percent to 10 especially severe for young rural adults: 19 percent in agriculture-based countries, percent of people 15­24 years old in 1990, from 15 percent to 4 percent in transform- the most productive age, are estimated to ing countries, and from marginally nega- have died by 2000. But agriculture also tive protection to net protection of 9 per- poses threats to the health of the rural cent in urbanized countries. However, a low poor. Irrigation can increase the incidence level of net taxation hides a combination of of malaria, and pesticide poisoning is esti- protection of importables and taxation of mated to cause 355,000 deaths annually. exportables (especially in the agriculture- Zoonotic diseases such as avian influenza based and transforming countries), which that arise from the proximity of humans can both be high (figure 5). Hence, consid- and animals pose growing threats to human erable room remains for further efficiency health. Better coordination of the agricul- gains through reforms in developing coun- ture and health agendas can yield big divi- tries' own trade policies. Liberalization of dends for productivity and welfare. imports of food staples can also be pro- poor because often the largest number of Make smallholder farming more poor, including smallholders, are net food productive and sustainable buyers. But many poor net sellers (some- Improving the productivity, profitability, times the largest group of poor) will lose, and sustainability of smallholder farming and programs tailored to country-specific is the main pathway out of poverty in using circumstances will be needed to ease the agriculture for development. What will transition to new market realities. this take? A broad array of policy instru- In sharp contrast, there has been relatively ments, many of which apply differently to little progress in the overall decline in pro- commercial smallholders and to those in ducer support in member countries of the subsistence farming, can be used to achieve Organisation for Economic Co-operation the following: and Development (OECD). Producer sup- port declined from 37 percent of gross value · Improve price incentives and increase the of farm receipts in 1986­88 to 30 percent in quality and quantity of public investment 2003­05. There has been a shift away from (chapter 4) support directly linked to product prices to · Make product markets work better other less-distorting forms such as cash (chapters 5 and 6) transfers "decoupled" from production, par- · Improve access to financial services ticularly in the European Union (EU). But and reduce exposure to uninsured risks such transfers are not always neutral for pro- (chapter 6) duction because they reduce aversion to risk · Enhance the performance of producer (wealth effect), reduce the variability in farm organizations (chapter 6) income (insurance effect), and allow banks · Promote innovation through science to make loans to farmers that they otherwise and technology (chapter 7) would not. Overview 11 Figure 5 Developing countries are taxing agricultural exportables less Exportables Importables Nominal rate of assistance, % Nominal rate of assistance, % 30 30 1980­84 2000­04 1980­84 2000­04 26 23 20 20 14 13 10 11 10 10 2 0 0 ­10 ­10 ­13 ­14 ­20 ­19 ­20 ­30 ­29 ­30 ­40 ­40 ­46 ­50 ­50 Agriculture-based Transforming Urbanized Agriculture-based Transforming Urbanized Source : Anderson, forthcoming. Note : The nominal rate of assistance is a measure of domestic output prices relative to border prices, which also takes into account domestic input subsidies. The estimated welfare impacts of full The political economy will determine trade liberalization are relatively large. By the pace and extent of further trade, price, removing their current level of protection, and public spending reforms. Membership industrial countries would induce annual in the World Trade Organization (WTO) welfare gains for developing countries esti- can help induce reform, and local media mated to be five times the current annual can expose taxpayer costs and unequal flow of aid to agriculture. But this impact incidence of gains. In some cases, bar- is heterogeneous across products and coun- gained compromises and compensation tries. With full trade liberalization, inter- schemes for the losers can be effective--as national agricultural commodity prices in Japan's rice policy reforms, the EU's are estimated to increase on average by 5.5 sugar reforms, and Mexico's 1990s reforms percent, while those of cotton are expected for food staples. Linking domestic agricul- to increase by 21 percent and oilseeds by 15 tural reforms to a broader set of economy- percent. This raises particular concerns for wide reforms can increase the likelihood of food-importing countries with tight for- success, as in many developing countries eign exchange constraints such as Burundi, in the 1980s and 1990s, but these reforms Rwanda, and Niger. Poor countries that tend to remain incomplete for agriculture. export cotton or oilseeds, such as Chad, Other subsidy reforms, such as free electri- Sudan, Burkina Faso, Mali, and Benin, stand cal power to Indian farmers, remain dead- to gain. Among the big expected gainers are locked in clientelistic bargains at high effi- Brazil, Thailand, and Vietnam. ciency and environmental costs. The Doha Round of trade negotiations The response to better price incentives must urgently be concluded, particularly depends on public investments in market to eliminate distortions, such as U.S. cot- infrastructure, institutions, and support ton subsidies, which are detrimental to services. But the quality of public spend- the poorest countries. Complementary ing is often low and needs improvement. policies and programs (including aid-for- In some countries, nonstrategic subsidies trade) are needed to compensate losers amount to as much as half of the public (transfer programs) and to facilitate rapid budget for agriculture. To mobilize politi- and equitable adjustments by smallhold- cal support for better use of public expendi- ers to emerging comparative advantages tures in agriculture, an initial step is greater (investments in public goods and institu- public disclosure and transparency of bud- tional reforms). get allocation, and analysis of impacts. 12 WORLD DEVELOPMENT REPORT 2008 Make product and input markets work this has been done, production and quality better. With major structural changes in have improved--as for cotton in Zambia, agricultural markets and the entry of pow- where production tripled. Critically impor- erful new actors, a key issue for development tant, too, is to increase the productivity of is enhancing the participation of small- exports, as exemplified by the recent suc- holders and ensuring the poverty-reducing cessful Ghana experience with cocoa. Qual- impacts of agricultural growth. Options ity improvements and fair trade can open differ across the spectrum of markets. new opportunities for more remunerative markets for some smallholders. Food staples markets. Reducing transac- tion costs and risks in food staples markets High-value markets. The participation of can promote faster growth and benefit the smallholders can also be enhanced in high- poor. Beyond investments in infrastructure, value markets, both global and domes- promising innovations include commodity tic, including the supermarket revolution exchanges, market information systems unfolding in many countries. High-value based on rural radio and short messaging markets for domestic consumption are systems, warehouse receipts, and market- the fastest-growing agricultural markets based risk management tools. in most developing countries, expanding A particularly thorny issue in food mar- up to 6­7 percent a year, led by livestock kets is how to manage price volatility for products and horticulture (figure 6). Fresh politically sensitive food staples in countries and processed fruits and vegetables, fish where they account for a large share of con- and fish products, meat, nuts, spices, and sumer spending. If the food staple is trad- floriculture now account for 43 percent of able, insurance through exchange-traded agrofood exports from developing coun- futures contracts can sometimes manage tries, worth about $138 billion in 2004. As price risks, as for countries or traders in incomes rise, supermarkets become more southern Africa that use the South African dominant in the domestic retail sales of commodity exchange. Risk management agricultural products--reaching 60 percent can also be enhanced by more open bor- in some Latin American countries. ders and private trade, as in the successful The poverty impacts of this growth management of flood-induced rice short- depend on how the rural population partic- ages in Bangladesh in 1998. But most food ipates in high-value markets, either directly staples in agriculture-based countries are as producers (as in Bangladesh) or through only partially tradable, and many countries the labor market (as in Chile). Enhancing subject to frequent climatic shocks man- smallholder participation needs market age public grain reserves to reduce price infrastructure, upgrading farmers' techni- instability--with mixed success. High risks cal capacity, risk management instruments, of price volatility remain for both farmers and collective action through producer and consumers in many agriculture-based organizations. Addressing the stringent san- countries and effective safety nets will con- itary and phytosanitary standards in global tinue to be important until incomes rise or markets is an even bigger challenge. Doing market performance improves. it well depends on joint public and private efforts in policy (food safety legislation), Traditional bulk exports. The long down- research (risk assessment, good practices), ward trend in world market prices of such infrastructure (export processing facilities), traditional exports as coffee and cotton and oversight (disease surveillance). threatens the livelihoods of millions of pro- ducers. Reduced taxation and greater liber- Input markets. Especially for seed and fer- alization of export markets has improved tilizer, market failures continue to be perva- incomes in many settings. But these liber- sive in Sub-Saharan Africa because of high alized markets require a new role for gov- transaction costs, risks, and economies of ernment, particularly in regulating fair and scale. As a result, low fertilizer use is one of efficient operations in marketing. Where the major constraints on increasing agricul- Overview 13 Figure 6 Domestic consumption and exports of high-value products in developing countries are growing rapidly Domestic consumption Exports Index, 1980 = 100 Index, 1980 = 100 350 350 300 300 250 250 200 200 150 150 100 100 50 50 0 0 1980 1985 1990 1995 2000 2005 1980 1985 1990 1995 2000 2005 Meat Horticulture Cereals Meat Horticulture Oilseeds Traditional exports Source: http://faostat.fao.org, accessed June 2007, and http://comtrade.un.org. tural productivity in Sub-Saharan Africa. to agriculture through public programs or The renewed interest in fertilizer subsidies state banks has left huge gaps in financial needs to focus on sustainable solutions to services, still largely unfilled despite numer- market failures. "Market-smart" approaches ous institutional innovations. to jump-starting agricultural input markets include targeted vouchers to enable farmers Rural finance. The microfinance revolu- to purchase inputs and stimulate demand tion, providing access to credit without for- in private markets, and matching grants to mal collateral, has opened access to loans for underwrite selected start-up costs of entry millions of poor people, especially women, of private distributors to input markets. but it has not reached most agricultural Like any subsidies, input subsidies must activities, except in high-turnover activities be used with caution because they have such as small livestock and horticulture. high opportunity costs for productive pub- However, the range of financial products lic goods and social expenditures and they available to the rural poor has broadened risk political capture and irreversibility. But to include savings, money transfers, insur- through the judicious use of subsidies, it is ance services, and leasing options. With possible to underwrite risks of early adop- the rise of integrated supply chains and tion of new technologies and achieve econ- contract farming, financial intermediation omies of scale in markets to reduce input through interlinked agents is becoming prices. Subsidies need to be part of a com- more common. Information technologies prehensive strategy to improve productivity are reducing transaction costs and making and must have credible exit options. loans less costly in rural areas, for example, using agricultural credit cards to purchase Improve access to financial services and inputs or cellular phones to complete bank- reduce exposure to uninsured risks. Finan- ing transactions. Credit reporting bureaus cial constraints in agriculture remain per- covering microfinance institutions and the vasive, and they are costly and inequitably lower tier of commercial banks also help distributed, severely limiting smallholders' smallholders capitalize on the reputations ability to compete. Financial constraints they establish as microfinance borrowers to originate in the lack of asset ownership to access larger and more commercial loans. serve as collateral (wealth rationing) and in Many of these innovations are still at the the reticence to put assets at risk as collat- pilot stage, requiring evaluation and scaling eral when they are vital to livelihoods (risk up to make a real difference for smallholder rationing). The demise of special credit lines competitiveness. 14 WORLD DEVELOPMENT REPORT 2008 Managing risk. Exposure to uninsured landless and women, and they produce 22 risks--the result of natural disasters, health percent of India's total milk supply. shocks, demographic changes, price vola- In spite of many successes, producer tility, and policy changes--has high effi- organizations' effectiveness is frequently ciency and welfare costs for rural house- constrained by legal restrictions, low man- holds. To manage exposure to these risks, agerial capacity, elite capture, exclusion of farmers have to forgo activities with higher the poor, and failure to be recognized as full expected incomes. Selling assets to sur- partners by the state. Donors and govern- vive shocks can have high long-term costs ments can assist by facilitating the right to because decapitalization (distress sales of organize, training leaders, and empowering land and livestock) creates irreversibilities weaker members, in particular women and or slow recovery in the ownership of agri- young farmers. However, providing this cultural assets. In addition, child educa- assistance without creating dependency tion and health can suffer long-term con- remains a challenge. sequences when children are taken out of school in response to shocks or are exposed Promote innovation through science and to early periods of malnutrition, leading to technology. Driven by rapidly growing intergenerational transfers of poverty. private investment in research and devel- In spite of multiple initiatives, little prog- opment (R&D), the knowledge divide ress has been made in reducing uninsured between industrial and developing coun- risks in smallholder agriculture. State- tries is widening. Including both public managed insurance schemes have proven and private sources, developing countries largely ineffective. Index-based insurance invest only a ninth of what industrial coun- for drought risk, now being scaled up by tries put into agriculture R&D as a share of private initiatives in India and elsewhere, agricultural GDP. can reduce risks to borrowers and lenders To narrow this divide, sharply increased and unlock agricultural finance. However, investments in R&D must be at the top of these initiatives are unlikely to reach a crit- the policy agenda. Many international and ical mass unless there is some element of national investments in R&D have paid off subsidy, at the very least to cover start-up handsomely, with an average internal rate costs. of return of 43 percent in 700 R&D proj- ects evaluated in developing countries in Enhance the performance of producer all regions. But global and national failures organizations. Collective action by pro- of markets and governance lead to serious ducer organizations can reduce transac- underinvestment in R&D and in innova- tion costs in markets, achieve some mar- tion systems more generally, particularly ket power, and increase representation in in the agriculture-based countries. While national and international policy forums. investment in agricultural R&D tripled in For smallholders, producer organizations China and India over the past 20 years, it are essential to achieve competitiveness. increased by barely a fifth in Sub-Saharan They have expanded remarkably rapidly Africa (declining in about half of the coun- in number and membership, often in an tries there).5 African countries are addi- attempt to fill the void left by the state's tionally disadvantaged by the fact that the withdrawal from marketing, input pro- specificity of their agroecological features vision, and credit, and to take advantage leaves them less able than other regions of democratic openings allowing greater to benefit from international technology civil society participation in governance. transfers and the small size of many of Between 1982 and 2002 the percentage of these countries prevents them from captur- villages with producer organizations rose ing economies of scale in agricultural R&D. from 8 to 65 percent in Senegal and from 21 Low investments in R&D and low interna- to 91 percent in Burkina Faso. The Indian tional transfers of technology have gone Dairy Cooperatives Network has 12.3 mil- hand in hand with stagnant cereal yields in lion individual members, many of them Sub-Saharan Africa, resulting in a widening Overview 15 yield gap with the rest of the world (figure and biological control more than pesti- 7). For these countries, sharply increased cides. Because most of these technologies investment and regional cooperation in are location-specific, their development R&D are urgent. and adoption require more decentralized Low spending is only part of the prob- and participatory approaches, combined lem. Many public research organizations with collective action by farmers and face serious leadership, management, and communities. financial constraints that require urgent Revolutionary advances in biotech- attention. But higher-value markets open nology offer potentially large benefits to new opportunities for the private sector poor producers and poor consumers. But to foster innovation along the value chain. today's investments in biotechnology, con- Grasping them often requires partner- centrated in the private sector and driven ships among the public sector, private sec- by commercial interests, have had limited tor, farmers, and civil society in financing, impacts on smallholder productivity in the developing, and adapting innovation. With developing world--the exception is Bt cot- a wider range of institutional options now ton in China and India. Low public invest- available, more evaluation is needed of what ment in biotechnology and slow progress works well in what contexts. in regulating possible environmental and A further challenge is to narrow the food safety risks have restrained the devel- income and productivity gaps between opment of genetically modified organisms favored and less-favored regions. Better (GMOs) that could help the poor. The technologies for soil, water, and livestock potential benefits of these technologies will management and more sustainable and be missed unless the international develop- resilient agricultural systems, including ment community sharply increases its sup- varieties more tolerant of pests, diseases, port to interested countries. and drought, are needed for the latter regions. Approaches that exploit biologi- Make agriculture more sustainable--and a cal and ecological processes can minimize provider of environmental services. The the use of external inputs, especially agri- environmental footprint of agriculture has cultural chemicals. Examples include con- been large, but there are many opportuni- servation tillage, improved fallows, green ties for reducing it. Since the 1992 Earth manure cover crops, soil conservation, Summit in Rio, it is generally accepted that and pest control that relies on biodiversity the environmental agenda is inseparable from the broader agenda of agriculture for development. And the future of agriculture Figure 7 The yield gap for cereals between Sub- Saharan Africa and other regions has widened is intrinsically tied to better stewardship of the natural resource base on which it Yield, tons per hectare 6 depends. Both intensive and extensive agriculture 5 face environmental problems--but of dif- ferent kinds. Agricultural intensification 4 has generated environmental problems 3 from reduced biodiversity, mismanaged irrigation water, agrochemical pollution, 2 and health costs and deaths from pesticide poisoning. The livestock revolution has its 1 own costs, especially in densely populated 0 and periurban areas, through animal waste 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 and the spread of animal diseases such as Developed countries South Asia avian influenza. Many less-favored areas East Asia & Pacific Sub-Saharan Africa suffer from deforestation, soil erosion, Latin America & Caribbean desertification, and degradation of pas- Source: http://faostat.fao.org, accessed June 2007. tures and watersheds. In the East African 16 WORLD DEVELOPMENT REPORT 2008 highlands, soil erosion can result in pro- ening property rights (as with agroforestry ductivity losses as high as 2­3 percent a parklands in Niger) and providing long- year, in addition to creating offsite effects term incentives for natural resource man- such as the siltation of reservoirs. agement with off-farm benefits (such as The answer is not to slow agricultural matching grants for soil conservation) are development, but to seek more sustain- necessary in both intensive and extensive able production systems and to enhance farming areas. Inappropriate incentives agriculture's provision of environmental that encourage mining resources--such services. Many promising technological as subsidies to water intensive crops that and institutional innovations can make cause groundwater overpumping--must agriculture more sustainable with mini- be reduced. mum tradeoffs on growth and poverty Reforms are often politically difficult. reduction. Water management strategies Better water measurement through tech- in irrigated areas must improve water nology (remote sensing), better quality of productivity, meeting demands of all irrigation services, and greater accountabil- users (including the environment), and ity to water users can generate political sup- reduce water pollution and the unsus- port for otherwise stalled reforms. tainable mining of groundwater. These Payments for environmental services strategies depend on removing incentives can help overcome market failures in man- for wasteful water usage, devolving water aging environmental externalities. Water- management to local user groups, invest- shed and forest protection create envi- ing in better technologies, and regulating ronmental services (clean drinking water, externalities more effectively. Decentral- stable water flows to irrigation systems, ized governance in irrigation manage- carbon sequestration, and protection of ment has a higher chance of success if legal biodiversity) for which providers should be frameworks clearly define the roles and compensated through payments from ben- rights of user groups and if the capacity of eficiaries of these services. Interest has been groups to manage irrigation collectively is growing, particularly in Latin America. In increased. Nicaragua, payments induced a reduction Better technologies and better ways of in the area of degraded pasture and annual managing modern farm inputs can also crops by more than 50 percent in favor of make rainfed farming more sustainable. silvopastoralism, half of it by poor farmers. One of agriculture's major success stories Environmental certification of products in the past two decades is conservation (or also allows consumers to pay for sustainable zero) tillage. This approach has worked in environmental management, as practiced commercial agriculture in Latin America, under fair trade or shade-grown coffee. among smallholders in South Asia's rice- wheat systems, and in Ghana. In less-favored The urgency of dealing with climate change. regions, community-based approaches to Poor people who depend on agriculture are natural resource management, such as the most vulnerable to climate change. Increas- watershed management program in Eastern ing crop failures and livestock deaths are Anatolia of Turkey, offer significant prom- already imposing high economic losses ise. As survey data from 20 countries show, and undermining food security in parts of women's active engagement in community Sub-Saharan Africa, and they will get far organizations improves the effectiveness more severe as global warming continues. of natural resources management and the More frequent droughts and increasing ability to resolve conflicts. water scarcity may devastate large parts of Getting incentives right is the first step the tropics and undermine irrigation and toward sustainable resource management. drinking water in entire communities of Widespread adoption of more sustainable already poor and vulnerable people. The approaches is often hindered by inappro- international community must urgently priate pricing and subsidy policies and the scale up its support to climate-proof the failure to manage externalities. Strength- farming systems of the poor, particularly Overview 17 in sub-Saharan Africa, the Himalayan Figure 8 Agriculture and deforestation are heavy contributors to greenhouse gas regions, and the Andes. Based on the pol- emissions luter-pays principle, it is the responsibility % of total GHG emissions of the richer countries to compensate the 63% poor for costs of adaptation. So far, global 60 Developed countries Developing countries commitments to existing adaptation funds have been grossly inadequate. 40 Developing-country agriculture and 15% 20 deforestation are also major sources of 11% 7% 4% greenhouse gas emissions: they contribute 0 an estimated 22 percent and up to 30 per- Energy Agriculture Deforestation Industrial Waste cent of total emissions, more than half of (excluding land processes use change) which is from deforestation largely caused by agricultural encroachment (13 million Source: WDR 2008 team, based on data from the United Nations Framework Convention on Climate Change, www.unfccc.int. hectares of annual deforestation globally) (figure 8).6 Carbon-trading schemes-- especially if their coverage is extended to provide financing for avoided deforestation efficient producers with current technolo- and soil carbon sequestration (for example, gies. Policy decisions on biofuels need to conservation tillage)--offer significant devise regulations or certification systems untapped potential to reduce emissions to mitigate the potentially large environ- from land-use change in agriculture. Some mental footprint of biofuels production. improvements in land and livestock man- Increased public and private investment agement practices (for example, conserva- in research is important to develop more tion tillage and agroforestry) are often win- efficient and sustainable production pro- win situations: after the initial investments, cesses based on feedstocks other than food they can result in more productive and sus- staples. tainable farming systems. Moving beyond farming: Biofuels--an opportunity and a challenge. a dynamic rural economy and Promising new opportunities for mitigating skills to participate in it climate change and creating large new mar- Creating rural employment. With rapid kets for agriculture have emerged through rural population growth and slow expan- the production of biofuels, stimulated by sion in agricultural employment, creating high energy prices. But few of the current jobs in rural areas is a huge and insuffi- biofuels programs are economically viable, ciently recognized challenge. Between 45 and many pose social (rising food prices) and 60 percent of the rural labor force is and environmental (deforestation) risks. engaged in the agricultural labor market To date, production in industrial countries and the rural nonfarm economy in Asia has developed behind high protective tar- and Latin America. Only in Sub-Saharan iffs on biofuels and with large subsidies. Africa is self-employment in agriculture These policies hurt developing countries still by far the dominant activity for the that are, or could become, efficient produc- rural labor force, especially for women. ers in profitable new export markets. Poor But with rapidly growing rural populations consumers also pay higher prices for food and declining farm sizes, the rural employ- staples as grain prices rise in world markets ment problem will need to be addressed directly due to the diversion of grain to bio- there as well. fuels or indirectly due to land conversion The rural labor market offers employ- away from food production. ment possibilities for the rural popula- Brazil is the world's largest and most tion in the new agriculture and the rural efficient producer of biofuels, based on its nonfarm sector. But opportunities are bet- low-cost production of sugarcane. But few ter for those with skills, and women with other developing countries are likely to be lower education levels are at a disadvantage. 18 WORLD DEVELOPMENT REPORT 2008 Migration can be a climb up the income fers, also have an insurance function in ladder for well-prepared, skilled workers, or protecting the most vulnerable against it can be a simple displacement of poverty shocks. These programs have to be orga- to the urban environment for others. nized so that they do not undermine the The policy priority is to create more local labor market and food economy and jobs in both agriculture and the rural non- do not create work disincentives for ben- farm economy. The basic ingredients of a eficiaries, but do reach those most in need dynamic rural nonfarm economy are a rap- "just in time." With the shift in emphasis idly growing agriculture and a good invest- of governments and donor programs over ment climate. Linking the local economy the past two decades toward transfers as to broader markets by reducing transaction an instrument for poverty reduction and costs, investing in infrastructure, and pro- the greater attention to impact evaluation, viding business services and market intel- much has been learned about how to bet- ligence are critical. Agro-based clusters-- ter target and calibrate these programs for firms in a geographic area coordinating to greater effectiveness. compete in servicing dynamic markets-- have been effective, with well-documented How can agriculture-for- experiences for nontraditional exports in development agendas best the San Francisco Valley of Brazil and for be implemented? dairy production in Peru and Ecuador. The real challenge is to assist the tran- Pursuing an agriculture-for-development sition of the rural population into higher- agenda for a country implies defining what paying jobs. Labor regulations are needed to do and how to do it. What to do requires that help incorporate a larger share of a policy framework anchored on the behav- rural workers into the formal market and ior of agents--producers and their organi- eliminate discrimination between men zations, the private sector in value chains, and women. Education, skills, and entre- and the state. How to do it requires effective preneurship can be fostered--by providing governance to muster political support and incentives for parents to better educate their implementation capacity, again based on children, improving the quality of schools, the behavior of agents--the state, civil soci- and providing educational opportunities ety, the private sector, donors, and global relevant to emerging job markets. institutions. Providing safety nets. Providing social Defining an agriculture-for- assistance to the chronic and transitory development agenda poor can increase both efficiency and wel- Opening and widening pathways out of fare. Efficiency gains come from reducing poverty. Rural households pursue port- the cost of risk management and the risk folios of farm and nonfarm activities that of asset decapitalization in response to allow them to capitalize on the different shocks. Welfare gains come from support- skills of individual members and to diversify ing the chronic poor with food aid or cash risks. Pathways out of poverty can be through transfers. In Brazil, South Africa, and most smallholder farming, wage employment in countries in Europe and Central Asia, rural agriculture, wage or self-employment in the noncontributory pension funds protect the rural nonfarm economy, and migration out aged, facilitate earlier land transfers to the of rural areas--or some combination thereof. younger generation, and relieve those who Gender differences in access to assets and work from the financial burden of support- mobility constraints are important determi- ing the elderly. These policies have been nants of available pathways. shown to have important spillover effects Making agriculture more effective in on the health and education of the pension- supporting sustainable growth and reduc- ers' grandchildren. ing poverty starts with a favorable socio- Safety nets, such as guaranteed work- political climate, adequate governance, fare programs and food aid or cash trans- and sound macroeconomic fundamentals. Overview 19 It then requires defining an agenda for each Figure 9 The four policy objectives of the agriculture-for-development agenda form a country type, based on a combination of policy diamond four policy objectives--forming a policy Preconditions diamond (figure 9): Macroeconomic fundamentals Governance · Objective 1. Improve access to markets Sociopolitical context and establish efficient value chains 1 · Objective 2. Enhance smallholder com- Improve petitiveness and facilitate market entry Demand for market access; Demand for agricultural establish efficient agricultural and · Objective 3. Improve livelihoods in sub- products value chains nonfarm products sistence farming and low-skill rural occupations 4 · Objective 4. Increase employment in agri- 2 Pathways out Increase employment in Enhance smallholder culture and the rural nonfarm economy, of poverty agriculture and the rural competitiveness; Farming, labor, nonfarm economy; and enhance skills facilitate market entry Income migration Income effects effects enhance skills In using agriculture for development, a country should formulate an agenda with the following characteristics: 3 Transition Transition Improve livelihoods to market to market · Established preconditions. Without social in subsistence agriculture and peace, adequate governance, and sound low-skill rural macro fundamentals, few parts of an occupations agricultural agenda can be effectively Source: WDR 2008 team. implemented. This basic premise was all too often missing in agriculture-based countries until the mid-1990s, particu- countries account for over 80 percent of the larly in Sub-Saharan Africa. rural population in the agriculture-based · Comprehensive. The agenda combines countries. For them, with both limited trad- the four objectives of the policy diamond, ability of food and comparative advantage depending on country context, and spec- in primary subsectors, agricultural produc- ifies indicators that help in monitoring tivity gains must be the basis for national and evaluating progress toward each pol- economic growth and the instrument for icy objective. mass poverty reduction and food security. · Differentiated. Agendas differ by country This poses a huge challenge to governments type, reflecting differences in priorities and the international community, but there and structural conditions across the three is little alternative to success in this under- agricultural worlds. The agendas must be taking, and there are new opportunities further customized to country specifics that provide a basis for optimism. through national agricultural strategies As macroeconomic conditions and com- with wide stakeholder participation. modity prices improved in Sub-Saharan · Sustainable. The agendas must be envi- Africa starting in the mid-1990s (figure ronmentally sustainable both to reduce 10), agricultural growth accelerated from the environmental footprint of agricul- 2.3 percent per year in the 1980s to 3.8 per- ture as well as to sustain future agricul- cent between 2001 and 2005. Rural poverty tural growth. started to decline where growth occurred-- · but rapid population growth is absorbing Feasible. To be implemented and have much of the gain, reducing per capita agri- significant impact, policies and pro- cultural growth to 1.5 percent. Faster growth grams must meet the conditions of polit- and poverty reduction are now achievable, ical feasibility, administrative capacity, but they will require commitments, skills, and financial affordability. and resources. Agriculture-based countries: achieving Diverse local conditions in Sub-Saharan growth and food security. Sub-Saharan Africa produce a wide range of farming 20 WORLD DEVELOPMENT REPORT 2008 Figure 10 Agricultural growth in Sub-Saharan Africa has increased as macroeconomic conditions improved Macroeconomic score Agriculture growth rate, % per year 10 6 5 8 4 6 3 2 4 1 2 1984­95 0 1995­2005 0 ­1 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 ­0.2 0 0.2 0.4 0.6 Average annual change in macroeconomic score Source: WDR 2008 team, based on data from International Country Risk Guide, http://www.icrgonline.com. Note: Macroeconomic score is the average of the budget balance score, inflation score, and exchange rate stability. Each point represents a country. systems and reliance on many types of food water management are needed to enhance staples, implying a path to productivity the resilience of farming systems, especially growth that differs considerably from that for people in subsistence farming in remote in Asia.7 Although diversity complicates and risky environments. And it requires the development of new technologies, it capitalizing on agricultural growth to acti- offers a broad range of opportunities for vate the rural nonfarm economy in produc- innovation. Dependence on the timing and ing nontradable goods and services. The amount of rainfall increases vulnerability agenda must recognize the often-dominant to weather shocks and limits the ability to role of women as farmers, agroprocessors, use known yield-enhancing technologies. and traders in local markets. But the untapped potential for storing The Sub-Saharan context implies four water and using it more efficiently is enor- distinct features of an agriculture-for- mous. Small and landlocked countries act- development agenda. First, a multisectoral ing alone cannot achieve economies of approach must capture the synergies between scale in product markets and in research technologies (seeds, fertilizer, livestock and training, which makes regional inte- breeds), sustainable water and soil manage- gration important. Low population density ment, institutional services (extension, that increases the cost of providing infra- insurance, financial services), and human structure services and loss of human capital development (education, health)-- resources because of HIV/AIDS impose all linked with market development. Second, additional constraints. agricultural development actions must be The agenda for Sub-Saharan Africa is to decentralized to tailor them to local condi- enhance growth by improving smallholder tions. These include community-driven competitiveness in medium and higher approaches with women, who account for potential areas, where returns on invest- the majority of farmers in the region, playing ment are highest, while simultaneously a leading role. Third, the agendas must be ensuring livelihoods and food security of coordinated across countries to provide an subsistence farmers. Getting agriculture expanded market and achieve economies of moving requires improving access to mar- scale in such services as R&D. Fourth, the kets and developing modern market chains. agendas must give priority to conservation It requires a smallholder-based productivity of natural resources and adaptation to cli- revolution centered on food staples but also mate change to sustain growth. including traditional and nontraditional This agenda will require macroeco- exports. Long-term investments in soil and nomic stability, policies to improve pro- Overview 21 ducer incentives and trade, and sharply Continuing demographic pressures imply increased public investment--especially in rapidly declining farm sizes, becoming infrastructure, roads, and communications so minute that they can compromise sur- to improve market access, and in R&D to vival if off-farm income opportunities are address Africa's distinct crops and agro- not available. Competition over access to ecologies, as proposed by the New Partner- water is acute, with rising urban demands ship for Africa's Development. and deteriorating quality from runoffs. As The recent surge in growth of Sub- nonfarm incomes rise, pressures to address Saharan agriculture has been induced by rural-urban income disparities through improved price incentives from macro and subsidies would compete for fiscal expendi- sectoral reforms and higher commodity tures, at a high opportunity cost for public prices. As the easy gains from price reforms goods and rural basic needs. On the other have been captured in many countries, hand, addressing those disparities through future growth will have to rely more on import protection would elevate food costs increased productivity. The increased will- for the large masses of poor consumers who ingness of governments, the private sector, are net food buyers. and donors to invest in Sub-Saharan agri- Because of demographic pressures and culture opens a window of opportunity that land constraints, the agenda for trans- should not be missed. forming countries must jointly mobilize all pathways out of poverty: farming, Transforming countries: reducing rural- employment in agriculture and the rural urban income disparities and rural poverty. nonfarm economy, and migration. Pros- In transforming countries, with 600 million pects are good for promoting rural incomes rural poor and 2.2 billion rural inhabitants, and avoiding the subsidy-protection trap, nonagricultural sectors have been the fast- if the political will can be mustered. Rap- est growing in the world. The main focus idly expanding markets for high-value of agriculture for development is to narrow products--especially horticulture, poul- rural-urban income disparities and reduce try, fish, and dairy--offer an opportunity rural poverty while avoiding the subsidy and to diversify farming systems and develop a protection traps, challenges poorly addressed competitive and labor-intensive small- thus far (figure 11). With growing political holder sector. Export markets for nontradi- attention to widening income disparities, tional products are also accessible because there are strong pressures to better use the transforming countries have a comparative powers of agriculture for development.8 advantage in labor- and management- In these countries, agriculture is almost intensive activities. Many countries have exclusively in the hands of smallholders. high levels of poverty in less-favored regions Figure 11 The urban-rural income disparity has increased in most of the transforming countries Ratio of urban to rural median income 4 Initial year End year 3 2 1 0 Guatemala Thailand Cambodia China Vietnam Indonesia Bangladesh India Pakistan 1989­2002 1990­2002 1997­2004 1985­2001 1992­2001 1993­2002 1991­2000 1989­99 1999­2001 Source: WDR 2008 team, based on nationally representative household surveys. 22 WORLD DEVELOPMENT REPORT 2008 that require better infrastructure and tech- of poverty with good agricultural poten- nologies adapted to these regions. tial. For regions without such potential, the To confront rural unemployment, a transition out of agriculture and the provi- complementary policy objective is promot- sion of environmental services offer better ing a dynamic rural nonfarm sector in sec- prospects. But support to the agricultural ondary towns, linked to both agriculture component of the livelihoods of subsis- and the urban economy. China has brought tence farmers will remain an imperative industry to rural towns, diversifying rural for many years. incomes, an approach that could be emu- lated in other transforming countries. In Implementing an agriculture-for- all transforming countries, the transfer of development agenda labor to the dynamic sectors of the econ- The agriculture-for-development agenda omy must be accelerated by massive invest- presents two challenges for implementa- ments in skills for this generation and the tion. One is managing the political econ- next. The momentous changes this restruc- omy of agricultural policies to overcome turing implies must be insured by effective policy biases, underinvestment, and mis- safety-net programs to allow households investment. The other is strengthening to assume risks in moving to their best governance for the implementation of agri- options. Successfully meeting the dispar- cultural policies, particularly in the agri- ity problem in transforming countries can culture-based and transforming countries make a huge dent in world poverty. for which governance gets low scores (fig- ure 12). Urbanized countries: linking smallhold- Insufficient attention to these political ers to modern food markets and providing economy and governance challenges was a good jobs. The broad goal is to capitalize major reason several key recommendations on rapid expansion of modern domestic of the 1982 World Development Report on food markets and booming agricultural agriculture were not fully implemented, subsectors to sharply reduce the remaining particularly those for trade liberalization, rural poverty, still stubbornly high. The increased investments in infrastructure urbanized countries, with 32 million rural and R&D in Africa, and better delivery poor--representing 39 percent of all their of health and education services to rural poor--are experiencing the supermarket populations. revolution in food retailing. For smallhold- ers, being competitive in supplying super- The future offers more promise for agricul- markets is a major challenge that requires ture for development. The prospects are meeting strict standards and achieving scale brighter today than they were in 1982. The in delivery, for which effective producer anti-agriculture bias in macroeconomic organizations are essential.9 Exceptionally policies has been reduced thanks to broader high land inequality in Latin America also economic reforms. Agriculture is likely to constrains smallholder participation. benefit from other general governance Increasing the access of smallholders reforms that are now high on the agenda, to assets, particularly land, and increas- such as decentralization and public sector ing their voice in unequal societies can management reforms. But reforms specific enhance the size and competitiveness of to using agriculture for development are yet the smallholder sector. Beyond farming, to be widely implemented. territorial approaches are being pursued There is also evidence that the politi- to promote local employment through cal economy has been changing in favor of interlinked farming and rural agroindus- agriculture and rural development. Both try, and these experiences need to be bet- rural civil society organizations and the pri- ter understood for wider application. Agri- vate sector in agriculture value chains are cultural growth is especially important to stronger than they were in 1982. Democra- improve well-being in geographic pockets tization and the rise of participatory policy Overview 23 Figure 12 Agriculture-based and transforming countries get low scores for governance Governance score 2.0 1.5 1.0 0.5 0 ­0.5 ­1.0 Voice and Political Government Regulatory Rule of law Control of accountability stability effectiveness quality corruption Agriculture-based countries Urbanized countries Transforming countries Developed countries Source: Kaufmann, Kraay, and Mastruzzi 2006. making have increased the possibilities for reforms to redefine their roles and develop smallholder farmers and the rural poor to new capacities. New models are starting to raise their political voice. The private agri- emerge. Uganda pioneered contracting out business sector has become more vibrant, agricultural advisory services, giving pro- especially in the transforming and urban- ducer organizations a say in awarding the ized countries. New, powerful actors have contracts. entered agricultural value chains, and they have an economic interest in a dynamic and Strengthening civil society and democracy. prosperous agricultural sector and a voice in The "third sector"--communities, pro- political affairs. Yet these improved condi- ducer and other stakeholder organiza- tions alone do not guarantee the more suc- tions, and nongovernmental organizations cessful use of agriculture for development-- (NGOs)--can improve representation of smallholders must have their voices heard the rural poor and, in so doing, governance. in political affairs, and policy makers and Producer organizations can give political donors must seize the new opportunities. voice to smallholders and hold policy mak- ers and implementing agencies account- New roles for the state. Market failures able by participating in agricultural policy are pervasive, especially in the agricul- making, monitoring the budget, and engag- ture-based countries, and there is a need ing in policy implementation. In Senegal, for public policy to secure desirable social the Conseil National de Concertation et de outcomes. The state has a role in mar- Coopération des Ruraux, an umbrella orga- ket development--providing core public nization of producer organizations, is active goods, improving the investment climate in the development and implementation of for the private sector--and in better natu- national agricultural strategies and poli- ral resources management by introducing cies. Freedom of association, a free press, incentives and assigning property rights. and investment in the social capital of rural Strengthening the capacity of the state in organizations, including women's organiza- its new roles of coordinating across sectors tions, are important for such demand-side and partnering with the private sector and strategies of improving governance. civil society is urgently needed for imple- menting the agriculture-for-development A mix of centralized and decentralized agendas. In most countries, ministries services. By bringing government closer of agriculture are in need of far-reaching to rural people, decentralization holds the 24 WORLD DEVELOPMENT REPORT 2008 potential to deal with the localized and Reforming global institutions. The agri- heterogeneous aspects of agriculture, espe- culture-for-development agenda cannot be cially for extension. But not all agricultural realized without more and better interna- services should be decentralized, as some tional commitments. And the overarching such as scientific research and animal dis- global tasks of the 21st century--ending ease surveillance have important economies hunger and poverty, sustaining the envi- of scale. Decentralized institutions need to ronment, providing security, and managing address local elite capture and social exclu- global health--will not be accomplished sion, often prevalent in agrarian societies. without agriculture. The global agricul- In India, the reservation of seats for women tural agenda has a multiplicity of dimen- in local councils has helped better target sions: establishing fair rules for interna- public investments to gender-specific needs. tional trade, agreeing on product standards Elsewhere corruption has been reduced by and intellectual property rights, providing grassroots monitoring systems, government new technologies for the benefit of the poor, audits with results diffused by the media, avoiding such negative externalities as live- and use of information and communica- stock diseases, conserving the world's bio- tion technologies to keep records and share diversity, and mitigating and adapting to information. climate change. Community-driven development (CDD) With their narrow sectoral focus, the can harness the potential of rural communi- global institutions created for agriculture ties--their local knowledge, creativity, and in the 20th century, despite their many social capital. Decentralization and CDD achievements, are inadequately prepared typically contribute to the agriculture-for- to address today's interrelated and multi- development agenda in a sequenced way, sectoral agendas. Institutional reforms and focusing on basic services and public goods innovations are needed to facilitate greater first, and engaging in income-generating coordination across international agencies activities once the most basic needs have and with the new actors in the global arena, been met. Territorial development can help including civil society, the business sector, manage economic projects with a broader and philanthropy. scale than the CDD approach. Implementing the global agenda requires a mix of institutional arrangements. Spe- Improving donor effectiveness. In the cialized institutions, such as the Consul- agriculture-based countries, donors are tative Group on International Agricul- extraordinarily influential. In 24 Sub- tural Research, the Food and Agriculture Saharan countries, donor contributions Organization of the United Nations, and represent at least 28 percent of agricultural the International Fund for Agricultural development spending--and more than Development, can provide long-term sup- 80 percent in some countries. Country- port and commitment by improving their led agricultural strategies and the broader efficiency and cross-agency coordination. poverty reduction strategies provide a Cross-sectoral, issue-specific networks can framework for donors to align their sup- react quickly to emergencies, such as con- port to the agricultural sector and with trolling avian influenza, and seize emerg- each other, using the government's public ing opportunities, such as biofortification expenditure and procurement systems as through nutrient-enhanced crops. In other mechanisms for program implementation. cases, mainstreaming global priorities, At the regional level, the Comprehensive such as adaptation to climate change, into Africa Agricultural Development Program increased donor aid to agriculture may work provides priorities for coordinating donor best. Delivering on the international agenda investments. Although these national and is a matter not only of self-interest, which regional efforts provide the institutional extends broadly in a global world, but also frameworks for donor support to agricul- of equity and justice between the developed ture, progress in implementation has been and developing worlds and between present slow. and future generations. Overview 25 What now? Toward implementation requires mobilizing political support, skills, If the world is committed to reducing pov- and resources. erty and achieving sustainable growth, the There is growing recognition among powers of agriculture for development must governments and donors that agriculture be unleashed. But there are no magic bullets. must be a prominent part of the develop- Using agriculture for development is a com- ment agenda, whether for delivering growth plex process. It requires broad consultations in the agriculture-based countries or for at the country level to customize agendas reducing rural poverty and addressing the and define implementation strategies. It also environmental agenda everywhere. Today's requires having agriculture work in concert improved opportunities and greater will- with other sectors and with actors at local, ingness to invest in agriculture provide national, and global levels. It requires build- optimism that agriculture-for-development ing the capacity of smallholders and their agendas can move forward. The window of organizations, private agribusiness, and the opportunity that this offers should not be state. It requires institutions to help agricul- missed because success will provide high ture serve development and technologies payoffs toward the Millennium Develop- for sustainable natural resource use. And it ment Goals and beyond. PART I Growth and poverty reduction in agriculture's three worlds What can agriculture do for development? 1 Three out of four poor people in develop- Africa, especially when contrasted with the ing countries--883 million people--lived green revolution in South Asia (figure 1.1). in rural areas in 2002.1 Most depend on In the mid-1980s, cereal yields were compa- c h a p t e r agriculture for their livelihoods, directly or rably low and poverty was comparably high. indirectly. So a more dynamic and inclusive Fifteen years later in South Asia, yields had agriculture could dramatically reduce rural increased by more than 50 percent and poverty, helping to meet the Millennium poverty had declined by 30 percent. In Development Goal on poverty and hunger. Sub-Saharan Africa, yields and poverty There are many success stories of agri- were unchanged. Food security remains culture as an engine of growth early in the challenging for most countries in Africa, development process and of agriculture as given low agricultural growth, rapid popu- a major force for poverty reduction. Most lation growth, weak foreign exchange earn- recently, China's rapid growth in agricul- ings, and high transaction costs in linking ture--thanks to the household responsi- domestic and international markets. bility system, the liberalization of markets, Important challenges persist for agricul- and rapid technological change--has been ture in other regions as well. Where growth largely responsible for the decline in rural in nonagricultural sectors has accelerated, poverty from 53 percent in 1981 to 8 per- especially in Asia, the reallocation of labor cent in 2001 (see focus A). Agricultural out of agriculture is lagging, concentrating growth was the precursor to the accelera- poverty in rural areas and widening rural- tion of industrial growth, very much in the urban income disparities. This becomes way agricultural revolutions predated the a major source of political tensions and industrial revolutions that spread across insecurity. Where agriculture's share in the temperate world from England in the the economy has shrunk significantly, as mid-18th century to Japan in the late-19th in Latin America, connecting poor rural century.2 households to agriculture's new dynamic Agriculture has also offered attractive subsectors, either as smallholders or as business opportunities, such as high-value workers, remains a challenge. And every- products for domestic markets (dairy farm- where, agriculture is a major user and a ing in Kenya, aquaculture in Bangladesh, frequent abuser of natural resources. By vegetables for supermarkets in Latin Amer- making better use of water and land and ica) and international markets (specialty providing such environmental services as coffee in Rwanda, horticulture in Chile, managing watersheds, agriculture can make Guatemala, and Senegal). There have also growth more environmentally sustainable. been successes in traditional crops with This chapter takes a macro perspec- new demands, such as feed-maize exports tive to show that in many settings it pays to China from Laos and sugar cane for bio- to rebalance incentives facing agriculture, fuels in Brazil. manufacturing, and services and to invest Parallel to these successes are numer- better and more in agriculture. To design ous failures in getting agriculture moving. appropriately differentiated policies across Most striking is the still-unsatisfactory settings, this chapter presents a typology of performance of agriculture in Sub-Saharan countries based on agriculture's contribu- 26 Growth and poverty reduction in agriculture's three worlds 27 Figure 1.1 Cereal yields are up and poverty is down in South Asia, but cereal yields and poverty were unchanged in Sub-Saharan Africa South Asia Sub-Saharan Africa Cereal yields, Cereal yields, tons per hectare Poverty incidence, % tons per hectare Poverty incidence, % 60 60 3.5 50 3.5 50 Poverty (right axis) Poverty (right axis) 40 40 2.5 2.5 30 30 Yields (left axis) 20 20 1.5 1.5 10 10 Yields (left axis) 0.5 0 0.5 0 1984 1987 1990 1993 1996 1999 2002 1984 1987 1990 1993 1996 1999 2002 Sources: Ravallion and Chen 2004; World Bank 2006y. tion to growth and poverty reduction: agri- cally despite almost no growth in per capita culture-based, transforming, and urban- GDP, as illustrated by Nigeria (figure 1.2). ized. It reviews past policies and investment The same is true for Latin America since patterns and introduces a framework to 1980, as illustrated by Brazil. This is con- understand the political economy behind sistent with the observed urbanization of agricultural policymaking. poverty in these two regions. By contrast, the reallocation of labor out of agriculture The structural transformation has been very slow in China, partly because The process of economic development is of restrictions on labor mobility, which, one of continuous redefinition of the roles given rapid growth outside of agriculture, of agriculture, manufacturing, and services. is consistent with an increase in the rural- Two empirical regularities characterize this urban divide.3 structural transformation. First, at low lev- Agriculture's essential but els of development, the shares of agriculture in gross domestic product (GDP) and in declining contribution to employment are large (up to 50 percent and growth as countries develop 85 percent, respectively), but they decline as Many poor countries still display high agri- countries develop (figure 1.2). Second, there cultural shares in GDP and employment (an is a large and persistent gap between the average of 34 and 64 percent, respectively, share of agriculture in GDP and the share in Sub-Saharan Africa).4 In countries in the of agriculture in the labor force. These two $400-to-$1,800 GDP per capita range, many stylized facts suggest an essential but evolv- of them in Asia, agriculture is on average 20 ing role for agriculture in fostering growth percent of GDP and 43 percent of the labor and reducing poverty. force. These ratios decline to 8 percent and These patterns of structural transfor- 22 percent, respectively, in countries in the mation have been observed historically $1,800-to-$8,100 GDP per capita range, in most developed countries and are cur- many of them in Eastern Europe and Latin rently taking place in developing countries America. Adding the forward and back- that experience growth. But there are note- ward links to agriculture (extended agri- worthy deviations. In most Sub-Saharan culture) typically increases the share in the countries over the last 40 years, the share of economy by half or more, especially in the labor in agriculture has declined dramati- middle-income countries.5 28 WORLD DEVELOPMENT REPORT 2008 Figure 1.2 As countries develop, the shares of GDP and labor in agriculture tend to decline, but with many idiosyncrasies Share of labor and GDP in agriculture 1.0 NPL Share of labor in agriculture (1990­2005, average) BFA BDI Share of GDP from agriculture (1990­2005, average) NER RWA Trajectories of the share of labor in agriculture, 1961­2003 ETH MLIUGA GIN TZA MWI MOZ 0.8 LAO TCD MDG KEN PNG KHM ZMB SEN AGO CHN VNM ZAR ZWE SDN China 1961­2003 TGO CMR IND 0.6 THA BEN ZAR GHA BGD CIV BDI LAO YEM IDN TUR ETH PAK UGA TZA BOL GTM RWA Nigeria 1961­2003 LKA MLI Brazil 1961­2003 GHA KHM 0.4 NPL PHL MAR NER BEN EGY TJK TGO NGA HND PRY SDN NGA CMR PER TCD TJK MDG PNG SYR SLV PRY IRN MWI MOZ BFA UZB VNM UZB AZE ECU TUN HND DZA BGD KEN CIV SYR COL POL IND PAK MYS MEX IRN GTM 0.2 GIN YEM LKA PHL DOM BLR ROM SEN EGY BRA ZWE CHN UKR MAR COL CHL IDN ROM TUN TUR ZMB AZE BOL SLV HUN BLR BGR DZA ZAF MYS ARG UKR VEN CZE AGO ECU DOM SVK CHL BGR PER BRA THA SVK MEX ARG ZAF POL HUN VEN CZE 0 90 150 245 400 670 1100 1800 3000 4900 8100 GDP per capita, constant 2000 US$ (log scale) Source: WDR 2008 team, based on data from World Bank 2006y. Note: The list of 3-letter codes and the countries they represent can be found on page xviii. The large share of agriculture in poorer Agriculture's power economies suggests that strong growth in for poverty reduction agriculture is critical for fostering overall The large and persistent gap between agri- economic growth. As GDP per capita rises, culture's shares in GDP and employment agriculture's share declines, and so does its suggests that poverty is concentrated in contribution to economic growth. This hap- agriculture and rural areas--and that as pens while agricultural output simultane- nonagricultural growth accelerates, many ously increases in absolute value, because the of the rural poor remain poor. nonagricultural sectors are growing faster. That the incidence of poverty among Increasingly, agriculture contributes to agricultural and rural households is per- shaping the environmental sustainability of sistently much higher is confirmed by the the growth process, across the development micro evidence from numerous country spectrum. It is a major user of scarce natu- poverty studies by the World Bank (see ral resources (85 percent of the developing focus A). Furthermore, where nonagricul- world's fresh water withdrawal and 42 per- tural growth has accelerated, rural-urban cent of its land) and a largely unrecognized income disparities widen. For example, in provider of environmental services (seques- East Asia, the ratio of rural-to-urban pov- tering carbon, managing watersheds, and erty increased from about 2:1 to more than reducing deforestation). 3.5:1 between 1993 and 2002, despite a sub- Growth and poverty reduction in agriculture's three worlds 29 stantial decline in absolute poverty. Even be ascribed to improved conditions in rural with rapid urbanization, the developing areas; migration accounted for only 19 per- world is expected to remain predominantly cent of the reduction.7 The comparative rural in most regions until about 2020 (box advantage of agricultural growth in reduc- 1.1), and the majority of the poor are pro- ing poverty is also supported by economet- jected to continue to live in rural areas until ric studies. Cross-country econometric evi- 2040.6 dence indicates that GDP growth generated The persistent concentration of (abso- in agriculture has large benefits for the poor lute and relative) poverty in rural areas and is at least twice as effective in reducing illustrates the difficulty of redistributing poverty as growth generated by other sec- income generated outside of agriculture and tors, controlling for the sector's size (box the deep inertia in people's occupational 1.2). However, as countries get richer, the transformation as economies restructure. superiority of growth originating in agri- Migrating out of agriculture to urban areas culture in providing benefits for the poor is often hampered by lack of information, appears to decline. cost, skill gaps, aging, and family and social ties. Consequently, many people remain The three worlds of agriculture in rural areas with expectations for bet- ter lives unfulfilled, generating social and for development political tensions that can jeopardize the In light of the evolving role of agriculture growth process. Broad-based growth in the in fostering growth and reducing poverty, rural economy appears essential for reduc- countries are classified in this Report as ing both absolute and relative poverty. agriculture-based, transforming, or urban- Indeed, from a simple decomposition, 81 ized, based on the share of aggregate growth percent of the worldwide reduction in rural originating in agriculture and the share of poverty during the 1993­2002 period can aggregate poverty ($2.15 a day) in the rural B O X 1 . 1 Rural population dynamics An estimated 2.5 billion of the 3 billion rural population is expected to continue to grow after 2030 at the earliest. But rural areas of inhabitants are involved in agriculture: 1.5 until 2020 and decline thereafter, due to slower Latin America and East Asia have been losing billion of them living in smallholder house- population growth and rapid urbanization in population since 1995. However, the share of holds and 800 million of them working in most countries (figure below). South Asia will the population living in rural areas is declining smallholder households. The size of the rural begin such a decline only after 2025, and Africa on all continents, including Africa. Populations in developing countries will remain predominantly rural until 2020 Billions of people Rural population, millions Share of rural population 4 1,500 100 1,200 80 3 900 60 Rural 2 600 40 1 300 20 Urban 0 0 0 1950 1970 1990 2010 2030 1950 1970 1990 2010 2030 1950 1970 1990 2010 2030 Africa Latin America & Caribbean South Asia Developing countries East Asia Source: United Nations 2004. 30 WORLD DEVELOPMENT REPORT 2008 contributes only a little to growth. Poverty B O X 1 . 2 Cross-country evidence on the effect of is no longer primarily a rural phenomenon, agricultural growth on poverty reduction although the $2.15-a-day poverty incidence is 63 percent higher than in urban areas. Among 42 developing countries over erty reduction in Sub-Saharan Africa and Agriculture acts like any other competitive 1981­2003, 1 percent GDP growth South Asia, but larger poverty-reducing originating in agriculture increased the tradable sector, and predominates in some effects of growth originating in other sec- expenditures of the three poorest deciles tors in East Asia and Latin America. locations. In these economies, agriculture at least 2.5 times as much as growth can reduce the remaining rural poverty by originating in the rest of the economy including the rural poor as direct producers (figure below). Welfare gains from growth originating Similarly, Bravo-Ortega and Lederman in agriculture are substantially larger and by creating good jobs for them. for households in the poorer five (2005) find that an increase in overall GDP There is no unique route for a country expenditure deciles coming from agricultural labor productiv- to move from an agriculture-based to an ity is on average 2.9 times more effective Expenditure gains induced urbanized and eventually to a high-income in raising the incomes of the poorest by 1% GDP growth, % quintile in developing countries and 2.5 country. However, the routes traveled by 8 times more effective for countries in Latin China (1981­85 to 1996­01), India (1965­ Agriculture America than an equivalent increase in 6 Nonagriculture 70 to 1989­94), Indonesia (1970­76 to GDP coming from nonagricultural labor 1990­96), and Brazil (1970­75 to 1990­96) productivity. Focusing on absolute pov- 4 erty instead, and based on observations are illustrative (figure 1.3). Both China and from 80 countries during 1980­2001, 2 India moved from the agriculture-based Christiaensen and Demery (2007) report category to the transforming category over that the comparative advantage of agri- 0 15 to 25 years, but with little change in the culture declined from being 2.7 times more effective in reducing $1-a-day pov- rural share in poverty. Indonesia, already ­2 erty incidence in the poorest quarter of 2 3 4 5 6 7 8 9 in the transforming category in the 1970s, countries in their sample to 2 times more Lowest Highest further reduced the share of rural poverty, effective in the richest quarter of coun- Expenditure deciles tries. Using cross-country regressions per as did Brazil, a country in the urbanized region and looking at $2-a-day poverty, Source: Ligon and Sadoulet 2007. category. Note: The two curves are significantly different Hasan and Quibriam (2004) find larger at the 95 percent confidence level for the lowest The three country types capture the effects from agricultural growth on pov- five expenditure deciles. major distinguishing features in the role of agriculture for growth and poverty reduction across countries and provide a sector. Three clusters of structurally differ- useful framework to focus the discussion ent economies emerge, each with distinct and help formulate broad policy guidance. challenges for agricultural policy mak- Even so, substantial variations remain ing (figure 1.3 and tables 1.1 and 1.2). In among (and within) the countries in each the agriculture-based economies (most of type (box 1.3). them in Sub-Saharan Africa), agriculture contributes significantly to growth, and the Agriculture-based countries poor are concentrated in rural areas. The In the agriculture-based countries, most key policy challenge is to help agriculture of them in Sub-Saharan Africa, agricul- play its role as an engine of growth and pov- ture accounted for about a third of overall erty reduction. growth over 1993­2005. More than half In transforming economies (mostly in a billion people live in these countries, 49 Asia and North Africa and the Middle East), percent of them on less than $1 a day and agriculture contributes less to growth, but 68 percent of them in rural areas (tables 1.1 poverty remains overwhelmingly rural. and 1.2). By its mere size, the agricultural The rising urban-rural income gap accom- sector is critical for development, at least in panied by unfulfilled expectations creates the medium term. Both the staple crop and political tensions.8 Growth in agriculture the agricultural export sectors play impor- and the rural nonfarm economy is needed tant, but distinct roles in fostering growth to reduce rural poverty and narrow the and reducing poverty. The staple crop sec- urban-rural divide. tor is typically the largest subsector and pro- In urbanized economies (mostly in East- duces mostly for the domestic market. The ern Europe and Latin America), agriculture nonstaple crop sector typically produces Growth and poverty reduction in agriculture's three worlds 31 Figure 1.3 Agriculture-based, transforming, and urbanized countries constitute agriculture's three worlds Agriculture's contribution to growth, 1990­2005, % 80 Actual poverty data Agriculture-based BDI Predicted poverty data countries Poverty data over time RWA 60 CMR MWI SDN PRY BEN NER INDIA NGA PNG 40 CMR (1965­94) BGR TGO LAO GHA AZE CIV SYR MDG KEN BFA ETH NPL ZMB MOZ GIN MLI TCD UGA 20 HND PAK IRN YEM KHM SEN DZA GTM CHINA BRA PHL BOL AGO IDN BGD PER EGY VNM (1981­2001) ARG UKR RUS DOM IND TUN VEN SVK TUR MAR THA LKA CHN BRAZIL CHL CZE SLV ECU INDONESIA 0 (1970­96) MEX ZAR POL ROM ZAF MYS (1970­96) TJK HUN COL Transforming countries BLR Urbanized countries ZWE ­20 0.0 0.2 0.4 0.6 0.8 1.0 Rural poor/total poor, 2002 Source: WDR 2008 team. Note: The contribution of agriculture to growth is defined as the agricultural growth rate times the sector average share over the period divided by the GDP growth rate (computed from World Bank DDP 2006). Rural shares in poverty marked with a green circle are from Ravallion, Chen, and Sangraula (2007), using the $2.15/day poverty line. Rural shares of poverty marked with an orange diamond are predicted with an estimated regression of the rural share of poverty on rural share of population, agricultural share in GDP, log of GDP per capita in 2000 US$ , and regional dummies. The dynamic paths are taken from Ravallion and Chen (2004) for China; World Bank (2000b) for India; the United Nations' Economic Commission for Latin America and the Caribbean; (http://www. eclac.org) for Brazil; and the Central Bureau of Statistics (http://www.bps.go.id) for Indonesia, with poverty rates based on their national poverty lines. Arrows show paths for Brazil, China, India, and Indonesia. The list of 3-letter country codes and the countries they represent can be found on page xviii. Table 1.1 Demographic and economic characteristics of three country types, 2005 Agriculture-based countries Transforming countries Urbanized countries Population Total (millions) 615 3,510 965 Rural (millions) 417 2,220 255 Share of rural population (%) 68 63 26 Annual population growth, 1993­2005 (%) 2.5 1.4 1.0 Geographical distribution of rural population (%) Sub-Saharan Africa 82.2 13.6 4.2 South Asia 2.2 97.8 0 East Asia and Pacific Islands 0.9 96.1 2.9 Middle East and North Africa 8 92 0 Europe and Central Asia 0 12 88 Latin America and Caribbean 2.2 9.7 88.1 Labor force (in 2004) Total (millions) 266 1,780 447 Agricultural (millions) 172 1,020 82 Share of agriculture (%) 65 57 18 Economy GDP per capita (2000 US$) 379 1,068 3,489 Annual GDP growth, 1993­2005 (%) 3.7 6.3 2.6 Agriculture Agriculture value added per capita (2000 US$) 111 142 215 Share of agriculture in GDP (%) 29 13 6 Agriculture's contribution to growth, 1993­2005 (%) 32 7 5 Annual agricultural GDP growth, 1993­2005 (%) 4 2.9 2.2 Annual nonagricultural GDP growth, 1993­2005 (%) 3.5 7 2.7 Sources: Labor force data: FAO 2006a. Other data: World Bank 2006y. Note: Averages are weighted and based on 74 countries with at least 5 million people, except for agriculture value added, which is based on 71 countries because of missing information. Data are for 2005 unless otherwise noted. 