36789 At Loggerheads? Agricultural Expansion, Poverty Reduction, and Environment in the Tropical Forests A World Bank Policy Research Report This report is dedicated to the memory of Ricardo Tarifa, who died tragically in an airplane accident in the Amazon forest on September 29, 2006. Ricardo was returning from Manaus, where he had visited and contributed to several major projects focused on conserving the Brazilian Amazon. Ricardo's life and work exemplifies, in a very practical way, the theme of this book--seeking ways to conserve the forest and better the lives of its people. Ricardo loved forests and the people that live in them. He felt at home with the communities on the banks of the Tapajós river and those in the Amazonas floodplains. A Yale-trained forest engineer, he moved among distant worlds: the world of forest dwellers, of academe, and of World Bank offices--but his preference was clear. Ricardo loved to work in the field. He believed in the power of local action to promote change, to seek local solutions to global problems. Ricardo, and the contributions he was yet to make, will be missed. At Loggerheads? Agricultural Expansion, Poverty Reduction, and Environment in the Tropical Forests Kenneth M. Chomitz with Piet Buys, Giacomo De Luca, Timothy S. Thomas, and Sheila Wertz-Kanounnikoff The World Bank ©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 9 8 7 This volume is a product of the staff of the International Bank for Reconstruction and Development / The World Bank. The findings, interpretations, and conclusions expressed in this volume do not neces- sarily 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 boundar- ies, 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 endorse- ment 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 Recon- struction and Development / The World Bank encourages dissemination of its work and will nor- mally 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. ISBN-10: 0-8213-6735-8 ISBN-13: 978-0-8213-6735-3 eISBN-10: 0-8213-6736-6 eISBN-13: 978-0-8213-6736-0 DOI: 10.1596/978-0-8213-6735-3 Library of Congress Cataloging in Publication Data Chomitz, Kenneth M. At loggerheads? : agricultural expansion, poverty reduction, and environment in the tropical forests / [written by Kenneth M. Chomitz]. p. cm. -- (World Bank policy research report) Includes bibliographical references and index. ISBN-13: 978-0-8213-6735-3 ISBN-10: 0-8213-6735-8 ISBN-10: 0-8213-6736-6 (electronic) 1. Deforestation--Tropics. 2. Deforestation--Social aspects--Tropics. 3. Deforestation--Environmental aspects--Tropics. 4. Forest policy--Tropics. 5. Agriculture--Tropics. I. Title. II. Title: Agricultural expansion, poverty reduction, and environment in the tropical forests. III. Series. SD418.3.T76C46 2006 2006032118 Cover image: © Michael Fay / National Geographic Image Collection. The cover photo shows the western boundary of Kilimanjaro National Park, Tanzania. Cover designer: Drew Fasick This book was printed using recycled paper. Contents Foreword xi Acknowledgments xv Abbreviations and Acronyms xvii Overview 1 Why Are Tropical Forests a Concern? 1 This Report's Aims, Audience, and Scope 2 This Report's Arguments and Structure 5 Setting the Stage: Two Contrasting Cases of Poverty, Wealth, Biodiversity, and Deforestation 19 Poverty, Biodiversity Loss, and Deforestation in Madagascar 19 Wealth, Biodiversity Loss, and Deforestation in Brazil's Cerrado 21 Part I: The Where and Why of Deforestation and Forest Poverty 25 1. Forests Differ 27 Three Stylized Forest Types 27 From Stylized Types to Mapped Domains 31 The Uneven Distribution of Forest Populations 37 Trends in Forest Change 40 Threatened Species--Concentrated in Less-remote Areas and Mosaiclands 47 Summary 50 AT L O G G E R H E A D S ? 2. Incentives and Constraints Shape Forest Outcomes 53 The View from the Forest Plot: Comparing the Returns to Forestry and Agriculture 54 How Do Agroclimate, Prices, Technology, Tenure, and Other Factors Affect Deforestation and Income? 60 Forest Trajectories: Roads, Markets, and Rights Shape Outcomes for Environment and Income 71 Summary 78 3. Poverty in Forests Stems from Remoteness and Lack of Rights 81 Poverty Rates and Poverty Density: Two Ways of Viewing Poor Areas 82 Remote Forests--High Poverty Rates, Low Poverty Densities 84 Incomes of Forest Dwellers Depend on Rights and Access to Forestlands 88 Forests, Poverty, and Deforestation: Ambiguous Relationships 93 Summary 104 4. Deforestation Imposes Geographically Varied Environmental Damages 109 Biodiversity Loss--A Local and Global Concern 110 How Does Deforestation Affect Water, Air, and Weather? 115 Deforestation Spurs Climate Change 125 Forest Loss--Sometimes Irreversible 129 Summary 131 Part II: Institutional and Policy Responses 135 5. Improving Forest Governance 137 Who Should Have Rights over Forests? Which Rights? 137 How Should Society Balance Environmental Services against Production of Food, Fiber, and Wood? 137 Balancing Interests while Enforcing Commitments 139 Catalytic Innovations in Institutions and Technology 140 Summary 150 i C O N T E N T S 6. Local and National Policies: Framing Rights and Incentives for Forest Management 153 The Challenge of Forest Conflict 154 Forest Rights and Restrictions--A Range of Possibilities 155 Public Management of Forests: Protected Areas and Concessions 161 Community Control of Forests--Balancing Rights and Responsibilities 171 Private Property--Especially in Mosaiclands 178 Other Development Policies with Forest Spillovers 188 Summary 193 7. Mobilizing Global Interests for Forest Conservation 195 Forest Carbon Finance: An Ungrasped Opportunity 195 Why Carbon Finance Makes Sense for Climate 196 Why Carbon Finance Makes Sense for Forests and Rural Development 196 Financing Avoided Deforestation: Problems and Solutions 197 Implementing Incentives for Avoided Deforestation 203 Related Opportunities for Biodiversity Conservation 209 Summary 209 8. Conclusions and Recommendations 211 International Level 213 National Level 215 Accelerating the Forest Transition 218 Appendix A: Tables 220 A.1 Findings of Studies Assessing How Road Proximity Affects Deforestation 220 A.2 Findings of Studies on How Roads Affect Development 230 A.3 Forest Management and Tenure 236 Appendix B: Data and Methods 240 References 247 Index 277 ii AT L O G G E R H E A D S ? Boxes 1 Unreliable Generalizations about Deforestation and Poverty 3 2 The World Bank's Forest Strategy 5 3 This Report's Recommendations 17 1.1 This Report's Geographic Scope 28 1.2 Mapping the Domains and Tallying Their Populations 37 2.1 The Forest Transition 77 4.1 Forest Fragmentation Can Trigger Local Ecological Collapse 114 4.2 Trees and Carbon: Lessons from Biology for Forest Policy 127 6.1 Cameroon: A Nexus of Institutional Reform 168 6.2 Self-assembling Biodiversity Corridors: Reconciling Voluntary Participation Decisions with Landscape-level Goals 186 8.1 This Report's Recommendations 212 Figures 1 Structure of This Report's Arguments 6 1.1 Forests Vary Greatly in Population Density, 2000 41 1.2 Africa and Latin America Have Higher Degradation on Better Soils, 1990­2000 46 1.3 The Incidence of Threatened Amphibian Species Is Much Higher in Nonremote Areas 48 1.4 Imminent Extinction Sites Are Concentrated Near Cities 49 2.1 Deforestation in Brazilian Amazônia Is Shaped by Rainfall and Farmgate Prices of Beef, 2001­03 63 2.2 A Stylized Model of How Land Use Changes with Remoteness 72 2.3 As Remoteness Increases, Mosaiclands Are Displaced by Forests, 2000 73 3.1 Extreme Rural Poverty Increases with Travel Time to Managua 85 3.2 Rural Population Density Decreases with Travel Time to Managua 86 3.3 Forest Cover Increases with Travel Time to Managua 86 3.4 Most Deforestation in Brazilian Amazônia Reflects Large- and Medium-scale Clearing, August 2000 to July 2003 95 iii C O N T E N T S 3.5 Illiteracy and Forest Cover Have No Clear Link in India 99 4.1 Guatemala Critical Watersheds Have High Poverty Rates 121 4.2 Deforestation Would Be Unprofitable in Many Land Systems at Modest Carbon Prices 129 5.1 Optimizing the Mix of Agricultural Output and Biodiversity 138 5.2 Indonesians Favor Some Restrictions on Forest Exploitation 142 6.1 Protected Areas Have Grown Rapidly in Tropical and Subtropical Forests 162 6.2 Recent Decades Have Seen Little Change in the Remoteness of New Protected Areas 166 8.1 Some Forested Countries Will See Shrinking Rural Populations 219 Maps 1.1 This Report's Focus: Tropical Forests and Savanna Woodlands 28 1.2 Domains in Africa's Tropical Forest Biomes 32 1.3 Domains in Africa's Tropical Savanna Biomes 33 1.4 Domains in Asia's Tropical Forest Biomes 34 1.5 Domains in Latin American and Caribbean Tropical Forest Biomes 35 1.6 Domains in Latin American and Caribbean Tropical Savanna Biomes 36 1.7 Hotspots of Tropical Deforestation 44 1.8 Imminent Extinction Hotspots 48 3.1a Poverty Rates for Brazil 82 3.1b Poverty Densities for Brazil 83 3.2 Amazônian Deforestation 2000­03 Showing Rates and Predominant Clearing Size 96 3.3 Amazônian Deforestation Rates and Rural Illiteracy Densities 97 3.4 Poverty, Forests, and Deforestation in Kalimantan 102 3.5 Poverty, Forests, and Deforestation in Sulawesi 103 3.6 Forest Cover, Deforestation, and Poverty in Madagascar 105 4.1 Mortality Risks from Landslides 123 ix AT L O G G E R H E A D S ? Tables 1 Alternative Bundles of Forest Rights 12 1.2 Stylized Forest Types Have Equivalents in Mapped Domains 32 1.3 Forest Populations and Areas Vary by Continent, Biome, Domain, and Remoteness, 2000 39 1.4 Estimated Annual Deforestation Is Highest in Latin America and Asia, 1990­97 42 1.5 During the 1990s Savannas and Asian Forests Experienced Considerable Degradation 45 2.1 Land Values in Forested Areas Vary Enormously 57 2.2 Predictions of How Changes in Local Variables Will Affect the Environment and Welfare 61 2.3 Five Trajectories of Forest Cover, Income, and Population 76 3.1 How Does Increasing Remoteness from Markets Affect Poverty and the Environment? 85 4.1 Externalities of Deforestation Vary by Location of Source and Impact 132 6.1 Examples of Forest Ownership and Use Restrictions 156 6.2 Integrated Conservation-Development Project Interventions Have a Mixed Record 164 6.3 Latin American Countries Impose Varying Restrictions on Deforestation of Private Property 178 7.1 Policies to Reward Avoided Deforestation Can Have Synergistic Effects 208 A.1 Findings of Studies Assessing How Road Proximity Affects Deforestation 220 A.2 Findings of Studies on How Roads Affect Development 230 A.3 Forest Management and Tenure 236 B.1 GLC2000 Land Cover Categories 241 x Foreword T hree billion people--almost half of humanity--live in rural areas of the developing world, and 1.5 billion of them on less than $2 a day. Forests are important resources for the rural poor, with over 800 million people living in forests and woodlands in the tropics alone. However, global deforestation continues at an alarming rate, with annual losses the size of Portugal, as forests are cleared for agriculture or harvested unsustainably. In addition to the implications for poor populations' welfare, forest destruction results in the loss of globally irreplaceable biodiversity and contributes to global climate change, which threatens both the rich and poor. Forests are integral to the Bank's mission of poverty reduction and commitment to mitigating global environmental problems. The Bank's forest sector strategy is founded on three mutually reinforc- ing goals of poverty reduction, economic development, and conser- vation of forest environmental values. While the Bank is committed to engagement in both forest-rich and forest-poor countries in all forest types, this report focuses on the causes, consequences, and connections of deforestation and forest poverty in the tropical world. Specifically, the report addresses the potential dilemma of trade-offs between poverty reduction and environmental protection. Deforestation causes environmental damage, but it also increases the supply of farmland and generates rural income and employment, sometimes sustainable and sometimes not. Overall, the report sug- gests that poverty alleviation and environment are not inherently at loggerheads, nor are they automatically aligned. Outcomes depend on the policies adopted and specific conditions on the ground. The report proposes a typology for three kinds of forests, which face differential kinds of environmental pressure and offer disparate opportunities for growth and poverty alleviation, to appraise policy xi AT L O G G E R H E A D S ? options. It identifies ample opportunities for "win-win" policies. In particular, anything that boosts labor demand outside agriculture will tend to reduce both poverty and deforestation. Additionally, promotion of some kinds of agroforestry can help to improve the ecological functions of degraded forests while boosting farm output and employment. Resolving many forest issues requires mediation between stake- holders with conflicting claims on forests. Sorting out and defending land and forest tenure is one key policy challenge. Millions of peo- ple live with limited or insecure rights to trees and land, unable to tap forest resources and without any motivation to preserve them. Another challenge is recognizing the environmental externalities associated with forest management. Communities at all levels, from local watersheds to the entire planet, need to find ways of reward- ing forest owners and managers whose actions benefit others. These challenges are difficult even for nations with relatively high capacities for governance, yet many tropical-forested nations rank low on governance measures. Nonetheless, the report is cau- tiously optimistic that these challenges can be tackled. It points to a number of innovations that could tip the balance toward improved governance and thus to deployment of better policies. The rapidly decreasing cost of information is a critical factor in the emergence of these innovations, as it becomes cheaper and easier to moni- tor forest conditions, communicate with forest populations, and scrutinize the actions of landholders and of government agencies. Together with new institutional mechanisms such as independent forests observers and third-party certification, these innovations can boost transparency in the sector and restrain environmentally and socially destructive resource grabs. Global finance for forests could accelerate these institutional changes while directly supporting conservation actions and live- lihood improvements. While noting the global demand for biodi- versity conservation, the report focuses particular attention on the potential opportunities offered by global carbon finance. This is a topic of current and increasingly intense international discussion. About 20 percent of global carbon dioxide emissions come from tropical deforestation. The costs of abating some of these emissions appear low in comparison to other options. International finance for carbon services could defray the direct opportunity costs of for- est conservation while also fostering sustainable agricultural and xii F O R E WO R D silvicultural development, which would relieve pressures on pro- tected forests. This is a long-term vision, but it could spur near- term institutional strengthening that would benefit forests and their inhabitants. The report offers a systematic framework for thinking about how to integrate forest management with rural development in a sustainable way. We hope that this report will help to shape the debate on how best to manage the rural landscape for local and global benefits. François Bourguignon Senior Vice President and Chief Economist World Bank Katherine Sierra Vice President, Sustainable Development Network World Bank xiii Acknowledgments T his report was written by Kenneth M. Chomitz (Senior Advi- sor, Independent Evaluation Group) while with the Develop- ment Research Group (DECRG), under the general supervision of Zmarak Shalizi (Senior Research Manager) and L. Alan Winters (Director, DECRG). Chomitz was assisted by a research team made up of Piet Buys, Giacomo De Luca, Timothy S. Thomas, and Sheila Wertz-Kanounnikoff. CIFOR produced two background papers: one by Arild Angelsen, and another by William Sunderlin, Sonya Dewi, and Atie Puntodewo. Dirk Kloss also wrote a background paper. Klas Sander contributed material on Madagascar. The team is grateful for guidance and feedback from an external advisory board consisting of Alain de Janvry, David Kaimowitz, José Sarukhan, and Sara Scherr. Thanks go also to Yasmin d'Souza and Julie Terrell for administrative support; to Paul Holtz for editing; to Susan Graham for managing the production of a particularly complex manuscript; to Nancy Lammers, Stephen McGroarty, and colleagues in the Office of the Publisher; and to Kavita Watsa and Maya Brahmam for assistance in dissemination. The team thanks management and colleagues from the World Bank's Forest Team, Environment Department, and Agriculture and Rural Development Departments for support and advice. The team benefited greatly from comments, discussions with, and help from many people, including Keith Alger, Eugenio Arima, Philippe Ambrosi, Deborah Balk, Garo Batmanian, Diji Chandrasekharan Behr, Jill Blockhus, Anne Branthomme, Mario Boccucci, Sampurno Bruijnzeel, Malcolm Childress, Chona Cruz, Richard Damania, Rob- ert Davis, Uwe Deichmann, Laurent Debroux, Gerhard Dieterle, Giovanna Dore, Ellen Douglas, Gershon Feder, Erick Fernandes, Douglas J. Graham, Theodore Greenberg, Armando Guzman, Mike Hoffman, Peter Holmgren, Miroslav Honzák, William Hyde, Nalin Kishor, Somik Lall, Nadine Laporte, Franck Lecocq, Daniel Leder- x AT L O G G E R H E A D S ? man, Karen Luz, Kathy MacKinnon, William Magrath, Edgardo Maravi, Grant Milne, Augusta Molnar, Adriana Moreira, Carlos Muñoz, Polly Means, Andy Nelson, Daniel Nepstad, Peter Ngea, Ste- fano Pagiola, Kent Redford, Jeffrey Richey, Klas Sander, Jeff Sayer, Robert Schneider, Gerardo Segura, Jim Smyle, Claudia Sobrevila, Fred Stolle, Simon Stuart, Jatna Supriatna, Giuseppe Topa, Juan Manuel Torres, Barbara Verardo, David Wheeler, Tony Whitten, Andy White, Stanley Wood, Greg Yetman, and Liang You. Apologies to anyone inadvertently omitted from this list. Support from the Knowledge for Change Program, the Trust Fund for Environmentally and Socially Sustainable Development, and the German Consultant Trust Fund is gratefully acknowledged. xi Abbreviations and Acronyms ASB Alternatives to Slash and Burn Project CIFOR Center for International Forestry Research CO2 carbon dioxide CONABIO Mexico's National Biodiversity Commission EMBRAPA Brazilian Agricultural Research Corporation ETS Emissions Trading Scheme EU European Union FAO UN Food and Agriculture Organization FEMA [state environmental agency of Mato Grosso] FRA Forest Resources Assessment FRA-RSS Forest Resources Assessment Remote Sensing Survey FSC Forest Stewardship Council GEF Global Environment Facility GHGs greenhouse gases GPS Global Positioning System ICDPs integrated conservation-development projects IMF International Monetary Fund INPE Brazilian National Institute of Space Research ITTO International Tropical Timber Organization IUCN The World Conservation Union LSMS Living Standards Measurement Survey NGOs nongovernmental organizations NPV net present value RISEMP Regional Integrated Silvopastoral Ecosystem Management Project RL reference level SLAPR Rural Property Environmental Licensing System (Mato Grosso) TREES Tropical Ecosystem Environment Observation by Satellite WWF Worldwide Fund for Nature/World Wildlife Fund 100 ha = 1 km2 1 ton carbon is equivalent to 3.67 tons CO2 xii Juan Pablo Moreiras / Fauna & Flora International / Comisión Centroamericana de Ambiente y Desarrollo photo archive. Overview O ver the past three decades tropical forests have captured the world's attention. There have been endless meetings, stacks of reports, demonstrations in the streets, and billions of dol- lars poured into forest projects. Why Are Tropical Forests a Concern? Two broad concerns have driven this attention. Tropical Forests Are Shriveling before Our Eyes Satellites allow us to watch forests burn in real time. The tropical forest estate, extraordinarily large at the middle of the 20th cen- tury, is shrinking at about 5 percent a decade. By the middle of the 21st century only shreds of this once-vast forest may be left. Unless trends change, the consequences will be severe: 3 billion tons of carbon dioxide (CO2) added to the atmosphere each year, intensify- ing climate change; loss not just of many species but also entire eco- systems; and across the tropics, widespread changes in water flows, scenery, microclimates, pests, and pollinators. These environmental damages would touch people near and far. Pressures on forests will not disappear soon. Croplands, pastures, and plantations are expanding into natural forests and will likely do so for the next 30­50 years. Expansion is driven by both wealth and poverty. A huge rural population relies on low-productivity agricul- ture for subsistence. A growing, increasingly wealthy urban popula- tion demands commodities produced at the forest's edge: beef, palm oil, coffee, soybeans, and chocolate. AT L O g g E r h E A d S ? The Food and Agriculture Organization predicts that the growth in such demand will slow--but still expects croplands in the devel- oping world to expand by a net 3.8 million hectares a year over the next three decades (Bruinsma 2003). gross expansion will be even greater, because some farmland is abandoned. And these estimates do not include expansion of pastures and planted forests. Forests are also under pressure from loggers. Poor people need fuelwood, and a wealthier world demands more wood and pulp-- demands only partly met by plantations. Logging thins and degrades forests and helps finance and provide access to farmers and entre- preneurs who burn unsellable trees to establish agriculture. Forests Are Home to Some of the World's Poorest People Forests play a crucial role in the lives of many poor people. Almost 70 million people--many indigenous--live in remote areas of closed tropical forests. Another 735 million rural people live in or near trop- ical forests and savannas, relying on them for much of their fuel, food, and income--or chopping them down for crops and pasture. From a policy viewpoint, what is distinctive about forest poverty relative to other rural poverty? how is it related to deforestation? When are forests a geographic poverty trap--and when are they a route out of poverty? This Report's Aims, Audience, and Scope despite the volume of published material, confusion remains about the causes of forest loss and forest poverty and about effective pol- icy responses. Forest discourse often relies on unreliable generaliza- tions (box 1). Although there is an element of truth in each of these generalizations, uncritical application of them can impede diagnosis of poverty and environmental problems--and without proper diag- noses, prescriptions can go awry. Two examples: · Kerinci-Seblat National Park, in Sumatra, Indonesia, is one of the world's richest, most distinctive biodiversity sites, containing 4,000 plant species and 3 percent of Earth's mammal species--including threatened ones such as the clouded leopard and small Sumatran rhi- noceros. A World Bank­global Environment Facility project sought to deter deforestation by boosting local incomes. But deforestation in Kerinci was driven not by Ov E rv I E W poverty, but by avarice and opportunity. The region's forests consist of prized hardwoods, and its cool cli- mate and volcanic soils make it one of the best places in the world to grow cinnamon. Modest assistance to local people didn't deter them from deforestation and had no effect on outsiders who sought to cash in on the region's wealth (MacKinnon 2005). · A Panamanian land-use plan envisioned reforesting 144,000 hectares of pasture to protect the Panama Canal watershed--a project that might cost more than $250 million. But a study found that this investment could have a result contrary to what was expected, reducing rather than increasing water available to the canal during the dry season. Over time, such a change would cost Panama $630 million in revenues and raise global shipping costs by $3 billion (Aylward 2002). Box 1 Unreliable Generalizations about Deforestation and Poverty Poverty causes deforestation Deforestation causes floods and reduces dry Poor people deforest, but so do the rich. Added season flows income may not deter poor people from defor- deforestation's impacts vary considerably estation. depending on the watershed's size and steep- ness and how the land is subsequently used. It Deforestation causes poverty often increases dry season flows, but in some depending on who does it and why, deforesta- cases could reduce them. tion can destroy or create assets for poor people. High timber prices promote forest conservation Highly forested areas tend to be very poor high timber prices motivate "mining" of unpro- Many factors muddle this relationship. remote tected old-growth forests--but can also increase areas tend to have high forest cover and high returns to regulated logging and stimulate man- poverty rates, but they also usually have low agement of secondary forests and plantations absolute numbers of poor people. Forest dwell- in areas already logged over. ers can prosper when they can profitably access forest resources--or suffer when those resources are meager or controlled by others. AT L O g g E r h E A d S ? This report seeks to improve the diagnosis of forest problems and facilitate the prescription and application of solutions. It offers tools for tackling two issues related to environmental management and regional development policies: · Some people benefit, and some people are harmed, when forest is degraded or converted to agriculture. how should society intermediate between these groups? · how pervasive is poverty among forest dwellers? What approaches are needed to address it? And how do poli- cies for reducing forest poverty relate to those for miti- gating deforestation? The report is directed at people concerned about environmen- tal and poverty policies in the tropical world--particularly those who have to straddle sectors or disciplines. It can help design- ers of conservation projects assess the plausibility of assumptions about links between conservation and poverty. Local governments and stakeholders might use it to think about their goals and tools for regional development in forested areas. At the national level, it seeks to provide a platform for discussions among environment, agriculture, forest, and finance ministries. It can inform policy- makers and voters in formulating equitable, enforceable regula- tions on land and forest use. Finally, it is intended to contribute to international discussions about the role of forest conservation in mitigating climate change. The report's cross-sector approach can inform implementation of the World Bank's Forest Strategy (box 2). The strategy recog- nizes that forests are undervalued because their environmental ser- vices fall outside markets and emphasizes the need to reward forest managers for these services. It also recognizes that tapping forests' potential for poverty reduction and sustainable economic devel- opment requires politically complex trade-offs between the differ- ent groups interested in conservation and production and involves cross-sector coordination. Though this report has broad ambitions, it is limited in scope. It focuses one spotlight on the causes and consequences of forest con- version to agriculture, and another, somewhat overlapping, spotlight on the nature and location of forest poverty. Those spotlights cover a lot of material, but leave many traditional forestry topics partially shadowed. Such issues--the economics of investing in plantations Ov E rv I E W Box 2 The World Bank's Forest Strategy The World Bank's 2004 Forest Strategy and 3. Protecting local and global environmental Operational Policy has three interdependent values by: parts: · Establishing protected areas. 1. harnessing the potential of forests to · Improving forest management in other reduce poverty by: areas. · Strengthening rights of people-- · developing markets and finance for especially marginalized groups--to international public goods such as forests and fostering their participation biodiversity and carbon sequestration, in forest management. and helping governments create · Promoting sustainable forestry, national markets for environmental community forestry, and agroforestry. services from forests. · Addressing cross-sector links that affect 2. Integrating forests in sustainable economic environmental values. development by: · Improving forest governance and introducing legal and institutional reforms. · Encouraging investments that catalyze production of forest products, including environmental services. Source: World Bank 2004. and building capacity among small sawmills and furniture enter- prises, the policy rationales for stimulating smallholders to grow pulpwood, introducing marketing interventions for community for- ests, promoting reduced-impact logging, and controlling illegal log- ging--enter the story, mainly as they affect incentives to maintain or convert natural forests. But readers should not expect detailed or operationally oriented discussions of these topics. This Report's Arguments and Structure The report has two parts. The first is diagnostic: it examines the drivers and consequences of deforestation and forest poverty. The second part steps back to see how governance, institutions, and pol- icies shape those drivers--leading to prescriptions. The line of argu- ment is summarized below by chapter and outlined in figure 1. AT L O g g E r h E A d S ? Figure 1 Structure of This Report's Arguments Part I: The Where and Why of Deforestation and Forest Poverty 1 Global variations in local forest context Chapter 2 Local context shapes environmental and livelihood outcomes Chapter determining determining 3 The relationship between 4 The environmental economics forests, deforestation, of deforestation--who gains Chapter and poverty Chapter and who loses how Part II: Institutional and Policy Responses how 5 affecting Governance affecting and institutions Chapter shaping shaping 6 7 Local and International national policies policies Chapter Chapter 8 Conclusions and recommendations Chapter Chapter 1. Forests Differ Forests differ in the deforestation pressures they face, the extent and depth of poverty they harbor, and the environmental consequences of their conversion. Understanding these differences is essential to prescribing appropriate institutions and policies. Chapter 1 distinguishes three stylized forest types: Forest-agriculture mosaiclands--where land ownership is usually better defined, population densities higher, and Ov E rv I E W markets nearer, and natural forest management often can- not compete (from the landholder's perspective) with agri- culture or plantation forestry. Frontier and disputed areas--where pressures for defores- tation and degradation are increasing, and control is often insecure and in conflict. Areas beyond the agricultural frontier--where there is a lot of forest, few but largely indigenous inhabitants, and some pressure on timber resources. The chapter maps proxies for these types, showing that most forests have few residents--and that most forest dwellers occupy a relatively small portion of the forest estate. The chapter also shows that frontier and mosaic forests have high deforestation rates and threatened or unique biodiversity. These different constellations of economic pressures, forest tenure security, and environmental cir- cumstances require different policy responses. Chapter 2. Incentives and Constraints Shape Forest Outcomes People clear and log forests because they gain from doing so. gains can be unsettlingly small or impressively large (see the "Stage-Setting" section that follows this overview), ephemeral or sustainable. Chapter 2 explores how local conditions, incentives, and constraints determine where and why deforestation occurs, and with what impacts. A simple economic framework applies to all forest actors: sub- sistence households and large companies; farmers, ranchers, and loggers. The framework revolves around the relative attractiveness of maintaining forest relative to converting it to agriculture. Land- holders and land claimants weigh cultural, economic, and legal con- siderations when making decisions about land use. A central issue for the report is that some may find agriculture a more profitable, attractive land use than sustainable management of forests for tim- ber and other products. Low wages, good soils, and higher prices for agricultural goods all motivate deforestation. In addition, high prices for timber can provoke mining of old-growth forests--though it can also stimu- late sustainable management of plantations and secondary forests. These relationships are strongly affected by governance and tenure conditions. Where governance is weak and tenure poorly defined, powerful interests can seize forest resources, and smallholders can AT L O g g E r h E A d S ? engage in conflict-ridden races for property rights. But even land- holders with secure tenure may choose deforestation if it offers higher returns. The report uses empirical examples to illustrate aspects of this general framework as they apply in different contexts. The frame- work helps explain and predict: · Where deforestation occurs. · Private gains and public costs of deforestation. · how a wide range of policies--involving trade, road expansion, forest tenure, and other areas-- are modulated by local conditions to affect poverty and environment. · Why some places experience forest transitions-- deforestation followed by recovery in forest cover-- while others follow immiserizing paths of deforestation and increasing poverty. Chapter 3. Poverty in Forests Stems from Remoteness and Lack of Rights The relationship between forests and poverty eludes simple general- izations, as shown by the examples in the Stage-Setting section that follows this overview. Asserting that poverty causes deforestation, or vice versa, doesn't provide a fruitful framework for understand- ing the issues. Empirically, this link is weak. Although poor subsistence farmers cut down trees, so do rich ranchers and plantation owners. defores- tation can deprive poor people of resources--but it can also provide them with sustainable incomes from cash crops. This chapter pres- ents new data showing weak, inconsistent geographic overlap among forest cover, deforestation, and poverty in several forested countries. So what is distinctive about forest poverty as opposed to other rural poverty? First, remoteness. Because the best, most accessible farming lands have long been cleared and tilled in many parts of the world, forests and their inhabitants tend to be relegated to remote or unfavorable areas. As a result, areas with high forest cover often have low population densities but high poverty rates. Second, forest dwellers may be unable to tap forest resources. People living in or near forests derive much of their income from Ov E rv I E W collecting fuelwood, food, and other forest products, or by practic- ing long-fallow agriculture. If they lack the right to harvest forest resources or to use forest land for cropping, their income can suf- fer. Sometimes this happens when governments or wealthy interests claim forests and restrict access. In other cases forests effectively belong to no one--with the result that their resources are degraded through overuse. Third, forest dwellers may lack the resources, capacity, and social organization to profit from managing forests for timber or nontimber products. Chapter 4. Deforestation Imposes Geographically Varied Environmental Damages Environmental problems are social problems, and society may be moved to intervene if one person's land-use decisions significantly affect other people's well-being. Chapter 4 traces those impacts, which operate through different channels and depend on the loca- tion of deforestation, for instance: · The most widespread impact--and arguably the one with the most costly damages--is the effect of forest loss on climate change through CO2 emissions. These greenhouse emissions are associated with all perma- nent losses of forest, regardless of location. Moreover, their physical impact is reasonably well understood, and society can place an increasingly well-defined eco- nomic value on reducing these emissions. · There is considerable global demand for prevent- ing extinctions and other biodiversity losses, and an increasingly good understanding of where biodiversity is richest and most threatened. · The impacts of forest loss on flooding, smoke pollu- tion, and water availability and quality are important in some locations but specific to local conditions and changes in land use. Protecting local watersheds can be important for maintaining urban water quality. · Some theories and evidence suggest that deforestation can cause local and global changes in weather patterns quite apart from its effects through CO2 emissions. AT L O g g E r h E A d S ? · direct economic benefits of forest conservation--such as pollination and pest control--are likely to be great- est in mosaiclands but have not been well quantified in physical and economic terms. Chapter 5. Improving Forest Governance diagnosis of forest poverty and environmental issues reveals two basic problems: · Many forests are nominally owned by governments, but actual control is unclear or disputed. Elsewhere, private and community rights are not respected. Who should have the right to use and manage forests? how can rights be reliably enforced? · how should the interests of forest owners in removing trees be balanced against the interests of others--near and far--in maintaining the environmental benefits of those trees? These are problems of governance that require balancing inter- ests between groups, negotiating solutions, and enforcing commit- ments. But these problems have been difficult to address. First, elites tend to capture the institutions that allocate forest resources. Second, there are strong asymmetries of information, power, and organization between the beneficiaries of deforestation and those who bear its burdens. The diffuse interest groups favoring forest conservation find it hard to organize themselves to counter- balance the concentrated interests of forest degradation. Building on a framework introduced in the World Development Report 2003: Sustainable Development in a Dynamic World (World Bank 2002), chapter 5 describes institutional and technological innovations that might help overcome these two barriers to collec- tive action, thereby facilitating implementation of the prescriptive policies described in chapters 6 and 7. These catalytic innova- tions--which include building constituencies for conservation and better governance, improving public monitoring and disclosure of forest conditions and management, certifying forest and agricultural products, and introducing more flexible, market-like approaches to environmental regulation--can help diffuse groups organize, check abuses of power, and cut the costs of reaching agreements. 0 Ov E rv I E W Chapter 6. Balancing Interests at the National Level Nations need to sort out who has the rights to manage forest and how stringently to regulate those rights in the public interest. The challenges play out differently in different types of forests: In mosaiclands, where agriculture and forests are in close contact, the challenge is to ensure that land managers take into account the benefits of forest maintenance for their neighbors. At the frontier and in disputed regions, to resolve conflicting claims to forestlands and determine where gains from forest conversion outweigh environmental damages. Beyond the agricultural frontier, to recognize and defend long-standing indigenous claims, tap and fairly share rents from timber exploitation while avoiding needless forest deg- radation, and avert disorderly races for property rights when the frontier arrives. To realize these goals, governments can deploy the following tools, often in combination: Tenure, zoning, and land-use regulation--revisiting the ownership and management of government lands, imple- menting systems to enforce property rights, regulating the exploitation of public and private forests, and promoting participatory planning for land management. Making forest management more attractive relative to agri- culture--by funding or facilitating markets for environmen- tal services, researching, developing, and disseminating environmentally friendly land management practices, and removing barriers to sustainable management of forests for timber and other products. Coordinating regional development interventions (such as road network expansion and agriculture policies)--to exploit synergies between, or minimize trade-offs between, environmental and livelihood goals. Table 1 shows possible ways to assign property and use rights in forests. Allocating and enforcing property rights and land use regulations is not easy, for reasons described in chapter 5: doing so AT L O g g E r h E A d S ? Table 1 Alternative Bundles of Forest Rights Ownership and/or management Use restrictions State Community Private No restrictions on State forests zoned Some common property Private lands conversion for conversion Conversion prohibited, Direct state Most community forestry Regulated private forests sustainable management management; forest allowed concessions Limited or no productive Strict protected areas Some indigenous lands Private reserves use requires settling disputes between groups and enforcing agreements. Chapter 6 reviews efforts to zone and regulate land use at regional and national levels. These efforts have often foundered due to fail- ure to motivate compliance by landholders and to create reliable institutions for resolving disputes and preventing resource capture by elites. Still, some positive examples are emerging. Chapter 6 also assesses the pros and cons of different tenure and management regimes: Protected areas. The establishment of protected areas is perhaps the longest-standing, most widely practiced, and best-funded approach to maintaining forest environmental services. Evidence suggests that such areas can reduce defor- estation even in weak institutional settings. Their effects on livelihoods are less documented, but they have been nega- tive when people have been excluded from protected areas that they relied on for forest products. But there is a trend toward permitting multiple uses for protected areas, and the World Bank has instituted strict social safeguards for their creation. Most new protected forest areas are beyond the agricultural frontier, where it is easier to accommodate local residents, and there is less competition from commercial interests. Indigenous areas. Management and ownership of remote forest areas is increasingly being transferred to indigenous control. Indigenous ownership is sometimes associated Ov E rv I E W with much lower deforestation rates than in compara- ble areas. But indigenous areas are sometimes prohibited from undertaking commercial logging or large-scale land transformation. Regulated logging concessions. In frontier areas, where land and forests can be profitably exploited, the biodiversity benefits of protected areas come at an opportunity cost. In principle, regulated logging concessions offer considerable biodiversity protection at a much lower opportunity cost. Where constituencies for protected areas are weak, estab- lishing regulated logging concessions may be a politically feasible alternative that could be far superior--in environ- mental terms--to agricultural conversion. Innovations in monitoring and control, including certification, can increase public capture of logging profits and reduce environmen- tal damages associated with logging. Efficient regulation-- streamlining regulations to ease the burden of compliance and monitoring costs--can also help. Community forest management. Communities are increas- ingly sharing management of or taking ownership of pub- lic forests. In principle, communities should be better than distant governments at managing and policing their forests, and better suited than individuals to exploit economies of scale in forest management. But successful community management depends on the strength of community orga- nization, the regulations facing communities, and economic and cultural incentives to maintain forests. Communities need strong social capital to enforce compliance with man- agement rules and avoid elite capture of forest resources. Communities may lack the ability to commercially exploit forests or effectively negotiate sales of logging rights to outsiders. Onerous regulations--such as requirements for detailed management plans--can be prohibitively costly. The economics of community management of natural for- ests can be unfavorable when market access is poor or the density of sellable tree species is low. Still, some communi- ties have overcome these obstacles. Policies to support com- munity forestry include building capacity, fostering markets for less-known wood species, and streamlining regulations. AT L O g g E r h E A d S ? Privately owned forests. reconciling agricultural and envi- ronmental services has proven difficult in frontier and mosaiclands. Some countries have applied zoning and for- est regulations of varying sophistication. Although rigor- ous impact evaluations are lacking, these regulations do not appear to be heavily enforced on wealthy interests-- but may impose costs on poor people, without generating environmental benefits. Systems for environmental service payments and tradable development rights can help secure landholder cooperation in achieving environmental goals. There may also be scope to organize communities in sup- port of land management goals. Many rural development and agricultural policies have spillover effects on forests and deforestation. Placement of rural roads is espe- cially important for policy attention because it is under direct policy control (though subject to political pressures). rural roads can have large effects on both rural incomes and deforestation pressures. Thus, careful planning and regulation of road construction, and coordina- tion of road policies with land and forest tenure regularization, can minimize trade-offs between rural incomes and environmental pro- tection. Similarly, policy-induced increases in agricultural commodity prices could benefit rural populations but will tend to increase pres- sures for forest conversion; these side effects need to be anticipated. Special attention needs to be paid to the challenges of reducing poverty in remote areas with low population densities. Protecting forest and land rights is a start. Innovative means of delivering ser- vices to these areas are also needed. Chapter 7. Mobilizing Global Constituencies for Forest Conservation While forests have many environmental benefits, only two com- mand a global constituency with potentially large willingness to pay for those benefits: carbon storage and conservation of globally significant biodiversity. Mobilizing global finance for these environ- mental services is a crucial long-term challenge. global carbon finance offers an ungrasped opportunity for miti- gating climate change, supporting sustainable land use, and conserv- ing forests. About a fifth of global CO2 emissions come from tropical deforestation--and the costs of abating some of these emissions appear low. In Latin America dense tropical forest is often cleared to Ov E rv I E W create pastures worth a few hundred dollars a hectare, while releas- ing 500 tons of CO2 per hectare. This implies, at a societal level, a CO2 abatement cost of less than $1 a ton. Meanwhile, some observers think that tackling climate change requires paying about $3 a ton for CO2 abatement--and European Union (EU) members are currently paying up to $20 a ton (though this price is volatile). In other words, deforesters are destroying a carbon storage asset theoretically worth $1,500­10,000 to create a pasture worth $200­500 (per hectare). Yet carbon markets, such as those under the Kyoto Protocol and EU Emissions Trading Scheme, do not reward forestholders for reduced emissions from avoided deforestation. The failure to grasp this opportunity reflects concerns about the implementation of incentive payments for reducing forest emissions. But a long-term approach could address those concerns by: Securing global commitment to mitigating climate change, so that reduced deforestation helps reduce the global cost of arresting rising atmospheric CO2. reduced deforestation must be part of a long-term global package that includes lower industrial and transport emissions and more rapid research and development on clean energy. Creating a system of financial incentives, funded by indus- trial countries, for developing countries to reduce their for- est CO2 emissions. This might be incorporated, for instance, into a change-climate regime whereby industrial countries could meet more stringent requirements for reducing CO2 emissions partly by funding national programs to reduce deforestation. developing countries would receive pay- ments tied to measured reductions in deforestation below some agreed level. Developing national infrastructure for forest carbon--that is, institutions and policies to monitor it and reduce deforesta- tion. In most countries this would be a gradual process. The institutional requirements, though significant, would be consistent with those already required for better forest gov- ernance. The policies would not necessarily involve direct payments to forestholders, but would support systems for preventing accidental forest fires, strengthening monitoring and enforcement of regulations, regularizing forest tenure, AT L O g g E r h E A d S ? and increasing returns to forest maintenance relative to agriculture. Stimulating production of food and timber on degraded lands. This is crucial to counteracting leakages (where pres- sures on protected forests are diverted to unprotected ones), and is an important avenue through which emission reduc- tion programs stimulate sustainable development. Providing small payments per ton of CO2 deferred each year, rather than large upfront payments for unenforceable com- mitments to perpetual reductions in CO2. The spatially concentrated nature of threatened biodiversity lends itself to environmental service payments in which landholders would be paid for maintaining habitat quality. Such markets might be particularly apt for mosaiclands--where biodiversity is highly threatened and land tenure is reasonably well defined--and for community forests--where buyers of conservation services might be able to outbid loggers for concession rights. Auction-based systems for purchasing conservation services have advantages of transpar- ency and efficiency. Such systems could elicit self-assembling bio- diversity corridors in biodiversity hotspots where forest remnants persist in areas less attractive