32 WORLD DEVELOPMENT REPORT 2008 Table 1.2 Poverty in three country types, 2002 Agriculture-based Transforming Urbanized countries countries countries Population (millions) Total 494 3,250 888 Rural 335 2,100 251 Poverty ($2.15 a day) Total poverty rate (%) 80 60 26 Number of rural poor (millions) 278 1,530 91 Share of rural poor in total poor (%) 70 79 39 Rural poverty rate (%) 83 73 36 Urban poverty rate (%) 73 35 22 Poverty ($1.08 a day) Total poverty rate (%) 49 22 8 Number of rural poor (millions) 170 583 32 Share of rural poor in total poor (%) 70 82 45 Rural poverty rate (%) 51 28 13 Urban poverty rate (%) 45 11 6 Source: Ravallion, Chen, and Sangraula 2007. Note: Averages are weighted and based on 60 countries among those of table 1.1 for which poverty is documented in the source. Poverty lines are defined in 1993 purchasing power parity dollars. for export and is often dominated by tradi- make the nonfood tradable sector competi- tional commodities, but increasingly it also tive.11 For major staples in Africa, there is includes new dynamic subsectors of high- evidence of a negative correlation between value products such as vegetables, flowers, per capita production and price for maize and fish. in Ethiopia and Ghana; sorghum in Burkina Faso, Mali, and Sudan; cassava in Ghana; The nontradable staple crop sector. Even and (weakly) millet in Burkina Faso, Mali, with globalization, the staple crop sector and Sudan. Only Kenya, with its significant remains largely nontradable in substan- price intervention, does not follow the pat- tial parts of the agriculture-based coun- tern. However, this transmission mecha- tries for two reasons. First, locally grown nism will be sustained only if the gains from staples such as cassava, yams, sorghum, total factor productivity rise faster than the millet, and teff, which are not internation- decline in food prices so that farmer prof- ally traded (although sometimes region- itability is maintained. If not, farmers may ally traded), often predominate in the local abandon the technologies that induced the diets. Second, the domestic food economy productivity gains in the first place. remains insulated from global markets by The poverty-reducing effects of enhanc- high transport and marketing costs, espe- ing production in the farm sector depend cially in the rural hinterlands9 and in land- on the net marketing position of the poor locked countries. In Ethiopia the price of and the price elasticity of food demand.12 maize can fluctuate from around $75 per Poor net-food-buying households benefit ton (the export parity price) to $225 per ton from lower food prices, as long as the gain (the import parity price) without triggering from reduced spending on food exceeds international trade. This nontradable staple the loss from reduced wage income. Poor crop sector represents 60 percent of agricul- net-food-selling producers, by contrast, tural production in Malawi and 70 percent gain only if productivity grows faster than in Zambia and Kenya.10 prices fall. Given that demand for staple When the staple crop sector is large and crops is usually price inelastic, producers nontradable, gains in staple crop productiv- may well lose. Even so, increasing staple ity increase the aggregate food supply and crop productivity usually reduces poverty reduce food prices. That keeps the nomi- overall, because in addition to the urban nal wages of unskilled workers as well as poor, more than half of poor rural house- the prices of all the inputs that have a large holds are typically net food buyers, a little labor content at lower levels, thereby helping appreciated fact (chapter 4). Growth and poverty reduction in agriculture's three worlds 33 B O X 1 . 3 Large countries have regional heterogeneity that replicates the three worlds of agriculture In very large countries, individual states may and a few urbanized states (figure below). this heterogeneity, all states of Brazil qualify fall into different categories. India, overall a Similarly, Mexico, an overall urbanized coun- as urbanized, and in China all provinces but transforming country, also has agriculture- try, also has some transforming states and Hainan are transforming. based states such as Uttar Pradesh and Bihar two agriculture-based states. In contrast with Transforming India has agriculture-based and urbanized states, and urbanized Mexico has transforming and agriculture-based states India Mexico Agriculture's contribution Agriculture's contribution to growth, 1993­2005, % to growth, 1993­2004, % 50 50 Agriculture-based 40 40 Agriculture-based 30 Jammu-Kashmir Zacatecas Bihar 30 UP Sinaloa Punjab 20 20 Chiapas AP Michoacán Guerrero Rajasthan Durango Haryana West Bengal 10 10 Jalisco Hidalgo Oaxaca Marashtra México Baja Goa Puebla Orissa Distrito Yucatán Querétaro California Pondicherry Federal 0 Tamil Nadu 0 Arunach P Chandighar Kerala Urbanized ­10 ­10 Transforming Urbanized Transforming ­20 ­20 0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0 Rural poor/total poor, 2002 Rural poor/total poor, 2002 Sources: Central Statistical Organization and Planning Commission for India, Instituto Nacional de Estadística Geografía e Informática (INEGI) for Mexico. Figure 1.4 Price and wage effects Microevidence from Madagascar is illus- effects onfarm incomes, which dominated dominated the long-run elasticity of trative. Although rice is usually tradable, it in the short run (figure 1.4). rural poverty to cereal yields in India, proved effectively nontradable in rural areas 1958­94 of Madagascar because of high transport The tradable agricultural sector. Global- Elasticity of poverty to cereal yield costs. Analysis of commune census data ization and new dynamic producers (for ­1 shows that doubling rice yields reduces the example, coffee in Vietnam) have increased ratio of the food insecure in the community competition in traditional exports. But the ­0.8 by 38 percentage points and shrinks the recent boom in smallholder cocoa produc- ­0.6 hungry period by 1.7 months (or one-third). tion in Ghana (from 390,000 tons in 2001 to Falling rice prices and rising nominal wages 740,000 tons in 2006)14 through new plant- ­0.4 of agricultural laborers boosted real wages, ings, new varieties, and better husbandry benefiting especially the poorest, who are following higher world market prices sug- ­0.2 often net rice buyers supplying labor. Poor gests that many African countries are com- net sellers also benefited, as productiv- petitive in primary agricultural commodi- 0 ity gains exceeded food price declines.13 Short run Long run ties. Tea in Kenya is another example. And Econometric studies of India for 1958­94, there is good potential to increase yields Direct Food Wage income price effect where many of the rural poor are landless, further. New markets have also opened effect effect report price and wage effects of food crop for traditional exports, such as premium productivity to be more important in reduc- coffees, and for nontraditional high-value Source : Datt and Ravallion 1998a. Note : The direct income effect includes that ing rural poverty in the long run than direct agricultural products, such as vegetables from higher yields and employment. 34 WORLD DEVELOPMENT REPORT 2008 (from Senegal), fish (from Uganda), and able agriculture, growth of agriculture can flowers (from Kenya).15 enhance growth in other sectors through con- The defining macroeconomic contribu- sumption and production links. When agri- tion of tradable agriculture to aggregate cultural incomes are spent on domestically growth is foreign exchange, which allows produced nontradable goods and services, imports of inputs and capital goods. Coun- it stimulates demand for domestic industry tries with mineral resources, such as Zam- and services. Production links proceed for- bia, obviously depend less on their agricul- ward by fostering growth in agroprocessing tural exports. But most agriculture-based and food marketing and backward through economies depend on agriculture for a large demand for intermediate inputs and services. share of their foreign exchange, as exempli- The availability of resources (entrepreneur- fied by tobacco exports in Malawi. ship, excess capacity) and a favorable invest- The poverty-reducing effects of developing ment climate that allow a supply response tradable agriculture depend on the participa- from the nonagricultural sector are critical tion of smallholders and poor households in for realizing such links. production. Labor intensive nontraditional Empirical evidence confirms these exports can also have substantial local pov- multiplier effects.17 The strength of the erty-reducing effects by generating employ- agricultural multipliers differs depending ment, as in Kenya and Senegal,16 despite the on a country's economic structure. Small tightening food standards and more verti- economies with large tradable sectors (for cally integrated market chains that tend to example, Lesotho) have smaller multipliers favor medium farms (chapter 5). than large economies with a high share of nontradable agriculture and services (for Links with sectors outside of agriculture. In example, Cameroon, Nigeria, and Tanza- addition to the macroeconomic channels nia). Most of these linkage effects occur through prices for nontradable agriculture through commerce and services. Hence and through foreign exchange for trad- globalization and inexpensive imports of manufactured goods in rural markets--say, from China--likely have limited effects on the strength of the links. They also enable B O X 1 . 4 Agriculture's comparative advantage new agro-based exports to create links. in Sub-Saharan Africa Agriculture's comparative advantage support the contention that indirect costs Agriculture as an engine for growth early comes from three sources: inherent in a poor business environment on. Agriculture is an effective engine for First, from factor endowments. Most are higher on average in Africa than in growth for most agriculture-based coun- African and agriculture-based economies their competitors in the developing world. tries because they need to produce most of are relatively rich in natural resources, but Third, from dynamic economies of poor in skilled labor, suggesting compara- scale. The very existence of economies of their own food, and they are likely to keep tive advantage for unprocessed primary scale puts late-comers at a disadvantage a comparative advantage in agriculture at products. In some countries, a combina- in competing with countries that have least in the medium term. Consider food tion of natural resources and human already developed their industrial base. production first. In low-income countries, capital endowments point to comparative Agriculture-based economies have largely advantage in processed primary com- missed the expansion of labor-intensive the demand for staple food is driven by modities, even though other factors may manufacturing that spurred development rapid population growth and high income have prevented the development of the in Asia in the 1980s. There is still debate elasticity. In Africa, demand for food is agricultural processing sector to date. on the likelihood that Africa will emerge Second, from the difference in produc- expected to reach $100 billion by 2015, dou- as a significant exporter of manufactured tivity and costs. These are determined by goods. But, based on current and emerg- ble its level of 2000.18 With staples mostly the business environment, infrastructure ing comparative advantage, a diverse nontradable, and frequent shortages of (roads, electricity, communications), and portfolio of processed and unprocessed foreign exchange for importing substitute institutions (legal, financial, regulatory) primary-based exports (including services that influence the efficiency of operations such as tourism) will remain the main cereals, food production in the agriculture- for firms and industries. The business option for generating foreign exchange in based countries has to keep up with domes- environment is more important for manu- the medium term. tic demand (see focus C). facturing and high-value services because Source: Collier and Venables (Forthcoming); Now consider exports. Beyond Mauritius they use these factors more intensively. Eifert, Gelb, and Ramachandran 2005; Wood World Bank Investment Climate surveys and Mayer 2001. and, more recently, apparel from Kenya and Madagascar under preferential trade agree- Growth and poverty reduction in agriculture's three worlds 35 ments (especially the African Growth and Opportunities Act), manufactured exports B O X 1 . 5 A role for agriculture in Africa's have not taken off in Sub-Saharan Africa. mineral-rich countries African exports are concentrated in unpro- cessed primary products, in sharp contrast Agriculture accounts for one-third of the ture, irrigation, agricultural credit, and economies of African mineral-rich coun- fertilizer subsidies. Nigeria, by contrast, with the manufactured goods exported tries. Between 1985 and 1999, agriculture squeezed agriculture, directly through the from the transforming countries of Asia. contributed on average twice as much as marketing boards, and indirectly through While some of that difference is related industry to their overall growth.21 Poverty its fixed exchange rate, which heavily remains widespread, however, despite taxed its agricultural exports and subsi- to macro and trade policies, this trade higher average per capita GDP than in the dized cheap imports. composition largely corresponds to the mineral-poor countries. The contrasting In Indonesia $1-a-day poverty declined comparative advantages for most African pre-1997 experiences of Indonesia and from 47 percent in 1981 to 14 percent in countries (box 1.4). Therefore, the growth Nigeria, both large oil-exporting coun- 1996. In Nigeria it increased from 58 per- tries, is telling. cent to 70 percent in the same period.22 strategy of agriculture-based economies for Indonesia supported agriculture, indi- The different treatment of agriculture many years to come has to be anchored in rectly through regular devaluations of the explains much of these widely divergent improving agricultural productivity. exchange rate that provided incentives outcomes. to its producers of agricultural tradables, and directly through investments of some What history shows. Higher agricultural Sources: Mwabu and Thorbecke 2004; World windfall oil revenues in rural infrastruc- Bank 1982. productivity generating an agricultural surplus, taxed to finance industrial devel- opment, and enabling lower food prices underpinned early development in Western land and forests, as a basis for sustainable Europe, the United States, and Japan, and agricultural growth. As shown by the con- later in Taiwan, China, and the Republic of trasting experiences of Indonesia and Nige- Korea.19 More recently, rapid agricultural ria, both large oil exporters, fostering agri- productivity growth in China and India has cultural growth is appropriate for reducing been widely credited with initiating indus- poverty in mineral-rich countries as well trialization and inducing rapid reductions (box 1.5). in poverty.20 The critical insight from these successful experiences is that the adverse Transforming countries effects of surplus extraction on agriculture More than 2 billion people, about three- were each time counterbalanced (or pre- quarters of the rural population in devel- dated) by public investment in scientific oping countries, reside in the rural areas research for agricultural technologies and in of transforming economies, encompassing rural infrastructure, including irrigation. most of South and East Asia, North Africa Premature and unduly high extraction and the Middle East, and some of Europe through an urban policy bias combined and Central Asia. Although agriculture with a lack of public investment in agri- contributed only 7 percent to growth dur- culture despite good growth potential are ing 1993­2005, it still makes up about 13 highlighted in the next section as key rea- percent of the economy and employs 57 sons for sluggish agricultural performance percent of the labor force. Despite rapid in many agriculture-based countries. Gha- growth and declining poverty rates in na's growth and poverty reduction in the many of these countries, poverty remains 2000s suggest that robust balanced agricul- widespread and largely rural--more than tural growth is still feasible today (see focus 80 percent of the poor live in rural areas. A). In countries, or regions within coun- Natural resources are also coming under tries, with poor agroecological conditions, growing pressure from agriculture and the agriculture's contributions to growth will competition for land and water from rap- be limited. Even so, agriculture is still likely idly growing urban populations and non- to play an important complementary role agricultural sectors. in reducing poverty and improving food security (see focus C). Agricultural intensi- Managing the rural-urban divide. A dis- fication will also be critical for reversing the tinguishing feature of transforming econo- degradation of natural resources, especially mies is the widening gap between rural and 36 WORLD DEVELOPMENT REPORT 2008 urban incomes. In China the incidence of make migration more attractive, inflating urban poverty declined twice as fast as that the pool of urban unemployed, leading to of rural poverty between 1980 and 2001; in urban congestion and the urbanization Indonesia, 2.5 times as fast over the same of poverty. Complementing these policies period; and in Thailand 3.7 times as fast with those that foster rural income growth between 1970 and 1999.23 and slow migration out of the traditional Nonagricultural sectors now account sector can provide important synergies.25 for most of the economic growth. But the Rural income growth can do much for transition of people out of agriculture and poverty reduction in the transforming rural areas is not keeping pace with the countries (see focus A). For example, 75­ restructuring of economies away from 80 percent of the dramatic drop in national agriculture. In China, longstanding policy poverty in China during 1980­2001 was impediments to labor mobility24 kept the the result of poverty reduction in the rural rural population behind while urban econ- areas. A similar pattern was observed in omies were expanding rapidly. In India, the Indonesia where the emergence of rural low level and quality of education of most towns ("urbanization without migration") rural workers is mainly responsible for was further emphasized.26 their inability to find jobs in the booming services economy. Reducing rural poverty through the new One policy response is facilitating faster agriculture and nonfarm employment. absorption of the agricultural labor force in Historically, there have been numer- the urban economy through investments in ous attempts to reduce rural poverty and human capital and labor market policies, address the rising income gap by increasing such as vocational training, transport ser- agricultural protection, often with limited vices, and job matching (see chapter 9). But success. The current call for agricultural the time lags in educating people are sub- subsidies in the face of weak fiscal capac- stantial. Moreover, the same policies also ity in the transforming countries is also unlikely to provide a sustainable solution to massive rural poverty (box 1.6). Increasing agricultural productivity, B O X 1 . 6 Supporting farmers without a strong fiscal including yields for staple crops, will be base: lessons from Thailand critical in countering pressures for agri- Before the 1960s, Thailand was an rice premium revenues fell with the decline cultural protection. Staple crops are still agriculture-based country with rice in world rice prices after the food crisis. the largest agricultural subsector (slightly accounting for the bulk of its export earn- This episode epitomizes the dilemma more than a third of agricultural output in ings. Rice exports were heavily taxed, in formulating sustainable policies to China and India, and slightly more than mainly through a duty levied proportional address rural-urban disparities. The pro- to export quantities (the rice premium), gram was contradictory because it tried half in Vietnam). In some countries that which hovered around 30 percent until the to support farmers based on the revenue are large players in international markets, mid-1970s. This served the dual purpose from taxing them, without a strong fiscal continuing to focus on food staples is also of raising government revenue for invest- base outside of agriculture. Even if the necessary to ensure national food secu- ment and securing cheap food for urban program had worked, increasing rice prices consumers. As GDP per capita doubled would have met strong resistance from rity. But rising incomes shift the compo- and exports from labor-intensive manu- poor urban consumers. sition of food expenditure from basic and facturing increased (40 percent by the As Thailand's economy advanced, the unprocessed staple foods to more varied end of the 1970s), widening rural-urban rice premium was gradually reduced and disparities pressured politicians to install diets with processed foods (chapter 2). So then abolished in 1986. New support pro- visible measures supporting farmers. grams have since been introduced, such as growth in agriculture is increasingly driven After some political instability, the the commodity credit program. Low-inter- by the rapidly expanding demand for live- Farmers' Aid Fund was established in 1974, est government loans are given against stock products and high-value crops, which based on large rice premium revenues from the pledge of rice, with the pledged rice sharp increases in world rice prices during canceling the debt if rice prices do not are also more labor intensive.27 the world food crisis of 1973­75. The fund meet a target. However, such programs The poverty impact of growth in the undertook several programs to support are unlikely to be sustainable or generous agricultural sector will thus depend increas- farmers, including price supports through enough to close income gaps. ingly on the poor connecting to these new government rice purchases. Yet the pro- gram was soon terminated, largely because Source: Hayami 2005. growth processes, either as smallholders or as laborers. Vertically integrated supply Growth and poverty reduction in agriculture's three worlds 37 chains may pose particular challenges for higher urban labor costs also stimulated them (see chapter 5), although recent evi- urban-to-rural subcontracting in vari- dence from China suggests that small and ous sectors throughout East Asia, both for poor farmers take an active part in China's domestic consumption and for export.36 rapidly expanding horticulture economy.28 Without the rapid expansion of rural non- farm employment through subcontracting Nonfarm employment. Agriculture alone in the export-oriented town and village cannot relieve rural poverty; rural nonfarm enterprises, rural poverty and inequality employment is also important. The poten- would have been much higher in China's tial of agriculture to contribute to rural central province of Hubei.37 poverty reduction differs across countries. Poverty reduction through rural non- In China, where land is relatively equally farm employment is often indirect. In distributed, the reduction in poverty was India and Bangladesh, relatively few of almost four times higher from GDP growth the poor gain access to nonfarm jobs.38 originating in agriculture than from GDP Yet by siphoning off nonpoor agricultural growth originating in industry or ser- wage laborers, nonfarm employment puts vices.29 Rapid agricultural development upward pressure on agricultural wages, also contributed substantially to the dra- benefiting the poor. matic poverty reduction in Vietnam over the past 15 years and is likely to remain an Urbanized countries important pathway out of poverty for many Agriculture makes up only 6 percent of the of Vietnam's poor.30 In India and Indone- urbanized economies and contributes about sia, however, growth in rural services was proportionately to growth, but the agribusi- estimated to contribute at least as much as ness and food industry, and services can growth in agriculture toward reducing pov- account for 30 percent of GDP. Although erty.31 In India the poverty-reducing effects almost three-quarters of the population of of nonfarm economic growth are greater in urbanized countries lives in urban areas, 45 states with higher initial levels of farm pro- percent of the poor are in rural areas, and ductivity and rural living standards.32 18 percent of the labor force still works in Growth in rural nonfarm employ- ment in many cases remains closely linked Figure 1.5 The ratio of food processing to agricultural value added rises with incomes to growth in agriculture, as agriculture Food processing value added/agriculture value added becomes a larger supplier of intermediate 0.6 inputs to other sectors such as processed HUN foods (forward linkages) (figure 1.5). Rural trading and transport, often of food, make up about 30 percent of rural nonfarm employment.33 Econometric estimates ARG from rural China also suggest significant 0.4 ROM BRA MEX cross-sectoral effects from growth in farm- ing to certain nonfarming activities, with less evidence of reverse linkages.34 ZWE ECU IRN MYS But with urbanization and globaliza- COL ZAF tion, growth in rural nonfarm employment BOL PER 0.2 SVK occurs increasingly independently from SEN PHL TUR agriculture. Regions in India with the slow- MAR MWI THA est growth in agricultural productivity had IDN the largest increase in the rural nonfarm EGY NPL BGD tradable sector.35 When capital and prod- IND ucts are mobile, investors seek low-wage 0 UGA 0 2,000 4,000 6,000 8,000 opportunities in areas that did not increase GDP per capita, constant 2000 US$ their incomes through higher agricultural Source : World Bank 2006y; UNIDO Industrial Statistics Database 2005. productivity. Urban overcrowding and Note : The list of 3-letter codes and the countries they represent can be found on page xviii. 38 WORLD DEVELOPMENT REPORT 2008 agriculture. Most countries in Latin Amer- good jobs in agriculture and the rural non- ica and many in Europe and Central Asia farm economy (chapter 10). The rapid con- fall into this category. centration in agribusiness and food retail- ing sharpens this challenge (chapter 5). Agriculture: a good business with poverty- reducing potential. In urbanized coun- Agriculture's development tries, agriculture acts like other tradable potential shortchanged sectors, often economically important in The agriculture-for-development connec- specific subregions. It provides growth tions revealed by the evidence reviewed here opportunities in subsectors with a com- have too often not been exploited. Certainly parative advantage as well as environmen- agriculture has yet to perform as an engine tal services--with new markets for biofu- of growth in most Sub-Saharan countries, els, carbon trading, and the preservation where populations are slowly urbanizing of biodiversity opening opportunities yet without a reduction in poverty. Even in the largely to be tapped. The main divide is now transforming countries, the rural poverty between the traditional rural sector and the and income disparity challenges remain modern rural and urban sectors. The pres- huge, despite spectacular growth in some sure for agricultural protection remains. countries. With agriculture mostly tradable, prices Four hypotheses could explain this do not decline from growth in productivity, divide between promise and reality: and landowners capture most of the sur- plus. The distribution of land and the labor · Agricultural productivity growth is intrin- intensity of production govern the pov- sically slow, making it hard to realize the erty-reducing effects. Poverty is increas- growth and poverty-reducing potential of ingly reduced through the employment of agriculture. unskilled labor. Much of the expansion of · Macroeconomic, price, and trade policies Chile's agricultural GDP can be attributed unduly discriminate against agriculture. to a labor-intensive agroexport boom over · There has been an urban bias in the allo- the past two decades. The rural poor ben- cation of public investment as well as efited indirectly through their employment misinvestment within agriculture. by large-scale farmers and agroprocessors, Official development assistance to agri- with many jobs taken by women. The pov- · culture has declined. erty-reducing impact has been substantial, despite vertically integrated supply chains. Each percent expansion of agricultural Is the agricultural sector and agroprocessing output is estimated to less productive? have reduced national poverty by 0.6­1.2 Some refer to the oft-observed slower percent.39 growth in agriculture than in the rest of the But success in agriculture does not always economy to argue that agriculture is inher- reduce poverty. Brazil experienced dramatic ently less dynamic. The argument goes as growth in agriculture during the 1990s, fol- far back as Adam Smith, who posited that lowing trade liberalization and an improve- productivity was bound to grow slower in ment of price incentives. But it is unclear agriculture than in manufacturing because how much the boom reduced rural poverty of greater impediments to specialization because agricultural employment declined and the division of labor in agricultural and shifted to higher-skilled wage workers as production. More recently it is argued, production became more capital intensive. especially for Africa, that rapid agricul- The reduction in rural poverty was largely tural growth will be difficult because of an the result of income transfers and employ- inherently unfavorable agroecological base, ment in the rural nonfarm economy.40 rapid soil degradation, low population den- The challenge of using agriculture for sity, poorly functioning markets, and com- development in the urbanized countries petition from the rest of the world.41 is to create opportunities for smallholders In this debate, it is important to dis- in supplying the modern food markets and tinguish the rate of growth in output (or Growth and poverty reduction in agriculture's three worlds 39 value added) in agriculture from the rate Figure 1.6 Labor productivity has been a more important source of growth in agriculture of growth in some measure of productiv- than in nonagriculture, 1993­2005 ity, such as labor productivity or total fac- Percenta Percenta Percenta tor productivity. Comparing the rate and 7 7 7 sources of growth in value added in agri- 6 6 6 culture and in the nonagricultural sectors over the past 15 years shows different pat- 5 5 3.68% 5 terns over the three worlds of agriculture 4 4 4 (figure 1.6). In transforming countries, the extraordinary dynamism of the nonagricul- 3 3 3 2.13% tural sector is reflected in its sustained high 2 2 2 0.55% growth rate based on both the increase in 4.39% 2.21% 3.16% employment and in labor productivity--as 1 1.79% 1 1 2.69% 1.96% evident from this decomposition of growth. 0.65% 0 0 0 But rates of growth in agriculture and non- ­0.63% ­1.01% agriculture are similar in the agriculture- ­1 ­1 ­1 based and urbanized countries. And labor Agriculture Nonagriculture Agriculture Nonagriculture Agriculture Nonagriculture productivity in agriculture grew faster than Agriculture-based countries Transforming countries Urbanized countries in nonagriculture in each of these two coun- Annual growth in employment, % Annual growth in labor productivity, % try categories. Moreover, total factor productivity Source: FAO (2006a). a. Annual sector growth rate (equal to the sum of the growth rates for employment and labor productivity). (TFP) has grown faster in agriculture than in industry in many settings. For 50 low- countries taxed agriculture relative to other and middle-income countries during 1967­ sectors. Interventions induced a 30 percent 92, the average growth in TFP was 0.5 to decline in the relative price of agricultural 1.5 percentage points higher in agriculture products with respect to a nonagricultural than in nonagriculture, with comparable price index. This policy bias was largest in differences observed across the develop- agriculture-based countries of Sub-Saha- ment spectrum.42 ran Africa, with overvalued exchange rates, These findings are not taken to claim high tariff protection in industry, and taxes superiority in agricultural TFP growth over on agricultural exports all contributing to the past decades, but to refute the notion the bias. It was estimated that a 10 percent- that agriculture is a backward sector, where age point reduction in total taxation to investment and policies are automatically the sector would increase overall annual less effective in generating growth. Brazil growth by 0.43 percentage points. and Chile--where agricultural commodi- Since then, most developing countries ties have become mostly tradable and where have substantially improved their macro- growth in agriculture has exceeded growth economic policy and reduced their biases in nonagriculture for more than a decade-- against agriculture (chapter 4). A com- confirm that agriculture can be a dynamic posite score comprising three key elements sector. But in many countries where agri- of sound macroeconomic policy (fiscal, culture is less tradable, it is likely to grow monetary, and exchange rate) shows a clear more slowly than nonagricultural sectors, improvement since the mid-1990s in almost given Engel's Law (as incomes rise, the pro- all Sub-Saharan African countries (figure portion spent on food falls). 1.7). A positive association is also observed between improvement in that score and the Are macroeconomic, price, and performance of agriculture. trade policies discriminating against Econometric evidence at the country agriculture? level shows that periods of rapid growth in There is considerable evidence that slower agriculture and substantial poverty reduc- growth in agriculture relates to the macro tion have followed reforms. In Uganda the and sectoral policy biases against it. The increase in coffee prices--largely brought landmark Krueger, Schiff, and Valdés about by domestic market liberalization, but (1991)43 study clearly documented how 18 also by the devaluation of the exchange rate 40 WORLD DEVELOPMENT REPORT 2008 Figure 1.7 Macroeconomic policy and agricultural growth have improved in Sub-Saharan Africa Macroeconomic score Agriculture growth rate, % per year 10 6 5 8 4 6 3 2 4 1 2 1984­95 0 1995­2005 0 ­1 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 ­0.2 0 0.2 0.4 0.6 Average annual change in macroeconomic score Source: http://www.icrgonline.com. Note: The macroeconomic score is the average of the budget balance score, inflation score, and exchange rate stability score provided by the International Country Risk Guide. Each point represents a country. and favorable world prices--substantially It was the heavy exploitation of agriculture reduced rural poverty during 1992­2000 by before meaningful (public) investment spurring a supply response. It is estimated in agricultural development that proved that a 10 percent increase in the price of lethal, especially in Africa. The goose was coffee reduces the poverty headcount by 6 often killed before it could lay its golden percentage points.44 In China 60 percent of egg. The share of public spending on agri- the dramatic expansion of agricultural out- culture in agriculture-based countries put and 51 percent of the reduction in rural (mostly in Africa) is significantly less (4 poverty from 33 to 11 percentage points percent in 2004) than in the transforming between 1978 and 1984 have been attrib- countries during their agricultural growth uted to institutional reforms, especially the spurt (10 percent in 1980) (table 1.3). The household production responsibility sys- low levels of agricultural spending in Sub- tem, and to price reforms.45 Saharan Africa are insufficient for sustained Even where macroeconomic and price growth. Recent advocacy by the New Eco- policies have been reformed, interna- nomic Program for African Development tional trade policies--especially protection to increase agricultural spending to 10 and subsidies of member countries of the percent of national budgets aims to reverse Organisation for Economic Co-operation this trend, bringing it to a level that is closer and Development (OECD)--continue to to that which brought success to the now impose substantial costs on developing- transforming countries. country agriculture. Overall trade policies To assess optimal cross-sectoral allo- depress prices of agricultural products in cations of public investment, the returns international markets by an average of 5 to spending across sectors would ideally percent (chapter 4). Only modest progress be systematically compared. Doing so is has been made to date in reforming these fraught with conceptual, methodological, policies, and much depends on a successful and data problems, indicating an important outcome of the Doha Round of trade talks. continuing research agenda. High returns to agricultural research and extension have been documented, with a meta-analysis Is public spending biased toward reporting rates of return in the range of 35 urban needs? percent (Sub-Saharan Africa) to 50 percent Successful countries have invested in agri- (Asia) for 700 studies, far above the cost of culture before taxing it (directly and indi- money accessible to developing countries rectly) to finance industrial development.46 (see chapter 7).47 While irrigation projects Growth and poverty reduction in agriculture's three worlds 41 Table 1.3 Public spending in agriculture-based countries is low Agriculture-based countries Transforming countries Urbanized countries 1980 2004 1980 2004 1980 2004 Public spending on agriculture as a share 6.9 4.0 14.3 7.0 8.1 2.7 of total public spending (%) Public spending on agriculture as a share 3.7 4.0 10.2 10.6 16.9 12.1 of agricultural GDP (%) Share of agriculture in GDP (%) 28.8 28.9 24.4 15.6 14.4 10.2 Source: Fan forthcoming. Note: Numbers for agriculture-based countries are based on 14 countries (12 from Sub-Saharan Africa), those for transforming countries on 12 countries, and those for urbanized countries on 11 countries. in Sub-Saharan Africa were often inef- where agricultural subsidies rose from 40 fective in the 1970s and 1980s, returns on percent of agricultural public expenditures projects now often reach the 15­20 percent in 1975 to 75 percent in 2002 (chapter 4). range commonly obtained in the rest of the Underinvestment in agriculture, especially world (chapter 2).48 Evidence from rural pronounced in the agriculture-based econo- Uganda shows agricultural R&D and rural mies, is further compounded by misinvest- feeder roads as profitable investments.49 ment, especially in the transforming and In Asia and Latin America, the decline urbanized countries. in public funding for agriculture partly reflects agriculture's diminishing impor- Development assistance to tance in the economy (table 1.3). There have agriculture declined dramatically been recent reversals in several countries The share of agriculture in official develop- though, including China, India, and Mex- ment assistance (ODA)52,53 declined sharply ico,50 motivated by the need to fight poverty over the past two decades, from a high of and narrow the rural-urban income gap. about 18 percent in 1979 to 3.5 percent in Agricultural spending has often been 2004 (figure 1.8). It also declined in abso- biased toward subsidizing private goods (fer- lute terms, from a high of about $8 billion tilizer, credit) and making socially regressive (2004 US$) in 1984 to $3.4 billion in 2004. transfers. These are overall substantially less The bigger decline was from the multilat- productive than investments in core public eral financial institutions, especially the goods such as agricultural research, rural World Bank. In the late 1970s and early infrastructure, education, and health.51 The 1980s the bulk of agricultural ODA went bias toward private goods often worsens as to Asia, especially India, in support of the countries' GDP per capita rises, as in India, green revolution, although this declined Figure 1.8 Official development assistance to agriculture declined sharply between 1975 and 2004 Share and level of ODA to agriculture Source of ODA to agriculture ODA to agriculture by region, 2004 US$ billions 2004 US$ billions Percentage 2004 US$ billions 2004 US$ billions 10 20 5 5 % of total ODA to agriculture Multilateral (right axis) Asia 8 16 4 4 Bilateral 6 12 3 3 Africa 4 ODA to agriculture 8 2 2 (left axis) 2 4 1 1 Latin America 0 0 0 0 1975 1980 1985 1990 1995 2000 2005 1975 1980 1985 1990 1995 2000 2005 1975 1980 1985 1990 1995 2000 2005 Source: OECD 2006a. Note: Data smoothed by locally weighted regressions. 42 WORLD DEVELOPMENT REPORT 2008 dramatically thereafter. Total ODA to agri- The political economy culture in Africa54 increased somewhat in of agricultural policy the 1980s, but it is now back to its 1975 level While the low-productivity beliefs may be of about $1.2 billion. This decline in atten- changing under the weight of evidence, and tion to agriculture is all the more striking the macroeconomic context has definitely because it happened in the face of rising improved, a better understanding of the rural poverty. political economy of agricultural policy A complex of reasons explains the making is necessary to address the continu- decline of donor support to agriculture and ing policy neglect and under- and misin- rural development: (1) falling international vestment in the sector. This understanding commodity prices that made agriculture will be used in chapters 4 to 8 to interpret less profitable in developing countries; (2) policy outcomes, and in chapters 10 and increased competition within ODA espe- 11 to design agriculture-for-development cially from social sectors; (3) emergency agendas that meet the political feasibility responses to numerous crises; (4) opposi- criterion. tion from farmers in some donor countries to supporting agriculture in their major export markets; and (5) opposition from The process of agricultural environmental groups that saw agricul- policy making ture as a contributor to natural resource Agricultural policy making can be seen as destruction and environmental pollution. the outcome of a political bargain between Failed agricultural development efforts politicians and their citizens.56 Citizens can also influenced the expectations of donors. be atomistic individuals who demand pol- The "agroskepticism" of many donors may icy action in exchange for political support well be related to their experience with past (votes) or they can be organized in lobbies unsuccessful interventions in agriculture, that defend special interests. such as large-scale integrated rural devel- opment and the training-and-visit system State objectives and policymaking. Politi- for extension, which were both promoted cians enjoy different degrees of autonomy. heavily by the World Bank.55 Poor under- They have their own objectives, for example, standing of agrarian dynamics, weak gover- to be reelected or to maintain legitimacy, to nance, and the tendency for donors to seek improve the welfare of their constituency, one-size-fits-all approaches contributed to or to pursue some vision for the country. the failures. Implementation difficulties are Institutions such as the structure of the especially challenging in agriculture with bureaucracy, alternative forms of represen- weak governance and the spatial dispersion tation, agenda-setting mechanisms, and of programs. This experience underlines reward systems condition their preferences the need to strengthen donor and country and power in the political game. There are capacity for program design and to invest many examples of major policy reforms led in governance and institutions for effective by a state with considerable autonomy in implementation (chapter 11). decision making. The green revolution in Since 2001, government and donor inter- Asia, for example, occurred in both demo- est in agriculture has increased, at least in cratic and nondemocratic political systems. discourse and modestly in support. This is In India, the driving force of the green happening because of a turnaround in the revolution was the political will to become reasons for the decline in support to agri- food self-sufficient, once the U.S. govern- culture, such as higher international com- ment decided in the mid-1960s to use food modity prices; higher priority of agricul- aid as an instrument of foreign policy.57 ture to developing-country governments; Indonesia (under Suharto) is an example of and new approaches to agricultural devel- a single-party regime that launched a green opment projects based on decentralization, revolution. participation, and public-private partner- Authoritarian regimes in Africa appar- ships, with greater likelihood of success. ently had fewer political incentives to sup- Growth and poverty reduction in agriculture's three worlds 43 port smallholder agriculture. African states capital, such as the skills to influence poli- used both coercion and the strategic sup- tics. And--importantly--they need social port of larger farmers to suppress opposi- capital such as strong membership orga- tion to agricultural pricing policies that nizations that can be mobilized for dem- taxed agriculture.58 There are also numer- onstrations and lobbying. In developing ous cases in which African states did make economies, farmers' transaction costs in serious efforts to intensify agricultural pro- collective action are high in view of their duction, but unlike in Asia, many focused large numbers, dispersed nature, high on large-scale production, without sus- transportation and information costs, pov- tained success.59 erty, and strong patronage relations with Economic crises can give policy mak- a landlord class that may pursue oppo- ers more autonomy to engage in reforms site interests. For this reason, smallholder that were difficult in normal times. Many interests tend to be poorly represented, and reforms of the role of the state in agricul- policy is biased toward urban interests and ture were introduced as part of structural those of the landed elite. adjustment made inevitable by the debt cri- The urban poor, by contrast, do not need sis--for example, the dismantling of mar- a high degree of organization to stage a pub- keting boards in Uganda (see box 4.4). lic protest, as illustrated by the food riots More often, policy makers seek to maxi- over the price of bread in Egypt. Industrial mize political support within their resource groups usually have more financial resources constraints. Political support is usually to influence politics, and they often belong related to the expected policy-induced to social elites, whose social capital facilitates changes in welfare. Hence politicians may lobbying. As countries urbanize and indus- rally support by favoring groups that are los- trialize, farmers face fewer challenges to col- ing ground relative to the others. Farm sub- lective action. Their numbers decrease and sidies were introduced in the 1930s in the their access to resources increases while the United States when farm incomes dropped widening income gap between the agricul- 50 percent more than those of their urban tural and nonagricultural sectors provides counterparts. Electricity subsidies in India a cause for action. Historically, in industrial are maintained partly as a compensation economies, farmers have formed astonish- for the increasing income disparity between ingly effective pressure groups to pursue the agricultural and nonagricultural sec- agricultural protection and subsidies, which tors. China's bold reforms launched in 1978 have proved extremely difficult to dismantle answered the imperative of restoring China's in spite of the rapidly decreasing number of food independence and a minimum living farmers (see chapter 4).61 standard for all its citizens. In democracies, Democratization in many developing the votes of farmers can be very influential. countries has increased the possibilities The 2004 elections in India, for example, for smallholders to form organizations were won by a party coalition that promised and influence politics. In West Africa, for to resolve "agrarian distress."60 example, producer organizations and par- liaments are increasingly involved in the Collective action and policymaking. Or- formulation of agricultural strategies and ganized groups of citizens can have strong policies (see chapter 11).62 They have influ- influence over the policy process. The power enced policy making in Senegal and Mali. of lobbies depends on their ability to over- Whether these agricultural policies will come the costs of organization and free- increase budget allocations to agriculture riding. Extensive empirical evidence shows remains to be seen. that small and more geographically con- centrated groups fare better, as do groups Why use inefficient better organized and with strong leader- policy instruments? ship. To be effective, lobbies need financial Imperfect information on welfare effects resources--for example, to contribute to implies that certain policy instruments are political campaigns. They also need human politically more effective than others, even 44 WORLD DEVELOPMENT REPORT 2008 if less efficient economically. As politicians Decentralization and closer proximity maximize short-run political support rather between the electorate and policy makers than their constituency's welfare, they pre- may be part of the answer. Increasing the fer the former instruments over the latter. autonomy of compensatory agencies or cast- For example, price supports are chosen over ing compensations into legislation--such as direct income transfers because self-suffi- Mexico's PROCAMPO to make the North ciency appeals to a nationalistic sentiment American Free Trade Agreement negotia- of voters, farmers fear an excessively visible tions politically acceptable to producers of "welfare" stigma, and information on the crops competing with imports--have been cost of direct transfers could lead to sub- used successfully, with the risk of irrevers- sequent policy reversal. Some instruments ibility once subsidies have been introduced. have benefits that are easier to target to A new role for agriculture political clients, such as investment projects or food aid. Broadly distorting export taxes in development may thus be maintained to provide fiscal The case for using the powers of agriculture revenues that can be used to reward politi- for poverty reduction and as an engine of cal clients and ethnic-group supporters.63 growth for the agriculture-based countries Certain instruments have costs that are is still very much alive today. Effective use easier to conceal--for example, trade taxes requires adjusting agendas to each country as opposed to land or value added taxes. type and within countries as well. However, Net social cost is exchanged for political despite convincing successes, agriculture feasibility and redistributive gains. has not been used to its full potential in The inability to make credible commit- many countries because of anti-agriculture ments in a dynamic policy process may fur- policy biases and underinvestment, often ther force the government into suboptimal compounded by misinvestment and donor policy. Groups losing from reform anticipate neglect, with high costs in human suffering. that they will be worse off in the long run, New opportunities for realizing this poten- even though compensation may be prom- tial are present today, but also coming are ised now. Lack of a commitment device to new challenges, particularly in pursuing a clinch compensation when there is a delay smallholder-driven approach to agricultural between policy implementation and redis- growth that reconciles the economic, social, tributive effects is a major hurdle to policy and environmental functions of agriculture. making. The resulting status quo bias has The following chapters explore the instru- been used to explain opposition to trade ments available to use agriculture for devel- reforms and to the removal of subsidies in opment and how to define and implement exchange for better future public services. agendas specific to each country type. focus A Declining rural poverty has been a key factor in aggregate poverty reduction Poverty rates in rural areas have declined over the past decade, mostly because of the impressive gains in China. But 75 percent of the world's poor still live in rural areas, and rural poverty rates remain stubbornly high in South Asia and Sub- Saharan Africa. Rural poverty reduction contributed more than 45 percent to overall poverty reduction in 1993­2002, with only a small share of that resulting from rural-urban migration. Rural-urban income gaps have narrowed in most regions except Asia, where the widening gap is a source of political tensions and a motive for new efforts to stimulate agricultural and rural development. P overty is concentrated in rural areas: other countries the declines in rural poverty growth. But from what is now known, it With an international poverty line of were unrelated to agriculture, such as in El appears very likely that the majority of the $1.08 a day, 75 percent of the develop- Salvador and Nepal, where rural poverty fell world's poor will still be in rural areas for ing world's poor live in rural areas whereas largely because of rising nonfarm incomes several decades. only 58 percent of its population is rural. and remittances.1 Poverty rates in rural areas have declined The urban population share for the The rural-urban income divide in the past 10 years, but remain extremely developing world is expected to reach 60 high (figure A.1). They declined from 37 per- percent by 2030.2 At that rate, the urban is large and rising in most cent in 1993 to 30 percent in 2002 for the share of $1.08-a-day poverty--now 25 per- transforming economies developing world as a whole, using a $1.08- cent--will reach 39 percent by 2030.3 These In almost all parts of the world, rural pov- a-day poverty line (box A.1). Outside China, projections are approximations because erty rates are higher than urban ones, and though, the results are less impressive, with a the pace of urbanization will depend on the depth of poverty is usually greater. In decline from 35 percent to 32 percent. The the extent and pattern of future economic 2002, the poverty rate for rural areas in number of poor people in rural areas fell only slightly, from 1 billion to 0.9 billion. With a Figure A.1 Rural poverty rates and number of rural poor ($1.08-a-day poverty line) higher poverty line ($2.15 a day), the poverty Rural poverty rate, % Rural poor, millions World rural poor, millions rates declined from 78 percent to 70 percent, 60 600 1,200 and the number of poor people slightly declined from 2.2 billion to 2.1 billion. World (right axis) 50 500 1,000 These global trends hide large variations in the evolution of poverty across regions 40 400 800 and countries. Rural poverty rates remain frustratingly high and tenacious in South 30 300 600 Asia (40 percent in 2002) and Sub-Saharan Africa (51 percent), and the absolute num- 20 200 400 ber of poor in these regions has increased since 1993. 10 100 200 Many countries that had fairly high agri- cultural growth rates saw substantial reduc- 0 0 0 tions in rural poverty: Vietnam, with land 1993 1996 1999 2002 1993 1996 1999 2002 reform and trade and price liberalization; Sub-Saharan Africa Latin America & Caribbean Moldova, with land distribution; Bangla- South Asia Europe & Central Asia desh, with rising farm and rural nonfarm East Asia & Pacific Middle East & North Africa earnings and lower rice prices resulting World from new technologies; and Uganda, with Source: Ravallion, Chen, and Sangraula 2007. economic reforms and a resulting boom in coffee production. Agriculture was also the key to China's massive and unprecedented B O X A . 1 A reestimation of international poverty levels reduction in rural poverty and to India's slower but still substantial long-term decline World rural and urban poverty statistics world are unsurprisingly higher than previ- (boxes A.2 and A.3). Ghana is Sub-Saharan for the period 1993 to 2002 have been con- ous estimates, by about 10 percent. With Africa's breaking story of poverty reduction sistently estimated for the first time by a the $1.08 poverty line, aggregate poverty over 15 years, with a decline in rural poverty World Bank team (Ravallion, Chen, and San- in 2002 is now estimated at 1,183 million, as the largest contributor (box A.4). graula 2007). The methodology includes compared with the 1,067 million previously But in some countries rural poverty did an adjustment of the poverty lines of $1.08 reported. The recent decline in aggregate and $2.15 a day, in 1993 purchasing power poverty, published in other World Bank not decline, despite agricultural growth: for parity (PPP), for the higher cost of living documents, is not reflected in this Report example, Bolivia and Brazil's agricultural in urban areas. With this adjustment, the because the rural-urban disaggregation is growth concentrated in a dynamic export- new estimates for aggregate poverty in the not available beyond 2002. oriented sector of very large farms. And in 46 WORLD DEVELOPMENT REPORT 2008 B O X A . 2 China's unprecedented reduction in rural poverty China's poverty reduction in the past 25 years varieties (Lin 1992). Rural incomes rose by 15 imbalances. Restrictions on internal labor is unprecedented. Estimates by Ravallion and percent a year between 1978 and 1984 (Von migration, industrial policies that favored Chen (2007) indicate that poverty fell from Braun, Gulati, and Fan 2005), and the bulk of China's coastal areas over the poorer inland 53 percent in 1981 to 8 percent in 2001, pull- national poverty reduction between 1981 and regions, and service delivery biases that ing about 500 million people out of poverty. 1985 can be attributed to this set of agrarian allowed the Chinese rural education and health Rural poverty fell from 76 percent in 1980 to 12 reforms. systems to deteriorate are all examples of poli- percent in 2001, accounting for three-quarters The role of agricultural growth in poverty cies that contributed to disparities in regional of the total. The evolution of poverty has been reduction remained important in subsequent and sectoral economic performances. very uneven over time, however. The sharpest years, as the reforms created the rural nonfarm reduction was in the early 1980s, with some sector, which provided employment and income Urban and rural poverty in China reversal in the late 1980s and early 1990s. to millions of people whose work was no longer needed on farms. The share of the rural nonfarm Poverty rate, % The role of institutional change 70 sector in GDP went from close to zero in 1952 to in poverty reduction more than one-third in 2004 (Von Braun, Gulati, The sharp decline in poverty from 1981 to 60 and Fan 2005). Considering the entire period, 1985 was spurred by agricultural reforms that 50 Ravallion and Chen (2007) concluded that started in 1978. The household responsibility growth in agriculture did more to reduce pov- system, which assigned strong user rights for 40 erty than did either industry or services. individual plots of land to rural households, 30 the increase in government procurement Rising inequalities Rural prices, and a partial price liberalization all Higher incomes for large parts of the popula- 20 National had strong positive effects on incentives for tion came at the cost of higher inequality. 10 Urban individual farmers. In the initial years of the Unlike most developing countries, China has reforms agricultural production and produc- higher relative income inequality in rural areas 0 1981 1984 1987 1990 1993 1996 1999 2002 tivity increased dramatically, in part through than in urban areas (Ravallion and Chen 2007). farmers' adoption of high-yielding hybrid rice There are also large regional and sectoral Source: Ravallion and Chen 2007. B O X A . 3 Reducing rural poverty in India The role of technological change relative terms. The rapid trade liberalization of inequality. But despite impressive growth in poverty reduction the 1990s had sharply differentiated regional and poverty reduction in the 1990s, the In the 1960s and 1970s the introduction of impacts. Rural districts with a higher concen- picture of overall welfare gains is nuanced, semidwarf varieties of wheat and rice--in tration of industries hurt by liberalization had because health outcomes have not improved. the green revolution--led to dramatic leaps slower progress in reducing the incidence and India's recent reforms, unlike China's, were in agricultural production and raised farm- depth of poverty because of the extremely not directed at agriculture. Today, there is a ers' incomes, especially in northwest India. limited mobility of labor across regions and renewed policy focus on agriculture in India, Rural poverty fell from 64 percent in 1967 industries. because many believe that the full poverty to 50 percent in 1977 and to 34 percent in Urban incomes and expenditures also reduction potential of agriculture in India has 1986. A large share of the gains came from increased faster than did rural incomes, yet to be unleashed. an increase in real wages and a decline in resulting in a steady increase in the ratio Sources: World Bank 2000b; Burgess and Pande grain prices. Growth in the agricultural sec- of urban-to-rural mean real consumption 2005; Chaudhuri and Ravallion 2006; Von Braun, tor reduced poverty in both urban and rural from just below 1.4 in 1983 to about 1.7 in Gulati, and Fan 2005; Topalova 2005; Ravallion and areas. This was true also of growth in services. 2000. Even then, India had fairly low income Datt 1996; Datt and Ravallion 1998a. But industrial growth did not reduce poverty. Land reform, rural credit, and education poli- Urban and rural poverty in India cies also played a role in the 1970s and 1980s, even if these programs might have cost some Poverty rate, % economic growth. 70 Beginning in 1991 India instituted sweep- 60 ing macroeconomic and trade reforms that spurred impressive growth in manufacturing 50 Rural and especially in services. Poverty data for 40 2004, comparable to the 1993 figures, show a continuing decline in poverty rates. 30 Green Urban Revolution Diverging patterns and a mixed picture 20 Period of rural welfare 10 Although there is a consistent poverty-reduc- ing pattern across almost all Indian states, 0 growth has been uneven. From 1980 to 2004 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 initially poorer states grew more slowly, result- Source: World Bank 2000b; 2007 National Sample Survey (NSS), Government of India. ing in income divergence in both absolute and Note: Poverty rates based on NSS data and the official poverty line. Declining rural poverty has been a key factor in aggregate poverty reduction 47 B O X A . 4 Ghana: African success in poverty reduction Ghana's growth and poverty reduction over modestly at 1 percent. Since 2001 a significant Urban and rural poverty in Ghana the past 15 years is Africa's new and important part of productivity gains has been in cocoa. Poverty rate, % success story. Real GDP has grown at more Cocoa production, although accounting for 80 than 4 percent a year since 1980 and at more only 10 percent of total crop and livestock than 5 percent since 2001. The poverty rate production values, contributed about 30 Rural savannah fell from 51.7 percent in 1991­92, to 39.5 in percent of agricultural growth. Ghana has 60 1998­99, and 28.5 in 2005­06. Poverty fell by also enjoyed strong growth in horticulture All rural about 17 points in the urban areas, and by 24 (almost 9 percent of total exports in 2006) in rural. If all rural-urban migrants are assumed driven mostly by pineapples. Both cocoa and 40 National to be poor, an estimated 59 percent of the pineapples are smallholder-based, and the Rural forest total poverty reduction was due to declining poverty reduction associated with recent 20 rural poverty. But there has been an increase in growth appears particularly strong among Urban inequality (the Gini coefficient rose from 0.35 cash-crop growers. Even so, the resource and to 0.39 over the 15 years), particularly at the export base of the economy remains narrow, 0 regional level, with Accra and the forest areas and the economy highly vulnerable to exter- 1991­92 1998­99 2005­06 experiencing more poverty reduction than has nal shocks. the rural savannah in the north. Ghana is one of the few Sub-Saharan Source: Coulombe and Wodon 2007. Ghana's accelerated growth is a result of Africa countries to register a sustained posi- better economic policy and a better invest- tive growth in per capita food production and ment climate as well as high commodity prices. declining food prices since 1990. But there is total factor and labor productivity and grow- In 2001­05 agriculture outperformed the ser- evidence of environmental degradation and ing fertilizer use over the past 10 years are vice sector, growing at 5.7 percent a year, faster unsustainable natural resource use. Food crop positive indicators of such a process. than overall GDP at 5.2 percent. and livestock production needs to intensify Agricultural growth has been mainly due to sustain current rates of agricultural growth Sources: Bogetic and others 2007; Coulombe and to area expansion, with yields increasing and to benefit more of the population. Rising Wodon 2007; Jackson and Acharya 2007. developing countries (30 percent) was more income in half the countries. Differences Why the poverty decline in rural than twice that for urban areas (13 percent), have been increasing in many countries. This areas--rural development or using the $1.08-a-day poverty line.4 Though increase is most notable in rapidly trans- migration? the gap has been closing in many parts of forming Asia (figure A.2). In India, rural and Higher urban incomes have pulled rural- the world, it has opened dramatically in East urban incomes were fairly similar in 1951, urban migration flows. But to what extent Asia and remained stable in South Asia. but the gap has since widened substantially are observed reductions in rural poverty Differences in income between rural and (box A.3). In China, the gap between rural caused by migration or by a genuine decline urban areas illustrate the rural-urban dispar- and urban incomes narrowed in the early in poverty among the nonmigrants who stay ity problem. In a sample of almost 70 coun- reform years, when rapid agricultural growth in rural areas? The answer depends on the tries, the median urban income (consump- drove overall economic growth, but it has pattern of migration. tion) is at least 80 percent higher than rural since opened again (box A.2).5 If migration is poverty-neutral--that is, the poor and nonpoor migrate at the same rate--the genuine decline in poverty of rural residents is equal to the observed decline in Figure A.2 The urban-rural income disparity has increased in most of the transforming countries the rural poverty rate. But if the nonpoor Ratio of urban to rural median income are more likely to migrate--as documented 4 for many countries--the reduction in pov- Initial year End year erty among nonmigrants is higher than the observed decline in poverty. If all migrants 3 are assumed to be poor, that sets a lower bound for the genuine reduction of poverty 2 in rural areas.6 If all those who migrate are poor, 81 per- cent of the reduction in rural poverty (6.9 1 percentage points of an 8.5 percentage point reduction) is still due to reduction of pov- 0 erty among rural residents, not to migration Guatemala Thailand Cambodia China Vietnam Indonesia Bangladesh India Pakistan (table A.1). Indeed, almost all the decline in 1989­2002 1990­2002 1997­2004 1985­2001 1992­2001 1993­2002 1991­2000 1989­99 1999­2001 South Asia and East Asia is because of a gen- Source: WDR 2008 team, from nationally representative household surveys. uine decline in poverty in rural areas. Even 48 WORLD DEVELOPMENT REPORT 2008 Table A.1 Even assuming that all migrants are poor, most poverty reduction in rural areas is due to tion gives a lower bound for the genuine declining poverty among rural residents reduction of aggregate poverty achieved in Rural poverty rate Change in rural poverty rate rural areas. ($2.15-a-day poverty line) for nonmigrants A lower bound for the contribution of Poverty- the rural sector to the decline in overall pov- neutral All migrants erty is 45 percent, and a more likely contri- Region 1993 2002 migration poor bution is more than 55 percent (table A.2). Sub-Saharan Africa 85.2 82.5 ­2.8 ­1.5 Outside China, the contribution of rural South Asia 87.6 86.8 ­0.8 ­0.4 areas is likely to be 80 percent (certainly not India 91.5 88.6 ­2.9 ­2.7 less than 52 percent), and in Sub-Saharan East Asia Pacific 85.1 63.2 ­21.9 ­20.0 Africa more than 80 percent. Rural develop- China 88.6 65.1 ­23.6 ­22.1 ment is thus essential to reduce poverty and Middle East and North Africa 35.8 37.6 1.9 6.1 achieve the Millennium Development Goal Europe and Central Asia 19.8 18.7 ­1.1 ­0.3 of halving the aggregate poverty rate. Latin America and Caribbean 47.3 46.4 ­0.9 7.8 Total 78.2 69.7 ­8.5 ­6.9 Within-country heterogeneity: Less China 73.7 71.3 ­2.4 ­1.6 less favored areas and poverty Source: WDR calculations, based on data in Ravallion, Chen, and Sangraula 2007. Note: Poverty rates are estimated using the 1993 $2.15-a-day poverty line. Beyond the rural-urban income divide, within-country heterogeneity in poverty when China is excluded from the sample, into the rural contribution, the urban con- across rural areas is a significant concern in 67 percent of the reduction in rural poverty tribution, and a population shift component many countries. It is commonly stated that is from causes other than migration. Note, (table A.2).7 In the first decomposition, the agricultural and rural investments should however, that this decomposition is an rural contribution is the reduction in the be directed to less favored areas because accounting exercise and thus does not speak rural poverty rate applied to the rural popu- poor people are concentrated there. Others to the indirect ways in which migration and lation in 2002. The urban contribution is the dispute this.8 Recent advances in geographic urban growth contribute to rural poverty reduction in the urban poverty rate applied information systems provide new opportu- reduction (such as remittances). to the 2002 urban population (the urban nities to answer basic questions about the population of 1993 plus the migrants). And spatial distribution of rural poverty in rela- the rural-urban migration contribution is tion to agriculture. Methods to estimate Rural areas contribute the poverty reduction corresponding to the welfare at the level of small communities, to a large share of the decline transition of migrants from the rural to the often referred to as "poverty mapping," pro- in national poverty urban poverty rate. vide basic information on the location of What, then, is the contribution of rural pov- A second specification assumes that all the poor. This information can be overlaid erty reduction to overall poverty reduction? migrants are poor. By attributing maximum with geographic information on agroeco- There are two ways to decompose aggregate contribution of migration to the reduction logical conditions and market access, such change in poverty between 1993 and 2002 of poverty in rural areas, this decomposi- as reported in chapter 2. Table A.2 Contribution of the rural sector to the aggregate poverty change Aggregate poverty rate Contribution of rural sector ($2.15-a-day poverty line) to aggregate poverty change change Poverty-neutral Region 1993 2002 1993­2002 migration All migrants poor Sub-Saharan Africa 79.8 77.5 ­2.2 81.1 44.6 South Asia 85.1 83.4 ­1.7 32.8 17.4 India 89.1 85.6 ­3.5 60.7 56.0 East Asia Pacific 70.6 45.6 ­25.0 53.4 48.8 China 72.8 44.6 ­28.3 52.0 48.8 Middle East and North Africa 23.5 23.5 0.1 n.a. n.a. Europe and Central Asia 16.6 13.6 ­3.0 14.1 3.5 Latin America and Caribbean 29.6 31.7 2.1 ­10.3 88.1 Total 63.3 54.4 ­8.8 55.5 45.1 Less China 59.6 57.9 ­1.8 78.8 52.4 Source: WDR calculations, based on data in Ravallion, Chen, and Sangraula 2007. Note: Poverty rates are estimated using the 1993 $2.15-a-day poverty line. n.a. = not applicable. Declining rural poverty has been a key factor in aggregate poverty reduction 49 Analyses for Brazil, Ecuador, Thailand, and medium agricultural potential areas, in Figure A.3 Incidence of poverty and Malawi, and Vietnam show that poverty Thailand and Cambodia more than 70 to geographic characteristics, Brazil rates tend to be higher in remote areas 80 percent live in good agricultural poten- Poverty rate, % than in more accessible areas (figure A.3). tial areas. 50 Poverty is also deeper and more severe in Where poverty incidence does not coin- remote areas. But at the level of disaggre- cide with poverty density, there are impor- 46 gation used for poverty, there is no general tant tradeoffs in the regional targeting of relationship between poverty rates and agri- policy interventions. The greatest impact on cultural potential.9 42 poverty may be through fostering growth The spatial patterns in the numbers of in more favored regions where most poor poor people (poverty density) are strikingly people live, especially growth that gener- 38 Good access Poor access different from those for poverty rates (pov- ates incomes for smallholders and creates erty incidence). In all the countries studied employment. Yet the extreme poor in more Share of rural poor, % the majority of the rural poor live in locali- marginal areas are especially vulnerable, and 30 ties with good access, as seen in Brazil (figure until migration provides alternative oppor- A.3).10 This is largely because less favored tunities, the challenge is to improve the sta- 20 areas are typically less densely populated bility and resilience of livelihoods in these than are favorable areas. In Brazil, for exam- regions. One concern with marginal areas 10 ple, 83 percent of the rural population lives is the possible existence of geographic pov- within two hours of a large city. By contrast, erty traps. Evidence of such traps has been 0 there is no clear pattern among countries shown for China, for example.11 In such a Good access Poor access for the distribution of the poor population case, reducing rural poverty requires either High Medium Low potential and agricultural potential. Whereas in Brazil a large-scale regional approach or assisting more poor people (75 percent) live in low the exit of populations. Source: Buys and others 2007. Note: High agropotential areas are those with very high or high agropotential under the GAEZ (Global Agro Ecological Zones) classification, which accounts for climate, soil, terrain, and land use, created by the Food and Agriculture Organization and IIASA in 2000. Medium agropotential areas are those with a medium or moderate GAEZ classification. Low agropotential areas are those classified as low, marginal, or very marginal in the GAEZ classification. Good access is defined as travel time of not more than two hours to the nearest city with a population of 100,000 or more. The share of the poor population is calculated for rural areas at the district level. Agriculture's performance, diversity, and uncertainties 2 A big question at the time of the last World productivity and income growth in the Development Report on agriculture, in 1982, face of declining prices for grains and tra- was whether agriculture would be able to ditional tropical exports. Rising demand c h a p t e r provide enough food for the world's grow- for high-value horticulture and livestock ing population. Twenty-five years later it in these rapidly growing economies offers is clear that world agriculture has met the farmers opportunities to diversify into new global demand for food and fiber. Increasing markets. per capita production, rising productivity, This chapter highlights emerging trends, and declining commodity prices all attest opportunities, and constraints that will drive to this success. But adequate global supplies future agricultural performance in response do not mean that countries or households to four challenges: the potential for a pro- have enough food--purchasing power mat- ductivity revolution in Sub-Saharan Africa, ters more than availability (see focus C). options for less-favored areas, diversification And the future world supply of food may in favored areas, and global uncertainties. be uncertain: increasing resource scarcity, The considerable diversity of agricultural heightened risks from climate change, higher production conditions underlines the com- energy prices, demand for biofuels, and plexity of these challenges. doubts about the speed of technical progress all have implications for future agricultural Productivity growth in performance. developing countries drove In addition, improved agricultural per- agriculture's global success formance has not been uniform through- out the world. Improvements have yet to Agriculture's performance has been impres- stimulate enough growth in agriculture- sive. From 1980 to 2004, the gross domes- based countries, especially in Sub-Saharan tic product (GDP) of agriculture expanded Africa, to allow them to achieve a sustained globally by an average of 2.0 percent a year, structural transformation (chapter 1). more than the population growth of 1.6 per- Environmental costs have often been high, cent a year. This growth, driven by increas- compromising the sustainability of future ing productivity, pushed down the real price production and affecting natural ecosys- of grains in world markets by about 1.8 per- tems and human health. cent a year over the same period. Poor agricultural performance in some areas relates to difficult agroclimatic condi- Developing countries have led tions or low investments in infrastructure agricultural growth that constrain market access. The agricul- Developing countries achieved much faster tural challenge in these less-favored areas agricultural growth (2.6 percent a year) is to sustainably intensify production in than industrial countries (0.9 percent a diverse farming systems, while improving year) in 1980­2004. Indeed, developing infrastructure and markets. countries accounted for an impressive 79 In the high-potential areas that have percent of overall agricultural growth dur- led the global increase in food production, ing this period. Their share of world agri- especially the transforming countries of cultural GDP rose from 56 percent in 1980 Asia, the challenge is different: sustaining to 65 percent in 2004. By contrast, they 50 Agriculture's performance, diversity, and uncertainties 51 accounted for only 21 percent of nonagri- continued to expand through the 1990s and cultural GDP in 2004.1 into the 2000s. Today 39 percent of the crop The transforming economies in Asia area in South Asia is irrigated, 29 percent in accounted for two-thirds of the develop- East Asia and the Pacific, but only 4 percent ing world's agricultural growth.2 The major in Sub-Saharan Africa. contributor to growth in Asia and the devel- Modern crop varieties of cereals began oping world in general was productivity to be widely adopted in the 1960s. The area gains rather than expansion of land devoted devoted to improved varieties has con- to agriculture. Cereal yields in East Asia tinued to expand, and by 2000 they were rose by an impressive 2.8 percent a year in sown on about 80 percent of the cereal area 1961­2004, much more than the 1.8 percent in South and East Asia, up from less than growth in industrial countries (figure 2.1). 10 percent in 1970. After a late start, Sub- Due to rising productivity, prices have been Saharan Africa is also expanding the use of declining for cereals--especially for rice, improved cereal varieties, which covered 22 the developing world's major food staple-- percent of the cereal area there in 2000.3 and for traditional developing-world export Chemical fertilizer use has also expanded products, such as cotton and coffee. significantly in most of the developing world, except Sub-Saharan Africa. The Better technology and better policy developing-country share of global fertil- have been major sources of growth izer use has risen from about 10 percent in Since the 1960s, rising cereal yields have the 1960s to more than 60 percent today. been driven by widespread use of irrigation, Asian farmers are the major users, with improved crop varieties, and fertilizer (fig- use up sharply from an annual average of ure 2.2). Although crop improvements have 6 kilograms per hectare in 1961­63 to 143 extended well beyond the irrigated areas to kilograms per hectare in 2000­02,4 more embrace huge areas of rainfed agriculture, than in developed countries. Higher fertil- Sub-Saharan Africa has not participated in izer use accounted for at least 20 percent of this agricultural success. the growth in developing-country agricul- For millennia Asian agriculture has ture (excluding dryland agriculture) over been intensified through irrigation, which the past three decades.5 Figure 2.1 Cereal yields rose, except in Sub-Saharan Africa Yields, tons per hectare Yields, tons per hectare 5 5 4 4 Developed countries 3 3 3 2 2 Developing countries 1 1 0 0 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 East Asia & Pacific Europe & Central Asia Latin America & Middle East & North Caribbean Africa South Asia Sub-Saharan Africa Source: FAO 2006a. 52 WORLD DEVELOPMENT REPORT 2008 Figure 2.2 Modern inputs have expanded rapidly but have lagged in Sub-Saharan Africa Irrigation Sub-Saharan Africa 4 South Asia 39 East Asia & Pacific 29 Middle East & North Africa 33 Europe & Central Asia 11 Latin America & Caribbean 11 0 10 20 30 40 Arable and permanent cropland, % 1962 1982 2002 Improved varieties of cereals Sub-Saharan Africa 24 South Asia 77 East Asia & Pacific 85 Middle East & North Africa 48 Latin America & Caribbean 59 0 20 40 60 80 Cereal area, % 1980 2000 Fertilizer use Sub-Saharan Africa 13 South Asia 98 East Asia & Pacific 190 Middle East & North Africa 73 Europe & Central Asia 34 Latin America & Caribbean 81 0 20 40 60 80 100 120 140 160 180 200 Kg of nutrients per hectare of arable and permanent cropland 1962 1982 2002 Sources: Evenson and Gollin 2003; FAO 2006a. Livestock expansion has also contrib- roads, and human capital from the 1960s, uted to the high agricultural growth rates. combined with better policies and institu- Livestock is one of the fastest growing sub- tions, were the major drivers that made the sectors in developing countries, where it agricultural productivity gains possible.9 already accounts for a third of agricultural Decompositions of productivity gains GDP.6 Production of meat has doubled over consistently point to investment in research the last 15 years, led by a 7 percent annual and development (R&D) as major sources increase in poultry production. of growth.10 Hybrid rice alone is esti- The combination of these breakthroughs mated to have contributed half of the rice produced steady growth in total factor pro- yield gains in China from 1975 to 1990.11 ductivity (TFP), especially in Asia at 1­2 Improved varieties contributed 53 per- percent a year.7 TFP growth was respon- cent of total factor productivity gains in sible for half of output growth after 1960 in the Pakistan Punjab from 1971 to 1994. China and India, and 30­40 percent of the Even in Sub-Saharan Africa, the impact of increased output in Indonesia and Thai- R&D has been identified as important in land, greatly reducing pressure on increas- its (limited) productivity growth.12 Infra- ingly scarce land.8 Investments in science, structure, especially roads, has also been Agriculture's performance, diversity, and uncertainties 53 an important factor in agricultural growth In most cases, countries with high in Asia. In India, investments in rural growth rates of agricultural value added roads contributed about 25 percent of the per capita of agricultural population--such growth in agricultural output in the 1970s, as China (3.5 percent annual growth rate), with high payoffs.13 Investments in human Malaysia (3.1 percent), and Vietnam (2.4 capital--improved education, health, and percent)--were also good performers in nutrition--have repeatedly been shown rural poverty reduction (see focus A). But to increase aggregate productivity.14 One Brazil (5.3 percent annual growth rate) and study for Sub-Saharan Africa found a sig- Pakistan (2.4 percent) have been less suc- nificant positive impact of calorie availabil- cessful in reducing poverty, mainly because ity on agricultural productivity, providing of the highly unequal ownership of and evidence of the interdependence of malnu- access to productive assets such as land and trition, hunger, and agricultural growth.15 irrigation water.18 Policy and institutional changes are also The distinguishing feature of Sub- likely to have been major sources of pro- Saharan growth is the high variability ductivity growth, although few studies have among countries and over time. Over the explicitly quantified the impacts. One such past 25 years, only Nigeria, Mozambique, study is the well-documented impact of the Sudan, and South Africa maintained agricul- household responsibility system in China, tural growth rates per capita of agricultural in which institutional and policy reform population above 2 percent a year, while was the dominant factor promoting agri- seven countries had rates below 1 percent a cultural growth and reducing rural poverty year and another six countries had negative during 1978­84.16 per capita growth. Many countries had sig- Despite this progress, long-term produc- nificant periods of negative growth associ- tivity growth could have been higher and ated with conflicts or economic crises. ecosystem and health costs reduced if the The growth rate of agricultural GDP per environmental costs of modern technology capita of agricultural population for the had been avoided. As much as a third of the region was close to zero during the early productivity gains from technical progress 1970s and negative through the 1980s and in China and Pakistan have been negated early 1990s. But with positive growth rates by soil and water degradation, and this does in the last 10 years, this trend has been not include the offsite pollution costs.17 reversed, suggesting that the stagnation in Growth across regions and Figure 2.3 Growth in agricultural GDP per countries has been uneven agricultural population is lowest in Sub-Saharan The progress in agricultural growth in Africa developing countries has been dominated Annual growth rate 1980­2004, % by the significant gains in Asia, especially in 4 China. Growth in Sub-Saharan Africa has averaged nearly 3 percent over the past 25 3 years, close to the average for all developing countries. But the growth per capita of agri- 2 cultural population in Sub-Saharan Africa (a crude measure of agricultural income) 1 has been only 0.9 percent, less than half that in any other region and well below the star performer, East Asia and the Pacific, at 3.1 0 Sub- South East Middle Latin percent. Latin America had lower agricul- Saharan Asia Asia East America & tural growth than Sub-Saharan Africa, but Africa & Pacific & North Caribbean with Latin America's declining agricultural Africa population, the growth per capita of agri- Agricultural GDP Agricultural GDP/agricultural population cultural population has averaged a healthy 2.8 percent a year (figure 2.3). Sources: FAO 2006a; World Bank 2006y. 54 WORLD DEVELOPMENT REPORT 2008 Figure 2.4 Stagnation in Sub-Saharan African factors can now be readily quantified and agriculture may be over mapped against agricultural areas and (Growth in agricultural GDP per capita of agricultural population in Sub-Saharan Africa) populations using geographical informa- tion systems. Annual growth rate, % 8.0 Both agroecological conditions and 6.0 market access matter Agricultural potential, especially that of 4.0 rainfed agriculture, is highly sensitive to 2.0 soil quality, temperature, and rainfall. Two-thirds (1.8 billion) of the developing 0.0 world's rural population lives in areas with favorable agroecological potential--that is, ­2.0 irrigated areas (42 percent of the rural pop- ­4.0 ulation) or humid and semihumid rainfed areas with reliable moisture (26 percent of ­6.0 the rural population) (map 2.1 and figure ­8.0 2.5).19 But one-third (820 million people) 1968 1972 1976 1980 1988 1984 1992 1996 2000 2004 live in less favored rainfed regions, char- acterized by frequent moisture stress that Sources: FAO 2006a; World Bank 2006y. Note: Line is for data smoothed by locally weighted regressions. limits agricultural production (arid and semiarid areas of map 2.1). Although these less-favored areas account for 54 percent of the agricultural area (45 percent of the Sub-Saharan African agriculture may be cropped area), they produce only 30 per- over (figure 2.4). Improvements in agri- cent of the total value of agricultural pro- cultural performance coincide with better duction. Latin America, the Middle East macroeconomic policies and higher com- and North Africa, and Sub-Saharan Africa modity prices (chapter 1). But food produc- all have fairly high shares of rural popula- tion is still lagging (box 2.1). tion in these moisture-stressed areas. Another characteristic of Sub-Saharan Performance also relates to access to Africa is the generally poor yields of food markets and services. Rural areas by defi- staples, even in the most recent period. The nition are spatially dispersed, which affects green revolution breakthrough in cereal the costs of transport, the quality of public yields that jump-started Asia's agricultural services, and the reliance on subsistence and overall economic growth in the 1960s production. In developing countries 16 and 1970s has not reached Sub-Saharan percent of the rural population (439 mil- Africa, where the adoption of productivity- lion people) lives in areas with poor mar- enhancing inputs has been low (figure 2.2). ket access, requiring five or more hours to There are many reasons for this: depen- reach a market town of 5,000 or more (map dence on rainfed agriculture, diverse food 2.2). About half the agricultural area in crops, poor infrastructure, policy discrimi- these remote regions has good agricultural nation against agriculture, and low invest- potential but lacks the infrastructure to ment (box 2.1). integrate into the wider economy. In Sub- Differences in performance Saharan Africa and the Middle East and North Africa, the percentage of rural pop- reflect different underlying ulation with poor market access is much conditions higher, more than 30 percent (figure 2.5). The different performances of countries In South Asia, only 5 percent live in remote and regions in part reflect the huge diversity areas, and 17 percent in East Asia and the of agricultural production systems--their Pacific. Poor market access reflects low agroclimatic potential, their population investments in infrastructure, often due to density, their infrastructure. Many of these low population density (box 2.2). Agriculture's performance, diversity, and uncertainties 55 B O X 2 . 1 The green revolution in food staples that didn't happen: Sub-Saharan Africa's variegated palette The expansion of food production has taken roads in Sub-Saharan Africa than there were kilograms of nitrogen-phosphorus-potassium quite different courses in Asia and in Sub- in Asia at the time of the green revolution. per hectare per year, five times the average Saharan Africa, where increases in food staples Sub-Saharan African countries are small, many application of 10 kilograms per hectare of were achieved largely by expanding the area of them landlocked, and barriers to trade are nutrients through chemical fertilizers.22 Soil cultivated, as shown in the figure below. relatively high because of high transport costs. nutrient mining is highest in areas of high pop- Population density--low? To some As already mentioned, Sub-Saharan African ulation density. For example, the estimated extent the extensification in Sub-Saharan investment in irrigation (4 percent of crop annual productivity loss in the Ethiopian high- Africa reflects differences with Asia in popula- area) is also only a fraction of that in Asia (34 lands from soil degradation is 2­3 percent of tion density and land availability. The popula- percent of crop area). agricultural GDP a year.23 Clearly the decline of tion density of 29 persons per square kilometer Geography and agroecology--diverse. soil fertility is a large part of the reason for Sub- in Sub-Saharan Africa is only one-tenth that in Other reasons for the differences in agricul- Saharan Africa's low yields, so reversing it must South Asia. Yet population densities in many tural productivity growth include Sub-Saharan be a high priority. areas of Sub-Saharan Africa have reached lev- Africa's intrinsically different agroecological Policies--historically distorted. To els at which growth through land expansion characteristics. The main green revolution reduce risks and increase profitability, Asia under rainfed conditions is no longer sustain- cereals in Asia were wheat and rice, largely provided credit, support prices, and input sub- able. When population density is adjusted for irrigated. Sub-Saharan Africa's diverse rainfed sidies to farmers. In Sub-Saharan Africa gov- land quality, densities in much of Sub-Saharan agroecologies use a wide range of farming ernments also intervened heavily in markets, Africa are similar to those in Asia. For example, systems and a broad number of staples (from but agriculture was taxed more than in other the land-quality-adjusted population density cassava in west and central Africa to millet regions--and it still is (chapter 4). Although in Kenya is estimated to be higher than that in and sorghum in the Sahel). What does such Kenya, Malawi, Zambia, and Zimbabwe initi- Bangladesh.20 heterogeneity in crop production and agro- ated maize-based revolutions using hybrid Infrastructure--undeveloped. Sub- ecological conditions mean? In Sub-Saharan seed and fertilizer, the programs have been Saharan Africa is massively disadvantaged in Africa improved varieties for many different difficult to sustain, due to high marketing infrastructure, increasing transaction costs crops will be needed to increase productivity. costs, fiscal drain, and frequent weather and market risks. In part due to low population Outside technologies often are not directly shocks. Macroeconomic policies and much densities, there are fewer and less-developed transferable, and Africa-specific technologies lower public investment in agriculture than in will be required to improve the region's agri- Asia have also reduced incentives to private cultural productivity (chapter 7). Yet the trend agents and limited supply of public goods Expansion of cereal production for R&D spending was stagnant in the 1990s. such as R&D and roads (chapter 1). has followed very different paths Fertilizer use--low. Largely because of Turning the corner? Recent evidence in Sub-Saharan Africa and Asia poorly developed markets, fertilizer use in Sub- suggests that Sub-Saharan Africa may be Yield (1961­65 = 100) Saharan Africa has stagnated at very low levels, turning the corner. There are many local suc- 260 one of the main reasons for the region's low cesses in food crop production, such as maize 2001­04 agricultural productivity relative to Asia. On in several West African countries, beans in average, Sub-Saharan African farmers must sell Eastern Africa, cassava in many countries, 220 about twice as much grain as Asian and Latin market-driven expansion of the use of fertilizer American farmers to purchase a kilogram of fer- on maize crops in Kenya, and many promising Asia tilizer, given its high price.21 Low volumes, high technological innovations in the early stages 180 prices, high transport costs, and undeveloped of adoption (chapter 7). The challenge is how private input markets are major barriers to fer- to achieve productivity gains in diverse rain- tilizer use in Sub-Saharan Africa (chapter 6). fed systems by coordinating investments in 140 Africa Soils--degraded. The combination of technology with investments in institutions 2001­04 shorter fallows, expansion to more fragile and infrastructure to promote development of land driven by rapid population growth, and input and output markets. 100 a lack of fertilizer use is degrading soils in 100 140 180 220 260 Sub-Saharan Africa. About 75 percent of the Area (1961­65 = 100) farmland is affected by severe mining of soil Sources: Borlaug and Dowswell 2007; Cummings 2005; Djurfeldt and others 2006; Harrigan 2003; nutrients. According to a recent report by the Source: FAO 2006a. InterAcademy Council 2004; Johnson, Hazell, Note: Each point represents a five-year average, International Fertilizer Development Center, and Gulati 2003; Mosley 2002; Sanchez 2002; starting with 1961­65 =100. the average rate of soil nutrient extraction is 52 Spencer 1994. Defining less-favored areas are those that are irrigated or have good The combination of agroclimatic potential rainfall and have medium to high access to and market access provides a working defini- markets. Sixty percent of the rural popula- tion of areas that are favored or less favored tion live in these areas. Less-favored areas for agriculture, at least for market-oriented are of two types--constrained by poor mar- production. In this Report, favored regions ket access, and limited by rainfall. Almost 56 WORLD DEVELOPMENT REPORT 2008 Map 2.1 Agroecological zones in agricultural areas Source: Sebastian 2007, based on GAEZ climate data from FAO/IIASA; GMIA irrigated area data from FAO; and cropping and pastureland data from Ramankutty/SAGE. Note: Agricultural areas include those with at least 10 percent irrigated, cultivated, or grazing lands. Map 2.2 Market access in agricultural areas of Africa, Asia, and Latin America Source: Sebastian 2007, based on market access data from A. Nelson, and extent of agriculture from IFPRI. Note: Agricultural areas include those with at least 10 percent irrigated, cultivated, or grazing lands. Data are not shown for Australia, Canada, Europe, and the United States. two-thirds of the Sub-Saharan rural popu- be considered, including the fragility of the lation are in less-favored areas with either or natural resource base (chapter 8) and social both low agricultural potential or poor mar- conditions. ket access, compared with only 25 percent These distinctions determine the choice for South Asia. Of course, many additional of farming systems and strategies. For elements of less-favored areas should also example, in Ethiopia a disproportionate Agriculture's performance, diversity, and uncertainties 57 Figure 2.5 There are big differences across regions in agricultural potential and access to markets Agricultural potential Sub-Saharan Africa South Asia East Asia & Pacific Middle East & North Africa Europe & Central Asia Latin America & Caribbean 0 20 40 60 80 100 Percentage of rural population Irrigated Humid/sub-humid Arid/semi-arid Market access Sub-Saharan Africa South Asia East Asia & Pacific Middle East & North Africa Latin America & Caribbean 0 20 40 60 80 100 Percentage of rural population Good (0­1 hour) Medium (2­4 hrs) Poor (5 hrs or more) Source: Sebastian 2007. share of vegetable production is in high- is to sequence cost-effective investments in access areas (63 percent of production, but areas that have low population density and only 38 percent of the rural population), little commercial activity. One option is to while cereals are concentrated in less- focus investments geographically to foster favored areas, whether defined by rainfall the development of growth poles. or by market access.24 Beyond infrastructure, agricultural These characteristics are not immutable. investments in new varieties to improve Investments can convert less-favored areas yield stability and in natural resource with low rainfall or poor roads into high- management can be effective in less- potential areas. The most common is irri- favored areas (chapter 8). Over the long gation, which has made some of the world's term, investments in human and social deserts bloom, transforming agricultural capital (education, health, and institutional systems and livelihoods. Likewise, invest- strengthening) to enhance income diversi- ment in transport infrastructure has allowed fication and out-migration may be the best Brazil's interior states to enter global mar- option for many areas (chapter 9). kets for soybeans and other crops. Although the conventional wisdom is that For much of Sub-Saharan Africa, poor most of the poor are in less-favored regions, market access is almost as important a overlapping maps of agroclimatic potential constraint (34 percent of the rural popu- and market access with poverty maps indi- lation) as rainfall (45 percent of the rural cate that this is not so (see focus A). Although population). In Ethiopia, 68 percent of the the poverty rate is often highest in more mar- rural population lives in medium- to high- ginal areas, the largest number of poor people rainfall areas, but farm households are on live in the more-favored areas. Lagging average 10 kilometers from the nearest road regions with high poverty rates are even and 18 kilometers from the nearest public found within countries with rapid economic transport. The challenge in such contexts growth (box 2.3). 58 WORLD DEVELOPMENT REPORT 2008 decline in prices of cereals and traditional B O X 2 . 2 Population density and the definitions exports. of "rural" Changes in consumer diets--brought about by rapid income growth and increas- Market access is closely related to popula- Co-operation and Development (OECD) ing urbanization--are already driving tion density. Worldwide there is enormous uses population density of 150 people per heterogeneity in population densities. In square kilometer to define rural. Applying diversification. Especially in the transform- India less than 1 percent of the population this definition to Brazil would increase ing and urbanized economies, dietary pat- live in areas with fewer than 50 people per its rural population to 25 percent. India terns are shifting away from cereals, roots, square kilometer, compared with 20 per- reports 72 percent of its population as tubers, and pulses to livestock products, cent in Brazil and 60 percent in Zambia (see rural, but the OECD definition would figure below). Zambia's population distribu- reduce that to only 9 percent. Even heavily vegetable oils, fruits, and vegetables (fig- tion is quite uneven, while Cambodia's is agricultural areas in India would not be ure 2.6). Consumer preferences in indus- fairly equal.25 This also means that national rural under the OECD definition. trial countries for specialty products and definitions of "rural" can have quite differ- Differences in population density year-round supplies of fresh produce create ent meanings in different countries. and distance to market towns imply very Rural areas can be defined by settle- different challenges for infrastructure, global markets for many of them. Horticul- ment size, population density, distance to service delivery, and rural development. ture, oilseeds, and livestock are expanding metropolitan areas, administrative divi- High population density makes it cheaper the fastest, with new markets also emerg- sion, and importance of the agricultural to provide public goods, such as roads. ing for feed grains, livestock, and biofuels. sector. Brazil uses administrative divisions Low population density increases the cost and reports 19 percent of its population of such investments but eases constraints Most food products in this new agriculture as rural. The Organisation for Economic of land resources. are perishable, and quality and safety stan- dards are tighter, thus increasing the verti- The distribution of population within a country varies widely cal integration of food systems. Share of population, % Share of population, % 100 100 The horticulture revolution Fruits and vegetables are one of the fastest 80 80 growing agricultural markets in developing countries, with production increasing by 3.6 60 60 percent a year for fruits and 5.5 percent for Zambia vegetables over 1980­2004.26 During this 40 40 period, 58 percent of the increase in world- Cambodia wide horticulture production came from 20 20 China, 38 percent from all other developing Brazil India countries, and the remaining 4 percent from 0 0 developed countries, suggesting that the 0 100 200 300 400 500 0 100 200 300 400 500 Population per sq. km Population per sq. km boom in horticulture is mainly benefiting developing countries. In India, fruits and Share below 150 per sq. km: Share of rural (UN): vegetables were the most important growth Brazil: 25% Brazil: 19% India: 9% India: 72% sector for crop production in the 1990s.27 Zambia: 65% Zambia: 65% The horticulture revolution boosts Cambodia: 31% Cambodia: 83% incomes and employment. Relative to cere- Sources: Center for International Earth Science Information Network (CIESIN) 2006; data on the als, horticulture increases the returns on share of rural population are from United Nations 2004. land about 10-fold. And it generates con- siderable employment through production (about twice the labor input per hectare of cereals) and more off-farm jobs in process- Opportunities for a new ing, packaging, and marketing (chapter agriculture through 9).28 Women hold many of these new jobs. diversification But horticulture also requires produc- Farmers in areas of good agricultural ers to adjust. It is management-intensive, potential and with access to markets-- with a variety of crops and heavy use of about 60 percent of the rural population in cash inputs and chemicals. It is risky, due the developing world--have good oppor- to both pest outbreaks and price volatility, tunities in new markets. By diversifying to and fruit production requires an invest- higher-value products, they can offset the ment of several years to recoup costs. It can Agriculture's performance, diversity, and uncertainties 59 B O X 2 . 3 Why are there lagging regions in countries with high agricultural growth? Even countries with strong overall agricultural access to land. Nearly two-thirds of its soils agricultural growth, which was largely concen- growth have lagging regions, where agricul- are not suitable for farming, a situation only trated in the irrigated coastal regions. tural productivity and household incomes are aggravated by centuries of use (particularly for India's Bihar: Meeting the challenges low. In many cases these regions have lower livestock) that degraded soils and limited their of governance in areas with high agricultural potential or poorer market access capacity to absorb rainfall. agricultural potential than other regions in the same country. But Peruvian Andes: Isolated areas have not Well endowed with fertile land and water lagging areas can also be the result of social participated in rapid agricultural growth resources, Bihar has the potential to achieve processes, with specific territories left aside by Recent economic growth in Peru has been productivity levels equivalent to the more- public policies or poor governance. The most driven by the mining and agricultural sectors, developed states of India.33 But the state's difficult regions are those that combine poor with annual growth rates of 7.9 percent and 3.8 agricultural performance lags seriously agroecological endowments, isolation, and percent, respectively, in 1997­2004. Growth behind the country's. Employing 80 percent social marginalization. in these sectors helps explain why rural areas of Bihar's workforce and generating nearly 40 Brazil's northeast: Low agricultural appear to have done better than urban ones in percent of its GDP, agriculture has performed potential next to a breadbasket reducing poverty after the 1998­99 economic particularly poorly, declining in the early Brazil's agricultural growth of 5.3 percent a crisis. But poverty reduction in rural areas has 1990s by 2 percent a year and growing by less year during 1990­2004 was led by agricultural been unequal across geographic regions. than 1 percent a year since 1995--half the exports from the south and center of the Rural poverty appears to be most respon- national average. country. Agricultural GDP growth there was sive to growth in the coastal regions (elasticity Bihar's agricultural sector has been impressive--Mato Grosso at 14.8 percent a between ­0.9 and ­1.3), and least responsive plagued by low productivity, slow diver- year, Goiás 6.8 percent, Paraná 6.7 percent, and in the sierra regions (elasticity between ­0.6 sification into higher-value crops, poorly Mato Grosso do Sul 5.3 percent. But this per- and ­0.9).31 This can be explained by the developed rural infrastructure, inadequate formance does not reflect the entire country. geography of the Andean region, which iso- investments to expand and maintain surface Alongside a rural Brazil that is a global leader lates towns from the rest of the economy. The irrigation systems, small and fragmented in several agricultural exports is another rural mountainous terrain increases the costs of farms with widespread illegal land tenancy, Brazil, with widespread poverty and depriva- road construction. In some areas it is necessary little transparency in product marketing, and tion affecting millions of people in semisubsis- to walk for several hours to get to a market inadequate public research and extension tence farming. town, health center, or public school. The dis- services. Bihar faces serious challenges to The northeast of Brazil has the country's tance to markets encourages subsistence farm- improve growth and strengthen the public highest rural poverty rates (76 percent) and ing using few purchased inputs, with about 20 administration, service delivery, and invest- the largest concentration of rural poor in Latin percent of agricultural production for personal ment climate. Government efforts to address America.29 States in the northeast were among consumption, labor exchanges characterized the needs of farmers and deliver support ser- the poorest agricultural performers in the by reciprocity, and poor opportunities for non- vices have had little success largely because country for 1990­2004, some with negative agricultural income despite the low productiv- of an unclear strategy, weak institutional agricultural growth rates (Ceará ­4.3 percent ity of the land. capacity, and little accountability, as well as a year, Rio Grande do Norte ­2.3 percent, and These isolated areas have the highest concerns about security and lawlessness. Sergipe ­0.5 percent).30 The northeast's pau- poverty rates in the country ($1-a-day poverty The cause of these problems: a semifeudal city of natural resources and climatic instability rates of more than 65 percent).32 Even though social structure divided by caste. Community (with droughts occurring on average every agricultural income represents more than 75 involvement and transfers of responsibility in five years) are accentuated by the fragile equi- percent of total income in the Andean areas, delivering agricultural technology and sur- librium of its ecosystems and highly unequal these regions did not benefit from recent face irrigation are enjoying some success.34 also inflict considerable harm to the envi- The livestock and ronment: horticulture crops account for 28 aquaculture revolutions percent of global pesticide consumption.35 The livestock and aquaculture revolutions The horticulture revolution, unlike the have been most notable in the transform- green revolution, has been driven largely by ing and urbanized countries of Asia and the private sector and the market. This has Latin America, driven by rising demand for implications for the organization of value poultry, pork, fish, and eggs with increasing chains, with specialized agribusinesses incomes. Beef and milk production have and supermarkets increasing their share in also risen steadily in rapidly growing coun- these markets, especially in the urbanized tries. In India the consumption of milk countries. Grades and standards make it nearly doubled between the early 1980s and more difficult for smallholders acting alone late 1990s.36 to participate in these markets, giving rise Livestock production is switching from to contract farming and collective action by extensive (grazing) to intensive (stall-fed producer organizations (chapter 5). poultry, pigs, and dairy cows), increasing 60 WORLD DEVELOPMENT REPORT 2008 Figure 2.6 Per capita food consumption in developing countries is shifting to fruits and providing more diversified diets. But inten- vegetables, meat, and oils sive production methods and the growing Annual growth rate of concentrations of animals near urban and kilocalories/person/day, % periurban areas of developing countries 4 can increase waste pollution and the inci- dence of diseases such as tuberculosis and 3 avian flu. The movement of live animals and aquatic products makes the accidental 2 spread of disease more likely. Globalization 1 may further widen the environmental foot- print from livestock (box 2.4) and aquacul- 0 ture, calling for policies to prevent irrevers- ible consequences (chapter 8). ­1 Cereals Sugar Oil crops Horticulture Meat Milk Diversifying through export markets 1961­75 1976­90 1991­2003 High-value products also make up a rapidly Source : FAO 2006a. growing share of international trade in agri- cultural products. Exports of horticulture, the demand for feed grains, including oil- livestock, fish, cut flowers, and organic prod- seeds. In developing countries, 28 percent ucts now make up 47 percent of all develop- of grain consumption was already used for ing-country exports, far more than the 21 feed in 2005. But the use of cereals for feed percent for traditional tropical products such is growing more slowly than the increase in as coffee, tea, and cotton (figure 2.7). Across meat production because other feedstuffs, a broad range of nontraditional export prod- such as oilseed meals and cassava, are sub- ucts, developing countries have been gaining stituted for cereal grains, and the share of market share--in 2004 they held 43 percent poultry in total meat production is grow- of the world trade in fruit and vegetables ing. (Poultry requires only 2­3 kilograms of (excluding bananas and citrus). feed per kilogram of meat, compared with Brazil, Chile, China, and Mexico domi- 10 kilograms for beef.)37 nate nontraditional agricultural export Aquaculture is the world's fastest grow- markets. But many countries, including ing food-production sector, increasing at some in Sub-Saharan Africa (Kenya, for an annual average rate of 10 percent since example), are now gaining shares in selected the mid-1980s. Aquaculture now represents product markets. The least-developed coun- more than 30 percent of total food-fish pro- tries have very limited participation--only duction.38 More than 90 percent of aqua- Niger is significant, with 2.6 percent of the culture production occurs in developing world's green bean exports by value39--but countries, and China alone accounts for there have been other recent successes, such 67 percent of global production. Aquacul- as cut flowers from Ethiopia. Despite the ture can provide an important source of expansion of nontraditional exports, prices livelihood for the rural poor, generating have held up well in real terms. Estimates income through direct sales of products of the elasticity of export revenues for non- and employment in fish production and traditional export products indicate there is services, especially in processing. In Asia, room for further market expansion.40 more than 12 million people are directly Even traditional export commodities employed in aquaculture. In Bangladesh provide opportunities for entering high- and Vietnam, more than 50 percent of value markets. The markets for premium workers in fish depots and processing plants quality goods such as coffee, organics, and are women, and although salaries are still Fair Trade products have grown consider- quite low, they are significantly higher than ably in the last decade, starting from a low wages from agricultural activities. base. The Fair Trade market is most devel- The livestock and aquaculture revolu- oped in Europe, less so in Japan and the tions are increasing the supply of protein and United States. But the market for organic Agriculture's performance, diversity, and uncertainties 61 produce has grown strongly in both Europe (retail sales of $10.5 billion in 2003) and the B O X 2 . 4 The global environmental footprint of United States ($12 billion).41 There is con- expanding livestock siderable scope for expanding exports to these emerging markets (chapter 5). During the decade 1994­2004, world trade area under soybean cultivation between in soybeans doubled. Seventy percent 1999/2000 and 2004/05 by planting an of the global increase in exports went to additional 54,000 square kilometers--an Biofuels--a revolution China, where total meat production rose area larger than Costa Rica--much of it in the making? from 45 million to 74 million tons over displacing ecologically important savanna the period, generating rapid expansion woodland (cerrado) and forest.45 The mean Biofuels could be the next revolution. Based in demand for feedgrains. Argentina and annual deforestation rate in the Amazon on maize, sugar, cassava, oil palm, and other Brazil responded rapidly to this market from 2000 to 2005 (22,392 km2 per year) crops, biofuels offer potentially major new opportunity, providing more than two- was 18 percent higher than in the previous thirds of the increased global exports of five years (19,018 km2 per year), partly the markets to agricultural producers. Some soybeans. result of agricultural expansion.46 Because countries have been aggressively encour- Rapid growth in exports from Argen- trees are being burned to create open aging biofuel production as oil prices have tina and Brazil has been supported by land in the frontier states of Pará, Mato risen and concerns over energy security bringing new land under cultivation, often Grosso, Acre, and Rondônia, Brazil has at the expense of forests and woodlands. become one of the world's largest emit- and the environment have increased. But In the northern Salta region of Argentina, ters of greenhouse gases. To mitigate the current economics, environmental issues, half the area under soybean cultivation in negative ecological impacts, an alliance and the prospects of alternative technolo- 2002/03 was previously covered by natural of private companies, nongovernmental gies and feedstocks make biofuels' future vegetation. Much of this area included organizations, and the government of the highly threatened Chaco ecosystem.44 Brazil signed a two-year moratorium on growth quite uncertain (see focus B). In Brazil the states of Goias, Mato Grosso, buying soybeans from newly deforested and Mato Grosso do Sul doubled the land in the Amazon.47 Future perspectives: confronting challenges and rising uncertainties Figure 2.7 High value exports are expanding rapidly in developing countries Even if agricultural and food systems have Billions of dollars (2000 prices) been globally successful over the past four 80 decades, can they meet the likely demand for food over the next 25 or 50 years? Can 60 they accommodate rapid urbanization and 40 changing diets, and will they do this in a sustainable and environmentally friendly 20 way? What are the main uncertainties that might compromise success? 0 Cereals Traditional Oilseeds, feeds, Livestock and Fruits and exports and others fish products vegetables A "business as usual" scenario 1980 1990 2000 2004 Projections of global future food supply and demand are always subject to wide mar- Source: United Nations Commodity Trade Statistics Database (COMTRADE). gins of error and generally influenced by Notes: Traditional exports include cocoa, tea, coffee, rubber, tobacco, sugar, cotton, and spices. prevailing market conditions: when prices are fairly high, as they are today, projec- tions of the impact of climate change and tions tend to be more "pessimistic."42 Both energy prices are especially difficult given the United Nations' Food and Agriculture current uncertainties--the IFPRI baseline Organization (FAO) and the International uses "medium" scenarios for both.48 Food Policy Research Institute (IFPRI) In the IFPRI models, the overall projec- have recently released "business as usual" tion is that global food consumption will projections to 2025­30 and 2050 that show increase more slowly in the future. Growth broadly consistent trends.43 Such projec- in cereal consumption will slow from 1.9 tions are inherently conservative; they percent annually in 1969 to 1999 to 1.3 assume no major changes in policies (such percent a year from 2000 to 2030; growth as trade) or policy responses to market in meat consumption will also slow from conditions (such as increased investment 2.9 percent a year to 1.7 percent annually in R&D induced by higher prices). Projec- (see figure 2.8).49 This slowdown reflects 62 WORLD DEVELOPMENT REPORT 2008 Figure 2.8 Slower growth in cereal and meat consumption is projected for the next 30 years Cereal consumption Meat consumption Annual growth rate, % Annual growth rate, % 8 8 7 1969­99 (actual) 7 1969­99 (actual) 6 2000­30 (projected) 6 2000­30 (projected) 5 5 4 4 3 3 2 2 1 1 0 0 ­1 ­1 Sub- South East Asia Middle East Europe & Latin Sub- South East Asia Middle East Europe & Latin Saharan Asia & Pacific & North Central America & Saharan Asia & Pacific & North Central America & Africa Africa Asia Caribbean Africa Africa Asia Caribbean Source: Rosegrant and others 2006b. a. Includes food, feed, and other uses. b. No data are available on meat consumption for Europe and Central Asia in 1969­99. two factors: an overall slowing of popula- The global projections hide widening tion growth to 1 percent a year (nearly all supply-demand imbalances in developing growth is in developing countries), and the countries. Net cereal imports by developing medium to high levels of food consumption countries of Asia, Africa, and Latin Amer- per capita already attained in some highly ica are projected to increase to 265 million populous developing countries (for exam- tons in 2030 from 85 million tons in 2000. ple, China). This reflects continuing high import depen- In developing countries overall, per cap- dence in the Middle East and North Africa ita consumption of cereals for food will fall and sharp increases in imports in Asia and slightly; together with continuing trends Sub-Saharan Africa (figure 2.9). in the efficiency of converting feed grain These trends greatly increase the impor- to meat, per capita cereal consumption for tance of developing countries in global food all uses in developing countries increases markets. The major exporting countries by only 0.1 percent a year. Slower demand are the developed countries and Brazil and growth leads to slower growth of cereal Argentina. Some countries in Europe and production in all regions. Meat consump- Central Asia are projected to become impor- tion also slows sharply, except in South Asia tant exporters. Only in Sub-Saharan Africa, and Sub-Saharan Africa, where meat con- with high transport costs and scarce foreign sumption will increase at a slightly faster exchange, is the growing import gap a con- rate, but from very low per capita consump- cern for food security. Again, the biggest tion levels. challenge is in Sub-Saharan Africa, where Despite the slowing growth in consump- even in 2030 the average per capita calorie tion, current projections reverse the long- consumption is expected to be around 2,500, term decline in cereal prices at 1.6 percent less than the 3,000+ in other regions. a year observed in previous decades. Cereal The assumptions underlying these pro- prices are projected to increase margin- jections show that supply constraints for ally at 0.26 percent a year to 2030 and to land, water, and energy; increased climate accelerate to 0.82 percent a year from 2030 variability and climate change; and persis- to 2050.50 The slight upward price trend for tent low investment levels in research pose cereals is a significant reversal from previ- formidable challenges in meeting future ous projections--land and water scarcity food demand. They suggest rising uncer- combined with slower technical progress tainty and the potential for larger and more (discussed below) explain this reversal. frequent shocks to global food prices. Agriculture's performance, diversity, and uncertainties 63 Looming land constraints Figure 2.9 Developing countries will become even bigger markets for cereals exported largely by developed countries Throughout most of history, agriculture grew by bringing more land under culti- Tons, millions 300 vation, driven by population growth and expanding markets. But in the more densely 2000 (actual) 200 2030 (projected) populated parts of the world, the land fron- tier has closed. In Asia land scarcity has 100 become acute in most countries, and rapid urbanization is reducing the area available 0 for agriculture.51 The urbanized countries of Latin ­100 America and Europe and Central Asia are ­200 relatively land-abundant because of lower Developed Europe & Latin Sub-Saharan South Middle East & East Asia population densities and a declining agri- countries Central Asia America & Africa Asia North Africa & Pacific Caribbean cultural population (see figure 2.10). In Latin America there is further scope for Source: Rosegrant and others 2006b. Notes: Negative values indicate net cereal imports, and positive values indicate net cereal exports. agricultural land expansion, driven by export markets, but this is often at the expense of cutting subtropical and tropical Even land now used for agriculture is forests and woodlands.52 In Sub-Saharan threatened. Productivity growth of avail- Africa high rural population growth drives able land is often undermined by pollu- expansion into forest or grazing land--cre- tion, salinization, and soil degradation ating conflicts with traditional users--or from poorly managed intensification, all into areas subject to human and animal dis- reducing potential yields (chapter 8). Some eases. Even so, there is considerable room sources suggest that globally, 5 to 10 mil- for land expansion in some Sub-Saharan lion hectares of agricultural land are being countries, but this will require large invest- lost annually to severe degradation.53 Soil ments in infrastructure and human and degradation through nutrient mining is animal disease control to convert these a huge problem in Sub-Saharan Africa, lands to productive agriculture. though much of it is reversible through Figure 2.10 Arable and permanent cropland per capita of the agricultural population is falling in Sub-Saharan Africa and South Asia Index of cropland per agricultural population (1961 = 100) 180 Sub-Saharan Africa 0.48 160 140 South Asia 0.27 120 East Asia & Pacific 0.23 100 80 Middle East & 0.74 North Africa 60 Europe & 3.53 40 Central Asia 20 Latin America & 1.55 Caribbean 0 1961 1967 1973 1979 1985 1991 1997 2003 0 1 2 3 4 Cropland per capita of agricultural Sub-Saharan Africa Middle East & North Africa population in 2003, hectares South Asia Europe & Central Asia East Asia & Pacific Latin America & Caribbean Source: FAO 2006a. Notes: Cropland represents both arable and permanent cropland. 64 WORLD DEVELOPMENT REPORT 2008 better soil management and fertilizer use water because of a lack of infrastructure or (see box 2.1). the human and financial capital to tap the available resources (chapter 8). The Middle Acute water scarcity East and North Africa and Asia face the Agriculture uses 85 percent of fresh water greatest water shortages, although there are withdrawals in developing countries, and pockets of severe water scarcity in all other irrigated agriculture accounts for about regions as well. 40 percent of the value of agricultural pro- Large areas of China, South Asia, and duction in the developing world.54 With- the Middle East and North Africa are now out irrigation, the increases in yields and maintaining irrigated food production output that have fed the world's growing through unsustainable extractions of water population and stabilized food production from rivers or the ground.57 The ground- would not have been possible. water overdraft rate exceeds 25 percent in Demand for water for both agricultural China and 56 percent in parts of northwest and nonagricultural uses is rising, and India.58 With groundwater use for irriga- water scarcity is becoming acute in much tion expected to continue rising, often of the developing world, limiting the future driven by subsidized or free electricity, the expansion of irrigation. The water avail- degradation of groundwater aquifers from able for irrigated agriculture in develop- overpumping and pollution is certain to ing countries is not expected to increase become more severe (chapter 8).59 because of competition from rapidly grow- Sub-Saharan Africa and Latin America ing industrial sectors and urban popula- have large untapped water resources for tions.55 New sources of water are expen- agriculture. But even in Sub-Saharan Africa, sive to develop, limiting the potential for almost a quarter of the population live in expansion, and building new dams often water-stressed countries, and the share imposes high environmental and human is rising.60 Even so, there now are many resettlement costs. opportunities for economically investing in According to the Comprehensive Assess- irrigation in Sub-Saharan Africa (box 2.5), ment of Water Management in Agricul- and the irrigated area there is projected to ture,56 approximately 1.2 billion people live double by 2030. in river basins with absolute water scarcity In other regions, the emphasis on water (figure 2.1); 478 million live in basins where for irrigation has already shifted to increas- scarcity is fast approaching; and a further ing the productivity of existing water with- 1.5 billion suffer from inadequate access to drawals by reforming institutions and removing policy distortions in agriculture Figure 2.11 Water scarcity affects millions of people in Asia and the Middle East and and in the water sector (chapter 8). With North Africa productivity growth and a modest growth Population living in areas of absolute water scarcity in irrigated area of 0.2 percent annually, Percent Millions irrigated production is projected to account 600 for nearly 40 percent of the increased agri- 60 Percent (left axis) 500 cultural production in the developing world Millions (right axis) by 2030. 400 40 300 Uncertain effects of climate change Global warming is one of the areas of great- 200 20 est uncertainty for agriculture. If emissions 100 continue at today's rate, the global average temperature is likely to rise by 2°C­3°C 0 0 Sub-Saharan South East Middle Europe & Latin OECD over the next 50 years, with implications Africa Asia Asia & East Central America & for rainfall and the frequency and intensity Pacific & North Asia Caribbean Africa of extreme weather events.61 The effects are not evenly distributed. While many regions Source: International Water Management Institute (IWMI) analysis done for the Comprehensive Assessment of Water Management in Agriculture (2007) using the Watersim model. have already become wetter, parts of the Agriculture's performance, diversity, and uncertainties 65 B O X 2 . 5 Substantial potential for expanding irrigation in Sub-Saharan Africa-- in the right way Sub-Saharan Africa has a large untapped of poorer market opportunities and higher These economic returns can be realized only potential for irrigation. Only 4 percent of the investment costs than in other regions. But if a significant share of the area is sown with total cultivated area is under irrigation, with with the new generation of better-designed higher-value crops. This underlines the need a mere 4 million hectares added in the last 40 irrigation projects, costs in Sub-Saharan for complementary investments in roads, years, far less than in any other region.62 Invest- Africa are now comparable to those in other extension services, and access to markets. ment in irrigation projects steadily declined in regions, thanks to improvements in institu- Small-scale irrigation is also showing recent the 1980s, partly in response to the many failed tions, technology, and market opportunities successes, especially in Niger and the Fadama irrigation investments and partly because for high-value products (see table below). program in Nigeria (chapter 8). Returns on irrigation in Sub-Saharan Africa 1970­74 1975­79 1980­84 1985­89 1990­94 1995­99 Sub-Saharan Africa Number of projects 3 9 11 15 4 3 Cost per hectare (2000 US$) 4,684 24,496 11,319 7,669 8,287 8,347 Average economic rate of return (%) 10 2 8 16 17 30 Non-Sub-Saharan Africa Number of projects 21 66 75 41 49 6 Cost per hectare (2000 US$) 3,433 4,152 5,174 2,252 3,222 3,506 Average economic rate of return (%) 19 15 15 18 21 17 Sources: African Development Bank and others 2007; Carter and Danert 2007; IFAD 2005a; International Water Management Institute (IWMI) 2005; World Bank 2006t. Note: Rates of return on externally financed irrigation projects in Sub-Saharan Africa and the rest of the world (two-thirds of which were in Asia) during 1970­99. Sahel, the Mediterranean, southern Africa, change after 2050 with further increases and parts of southern Asia are becoming in temperature.65 But stronger impacts drier--and this trend will continue. Water are expected at the regional level. Relative scarcity will increase in many areas, partic- to the scenario of no climate change, agri- ularly in the already-dry parts of Africa and cultural GDP in Sub-Saharan Africa (the in areas where glacial melt is an important region with the highest impact from cli- source of irrigation water. mate change) could contract by anywhere With moderate warming, crop yields are from 2 to 9 percent.66 expected to increase in temperate areas and The major implications of climate change decline in the tropics. Crop-climate models are thus largely for the distribution of agri- predict an increase in global crop produc- cultural production. In a globalizing world, tion in slight to medium warming scenarios some of the adaptation can be accommo- of less than 3°C.63 But the combined effects dated by trade, if measures are in place to of higher average temperatures, greater ensure alternative livelihoods of those most variability of temperature and precipita- affected. But for much of the tropics, espe- tion, more frequent and intense droughts cially areas of Sub-Saharan Africa nega- and floods, and reduced availability of tively affected by climate change, trade can water for irrigation can be devastating for only partially fill the gap. agriculture in many tropical regions (see focus F). One-third of the population at High energy prices: pressure on food risk of hunger is in Africa, one-quarter in prices from two sides Western Asia, and about one-sixth in Latin Although there is considerable uncertainty America.64 about future energy prices,67 there is little The impact of climate change on food doubt that energy prices will be higher prices at the global level is predicted to than in the past 20 years and that this will be small through 2050. Some models pre- increase agricultural production costs, dict more substantial effects from climate placing upward pressure on food prices. 66 WORLD DEVELOPMENT REPORT 2008 On the demand side, the greatest uncer- gated wheat in the Indian Punjab in 2002, tainty is the pace of expansion of biofuels and 34 percent of soybean costs in Mato using agricultural feedstocks in response Grosso, Brazil.74 Sharply higher fertilizer to high energy prices. The magnitude of prices could have far-reaching effects on the expansion of use of feedstocks and its developing-country agriculture--push- impact on food prices is uncertain. Recent ing down fertilizer application rates and projections indicate real price increases of crop yields and raising food prices--unless as much as 40 percent for maize by 2020, rapid advances are made in tapping nutri- with spillover effects on substitute grains ent sources that do not depend on fossil (wheat), given rapid growth in biofu- fuels, such as biological nitrogen fixation els demand.68 But over the long run, the by including legumes in farming systems or prices of feedstocks such as maize and biotechnological advances that fix nitrogen sugar cannot rise faster than real energy in cereals (chapter 7). prices if biofuels are to be competitive, so Beyond the farmgate, other energy- the impacts are likely to be much lower.69 dependent food production inputs, such Major uncertainties then relate to the price as transport and refrigeration costs, will be of oil, the technical progress in conversion affected by higher energy costs. Four per- efficiency of agricultural feedstocks and cent of U.S. food costs are attributable to biomass, and the extent that governments transport expenses alone.75 Long-distance subsidize or mandate biofuel production air freight for global food markets may be (see focus B). most affected--aviation fuel represents On the supply side, much of today's agri- about 7 percent of the retail price of a basket cultural production is fairly energy inten- of high-value products in a U.K. supermar- sive, more so in the developed world than ket.76 These costs are stimulating interest in in the developing. Estimates by the FAO local food markets in industrial countries indicate that 6,000 megajoule (MJ) of fossil to minimize "food miles"; however, there is energy--equal to 160 liters of oil--are used not always a strong association between the to produce one ton of maize in the United distance that food travels and the combined States. One ton of maize grown in Mexico use of nonrenewable energy in food produc- under traditional methods uses only 180 MJ tion and transport.77 of energy inputs, equal to 4.8 liters of oil.70 Energy is required directly for the opera- Will science deliver? tion of machinery and indirectly for fertil- With growing resource scarcity, future food izers and other chemicals. Fertilizer prices, production depends more than ever on for example, are linked to energy prices increasing crop yields and livestock pro- because natural gas, a primary component ductivity. But the outlook for technological in nitrogen fertilizer production, repre- progress has both positive and negative ele- sents 75 to 90 percent of the production ments that raise uncertainty. For the major costs.71 In the United States, energy costs cereals--rice, wheat, and maize--the accounted for 16 percent of agricultural growth rate of yields in developing coun- production costs in 2005, about one-third tries has slowed sharply since the 1980s for fuel and electricity and two-thirds indi- (figure 2.12); the easy gains from high use rectly for energy to produce fertilizer and of green-revolution inputs have already chemicals.72 Econometric analyses suggest been made, except in Africa. Plant breed- that U.S. grain prices (which determine ers continue to increase the yield poten- world prices) would rise by 18­20 percent of tial of wheat by about 1 percent annually, any increase in crude oil prices, not includ- but less for the world's major food crop, ing effects on the demand side through rice.78 Slowing of R&D spending in many biofuels.73 countries raises concerns about the pace of In developing countries, fertilizer costs future gains (chapter 7). are a growing share of production costs-- Historically, a significant part of yield 18 percent of the variable costs for irri- gains has been achieved by narrowing the Agriculture's performance, diversity, and uncertainties 67 Figure 2.12 Growth rates of yields for major whether this revolution will become a real- cereals are slowing in developing countries ity for food production in the developing Average annual growth rate, % world because of low public investment in 6 these technologies and controversies over 5 their possible risks (see focus E). However, Wheat biotechnology applications using genomics 4 and other tools are not controversial, and 3 Maize their declining costs and wider applica- tion should ensure continuing yield gains 2 through better resistance to disease and Rice 1 tolerance for drought and other stresses 0 (chapter 7). 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 Source: FAO 2006a. The bottom line: a more Note: Data smoothed by locally weighted regressions. uncertain future? Future trends could be accentuated if sev- eral adverse outcomes eventuate. High gap between average farm yields and the energy prices combined with more biofu- experimental yield potential of the crop, up els production from food crops could lead to a point where average farm yields reach to large food crop price increases through about 80 percent of experimental yields. effects on both supply and demand. Global China's major rice-producing provinces warming could occur faster than expected and much of the wheat and maize pro- and add to water shortages, hitting irrigated duced in industrial countries have already agriculture with lower yields and increasing reached this point, so the gap is closing.79 risk in rainfed agriculture. Rapid income Other rice-producing areas of Asia are well growth in Asian countries with limited land below 80 percent of experimental yields, and water resources could lead to a surge in and their yield growth has slowed because food imports that, combined with higher of deteriorating soil and water quality and energy and fertilizer prices, drive up food imbalanced nutrient use.80 prices. Or, all three could happen together. Exploitable yield gaps are especially Interdependence also implies likely trad- high in medium- to high-potential areas of eoffs between poverty, food security, and agriculture-based countries. Onfarm dem- environmental sustainability. For example, onstrations using available "best bet" tech- land constraints can be relaxed in many nologies suggest a wide yield gap for maize in Sub-Saharan Africa (figure 2.13). But Figure 2.13 Exploitable yield gaps are high for maize in Africa closing the gaps is a matter not just of trans- ferring these technologies to farmers, but Tons per hectare 6 of putting in place the institutional struc- Average national yield tures--especially well-functioning input 5 Average yield in farm demonstrations and output markets, access to finance, and 4 ways to manage risks--that farmers need to adopt the technology (chapters 5 and 6). 