to agriculture. To date there has been no large-scale financing mechanism for payments of this kind. Existing conservation funders--including the global Environmental Facility (gEF) and nongovernmental organi- zations (NgOs)--could direct part of their portfolios to such pay- ments. Supplemental funds could be raised if nations, individually or together, create markets for biodiversity offsets to compensate for damages associated with construction, mining, and other projects that harm the environment. In addition to these long-horizon initiatives, the international community could immediately fund the compilation of information that is critically needed to plan and execute policies for reducing for- est poverty and deforestation. Severe but readily remediable infor- mation gaps include: · rates, locations, and types of deforestation and degradation. · Poverty levels of forest-dwelling and forest-using populations. Ov E rv I E W · Monitoring and evaluation of the environmental and economic outcomes of forest conservation projects and policies for devolving forest control. · Physical and economic impacts of forest conservation on environmental service flows. Chapter 8. Conclusions and Recommendations In the long run, rising wages and urbanization will pull rural peo- ple away from marginal lands at the forest edge, halting deforesta- tion and in some cases resulting in forest regrowth and recovery. But some forests may never recover, and others may irretrievably lose some of their biodiversity. Better institutions for forest manage- ment can help bridge the forest transition--preventing deforestation for small and ephemeral gains while providing more sustainable livelihoods. Box 3 This Report's Recommendations International level land and forest allocations and regulations · Mobilize carbon finance to reduce defores- more transparent, and support civil society tation and promote sustainable agriculture. organizations that monitor regulatory com- · Mobilize finance for conservation of glob- pliance by government, landholders, and ally significant biodiversity. forest concessionaires. The prospect of car- · Finance national and global efforts to moni- bon finance can help motivate these efforts. tor forests and evaluate the impacts of forest · Make forest and land use regulations more projects and policies--including devolution efficient, reformulating them to minimize of forest control. monitoring, enforcement, and compliance · Foster the development of national-level costs. Economic instruments can help. research and evaluation organizations through twinning with established foreign Areas beyond the frontier partners. · Avert disruptive races for property rights by equitably assigning ownership, use rights, National level and stewardship of these lands. · Create systems for monitoring forest con- · Options for forest conservation include ditions and forest dwellers' welfare, make combinations of indigenous and community (continued on next page) AT L O g g E r h E A d S ? Box 3 (continued) rights, protected areas, and forest conces- Disputed areas sions. Still, some forest may be converted · Where forest control is transferred to local to agriculture where doing so offers high, communities, build local institutions with sustainable returns and does not threaten upward and downward accountability. irreplaceable environmental assets. · Where community rights are secure and · Plan for rational, regulated expansion of markets are feasible, provide technical assis- road networks--including designation of tance for community forestry. roadless areas. · Make landholder rights more secure in "for- · Experiment with new ways of providing ests without trees." services and infrastructure to low-density · When forest tenure is secure, use carbon populations. markets to promote forest regeneration and maintenance. Frontier areas · Assign and enforce property rights equitably. Mosaiclands · Plan and control road network expansion. · reform regulations so that they don't penal- · discourage conversion in areas with hydro- ize tree growing. logical hazards, or encourage community · Promote greener agriculture--such as inte- management of these watersheds. grated pest management and silvopastoral · Use remote sensing, enhanced communica- systems--through research and develop- tion networks, and independent observers ment, extension efforts, community organi- to monitor logging concessionaires and pro- zation, and reform of agriculture and forest tect forestholders against encroachers. regulations. · Consider using carbon finance to sup- · develop a wide range of markets for environ- port government and community efforts to mental services--carbon, biodiversity, water assign and enforce property rights. regulation, recreation, and pest control--to · Encourage markets for environmental ser- support more productive, sustainable land vices in community-owned forests. management. SETTING THE STAGE Two Contrasting Cases of Poverty, Wealth, Biodiversity, and Deforestation This report emphasizes the need for understand- real GDP per capita has fallen from $383 to $246. ing local links among deforestation, environ- About 70 percent of the total population and 77 ment, and poverty. To set the stage and illustrate percent of the rural population live below the the diversity of forest situations, consider two national poverty line. sharply contrasting cases of deforestation. In Meanwhile, Madagascar's natural capital, Madagascar much deforestation is undertaken by priceless but unsellable, is being run down. Dur- poor people for paltry and unsustainable gains. ing the 1990s deforestation proceeded at 0.86 In Mato Grosso, Brazil, large commercial farmers percent a year, and habitat fragmentation threat- realize substantial monetary gains from defores- ened the survival of forest species. But forests are tation. In both cases the environmental impacts not being used to create productive, sustainable are widely felt. assets: they are being converted to low-productiv- ity maize and rice cultivation (photo 1). Averag- Poverty, Biodiversity Loss, and ing 2 tons a hectare, rice yields are barely half Deforestation in Madagascar the world average (Randrianarisoa 2003; Uphoff 2003). Some fields are rapidly degraded and If there were a real market for biodiversity, Mada- abandoned, and some forest irreversibly lost. And gascar would be rich--the sole owner of 11,200 with stagnant productivity and a rapidly expand- endemic plant species and 144 endemic mam- ing (2.8 percent annual growth) but still largely mals, including charismatic lemurs. In fact, Mada- rural population, pressures on the forest seem gascar has one of the world's largest assemblages likely to continue. of unique plants and animals. Is poverty responsible for deforestation in But there is no organized market for biodi- Madagascar? At the broadest, most macro level versity, and Madagascar is poor. Despite its $75 the answer must be yes: people persist in forest- million investment in protected areas since 1991-- degrading activities with low returns because the much of it supported by donors--the country's economy offers no better alternative. But the pov- tourism revenue has not risen much. Since 1960 erty-deforestation link blurs if we look closer for 19 AT L o G G e R h e A D S ? Deforestation in Madagascar's moist forests is rural incomes by stimulating expansion of farms undertaken for low-yielding upland rice cultivation. into forest--leading to a trade-off between poverty reduction and environmental protection. export- oriented maize farming has been responsible for intense hotspots of deforestation in the country's dry southwestern region. For more subsistence- oriented farmers in the country's humid forests, Ferraro (2002) and Shyamsundar and Kramer (1996) show that restricting their ability to con- vert forest would translate into lower incomes. Although policy solutions for deforestation and poverty have been elusive, the search contin- ues. Protected areas seem to be effective in reduc- ing deforestation (Gorenflo and others 2006), but protected area revenues and integrated conserva- tion development projects (often associated with protected areas) have done little to augment local incomes. experiments with community co-man- agement of state forest areas have also not lived up to expectations (Antona 2002). one hope is that intensifying agriculture in lowland irrigated areas could reduce migration to forest frontiers. In some cases control of upland deforestation may reduce sediment flows that clog lowland irrigation canals--making irrigated rice more productive. Although the obstacles are formidable, mar- kets for environmental services may offer a partial © Rickey Rogers / Reuters / Corbis. long-run solution. Deforestation in Madagascar releases carbon dioxide (Co2) into the atmo- diagnoses that will help determine policy solu- sphere, contributing to global warming. Although tions. Comparing localities and holding other the country's contribution to global warming is things constant (such as road access and topog- tiny, a global market for Co2 emission reductions raphy), there is not a strong correlation between could provide financing for sustainable land man- local poverty and deforestation rates. But defor- agement in Madagascar. A program for reducing estation is closely associated with proximity to global emissions might be able to offer farmers roads (Gorenflo and others 2006), while poverty more for forest conservation than the proceeds of is associated with distance from them (INSTAT low-yield rice production. And over the very long and others 2003). run, the uniqueness of Madagascar's biodiversity These findings suggest that improving the assets and growth in ecotourism demand could country's dilapidated road network could raise yield a lucrative income stream. 20 S e T T I N G T h e S TAG e Wealth, Biodiversity Loss, and are not available, the 1995­96 agricultural census Deforestation in Brazil's Cerrado found that 72 percent of Mato Grosso farmland (and 58 percent in the entire center-west region) Cattle and soybeans are displacing cerrado is in establishments larger than 1,000 hectares (savanna woodland) and forest in Brazil's center- (IBGe 1998). west region (photo 2). Between 1999/2000 and Cerrado conversion comes at an environ- 2004/05 the states of Goias, Mato Grosso, and mental price that is large but difficult to quan- Mato Grosso del Sul planted an additional 54,000 tify. The cerrado is home to 4,400 plant species square kilometers to soy--an area slightly larger found nowhere else and is one of the planet's than Costa Rica--doubling the area under soy most important biodiversity hotspots. Yet this cultivation. At the same time, cattle herds in these states soared from 57 million in 1999 to 71 Recently cleared farmland abuts Amazônian forest in million in 2004. While some of the soy expanded Mato Grosso state, Brazil. into former pasture, the combined effect was loss of savanna and forest. The Brazilian National Institute for Space Research (INPe) estimates that 38 percent of total Amazonian deforestation over 1999­2003 occurred in Mato Grosso (INPe 2006). About 5.6 percent of remaining Mato Grosso cer- rado woodlands were deforested between 1998 and 2002. Rising profits catalyzed the boom. In 1999 Brazil devalued its currency, which fell against the U.S. dollar by 50 percent or more, making exports more attractive. At the same time, the price of soybeans rose from $184 to $277 in 2004 (USDA 2006), and control of hoof and mouth dis- ease boosted the value of beef exports (Kaimow- itz and others 2004). In effect, there is money lying on the ground for the taking--if trees are removed. In Mato Grosso the price of fields in local currency quadru- pled between mid-1999 and the end of 2004. Pas- tures were worth less, but more favorable lands command higher prices: more than $3,000 a hect- are for fields in the most productive parts of Goias state. In 2002 forest conversion in Mato Grosso created farmland with a gross value of about $100 million (Chomitz and Wertz-Kanounnikoff 2005). Most of these values are being appropriated by large farmers and ranchers. Although recent data © Louise Cobb / Corbis SABA. 21 AT L o G G e R h e A D S ? irreplaceable biodiversity is hard to mone- erty environmental Licensing System (SLAPR), it tize. It can't compare with African savannas in promised a technological and institutional revolu- terms of large, charismatic mammals (though it tion in regulating land use on private rural prop- does boast a splendid anteater) and so cannot erties. The system used a multipronged approach support an extensive ecotourism industry. Its to encourage compliance with land use regula- unique plants may contain commercially valu- tions: deterring deforestation on unlicensed prop- able genetic information, but there is no current erties, encouraging landowners to license their market for this information, and a future market properties, and enforcing regulations on licensed probably wouldn't support conservation of the properties. The central innovation was to license entire area--a modest area might provide all the large landholders, requiring them to precisely samples needed. The cerrado also has subtle but map their allowed land use using a geographic important environmental links to the rest of Bra- information system. Satellite images were then to zil. Deforestation and land degradation in the cer- be used to monitor compliance by licensed land- rado, for instance, have resulted in higher rainy holders, as well as look for illegal forest burning season flows on the Tocantins River (Costa, Botta, and clearing on unlicensed properties. Because and Cardille 2003). Sediment and agrochemicals large properties (larger than 1,000 hectares) were run off fields, polluting downstream drinking few in number but accounted for most private water. land, this approach was potentially cost-effective The result is that landholders face a rather (Fundação estadual do Meio Ambiente 2001). easy choice, Brazilian society a more difficult The system's design called for all licenses to be one. From a landholder's perspective, converting posted on the Internet to facilitate public over- a few hundred hectares of cerrado to pasture or sight of compliance and enforcement--a feature soy brings substantial personal gain. From soci- never implemented. ety's perspective, those palpable immediate bene- early studies of SLAPR credited it with reduc- fits, and any knock-on local development effects, ing deforestation (Fundação estadual do Meio have to be weighed against large but unmonetiz- Ambiente 2002; Fearnside 2003). But it is difficult able environmental damages. to separate the system's effects from the annual Brazil's long-standing forest code enshrines ups and downs driven by economic factors such a pragmatic solution to this problem, requiring as soy and beef booms. Using disaggregated, landholders to set aside 20­80 percent of their spatially explicit FeMA deforestation data from properties as forest reserve (depending on loca- 1999­2002, Chomitz and Wertz-Kanounnikoff tion). They also must maintain forests by river- 2005 found that before 2002, the system did shift banks and on hills. But landholders face strong landholder behavior in a direction consistent with incentives to flout the rules, and authorities have reduced illegal deforestation. had a hard time monitoring and enforcing these But subsequent developments showed SLAPR regulations over Brazil's vast territory. to be ineffective. During 2003­04 deforestation In 2000, in response to high deforestation soared in Mato Grosso. Instituto Socioambiental rates, the state environmental agency of Mato (2005) found that during this period SLAPR failed Grosso (FeMA) implemented an innovative envi- to achieve what was arguably its main goal: pre- ronmental control system. Called the Rural Prop- venting illegal deforestation on licensed proper- 22 S e T T I N G T h e S TAG e ties, where such activity should be easy to detect sistent state and federal environmental legis- and prosecute. More spectacularly, in June 2005 a lation, missing institutional cooperation, and huge network of illegal deforestation and timber inadequate transparency and control of FeMA trade was uncovered during a federal anticorrup- activities (Lima and others 2005; Barbosa 2006). tion operation. Presumed to have existed for at FeMA has since been overhauled, and it least a decade, the corruption scheme involved intends to reinvigorate SLAPR (Araújo 2006). 600 timber companies, dozens of intermediary Meanwhile, the neighboring state of Goias is con- traders, and numerous public servants from the sidering an alternative, possibly complementary federal and Mato Grosso environmental agencies approach to enforcing the forest reserve require- (Diário de Cuiabá, December 31, 2005). About ment. Like Mato Grosso, Goias has experienced 200 people were arrested during the sting-- an agriculture-led boom. Its private sector is including the director of FeMA, who was charged interested in finding ways to continue to grow with contributing to illegal deforestation by issu- while complying with environmental laws. Com- ing environmental licenses in protected areas. pliance would not only remove legal uncertain- (Three main perpetrators have been jailed; other ties, it could also facilitate financing as well as prosecutions are proceeding.) exports to green global markets. Moreover, FeMA officials in the state govern- one possibility is to allow trading of forest ment that entered office in 2003 were accused of reserve obligations. Doing so would allow farm- issuing environmental licenses and deforestation ers who lack enough forest reserves to meet their permits arbitrarily (such as by using alternative obligations by paying to protect an equivalent land cover maps to define the required propor- amount of forest elsewhere. In principle, this tions of legal forest reserves) and of being neg- approach could greatly reduce the cost of compli- ligent in enforcing environmental laws. Major ance while boosting the environmental quality of problems with implementation of SLAPR include the reserves (see chapter 6). But making it work insufficient spatial coverage and automation, a will require building more reliable institutions for possibility of falsified satellite images, FeMA's monitoring trades and tracking compliance. insufficient human and technical capacity, incon- 23 PART I The Where and Why of Deforestation and Forest Poverty In an Assam (India) forest severely degraded by illegal logging, inhabitants eke out a living selling firewood. Grant Milne / World Bank. Frontier areas: A mounted forest patrol returns from a day in the field, northern Sumatra. Photo by Josef Leitmann, World Bank. CHAPTER 1 Forests Differ F orests differ: in the pressures they face, the people they sup- port, and their environmental functions. Three stylized types (forest-agriculture mosaiclands, frontier and disputed areas, and areas beyond the agricultural frontier) in two biomes (forests proper and savanna woodlands) capture much of the social, envi- ronmental, and economic variation of tropical forests (box 1.1). This chapter describes these stylized forest types in the tropical world and the challenges they face. It maps the extent of these types, tallies their populations, and assesses their biodiversity. The analysis here sets the stage for subsequent discussions of appropriate policies for different kinds of forests. Three Stylized Forest Types How are we to make sense of the staggering diversity of tropical for- ests? A few factors shape forests' environmental characteristics and human pressures on them--but these generalizations are riddled with exceptions. A first cut is the distinction between "true" forests and savan- nas. Forests typically receive more rainfall and (if undisturbed) have unbroken canopies, high densities of wood, and more diverse tree species. These features make them susceptible to selective "min- ing" for timber, harder to clear for agriculture, and major sources of carbon dioxide emissions and biodiversity loss if cleared. Savannas are grasslands and woodlands dotted with smaller trees and shrubs. Easier to clear, less diverse, and more resilient than rainforests, and less valuable for fine woods, these woodlands are often cut for fuel- wood and charcoal. 27 AT L o g g E r H E A D S ? Box 1.1 This Report's Geographic Scope To keep this report's scope manageable, it here where almost all deforestation and forest- focuses on the developing world's tropical for- related carbon emissions occur. ests and savanna woodlands, with an empha- This report looks broadly at ecosystems, sis on the former. This approach was chosen rather than narrowly at trees. The focus is on somewhat reluctantly, since there are impor- tropical and subtropical forest and savanna tant challenges related to the management of biomes--that is, areas originally covered boreal and temperate forests, and some for- by these types of vegetation, as mapped by est governance issues cut across all forest WWF (map 1.1). Excluded from these areas types. But tropical forests face distinct issues but sometimes included in the discussion are and challenges. They are home to most of the xeric shrublands (such as in southwest Mada- world's poor forest dwellers and contain the gascar or the Brazilian caatinga) and mon- bulk of its forest-based biodiversity, and it is tane grasslands and shrublands (as in the Map 1.1 This Report's Focus: Tropical Forests and Savanna Woodlands Biomes Tropical forest biome Tropical savanna biome Source: Authors' mapping of data from WWF 2001. 28 F o r E S T S D I F F E r Box 1.1 (continued) highest elevations of the Andes). The report to concerns in civil society and government, devotes relatively little attention to plantation Brazil has extensive experience grappling with management. land and forest use regulation--experience The report often uses data and examples from which other countries can learn. Finally, from Brazil, for several reasons. Brazil con- Brazil has superb statistics on its popula- tains a large portion of the world's tropical for- tion, economic conditions, and deforestation. ests, including disparate types: remote, dense Indeed, it is partly because of the Brazilian rainforests; savannas; and highly fragmented, government's bold decision to regularly moni- long-settled forests of extreme biodiversity. It tor and publicize deforestation rates that the also has a diverse array of forest actors: indig- world pays more attention to deforestation in enous people, subsistence-oriented small- Brazilian Amazônia than in other, less trans- holders, extractivists, agrobusinesses. Due parent regions. 29 AT L o g g E r H E A D S ? Human pressure on forests is shaped by their market access, suitability for farming, and tenure security. Moreover, these three factors tend to be intertwined, for reasons explored in more depth later. These issues motivate a stylized, three-part typology. · Forest-agriculture mosaiclands with better-defined tenure are settled agricultural areas interspersed with woods, often close to urban centers. They have rela- tively high population densities and depleted, frag- mented forests. These areas account for a minority of the world's forest estate but contain a substantial por- tion of its forest dwellers, a large share of threatened species, and the bulk of locally valued forest services. Consequently, they are where interactions between people and trees are most intense, and where trees are important sources of incomes and environmental ser- vices. Here the potential for both poverty reduction and environmental protection is great--but so is the poten- tial for trade-offs between these goals. · Frontier and disputed areas often suffer from conflicts over land and forest resources. In Latin America and parts of Africa these are places where waves of agri- cultural expansion are crashing on a broad front of relatively undisturbed forest. In parts of Indonesia large timber and plantation interests, small commercial farm- ers, and long-time residents are battling for control of forest areas. Elsewhere in Asia and in parts of Africa, ineffectual government control of nationalized forests creates a tragedy of the commons where communities, forest services, and fuelwood extractors dispute wood- land control and create degraded landscapes. · Areas beyond the agricultural frontier are outside the reach of most agricultural markets, though not beyond human influence. These include the last great expanses of tropical forest: the Amazon and Congo basins, and some scattered smaller areas. A minority of forest dwellers live here, but they include indigenous popula- tions and some of the world's poorest people. These types and the challenges facing each are summarized in table 1.1. These categories are meant to help organize thought; 30 F o r E S T S D I F F E r Table 1.1 Forest Types and Their Challenges Poverty and development Environmental Governance Type of area Features challenge challenges challenges Mosaiclands with High land value; Managing landscapes for production Agreeing on, better-defined tenure contain many of and environmental services; committing to, the world's forest preventing extinctions of threatened and enforcing dwellers but a species; mitigating carbon dioxide property rights small fraction of (CO2) emissions; fostering carbon over land, trees, the forest sequestration and environmental services Frontier and disputed Agricultural Fostering more Avoiding Restraining areas expansion; rapidly intensive rural irreversible resource grabs increasing land development and degradation; by large actors; values in frontiers; access to off-farm mitigating CO2 averting races for conflicts over employment emissions; property rights forest use in avoiding forest by smallholders; disputed areas fragmentation equitably adjudicating land claims Areas beyond the Most of the Providing services Maintaining Protecting agricultural frontier world's tropical for dispersed large-scale indigenous people's forests; contains a populations environmental rights; averting minority of forest processes disorderly frontier dwellers but many expansion indigenous people they are not intended to be rigid pigeonholes into which every for- est neatly fits. rudel (2005) presents a similar but simpler typology, focusing on the distinction between large and small forests--the unbroken blocs of the Amazon and Congo basins and Indonesian forests relative to the fragmented forests of other regions. rudel also describes deforestation processes by continent--a valuable resource when contemplating forest types. From Stylized Types to Mapped Domains It is impossible to map the three stylized forest types because geo- graphic data on forest tenure are weak. But it is possible to map some rough proxies (maps 1.2­1.6). Table 1.2 shows the correspon- 31 AT L o g g E r H E A D S ? Map 1.2 Domains in Africa's Tropical Forest Biomes Domains Mosaiclands Forest edge Forest core Source: Authors' calculations; see appendix B. Table 1.2 Stylized Forest Types Have Equivalents in Mapped Domains Stylized type Mapped domain Mosaiclands with better Mosaiclands: agricultural lands, defined tenure agriculture-forest mosaics, and small forest patches Frontier and disputed areas Forest (and savanna) edges: the forested borders of mosaiclands Areas beyond the Forest (and savanna) cores: forested agricultural frontier areas well away from mosaiclands 32 F o r E S T S D I F F E r dence between the idealized types and their mapped equivalents-- called domains here. Box 1.2 and Appendix B provide more details on the construction of the domains. The maps, especially those for the "true" forest, capture much of the spirit of the typology. Map 1.5, for instance, clearly shows the mosaiclands that characterize much of Central America and the South American Atlantic Forest. Its depiction of forest edges high- lights the expansion of the Amazônian frontier from the south and east, and along roads and rivers. Map 1.4 shows forests remaining beyond the frontier on the islands of Borneo and New guinea, but Map 1.3 Domains in Africa's Tropical Savanna Biomes Domains Mosaiclands Savanna edge Savanna core Source: Authors' calculations; see appendix B. 33 AT L o g g E r H E A D S ? Map 1.4 Domains in Asia's Tropical Forest Biomes Domains Mosaiclands Forest edge Forest core Source: Authors' calculations; see appendix B. mostly already transformed to frontier in Sumatra and Sulawesi, Indonesia. The correspondence between stylized types and mapped domains, however, is inexact. Nonremote forest edges could be akin to the mosaiclands they border. In long-settled parts of Asia these edges may be static interstices between scattered settlements. Some of these interstitial forests may be under disputed or insecure tenure. other forest edges are near cities and have high population densities. These may represent agroforestry rather than frontiers or disputed areas. Thus the maps and tables in this chapter should be considered indicative of broad tendencies, and are not meant to give a definitive classification to any particular spot. They are intended to inspire more detailed work at the national level. 34 F o r E S T S D I F F E r Map 1.5 Domains in Latin American and Caribbean Tropical Forest Biomes Domains Mosaiclands Forest edge Forest core Source: Authors' calculations; see appendix B. 35 AT L o g g E r H E A D S ? Map 1.6 Domains in Latin American and Caribbean Tropical Savanna Biomes Domains Mosaiclands Savanna edge Savanna core Source: Authors' calculations; see appendix B. 36 F o r E S T S D I F F E r Box 1.2 Mapping the Domains and Tallying Their Populations To map tropical forest domains, this report household extractive activities or shifting cul- uses gLC2000, a global map of land cover tivation around a small settlement. Finally, for- based on satellite data (ECJrC 2003). The est cores (a proxy for areas beyond the frontier) gLC2000 maps the world in 1-square kilome- are more than 6 kilometers from mosaiclands. ter cells based on predominant vegetation. The populations of these cells were esti- Although the gLC2000 is the best available mated by overlaying them with data from the map of global land cover, it has limitations. global rural-Urban Mapping Project (CIESIN remote sensing data at this resolution has dif- and others 2004a, b, c). This data set compiled ficulty distinguishing agroforestry from forests, global census data at a fine administrative and relatively undisturbed savanna woodlands level--the equivalent of districts or smaller from forest-agriculture mosaiclands. units, where possible. residents of towns and Mosaiclands are mostly agriculture, mix- cities were distinguished from rural dwell- tures of forest and agriculture, and small ers, who were then assumed to be spread out clumps of forest surrounded by agriculture. So evenly across the rest of the census district this domain consists of mosaic forests embed- and mapped into 1-square kilometer cells. ded in a sea of agricultural lands. The assumption of even distribution prob- Forest edges (a proxy for frontier areas) are ably results in an overestimate of the number forests and woodlands outside mosaiclands, of people living in forests, because population but within 6 kilometers of them. This defini- densities are higher in cleared areas than in tion might roughly approximate the radius of forests. The Uneven Distribution of Forest Populations In the great forest areas beyond the frontier, people live at densities of less than 1 per 10 square kilometers--while some Asian forests have population densities hundreds of times that. options for for- estry and agriculture are very different in these settings. So is the environmental relationship between people and trees. This section, and much of this report, elaborates on an obvious but crucial dis- tinction: most forest dwellers live in a small part of the forest, and most of the forest has few inhabitants. It is important to understand the limitations of the data used to map forest populations (see box 1.2). The data assume that rural populations are evenly spread throughout their counties, districts, or municipios. We then tally the number of people in forested areas. This approach results in overestimates, because people tend to 37 AT L o g g E r H E A D S ? occupy unforested areas. Still, the data roughly indicate the number of people living in or around the three domains. The estimates are least reliable for African savannas, where remote sensing sources have difficulty distinguishing small-scale agriculture from savanna vegetation. The resulting estimates provide guidance on the relative size of the three domains (table 1.3). These tabulations refer only to dis- persed rural populations, excluding even small towns with a few thousand people. Some key findings: · Forest dwellers outnumber dwellers on purely agricul- tural lands. In tropical regions more than 800 million rural people live in or near forests (considered at a fine geographic scale), while about 460 million live on lands that are mostly agricultural. Even keeping in mind that small towns are excluded from this cal- culation, and that forest dwellers are overstated, this estimate highlights the importance of woodlands in the populated rural landscape. · Latin America and the Caribbean have the most forest area, Africa the most savanna. The area of nonsavanna forest in Latin America--about 10 million square kilo- meters--exceeds that in Africa and Asia combined. The savanna woodlands of Africa occupy about 11 million square kilometers. · Asia has by far the most forest dwellers and the highest population density in forests. About 350 million people live in the forest edges and cores of Asia, with another 90 million in the Asian mosaic forests. The population densities of less remote Asian edge forests, at about 85 per square kilometer, are high enough to suggest long-settled areas dependent on agroforestry or planted forests. While most of these areas are unlikely to be expansion frontiers, some could be disputed. Africa and Latin America have about 165 million people liv- ing in nonsavanna forests, and Africa has 185 million people in savanna woodlands. · Forest edges contain most forest dwellers. In the nonsa- vanna forests of Asia and Latin America, forest edge 38 F o r E S T S D I F F E r Table 1.3 Forest Populations and Areas Vary by Continent, Biome, Domain, and Remoteness, 2000 Population (millions) Mosaiclands Forest edges Forest cores Agricultural Mosaic lands forest Hours to major Hours to major Hours to major Hours to major city city city city Continent Biome < 8 > 8 < 8 > 8 < 8 > 8 < 8 > 8 Africa Forests 13.2 2.9 25.5 3.6 22.6 7.9 18.3 12.0 Savannas 55.4 6.9 28.5 3.6 54.3 11.6 58.9 28.8 Asia Forests 324.1 12.6 71.5 18.6 256.5 29.5 60.9 6.1 Savannas 4.7 0.0 0.2 0.0 1.8 0.1 0.0 0.0 Latin America and Caribbean Forests 31.2 3.2 18.2 1.8 34.8 7.2 7.5 3.9 Savannas 5.2 0.5 2.8 0.3 4.1 1.0 0.7 0.3 All Forests 368.5 18.6 115.2 24.0 313.9 44.7 86.7 22.0 Savannas 65.3 7.4 31.5 4.0 60.2 12.7 59.5 29.0 Area (thousands of square kilometers) Mosaiclands Forest edges Forest cores Agricultural Mosaic lands forest Hours to major Hours to major Hours to major Hours to major city city city city Continent Biome < 8 > 8 < 8 > 8 < 8 > 8 < 8 > 8 Africa Forests 114 54 440 130 480 466 693 1,206 Savannas 1,189 480 778 284 1,446 1,012 3,024 4,307 Asia Forests 1,684 169 636 517 2,045 1,527 594 563 Savannas 15 1 1 13 8 12 0 3 Latin America and Caribbean Forests 993 222 922 331 1,622 1,947 647 4,458 Savannas 566 257 324 170 749 636 259 323 All Forests 2,792 445 1,998 978 4,148 3,941 1,934 6,226 Savannas 1,770 737 1,104 467 2,203 1,660 3,283 4,633 Source: Authors' calculations based on CIESIN 2004a, b, c and ECJRC 2003; see appendix B. 39 AT L o g g E r H E A D S ? populations outnumber core and mosaic forest popula- tions combined. Africa's nonsavanna forest populations are about equally split between the three domains. · Remote forest dwellers are relatively few. globally, about 22 million people live in nonsavanna forests more than 6 kilometers from the nearest agriculture and more than 8 hours' travel from the nearest city of 100,000 people. An additional 45 million live at similar remote- ness, but within 6 kilometers of agriculture. Figure 1.1 shows how unevenly people are distributed through- out forests. The middle panel shows the cumulative number of for- est dwellers in Africa, Asia, and Latin America and the Caribbean as population densities increase, up to 500 per square kilometer. As noted, densities at this level include people who are near forests but not necessarily under forest canopy. An enormous share of these forest dwellers are Asian. The top of figure 1.1 zooms on the most sparsely populated forests, with population densities below 50 per square kilometer. Dwellers in these forests are presumably highly dependent on for- est resources and include shifting cultivators. Most of the 20 million inhabitants of the lowest-density forests (those with fewer than 10 people per square kilometer) are African and Latin American. The bottom panel of figure 1.1 shows the cumulative area of forest as a function of population density. Most African and Latin American forests (excluding savannas) are occupied at densities of fewer than 20 people per square kilometer. In contrast, most Asian tropical forest is occupied at population densities above 50. These widely varying population densities have divergent implications for livelihoods and forest management. Trends in Forest Change With all the attention devoted to loss of tropical forests, one might think that the basic dimensions of the problem--the rate, location, and nature of forest degradation--would be well quantified and understood. They are not. only a few countries, notably Brazil and India, are measuring and reporting forest status on a regular basis using remote sensing. Some countries, such as Mexico, have under- taken detailed forest inventories that permit assessment of changes in land use over specific periods. For many other important forest 40 F o r E S T S D I F F E r Figure 1.1 Forests Vary Greatly in Population Density, 2000 A 100 90 80 Asia 70 (millions) 60 50 population 40 Latin America 30 and Caribbean 20 Cumulative Africa 10 0 0 5 10 15 20 25 30 35 40 45 50 Population density (people per sq. km.) 500 450 Asia 400 (millions) 350 300 250 population 200 150 A Cumulative 100 Latin America and Caribbean 50 Africa 0 0 50 100 150 200 250 300 350 400 450 500 Population density (people per sq. km.) 18 16 km.) Asia 14 sq. of 12 10 Latin America (millions and Caribbean 8 area 6 4 Cumulative 2 Africa 0 0 50 100 150 200 250 300 350 400 450 500 Population density (people per sq. km.) Source: Authors' calculations; see appendix B. 41 Note: Excludes savannas. AT L o g g E r H E A D S ? countries, consistent and comprehensive data on forest change do not exist. And overexploitation of forest plants and animals, a seri- ous ecological threat, is hard to monitor anywhere. Studies on Tropical Deforestation There have been attempts to estimate tropical forest loss on a global scale. Each has advantages and disadvantages. The most authorita- tive source is the United Nations Food and Agriculture organization (FAo 2001b) Forest resources Assessment (FrA), which produces two sets of estimates. The first, oft-cited set is compiled from national inventories. While the quality of national reports increased substan- tially between 1990 and 2000, estimates of deforestation rates are affected by low-quality, incompatible, or missing inventories from 1990 or earlier and by inconsistencies between countries. The less-known FrA remote sensing survey (FrA-rSS) esti- mates deforestation using high-resolution (30-meter pixels) satellite data covering 10 percent of the world's tropical forests. The results are representative at the continental but not national level. The FrA-rSS also tries to distinguish between kinds of forest degrada- tion. The Tropical Ecosystem Environment observation by Satellite (TrEES) project uses high-resolution sample data to estimate forest loss in humid tropical forests. Achard and others (2004) combined FrA and TrEES data to produce the estimates of forest loss shown in table 1.4. Table 1.4 Estimated Annual Deforestation Is Highest in Latin America and Asia, 1990­97 (millions of hectares) Humid forests Dry forests Latin Latin Type of America and America forest Caribbean Brazilian and change except Brazil Amazônia Africa Asia Caribbean Africa Deforestation 1.08 ± 0.55 1.43 ± 0.88 0.85 ± 0.30 2.84 ± 0.90 1.9 ± 1.1 1.5 ± 0.6 Other degradation 0.61 ± 0.46 0.22 ± 0.21 0.39 ± 0.19 1.07 ± 0.44 n.s. n.s. Regrowth 0.20 ± 0.11 0.08 ± 0.11 0.14 ± 0.11 0.53 ± 0.25 n.s. 0.07 ± 0.05 Source: Achard and others 2004. n.s. Not significant. 42 F o r E S T S D I F F E r In contrast, Hansen and DeFries (2004) used globally compre- hensive but low-resolution (8-kilometer pixels) data to map defores- tation for the entire tropical realm (map 1.7). In addition to its more comprehensive coverage, their study assesses changes over 17 years instead of the 7 years of Achard and others (2004). But its methodol- ogy is less accurate for the dry forests and savannas of Africa and may not be able to detect subtle, small-scale patterns of degradation in forests and mosaiclands. There is rough agreement between these two sources on the magnitude of gross deforestation on two continents during the 1990s: about 4.4 million hectares a year in Latin America and 2.8 million in Asia. FrA-rSS and TrEES also detect substantial forest degradation: about another 1 million hectares a year for each of those continents. DeFries and others (2002) report much more forest regrowth than do the other sources. And the greatest disagreements are for the dry forests and savannas of Africa. The FrA country studies report 5.2 million hectares of annual net forest loss in tropical Africa, while DeFries and others (2002) estimate a net annual loss of just 376,000 hectares. There is even more uncertainty about the actors responsible for deforestation. What are the roles of loggers, shifting agricultural- ists, sedentary farmers, large-scale ranchers, and plantation own- ers? Because tropical forests are species-rich, loggers are selective in removing trees and rarely engage in clear-cutting. of the 152 cases of deforestation reviewed by geist and Lambin (2001), only 5 involved logging without follow-on agriculture. However, logging provides access roads to follow-on settlers, and log sales can help finance the cost of clearing remaining trees and preparing land for planting of crops or pasture. Thus logging can catalyze deforestation. Moreover, logging can seriously degrade forests. In Indonesia and Malaysia, which have a high density of saleable trees, extrac- tion rates of 150 cubic meters per hectare have been reported (Putz and others 2001). Using new remote sensing techniques, Asner and others (2005) found that the area degraded by logging exceeded the clear-cut area in five Brazilian states. While much of the logged area might eventually be converted to agriculture, only about 19 percent of degraded areas were deforested within three years after logging. In Africa, forest loss around large cities reflects agricultural expansion and overharvesting for charcoal. The Stockholm Envi- ronment Institute (2002) used remote sensing to study forest deg- radation around Dar es Salaam (Tanzania), Lusaka (Zambia), and 43 AT L o g g E r H E A D S ? Map 1.7 Hotspots of Tropical Deforestation FRA-RSS survey scenes Deforestation hotspots Source: Hansen and DeFries 2004 (hotspots); FAO 2001b (survey scenes). Maputo (Mozambique) during the 1990s. In all three areas there was a reduction in open woodlands and an increase in thicket, bush, and grasslands, due to a combination of charcoal extraction and shrink- ing fallow periods. Around Lusaka nearly a third of deforestation was attributable solely to charcoal. The relative importance of small- and large-scale agriculturalists is also debated. A lot of Brazilian and Indonesian deforestation is undertaken by large commercial interests (see chapter 3), and most African and mainland Southeast Asian deforestation is thought to be carried out by smallholders. The only hints to the relative global magnitudes of corporate and "populist" deforestation are provided by the FrA-rSS. Using this data, FAo (2001b) estimated that expan- sion of shifting cultivation into undisturbed forest represented only about 5 percent of all pan-tropical changes in land use. Intensifi- cation of agriculture in shifting cultivation areas represented more than 20 percent of tropical land use change in Asia and less than 10 percent in Africa. Direct conversion of forest area to small-scale permanent agriculture accounted for 60 percent of land use change in Africa, but only a small portion elsewhere. Direct conversion of forest to large-scale permanent agriculture represented about 45 per- cent of tropical land use change in Latin America and about 30 per- cent in Asia. 44 F o r E S T S D I F F E r Where Are Forest Degradation Rates the Highest? Deforestation doesn't occur randomly. Assessing geographic risk factors for deforestation can help in the design of policies that pro- tect threatened forests with high environmental value but little to offer for agriculture. Table 1.5, based on an original analysis using the FrA-rSS, relates forest degradation rates (including both deforestation and Table 1.5 During the 1990s Savannas and Asian Forests Experienced Considerable Degradation (percent) Tropical forests Tropical savannas Latin America Latin America and the and the Domain Africa Asia Caribbean Africa Caribbean Mosaiclands 11.1 16.8 20.2 11.8 18.4 Forest edges 4.7 9.9 4.3 9.2 8.5 Forest cores 2.7 4.4 0.6 9.6 0.8 Total 5.4 10.9 3.6 9.9 10.8 Source: Authors' calculations based on CIESIN and others 2004a, b, c, ECJRC 2003, and FRA-RSS; see appendix B. Note: The table shows the percentage of forested 2- by 2-kilometer cells, by condition in 2000, that experienced a reduction in forest cover since 1990. 45 AT L o g g E r H E A D S ? forest thinning) to the three forest domains.1 Two features stand out. First, degradation rates are high--about 10 percent--in savan- nas and Asian forests. Second, degradation rates are quite high in mosaiclands. In Latin American forests, for instance, about 20 percent of mosaicland areas with trees experienced degrada- tion in the 1990s.2 This includes degradation of forest fragments on largely agricultural lands. rates are lower on the frontier and lower still on areas beyond the frontier. Because there is relatively little forest left within mosaiclands, the total area degraded is as large or larger outside mosaiclands, even though the degradation rate is lower. Does deforestation yield valuable agricultural land? or does it occur on marginal soils with poor prospects for sustainability? Figure 1.2 shows forest degradation rates by suitability for rainfed annual cropping, according to the global Agro-Ecological Zones assessment (FAo and IIASA 2000). In Africa and Latin America degradation Figure 1.2 Africa and Latin America Have Higher Degradation on Better Soils, 1990­2000 0.15 Asia 0.12 degradation forest 0.09 Latin America with 0.06 and Caribbean Africa cells of 0.03 Proportion0.00 Very High Good Medium Moderate Marginal Very Not High Marginal Suitable Soil suitability Source: Authors' calculations based on FRA-RSS and GAEZ; see appendix B. Note: The figure shows the proportion of forested 2-by-2-kilometer cells experiencing loss of forest cover. 46 F o r E S T S D I F F E r rates are higher on better soils, as a simple economic model would predict. This portends significant costs to restraining deforestation in some places. But a lot of forest degradation is occurring on lands considered marginal for annual agriculture. In Asia there is no rela- tionship between agroclimatic conditions and degradation rates. It is possible that in these areas, forest degradation is driven by logging or by conversion to perennials rather than annual cropping. In all three continents this analysis suggests that a substantial amount of deforestation and degradation occurs in areas with little agricultural value--suggesting that it may be possible to reduce some deforesta- tion at relatively low cost. Threatened Species--Concentrated in Less-remote Areas and Mosaiclands To support conservation, it is essential to identify and locate species threatened with extinction. The World Conservation Union's red List is a systematic effort at identifying those species (www.redlist. org). The list classifies species as endangered (extinction probability of 20 percent within 20 years) or critically endangered (extinction probability of 50 percent within 10 years) based on several criteria, including limited or declining ranges or populations. So it is not sur- prising that threatened species are more likely to be found in non- remote areas with higher human populations and more fragmented habitats. The World Conservation Union and other conservation groups are also making massive efforts to map the ranges of threatened and nonthreatened species, but are hampered by spotty data. observa- tions may be outdated. observers may have favored locations near roads or parks and neglect to look in mosaiclands. And imputed ranges may not take into account actual habitat conditions. Still, the creation of these range maps is a major step forward and provides at least a rough look at the distribution of species. Information from the red List and the global Amphibian Assessment made it possible to map the location of all tropical for- est biome cells with at least one threatened amphibian species. The incidence of threatened species is much higher in nonremote areas, at least in Africa and Latin America (figure 1.3). ricketts and others (2005) apply a stricter criterion to a broader set of species, includ- 47 AT L o g g E r H E A D S ? Figure 1.3 The Incidence of Threatened Amphibian Species Is Much Higher in Nonremote Areas 0.25 Latin America Asia and Caribbean 0.20 0.15 (proportion) Africa 0.10 Incidence 0.05 0.00 <1 1­2 2­3 3­5 5­8 8­11 11­16 16­26 >26 Remoteness (hours to city) Source: Authors' calculations based on World Conservation Union Red List and Global Amphibian Assessment; see appendix B. Map 1.8 Imminent Extinction Hotspots Imminent extinction hotspots Source: Updated from Ricketts and others 2005 using data from Alliance for Zero Extinction. 48 F o r E S T S D I F F E r Figure 1.4 Imminent Extinction Sites Are Concentrated Near Cities 6 Latin America 5 and Caribbean 4 3 Africa Incidence 2 1 Asia 0 <1 1­2 2­3 3­5 5­8 8­11 11­16 16­26 >26 Remoteness (hours to city) Source: Authors' calculations based on Alliance for Zero Extinction (data set v 2.1). Note: For each remoteness category, the figure shows the number of imminent ex- tinction sites per 100,000 square kilometers. 