3 The world is poised for another tech- 2 nological revolution in agriculture using the new tools of biotechnology to deliver 1 significant yield gains (chapter 7). Already 0 100 million hectares of crops, or about 8 Malawi Ethiopia Nigeria Uganda Mali Mozambique percent of the cropped area, are sown with (4566) (310) (2501) (461) (163) (508) transgenic seeds (often known as geneti- Source: Sasakawa Africa, personal communication. cally modified organisms or GMOs). But Notes: Number of plots in parentheses. Open pollinated improved varieties in all cases except Nigeria, which uses hybrids. Data for 2001 for Ethiopia, Mozambique, Nigeria, and Uganda; 2002 for Malawi; and an average of there is considerable uncertainty about 2001, 2002, and 2004 for Mali. 68 WORLD DEVELOPMENT REPORT 2008 regions in response to rising prices, but growth, in a context where food produc- only at significant environmental cost. tion is important for food security (chapter Because of these uncertainties, global, 1). With limited tradability because of the national, and local production shocks could types of food consumed and high transac- become more frequent. Countries will need tion costs, the need for Sub-Saharan Africa to increase their capacity to manage shocks to feed itself based largely on its own pro- through production risk mitigation (better duction remains a stark reality. Poor per- water control or drought-tolerant varieties), formance is a source of food insecurity only trade, and insurance (chapter 5). Countries partially compensated by food imports and with rising incomes will be best able to food aid. manage these shocks because higher food Faster growth of agricultural production prices will have less impact on real incomes. in Sub-Saharan Africa is also essential for The least-developed countries would be hit overall growth and poverty reduction in hardest. the region, as seen in chapter 1. The recent progress in accelerating growth in Sub- A growing divide among regions? Saharan Africa must be sustained in coun- Differences in agricultural performance tries already experiencing rapid growth and among countries are projected to persist broadened to (often conflict or post conflict) and even deepen under a business-as-usual countries that have not yet participated. scenario, especially between the agriculture- The second reason for a continued focus based countries and the rest. Within Sub- on agricultural production is the poor agri- Saharan Africa, continuing rural population cultural performance across all country growth greater than 1.8 percent a year in types in areas with difficult agroclimatic some countries adds to already serious pres- conditions or inadequate infrastructure sure on available land.81 Together with poor that constrains market access. In these agricultural resources and a high depen- regions, livelihoods depend on agricultural dence on domestic agriculture, the risks of production, either as a source of income food insecurity in such landlocked countries or for food for home consumption. The as Burundi, Ethiopia, and Niger will greatly challenge is to improve the productivity increase unless massive efforts are mounted of subsistence agriculture, diversify to new to intensify production on existing land.82 markets where possible, and open opportu- IFPRI projections highlight the close link nities for nonfarm work and migration as between agricultural productivity and nutri- pathways out of poverty (chapter 3). tional outcomes in Sub-Saharan Africa-- The third reason is that even high- and the urgency of increased investments to potential areas that led the global increase reach the Millennium Development Goal of in food production (such as the transform- cutting hunger by half. ing countries of Asia) are facing a triple production challenge. They must sustain productivity and income growth in the Conclusion--a continuing face of declining prices in grains and tra- production challenge ditional tropical exports, they must seize Does success over the past three decades the opportunity to diversify in high-value in meeting rapidly growing food demands horticulture and livestock in response to mean that food production is no longer a rapidly growing domestic and interna- problem? The review of food and agricul- tional demand, and they must reduce the tural production trends and challenges in environmental footprint of intensive crop this chapter suggests four reasons why the and livestock systems. production problem still belongs on the The last reason is more speculative, but development agenda. still important. Even at the global level, The first is the lagging performance of future agricultural success may be com- agriculture-based countries, especially in promised by greater resource scarcity, Sub-Saharan Africa, relative to population heightened risks from climate change, Agriculture's performance, diversity, and uncertainties 69 higher energy prices, competition for land of rising food prices. The world food supply between food and biofuels, and under- requires close monitoring and new invest- investment in technical progress. For the ments to speed productivity growth, make first time since the world food crisis in the production systems more sustainable, and 1970s, global models predict the possibility adapt to climate change. focus B Biofuels: the promise and the risks Biofuels offer a potential source of renewable energy and possible large new markets for agricultural producers. But few current biofuels programs are economically viable, and most have social and environmental costs: upward pressure on food prices, intensified competition for land and water, and possibly, deforestation. National biofuel strategies need to be based on a thorough assessment of these opportunities and costs. Biofuels could become big markets a fivefold increase of the share of biofuels Are biofuels economically viable without for agriculture--with risks in global transport energy consumption-- subsidies and protection? The breakeven With oil prices near an all-time high and from just over 1 percent today to around 5 price for a given biofuel to become eco- few alternative fuels for transport, Brazil, to 6 percent by 2020.4 nomical is a function of several parameters. the European Union, the United States, and The most important determining factors are several other countries are actively support- the cost of oil and the cost of the feedstock, Economic viability of biofuels ing the production of liquid biofuels (etha- which constitutes more than half of today's and the impact on food prices nol and biodiesel).1 The economic, envi- production costs. ronmental, and social impacts of biofuels Governments provide substantial support Biofuel production has pushed up feed- are widely debated. As a renewable energy to biofuels so that they can compete with stock prices. The clearest example is maize, source, biofuels could help mitigate climate gasoline and conventional diesel. These sup- whose price rose by 23 percent in 2006 change and reduce dependence on oil in the ports include consumption incentives (fuel and by some 60 percent over the past two transportation sector. They may also offer tax reductions), production incentives (tax years, largely because of the U.S. ethanol large new markets for agricultural producers incentives, loan guarantees, direct subsidy program.6 Spurred by subsidies and the that could stimulate rural growth and farm payments), and mandatory consumption Renewable Fuel Standard issued in 2005, incomes. On the downside are environ- requirements. According to recent esti- the United States has been diverting more mental risks and upward pressure on food mates, more than 200 support measures maize to ethanol. Because it is the world's prices. These impacts, which depend on the costing around $5.5­7.3 billion a year in largest maize exporter, biofuel expansion type of feedstock (raw material), production the United States amount to $0.38­0.49 per in the United States has contributed to a process, and changes in land use, need to be liter of petroleum equivalent for ethanol decline in grain stocks to a low level and carefully assessed before extending public and $0.45­0.57 for biodiesel.5 Even in Bra- has put upward pressure on world cereal support to large-scale biofuel programs. zil, sustained government support through prices. Largely because of biodiesel produc- Of the global fuel ethanol production direct subsidies was required until recently tion, similar price increases have occurred of around 40 billion liters in 2006, about to develop a competitive industry, despite for vegetable oils (palm, soybean, and rape- 90 percent was produced in Brazil and the uniquely favorable sugarcane-growing con- seed).7 Cereal supply is likely to remain United States, and of over 6 billion liters of ditions, a well-developed infrastructure, and constrained in the near term and prices will biodiesel, 75 percent was produced in the a high level of synergy between sugar and be subject to upward pressure from fur- EU--mainly in France and Germany (figure ethanol production. Domestic producers in ther supply shocks.8 Provided there is not B.1). Brazil is the most competitive producer the European Union and the United States another major surge in energy prices, how- and has the longest history of ethanol pro- receive additional support through high ever, it is likely that feedstock prices will rise duction (dating back to the 1930s), using import tariffs on ethanol. less in the long term as farmers respond to about half its sugarcane to produce etha- nol and mandating its consumption. With tax incentives, subsidies, and consumption Figure B.1 Fuel ethanol and biodiesel production is highly concentrated mandates for biofuel production, the United Ethanol (40 billion liters) Biodiesel (6.5 billion liters) States used 20 percent of its maize crop to produce ethanol in 2006/07 (forecast).2 Other European 8% Union New players are emerging. Many devel- 4% oping countries are launching biofuel pro- grams based on agricultural feedstocks: Other 12% biodiesel from palm oil in Indonesia United States and Malaysia, ethanol from sugarcane in 13% Mozambique and several Central Ameri- United States can countries, and ethanol from sugarcane 46% and biodiesel from such oil-rich plants European Brazil Union as jatropha, pongamia, and other feed- 42% 75% stocks in India.3 Although assessments of the global economic potential of biofuels have just begun, current biofuels policies Source: F.O.Licht Consulting Company, personal communication, July 17, 2007. could, according to some estimates, lead to Note : Percentages of global production of fuel ethanol and biodiesel in 2006. Biofuels: the promise and the risks 71 higher prices (chapter 2), and biofuels pro- Potential environmental impacts: Global technologies are likely to require even larger duction will be moderated by lower profits environmental benefits from using renew- economies of scale, with investment costs in because of higher feedstock prices.9 able fuels--reducing greenhouse gas emis- the hundreds of millions of dollars just to Rising agricultural crop prices from sions (GHGs)--are frequently cited as rea- build one plant. demand for biofuels have come to the sons for policy support to biofuels. Although Although most biofuel production is forefront in the debate about the potential possibly significant, those benefits cannot be large in scale, small-scale production of conflict between food and fuel. The grain assumed. The emissions from growing feed- biodiesel with current technologies could required to fill the tank of a sport utility stocks (including emissions from fertilizer meet local energy demand (for example, vehicle with ethanol (240 kilograms of maize production), manufacturing biofuels, and biodiesel use in stationary electricity gener- for 100 liters of ethanol) could feed one per- transporting biofuels to consumption cen- ators). For wider markets and for biodiesel son for a year, so competition between food ters, as well as those from changes in land use for transportation, meeting consistent and fuel is real. Rising cereal prices will have use, also have to be evaluated.14 quality standards in small-scale production an adverse impact on many food-importing Using existing crop land, Brazilian sug- is a problem.20 countries. Even in the short term, higher arcane is estimated to reduce gasoline emis- prices of staple crops can cause significant sions by about 90 percent. Biodiesel is also welfare losses for the poor, most of whom relatively efficient, reducing GHGs by 50 to 60 Defining public policies for biofuels are net buyers of staple crops.10 But many percent. In contrast, the reduction of GHGs To date, production in industrial countries poor producers could benefit from higher for ethanol from maize in the United States has developed behind high protective tariffs prices (chapter 4). is only in the range of 10 to 30 percent.15 In on biofuels and with large subsidies. These Future biofuels technology may rely on such cases, demand-side efficiency measures policies are costly to those developing coun- dedicated energy crops and agricultural in the transport sector are likely to be much tries that are or could become potentially and timber wastes instead of food crops, more cost-effective than biofuels in reduc- efficient producers in profitable new export potentially reducing the pressure on food ing GHGs. The cost of reducing one ton of markets.21 Poor consumers may pay higher crop prices and contributing to the supply carbon dioxide (CO2) emissions through the prices for food staples as grain prices rise in of more environmentally friendly supplies production and use of maize-based ethanol world markets. Food prices may rise directly of liquid biofuels. But technology to break could be as high as $500 a ton, or 30 times because of the diversion of grain to biofu- cellulose into sugars distilled to produce the cost of one ton of CO2 offsets in the Euro- els or indirectly because of land conversion ethanol or gasify biomass is not yet com- pean Climate Exchange.16 away from food when induced by distor- mercially viable--and will not be for several According to the 2006 EU Biofuel Strat- tionary policies. years.11 And some competition for land and egy, a change in land use, such as cutting Can developing countries, apart from water between dedicated energy crops and forests or draining peat land to produce Brazil, benefit from production of biofuels? food crops will likely remain. feedstocks such as oil palm, can cancel the Favorable economic conditions and large GHG emission savings "for decades."17 environmental and social benefits that justify Reducing potential environmental risks significant subsidies are probably uncom- Nonmarket, context-specific from large-scale biofuels production could mon for the first-generation technologies. benefits need to be evaluated be possible through certification schemes to In some cases, such as landlocked countries Whether the financial costs, efficiency losses, measure and communicate the environmen- that are importers of oil and potentially effi- and the tradeoffs between food and fuel asso- tal performance of biofuels (for example, a cient producers of sugarcane, the high costs ciated with these various support measures Green Biofuels Index of GHG reductions).18 of transport could make biofuel produc- are justified depends on the environmental Similar standards exist for organic products tion economically viable even with current and social benefits and risks of biofuels and and for the sustainable production of forest technologies.22 The much higher potential their contribution to energy security. products (Forest Stewardship Council). But benefits of second-generation technologies, Potential to enhance energy security: Cur- the effectiveness of certification schemes at including for small-scale biodiesel produc- rent-technology biofuels can only margin- reducing environmental risks from biofuels tion, justify substantial privately and pub- ally enhance energy security in individual will require full participation from all major licly financed investments in research. countries because domestic harvests of feed- producers and buyers as well as strong mon- The challenge for developing country stock crops meet a small part of the demand itoring systems. governments is to avoid supporting biofuels for transport fuels, with few exceptions (for Benefits to smallholders: Biofuel can ben- through distortionary incentives that might example, ethanol in Brazil). In 2006/07, efit smallholder farmers through employ- displace alternative activities with higher around one-fifth of the U.S. maize harvest ment generation and higher rural incomes, returns--and to implement regulations was used for ethanol but displaced only but the scope of these impacts is likely to and devise certification systems to reduce about 3 percent of gasoline consumption.12 remain limited. Ethanol production with environmental risks. Governments need According to recent projections, 30 percent current technologies requires fairly large to carefully assess economic, environmen- of the U.S. maize harvest would be used for economies of scale and vertical integra- tal, and social benefits and the potential to ethanol by 2010, but it would still account tion and may do little to help small-scale enhance energy security. Other often more for less than 5 percent of U.S. gasoline con- farmers. In some parts of Brazil, however, cost-effective ways of delivering environ- sumption.13 Second-generation technologies producer cooperatives have succeeded in mental and social benefits need to be con- could potentially make a higher contribution ensuring smallholder participation.19 Sec- sidered, especially through improvements in to energy security. ond-generation biofuels using cellulosic fuel efficiency. Rural households and their pathways out of poverty 3 Agriculture is a major source of livelihoods oriented smallholder entrepreneurs and for people in developing countries, but smallholders largely engaged in subsistence rural areas are a large harbor of poverty. To farming. There is a parallel dualism in the c h a p t e r understand how agricultural growth can labor market between high-skill and low- reduce rural poverty, this chapter identi- skill jobs, and between migration with fies three pathways out of rural poverty. It high and low returns. Nor is diversification characterizes the livelihood strategies of always a sign of success. Chapter 9 analyzes rural households and identifies challenges the factors underlying the heterogeneity in to defeating rural poverty through these labor market and migration outcomes, with pathways.1 a focus on policy measures to improve these Many rural households move out of outcomes for the rural poor. poverty through agricultural entrepreneur- Rural households design livelihood ship; others through the rural labor market strategies to suit their asset endowments and the rural nonfarm economy; and oth- and account for the constraints imposed by ers by migrating to towns, cities, or other market failures, state failures, social norms, countries. The three pathways are comple- and exposures to uninsured risks. They mentary: nonfarm incomes can enhance may not use those terms, but they certainly the potential of farming as a pathway out understand the constraints. Their strategies of poverty, and agriculture can facilitate the can reflect joint decision making by men labor and migration pathways. and women in the household, or can be bar- Inspecting what individuals and house- gained outcomes when members each pur- holds do in rural areas helps dismiss two sue their own advantage. But their strategies frequent misconceptions about rural pop- compensate for only part of the constraints ulations. The first is the belief that rural they operate under, leaving important roles households are either all farmers or all for improvements in their access to assets diversified. To the contrary, there is a con- and in the contexts for using these assets.2 siderable heterogeneity in what they do and The key, then, is to enhance collective action in the relative importance of what they do and mobilize public policy to maximize the for their incomes. A large majority of rural likelihood of success for rural households to households are engaged in some agricul- travel a pathway out of poverty. tural activity, but many derive a large part Policy makers thus face daunting chal- of their income from off-farm activities and lenges. The asset endowments of rural from migration. Individuals participate in a households have been low for generations, wide range of occupations, but occupational and they continue to decline in places. diversity does not necessarily translate into Market and government failures affect- significant income diversity in households. ing the returns on those assets are perva- The second misconception is the belief sive. Adverse shocks often deplete already- that the type of activities households pur- limited assets, and the inability to cope with sue determines their success in moving out shocks induces households to adopt low- of poverty. This is not so because of the risk, low-return activities. Recent changes considerable heterogeneity within activi- in the global food market, in science and ties. Livelihood strategies in agriculture are technology, and in a range of institutions characterized by dualism between market- that affect competitiveness are also creat- 72 Rural households and their pathways out of poverty 73 ing new challenges to the competitiveness sion helped many households move up the of smallholders. Understanding these chal- socioeconomic ladder. Others benefited lenges is essential in designing public poli- from greater trade in food crops.4 cies that can help rural men and women In Vietnam, liberalizing agricultural pull themselves out of poverty. The chal- markets induced many subsistence farm- lenges differ across countries and subna- ers to become more market oriented (table tional regions, and thus demand context- 3.1). Two-thirds of smallholders previously specific agendas to reduce rural poverty. engaged primarily in subsistence farming entered the market. Their poverty rates fell drastically, and their incomes almost dou- Three complementary pathways bled, while the production of high-value out of rural poverty: farming, and industrial crops rose. Agricultural sales labor, and migration increased more for households with larger Rural poverty rates have declined in many land endowments and those closer to mar- countries (see focus A). But how exactly has kets or with nonfarm industries in their this happened? Is it that poor households communities. Households engaged in sub- leave rural areas, or that older, poor gen- sistence farming that did not enter the mar- erations are replaced by younger, less-poor ket were more likely to diversify their income generations? Have specific households been sources outside of agriculture, with poverty able to escape poverty by gradually improv- rates in those groups falling as well. ing the earnings from whatever they do, or In India, income from the nonagricul- has this happened by drastically changing tural sector--the labor pathway out of pov- activities? Success stories help illustrate erty--was an important driver of growth in how rural households have exited poverty rural areas between 1970 and 2000. Nonag- through the three pathways of farming, ricultural employment also had important labor, and migration. indirect effects by increasing agricultural In Tanzania, those most successful in wages. In Indonesia, agricultural house- moving out of poverty were farmers who holds that shifted into the nonfarm econ- diversified their farming activities by grow- omy between 1993 and 2000 were likely ing food crops for their own consumption to have exited poverty. In Tanzania, too, and nontraditional cash crops (vegetables, business and trade provided an important fruit, vanilla) as well as raising livestock. pathway out of poverty, but only for those People who remained in poverty were those with networks in well-connected commu- who stuck to the more traditional farming nities. In addition, remittances from both systems. In Uganda, escaping from poverty domestic and international migration have was linked to improving the productivity reduced rural poverty, as happened in rural of land and diversifying into commercial China and Nepal.5 Migration can offer a crops. Qualitative evidence for Niger shows pathway out of poverty for those who leave that shifts to more sustainable cultivation and for those who stay behind (chapter 9). practices by small-scale farmers led to better Several pathways often operate at the soil conservation, increased income from same time. In Bangladesh and Tanzania, the agroforestry, and lower vulnerability.3 farm, nonfarm labor, and migration path- Some policy reforms have greatly ways were all successful. In Indonesia, some enhanced the capability of smallholder people moved out of poverty through the entrepreneurs to lift themselves from pov- farming pathway, others through the non- erty. This was clearly a key to China's early farm pathways. And in 35 villages in Andhra agricultural success story (see focus A). In Pradesh, diversification of income sources is Malawi, reforms reducing differential pro- correlated with moving out of poverty. 6 tection of large estates dramatically shifted These careful studies using longitudinal the structure of agricultural production. data have shed light on the strong potential Smallholders rapidly diversified into cash relationships between poverty reduction crops and now produce 70 percent of burley and each of the pathways. However, estab- tobacco, a major export crop. The expan- lishing causality is difficult, and there is no 74 WORLD DEVELOPMENT REPORT 2008 Table 3.1 Changing market participation among farming households in Vietnam Subsistence oriented Market entrant Market oriented 6a 13a 28a Household characteristics 1992/3 1997/8 1992/3 1997/8 1992/3 1997/8 Assets Land owned (ha) 0.37 0.43 0.50 0.57 0.60 0.72 Land used (ha) 0.55 0.43 0.59 0.58 0.71 0.75 Education of household head (years) 4.6 -- 6.3 -- 6.3 -- Context Market in community (%) 31 -- 40 -- 47 -- Commercial enterprise in community (%) 34 -- 43 -- 42 -- Outcomes Real income per capita (1998 dong 1,000) 893 1,702 1,138 2,042 1,359 2,978 Share of agricultural income in total income (%) 80 62 83 66 83 73 Share of households below the poverty line (%) 86 62 73 48 64 37 Shares of gross agricultural income by crop type Staple crops (%) 78 73 70 61 63 54 High-value and industrial crops (%) 14 13 21 31 29 39 Source: WDR 2008 team using VLSS 1992/93 and 1997/98. Note: Subsistence-oriented farming households are defined here as selling less than 10 percent of their agricultural production in both years; market-entrant households as selling less than 10 percent in 1992/3 and more than 25 percent in 1997/8; and market-oriented households as selling more than 25 percent in both years. Rural farming households are households with more than 50 percent of income from agriculture. a. Percent of rural farming households. -- = not available. systematic evidence on the relative impor- have a large impact on agricultural land tance and success of these strategies, a result purchases, and returning migrants are more of conceptual challenges in understanding likely to set up a nonfarm business.8 the dynamics of poverty (box 3.1). While transfers from migrants back to the farm household can relax capital and Pathways often enhance each other risk constraints, the relationship between The complementing effects of farm and migration and agricultural productivity nonfarm activities can be strong. In Ban- is complex. The (temporary) absence of gladesh and Ecuador, farm households with household members reduces the agricul- better market access or in areas with higher tural labor supply. Agricultural productiv- agricultural potential earn more from agri- ity can therefore fall in the short run but culture, but they also diversify more into rise in the long run as households with nonfarm activities. In Asia, high rural sav- migrants shift to less labor intensive, but ings rates from rising incomes during the possibly equally profitable, crops or live- green revolution made capital available for stock.9 Male out-migration can transfer investment in nonfarm activities.7 Diversi- responsibility for farm management to fication into nonfarm activities can relax women. And where women have less access credit and liquidity constraints on own- to credit, extension, and markets, as is fre- farm agricultural production and enhance quently the case, farm productivity might the competitiveness of the family farm on fall as a result. The transfer of responsibility the agricultural pathway. may also be only partial, limiting women's The farming, labor, and migration path- possibilities to take advantage of emerging ways have often enhanced each other. In the opportunities to improve competitiveness. Philippines, the green revolution allowed The variation in rural children of land reform beneficiaries and large farmers--especially daughters--to households' income strategies attain high levels of education. These highly Contrary to the prototypical image of educated offspring are now sending large smallholders as pure farmers, landed rural transfers back to farm households. In Paki- households rely on many activities and stan, remittances from temporary migrants income sources. Besides farming, they par- Rural households and their pathways out of poverty 75 ticipate in agricultural labor markets, in self-employment or wage employment in B O X 3 . 1 Establishing the relative importance the rural nonfarm economy, and they might of the different pathways receive transfers from household members who have migrated. Moving out of poverty is a process households. It is thus not possible to know that can take a very long time. Many whether they would have been poor had Diversification has several dimensions shocks can occur during that time, and a they not migrated (see focus A). This is that should not be confounded. The rural household's income fluctuations may be particularly important because many economy is diversified, even if many non- similar in magnitude to long-term income migrants are more educated than those changes. So, in the short-term, it is seldom agricultural activities are indirectly linked who stay behind, and they would prob- clear whether observed income changes ably not have been among the poorest. to agriculture. Within this diversified rural reflect transitory movements in and out Nor is it easy to disentangle why economy, a large part of household income of poverty, or long-term trends. Only by households chose a particular strategy diversification comes from combining interviewing the same households many from what made the pathway successful. times over long periods might it be pos- More entrepreneurial households might incomes from the different household sible to gauge the relative importance of choose "better" strategies, but they might members, each often specializing in one different pathways in a particular context. also be more successful in moving out of occupation. In Malawi, 32 percent of farm Consider trying to capture the full poverty independently of the strategies households have two sources of income, and effects of the migration pathway on those they choose. Some migration studies have who migrated. When people migrate, they addressed this selection issue and estab- 42 percent have three or more, but among typically disappear from surveys, unless lished the effects of migration on the pov- household heads only 27 percent engages in one manages to track them down in their erty of household members left behind. more than one activity. In China, 65 per- new locations, which can be difficult. But doing this for the other pathways cent of rural households operate in both Moreover, a lot of migration is by young remains unresolved. people, before they form independent the farm and nonfarm sectors, while only a third of individuals do so.10 These patterns imply that household income diversification can fluctuate considerably with households depend on farming for their livelihoods, life cycles, and the number of working-age but use the majority of their produce for individuals in the household. Further, the home consumption (subsistence-oriented returns on many of these activities are low, farmers).12 Still others derive the larger and the diversity of occupations does not part of their incomes from wage work in always translate into income diversifica- agriculture or the rural nonfarm economy, tion: one activity is often the dominant or from nonagricultural self-employment source of income. (labor-oriented households). Some house- To design policies that help households holds might choose to leave the rural sec- along successful pathways, it is crucial tor entirely, or depend on transfers from to understand which income strategies members who have migrated (migration- they currently pursue and why they chose oriented households). Finally, diversified to pursue them. This allows evaluating households combine income from farming, whether policies should aim at enhancing off-farm labor, and migration. their current strategies or at helping them Income sources can be used to classify to pursue more remunerative ones. Further- rural households according to the five liveli- more, understanding why some households hood strategies (table 3.2 and box 3.2). The remain poor despite choosing strategies relative importance of each differs across that are optimal, given their assets and con- the three country types: agriculture-based, straints, helps to identify policy options. transforming, and urbanized. It also differs across regions within countries. Farming- A typology of rural households led strategies are particularly important Rural households engage in farming, in the agriculture-based countries, where labor, and migration, but one of these farming is the main livelihood for a large activities usually dominates as a source share of rural households, as many as 71 of income. Five livelihood strategies can percent in Nigeria and 54 percent in Ghana be distinguished. Some farm households and Madagascar. Many of those households derive most of their income from actively are subsistence oriented. engaging in agricultural markets (market- In the transforming and urbanized coun- oriented smallholders).11 Others primarily tries, the labor- and migration-oriented 76 WORLD DEVELOPMENT REPORT 2008 Table 3.2 Typology of rural households by livelihood strategies in three country types Farm oriented Market Subsistence Labor Migration oriented oriented Total oriented oriented Diversified Total Country Year (Percentage of rural households in each group) Nigeria 2004 11 60 71 14 1 14 100 Madagascar 2001 -- -- 54 18 2 26 100 Agriculture- Ghana 1998 13 41 54 24 3 19 100 based Malawi 2004 20 14 34 24 3 39 100 countries Nepal 1996 17 8 25 29 4 42 100 Nicaragua 2001 18 4 21 45 0 33 100 Vietnam 1998 38 4 41 18 1 39 100 Pakistan 2001 29 2 31 34 8 28 100 Albania 2005 9 10 19 15 10 56 100 Transforming Indonesia 2000 -- -- 16 37 12 36 100 countries Guatemala 2000 4 7 11 47 3 39 100 Bangladesh 2000 4 2 6 40 6 48 100 Panama 2003 1 5 6 50 6 37 100 Urbanized Ecuador 1998 14 11 25 53 2 19 100 countries Bulgaria 2001 4 1 5 12 37 46 100 Source: Davis and others 2007. Note: Farm-oriented household: more than 75 percent of total income from farm production. Farm, market-oriented household: more than 50 percent of agricultural production sold on market. Farm, subsistence-oriented household: less than or equal to 50 percent of agricultural production sold on market. Labor-oriented household: more than 75 percent of total income from wage or nonfarm self-employment. Migration/transfers-oriented household: more than 75 percent of total income from transfers/other nonlabor sources. Diversified household: Neither farming, labor, nor migration income source contributes more than 75 percent of total income. -- = not available. strategies are more common, with shares of ute a large share of average labor income labor-oriented households varying from 18 (as in Indonesia, Pakistan, and Panama), percent in Vietnam to 53 percent in Ecua- while nonagricultural self-employment dor.13 Among these households, wages earnings are more important in labor- from nonagricultural labor often contrib- oriented households in Ghana and Viet- nam. In Bulgaria, Ecuador, and Nepal, agri- cultural wages are important for the income of labor-oriented households. Despite the B O X 3 . 2 Constructing comparable measures of income across countries importance of the labor pathway in trans- forming countries, market-oriented farm- The analysis of sources of rural income pre- methodologies would make it impossible ing households remain the largest rural sented here is based on income aggregates to compare results between different group in Vietnam. from the Rural Income Generating Activity countries. Even if most households are special- database. For each country the income While the standardized calculations components include wages (separately across countries enhance comparability, ized--that is, they derive the vast major- for agriculture and nonagriculture), self- the analysis of sources of rural income ity of their income from only one of the employment, crops, livestock, transfers, is constrained by the pervasive weak- three income sources (farming, labor, or and a final category of all remaining non- ness of the raw income data in many of migration)--a substantial remaining share labor income sources (excluding imputed the surveys analyzed. Many household rent), as reported in each country question- surveys likely underestimate income of households in all countries has diversi- naire. All aggregates are estimated in local because of underreporting, misreport- fied income strategies. In the 15 countries currency at the household level and annu- ing of the value of own consumption, of table 3.2, 14 to 56 percent of households alized and weighted. Some of the country income seasonality, and the difficulty do not derive more than 75 percent of their results may differ from results previously of obtaining reliable income data from published in poverty assessments and households that do not usually quantify income from one of these three sources, but other country reports because of efforts to their income sources. instead have a more mixed income portfo- ensure comparability across countries in See Davis and others (2007) and lio.14 These diversified households derive the results presented here. www.fao.org/es/esa/riga/ for further between 20 percent (in Bangladesh) and 46 Analyses that draw on income aggre- information on methodology. gates from different sources using different percent (in Ghana, Malawi, and Vietnam) of their income from farming. Rural households and their pathways out of poverty 77 Heterogeneity of the agriculture provides employment to 1.3 bil- household strategies lion people worldwide, 97 percent of them in developing countries.16 It is also a major A household's income structure does not source of income for rural households. tell whether it is engaged in a successful Between 60 and 99 percent of rural house- income strategy. Each of the strategies can holds derive income from agriculture in become pathways out of poverty, but many 14 countries with comparable data (figure households do not manage to improve their 3.2). In the agriculture-based countries in situation over time, reflecting the marked figure 3.2, farm crop and livestock income heterogeneity in each of the activities and and agricultural wages generated between the fact that income varies widely for each of the strategies (figure 3.1).15 42 and 75 percent of rural income. Onfarm income comes both from production for Rural occupations and self-consumption and from sales of agricul- tural products to the market. In the trans- income sources forming and urbanized countries, the share The heterogeneity in each of the household of rural income from onfarm activities and strategies reflects differences in the returns agricultural wages is between 27 and 48 on the various activities of rural households percent. So, participating in agricultural and individuals. The economic activities activities does not always translate into high and the sources of income themselves also agricultural income shares. differ substantially across regions, between For the poorest households, onfarm poor and rich households, between house- income and agricultural wages typically holds with different asset endowments, and account for a larger share of household between men and women. income, ranging from 77 percent in Ghana to 59 percent in Guatemala, than for richer Agriculture: a major occupation households (figure 3.3). In Asia, Latin for rural households, America, and some countries in Africa especially for the poor (Malawi and Nigeria), agricultural wages The Food and Agriculture Organization of are more important for low-income than the United Nations (FAO) estimates that for high-income households. Onfarm Figure 3.1 Real per capita income varies widely for each livelihood strategy Ecuador 1995 Nepal 1996 Frequency Frequency 0 0 ­1,000 0 1,000 2,000 3,000 4,000 ­2,500 0 2,500 5,000 7,500 10,000 Sucres (thousands) Rupees Market-oriented farming Labor-oriented Subsistence-oriented Migration-oriented farming Diversified Source: Davis and others 2007. 78 WORLD DEVELOPMENT REPORT 2008 Figure 3.2 In most countries, the vast majority of rural households participate in agriculture Percent 100 80 60 40 20 0 2004 2004 1998 1996 2000 2001 1998 2000 2005 2001 2001 2000 1998 2003 Malawi Nigeria Ghana Nepal Pakistan VietnamIndonesia AlbaniaBulgaria EcuadorPanama Bangladesh Nicaragua Guatemala Sub-Saharan South Asia East Asia Europe & Latin America Africa & Pacific Central Asia & Caribbean Participation in agriculture Income share from agriculture Source: Davis and others 2007. income often declines as overall expen- Yet asset endowments are not always ditures increase (in Ghana, Guatemala, good predictors of market orientation. Dif- and Vietnam, for example), but it is most ferences in land endowment between mar- important for households in the middle of ket- and subsistence-oriented farmers are the distribution of income in Nepal. much less pronounced in Bangladesh, Gua- In most countries, there is a marked temala, and Malawi. In Ghana and Nigeria, dualism in the smallholder sector, between female-headed households are more likely market-oriented farmers and smallholders to be market oriented than subsistence ori- engaged in subsistence farming. Only a very ented. This shows that market orientation small share of all marketed agricultural can also be conditioned by many other fac- products is produced by the subsistence- tors, such as land quality, access to markets, oriented households. In Malawi, subsistence or agricultural potential affecting crop and farmers sell about 9 percent of the marketed livestock choice and productivity. agricultural products, but in Nepal and Within the household, market orienta- Vietnam, less than 2 percent.17 The dual- tion can differ with the gender of the cul- ism in household farming strategies usually tivator, and women are often more likely to reflects differences in asset endowments. be engaged in subsistence farming and less Farmers with larger land endowments are likely to cultivate cash crops. Large-scale more likely to be market-oriented. Market- production of nontraditional and high-value oriented farmers own almost twice as much agricultural exports has, however, increased land as subsistence farmers in Nicaragua women's wage work in fields, processing, and and Panama, and four times more land in packing. This does not hold everywhere. In Pakistan. The human capital endowments China, for example, the evidence suggests of rural households are also correlated with there is no feminization of agriculture.18 their market orientation. Educated house- More generally, women's participation in hold heads are often more likely to sell a agricultural self-employment differs across large share of their products to the mar- regions. In Africa, Europe and Central kets, while female-headed households more Asia, and some East Asian countries, men often produce for self-consumption. and women work equally in agricultural Rural households and their pathways out of poverty 79 Figure 3.3 Sources of income vary between poor and rich Income, % 100 80 60 40 20 0 2nd 3rd 4th 2nd 3rd 4th 2nd 3rd 4th 2nd 3rd 4th Poorest Richest Poorest Richest Poorest Richest Poorest Richest Ghana 1998 Guatemala 2000 Vietnam 1998 Nepal 1996 Onfarm income Agricultural wage labor Nonagricultural income Transfers & other Source: Davis and others 2007. Note : For each country, columns represent the bottom fifth to the top fifth of the expenditure distribution. self-employment19 (figure 3.4). In Mozam- regions may diversify into nonagricultural bique, Rwanda, Uganda, and Egypt, women activities to take advantage of attractive are even more likely to participate in agri- opportunities. Those in less-favored envi- cultural self-employment. By contrast, ronments may shift into low-value nonag- in Latin America and South Asia, women ricultural activities to cope with the risks. reportedly work less in agricultural self- Households with good asset endowments employment. But in these regions, as well may seize remunerative opportunities in as in Africa, women have broadened and the nonfarm sector. Those lacking land or deepened their involvement in agricultural livestock may be driven into low-value non- production in recent decades.20 Yet many farm employment. Labor market income development policies continue to wrongly can also be important where population assume that farmers are men. The impor- pressures on limited land resources are high tant role of women in agriculture in many or where seasonal income from farming is parts of the world calls for urgent attention insufficient for survival in the off-season, to gender-specific constraints in produc- possibly because of chronic rainfall deficits, tion and marketing. prices, or diseases.21 Off-farm income can be important for both poor and rich households. Yet, the Income diversification rich often dominate lucrative business and specialization in wage niches. The poor, lacking access to capital, employment and nonagricultural education, and infrastructure, are not the self-employment main beneficiaries of the more lucrative Market-oriented smallholders can be highly sources of nonfarm income. This is, in part, successful in food markets and in the new because of the differential access to high- agriculture. But for many smallholders, skill and low-skill jobs (chapter 9). Illiterate agriculture is a way of life that offers secu- adults are more likely to be working in agri- rity and complements earnings in the labor cultural wage and self-employment. Liter- market and from migration. Other rural ate adults are more likely to have nonagri- households specialize in wage employ- cultural wage jobs. And older cohorts are ment or nonagricultural self-employment. less likely to be working in nonagricultural Households in prosperous agricultural wage employment than younger cohorts.22 80 WORLD DEVELOPMENT REPORT 2008 Figure 3.4 Women's reported participation in agricultural self-employment relative to men's varies by region Sub-Saharan Africa South Asia East Asia & Pacific Middle East & North Africa Europe & Central Asia Latin America & Caribbean 0 20 40 60 80 100 Percentage of men Agricultural self employment Agricultural wage employment Nonagricultural self employment Nonagricultural wage employment Sub-Saharan Africa South Asia East Asia & Pacific Middle East & North Africa Europe & Central Asia Latin America & Caribbean 0 20 40 60 80 100 Percentage of women Agricultural self employment Agricultural wage employment Nonagricultural self employment Nonagricultural wage employment Source : Regional averages based on available household surveys for 66 countries (ages 15 to 64). Note: The omitted group includes individuals out of the labor force and individuals whose economic activity is not defined. Activity refers to the individual's reported principal activity. For a more detailed explanation, please see endnote 19 on page 272. See also WDR 2008 team 2007. Exiting, coping, and acquiring Migration responds to income gaps capital through migration between the origin and the destination. Where access to nonagricultural employ- It can occur because people are pushed ment is limited or where the climate (or out of rural areas by negative shocks or a technology) prevents continual cultiva- deteriorating resource base--or are pulled tion, seasonal migration can supplement out by attractive employment opportuni- income, smooth consumption, and pro- ties elsewhere. In Chile, the local unem- tect household asset bases during the lean ployment rate is positively correlated with season. Laborers migrate seasonally to out-migration, but the expansion of agri- other regions in their own country, often cultural employment and jobs in agropro- attracted to large export crop estates that cessing slowed migration. Cohort analyses provide income in the off-season or during with population censuses between 1990 emergencies. They also migrate across bor- and 2000 for Ecuador, Mexico, Panama, ders, and a large part of south-south migra- and Sri Lanka suggest that people move tion is seasonal.23 out of localities that are more remote, with Where migration is more or less perma- less infrastructure, and with poorer living nent, income from migration depends on conditions. Yet areas with high agricultural the success of the migrant and the reason potential can also have high out-migration, for migration. So migration is not a guar- as in Guatemala. Rural migrants often go anteed pathway out of poverty (chapter 9). abroad or to urban areas that offer bet- Nor is it available to all. High migration costs ter income opportunities. However, many often prevent the poorest-of-the-poor from choose to migrate to urban areas that are migrating, or limit their migration to nearby relatively close by or move to other rural areas, where the returns might be low. areas (box 3.3).24 Rural households and their pathways out of poverty 81 B O X 3 . 3 The challenge of drastic demographic changes from selective migration Migration can be an important source of Almost a quarter of the 15­24 cohort from 1990 had left rural Mexico by 2000 remittance income (money sent home by household members who have left to find Urban population 1990­2000 by cohort Rural population 1990­2000 by cohort work), but it often drastically changes the Millions Millions composition of the rural population. This can 12 12 pose its own challenges for rural develop- ­2% ment, because migration is selective. Those 10 1990 2000 10 1990 2000 who leave are generally younger, better 4% educated, and more skilled.25 Migration thus 8 8 can diminish entrepreneurship and education level among the remaining population.26 In 6 6 ­3% ­24% addition to changing the skill and age com- ­6% 4 4 position of those staying behind, migration ­7% ­11% can change the ethnic composition of rural ­12% 2 2 populations. Migration rates of indigenous populations are often lower, because they are 0 0 attached to land as ancestral territories and 15­24 25­34 35­44 45­54 15­24 25­34 35­44 45­54 because they may be discriminated against Age in 1990 Age in 1990 in labor markets. There are also clear gen- Source: Buck and others (2007)), from information in the 10 percent sample of Mexico census (1990 and 2000). der differences in migration, but they differ Note: Columns represent the same cohort of people observed in the 1990 and 2000 censuses with a 10-year across countries, even within the same region. difference in age. The population reported for 1990 is corrected for location and age-specific mortality rate International migration out of rural areas is during the decade. The residual change is thus due to net out-migration. male-dominated in Ecuador and Mexico, but female-dominated in the Dominican Republic, migrate. People at all education levels moved than women (27 percent versus 21 percent) and Panama, and the Philippines.27 to both urban and rural areas, but the highly among nonindigenous than indigenous (25 Analyses of the population censuses of educated were much more likely to move to percent versus 18 percent). Until 2000 women Brazil and Mexico illustrate some of the regu- out-of-state urban centers (see figure below). were more prone to migrate to semiurban and larities. In Brazil between 1995 and 2000, rural Almost a quarter of those ages 15­24 in urban centers within the country, and men to men and women ages 20­25 were most likely 1990 had left rural Mexico by 2000, migrating to the United States. Indigenous migration has its to migrate, and young women migrated more urban centers or abroad (see the figure above). own dynamics, responding to seasonal agricul- than men (the first figure below). Illiterate indi- Among the older cohorts, migration was also tural cycles within Mexico, though international viduals were least likely to migrate, and highly high, reaching 6­12 percent. Rural emigration migration among indigenous groups steadily educated individuals were twice as likely to is much more common among Mexican men increased in the 1990s. Young Brazilian women migrate more than young men--and the less educated migrate less Share of migrants in each category, % Education level 25 Higher 20 Women, rural Secondary 15 Primary Men, rural 10 None Men, urban Women, urban 0 10 20 30 5 Share of rural population migrating, % To urban in municipality To urban in state 0 To urban out of state 15 25 35 45 55 65 To rural Age Source: Buck and others 2007; Lopez-Calva 2007; from information available in Brazil's 2000 census on residence in 1995. 82 WORLD DEVELOPMENT REPORT 2008 Income from remittances sent by for- Living in a poor area can itself be a causal mer household members often increases factor in perpetuating poverty because of the land, livestock, and human capital base geographical externalities.30 The strategies of rural household members who stayed of rural households are conditioned by the behind. Remittances can also offset income agricultural potential and natural resources shocks, protecting households' productive available in their environment (chapter 2). asset base. Evidence from the Oportunidades Recent work on the geography of poverty program in Mexico suggests that public sheds light on how these factors relate to transfers can similarly lead to investments household strategies and rural poverty (see in productive activities and risk coping.28 focus A). Population density and access to Private and public transfers account for markets, strongly correlated with transac- a surprisingly large share of rural income, tion costs and asymmetric information, also particularly in transforming and urbanized determine household strategies. With good economies. In some countries there have information, farmers are more equipped been major increases in transfers. In Bul- to make relevant decisions and learn about garia, households became more dependent additional diversified employment oppor- on public transfers as government spending tunities. New information technologies can on social protection rose to offset economic help address some of these information dis- hardships. In Brazil and Mexico, conditional advantages (chapter 7). cash transfers have become important for When market failures coincide, house- rural household income and are major con- holds need to consider their consumption tributors to rural poverty reduction. needs in making production decisions, and Urban-to-rural migration highlights vice versa. This can explain many aspects agriculture's role as a safety net, showing of rural households' livelihood strategies, that many urban residents are still part of including some that might otherwise appear a broader rural kinship network. During irrational.31 Consider a few examples. the 1997 financial crisis in Indonesia and Farm households that produce food and Thailand, and during the early transition cash crops will not always be able to respond years in the Caucasus and Central Asia, to an increase in the price of the cash crop. reverse migration helped people deal with When transaction costs in food markets are economic shocks. There is also evidence of high and labor markets function imperfectly, return migration in parts of Africa, related a household might not be able to employ to economic shocks and AIDS. Agriculture more labor to increase cash-crop produc- thus provides "farm-financed social wel- tion while maintaining the necessary food fare" when public welfare services are defi- production for its own food security.32 It is cient or nonexistent.29 thus confined to responding to price incen- tives through technological change or more use of fertilizer, but capital market imperfec- Household behavior when tions can limit these possibilities. As a result, markets and governments fail: the response to price incentives in cash crops rational, despite appearances is often limited, shrinking the benefits from Rural men and women determine their price and trade policies that increase pro- livelihood strategies in a context of failed ducer incentives (chapter 4).33 markets. Many markets in rural settings Market imperfections, combined with do not support efficient outcomes because differences in asset endowments, includ- of high transaction costs, insufficient and ing social capital, can also shed light on unequal access to information, imperfect technology adoption (chapter 7). Evidence competition, externalities, and state failures from Ghana, India, and Mozambique sug- to provide public goods. With such market gests that social learning may be important and state failures, initial asset endowments for adopting new technologies. Farmers' affect the efficiency of resource use and decisions are influenced by the experiences thus the well-being of households. of farmers in their social networks, which Rural households and their pathways out of poverty 83 can help reduce asymmetric information In many cases, collective action alone on the new technology. New technologies cannot correct market failures; that is a often involve uncertainties about appropri- crucial role for policies and the state. Yet ate application or suitability for a particu- in many developing countries, the state lar environment. Consequently, adoption has failed to play this role. To the contrary, patterns can be slow, as individual farm- many policies have been detrimental to ers gain from waiting and learning from rural households' livelihoods. Taxation of others' mistakes. Sometimes all farmers the agricultural sector, policy biases favor- can deem the evaluation costs too high or ing large farms, and failure to provide uncertain, choosing to stay with the status education and health services severely con- quo, behavior that can appear inefficient to strain the potential of rural households to an outsider. Recent evidence from Kenya pull themselves out of poverty through the suggests that households might also have a farming pathway. Reversing such policies saving commitment problem and thus do can enhance existing household strategies not put money aside after the harvest to or open the potential for new and success- buy fertilizer for the next season, another ful ones. explanation for the limited adoption of otherwise profitable strategies.34 Mutual influence of household The household is the domain of complex strategies and social norms interactions of cooperation and power plays. Social norms often have a strong influence A woman's power is affected by her partici- on household strategies and on the roles of pation in economic activity, which itself men and women in the household. In Côte depends on her asset endowment (including d'Ivoire, social norms not only dictate that human capital) and her access to the house- food crops should be grown by women hold's assets. Intrahousehold differences in and cash crops by men, but also influence control over assets and cash can thus affect the use of profits from different crops for cultivation and technology decisions, as well household expenditures.37 Social norms as a household's market orientation. A study often dictate that most of the childrear- in southern Ghana found that soil fertility, ing, cooking, and household chores are the tenure security of plots, and participation responsibilities of women, limiting their in the credit market were lower for women potential to take advantage of new farming, than for men; consequently, women were labor, or migration opportunities, reinforc- much less likely to plant pineapples than ing inequalities. Or increased labor force men. Pineapples were more profitable than participation by women, combined with the subsistence crops that women tended to these traditional roles at home, mean much cultivate. Evidence from Burkina Faso sug- longer workdays for women than for men. gests that output of crops grown by both Yet in some contexts women's wage men and women could increase by 6 percent jobs, and the income they generate, can if some labor and manure were reallocated shift the balance of power and work inside to women's plots.35 the house. Women's employment in the To the extent that these factors prevent growing export flower industry in Ecua- households from maintaining soil fertility dor increased the participation of men in or otherwise adopting sustainable practices, housework.38 Traditional time allocation they can have important repercussions for patterns can also be affected when house- natural resource management. Unsustain- holds move to more market-oriented cash able outcomes can also be the result of collec- crop production. Gender divisions between tive action problems, with the "tragedy of the crops can shift with new technology, as commons" looming where household liveli- occurred with rice growing in The Gambia. hoods depend on open access to resources In Guatemala, labor shortages associated (chapter 8). Empirical evidence suggests, with high-value export production forced however, that cooperative resource manage- women to reduce the time they devoted to ment often emerges in such settings.36 independent income-producing activities 84 WORLD DEVELOPMENT REPORT 2008 or to cultivating crops under their own con- Lacking a minimum asset endowment trol. Labor constraints also encroached on can thus trap households in long-term pov- the time that women could allocate to food erty. The asset endowments of many rural crops. Where men control income from households have been low for generations, cash crops, power imbalances in the house- explaining the persistence of rural poverty, hold can be reinforced when new market and the tighter asset squeeze on many small- opportunities open.39 Shifts in household holders challenges their survival. Increasing strategies that might lead to pathways out the asset base of the poor is a major chal- of poverty are not gender neutral. lenge for policy makers in implementing an agriculture-for-development strategy. Rural household asset positions: often low and unequal Human capital endowments Household asset positions determine Rural households' human capital endow- household productivity. More generally, ments tend to be dismally low. Rural-urban household asset endowments condition gaps in educational attainment and health livelihood strategies. Education and health outcomes remain large in most regions. status affect a person's potential to engage Regional averages for Sub-Saharan Africa, in high-value nonfarm jobs as well as the South Asia, and the Middle East and returns on agriculture. Education might North Africa show that rural adult males facilitate learning about new technologies, have about 4 years of education, and rural and given the physical intensity of most adult females have 1.5 to 4 years (figure agricultural labor, health and nutrition can 3.5). Only in Europe and Central Asia are affect agricultural productivity. The size education levels notably higher. Inequality and quality of landholdings condition crop in access to education by ethnic group is and technology choices and the potential of also high in many countries. Differences producing marketable surplus. Households between rural and urban areas are even without any access to land are excluded larger, with adult males in rural Africa and from the farming pathway. Owning work Latin America having about 4 years less animals can affect the timing of cultivation education than their urban counterparts practices. And livelihood strategies rely on (figure 3.5). social networks for trust, social learning, In some countries, such as Mexico, adult and collective action. education programs have boosted rural literacy rates. In many countries school enrollment rates have increased consider- Figure 3.5 Rural-urban gaps in educational attainment are large ably over the last decade. Yet differences in Years of education school attendance for children by wealth 12 categories and ethnic groups remain large, and gender differences are still significant 9 in most countries. In Latin America, the 6 returns to education were lower for indig- enous groups. Moreover, the quality of 3 education is often drastically lower in rural 0 areas (chapter 9).40 Access to quality health services is ­3 also much lower in rural areas. In many ­6 countries the imbalance between rural Europe & East Asia Middle East Sub-Saharan Latin America South Asia Central Asia & Pacific & North Africa Africa & Caribbean and urban areas in skilled health workers is extreme. In Africa only half the rural Rural men Rural women Rural-urban difference: men Rural-urban difference: women population has access to improved water Source: WDR 2008 team. or improved sanitation, and in Asia only 30 Note : Average education levels for adult populations, 25­64 years old, for countries in each region. Calculations percent.41 Poor health reduces agricultural based on 58 countries (excluding China and India) with recent household survey data with information on years of education, weighted by 2000 population. See Background Note by WDR 2008 team (2007) for details. productivity, and some agricultural prac- Rural households and their pathways out of poverty 85 tices contribute to health problems such as periods show declines in average farm sizes malaria, pesticide poisoning, and zoonotic and increases in landlessness. In many diseases (see focus H). high-population-density areas of Africa, AIDS takes a heavy toll on rural popu- average farm sizes have also been declining. lations in Africa, with mortality among Such land pressure in economies still heav- young adults rising sharply. Life expectancy ily reliant on agriculture is a major source is declining in many countries--in Malawi, of rural poverty, and it can also produce for example, from 46 years in 1987 to 37 social tensions contributing to civil con- years in 2002. HIV incidence early in the flict.46 This is true even if the division of epidemic is often higher for the educated, landholdings may have an equalizing effect, decimating human capital.42 AIDS also as the declining land Gini coefficients (less reduces adults' capabilities to work, diverts inequality) for India, Malawi, and Tanzania the labor of others to caregiving, and suggest (see table 3.3). breaks the intergenerational transmission By contrast, agricultural land is still of knowledge. All these factors can result in expanding in some African and Latin reduced agricultural production. Evidence American countries, and farm sizes are from rural Kenya suggests that antiretro- increasing (table 3.3 and chapter 2). In cash- viral treatment can sustain the adult labor cropping regions of Mozambique, such area force, leading to less child labor and better expansion was found to reduce poverty.47 child nutrition outcomes.43 Greater access to land for the rural poor, AIDS can also severely affect the particularly where off-farm income and demographic profile of rural populations migration opportunities are lacking, is a through the direct effects on mortality major instrument in using agriculture for and through migration that helps people development. cope. In its 2003 World Health Report, the In Latin America and some countries of World Health Organization (WHO) (2003) Africa and South Asia, unequal land access reported a shift of orphans to rural areas.44 is often perpetuated through social mecha- Analysis based on population censuses nisms--leaving many households, often suggests that African countries with high ethnic minorities or indigenous people, HIV prevalence (Botswana, Swaziland, and without access to land or with land plots Zimbabwe) have higher dependency ratios too small to meet their needs. Most of the than would be predicted for their level of land is in large farms, while most farms development.45 These changes in rural are small.48 This bimodal pattern has been household composition are likely to affect increasing in Brazil over the last 30 years, household income strategies, as well as the where the number of medium-size farms potential of rural households to benefit declined while the numbers of both small from agricultural and rural growth. The and very large farms increased. Small farms changes also have implications for the role control a declining share of the land, while of subsistence farming for household sur- large farms control a growing proportion vival (box 3.4). (figure 3.6). In Bangladesh the number of farms doubled in 20 years, and the number Land pressures and the persistence of farms smaller than 0.2 hectares increased of bimodal land distributions affect more than proportionally--but most of household landholdings the land is in larger farms.49 Moreover, As land gets divided through inheritance in a large share of rural households in these a growing population, farm sizes become regions do not have any access to land.50 smaller. In India the average landholding Land concentration thus contributes to the fell from 2.6 hectares in 1960 to 1.4 hect- asset squeeze on smallholders and landless ares in 2000, and it is still declining. Panel households. data that followed household heads and Mechanisms that perpetuate land their offspring in Bangladesh, the Philip- inequality include segmented land mar- pines, and Thailand over roughly 20-year kets when property rights are insecure, 86 WORLD DEVELOPMENT REPORT 2008 B O X 3 . 4 Returning to the farm in Zambia--subsistence agriculture, AIDS, and economic crisis Cohort analysis with the Zambia census data The most striking observation is the high by the attrition rates in both urban and rural sheds light on changes in the age composi- mortality rate between 1990 and 2000. areas, indicates high mortality. tion of the urban and rural populations in Because international migration is very low, In urban Zambia, large population declines a country with high HIV prevalence rates. the declining size of each cohort, indicated have occurred across all age groups, except the youngest. This contrasts with rural Zambia, where declines are especially large among Following 1990 population cohorts to 2000 shows high mortality rates, particularly young adults (19 percent for those 15­24 in among young adults 1990), indicating high mortality rates for this Population 1990­2000 by cohort, rural Population 1990­2000 by cohort, urban group.51 Similar population analysis also sug- Thousands Percent Thousands Percent gests higher mortality rates among the literate 50 50 population, confirming trends observed else- where in Africa. 2000 40 2000 40 Economic shocks that induced domes- 30 30 (right scale) (right scale) tic migration help explain the differences 1000 20 1000 20 between rural and urban patterns. In 2000 10 10 many more rural residents, of all age groups, reported having moved from the urban areas. 0 0 0 0 By contrast, fewer urban residents had rural 5­14 15­24 25­34 35­44 45­54 5­14 15­24 25­34 35­44 45­54 origins, particularly among older age groups Age in 1990 Age in 1990 (figure below). This indicates that net migra- tion reversed from rural-to-urban in 1990 to 1990 2000 Attrition rate urban-to-rural in 2000. Rural-to-urban migra- tion slowed considerably between 1990 and Source: WDR 2008 team, based on Zambia population census. Note: Columns represent the same cohort of people observed in the 1990 and 2000 censuses with a 10-year 2000, but urban-to-rural migration increased. difference in age. The attrition between the two observations includes both net out-migration and death. These patterns have been linked to the dearth Ages refer to cohort ages in 1990. of employment opportunities in towns and cities and the stagnation in the (largely urban) Migration patterns have reversed, with a recent increase in rural-to-urban migration copper mining industry triggered by a global slump in copper prices. Population 1990­2000 by cohort Population 1990­2000 by cohort Another explanation of the rural-urban born urban, rural now born rural, urban now differences in attrition rates among adults Thousands Thousands is return migration by HIV-affected people. 300 300 A higher proportion of rural households has 1990 2000 1990 2000 elderly household heads (12.9 percent versus 4.8 percent in urban areas). These households 200 200 rely more on subsistence agriculture and have considerably less access to income from non- 100 100 farm sources, including transfers, than other rural households. The majority of the rural 0 0 elderly households have (AIDS) orphans living 5­14 15­24 25­34 35­44 45­54 5­14 15­24 25­34 35­44 45­54 with them (on average, 0.8 orphans per elderly Age in 1990 Age in 1990 rural household). Note : Columns represent the number of people from the same cohort of age and born in urban (rural) areas Source: Potts 2005; World Bank 2005p; calculations that lived in rural (urban) areas in 1990 and 2000. The difference between the two observations includes of WDR 2008 team, based on Zambia population both migration and death. census. and unequal access to capital and other likely to own land, and female landowners input or output markets. More generally, tend to own less land than men. Evidence the inequality in many rural societies is from a sample of Latin American coun- perpetuated by elite capture in public ser- tries shows that only 11 to 27 percent of all vices; intergenerational transfers of poverty landowners are women. In Uganda women through low education, ill health, and poor account for the largest share of agricultural nutrition; and a deeply entrenched culture production but own only 5 percent of the of poverty (box 3.5).52 land, and they often have insecure tenure Women's access to land is often limited rights on the land they use.53 by unfavorable marital and inheritance Country examples shed light on some of laws, family and community norms, and the underlying mechanisms. Until a recent unequal access to markets. Women are less law change, a woman in Nepal could not Rural households and their pathways out of poverty 87 Table 3.3 Changes in farm size and land distribution Land distribution Average farm size Change in (Gini) (hectares) total number Change in Farm size of farms total area definition Country Period Start End Start End % % useda Smaller farm size, more inequality Bangladesh 1977­96 43.1 48.3 1.4 0.6 103 ­13 Total Pakistan 1990­2000 53.5 54.0 3.8 3.1 31 6 Total Thailand 1978­93 43.5 46.7 3.8 3.4 42 27 Total Ecuador 1974­2000 69.3 71.2 15.4 14.7 63 56 Total Smaller farm size, less inequality India 1990­95 46.6 44.8 1.6 1.4 8 ­5 Total Egypt 1990­2000 46.5 37.8 1.0 0.8 31 5 Total Malawi 1981­93 34.4 33.2b 1.2 0.8 37 ­8 Cultivated Tanzania 1971­96 40.5 37.6 1.3 1.0 64 26 Cultivated Chile 1975­97 60.7 58.2 10.7 7.0 6 ­31 Agricultural Panama 1990­2001 77.1 74.5 13.8 11.7 11 ­6 Total Larger farm size, more inequality Botswana 1982­93 39.3 40.5 3.3 4.8 ­1 43 Cultivated Brazil 1985­96 76.5 76.6 64.6 72.8 ­16 ­6 Total Larger farm size, less inequality Togo 1983­96 47.8 42.1 1.6 2.0 64 105 Cultivated Algeria 1973­2001 64.9 60.2 5.8 8.3 14 63 Agricultural Sources: Anríquez and Bonomi (2007). Calculations based on agricultural censuses. a. Total land area, agricultural (arable) land area, or cultivated (planted) crop area. b. Inequality obtained from the Malawi 2004/05 household survey. Figure 3.6 Farm size distributions are often bimodal Bangladesh India Brazil Percent of farms Percent of farms Percent of farms 50 50 20 40 40 1995 16 1977 30 30 12 1970 1996 20 20 1970 8 1996 10 10 4 0 0 0 0.01 0.1 1 10 0.01 0.1 1 10 100 0.1 1 100 1,000 100,000 Farm size, hectares (log scale) Farm size, hectares (log scale) Farm size, hectares (log scale) Bangladesh India Brazil Percent of land Percent of land Percent of land 60 25 20 1970 50 20 1977 15 40 15 1995 30 10 1996 10 1970 20 1996 5 5 10 0 0 0 0.01 0.1 1 10 0.01 0.1 1 10 100 0.1 1 100 1,000 100,000 Farm size, hectares (log scale) Farm size, hectares (log scale) Farm size, hectares (log scale) Source: Estimations based on agricultural census (Anríquez and Bonomi 2007). Note: Farm size in log scale. 88 WORLD DEVELOPMENT REPORT 2008 such contexts, household welfare depends B O X 3 . 5 New technologies and positive discrimination on herd size and the shocks that might policies reduce social inequalities in India affect it. The rapidly growing demand for livestock products in developing countries Inequalities across cultural, social, and India's recent shift to the panchayat reinforces the value of livestock as part of ethnic groups often reflect differences system of local government includes in access to economic opportunities. reserved council seats for women and household asset portfolios and its potential Consider the persistence of caste-based members of scheduled castes and tribes. to reduce poverty.55 inequalities in the Indian economy. The new emphasis on participatory and In 14 countries analyzed, the majority of Members of underprivileged "scheduled" community approaches has created possi- rural households own some livestock, with castes and tribes typically live in sub- bilities for marginal groups to gain power, habitations of a village, geographically challenging cultural norms while shifting shares above 80 percent in Albania, Ecuador, distinct from the main village. Residential structures of traditional authority. Nepal, and Vietnam. Even among the poor- segregation means that the public goods New technologies that link villages est households, more than 40 percent own consumed by members of scheduled with world production, consumption, and livestock, except in Pakistan. Many live- castes and tribes--such as sanitation governance further reduce the depen- facilities, drinking water, local roads, and dence on traditional norms. Television and stock holdings consist of small animal spe- even schools--are distinct from those communications have changed rural con- cies; fewer than 40 percent of rural house- consumed by better-off castes and are sumer preferences. Technological changes holds own cattle. The share of livestock generally of very poor quality. in agriculture, information technologies, owned by the top fifth of livestock holders Governments can reduce inequalities trade, and transportation have expanded by targeting funds toward areas popu- opportunities for many rural people. The varies between 42 percent and 93 percent, lated by the poor. Indeed, many Indian access to new knowledge does not neces- showing that livestock holdings tend to be government programs require funds to sarily correlate with traditional social hier- quite unequal. Indeed, these inequalities are be spent on scheduled-caste habitations. archies, so it can help break the traditional similar to those for landholdings.56 Recent research suggests that such man- inequality traps. But it can also lead to new dates ensure a higher level of investment inequalities as access to information and in poor habitations. However, it also shows capital come to matter more than tradi- Differential access to formal and that these policies cannot significantly tional norms. informal social capital reduce the prevailing bias of village gov- ernments to devote far more resources to Membership in formal and informal orga- the main village complexes. Sources: Kochar 2007; Rao 2007. nizations--and in community or ethnic networks--is a major asset of the rural poor, important for access to input and output markets, insurance, trust in transac- inherit land from her parents. In Malawi tions, and influence over political decisions. widows can lose their land from land grab- Social networks can also foster technology bing by the husband's family. Women's land adoption through social learning. Exclusion rights under customary tenure regimes are from such networks can severely limit the also much weaker than men's. Evidence choices of many, and the poorest are most from Ghana suggests that shifts to individ- likely to be excluded. Social capital is not ual ownership in such contexts can some- only important for farmers; it also deter- times strengthen women's land rights. Yet mines opportunities in the nonagricultural in other cases, titling programs, by con- sectors (for traders or for job referrals) and ferring titles to the male household head, for migration. For agricultural workers in contribute to the breakdown of custom- (often isolated) large estates in Sri Lanka ary systems that helped guarantee married and elsewhere, the lack of networks is a women's access to land.54 major constraint on upward mobility.57 Producer organizations can be part of the Livestock: a key asset for the social capital of many smallholders, con- poorest, particularly in arid tributing to smallholder competitiveness. and semiarid settings Between 1982 and 2002, the proportion of Livestock is often the largest nonland asset villages with a producer organization rose in rural household portfolios. In Burkina from 8 percent to 65 percent in Senegal and Faso and Ethiopia, livestock accounts for from 21 percent to 91 percent in Burkina more than half of rural households' wealth. Faso. Overall, 69 percent of Senegal's rural In arid and semiarid settings of Africa and households and 57 percent of Burkina Asia, livestock can offer the only viable Faso's are now members of producer orga- household agricultural strategy (box 3.6). In nizations. Data for other African and Latin Rural households and their pathways out of poverty 89 American countries, although fragmented, also indicate a rapid increase in the number B O X 3 . 6 Pastoralists' precarious livelihoods of such local organizations.58 Pastoralism and agropastoralism are the linked to weather conditions and thus are Exclusion from formal networks typi- main agricultural production systems in particularly vulnerable. cally affects women more than men, and dryland areas, supporting the livelihoods Pastoral strategies of herd diversity, women are less likely to be members of of 100 to 200 million people worldwide. flexibility, and mobility reflect rational producer organizations, their member- The number of extremely poor pastoralists and crucial survival mechanisms in erratic and agropastoralists is estimated at 35 to environments. Such strategies can be ship constrained by cultural norms. But 90 million. More than 40 percent of the enhanced by policy, and some Sahelian there are exceptions. In Senegal women pastoralists live in Sub-Saharan Africa, 25 countries (Burkina Faso, Mali, Mauritania, participate more than men in producer percent in Middle East and North Africa, and Niger) have been promoting policy 16 percent in East Asia, 8 percent in South organizations. In Bangladesh and India, reforms aimed at legally recognizing the Asia, and 4 percent each in Latin America rights of pastoralists and improving the self-help and microlending groups consist and in Europe and Central Asia. management of rangeland resources. But primarily of women. In Andhra Pradesh, Itinerant herding, moving animals recent efforts to set aside extensive areas poverty-reduction programs reaching from place to place to follow water and of marginal lands as national parks and pasture availability, has evolved over cen- biodiversity reserves, particularly in Africa, more than 8 million women have built turies and is well suited to sustaining life pose new challenges to pastoralism. on and enhanced such self-help groups, in areas where rainfall is unpredictable. Sources: Blench 2001; Rass 2006; Thornton and increasing the access to group loans and Yet, pastoralist livelihoods remain closely others 2002. collective marketing for agricultural com- modities and input supplies.59 also worry about abrupt changes in rules Pervasive risks and for land tenure or regulations for trade; for costly responses them, the state can be an additional source Agriculture is one of the riskiest sectors of uninsured risk. Rural political violence of economic activity, and effective risk- and crime can also cause considerable farm reducing instruments are severely lacking productivity losses, as in Colombia.61 in rural areas. Negative shocks can deplete The lack of access to insurance and credit assets through distress sales of land and markets makes agricultural producers par- livestock. It can take a very long time for ticularly vulnerable. Households thus often households to recover from such losses. reduce their consumption risk by choosing When income and asset shocks coincide, low-risk activities or technology, which households have to choose between reduc- typically have low average returns. In rural ing consumption or depleting assets.60 areas of semiarid India, such self-insurance This suggests a role for policies to enhance produces returns for the poor that are 35 household's ability to manage risk and to percent lower than if they did not need to cope when hit by a shock. self-insure.62 Rural households often identify weather- Shocks can be idiosyncratic--when one related and health shocks as their biggest household's experience is weakly related, if risks. The immediate production and wel- at all, to that of neighboring households-- fare losses associated with drought can or covariate--when households in a same be substantial. In Kilimanjaro, Tanzania, geographical area or social network all suf- farmers who reported rainfall patterns well fer similar shocks. Idiosyncratic shocks can below normal in the year prior to the sur- arise from microclimatic variation, local vey experienced a 50 percent reduction in wildlife damage or pest infestation, ill- their agricultural revenues and a 10 per- ness, and property losses from fire or theft. cent reduction in their consumption. Ill- Such shocks can, in principle, be managed nesses and injuries in a family simultane- by insurance within a locale. By contrast, ously reduce income because of lost time covariate shocks, arising from war, natu- working and deplete household savings ral disasters, price instability, or finan- because of spending on treatment. Studies cial crises, are difficult to insure locally for Africa, Asia, and Latin America suggest and require some coordinated external that health shocks contribute to more than response. Yet, even idiosyncratic risk often half of all descents of previously nonpoor has large effects, indicating the potential households into chronic poverty. Farmers for better local risk management. 90 WORLD DEVELOPMENT REPORT 2008 Are agricultural risks increasing? Recent herd size recovered to 50 percent of pre- empirical evidence suggests that heightened drought levels, but there was little recovery volatility attributable to apparent increases for households that lost their entire breed- in climate variability (drought, flooding, ing stock.65 and other natural disasters) has been off- Coping with shocks often comes at the set by reduced volatility from greater use of expense of investments in the next genera- irrigation and livestock.63 Yet the costs of tion. In addition to the higher infant mor- each meteorological event or other natural tality rate in drought years, survivors are disaster are rising, reflecting the expansion often stunted, which in turn affects future of population and cultivation into more educational attainment and lifetime earn- vulnerable areas. Moreover, the economic ings. Rural households often also respond costs of extreme weather events increase as to low rainfall or unemployment shocks production systems use more capital, unless by withdrawing children from school or that capital allows the use of risk-reducing decreasing their attendance so that they technology. Higher investments can thus can help at home and on the farm. Children increase asset-risk exposure, one obstacle to taken out of school for even a short period expanding credit use by poor households. are much less likely to return to school.66 This also helps explain why many farm- Negative shocks can have differential ers who are not poor remain vulnerable effects along gender lines, and women (or to shocks in the absence of risk-mitigating girls) in poor households often bear the measures. largest burden. Meeting current consump- Poor areas generally are also riskier. tion after a shock can also degrade the Prices tend to be more variable in more environment at a cost of future livelihoods. remote areas, often the poorest regions, Shocks can intensify pressures on common because limited market access and greater property, increase poaching and encroach- costs of getting to market make it more ing on protected areas, and augment con- difficult to offset local supply and demand flicts between pastoral and farming com- shocks. Poor households also have fewer munities.67 So protecting rural households means to insure against bad weather, and against uninsured risks is an area for greater they face more weather-related disasters-- policy attention (chapter 6). aggravated by inequality in the coverage and effectiveness of infrastructure. People in Smallholder challenges low-income countries are four times more to compete likely to die in natural disasters than those The potential of agriculture to contribute in high-income countries.64 Uninsured to growth and poverty reduction depends risks and poverty can thus create downward on the productivity of small farms. The vast spirals of perpetual impoverishment. majority of farmers in developing coun- tries are smallholders, and an estimated 85 Lack of insurance and percent of them are farming less than two asset depletion hectares. In countries as diverse as Bangla- The inability to protect a household from desh, China, Egypt, and Malawi, 95 percent income and asset shocks can result in long- of farms are smaller than two hectares, and term consequences across generations in many other countries the great major- through reduced investments in health, ity of farms is under two hectares.68 The nutrition, and schooling. In many circum- literature linking household's asset endow- stances, recovering from a shock is slow ments to agricultural productivity has long and often incomplete by the time the next emphasized an inverse relationship between shock occurs. And after an income shock, farm size and factor productivity. Both the poor recover more slowly than the non- theory and empirical evidence have shown poor. Households in an isolated community that such a relationship is common when in Zimbabwe lost 80 percent of their cattle imperfections in both land and labor mar- in the 1992 drought. By 1997, the average kets are large.69 The inverse relationship is Rural households and their pathways out of poverty 91 a powerful rational for land access policies While smallholders have an advantage in that redistribute land toward smallholders, overcoming labor supervision problems, increasing both efficiency and equity. other factors can erase their competitive Smallholder farming--also known as advantage. Yields on land allocated to crops family farming, a small-scale farm operated might be higher on larger farms, which tend by a household with limited hired labor-- to apply more fertilizer or other inputs. remains the most common form of orga- And the gap might be increasing over time. nization in agriculture, even in industrial For example, gains in cereal yields on small countries. The record on the superiority of farms are lagging behind gains on larger smallholder farming as a form of organi- farms in both Brazil and Chile (figure 3.7). zation is striking. Many countries tried to Yield gaps can arise because imperfec- promote large-scale farming, believing that tions in credit and insurance markets pre- smallholder farming is inefficient, back- vent small farmers from adopting more ward, and resistant to change. The results productive capital-intensive techniques or were unimpressive and sometimes disas- higher-value products. Evidence from Brazil trous. State-led efforts to intensify agricul- indicates that price changes following mar- tural production in Sub-Saharan Africa, ket liberalizations favored technologically particularly in the colonial period, focused more advanced producers who were better on large-scale farming, but they were not able to cope with price and yield variability sustainable. In contrast, Asian countries and deal with the demands of agroprocess- that eventually decided to promote small ing. Imperfections in capital and insurance family farms were able to launch the markets, combined with transaction costs, green revolution. They started support- can also prevent markets for land sales and ing smallholder farming after collective rentals from allocating land to the most farms failed to deliver adequate incentives efficient users.70 Moreover, imperfect com- to produce, as in China's farm collectiviza- petition in those markets might favor land tion, or on the verge of a hunger crisis, as concentration in larger farms. These com- in India and Indonesia. Countries that pro- plexities indicate the need to jointly con- moted smallholder agriculture--for vari- sider policies targeting land, capital, and ous political reasons--used agriculture as risk for smallholders (chapter 6). an engine of growth and the basis of their Moreover, while there may be constant industrialization. returns to scale in production, economies of Even if small farmers use their resources scale in the "new agriculture" often are the more efficiently than larger farmers, there key for obtaining inputs, technology, and may still be disadvantages in being small. information and in getting products to the Figure 3.7 Yields on small farms lag behind large farms in staples in Brazil and Chile Brazil Chile Maize yield, tons per hectare Wheat yield, tons per hectare 4 6 5 3 1996 4 1997 2 3 1980 2 1 1 1976 0 0 1 10 100 1,000 1 10 100 1,000 10,000 Farm size (harvested area), hectares (log scale) Farm size (harvested area), hectares (log scale) Source: Anríquez and Bonomi 2007. 92 WORLD DEVELOPMENT REPORT 2008 market (chapter 5). As agriculture becomes to become market participants in staples more technology driven and access to con- and high-value crops. sumers is mediated by agroprocessors and Smallholders can act collectively to supermarkets, economies of scale will pose overcome high transaction costs by form- major challenges for the future competi- ing producer organizations (chapter 6). tiveness of smallholders. Cooperation between larger commercial These different mechanisms can all farmers and smallholders is another pos- reverse the small farm labor advantage, or sibility. Smallholders sometimes can also make it irrelevant, leading to a potential benefit from economies of scale in input or decline of the family farm (box 3.7). The output markets by renting out their land perceived "crisis" in smallholder agricul- and working on the larger farms.71 Increas- ture is epitomized by the rash of suicides ing the bargaining power of smallholders in of heavily indebted farmers in India, the this type of arrangement can help guaran- long-term stagnation of productivity of tee that benefits are shared by smallholders food crops in Africa, the role of poor (indig- and the larger farms. enous) farmers in the political instability of many Latin American countries, and the Conclusions increasing rural-urban income disparities Three powerful and complementary path- in South and East Asia. But there are many ways out of poverty are smallholder farm- policy instruments to help smallholders ing, off-farm labor in agriculture and the increase their competitiveness, as long as rural nonfarm economy, and migration. governments do not tilt the playing field The following chapters discuss policies and against them. programs that can open and widen these pathways for the rural poor by increasing Smallholder entrepreneurs their asset holdings and by improving the and cooperation context that determines the level and vola- Heterogeneity in the smallholder sec- tility of the returns on assets. Chapters 4 to tor implies that a group of entrepreneur- 8 explore how farming can be made more ial smallholders is likely to respond when effective in providing a pathway out of pov- markets offer new opportunities. Improved erty. Chapter 9 looks into the possibilities access to assets, new technologies, and bet- offered by the agricultural labor market, the ter incentives can allow more smallholders rural nonfarm economy, and migration. B O X 3 . 7 Are farms becoming too small? Population pressures, unequal landholdings, A related question is whether declining small, inefficient farms. In China, greater ten- and inheritance norms favoring fragmenta- farm sizes widen rural-urban income gaps. ure security has been advocated to facilitate tion are leading to rapid declines in farm sizes With urban wages increasing in many Asian moves to the nonfarm economy. Without such in many parts of Asia and Africa. In China and countries, labor productivity in agriculture a policy change, the trend of declining farm Bangladesh, average farm size is about 0.5­0.6 might have to increase to avoid widening the sizes in China might continue. hectares, and in Ethiopia and Malawi about 0.8 gap. One way of achieving such productivity In other places, policy-led land consolida- hectares. Have farms become "too small"? gains might be through farm consolidation tion has been considered. The advantages The farm-size debate is motivated by a and mechanization. are not always clear, however, because some number of concerns. First, some argue that the Policies activating land rental and sales households will lose their access to land.72 But inverse farm size­efficiency relationship might markets can promote such consolidation. where consolidation occurs through the land not hold at very small farm sizes, or that even Increases in land inequality and landlessness rental market, win-win situations can occur. if such farms are efficient, they might be too can then coincide with a pro-poor process of Alternatively, increasing the productivity of small for rural households to escape poverty change, as in Vietnam, where rural economic small farms--through high-value crops or based on the income of the farm alone. Others development and greater diversification in the higher-yielding technologies for food crops-- argue that small farms disguise unemployment sources of income sharply reduced poverty. can increase the incomes from small farms. if labor markets do not work properly. The rel- Conversely, tenure insecurity can prevent land evance of these arguments depends in part on reallocation through sales or rental markets, Sources: Anríquez and Bonomi 2007; Deininger and the availability of alternative income sources preventing such gains. In Japan, government Jin 2003; Otsuka 2007; Ravallion and van de Walle and on the safety-net value of small farms. intervention in land rental markets preserves forthcoming. Rural households and their pathways out of poverty 93 The heterogeneity of smallholders, some wide range of institutions that affect com- market oriented and some subsistence ori- petitiveness are creating new challenges for ented, calls for differentiated agricultural smallholder entrepreneurs. They are also policies that do not favor one group over opening new opportunities. By addressing the other, but that serve the unique needs these challenges and seizing these oppor- of all households while speeding the pas- tunities, smallholders can escape poverty sage from subsistence to market-oriented through the farming pathway, especially farming. Recent changes in the global food when policies reverse traditional biases market, in science and technology, and in a against the smallholder. focus C What are the links between agricultural production and food security? Today, the world has more than enough food to feed everyone, yet 850 million are food insecure. Achieving food security requires adequate food availability, access, and use. Agriculture plays a key role in providing (1) food availability globally (and nationally and locally in some agriculture-based countries); (2) an important source of income to purchase food; and (3) foods with high nutritional status. I n the mid-1970s, as rapidly increasing Secure world, insecure households the 1970s, 37 million people were removed prices caused a global food crisis, food The world is generally food secure, produc- from the ranks of the undernourished, and security emerged as a concept. Attention ing enough food to meet the dietary needs 100 million in the 1980s, but in the 1990s, focused first on food's availability but then of today's global population--although only 3 million were removed. quickly moved to food access and food use-- future global food security should not be What accounts for these millions of and, most recently, to the human right to taken for granted because of uncertainties food-insecure individuals? Food security adequate food. The International Covenant from growing resource scarcity and cli- depends on adequate and stable food avail- on Economic, Social, and Cultural Rights, mate change (chapter 2). Yet 850 million ability, access to adequate and appropriate ratified by 153 states, obligates these states people remain undernourished.2 Accord- food, and proper use and good health to to progressively realize the right to food. ingly, the first Millennium Development ensure that individual consumers enjoy the The commonly accepted definition of Goal includes the target of halving hunger full nutritional benefits of available, acces- food security is-- as tracked by the measure of undernour- sible food. Availability is necessary but not enough to ensure access, which is necessary when all people, at all times, have physi- ishment given by the Food and Agriculture but not enough for effective use. cal, social, and economic access to suf- Organization of the UN (FAO) .3 ficient, safe, and nutritious food to meet The highest incidence of undernourish- Food availability--producing their dietary needs and food preferences ment is in Sub-Saharan Africa, where one for an active and healthy life.1 in every three persons suffers from chronic enough to eat hunger (figure C.1). The greatest number The price increases in the mid-1970s world The chronically food insecure never of undernourished is in South Asia (299 food crisis were exacerbated by low foreign have enough to eat. The seasonally food million), closely followed by East Asia (225 exchange reserves, limiting food imports insecure fall below adequate consumption million). in many food-deficit countries. This rise levels in the lean season. And the transitory East Asia has reduced the prevalence in prices prompted some countries to look food insecure fall below the food consump- of undernourishment in the past decade inward, striving for food self-sufficiency tion threshold as a result of an economic or by more than 3 percent a year and South through domestic production. But today natural shock such as a drought, sometimes Asia by 1.7 percent a year, but the failure with deeper international markets, lower with long-lasting consequences. to reduce the absolute number of under- real prices, and more countries with con- Investments in agriculture are impor- nourished remains a cause for concern. In vertible exchange rates, trade can stabilize tant to increase food security. The channels are complex and multiple. Rising produc- Figure C.1 Undernourishment is highest in Sub-Saharan Africa tivity increases rural incomes and lowers food prices, making food more accessible Prevalence of undernourishment, 2003 (% of total population) to the poor. Other investments--such as 40 improved irrigation and drought-tolerant SSA crops--reduce price and income variability by mitigating the impact of a drought. Pro- 206 30 SA ductivity gains are key to food security in countries with foreign exchange shortage or limited infrastructure to import food. The 299 20 same applies to households with poor access EAP to food markets. Nutritionally improved LAC MENA crops give access to better diets, in particular 225 10 52 36 ECA through biofortification that improves crop 27 nutrient content. The contributions that agriculture makes to food security need to 0 be complemented by medium-term pro- ­4 ­3 ­2 ­1 0 1 2 Annual average change in prevalence of undernourishment, 1992­2003, % grams to raise incomes of the poor, as well as insurance and safety nets, including food Sources: http://faostat.fao.org; FAO 2006c. aid, to protect the chronic and transitory Note: The size of bubbles represents millions of undernourished people. EAP--East Asia and Pacific, LAC--Latin America and Caribbean, SA--South Asia, SSA--Sub-Saharan Africa, MENA--Middle East and North Africa, poor (chapter 9). ECA--Europe and Central Asia. What are the links between agricultural production and food security? 95 food availability and prices for most coun- than 40 percent in Guinea-Bissau, Haiti, and ran Africa. Yet it remains home to 210 mil- tries (chapter 5). And most countries have the Republic of Yemen. With such levels of lion undernourished people and 39 percent diversified their export base, increasing their dependency and food imports often repre- of the world's underweight children.8 Ban- capacity to import. senting more than 20 percent of the avail- gladesh, India, and Nepal occupy three of the However, food availability is still a concern able foreign exchange, world price fluc- top four positions in the global ranking of in some agriculture-based countries. Many tuations place additional strain on import underweight children. Ethiopia is the fourth, countries have declining domestic production capacity and therefore domestic food avail- with the same incidence of underweight per capita of food staples.4 Burundi, Ethiopia, ability. World price variability remains high, children as India. Many believe that the infe- Kenya, Madagascar, Nigeria, Sudan, Tanzania, with a coefficient of variation of around 20 rior status of women in South Asia has to and Zambia all had negative per capita annual percent. some extent offset the food security benefits growth rates in staple food of ­1.0 to ­1.7 Because of the low price elasticity of of agriculture-led poverty reduction. percent from 1995­2004. In addition, staple demand for food staples and the thinness of food production in many agriculture-based markets, problems in food availability (from Food use--ending hidden hunger countries is largely rain fed and experiences low domestic production or lack of imports) Food use translates food security into nutri- large fluctuations caused by climatic variabil- translate into large spikes in domestic prices tion security. Malnutrition has significant ity. In Sudan, for example, the coefficient of and reductions in real incomes of poor con- economic consequences, leading to estimated variation of domestic staple food production sumers (many of whom are farmers). Even individual productivity losses equivalent to is 25 percent. This means that a shortfall of at in countries that engage in trade, transpor- 10 percent of lifetime earnings and gross least 25 percent of average production occurs tation and marketing costs result in a large domestic product (GDP) losses of 2 to 3 every six years. And many other countries wedge between import and export parity percent in the worst-affected countries.9 But have similarly high coefficients: Niger and within which domestic prices can fluctuate malnutrition is not merely a consequence Malawi at 18 percent; Rwanda at 15 percent; without triggering trade. Price variability, of limited access to calories. Food must not and Burkina Faso, Chad, Kenya, Uganda, and which is already high even in capital cities only be available and accessible, but also be the Republic of Yemen above 10 percent. with mostly liberalized markets, is exacer- of the right quality and diversity (in terms Stagnation or decline in domestic pro- bated in inland and more remote regions. of energy and micronutrients), be safely pre- duction and large fluctuations clearly raise pared, and be consumed by a healthy body, a potential problem of food availability Food access--having enough as disease hinders the body's ability to turn at the national level. Can this problem be to eat food consumption into adequate nutrition. addressed through imports? In many coun- But for most of the malnourished, the lack Lack of dietary diversity and poor diet tries the answer is yes. In other countries, of access to food is a greater problem than quality lead to micronutrient malnutrition or however, the main staples consumed have food availability. Nobel Laureate Amartya hidden hunger,10 even when energy intakes a low degree of tradability and are hardly Sen famously wrote that "starvation is a are sufficient. Hidden hunger can cause ill- traded internationally (chapter 1). Poor matter of some people not having enough ness, blindness, and premature death as well infrastructure imposes high costs for food food to eat, and not a matter of there being as impair the cognitive development of survi- to reach isolated areas, even when the capi- not enough food to eat."5 The irony is that vors. In the next 12 months, malnutrition will tal city and coastal cities are well served by most of the food insecure live in rural areas kill 1 million children before the age of five.11 international markets. where food is produced, yet they are net Iron deficiency among female agricultural Beyond tradeability issues--with ade- food buyers rather than sellers (chapter 4). workers in Sierra Leone will cost the economy quate infrastructure and internationally Poverty constrains their access to food in the $100 million in the next five years.12 traded staples--low foreign exchange avail- marketplace. According to the UN Hunger Although increased production of hor- ability often limits the capacity to import. Task Force, about half of the hungry are ticulture products and livestock has been Consider the case of Ethiopia that would smallholders; a fifth are landless; and a tenth agriculture's main avenue to improve diet import on average 8 percent of its staple are agropastoralists, fisherfolk, and for- quality, agriculture now offers an additional food consumption (assuming no food est users; the remaining fifth live in urban pathway to address hidden hunger. Biofor- aid) to maintain current levels. Addition- areas.6 Today, agriculture's ability to gener- tification is enhancing staple crop varieties ally, a 9 percent shortfall in production, ate income for the poor, particularly women, and improving diet quality with higher levels which occurs on average every six years, is more important for food security than its of vitamins and minerals through conven- could only be compensated by a doubling ability to increase local food supplies. Women, tional crop-breeding and biotechnology. of imports. But in the absence of food aid, more than men, spend their income on food. In the future, agriculture will continue to Ethiopia would already be spending 16 per- In Guatemala, the amount spent on food in play a central role in tackling the problem of cent of its foreign exchange earning on food households whose profits from nontradi- food insecurity. It can maintain and increase imports, leaving little scope for the neces- tional agricultural exports were controlled global food production, ensuring food avail- sary increases in imports. by women was double that of households ability. It can be the primary means to gener- Almost all the agriculture-based coun- whose men controlled the profits.7 ate income for the poor, securing their access tries are net importers of food staples, India has moved from food deficits to to food. And through new and improved importing on average 14 percent of their food surpluses, reducing poverty signifi- crop varieties, it can improve diet quality total consumption over the past 10 years, cantly and reaching a per capita income and diversity and foster the link between but reaching high dependency levels of more higher than that in most parts of Sub-Saha- food security and nutrition security. PART II Reforming trade, price, and subsidy policies What are effective instruments for using agriculture for development? 4 Agriculture is a cause of contention in countries, these economic and social costs international trade negotiations as well as in remain significant and perpetuate global domestic debate on price and subsidy poli- income disparities. Correcting those pol- c h a p t e r cies. It is often the cause of delays to multi- icy and investment failures can accelerate lateral trade negotiations, as in the Uruguay growth and reduce poverty. and Doha Rounds; is a source of political This chapter reviews the recent policy tension, especially in transforming coun- shifts across developed and developing tries; and is a challenging area for policy countries; the potential gains from further dialogue with development partners, par- reforms; who gains and loses from reform; ticularly in the poorest countries. Reforms and the pace, sequencing, and complemen- are usually politically sensitive with strong tary support needed in advancing these vested interests and, hence, are often diffi- reforms to enhance growth and reduce cult to achieve. Yet significant gains can be poverty. The political economy framework made from further agricultural trade, price, from chapter 1 helps in understanding the and subsidy policy reforms. Such gains determinants of policy choices for selected will not come easily, however, for reforms cases--and the ways to further improve require addressing the political economy of trade and price incentives and the efficiency difficult policy choices. There will be both of public spending. gainers and losers from reforms. Agricultural policies vary widely across Agricultural protection and countries. They have historically tended to shift from net taxation to subsidies as a subsidies in developed countries country's per capita income rises (chapter Much attention has been given to reducing 1).1 Low-income countries tend to impose the negative impacts of developed country relatively high taxes on farmers in the policies on developing countries--par- export sector as an important source of fis- ticularly through efforts to open markets cal revenue, while developed countries tend and to remove developed-country subsidy to heavily subsidize farmers. These differ- policies that have induced production and ences often create a policy bias against the depressed world prices (box 4.1). Rising poor in both domestic and international agricultural protection in developed coun- markets. tries and concerns about its impact on The economic and social costs of today's poorer developing countries spurred inter- trade, price, and subsidy policies in world national efforts in the 1980s to reduce dis- agriculture are large. They depress inter- torted prices in world markets. At the start national commodity prices by about 5 per- of the Uruguay Round of trade negotiations cent on average (much more for some com- in 1986, some agricultural exporting coun- modities) and suppress agricultural output tries formed the Cairns Group and ensured growth in developing countries. They con- that members of the General Agreement on sume a large share of the government bud- Tariffs and Trade put agricultural trade and get and distract from growth-enhancing subsidy reform high on the Uruguay Round investments. Although reduced over the agenda. Developing countries also formed last two decades, especially in developing the G-20 group at the time of the Cancun 96 Reforming trade, price, and subsidy policies 97 Ministerial conference in the Doha Round in 2003 to secure reductions in developed- B O X 4 . 1 Types of instruments that distort trade country protection. Three main types of instruments distort of the costs of exporters such as market- trade: market access, export subsidies, and ing expenses, special domestic transport Reform progress is slow, with little domestic support. charges, and payments to domestic change in overall support Market access: These include import exporters to make sourcing products from tariffs and quotas that protect local pro- domestic producers competitive. Member countries of the Organisation for ducers from competing imports. Protec- Domestic support: These include Economic Co-operation and Development tion induces local production to be higher direct support to farmers linked to the (OECD) are reforming their agricultural than would be the case at market prices, type, price, and volume of production. policies, but progress is slow. The average at the expense of international producers Depending on the level of support, local and exporters. production is usually higher and compet- support to agricultural producers fell from Export subsidies: These include ing imports lower than in the absence of 37 percent of the gross value of farm receipts government payments that cover some subsidies. in 1986­88 (the beginning of the Uruguay Round) to 30 percent in 2003­05. This esti- mate, referred to as the producer support estimate (PSE), measures the annual mon- Figure 4.1 Progress has been slow in reducing overall support to agricultural producers in etary value of gross transfers from consum- the OECD, but there has been some move to less-distorting "decoupled" payments ers and taxpayers to agricultural producers, Producer support estimate, %a measured at the farmgate level as a share of 80 70 the gross value of farm receipts. It arises 64 62 from policy measures that support agricul- 60 58 ture, regardless of their nature, objectives, or impacts on farm production or income.2 41 40 37 34 While the 7-percentage-point decline in 30 support is progress, the amount of support 22 increased over the same period from $242 20 16 8 billion a year to $273 billion. 13 5 3 8 1 2 2 0 3 8 6 More than 90 percent of the dollar value 2 0 2 OECD EU Japan Rep. of Korea United States Australia of agricultural support in OECD countries is provided by the European Union (which PSE due to coupled payments 1986­88 PSE due to coupled payments 2003­05 alone provides about half); Japan; the PSE due to `decoupled payments' 1986­1988 PSE due to `decoupled payments' 2003­05 United States; and the Republic of Korea. In Source: OECD 2006b. all four, the PSE remains high (figure 4.1).3 Note: Coupled payments include market price supports and payments tied to output level and input use. OECD countries include EU, Australia, Canada, Iceland, Japan, Republic of Korea, Mexico, New Zealand, Norway, In contrast, two OECD countries--Austra- Switzerland, Turkey, and the United States. lia and New Zealand--provide little sup- a. Transfers to agricultural producers as a share of the gross value of farm receipts. port to their farmers. OECD countries have increased preferen- Price support to farmers in OECD coun- tial access to their markets for some devel- tries creates incentives to produce more. oping countries. For example, in 2000, the The recent shift to separate or decouple United States signed the African Growth support from the type, volume, and price and Opportunity Act, which offers preferen- of products is an effort to reduce the trade- tial access to Africa's products in U.S. mar- distorting effects on current or future kets. The EU continues to provide extensive production while maintaining support to nonreciprocal preferential market access to farmers. Twenty-eight percent of the PSE countries in Sub-Saharan Africa, the Carib- in 2003­05 was decoupled from produc- bean, and the Pacific under the Cotonou tion and input use, up from 9 percent in Agreement. In 2001 the EU also provided 1986­88 (figure 4.1). duty-free and quota-free access to its markets Decoupled payments are less distorting to UN-designated Least Developed Coun- than output-linked forms of support such tries for "Everything But Arms," although it as tariff protection, but they can still influ- excluded services and delayed opening sensi- ence production. They can reduce farmers' tive markets for bananas, rice, and sugar. aversion to risk (wealth effect) and reduce 98 WORLD DEVELOPMENT REPORT 2008 the variability in farm income (insurance Agricultural taxation effect). Banks often make loans to farmers in developing countries that they would not make to other borrow- ers, keeping farmers in agriculture.4 Policies in developing countries have also blunted the incentives for agricultural pro- Most programs of decoupled payments ducers. Macroeconomic policies historically have no time limit, as in the EU and Tur- taxed agriculture more than agricultural key. The United States had a program with policies did, but both were important in a time limit in the 1996 Farm Bill, but it poorer countries. The indirect tax on agri- was not enforced. Mexico's decoupled pro- culture, through overvalued currencies gram initially had a time limit; the program and industrial protection, was nearly three was supposed to expire when the North times the direct tax on the sector at the time American Free Trade Agreement phase-in of the last World Development Report on is completed in 2008, but the government agriculture (1982). In a study that included has already announced that the program 16 of today's developing countries from the will be retained in some form. Unless 1960s to mid-1980s, average direct taxation these programs have time limits with cred- was estimated at 12 percent of agricultural ible government commitments to stick to producer prices and indirect taxes at 24 them, decoupled payments risk becoming percent. High taxation of agriculture was more distorting and costly than commonly associated with low growth in agricul- assumed. In addition, continuing output- ture--and slower growth in the economy.5 linked programs along side decoupled sup- The poorest developing countries taxed port can significantly dampen the less-dis- agriculture the most, and reinvestments of torting effects of decoupled programs. tax revenues in agriculture were low and Progress on decoupling has varied sig- inefficient (chapter 1). nificantly by commodity, with most prog- With reforms in the 1980s and 1990s to ress on grains--although recent initiatives restore macroeconomic balance, improve to expand the use of biofuels in OECD resource allocation, and regain growth in countries may indirectly reverse some of many of the poorest countries, both direct this progress. Needed now is a rapid shift and indirect taxes were reduced. The reform to less-distorting decoupled support for of overvalued currencies, which taxed agri- export products important to developing cultural exports (usually exported at the countries, particularly cotton. There have official rate) and subsidized food imports, been some recent changes to rice, sugar, is reflected in the huge reduction in the and cotton policies in Japan, the EU, and parallel market premiums for foreign cur- the United States, respectively, all at an rency in developing countries. For 59 devel- early stage of implementation. oping countries, the trade-weighted average premium fell from more than 140 percent Political economy factors matter in the 1960s to approximately 80 percent in for further reform the 1970s and 1980s and to just 9 percent in Political economy factors in each coun- the early 1990s, with wide variation across try have determined the pace and extent countries.6 of reforms. U.S. cotton policies, EU sugar policies, and Japan rice policies indicate Agriculture-based countries are that the impact of the World Trade Orga- taxing agriculture less nization (WTO) in inducing reform is real Reforms in agriculture-based countries, and that media pressure can complement particularly in Sub-Saharan Africa, more it (box 4.2). The cases show that reforms than halved the average net taxation of are not easy and often require bargained agriculture from 28 percent to 10 percent compromises and compensation schemes between 1980­84 and 2000­04 (simple for the losers to get agreement on further average across countries included in figure reducing high levels of agricultural protec- 4.2). The approach used to measure the tion (as in the Japanese rice policy reforms change in net taxation in developing coun- and the EU sugar policy reforms). tries is through calculation of a nominal Reforming trade, price, and subsidy policies 99 B O X 4 . 2 The political economy of agricultural reforms in developed countries Agricultural subsidies and tariffs on rice and get agreement to a reform program while not EU structural and rural development funds. sugar, aggregated across all countries, are undermining, but perhaps slowing, the even- Africa, Caribbean, and Pacific countries that estimated to account for 20 percent and 18 tual shift to larger-scale production. Larger- received higher-than-world-market prices for percent, respectively, of the global cost of all scale farmers are already exiting the Japan their quota of sugar produced for sale in the agricultural trade policies--the highest of all Agricultural Cooperatives marketing system, EU market were eligible for an assistance plan commodities. Although the equivalent global exits expected to accelerate under the direct- worth 40 million for 2006. cost of cotton subsidies and tariffs is much payments program, reducing the political smaller, the absolute cost to developing coun- power of Japan Agricultural Cooperatives and U.S. cotton policy reform: tries is large, an estimated $283 million a year. its resistance to reform. WTO and local media pressure For Sub-Saharan Africa, the developed-country to offset industry lobby power cotton subsidies and tariffs account for about EU sugar policy reform: compensation The United States accounts for 40 percent 20 percent of the total cost of trade policies on and restructuring to complement reform of world cotton exports and 20 percent of all merchandise goods. EU domestic sugar prices--supported by high world cotton production. Subsidies have import tariffs--are three times higher than been equivalent in value to about two-thirds Japanese rice policy reform: bargained world market levels, increasing incentives to of the market value of production over the compromise to agree on decoupled produce sugar in the EU and depressing the 2000­05 period. The additional U.S. produc- support world market price of sugar at the expense of tion prompted by these subsidies is estimated Japan protects rice producers, a traditional many developing-country exporters. However, to reduce the world cotton price by 10 to 15 source of political support, through a 778 some African, Caribbean, and Pacific countries percent, at significant cost to developing percent ad valorem tariff equivalent on rice benefit from these higher prices under the countries. imports. In 2007 Japan introduced a less- Everything But Arms trade agreements. U.S. cotton policy is heavily influenced by distorting direct payment to farmers linked The European Union agreed to reform its a strong interest group, the Cotton Council of to farm size, not production. The payment is sugar regime in February 2006; reforms began America (representing the 24,721 cotton grow- expected to be bargained against a decline in July 2006 and extend for four years. If fully ers, according to the census in 2002, as well in tariff levels for rice--making payments to implemented, the reforms would radically as ginners, exporters, bankers, and suppliers). farms larger than a certain size to target "prin- change the sugar regime, in place for almost The council is one of the most powerful U.S. cipal" rather than "part-time" farmers. The new 40 years. For years, the policy had encountered commodity lobbies, winning disproportion- scheme is viewed as a less-distorting alterna- discontent from the food processing industry, ately higher support relative to other sectors, tive to border protection and as a mechanism paying three times the world price for sugar. particularly since the enactment of the 1996 to induce larger-scale production. But two main factors led to the initiation of Farm Bill (an average equivalent of $120,000 a Why did politicians agree to the proposed reforms. First, the EU's sugar export subsidy year per farmer). scheme despite the apparent risk of undermin- system was ruled noncompliant with agreed Four West African cotton-producing ing their political support from rural areas? commitments under the WTO. Second, the EU's countries (Benin, Burkina Faso, Chad, and Three factors. One is the ever-strengthening Everything But Arms initiative was introduced Mali) submitted a joint proposal to the WTO in voices from nonfarm sections of the economy. in 2001 to open the EU sugar market to duty- May 2003, demanding removal of support to A second is media pressure: fearing Japan's free and quota-free imports from the world's 50 the cotton sector by the United States, China, increasing isolation in the global economic Least Developed Countries from 2009 onward. and the EU and compensation for damages community for its rice policies. Third is the This was expected to lead to a surge in imports until full removal of support. Brazil initiated a view that agriculture should be part of the and the destabilization of the EU sugar regime comprehensive case against the United States broader economic reforms. unless the sugar price was reduced. Adding to for noncompliance with its WTO obligation The system of protection of agriculture has these determinant factors was the campaign on cotton subsidies. In March 2005, the WTO been kept in place by a strong pro-agricultural of an international nongovernmental organiza- Dispute Settlement Body instructed the United coalition of the Ministry of Agriculture, Forestry, tion coalition that emphasized the negative States to bring the offending cotton subsidy and Fisheries; the ruling Liberal Democratic effects of the EU sugar policy for developing measures into compliance with its WTO obliga- Party; and the Japan Agricultural Cooperatives, countries. The reform became imperative. tions. The United States made adjustments in which implements the farm subsidies pro- While the political equilibrium turned response to the WTO decision, but in Decem- grams. But the Ministry of Agriculture, Forestry, against the sugar producers, measures were ber 2006 Brazil formally expressed its dissatis- and Fisheries has gradually shifted to more put in place to address the expected loss of faction with the extent of U.S. policy changes market-oriented policies. The Liberal Demo- revenues that the reform will induce and to and asked the WTO panel to find the United cratic Party has shifted its balance of interest counter the producers' opposition. Compensa- States "out of compliance" with the original rul- toward urban areas because of growing sup- tion and a restructuring fund (financed partly ing. The compliance phase of the case is now port from cities in recent elections, an indica- by producers) to encourage uncompetitive proceeding. While the reduction in U.S. cotton tion that nonagricultural groups are gaining producers to leave the industry were agreed to subsidies was a response to the legal case at political capital in this policy arena. in February 2006. EU farmers are expected to the WTO, the U.S. media and reform-minded While reform seems inevitable, opposi- receive compensation for an average of 62 per- groups also pressured the U.S. Congress to tion by Japan Agricultural Cooperatives cent of the price cut phased over four years. reduce support. led to a compromise in the coverage of the The four-year restructuring fund has three direct-payment scheme, expanded to include main objectives: to encourage less-competi- Sources: Anderson, Martin, and van der direct payments to small part-time farmers if tive producers to leave the industry, to cope Mensbrugghe 2006a; Anderson and Valenzuela they organized into a collective farming unit. with the social and environmental impacts of forthcoming; Masayoshi Honma, Yujiro Hayami, Although viewed as weakening the efforts factory closures, and to help the most affected Dan Sumner, Don Mitchell, and John Baffes, all at structural change, it seemed necessary to regions develop new businesses in line with personal communication 2007. 100 WORLD DEVELOPMENT REPORT 2008 rate of assistance to farmers (box 4.3). Nine Figure 4.2 For agriculture-based countries, net of 11 countries in a recent study had lower agricultural taxation fell in 9 of 11 countries net taxation in the second period (figure Kenya 4.2). Only Nigeria and Zambia had higher net taxation between the two periods, with Uganda the highest net taxation in 2000­04 in Côte Madagascar d'Ivoire (about a ­40 percent nominal rate Cameroon of assistance). Despite macroeconomic adjustment, Sudan real domestic prices for agricultural Ghana exports across these countries did not Nigeria change much on average over the 1980s as Ethiopia the macroeconomic improvements barely offset the declines in world commodity Tanzania prices. The situation changed during the Zambia 1990s--more favorable world commodity Côte d'Ivoire prices, continued macroeconomic reforms, and agricultural sector reforms led to larger ­80 ­60 ­40 ­20 0 20 40 increases in real domestic prices of agricul- Nominal rate of assistance, % tural exports.7 The stronger price incen- 1980­84 2000­04 tives explain part of the higher agricultural growth in many of the agriculture-based Source: Anderson (Forthcoming). countries since the mid-1990s (chapter 1). The aggregate nominal rates of assistance decline in taxation is the result of improved mask significant differences in taxation and macroeconomic policies. protection between agricultural imports For the agriculture-based countries, and exports and among products. An aver- tobacco, groundnuts, and cocoa were still age nominal rate of assistance close to zero heavily taxed over 2000­04. The net taxa- at the country level simply indicates no net tion of coffee declined from 53 percent to taxation, but it could be the result of large 7 percent, and for cotton it declined from import tariffs offsetting large export taxes. 32 percent to 15 percent over the two peri- On average between 1980­84 and 2000­04, ods. Sugar shifted from being heavily taxed agriculture-based countries lowered pro- (nominal rate of assistance of ­36 percent tection of agricultural importables, from a in 1980­84) to being heavily protected (76 14 percent tariff equivalent to 10 percent, percent in 2000­04) (table 4.1). and there has been a significant reduction in taxation of exportables, from 46 per- Transforming and urbanized cent to 19 percent (figure 4.3). Most of the countries are protecting agriculture more Net taxation in transforming countries B O X 4 . 3 Nominal rates of assistance declined on average from 15 percent to 4 per- cent, but with significant variations across The nominal rate of assistance to farmers input subsidies and differences between countries (simple average across countries is defined as the price of their product the international prices of inputs and the included in figure 4.4). Some countries in the domestic market (plus any direct prices that farmers pay for these inputs. output subsidy) less its price at the border, If a country distorts its market for foreign shifted to protect the sector more (Indone- expressed as a percentage of the border currency, efforts are made to account for sia, India, Malaysia, and Thailand), while price (adjusting for transport costs, quality the difference between the exchange rate others continued to tax it, although at lower differences, and so forth.). The nominal used by the importers (assumed to be the levels than in the 1980s (as in Egypt and rate measures differences in output prices, parallel exchange rate) and the exporters but there may also be distortions on the (a weighted average of the parallel and Senegal) (figure 4.4). Zimbabwe is the only input side. To capture those distortions in official exchange rates) and an estimated country of this group that had a higher net countries where they are important, the equilibrium exchange rate. tax on the sector, mainly because of a highly nominal rate is adjusted (expressed as out- overvalued currency. There has also been a put price equivalent) to account for direct Source: Anderson (Forthcoming). significant shift in the relative rate of assis- Reforming trade, price, and subsidy policies 101 Figure 4.3 Developing countries are taxing exportables less Exportables Importables Nominal rate of assistance, % Nominal rate of assistance, % 30 30 1980­84 2000­04 1980­84 2000­04 26 23 20 20 14 13 10 11 10 10 2 0 0 ­10 ­10 ­13 ­14 ­20 ­19 ­20 ­30 ­29 ­30 ­40 ­40 ­46 ­50 ­50 Agriculture-based Transforming Urbanized Agriculture-based Transforming Urbanized Source: Anderson (Forthcoming). Note: The countries used for each category are shown in figures 4.2, 4.4, and 4.5, respectively. The aggregates are simple unweighted averages. Value-weighted averages show a similar pattern, although the NRA for exportables in transforming countries in 2000-04 was close to zero, given the dominance of China in the weights. Value-weighting also reduced the NRAs for importables in urbanized countries over the two periods. tance to agriculture versus nonagriculture in underestimate actual taxation as currency some countries, with a remaining challenge overvaluations were not included in the to keep sectoral biases low (box 4.4). estimates.8 (The official exchange rate was There are also differences across agri- used for both time periods.) Six of seven cultural imports and exports. On average countries analyzed (Argentina, Chile, between 1980­84 and 2000­04, transform- Colombia, the Dominican Republic, Ecua- ing countries slightly reduced protection of dor, and the Philippines) had higher pro- agricultural importables from a 13 percent tection or lower taxation in 2000­04 than tariff equivalent to 11 percent, and reduced in 1980­84 (figure 4.5). Rice and sugar are the taxation of exportables from 29 percent the most-highly-protected products in the to 13 percent (figure 4.3). urbanized countries (table 4.1). Between In urbanized countries, the average net 1980­84 and 2000­04, urbanized coun- taxation shifted from marginally negative tries slightly lowered their level of protec- in 1980­84 to a net protection rate of 9 tion of agricultural importables from an percent in 2000­04 (simple average across average tariff equivalent of 26 percent to 23 countries included in figure 4.5). The net percent, and shifted from a tax on export- taxation estimate for Latin American coun- ables of 14 percent to a subsidy equivalent tries, particularly in the earlier period, may of 2 percent (figure 4.3). Table 4.1 Nominal rates of assistance by commodity in developing countries (percent) Agriculture-based Transforming Urbanized Product 1980­84 2000­04 1980­84 2000­04 1980­84 2000­04 Sugar ­36 76 33 35 ­11 52 Rice ­4 5 ­12 4 ­4 44 Wheat ­12 ­3 ­4 8 8 ­8 Coffee ­53 ­7 -- -- ­38 4 Maize ­11 ­7 ­23 8 ­14 ­1 Cotton ­32 ­15 ­20 ­2 -- -- Cocoa ­51 ­36 -- -- -- -- Groundnuts ­19 ­38 9 9 -- -- Tobacco ­49 ­50 -- -- -- -- Source: Anderson (Forthcoming). Note: The nominal rate of assistance is weighted by the value of production across countries in each of the three country categories, and estimates are included only if data were available for three or more countries. -- = not available. 102 WORLD DEVELOPMENT REPORT 2008 Figure 4.4 For transforming countries, 9 of 10 Net protection has on average increased either increased protection or reduced taxation from 4 percent in 1992/93 to 31 percent in 2002/03 (simple average across countries).10 Indonesia There are large differences across countries. India For example, Estonia, Latvia, and Lithuania Thailand imposed about a 30 percent tax equivalent on the sector in 1992/93, while Slovenia Malaysia protected the sector. Between 1992/93 and China 2002/03, protection on agricultural imports Pakistan increased on average from a 13 percent to a Sri Lanka 38 percent tariff equivalent. Exports were taxed at 2 percent on average in 1992/93, Egypt but in 2002/03 they were protected with an Senegal average tariff equivalent of 24 percent. The Zimbabwe increase in protection is in part a result of EU accession by many of these countries ­80 ­60 ­40 ­20 0 20 40 over the period analyzed, resulting in a shift Nominal rate of assistance, % to the higher protection levels of the EU. 1980­84 2000­04 Still space for further efficiency gains Source: Anderson (Forthcoming). While there is less domestic price and trade policy exploitation of farmers in develop- Urbanized countries in Eastern and Cen- ing countries now than in the 1980s, it has tral Europe have on average increased agri- not disappeared. Net taxation of agricul- cultural protection.9 (Comparative statistics ture is low in all but a few countries. But are not included in the figures here because disaggregating net taxation by exportable the earliest data available are from 1992.) and import-competing products shows B O X 4 . 4 Significant progress in reducing the antiagricultural bias in China and India As developing countries become richer, they omies of Northeast Asia when they had similar ing upward in both countries. China bound its generally protect agriculture more. Both China per capita incomes, the trends are strikingly agricultural tariffs at relatively low levels when and India have reduced their antiagricultural similar. China has reduced its antiagricultural it joined the WTO in 2001. The challenge now is bias substantially over the past three decades, bias at a later stage of economic development to keep sectoral biases low and not follow the not only directly but also indirectly via cuts than India, but the assistance to agriculture rela- trend to heavily protect agriculture that other to manufacturing protection (figures below). tive to nonagriculture (measured by a relative countries followed when they were at similar When compared with the more-advanced econ- rate of assistance [RRA] index) has been trend- levels of development. India--assistance to tradables, 1965­2004 China--assistance to tradables, 1980­2005 Rate of assistance, % Rate of assistance, % 120 120 NRA--nonagriculture 80 80 NRA--nonagriculture 40 40 NRA--agriculture 0 0 NRA--agriculture ­40 ­40 Relative rate of assistance Relative rate of assistance ­80 ­80 1965­69 1970­74 1975­79 1980­84 1985­89 1990­94 1995­99 2000­04 1980­84 1985­89 1990­94 1995­99 2000­05 Source : Anderson (Forthcoming). Note : The relative rate of assistance to agriculture is 100*[(100 + NRAagt)/(100+NRAnonagt) ­ 1] , where NRAagt is the nominal rate of assistance to producers of tradable agricultural goods and NRAnonagt is the nominal rate of assistance to nonagricultural tradables (mainly mining and manufacturing). The index is bound from below at ­100 and is zero when the agricultural and nonagricultural tradables sectors have identical nominal rates of assistance. Reforming trade, price, and subsidy policies 103 Figure 4.5 For urbanized countries, 6 of 7 either The power of outside actors is real, as dem- increased protection or reduced taxation onstrated by the impact of WTO accession on protection in transforming and urban- Colombia ized countries and by the impact of foreign Philippines assistance on taxation in agriculture-based Ecuador countries. However, lasting change occurs only with a strong domestic constituency. Chile Strengthening local constituencies to build Dominican Republic coalitions for remaining policy reforms South Africa can help--particularly as political systems become more open and competitive. Argentina ­80 ­60 ­40 ­20 0 20 40 Simulated gains Nominal rate of assistance, % from trade liberalization 1980­84 2000­04 Agricultural policy reform in both devel- Source: Anderson (Forthcoming). oped and developing countries offer signifi- cant potential welfare gains, including from trade reforms. The magnitude of the costs that exports are still heavily taxed in many of current trade policies and correspond- countries, while some imports are heavily ing potential gains from further reforms protected. This suggests room for further have been quantified through simulations welfare gains. Further reforms should be of global computable general equilibrium designed in the context of a country's level models. These models are based on a sim- of development. Many developing coun- plified but consistent representation of tries where agriculture is a large share of production, income, and demand in each gross domestic product (GDP) will need to country or group of countries and of inter- continue to tax agriculture (although not national markets. While the models require disproportionately) to provide a surplus for strong assumptions, they remain a power- broader development programs (see transi- ful tool for analysis of global trade scenar- tional support section). ios (box 4.6). Political economy factors matter The costs to developing countries of for further reform current trade policies are substantial Agricultural reforms in many of these The global welfare costs of current trade countries, particularly the agriculture- policies fall on both developed and develop- based ones, came after the macroeconomic ing countries. Recent estimates show that reforms of the 1980s. They were heavily the global costs of trade tariffs and subsidies supported by external donors through pol- would reach about $100 billion to $300 bil- icy advice and conditional lending. Other lion a year by 2015.11 About two-thirds of important elements of the reforms, reflect- the costs are estimated to come from agri- ing the political economy in countries (box cultural tariffs and subsidies (the remainder 4.5), include leadership and exploiting win- from tariffs and subsidies in other sectors), dows of opportunity (as in Uganda), tying much higher than agriculture and processed the fortunes of local leaders to the success food's 6 percent share of global GDP and 9 of the local economy, building on local sup- percent share of international trade. While port, using WTO accession (as in China), these costs are a modest share of global GDP and bargained complementary policies to for developing countries, they are substantial support free trade (as in Mexico). relative to current aid flows for agricultural Reforms are not easy, because there will development. Developed-country agricul- be both gainers and losers. Reducing heavy tural policies cost developing countries about taxation and protectionist biases in devel- $17 billion per year--a cost equivalent to oping countries requires understanding about five times the current levels of overseas the political economy aspects of reform. development assistance to agriculture.12 104 WORLD DEVELOPMENT REPORT 2008 B O X 4 . 