49 AT L o g g E r H E A D S ? ing birds, mammals, amphibians, some reptiles, and conifers. The authors focus on threatened species found in only one, well-defined location in the world, where the loss of habitat in that location prob- ably implies loss of the species. These locations--dubbed "immi- nent extinction" locations--tend disproportionately to be on islands and mountaintops, partly because of evolutionary processes (map 1.8). But again, outside Asia the incidence is closely related to urban proximity (figure 1.4). The incidence is much higher in agricultural and forest frontier land than on areas beyond the frontier. Summary The world's tropical forests are diverse, with important differences between continents. Asia has most of the world's tropical forest population, living at relatively high population densities and closer to major cities on average than in Latin America and Africa. Asia's deforestation rates are also higher. Latin America has most of the nonsavanna tropical forest area, but only a small proportion of for- est inhabitants. Africa has vast areas of savanna. Populations in nonsavanna forests are quite concentrated, with more than 400 million people occupying the most crowded 2 million square kilometers, and only about 10 million in the least crowded 10 million square kilometers. About a quarter of forest dwellers are in mosaiclands, where agriculture and forests are closely intermin- gled, though these lands constitute only a small portion of the forest estate. The deforestation rate is particularly high on these frag- mented lands. Most forest dwellers live on the edges of these mosa- iclands, some in frontier conditions. Biodiversity is most threatened in less-remote mosaic and edge forests. relatively few people live in core forests or those beyond the frontier. The next chapter examines the geographic and economic forces that shape these patterns. Endnotes 1. The FrA-rSS estimates are based on a stratified random sample of satellite scenes; sampled areas are represented by the blue rectangles in map 1.7. Note that there are no sample scenes in China, where the map shows deforestation hotspots. Degradation incidence is the proportion of tree-bearing sample points that expe- rience some loss of canopy. 50 F o r E S T S D I F F E r 2. This calculation is somewhat biased because the mosaicland classification is based on land cover in 2000, at the end of the period over which change was measured. But the results are qualitatively similar when the outermost cells of the mosaiclands--those in con- tact with the frontier and thus the most likely to have undergone recent change--are omitted. 51 Roads often trigger forest conversion to agriculture. This road was opened in Kalimantan, Indonesia as part of a 1980s transmigration program. © Michael Nichols / National Geographic Image Collection. CHAPTER 2 Incentives and Constraints Shape Forest Outcomes P ut yourself in the place of a farmer. You have some forest, or are thinking about claiming some forest. Should you log it? If so, should you extract as much as you can now, or plan for sustainable harvesting over the coming decades? Or should you simply clear cut the forest and replace it with crops, pasture, or tree plantations? Your choices will be shaped by your constraints and abilities, the characteristics of the forest, your rights over it, and the wider social, economic, and political context. Your choices will affect your livelihood--as well as stream flows of your downhill neigh- bors and climates of people in distant lands. When your interests and other people's diverge, there could be a mediating role for public policy. Understanding landholders' behavior is essential to understand- ing how policies and context affect deforestation and forest poverty. Attempts to understand the effects of sweeping policies (such as structural adjustment) on sweeping outcomes (such as aggregate forest loss in a country) are doomed to inconclusiveness. Policy changes typically pull many economic and social levers--changing prices and wages; stimulating one sector, dampening another. Each lever could have a distinctive impact on deforestation and forest poverty, and those impacts might differ between regions. At the 53 At L O g g E r h E A d S ? national level these impacts might be difficult to disentangle. So this report's strategy is to try to understand how each potential lever works. this chapter, which draws heavily on Angelsen (2006), offers a simple but powerful model of land use decisions at a particular point in space and time.1 It then uses that model to examine how forest cover and poverty might evolve over time for entire regions. The View from the Forest Plot: Comparing the Returns to Forestry and Agriculture the International tropical timber Organization (2006, p. 46) describes the dilemma facing sustainable forest management: "alter- native land uses, which usually involve a much more intensive use of the land, are more profitable or provide quicker returns." how and why does this dilemma arise? Is Sustainable Forest Management Appealing to Landholders? Culture and experience may impel long-time forest dwellers to main- tain forest even if other land uses are potentially more lucrative. Shifting cultivators, for instance, have a long history of sustainable forest management, temporarily clearing small plots for agriculture and cycling over long periods through large tracts of forest. Some forest-owning Mexican communities harvest less than regulations permit or profits might dictate (Bray and others 2003). And cultures around the world protect sacred forest groves. Still, economics is likely to intrude on the decisions of most forestholders. there are few long-cycle shifting cultivators left in the world--reflecting rising population densities, accelerating fal- low cycles, and forests degrading into bush. Elsewhere, as markets approach, forestholders (or would-be forest claimants) balance returns from sustainable timber production against predatory extrac- tion, followed by agricultural conversion. though there are exceptions, sustainable timber management is often less lucrative than other options. Exceptional cases involve forests with precious woods, many saleable trees, fast-growing trees, or soils unsuited to agriculture. For instance, sustainable manage- ment of Indian teak forests is estimated to confer a land value of more than $5,000 a hectare in net present value (World Bank 2005, vol. II, p. 76).2 Coniferous forests in Mexico, where nearly all trees are commercially valuable, are another example. 54 I n C E n t I v E S A n d C O n S t r A I n t S S h A P E F O r E S t O U tC O M E S But in old-growth rainforests with diverse, slow-growing species, biological and financial considerations could push landholders away from sustainability. An analysis of logging economics at a Brazilian site by Boltz and others (2001) illustrates a general pattern.3 reduced impact logging could net $128 a hectare from an initial selective har- vest, leaving the residual forest in reasonably good condition. Left alone (without silvicultural treatment), the forest regenerates, but its value grows by just 2 percent a year--a bad investment. Another har- vest is possible in 30 years, but the present value of that harvest, eval- uated at a 20 percent discount rate (a reasonable approximation of the discount rate in many developing countries), is only $0.24 a hectare. Even low-return pasture or staple crops offer higher returns to land- holders. Of course society, with a lower discount rate and a demand for forest environmental services, may view things differently. Private Gains from Deforestation: Sometimes Minuscule, Sometimes Huge how big are the private gains to deforestation? Knowing this is essential to assessing the economic and political costs of encourag- ing sustainable forest management. the answer--not surprising, but important--is that these gains vary tremendously between places, technologies, and land use systems. Profits from deforestation range from near zero to thousands of dollars a hectare. Profits are the benefits to landholders from sales of timber and agricultural products, after costs of conversion and production, including labor. For smallholders dependent on unpaid family labor, this concept of profits can be interpreted as income above what fam- ily members could earn elsewhere. In other words, a strict measure of profit deducts the opportunity cost of family labor. the resulting measure of net profits per hectare is a convenient measure of the economic pressure for forest conversion--or of the opportunity cost of conservation. however, where labor markets are imperfect, work- ers and policymakers may consider labor absorption a benefit. So employment per hectare is another way of assessing the benefits of forest conversion. It is challenging to document the value of forested land in the tropics. In a few places, mostly in Latin America, markets provide a clear indicator of the profitability of land. In theory, prices for pas- ture or prepared fields in these areas should reflect the net present value of future revenue from farming, including expected gains from road construction and improvements in tenure security. 55 At L O g g E r h E A d S ? Elsewhere in the developing world, where land, labor, and prod- uct markets are thin, estimates of profitability come from farm stud- ies. the Consultative group on International Agricultural research's Alternatives to Slash and Burn (ASB) program has undertaken espe- cially rigorous measurements of economic benefits and environmen- tal impacts of forest conversion in Brazil, Cameroon, and Indonesia (tomich and others 2005). these measurements, along with other reported land values from forested areas, appear in table 2.1. Although the land values provide a useful benchmark, they typ- ically overstate the private gains to forest conversion, for two rea- sons. First, it is necessary to account for the upfront costs of clearing logged-over forest and preparing the land for crops or pasture. In Bolivia, for instance, the cost of clearing and pasture establishment averages $480 a hectare, defrayed only in part by after-tax timber revenues of $227 (Merry and others 2002).4 these upfront costs are factored into the ASB estimates, but they also need to be deducted from some of the others. Second, most analyses that compute net present values adopt a 10 percent discount rate, which is lower than typical private discount rates--especially among poor people. At a higher discount rate, the returns to conversion would fall substantially. In ninan and Sathya- palan (2005) increasing the assumed discount rate from 8 percent to 12 percent cut the net present value in half. naidoo and Adamo- wicz (2006) present evidence supporting a discount rate of 15­25 percent for Paraguay; gEF (2006) suggests that discount rates in the developing world are typically even higher. For these reasons the net present values reported in table 2.1 might be two or three times greater than landholder perceptions of returns to forest conversion. Some highlights from these studies: · In some places there are huge incentives to degrade or convert forest. In Cameroon oil palm and intensive cocoa cultivation has a net present value of more than $1,400 a hectare. In Brazil's cerrado (savanna) region, converting native woodlands to soy results in land worth over $3,000 a hectare. India offers extraordi- narily high values for land devoted to coffee cultivation in the Western ghats, a biodiversity hotspot. · In contrast, mean land values are just $400 a hectare in another hotspot, the Atlantic forest of Bahia, Bra- zil--one of the world's most important places for bio- 56 I n C E n t I v E S A n d C O n S t r A I n t S S h A P E F O r E S t O U tC O M E S Table 2.1 Land Values in Forested Areas Vary Enormously a. Studies Reporting Land Prices or NPV Price or net Land use, type, present value Study Location Year(s) or location (per hectare) Notes Batagoda and Sinharaja, 1995 Tea $4,281 NPV at 8 percent others (2000) Sri Lanka Timber potential $1,129 Chomitz and Bahia, Brazil 2000 Median land $400 Price others (2005b) value Davies and Bolivia 1992 Mechanized $1,500 NPV at 10 Abelson (1996) soybeans and percent maize Traditional farm $270 excl. coca Traditional farm $385 with coca FNP Goias, Brazil 2004 Cerrado $140­1,290 Price Consultoría & (various (savanna) Agroinformativos subregions) High-productivity $1,950­3,150 agricultural land Fundacão Brazil 2004 Pará $200 Price of pasture Getulio Vargas Rondonia $318 Grimes and Ecuador 1987­91 Cattle ranching $57­287 NPV at 5 percent others (1994) (Amazon region) Timber $189 NPV at 5 percent Agriculture <$500 NPV at 5 percent Land price $50­220 Price Howard and Costa Rica 1994 Cattle ranching NPV at 10 Valerio (1996) Atlantic $1,239 percent South $1,433 North $880 Bean crops South $2,716 North $2,163 Corn (Atlantic) $2,281 Kazianga and Cameroon 2001 Land at the $86 Price Masters (2005) frontier Kishor and Costa Rica 1989 Cattle ranching $1,319 NPV at 8 percent Constantino Clear felling $1,292 (without taxes (1993) Plantations $3,223 and subsidies; Managed forestry $854 includes timber revenue) (continued on next page) 57 At L O g g E r h E A d S ? Table 2.1a (continued) Price or net Land use, type, present value Study Location Year(s) or location (per hectare) Notes Merry and Bolivia Not Pasture $24­500 Price; range others (2002) available reflects accessibility Ninan and Ghats, India 2000 Coffee on farm $1,593 NPV at 10 Sathyapalan <2.5 acres percent; small (2005) 2.5­5.0 acres $1,819 farms more likely 5­10 acres $4,834 to be in forests >10 acres $8,280 Olschewski and Ecuador 2001 Grazing land Price Benitez (2005) North $150­500 Coast $400­1,000 Nearest Quito $800­2,000 Pinedo-Vasquez, Peruvian 1988­89 Swidden $1,627 NPV at 10 Zarin, and Jipp Amazon agriculture percent (1992) (rice, cassava, plantains, fallow) Ricker and Veracruz, Mexico 1998 Pasture $210­1,052 Price others (1999) Tomich and Brazil 1996 Pasture $2 NPV at 9 percent others (2005) Amazônia Tomich and Cameroon 1990s Food crop $283­623 NPV at 10 others (2005) Cocoa $424­1,409 percent Oil palm $722­1,458 Tomich and Sumatra, 1997 Rubber $1 NPV at 20 others (2005) Indonesia agroforestry percent Community forest $5 management Oil palm $114 Unsustainable $1,080 logging Wunder (2000) Ecuador 1994­96 Deforestation $1,721 NPV at 10 cycle (wood, percent; includes crops, cattle) initial timber revenue Yaron (2001) Cameroon 1997­98 Small farming $2,380­4,275 NPV at 10 Oil palm and ­$2,838 to $96 percent rubber Sustainable $45­470 timber production I n C E n t I v E S A n d C O n S t r A I n t S S h A P E F O r E S t O U tC O M E S Table 2.1 (continued) b. Studies Reporting Annual Net Returns Land use, type, Annual net returns Study Location Year(s) or location (per hectare) Naidoo and Uganda 1993­2001 Farming $114 Adamowicz (2005) Norton-Griffiths and Kenya 1989­93 High potential zone $151 Southey (1995) Medium potential $91 Per humid $38 Arable $54 Olschewski and Ecuador 2001 Cattle ranching Benitez (2005) North $25a Coast $42a Nearest Quito $110a Zelek and Shively Philippines 1994­96 Low-input maize $260 (2003) NPV stands for net present value. a. Returns are net of costs except labor. diversity conservation. Only small fragments of forest remain in this long-settled region. the study also finds that remaining forest sells at a steep discount relative to other land with similar characteristics. this disparity may reflect the effect of laws, even though imperfectly enforced, against deforestation. It may also reflect relegation of the poorest-quality land to forest; after decades of occupation, most agriculturally suitable land has already been cleared. Both effects may be present in other biodiversity hotspots where forests have been heavily fragmented. · At the Latin American frontier, forest is being converted to low-value uses that generate little employment. Conversion of forest to traditionally managed pasture in Amazônia yields pasture worth only a few hundred dollars a hectare. Pasture at the Ecuadorian frontier is worth $150­500 a hectare; at the Bolivian frontier, $24­500. After accounting for costs, ASB estimates that converting a hectare confers a net present value of only $2 and provides just 11 days of employment a year. But 59 At L O g g E r h E A d S ? values are much higher near cities and on well-man- aged farms using improved production systems. · Low-value land uses are also reported in Indonesia, Uganda, and the Cameroonian forest frontier. · Sustainable forest management typically provides lower returns and employment than does commercial agri- culture. In Sumatra, for instance, management for non- timber forest products employs 0.3 people a hectare per year and returns a net present value of just $5 a hect- are--while oil palm cultivation employs 108 people a hectare per year and returns $114 a hectare. Agricultural returns outstrip those from sustainable forest manage- ment in Cameroon, Costa rica, India, and Sri Lanka. In summary, there is great variation across pantropical forest margins in the strength of incentives for deforestation. Where condi- tions are amenable to crops such as soybeans, oil palm, or cocoa, and where old-growth timber is still standing, deforesters are rewarded with thousands of dollars a hectare. On marginal lands, lands far from markets, or where agricultural technologies are unavailable, there may be little incentive beyond the ability to eke out a living at the going wage. How Do Agroclimate, Prices, Technology, Tenure, and Other Factors Affect Deforestation and Income? this section considers how the environmental, social, and economic context of a forest plot affects the relative returns to forest mainte- nance and agriculture. the discussion here helps in understanding how policy levers affect outcomes in the forest domains described in chapter 1. table 2.2 summarizes the discussion. Richer Farmers Are Better Able to Finance Deforestation A poor household can't afford to clear much forest. In Bolivia clear- ance and land preparation costs range from $350­605 a hectare (Merry and others 2002); in Costa rica clearance costs $78 a hectare (howard and valerio 1996). Sometimes these costs can be partly or fully defrayed by sales of timber; sometimes wealthy interests are willing to finance clearing by smallholders on their behalf. Where these markets are lacking, successful deforesters must be able to mobilize a lot of family or community labor--50 to 70 person-days 60 I n C E n t I v E S A n d C O n S t r A I n t S S h A P E F O r E S t O U tC O M E S Table 2.2 Predictions of How Changes in Local Variables Will Affect the Environment and Welfare Effect on environment Effect on welfare ­ Promotes deforestation ­ Reduces welfare Element + Inhibits deforestation + Enhances welfare Access to credit ­ + markets; own assets Lower discount rates + with exceptions + Good soils, moderate ­ + rainfall Higher prices for ­ + extensive farm output Higher prices for + Where labor markets are imperfect, could + intensive farm output decrease deforestation by attracting labor away from extensive production ­ Where capital markets are imperfect, could increase deforestation by funding forest conversion Higher prices for ­ Spur deforestation of old-growth timber ± Effect on local poverty timber ­ Increase deforestation in open access areas depends on who extracts the + Encourage sustainable management of timber and wider economic secondary forests where there is secure tenure effects; poverty may increase + Spur reforestation in forest-poor areas if outsiders degrade forests on which locals depend Higher off-farm wages + Where labor markets are imperfect or in- + migration is limited, draw labor away from deforestation of marginal areas ­ Could fund deforestation Higher-yielding ­ If labor and capital can migrate to forest + (though indirect negative agricultural margins effects are possible) technologies + If marketwide effects lower prices + If technologies absorb labor and in-migration is limited More secure land + Reduces deforestation as a means of + tenure claiming land + Makes sustainable forest management more attractive ­ Makes investments in land improvements (including perennial crops) more attractive Road extension or ­ Increases farmgate prices of outputs, lowers + (unless outsiders displace improvement prices of inputs, makes in-migration more locals) attractive 61 At L O g g E r h E A d S ? a hectare--or to hire workers, chainsaws, and bulldozers. this point suggests that cash and credit constraints hamper smallholder deforestation. relaxing those constraints--through transfers, stron- ger credit markets, and better opportunities for off-season employ- ment--could increase both incomes and deforestation. Good Land Is Cleared First Soils, topography, and climate (agroclimate, for short) strongly affect land rents. differences in soils and climates explain most county-level variations in land values in Brazil, India, and the United States (Men- delsohn, dinar, and Sanghi 2001). In Bahia rural land prices increase with soil quality but decrease with slope, holding constant other characteristics such as road access (Chomitz and others 2005b). deforestation occurs more quickly on lands that offer higher rents. Studies of deforestation at the farm or local level generally find that deforestation rates are lower on hillsides, other things constant (appendix table A.1). these studies also find a strong correlation between soil quality and deforestation. In periurban areas of Latin America and Asia tree cover is about twice as high on the poorest soils as on the best soils for rainfed agriculture (see chapter 1). high densities of saleable trees can also promote deforesta- tion. roads built by loggers and revenue from timber sales can help finance agricultural clearing. If the density is high enough, extrac- tion can lead to deforestation even in the absence of agriculture. this is thought to be true in Southeast Asia, where lowland for- ests have high densities of valuable dipterocarp trees. For instance, logging is blamed for deforestation in sparsely populated, protected areas of Kalimantan, Indonesia (Curran and others 2004). deforestation skirts areas with high rainfall, which is inimical to cultivation of annual crops and discourages cattle ranching--espe- cially when there is no dry season. A study of Brazilian Amazô- nia by Chomitz and thomas (2003) found that, controlling for road access, higher rainfall is associated with lower deforestation, more land abandonment, and lower grazing densities. Higher Prices for Farm Output Induce Forest Conversion and Benefit Farmers Other things being equal, higher prices for crops and lower prices for farm inputs will spur faster deforestation. this prediction is important because many policies can affect farmgate prices, including taxes, tariffs, subsidies, road improvements, and exchange rate policies. 62 I n C E n t I v E S A n d C O n S t r A I n t S S h A P E F O r E S t O U tC O M E S the prediction can be tested by looking for variations in prices across the landscape within a country, between countries, or over time, and correlating prices with deforestation rates. doing so is dif- ficult. Within a country, at a single point in time, there may be little price variation. Comparisons between countries and over time are problematic because there are many other confounding influences, and because measurements of deforestation may be inconsistent. So there are only a few relevant studies. Most of these studies find a strong link between higher agricul- tural prices and more rapid or extensive deforestation. the degree of price sensitivity varies but tends to increase with more localized measurements. For instance, an analysis of remote sensing data shows that, after controlling for other factors, deforestation rates in Brazilian Amazônia are closely linked to farmgate prices of beef (fig- ure 2.1). this analysis focuses on unprotected lands (excluding land reform settlements) and shows the strong effect of rainfall levels and farmgate prices on deforestation rates. In areas with moderate rain- fall (less than 2,000 millimeters a year) near roads, deforestation over 2001­03 was 8 percent where the beef price was above r$600 a ton, 4 percent where the price was r$400­600, and 0.5 percent where the price was below r$400. In a study of Mexico, deininger and Minten (1999) examined the relationship between deforestation and proximity to buyers of Figure 2.1 Deforestation in Brazilian Amazônia Is Shaped by Rainfall and Farmgate Prices of Beef, 2001­03 10 7.97 % 8 price < R$400/t price R$400­600/t rate, 6 price > R$600/t 4.02 4 2.94 Deforestation 2 0.49 0.96 0.15 0 Rainfall < 2000 mm/year Rainfall 2000­2500 mm/year Source: Authors' calculations; see Appendix B. Note: Rate is deforested area/initial forest area. Excludes protected areas and land reform settlements. 63 At L O g g E r h E A d S ? maize--the main forest-competing crop. Because maize is bulky, closer proximity translates into lower transport costs and higher farmgate prices. the authors found that an increase of one standard deviation in buyer density corresponded to a 40 percent increase in the deforestation rate. Barbier and Cox (2004) examined mangrove deforestation (due to shrimp farming) in thai provinces and found that a 10 percent hike in shrimp prices would boost deforestation a modest 1.6 percent--while a similar hike in the price of ammonium phosphate (an input) would reduce deforestation by 4.5 percent. But not all studies find strong effects. gbetnkom (2005), for instance, finds that changes in prices of coffee, cocoa, and food have negli- gible effects on forest clearance in Cameroon. the impacts of price changes become more complicated when two other land uses compete with forest. Suppose that one use is exten- sive: long-fallow cultivation of a staple food (such as cacao, irrigated rice, or coca). Suppose that the other is much more intensive, using far more labor per hectare (say, shifting cultivation of maize, rice, or plantains). Suppose too that the labor supply is limited, and outsiders cannot easily move in to exploit new opportunities. then, theory says, an increase in the returns to the intensive land use could absorb labor from the extensive one, at least in the short to medium run. there is evidence that this happens. Coxhead and demeke (2004), in a study of upland farmers in the Philippines, find strong cross-effects between vegetable and maize production. An increase in the price of vegetables, the more intensive crop, is predicted to slightly reduce the total area under cultivation. higher prices for farm products benefit land owners and increase employment. So in general, higher prices for outputs and lower prices for inputs will reduce rural poverty--with two exceptions. First, because farmers with tiny plots might be net buyers of food, higher food prices will hurt them. Second, substitution between crops could indirectly hurt poor people. For instance, higher prices for beef or soy--which use relatively little labor--could divert land away from more intensive cultivation. Higher Timber Prices Put Pressure on Old-growth Forests but Create Incentives for New Ones do high timber values promote or undermine sustainable forest management? the answer depends on the state of the forest (von Amsberg 1998) and how it is regulated. new roads or new markets 64 I n C E n t I v E S A n d C O n S t r A I n t S S h A P E F O r E S t O U tC O M E S can confer enormous value on old-growth forests. Individual trees can be worth thousands of dollars. In the absence of regulation, rising prices induce loggers to sweep deeper into old-growth forests, mining sellable trees (Stone 1998). But where societies are willing and able to require forest own- ers to practice sustainable forest management, higher timber prices make such regulation less onerous. And where forests have already been depleted, higher timber prices make it more attractive to raise trees--especially plantations of fast-growing ones--as a crop. Higher Off-farm Wages Discourage Deforestation in Marginal Areas Many, though not all, forest dwellers have opportunities to earn wages. the opportunities may be on neighboring farms or planta- tions, in nearby market towns, or in distant cities. As these oppor- tunities become more lucrative, there is less incentive to use forest for subsistence or low-value crops. Kaimowitz and Angelsen (1998) found broad support for this proposition. A dramatic long-run example of this is the abandonment of the hillsides of Puerto rico. By 1950 almost all the island's hillside for- ests had been converted to coffee plantations or other agriculture, leaving only 9 percent of the island under forest. Subsequently, there was massive out-migration from the hillsides as people sought better-paying employment in San Juan and the United States. the result was regeneration of the deforested area: by 1990, 37 percent of the island was under forest (rudel, Perez-Lugo, and Zichal 2000; Lamb, Erskine, and Parrotta 2005b). Between 1994 and 2002 Coxhead and demeke (2004) observed a wage rise of about 50 percent among hillside farmers in the Philip- pines, as transportation and communications improved. According to their analysis, this increase would by itself reduce land cultivation by about 20 percent. But wage increases can also affect deforestation in other ways. Barbier and Cox (2004) found that higher wages were associated with higher clearance of mangroves for shrimp farming in thailand. they suggest that this was because shrimp growers, faced with higher wages, had ways of substituting land for labor. Wage increases can also increase the demand for fuelwood and food, spurring additional deforestation. Whatever their effects on deforestation, increases in off-farm wages are essential to poverty alleviation. A growing literature docu- 65 At L O g g E r h E A d S ? ments the potential role of off-farm employment in alleviating rural poverty (reardon, Berdegue, and Escobar 2001). Agricultural Technology Promotes Growth --With Ambiguous Implications for Deforestation technological improvements in agriculture are crucial to raising rural welfare (through higher farm incomes) and consumer welfare (through lower food prices). But the gains from these improvements may be unequally shared. And except in special circumstances, tech- nological improvements are likely to increase pressures on forest. to explain why, this section draws on Angelsen (2006) and Angelsen and Kaimowitz (2001). to be adopted, a technical innovation generally has to save a farmer's time or increase farm output. But any innovation that makes farming more profitable is likely to prompt the expansion of farms into forests or attract new farmers to the forest frontier. And anything that reduces labor requirements could release unemployed farmers to search for new frontiers. For instance, ruf (2001) claims that in Sulawesi, Indonesia, the introduction of herbicides and mechanical cultivators in lowland rice production released workers to engage in upland deforestation. Consider too the impact of improved soybean varieties in Bra- zil's cerrado (savanna) region. the region's poor soils and short days had been unsuitable for cultivating traditional soy varieties. So EMBrAPA, the Brazilian agricultural research agency, bred variet- ies adapted to the region. As a result soy cultivation exploded--at the expense of pasture, biodiversity-rich cerrado, and, recently, dense forests. the area cultivated jumped from nearly zero in 1970 (Warnken 1999) to 117,000 square kilometers in 2004 (IBgE 2006). Soybean and soy product exports were $9.8 billion in 2004 (Econo- mist Intelligence Unit 2005). For a technological innovation to simultaneously increase farmer welfare and reduce forest pressure, one of the following conditions must apply: · the innovation increases food production so much that food prices fall, easing pressure to convert forested uplands. this might happen in isolated locales cut off from markets. Or it might happen if the productivity increase is so large that it depresses national or even global markets. Some analysts think that the green 66 I n C E n t I v E S A n d C O n S t r A I n t S S h A P E F O r E S t O U tC O M E S revolution is an example, positing that improvements in irrigated rice yields reduced pressures for upland deforestation. · the innovation boosts the productivity of subsistence farmers not closely linked to food markets. this could reduce their need for clearing and might occur in areas beyond the frontier. · the innovation boosts both productivity and labor use per hectare. Moreover, labor supply is limited, either because of remoteness or because local residents have secure tenure over large amounts of land and prefer not to rent or sell to newcomers. In these condi- tions--more characteristic of frontier areas than mosa- iclands--some intensive farming systems could absorb labor away from extensive, more forest-damaging ones. holden (2001), Shively and Pagiola (2004), Shively and Martinez (2001), and Coxhead and demeke (2004) present examples of how expansion of intensified land use systems can draw labor away from extensive, deforesting land uses. It is uncertain, though, whether over the long run inflows of labor might counteract this effect. · the innovation stimulates nonfarm employment. returning to the example of soy in Brazil's cerrado, the direct beneficiaries were soybean farmers, includ- ing large and industrial growers. But related growth in services, transportation, and processing has contrib- uted to the rapid development of urban centers in the soybelt, and during the 1990s these cities accounted for substantial employment growth. however, the size of the link between soy expansion and urban employment has not been quantified. Tenure Is Good for Landholders, but Has Uncertain Effects for Deforestation Landholders with secure tenure are more likely to make physical improvements, invest in perennial crops, and plant and maintain forests. they worry less about defending their property and lives from thieves. they are better able to tap credit markets. And large 67 At L O g g E r h E A d S ? landholders with secure tenure are more inclined to rent out land to tenants or sharecroppers, rather than keeping it idle or under pas- ture. For all these reasons, tenure security boosts incomes of rural landowners and workers (deininger 2003). Poorly defined tenure is generally bad for people and forests. In many parts of the world, governments have nominal control of for- ests but are too weak to effectively regulate their use. this can lead to a tragedy of the commons where forest resources are degraded. the relationship between tenure and deforestation is more ambiguous. In frontier areas deforestation is a common way of lay- ing claim to land and securing tenure, in both practice and law. this setup encourages a destructive race for property rights at the fron- tier (Schneider 1995), where land is prematurely deforested--that is, before it generates any economic rent--in speculation that roads or government will eventually confer value on it. And in countries with pressure for land reform, large landholders will feel pressured to deforest just to demonstrate "productive use" of the land and so avoid invasion or expropriation. that has been especially common in Brazil, where uncertainty over land rights has led to violent fights over forested properties. But secure tenure does not guarantee that landowners will spare forests. As noted, landholders will likely first extract and sell large, mature, slow-growing trees. Landholders will then weigh the rela- tive advantages of forest maintenance and cropping. With secure tenure, investments in perennial crops such as black pepper or oil palm may be more attractive. Roads Provide the Path to Rural Development --and Forest Clearance Providing road access is the most effective determinant of defor- estation that is under policy control. the theoretical argument is strong: it says that road provision increases farmgate prices for out- puts and decreases farmgate prices for inputs, with all the effects just reviewed. Property-level studies of land values in nepal (Jacoby 2000) and the Atlantic forest area of Brazil (Chomitz and others 2005b) support this linkage. this means that improving access to a forest plot generally creates strong pressures to deforest it. the theory allows for exceptions. In rural areas where tenure is strong and immigration is limited, better road access might allow 68 I n C E n t I v E S A n d C O n S t r A I n t S S h A P E F O r E S t O U tC O M E S residents to work in towns, or shift them from extensive production of subsistence crops to more intensive production of commercial crops. deforestation might then fall as long as residents can and will exclude in-migrants. road links to nearby towns might boost local wages more than farmgate prices, attracting farmers away from mar- ginal lands. And where forests are already exhausted, better road access could trigger tree planting for poles, firewood, and timber. But an extensive empirical literature strongly supports the prop- osition that roads tend to promote, rather than inhibit, deforestation. A major challenge for this literature is determining causality when road development and deforestation occur together. did the roads facilitate deforestation? Or were they built in response to settlement that would have occurred in their absence? One approach to answering these questions is through case studies of deforestation (for example, Arima and others 2005). One analysis of 152 case studies finds that road access was a driver of deforestation in 93 cases (geist and Lambin 2001), and another metareview concurs on the importance of road access (Kaimowitz and Angelsen 1998). Another approach uses spatial econometric analysis to relate the incidence of deforestation to road proximity. Investigators com- pare small geographical regions, or even individual points on the landscape, in order to account for confounding factors such as soil fertility, climate, slope, or elevation. this helps to control for the possibility that roads are a symptom rather than cause of deforesta- tion. this report reviewed 33 such studies, most of them at the finest level of geographic analysis (appendix table A.1). twenty-one found a statistically significant, positive relationship between road proxim- ity and deforestation. Eight found complex or ambiguous patterns, for instance when several measures of remoteness were used, or when there were differential effects on different groups. the remain- der were inconclusive. road access also facilitates hunting of large mammals. In cen- tral African forests this is a more severe environmental threat than deforestation, and a study in gabon found fewer mammals near roads (Laurance and others 2006). rural roads are generally believed to raise rural incomes and alleviate poverty, for the same reasons our model suggests they pro- mote deforestation: by raising farmgate prices, lowering prices of 69 At L O g g E r h E A d S ? urban manufactured goods, and promoting more intensive demand for labor. rural roads also facilitate access to nonfarm employment in towns, which is often crucial to alleviating poverty in rural areas. For these reasons rural road provision is a mainstay of rural devel- opment strategies. Considering the importance of rural roads to development strate- gies, the literature on their impact is thin. this report reviewed 26 studies and two metareviews covering 56 others. though they were almost unanimous in finding positive impacts, the magnitude of the impacts varies greatly. Few of the studies used rigorous, quasi-exper- imental evaluations of how roads affect income and welfare. One of the most rigorous evaluations compared Peruvian villages that had received rehabilitated road links with similar control villages (Insti- tuto Cuanto 2005). After five years, male (but not female) wages in the villages with rehabilitated roads rose by 20 percent relative to the control villages. In subsequent hard economic times, poverty in the control villages increased by 4­6 percentage points more than in the villages with rehabilitated roads. two recent simulations are of particular interest because they examine countries with extensive forest cover. In Papua new guinea a study assessed the impact of reducing to three hours the access time to a road of all households that required more time (gibson and rozelle 2003). this potentially expensive undertak- ing would cut the number of poor people by 12 percent. the other study found that providing all-season roads to the 50 percent of Laotians lacking them would release 5 percent of the population from poverty (Warr 2005). Other studies involve econometric analysis of district or pro- vincial data, attempting to control for other potentially confound- ing factors. Fan and Chan-Kang (2004) summarize some of these studies, reporting astounding returns to road investment--hundreds of percentage points--in China, Uganda, and rain-fed regions of India. Other reported returns are far more modest, but still positive (appendix table A.2). the inconsistent relationship between rural roads and poverty alleviation reflects a variety of factors. First, it may be modulated by other policies and conditions. Finan, Sadoulet, and de Janvry (2005) find that rural Mexicans with both road access and primary educa- tion earn about 10 times more from an extra hectare of land than do those without either asset. Second, where immigration is possible, roads may cause an increase in workers rather than wages. 70 I n C E n t I v E S A n d C O n S t r A I n t S S h A P E F O r E S t O U tC O M E S Forest Trajectories: Roads, Markets, and Rights Shape Outcomes for Environment and Income Astronomers teach us that the farther into space we peer, the farther back in time we see. So too, when we stand in an urban center and look toward the remote forest frontier, we see not only a changing spa- tial pattern of forests on today's landscape, but also a history of how that landscape evolved. Seen from the other direction, conditions near contemporary towns--old frontiers--provide hints about the future prospects of today's frontier regions. this section builds on our under- standing of landholders' behavior, expanding from a single plot to an entire landscape, and from a snapshot to an evolving pattern. From Urban Center to Forest Frontier: A Stylized View of the Landscape Let's first take a stylized journey from an urban center to a forest frontier, at a moment in time. Our guide is Johann heinrich von thünen, the 19th-century economist. von thünen's enduring insight was that farms and forests closer to towns are more valuable, other things (such as soils and topography) being equal. the reason is simple: if the price of rice or wood is determined in a town's mar- ket, then nearby farmers bear lower costs in getting their products to market. Because they make higher profits, their land is worth more--that is, its rent for agriculture is higher. rents fall with distance to town, rapidly for bulky or perishable commodities (vegetables, milk) and more slowly for others (beef, coffee, hardwood timber; see figure 2.2). As land values decrease, land uses become more extensive, with pastures displacing crops and rotating fallows replacing permanent fields. After a certain dis- tance farmers can no longer profitably supply crops to market, and their land has no agricultural rent. this is the agricultural frontier; beyond it there are only subsistence farmers and standing forests. thus this stylized model predicts concentric rings of land uses cen- tered on urban areas. there is evidence that this model, inspired by german landscapes of the early 19th century, describes landscapes across the developing world (Chomitz and thomas 2003; Barnes, Krutilla, and hyde 2005). how does forestry fit into this picture? there is an important distinction between one-time extraction of old-growth trees and sus- tainable management of planted or natural forests. Big, valuable, old trees tend to get extracted as soon as they are accessible. Smaller, 71 At L O g g E r h E A d S ? Figure 2.2 A Stylized Model of How Land Use Changes with Remoteness Rent for intensive agriculture Rent for extensive agriculture and/or managed forestry Value of mined timber Distance from central market Intensive Extensive agriculture Open access Old-growth agriculture and/or managed forestry forests forests Timber extraction followed Timber extraction by agricultural conversion without conversion Source: Authors, adapted from Angelsen 2006. less valuable trees are often sold as a byproduct of clearing for agri- culture in von thünen's inner rings--especially if the central town has an appetite for fuelwood or charcoal. Only when natural forests are depleted does it become attractive to manage them, or plant new ones, for sustained harvest over time. When that happens, a forest ring can emerge at the edge of the agricultural ring. Of course, real landscapes don't look like archery targets. two elaborations are needed to make the model more realistic. First, as noted, the effects of distance are strongly modulated by soil, climate, and topographical features. Forests may remain on steep slopes near cities. remote areas with excellent soils may attract early coloniza- tion. And different combinations of accessibility, soil characteristics, and topography may appeal to different land uses and users. Chomitz and gray (1996), for instance, used extremely detailed land cover, topography, and soil data for Belize to elucidate the determinants of 72 I n C E n t I v E S A n d C O n S t r A I n t S S h A P E F O r E S t O U tC O M E S land use. they found that semisubsistence shifting cultivators--those who can't afford fertilizers and don't sell much in the market--favor hilly areas with nitrogen-rich soil and are only moderately sensitive to distance from town. In contrast, commercial cultivators--those who can afford fertilizers but rely on tractors--favor flat lands, regardless of soil fertility, and tend to be closer to markets. Second, security of land tenure is a crucial part of the picture. Although the determinants of land tenure are complex and rooted in history and institutions, they follow an important geographic pattern. typically, the more remote a plot of forest from settled areas, the more difficult it is to establish and defend property rights. So, elabo- rating von thünen's model, the cost of defending property likely rises with distance from town. (Moreover, defending a managed forest is typically more costly than defending a pasture.) At some point--the frontier--the cost of defending property rights exceeds the profitabil- ity of land. Beyond that point it doesn't make economic sense to invest in establishing a farm or actively managing a forest plot. In sum, von thünen's theory tells us that agricultural lands give way to forests with increasing remoteness. Figure 2.3, based on pan- tropical data, shows that the theory does a good job of describing today's tropical world. Figure 2.3 As Remoteness Increases, Mosaiclands Are Displaced by Forests, 2000 80 70 60 Asia 50 Latin America and Caribbean 40 mosaiclands in 30 Africa % 20 10 0 <1 1­2 2­3 3­5 5­8 8­11 11­16 16­26 >26 Hours to major city Source: Authors' calculations based on ECJRC 2003; see appendix B. Note: Covers only nonsavanna areas. 73 At L O g g E r h E A d S ? From Forest Frontier to Urban Center: A Stylized View of Forest Dynamics Let's now take a return journey, starting at the frontier. But this time we'll take the trip in a time machine, looking in a stylized way at the dynamics of change and the role of institutions, markets, and geography in shaping the trajectories of poverty, development, and environment. Some of these trajectories will end up at an urban center; others will not. Arrival of the Frontier the journey begins beyond the agricultural frontier. Population is sparse, and inhabitants are mostly long-residing indigenous peoples. An increase in forest rents triggers the arrival of the frontier. grad- ually or suddenly it becomes worth mining forests for timber, or worth defending plots of land to establish farming or pasture. Areas that had been beyond the frontier are now under contention. A race for property rights--or a dispute--begins. there are a number of triggers, some linked. Sometimes, as in Madagascar, the trigger is the growth of populations engaged in sub- sistence farming. this increases demand for land and lowers effec- tive wages and can be visualized as a shift upward in the rent curve for agriculture. But the most important trigger is the construction or substantial improvement of major roads, which make it possible to exploit new areas for timber and agricultural products. In the von thünen dia- gram the impact of new roads can be visualized as a counterclock- wise rotation of rent curves. the cost of transport falls and the reach of property rights is eventually extended, shrinking the rent penalty associated with remoteness. there are several spurs for the construction of major new roads, which may coincide with other triggers. First, it may be worthwhile to finance roads precisely because they offer returns in exploitable timber and raise land value. Farmers do this on a small scale with local road construction. Mahogany loggers, seeking lucrative stands of timber, can finance forest roads hundreds of kilometers long. Min- ers can open new roads. And state or national governments may find it beneficial to open new areas to forest extraction and conversion. At the national level, economic considerations blur with political ones. In Brazil and Indonesia between the 1960s and 1980s, roads were built in forested areas to promote colonization by landless farmers. road expansion, though without organized colonization schemes, was important in the opening of the Bolivian and Peruvian 74 I n C E n t I v E S A n d C O n S t r A I n t S S h A P E F O r E S t O U tC O M E S Amazon during the same period. Forest road construction is some- times geopolitically motivated--aimed at increasing government or military presence in remote and border areas. Elite interests and cor- ruption also play a role, if the rents created by road construction are funneled to politically connected interests (ross 2001). Finally, frontier expansion can be triggered by market and tech- nological changes. these can include booming markets for forest- competing commodities such as cacao, oil palm, coffee, and beef. Agronomic technology can also change incentives for deforestation. As noted, the breeding of soybean varieties adapted to low latitudes facilitated conversion of Brazilian savanna areas to cultivation. Trajectories Out of the Frontier: Disappearing or Rebounding Forests, Immiserization or Growth When the frontier arrives, people jockey for rights to trees and land. depending on who obtains possession of those resources, under what circumstances, and how they dispose of them, different trajectories of forest cover, income, and population evolve (table 2.3). Some of these trajectories correspond to the forest transition (box 2.1). · Intensification with deforestation. In this trajectory, changes in markets or roads increase the value of both standing timber and agricultural land in areas with favorable soils and climate. the resulting rush to claim timber and land often leads to conflicts between large and small actors. Profits from timber sales are used to finance the costs of clear cutting and of establishing crops. Agricultural development and timber harvest- ing may stimulate the growth of market towns with sawmills, slaughterhouses, and other agriculturally oriented service and processing businesses. this in turn increases the local population and demand for land. Land values rise, benefiting landholders; the results may be good or bad for equity depending on whether large or small landholders appropriate the land. Labor demand rises, either on farms or in pro- cessing and servicing centers, with possible benefits for poverty alleviation. Forest cover stabilizes at a low level, with remaining forest occupying slopes or poor- quality land. Agriculturally favorable areas, especially near cities, would be expected to follow this trajectory. the soybean areas of the Brazilian savanna provide an example. 