5 The political economy of agricultural reforms in developing countries Three examples, one from each country more than 300 million to fewer than 50 million. Nacional para el Campo (National Agreement category, illustrate the political economy of Why was China able to make these tough for the Countryside), greatly increasing public reform in developing countries. In Uganda decisions when leaders in many other nations resources funneled to rural areas. (agriculture-based) and China (transforming), falter? The 1990s reforms eliminated state trading net taxation of agriculture declined signifi- Much of the pressure for reform came from enterprises in agricultural products and sup- cantly between 1980­84 and 2000­04, while in the failed policies and poor performance of port prices. In exchange, they provided com- Mexico (urbanized) there was a shift to protec- agriculture. China's leaders were committed to mercial producers with brokerage services and tion over the same period. becoming a secure and independent country. market information for price-risk management, There was also an imperative to worry about and substituted support prices with compen- Uganda: leadership and a window equity and provide citizens with a minimum satory payments based on target incomes. The of opportunity standard of living. Central planning was not government complemented market support Uganda's agricultural reforms disbanded the proving effective. with decoupled, per-hectare payments to Coffee Marketing Board and the Lint Marketing The decentralization reforms in China producers of basic grains and oilseeds, under a Board monopolies in 1991 and the Produce tied the fortunes of local leaders significantly new program called PROCAMPO. The govern- Marketing Board in 1993--all had heavily taxed to the success of the local economy. Hence, ment strengthened land property rights in agriculture. Cross-district product movement policy initiatives that tied local revenues, local rural areas. Major grants and subsidized credit- restrictions were also removed. The reforms investment spending, and cadre salaries to based programs assisted the agricultural sec- significantly increased the share of the border the increases in agricultural output and the tor's transition toward greater efficiency and price received by farmers and contributed to transformation of the economy toward rural global competitiveness, through the Alianza the large 1990s decline in the percentage of industrialization had local support. That the Contigo (Alliance with You). In 2004 roughly people below the national poverty line. reforms were introduced in a gradual process 80 percent of the Ministry of Agriculture's The reforms followed a broader set of of local experimentation and learning reduced $3.7 billion budget was devoted to marketing macroeconomic reforms by the National Resis- the political risks associated with the reform. support, PROCAMPO, and Alianza Contigo, tance Movement government, which came to Moreover, the grassroots pressure built in the roughly a third of Mexico's public spending on power in 1986. The macroeconomic reforms process helped the reformers in the Chinese rural development. (by reducing the overvalued currency) had a government win the battle with conservative The reforms have not eliminated distor- greater impact on agricultural export prices reform critics. tions in the allocation of production factors. than the agricultural reforms, although both Market interventions under the new policy were significant. Following the armed struggle Mexico: delicate balance between regime, while greatly increasing the role of to power, popular legitimacy formed the complementary programs to facilitate the private sector, have perpetuated or even bedrock of the regime, enabling the president agricultural policy reform and exacerbated such distortions, hampering the to pursue difficult and potentially unpopular protection traps adjustment toward more efficient use of pri- reforms, including those in agriculture. Groups During the 1990s, following the North Ameri- vate and public resources. Although interven- with vested interests in the marketing boards can Free Trade Agreement, which established tions were initially established as temporary lost their political weight in the regime change. the (gradual) elimination of tariff and nontariff measures to ease adjustment to a market- barriers to agricultural imports by 2008, the based food sector, the economic interests China: tying the success of local leaders Mexican government implemented wide- created by these interventions and the export to the success of the local economy ranging agricultural market-oriented policy subsidies in developed countries have made it China launched a bold but gradual set of reforms. The reforms were designed in ways politically infeasible for Mexican policy makers reforms in 1978, first raising prices for agri- that avoided major political opposition from to justify an exit strategy. cultural commodities; then decollectivizing domestic agricultural producers with signifi- agricultural production, making the farm cant political power. household the residual claimant; and finally The power of farmer organizations in Sources: Avalos-Sartorio 2006; Huang, Rozelle, beginning to slowly but steadily dismantle Mexico was evident in 2002 with a horseback and Rosegrant 1999; Lin 1992; McMillan, Waley, and Zhu 1989; Opolot and Kuteesa 2006; Qian the state-run procurement and input supply incursion into Mexico's congressional build- and Weingast 1996; Robinson 2005; Rosenzweig systems. In response, the rural economy took ing as a way to influence policy. The mes- 2003; Rozelle 1996; Swinnen and Rozelle 2006; off. Agriculture boomed. Productivity nearly sage, reminiscent of the Mexican Revolution World Bank 2002a; Yang 1996; Yunez-Naude and doubled. The number of rural poor fell from of 1910, paid off with a negotiated Acuerdo Barceinas Paredes 2004; Zahinser 2004. Developing countries are estimated to Vietnam and 3.2 percent for Thailand. For share 30 percent of the welfare costs of cur- agricultural and nonagricultural liberal- rent trade policies, whether from agricul- ization alike, half of the costs to developing tural policies or from policies in the other countries are estimated to come from poli- sectors (table 4.2). These lower absolute cies in developed countries, the other half costs on developing countries translate from policies in developing countries as a into a higher percentage of income because group (table 4.2). of their smaller economies. As a group, the More than 90 percent of the global costs estimated cost by 2015 is 0.8 percent of real are estimated to come from market access GDP--but for some countries it is esti- restrictions through tariffs rather than mated to be much higher: 5.2 percent for from export subsidies or domestic support. Reforming trade, price, and subsidy policies 105 However, their relative importance varies significantly by product.13 For example, B O X 4 . 6 Simulating the effects of trade liberalization the reverse is true for cotton, where 89 per- with global models cent of the costs are expected to come from export subsidies and domestic support pro- The general equilibrium models used by flexibility in resource reallocation across grams and 11 percent from tariffs.14 different studies to analyze global trade sectors of production; and the character- scenarios are conceptually similar: disag- ization of the competitive market structure. Trade reforms offer significant scope gregating the world into a number of Particular attention is given to modeling to reduce the global costs of current poli- countries or groups of countries, modeling sources of price distortion, including bilat- in each case supply and demand for a large cies through raising international agricul- eral tariffs and subsidies and domestic number of commodities, deriving import subsidies to agriculture, but modeling the tural prices, which is expected to increase demand and export supply, and solving distortionary effects of specific measures developing-country agricultural trade for the world equilibrium prices that clear such as tariff-quotas, various forms of quan- shares and agricultural output growth the international market. The World Bank tity restrictions, and so-called decoupled LINKAGE model, for example, comprises support is extremely difficult at a global rates in the aggregate. However, not all 27 regions or countries, with a focus on level. There is little empirical evidence on developing countries will gain. isolating the largest commodity exporters which to base specification of investment and importers, and 25 sectors, of which 13 and productivity effects, and thus these Large price increases expected are agriculture or food. One of the great are largely ignored, (although they could for some commodities from trade strengths of general equilibrium models is presumably be important). The level of that they impose consistency: all exports disaggregation by income groups within reforms: a gain for exporters, a loss are imported by another country, total countries also tends to remain low, if at all. for importers employment never exceeds labor supply, As recognized by the authors, the many and all consumption is covered by produc- assumptions underlying these models can According to the 2006 World Bank study, tion or imports. However, they must rely on lead to large over- or underestimates of the full trade liberalization is estimated to strong assumptions--particularly on the impacts of merchandise trade reforms on increase international commodity prices on adjustments to changes in trade policies net real household income, although with as captured by key supply and demand much more consensus on the structural average by 5.5 percent for primary agricul- elasticities, for which empirical validation is impacts. Yet, there is no real alternative to tural products and 1.3 percent for processed often inadequate. Key features of the mod- using these models when analyzing reform foods.15 Developing countries are estimated els are the degree of tradability of com- with many indirect effects, and comparison to gain 9 percentage points in their share modities in each country, which determines of outcomes across models is important to the passthrough of international prices get a sense of their validity. of global agricultural exports--increasing to domestic prices; the supply response from 54 percent to 65 percent. to price changes, which depends on the Sources: Francois and Martin 2007; Hertel and But these aggregate results hide big dif- availability of resources in the country and others 2006; van der Mensbrugghe 2006. ferences across commodities and, there- fore, countries. The largest estimated price increases are for cotton and oilseeds (figure Table 4.2 Estimated cost distribution of current trade policies 4.6), with significant estimated trade share (percent of costs of current global trade policies in 2015 relative to a full trade liberalization scenario) gains to developing countries exporting Distribution of welfare costs these products (figure 4.7). Liberaliza- Developing Developed tion of cotton and oilseeds is estimated countries countries Total to induce a shift of world production to Source of welfare costs: the developing countries, with an even- Developing countries policies greater shift in export shares. Developing Agriculture and food 9.8 6.6 16.4 countries' share of exports is estimated to Other sectors 5.2 23.0 28.2 increase from 49 percent to 83 percent for Developed countries policies cotton, and from 55 percent to 82 percent Agriculture and food 9.1 38.0 47.0 Other sectors 5.9 2.4 8.4 for oilseeds. The direction of change in All countries trade policies (sum of the above) 30.0 70.0 100.0 international prices is unambiguous, but Real GDP cost 0.8 0.6 0.7 the magnitude of the price changes differs Source: Anderson, Martin, and van der Mensbrugghe 2006a. across studies. For example, a review of 11 Note: The full trade liberalization scenario is based on estimates of bilateral tariffs and domestic and export subsidies studies estimating the changes to interna- as of 2001. Bilateral trade preferences are included. tional cotton prices from full trade liberal- ization suggests an average price increase Oilseed production subsidies in the of 10 percent16 (lower than the 21 percent OECD and import tariffs in some develop- estimated in the 2006 World Bank study), ing countries are the main causes of the cur- and estimates of cereal price increases rent oilseed trade share loss to developing range from 4 to 8 percent.17 countries as a group. While OECD country 106 WORLD DEVELOPMENT REPORT 2008 Figure 4.6 Estimated real international commodity price increases following complete countries also provide significant direct trade liberalization assistance to cotton producers (for example, China) and apply import tariffs of up to 10 Cotton 20.8 percent (Argentina, Brazil, Egypt, India, Oilseeds 15.1 Dairy products 11.9 and Uzbekistan).19 Full trade liberalization Coarse grains 7.0 would increase international prices and pro- Wheat 5.0 duction in Sub-Saharan Africa. West African Processed meat 4.3 cotton exports are estimated to increase by Rice 4.2 60 percent.20 Removing U.S. cotton subsidies Fruit and vegetables 2.8 alone is estimated to increase the incomes of Other crops 2.6 West African cotton producers by 8 to 20 Sugar 2.5 percent.21 Production in OECD countries Livestock 2.5 is estimated to decline significantly in the Vegetable oil and fats 1.9 absence of current producer subsidies. % change in real price With international food prices expected Source: Anderson, Martin, and van der Mensbrugghe 2006a. to increase, there is particular concern for food-importing developing countries.22 Because many of the poorest countries Figure 4.7 The corresponding gain in the estimated trade shares of developing countries spend a large part of their incomes on cereal imports, they may incur an overall welfare Cotton 27 loss despite gains from price increases in Oilseeds 34 nonfood commodities such as cotton.23 Dairy products 7 Almost all of the agriculture-based Coarse grains 5 countries are net importers of cereals, with Wheat 21 a large share of their export earnings spent Processed meat 18 on cereal imports--more than 10 percent Rice 2 Fruit and vegetables 4 over the past 10 years in Benin, Burundi, Other crops ­3 Ethiopia, Mozambique, Niger, Rwanda, and Sugar 9 Sudan, and 20 percent in Burkina Faso. An Livestock 2 increase of cereal prices by about 5 percent Vegetable oil and fats 1 (the change expected from full liberaliza- Percentage point gain tion) would negatively affect these cereal importers. This expected long-term price Source: Anderson, Martin, and van der Mensbrugghe 2006a. change is small relative to short-term cereal- price movements, as experienced for maize tariffs on oilseeds are low, many countries with the more than 50 percent increase in provide support for domestic production international prices over the past two years. through farm subsidies. India and China, A cereal price increase may also accentu- the largest importers of oilseeds, impose ate the problems associated with fluctua- significant import tariffs. Full trade liber- tions in domestic production (food security alization is estimated to raise international focus). Yet, many of the same countries are oilseed prices and production in Latin net exporters of oilseeds and cotton. Sudan American and Sub-Saharan Africa, reduce earns on average 12 percent of its foreign oilseed production in OECD countries exchange from oilseeds exports and 7 per- (from subsidy removal), with little aggre- cent from cotton exports. Over the past 10 gate net impact in South and East Asia as years, cotton exports on average accounted price effects of lower import tariffs (mainly for 40 percent of total exports from Benin, in India and China) would be offset by 25 percent from Chad and Mali (although higher international prices.18 these shares have been decreasing), and OECD cotton production subsidies, pri- 30­60 percent from Burkina Faso. Trade marily in the United States, significantly reforms that increase the price of cotton reduce the share of cotton exports from and oilseeds simultaneously with that of developing countries. Several developing cereals appear to more than compensate Reforming trade, price, and subsidy policies 107 these countries for the foreign exchange Figure 4.8 Latin America and Sub-Saharan Africa are expected to have higher agricultural loss on cereal imports. There are, however, output growth under global trade reforms food-importing countries that produce Change in annual agricultural output growth, %a little or no cotton and oilseeds--such as 3 Burundi, Kenya, Niger, and Rwanda--and 2.0 they would remain vulnerable to cereal 2 price increases. Additional investments in domestic agriculture to raise the produc- 1 0.3 0.4 tivity of food staples may be needed for the 0.0 0.0 0 most vulnerable countries. ­0.2 ­0.1 ­1 Faster agricultural output ­1.7 growth in Latin America ­2 Developed Developing Sub- South Asia East Asia Middle East Europe Latin and Sub-Saharan Africa countries countries Saharan & Pacific & North & America & In the World Bank study, agricultural out- Africa Africa Central Caribbean Asia put growth in developing countries is esti- Source: Derived from Anderson, Martin, and van der Mensbrugghe 2006a. mated to increase from an annual rate of 3.9 a. Difference between estimated average annual agricultural growth to 2015 under full liberalization in 2005 and percent in the baseline scenario to 4.2 per- the baseline without liberalization. cent under the full liberalization scenario, an 8 percent increase in the growth rate or countries, but worsens for net importers of a 4.3 percent increase in agricultural output these commodities. Subsequent changes in over a 10-year period. Latin America and national welfare usually follow the direction Sub-Saharan Africa share the largest gains, of these terms of trade changes, but changes while developed countries, South Asia, and in poverty often do not follow this pat- Europe and Central Asia are estimated to tern. A fall in poverty can occur even with lose on average (figure 4.8). worsening terms of trade (as estimated for Most of the gains to developing coun- Bangladesh), and vice versa (as estimated tries are estimated to come from efficiency for Vietnam) (table 4.3). In contrast to the gains.24 Hence, complementary investment dominance of the terms-of-trade effects support will be needed to facilitate adjust- from developed-country reforms, the gains ment to realize these efficiency gains from from developing-country agricultural trade trade reforms. reforms are estimated to come mainly from efficiency gains from their own country Poverty declines in many countries, reforms. These gains are estimated to have but not in all positive poverty-reducing effects. However, Not everyone will gain from agricultural the magnitude of these effects varies across trade liberalization; there will be losers countries, depending on the size of the pre- across and within developing countries. vailing distortions. Tracing the overall welfare effects of trade The transmission of global trade reforms policy reform on poverty requires a compre- to poverty reduction involves many chan- hensive approach that links a broad general nels, and the specific effects are as varied equilibrium macroeconomic model with as the countries themselves. Some devel- detailed household survey data. A recent oping countries are estimated to benefit study of 15 developing countries takes this from large terms-of-trade improvements approach.25 following developed country reform, such Several broad regularities emerge from as Brazil (competitive in heavily protected the study. Removal of trade-distorting agri- agricultural products such as sugar, oil- cultural policies in developed countries seeds, and beef) and Thailand (an exporter has mixed terms-of-trade effects on devel- of rice) (table 4.3). The terms-of-trade oping countries. Term of trade improves improvements translate into higher levels of for developing countries exporting com- national welfare in Thailand than in Brazil modities currently protected in developed as the former is more trade dependent. The 108 WORLD DEVELOPMENT REPORT 2008 Table 4.3 Illustrative poverty effects from agricultural trade reform in developed and developing countries Brazil Thailand Vietnam Mexico Mozambique Bangladesh Developed countries liberalize Change in: Terms of trade (percent) 4.9 1.1 0.3 ­0.2 ­0.4 ­0.5 Welfare (percent) 0.7 0.8 0.2 ­0.2 ­0.6 ­0.2 Poverty (percent) ­1.8 ­6.6 0.2 0.3 0.1 ­0.1 Developing countries liberalize Change in: Terms of trade (percent) 0.6 0 ­0.4 ­0.3 0.6 ­0.4 Welfare (percent) 0.1 0.5 1.1 0.1 1.8 0.3 Poverty (percent) ­0.2 ­4.6 ­1.7 0.6 ­1.1 ­0.2 Both developed and developing countries liberalize Change in: Poverty at $1 a day (percent) ­1.9 ­11.2 ­1.5 0.9 ­1.0 ­0.3 Poverty at $1 a day (thousands of people) ­445 ­133 ­23 86 ­62 ­128 Source: Hertel and others 2007. Note: Six of the 15 countries are presented in the table above, selected to illustrate the different transmission magnitudes from terms of trade, to welfare, to poverty reduction across countries. Of the 15 countries studied, 2 were estimated to experience an increase in poverty from agricultural trade liberalization in both developed and developing countries. terms of trade are estimated to worsen for import tariffs lowers the price of food for countries such as Bangladesh (an importer poor consumers and lowers the income of of cotton, wheat, and oilseed) and Mozam- surplus food producers. For example, in bique (an importer of wheat and rice and Mexico poverty in rural households is esti- an exporter of seafood, the international mated to rise from domestic tariff cuts. By price of which is expected to decline with contrast, in Vietnam both real agricultural global trade reforms). incomes and real wages are estimated to The poverty effect of terms-of-trade rise following reforms, generating broad- changes from developed-country agricul- based poverty reductions. ture reforms depend on where the poor Overall, when developed and develop- are, what they do for a living, and what ing country agricultural trade reforms are they consume. For example, smaller terms- combined, the extent of poverty reduction of-trade changes for Thailand are estimated tends to be enhanced--and the proportion to lead to larger poverty impacts relative to of the population experiencing a poverty Brazil. The reason: one-third of the extreme rise diminishes. poor (below $1 per day) in Brazil mostly live off transfers and lose from food price Gainers and losers among the poor increases, which dampen the employment within countries and income gains of the other two-thirds A particular concern with trade policies for of the extreme poor, mainly unskilled agri- staple foods is their potential welfare impact cultural workers and self-employed. In on the poor. While most poor are net buyers contrast, the extreme poor in Thailand are of food, others are net sellers. Any change predominantly rural households with diver- in price will therefore produce gainers and sified income sources and are estimated to losers among the poor. Considering only the gain from price increases. In Bangladesh, average poverty effect (as presented in table the estimated terms-of-trade loss translates 4.3) may hide important consequences of into lower poverty levels as the poor are policy reform on poverty across households heavily reliant on unskilled wage income (box 4.7). The distribution of gainers and and benefit from lower food prices. losers is country specific. Developing-country agricultural trade In assessing the impact of food import reforms are estimated to have a much prices on household welfare, the degree of smaller impact on their own terms of trade transmission of international prices to rural than developed-country policy changes households also matters. The degree of (table 4.3). Removing developing-country transmission varies significantly by coun- Reforming trade, price, and subsidy policies 109 B O X 4 . 7 Net buyers and net sellers of food staples within a country The vulnerability of poor people to food price households, and they spend 27 percent of their net sellers of food staples. As rice sales (and increases varies across countries (table below). total budget on purchasing rice. Poor Bangla- maize in Madagascar) represent a large share In Bolivia and Ethiopia, the diet includes staples deshis are the most vulnerable to increases in of household income in these countries--up such as potatoes, sorghum, and teff that are rice prices. Only 8 percent of the poor are net to 70 percent in Madagascar--net sellers are not traded by these countries on international sellers of food, so the aggregate welfare effect sensitive to any changes in rice prices. Aggre- markets. As a result, poor people are less vul- of a change in rice prices is dominated by its gate income gains to sellers from an increase nerable to variation in prices of imported cere- effect on net buyers. Zambia has few land- in rice prices overwhelm the loss to buyers. als. In the five other countries in the table, trad- less poor people but many smallholders who Similarly, in Morocco 35 percent of poor rural able products (rice, wheat, maize, and beans) are net buyers, and they are affected by price households are net sellers and lose more in the represent between 40 percent and 64 percent changes of imported maize and wheat. aggregate than net buyers from cereal price of food expenditures. In Bangladesh, more In contrast, Cambodia, Madagascar, and declines.26 than 50 percent of the poor are in rural landless Vietnam have many smallholders who are A majority of the rural poor are not net sellers of tradable food staples. Bolivia Ethiopia Bangladesh Zambia Cambodia Madagascar Vietnam 2002 2000 2001 1998 1999 2001 1998 Share of internationally traded staples 25.5 24.1 41.2 40.4 56.3 62.7 64.4 in food consumption of the poor (%) Distribution of poor (%) Urban (buyers) 50.9 22.3 14.9 30.0 8.4 17.9 6.1 Rural landless (buyers) 7.2 -- 53.3 7.4 11.5 14.8 5.8 Smallholders net buyers 29.1 30.1 18.8 28.8 25.8 18.9 35.1 Smallholders self-sufficient 7.1 39.5 4.6 20.8 18.0 27.3 19.4 Smallholders net sellers 5.6 8.0 8.4 13.0 36.3 21.1 33.6 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Share of net purchase/sale of staples by specific groups of the poor (% of the total expenditures of the specific groups) Purchase per net urban buyer 12.0 9.4 22.7 11.5 5.9 4.8 13.1 Purchase per net rural buyer 12.9 28.4 27.3 18.9 20.8 10.7 19.9 Sales per net seller 37.6 35.1 39.7 21.0 39.0 70.3 37.4 Share of net purchase/sale of staple aggregated across all the poor (% of the total expenditure of all poor) Purchase by all poor net buyers 11.3 10.2 22.0 10.3 8.1 3.6 8.8 Sales by all poor net sellers 1.4 2.8 4.0 2.3 14.4 18.4 12.5 Source : Authors' calculations, based on data provided by Ataman Aksoy and Aylin Isik-Dikmelik, personal communication. Note : Data are only for those people below the national poverty lines. Tradable staples included are rice, wheat, maize, and beans. Excluded staples are cassava, potatoes, plantains, sorghum, and teff. -- = not available. try, affected by transaction costs and trad- through the employment and wage effects ability within the country. For example, a induced by the price changes.29 In many recent study of eight developing countries countries, such as Mali and Burkina Faso, a indicates low price transmission to farmers large number of smallholders produce both in Colombia, Egypt, Ghana, Indonesia, and food and export commodities and may Madagascar. However, in Argentina, Chile, benefit from trade liberalization, which and Mexico about 60 percent of domestic would result in a rise in cereal and cotton price variability can be explained by world prices. The ability of farmers to respond to price changes.27 Price changes at the house- new market opportunities depends on such hold level determine the magnitude of wel- nonprice factors as market infrastructure, fare impacts.28 institutions, and services. Broad-ranging Beyond the first-order food price effects, trade reform in Vietnam in the early 1990s trade liberalization affects the poor through induced a large supply response and welfare the creation and loss of markets and gain among poor farmers.30 110 WORLD DEVELOPMENT REPORT 2008 Rising or falling prices of staples and ing countries. Tariff reduction is expected other agricultural products can also induce to have a greater impact on global welfare changes in employment and wages. The and poverty reduction than the removal of direction and magnitude of these effects subsidies in developed countries, although are case specific and depend on labor mar- both are important.36 There are excep- ket conditions. In countries with a large tions (for example, cotton) where reducing share of a landless rural population work- export subsidies are expected to have large ing in agriculture for wages, as in South impacts and where important gains from Asia, labor market impacts can be signifi- the Doha round can be made.37 cant. A study of Bangladesh concluded that The suspension of the Doha Round of the average landless poor household loses trade negotiations between July 2006 and from an increase in rice prices in the short January 2007, and the fitful progress follow- run, but gains in the long run as wages rise ing the resumption of talks, raise important over time.31 An opposite result is obtained questions about the prospects for further in Mexico, where the reforms of the 1990s reforms through multilateral agreements. induced a decline in unskilled wages and There are several possible scenarios. agricultural profits that offset the gain from lower prices of consumption goods.32 A Doha Round agreement--content Decompositions of incomes in Vietnam, matters. The best outcome would be an Bangladesh, and Uganda reveal that labor agreement on further reforms, particu- market effects are indeed important chan- larly on agricultural products important to nels for trade reforms to affect welfare.33 the poorest countries, such as cotton. The impact would depend on the following: Scope for achieving · The extent to which applied or actual potential gains tariffs are below their upper-bound rates Advancing global trade liberalization is not agreed upon at the WTO. Current applied easy, as demonstrated by the Uruguay and rates are generally below bound rates, Doha Rounds of trade negotiations. Vested requiring larger cuts in bound rates if interests strongly defend many current applied rates are to be cut. Average bound policies and are reluctant to change. Most tariffs are almost double applied rates in past policy reforms have come from unilat- developed countries, and over two and eral reform efforts, which will continue to a half times applied rates in developing be important in the future, but multilateral countries.38 and regional agreements remain important · The level of developed-country subsidy instruments to remove distortions in inter- reduction for key export crops, such as national and regional markets.34 cotton. As domestic support programs account for 89 percent of the global Multilateral agreements: welfare costs of cotton trade policies, the Doha Round reducing these subsidies could be an The Doha Development Round of trade important gain to developing countries, negotiations provides an opportunity to particularly the cotton-producing coun- realize at least part of the potential gains of tries in Sub-Saharan Africa. Again, the full trade liberalization. While the poten- limits agreed at WTO greatly exceed tial gains from full trade liberalization as current support levels. a share of GDP are larger for developing · The treatment of "sensitive products," countries than for developed countries which if not tightly constrained can (table 4.2), the estimated impacts of a undercut reform impacts. Developed potential Doha agreement suggest the gains countries are seeking smaller tariff and are smaller for developing countries.35 Part subsidy reductions for self-selected sen- of the reason: Doha places heavier empha- sitive products than implied by a general sis on eliminating export subsidies and on formula approach. Estimates show that if cutting domestic subsidies than on reduc- only 1 percent of all tariff lines in the EU ing tariffs in both developed and develop- were exempt, the expected overall aver- Reforming trade, price, and subsidy policies 111 age tariff reduction estimated under the need to use bilateral and regional agree- Doha Round, with no exemptions, could ments to advance reforms. More bilateral halve.39 The United States proposal is to and regional trade agreements on agricul- limit sensitive products to 1 percent of ture would be a less-efficient and more- all tariff lines, while the EU proposal is costly outcome than further global reform, 8 percent. perhaps delaying and complicating it. But · The treatment of "special products." regional agreements can often be useful for Developing countries are seeking small addressing issues not on the multilateral or no tariff cuts on special products-- agenda (see below). deemed important for food security, The worst outcome of a Doha Round livelihood security, and rural devel- failure would be a spiraling back to global opment. The potential impact of any protection, including in developing coun- exemptions will likely be country spe- tries, reversing past efficiency gains and cific. Net buyers of food, especially the impacts on poverty reduction. OECD sub- very poor, will likely be hurt by tariffs on sidies are already inducing some developing food staples that raise prices above what countries to call for higher protection rates they would be without tariffs (box 4.7). on a range of agricultural products (as at Net sellers would benefit. Some devel- the 2006 Food Security Summit for Sub- oping countries exporting products that Saharan Africa). may be deemed "special" by other coun- Regional trade agreements tries are concerned about the potential restrictions on developing-country mar- As trade among developing countries is ket access for these products. These fac- a growing share of their overall trade, tors need to be considered in any agree- improving developing-country access to ments on special products. (See also the developing-country markets can have a section on transitional support.) significant effects. Regional agreements can address · Special and differential treatment for regional collective action issues that are not developing countries. Developing coun- on the agenda in multilateral trade discus- tries are required to make smaller cuts sions. For example, regional agreements in protection than developed countries can reduce political tension and take advan- under the current development round tage of economies of scale in infrastructure of trade negotiations (under special and provision. Greater regional integration and differential treatment agreements). While opening regional markets can be important developed-country agricultural trade in regions with many small countries (Sub- reform will likely have a larger poverty Saharan Africa, for example).41 impact on many countries than devel- More than a third of global trade is oping-country reforms, the latter can between countries that have some form potentially reduce poverty more consis- of reciprocal regional trade agreement.42 tently across a large number of develop- ing countries--both are important.40 These agreements have usually been eas- ier to reach than multilateral agreements, Following the above, a Doha agreement with fewer participants involved, and they would capture some of the benefits of full usually extend beyond tariff reductions to liberalization if that agreement lowers tariff reduce impediments associated with border bindings significantly below actual levels, crossings, regulations, and standards. Not reduces developed-country subsidies where all such agreements create new trade and they matter most for developing countries investment--some instead divert them. (such as for cotton), limits sensitive-product (For example, countries with high exter- tariff lines, and reflects the net-buyer status nal border protection may actually reduce of the poor in special-product agreements. members' trade overall, even through trade within the group increases.) Scenarios in the absence of an agreement. African countries have four regional In the absence of a Doha Round trade trade agreements on average, and Latin agreement, developing countries would American countries have seven, adding to 112 WORLD DEVELOPMENT REPORT 2008 the complexity of trade. A recent World increase over the last two years. Moreover, Bank review of regional agreements con- because of infrastructure and transport cluded that agreements most likely to costs, the transmission of world food staple increase national incomes are those with prices to domestic producers is very imper- low external "most-favored nation" tariffs, fect, especially in agriculture-based coun- few sectoral and product exemptions, non- tries.44 In fact, most food staples in most restrictive rule-of-origin tests, measures to agriculture-based countries are not traded facilitate trade, rules governing investment internationally, but only locally and in the and intellectual property that are appro- region (see focus C). So the overall effect of priate to the development context, and trade distortions on farm incomes of food implementation schedules put into effect staple producers in the poorer developing on time.43 Implementation has proven dif- countries is likely to be small. ficult in many countries: volumes of formal In the case of a tradable food staple with documents legalize free movement of goods high price transmission, a case for protec- and people across borders, but implemen- tion could be made for modest, short-term tation remains weak. Efforts are needed to protection where there is a high likelihood ensure policy harmonization, reduce non- of reduced protection in world markets in tariff barriers, reduce border formalities the short to medium term that would cause and corruption, address problems of cur- world prices to rise, and where the domes- rency transfers, and capitalize on econo- tic industry would be clearly competitive mies of scale in infrastructure. with undistorted prices. But even in these cases, protection would be modest (that is, of a magnitude close to the expected rise Transitional support in world prices, which for cereal products Transitional support may be needed to facil- is about 5­10 percent). The political dif- itate further reforms and sector adjustment. ficulties of adjusting policy once the trade Important issues are the role of transitional distortion is removed must be considered. protection, the ability to shift to alternative Consequently, credible exit strategies should forms of revenue, and the needed public be specified if protection is introduced. spending to support transitions. Food security. Aside from arguments Arguments for and against about distorted world prices, the case is protection of food staples in sometimes made for protecting domestic developing countries food staple industries in the name of food OECD policies. There have been recent security. This should be considered with calls by some developing countries for caution. First, consumers bear the cost of interim import protection in response to protection, particularly poor consumers current OECD trade policies. The argu- who spend a high share of income on food ments are that OECD protection reduces staples, and many rural poor are net food international prices below the long-term buyers in many countries (see box 4.7). trend, which harms the competitiveness of Second, poverty and insufficient purchas- import-competing food sectors and leads ing power rather than lack of food supply is to the decapitalization of agriculture and usually the main cause of food insecurity, to rural-urban migration. Therefore, it is although there are important exceptions in argued, import protection is justified to the agriculture-based countries (focus C). maintain the domestic industry. For example, in 2004 Indonesia enacted But there are several counterarguments. a temporary import ban on rice--which The average distortion in world prices from has now become permanent--to increase trade policies is about 5 percent for food domestic production. Two-thirds of the staples, as discussed elsewhere in this chap- poor are net consumers of rice and are ter. This long-term effect is small relative hurt by the rice price increases induced by to recent price changes, as reflected by the the ban. The impacts of the ban have been more than 50-percent world maize price identified as the main cause of the increase Reforming trade, price, and subsidy policies 113 in poverty headcount from 16 percent in as alternative policies such as increasing 2005 to 18 percent in 2006.45 access to assets and productivity-enhanc- If an industry is already protected, rapid ing investments in research, education, liberalization for a sector that is a large and extension, and rural infrastructure. But tradable part of the economy can generate in recognition of the political sensitivity significant unemployment and hardship of these markets and country specificity of in the short term, especially for the poor, trade policy impacts, providing flexibility who lack the assets or knowledge to take within trade rules makes sense if it is done advantage of new opportunities.46 In this in a way that encourages the shift to market case, it is imperative to include transitional liberalization. support for vulnerable groups to ensure that they benefit from growth, and to sus- Transitioning to alternative forms tain political support for trade reform (see of taxation below). For those with productive assets, Further reducing the protection of imports this transitional support should be pro- and the taxation of agricultural commodity vided not only for income support (as in exports can pose a fiscal dilemma for many PROCAMPO in Mexico), but also to facili- agriculture-based countries that depend tate transition to competitive activities. on these revenues for public investment. In Sub-Saharan Africa, trade taxes account for Safeguard policies. Governments that about a quarter of all government revenues; require a safety net to increase their com- in the developing countries of Asia and the fort level when they liberalize markets and Pacific, they account for about 15 percent.48 reduce applied tariffs, may consider price Agriculture remains the dominant sector bands to reduce exposure to world price in most agriculture-based countries and variability, if such safeguard policies are so will have to continue to contribute to allowed in the new round of WTO nego- national and local government revenues-- tiations. Price floors implemented through consistent with their current level of eco- a temporary increase in the import levy nomic development. Four key principles may help to prevent extreme hardship to to guide agricultural taxation, highlighted producers in years when world prices are in a previous analysis of Africa, remain extremely low. Similarly, temporary reduc- valid:49 they should be nondiscriminatory, tions in tariffs could be implemented when minimize efficiency losses, and consider the world prices are very high. (It must be rec- effectiveness of fiscal capture and capacity ognized, however, that the ability of this to implement. mechanism to significantly reduce upward Agriculture should not be taxed at a price volatility is limited, unless there is higher rate than other sectors, and agricul- significant initial tariff protection, which is tural taxes should be integrated with general not likely to be either efficient or equitable.) value added, profit, and income taxes. Out- To minimize the economic costs of any such put and input taxes should be minimized. variable levy schemes, and to ensure that Land taxes can minimize efficiency losses they do not become permanent increases and induce production, although these do in protection, it is important to have clearly not generally exist in agriculture-based defined rules for safeguard interventions countries. Output taxes can be replaced by that cannot be captured by vested inter- consumption taxes (sales or value added ests, and that temporary tariff increases are taxes) in countries with the administrative infrequent and of short duration.47 To date, capacity to implement them.50 Capacity to there are few, if any, successful examples of implement new systems will have to be built using such safeguards and some examples over many years. In the interim, it may be in which they clearly did not work well. necessary to rely partly on commodity and In sum, trade policy on food staples input taxes for revenue. must recognize that protection of domes- Recent evidence shows a mixed picture tic production is often not pro-poor. Nor in shifting to alternative sources of revenue is protection as efficient in helping farmers but provides some lessons on how to deal 114 WORLD DEVELOPMENT REPORT 2008 with trade revenue losses. Developed coun- needed adjustments, the government tries have recovered all revenue lost from may have to make cash payments and previous trade reforms. Middle-income provide social safety nets, as in Mexico countries have recovered 45­60 cents of through the PROCAMPO program (see each dollar of lost revenue. Low-income box 4.5).53 However, cash transfers to countries have recovered only 30 cents compensate for losses are insufficient to of each dollar of lost revenue. Experience induce supply response. Targeted invest- across low-income countries varies widely. ments, such as infrastructure investments Malawi, Uganda, and Senegal have man- and extension services, are needed to aged to recover most revenue losses. What improve productivity or education and to makes this possible? Efforts to broaden tax facilitate transition (see next section).54 bases by reducing exemptions, simplify- The challenge is to ensure an adequate ing rate structures, and improving revenue balance among the complementary income administration can help, as can excise and support for transitions and core public pro- broad-based value added taxes on con- sumption.51 By contrast, value added sys- grams to spur long-term agricultural growth and poverty reduction. The risks of falling tems with multiple rates and exemptions into protection and subsidy traps induced and weak administrative capacity have led by a dominant focus on transitional sup- to low recovery. Trade reform may need to port at the expense of long-term growth are be sequenced with complementary domes- high. Governance problems that may limit tic tax reforms and significant improve- the capacity to implement these programs ments in the quality of agricultural public must also be addressed (chapter 11). spending. Public investment Policies and public spending to support transitions for long-term development Too often trade liberalization is discussed The magnitude of smallholder supply without considering the important role of response to trade and price policy reforms complementary policies and programs to depends on, among other factors, rural facilitate transitions and support the los- infrastructure (irrigation, roads, transport, ers. Complementary policies include pub- power, and telecommunications), markets, lic investment and other policies that will rural finance, and research.55 Where these facilitate response to the new market signals are deficient, complementary investments for long-term growth (discussed in the next will be necessary to take advantage of trade section). It is necessary to recognize the het- reforms. Similarly, if these nonprice factors erogeneity in the groups adversely affected, are in place but domestic macroeconomic examine their distinguishing demographic and sectoral policies depress incentives to and geographic characteristics, and analyze produce, the supply response may be lim- the magnitude of the losses and potential ited. In many countries, particularly the gains. Transitional support may include the agriculture-based ones, these nonprice fac- following: tors are undeveloped and need significant investment, particularly in market infra- · Grants to facilitate production shifts. An structure, institutions, research and exten- example is the Turkey program to reduce sion, and natural resource management. agricultural subsidies. Per-hectare grants Over the long term, these investments are were paid to farmers to facilitate their likely to be more important than trade transition out of tobacco and hazelnut reforms in using agriculture for develop- production and into more efficient alter- ment. Details of investment priorities will natives such as maize, soybean, sunflower, be the topics of subsequent chapters. and vegetables. Complementary support Public spending has often been diverted was provided to improve the efficiency of from these needed long-term investments the cooperative marketing channels.52 to agricultural subsidies. Subsidies are usu- · Cash transfers and social safety nets. To ally economically inefficient and often pro- sustain the extreme poor and to support mote wasteful use of resources at a high cost Reforming trade, price, and subsidy policies 115 to farmers in terms of foregone growth and subsidies. Overall public expenditures on incomes. Where long-term capital invest- agriculture have remained at approximately ments have been made, too few resources 11 percent of agricultural GDP, while the are allocated to operations and mainte- share of subsidies for fertilizer and elec- nance to ensure the sustainability of these tricity (see box 4.8) and for support prices investments. for cereals, water, and credit has steadily Agricultural subsidies are defined here as risen--at the expense of investments in payments from the public budget for essen- public goods, such as research and devel- tially private goods such as agricultural opment, irrigation, and rural roads. Agri- inputs. Subsidies can help overcome tem- cultural spending is about 4 times greater porary market failures (as part of a broader on subsidies than on public goods (figure strategy), offset fixed costs of infrastruc- 4.9). Moreover, the returns on subsidies in ture, and reduce risk (chapter 6). But they India have declined.57 These findings and have seldom been used for these purposes, the results from a related study suggest the have mostly benefited richer farmers, and potential for significant efficiency gains are often difficult to remove once estab- from reallocating public expenditures in lished--all leading to inefficient and ineq- agriculture in India.58 uitable resource use. Thus the quality of Reforms to improve the efficiency public spending--the efficiency of resource of rural public spending use--is often an even more important issue to address than its level. Understanding why public rural expen- ditures are allocated to unproductive Inefficiency of current spending interventions requires understanding the A large share of public spending has been political economy of government policies. used to provide private goods at high cost. Institutional, demographic, and economic Public expenditure reviews suggest that variables jointly shape the size and quality of agricultural budget allocations to private public spending. One factor affecting quality goods are high: 37 percent in Argentina is information. The lack of a formal program (2003), 43 percent in Indonesia (2006), of expenditure evaluations--combined 75 percent in India (2002), and 75 percent with a lack of access to public information in Ukraine (2005). Transfers to parastat- on expenditures and their beneficiaries-- als and subsidies in Kenya in 2002/03 dilutes the effectiveness of any formal accounted for 26 percent of total govern- accountability mechanisms that might be ment expenditures in agriculture, and in provided by political checks and balances, Zambia in 2003/04, about 80 percent of a free press, or well-intentioned civil society nonwage spending went to subsidies to organizations. With such information gaps, farmers for fertilizer and maize prices. public debates about public policies tend to Allocations to subsidies often divert be manipulated by special interest groups. funds from high-return investments in public goods. In Zambia only about 15 percent of the 2003/04 agricultural budget Figure 4.9 Subsidies have risen while public was spent on research, extension services, goods investments have declined in India and rural infrastructure--investments % of agricultural GDP that have shown high payoffs (chapter 7). 7 Reallocating spending on private subsi- 6 dies to public goods can increase growth.56 5 However, although these subsidies are eco- Subsidies nomically inefficient, they are often politi- 4 cally expedient. Improving the efficiency of 3 resource use thus requires addressing the 2 Public good investment political economy pressures determining 1 budget allocations (box 4.8). 0 In India, too, the trend has been to move 1975­79 1980­84 1985­89 1990­94 1995­99 2000­02 away from public goods investments toward Source: Chand and Kumar 2004. 116 WORLD DEVELOPMENT REPORT 2008 B O X 4 . 8 Examples of subsidies in India and Zambia Electricity subsidies to agriculture to the agricultural sector. Unlike other policy research and extension. Spending on agri- in India: can greater local accountability instruments, it does not require implementa- cultural research and development fell from induce reforms? tion by the (often ineffective) public admin- about 1.2 percent of agricultural GDP in 1985 With 55­60 percent of India's irrigated land istration. Farmers who buy water from pump to about 0.5 percent in 2000. supplied by groundwater, electricity for tube- owners--a considerable proportion of farmers Why is spending on fertilizer subsidies so well pumps is an important input. Most state in most states--potentially benefit from the high? There are no powerful groups that would governments provide electricity to farmers at subsidy, too, which increases the attractiveness benefit from its elimination, despite its being a subsidized flat rate--often for free. But the of this policy instrument for politicians who an economically unproductive use of public quality of service is poor because of erratic and want to win state elections. resources. This contrasts with early reforms in limited supply and voltage fluctuations, which Addressing jointly the quality of electricity maize milling, where the private sector gained can result in crop losses from forgone irriga- supply and its cost is a key element of reform- significantly from privatization and strongly tion and damaged pumping equipment. ing the subsidy policy. However, because of supported the reform. Under the fertilizer pro- The electricity subsidies to agriculture widespread power theft and losses, states gram, traders often benefit. are also fiscally draining and environmentally lack the credibility to deliver better service in A 2002/03 household survey showed that damaging. In Punjab electricity subsidies to exchange for higher prices. One option would only 29 percent of farmers acquired fertilizer, agriculture in 2002/03 were 7 percent of state be to decentralize energy supply to local 59 percent of them through private dealers expenditures. Together with other policies that governments or community groups, relying and 36 percent through the government promote water-intensive crops such as rice, the on local accountability to improve electricity Fertilizer Support Program. Both groups had electricity subsidies contribute to the overex- quality. Elite capture must still be prevented, higher income and wealth and were close to ploitation of groundwater. About 60 percent of but this community-oriented option has the tarmac roads and district centers. However, the state's groundwater resources are already potential to break the political impasse. This those receiving fertilizer through the gov- overexploited, with extraction rates exceeding exemplifies a tradeoff between potential effi- ernment program were predominantly civil recharge rates--clearly not sustainable. ciency cost from the loss of economies of scale service employees, in a program intended to Increasing electricity prices and introduc- in decentralized generation, and not making be targeted at the poor. Parliamentarians also ing metering are technically and economically any progress at all. benefited, sometimes informing groups of sound, but they are not politically feasible, farmers that there was no need to repay loans so far. Larger farmers obviously benefit more Zambia fertilizer subsidies: no strong on fertilizer received. from the subsidy, and they have political influ- opposing coalitions The economic costs of the program are ence, but there is more to these subsidies. About 5 percent of Zambia's national budget high--both from lower spending in higher- Their introduction followed massive farmer goes to agriculture. In fiscal 2005 more than productivity areas such as agricultural protests against electricity price increases in half the agriculture budget was spent on the research, extension, and infrastructure, and the 1980s. Now, their continuation responds Fertilizer Support Program (37 percent) and from slower diversification away from maize to the increasing income disparity between crop marketing (for maize) under the Food production. the agricultural and nonagricultural sectors, Reserve Agency (15 percent). Only 3 percent worsened by India's relatively low agricultural of the budget went to irrigation development Sources: Beintema and others 2004; Birner, Sharma, growth rate. Making electricity free is a politi- and other rural infrastructure, and 11 percent and Palaniswamy 2006; Govereh and others 2006; cally convenient instrument to transfer income to operating costs, which included agricultural Pletcher 2000;. World Bank 2003d. Rigorous evaluations, their wide dissemi- trative decentralization accompanies polit- nation, and increasing transparency could ical democratization.60 Still, the challenge reduce this information gap. is to avoid elite capture, and so far the evi- Special interest groups also influence dence on the effect of decentralization on patterns of public spending. In Latin Amer- corruption is mixed.61 ica the share of rural subsidies provided by governments is higher where there is more income inequality.59 Economic sectors or Conclusions groups of producers that control a large por- Recent policy reforms have improved price tion of national wealth also have the means incentives for agricultural producers in to influence public policies to their benefit. developing countries. Net agricultural taxa- If the ineffectiveness of public expenditures tion across these countries has, on average, is a result of the influence of special interest declined sharply. Between 1980­84 and groups, the solution might be to link budget 2000­04, it declined from 28 percent to implementation to participatory decision 10 percent in agriculture-based countries, making in which poor rural households from 15 percent to 4 percent in transform- have a voice (chapter 11). This might work ing countries, and from marginally negative best for local expenditures where adminis- to a net protection of 9 percent in the urban- Reforming trade, price, and subsidy policies 117 ized countries. But changes in net taxation opment, education, and associated insti- in some countries are the result of rising tutional support. But public investments protection of agricultural imports with con- in agriculture are too often squandered tinuing taxation of exports. These differ- on regressive subsidies. Significant room ences suggest considerable space for further remains for improving the efficiency of policy improvements, but with potential public resources by increasing investments distributional impacts within countries. In on high-priority public goods. Needed are contrast there has been relatively little prog- actions to increase information, account- ress in the overall decline in OECD producer ability, and commitment. Information gaps support. However, there has been a shift in public knowledge of budget allocations away (decoupling) from support directly and impacts of public spending on agricul- linked to product prices, volumes, and area ture have to be closed through greater pub- planted to other less-distorting forms such licity and transparency of budget allocation as cash transfers, particularly in the EU. and evaluation. The estimated impacts of full trade Political economy determines the pace liberalization are substantial for develop- and extent of reform and has to be addressed ing-country trade and agricultural output in both developed and developing countries. growth. Full trade liberalization is expected Building coalitions to support and sustain to increase international commodity prices reforms can help. The WTO has induced by 5 percent on average, developing-coun- reform, and local media have played sup- try share in global agricultural trade by portive roles (as in the U.S. cotton industry). about 9 percentage points, and agricultural In some cases, bargained compromises and output growth in developing countries on compensation schemes for the losers may be average by about 0.3 percent a year. Urban- needed--as in the new Japanese rice policy ized countries, particularly those in Latin reforms, the EU sugar reforms, and Mexico's America with competitive advantage in 1990s reforms. Linking domestic agricul- many of the currently protected products, tural reforms to a broader set of economy- stand to benefit the most. Not everyone will wide reforms can strengthen reform coali- gain from liberalization: net-selling farm- tions and increase the likelihood of progress, ers will benefit, while households that are as happened in many developing countries net buyers of food may lose from higher in the 1980s and 1990s. food prices if their wages or other earnings Key elements of the future agenda are to do not increase enough to compensate. continue to get prices right through trade Further trade liberalization in develop- and domestic policy reform, to ensure com- ing countries may need to be sequenced with plementary tax reforms to replace lost trade tax reforms to reduce tax losses from trade revenues for reinvestment in the sector, to revenues and subsequent public investment ensure that the quality of public spending in the agriculture sectors in these countries. improves, to provide support to comple- Complementary policies and programs are mentary programs to facilitate transitions, needed to compensate losers in developing and to invest massively in core public goods countries and to facilitate rapid and equi- for longer-term sustained growth. All of this table adjustment to emerging comparative requires a comprehensive approach beyond advantages. price and adjustment; governments must Supply response to trade reforms depends focus on improving market infrastructure, on public investments in core public goods institutions, and support services--topics such as irrigation, roads, research and devel- of the subsequent chapters. Bringing agriculture to the market 5 Far-reaching changes in domestic and alone do not promote equitable outcomes. global markets are creating big opportu- So smallholders may need to build their nities for farmers and agribusiness entre- bargaining power through their producer c h a p t e r preneurs. The demand for high-value organizations, assisted by public policy. primary and processed products is rapidly The nature and pace of market develop- increasing, driven by rising incomes, faster ment differs across food staples (cereals), urbanization, liberalized trade, foreign traditional bulk export commodities (cof- investment, and advancing technology. fee, cocoa, tea, cotton), and higher-value These developments are expanding market products for domestic and export markets opportunities, which is important for faster (dairy, meat, fruits, vegetables). This chap- agricultural and nonfarm growth and for ter examines the new opportunities and greater employment and rural incomes. But challenges for smallholders in the mar- the new markets demand quality, timely kets for each of these important commod- deliveries, and economies of scale, posing ity groups. It highlights the broad array of special challenges for smallholders. private, public, and civil society initiatives Still in many agriculture-based and that have been pursued to make markets transforming countries, food staples remain work better for development and poverty a mainstay for a major share of households, reduction. many of them poor. But the performance of food staple markets is often hampered Food staples: improving by poor infrastructure, inadequate support commodity trading and services, and weak institutions, pushing up transaction costs and price volatility. How risk management markets for food staples function thus The market for food staples remains by far affects livelihoods, welfare, and food secu- the most important in many agriculture- rity, especially for poor households. based and transforming countries, because Well-functioning agricultural marketing staples take up a major share of household systems can reduce the cost of food and the food expenditures and account for the bulk uncertainty of supply, improving the food of agricultural gross domestic product security of poor and nonpoor households. (GDP). Growing populations sustain By linking farmers more closely to con- demand, supplemented by the rapidly grow- sumers, these marketing systems transmit ing demand for livestock feed in middle- signals to farmers on new market opportu- income countries. Inhibiting the market for nities and guide their production to meet food staples are high transaction costs, changing consumer preferences for quan- product wastage and losses, wide marketing tity, quality, variety, and food safety. margins, poor market integration, limited Efficient markets require good gover- access to trade finance, and weak regulatory nance and public policy--infrastructure, institutions. Better markets for food staples institutions, and services that provide have broad implications for agricultural market information, establish grades and growth because they raise farmgate prices, standards, manage risks, and enforce con- build the confidence of farmers in their reli- tracts--a continuing challenge in many ability, and allow farmers to diversify to countries. However, efficient markets higher-value products. 118 Bringing agriculture to the market 119 Figure 5.1 Layers of intermediaries characterize Ghana's maize markets $ per 100 kilograms, 1998 1.9 31.0 0.9 1.2 0.7 27.0 0.9 1.7 0.8 1.5 22.8 fee profit losses 1.1 0.3 profit 0.3 mkt 1.7 0.6 losses 17.3 Retailer fee others fee profit interest, price Transport losses fee, agent interest, handling, costs mktg Wholesaler price Transport & mktg Storage, other interest, price Wholesaler price wholesale Storage, & Transport, Wholesale wholesale retail Handling, Commission Storage, Farmgate Handling Techiman Accra Accra Source: Natural Resources Institute, personal communication 2006. In agriculture-based and transforming nections is thus critical to strengthening the countries, small and medium-size traders links of farmers and the rural economy to and layers of intermediaries are common local, regional, and international markets in the marketing of food staples and other (box 5.1). agricultural commodities (figure 5.1). Often one-person businesses dealing in Market information systems several commodities, the traders and inter- Market information keeps farmers and mediaries are mainly self-funded because of traders attuned to the demands and chang- limited access to credit. They maximize the ing preferences of consumers, guiding returns on their working capital by rapidly farming, marketing, and investing. Mar- turning over small quantities, with little ket information encompasses timely and storage. Quality grades are rarely standard- accurate prices, buyer contacts, distribu- ized, nor are weights and measures, mak- tion channels, buyer and producer trends, ing personal inspection by buyers essential. import regulations, competitor profiles, This requires that traders travel extensively, grade and standards specifications, post- increasing transaction costs. harvest handling advice, and storage and Improving and modernizing the mar- transport recommendations.2 keting system can increase market effi- Public market information systems have ciency, foster competitiveness with imports, often been disappointing, with information and reduce losses and risks. Market mod- disseminated too slowly, in the wrong form, ernization, beyond improving basic trans- or too infrequently to be of real use to market port, includes marketing information sys- participants.3 Several innovative approaches tems, commodity exchanges, and price-risk are being piloted in different parts of the management. world, building on advances in communi- cations technology (radio, cell phone, tele- Poor road connections vision, Internet) and the liberalization of Inadequate transport infrastructure and ser- telecommunications and broadcasting. In vices in rural areas push up marketing costs, India, the Ministry of Agriculture operates undermining local markets and exports. This AgMark Net, which collects price informa- is particularly the case in Africa, where less tion from wholesale markets nationwide than 50 percent of the rural population lives and disseminates it through the Internet. close to an all-season road. Trader surveys The private sector in India is investing in in Benin, Madagascar, and Malawi find that telecommunications infrastructure, such as transport costs account for 50­60 percent of mobile phone networks and Internet-linked total marketing costs.1 Improving road con- rural kiosks, which aid in strengthening 120 WORLD DEVELOPMENT REPORT 2008 B O X 5 . 1 Impacts of road infrastructure on markets and productivity Rural road development has the potential one study in Vietnam found that four to six need to be provided as part of a package of to reduce transport costs and generate mar- years after road rehabilitation, road transport interventions that helps certain groups benefit ket activity. In Vietnam, road rehabilitation services were more likely to respond where more than they would have and that protects increased the variety of goods that households markets were already established and natural or compensates those who may lose. sold in the market--primary fruits, vegetables, disasters were relatively infrequent. Policy The governance and institutional settings and meat--and encouraged greater participa- needs to consider more than the absence or are also important in determining impacts. tion in trade and services. In Georgia, the con- dire condition of a road before deciding that a Road project funds may not end up funding struction and rehabilitation of roads increased new road is critical. In each specific case, policy what was intended and hence have no impact. the opportunities for off-farm and female should ask whether roads are the right instru- Infrastructure expenditures present opportu- employment. In Madagascar, simulations sug- ment for overcoming the constraints to a given nities for graft and the diversion of resources. gest that a 50 percent reduction in travel time welfare outcome and if so, what other policy This can change when incentives change. One per kilometer on roads would increase rice initiatives and investments are needed. study found that the threat of an audit on road production by 1 percent. Heterogeneities across households will projects in Indonesia significantly increased However, these effects will be mediated by determine who gains and who loses. Hold- the actual amounts spent on labor and build- specific geographic, political, and economic ing community characteristics constant, ing materials for roads, thereby bringing the settings. Complementary inputs and policies some households will be better placed to quality of the roads nearer to that originally may be required to achieve the full benefits take advantage of a new road, based on their intended. Fungibility can also dull impacts, as from improved roads. Even if aggregate out- endowments and the nature of their occupa- aid or central government funding for road put gains are forthcoming, there will almost tions. Households differ in what they buy and projects may substitute for local government certainly be losers too. How one weighs the sell and hence how much they will gain or lose infrastructure spending. Finally, the lack of gains and losses and whether poverty falls from the changes in prices induced by better funding and institutional arrangements for is ultimately an empirical question. Recent roads. Poor households are more likely to rely routine maintenance can significantly reduce work using impact evaluation methods shows on the production of nontraded goods and the impacts of newly improved roads. mixed results, suggesting that to be effective, services that may actually be displaced by bet- How much roads matter depends on a rural road policy needs to adapt to context ter roads fomenting increased competition. range of factors. Of course, roads matter to and setting. On the other hand, road improvement has a economic development, but how much they Policy should focus more on the comple- general income effect that could generate matter depends on a number of other factors. mentary role of rural roads. Past policy has demand for services from poor providers. The Comprehensive approaches are needed that fixated on the supply of rural roads as a cata- net effect is an empirical question. The picture are compatible with how local institutions lyst to development and market activity. Poor that emerges from recent, more methodologi- work in practice, including what they are capa- road conditions often coincide with a number cally rigorous impact evaluations is a complex ble of delivering. This may require fashioning of other bottlenecks inhibiting agricultural one. In Nepal, better road access benefited the a whole package of cross-sectoral investments productivity and economic development, poor and the nonpoor, but the proportionate (roads and complementary investments) including poor agroclimatic endowments, low gains were higher for the nonpoor. Dercon and and policy changes that will ensure a higher population density, no transport services, low others (2006) find that access to all-weather efficiency impact, as well as more desirable education levels, a lack of electricity, and risk, roads in 15 villages in Ethiopia reduced the poverty and equity outcomes. credit, and other market failures. Road benefits incidence of poverty by 6.7 percent. Given depend heavily on interactions with other the heterogeneity of impacts, more attention infrastructure and geographical, community, needs to go to beneficiary selection, recogniz- Sources: Limao and Venables 2001; Van der Walle and household characteristics. For example, ing that tradeoffs exist. Moreover, roads may 2007. market information, extension, and other SMS) and the willingness of mobile phone services to farmers. In West Africa, a public- companies to invest in these systems, sup- private partnership set up TradeNet, a trad- ported by initial donor funding. The new ing platform that allows sellers and buyers to systems have the potential to significantly get into contact over the Internet and by cell reduce transaction costs, especially search phones (box 5.2). and transport costs, and warrant continued Market information systems also dis- investment and evaluation. seminate price information in Kenya, Mozambique, and Senegal, using a mix Commodity exchanges: of Internet, short message service (SMS), fast and low cost voicemail, radio, and market chalkboards.4 Commodity exchanges offer a fast and Local FM radio broadcasts market infor- low-cost mechanism for discovering mation in Mali and Uganda.5 It is still too prices, trading, and resolving contractual early to judge the long-term viability and disputes. A physical exchange is often a impact, but anecdotal evidence points to first step to more sophisticated trading the interest of farmers (with rising use of contracts--initially contracts for forward Bringing agriculture to the market 121 B O X 5 . 2 Innovative uses of information technology to link farmers to markets in India and West Africa E-Choupal and its rural Internet kiosks its delivery to ITC. In 2006/07 ITC purchased receive real-time prices for more than 80 com- Between 2000 and 2007, the agribusiness about 2 million tons of wheat, soybeans, cof- modities from 400 markets across West Africa. division of ITC Limited set up 6,400 Internet fee, shrimp, and pulses valued at $400 million Individual users can advertise their goods and kiosks called e-Choupals in nine Indian states, through the e-Choupal network. This direct offers on free Web sites with their own Internet reaching about 38,000 villages and 4 million purchasing cuts marketing costs for both farm- addresses, and farmer and trader groups can farmers. ITC establishes an Internet facility in ers and ITC. It improves price transparency set up Web sites to manage all these services a village and appoints and trains an operator and allows better grading of produce. It also for their members. (sanchalak) from among the farmers in the allows farmers to realize a bigger share of the The Ghana Agricultural Producers and village. The sanchalak operates the computer final price. Traders Organization (www.tradenet.biz/ to enable farmers to get free information on gapto) is a major beneficiary. In 2006 it con- local and global market prices, weather, and TradeNet, a West African trading platform cluded trade deals worth $60,000 with other farming practices. The e-Choupal also allows with Internet and mobile phones producer and trader organizations in Burkina farmers to buy a range of consumer goods and TradeNet, a Ghana-based trading platform, Faso, Mali, and Nigeria. These deals involved agricultural inputs and services (sourced from allows users to sign up for short message ser- purchasing tomatoes, onions, and potatoes other companies). vice (SMS) alerts for commodities and markets without middlemen, reducing the transaction The e-Choupal serves as a purchase center of their choice and receive instant alerts for costs substantially. for ITC for 13 agricultural commodities, with offers to buy or sell as soon as anyone else on Source: Kofi Debrah, personal communication, the sanchalak acting as the commission agent the network has submitted an offer on their 2007; DeMaagd and Moore 2006; Shivakumar, in purchasing the produce and organizing mobile phone. Users can also request and personal communication, 2007. delivery, and perhaps later, contracts for from the exchange by high transport and futures, options, and swaps. China, India, transaction costs or by quality differences. South Africa, and Thailand have agricul- Establishing exchanges in Africa is chal- tural futures exchanges to facilitate a wider lenging because of continuing govern- range of financing and risk management ment intervention in grain markets, small transactions.6 All four have large domestic markets, and weak systems for warehouse markets and fairly well-developed financial receipts and grades and standards. sectors. India's commodity futures exchanges Price-risk management: expanded rapidly after the government a role for governments? eliminated the ban on their operations in Because of the vulnerability of poor pro- 2004.7 Three national electronic and 21 ducers and consumers to price shocks for regional futures exchanges trade contracts food staples, governments often seek to sta- for cereals, sugar, cotton, potatoes, oilseeds, bilize prices, countering efforts to liberal- and spices.8 The fortnightly turnover totaled ize markets. The variability in world grain $8.7 billion on the three national exchanges prices remains significant, with coefficients in a two-week period in September 2005.9 of variation 20­30 percent for rice, wheat, The South Africa Futures Exchange (SAFEX) and white maize. Domestic price instabil- offers futures contracts on white and yellow ity tends to be high in Africa, especially in maize, wheat, sunflower, and soybeans, and land-locked countries (such as Ethiopia), it traded more than 1.9 million contracts in where the wedge between the export and 2006. Traders throughout southern Africa import parity price is large and drought use SAFEX as a benchmark for pricing increases the impact of domestic shocks physical trades. In 2006 the government of (figure 5.2).11 Malawi used a SAFEX-based call option to The appropriate role of government protect itself from the risk of international in managing food-price risk continues price increases when a bad harvest would to be debated. Opponents of government require significant imports.10 intervention note that price stabilization Futures trading requires good financial policies often lead to economically ineffi- and legal structures and supportive gov- cient production decisions and discourage ernment policies. The benefits diminish if incentives to search for cost-reducing tech- the markets for smallholders are separated nical and institutional innovations. Most 122 WORLD DEVELOPMENT REPORT 2008 Figure 5.2 Wholesale prices in Ethiopia fluctuate within a wide import-export parity band Many developing countries have agen- cies to maintain publicly owned strategic US$ per ton 325 reserves that aim to reduce price instability, but the agencies instead often destabilize Import parity (Addis Ababa) 275 Wholesale price at Addis Ababa prices through unpredictable market inter- Export parity (FOB Djibouti) ventions, border closings, and poorly timed 225 imports.16 Safeguards are needed to prevent this. They include arm's length, central bank­type autonomy; highly professional 175 management and analytical capacity; strict rule-based and transparent market opera- 125 tions to meet a narrowly defined objective; and tendering procurement and storage to 75 the private sector.17 25 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Traditional bulk export Source : Rashid, Assefa, and Ayele 2006. commodities: maintaining international competitiveness often, the government agencies implement- Maintaining international competitiveness ing the policies are subject to inefficiencies, in bulk agricultural commodity exports corruption, and vested interests, resulting is a major challenge for many low-income in huge fiscal costs.12 Proponents of gov- countries, especially in Africa. Competi- ernment intervention, by contrast, show tiveness is important, because exports of that the net welfare effects of food-price coffee, cocoa, tea, cotton, and other bulk instability can be significant for economic commodities are their main source of for- growth and for household food and nutri- eign exchange. For Benin, Burkina Faso, tion security.13 Burundi, and Mali, one such commodity Another view is that the nature and accounts for more than half of the value of extent of price-stabilization interventions total exports. will depend on country-specific factors.14 Producers of these commodities, how- Food-price stabilization is more relevant in ever, have faced a long-term downward low-income countries where food staples trend in prices as global supply outpaced are a large share of the incomes of poor pro- demand (figure 5.3). Productivity increased ducers and the expenditures of poor con- among traditional producers and export- sumers, where one food staple dominates, ers, and new players, such as Vietnam in where domestic production is highly vari- coffee and tea, further expanded supply.18 able, and where poor infrastructure and Increasing productivity to cope with declin- location restrict tradability. ing prices helped some countries in the Lessons over several decades suggest that short term but added to the long-term the design of food-price risk interventions downward pressure on world prices, with should be part of a long-term strategy that consumption stagnating in the major mar- emphasizes measures to raise productiv- kets (Western countries) and growth lim- ity of food staples, improve the efficiency ited in the "new" markets (Eastern Europe, of markets (infrastructure, market infor- the Middle East, and the former Soviet mation, grades and standards, warehouse Union).19 Cotton subsidies in member receipts), and minimize the impact of price countries of the Organisation for Economic shocks (weather-based insurance and safety Co-operation and Development (OECD) nets) (chapter 6).15 Liberalizing trade, espe- further depressed prices (chapter 4). Pro- cially by promoting regional trade, can be jections for coffee, cocoa, and tea indicate a source of "quick wins" for reducing price continuing price declines.20 volatility, especially in small and medium- Another major challenge is the declin- size countries (box 5.3). ing global demand for higher-priced grades Bringing agriculture to the market 123 of cocoa, coffee, and tea, as demand shifts to lower-quality products. Technological B O X 5 . 3 Price stabilization through international trade: advances in processing technology and saving $200 million in Bangladesh bulk transport permit international cocoa grinders to use, and cost-effectively com- To stabilize domestic prices and the urban centers, the total fiscal cost would pensate for, lower-quality cocoa beans.21 availability of food, many countries have have been $160­210 million. accumulated large national stocks as The liberal trade policy helped the And the technology advances in roasting emergency reserves. But market develop- government stabilize prices without large lower-quality robusta coffee allow its sub- ment and trade liberalization provide government stocks. Other factors were another option, potentially less costly and stitution for higher-quality arabica. Chang- also important. The large expansion in the more effective. winter season boro rice and wheat har- ing consumer preferences, such as the shift As floods spread across much of Ban- vests over more than two decades helped toward instant and flavored coffees and gladesh in mid-1998, rice prices rose to reduce the importance of the monsoon convenience teas, further shift demand import parity levels (the export price of rice crop in total domestic production and rice in the exporting country plus trans- minimize the time between major domes- toward lower-quality products. Cotton port and normal marketing costs). The tic harvests. Moreover, large-scale imports is the exception, where the importance liberalization of rice trade induced mas- would not have been possible without attached to lint quality has risen in recent sive imports of rice by hundreds of small market infrastructure, particularly roads years with the widespread use of high-speed traders. Private rice imports, estimated and bridges. Nor would they have been at 2.42 million tons, were 6.1 times larger possible without pro-market policies, spinning machines with demanding quality than government rice distributions. If the including no limits on private stocks and requirements. government had imported this grain, the no restrictions on movement and access to Specialty markets (organic, gourmet, added cost of the imported rice delivered foreign exchange. Fair Trade) offer an alternative higher- to local delivery points would have been $50­100 million. And if the government priced market, but they account for only a had subsidized this rice by selling it at the Sources: Del Ninno and others 2001; Dorosh small share of the global market (see section price used for limited government sales in 2001. on specialty markets). Currently, the spe- cialty coffee sector accounts for only about 6­8 percent of global consumption.22 Many countries, such as Tanzania and its Kiliman- Figure 5.3 World prices for traditional bulk exports continue to decline jaro specialty coffee or KILLICAFE initia- tive, are targeting these markets to expand Price (cents per kilogram) export markets and increase revenues.23 1000 Cocoa Coffee arabica Different paths to liberalizing Coffee robusta Tea (3 auctions) 800 Cotton Rubber domestic markets Bulk export commodity markets in Africa 600 were traditionally controlled by parastatal agencies, which often had monopoly pow- ers in domestic marketing, exporting, 400 and providing inputs to farmers (seeds, fertilizer, credit, extension services). The 200 parastatals also aimed to stabilize prices received by farmers. In many instances 0 the vertical coordination arrangements in 1970 1975 1980 1985 1990 1995 2000 2005 2010p 2015p production and marketing enabled farmers Source: World Bank data group. to overcome market failures in the input, Note: Prices are in constant 1990 dollars. Prices from 2007 onward are projected. credit, and insurance markets. They also ensured a steady supply of products of assured quality for export. But the agencies and 1990s, and many parastatal agencies were widely criticized for inefficiencies and were abolished or restructured. mismanagement that lowered the prices African countries that restructured their paid to farmers and raised the fiscal costs to bulk commodity markets followed differ- government.24 To redress these failures, the ent paths, ranging from restructuring the bulk commodity markets in many coun- parastatal ownership to include the private tries in Africa were liberalized in the 1980s sector and farmers (for example, cotton in 124 WORLD DEVELOPMENT REPORT 2008 Burkina Faso), to market zoning (for exam- producing countries in Africa. The absence ple, cotton in Ghana), to full market liber- of a clear legal and regulatory framework alization (cotton in Uganda, and cocoa and to guide private sector and farmer behav- coffee in Cameroon and Côte d'Ivoire).25 ior in the context of free market competi- Overall the liberalization programs gener- tion or weak contract enforcement created ated immediate benefits: an influx of pri- confusion and allowed some malpractices vate capital, management, and marketing to persist (box 5.4). To help private trad- expertise; and market competition reduc- ers enforce contracts, Côte d'Ivoire and ing transaction costs, increasing prices Zambia adopted zoning arrangements to received by farmers and typically leading to regulate cotton marketing that have worked prompter payment for crops purchased.26 reasonably well.28 However, competition One study found that 85 percent of coffee from new buyers in Zimbabwe and Tanza- producers in Tanzania were better off as nia weakened quality enforcement.29 the gains from higher producer prices more What contributed to these second- than offset the loss from reduced access to generation problems? The weaknesses and credit through public sources.27 lack of credibility of public institutions to enforce appropriate rules of behavior for After liberalizing: addressing the private sector is part of it. Public inter- second-generation problems vention in grades and standards and in In many countries, the restructuring of the contract enforcement is essential to ensure market brought second-generation prob- that private markets work. Liberalization lems, aptly illustrated by cotton in major also exposed the underdevelopment of rural financial systems, which need to be addressed (chapter 6). The African expe- B O X 5 . 4 Zambia and Burkina Faso: contrasting rience also highlights the potential for experiences in liberalizing domestic associations and professional organiza- cotton markets tions (farmer groups in Tanzania) to over- come the shortsightedness of individual Zambia--production triples, after some collecting produce on behalf of Dunavant. farmers and buyers.30 Partial privatiza- fixes. Zambia's cotton sector continues The distributor's remuneration was directly tion in Burkina Faso has given farmers to evolve after market liberalization, with tied to the amount of credit recovered, on more ownership, but it led to heavy fiscal significant impacts on productivity and an increasing scale. Dunavant established quality. In 1995 the government sold the inspection points in all buying stations to outlays (box 5.4). Lint Company of Zambia, the government enforce quality standards. National pro- parastatal, to two private companies, duction tripled between 2000 and 2003, Higher-value urban markets: Clark Cotton and Lornho, later acquired and credit repayments improved from by Dunavant. To ensure access by partici- linking producers to modern about 65 percent to more than 90 percent. pating farmers to extension services and There are now more than 300,000 cotton- supply chains inputs (on loan), the two companies imple- producing farmers in Zambia. mented outgrower schemes, contracting Burkina Faso--losses of $128 million. Rising incomes, urbanization, greater with smallholders. The costs of the inputs The government tried to reduce inefficien- female participation in the workforce, were to be paid by farmers upon sale of cies by changing the structure of ownership wider media penetration--all are driv- their seed cotton. But the rapid entry of of SOFITEX, the cotton parastatal, in 1999. ing the demand for higher-value products, other buyers created overcapacity in gin- It allowed producers, represented by the ning and fierce buyer competition. The Union Nationale des Producteurs de Coton du semiprocessed and processed products, outgrower schemes began to fail because Burkina Faso, to take up 30 percent owner- and convenience foods (figure 5.4). They of rampant side-selling by farmers to ship, empowering farmers to oversee the are also increasing consumer attention to other traders offering high prices without management of SOFITEX and ensure profes- food quality and safety. Diets are global- grading and defaults on input loans. As sional management. But the institutional the defaults increased, the cost of credit changes at SOFITEX did not improve its izing too, with local consumer preferences increased, which led to more defaults or financial position. Supporting and stabiliz- influenced by international tastes. These exits from the outgrower program. Produc- ing domestic cotton prices as world prices trends open new markets for a wide range tion in 2000 was less than half that in 1998. declined produced financial losses of $128 of higher-value agricultural products and After 2000 many agents and buyers million from 2004/05 to 2006/07. exited the industry, leaving two dominant propel the evolution of the marketing sys- Sources: Bonjean, Combes, and Sturgess companies. Dunavant used distributors to tem in many developing countries, with 2003; Food Security Research Project improve credit repayments. Distributors (FSRP) 2000; Christopher Gilbert, personal the entry and rapid growth of supermarket were responsible for identifying farmers, communication, 2007; Tschirley, Zulu, and chains and the food processing and food providing inputs and technical advice, and Shaffer 2004. service industries. Bringing agriculture to the market 125 Figure 5.4 Food consumption expenditures in in developing countries, either directly or Indonesia are shifting from cereals to higher-value through joint ventures with local firms. and prepared foods Changing consumer demand is also Share in per capita food expenditure, % driving the growth of the food process- 40 39 ing and food service industries. Processed 1981 foods account for about 80 percent of 30 2005 global food sales, estimated at $3.2 trillion 22 22 22 22 in 2002.31 Although spending on processed 20 20 19 foods is still low in developing countries 14 15 ($143 per capita per year in lower-middle- income countries and $63 per capita in 10 6 low-income countries), it is growing fast- est in these countries--28 percent a year 0 Cereals Beverages Meat, Fruits Prepared in lower-middle-income countries and 13 and and eggs, and foods percent a year in low-income countries. tubers other fish, vegetables "Eating out" is also becoming popular. For foods and dairy example, spending on food services now Source : Badan Pusat Statistik Indonesia, http://www.bps.go.id. accounts for 22 percent of food budgets in Brazil and Indonesia and 15 percent of urban food spending in China. For many developing countries, the supermarket revolution began in the early Infrastructure impediments to mid-1990s. By the early 2000s, retail food The perishability of most high-value agri- sales in supermarkets exceeded 50 percent cultural products requires careful handling, of total retail food sales in many countries special facilities (packhouses, cold storage, in Latin America and in major urban cen- and refrigerated transport), and rapid deliv- ters elsewhere (figure 5.5). Accelerating the ery to consumers to maintain quality and expansion: significant foreign direct invest- reduce physical and nutritional losses. In ment by multinational supermarket chains many developing countries, the long supply Figure 5.5 Rising per capita incomes drive supermarket growth Share of supermarkets in retail food 80 United Kingdom Brazil France United States 70 60 Argentina Mexico S. Africa Czech Republic Thailand 50 Chile Korea, Rep. of Philippines Costa Rica Colombia 40 Ecuador Guatemala 30 Indonesia Bulgaria Kenya 20 Nicaragua China 10 Bangladesh 0 India 0 5 10 15 20 25 30 35 40 45 GDP per capita, thousand PPP $ Sources : Reardon and Berdegué 2006; World Bank 2006y. 126 WORLD DEVELOPMENT REPORT 2008 chain, poor access to roads and electricity, and convenience of small produce shops and inadequate infrastructure and services and wet markets offer a competitive alter- in physical markets add to the transaction native. Fresh fruits and vegetables generally costs and cause quality deterioration and account for the lowest share in supermarket high spoilage losses. In India it is estimated sales, and small shops and wet markets will that fruit and vegetable postharvest losses likely remain important marketing chan- amount to about 40 percent of total annual nels for these products for years to come.36 production, equal to a year's consumption Significant inefficiencies in the tradi- in the United Kingdom.32 tional wholesale marketing systems and Market infrastructure and facilities in competition encourage supermarkets, food developing countries are often limited and processors, and food service providers to congested, increasing the difficulty of trad- use supply chains to reduce coordination ing perishable goods. A survey of wholesale costs, capture economies of scale, and markets handling fresh produce in four increase food safety and quality. This is states in India found that 17 percent had profoundly changing the structure of pro- no covered shops, about half did not have duction and wholesale marketing in many paved roads in the market yard, about 40 developing countries. Recent studies show percent of the shops had no electricity, and that procurement systems change earliest only 6 percent of the markets had a cold- for processed foods, meat, and dairy prod- storage facility.33 In Tamil Nadu, India, a ucts, eventually extending to fresh fruits related study found that wealthier farmers and vegetables.37 tend to capture a disproportionate share of Procurement takes many forms, varying the benefits of facilities in congested whole- by supermarket chain, product, and coun- sale markets.34 Nonetheless, investments in try.38 It can involve centralized procurement, market facilities would be pro-poor because which shifts from fragmented per-store sales by poorer farmers would increase pro- purchases to operating a distribution center portionally more than those by the wealthy catering to a district (as in China), the whole farmers. country (as in Mexico), or whole region (as in Central America). It can also involve Modern procurement systems shifting from purchases in traditional spot Supermarket growth in most countries wholesale markets to relying on specialized follows similar diffusion patterns across or dedicated wholesalers and logistics firms space, consumer segments, and product (as in Central America and East Asia) or to categories.35 From a base in large cities, direct contracting (as in East Asia and East- supermarkets initially spread to interme- ern Europe)--to cut transaction, coordina- diate cities and towns, and later to small tion, and search costs and ensure greater towns in rural areas--in response to mar- control over quality and consistency of sup- ket competition and saturation. They often ply.39 China Resources Enterprise estimates first target the upper-income consumer that it is saving 40 percent in distribution (national and expatriate), followed by the costs by combining modern logistics with middle class and later the urban lower- centralized distribution in its two large new income households. centers in southern China.40 Dominating the supermarket's product Modern procurement can also involve selection in the early stages are processed contracting with processors and farmers foods (canned, dry, and packaged food or using preferred-supplier lists. This is items), motivated by economies of scale often done where farmers or processors are in procurement and direct relations with grouped or are individually large (as in the processed-food manufacturers. Product Philippines, Russia, and Thailand).41 The selection gradually expands to semipro- contracts are incentives for suppliers to stay cessed foods (dairy, meat, and fruit prod- with the buyer and invest in assets that fit ucts). The last category to be added is fresh the retailer's specifications for products. The fruits and vegetables, as consumer prefer- arrangements may include direct or indirect ence for fresh produce and the proximity assistance for farmers to invest in training, Bringing agriculture to the market 127 management, inputs, and basic equipment. twice the farm size (two hectares versus Modern procurement also often involves one) and 40 percent more education than private standards and their enforcement-- nonparticipating farmers, and are nearly standards that serve two main functions.42 twice as likely to have irrigation, four times They help coordinate supply chains by stan- as likely to have a truck, and twice as likely dardizing product requirements for suppli- to be close to paved roads and be in a farmer ers over many regions or countries, enhanc- organization. Participating farmers use ing efficiency and lowering transaction costs. much more labor-intensive practices because And they help ensure that public food-safety of requirements for field practices, sorting, standards are met in all markets served by and packing. Because they are more likely the retail chain or food-processing firm, to double-crop over the year, participating distinguishing one's products from com- farmers hire 2.5 times more labor (typically petitors through signaling.43 As these private from local asset-poor households). So even standards are more widely adopted, there is if small farmers do not participate directly, growing concern about the capacity of small they can benefit through farm employment farmers to meet them. (chapter 9). Studies of tomato growers in Indonesia and kale growers in Kenya find Impact on smallholders and retailers similar results. The modernization of procurement systems Participation in modern supply chains affects farmers differently across coun- can increase farmer income by 10 to 100 tries and products. Some recent studies of percent (Guatemala, Indonesia, Kenya).48 selected commodities find that the modern Recent studies of contract farmers show procurement systems exclude asset-poor that they have significantly higher incomes farmers. Supermarket buying agents prefer than other farmers.49 Because participating to source from large and medium-size farm- farmers tend to reap substantial benefits, ers if they can (for example, for tomatoes in the payoff from assisting farmers to make Mexico and potatoes in Indonesia); if large the necessary "threshold investments" can and medium-size farmers have sufficient be high. quantities, smallholders are not included.44 Some studies have found that smaller Where small farms are the dominant struc- processing firms were left out of the supply ture, supermarkets have no choice but to chain, with medium-size and large proces- source their produce from them. Super- sors preferred for long-term contracts.50 The markets may also rely on small farmers to number of small retail stores often declined satisfy consumers' demand for specialty or with rising market share for supermar- niche products that only small farmers with kets--with implications for employment. abundant labor produce. Sometimes super- In urban Argentina, from 1984 to 1993, the markets need an advertising tool to promote most intense period of supermarket takeoff, sales with socially conscious consumers: the number of small food shops declined "buying local, from smallholders."45 from 209,000 to 145,000.51 But the competi- The most important determinant of small tion is also driving some small retail stores farmers' participation is not always farm size. and processors to grow and upgrade their Instead, it can be access to physical, human, services (as in India).52 and social assets: to education, irrigation, transport, roads, and such other physical Helping smallholders keep up with assets as wells, cold chains, greenhouses, the requirements good quality irrigation water (free of con- The government and the private sector can taminants), vehicles, and packing sheds.46 help smallholders expand and upgrade their An effective producer organization--another range of assets and practices to meet the new major asset--can also help small farmers requirements of supermarkets and other enter the high-value supply chains. coordinated supply chains (table 5.1). The Most farmers lacking these assets are options include public good investments to excluded.47 In Guatemala, lettuce farmers increase farmers' productivity and connec- participating in modern supply chains have tivity to markets, policy changes to facilitate 128 WORLD DEVELOPMENT REPORT 2008 Table 5.1 Public and private options for strengthening farmer links to the market Public sector Issue Public investments Policy environment Private sector Lack of access to markets Invest in education, rural Liberalize domestic trade; foster Assist farmers in forming producer infrastructure (roads, markets, development of input and credit organizations electricity, irrigation); support markets formation of producer organizations Weak technical capacity Support market-oriented extension Foster environment for private Provide extension and key inputs to extension to emerge farmers Meeting quality standards Support farmer training on good Establish grades and standards Supply inputs and train farmers on agricultural practices for quality quality management and food safety enhancement and food safety Meeting contract conditions Train firms in contract design and Foster institutions for dispute Foster trust; develop contracts that management; train farmers on their resolution; strengthen producer are self-enforcing rights and obligations organizations Farmer exposure to risk Foster development of commodity Create enabling environment for Use contracts that share risk equally and futures exchanges; train firms insurance market among parties; assist farmers to on use of market instruments to access insurance hedge risk Source: Adapted from World Bank 2007e. trade and market development, and public- with producer organizations or commercial private efforts to promote collective action farmer "leaders" that supplement their own and build the technical capacity of farmers production with that from individual small to meet the new standards. farmers (box 5.5). The producer organiza- Some supermarkets and processors or tions or farmer leaders provide technical their agents help farmers overcome their assistance to ensure quality, quantity, and asset constraints and improve their busi- timing of delivery. In addition, the pre- ness image by providing technical assis- ferred supplier or wholesaler often expects tance, in some instances through public- the producer organizations or farmer lead- private partnerships.53 Examples include ers to assemble the products (washing, joint extension by supermarket field staff sorting, grading, packaging, and labeling), and government extension officers, techni- ready to be placed on supermarket shelves. cal assistance to acquire inputs and obtain Many producer organizations do not certification, and training to improve prod- have the capacity to provide their mem- uct quality and food safety. bers with the technical assistance required Other supermarkets and processors enter for ensuring collective compliance with into production contracts, which sometimes quality, quantities, and timing (chapter 6). include the supply of inputs, credit, and Well-targeted technical and financial sup- extension services (for example, in Mada- port from donors, governments, or nongov- gascar and Slovakia).54 For many small ernmental organizations is often necessary farmers, these contracts are the only means for producer organizations to overcome to acquire inputs and use support services. these initial hurdles and become profes- By supplying inputs and providing assured sional entrepreneurs.55 The support must markets and prices, contracting firms share be provided with a long-term commitment production and marketing risks with farm- but with a clear phase-out strategy and a ers. Reducing these risks helps stabilize view to empower (chapter 6). farmers' incomes, critical in the absence of insurance markets. The technical assistance Higher-value exports: to farmers also generates indirect benefits, as farmers apply the improved farm prac- meeting product standards tices for the contract crops to other crops, Agricultural exports diversified signifi- increasing their productivity. cantly in the last two decades, particularly Supermarkets also procure through pre- into high-value fresh and processed prod- ferred suppliers or wholesalers that contract ucts, fueled by changing consumer tastes Bringing agriculture to the market 129 B O X 5 . 5 Linking small farmers to high-value chains: Three approaches The Philippines: a farmer leader and assorted vegetables--from 30 to 40 tons when members. With its marketing success, the small-farmer clusters it started to operate in May 2006, to 80 tons cooperative's membership increased from NorminVeggies is a multistakeholder associa- two months later. 29 to 152, its farmed area increased from 0.2 tion supplying vegetables to the fast-food hectare in 1992 to thousands of hectares in industry, supermarkets, and vegetable proces- China: farmer marketing through 2005, with total capitalization reaching RMB 21 sors in the Philippines. In December 2003 it a cooperative million in 2005. started Normincorp, a marketing company Supported by local government, a group of that links the farmer directly to the buyer, in small-scale growers registered the brand Croatia: supermarket assists farmers exchange for a 6 percent facilitation fee. The "Yulin" for their watermelons, with production to obtain investment loans farmer, liable for the product, retains owner- standardized through coordinated plant- In Croatia the supermarket chain Konzum ship over it all along the chain. Normincorp ing, quality inspection, and packaging. They established preferred-supplier programs to forms production clusters: a group of 10 small formed the Ruoheng watermelon cooperative procure strawberries. It encourages suppliers farmers allied with a commercial lead farmer to ensure their proprietary techniques and to use irrigation and greenhouses to reduce who helps jump-start quality production. expand their marketing network. The coopera- the seasonality of strawberry production and The clusters commit to undertake a common tive sells directly to wholesalers (40 percent), improve the quality of produce. Such invest- production and marketing plan for a particular supermarkets (25 percent), and retailers (35 ments require significant capital, which many product for an identified market. The lead percent), which buy from the cooperative farmers did not have, nor did they possess farmer coordinates the production processes because it can deliver large volumes on a regu- enough collateral to secure bank loans. So of the cluster farmers and is responsible for lar and timely basis and ensures food safety Konzum negotiated with the local banks to use training them to ensure the quality specified and quality standards. The "Yulin" watermelon the farmers' contracts with the supermarket as by the market. Normincorp has become the high-quality brand image commanded a a "collateral substitute." preferred supplier for several clients thanks higher price than other watermelons (3.0 yuan Sources: Concepcion, Digal, and Uy 2006; Dries, to its ability to respond to changes in market per kilogram versus 1.2 yuan per kilogram), Reardon, and Swinnen 2004; Zuhui, Qiao, and Yu requirements. It doubled monthly sales of increasing the income of the cooperative's 2006. and advances in production, transport, and Figure 5.6 Transport, handling, and packaging are other supply-chain technologies (chapter major costs for French bean exports in Bangladesh 2). Comparatively low and declining tar- $ per 100 kilograms, 2004 iff barriers and year-round supplies also 4.6 34.1 increased the competitiveness of develop- ing-country exports.56 Fresh and processed 13.5 fruits and vegetables, fish and fish products, meat, nuts, spices, and floriculture account airport for about 47 percent of the agricultural to exports from developing countries, which in 2004 amounted to $138 billion (chapter Transport 2). Continued growth of these high-value 2.3 0.4 exports will require efficient value chains, 0.9 0.3 particularly domestic transport, handling, 12.1 Cooling Packing and packaging, which make up a large share of the final costs (figure 5.6). center Grading to cleaning price Meeting sanitary and price Sorting, phytosanitary standards Farm Transport processing FOB For agrofood products, sanitary and phy- Source: Global Development Solutions LLC data 2004. tosanitary (SPS) standards govern inter- national trade to address food safety and agricultural health risks associated with fication in agricultural exports focus atten- pests (fruit flies), food-borne and zoo- tion to how widely the standards for food notic diseases (foot and mouth and mad safety and animal and plant health diverge cow diseases), and microbial pathogens across countries--and the different capaci- and other contaminants (mycotoxins and ties of governments and commercial supply pesticides). The rapid growth and diversi- chains to manage them. 130 WORLD DEVELOPMENT REPORT 2008 In reaction to the periodic "food scares" management functions of government. The in industrial countries, coupled with bet- greater attention to good practices in agri- ter scientific knowledge and greater public culture and food processing may not only concern about these various risks, many improve export competitiveness, but also countries tightened their SPS standards or generate spillover benefits to domestic con- extended their coverage to new areas. Pub- sumers. Although there will inevitably be lic standards were also introduced to ensure winners and losers, this view suggests that fair competition, reduce information costs enhanced capacity to comply with stricter to consumers (organic foods), and promote standards can provide the basis for more competition based on quality.57 In paral- sustainable and profitable agrofood exports lel, the private sector developed standards in the long term. and supplier protocols to ensure compli- There is general agreement that SPS ance with official regulations, fill perceived standards affect agrofood trade, but there gaps in such regulations, differentiate their is no consensus on the relative importance brands in a competitive market place, and of individual measures in relation to other otherwise manage their commercial and trade-distorting measures, or on the aggre- reputational risks.58 These standards tend gate net effects of those measures. The lack to blend food safety and quality manage- of consensus is not surprising, because esti- ment concerns--or to have protocols that mating the impact of such standards pres- combine food safety, environmental, and ents enormous empirical difficulties. Sev- social parameters (child labor, labor condi- eral studies based on econometric models tions, and animal welfare). An example is have estimated very large potential losses protocols developed by the transnational in trade.62 In contrast, most industry case Euro-retailer, Produce Working Group studies identify an array of competitive fac- for Good Agricultural Practices, which tors affecting trade (of which standards are includes 33 retail and food service compa- only one) and typically point to both "win- nies in Europe and Japan.59 ners" and "losers," not to absolute declines A concern for developing countries is in trade. When the Guatemalan raspberry the proliferation and stringency of food- industry faced official and private market- safety and health measures being adopted access problems following an outbreak of in export markets. Many fear that the food-borne illness in the United States, emerging standards will be discriminatory many leading operators shifted their pro- and protectionist. Developing countries duction base across the border into Mexico. worry that they will be excluded from the While the Guatemalan industry has never export markets because they lack in-coun- recovered, exports from Mexico and Chile try administrative and technical capacities have served an expanding market.63 to comply with the requirements or that the costs of compliance will erode their com- Meeting the costs of compliance petitive advantage. The standards could Despite the worry that SPS standards and further marginalize weaker economic the cost of compliance will disadvantage players, including smaller countries, enter- developing countries, recent studies find that prises, and farmers. Both anecdotal cases compliance costs64 tend to be small relative and research lend some evidence to support to the scale of most export industries. Fixed, this "standards as barriers" perspective.60 nonrecurrent costs are generally 0.5 percent An alternative view highlights the to 5.0 percent of three-to-five-year exports, opportunities in the evolving standards while recurrent costs tend to be 1 percent environment and the scope for capitaliz- to 3 percent of annual exports.65 The focus ing on them.61 Common public and private on compliance costs can distract countries standards across international markets can from the benefits, many of them long term reduce transaction costs. Standards can and intangible. Productivity gains, reduced also provide incentives for modernizing wastage, worker safety, environmental ben- developing-country supply chains and help efits, and the value of continuing market clarify the necessary and appropriate risk access can be underestimated or not counted Bringing agriculture to the market 131 as benefits. Compliance can also generate export standards that led to the shift from spillover benefits to domestic consumers smallholder contract farming to large-scale from greater awareness of food-safety risks integrated estate production, the higher and access to safer products. horticulture exports increased incomes Empirical work on the impact of more and reduced regional poverty by about 12 stringent standards on smallholder partici- percentage points and extreme poverty pation in higher-value supply chains show a by half.67 Poor households benefited more mixed picture. In theory, there are economies through labor markets than through prod- of scale in product traceability, certification, uct markets (box 5.6). and testing that tend to provide a competi- tive advantage to larger production units. Looking at the benefits and choices Yet there are examples from many countries Developing-country suppliers rarely face where, because of limits on land acquisition all-or-nothing choices when determining or other features of the agrarian structure, the changes and investments to conform smallholders remain the dominant suppli- to emerging standards. They have a range ers for export firms.66 Consequently, insti- of choices. One is compliance--adopting tutional arrangements have been developed measures to meet the standards. Another is to manage the attendant risks and transac- voice--seeking to influence the rules of the tion costs of sourcing exports with exacting game. A third is redirection--seeking other standards from smallholders. markets and countries or changing the mix Also important is the large increase in of products.68 Suppliers need to weigh the off-farm work opportunities with expanded costs and advantages for different products agrofood exports. In Senegal, despite tight and market segments. In some cases, there B O X 5 . 6 Employment gains and reduced poverty in rural Senegal Fresh fruit and vegetable exports from Sen- downstream buyers (to ensure markets) of households to 34 percent). While con- egal to the European Union (EU) increased and upstream suppliers (to guarantee food tract farming favored larger farmers, poorer significantly in the last 15 years, despite the safety, quality, and the timing of production). households participated as farm workers. Par- tightening of SPS standards. Senegal's main Increased vertical coordination led to the ticipation in fresh fruit and vegetable export export is French beans, which account for 42 shift from contract farming with smallholders production, whether as a worker or contract percent of fresh fruit and vegetable exports, to large-scale estate production in agroin- farmer, raised household incomes (figure more than doubling from 3,000 metric tons in dustrial farms. below). Estate farm workers had incomes 1991 to 7,000 metric tons in 2005. Changing The incidence of contract farming 1.2­2.3 million CFA francs higher than non- EU SPS standards put pressure on export- declined (from 23 percent of participating participating households, while contract ers to invest more to meet these standards households to 10 percent), but employment farmers had incomes between 2.4 million and and to increase vertical coordination with in estate farms increased (from 10 percent 4.1 million CFA francs higher. Household participation in French bean export production in Senegal Incidence of poverty and extreme poverty among participating households, 2005 Share of participating households, % Share of households, % 40 60 Poverty Total participants Extreme poverty 30 40 20 20 10 Contract farmers Estate farm workers 0 0 1991 1993 1995 1997 1999 2001 2003 2005 Non-participants Estate farm Contract workers farmers Source : Maertens and Swinnen 2006. 132 WORLD DEVELOPMENT REPORT 2008 Table 5.2 Public and private sector roles to enhance trade-related SPS compliance and quality management capacity Public sector Private sector Policy and regulatory environment Good management practices Pursue international dialogue; adopt domestic food safety legislation and Implement appropriate management practices (hazard analysis and critical standards consistent with local conditions and preferences, WTO, and other control point, "good" agricultural practices); obtain formal certification where trade obligations viable Risk assessment and management Traceability Strengthen national or subnational systems for pest, animal disease, and market Develop systems and procedures to enable traceability of raw materials and surveillance; support research on food safety and agricultural health concerns intermediate and final products Awareness building and promoting good practices Develop training, advisory, and conformity assessment services Support consumer awareness campaigns on food safety; promote good Strengthen human capital, physical infrastructure and management systems agricultural hygiene, and food processing practices to be integrated into to supply support services to agriculture, industry, and government related to extension programs; invest in appropriate laboratory infrastructure; accredit quality and food-safety management private laboratories Collective action and self-regulation Infrastructure investments Self-regulate through adoption and oversight of industry "codes of practice"; Improve water supply and sanitation and marketing facilities alert government to emerging issues; advocate for effective government services Source: Adapted from World Bank (2007e). may be larger and more profitable oppor- from developing countries. Geographic tunities to serve the domestic market, a indications (labeling such as Blue Mountain regional market, or industrial-country seg- coffee from Jamaica), which capitalizes on ments that impose less stringent standards local know-how and special agroecologi- or allow more time to implement them. cal conditions to establish brand identity, Addressing the export challenges of SPS are one example. Organic, Fair Trade, and standards requires joint public and private Rainforest Alliance­certified products are efforts. The public sector should take the others. Organic products are grown without lead in policy (standards and food-safety the use of conventional pesticides, artificial legislation), in research on risk assessment fertilizers, or sewage sludge--and pro- and good management practices, and in cessed without ionizing radiation or food disease surveillance (table 5.2). The pri- additives.71 Fair Trade seeks greater equity vate sector should take the lead in building in international trade and aims to contrib- awareness, training, and complying with ute to sustainable development by offering food-safety and agricultural-chemical-use better market conditions and securing the requirements, either individually or collec- rights of marginal producers and workers.72 tively through trade associations. Rainforest Alliance­certified products meet There is growing evidence that countries stringent environmental and social stan- staying abreast of technical and commer- dards for production. cial requirements and anticipating future Retail sales, mainly to meet the grow- changes have repositioned themselves in ing demand in high-income countries, and more remunerative market segments.69 To area planted under these products have strengthen local capacity to meet these stan- expanded significantly. The area planted to dards, developing countries can draw sup- organic crops reached 31 million hectares port from the Standards and Trade Devel- in 2005, with retail sales reaching $23.9 opment Facility, a global program aimed at billion in the EU, Canada, United States, providing financial and technical assistance and Asia in 2006.73 The biggest developing- to countries to enhance their expertise and country producers of organic products are capacity to analyze and implement SPS China and middle-income Latin American standards and improve their human, ani- countries. Sub-Saharan countries account mal, and plant health situation.70 for a large proportion of organic cotton production, while Asia and Latin America Decommodification dominate production of organic coffee and in specialty markets cocoa. Retail sales of certified Fair Trade The "decommodification" of some tradi- products in high-income countries reached tional agricultural products opens alter- $1.4 billion in 2005. Bananas and coffee are native markets for higher-value products the most traded products of Fair Trade.74 Bringing agriculture to the market 133 Fair Trade: How fair? defined by the World Trade Organization Most case studies highlight the positive reduce this risk. The risk of protection is impact of Fair Trade on producer prices, lowest for tropical products with limited incomes, and well-being. Some benefits of developed-country domestic competi- Fair Trade include building capacity (sup- tion and highest for in-season temperate port services, improved market informa- products.78 tion and awareness), empowering local An expansion of developing-country non- actors, mitigating gender imbalances, and traditional exports could create an adding- providing clear environmental benefits.75 up problem if several countries rapidly There are concerns, however, about the expand production, perhaps so much that sustainability of Fair Trade. Producers in export revenues decline. The potential for some developing countries face problems of this is greatest in commodity markets for rationing, because Fair Trade prices are set unprocessed foods.79 The potential compe- above market clearing levels and potential tition posed by efficient large producers-- supply is exceeding demand. There are also such as Brazil and China--can also be sig- concerns about long-term effects on invest- nificant.80 The Food and Agriculture ment and productivity and the efficiency of Organization of the United Nations esti- Fair Trade channels. But few evaluations mates that an increase in China's exports of have been carried out. green beans is likely to reduce world market Recent studies show that the costs and prices, with adverse effects on the export margins for coffee sold through Fair Trade revenues of other developing countries.81 So are high, and that intermediaries, not farm- under some circumstances, the expansion ers, receive the larger share of the price pre- of agricultural exports by some market par- mium. One estimate is that growers receive ticipants could curtail market potential. 43 percent of the price premium paid by the A close eye needs to be kept on export consumer for Fair Trade roasted coffee and products dominated by one or two coun- 42 percent for soluble coffee.76 The higher tries--or when smaller countries simul- cost of processing and marketing is partly taneously expand their export market explained by the diseconomies of scale shares.82 This emphasizes the need for related to the small volumes and high asso- export-promotion agencies in developing ciated costs: certification of supply-chain countries to build stronger capacities in actors, membership fees, advertising, and market intelligence. campaigning.77 Conclusion Market saturation: more production Markets are good for efficiency, and much at lower prices? progress has been made in market devel- There is also concern about export market opment, especially under private sector saturation for high-value exports, as devel- leadership. But further efficiency gains will oping countries jump onto the same export require public sector support to deliver the bandwagon, often referred to as the "adding necessary public goods, foster institutional up" or "fallacy of composition" problem. If innovation, and secure competitiveness. all countries, and especially large countries, Because efficient markets do not always try to substantially increase their exports secure socially desirable outcomes, comple- of a product, there is a risk that they will mentary policies are often needed to ensure encounter rising protection from industrial smallholder participation. countries--or that the terms of trade will A large agenda remains in improving the decline so much that the benefits of any performance of the marketing systems in increased export volume are more than developing countries. Public investments to offset by lower export prices. While there expand access to rural infrastructure and is some evidence that developing countries services--such as rural roads and transport face protectionist tendencies from indus- services, physical markets, telecommunica- trial (and also some developing) countries tions, and electricity--will be critical to when exports pass a threshold, the rules reducing transaction costs and physical losses 134 WORLD DEVELOPMENT REPORT 2008 and to enhancing transparency and competi- tions to guard against opportunistic and tiveness in traditional markets. Technical uncompetitive behavior in the marketing and institutional innovations reducing trans- system. Public-private partnerships can action costs and risks also show promise, also be important in conducting research especially the wider use of information tech- and capacity building to develop good agri- nologies (mobile phones, the Internet, and cultural practices, meet the new domestic commodity exchanges) and vertical coordi- and international SPS standards, and train nation arrangements with individual farmers and assist farmers to adopt them. or producer organizations. The public sector can facilitate small- Rapidly growing local and international holder access to the big opportunities offered demand for high-value agricultural products by market development. Greater access to opens important growth opportunities for assets for smallholders (as has clearly been the agricultural sector in developing coun- seen in the procurement preferences of super- tries. However, modern procurement sys- markets), level playing fields, and strong tems for integrated supply chains and super- producer organizations to achieve scale and markets with stringent food-safety standards market power are necessary elements. The raise concerns about how to ensure that opportunities offered by major changes in developing countries in general, and small markets will work for the poor only if these farmers in particular, share in these growth complementary policies are in place. opportunities. The private sector can enable smallhold- International experience highlights the ers to participate as partners in modern pro- respective roles of the government and the curement systems and exports. It can setup private sector to meet these challenges. A innovative vertical coordination arrange- priority area for public action is to establish ments with farmers or producer groups. It an enabling policy environment (competi- can facilitate farmer access to credit, inputs, tion policy, contract enforcement, setting extension, and certification. It can support grades and standards, food-safety legisla- the training of farmers in good agricultural tion). It will also involve developing cred- practices to meet quality, food-safety, and ible public institutions to enforce regula- international sanitary standards. focus D Agribusiness for development A dynamic private agribusiness sector linking farmers and consumers can be a major driver of growth in the agricultural and the rural nonfarm sectors. But growing agribusiness concentration may reduce its efficiency and poverty reduction impacts. A better investment climate for small and medium enterprises can improve competitiveness. Targeted public-private sector partnerships and corporate social responsibility initiatives are instruments to promote smallholder participation. A gribusiness is the off-farm link in small, mostly in rural market towns, and competitive small and medium agroenter- agrofood value chains. It provides operated by households that often have prises in rural areas can link smallholders to inputs to the farm sector, and it wage labor and farming as other sources value chains and urban demand. links the farm sector to consumers through of income.3 Medium and large agroenter- the handling, processing, transportation, prises are mainly urban based because of The agribusiness revolution: marketing, and distribution of food and the requirements for economies of scale Are there tradeoffs? other agricultural products.1 Thus, there are and infrastructure. The large enterprises are Growing concentration in the strong synergies between agribusiness and often dominated by multinational corpora- agribusiness sector the performance of agriculture for develop- tions that have consolidated through verti- Driven by gains from economies of scale ment. Dynamic and efficient agribusiness cal and horizontal integration.4 and globalization of the food chain, multi- spurs agricultural growth. And a strong link In recent years, influenced by changes in national agroenterprises increasingly domi- between agribusiness and smallholders can consumer demand and rapid technological nate the agribusiness sector along the value reduce rural poverty. and institutional innovations, the struc- chain. They provide inputs such as pesti- Agribusiness has a large and rising share ture of agribusiness has changed dramati- cides, seeds, and crop genetic technologies of gross domestic product (GDP) across cally and its performance has been highly that have consolidated horizontally and ver- developing countries (figure D.1). Though dynamic. Two major challenges need to be tically into a small number of multinational agriculture declines from 40 percent of GDP addressed in considering the role of agri- firms (table D.1). On the marketing side, a to less than 10 percent as GDP per capita business for development: Market forces few multinational enterprises are broadly rises, agribusiness (including agricultural do not guarantee competitiveness, nor do diversified from seeds, feeds, and fertiliz- trade and distribution services) typically they guarantee smallholder participation, ers to product handling and processing of rises from under 20 percent of GDP to more both essential to link agricultural growth to sweeteners and biofuels. Food processing than 30 percent before declining as econo- development. For these reasons, promoting firms are integrating backward to primary mies become industrial (13 percent in the competitiveness and enhancing smallholder product handling and forward to retail dis- United States).2 participation are two priorities of the agri- tribution. Retailing has been transformed Agribusiness comprises diverse private culture-for-development agenda (chapter by the "supermarket revolution" (chapter agroenterprises, a majority of which are 10). The two complement each other as 5). National, regional, and global supply chains are being radically altered, bypass- ing traditional markets where smallholders Figure D.1 The relative shares of agriculture and agribusiness in GDP change as incomes rise sell to local markets and traders. Supermar- % share of GDP kets control 60 to 70 percent of food sales 60 in Argentina and Brazil, and are expanding Agriculture rapidly in China, India, and urban Africa. 50 Agribusiness Though these trends in agribusiness con- solidation have been going on for years in Thailand 40 industrial countries, they are now becoming Malaysia common in developing countries as well.5 Indonesia Chile In 2004 the market share for the four 30 Agribusiness largest agrochemical6 and seed companies Brazil Argentina Uganda Côte d'lvoire Kenya (the concentration ratio of the top four, or 20 Agriculture CR4) reached 60 percent for agrochemi- Nigeria cals7 and 33 percent for seeds, up from 47 10 percent and 23 percent in 1997, respec- tively.8 The CR4 in biotechnology patents 0 was 38 percent in 2004 (table D.1). In some 150 400 1100 3000 8100 subsectors, global concentration is much GDP per capita, 2000 US$ (log scale) higher--in 2004 one company had 91 per- Sources: WDR 2008 team. Data from Jaffee (1999) as cited in World Bank (2003f); and from Pryor and Holt (1999). cent of the worldwide transgenic soybean Note: Agribusiness includes the value added for agro-related industries and for agricultural trade and distribution area.9 It is generally believed that when an services. Data are for Argentina, Brazil, Cameroon, Chile, Côte d'Ivoire, Ghana, India, Indonesia, Kenya, Malaysia, Mexico, Nigeria, Republic of Korea, South Africa, Tanzania, Thailand, Uganda, and Zimbabwe. industry's CR4 exceeds 40 percent, market 136 WORLD DEVELOPMENT REPORT 2008 Table D.1 Major suppliers of agricultural inputs and growing concentration Agrochemicals Seeds Biotechnology 2004 sales Market share 2004 sales Market share Number Patent share Company ($ million) (%) ($ million) (%) of U.S. patentsa (%) Monsanto 3,180 10 3,118 12 605 14 Dupont/Pioneer 2,249 7 2,624 10 562 13 Syngenta 6,030 18 1,239 5 302 7 Bayer Crop Sciences 6,155 19 387 2 173 4 BASF 4,165 13 -- -- -- -- Dow Agrosciences 3,368 10 -- -- 130 3 Limagrain -- -- 1,239 5 -- -- Others/Private 7,519 23 16,593 66 1,425 34 Public Sector -- -- -- -- 1,037 24 Market concentrationb CR4 (2004) 60 33 38 CR4 (1997)c 47 23 Sources: UNCTAD 2006b; International Seed Federation at http://www.worldseed.org. a. Number of U.S. agricultural biotechnology patents issued during the 1982­2001 period. b. Market concentration is measured by the concentration ratio CR4, which indicates the market share of the four largest firms participating in the market. c. Fulton and Giannakas 2001. -- = not available. competitiveness begins to decline,10 leading tor is not easy, and there are tradeoffs. Con- to have access to private financial services, to higher spreads between what consumers trols and administrative requirements will risk-sharing institutions, and business pay and what producers receive for their increase transaction costs, commercial and development services. produce.11 political risks, and opportunities for rent Rules and regulations for intellectual The high concentration in multinational seeking. Interventions protecting weak mar- property rights, employment conditions, agribusiness is evident in coffee, tea, and ket players may do more harm than good-- contracting, and product standards also cocoa. Coffee is produced by an estimated 25 consumers, small farms, and small enter- affect the profitability of agroenterprises million farmers and farm workers, yet inter- prises may lose out if private enterprises and the distribution of benefits from agri- national traders have a CR4 of 40 percent, and vote with their feet or pass on increased business development. Barriers to entry coffee roasters have a CR4 of 45 percent. There transaction costs to them. But support to in establishing businesses are particularly are an estimated 500 million consumers. agroenterprise development can increase strong for small businesses, which suffer The share of the retail price retained by competitiveness by favoring entry of small more from poor access to finance and weak coffee-producing countries--Brazil, Colom- and medium enterprises (SMEs) and facili- business skills.15 bia, Indonesia, and Vietnam account for 64 tating the inclusion of smallholders. Addressing bottlenecks percent of global production--declined Instruments to address the bottlenecks to from a third in the early 1990s to 10 percent Small and medium small and medium agroenterprise devel- in 2002 while the value of retail sales dou- agroenterprise development opment include matching grants, chal- bled. Similar concentrations are observed in Two complementary approaches can be fol- lenge funds in public-private partnerships, the tea value chain where three companies lowed to support agroenterprise develop- preferential access to finance, partial loan control more than 80 percent of the world ment for competitiveness and participation. guarantees, tax breaks, and assistance in the market. Cocoa has a CR4 of 40 percent for One is to improve the investment climate to formation of agroindustrial networks. For international traders, 51 percent for cocoa induce the entry of private investors, partic- developing smaller agroenterprise in rural grinders, and 50 percent for confectionary ularly SMEs. Surveys of the rural investment areas, the focus has usually been on direct manufacturers. Developing countries' claim climate in Indonesia, Nicaragua, Sri Lanka, interventions rather than on improving the on value added declined from around 60 and Tanzania indicate that the lack of rural investment climate that could have wider percent in 1970­72 to around 28 percent in finance, infrastructure, and business and and more sustained impacts. The reason is 1998­2000.12 public services is particularly binding.14 The that ministries of agriculture have no man- Concentration widens the spread other approach targets bottlenecks in small date in generic issues of investment climate, between world and domestic prices in com- and medium agroenterprise development, whereas ministries mandated with economic modity markets for wheat, rice, and sugar, particularly in value chains. policies have limited interest in agricultural which more than doubled from 1974 to Improving the investment climate value chains in rural areas.16 1994. A major reason for the wider spreads The investment climate's four main compo- There is some debate over matching grants is the market power of international trading nents can all contribute. First is to ensure a to promote agribusiness because they have companies.13 sound macro policy environment. Second is been linked to market distortions and favorit- Balancing private investment to provide public goods such as infrastruc- ism, and they do not always promote growth- and competitiveness ture. Third is to have a legal and regulatory oriented SMEs. But they have had some suc- Designing and implementing policies to framework that fosters competition, busi- cesses in increasing the capacity of smallholders induce competition in the agribusiness sec- ness integrity, and fair practices. Fourth is to link to value chains (box D.1). Matching What are the links between agricultural production and food security? 137 even though they can boost market develop- B O X D . 