75 At L O g g E r h E A d S ? Table 2.3 Five Trajectories of Forest Cover, Income, and Population Poverty and Location or Agricultural Managed forest Forest cover population identifying Trajectory rent curve rent curve trend trend characteristics Intensification Shifts up due Is everywhere Deforestation Landowners Periurban, good with to increasing dominated by continues and prosper, soils, high-input deforestation urban or agricultural rent stabilizes at low labor demand agriculture, international forest cover probably and higher demand and increases, population improved wages, and/or density tenure workforces increase, with labor growth possibly in towns Intensification Shifts up due Shifts up due Decreases, Landowners Periurban, with to increasing to increased then rebounds prosper, medium to reforestation urban demand, demand, labor demand good soils, increasing exhaustion of increases, and medium- to returns, and mined sources, wages and/or high-input improved and demand for workforces agriculture, tenure environmental increase and medium to services high population density Abandonment Shifts up due Shifts up due Decreases, Poverty Likely on with regrowth to increasing to improved then rebounds decreases due marginal lands: urban demand, tenure and to out-migration hillsides and/or then down due increased semiremote, to rising wages demand for forested, or wood and low population environmental density services Abandonment Shifts up, Never surfaces, Decreases Out-migration Marginal lands, and irreversible then down either because toward zero without poverty not near cities; degradation due to land of high costs alleviation nutrient-poor degradation of tenure or soils, slopes, or irreversibility of high incidence degradation of fire; grasslands in forest biomes Immiserizing Shifts up due Shifts down Decreases Larger but Probably not deforestation to falling wages due to soil toward zero poorer near cities; and increasing degradation, population anomalously food demand increases high population disputes over density given forest tenure remoteness and agroclimate 76 I n C E n t I v E S A n d C O n S t r A I n t S S h A P E F O r E S t O U tC O M E S Box 2.1 The Forest Transition the concept of the forest transition, introduced A number of developing economies appear by Mather (1992), describes a tendency for for- to be experiencing this transition. According est cover to decrease in response to coloniza- to rudel and others (2005), rebounds in for- tion, development, and population growth, est cover have been documented in Bangla- then rebound--a process that has occurred desh, China, Costa rica, Cuba, the dominican over the past two centuries in Western Europe, republic, the gambia, the republic of Korea, Japan, and the United States. rudel and oth- peninsular Malaysia, Morocco, Puerto rico, ers (2005) describe the two forces behind such and rwanda. India and vietnam may also be a turnaround. the forest transition can arise experiencing a forest transition. note that it is because higher wages, associated with the possible for forest cover to show a net increase opening of more productive farmlands, induce due to planting or secondary forest regrowth the abandonment of marginal farmlands, lead- even while old-growth natural forest is being ing to forest regrowth. the second route occurs lost in another part of the country. when deforestation makes wood so scarce that it is worth replanting trees. · Intensification with reforestation. the dynamics of intensification with reforestation are similar to those of the previous trajectory. But here, forest depletion leads to wood scarcity, and better tenure makes it possible for households and communities to manage forests. Under some conditions it becomes profitable to convert fields and pastures to woodlots or to tend and manage secondary forests. the result is a mosaic of croplands and managed forests. Examples include India (Foster and rosenzweig 2003), Kenya (tiffen and Mortimore 1994), and tanzania (Monela and others 2004). this is one route to the forest transition described in box 2.1. · Abandonment with regrowth. here one possible trig- ger may have been population expansion onto mar- ginal lands. After this trigger, rents are low and barely provide subsistence livelihoods for landholders. So if development elsewhere in the economy leads to higher wages, local populations migrate to better opportunities and these marginal areas are abandoned to natural for- est regeneration. this is the most familiar manifestation of the forest transition, and it summarizes the forest 77 At L O g g E r h E A d S ? experiences of Western Europe, Japan, and the United States. For instance, the U.S. state of vermont was largely cleared for agriculture in the early 19th century, despite its unfavorable terrain and climate. vermont's fields were then abandoned as western frontier expan- sion and better transportation brought new more pro- ductive farmlands into the market. Among tropical areas, Puerto rico is a striking and well-documented example, noted earlier. Other potential reasons for abandonment include a decline in the size of the youth cohort or in the price of agricultural commodities. Costa rica's strong forest regrowth during the 1990s may be an example of the latter, if pastures were abandoned in response to declining beef markets. · Abandonment with irreversible degradation. this tra- jectory is similar to the previous one, except that the land uses of in-migrants prove unsustainable. Soil fer- tility collapses due to nutrient exhaustion, compaction, or invasion by persistent weeds. the rent curve col- lapses, but natural regrowth doesn't occur. Examples include millions of hectares of imperata grasslands in Southeast Asia and large areas of apparently aban- doned pastures near Belem, Brazil. · Deforestation and immiserization. here the trigger could be population expansion. A combination of stagnant technologies and immobile labor contin- ues to push the rent curve out, but is combined with declining returns to labor and increased poverty. Poor agronomic conditions and inappropriate land use may further reduce incomes and increase pressure for nutri- ents from fresh deforestation. In environmental terms the outcome is similar to the abandonment with degra- dation trajectory. It differs in having a larger population and higher poverty rates. the humid forest of Mada- gascar exemplifies this scenario. Summary Soils, climates, markets, and governance shape pressures for defor- estation across space and over time. Changes can be driven slowly, 78 I n C E n t I v E S A n d C O n S t r A I n t S S h A P E F O r E S t O U tC O M E S as when population and income growth boost demand for food; or abruptly, as when new roads, crop varieties, or markets create pres- sure to convert forests. Formerly valueless land becomes more valu- able without forest cover than with. the resulting forest rents can range from barely more than zero to thousands of dollars a hect- are. Landholders, especially newcomers, respond rationally to these incentives, deforesting their lands to capture the rents. Positive feedbacks kick in: for instance, burgeoning populations demanding food, fuel, and secure land rights. So do negative feedbacks, such as deteriorating soil quality. the balance of these forces determines the regional trajectory of environment, income, and population. different trajectories are possible and imply different associations between poverty and deforestation (Sunderlin and others 2005). A prominent win-lose trajectory has historically been associated with rural development: the conversion of forest to intensive agriculture. here forests shrink but employment and incomes increase. Sometimes forest cover will rebound as wood becomes scarce, approximating a win-win outcome, but the recovered forest may not be equivalent in biodiversity or carbon storage to the previous forest. Alternatively, forest conversion can result in stagnant agriculture, providing sub- sistence income to a poor population that might be even worse off if denied access to this land. And in the worst, lose-lose case, forest conversion provides only an ephemeral income. this chapter stresses that policies and conditions that make forestland valuable for agriculture will result in a negative associa- tion between deforestation and poverty. More valuable land tends to result in more rapid deforestation but also higher incomes. Endnotes 1. It draws also on Chomitz and gray (1996), hyde and others (1997), and hyde (forthcoming). 2. Land values in this report are net present values unless explic- itly qualified as annual flows or as market prices or rentals. 3. See Boscolo and vincent (2000) for a similar bioeconomic analysis from Malaysia, and Pearce, Putz, and vanclay (2003) for a literature review. 4. the total value of timber was $324. But since the landholder may have the option to sell selectively extracted timber without clear cutting, the gross conversion cost is probably more relevant than the net cost in assessing the profitability of forest conversion. 79 Rural residents depend on forests and woodlands for fuelwood and other resources. Here, women carry firewood in the spiny forest region of Madagascar. © Frans Lanting / Corbis. CHAPTER 3 Poverty in Forests Stems from Remoteness and Lack of Rights P overty is pervasive in the tropical world--especially in rural areas. What is special about forest poverty? Why does it de- serve policy attention? Do forest dwellers constitute a sub- stantial proportion of all poor people? Are poor people the majority among forest dwellers? What poverty reduction policies might be tailored to forest dwellers? This chapter argues that it is fruitless to seek simplistic connec- tions between forests and poverty. Empirically, the links are weak. Some people derive wealth from forests, others from converting for- ests to agriculture. Many poor people live in marginal lands without trees. There are three distinctive forest poverty syndromes, with dif- ferent causes, locales, and possible remedies. First, remote areas tend to have high forest cover, high poverty rates, and low popu- lation densities. This remote forest and poverty syndrome poses a challenge for development because most standard approaches are inapplicable or extremely costly. A corollary is that forest-poverty relationships are quite different in remote and nonremote areas. Sec- ond, forest dwellers depend on forest resources for food, fuel, medi- cine, and income. But many interests compete to control or exploit forest resources. So changes in rights or access to forest resources can profoundly affect the livelihoods of people who live in and near forests. Third, there can be impediments--in policy, technology, or marketing--to commercializing forest products. Building on chapter 2 and Sunderlin, Dewi, and Puntodewo (2006), the discussion here uses a geographic lens to examine spa- 81 AT L O G G E R H E A D S ? tial overlaps between forests and poverty. This raises another ques- tion that may seem simple but is not: what do we mean when we say that an area is poor? Poverty Rates and Poverty Density: Two Ways of Viewing Poor Areas Let's set aside, for the moment, the question of how poverty should be measured at the level of individuals and households. Whatever the measure, a common approach to identifying high-poverty areas is to map, by province or district, the poverty rate: that is, the pro- Map 3.1a Poverty Rates for Brazil, 2000 Extreme poverty rate (percent) 0­5 6­7 8­9 10­13 14­20 21­29 30­37 38­44 45­51 52­82 N 0 500 1,000 1,500 2,000 Kilometers Source: Authors' mapping based on UNDP. 82 P O V E R T Y I N F O R E S T S S T E M S F R O M R E M O T E N E S S A N D L A C K O F R I G H T S portion of inhabitants who are poor. Map 3.1a shows this strikingly for Brazil: poverty rates are very high in Amazônia. But high-poverty areas can also be defined as places where the poverty density is high. Places with high poverty densities have a lot of poor people per square kilometer. Map 3.1b presents this measure for Brazil--where, as in many places, the two maps are like photo- graphic negatives of one another. Areas with high poverty rates tend to have low poverty densities, and vice versa. Which is the better definition of high-poverty areas? Later this chapter argues that each type of high-poverty area has distinct needs and policy implications. But first it examines the forces that shape Map 3.1b Poverty Densities for Brazil, 2000 Extremely poor people per km2 0­0.69 0.70­1.27 1.28­2.02 2.03­2.90 2.91­4.10 4.11­5.77 5.78­8.45 8.46­13.56 13.57­26.99 27.00­1,212.32 N 0 500 1,000 1,500 2,000 Kilometers Source: Authors' mapping based on UNDP. 83 AT L O G G E R H E A D S ? the geographic distribution of forest cover, deforestation, and poor people. Remote Forests--High Poverty Rates, Low Poverty Densities Remoteness mediates strong connections between forests, poverty, and population (see chapter 2). Because it is expensive to send pro- duce to markets from remote areas, it is rarely worth growing crops or harvesting timber for commercial use--meaning that deforesta- tion is low and forest cover high. The situation is even worse if areas have remained remote because they offer poor prospects for agricul- ture. In remote areas low land rents lead to low incomes because farm profits are negligible and off-farm employment opportunities missing. Hence poverty rates are high. Because farmgate prices are low, it is not worth applying much labor to a plot of land. Only extensive land uses such as pasture, shifting cultivation, and forest extraction are feasible. This means that population density is low--probably so low that poverty den- sity is also low. Low population densities, together with distance from administrative centers and poor communications, mean that forest dwellers have little voice in regional and national affairs. The problem is compounded if, as is likely, they are indigenous people not yet displaced by farmers or ranchers. Disempowered, they are subject to neglect or exploitation by elites seeking timber or mineral wealth. Finally, remoteness from law and communications and low population density mean that land and forest tenure are likely to be insecure. Table 3.1 summarizes predictions about remoteness and its effects on poverty and the environment. Evidence These relationships are evident in Nicaragua, a small country with a dominant city (Managua) and a forest frontier (Chomitz 2004). Extreme rural poverty rises sharply and smoothly with increasing travel time to Managua (figure 3.1). Population density falls just as smoothly and even more sharply, causing poverty density to decrease with increasing distance (figure 3.2). In addition, the ratio of rural workers to farmed land falls with remoteness, as expected. Forest cover rises with remoteness--though not as smoothly, partly because some nonremote areas are on slopes (figure 3.3). Tenure is less secure in frontier areas on the Atlantic side of the country: 84 P O V E R T Y I N F O R E S T S S T E M S F R O M R E M O T E N E S S A N D L A C K O F R I G H T S Table 3.1 How Does Increasing Remoteness from Markets Affect Poverty and the Environment? Indicator Effect Population density Decreases with remoteness Poverty rate Increases with remoteness Poverty density Decreases with remoteness Land productivity Decreases with remoteness Labor intensity Decreases with remoteness Tenure security Decreases with remoteness Forest cover Increases with remoteness a substantially lower proportion of farms have titles. This part of the country also has a prominent indigenous population. And unlike western Nicaragua, the Atlantic side has uniformly poor soil quality, as measured at the municipio level. Areas near Managua (within about four hours' imputed travel time) make up only a quarter of the nation's area but contain half of its extremely poor rural population. The most remote areas (those Figure 3.1 Extreme Rural Poverty Increases with Travel Time to Managua 60 % 50 rate, 40 poverty 30 extreme Rural20 0 20 50 100 200 500 1,000 2,000 Travel time to Managua in minutes Source: Chomitz 2004. Note: Excludes Managua department. 85 AT L O G G E R H E A D S ? Figure 3.2 Rural Population Density Decreases with Travel Time to Managua 500 200 50 20 km density, 10 sq. 5 per 2 1 population 0.5 people 0.2 Rural 0.1 0.05 0.02 20 50 100 200 500 1,000 2,000 Travel time to Managua in minutes Source: Chomitz 2004. Note: Excludes Managua department. more than 16 hours' journey) occupy about a third of the nation's area but contain only about 10 percent of its extremely poor rural population. The most remote areas have abundant forests; areas near Managua, sparser forests except on mountainsides. Figure 3.3 Forest Cover Increases with Travel Time to Managua 500 80 forest in 60 40 municipio of % 20 0 20 50 100 200 500 1,000 2,000 Travel time to Managua in minutes Source: Chomitz 2004. 86 P O V E R T Y I N F O R E S T S S T E M S F R O M R E M O T E N E S S A N D L A C K O F R I G H T S Implications The remoteness connection points to a distinct poverty-forest syn- drome. At the extreme end are places with relatively undisturbed forest cover and low population densities--perhaps 1 or 2 people per square kilometer, or less. Limited empirical and anecdotal evi- dence suggests that these people are extremely poor in terms of con- sumption, assets, and health indicators (such as child mortality). For instance, detailed measures of poverty for Vietnam in 1998 found that 73 percent of northern upland minority groups and 91 percent of central highlands minority groups lived below the poverty line, compared with 30 percent of the majority population (Baulch and others 2004, p. 278). Although data are lacking, indigenous people account for a large share of remote forest dwellers, and a disproportionate num- ber of indigenous people live in remote forest areas. For instance, Baulch and others (2004, p. 291) found that Vietnamese upland and highland minority members were four times farther from a market and six times farther from a telephone than were majority group members. Being indigenous compounds the difficulties associated with remoteness. Indigenous people have historically been subject to severe discrimination and exploitation. Despite legal and social progress in some countries, indigenous people remain disadvan- taged. A recent study of indigenous people in Latin America found that: · Indigenous children in Ecuador, Guatemala, and Mex- ico are twice as likely to be stunted (an indicator of severe malnutrition) as nonindigenous children. About half of indigenous children are stunted. · Indigenous adults have 2.3­3.7 fewer years of school- ing than do nonindigenous. · Indigenous people earn significantly less than do non- indigenous, and about half the gap cannot be explained by differences in education or other personal character- istics (Hall and Patrinos 2005). These differences would presumably be even larger if attention were limited to remote forest dwellers. Remote communities, indigenous or not, face enormous chal- lenges. For example, providing education and health care is difficult and expensive in remote areas (Chomitz and others 1998). Infra- 87 AT L O G G E R H E A D S ? structure is also difficult to provide in remote communities. Water and electricity systems cannot exploit economies of scale or den- sity. Building and maintaining feeder roads is expensive in rainy, swampy, or mountainous environments, and their unit costs sky- rocket if they serve few people and little traffic. Lack of roads is also an obstacle to realizing the potential of for- estry. Community forestry is often considered a development option that combines environmental and livelihood benefits for remote forest communities. Timber is, apparently, the main commercial resource that these communities have in abundance. But to benefit communities, that timber has to get to market. Poor roads mean high transport costs, and high transport costs reduce the stumpage value of timber--the value received by communities. Roper (2003) identifies poor roads as one of the main barriers for commercializ- ing the forests owned by indigenous people of Nicaragua's Atlantic region. Transport costs of $0.34 a cubic meter per kilometer eat into wood values, for these communities, of about $20 a cubic meter. In sum, there is not necessarily a strong relationship between forest cover and poverty rates, though poverty densities tend to be lower in forested areas. But some forest areas suffer from poverty because of their remoteness from agricultural markets and because low population densities make it difficult to deliver services and infrastructure. Incomes of Forest Dwellers Depend on Rights and Access to Forestlands Forests provide food, fuel, fodder, wood, and medicine to their inhab- itants and neighbors, for personal consumption and for sale. Though these resources represent a substantial portion of forest dwellers' income, it is difficult to measure forest income and dependence. It seems reasonable to suppose that households' degree of reliance is inversely related to population density. People living at extremely low densities--say, fewer than five per square kilometer--probably rely heavily on the forest for their livelihoods. These people are numer- ous in aggregate, but spread thinly across the world. For logistical and cultural reasons, they are hard to survey. Because Living Standards Measurement Surveys (LSMS) usu- ally omit remote, low-density districts and provinces, there is little quantitative information about this most forest-dependent popula- tion. On the other hand, there are hundreds of millions of people in 88 P O V E R T Y I N F O R E S T S S T E M S F R O M R E M O T E N E S S A N D L A C K O F R I G H T S high-population-density forests and forest-agriculture mosaics. Here the problem is accurately enumerating, measuring, and attributing cash values to the extraction of dozens of forest products. Standard survey instruments probably underestimate this income stream. A further issue is that forest products serve as a safety net, relied on more heavily in times of crop failure and other hardship. One-time surveys could easily miss this feature. With these concerns in mind, Vedeld and others (2004) con- ducted a metareview of 54 case studies that measured income from forest products. The studies are not a representative sample, so their data are merely indicative. Forest income (averaging $678 a year, adjusted for purchasing power parity) accounted for about a fifth of household income in the sample--a significant contribution, par- ticularly for families near the survival line. Wild food and fuelwood were the most important products, accounting for 70 percent of for- est income (although some products, such as fodder, are probably underreported in the sample). Forest income was higher the farther that households were from markets--suggesting that for remote communities, a lack of alter- native income opportunities and an abundance of forests lead to greater dependence on such resources. Reinforcing this, the most forest-dependent half of the sample cases (earning an average of 42 percent of their income from forest products) lived in more remote areas, had less education and livestock, and averaged only about half as much income per household. The few studies that examined the distribution of income within communities found that because poor people depended more on forest products, forest income reduced inequality. The average Gini coefficient (a common mea- sure of inequality) was 0.51 when forest income was excluded but fell to 0.41 when it was included. Forest Control and Tenure Can Affect Income Because rural poor people are dependent on forest resources, any- thing that affects their rights or access to those resources merits attention. Three policy concerns arise here. The first is a potential tragedy of the commons. Although some forests are effectively man- aged by communities as common property resources, others are open access--managed by no one, exploited by all. If these forests are degraded, local income streams are destroyed. Second, forest regulations from colonial times, or recently imposed on environmental grounds, may restrict forest dwellers' 89 AT L O G G E R H E A D S ? ability to gather fuelwood, food, and other forest products. Forest officials can also use these regulations as a source of rents, extract- ing bribes from poor forest dwellers. Third, changes in legal or de facto ownership of forests affect local dwellers' ability to undertake commercial forestry and agricul- ture, both of which can provide a route out of poverty. Open Access Forests Suffer Degradation There is a long history of concern that rural households, dependent on woodlands for fuel, suffer when those woodlands are depleted. A thorough recent literature review by Arnold, Kohlin, and Persson (2006, p. 604) concludes that "the body of information now available suggests that the greater part of rural populations in both Africa and South Asia do not face serious welfare implications due to decreas- ing access to biomass, but resource poor areas and households can face a problem, in particular landless people without access to com- mon pool biomass stocks." It is difficult to measure the extent and depth of deprivation due to degradation of open access forests. Given the substantial propor- tion of income derived from forests, forest degradation may trans- late into lower consumption or increased workloads as it becomes harder to glean resources from thinned-out woodlands. But it is hard to measure consumption of forest resources and local access to them, and to control for other correlates of resource availability. This point is illustrated by Bandyopadhyay, Shyamsundar, and Baccini (2006), who marshal unusually detailed and comprehensive data to assess the impacts of biomass scarcity in Malawi. They report that deforestation and forest degradation reduced biomass, nation- wide, by 16 percent over 1990­2004, and that fuelwood accounts for about 12 percent of the value of household consumption. Gathering fuelwood takes an average of 1.5 hours a day--and 84 percent of this burden falls to women. In this setting one might hypothesize a vicious circle of poverty and degradation. As forests thinned, people would be expected to reduce their consumption of fuelwood or to devote more time to gathering it. But Bandyopadhyay, Shyamsundar, and Baccini (2006) find that, other things being equal, households in lower-density forests did not spend more time gathering fuelwood, and that fuel- wood gathering did not come at the expense of agriculture. In the rural south of Malawi, where forests are more degraded, the authors found that a 10 percent lower biomass density was associated with 90 P O V E R T Y I N F O R E S T S S T E M S F R O M R E M O T E N E S S A N D L A C K O F R I G H T S 0.2 percent lower consumption. This modest association suggests that households adapt their fuel sources or strategies as forest resources dwindle. But because this effect applies to every house- hold in a neighborhood, forest degradation might be significant in areas with higher population densities. On the other hand, the study found that consumption actually fell with higher levels of biomass in less degraded areas. This may be a spurious association: high-biomass areas are likely more remote and less suitable for agriculture. But the study's bottom line is that low biomass densities are not associated with drastic poverty bur- dens--meaning that people are resourceful in adapting to exhausted biomass or that proper measurement is extremely difficult. More studies like this are needed in a wide variety of settings before gen- eral conclusions can be drawn. Another potential tragedy of the commons could result from overexploitation of bushmeat. Some 2.2 million tons of wild mam- mals and other animals are exploited for food each year in the Congo Basin, representing a major source of animal protein for the region (Fa, Currie, and Meeuwig 2003). Bushmeat accounted for about 10 percent of household production in a very poor village surveyed by de Merode, Homewood, and Cowlishaw (2004), and was especially important in the lean agricultural season. But bushmeat extraction already exceeds the sustainable supply by more than 25 percent in Cameroon and the Democratic Republic of Congo (Fa Currie and Meeuwig 2003). Population and income growth, increased road access, and shrinking animal populations threaten to make exploita- tion increasingly unsustainable. Regulations Can Limit Forest Use Forest regulations may restrict forest dwellers' ability to gather fire- wood or other forest products, to market timber, or to convert for- ests to agriculture. For instance, Cameroonians cannot legally sell trees they grow as part of a cocoa agroforest (Gockowski and others 2006). Facing depressed prices, their timber stock is undervalued by $1,460 a hectare. In Indonesia regulations discourage farmers from selling rubber trees they cultivate in an agroforestry system. The potentially valuable wood is burnt instead (Joshi and others 2002). Dwellers Are Often Dispossessed of Land and Forests When wealthy interests seize or degrade forests, poor local popula- tions can suffer. These situations do not lend themselves to con- 91 AT L O G G E R H E A D S ? trolled study, so evidence is anecdotal. For instance, Davis (2005) estimates that 100,000 Cambodians depend for their livelihoods on tapping forest dipterocarp trees for oleoresin, a commercially valu- able product. Davis, McKenney, and others (2004) and McAndrew and others (2004) report that illegal logging and conversion of for- ests to acacia plantations have deprived resin tappers of access to trees. Establishment of protected areas has sometimes involved dis- placement of and loss of assets by local populations (Ghimire and Pimbert 1997; Geisler and De Sousa 2001). (See chapter 6 for a discussion of efforts to emphasize comanagement of parks as an alternative to displacement.) Cernea and Schmidt-Soltau (2003) and Schmidt-Soltau (2003) review the establishment of nine national parks in central Africa and conclude that about 51,000 people were displaced. In only two of the nine cases were there formal resettle- ment policies. In two cases no compensation was made to the dis- placed populations, and in most of the other cases compensation was inadequate. The loss of assets could be thousands of dollars per capita depending on the potential surrendered stumpage value of timber, but this valuation is complicated by the need to estimate transport costs from these remote areas. Ferraro (2002) analyzes how the establishment of Ranomafana Park in Madagascar affected its inhabitants, who were subsequently denied park access and forced to rely on buffer zones for agriculture and forest extraction. The analysis accounts for different time paths of resource degradation under the unsustainable agricultural tech- nologies used by the residents. Access to the park allows them to defer the long-term effects of soil fertility decline and timber exhaus- tion. Ferraro finds that park exclusion imposed a mean annual cost of $39 a household--equivalent to 14 percent of household income. This is consistent with a survey by Shyamsundar and Kramer (1996) that asked households how much they would require in compensa- tion for resettlement. Forests without Trees Are a Widespread Dilemma Large swathes of tropical Asia are legally forestland but devoid of trees: · In Indonesia between 333,000 square kilometers (Contreras-Hermosilla and Fay 2005) and 370,000 square kilometers (Boccucci, Muliastra, and Dore 2005) 92 P O V E R T Y I N F O R E S T S S T E M S F R O M R E M O T E N E S S A N D L A C K O F R I G H T S of land are under Forest Department control but devoid of forests. Of this, about 100,000 square kilometers was designated for conversion to oil palm, timber, or pulp plantations that never materialized. The rest repre- sents deforestation of forests gazetted for conservation, watershed protection, or sustainable timber production. · In India 20 percent of reserved forest--at least 100,000 square kilometers--is without trees (Ministry of Envi- ronment and Forests 2005). · In the Philippines a 1981 presidential decree declared all land with a slope greater than 17 percent to be within the public domain. Today only a small fraction of these 150,000 square kilometers retains any forest (Fay and Michon 2005). · In Thailand 70,000 square kilometers of state forestland was treeless when a 1992 reform sought to rezone the forest domain (Fay and Michon 2005). Treeless forests are problematic because they are inhabited by people without secure land rights. The populations are large--at least 40 million in Indonesia alone (Boccucci, Muliastra, and Dore 2005). But lack of security makes it hard for them to invest in land improvements, such as reclaiming degraded grasslands or plant- ing trees. Weak tenure depresses land values and reduces access to credit (Deininger 2003). Forests, Poverty, and Deforestation: Ambiguous Relationships The search for win-win solutions to poverty and environment dilem- mas motivates the hypothesis that there is substantial spatial over- lap between areas with high poverty rates and areas with high forest cover, high deforestation, or both. This chapter and the previous one offer several reasons to expect those relationships to be muddled: · Remoteness is associated with high poverty rates and forest cover, but low deforestation. · Insecure tenure may be associated with high deforesta- tion and either low or high poverty rates, depending on the deforestation process. 93 AT L O G G E R H E A D S ? · Deforestation is sometimes undertaken by wealthy commercial interests. · Deforestation can create valuable agricultural assets for smallholders. · On the other hand, deforestation may reflect the expan- sion of subsistence-oriented populations onto increas- ingly unsuitable lands. Empirical studies reflect this ambiguity. Sunderlin, Dewi, and Puntodewo (2006) analyze associations between poverty rates, pov- erty densities, and forest cover in seven countries. In three of the seven they find a significant positive correlation, at the district level, between poverty rates and forest cover. Vietnam is a clear exam- ple, with high poverty rates, low population densities, and high forest cover in the remote mountain regions of the north and cen- tral parts of the country. In three countries there is no significant relationship. Only one country has a significant negative relationship: Brazil. At the national level, across all forest types, the relationship is nega- tive because the semiarid region of northeast Brazil has high poverty rates and low forest cover--while the wealthy southernmost part of the country, well into the forest transition, has low poverty rates and high forest cover. This national-level correlation result obscures the relationship evident in map 3.1: remote western Amazônian forests have high poverty rates and high forest cover. Deininger and Minten's (1999) study of Mexico related munici- pal poverty to deforestation, controlling for a host of biophysical and socioeconomic factors including slope, elevation, rainfall, indig- enous proportion of population, and land tenure. They found a strong partial relationship, statistically and quantitatively, between deforestation over 1980­90 and poverty rates in 1990. Other things constant, a one standard deviation increase in poverty was associ- ated with an increase of almost 3 percentage points in the annual deforestation rate. But rainfall and hilliness were strongly negatively associated with deforestation and positively associated with poverty. So it is likely that the simple correlation between poverty and defor- estation is negative. The rest of this section uses newly available, fine-scale data to examine spatial relationships between poverty rates, population densities, forest cover, and deforestation in four important forested countries: Brazil, India, Indonesia, and Madagascar. Keep in mind 94 P O V E R T Y I N F O R E S T S S T E M S F R O M R E M O T E N E S S A N D L A C K O F R I G H T S that measures of poverty, forest cover, and deforestation are not comparable between countries. Deforestation and Poverty in Brazilian Amazônia Are Largely Unrelated Deforestation of Brazilian Amazônia is sometimes blamed on poor people. But evidence suggests that poverty and deforestation prob- lems in Amazônia are largely separate problems requiring separate approaches: · Poverty and deforestation are spatially localized, with limited overlap. · Most deforestation is undertaken by large-scale, well- capitalized actors. · Much of this large-scale deforestation occurs on public land and so represents a regressive transfer of public resources. · Deforestation is profit-driven, but typically yields mod- est profits per hectare. Scale of Deforestation Remote sensing data suggest that poor people are responsible for less than a fifth of deforestation in Brazilian Amazônia (figure 3.4).1 Because clearing is expensive and large clearings require mechanical Figure 3.4 Most Deforestation in Brazilian Amazônia Reflects Large- and Medium-scale Clearing, August 2000 to July 2003 Small 10,751 km2 Large (19%) 21,475 km2 Small (< 20 ha) (39%) Medium (20­200 ha) Medium Large (> 200 ha) 23,541 km2 (42%) Source: Wertz-Kanounnikoff 2005. Note: Categories reflect size of clearings, not properties. 95 AT L O G G E R H E A D S ? equipment, there is a strong correlation between clearing size and the deforester's wealth or access to capital. Subsistence farmers are unlikely to be able to afford to clear more than 20 hectares a year, and most probably clear far less. About 39 percent of deforestation occurs in incremental clear- ings larger than 200 hectares, which likely represent wealthy inter- ests. This finding is consistent with Chomitz and Thomas (2003), who find that agricultural establishments of 2,000 hectares or more contain 53 percent of privately owned, cleared land in Amazônia. It is also consistent with the description by Margulis (2004) of large- scale ranching activities in the Amazon. Map 3.2 Amazônian Deforestation 2000­03 Showing Rates and Predominant Clearing Size Small clearings Medium clearings (< 20 ha) (20­200 ha) > 0%, < 0.1% > 0%, < 0.1% 0.1­0.5% 0.1­0.5% > 0.5% > 0.5% Large clearings (> 200 ha) > 0%, < 0.1% Forest 2000 0.1­0.5% > 0.5% Source: Authors' calculations, see Appendix B. Note: Rate = deforested area/total area. 96 P O V E R T Y I N F O R E S T S S T E M S F R O M R E M O T E N E S S A N D L A C K O F R I G H T S Location of Poverty and Deforestation Map 3.2 shows the concentration of deforestation in Brazil between 2000 and 2003, following a broad arc extending from Maranhão to Rondônia. The map, based on data from the Brazilian National Institute for Space Research (INPE), includes only deforestation of mature Amazônian forest, excluding deforestation of cerrado (savanna) woodland and secondary regrowth. Darker colors corre- spond to more rapid deforestation. The colors represent predomi- nant shares of deforestation by size of incremental clearing, a proxy for the scale of the actors involved. The map shows that large-scale clearings predominate in Mato Grosso and southern Pará along the Map 3.3 Amazônian Deforestation Rates and Rural Illiteracy Densities Rural adult Deforestation illiteracy rate 2001 (%) density 2000 (people/km2) > 0, < 0.03 0.04­0.07 < 0.01 0.08­0.15 0.01­0.1 > 0.16 0.1­1 1­2 > 2 Source: Authors' calculations, see Appendix B. Note: Rate = deforested area/total area. 97 AT L O G G E R H E A D S ? forest-cerrado boundary. Small-scale clearings--and thus, presum- ably, small-scale landholders--are scattered throughout but are most prominent in Rondônia and parts of Pará. Map 3.3 shows the density of rural adult illiteracy in Brazil in 2000, overlaid by deforestation rates in 2001. Most of the deforesta- tion hotspots of Mato Grosso and Pará are in areas where rural adult illiteracy density (a proxy for poverty density) is extremely low-- from 0.01­0.1 per square kilometer. These densities are too low for poverty to be a plausible cause of deforestation, reinforcing the con- clusions drawn from the predominance of large-scale clearings in these spots. But there are places where deforestation hotspots and higher illiteracy densities coincide, as in central Rondônia. About 12 percent of deforestation in Brazil between 2000 and 2003 occurred on lands known as terras arrecadadas--unambigu- ously public lands. This represents private appropriation of public lands. It is not known how much of this transfer was legal. Some, perhaps most, took place through an opaque process called grila- gem,2 which results in an award of title to land of uncertain status (Margulis 2004). What is clear is that about half of this deforestation occurred on incremental clearings of 20­200 hectares, and another quarter on clearings more than 200 hectares. (The properties them- selves are presumably much larger than the clearings.) The break- down of deforestation by size is similar in the regions for which tenure status is unknown. These regions contain terras devolutas-- unallocated and undemarcated public lands. So it is plausible that much deforestation in Brazilian Amazônia constitutes the appropri- ation of public lands by large private actors, in nontransparent and possibly illegal ways. India Contains Net Reforestation with Patches of Deforestation Despite its huge poor rural population and limited arable land, India has experienced a forest transition (see box 2.1). Between 1971 and 2003 forests grew from 10 to 24 percent of national area (Foster and Rosenzweig 2003). This expansion conceals a welter of local processes. One expla- nation may be a supply response to a long-term increase in the price of fuelwood. While some of this response may have occurred on pri- vate lands or in regenerating forests under joint forest management, it is also due to the Indian government's massive investments in tree plantations--on a nominal scale of about 1 million hectares a year 98 P O V E R T Y I N F O R E S T S S T E M S F R O M R E M O T E N E S S A N D L A C K O F R I G H T S since 1980. At the same time, forests dwindled in villages where the green revolution increased the value of putting land into agriculture (Foster and Rosenzweig 2003). A 1980 decree forbade deforestation for agriculture and probably restrained large government-sponsored projects (Rudel 2005). Still, deforestation continues in places. Many forests are thinning under human pressure, so the proportion of very dense forest is only 7.5 percent, or 1.5 percent of the national area (Ministry of Envi- ronment and Forests 2005). In sum, there is continuing conversion and degradation pressure on India's remaining native forests, while planted forests--already about half of the forest estate--expand. Figure 3.5 shows the relationship in India, at the district level, between forest cover and illiteracy--a rough indicator for poverty.3 There is no clear relationship. Districts with more than 50 percent forest cover contain just 3.6 percent of the country's illiterates. Addi- tional analysis finds examples of both coincidence and divergence of forest cover, illiteracy, and tribal populations. The role of joint for- est management in stimulating reforestation and reducing poverty remains to be comprehensively investigated. Figure 3.5 Illiteracy and Forest Cover Have No Clear Link in India 1.0 0.8 2003 in 0.6 forest 0.4 Proportion 0.2 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Illiteracy rate, 2001 Source: Authors' calculations based on Ministry of Environment and Forests 2005 and Census India 2001. Note: Bubble sizes are proportional to population. 99 AT L O G G E R H E A D S ? Indonesia Is a Complex Pastiche of Deforestation Processes In Indonesia the relationship between corporate interests and small- holders, poverty, and deforestation is complex and varies across the archipelago. FWI and GFW (2002), like other sources, surmise that most deforestation is due to clearance by timber, pulp, and oil plantation interests. In some, perhaps most, cases these parties used conversion permits to obtain timber or pulp, but failed to install promised plantations. There are no nationwide, reliable, quantitative estimates of the importance of corporate relative to smallholder deforestation. In a detailed study combining ethnography and remote sensing data, Dennis and others (2005) examine nine disparate sites in Sumatra and Kalimantan, sometimes finding multiple agents of deforestation at a single site. They find smallholders converting forest to com- mercially oriented permanent agriculture in two sites, smallhold- ers engaged in short-rotation shifting agriculture in six sites, and land clearance by large plantation companies in six sites. Deforesta- tion was also caused by arson connected with land tenure disputes between communities and companies, and by escaped fires lit by hunters in search of easier paths to deer, fish, and turtles. This kalei- doscope of actions by rich and poor actors illustrates the futility of seeking easy generalizations about the relationship between poverty and deforestation. Here is a tale of two islands: one where poverty and forests coincide, another where deforestation appears to accompany rela- tive prosperity (see maps 3.4 and 3.5). The tale uses new subdistrict data on poverty, forest cover, and deforestation over 1990­2000. (The deforestation data, although the best available, were assembled from disparate and possibly inconsistent sources and must be interpreted with caution.) Consider first the island of Sulawesi. Panel A of map 3.5 shows the relationship between its forest cover and poverty rate in 2000, with the bubble sizes indicating the population of each subdistrict (kecamatan). In many subdistricts that are mostly (more than 50 percent) forested, the poverty rate exceeds the national average of 17 percent. These subdistricts contain 95 percent of the poor people in Sulawesi's mostly forested subdistricts, and about a third of its poor people. 100 P O V E R T Y I N F O R E S T S S T E M S F R O M R E M O T E N E S S A N D L A C K O F R I G H T S Panel B of map 3.5 shows that most of these high-forest, high- poverty districts are in the remote central portion of the island, far from the urban centers at the tips of its "arms." Panel C shows the slightly negative relationship between the deforestation rate and pov- erty rate, with the bubble sizes showing initial forest cover. In sum, Sulawesi conforms with the remoteness­high poverty rate­high for- est cover syndrome and does not show a strong positive association between poverty rates and deforestation. Kalimantan, to the west, presents a different picture (map 3.4). Some areas have high forest cover and high poverty rates (panel A). Again, these are mostly in the remote center of the island (panel B) and are large in area but have few people. A much larger popula- tion lives in subdistricts that are mostly forested and have lower than national average poverty rates (green areas of panels A and B). Overall, the poverty rate in the high-forest areas is 19 percent-- scarcely more than the national average. Panel C of map 3.4 shows that the subdistricts undergoing the most rapid deforestation tend to have much lower poverty rates than more stable subdistricts. Why the difference between the two islands? Kalimantan has a much more active logging industry than Sulawesi. An hypothesis, to be confirmed, is that there is a pulse of income and deforestation at the logging frontier. This would be consistent with case studies of two forest communities, one on each island, by Deschamps and Hartman (2005). The remote Kalimantan site is inhabited by forest- and agriculture-dependent groups; the forest is threatened mostly by logging. The Sulawesi site abuts a national park, and here the threat is conversion to rice, cocoa, cloves, and other cash crops. The authors find that all three groups (the two in Kalimantan and the one in Sulawesi) receive similar agricultural incomes per household. The agriculturally oriented groups earn about as much again from forest extraction. The forest-oriented Kalimantan group, however, earns three times as much from forest extraction than from agriculture. It can draw on commercially valuable wood and gaharu (a prized nontimber forest product)--neither of which is plentiful at the Sulawesi site. Further investigation is needed to determine whether these income gains are sustainable. High poverty rates in low-forest areas hint at unsustainability and may reflect populations living in degraded forests without trees, with insecure land rights. 101 AT L O G G E R H E A D S ? Map 3.4 Poverty, Forests, and Deforestation in Kalimantan A Forest cover and poverty High forest, low poverty 1.0 High forest, high poverty Low forest, high poverty 2000, 0.8 Low forest, low poverty no data 0.6 kecamatan in 0.4 forest 0.2 C Proportion 0.0 1.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 2000 Poverty rate 0.8 deforested, 0.6 0.4 kecamatan of B 0.2 0.0 Proportion 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Poverty rate 102 P O V E R T Y I N F O R E S T S S T E M S F R O M R E M O T E N E S S A N D L A C K O F R I G H T S Map 3.5 Poverty, Forests, and Deforestation in Sulawesi A Forest cover and poverty 1.0 High forest, low poverty High forest, high poverty 2000, 0.8 Low forest, high poverty Low forest, low poverty 0.6 no data kecamatan in 0.4 forest 0.2 Proportion 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Poverty rate B C 1.0 2000 0.8 deforested, 0.6 0.4 kecamatan of 0.2 Proportion 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Poverty rate Note: The maps and graphs on these two pages show data by kecamatan (subdistrict). In panel A each subdistrict is shown by a bubble. The bubble size is proportional to the subdistrict's population. The graph shows the proportion for- est cover and poverty rate of each subdistrict. Subdistricts are classified into four color-coded categories, based on their combination of high versus low forest cover and above or below average (for Indonesia) poverty rate. Panel B maps the subdistricts according to these categories. Panel C shows the poverty rates and deforestation rates for each subdistrict. The rate is defined as forest loss/total subdistrict area. 103 AT L O G G E R H E A D S ? Madagascar Shows a Spatial Association between Forests and Poverty, with Other Factors at Work In Madagascar much deforestation is undertaken by people who are extremely poor by absolute standards. About three-quarters of the population lives below the national poverty line. How, then, are we to understand the relationship between poverty and forests within the country? Map 3.6 presents some perspectives, using district data on poverty in 1993 and deforestation over 1990­2000. About 10 percent of the country's poor people lived in districts with high forest cover (more than 50 percent forested). Almost all lived in districts that were mostly forested and had higher than aver- age poverty rates (red areas in map 3.6). Areas with the highest poverty rates also tended to have higher deforestation rates. But the overwhelming majority of Madagascar's poor people live in areas with low forest cover--including formerly forested areas that have become degraded. The link between poverty and deforestation in Madagascar unravels, however, when other factors that might affect forest clear- ance are taken into account. Gorenflo and others (2006) assessed the impact of poverty on deforestation, controlling for road access, topography, and the presence of protected areas. These factors were powerful correlates of deforestation. Holding them constant, there was a mild partial correlation between poverty and deforestation in most of the country's subregions. But in the southwest, where commercially oriented maize cultivation prevails, poverty was nega- tively associated with deforestation--suggesting that deforestation is at least temporarily associated with higher incomes. Summary Beware of facile generalizations about poverty, forests, and defores- tation. In general, forest cover is an unreliable indicator of poverty rates, and poverty is a poor proxy for deforestation. In Brazil, India, Indonesia, and Madagascar only a small proportion of poor people live in mostly forested districts. In India and Indonesia there are for- ested places with low poverty rates (by national standards) as well as high. A more reliable generalization is that highly forested areas tend to have low densities of poor people. But there are several important forest-poverty linkages that can guide policy. First, remote, forested areas in transfrontier zones often 104 P O V E R T Y I N F O R E S T S S T E M S F R O M R E M O T E N E S S A N D L A C K O F R I G H T S Map 3.6 Forest Cover, Deforestation, B and Poverty in Madagascar Population in cities and towns 500,000­2,000,000 100,000­500,000 50,000­100,000 25,000­50,000 Provinces Forest cover and poverty Low forest, low poverty Low forest, high poverty High forest, low poverty High forest, high poverty A 1.0 2000, 0.8 0.6 firaisana in 0.4 forest 0.2 Proportion 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Poverty rate C 0.4 0.3 2000 deforested, and 0.2 firaisana 1990 of 0.1 between Proportion 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Poverty rate Note: The maps and graphs on this page show data by firaisana (subdistrict). In panel A each subdistrict is shown by a bubble. The bubble size is proportional to the subdistrict's population. The graph shows the proportion forest cover and poverty rate of each subdistrict. Subdistricts are classified into four color-coded categories, based on their combination of high versus low forest cover and above or below aver- age (for Madagascar) poverty rate. Panel B maps the subdistricts according to these categories. Panel C shows the poverty rates and deforestation rates for each sub- district. The rate is defined as forest loss/total subdistrict area. 105 AT L O G G E R H E A D S ? have high poverty rates, especially when timber markets are distant. Providing services and development options in these areas is a great challenge, but they may benefit some of the world's poorest people. Second, forest-dwelling populations may face legal or bureaucratic obstacles to using forest assets. The scope of this problem is not well quantified, but it could be quite large. Third, tens of millions of people occupy hundreds of thousands of square kilometers of for- ests without trees. More secure tenure in these areas could improve both livelihoods and the environment. Endnotes 1. The satellite sensor used by the Brazilian National Institute for Space Research (INPE) may be unable to detect extremely small clearings (on the order of a hectare), leading to an underestimate of smallholder clearings. But the incremental expansion of such small clearings might be detected over two or three years. Thus on a statistical basis the area of small clearings might be approximately correct. It is also possible that some large clearings represent neigh- boring small clearings. 2. The word is said to derive from the practice of using crickets to soil forged documents to make them look antique. 3. Population data are from Census of India (2001). Forest cover data are from Ministry of Environment and Forests (2005) and include forests outside the tropical forest biome and both native and planted forests. Data are missing for some states and districts. 106 Harpy eagles, found in Latin America, exemplify the need to manage land- scapes to ensure biodiversity survival. A nesting pair of harpy eagles re- quires 100 km2 of forest to provide enough prey for sustenance. A viable population of harpy eagles requires dozens of times as much. Juan Pablo Moreiras / Fauna & Flora International / Comisión Centroamericana de Ambiente y Desarrollo photo archive. 107 Forest loss typically increases total annual water flows, potentially exac- erbating chronic (but not necessarily catastrophic) flooding. Juan Pablo Moreiras / Fauna & Flora International / Comisión Centroamericana de Ambiente y Desarrollo photo archive. CHAPTER 4 Deforestation Imposes Geographically Varied Environmental Damages W hat environmental problems are associated with defores- tation--where and when do they occur, and who suffers from them? Forests have diverse environmental values and functions. This chapter unbundles those functions, which in- clude provision of biological goods, maintenance of genetic diversity, regulation of water flows, and storage of carbon. The approach used here helps explain who suffers when those functions are impaired by forest degradation and the costs of maintaining or substituting for those functions. This review is necessarily highly selective. The Millennium Eco- system Assessment (2005), on which this report draws, provides a comprehensive synthesis of ecological relationships between forests and people. A motivation for that assessment, and for this report, is that social and economic policies often do not incorporate scientific insights. That shortcoming can result in poor prioritization of prob- lems and poor choices of instruments for addressing problems. Two questions underscore this chapter's discussion: What are the most compelling reasons for reducing deforestation, and where do they apply? Before starting the discussion, it is useful to offer a framework for thinking about environmental damages. The essence of environ- mental problems is the concept of an externality--where one per- son's actions unintentionally benefit or harm another person. For instance, I may remove the trees along my river banks, heedless that the consequent erosion will pollute your drinking water with sedi- 109 AT L o G G E r h E A D s ? ment. In principle, society as a whole would be better off if exter- nalities were factored into environmental decision making. If I gain $100 from felling my trees but you have to pay $1,000 to filter your drinking water, the potential damages from pollution far outweigh the potential profits from deforestation. There should be a way to arrange to keep the trees standing. The nature of that arrangement depends on whether society assigns me the right to cut the trees or guarantees you the right to clean water. If the latter, your right trumps mine. My potential $100 gain is too little to compensate you for your $1,000 loss, so I won't bother to cut the trees. But if I have an absolute right to cut my trees, you would find it worthwhile to pay me up to $1,000 for the environmental service of preserving the trees and reducing erosion. Either way, the trees would remain standing. And the assignment of rights determines who benefits and who loses. But there are transactions costs involved in enforcing rights and negotiating payments. Those costs can be high if there are diffuse sources of the externality and many on whom it impinges. Costs escalate further if sources and impacts are distant, with no shared institutions to help mediate the problem. Unfortunately, many of the environmental externalities associated with deforestation are charac- terized by a difficult combination of diffuse sources, diffuse impacts, and lack of intermediating institutions. Biodiversity Loss--A Local and Global Concern Biodiversity is an ambiguous term, and ambiguity can create con- fusion. The term is used in several senses. sometimes it refers to biological resources: timber, fuelwood, fish, medicinal plants, polli- nating bees. sometimes it describes the diversity of microorganisms or cultivars within a plot of cropland. And sometimes it denotes the diversity of species, genes, or ecosystems considered from a local or global viewpoint. These different senses of biodiversity all refer to valuable natu- ral assets, and they overlap to some degree. But failure to distin- guish between them can lead to conceptual and policy errors. For instance, there is evidence that diversity of rice cultivars within a farm plot increases yield and reduces costs. But that doesn't mean that conservation of plant and animal biodiversity in a nearby forest will necessarily boost productivity in a farmer's field. 110 D E F o r E s TAT I o N I M P o s E s G E o G r A P h I C A L LY VA r I E D E N V I ro N M E N TA L DA M AG E s Biological Resources Provide Locally Valuable Services Local biological resources and local biodiversity are probably the aspects of forest biodiversity most immediately relevant to liveli- hoods and welfare. The most obvious benefits are related to extrac- tion of biological resources such as fuelwood, timber, food, and fodder. These constitute a significant though inaccurately known share of income for people living in or near forests (see chapter 3). other local biodiversity services may be extremely important but inadequately recognized. According to Cassman and Wood (2005, p. 759) at least 80 percent of the world's 100 most important food crops are pollinated by wild pollinators. ricketts and others (2004), in a study of Costa rica, show that bees from forest fragments con- tribute substantially to coffee productivity and profits on adjacent farm plots--a pure, uncompensated externality. Forests may also support natural antagonists of crop pests. or they may provide food for those beneficial creatures during the fallow season when there are no pests to feed on (Cassman and Wood 2005, p. 759). These subtle relationships can be easy to miss--and easy to dis- rupt unintentionally. In principle, because most occur at a very local scale and contribute directly to livelihoods, it should be possible to set up local management institutions to handle them. In practice, these ecological benefits may not be immediately apparent to land managers, and solid scientific quantification is often lacking. Extinction Threats Draw Attention to Globally Significant Biodiversity This chapter focuses on conserving globally significant biodiversity: genes, species, and ecosystems at risk of extinction. Given the global nature of the externalities involved, this is a serious challenge. Con- servation of globally significant biodiversity is motivated by the growing threat of irreversible loss. According to the Millennium Eco- system Assessment (2005), current extinction rates are about 100 times the rate they were before humans existed and could increase by another 10­100 times. There are two reasons for the world to be concerned about this looming, irreversible loss. The first is instrumental: conserve bio- diversity because it provides specific economic services or averts specific risks. The second is intrinsic: conserve biodiversity because people attach aesthetic and spiritual values to it, or because their values demand it. Proponents of this rationale justify conservation 111 AT L o G G E r h E A D s ? of diversity as an inherent goal, of the same kind as reducing child mortality or preserving the great artworks of earlier civilizations. Can conservation of globally significant biodiversity be justi- fied on instrumental grounds? Many valuable pharmaceutical prod- ucts are derived from tropical plants. But this has so far failed to spark a significant market for bioprospecting rights: drug companies have not been willing to pay much for the right to prospect in par- ticular areas. simpson, sedjo, and reid (1996) provide a convinc- ing explanation for why. A standard approach to bioprospecting is unfocused--simply hoping that randomly selected organisms will yield promising chemicals for treatment of target diseases. But given a low rate of success and substantial overlap in genetic contents between forest plots, no individual plot is unique enough to com- mand much of a premium. Drug companies could continue to find ample specimens for evaluation, for decades to come, even with rapid deforestation. Accordingly, such bioprospecting has failed to provide sufficiently "bankable" benefits to pay for conservation. Better-focused bioprospecting might confer higher economic values for conservation. For instance, forests that harbor the wild relatives of commercially exploited species such as coffee, vegeta- bles, and fruits might be an enduring source of useful genetic infor- mation for global agriculture, including information about pests and their natural enemies. More systematic searches for particular kinds of biological activity, combined with better information about bio- diversity distribution, might generate high bioprospecting rents for particular locales. Finally, the Millennium Ecosystem Assessment (2005) argues that biodiversity loss and associated large changes in forest cover could trigger abrupt, irreversible, harmful changes. These include regional climate change, including feedback effects that could the- oretically shift rainforests to savannas; and the emergence of new pathogens as the growing trade in bushmeat increases contact between humans and animals. These instrumental arguments for conserving global biodi- versity are still rather speculative and unfocused. In contrast, the intrinsic rationale for conservation--conservation as a fundamental value--has deep resonance in ethics, aesthetics, and religion. These different approaches lead to a formulation of biodiversity policy that places an "existence value" on maintenance of diversity. surveys suggest substantial willingness to pay for this diversity, at least in industrial countries, though the validity of these stated preferences 112 D E F o r E s TAT I o N I M P o s E s G E o G r A P h I C A L LY VA r I E D E N V I ro N M E N TA L DA M AG E s has been questioned. Funding for the Global Environment Facility and for large, conservation-oriented nongovernmental organizations (NGos) is a palpable demonstration of willingness to pay. The science of biogeography provides important insight into the design and geographic targeting of forest policies aimed at reducing extinctions. First, biogeography tells us that species and other aspects of biodiversity are unevenly distributed across the Earth (Mace, Masundire, and Baillie 2005, pp. 90­91). species of vertebrates are much richer near the equator. Endemic species (those with limited ranges) tend to cluster on islands and mountaintops. And an analy- sis of biodiversity hotspots has found that half the world's vascular plant species are located on just 1.4 percent of Earth's land surface (Myers and others 2000). other conservation scientists emphasize the desirability of preserving places with distinctive ecological pro- cesses, such as mass migrations of wildlife, or locations that appear to be generating new species (Burgess and others 2006). second, one of the most well-established regularities in ecology links habitat area to number of species in the habitat. As habitats shrink, there are fewer niches for specialized species, and there is less room for predators that need large ranges to maintain viable populations. As usually formulated, a 90 percent reduction in area is associated with about a 30 percent reduction in the number of supported species. For an optimist, this is not too bad a result: the mostly empty glass is still, in effect, more than half full. Conserving just 10 percent of the original forest biome could theoretically main- tain more than half of its original biodiversity. Third, however, forest fragmentation further reduces the sur- vival prospects of species and their ecosystems. relative to a large chunk of forest, small chunks with the same total area are more exposed to natural and human pressures such as wind, fire, and hunting. Forest-specialist species requiring large contiguous blocks of habitat fare poorly in fragments that are widely scattered through an inhospitable matrix of fields or settlements (Laurance and others 2002). As fragmentation increases and connectivity weakens, pros- pects for these species get dramatically worse (box 4.1). Fourth, extinctions respond with a lag to loss of habitat. The relation between species and area holds over the long run. Places that have suffered rapid habitat loss may still contain their origi- nal complement of species, but may not be able to for much lon- ger. Brooks, Pimm, and oyugi (1999) estimate that a newly isolated 1,000 hectare fragment experiences about half its eventual species 113 AT L o G G E r h E A D s ? Box 4.1 Forest Fragmentation Can Trigger Local Ecological Collapse Imagine a checkerboard arrangement of for- when the proportion is above that thresh- ested properties or plots of land in an area that old. Although this is a highly stylized result, is a biodiversity corridor. Initially the land is it points to the possibility of rapid and unex- entirely forested, and a key animal species is pected ecological collapses as deforestation free to roam from one side to another. Then proceeds. It also underscores the potential settlers arrive and randomly convert some importance of landscape-level management for properties to fields that are inhospitable to the biodiversity in places where forests and farms animal. how does the amount of converted are intermixed. Farm-level decisions about land affect the animal's ability to traverse the cropping, maintenance of gallery forests, and corridor? establishment of living fences can make a big A remarkable mathematical result is that difference for biodiversity conservation. the animal can always find a path when the proportion is below 41 percent, but never Source: Based on Forman 1995. losses in 50 years. rosenzweig (2001, 2003) takes a longer, starker view. he says that in the very long run, species loss is proportional to habitat loss because climate and disease shocks continue to cause extinctions, while the rate of creation of new species is proportional to habitat area. so he argues that a 90 percent habitat loss implies an eventual 90 percent species loss. he doesn't specify whether that adjustment takes centuries, millennia, or longer. Finally, the rapid pace of climate change provides another rea- son why connectivity is important, and why rosenzweig's drastic adjustment may come sooner rather than later. Plant and animal species are adapted to particular temperature ranges. As global tem- peratures rise, the survival of some species will depend on their ability to migrate to cooler areas, on higher slopes, or at higher lati- tudes. If there is insufficient connectivity in existing habitats, these species may be unable to migrate and will get caught between rising temperatures and inhospitable surroundings. Climatically triggered diseases could make things worse. This foray into biogeography provides important policy lessons. · Conservation of globally significant biodiversity requires focus- ing attention and resources on certain places. This approach may clash with other norms for allocating resources between 114 D E F o r E s TAT I o N I M P o s E s G E o G r A P h I C A L LY VA r I E D E N V I ro N M E N TA L DA M AG E s countries and places. A country's share of global biodiversity, for instance, may differ from its share of global population or poverty. · The most urgent extinction threats will largely be in and around fragmented mosaiclands where habitat has drastically shrunk and become fragmented. here, species and ecosystems are living on borrowed time. There is a brief opportunity--the next few decades--to refurbish these landscapes and make them more habitable to most threatened species, while main- taining their usefulness for agriculture and human habitation (McNeely and scherr 2003; rosenzweig 2003). reducing frag- mentation and encouraging connectivity are important parts of this program. · Threats are low in large, unfragmented forest tracts. But these species-rich areas are the last irreplaceable examples of large-scale ecological processes in quasi-natural habitat. They also offer insurance against climate change, allowing species uninterrupted pathways for migration in response to rising temperatures. Policy interventions today to head off a dynamic of uncontrolled conversion could determine whether, over the coming century, these tracts retain ecological vitality or whether they grow fragmented, placing their species at threat of extinction. How Does Deforestation Affect Water, Air, and Weather? Clean water flows from forested hillsides, muddy torrents from steep denuded slopes. These observations have often been used reflexively to justify forest conservation. But in recent decades scientific research has refined people's understanding of how forests and land use affect flooding, sedimentation, landslides, and dry season flows. Although some aspects of these relationships remain debated, the overall mes- sage is clear: the forest-hydrology relationship is highly nuanced. The effects on hydrology of changes in forest and land use depend, sys- tematically and explicably, on how and where changes occur. It is important to get the science straight. For instance, Aylward (2005) reports on a study of the potential effect of reforesting pas- turelands around Costa rica's main hydropower reservoir to reduce sediment inflow. The study found that reforesting 1 hectare would 115 AT L o G G E r h E A D s ? reduce sedimentation and estimated the benefit, in increased reser- voir life span, to be $74 per hectare reforested. But the study also found that higher water consumption by the trees would draw down the reservoir--imposing potentially larger, countervailing costs due to decreased electric output during dry years. Consider too the diagnosis of the catastrophic 1998 floods on China's Yangtze river. The floods were blamed on deforestation, and a swift policy response was the shutdown of the Chinese log- ging industry. This move disrupted domestic employment and placed increased extractive pressures on forests of high biodiversity significance in southeast Asia and elsewhere. But to what extent was deforestation to blame? subsequent analysis suggests a complex picture (Yin and Li 2001). Beginning with the construction of the Great Jinjiang Levee in 1548, and accelerating in the past 50 years, many land use changes have reduced the ability of the Yangtze watershed to handle peak water flows. The flows, which formerly covered vast floodplains, have increasingly been constricted by levees and dikes. since 1949, 50 cubic kilometers of lakes have been reclaimed for agriculture, reducing lake storage capacity by a third and thus crippling a major buffer against flooding. heavy siltation has raised the river bed, increasing the risk of flooding. Deforestation and other land use changes have increased the proportion of the basin subject to erosion, and so over the long run have presumably contributed to siltation. But observations over a 30-year period did not show any link between siltation and deforesta- tion, suggesting that it may take decades for erosion to end up as sedi- ment in the river. Thus in the context of the complex hydrodynamics of a large basin, it is not at all clear to what degree the logging ban has reduced the risk of future floods or how it compares to alternative watershed management strategies. To help get the science straight, this section relies heavily on comprehensive and incisive reviews by Bruijnzeel (2004) and Bruijnzeel and others (2005). It also draws on Bonell and Bruijnzeel (2005), Calder (2005), CIFor and FAo (2005), and van Noordwijk and others (2006). Interested readers can refer to these works for more detailed treatment. From Farmer's Field to River Basin: Policy at Different Scales Interactions among people, precipitation, soils, and vegetation play out differently at different scales. First, many phenomena impor- tant at smaller scales become attenuated at larger ones (Kiersch and 116 D E F o r E s TAT I o N I M P o s E s G E o G r A P h I C A L LY VA r I E D E N V I ro N M E N TA L DA M AG E s Tognetti 2002; Calder 2005). Consider erosion and sedimentation. Erosion can be severe on steep fields, clogging local streams with sediment. At this scale, actions to prevent erosion and sedimenta- tion can have more or less immediate effect. But at the level of a large watershed, new upland erosion doesn't translate immediately to sedimentation far downstream. This is because any individual bit of sediment has to follow a long journey of short trips from moun- tainside to river mouth. That journey could take decades--or it may never be completed--because the sediment gets lodged somewhere along the way (Chomitz and Kumari 1998). second, the economics and politics of watershed management vary with scale. In small watersheds (10­100 square kilometers) it may be relatively easy to organize local communities to deal with clearly perceived issues such as erosion or landslide risk. Larger basins require more complex, wide-ranging institutions to negotiate interests between upstream and downstream populations. But the payoffs to cooperation at this scale might be considerable. Urban populations might be willing to pay substantial sums to reduce flood risk, sedi- ment damage to reservoirs, or pollution of urban water supplies. If they could do so by paying poor upland populations to conserve for- ests, it would be a happy outcome on many grounds: reducing flood- ing and poverty, with biodiversity conservation as a by-product. Against this background, consider how changes in forest cover and land use affect hydrological functions that people care about, at two scales: local and far-field. Local Hazards Depend on Many Variables Forests modulate water flows in various ways. In the popular concep- tion, trees are sponges, soaking up water and releasing it later. But this is an inadequate and incomplete metaphor (Bruijnzeel 2004). Forest floors, with their leaf litter and porous soils, easily accommo- date intense rainfall. Water infiltrates the ground until soils are satu- rated. In this sense forest soils act like sponges. But trees behave like fine-misted fountains, pumping water out of the ground and transpir- ing it as water vapor into the air.1 rain also clings to tree leaves, from which it evaporates without ever touching the ground. The effects of deforestation on water availability, flash floods, and dry season flows depend on what happens to these countervailing influences of infil- tration and evapotranspiration--the sponge versus the fountain. replacing a mature forest with a mature agroforest doesn't much change evapotranspiration and so has little effect on water yield. But 117 AT L o G G E r h E A D s ? permanent conversion of forest to pasture, annual crops, or short perennial crops reduces evapotranspiration and thus increases the water yield from a plot. on this point there is strong scientific con- sensus. Converting a tropical moist forest is roughly equivalent, in water yield, to increasing rainfall by 300 millimeters a year. That's why south Africa's Working for Water program pays for the removal of invasive tree species--to increase water availability in parched regions. The program employs 21,000 poor people and provides water to Capetown at a cost 90 percent below the alternative: con- struction of a dam (van Wilgen and others 2002). Floods and Flow Regularity Deforestation's effect on the timing of flows--on floods and dry sea- son flows--is more difficult to predict and is sensitive to the balance between infiltration and evapotranspiration effects. Deforestation tends to increase flooding for two reasons. First, with a smaller "tree fountain" effect, soils are more likely to be fully saturated with water. The "sponge" fills up earlier in the wet season, causing additional precipitation to run off and increasing flood risk. second, deforesta- tion often results in compacted soils with little ability to absorb rain. Locally, this causes a faster response of streamflows to rainfall and thus potential flash flooding. That is why some Costa rican run-of- river hydropower plants invest $1.50­5.00 in watershed protection per kilowatt generated each year (rojas and Aylward 2002). These small plants (6­17 megawatts) have no storage reservoirs, so their output is greatest when water flows evenly at their turbines' capacity. Dry Season Flows More controversial is the impact of deforestation on dry season flows. here there is strong divergence between the popular view that deforestation dries up springs, and scientific evidence that strongly indicates higher--not lower--flows after deforestation. A thorough review by Bruijnzeel (2004) finds only a couple documented cases of lower flows. But he stresses the need for more observations. In theory, deforestation could decrease dry season flows under certain conditions: · New land use patterns result in severely compacted soils, so losses of rain to runoff exceeds gains from shutting off the "fountain" (Bruijnzeel 2004). This might happen where cattle or machinery have caused severe compaction, where there has been extensive road building, or where fires have degraded the 118 D E F o r E s TAT I o N I M P o s E s G E o G r A P h I C A L LY VA r I E D E N V I ro N M E N TA L DA M AG E s landscape. In other words, postdeforestation land use matters more than just deforestation. · Annual rainfall is high and concentrated in the wet season. · soil has considerable water holding capacity or is in an impor- tant recharge zone. Sedimentation and Erosion Think about sediment, and you will understand why watershed man- agement involves more than simple decisions about how many trees to retain or plant (Van Noordwijk and others 2006). First, people place different values on sediment. reducing sediment is a service to downstream irrigators, reservoir owners, and water filtering plants-- but not to farmers who depend on it for renewed soil fertility. second, deforestation doesn't necessarily increase erosion, the main source of sediment. As with flooding, what matters is how land is used after forest is removed, and especially whether leaf lit- ter is maintained. Typical erosion rates are 0.2 tons a hectare under forest and 0.6 under plantations with ground cover, but more than 50 tons a hectare may be observed under plantations without leaf litter (Wiersum 1984, quoted in Bruijnzeel 2004). Forest roads gen- erate far more erosion per hectare than do agricultural uses. Ziegler and others (2004), in a study of a northern Thai watershed, found that unpaved roads delivered as much sediment to streams as did agricultural fields--though the fields occupied 24 times more area. Third, the spatial arrangement of land use matters. Van Noord- wijk and others (2006) simulated the effect on sedimentation of dif- ferent combinations of clean-weeded coffee plantation and forest on a sumatran hillslope. retaining 25 percent forest at the bottom of the slope reduced sedimentation by 93 percent relative to no forest. For- est retention elsewhere on the hillslope was much less effective. In sum, watershed management for sediment reduction involves many choices, with different consequences for incomes, biodiversity, and other environmental outcomes. Landslides Forests can provide protection against shallow landslides. Perotto- Baldiviezo and others (2004) studied the incidence of landslides after hurricane Mitch in a Nicaraguan watershed. They found that less than 1 percent of forested lands were affected by landslides, regard- less of slope. For plots under bare soil (recently harvested) this inci- dence jumped from near zero on flat land to 7.5 percent on land with 119 AT L o G G E r h E A D s ? 10 percent slope, and 10 percent when slope was 20 percent. But some deep-seated landslides occur regardless of forest cover. Water Quality Urban water protection is potentially one of the most important ser- vices that forests provide. Filtering and treating drinking water is expensive. Forests can reduce the costs of doing so--either actively, by filtering runoff, or passively, by substituting for housing or farms that generate runoff. An example is New York City's watershed (National research Council, Water science and Technology Board 2004, pp. 156­58). For many decades the city had drawn its water, untreated, from its 5,000 square kilometer watershed. In the mid- 1990s water quality began to deteriorate, and authorities were faced with the prospect of committing $6­8 billion (in operating and future maintenance costs) for a treatment plant to meet safety standards. Instead the city developed an innovative program for watershed protection, at a cost of $1.0­1.5 billion. About $250 million was used to purchase and protect land (though not necessarily forest- land). But the plan also involved activities to reduce water pollu- tion. Importantly, it worked with dairy farmers to manage manure and nutrient runoff. This example may be widely applicable. Dudley and stolton (2003) found that 18 of the world's 42 largest tropical cities draw their water directly from protected areas. An extension of this study might usefully identify cities that draw their drinking water from small, steep, forested watersheds. Under these conditions the pub- lic value of watershed protection is likely to counterbalance private rewards to forest conversion or degradation. Geography of Local Hazards Nelson and Chomitz (2006) examine the potential spatial coinci- dence of local hazard risk and poverty in two hilly Central American countries, honduras and Guatemala. The authors define a water- shed's critical zone as where forests meet agriculture on slopes. This is where deforestation might be most rapid and might lead most rapidly to erosion. They then define a watershed's "sensitivity" as its proportion in a critical zone. In Guatemala highly sensitive watersheds (more than 20 percent critical) occupy 22 percent of the country but contain 43 percent of its poor people and 54 percent of its montane forest. The poverty rate in these watersheds is 70 percent, compared with 53 percent for 120 D E F o r E s TAT I o N I M P o s E s G E o G r A P h I C A L LY VA r I E D E N V I ro N M E N TA L DA M AG E s Figure 4.1 Guatemala Critical Watersheds Have High Poverty Rates 100 90 80 70 % 60 rate 50 Poverty 40 30 20 10 0 0 10 20 30 40 50 60 Critical zone area as a % of watershed area Source: Nelson and Chomitz (forthcoming). Note: Bubble size indicates absolute number of poor people in watershed. the country as a whole. Figure 4.1 shows that all of the most sensi- tive watersheds have high poverty rates. Nelson and Chomitz (2006) also find that the relationship between hydrological vulnerability and poverty is stronger when smaller watersheds are considered. This focuses attention on highly local externalities, with poor people both causing them and bearing their burden. At the global scale, Dilley and others (2005) used topographic and geological criteria to identify areas around the world at risk of landslides. Areas with the highest imputed mortality risk from land- slides--including along the mountainous spine of Latin America, the islands of sumatra, New Guinea, and the Philippines, and the bor- der between India and Myanmar--are shown in map 4.1.2 Because mountain peaks tend to host distinctive, restricted range species, the landslide risk map is strikingly similar to the imminent extinctions map (see map 1.8). Far-field Impacts Can Be Major We turn now from local impacts of deforestation to far-field impacts--those felt tens or hundreds of kilometers away. 121 AT L o G G E r h E A D s ? Distant Floods Because urban floodplain populations are growing rapidly, human and economic exposure to flood risk is growing as well. But some hydrologists doubt whether large-scale upstream deforestation has much impact on distant downstream populations. They argue that small rainstorms, passing over a large river basin, affect only one tributary at a time, so any flooding effect is diluted by the time it reaches a city down the river's main branch. The rare storm large enough to drench the entire basin, they argue, would probably over- whelm the basin's ability to absorb water into the soil. A storm that big would cause a flood regardless of tree cover. This argument is hard to test. There are two ways to study links between deforestation and floods in large basins: through long-term historical studies and by simulation. Each has limitations. historical studies may face confounding trends--such as increases in irrigation or construction of dams--with impacts that are difficult to disentan- gle from those of deforestation. simulation studies are sensitive to the reliability and detail of data on rainfall, soils, and river flows. historical studies yield contrasting results. several studies reviewed by Bruijnzeel (2004), mostly in southeast Asia, found no marked increase in river flows following basinwide deforestation. In contrast, Costa, Botta, and Cardille (2003) found a substantial impact on far- field flows of deforestation in the 175,000 square kilometer Tocantins river basin. They compared flows over 1949­68 and 1979­98. only 6 percent of the basin had been converted to planted pasture or cropland in 1960, but by 1995 it was 49 percent. The study found that, despite similar rainfall, wet season river flows increased 28 percent between the two periods. The authors speculate that the difference between their results and those in southeast Asia reflect the faster natural regeneration in the Asian study areas. A change from primary forest to plantations or secondary regrowth may have little effect on hydrologi- cal flows (though possibly a profound effect on biodiversity). Because it is hard to isolate the effect of long-term, large-scale changes in land use through observations, researchers have turned to hydrological simulations. Advances in hydrological modeling have resulted in tools that can reproduce watershed behavior with some accuracy, such as the Distributed hydrology soil Vegetation Model (Wigmosta, Vail, and Wittenmaier 1994). These models trace water and sediment flows over the landscape, incorporating effects of vegetation and geology. Their accuracy is validated by comparing model predictions, based on historical precipitation records, with streamflow records. 122 D E F o r E s TAT I o N I M P o s E s G E o G r A P h I C A L LY VA r I E D E N V I ro N M E N TA L DA M AG E s Map 4.1 Mortality Risks from Landslides B A B C A C Landslide Mortality Risk Deciles 1st­4th 5th­7th 8th­10th Source: Dilley and others 2005. scientists are beginning to use these models to assess the impacts of land cover change in the tropical world. An example is Thanapakpawin and others (2006), who constructed and validated a model of the 3,853 square kilometer Mae Chaem watershed (in northern Thailand) using limited data on soils and precipitation. The 123 AT L o G G E r h E A D s ? authors used the model to assess, retrospectively, the impact of the loss of about 10 percent of the watershed's forest cover. other things being equal, they found that deforestation would have increased wet season flows by 2 percent and dry season by 4 percent. But expanded irrigation more than counterbalanced the deforestation effects, reducing net outflows by up to 6 percent in the wet sea- son and 16 percent in the dry. ongoing work extends this to larger basins and shorter periods. Douglas and others (2005, forthcoming) have conducted similar simulation studies in large river basins throughout the tropics. Their work seeks to determine whether and where forest losses would have significant effects on biodiversity and water flows. For this purpose, they simulated the hydrological impacts of the complete conversion to agriculture of all forests deemed "critical or endangered" by WWF.3 They calculated the resulting change in river flows and identified areas that would be expected to experience an increase in mean annual flows of more than 25 percent. The assumption was that increases of this magnitude could be associated with chronic if not catastrophic flooding. (The exact impact is highly sensitive to the local geography of the floodplains and to daily and hourly peak flows.) The authors found that the hypothetical catastrophic loss of these forests would increase annual river flows by more than 25 percent for about 100 million people, most of them on floodplains. In nine basins containing 55 million people, more than 100 people would be affected by each square kilometer of forest conversion--a crude indicator of the poten- tial for mobilizing downstream interest in upland forest conservation. The hydrological cost-effectiveness of forest protection might be increased by concentrating on hydrological hotspots--those where deforestation might have the greatest downstream impact. Combin- ing spatial models of deforestation and hydrological functions could help pinpoint these locations. Water Quality Does it make sense to manage land over a very large basin to reduce sedimentation for downstream users? Although sediment travels slowly over slopes, it can be mobilized fairly rapidly from riverbanks over a wide area. A simulation study found that sediment load in a 2,500 square kilometer watershed could be cost-effectively reduced by revegetating steep croplands close to the river (Khanna and others 2003). Although this finding sounds obvious and mirrors the local area results noted earlier, it contrasts with a strategy of protecting 124 D E F o r E s TAT I o N I M P o s E s G E o G r A P h I C A L LY VA r I E D E N V I ro N M E N TA L DA M AG E s forested uplands far from major rivers. In generalizing this analysis, it is important to keep in mind that vegetation other than trees can intercept sediment without using as much water as trees do. Effects on Air and Climate Forest fires release noxious smoke and smog, disrupting transport and industry and triggering respiratory illness. These are chronic problems. In Brazil smoke from forest and land fires is cited as a major environmental problem by municipio governments that represent 39 million people. These problems are worst in dry (El Niño) years. Tacconi (2003) estimates that the Indonesian fires of 1997­98 affected 110,000 square kilometers in Indonesia and Malay- sia, imposing costs in damaged health and industrial disruption of $676­799 million. Deforestation affects wind flows, water vapor flows, and absorp- tion of solar energy, so it's plausible that it affects local climate. But it's difficult to assess these impacts, which may operate differently at different scales, from field to continent. one study found that deforestation on lowland plains moved cloud formation and rainfall to higher elevations (Lawton and others 2001). Another study found strong changes in land-sea breeze patterns affecting cloud formation and upland rainfall in tropical island and coastal settings (van der Molen and others 2006). Yet another study simulated the impact on regional climates of a plausible scenario for global deforestation dur- ing this century. It predicted that Amazônian temperatures will rise by 2 degrees Celsius, in addition to effects from global warming. It also predicted possible disruptions of Asian monsoon patterns (Fed- dema and others 2005). Finally, van der Molen and others (2006) and other researchers intimate the possibility of global impacts of widespread deforestation--especially for coastal and island defores- tation, which disrupts atmosphere-wide wind patterns. Deforestation Spurs Climate Change There's no need to repeat here the vast literature on climate change; for authoritative summaries, see the work of the International Panel on Climate Change (Watson and Core Writing Team 2001) and the Mil- lennium Ecosystem Assessment (2005). still, three points are crucial. First, climate change is a real and growing threat to people, economies, and the environment. Arctic communities already face permafrost that is no longer permanent. Andean populations need 125 AT L o G G E r h E A D s ? to begin planning for the impending loss of glacial icepacks on which their water supply depends. Poor sahelian farmers and pasto- ralists, already coping with a difficult and volatile climate, may soon experience deteriorating conditions. Looming behind the predictable threats, however, is the real but unquantifiable possibility of rapid, catastrophic changes--such as a shutdown of Atlantic ocean cur- rents or massive changes in regional climates. second, tropical deforestation is an important source of green- house gases, releasing 3.8 billion tons of carbon dioxide (Co2) a year (Achard and others 2004). such deforestation also accounts for about 20 percent of human-generated Co2 emissions (house and others 2006). Third, preventing deforestation and encouraging forest regenera- tion have the same effect on atmospheric Co2, no matter where they occur. This is in marked contrast to other environmental benefits of forest conservation, which depend on local conditions. Forest management has global effects on greenhouse gases. But does it make economic sense to actively manage forests to reduce atmospheric Co2 emissions (box 4.2)? What might be the costs and benefits to landholders? how do the costs of reducing Co2 through forest management compare with those of abating Co2 emissions from transportation, electricity generation, and manufacturing? Consider first the cost of reducing Co2 emissions from defores- tation. The cost to landholders depends on: · The per hectare profits forgone by maintaining forest rather than converting it. · The difference in carbon storage between a conserved forest and a field or pasture. These two considerations vary tremendously depending on the factors discussed in chapter 2: agroclimate, market opportunities, and technology. rigorous data on these trade-offs have been assem- bled by the Alternatives to slash and Burn program for a number of land use systems in the moist forests of Brazil, Cameroon, and Indonesia (Tomich and others 2005). For each land use, researchers calculated the net present value of profits (a measure of the value of land devoted to that use), the amount of carbon stored, and the level of biodiversity conserved. Drawing on these data, figure 4.2 shows the implicit costs of reducing carbon emissions through forest conservation. These costs were calculated by comparing the profits and carbon storage of each 126 D E F o r E s TAT I o N I M P o s E s G E o G r A P h I C A L LY VA r I E D E N V I ro N M E N TA L DA M AG E s Box 4.2 Trees and Carbon: Lessons from Biology for Forest Policy Many of us vaguely remember learning, in a · reforestation and afforestation absorb Co2 biology class long ago, that trees absorb carbon out of the atmosphere, storing it as wood dioxide (Co2) and produce oxygen. But that's and biomass as long as the forest endures. only part of the story, because trees also use · oxygen for respiration and release Co2, just as Pulp and timber plantations absorb Co2 as they grow. The carbon they embody animals do. A growing tree absorbs more Co2 is transformed into wood or paper after from the atmosphere than it emits, embodying harvest. It may be released rapidly into the carbon as wood, leaves, and other biomass. the atmosphere if these products are dis- A tree that dies, or is cut down, rots or carded or incinerated. But timber used for burns. The carbon in its biomass is then enduring structures may stay out of the released into the atmosphere as Co2. soil car- atmosphere for a long time. bon may also be exposed and lost into the atmosphere. In a regenerating forest, growing · Plantations can create fuel from thin air trees outnumber dying ones, so carbon accu- by absorbing Co2 and transforming it to mulation is vigorous. But as forests mature, biomass. When burned as charcoal or bio- their net accumulation of Co2 slows. (There fuel, the same amount of Co2 is returned are debates about the rate at which old-growth to the atmosphere, in a closed cycle. so forests continue to sequester carbon.) sustainable biofuel plantations are carbon What does this mean for the impact of neutral--they don't add net Co2 to the alternative land uses on carbon storage? atmosphere--and can substitute for fossil fuels, which do augment global warming. · Converting forests to agriculture or pas- ture releases Co2 into the atmosphere, · Logging releases Co2 from damaged trees, so protecting a threatened forest could though forest recovery may partly offset reduce greenhouse gas emissions. this effect over time. The harvested tim- ber may release its carbon quickly to the · Mature standing forests maintain carbon atmosphere, if burned or discarded to rot, stocks, but have at most a low per hectare or slowly if used for construction. rate of accumulation. land use system and an assumed forest baseline. (The Cameroon and Indonesian examples assume a baseline of already logged, depleted forest.) The figure shows the tremendous variation in the potential costs of conserving carbon. At one extreme, traditional pasture management in Acre, Brazil, entails a loss of 145 tons of carbon per hectare but creates only $2 a hectare in land value (in net present value of all future earnings). so the cost of conserving carbon is, in principle, just $0.03 a ton C (or 127 AT L o G G E r h E A D s ? less than $0.01 per ton of Co2). similarly, traditional rubber agrofor- estry in Indonesia provides lower per hectare profits and a signifi- cant loss of carbon relative to community agroforestry--although it creates much more employment. In contrast, the most profitable land use, for intensive cocoa in Cameroon, entails a carbon loss of 103 tons per hectare, confers a land value of $1,149, and provides 93 days of employment. here the theoretical cost of conserving carbon is $11 a ton ($3/ton Co2). These calculations assume that the costs are borne by land- holders, in forgone profits, and that workers can find alternative employment at the same wage. This may be a reasonable assump- tion when deforestation is related to frontier migration. otherwise, workers bear