1 Opening export markets to small-scale organic ment for the industry the firm is engaged in. cocoa producers in the Dominican Republic An example is the school milk feeding and dairy development programs sponsored by The Department for International Develop- Growers Association and European cocoa the TetraPak Food for Development Office, ment's (DFID) Business Linkages Challenge buyers. The project also created a better with the objective of improving the health Funds (BLCF) provides cost-sharing grants to quality product that opened new types and academic performance of children while promote business linkages, market develop- of markets for gourmet cocoa producers. creating demand for milk products and sup- ment, and pro-poor impact for smallholders. These investments paid a differential of porting smallholder participation.17 In the Dominican Republic, a 2002 BLCF $405 per ton to small-scale growers, gener- grant, matched by the private sector, funded ated a 25 percent increase in employment At the other end of the spectrum are a two-year organic chocolate production benefiting women, spread computer and programs delivering strong benefits to the project that improved the competitiveness Internet technology across communities, poor while allowing enterprises to break of smallholder organic cocoa producers by and increased the purchasing power of the even, with profits reinvested. An example obtaining higher and more stable prices for broader community--all reducing poverty. is the yogurt-producing venture of Groupe their product. It created stronger relation- Danone in cooperation with the Grameen ships up and down the value chain and forged new links between the Small Cocoa Group in Bangladesh. It extends loans Source: www.businesslinkageschallengefund.org. and technical assistance to smallholders to acquire dairy cows and invests in a process- grants are best used for business opportuni- holder participation. There is a growing ing plant for dairy products that meet local ties that can be profitable in the long run tendency among large enterprises to pursue nutritional needs and create employment in but face high startup costs. Oversight from business ventures that not only appeal to processing and distribution.18 independent peer review boards is essential corporate interests but also deliver a social In between are public-private partner- to ensure fairness and transparency. return, often benefiting the poorest of the ships that can be advantageous for inves- poor. These activities can take a variety of tors but may not initially yield a return. Corporate social responsibility forms depending on their direct economic The Mars Corporation is taking the lead Smallholder inclusion in agrofood value payoff, but are largely public-private or in Indonesia in coinvesting with the public chains can also occur through agribusiness civil society-private partnerships, where the sector and donors in a research and devel- initiatives that are motivated by more than driver is the private sector. opment program to promote cocoa quality just profits. Global agroenterprises can use At one end of the spectrum are programs and a sustainable supply while paying small- their resources and expertise to help develop delivering social benefits, but with no short- holders a premium for growing the higher agrofood value chains and promote small- run profit-making value for the enterprise, quality product.19 Supporting smallholder competitiveness through institutional innovations 6 Structural adjustment in the 1980s disman- · Innovations in finance can provide tled the elaborate system of public agencies smallholders with better access to credit, that provided farmers with access to land, savings facilities, money transfer mecha- c h a p t e r credit, insurance, inputs, and cooperative nisms, remote payments, and leasing. organizations. The expectation was that · Weather-indexed insurance can pro- removing the state would free the market vide new ways of reducing problems of for private actors to take over these func- imperfect information in mitigating tions--reducing their costs, improving farmers' risks. their quality, and eliminating their regres- · Institutional innovations can also pro- sive bias. Too often, that didn't happen. mote more efficient input markets, as In some places the state's withdrawal was new local agrodealers have emerged and tentative at best, limiting private entry. market-smart subsidies are tried. Elsewhere, the private sector emerged only slowly and partially--mainly serv- · Producer organizations can engage in ing commercial farmers but leaving many more effective collective action to access smallholders exposed to extensive market services, achieve economies of scale in failures, high transaction costs and risks, markets, and acquire voice in policy and service gaps. Incomplete markets and making. institutional gaps impose huge costs in for- gone growth and welfare losses for small- Land policies for secure rights holders, threatening their competitiveness and reallocating resources and, in many cases, their survival. Institutions governing land rights and The last 10 years have seen a broad effer- ownership affect the efficiency of land use. vescence in institutional innovations to fill If those who farm lack secure rights to the deficits in land markets, financial ser- land, they have less incentive to exert effort vices, input markets, and producer organi- to use it productively and sustainably or to zations. Although significant progress has carry out land-related investments. And if been made, this institutional reconstruc- women--who cultivate much of the land tion of agriculture is still incomplete, espe- in Africa--have few vested rights, house- cially for smallholders and more marginal holds tend to produce less than their asset areas. Moving forward requires more clar- base could otherwise provide. Secure and ity on the roles of the state and the private unambiguous property rights also allow sector--and more analysis of what works markets to transfer land to more productive and how it could be improved. This chapter uses and users. Cost-effective systems of documents how: land administration facilitate agricultural · New mechanisms can increase the secu- investment and lower the cost of credit by rity of property rights, facilitate land increasing the use of land as collateral, thus reallocation as rural households adjust reducing risk for financial institutions. their livelihood strategies or leave for the Institutions governing access to land city, and facilitate access to land for the have a long history of adapting to social, landless. natural, and economic factors. Their diver- 138 Supporting smallholder competitiveness through institutional innovations 139 sity reflects land's value not only as a factor land institutions.2 With greater knowledge of production but as a source of status, cul- of such laws, land-related investments and tural identity, and political power. Design- productivity increase, as evidence from ing property rights that support efficient Uganda suggests. With fewer than a third of land use and recognize the multiplicity of households informed about the law, further rights, particularly for women and indige- efforts to disseminate information could nous groups, is a highly complex issue that have a large impact.3 requires further exploration. Land policies Communal lands and common property were often adopted less to increase effi- resources, including grazing and indig- ciency than to further the interests of dom- enous lands, are a special case of custom- inant groups, making land issues politically ary tenure. In addition to their productive charged. This section addresses how recent value, they are often important as safety institutional and technological innovations nets for the poor because of the cultural can help deal with such legacies, increase values embodied in them. But they are vul- the security of tenure, and provide broad- nerable to degradation and appropriation based access to land to maximize its contri- by powerful chiefs, outsiders, and bureau- bution to agricultural competitiveness and crats. Increasing access to and the produc- economic development. tivity of such resources can be achieved by the following: Enhancing tenure security · Formalizing customary laws in ways that Providing land owners or users with secu- are participatory and reflect the diversity rity against eviction enhances their com- of the ethnic, historical, and social con- petitiveness by encouraging land-related struction of land.4 Delineating legally investment, as numerous studies show.1 valid boundaries, identifying existing Earlier interventions to improve tenure rights that may overlap or be of a seasonal security focused almost exclusively on indi- nature (between herders and sedentary vidual titling, but this can weaken or leave agriculturalists), and registering them as out communal, secondary, or women's appropriate. rights. Moreover, the process of titling can be used for land-grabbing by local elites and · Vesting day-to-day management deci- bureaucrats. So, although individual titling sions in an accountable body that func- is still appropriate in many cases, it needs tions transparently--say, as a legally to be complemented by new approaches to incorporated user group with clear rules securing tenure. for conflict resolution that are respected by all involved. Recognizing customary tenure. In many · Making evolution to more formal types countries, vast expanses of land held under of tenure possible through a well-defined customary tenure do not enjoy legal pro- and transparent process. In Mexico cer- tection, often because of legislation from tified individual land plots in ejido com- colonial times. For example, many Afri- munities can become fully transferable can jurisdictions considered most land to freehold land through a qualified vote be "state land." Those who had cultivated by the assembly. But the fact that fewer such land for generations received only pre- than 15 percent of ejidos chose full titling carious tenure rights and could lose their shows that many users see that the ben- land--say, to make room for "strategic" efits of maintaining communal relations investments--with little or no compensa- can be greater than those from individu- tion. Over the last decade, a large number alization of rights. of African countries adopted a wave of new land laws to recognize customary tenure, Documenting land rights. While legal make lesser (oral) forms of evidence on recognition of existing rights is an indis- land rights admissible, strengthen women's pensable first step, there is often demand land rights, and establish decentralized to demarcate plots and issue certificates 140 WORLD DEVELOPMENT REPORT 2008 to reduce boundary disputes and facili- key asset, land rights are critical for women's tate land transactions. High survey stan- bargaining power within the household, dards and the associated costs under tradi- their broader economic opportunities, and tional technology--between $20 and $60 their long-term security in cases of divorce per parcel5--have been a major obstacle or the death of a family member. Recog- to broader implementation. But recent nition of the adverse consequences of dis- advances in technology--particularly the crimination against women in this area has widespread availability of satellite imag- led to changes in constitutional provisions ery and handheld global positioning sys- and more specific legislation to require gen- tem (GPS) devices, together with institu- eral equality of men and women, mandate tional arrangements that put local actors issuance of joint titles, modify inheritance in charge of systematic adjudication--can legislation, and ensure female representa- greatly reduce the cost of issuing certificates tion on land administration institutions.8 for boundaries with reasonable accuracy. Such measures can have a positive impact. Experience points to considerable demand But legal reforms that clash with traditional for these land certification programs, as in power arrangements may be indifferently Ethiopia (box 6.1). enforced. Examples, many of them from Where women have a main role in culti- Asia or Latin America, show that to mini- vation, their land rights affect productivity mize clashes, a mix of mediation and raising and investment.6 In addition, with land as a awareness can complement other programs to allow landholders to effectively exercise their rights. For example, Mexico's ejido sys- tem now includes mediation to protect the B O X 6 . 1 Benefits from community-driven land property rights of women. In Nicaragua a certification in Ethiopia program to title land rights in the names of Thanks to the promising results from issu- reduce encroachment (76 percent) and both spouses included consultations with ing land-use certificates to about 632,000 increase soil conservation (66 percent). the indigenous population to clarify both households in Tigray in 1998/99, other A rough estimate puts the cost of communal and collective rights. Ethiopian regions have embarked on a certificates at only $1 a plot, in large part large-scale certification effort, issuing because local inputs to conflict resolution land-use certificates to about 6 million and surveying are voluntarily provided by Expanding options for conflict resolution. households (18 million plots) in 2003­05. local land-use committees. Adding hand- In many developing countries a large share The process starts with local aware- held GPS with accuracy to less than one of court cases involve land-related disputes. ness campaigns, sometimes with the dis- meter to record corner coordinates would Apart from clogging courts and stifling tribution of written material, followed by increase these costs by about 60 cents. elections of land-use committees in each With modern technology making low-cost investment, unresolved conflicts can depress village. After a period of training, these approaches more feasible, systematic cer- the productivity of land use. In Uganda pro- committees resolve existing conflicts, tification could help implement new land ductivity on plots under dispute is less than referring cases that cannot be settled legislation in Africa and beyond. Without amicably to the courts. This is followed by mechanisms to keep records up to date, a third that on undisputed plots.9 Tradi- demarcation and surveys of undisputed however, the effect may be short lived. tional institutions can resolve some local- plots in the presence of neighbors, with Estimates for the Amhara Region suggest ized disputes, but they are not well equipped subsequent issuance of land-use certifi- that updating should be possible at about to address disputes that cut across groups cates that, for married couples, include 65 cents per transaction. names and pictures of both spouses7 but Demand for certificates is strong: 95 belonging to different communities--for no sketch map or corner coordinates. percent of households outside the program example, between nomads and sedentary Because land remains state owned would like to acquire one, 99 percent of agriculturalists, across ethnic boundaries, with strong restrictions on transfers, cer- those with a certificate would be willing to or between individuals and the state. Tradi- tificates document only inheritable use pay an average of $1.40 to replace a lost cer- rights. Even so, more than 80 percent of tificate, and 90 percent (most of them will- tional institutions also tend to be under the respondents in a nationwide survey indi- ing to pay) would like to add a sketch map. control of men and favor men in disputes cated that certification reduced conflicts, Although the positive impact of certifi- with women, such as those over inheritance encouraged them to invest in trees and cates is likely reduced by current policies rights.10 Expanding the options to resolve soil conservation and to rent out land, that restrict land rental and prohibit sales and improved women's situations. They or mortgaging of land, certification can be land conflicts systematically and out of also felt that having a certificate would a step toward a broader process of land court can have large benefits, especially for increase the possibility of getting com- policy reform. the poor and for women who otherwise are pensation in cases of land taking. Many seldom able to enforce their legal rights, as expect demarcation of communal land to Source: Deininger and others 2007. demonstrated in Ethiopia and India.11 Supporting smallholder competitiveness through institutional innovations 141 Modernizing land administration. In many countries, land administration is one B O X 6 . 2 Improving the efficiency of land of the most corrupt public services. Irregu- administration services in Georgia larities and outright fraud are frequent in allocating and managing public lands. The Georgia established a single national land The time for property registration fell administration agency, made all informa- from 39 days to 9 days, and the associated rents can be large. In India, bribes paid tion publicly available on the Internet, cost decreased from 2.4 percent to 0.6 annually by users of land administration put licensed private surveyors in charge percent of property value, with attendant services are estimated at $700 million,12 of conducting surveys, and drastically cut benefits for land users--evidenced by staff (from 2,100 to 600) while increasing greater rental and sales market activity three-quarters of the public spending on salaries eightfold. To keep the registry and more mortgages and credit by private science, technology, and environment. In financially independent, the registry law and agricultural lenders. Kenya, land grabbing by public officials, was revised, a free legal consultation systemic during 1980­2005, was "one of established, and the fee structure adjusted. Source: Dabrundashvili 2006. the most pronounced manifestations of corruption and moral decadence in our society."13 Modern technology and part- If tenure is insecure or restrictions con- strain land leasing, productivity-enhancing nerships with the private sector can yield rental transactions will not fully material- quick benefits. One example: computeriz- ize or the poor may be excluded. In the ing records in the Indian state of Karnataka Dominican Republic, Nicaragua, and Viet- is estimated to have saved users $16 million in bribes.14 Automating registration and nam, insecure land ownership reduced the propensity to rent and limited transactions the associated land valuation allowed out- to preexisting social networks.18 In Ethio- sourcing to the private sector, which sig- pia, fear of losing the land, together with nificantly improved access to the service explicit rental restrictions, was the main and cut stamp duties from 14 percent to reason for suboptimal performance of rental 8 percent, while quadrupling tax revenue from $120 to $480 million.15 markets.19 In India, tenancy restrictions reduce productivity and equity (box 6.4). Land administration institutions will Replacing them with policies that facilitate be viable in the long term and independent renting would improve access to land by from political pressure only if they can sus- those remaining in the rural sector. tain their operations financially, without charging more than users are willing to pay. Strengthening land sales' markets. Sales Although the reforms to make them more markets for acquiring land increase invest- efficient are well known, with their effec- ment incentives and provide a basis for tiveness repeatedly shown (box 6.2), imple- using land as collateral in credit markets. mentation faces strong resistance from However, imperfections in other mar- interests benefiting from the status quo. kets, and expectations of future land price Access to land increases, affect markets for land sales more than those for rentals, implying that sales Enabling land rental markets. Getting land would not necessarily transfer land to the markets to work is fundamental where new most productive producers. Historically, options emerge for households to diversify most land sales happened under distress, livelihoods and eventually leave agriculture. requiring defaulting landowners to cede In developed countries, about 50 percent of their land to moneylenders, who could farmland is rented, often under sophisticated amass huge amounts of it.20 contracts. In most developing countries, by Data on land sales over 20 years in India contrast, land rental markets are atrophied. reveal some peculiar features of land sales However, land rentals are increasing where markets: they had not been practiced extensively ear- lier--as in Eastern Europe;16 in Vietnam, · Land went to better cultivators and from where rental participation quadrupled to 16 land-abundant to land-scarce households, percent in five years;17 and in China, where allowing the land-scarce to improve their rentals allow rural communities to respond welfare without making sellers worse to large-scale out-migration (box 6.3). off. But sales markets are thinner, more 142 WORLD DEVELOPMENT REPORT 2008 B O X 6 . 3 How land rentals can increase productivity and equity in China Land rental markets can contribute much to sample with information on the two parties in the gains, with the rest going to landlords rural diversification and income growth in a land transactions highlights the impact of land in rents. rapidly growing economy. Look at China. After rentals on occupational structures, land pro- · Net income for both renters and landlords the introduction of the household responsibil- ductivity, and welfare: increased--respectively by 25 percent ity system in 1978, land-use rights were allo- · Land rentals transformed the occupational and by 45 percent (partly due to migration cated on a per capita basis, leading to an egali- structure. While almost 60 percent of those income)--in a very equitable way. tarian land "ownership" structure, with land renting out their land relied on agriculture as also functioning as a social safety net. Although This shows the importance of well- their main source of income before entering households held 15-year land-use contracts, functioning land rental markets in a context rental markets, only 17 percent continued to administrative reallocation--in clear breach of strong nonagricultural growth and migra- do so--while 55 percent migrated (up from of contractual obligations--was regularly tion. But many producers still feel constrained 20 percent) and 29 percent engaged in local practiced in response to population growth or by insecure property rights. To allow land nonfarm activity (up from 23 percent). to make land available for nonagricultural pur- markets to better respond to the needs of a poses. But with rural-urban migrants tripling · Land rentals also increased productivity. Net changing economy, recent initiatives, espe- from 5 percent of the total labor force in 1988 revenue on rented plots rose by about 60 cially the 2003 Rural Land Contracting Law, aim to 17 percent (or 125 million migrants) in 2000, percent, supporting the notion that rental at strengthening farmers' property rights and the limits of exclusive reliance on administra- markets, by transferring land to better farm- reducing the scope for discretionary interven- tive allocations became obvious. ers from those with low ability or little inter- tion by officials. Decentralized land rentals, which comple- est in agriculture, can improve rural welfare. mented and eventually replaced administrative Renters--who generally had less land, more Sources: Benjamin and Brandt 2002; Brandt, reallocations, have proven just as equitable family labor, and lower levels of assets and Rozelle, and Turner 2004; Cai 2003; Deininger and but significantly more productive. A national education--received about two-thirds of Jin 2005; Kung and Liu 1997. affected by life-cycle events, and less redis- ership, land markets are no panacea for tributive than those for rentals. addressing structural inequalities that · Climate shocks increased the probabil- reduce land productivity and hold back ity of distress land sales, although miti- development.24 To overcome such inequali- gated by local safety nets (employment ties, ways of redistributing assets, such as guarantees) and access to credit from land reform, are needed. Postwar Japan, banks.21 the Republic of Korea, and Taiwan (China) show that land reform can improve equity · Although land ownership ceilings im- and economic performance. But there are posed by "reform" may have played a many cases where land reform could not role, land sales and purchases did more be fully implemented or even had nega- than land reform to equalize India's land ownership.22 tive consequences. Evictions of tenants or changes of land use ahead of legislation that This implies little justification for policy would have given greater security to tenants measures to restrict land sales, especially or allowed expropriation of underused land because they tend to drive transactions often made prospective beneficiaries worse underground and undermine access to for- off or prompted land owners to resort to mal credit without addressing the underlying even less-efficient techniques.25 problems of asymmetries in power, informa- If land is transferred through redis- tion, and access to insurance. Safety nets and tributive land reform, improvements in other measures, including redistributing access to managerial skills, technology, land, are more appropriate than constraints credit, and markets are essential for the on sales to deal with these problems and new owners to become competitive. Some prevent distress sales. Land taxes can curb tenancy reforms have proved highly effec- speculative demand and encourage better tive,26 but measures to clarify ownership land use, while providing revenue for local rights are needed to avoid disincentives for governments to fulfill their functions.23 investments. Land reform through market exchange assisted by grants and technical Making land reform more effective. In assistance to selected beneficiaries shows countries with highly unequal land own- promise, with Brazil the leading innovator, Supporting smallholder competitiveness through institutional innovations 143 but this approach deserves further analysis of costs and impacts. To be effective, any B O X 6 . 4 Rental markets and the impact of restrictions approach to land reform must be integrated in India into a broader rural development strategy-- using transparent rules, offering clear and Where tenants had few alternatives, land- Higher village incomes increase the lords used land rentals to extract as much propensity to rent, because wealthier unconditional property rights, and improv- as possible. This led Indian policy makers households are more likely to move out ing incentives to maximize productivity to impose rent ceilings to protect tenants of agriculture and rent out their land. gains. Yes, it can enhance access to land and to prohibit tenancy in many states. The equity impact of rental restric- Partly as a result, reported land rental for the rural poor. But to reduce poverty tions is shown by comparing the marginal activity in India declined sharply, from 26 product of one day of labor in agricultural and increase efficiency, reform requires a percent in 1971 to less than 12 percent in self-cultivation (Rs 150 for males and commitment by government to go beyond 2001, contrary to trends in other countries. females) with daily wages in the casual providing access to ensuring the competi- Still, renting continues to be an important labor market (Rs 46 for males and Rs 34 means of accessing land. More households for females). The (statistically significant) tiveness and sustainability of beneficiaries rented land in 2001 than the total number difference implies that, even after sub- as market-oriented smallholders. that have benefited from land reforms tracting payments to the landlord, renting since independence. can improve household welfare consid- Financial services The assumptions underlying inter- erably. Gender discrimination in casual ventions in land rental markets may no labor markets would make renting par- for smallholders longer hold, as a national survey that ticularly attractive for women, consistent The ability of agricultural enterprises and allows comparisons over time suggests. with anecdotal evidence of rural women's Instead of causing reverse tenancy, rental use of self-help groups to rent land, often rural households to invest for the long markets help land-scarce and labor- against the law. And eliminating land term and make calculated decisions for abundant households with agricultural rental restrictions would facilitate moves risky and time-patterned income flows is skills but little education--37 percent of into the rural nonfarm economy. shaped by an economy's financial services. them landless--to rent land from land- abundant and wealthy households that Despite the rapid development of financial take up nonagricultural employment. Source: Deininger, Jin, and Nagarajan 2006. services, a majority of smallholders world- wide remain without access to the services they need to compete and improve their livelihoods. Broader access to financial Lifting the pervasive financial services--savings and credit products, constraints that perpetuate poverty financial transactions, and transfer ser- Financial constraints are more pervasive vices for remittances--would expand their in agriculture and related activities than opportunities for more efficient technol- in many other sectors, reflecting both the ogy adoption and resource allocation. nature of agricultural activity and the aver- Financial services are delivered to rural age size of firms. Financial contracts in rural populations by organizations that exist areas involve higher transaction costs and along a continuum from informal to for- risks than those in urban settings because mal, with the boundaries between catego- of the greater spatial dispersion of produc- ries often blurred. In general, formal finan- tion, lower population densities, the gener- cial institutions are licensed and supervised ally lower quality of infrastructure, and the by a central authority. They include public seasonality and often high covariance of and private commercial banks; state- rural production activities. So banks and owned agricultural or rural development other traditional for-profit financial inter- banks; savings and loan cooperatives; mediaries tend to limit their activities to microfinance banks; and special-purpose urban areas and to more densely populated, leasing, housing, and consumer finance more affluent, more commercial areas of companies. Informal providers of finan- the rural economy. Operating costs there cial services include rotating savings and are lower, loan sizes large enough to cover credit associations, money lenders, pawn- fixed transaction costs, and legal contracts shops, businesses that provide financing to more easily enforced. their customers, and friends and relatives. The rural reality: few households and In between stand financial nongovernmen- small firms can meet their need for credit tal organizations (NGOs), self-help groups, and other financial services. In India a recent small financial cooperatives, and credit survey of 6,000 households in two states unions. showed that 87 percent of the marginal 144 WORLD DEVELOPMENT REPORT 2008 farmers surveyed had no access to formal The root of the problem is that lenders credit, and 71 percent had no access to a sav- tend to offer only a limited menu of prod- ings account in a formal financial institu- ucts, mainly with heavy collateral require- tion.27 Informal financial arrangements ments. Wealthier farmers can obtain larger serve rural communities, but they tend to loans at lower cost from formal lenders fragment along lines of household location, because they can credibly pledge assets or asset ownership, or membership in kin- or future cash flows. Asset-poor households, by ethnic-based networks, all affecting the contrast, are limited to considerably smaller transaction costs of contracting, the size of loans at much higher rates because they have the possible transactions, and the rate of to turn to lenders who must substitute costly interest charged.28 There is thus a tremen- monitoring for collateral. Poor farmers may dous need for financial innovations that can also turn down loans, even if they qualify, place smallholders on a ladder of ascending because they are unwilling to bear the risk financial market access--as well as for inno- of losing collateral, termed "risk ration- vations that can complement financial ser- ing."31 In the studies of Honduras, Nica- vices by managing the systemic risks that ragua, and Peru, 20, 40, and 50 percent of undercut their supply. credit-constrained borrowers, respectively, The costs of financial constraints for are risk-rationed. Access to credit and insur- smallholders are huge--in forgone oppor- ance are thus closely tied conceptually and tunities and in their exposure to risk. In empirically and must be jointly improved to rural Honduras, Nicaragua, and Peru, the enhance access to credit. credit-constrained population constitutes The skewed access to credit can blunt some 40 percent of all agricultural produc- employment and contribute to worsening ers. Producers lacking credit use on average the income distribution. Land market poli- only 50 percent to 75 percent of the pur- cies also become less effective if there are chased inputs of unconstrained producers wealth-biased financial market constraints. and earn net incomes (returns on land and family labor) between 60 percent and 90 Adapting microfinance percent of the unconstrained (figure 6.1).29 to reach smallholders In Central and Eastern Europe, nearly 50 The inadequacies of rural financial markets percent of smallholders in five countries reflect real risks and real transaction costs report financial constraints to be the major that cannot simply be wished, or legislated, barrier to the growth and expansion of away. Innovations are required to permit their enterprises.30 more flexible forms of lending while guar- anteeing that borrowers repay loans. One approach to resolve these prob- Figure 6.1 Credit-constrained rural households lems follows from the pioneering efforts use fewer inputs and have lower incomes of the Grameen Bank. Microfinance insti- Ratio of constrained households to unconstrained, % tutions (MFIs) open the menu of avail- 100 able contracts with new arrangements that substitute for collateral. They often 80 have guidelines to favor groups--particu- larly women--excluded from borrowing 60 through other channels. Many MFIs lend 40 to local groups whose members select one another and share the liability for repaying 20 loans, so local social capital substitutes for wealth as collateral. MFIs often target rural 0 Nicaragua Peru Honduras areas, where social capital is stronger. Input per hectare This shared liability creates powerful Net income per hectare incentives for rigorous peer selection and Total productive wealth borrower monitoring, and it can work well Source : Boucher, Carter, and Guirkinger 2006. when loans are used for a diversity of (quick Supporting smallholder competitiveness through institutional innovations 145 turnaround) activities. However, it works survive without resorting to nontranspar- less well for crop activities, where all produc- ent fees.34 Other regulations make it nearly ers are subject to a common set of weather impossible for MFIs to mobilize savings risks (when one cannot pay, often no one can and accept deposits. Recognizing this, pay) and where project gestation periods are India recently passed a new microfinance long and share the same timing. Weather law reducing the amount of start-up capital risk also undermines the financial stability an MFI was required to have before it could of local MFIs, and most explicitly limit their take deposits. Such reforms need to bal- share of lending to agriculture to reduce ance protecting small-saver deposits with exposure to risk. Moreover, many microfi- expanding the menu of opportunities they nance organizations have targeting criteria face. One possibility is a well-structured for maximum landholding that restrict their insurance scheme for deposits. lending to agricultural activities. To meet the underserved agricultural Reshaping financial services market, MFIs have begun to innovate.32 for smallholders and the rural FUNDEA in Guatemala has offered indi- nonfarm economy vidual loans to agricultural producers of MFIs cannot, however, provide the main- short-cycle tomatoes and other vegeta- stay of rural finance. Promoting, improv- ble crops. It has adopted the value chain ing, or even creating rural institutions to approach to financing inputs and outputs, support a wide range of rural financial using standing crops as collateral. Caja los transactions remains one of the funda- Andes in Bolivia began to accept nonstan- mental challenges facing developing-coun- dard collateral assets and lend to farmers try governments. The range of alternatives well diversified across a range of agricul- is broad. Government-sponsored agricul- tural and nonagricultural activities.33 In tural lending institutions have been suc- 2006 it became a bank, Banco Los Andes cessful in many now-developed economies Procredit, and agricultural loans now con- such as the Republic of Korea and Taiwan stitute 10 percent of its portfolio. (China). But in many developing countries, In short, while microfinance lending in government efforts to improve rural finan- agriculture is still small, there are hope- cial markets have a record of doing more ful signs that innovation will permit the harm than good, heavily distorting market microfinance movement to partially fill the prices; repressing and crowding out private agricultural void, at least for producers with financial activities; and creating central- small enterprises engaged in high-value ized, inefficient, and frequently overstaffed activities, particularly animal husbandry bureaucracies captured by politics.35 There- and horticulture. There is a strong case fore it is not surprising that public agricul- for public policy support to search for, and tural and development banks came under pilot test, technological and institutional heavy criticism in the 1980s.36 Bolivia and innovations that reduce the costs and risks Peru simply closed their traditional agricul- of doing business. Many of the newly devel- tural banks, while The Gambia and many oped innovations may have the character of of the former Soviet republics sold off and a public good, because innovations by one privatized all or part of their state banking lender may be quickly adopted by another. programs.37 This justifies public support for promising start-ups to enable them to reach scale and Reforming public agricultural banks. become financially viable within predeter- Unless state-owned agricultural banks mined time periods. undergo a radical transformation in gover- Reformed financial regulations, coupled nance arrangements that can insulate them with better financial infrastructure, could from political capture, they are unlikely also boost access to financial services in to function in a commercially sustainable many countries. Forty developing and manner and serve the needs of smallhold- transition economies still have interest rate ers. What's needed is some form of privati- ceilings that make it difficult for MFIs to zation. Banrural in Guatemala shows how 146 WORLD DEVELOPMENT REPORT 2008 interference to transform financial institu- B O X 6 . 5 Banrural SA: from ill-performing tions. But such a transformation is hardly agrarian bank to profitable automatic or ensured, because state banks public-private financial institution remain vulnerable to political capture. Key elements of reform include those advocated Banrural SA in Guatemala shows that Bandesa employees (1). Each group elects to improve governance and accountability financial and development goals can its own directors and can sell stock only to of many state functions: transparency and be combined and that a large bank can other members of the group. This unusual remain highly profitable while offering governance model has empowered the professionalization. Financial objectives financial services to poor, rural, and agri- private stakeholders and balanced goals must be promoted by clear incentives for cultural clients. Banrural was created in of profitability and rural development. It is management and staff that tie rewards to 1997, when Guatemala closed Bandesa, sustainable because the board and equity the financial performance of branches. its poorly performing public agricultural makeup cannot be altered significantly bank. With 200,000 credit clients, Banrural over time. has a default rate of less than 1.5 per- A focus on rural areas and poor cli- Providing financial services through self- cent. With 1 million savings accounts, it ents. Banrural's profits come from a high help groups and financial cooperatives. In facilitates the transfer of more than $1.3 volume of small transactions, mostly in several Indian states, a separate movement billion in remittances. It works mainly rural areas. Having learned the lessons outside of Guatemala City. Half its clients of the microfinance revolution, it adapts has emerged, based on village-level women are women, and it provides biometric and financial technologies to its clientele-- self-help groups and their federations at the multilingual devices to serve illiterate and loan officers visit all clients, decisions are village, mandal, and district levels. These indigenous clients. based on an evaluation of business and estimated 2.2 million groups collect sav- An innovative governance model. household income flows, and use of tradi- Banrural is controlled by private share- tional collateral is limited--without losing ings from their members and either deposit holders. The public sector owns less than its identity as a bank. Its lending portfolio them in rural banks or lend them to mem- 30 percent of the equity and provides to agriculture has more than doubled since bers. After demonstrating their capacity no direct subsidies. The remaining 70 it was privatized. To increase its reach to percent is divided among five types of to collect on loans over a six-month time smallholders and rural microenterprises, stock, each represented on the board of Banrural functions as a second-tier bank, period, rural banks will typically leverage a directors. The 10 board seats are divided providing credit lines to more than 150 group's savings by a factor of four, provid- among the public sector (3), unions institutions, such as credit unions and ing additional capital that is mostly used for (mostly agricultural producer unions, not financial NGOs. To build strong community credit unions) (2), Mayan organizations bonds, it provides health care and scholar- agricultural purposes. It is often easier for (2), NGOs (1), small and micro enterprises ships and supports community activities. self-help groups to obtain loans than it is (including microfinance organizations) for larger farmers, many of them poor cus- (1), and the general public and former Source: Trivelli 2007. tomers for rural banks. With the self-help groups responsible for all screening, pro- cessing, and collection activities, the trans- firm budget constraints and appropriate action costs for loans are greatly reduced. governance mechanisms can create a pub- Financial cooperatives and their net- lic-private institution that meets the needs works are reemerging as promising institu- of rural and agricultural finance (box 6.5). tions in rural finance in many countries, Other reforms of state-sponsored lenders combining the advantages of proximity have produced some of the most success- with modern management tools.38 Locally ful agricultural-oriented finance programs, based, their transaction costs are typically including Bank Rakyat Indonesia and lower than those of other financial institu- BAAC Thailand. tions. But because they are members of a Building on existing (but perhaps failed) larger network, they can offer the variety public banks offers the opportunity of and volume of financial services that rural using their branch networks to establish a customers require, and they can pool risks presence and take advantage of scale and as well as costs. In Burkina Faso, RCPB, the spatial dispersion to reduce costs. The suc- largest network of financial cooperatives, is cessful restructuring and later privatization establishing rural service points and very of the former agricultural bank of Mongolia small village-based credit unions, managed (renamed KhanBank in 2006) and of NMB and supervised by financial cooperatives in in Tanzania demonstrate the potential of an larger villages.39 existing branch bank infrastructure, inno- vative and independent management and Expanding the reach of rural finance. In- oversight, and strong barriers to political formation technologies offer a broad array Supporting smallholder competitiveness through institutional innovations 147 of new ways to extend financial services to inadequate legal frameworks, which often rural areas, for value chains and for agri- prevent the collateralization of less con- culture more broadly. The use of mobile ventional assets (such as an input supplier's phones for banking is being pioneered by contract for a standing crop).44 Further Wizzit in South Africa and by Globe Tele- undercutting collateralized lending are com and Smart in the Philippines. The legal systems that fail to provide clear rules phones can be used to pay for purchases in for priority claims on assets and prompt stores and to transfer funds, significantly redress in the event of default. Without reducing transaction costs. With legal collateral, high risks cannot always be com- frameworks in place, m-banking could be pensated by higher interest rate premiums, one of the major breakthroughs in extend- so many smallholders are simply rationed ing outreach to poor customers.40 Branch- out of the credit market. less banking--using post offices, stores, gas stations, and input providers--is another Reputational collateral through microcredit successful approach to reaching rural cus- reporting bureaus. Microcredit reporting tomers at low cost. Brazil, India, Kenya, the bureaus that establish individual reputa- Philippines, and South Africa demonstrate tions can help small farmers use their past its financial viability, although there are credit histories as an asset. A smallholder issues in regulating such endeavors.41 begins by establishing a credit history in the Rural leasing is another financing option MFI sector, often using credit for nonagri- for rural entrepreneurs, in agriculture and cultural purposes. In some instances, sav- in the rural nonfarm economy. Commercial ings records are also accepted as proof of providers in Mexico, Pakistan, and Uganda good financial behavior. The credit bureau show that leasing can finance the acquisi- establishes a reliable, portable signal of the tion of productive assets.42 Now running borrower's reputation. Armed with this sig- profitably, these commercial providers all nal, a borrower should then be able to climb benefited from access to government and a lending ladder, moving from the more donor funds to jump-start their opera- restricted purposes and term structures of tions, demonstrating the potential benefits MFI credit to standard loan contracts from of public-private partnerships. institutions able to bear the portfolio risk and term structures required for agricul- Financing through interlinked agents. Yet tural loans. another way to increase agricultural For a lending ladder to work, two things access to capital is financial intermedia- must happen. First, a credit report must tion through linked agents in value chains help lenders select clients and induce clients (input suppliers or output processors) to repay loans. This becomes all the more (chapter 5). Those agents are often more essential as competition among lenders able to cost-effectively monitor on-farm rises. Second, information on a borrower's behavior (eliminating information asym- credit worthiness and reputation must flow metries), thus reducing monitoring costs up the rungs from MFI to commercial lend- and enabling financial institutions to accept ers. A study of a credit bureau that includes nonstandard forms of wealth as collateral, MFIs in Guatemala shows that both can such as standing crops or, for warehouse happen.45 However, a client's credit history receipt financing, harvested crops.43 addresses risks related to the borrower's Further work is needed to determine financial behavior--but it does not, and whether these (often spatially monopolis- cannot, address business risks related to tic) practices offer finance at competitive weather and prices in agriculture. rates and whether transaction costs con- tinue to bias them against smallholders. Insurance to manage risk As mentioned, some MFIs and coopera- Risk distorts investments and puts assets tives have themselves begun to adopt this in jeopardy. Insurance can assist farmers in form of secured lending. But their success taking more risks in production and prevent has in many instances been undercut by shocks from depleting their assets. It can 148 WORLD DEVELOPMENT REPORT 2008 also reduce interest rates needed to offset Managing risk through the risk of default and increase the availabil- microfinance ity of agricultural credit by making traders As discussed, the absence of insurance limits and other intermediaries more willing to access to credit. Conversely, accessible credit put their assets into an agricultural loan can help a household smooth consumption portfolio.46 And in addition to enhancing and avoid distress sales. But shouldn't house- the supply of agricultural credit, insurance holds save in anticipation of future needs can make potential borrowers more willing and use their savings to self-insure? House- to bear the risk of conventional collateral- holds do, of course, save grain and cash, but ized loans. As always, there is a tradeoff. less than might be expected. Just as there are Insurance is costly and leads to higher over- credit constraints, households have limits all costs when added on to a loan. to saving because of low (or even negative) Individuals and local networks can do real interest rates, security concerns, and the much to manage risk, but such strategies inaccessibility of banks. In addition, fam- often founder on systemic risk, beyond the ily obligations and gender roles hinder the capacity of the individual and community accumulation of cash. On the supply side, to manage. Innovations to address systemic many banks find that transaction and regu- risk can complement the local capacity to latory costs make small deposits unprofit- manage idiosyncratic risks. By so doing, able. MFIs partially address this. In addition the expectation is that the innovations will to their well-known extension of credit to underwrite a more productive and sustain- households with limited collateral, many able pattern of agricultural and human MFIs offer secure and convenient ways of capital investment. saving small amounts, often requiring a sav- ings history before granting a first loan. Individual and community MFIs can serve an additional role in risk responses to risk management: they can reduce the market- One element of any strategy to address the ing and monitoring costs of insurance by cost of risk is to expand a household's risk being intermediaries for insurance to their management opportunities. Communities clients. MFIs often require insurance on have developed informal systems of mutual the assets purchased when a loan is taken insurance and contingent loans to respond out--for example, to insure against the loss to shocks based on traditional norms47 and of a cow. They may also require clients to local information. For example, pastoralists insure against external factors that inter- in Kenya provide cattle to neighbors who fere with the ability to repay on schedule or have lost a portion of their herds to repay offer loan-protection insurance to ensure past assistance and to create future obliga- that debts are not passed on to survivors. tions.48 But these systems tend to fail poor MFIs can serve as intermediaries for families, for several reasons. One is the other types of insurance covering individ- inherent limitation of insuring for covari- ual risks, taking advantage of their ability to ate shocks: one's neighbors cannot provide collect small amounts regularly and in keep- assistance if they are also under stress. ing with the transformation of some MFIs Another is that such systems entail transac- from lending institutions to providers of a tion costs of searching for partners, coordi- broader range of financial services, includ- nating activities, and monitoring reciprocal ing savings accounts. The marginal costs for arrangements. As these costs increase, the collecting payments are reduced when staff optimal size of a mutual-support network is networks are already in place, opening the reduced, also reducing risk sharing. More- possibility of providing death and disability over, individuals tend to form networks insurance as well as health and crop insur- with others of their own caste, ethnicity, ance. Indeed, the lives of more than 1.6 mil- and gender, as well as a similar asset base. lion Africans were insured in 2004 through Mutual insurance, though useful, tends to a profit-making microinsurance product be weakest for the poorest and to fall short marketed though 26 NGO-managed MFIs, when it is most needed. 24 of them in Uganda.49 Supporting smallholder competitiveness through institutional innovations 149 Meeting the promise of weather- indexed insurance B O X 6 . 6 Mongolia's index-based livestock insurance MFIs cannot necessarily address moral haz- Since 2005, Mongolia has piloted index- windy summers and cold, high-snowfall ard or adverse selection, two major obstacles based livestock insurance to share risks winters, the insurance index is linked not to providing insurance. One innovation among herders, insurance companies, and to a weather event, but to historical live- the government. The project combines stock mortality data. Insurance payments that might do so is insurance indexed to an self-insurance, market-based insurance, are thus not directly linked to individual objective indicator of weather, such as rain- and social insurance. Herders retain small herders' livestock losses; payments are fall or temperature. Because weather is not losses that do not affect the viability of instead based on local mortality. This affected by individual behavior, indexed their business (self-insurance), while larger should avoid or reduce moral hazard and losses are transferred to the private insur- adverse selection--and reduce costs. insurance can address both monitoring ance industry (market insurance through A key to the approach is having good costs and moral hazard. The choice of indi- a base insurance product). This is not a data to develop the livestock mortality cator depends on both the type of cover- purely commercial program, however. The index. Mongolia has a 33-year time series government bears the final layer of cata- age and the cost and availability of data on adult animal mortality for all regions strophic losses (social insurance through a and for the four major species of animals for estimating the probability of a payout. disaster-response product). (cattle and yak, horse, sheep, and goat). Cumulative rainfall or the date of the start Herders pay a market premium rate The mortality index provides the basis for of a rainy season is often proposed as the for the base insurance product, which determining the specific mortality rates pays out to individual herders whenever that would trigger indemnity payments. indicator; the number of days with tem- the livestock mortality rate in a local peratures below or above a cutoff is also in region exceeds a threshold. As excess common use. mortality reflects a combination of dry, Source: World Bank 2005l. One concern is basis risk--the corre- spondence of the indicator and the actual losses incurred by a policyholder. The companies, and the government (box 6.6). more specific the indicator, the lower the In Malawi, weather-indexed insurance cov- basis risk and more responsive it will be ers the loans necessary to finance improved to farmers' needs. But a diverse range of seeds and fertilizer, with insurance payouts products--including separate rainfall con- going directly to banks to settle the farmers' tracts for planting, growing, and harvesting loans. In India, an MFI, BASIX, intermedi- stages--would make their marketing more ates between insurance companies and its difficult because individuals often find it clients. The entry of private investors and hard to assess the probabilities of an event. the number of repeat customers for unsub- Furthermore, addressing individual shocks sidized weather insurance indicates the increases monitoring costs. So, index-based potential for a private market. insurance may have its greatest potential in addressing broad covariate shocks. Defining government's role Several approaches are being tried to in agricultural insurance adapt indexed insurance to diverse condi- The track record of agricultural insur- tions. Because they are still in pilot stages, ance directly supplied by governments is no definitive statement about their sustain- not encouraging. In Brazil, costs exceeded ability or their impact on credit rationing, premiums by more than 300 percent.50 input use, and portfolio choice is avail- However, governments may have a role in able. Mexico determines the timing of inducing insurance services. In Tanzania, assistance to small farmers after weather- what farmers were willing to pay for insur- related shocks on the basis of a weather ance was less than the actuarial fair cost index. The payment amount is based on of providing coverage, particularly among proxies for chronic poverty. In 2006, 28 low-income farmers.51 Indeed, the tendency percent of the nonirrigated cultivated area for wealthier households to purchase more was covered through an insurance contract insurance is a general pattern, with impli- with the federal and state governments, cations for income distribution.52 Targeted with the availability of weather stations the subsidies might thus be warranted for vari- main limitation. Mongolia, by contrast, able costs to induce learning, especially when promotes private livestock insurance, with insurance premiums are less costly than ex the government addressing reinsurance to post assistance. Subsidies can also offset the share risks among herders, the insurance fixed costs of establishing a market. 150 WORLD DEVELOPMENT REPORT 2008 Governments can also improve ex post Similarly, demand for fertilizer used on risk mitigation by improving the data noncommercial crops is generally weak and necessary for privately provided market unstable, for many of the same reasons: lack insurance. For example, insurers may be of knowledge, information asymmetries, unable to estimate the costs of rare events: liquidity constraints, risk and uncertainty, a 1-in-100 event is hard to distinguish from and high opportunity costs.53 Profitability a 1-in-80 event. Similarly, risks are hard tends to weigh heavily in farmers' deci- to quantify in a changing climatic or eco- sions, because the cost of fertilizer often nomic environment. Thus, insurers may represents a large share of cash production require higher premiums to accommodate costs.54 When cost factors and risk factors such ambiguity of risk. When governments act in tandem, as they do in most rain- assemble information that can be employed fed environments, the impact on fertilizer in index-based insurance, they provide a demand can be significant.55 public good that can improve the efficiency How do the distinctive features of of markets and reduce costs. demand for seed and fertilizer affect supply? The incentives for private firms to invest in Developing efficient producing and distributing seed depend on the potential profitability of these activi- input markets ties. In industrial countries, where eco- Agricultural productivity has grown rapidly nomic incentives (and the expanding use of where modern varieties and fertilizers have intellectual property rights) make it more been widely adopted, but not where adop- likely that farmers will regularly purchase tion has lagged (chapter 2). In much of Asia seed, plant breeding is done mainly by seed and parts of Latin America, promoting seed companies. But in smallholder agriculture and fertilizer use was accompanied by com- in developing countries, seed companies plementary investments in irrigation, rural depend on public research programs to pro- roads, marketing infrastructure, financial vide varieties. This makes the pipeline for services, and other factors that made using new products uncertain. Private seed com- seed and fertilizer profitable and paved the panies usually have incentives to serve the way for dynamic commercial input mar- needs of business-oriented farmers when kets. But throughout most of Africa, these the predominant seed technology is hybrid, complementary investments are small or when onfarm seed production is difficult, nonexistent, and private input markets or when output markets demand a uni- have yet to emerge on a large scale. Recent form product that depends on genetically initiatives to build seed and fertilizer mar- uniform, high-quality seed.56 When these kets provide lessons that can inform future conditions are absent, as is often the case in policy design. smallholder farming systems, the incentives for private seed companies are low. Special challenges in seed and For fertilizer, seasonally variable and fertilizer markets geographically dispersed demand discour- Why are efficient markets for seed and fer- ages potential suppliers because markets tilizer so difficult to develop? To begin with, are small, making low-cost procurement demand for both inputs is highly variable in difficult. Producing, importing, and trans- time and space. In developing countries, the porting fertilizer entail major economies of demand for seed is strongest when farmers scale.57 Importing fertilizer, for example, is are growing hybrids, whose seed must be most cost effective in lots of 25,000 tons, replaced regularly. It is weakest when farm- considerably above the annual demand ers are growing varieties whose seed can be in most Sub-Saharan African countries. saved from the harvest and replanted for Transport costs are particularly high in several cropping seasons. In addition, the Africa because of the generally poor road quality of seed found in the market may be and rail infrastructure. Because of domes- unknown as quality cannot be determined tic transport costs, fertilizer use is higher through visual inspection. in coastal African countries than in land- Supporting smallholder competitiveness through institutional innovations 151 Figure 6.2 Transport costs make up about one- many smallholders with no reliable access third of the farmgate price of urea fertilizer in to seed. African countries, 2005 Initiatives to promote fertilizer use have Procurement costs, US$/ton usually encouraged cost-effective import- 400 ing. Many Sub-Saharan countries do not have access to the raw materials to manu- 300 facture fertilizer, and few have a domestic market big enough to support an efficient manufacturing facility. Government ini- 200 tiatives have often sought to make fertilizer more affordable at the farm level, com- 100 monly through subsidies, which are enjoy- ing new popularity.60 Subsidies remain 0 controversial, however, in part because of United States Nigeria Malawi Zambia their high cost. To cite a possibly extreme Other domestic costs example, in Zambia 37 percent of the public Port charges, and inland and domestic transport International shipping and insurance budget for agriculture in 2005 was devoted Sourcing to fertilizer subsidies (figure 6.3). Subsidies may also heighten inequality by benefiting Source: Gregory and Bumb 2006. mainly the larger farmers.61 There are situations where fertilizer can locked ones.58 In Malawi, Nigeria, and be productively subsidized, but they need to Zambia, international and domestic trans- be carefully identified (box 6.7). When used port costs make up about one-third of the as part of a broader strategy to address the farmgate price (figure 6.2). binding constraints on supply and demand, Adding to the high logistics costs are well-designed fertilizer subsidies can help to high financing costs. Fertilizer purchases overcome temporary market failures. But typically involve large volumes, and a year or they should be "market smart," contribut- more can elapse between the time advance ing to the development of viable private- payments are made to a supplier and the sector-led input markets.62 Market-smart time proceeds are received from retail sales. subsidies should be targeted to poor farmers Just as producers face risk, so do input sup- to encourage incremental use of fertilizer pliers. If rains fail early in the season, sales by those who would otherwise not use it. As of fertilizer can plummet as farmers scale volumes increase, the market price of fertil- back their planting. And if rains fail late in izer will come down to the true economic the season, credit recovery can become dif- price and reduce the need for subsidies. ficult as farmers experience crop failures and are unable to repay their loans. Figure 6.3 More than a third of Zambia's 2004/05 public budget for agriculture went to fertilizer Promoting seed and fertilizer use subsidies in Africa Food security Given the market failures that lead to Food Reserve Agency and drought relief Irrigation 11% socially suboptimal use of seed and fer- (maize marketing) investment 15% 3% tilizer, governments frequently step in to distribute them directly. Government-led Infrastructure 2% distribution programs have often increased input use, but the fiscal and administrative Personal costs are usually high and the performance emoluments erratic.59 Recent cutbacks in public seed 20% Fertilizer multiplication schemes and public seed dis- Support Program tribution programs have saved money for 37% Operational funds governments, but private companies have 11% not always stepped in to fill the gap, leaving Source: Jayne and others 2006a. 152 WORLD DEVELOPMENT REPORT 2008 B O X 6 . 7 Is there a rationale for fertilizer subsidies? Fertilizer subsidy programs have tried to rem- · They can stimulate adoption by encouraging tilizer subsidies would have to be the most edy low fertilizer use by small-scale farmers in farmers to use fertilizer and learn about its cost-effective option for achieving the desired Africa. Various benefits are cited in justifying benefits, creating positive externalities for social objective, compared with such alter- the subsidies--economic (real productivity others. natives as food aid, food for work, and cash increases), environmental (reductions in land · They can overcome missing or imperfect transfers. degradation), and social (poverty alleviation or credit or insurance markets for farmers that Whether fertilizer aid is cheaper than food emergency relief). Despite having some obvi- cause farmers to use suboptimal amounts of aid depends on the relative costs for govern- ous drawbacks--the high cost, difficult target- fertilizer. ments to acquire fertilizer and food, and to ing, and crowding out of commercial sales-- deliver the items to needy households. It also · They can offset taxes or output price controls fertilizer subsidies continue to have strong depends on the additional food crop output that make fertilizer financially unprofitable, support from farmers and from politicians who likely to be generated per dollar of fertilizer when removal of taxes or price controls is not view farmers as an important constituency. distributed to and applied by farmers--and feasible. Two questions should be addressed in other cost savings associated with fertilizer aid, · They can generate environmental externali- considering whether subsidies are appropri- such as avoiding farm-to-market transport and ties associated with higher soil fertility-- ate for promoting increased fertilizer use. handling costs incurred when farmers must sell reducing soil erosion, deforestation, and First, can fertilizer subsidies bring economic a portion of their crop to repay fertilizer loans. carbon emissions. benefits to societies that exceed their costs? Fertilizer aid would be appropriate if food Second, are there circumstances when subsi- In practice, it has been difficult to imple- markets are working poorly. However, cash dies are justified to achieve social rather than ment subsidies and avoid undesirable market transfers to enable households to purchase economic goals? and distributional effects. food may be more appropriate if food markets are working well, especially in marginal areas For efficiency For welfare where food production payoffs for fertilizer Fertilizer subsidies can bring economic ben- If it would not be economical to use fertilizer use are risky. efits to society in several ways: even when input, output, finance, and risk markets are working well, is there a rationale · They can kick-start fertilizer markets by off- Sources: Conley and Udry 2001; Foster and for using subsidies to achieve noneconomic setting high initial distribution costs until the Rosenzweig 1995; Gramlich 1990; Morris and or social safety-net objectives, such as food market expands, economies of scale are real- others 2007; Sachs 2003; Pedro Sanchez, personal security or emergency income support? Fer- communication, 2007. ized, and prices decline. Market-smart fertilizer subsidies can be On the demand side, efforts to encour- justified, but the conditions for using them age greater use of seed and fertilizer have efficiently are demanding. They should often focused on strengthening the ability stimulate new demand for fertilizer without of farmers to acquire inputs. To stimulate displacing existing commercial sales. They market development, vouchers have been should encourage competition in fertilizer- distributed rather than the inputs them- distribution channels. And they should be selves. In Malawi, under a scheme known temporary, introduced for a limited period, as Inputs for Assets, vouchers were distrib- with a clear schedule for phasing out when uted only to those who had participated they've achieved their purpose. Fertilizer in a public works project, providing some subsidies used as a safety-net measure in self-targeting because wealthier farmers marginal production environments can were less likely to participate in building rarely be recommended, because other roads. Vouchers were redeemable with instruments for providing income sup- local agrodealers, which strengthened port or ensuring food security will almost effective demand for inputs and increased always be more effective. sales--and profits--of private distribu- tors.63 More recently the government of What shows promise? Malawi has sought to increase demand sig- Because public interventions in seed and nificantly through large-scale distribution fertilizer markets have so often failed, of coupons (about 3.5 million in 2006/07), attention is turning to new approaches to with farmers expected to pay a cash price establish sustainable private-sector-led when redeeming the coupon equivalent to input distribution systems. What can be about one-third the retail price of fertilizer. done to overcome the weak demand and With the help of favorable weather, aggre- inadequate supply for seed and fertilizer? gate maize production increased sharply Supporting smallholder competitiveness through institutional innovations 153 after the program was launched, but the budgetary costs of the program have been B O X 6 . 8 Thriving rural input supply retailers as very high and difficult to control, and there agrodealers in Africa has been a high level of displacement of smallholder commercial fertilizer sales. The Rockefeller Foundation has led the dealers, not from the government-owned development of agricultural input sup- Agricultural Development and Marketing In an experimental pilot scheme in ply pipelines in rural Kenya, Malawi, and Agency or from large commercial distribu- Kenya, fertilizer vouchers were sold to Uganda. Working with global partners tors in urban areas. farmers at harvest time as a commitment such as the International Fertilizer Devel- With the number of agrodealers opment Center (IFDC) and local organiza- device to ensure that funds were reserved expanding, the distances traveled by small- tions, it has piloted: holder farmers in search of inputs have for fertilizer rather than drawn away to been drastically reduced in many districts. meet other demands--with good results.64 · Training rural retailers to develop their The range, volume, quality, and price of technical, product, and business man- In Mali and Nigeria, matching grants were agricultural inputs supplied into rural areas agement skills. After being trained, the have also improved significantly. provided to producer organizations dur- retailers become certified as agrodealers. Meanwhile, the default rate on the ing an initial period for use in testing and · Linking certified agrodealers to major credit guarantees was less than 1 percent learning about new technologies.65 agricultural input supply firms, using in the first three years of the program. partial credit guarantees that cover 50 On the supply side, the international The low default rate is attributed to the percent of the default risk. high quality of the technical and busi- research centers of CGIAR have promoted · Repackaging seed and fertilizer into ness management training for the agro- partnerships in eastern and southern Africa small packs (as small as 1 kilogram for dealers--and their acting together to between public plant-breeding programs seeds and 2 kilograms for fertilizer) to ensure repayment. As a result of greater increase the affordability for farmers. and private seed producers. In West Africa, involvement in seed and fertilizer sales, · Organizing agrodealers into purchasing agrodealers have become important Sasakawa Global 2000 has supported small- groups to facilitate bulk purchasing from extension nodes, and several seed, fertil- scale private seed producers by provid- suppliers. The group members provide izer, and agrochemical companies now ing technical training, business advisory joint collateral to guarantee repayment. use the agrodealers to conduct demon- services, and access to credit. In Kenya, strations of new technologies. These efforts to strengthen rural dis- Malawi, and Uganda, the Rockefeller Foun- tribution networks are beginning to bear dation has teamed with local NGOs to build fruit. In Malawi a recent survey of rural Source: Morris and others 2007; Kelly, networks of rural agrodealers (box 6.8).66 markets showed that the majority of farm- Adesina, and Gordon 2003; International ers now buy their inputs from local agro- Fertilizer Development Centre (IFDC) 2005. In Angola, Mozambique, and other coun- tries where farmers lost their seed stocks during civil conflicts, NGOs such as Seeds of Hope have sponsored seed fairs and seed tural input markets. Progress in improving exchanges to supplement emergency seed distribution.67 seed and fertilizer distribution systems will not be sustainable, however, unless there is Another avenue for improving input strong, effective demand for both inputs, supply systems is to strengthen the capacity assured only as long as investment in seed of producer organizations to take respon- and fertilizer is profitable for farmers. That sibility for the final stages of distribution. will be the case only if they have access to For smallholders, purchasing inputs in bulk reliable markets for selling their products and organizing distribution through their at remunerative prices (chapters 4 and 5). own organizations is a way to compensate Building input markets must go hand-in- for inadequate private sector delivery. For hand with building output markets and input suppliers, dealing with producer linking farmers to those markets. organizations presents considerable advan- tages over dealing with geographically dis- Producer organizations persed farmers who individually purchase in a context of value chains only very small quantities of inputs. In Ethiopia, producer organizations are tak- and globalization ing over retail fertilizer distribution from A prosperous smallholder sector is one of government and parastatal companies. the cornerstones of an agriculture-for- These and other innovative efforts to development strategy. Yet, smallholders stimulate greater use of improved seed and typically face high transaction costs and fertilizer provide lessons about state and low bargaining power in factor and prod- donor support to private-sector-led agricul- uct markets. They have limited access to 154 WORLD DEVELOPMENT REPORT 2008 public services, and their voices are often cooperatives.69 In France, 9 of 10 produc- not heard in policy forums where issues ers belong to at least one cooperative, with that affect their survival are being decided. market shares of 60 percent for inputs, 57 In a world increasingly dictated by value percent for products, and 35 percent for chains and the rules of globalization, com- processing.70 petitiveness is the condition for survival. To In the 1960s, many developing-country confront this situation, smallholders have governments initiated cooperative develop- formed various types of producer organi- ment programs, often to ensure quotas for zations to better compete. These organiza- cash crops and distribute subsidized credit tions have expanded rapidly in developing and inputs. Cooperatives were largely gov- countries, and there are dispersed successes ernment controlled and staffed. So farmers on three fronts: markets, public services, considered them as an extended arm of the and voice. However, the world of value public sector, not as institutions that they chains and global market forces is creat- owned. This form of cooperative was rarely ing new challenges for their organizations. successful. Political interference and elite The challenge for the organizations is how capture resulted in poor performance and to respond; for governments and donors it discredited the movement. For example, is how to assist without undermining the in the case of the Indian sugar cane coop- organizations' autonomy. eratives, large growers depress the price of sugar cane to the detriment of small farm- Producer organizations have ers. This generates retained earnings within increased rapidly in developing the cooperatives that large farmers can then countries siphon off through various means.71 Producer organizations are membership- This situation changed radically in b