a burden--because of lower-paying work--that should be included when calculating the cost of emission reductions. For extensive land uses this makes little difference. Traditional Brazilian pasture, for instance, provides only 11 days of employment a hectare per year. But rubber agroforestry, intensive cocoa, and intensive palm oil provide about 100 days of employment, at roughly $1.50 a day. Counting a portion of this as a cost would somewhat increase the cost of conserving carbon. one way of looking at these calculations is to ask whether soci- ety should incur these costs. Are they justified by the benefits from mitigating climate change? Making this assessment requires assign- ing a value to abating emissions. Yohe, Andronova, and schlesinger (2004) suggest that to mitigate climate change, the global commu- nity needs to value carbon abatement at $10 a ton now (and rising over time based on the interest rate). At that value, converting for- est to intensive cocoa barely breaks even from a social viewpoint-- though from a private viewpoint it is the most profitable land use shown in figure 4.2. Efforts are under way to scale up estimates of this type to the global level. These estimates must be considered tentative, because comprehensive information on land use systems is lacking. still, the estimates are useful for assessing the potential scale of the contribu- tion of land management to carbon abatement. sathaye and others (forthcoming) estimate the potential contribution of avoided defor- estation to carbon abatement for different levels and trends of car- bon prices. (Anticipating the discussion in chapter 7, imagine that per hectare monetary incentives are offered to nations if they reduce planned forest conversion.) relatively modest carbon prices ($5­10 a ton in 2010, rising 5 percent a year) could, in principle, deter con- 128 D E F o r E s TAT I o N I M P o s E s G E o G r A P h I C A L LY VA r I E D E N V I ro N M E N TA L DA M AG E s Figure 4.2 Deforestation Would Be Unprofitable in Many Land Systems at Modest Carbon Prices Cameroon intensive cocoa vs. secondary forest Cameroon extensive cocoa vs. secondary forest Cameroon short fallow food crop vs. secondary forest baseline Cameroon long fallow food vs. crop vs. secondary forest Cameroon oil palm vs. system secondary forest use Brazil improved pasture vs. forest with nut extraction Land Brazil traditional pasture vs. forest with nut extraction Indonesia rubber agroforestry vs. community forestry Indonesia oil palm vs. community forestry ­1 0 1 2 3 4 5 6 7 8 9 10 11 12 Implicit cost of reducing emissions ($/ton C) Source: Authors' calculations using data from Tomich and others 2005. version of 1­2 million square kilometers of forest by 2050, prevent- ing the release of 8­15 billion tons. A price of $100 per ton of carbon would induce conservation of 5 million square kilometers by 2050, abating the release of 47 billion tons. Forest Loss--Sometimes Irreversible In an uncertain world, it's good to have options. Even if an option isn't worth exercising today, there's a chance it may be worth a lot tomorrow. That possibility is enough to give the option real value today. The theory of options finds sophisticated application in finan- cial markets and investment decision making. It also applies to forests. A forest owner has two options: converting the forest to agricul- ture or maintaining it. In some cases conversion is irreversible, so individuals and nations need to exercise this option wisely. sometimes deforestation is reversible. Deforestation for exten- sive pasture, subsistence cropping, or perennial crops is often 129 AT L o G G E r h E A D s ? followed by field abandonment and spontaneous regrowth. regen- erated forests often recover their biomass and carbon content and the species richness of the original forest--though some or much of the original biodiversity may be lost. That happened in Puerto rico, which lost nearly all its forest only to see it rebound. Carbon densi- ties are now nearly at their original levels, and species diversity is high. But the set of species has changed, with some of the originals lost (Lamb, Erskine, and Parrotta 2005a). But sometimes the outcome is the worst possible: deforesta- tion for short-lived, low-value agriculture or extraction, followed by degradation to a persistent, low-biodiversity, low-carbon grassland or shrubland. These degraded lands do not spontaneously revert to forest, though regeneration can sometimes be induced (Chazdon 2003; Lamb, Erskine, and Parrotta 2005). This destructive pattern can result from "vicious circles"--especially those involving fire (Cochrane and others 1999; Nepstad and others 2001). Deforestation results in fragmented forests with high ratios of edge to area and greater exposure of soil to direct sunlight--leading to drier soils and greater susceptibility to wild and anthropogenic fires. Fires result in higher fuel loads, further increasing susceptibil- ity. repeated fires favor the growth of grass and inhibit forest regen- eration. smoke may also inhibit rainfall, further drying out the soil and increasing flammability (Nepstad and others 2001). The result, in Latin America, southern Africa, and southeast Asia, is a relatively stable grassland system with no tendency to revert to forest (Lamb, Erskine, and Parrotta 2005a). Another vicious-circle mechanism involves loss of mammals that disperse large seeds. In Madagascar, for instance, lemurs are important seed dispersers. Fragmented for- ests support fewer lemurs; fewer lemurs means less dispersal of tree seedlings and hence more fragmentation of forests. These degraded areas cover an appreciable portion of the Earth's surface. Imperata grasslands in southeast Asia are estimated to cover about 350,000 square kilometers (Garrity and others 1996), a bit less than the area of Paraguay. Degraded areas are thought to be large in parts of Latin America. For instance, in long-settled parts of the eastern Amazon, extensive tracts are reported abandoned but unused and appear to be the degraded remains of former forests. Ecologists have identified risk factors for persistent degradation (Chazdon 2003; Lamb, Erskine, and Parrotta 2005a). Geographic risks include areas with poor soil fertility and high susceptibility to ero- 130 D E F o r E s TAT I o N I M P o s E s G E o G r A P h I C A L LY VA r I E D E N V I ro N M E N TA L DA M AG E s sion, due to soil or slopes. risks are also related to the cause and conduct of deforestation. soil compaction from bulldozing or cattle is an important risk. so are repeated fires. Large expanses of deforesta- tion contribute to irreversibility, because natural reseeding is vigorous only within 100 meters of existing forest. Low-productivity pastures-- characterized by fire use, compaction, and large clearings--may be at particular risk of irreversibility. They represent a particularly bad bar- gain: low and temporary returns, little employment generation, large environmental damage, and high probability of irreversibility. Summary Environmental externalities associated with forests are diverse, unevenly distributed, and understood with varying degrees of sci- entific consensus and precision. Table 4.1 arranges environmental externalities in rough order of scale of impact, from global to local. It shows that, carbon aside, most externalities are generated by distinctive and often narrow places and circumstances, ruling out one-size-fits-all responses. Most cases involve different people at the sending and receiving ends of the externalities. Mosaiclands are a hotbed of externality-generating forests, reflecting their rapid defor- estation and the close interaction between forests, agriculture, and people in these areas. Carbon emissions and extinction risk rank at the top of the list of globally important externalities. returning to the example that opened this chapter, this opens a path for the global beneficiaries of forest conservation to compensate those who bear its costs. For instance, a serious global commitment to implementing the Frame- work Convention on Climate Change--which calls for stabilizing greenhouse gases in the atmosphere--would imply benefits for forest conservation that exceed the profits of most current forest conversion processes. Chapter 7 explores the implications of the for- est-carbon connection for global policy. The hydrological impacts of deforestation are extremely sensi- tive to local conditions. In the past, policy was influenced by hydro- logical myths, such as the one that forests generate water. reliance on these myths has led to reforestation with perverse outcomes and may have undermined efforts to mainstream forest protection. Current knowledge suggests that the highest payoff to watershed management occurs within small watersheds, in small steep basins 131 AT L o G G E r h E A D s ? Table 4.1 Externalities of Deforestation Vary by Location of Source and Impact Type of damage Location of deforestation Burden/location of impact Global climate change All deforesting locations; higher Global per hectare damages come from dense humid forests Imminent risk of globally significant Specific areas in mosaiclands and Global, but especially on high- biodiversity loss nonremote frontier forests income populations and future generations Long-term risk of globally Frontier and transfrontier forests Global, but especially on high- significant biodiversity loss income populations and future generations Local and regional climate change Unclear, possibly widespread Unclear, possibly widespread Smoke and smog from forest fires Most areas of rapid deforestation Populated areas downwind of large, rapid deforestation Local flooding, erosion, and Small, steep upper watersheds Small, steep lower watersheds in diminished dry season flows in mosaiclands, nonremote mosaiclands; coral reefs frontier forests, and short littoral watersheds Reduced water quality for drinking Small, steep watersheds near Downstream cities and reservoirs and irrigation cities and reservoirs Loss of pollination, pest control, Mosaiclands; high-density frontier Fields near deforesting locations; and other biological services forests possible far-field effects from which cities draw their water, or along the erodible margins of rivers. Fine-tuning the behavior of a watershed requires attention not just to the presence or absence of trees but also to their place- ment, to agricultural activities, and especially to road placement and maintenance. Native trees will not necessarily be superior to agro- forestry or other kinds of vegetation in achieving hydrological bene- fits. so biodiversity conservation may not be the best way to achieve hydrological benefits. on the other hand, forest conservation moti- vated primarily by biodiversity could pay dividends in hydrological benefits--perhaps even in large river basins. Nonetheless, scientific understanding of hydrological processes is incomplete. Large-scale deforestation could affect regional climate in some circumstances. And the extent to which deforestation could lead to reduced dry-season flows is debated. More scientific and economic research is needed to pinpoint situations where deforesta- tion poses these risks. 132 D E F o r E s TAT I o N I M P o s E s G E o G r A P h I C A L LY VA r I E D E N V I ro N M E N TA L DA M AG E s This is also true for other externalities in table 4.1. Because these externalities can be both subtle and important, solid demon- strations of their magnitude will be needed to motivate policy mak- ers and their constituencies to take action to correct them. Endnotes 1. This point is an adaptation of hamilton and King's (1983) metaphor: "roots may be more appropriately labelled a pump rather than a sponge," quoted in Bruijnzeel (2004). 2. The map does not distinguish areas prone to shallow land- slides, argued above to be most sensitive to loss of forest cover. 3. This term means that, in the absence of intervention, the habitat has a low to medium probability of surviving over the next 15 years. But the WWF classification does not necessarily imply the complete loss of trees, as the simulations assume. 133 Dantanpalli village is inhabited by the Gond tribal group and is part of the Andhra Pradesh (India) Community Forestry Project. In this meeting, the village community and Forest Department staff discuss forest management plans for the coming year. Grant Milne / World Bank. PART II Institutional and Policy Responses Villagers in East Cameroon carrying drinkable water from a facility constructed with funds from forest royalties. © WWF-Canon / Olivier VanBogaert. Inhabitants of this village, near the Lobeke National Park in East Cameroon, usually clear forested land to plant plantains, cocoa and manioc. © WWFCARPO / Peter Ngea. CHAPTER 5 Improving Forest Governance E merging from the first part of this report are two overarching public policy issues that affect equity, incomes, and the envi- ronment: forest ownership and environmental externalities. Finding institutions to grapple with these issues is at the core of bet- ter forest governance. Who Should Have Rights over Forests? Which Rights? Much of the world's tropical forest is under nominal state owner- ship--ownership sometimes disputed by indigenous groups and other forest dwellers. But even forests under community and pri- vate ownership are typically subject to some restrictions on timber extraction or against forest clearance. At stake is a vast amount of real estate, considerable timber wealth, and other assets including minerals, genetic information, and carbon rights. The public policy question is how to equitably adjudicate and efficiently defend these rights. How Should Society Balance Environmental Services against Production of Food, Fiber, and Wood? At all scales of land management, from the farmer's plot to the planet, there are trade-offs and complementarities between produc- tion of food and maintenance of environmental services. Consider a simplified example (with just one environmental service) that 137 AT L o G G E r h E A d s ? Figure 5.1 Optimizing the Mix of Agricultural Output and Biodiversity B production ? C Agricultural A Forest biodiversity Source: Authors. can be applied to all scales (figure 5.1). This production-possibil- ity curve shows different combinations of agricultural production and biodiversity conservation arising from different ways of man- aging land. The curve shows the ultimate technical and biological trade-offs: the maximum production consistent with a given level of biodiversity. At point A all land is devoted to undisturbed forest. Movement along the curve upward and to the left represents conver- sion of forest to agriculture. At first, production is gained with little loss of biodiversity--for instance, by substituting forest gardens for native forest. As more forest is affected, the trade-off becomes steeper. Increas- ingly marginal land is brought into production, increasingly critical habitats are disturbed, and more intensive production results in pol- lution from agrochemicals. Eventually at point B, further conversion to agriculture results in so much environmental damage that agri- cultural production suffers. What combination of production and biodiversity should soci- ety pursue? A society that did not value biodiversity would choose point B, the point of maximum production. A society that valued biodiversity would choose some point along the curve between B and A. The precise point would depend on preferences for agricul- tural production relative to biodiversity. But in real life, societies have incomplete control over landholder behavior and are likely to end up at point C--inefficient for both agricultural production and biodiversity. The policy challenge is to reach societal consensus on 138 I M p rov I n G F o r E s T G ov E r n A n C E a target point between A and B, and then use carrots and sticks to urge land managers toward that target. This challenge occurs at all scales. At the continental and global scale, scientists have worked on prioritizing the world's most unique spots for conservation, seeking a portfolio of locations that occu- pies little area but contains as many different species, ecosystems, and ecological processes as possible (rodrigues and others 2004; Ceballos and others 2005; Burgess and others 2006). This is a search for the flat part of the curve to the right of point B, where major increases in biodiversity are secured for negligible opportunity costs in forgone agricultural production. At the national level, agroecological zoning efforts follow a simi- lar logic, seeking to allocate the most productive lands to agriculture while restricting agricultural uses on land that is marginal for agri- culture and crucial for biodiversity. decisions on regional develop- ment, such as road placement, also determine where a society ends up on the production-possibility curve. And at the local level, many interventions in community management of natural resources and in diffusion of land management technologies can be seen as seeking ways to push landholders away from inefficient points such as C. Balancing Interests while Enforcing Commitments To address these two big issues, society has to find fair ways to balance opposing interests, forge agreements, and commit to those agreements. This is the essence of dealing with environmental exter- nalities (see chapter 4)--a problem particularly salient in less remote mosaiclands. The second problem is fair allocation of property rights. As the frontier expands into the forest, undefended trees, land, and environmental services take on value, and people scramble to claim them. Who should get the rights to these goods? Who or what will guarantee those rights? The same questions arise when communi- ties challenge nominal ownership of forests by governments. These are essentially institutional problems, and they are dif- ficult to address due to imbalances of power, lack of information, and lack of checks and balances. With environmental externalities, typically a relatively few people benefit a great deal from logging or agricultural conversion. Those people are typically influential--often a wealthy elite of loggers or ranchers with close ties to politicians, with continuing deforestation at the top of their agendas. 139 AT L o G G E r h E A d s ? The losers from deforestation--those who bear the burden of environmental externalities--are a large, diffuse, unorganized group. They may not be well informed about the losses because it is diffi- cult to monitor forests, and the environmental impacts of deforesta- tion are hard to track. And even if this group suffers large losses, deforestation may not be at the top of the agenda for each member of the group. These asymmetries of power, information, cohesion, and priority create hurdles to collective action. The hurdles may be even greater when there is a contest for resources between the pow- erful and the voiceless. Institutions exist to mediate these interests and implement agreements: forest codes, zoning laws, logging regulations, courts, and forest services. But sometimes these institutions appear hope- lessly broken. With wealth to be made in forests, regulators can become captured by powerful interests or powerless to intervene. The voices and interests of forest dwellers, far off and disconnected, are not heard. Constituencies for conservation are dispersed and dif- ficult to organize. In many forests remoteness and poor communica- tions have cloaked resource grabs, conflicts, and inequities. Catalytic Innovations in Institutions and Technology A combination of institutional and technological innovations has started to offer some hope for correcting this situation. These new approaches seek to catalyze change by organizing dispersed constitu- encies, improving transparency and information flows, and marshal- ing new counterweights against resource seizures. none of them is, by itself, a panacea. But together they provide an expanded portfolio of tools for addressing what have been almost intractable problems. How Can Institutions Mobilize Domestic Constituencies? In 1995­97 the World values survey asked people in 43 countries if they actively participated in an environmental organization (stein- berg 2005). The top-ranking countries were nigeria (12.3 percent participation) and Ghana (11.5 percent). Environmental participation rates in these and 13 other developing countries surpassed those in Finland, Germany, norway, spain, and sweden. other surveys rein- force this finding: the developed world does not have a monopoly on environmental concern (steinberg 2005). A tougher question is whether the public is concerned specifi- cally about forest conservation. Urban residents may be more con- 140 I M p rov I n G F o r E s T G ov E r n A n C E cerned about local environmental issues such as air pollution, while rural dwellers may favor forest exploitation over conservation. some insight is provided by an opinion survey conducted in connection with Indonesia Forest and Media (InForM), a campaign to promote conservation (Insan hitawasana sejahtera 2003). A purposive, strat- ified sample of three forested provinces and metropolitan Jakarta was evenly split by gender and location (urban or rural). Two-thirds of the 926 respondents were community leaders, the remainder high school and college students. nearly all agreed that "Indonesian for- ests are mostly destroyed," and 90 percent considered local indi- viduals and businesses responsible. At least 90 percent agreed that deforestation was linked to floods, fires, landslides, droughts, higher temperatures, and biodiversity losses. respondents expressed strong opposition to forest burning and were inclined to oppose logging by local governments to raise local revenue (figure 5.2). The group was mildly inclined to permit forest clearance for agriculture, with stronger support in the forested provinces. And despite the publicity attached to corruption in forestry, respondents overwhelmingly supported government control of forests. Less than 15 percent would make forest corruption their first choice for a media campaign. About half said they were willing to sign a petition opposing forest destruction, half said they would boycott products of forest-destroying companies, and a third said they would be will- ing to participate in a demonstration. These proportions were lower but not negligible in the forested provinces. In sum, local Indone- sian opinion leaders are aware of forest loss, concerned about the environmental impacts of forest fires, and often support restrictions on clearing and especially logging. To be heard, though, environmental interests must extend their bases and mobilize political resources. Environmental education is important. one subtle but perhaps catalytic intervention has been the creation of local-language guides to animals and plants. Appre- ciation of the importance of biodiversity conservation is difficult if people do not know what is at risk. The World Bank has sponsored about 100 of these field guides. one way that they may be effective is by increasing both the local demand for ecotourism and the sup- ply of nature guides. Another way to mobilize public support is through individuals and organizations that can frame environmental issues. steinberg (2001) describes how "policy entrepreneurs" and "coupling" institu- tions catalyzed path-breaking environmental policy innovations in 141 AT L o G G E r h E A d s ? Figure 5.2 Indonesians Favor Some Restrictions on Forest Exploitation 1.7 Forest clearance by 2.6 Jakarta / West Java burning should be allowed 2.2 Central Kalimantan 2.2 Jambi North Sumatra 3.5 Forest clearance for 3.4 mining should be allowed 4.5 3.7 Forest clearance for 4.7 agricultural land or 5.2 infrastructure should 5.7 be allowed 4.9 Large-scale logging 2.5 by local government 2.8 to raise local revenue 3.4 should be allowed 2.8 People should be 3.0 4.4 allowed to cut trees 4.6 for a living 3.6 1 2 3 4 5 6 7 8 Disagree Agree Source: Insan Hitawasana Sejahtera 2003. Bolivia and Costa rica. (A similar argument could be made for Bra- zil.) Internationally linked scientific research organizations served as incubators of expertise and action. They fostered the develop- ment of local capacity in ecology, created networks of domestic and foreign scientists and environmentalists, and fostered a nonpartisan atmosphere where policy entrepreneurs could draw on scientific findings to formulate locally relevant proposals. The results? sky- rocketing local appreciation of conservation--and a long list of glob- ally influential local policy innovations in environmental finance and management. Forging links between civil society and government is another way to mobilize environmental constituencies. one expression of this is the rapid rise in the number and prominence of environmen- tal and other nongovernmental organizations (nGos) over the past decade (steinberg 2005). 142 I M p rov I n G F o r E s T G ov E r n A n C E Another way to mobilize is by incorporating civil society input into government activities. In the philippines provincial and local multisector forest protection committees were created as part of a World Bank environmental adjustment loan (Cruz and Tapia 2006). The committees, which included participation by civil society groups and the national resources department, were charged pri- marily with monitoring forests but also with evaluating policies and operations and conducting information campaigns. Credited with reducing illegal logging, there were 314 such committees in 1999, when the loan closed. But many subsequently collapsed when fund- ing ended, raising questions about the depth of popular support. Brazil's local environmental councils offer an interesting view on the links between civil society and government. Local govern- ments in Brazil, urban and rural, are organized around the country's more than 5,000 municipios governed by locally elected mayors and municipal councils. Municipal government also allows for advisory councils focused on certain sectors, including environment. Because Brazil presents continental-level variation among municipios in aver- age income, education, size, rurality, and environmental conditions-- while holding constant national laws and institutions--it provides an opportunity to look for evidence on the determinants and impacts of local environmental institutions. The presence of an active environmental council is strongly related to income and education. (An active council is defined here as one that meets at least once a year, and at least half of whose members are from civil society.) Assume that the presence of an active municipal environmental council indicates environmental participation. About 14 percent of Brazil's 5,500 municipios meet this description. Councils are far more common in wealthier, better- educated municipios (29 percent) than in the poorest, least-educated municipios (6 percent). Multivariate analysis confirms that this is not merely because rich municipios are larger and thus have a larger pool of recruits. Although size and urbanization affect the likelihood of an active environmental council, mean income and education have a strong independent effect. It is difficult to determine whether active environmental councils are effective in bringing forest-related environmental issues to the attention of local governments. one problem is forest and land fires. Used for forest conversion and pasture management, fires can create serious problems when they get out of control, damaging neighbor- ing fields, fences, and woods (nepstad and others 2001). Among 143 AT L o G G E r h E A d s ? municipios that experienced fires (based on remote sensing data), a substantially higher share of those with active environmental coun- cils reported a fire problem (28 percent versus 18 percent) and a smoke problem (23 percent versus 15 percent). ongoing research is examining whether the council has a causal impact in prompting recognition of these problems. Revolutions in Monitoring Have Raised Awareness and Accountability For a long time forests have been invisible and their dwellers inau- dible. It has not been easy for the public and the law to detect defor- estation, logging, or mining deep in the forest. The scope of forest destruction and private appropriation of public property has gone unnoticed. so too has the extent to which public agencies charged with protecting forests have done their jobs. For most large for- ested nations in the developing world, reliable data on deforestation are lacking even at the aggregate level--let alone the provincial or regional level. All that is changing due to synergistic developments in institu- tions and technology. These have the potential to drastically cut the cost of monitoring forest activities and to empower civil society to use this information to more fairly balance forest interests. The first revolution involves remote sensing. satellite images can detect deforestation and logging. since the debut of Landsat in 1972, image quality and frequency have improved while the costs of acquiring, interpreting, and using images have plummeted. For detailed monitoring of particular sites, it is now possible to order snapshots with 1 meter resolution. For monitoring of global forests, ModIs images cover the entire world daily, can detect land cover changes as small as 25 hectares, and are available free of charge. The costs of hardware and software for analyzing and using images have also plummeted, placing them within reach of small nGos. The social and political impacts of this technology are becoming evident. In the developing world, Brazil has led the way in technol- ogy and applications. Its national Institute of space research (InpE) has long published annual or biannual reports on Amazônian defor- estation by state. These reports have helped focus national and international attention on Amazônian deforestation. More recently, InpE has started publishing on the Internet real-time images of fire locations and detailed (30-meter resolution) maps of annual 144 I M p rov I n G F o r E s T G ov E r n A n C E deforestation. It also provides south Americans with free data from CBErs2, the sino-Brazilian satellite. These data are used by govern- ment agencies to enforce land regulation and by local nGos to draw attention to forest issues and galvanize public support. Land and forest fire monitoring uses an even more accessible technology, providing nearly real-time results that can be used for fire prevention and control. ConABIo, Mexico's national Biodiver- sity Commission, began monitoring after the disastrous fires of 1998. Indonesia's space agency also provides nearly real-time information on fire occurrences and risks. Monitoring information of this kind can become much more valuable when combined with information on forest ownership and control. A complementary technological revolution--geographic positioning systems--makes it possible to identify the boundaries of properties and concessions at minimal cost. A third revolution-- cheap geographic information systems--makes it easy to overlay maps of deforestation on maps of property boundaries. This tech- nology enables government enforcement agencies to do their jobs better, and civil society to make sure that they're doing their jobs. For instance, prosecutors can use remote sensing images as evi- dence of illegal deforestation. The Brazilian state of Mato Grosso has set up a system that registers the location of large properties and uses remote sensing to track their compliance with land use regulations. In Cameroon nGos are using remote sensing to cor- relate the construction of new logging roads with logging conces- sionaires' reports of timber extraction (Global Forest Watch 2005). Mismatches may indicate mischief. roads without logs may mean that producers are evading taxes. Logs without roads suggest timber laundering--for instance, taking timber from a protected area but claiming it comes from a legal concession. In neither case has offi- cial enforcement been entirely successful. But the ability of outside groups to monitor the behavior of private parties and government may put pressure on both to comply with laws. As information gets better and cheaper, new possibilities emerge. several groups are working on ways to use ModIs to cheaply detect large-scale deforestation, at an annual frequency or better, for entire nations or even the world. At the global level this would be a quan- tum leap in tracking deforestation. Among developing countries only Brazil and India regularly report remote sensing information on forest cover, though Indonesia is creating such a system. 145 AT L o G G E r h E A d s ? At the national level such a monitoring system could detect hot- spots of deforestation rapidly enough to trigger action. It could be used, for instance, to track the impacts of new road construction or macroeconomic policy changes. It could also be used to direct higher-resolution monitoring for enforcement purposes. Technolo- gies to detect hidden logging have been demonstrated by Asner and others (2005), and progress is being made in the use of satellite- based cloud-piercing radar. Complementing the use of remote sensing is the rapidly growing potential for participatory, ground-based observations by citizens. These observations could be used to help interpret satellite images and to provide information unavailable from the sky. The Confluence project (www.confluence.org) provides a hint of the possibilities. It has asked for volunteer observations of the world's latitude and lon- gitude intersections, at 1-degree intervals. The map is rapidly filling in. The explosive growth of cell phone coverage is rapidly putting a lot of mosaic forest within instant communication and reporting range. Already, about a quarter of the world's "imminent extinction spots" (see map 1.8) are covered by GsM cell phones.1 Weak State Institutions Can Be Aided by Better Checks and Balances and Transparency Institutions charged with enforcing forest laws are often ineffective. Even worse, they may be captured by the interests they're supposed to regulate. This is a grave risk when large amounts of money are at stake--as when state agencies allocate land or forest concessions or are charged with ensuring that industrial loggers and large land- owners comply with environmental regulations. Corrupt officials, legislators, and military officers can form alliances with large actors (including timber companies, pulp mills, ranchers, and plantation owners) to allocate land and forests for exploitation. The result is private appropriation of wealth that belongs to the public or local communities, conflict with forest dwellers, and unregulated forest destruction. In response, an efflorescence of institutional innovations have been created to bolster the performance and accountability of gov- ernment agencies and the interests they oversee. It is possible to strengthen a system from within. Akella and Cannon (2004) explain why forest law enforcement often fails. Landowners are deterred from illegal deforestation or logging only if they perceive a signifi- cant probability of a significant penalty. In a system that relies on 146 I M p rov I n G F o r E s T G ov E r n A n C E criminal penalties, a long chain of events must occur before a mis- creant is punished: detection of the legal violation, citation, pros- ecution, conviction, and execution of the penalty. If landholders perceive low probabilities of progressing from any link in this chain to the next, the level of deterrence is low. Links may be weak by design or for lack of capacity. palmer (2005) describes how Indonesian logging regulations motivate enforcers to seek bribes from log smugglers, rather than prosecute them. on the other hand, Brito, Barreto, and rothman (2005), in a review of Brazilian environmental crime law enforcement, identify fixable logistical problems as an impediment to prosecution. Against a global backdrop of failed systems of internal checks and balances, Brazil's public Ministry provides an interesting model for a possible solution. The ministry, which exists at both state and federal levels, is a prosecutorial agency charged with ensuring legal compliance by both citizens and the executive branch. A meritocratic institution, its staff are selected through competitive exams that only a few percentage of applicants pass. As a result it attracts extremely qualified and idealistic staff, many of whom are interested in envi- ronmental issues. prosecutors have considerable autonomy in choos- ing cases to pursue. This promotes independence but impedes focus. Typically the ministry seeks to resolve problems through negotia- tions, holding the threat of prosecution in reserve. Brito, Barreto, and rothman (2005) call the public Ministry the most powerful institu- tional force for environmental protection in Brazil. A new institution of independent monitors stands at the bound- ary between strengthening internal controls and enabling external ones. The governments of Cambodia and Cameroon, under inter- national and domestic pressure to strengthen oversight of forest resources, have employed donor-funded independent monitors of forest law enforcement. A concern in both cases was that timber wealth was being nontransparently and inequitably allocated, and that loggers were not adequately supporting the sustainability of for- est resources. In Cameroon a monitor observed the conduct of concession auc- tions, spurring an increase in bids and better application of techni- cal standards for prequalifying bidders. In both countries monitors examined logging operations and enforcement actions. A review of these and similar experiments by Brown (2004) found strong posi- tive impacts on transparency, but questioned the sustainability of the monitoring institutions. The fundamental issue is whether there 147 AT L o G G E r h E A d s ? is a domestic constituency that values and demands the information provided by the monitors. A variant of public disclosure policies from the field of industrial pollution control may offer lessons. Indonesia's propEr program rates the pollution control efforts of industrial firms. The program was set up by the country's environmental protection agency in response to difficulties in enforcing pollution laws. Based on audited self-reports, it classifies firms on a five-point scale: completely non- compliant and making no effort to comply, some environmental effort but inadequate to meet standards, minimally compliant, good practice, and best practice. These ratings are easily understood by the public and have induced firms to improve their performance. The ratings' interesting feature, relative to current practice in forest law, is the recognition that propEr gives to better-performing firms. This may help defuse opposition to the program. voluntary certification systems share similarities with public disclosure systems. From a policy perspective, certification systems are appealing because they can encourage better forest manage- ment even where local institutions are ineffective at enforcing regu- lations. Like public disclosure systems, they seek to reward good performers--though usually only on a pass/fail basis, without the finer distinctions made by propEr. The best-known examples are for forest management, where the Forest stewardship Council and other standards-setting organizations have developed standards for responsible, sustainable forest management. These standards include compliance with national laws, respect for indigenous rights, conservation of biodiversity, and establishment of and com- pliance with a management plan. systems have also been proposed to certify that commodities such as beef and soybeans are produced without illegal deforestation. Certification is conducted under con- tract by accredited private, third-party certifiers. The integrity of the certification process rests on the desire of the certifiers to maintain their reputations. Can certification make a big difference in forest management? The main question is whether forest owners will find it worthwhile to seek certification. Certification imposes substantial direct and indirect costs. Because there are direct fixed costs associated with filling out paperwork and paying for a certifier's visit, community forests and other small producers are at a serious disadvantage. Indirect costs are those associated with compliant behaviors-- such as refraining from cutting timber on slopes. These costs can 148 I M p rov I n G F o r E s T G ov E r n A n C E be substantial, depending on the stringency of regulations and the nature of the forest. The most widely cited benefit for forest own- ers is increased access to export markets, possibly with a price pre- mium for certified products. skeptics doubt whether this benefit is wide enough and deep enough to motivate widespread change in forest management. demand for certified products is only a small (but rapidly growing) portion of export markets, accounting for 12 percent of wood production in Africa and 18 percent in Asia. Moreover, it is controversial whether there is any price premium for certified wood. A survey in the U.K. market finds price premiums of 2­3 percent for some tropical woods--and premiums of 20 per- cent in thin markets where demand at such prices may be lacking (robinson 2006). But if passed back to the producer, even a small retail or wholesale premium can translate into a large stumpage pre- mium, and certification may be important for certain markets. still, questions remain on whether certification can be expected to influ- ence producer behavior on a large scale. But certification may change firms' behavior through another mechanism. Because certification criteria are consistent with risk and liability reduction and with the existence of good internal man- agement controls, certified loggers and landowners may find it easier to obtain insurance and financing. For the same reason certi- fication--seen as a proxy for good management and low risks--may increase the value of a forest concession or property, or of a log- ging company. This avenue may prove to be a stronger incentive than a consumer price premium. And there could be indirect effects through local politics, as certified companies seek to ensure that uncertified competitors also comply with local regulations.2 Finally, anti­money laundering laws are beginning to attract attention as a tool against illegal logging and forest conversion. Intergovernmental bodies--the Financial Action Task Force and associated regional bodies--have offered recommendations on these laws, which designate certain crimes as "predicate" crimes. disguis- ing the movement of gains from predicate crimes is a money laun- dering offense. Indonesia has explicitly designated illegal logging as a predicate crime; in many other countries violations of forest or land use laws could be interpreted as predicate crimes. What this might mean is that a much broader net can be cast for violators of forest law. domestic law enforcement agencies have another tool at their dis- posal: money laundering crimes may be easier to detect and pros- 149 AT L o G G E r h E A d s ? ecute than forest law violations. Money laundering crimes require domestic banks to exert closer scrutiny of their clients, deterring crime. Foreign banks must scrutinize their correspondent banks as well as deposits by offshore clients. But application of money laun- dering laws to forest law enforcement is still at an early, speculative stage. Summary This chapter looked upstream of the policy process. If societies are to maintain environmental services in the face of strong pressures for forest degradation, there must be vocal and effective constitu- encies for such services. And if societies are to fairly allocate and defend rights to forest resources, they must prevent powerful elites from seizing them. new institutions and technologies for transpar- ency, monitoring, and incentives can help address these challenges. With this context, the next chapter examines successes and failures of policies affecting forest management and protection. Endnotes 1. Authors' calculation based on 2005 coverage data. 2. The contrary is also possible. As pressures for certification increase on large companies, they may transfer forest assets to small companies that evade certification or are subject to laxer standards. 150 Road building and agricultural development projects triggered forest clearance in Santa Cruz, Bolivia. UNEP / GRID-Sioux Falls. Most of Mexico's forest is owned by communities, and many have developed viable commercial forest enterprises. Heriberto Rodriguez. CHAPTER 6 Local and National Policies: Framing Rights and Incentives for Forest Management W e turn now to twin governance challenges introduced in chapter 5. These challenges revolve around the assignment and enforcement of rights: · In frontier and transfrontier areas much of the for- est is nominally owned by the state--but the reach of government and the rule of law are weak and property rights insecure. Who should control these vast tracts and receive revenues from them? And as demands on these areas increase, how can large-scale environmen- tal disruption be prevented? · Within mosaiclands, how can land and forest be man- aged productively and equitably? For instance, how should society balance landowners' desire to exploit rich riverfront soil against downstream neighbors' interest in maintaining riverrine forests as a bulwark against sediment? This chapter examines ways of improving forest outcomes, such as: · Sorting out who has rights to use forests for different purposes, and how stringently to regulate those rights in the interests of sustainability and environmental benefits. 153 AT L o g g e R h e A d S ? · Making sustainable forest management more finan- cially attractive relative to agricultural conversion. · Coordinating regional development and agricultural policies. The chapter first reviews the challenge of forest conflict, then assesses zoning and land use allocation mechanisms at the national and regional scales. After that it analyzes the pros and cons of various property rights schemes--government management, com- munity management, private management--under different circum- stances, and for each examines ways of making forest management more attractive. Finally, the chapter considers how to manage road building and other development policies to take into account their impacts on forest management. The Challenge of Forest Conflict Violent conflicts are endemic to forests. According to the FAo (2005, p. 117), over the past 20 years at least 26 tropical countries have experienced armed conflicts in forested areas; another 4 experi- enced "substantial social violence." Some of these have been civil wars, reflecting failed states. Many rebels and insurgents have used forests as a base of operations and timber sales to finance fighting. Resolving these disputes involves governance issues far beyond the forest agenda. In stronger states conflicts can erupt in frontier areas over access to land and timber. In an analysis of violence in Brazilian Amazônia, Alston, Libecap, and Mueller (2000) illustrate some basic challenges in allocating property rights at the frontier. during the 1990s there were frequent clashes as landless groups and large landholders dis- puted property ownership. According to the authors, these clashes stemmed from legal ambiguity. one set of laws guaranteed land ownership to the property holders, while another allowed redistri- bution of "underutilized" land (including forest) to landless people. The authors asked why this dilemma leads to conflict in some places and not others. They concluded that clashes are more likely when land is valuable and there is uncertainty about which laws will prevail. It's not worth fighting over valueless land, and it's not worth fighting for a lost cause. good laws and good judges are the basis for property rights at the frontier. But this maxim, easily enun- ciated, is hard to put into practice. 154 LoCAL ANd NATIoNAL PoLICIeS Forest Rights and Restrictions--A Range of Possibilities "Who owns the world's forests?" asked White and Martin (2002) in a seminal piece. Appendix table A.3, drawn from International Trop- ical Timber organization (ITTo) (2006) and United Nations Food and Agriculture organization (FAo) (2005), updates their answer, but reaches many of the same conclusions. First, much of the world's tropical forest is under insecure state ownership and much is managed unsustainably--or not at all. gov- ernments own almost all forest in Africa and Asia and most in Latin America. Some 3.5 million square kilometers of forest in ITTo mem- ber countries is zoned for timber production, with about half under government concession. But only 7 percent is under sustainable management, and only 3 percent is under certified management. Another 4.6 million square kilometers of ITTo forest is zoned for protection (some on private lands), but just 4 percent is under some kind of management plan. Second, a large and growing portion of the forest estate is owned or managed by communities. Local and indigenous groups own most forests in Fiji, Mexico, and Papua New guinea; own large tracts in Bolivia, Brazil, Colombia, ecuador, guyana, Panama, and Peru; and co-manage forest in guatemala, India, and the Philip- pines. In Southeast Asia communities manage 10 percent of pub- licly owned forest and have limited user rights to another 54 percent (Romano and Reeb 2006). But forest ownership and management rights are almost always restricted, and restrictions on ownership and use define alternative tenure systems (table 6.1). These two dimensions mirror the two main challenges identified in chapter 5--assigning ownership and recognizing environmental externalities. The balance of rights can be tilted strongly toward society, in the form of publicly owned, strictly protected areas. or state ownership and management can be retained, but with sustainable timber extraction allowed. Much of the world's tropical forest falls into this second category, under either direct (often ineffective) state supervision or concessions. But community participation in forest ownership and management is also growing quickly, though still with restrictions on extraction and conversion. In contrast, most or all forest rights--including to conversion-- can be allocated to private owners, as in many mosaiclands. In such 155 AT L o g g e R h e A d S ? Table 6.1 Examples of Forest Ownership and Use Restrictions Ownership or management type Allowed forest uses State Community Private Any, including full -- -- Many mosaiclands conversion Limited permanent Zoned areas for -- Brazil, Paraguay conversion permitted; conversion (Indonesia) remainder must be managed sustainably Conversion prohibited; Direct state management Community concessions Regulated private forests sustainable management (often ineffective) (Guatemala), Joint Forest for commercial forest and industrial timber Management (India), products permitted concessions; includes a community forests lot of Asian and African (Mexico), indigenous forests lands (Papua New Guinea) Limited or no extractive Strictly protected areas Some indigenous areas Private reserves and uses permitted conservation easements Source: Authors. cases society must compensate landholders to get them to mitigate carbon emissions, sedimentation, and other consequences of defor- estation. Where conversion rights are limited, the rights of private forestholders are circumscribed so that they share with society the costs of environmental protection. In Brazil and Paraguay, for instance, landholders must keep a proportion of their property under forest cover. Making this obligation tradable could, however, reduce the costs of achieving environmental goals. Finally, in some places landholders retain only a few rights to forest use, such as establish- ing a residence or conducting ecotourism. examples include private reserves such as Brazil's Reserva Particular do Patrimônio Natural (RPPN). Choices among these tenure systems will reflect an area's his- tory, the relative power of different interest groups, the efficiency of different groups in managing forests, the importance of environmen- tal protection relative to agriculture, and societal attitudes about pri- vate and social interests. The next few sections discuss mechanisms for making those choices and examine issues related to implement- ing these alternative systems. 156 LoCAL ANd NATIoNAL PoLICIeS Zoning Has Technical Appeal but Poses Practical Difficulties Zoning has a sensible premise: efficient land allocation and man- agement. Some land is suited to agriculture, with flat terrain, fertile soils, favorable climates, and available water. Some is terrible for agriculture, as on erosion-prone hillsides. Similarly, some areas have unique animal or plant species or play a crucial role in moderat- ing water flows. So land use planners suggest dedicating good agri- cultural land to farming, keeping crops and cows away from easily degraded lands and protecting areas of high biodiversity. Sophisticated land use planning methodologies have been developed, at scales ranging from continent to nation to province to watershed. There are at least two strands of technical planning, though in practice they may be combined. one is rooted in agricul- tural science and forestry. Information on topography and soils is used, together with crop modeling, to indicate the "vocation" of the soil--that is, the recommended land uses. data on forest cover and distributions of tree species and human populations are also used to indicate areas for sustainable timber management. This approach is common in Latin America and was also used in Malaysia. The second approach comes from systematic conservation plan- ning (Margules and Pressey 2000; Cowling and others 2003; Stoms, Chomitz, and davis 2004). This highly technical approach is framed as a mathematical optimization problem: finding the landscape con- figuration that achieves specific environmental goals at minimum cost. For instance, given a set of potential reserve sites in a region, a planner may try to identify the smallest number containing all of the region's threatened species. More sophisticated formulations try to ensure that there is enough contiguous habitat to ensure the long-term survival of those species and to maintain broader ecological processes. The resulting zoning plans can be indicative or prescriptive. Indicative plans can be used by landholders to choose appropriate land uses, or by governments to decide on road placement, estab- lishment of protected areas, granting of permits for mines and plan- tations, and other regional development issues. Prescriptive plans dictate which land uses are permitted or prohibited at each spot on the landscape. When environmental externalities are the motiva- tion, plans are often prescriptive, identifying specific hillsides, flood- plains, biodiversity corridors, and wildlife habitats for protection or restrictions on use. 157 AT L o g g e R h e A d S ? The promise of prescriptive zoning is also its pitfall: to achieve its social and environmental goals, it must restrict the rights of cur- rent or prospective landholders to use the land or forest. The legit- imacy and effectiveness of zoning are thus closely linked to land tenure and depend on securing landholder consent and cooperation. Poor people can suffer if zoning is imposed on them without con- sent or compensation, while wealthier or more powerful interests may flout the rules with impunity--or there may be no political will to impose zoning on anyone. For this reason, implementation of zoning has been problematic at the national level (hoare 2006): · In Suharto-era Indonesia, the government asserted claims to a forest domain encompassing about three- quarters of the country, superseding the rights of traditional communities to their forests. A large-scale zoning plan delineated areas for protection, timber management, and conversion. But the plan was often disregarded or manipulated. Communities were denied rights to agroforests they had created, and the plan did not prevent deforestation of protected areas. either because of poor planning or subsequent deforestation, 40 million Indonesians live in areas zoned for forest but lacking trees--areas where agriculture is not allowed. · In Brazil two expensive World Bank­funded exercises developed zoning plans for the states of Rondônia and Mato grosso. The plans were devised without much popular participation or political buy-in. Because they placed significant areas off-limits to powerful ranching and timber interests, they failed to gain widespread support and apparently did not have much effect on land use (though no rigorous evaluation has been conducted). A revised plan is being implemented in Rondônia (Mahar 2000; World Bank 2003). · Agroecological zoning has been undertaken at the pro- vincial or state level in several Latin American coun- tries, often with support or methodology from the FAo. These include large exercises in Bolivia and Peru and regional or local exercises in Chile, Colombia, Costa 158 LoCAL ANd NATIoNAL PoLICIeS Rica, Nicaragua, and Paraguay. No formal evaluations of these exercises have been produced. · Conservation science has produced a number of elabo- rate land use plans that try to reconcile biodiversity, agricultural, and commercial forestry goals, such as one for Papua New guinea (Faith, Walker, and Mar- gules 2001). Yet according to Faith and others (2003 p. 313), "In spite of a decade or more of work on reserve selection methods, no complete set of areas produced by such computer algorithms, to our knowledge, has been implemented anywhere in real-world regional bio- diversity planning."1 In sum, these plans have failed because they did not muster popular support and did not consider how people with claims or designs on forests would react. Zoning with a Human Face? It is not enough to draw up a rational zoning map at 1:250,000 scale. Planners must also induce land users to conform to the map--but how? They can try exhortation, which may work if the maps pro- vide novel information about agricultural suitability or if the target audience is a tight-knit group with strong social controls. They can try legal compulsion, which requires public consensus on means and ends if it is to be legitimate and effective. They can provide incentives for compliance, including compensation for accepting restrictions. And perhaps most important, they can draw on popu- lar participation when shaping plans and negotiating the land rights that often underlie them. That means adjusting the map to recog- nize reality on the ground, rather than vice versa. An example is Cameroon, which has zoned its permanent forest estate to reflect land use patterns. The 1994 forest code mandated reserving 30 percent of the country as permanent forest estate for conservation and sustainable timber production. A preliminary zon- ing plan achieved that goal largely by reserving areas with dense forest and few people. Although the permanent estate is state prop- erty, communities inside it can manage local forests, are entitled to half the revenue from nearby timber concessions, and can challenge and redraw the boundaries of those concessions. 159 AT L o g g e R h e A d S ? The zoning plan is thought to have deterred agricultural con- version in the permanent forest estate, though no formal evalua- tions exist and pressures may be low. But critics say that the plan was insufficiently participatory, did not adequately recognize the needs of indigenous forest dwellers, sometimes deprived commu- nities of traditional rights, and was sometimes treated as immu- table rather than subject to objection and renegotiation (hoare 2006). oyono (2005), however, suggests that the zoning plan and associated legal changes are improvements over the previous de facto rights regime. Another approach is more systematically participatory. It starts, as do the technical exercises, with maps of the landscape and its resources. Thanks to new technology, such maps are becoming rela- tively cheap and easy to assemble. They may vary in sophistication from simple paper maps to complex geographic information systems with decision support software. Residents and claimants review the maps, delineate historical claims, negotiate boundaries, and discuss issues that require coordination. Mediation and conflict resolution are important parts of the exercise. There has been an efflorescence of this participatory land use planning. It is often used to help demarcate indigenous lands. For instance, it is being used to resolve conflicts between forest dwellers and plantation interests in Papua province, Indonesia, and to delin- eate community boundaries in Vietnam. Successful applications have also been reported in Cameroon (Lescuyer and others 2001) and Madagascar (Cowles and others 2001). over the past decade Australia has instituted 10 regional forest agreements to zone public forests with commercial timber poten- tial (hoare 2006). The agreements set conservation goals that try to achieve a "comprehensive, adequate, and representative reserve system" (www.rfa.gov.au). To realize conservation goals while also taking into account the interests of forest dwellers, indigenous people, and forest industry groups, the agreements were created through extensive consultations. The process started with substan- tial investment in gathering and mapping information on social and environmental values of forests. This information was gathered in a participatory fashion and provided the basis for stakeholder nego- tiations. hoare (2006) cites studies (published before 2000) that credit the regional forest agreements with increasing the nation's reserve network and increasing stakeholder involvement, especially of indigenous people. But she concludes that stakeholder participa- 160 LoCAL ANd NATIoNAL PoLICIeS tion could be further improved and final decision making on forest allocations made more transparent. While participatory land use plans are promising, they face two implementation challenges. The first is legitimacy. Who participates in the participation, and who do they represent? If local govern- ments are deemed unrepresentative, what alternatives are better? how can capture by elites be avoided? Is there a solid legal basis for land right allocations and restrictions? The second challenge involves commitment and enforcement. how will agreements be enforced, and how will future disputes be negotiated? Actions such as physically marking boundaries can help prevent disputes. But if there is a disagreement between local com- munities and powerful industrial groups, what will keep the power- ful from capturing the machinery for dispute resolution? There are no easy answers to these questions. For a while, anyway, partici- patory land use planning will remain a novel endeavor--one from which it is essential to quickly learn lessons. Public Management of Forests: Protected Areas and Concessions Publicly owned forests are often poorly regulated and administered. This section examines the two most prominent approaches to public administration of forests: protected areas and regulated concessions. The next section considers an alternative: devolving ownership or some management responsibilities to communities. Protected Areas Are Expanding Quickly Protected areas represent, in extent and financing, the largest policy intervention for conservation and active management of tropical for- ests. FAo (2001b) estimates that 3.46 million square kilometers of tropical and subtropical forest have protected status--about a sev- enth of the world's forest and approximately equivalent in area to India. over the past 20 years the number of protected areas and the area under protection have grown rapidly (figure 6.1). A full accounting on spending to establish protected areas in tropical forests is unavailable. But during 1992­2002 the global envi- ronment Facility financed $3.6 billion in projects for protected areas, covering about a quarter of the world's protected areas. Across the developing world, total annual spending (including recurrent spend- ing) on protected areas is roughly $800 million. 161 AT L o g g e R h e A d S ? Figure 6.1 Protected Areas Have Grown Rapidly in Tropical and Subtropical Forests 350 VI ­ Managed resource protected area 300 V ­ Protected landscape/seascape IV ­ Habitat/species management area ha 250 III ­ Natural monument of II ­ National park 200 Ia ­ Strict nature reserve/wilderness area millions in 150 Area100 50 0 1920 1930 1940 1950 1960 1970 1980 1990 2000 Year Source: Kloss 2006. Note: Excludes areas for which establishment dates are not available. Are Protected Areas Protected? Though sometimes denigrated as "paper parks" because of their poor funding or management, protected areas may be more effective than is commonly thought. Remarkably, despite their flagship role in conservation, there is little quantitative analysis of their effec- tiveness in protecting biodiversity. But some remote sensing studies suggest that protected areas may deter deforestation. Nepstad and others (2006) compare deforestation inside and outside the bound- aries of protected areas in Brazil. (This is a clever way of controlling for differences in soils, market access, and other confounding fac- tors.) The authors consistently find much higher deforestation rates outside, suggesting a strong protective effect. A similar study of Costa Rica also found a strong differential (Sanchez-Azofeifa and others 2003). Remote sensing images such as those of Laporte show intact protected areas surrounded by a sea of agriculture.2 Many of the spatial analyses reviewed in chapter 2 also find that deforestation is lower in protected areas, holding constant accessibility, agroclimatic conditions, and other factors. gorenflo and others (2006), for instance, find that even in Madagascar's weak institutional setting, parks appear to reduce conversion. 162 LoCAL ANd NATIoNAL PoLICIeS But against this generally positive view of park effectiveness, there are examples of ineffectiveness. Curran and others (2004) document rapid, massive deforestation in the protected areas of Kalimantan, Indonesia. Moreover, biodiversity can be damaged in ways that are undetectable by remote sensing--as with hunting of large mammals. only a couple studies explore reasons for variations in the effec- tiveness of protected areas. Bruner and others (2001) and dudley and others (2005) survey such areas, correlating management prac- tices with self-reported measures of park conditions. The clearest result is a correlation between staffing and effectiveness, suggest- ing that guards are an important part of the transformation between "paper parks" and working parks, though staff may also be impor- tant in working with local residents. Can Protected Forests Sustain Livelihoods? Park creation has sometimes been associated with reduced for- est access for local people (see chapter 3). In response, there has been a trend to actively engage local and indigenous populations in comanagement and sustainable use of protected areas. Kloss (2006) shows that new protected areas are less likely to have strict pro- tection (World Conservation Union categories I­III) than to allow multiple uses (IV­VI; see figure 6.1). The 1990s saw striking growth in category VI, which is land managed not just for biodiversity but also for "a sustainable flow of natural products and services to meet community needs," according to the IUCN (1994, p. 23) definition. There has also been an extensive--but largely unevaluated-- effort to seek win-win outcomes through integrated conserva- tion-development projects (ICdPs). These projects aim to boost development in forest communities, often those in or near protected areas. development of an ICdP is often based on several premises: that poor people are the main agents of forest degradation, that pro- vision of higher incomes or alternative income sources will reduce deforestation by poor people, that project-based interventions can stimulate long-term sustainable improvements in livelihoods, and that communities can credibly commit to relinquish future use of a forest in exchange for current compensation. All these premises are subject to debate (Fisher and others 2005). First, ICdPs won't reduce deforestation if targeted communities are not to blame for deforestation. A review of Indonesian ICdPs found that local communities were bystanders to ongoing defores- tation by wealthy timber and plantation interests (Khan and others 163 AT L o g g e R h e A d S ? 1999). Second, there is no strong reason to expect that unconditional provision of alternative livelihoods will automatically reduce a com- munity's pressure on forests and other natural resources. Chapter 2 shows that higher incomes and increased agricultural productivity often increase deforestation, not retard it. Third, while ecotourism and nontimber forest products can motivate conservation and raise incomes, it can be difficult to set up these businesses. Some research- ers have concluded that ICdPs can succeed only if there is a specific quid pro quo bargain--such as periodic payments to communities based on measured conservation outcomes (Ferraro and Kiss 2002). A recent review by the global environment Facility supports these propositions (geF 2006). The review examined the impact on local incomes of 88 biodiversity projects, mostly in protected areas (but not all forests). Less than half of projects for which information was available succeeded in boosting incomes (table 6.2). Not sur- prisingly, alternative income generating programs often failed when they were not financially viable. Moreover, financial success did not guarantee environmental success when the new business was unre- lated to the natural resource at risk. ecotourism ventures were more likely to prosper in areas with tourism infrastructure. Such ventures required sophisticated skills and often benefited wealthier community members. Forests, with their shy and elusive wildlife, tend to offer less spectacular tour- ism experiences than savannas with large mammals. Projects based Table 6.2 Integrated Conservation-Development Project Interventions Have a Mixed Record Type of intervention Outcome Alternative income Ecotourism Sustainable generating activities resource use Success in boosting 17 21 11 incomes Failure 19 25 22a No information 15 23 Not applicable 37 19 55 Source: GEF 2006. Note: Shows the results of an evaluation of 88 biodiversity projects. Some projects supported more than one type of intervention. a. Breakdown between "failure" and "no information" not provided. 164 LoCAL ANd NATIoNAL PoLICIeS on sustainable resource management were successful when they built capacity to care for attractive common property. For instance, a Ugandan project supported regulated beekeeping in the forest and so motivated the community to prevent forest fires. elsewhere, strict park regulations were often an insuperable legal barrier to sustain- able resource use. overall, though, the evidence on ICdP impacts is weak, reflecting a lack of systematic monitoring and evaluation (Agrawal and Redford 2006). Protecting Additional Areas Is Difficult--And Other Options Exist What is the scope for creating additional protected areas? Where are they most appropriately situated? From a conservation view- point--the demand side--the need is most urgent where unique biodiversity (species, ecosystems) is under threat and fares poorly in human-dominated landscapes. gap analysis points to places with these characteristics. Many are in nonremote areas where economic pressures on forests are likely high. (Though Brandon and others [2005] argue that opportunity costs are low in "gap" areas in Mex- ico, which is rich in biodiversity.) on the supply side, protected areas have traditionally been cre- ated in remote areas where economic demands on the land have been weakest. This trend has not changed over the past three decades (figure 6.2). While sometimes considered opportunistic, protecting large remote forests under low current threat may be a far-sighted and cost-effective way of averting their long-run fragmentation and degradation. In addition, maintaining large forest blocs provides an important hedge against climate change, because large contiguous blocs allow plants and animals to migrate in response to tempera- ture or precipitation changes. But in both remote and nonremote areas, the economic and social costs of creating new protected areas must be weighed against those of other forms of ownership and management. In mosaiclands and nonremote frontiers, where land values are high and tenure rel- atively well defined, protected areas are likely to be small and have to be purchased from landowners and managed against encroach- ment by fairly high-density populations. Since protected areas are already poorly funded, acquiring such areas may be difficult. envi- ronmental service payment systems are a potential alternative (see below). In remote transfrontier areas creating new protected areas on public lands must address the potential claims of indigenous 165 AT L o g g e R h e A d S ? Figure 6.2 Recent Decades Have Seen Little Change in the Remoteness of New Protected Areas 90 Travel time to nearest city 80 < 3 hours 3­8 hours > 8 hours 70 areas 60 park 50 created 40 newly 30 of % 20 10 0 1930­39 1940­49 1950­59 1960­69 1970­79 1980­89 1990­99 Decade of park establishment Source: Kloss 2006. people, whose rights are increasingly recognized and supported (for instance, through the International Labour organization's Conven- tion 169 on Indigenous and Tribal Peoples). In such places coman- agement of protected areas (Redford and Painter 2006) and transfer to indigenous ownership are alternative management forms. Forest Revenues, Concessions, and Regulations The owners of a forest--whether a nation, community, or indigenous group--may find it politically or economically infeasible to conserve it without any revenue (point A in figure 5.1). Sustainable timber extraction represents a move up and to the left on the curve of figure 5.1. Although revenue comes at an environmental cost, the damages associated with logging are typically much smaller than those with clearing for permanent agriculture. In some cases sustainable log- ging may provide more income than agriculture; in others it lowers the opportunity cost of forgoing agriculture. In short, forest man- agement potentially provides a mechanism that can defend against pressure to convert forests to agriculture. And for many forest com- munities, it is the main source of income. 166 LoCAL ANd NATIoNAL PoLICIeS Many forest owners, both governments and communities, lack the technical skills or inclination to undertake commercial for- est management directly. When they look for a logger to do it for them, they face a problem with contracting and supervision: how to maximize the extraction of rent (that is, profit from timber sales) while maintaining the quality of the resource (for instance, avoid- ing damage to the forest as a result of sloppy logging, poaching of rare animals, or follow-on invasions of farmers) and meeting social goals. This is a difficult problem. Unregulated, a logger will not have strong incentive to log sustainably or responsibly. It is hard to moni- tor and supervise loggers in the expansive depths of the forest. It's difficult to know how much to charge them for the right to log. And public or community officials may collude with loggers, award- ing cheap contracts, failing to enforce regulations, and sharing the resulting gains. The traditional approach to logging concessions involves high transactions costs and is conducive to corruption. It relies for rev- enue on taxes per cubic meter of extracted timber, sometimes dif- ferentiated by species. These fees are difficult to set and don't reflect variations in profitability associated with different road access. Collecting these fees requires monitoring the flow of logs over the landscape--a massive task that lends itself to petty corruption by inspectors. Forest laws may require complex management plans to regulate logging and protect resources. Stringent on paper, these regulations tend to be unenforced. Monitoring and inspection are costly, and inspectors are easily bribed. An emerging approach strives for efficient regulation, combined with transparency and public disclosure. Instead of trying to tax each log, it auctions concession rights by the hectare. If the auction can be made competitive, that extracts more revenue for the forest owner, and the revenue is easy to collect. efficient regulation of log- ging looks carefully at the enforceability of regulations, preferring imperfect but easily monitorable criteria to ideal but unenforceable ones. For instance, it can use remote sensing or ground verifica- tion to check requirements such as maintaining forests on slopes and proper siting and construction of logging roads. Above all, the emerging approach relies on transparent information to enlist public oversight of loggers and officials. This new approach is exemplified in Cameroon (box 6.1), in an experiment that will be of wide inter- est for nations and communities with forest resources. 167 AT L o g g e R h e A d S ? Box 6.1 Cameroon: A Nexus of Institutional Reform Cameroon's forest experience over the past 10 of Cameroon by Transparency International in years exemplifies a central theme of this report: 1998 as the most corrupt of 85 countries sur- how can diffuse public interests in forest con- veyed. These catalysts arguably helped induce servation and in a broad allocation of forest change partly because they provided leverage wealth counterbalance narrow vested interests for domestic reformers. in forest exploitation? Cameroon has set up far-reaching, complementary institutional and Reform elements policy innovations that try to do that. These · Allocation of forest concessions by trans- innovations illustrate many of the policy and parent auctions. This approach replaced institutional issues discussed in this chapter discretionary procedures and awarded con- and chapter 5: zoning, mobilization of the cessions based on competitive bidding for public interest, promotion of community for- area-based fees. estry, efficient design of concession contracts, · Appointment of an independent observer. and checks and balances. The initial auctions resulted in concession awards that were inconsistent with the rules. The emergence of reform In response to pressure from the World Bank In the 1980s Cameroon was rich in timber, and others, an independent observer began but timber concessions and revenues were reporting on subsequent auctions. allocated opaquely, based on political patron- age. Concessions were awarded for only five · Forest tax reform. A new forest tax system years, so concessionaires had no incentive to prominently relied on the auction-deter- care for forests. Meanwhile, agriculture-driven mined area tax. Independent of production deforestation threatened the country's rich volumes, this tax is predictable and easily biodiversity. administered. Reforms also simplified the The economy fell into crisis in 1986 due to tax system, slashing export taxes and using a fall in prices of the country's main exports: harvesting and factory entry taxes to help oil, coffee, and cocoa. As a condition of ongo- monitor forestry and reduce waste. ing World Bank and International Monetary · Allocation of timber rents to communities. Fund (IMF) assistance, in 1994 the govern- The law requires that 40 percent of timber ment adopted a new forest policy allowing royalties go to rural councils (municipali- for the award of timber concessions by auc- ties) and 10 percent to local communities. tion. When the auctions failed to meet the In theory, this should promote local political standards embodied in the reform, additional support for the reforms. institutional reforms were made a condition of a 1998 World Bank structural adjustment · Concession management plans required, with loan. Another external catalyst was the listing another independent observer. Concession- 168 LoCAL ANd NATIoNAL PoLICIeS Box 6.1 (continued) aires are now required to create and follow Illegal logging by the informal sector has management plans. A second independent reportedly increased, reflecting a scarcity of observer's office monitors enforcement of legally cut timber for local markets. these plans, taking missions with and inde- · Effect on government revenue. The new auc- pendently from government authorities. In tion and tax system mitigated what would addition, remote sensing is used to monitor otherwise have been a severe drop in gov- logging activity. ernment revenues due to a ban on log · Definition of the forest estate and provision exports introduced in 2000. The reduction for community forests. A forest zoning plan in tax revenues since the reforms should be identified and provided preliminary bound- viewed against the improvements in sector aries for conservation and production for- governance and reductions in areas under ests. outside the permanent forest estate, it logging and annual volumes harvested. Vin- provided for community forests that could cent, gibson, and Boscolo (2003) note the be managed for timber or converted to non- unique character of Cameroon's system, forest uses based on local needs and prefer- which fixes area-based taxes for a long-term ences. The law allowed forest communities (15-year) contract. According to the authors, to challenge and redraw the provisional the lack of adjustments for timber price vol- zoning boundaries. atility exposes concessionaires to risk and may depress bids relative to a system index- Outcomes ing taxes to international timber prices. · Transparency and monitoring. The inde- · Effect on local incomes and capacity. By pendent observers' offices have greatly 2004 a total of $53 million in forest royal- increased public scrutiny of concession ties had been distributed to communities, operators and the government. one impor- and additional money went to community tant consequence was an increase in auction forests. There were no such distributions bids above the floor price after the creation before the forest tax reform. Annual audits of the independent observer. The precedent of these funds are produced with donor for transparency and oversight may spill support. oyono (2005) gives a mixed but over into other parts of government. mostly disappointing account of the use of · Greater potential control over large-scale log- these funds. Positive impacts include stimu- ging. With the introduction of better moni- lating community organization, growing rec- toring and a performance bond requirement, ognition of Pygmy rights, some productive there has been a reported reduction in ille- social investments, and retention of youth gal logging by large-scale concessionaires. in rural areas. Negative impacts revolve (continued on next page) 169 AT L o g g e R h e A d S ? Box 6.1 (continued) around the emergence of elites that control 13 percent to roads, and 10 percent to educa- forest revenues sent to locally elected bodies tion and health. There is less accountability and communities. These elites conflict with at the community level, where only about traditional authorities and are poorly super- half of expenditures could be tracked. edu- vised. The result is social conflict and mis- cation, housing, water management, and appropriation of funds, according to oyono. culture and sport were spending priorities. Further insight is provided by official · Industry impacts. About 15 companies audits on what may be an evolving situa- changed ownership, with a trend toward tion as institutions mature (Ndjanyou and more efficient and law-abiding companies Majerowitz 2004). There are strong efforts and increased domestic ownership. to impose transparency on the system, start- ing with public transfer of royalty checks to · Environmental impacts. There have been local officials. As a result the share of veri- no studies on how the zoning of the for- fiable expenditures by rural councils rose est estate and other reforms have affected from 49 percent in 2003 to 72 percent in deforestation or forest degradation. 2004. About a quarter of this revenue went to recurrent expenses, half of it salaries. of investment expenditures, about 60 percent Source: Vincent, gibson, and Boscolo 2003; World went to municipal buildings and vehicles, Bank staff. efficient regulation also calls for a reexamination of logging rules. Stringent rules appear to favor sustainability, but they also have disadvantages. Increasingly stringent rules provide smaller and smaller environmental benefits at escalating costs to forest owners. Boscolo and Vincent (2000) use a bioeconomic model to examine the costs and benefits of logging regulations in Malaysia. They find that shifting from unregulated to regulated logging, with a 40-centimeter minimum cutting limit, reduces the net present value of the timber stand by $510 a hectare (12 percent) but con- serves carbon and biodiversity. Raising the minimum cutting limit from 40 to 60 centimeters reduces the value of the stand by another $1,223 a hectare and yields only a modest additional benefit for carbon and biodiversity conservation. Because stringent regulations impose large burdens on loggers and forest owners, and because they are more difficult and expen- 170 LoCAL ANd NATIoNAL PoLICIeS sive to enforce, compliance is likely to be low. Burdensome and unenforceable laws spawn evasion, illegality, and corruption. A dis- tressing example of a lose-lose consequence occurs when it is easier to get a permit for forest conversion than for forest management-- reportedly the case in Brazil, Indonesia (FWI and gFW 2002), and elsewhere. This leads loggers to clear-cut and abandon plots that they would have been content to harvest selectively. Community Control of Forests--Balancing Rights and Responsibilities Centralized control of forests is increasingly considered untenable. Central authorities have difficulty defending forests against residents with little incentive to maintain someone else's resource. Central authorities are also viewed as being subject to capture by loggers and other vested interests. devolving forest control into local hands is seen as having equity dimensions--locals get a larger share of resource rents and efficiency benefits--with tenure, and locals are more likely to man- age the resource for long-term yields. In addition, locals have a deep understanding of local forest resources.3 Finally, indigenous and local people may hold traditional rights to these forests. on the other hand, there are potential trade-offs in devolving forests to local management. devolution processes may be imper- fect, with national authorities relinquishing poor-quality forests but maintaining control over rich ones. Moreover, local residents often lack management capacity. Local elites may wrest forest con- trol away from the poorest, and local leaders may be as susceptible to corruption and capture as agents of the national government. Although communities may have an incentive to maintain local environmental goods, such as watershed protection, they cannot necessarily be expected to take into account regional, national, and international externalities such as biodiversity loss and carbon emis- sions. Communities also might not be expected to engage in sustain- able management of resources, such as old-growth timber or rare animals, that fetch high prices but reproduce slowly. This section discusses several types of decentralization and devolution, following a spectrum of increasing rights transfer and drawing in part on a recent review by Shyamsundar, Araral, and Weeraratne (2005). 171 AT L o g g e R h e A d S ? Indigenous Populations Seem to Limit Deforestation devolution to indigenous communities is a case of special impor- tance. The International Labour organization's Convention 169 (on Indigenous and Tribal Peoples), which went into force in 1991, reflects the growing attention paid to indigenous rights and issues. The convention stresses the central importance of land to such pop- ulations, and requires that "the rights of ownership and possession of the peoples concerned over the lands which they traditionally occupy shall be recognized." But the convention has only been rati- fied by 17 countries. The convention has had particular resonance in Latin America, where it has been associated with shifts toward devolution in Bolivia, Brazil, Colombia, Costa Rica, Panama, Para- guay, and Peru (Roldan ortega 2004). As noted, indigenous land rights have also been recognized in a number of nonsignatory countries. Limited evidence suggests that, other things being equal, indig- enous forest ownership can be associated with significantly lower deforestation. The reason is debated. It could be that national gov- ernments treat indigenous areas as if they were protected, guarding them against encroachment by outsiders but also restricting conver- sion and degradation by residents. It could also be that indigenous people place a higher value on forest maintenance than outside colonists, use more benign and appropriate technologies for land and forest management, or have less contact with markets. or indig- enous lands may simply have extremely low population densities, and indigenous people may be unwilling (or not permitted) to sell or rent their land to others. The Brazilian study by Nepstad and others (2006) shows that indigenous areas have much lower deforestation rates than surround- ing areas. The protective effect of these areas shows up clearly in maps of fire incidence, which show, in effect, flames lapping at edges of reserves. The authors show that the protective effect of indigenous reserves declines with increasing population density and increas- ing time that indigenous groups have been in contact with Western society--but remains significant even for higher population densities and longer contact times. Stocks, McMahan, and Taber (2006) argue that in Nicaragua, indigenous people have different and more forest- friendly land use technologies than do colonists and can defend their land from colonists even in the absence of state support. 172 LoCAL ANd NATIoNAL PoLICIeS Decentralization to Local Governments Has Mixed Results Some countries are moving forest management authority from cen- tral to local governments, often as part of broader decentralization programs. examples include Bolivia, guatemala, honduras, Indone- sia, Malawi, and Zimbabwe. Results are mixed, especially for for- est conservation. given the chance, local populations may prefer to rapidly exploit forest resources if returns are high (Pacheco 2002; Colchester 2006). And local governments may be as incapable, or as subject to elite capture, as national ones. Andersson (2003) examined 50 Bolivian municipios with the mandate and funding to administer local forests. he found that the governments were much more likely to pay attention to forest administration if pressured by central government oversight, non- governmental organizations (Ngos), or communities. But forestry issues ranked low on community priorities, so such pressures were typically not exerted. early accounts of Indonesian decentralization paint a picture of local officials with little accountability to their constituency, resulting in increased logging (Larson 2004). A more recent study paints a much brighter picture, showing a tremendous increase in the proportion of community members who benefit from logging (Palmer and engel 2006). Community Forests Vary a Lot in Structure and Outcomes In a recent trend, some countries are sharing responsibility for administration of state-owned forests with local communities--or have transferred ownership outright. This represents a transfer of management of a significant portion of the earth's surface. Although some prominent examples provide a sense of the achievements and shortcomings, rigorous evaluation data are almost entirely lacking. The largest and best-known example is probably India's Joint Forest Management program--a complex experience that eludes easy summary. Although there were precursors, nationwide adop- tion of the program grew out of the National Forest Policy of 1988. By 2005 Joint Forest Management covered 27 percent of the national forest area across 27 states (17.3 million hectares) and included more than 8 million families--half belonging to scheduled tribes and castes. Although program rules differ by state, they give com- munities access to forests for fuelwood, fodder, and other extractive products and grant them a proportion of revenue from commercial 173 AT L o g g e R h e A d S ? timber sales. But more degraded, less commercially valuable forests are the most likely to be put under the program. one review paper found that Joint Forestry Management improves forest regeneration in these settings (Murali, Murthy, and Ravindranath 2002). Several papers suggest that it has a positive impact on livelihoods (Sarin and others 1998; Shyamsundar and Bandyopadhyay 2004; Köhlin and Amacher 2005). A recent study suggested institutional reforms that might increase the program's benefits for communities--suggestions that might apply even out- side India (World Bank 2005). These include providing longer-term, more secure tenure arrangements, giving communities a larger share of revenue from commercial forest products, and strengthen- ing the legal basis for the program, which in many states is based on a mutable executive order rather than legislation. Nepal's community forestry arrangements give more control to communities. gautam, Shivakoti, and Webb (2004) studied the Nepalese experience over 1976­2000. They found that the highest net improvement and gain in forest cover occurred in semigovern- ment forests (area legally under the forest department but with de facto control and ownership claims by local communities or municipalities), followed by formalized community forests (includ- ing leasehold), with government-run forests faring least well. Sch- weik, Nagendra, and Sinha (2003) similarly found that community management explained the persistence of forests in areas that would otherwise be under deforestation pressure. Somanathan, Prabha- kar, and Metha (2005), studying an unusual Indian setting akin to Nepal's, found that community-run forests fared much better than open-access forests and as well as government-run forests, despite being much cheaper to administer. But Malla (2000) shows that after implementation of community forestry, poor Nepalese lost their privileged access to forest products (such as fuelwood) because the forest user group shared those products equally among all house- holds. This observation is common among observers of community devolution and highlights the reality of intracommunity political and economic inequality. In Mexico about 80 percent of forests are owned by indigenous communities and by nonindigenous communities called ejidos. each forest is owned as common property by the formal members of the communities. Although ownership dates to the early 20th century or before, the Mexican government has transferred substantial manage- 174 LoCAL ANd NATIoNAL PoLICIeS ment authority to the communities only over the past two decades. By 2002 more than a quarter of the 8,500 communities with forests were engaged in formally recognized commercial forestry, with lack of technical skills impeding its wider diffusion (International Tropi- cal Timber Council 2005). But Antinori and Bray (2005) report that some communities have mastered a progression of skills, moving from sales of standing timber to felled logs to sawn wood to wood products. despite these successes, deforestation remains high in the forest communities, at about 1.7 percent a year over 1993­2000 (Fernan- dez and Munoz 2006). Most deforestation results from conversion to pasture or cropland, not from timber extraction. In many cases this reflects a failure of communities to restrict activities by people who are fellow residents but not formal community members and thus difficult to control. during the 1990s guatemala placed almost 500,000 hectares of forest in the Petén region under 13 community-managed con- cessions, with substantial donor funding (Nittler and Tschinkel 2005). Although corruption and incapacity have been a problem in this remote region lacking strong community organization, these enterprises are profitable--often highly so, due to valuable stands of mahogany. Almost all the enterprises are certified by the Forest Stewardship Council (FSC), although the mahogany extraction rate may not be fully sustainable. A striking impact is that deforestation within the concession areas appears to be much lower than outside them or in protected areas. In highly degraded areas of Tanzania, new government policies and law reforms enabled local villagers to redevelop their ngitili (fod- der and grazing reserves; Monela and others 2004). As a result fuel- wood collection times fell by two to six hours a day per household, fodder and wood availability increased (including for poor people), flora and fauna became more abundant, and local income and invest- ment rose significantly. WRI (2005) cautions, however, that tenure insecurity could threaten the program's long-term sustainability. Finally, Papua New guinea presents a warning that formal legal tenure is insufficient to guarantee favorable outcomes. All the coun- try's land is held by indigenous communities, who are permitted to grant industrial loggers access under what are supposed to be highly regulated conditions. But a summary of government-commissioned, independent reviews found poor enforcement of and compliance 175 AT L o g g e R h e A d S ? with regulations (Forest Trends 2006). None of the loggers secured informed consent from landowners, and none set up a sustainable timber harvest system. Cash royalties are not equitably or transpar- ently distributed and are not being invested for community welfare. Moreover, promised infrastructure is generally not fully delivered. Community Forest Management Presents Challenges Timber is, for the most part, the most valuable resource available to forest communities. But even if communities have rights to that tim- ber, they face hurdles if it is to both provide income and motivate forest conservation. The first hurdle involves geography and markets, which deter- mine the value of the resource and the cost of defending it. In fron- tier and transfrontier areas, big and valuable trees remain, but it may be difficult to get them to market or protect them against poachers. Some places are blessed with precious woods such as mahogany, or with high densities of less valuable but still marketable trees. But many tropical forests are a heterogeneous mix of species without established markets. Markets are closer in mosaiclands, boosting timber values, but forests are more depleted and conversion to crops or plantations is a greater threat (or opportunity). Second, the community has to prefer sustainable manage- ment--and not everyone shares that preference (see chapter 2). dis- count rates in developing countries are typically 25 percent or more (geF 2006), while logged-over forests tend to grow in value more slowly. So rather than spare seed trees or small trees from the axe, forest dwellers may prefer to liquidate these resources, using the proceeds to finance their children's education, migration, or other higher-return investments. But communities with abundant forest, or strong cultural and subsistence ties to it, may be content to man- age it sustainably for a wide range of benefits. Third, communities need people with management and market- ing skills. If communities decide to sell stumpage to outside loggers, they need legal and negotiating skills. If they manage their forests themselves, and especially if they set up sawmills or other process- ing facilities, they need sophisticated technical and financial skills. These capabilities are often lacking in remote forest areas with poor access to education. Finally, communities need effective and equitable ways of orga- nizing themselves. They need to set rules for accessing forests and 176 LoCAL ANd NATIoNAL PoLICIeS sharing benefits, and monitor and enforce compliance (ostrom 1990; Baland and Platteau 1996; Agrawal 2001; gibson, Williams, and ostrom 2005). While some forest communities have centuries- old traditions to draw on, others are communities in name only-- assemblages of recent migrants who face greater challenges in building the needed trust and internal organization. elite capture is a grave risk in both settings. It is inevitable that wealthier, bet- ter educated, more politically connected community members will exercise disproportionate control over forests, but this control can be more or less benign. In the worst cases, corrupt leaders sell or seize community resources for private gain, often in collusion with outside interests. But as discussed, some communities have surmounted these obstacles. While a community's social capital appears to play a deep role in its success, public policies can help. Scherr, White, and Kai- mowitz (2003) provide a detailed overview of options. Provision of secure forest rights is a basic prerequisite. Through training, govern- ments and aid agencies can build technical capacity for forest man- agement. There could be a public role in marketing. For instance, popularizing less-known tree species can increase the value of for- ests. Reducing transport costs can also make forestry more profit- able, though it could tip landholders toward agriculture. Improving forest governance at the local and national levels is also crucial (Ribot 2002, 2003). This could involve interventions at the community level to deter elite capture by making local leaders downwardly and upwardly accountable. A thicker web of report- ing, transparency, and accountability relationships is presumed to help, as is local democracy. Again, the forest agenda merges with the broader governance agenda. Regulating community forests poses special challenges. Colches- ter (2006) shows how onerous regulations can burden communities. In Bolivia complying with logging regulations can cost a community $20,000 to start and $8,000 a year after that. Avoiding such regu- lations imposes costs, either in limited marketing opportunities or bribes to officials. An obvious answer is to minimize the burden on communities through simpler rules. Contreras and Peter (2006) provide examples from guatemala, The gambia, and elsewhere. But when community forests are near more heavily regulated industrial concessions, it can lead to "leakage" of industrial logging into com- munity forests. 177 AT L o g g e R h e A d S ? Why should communities be subject to any kind of forest regula- tion? Regulation could be justified as protecting community forests against exploitation by outsiders or community elites. or it could be a protection against myopia--a public decision that people should not allow their high discount rates to induce them to sacrifice future pro- duction for current consumption. or sustainability restrictions could reflect a decision that communities should bear part of the social cost of providing the environmental benefits of forest. These rationales, and the trade-offs among social protection, environmental protec- tion, and income potential require careful consideration in framing restrictions on the ability of communities to use their forests. Private Property--Especially in Mosaiclands Rights to land are not the same as rights to trees. Landowners, in mosaiclands and beyond, may have secure tenure but still face restrictions on tree cutting, for the public good. For instance, there is a strong environmental rationale for maintaining vegetation near riv- ers and streams and on steep hillsides (see chapter 4). Many coun- tries limit landowners' rights to clear riverrine or hillside vegetation, and may require permits or management plans for any kind of tree cutting (table 6.3). But these restrictions are often poorly enforced. Carrots often work better than sticks--if you can afford the car- rots. A wide range of programs and policies seek to change land- holders' incentives for forest maintenance over degradation by Table 6.3 Latin American Countries Impose Varying Restrictions on Deforestation of Private Property Width of protected buffer around Degree of protected Size of property requiring Country watercourses (meters) slopes forestry plan Argentina 100 > 20 No information Brazil 30­500 > 45 No information Bolivia 10­100 > 45 > 3 hectares Chile 100­200 > 45 20­1,000 hectares Costa Rica 10­50 n.a. > 2 hectares Ecuador 50 n.a. All properties Peru 50 n.a. All properties Source: Environmental Law Institute 2003, p. 32. Note: n.a. = not applicable 178 LoCAL ANd NATIoNAL PoLICIeS offering them money or technical assistance. This section discusses two such approaches: · Promotion of greener agricultural technologies to make mosaiclands more like forests. · environmental service payments and other payments to landholders conditioned on forest status or environ- mental conditions. Greener Agriculture: A Technical Fix? Can farmers benefit by enhancing the amount of biodiversity on their lands? And if so, are they aware of it? This debate often gets muddled by the ambiguity of the term biodiversity. Farmers can benefit from having larger portfolios of cultivated biodiversity. Rice crops become more profitable, for instance, when different strains are mixed together. Less clear is whether there is an advantage to having more "wild" biodiversity. Some people (Rosenzweig 2001; McNeely and Scherr 2003) argue, plausibly, that some agricultural technologies favor wild biodiversity while also promoting farm profitability or reducing risk. It's further plausible that market forces might not, unassisted, spur the invention and diffusion of these technologies. If so, policy might fill these gaps. To return to figure 5.1, the goal is to improve over point C. Although systematic evidence is lacking, there are indica- tions that these win-win technologies exist. But policies might be needed to overcome barriers to their adoption. Vosti, gockowski, and Tomich (2005) contrast a land use system in the Sumatran for- est with a potential alternative. Sumatrans have long practiced rub- ber agroforestry, enriching secondary forest with rubber trees. This maintains far higher biodiversity than in monoculture oil palm, a competing land use. But returns to land planted with traditional rub- ber material are low. Improved rubber clones could drastically boost the profitability of rubber agroforestry, allowing it to compete with oil palm plantations. But problems in creating markets for seedlings, together with credit market failures, are blamed for the failure of this model to take off. The pastures of Central and South America provide another example. Pastures are poor in biodiversity and carbon storage rela- tive to the forests they replace. Yet there is more biodiversity in them than meets the eye. Farmers plant trees as living fences, allow them 179 AT L o g g e R h e A d S ? to persist on hillsides, and suffer a few to remain in pasture. harvey and others (2005) have found that these trees harbor substantial bird life while providing shade that improves the health of livestock. Pagi- ola and others (2004) hypothesize that silvopastoral systems (which involve planting more trees in pasture) could improve ranch profits, sustainability, and labor demand by providing fodder, fruit, nutri- ents, and nitrogen fixation. But ranchers--facing large upfront costs of planting, a four-year wait for the trees to mature, and uncertainty about the viability of the systems--see mediocre investment returns and substantial risk. The Regional Integrated Silvopastoral ecosystem Management Project (RISeMP), discussed below, is testing whether one-time payments to adopt these systems will lead to their retention and diffusion. The hypothesis is that once the trees are in place, the higher income they produce will ensure that they are maintained. Integrated pest management is another win-win technology that faces barriers to adoption. Using pesticides in forest-agricul- ture mosaiclands not only damages biodiversity, it also sickens and kills farmers. Integrated pest management is an appealing alterna- tive. It uses natural antagonists to get rid of pests--at low cost and with a substantial reduction in pesticide use. But adoption has been slow. one problem is coordination: it is nearly useless to be the only farmer in a neighborhood using integrated pest management. Neighbors' pesticides will kill your beneficial bugs. everyone has to adopt at the same time. The need for collective action is a hurdle that not all communities can surmount. (one way of reducing that hurdle would be to ensure that prices of pesticides and herbicides fully reflect their societal costs.) In sum, there could be a range of win-win technologies that improve incomes and environmental outcomes, especially in agri- culture-forest mosaiclands. There is inadequate investment in research, development, and diffusion of these technologies, because they could be difficult to protect through patents or other means. (genetically engineered plants could be an exception.) Much more research is needed to compile, develop, and test such technologies. Economic Instruments and Markets for Environmental Services Provide Alternatives to Command and Control The starting point for an approach to conservation based on direct incentives assumes that landholders have some rights to modify or cut their trees. It is then up to others--perhaps downstream res- idents, perhaps society as a whole--to provide incentives for the 180 LoCAL ANd NATIoNAL PoLICIeS landholders to manage their land and trees in a way that provides benefits (or reduces costs) to others in society. For instance: · Urban residents may pay upstream farmers to reduce sediment in drinking water sources, to lower costs of water treatment. · Society may compensate landholders for maintaining or improving biodiversity-friendly habitats. · Firms required to reduce carbon emissions may pay landholders to reduce emissions from deforestation or to sequester carbon in regenerating forests. · "Run of river" hydroelectric plant owners may pay landowners to maintain forests in a way that promotes stable water flows. These incentive-based or economic instrument strategies could yield more efficient land management than do command-and-control approaches. They do so by eliciting information from landholders on the true value of land under alternative uses, then motivating conservation only if the value to society of doing so is more than the value to landholders of putting the land into agriculture. If society has flexibility in meeting its environmental goals, economic instru- ments can reduce the cost of meeting them. Because the use of economic instruments for land management is still new in the developing world, reviewing some actual and pro- posed examples will help ground the discussion of their potential implementation. Transferable Forest Protection Obligations in Brazil4 An emerging topic of policy discussion in Brazil is the possibility of shifting from a command and control system of forest regulation to an economic system of tradable rights found in other environmen- tal spheres, such as fishing management and pollution regulation. For more than 70 years Brazilian landholders have been obliged to maintain a portion of each rural property under natural vegetation. This requirement is 20 percent in southern states and 80 percent in Amazônian forest. The legal reserve requirement supplements a separate requirement to maintain riverrine and hillside forest. But the reserve requirement has not been strictly enforced. In many agriculturally dynamic locales aggregate forest cover has dropped well below 20 percent. These areas now face increasing 181 AT L o g g e R h e A d S ? pressure to enforce the regulation. But compliance will be expen- sive if landholders are required to abandon valuable plantings. And on heavily farmed properties with little remaining natural vegeta- tion, the rate and quality of natural regeneration might be extremely slow, so biodiversity gains might be minimal. on the other hand, forests and woodlands--often with high biodiversity values--remain in more remote and less favorable regions. Yet deforestation contin- ues there, often for low-value extensive pasture, with charcoal as a by-product from clearing. In many of these forests, deforestation remains legal down to the reserve requirement. As enforcement pressure increased in the 1990s, it occurred to many people that the property-wise legal reserve requirement, while well motivated, was neither economically nor environmentally effi- cient. Why not allow out-of-compliance landholders to meet their obligations by protecting land of more biodiversity value but less agricultural value? Landholders with more than 20 percent forest cover could be rewarded with the right to sell legal reserve services (equivalent to development rights) from their "excess" forest. Chomitz, Thomas, and Brandao (2005) simulated the impact of such a plan for the state of Minas gerais. They found that relative to command and control rules, a tradable rights scheme would cut compliance costs by up to two-thirds and protect up to one-third more forest designated as being a high conservation priority. gains were greater when the ambit of permitted trades widened from microwatershed to river basin to biome. Could this kind of tradable obligation system be widely adopted? In Brazil, where other states are examining this approach, the long history of the legal reserve obligation is an important contributing factor. other countries might start with long-standing but sporadi- cally enforced prohibitions on tree cutting. Relaxing and making flexible these obligations might be welcomed as an improvement by landholders and lead to preferred environmental outcomes. A point in favor of the approach is that it represents a social compromise between the extremes of assigning landholders complete rights to their trees versus none. Examples of Environmental Service Programs In contrast, payments for environmental service programs take as their starting point landholders' full rights--in fact or in law--to plant, maintain, or cut trees on their property. People affected by those decisions offer conditional payments for maintaining trees. An 182 LoCAL ANd NATIoNAL PoLICIeS environmental rationale for these programs distinguishes them from the more frequent, well-funded practice of subsidizing plantation forestry as an industrial or development strategy. A diverse set of environmental service programs are under way in the developing world. A few snapshots: · Costa Rica's pathbreaking system is probably the most famous (Chomitz, Brenes, and Constantino 1999; Pagi- ola 2005). It grew out of a 1996 forestry law that rec- ognized forest environmental services such as carbon sequestration, biodiversity, hydrological regulation, and provision of scenic beauty. Landholders volunteer to participate; those selected receive about $45 a hectare per year to maintain forests. FoNAFIFo, a government agency, funds the program from a variety of sources. Some of these conform to the model of paying for envi- ronmental service: a grant from the global environ- ment Facility to conserve biodiversity and payments from hydropower plants to maintain watersheds. But a national fuel tax provides most of the funding. About 240,000 hectares are under contract. · Mexico introduced payments for hydrological environ- mental services in 2003 (Muñoz-Piña and others 2005). Similar to the Costa Rican program, but motivated by concerns about water scarcity and quality, it rewards landholders for forest conservation, paying $27 a hect- are per year to conserve cloud forest and $18 for other types. The program was initially funded by an $18 mil- lion earmark on water fees. · China's sloping land program is one of the world's largest environmental service payment programs (Xu and others 2004; Bennett 2005). The program, initi- ated in 1999, was motivated by concern over severe sedimentation: sloping farmlands generate 1.3 billion tons of sediment a year in the Yangtze and Yellow riv- ers, which was believed to have been a major cause of the Yangtze floods of 1998. The program offers seed- lings, cash, and grain to farmers who retire marginal or steep, erosion-prone farmland, replanting it with grass, fruit-bearing trees, or trees for timber. Nominal per-hectare incentives are two to three times the mean 183 AT L o g g e R h e A d S ? value of payments in the analogous U.S. program (Con- servation Reserve Program), though a survey found shortfalls in payments (Xu and others 2004). More than 7 million hectares were enrolled in the program's first five years, with another 7 million planned. · RISeMP, which began in 2002 in Colombia, Costa Rica, and Nicaragua, is testing whether payments for envi- ronmental services can catalyze permanent adoption of silvopastoral systems (see above; Pagiola and others 2004, 2006). The project pays landholders for environ- mentally beneficial changes in land cover. An ecologi- cal index establishes a point value for each type of land cover based on a rough assessment of its significance for biodiversity and carbon storage. For instance, degraded pasture is worth 0 points a hectare, native pasture without trees 0.2, native pasture with trees 1.0, and young secondary vegetation 1.4 points. Landhold- ers receive $75 a point per year for changes, for up to four years. After two years the share of the project area considered to be improved "pasture with high tree den- sity" had tripled from its initial 5 percent. In addition, the ecological index for the area rose 42 percent. These, and a host of smaller programs worldwide, are pioneer- ing efforts to solve the complex questions of collective action run- ning through this chapter. Their early successes and shortcomings offer lessons for current and would-be designers of systems provid- ing payments for environmental services. Designing Environmental Service Payments Three sets of design issues shape the feasibility and impact of a sys- tem to provide payments for environmental services: · Financing: is there demand for the environmental ser- vice, and can people organize themselves to pay for it? · Structuring the payments: who is eligible, for how much, and under what conditions? · Logistics: is it possible to cost-effectively collect funds, make payments and monitor compliance with condi- tions for payment? 184 LoCAL ANd NATIoNAL PoLICIeS Financing. Ultimately, buyers have to believe in the services they are buying. The case for domestically funded payments, as the examples suggest, is largely based on the promise of hydrological benefits. Sensible enough: clean, plentiful water commands a large constituency. But the link between tree protection and water ben- efits can be tenuous. Systems predicated on questionable assump- tions--for instance, that forest maintenance will increase water flows--could backfire if they fail to deliver promised services. Moni- toring, modeling, learning, and adjusting the system are essential if programs based on subtle services are to succeed. Suppose people are convinced that they will benefit from an environmental service. They still might not be willing to pay for it if they think it is possible to free ride. For instance, all the residents of a broad plain can enjoy scenic vistas of a forested mountainside, even if they individually fail to contribute to forest conservation. To overcome this obstacle, beneficiaries need to organize themselves and commit to sharing the costs. Sometimes there will be a ready-made institutional solution. When the New York City Water Authority buys water on behalf of its clients, it can build the cost of watershed protection into the rates it charges. often government is a natural choice for financing the public good, especially one whose benefits are widely enjoyed. But where internal controls are weak, government administration can be problematic. Structuring the payments. defining payment rules for a govern- ment-coordinated environmental services payment program can entail a struggle between technical efficiency and political viability. economic efficiency requires keeping forest on properties whose environmental value exceeds their value as cropland. Fiscal effi- ciency requires minimizing payments to landholders who would have maintained their services in any case. Both kinds of efficiency can be approached through auctions. For instance, in the U.S. Con- servation Reserve Program, landholders submit bids specifying the environmental services they can provide and the lowest rate they will accept. The government ranks the bids by cost-effectiveness, funding from the top of the list until the budget is exhausted (box 6.2). But efficiency doesn't always win support. A focus on efficiency directs attention to special spots--those with the most threatened species or highest erosion rates. If these places are in the minority, there will be pressure to extend the benefits more broadly. There 185 AT L o g g e R h e A d S ? Box 6.2 Self-assembling Biodiversity Corridors: Reconciling Voluntary Participation Decisions with Landscape-level Goals Biodiversity survival depends on connected services, modeled after the U.S. Conservation habitat (box 4.1). But what kind of policies Reserve Program and Australia's BushTender can induce landholders to create these connec- (Stoneham and others 2003). In this system, a tions? Conservation planners have developed conservation agency solicits offers from land- sophisticated methods for laying out biodiver- owners. The landowners name a price at which sity corridors and reserve networks that allow they would be willing to accept a conservation thriving wildlife populations. But it is difficult contract for their property, retaining ownership to make landholders comply with such plans. but protecting existing forest and promoting for- even if funding is available for compensation, est regrowth. Their property is also rated, on a an obstinate landholder could block the forma- point system, for environmental benefits. The tion of a planned corridor. Programs that make agency then ranks the landowners' bids on the payments for environmental services, in con- basis of environmental benefit points per dollar, trast, have the virtue of voluntarism. But how funding the highest ranked bid until its budget can uncoordinated responses result in a con- is exhausted. nected, viable corridor? Chomitz and others (2006) found that rela- A simulation by Chomitz and others tively modest budgets could theoretically elicit (2006) suggests that this might be easier than the formation of self-assembling biodiversity it seems in conditions typical of biodiversity corridors, as contracted properties coalesced hotspots. The simulation addressed conserva- into viable, connected forest patches. The rea- tion in the Atlantic forest of Bahia, Brazil--a son is that in Bahia, which has suffered sub- place with high species richness and endemism stantial deforestation, remaining forest is often and extreme fragmentation. The authors intro- a marker for plots with low land value. Because duced flexibility into the conservation problem of isolation or poor soil quality, such places by specifying a biological goal: increase the occur in clumps. By focusing on goals (viable number and geographic diversity of connected habitats) rather than means (a prespecified cor- forest patches large enough to support a viable ridor plan), it might be possible to reconcile population of primates. voluntarism with conservation planning. Theyproposedahypothetical,auction-based system to make payments for environmental Source: Chomitz and others 2006. will be strong temptation to use the program to meet social goals such as poverty alleviation and strong pressure to direct it toward elites and well-connected groups. Another targeting dilemma is related to moral hazard. Payments will be most effective in delivering services if they focus on areas most at risk of deforestation. This tends to exclude landholders 186 LoCAL ANd NATIoNAL PoLICIeS who, out of conviction or by default, have preserved their forests. Finally, auction systems and targeting systems result in differenti- ated payments. Some landholders get more money, either because they are providing more services or because their opportunity costs are higher. But despite firm economic justification, differentiated payments may be perceived as inequitable, nontransparent, costly to administer, or subject to manipulation. The tension between implementation and efficiency issues was evident in the initial design of Mexico's payment program for hydro- logical services (Muñoz-Piña and others 2005; Alix-garcia and oth- ers 2005) and Costa Rica's (Chomitz, Brenes, and Constantino 1999). Although the Mexican program's rationale emphasized protecting forests in recharge areas of water-scarce basins, at least 85 percent of payments went to basins known to be underexploited (Alix-gar- cia and others 2005). Both countries used simple flat payment sys- tems rather than auctions or highly differentiated payments. Flat payments might be expected to result in adverse selection, attracting landholders with no intention or opportunity to convert their land to agriculture. In fact, only about 10 percent of the Mexican contracts went to the 20 percent of areas with the highest predicted risk of deforestation. There is evidence of similar outcomes in Costa Rica. In both countries demand for payments far exceeds supply. Logistics and administration. Although the logistics of environ- mental payment systems are daunting, Costa Rica and other coun- tries have shown that it is possible to create such systems. A basic remaining problem is reducing transactions costs. There are fixed costs associated with drawing up contracts, making payments, and verifying compliance. Consequently, it is cheaper to enroll large properties. Zbinden and Lee (2004) show that such properties are overrepresented in Costa Rica's program. There are two related ways to cut these costs. one is to explore the tradeoffs involved in detailed specification of management plans and payment criteria--do gains from theoretical efficiency outweigh added transactions costs? RISeMP considers this issue, assessing the benefits of tying payments to detailed measurements of carbon and biodiver- sity benefits. Many more such experiments are needed. The second way to reduce costs is by using modern technolo- gies. Mato grosso's SLAPR program (see "Setting the Stage" after the overview) used global Positioning System (gPS) technology to map the locations of enrolled properties, and satellite monitoring to check compliance with forest obligations. The original program design envisioned public disclosure of enrollment and monitoring 187 AT L o g g e R h e A d S ? information. Failure to use such oversight may have contributed to the program's problems. The three design questions--finance, structure of payments, logistics--all bear on the implications of such payment programs for poor people (Pagiola, Arcenas, and Platais 2005). It would be ineq- uitable to ask the poor to finance these programs if wealthier people were the source of the externality in question. In that case there may be an argument for government to finance the services from general revenue, as a means of improving poor people's income or decreasing their vulnerability. Poor people can also benefit as ser- vice providers. Pagiola, Arcenas, and Platais (2005) argue convinc- ingly that environmental services, not poverty, must be the basis for payments. otherwise a program could fail to provide the services it promises. The challenge is to reduce barriers to participation by qualified poor people. Those include information barriers and econ- omies of scale in participation. Community groups can help solve those problems. Mexico's program provides an interesting example of reduced transactions costs and possibly of the spatial coincidence of poverty and environmental services. As noted, most forests in Mexico are owned by communities, and for these forests the program paid entire communities--lowering transactions costs relative to dealing with many smallholders. Although early suggestions to make poverty an explicit criterion for payments were rejected, 83 percent of contracts (weighted by enrolled area) went to communities with high or very high marginalization (Alix-garcia and others 2005). But it is not yet clear whether poor communities are more likely to provide real ser- vices--that is, an actual reduction in deforestation combined with a real link between forest cover and hydrological services. Other Development Policies with Forest Spillovers often policies conceived outside the confines of forestry have important implications for land use and for the incomes of forest dwellers. Road Planning Can Ease Poverty-Environment Tradeoffs Planning and regulating rural roads may present a trade-off between poverty alleviation and conservation. Building roads in forest areas can lead to deforestation. It also accelerates deforestation, as new roads branch out from old ones. 188 LoCAL ANd NATIoNAL PoLICIeS Yet roads also stimulate agricultural production and can allevi- ate poverty. Roads are favorite investments for local communities and politicians, especially in remote areas. Consider Papua, Indone- sia, a forest-rich province with few people and few roads. After fis- cal reforms in 1999, the province was flush with cash from increased revenue sharing and rents from oil and mining operations. The prov- ince devoted more than a fifth of its development spending to roads, extending its paved network by 20 percent in just three years. grappling with this trade-off requires political decisions at lev- els corresponding to the road under consideration. Building new or improved roads through forest areas can profoundly affect regional development and the environment by triggering follow-on offshoots of roads in forests. These roads could set the world's few large, rel- atively undisturbed tracts of tropical forest on a trajectory toward long-term fragmentation and loss. examples of such roads under dis- cussion or planning include BR163 in Brazil and trans-Kalimantan highways in Indonesia. In the democratic Republic of Congo reha- bilitation of the collapsed road infrastructure is sure to be part of the long-run development strategy. elsewhere, mining and logging can open new roads in trackless areas. And it is important to remember von Thünen's lesson: that road improvements near markets or ports are transmitted through the entire network, affecting deforestation hundreds of miles away (see chapter 2). Now is the time for long-term planning, including road plan- ning, for remaining transfrontier areas. This planning should steer conversion pressures toward areas with higher agricultural poten- tial and away from areas with unique biodiversity or environmental characteristics. If planning starts now, it may be possible to develop a rational, widely shared vision of a forest region's potential and to build constituencies for conservation and sustainable use. But if planning is delayed, it may be too late. After colonists arrive, and after forestlands begin to offer significant rents, strong political con- stituencies form for deforestation. It is easier to clarify ownership while the forest still has negligible value, before disputes arise. In frontier areas there may be opportunities to tie road construc- tion or upgrades to changes in forest tenure or protection (Ledec and Posas 2006). A good example is the construction of the San José­ guápiles highway in Costa Rica. Planned when deforestation was high, this major road had the potential to trigger it on hydrologically sensitive slopes. But establishment of the Braulio Carrillo National Park, before the road was constructed, has protected the area. 189 AT L o g g e R h e A d S ? In mosaiclands road construction may tend to target remaining forest stands because doing so requires less disturbance of fields and settlements. here a possible solution is to apply compensatory rules, requiring set-asides or regeneration of sites equivalent to (or larger than) the forest to be cleared. Can Local Development Reduce Local Forest Pressure? As a rule, development and deforestation propagate outward from towns and intensive agricultural areas. But as with many rules, there are possible exceptions. Local development could attract work- ers away from nearby marginal forestlands. If local people maintain reasonably secure tenure over those lands or there are barriers to migration, it could result in abandonment of farming in the marginal forestlands. There are indications of such processes at work in Southeast Asia. Shively and Pagiola (2004) describe an interesting natural experiment in southern Palawan, a forest frontier of the Philip- pines, using panel data on household income and activities. There, upland dwellers at the forest margin live within commuting distance of lowland populations in fertile, deforested river valleys. Between 1995 and 1999 nearly all the lowland farms were converted to irri- gated rice production, enabling a shift to multicropping. As a result lowland farmers nearly tripled their employment of uplanders. The uplanders, now better employed, reduced by about half their rate of forest clearance and intensified production on their plots. At last report there was not a compensating influx of immigrants to aug- ment forest clearing on the upland frontier. Müller and Zeller (2002) report similar dynamics for two dis- tricts of the Central highlands of Vietnam, using remote sensing data on land cover. There too there was paddy rice production in valleys, surrounded by shifting cultivation in hillsides and uplands. From 1975­92 paddy and shifting cultivation expanded at the expense of open forest. during the 1990s substantial investments in roads, irri- gation, and rice technologies led to a tripling of the rice yield. In addition, new crops were introduced. Shifting cultivation apparently gave way to permanent cropping. grasslands reverted to secondary forest. Müller and Zeller attribute these developments to the labor- absorbing impact of intensified production techniques, despite pop- ulation growth. But the authors note that this outcome contrasted with continued deforestation in nearby districts, where soils were better suited to high-value crops such as coffee and pepper. This 190 LoCAL ANd NATIoNAL PoLICIeS pattern underscores the lesson that increased agricultural productiv- ity reduces deforestation only where the labor supply or commodity demand is rather limited. Policies Affecting Agricultural Prices Macroeconomic policy making involves balancing interests between sectors and between producers and consumers. Because of the macroeconomy's complexity, it can be difficult to trace the reper- cussions of policy shocks. Perhaps the easiest to trace, and most relevant to this report, are those that affect the prices of forest-com- peting commodities. higher prices for forest-competing commodities should increase pressure for deforestation, unless those prices attract deforesters away from yet more forest-damaging activities. These effects should be most evident at the forest frontier, because a small change in prices at an urban port or market translates into a large propor- tional change at the frontier. Consider again the example of Brazil- ian Amazônia (see figure 2.1). deforestation rates were much higher near markets, where beef fetches up to 800 reais a ton, than at the frontier, where it commands only 400 reais a ton because of trans- port costs. At 400 reais a ton, farmers barely break even (Arima, Barreto, and Brito 2006). If the price of beef rises by 80 reais a ton at the market, profits of periurban farmers increase 20 percent. Trans- mitted down to the frontier, this increase transforms worthless (for agriculture) transfrontier forest into land worth ranching. Arima, Barreto, and Brito (2006) calculate that a 10 percent increase in the urban price would extend the frontier of cattle ranching by 260,000 square kilometers. And it would increase the area of high profitabil- ity and high deforestation (where the farmgate price of beef exceeds 600 reais a ton) by almost 600,000 square kilometers. In Brazil--the only tropical country with good annual data on deforestation--this prediction is consistent with recent macroeco- nomic and price changes. A substantial devaluation at the beginning of 1999 was followed by increases in world prices of three important export commodities: beef, soybeans, and pig iron (which in Brazil is produced, in part, using forest charcoal rather than mineral coke). At the same time, hoof and mouth disease restrictions on Amazô- nian cattle movement have been eased (Kaimowitz and others 2004). Together these changes greatly increased incentives for deforestation, with annual deforestation rates rising from 18,000 square kilometers in 1999 to more than 26,000 square kilometers in 2003­04.5 191 AT L o g g e R h e A d S ? Wunder and Sunderlin (2004), using less rigorous data, tell the converse story for gabon--where currency overvaluation, the result of oil revenue, has made agriculture and deforestation relatively unattractive. Similarly, they describe how a Venezuelan oil boom over 1930­50 resulted in pasture abandonment and forest regrowth as agriculture became unattractive. Later, though, Venezuelan invest- ments of oil revenue in frontier roads and transport subsidies led to renewed deforestation. But Jensen, Robinson, and Tarp (2004) warn that changes in exchange rate can have counterintuitive effects due to intersectoral shifts. Wunder and Sunderlin's account of Cameroon's volatile econ- omy bears out this point. during 1978­85, when high prices for oil, coffee, and cocoa created a boom, the net effect was an urban bias that accelerated rural-urban migration and slowed deforestation. After 1986 low prices for oil and cash crops, combined with a fixed and overvalued exchange rate, precipitated a crisis. That reversed the migration and significantly increased deforestation for crop pro- duction, more than offsetting a decline in clearing for export crops. Madagascar also shows how changes in agricultural prices can affect deforestation and welfare. The country's southwest region experienced a boom in maize exports when european subsidies stimulated hog production on nearby Reunion Island--until Argen- tina's devaluation made it a lower-cost supplier of hog feed (Moser, Barrett, and Minten 2005; Minten and Méral 2005). The maize boom accounted for about a fifth of deforestation in Madagascar's south- west region; deforestation there is thought to lead to irreversible land degradation and abandonment after a few years. The boom may have contributed, at least temporarily, to higher incomes in this extremely poor region. These examples suggest that price levers put deforestation and poverty alleviation at odds. A look at Madagascar's main staple crop, however, intimates the possibility of a nearly win-win combi- nation, at least in theory. even though 60 percent of the households grow rice, most of them buy more than they sell, and the country is a net rice importer (Minten and dabat 2005). A tariff protects rice growers. Reducing the tariff would reduce the price of rice with little effect on the purely subsistence population, but it would benefit the larger part of the population who are net buyers. A lower price and more imports would ease pressures for expanding the area under rice cultivation. Although net sellers of rice would tend to lose, some ameliorative options are available. Large areas of Madagascar are constrained by transaction costs from participating in the national rice market (dorosh and Minten 2005). Improving the country's 192 LoCAL ANd NATIoNAL PoLICIeS dilapidated roads and mitigating other market failures could boost the farmgate price for growers in degraded areas (Moser, Barrett, and Minten 2005). Summary Managing forests is difficult because it requires balancing weak and powerful interests, concentrated and diffuse interests, and today's certainties and tomorrow's possibilities. Creating institutions that do this fairly and efficiently requires mobilizing constituencies for conservation and sustainable land management and giving voice to poor and indigenous groups. As communications and monitoring become cheaper, these groups can be empowered with information and may be able to resist resource grabs by large actors at the forest frontier. Placing frontier and transfrontier forests under clear and guaranteed stewardship--of indigenous groups, other local popula- tions, protected area managers, or regulated concessions--is neces- sary but insufficient. It is essential to recognize that the steward's interests may not be perfectly aligned with the wider public inter- est, making some form of regulation or incentive necessary to align interests. efficient regulation can minimize burdens on forestholders as well as costs of monitoring and enforcement. Within the frontier, it is necessary to arrive at a workable bal- ance between landholders' rights and responsibilities. Attempts at complete regulation of tree and forest management are unworkable. But it may be equitable and efficient to enforce some land man- agement responsibilities on some groups of forestholders. domesti- cally financed payments for environmental services must be tightly focused on clearly demonstrable, locally valued services, because the ability to raise domestic funds will be limited. Turning to an international arena, however, can create new possibilities. Endnotes 1. An emerging exception may be conservation plans for the Cape region of South Africa, which compensate landowners for reserving areas. 2.http://www.whrc.org/africa/INFoRMS/study_sites/Kasyoha_ Kalinzu.htm 3. ostrom 1999; Ribot 2002; Molnar, Scherr, and Khare 2004; Colfer and Capistrano 2005; Ribot and Larson 2005. 4. This section is based on Chomitz, Thomas, and Brandao (2005). 5. data are from INPe (http://www.obt.inpe.br/prodes/). 193 Clearance of tropical forest for agriculture is a major source of CO2 emissions, contributing to global warming, and also causes local air pollution. This satellite photo shows vegetation burning near an oil palm plantation in eastern Sumatra. Image acquired and processed by CRISP, National University of Singapore IKONOS image © CRISP 2005. CHAPTER 7 Mobilizing Global Interests for Forest Conservation T wo rationales for forest conservation attract a large, wealthy, worldwide constituency. All forests store carbon, so preventing deforestation can mitigate global damages from climate change. In addition, some forests harbor unique biodiversity whose survival is threatened by deforestation. The challenge for international policy is to find ways to tap these global interests to finance forest conserva- tion, using approaches acceptable to forested countries. Forest Carbon Finance: An Ungrasped Opportunity Chapter 4 presented a paradox. Throughout the developing world, farmers fell trees for sometimes small and ephemeral gains, creat- ing croplands and pastures worth perhaps a couple hundred dollars a hectare. As those trees burn and rot, they release carbon diox- ide (CO2) to the atmosphere--perhaps 500 tons a hectare in dense rainforests. Meanwhile, the European Union (EU) market values CO2 abatement at $20 a ton.1 In other words, farmers are destroy- ing a $10,000 asset to create one worth $200. (While the $20 price is highly volatile, the disparity would remain even if prices plum- meted. And the sum doesn't include the value of biodiversity and other environmental attributes.) There seems to be a great opportunity for arbitrage here. Indus- trial countries could pay the poor farmers for forest conservation, at some amount between $200 and $10,000 a hectare, and both par- ties would gain. That would be a good deal for the farmers even if industrial countries' willingness to pay were at the modest price of 195 AT L O G G E r h E A d s ? $2.70 a ton suggested by Yohe, Andronova, and schlesinger (2004) as a target for global policy. Yet this opportunity remains ungrasped. Why? What are the obstacles? And can they be overcome? Why Carbon Finance Makes Sense for Climate The United Nations Framework Convention on Climate Change, signed by 189 countries, aims to stabilize the amount of greenhouse gases (GhGs) in the atmosphere. Greenhouse gases are increasing largely because people are burning more fuel. Thus stabilization requires a long-term shift to cleaner energy. halting all deforestation, even if possible, would not by itself achieve the convention's goal. But no single line of action will be sufficient to achieve that goal. Pacala and socolow (2004) outline 15 options for reducing CO2 emissions over the next half-century. Each option would cut emis- sions by about 25 billion tons during this period. seven to ten activi- ties of this magnitude would be required to stabilize GhGs in the atmosphere (that is, hold atmospheric concentrations of CO2 to 500 parts per million). reduced deforestation and increased reforesta- tion are one option. so while containing forest carbon is no panacea for climate change, it could be part of the solution. And it could be an important part if it is cheap. Cheapness is a virtue. We don't know how much it will cost to mitigate climate change. And we don't know how much mitigation is needed. More stringent targets for atmospheric CO2 concentrations provide better insurance against catastrophic climate changes, but each part per million reduction will cost more than the last. Because the risks are difficult to quantify, it is hard to achieve global agreement on how much to spend and how to split the bill. Thus anything that reduces the cost of a global mitigation strategy will increase the chance that the strategy is embraced. Why Carbon Finance Makes Sense for Forests and Rural Development Forests may play a relatively small role in mitigating climate change, but climate change mitigation could play a large role in financing forest maintenance. Among the potential environmental services of forests, carbon sequestration has the widest applicability. That is because any action that keeps a ton of carbon out of the atmosphere has the same climatic impact no matter where it occurs. In contrast, 196 M O B I L I z I N G G L O B A L I N T E r E s T s F O r F O r E s T C O N s E rvAT I O N many of the environmental services enumerated by the Millennium Ecosystem Assessment are location-specific and idiosyncratic: eco- tourism, hydrological regulation, or maintenance of globally signifi- cant biodiversity. Carbon payments might provide significant benefits to tropical countries. sathaye and others (forthcoming) find that over 40 years, paying $10 per ton of carbon (rising with the interest rate) would have a net present value of $150 billion in payments to developing countries for avoided deforestation. Containing forest carbon would also provide local and global benefits that would otherwise be diffi- cult to finance--including conservation of globally significant biodi- versity and of forests with spiritual or other values that are difficult to monetize. Forest carbon control might also help finance agrofor- estry and agricultural intensification in unforested areas. Financing Avoided Deforestation: Problems and Solutions The UN Framework Convention on Climate Change is responding to a submission by Costa rica and Papua New Guinea to examine options for providing countries with incentives to avoid deforesta- tion through forest carbon. To gain acceptance, these options will have to address, in practical ways, the objections that kept defor- estation out of the Kyoto Protocol. This section lists the main con- cerns--and ways to deal with them. (The discussion here draws on and expands Chomitz 2002.) "Forest Carbon Makes Mitigation Too Cheap" Problem: At first glance this objection is hard to understand. Cheap- ness, as noted, is a virtue. Getting people, firms, and countries to take actions for global benefit is easier if those actions are cheaper. What drives this objection is a fear that introducing forest carbon into the Kyoto Protocol would swamp the emerging carbon mar- ket--driving prices toward zero and reducing industrial countries' incentives to shift to clean energy. But that outcome would arise not from overly cheap mitigation. rather, it would be the result of a timid, ineffective mitigation goal. The Kyoto Protocol currently places only moderate limits on green- house gas emissions from participating industrial countries. The limits for 2008­12 are perhaps a billion tons a year (CO2 equivalent) less than would be emitted in the absence of the agreement. 197 AT L O G G E r h E A d s ? Countries can try to reduce their emissions by this amount, or they can buy offsetting emission reductions abroad. Either way, the total Kyoto limit is still met. developing countries can reduce CO2 emissions, for instance, by switching from coal to wind power-- then sell the reductions. This approach creates a market for emission reductions. The Kyoto emission limits determine the demand for these reductions, and opportunities for switching fuels and increas- ing efficiency in the developing world largely determine the supply. The Kyoto Protocol doesn't allow developing countries to create emission reductions from avoided deforestation. But suppose it did and that countries could instantly create the institutions needed to reduce deforestation and that the protocol did not change its limits on total CO2 emissions. In this unlikely set of contingencies, the supply of emission reductions would increase and their price would fall. The Kyoto emission limits would still be satisfied, and the cost of meeting them would be reduced. But the resulting low prices for CO2 reductions would provide little benefit to developing countries and little stimulus for energy research and development. But because the Kyoto limits are so slack, this scenario is not very relevant to policy. As it stands, Kyoto is just a pilot program. If all industrial countries (currently participating or not) met the nego- tiated Kyoto limits, it would merely delay the buildup of greenhouse gases by a few years. To limit CO2 buildup to prudent levels, reduc- tions of tens of billions of tons a year are needed by mid-century. To attempt meaningful mitigation of climate change, the protocol would have to drastically limit emission allowances. But doing so might drive the price of compliance so high that countries would refuse to sign on. Solution: This is where cheapness comes in. By incorporating avoided deforestation into the global climate strategy, the world could afford to set a more ambitious goal for reducing CO2 buildup. In the Kyoto context that would mean tightening emission allowances while allowing avoided deforestation as a source of emission reductions. By increasing both demand and supply, the price can stay around acceptable levels for all parties, but the climate impact is greater. "Deforestation Avoidance Has to Be Permanent to Be Useful--but It Is Impossible to Secure Permanence" Problem: Buyers of forest carbon want permanent agreements, while sellers want temporary ones. For buyers the problem is this. Because mitigating climate change requires stabilizing CO2 concentrations, 198 M O B I L I z I N G G L O B A L I N T E r E s T s F O r F O r E s T C O N s E rvAT I O N many people assume that every project to reduce CO2 emissions must have a permanent effect. Many energy projects do have permanent effects. replacing a diesel generator with a windmill today means that less fuel will be burned this year. Even if the windmill breaks and the generator is put back in service next year, CO2 emissions will have been reduced--the atmosphere is a little cleaner than it would have been without the windmill. But forest conservation is riskier. Forests can burn. Climate change may imperil tropical forests if temperatures rise and rainfall decreases. And drastic changes in politics or markets may lead the heirs of today's forest owners to repudiate decades-old conservation commitments. Given these risks, buyers worry that it is impossible to sign an agreement today that securely guarantees carbon sequestra- tion into the distant future. And without such a guarantee, they see no benefit from sequestration or reduced deforestation. sellers, on the other hand, may not want to sign such an agree- ment precisely because it closes off future options. Agricultural technologies and markets change rapidly, and expanded transport networks can transform development possibilities for once remote regions. so forest owners may not want to commit to conservation forever. Solution: recognize that avoided deforestation is valuable even without a guarantee of permanence. Carbon sequestration doesn't have to be permanent to be part of a climate change mitigation pro- gram. There are three ways that temporary commitments to carbon sequestration buy time to act on climate change: · Temporary sequestration buys insurance against catas- trophe in the face of uncertainty. The climate system is unstable. small changes can trigger large and irrevers- ible impacts, such as those that apparently shifted the sahara from being heavily vegetated to desert (Foley and others 2003; schneider 2004). There's a fear that too much CO2 in the atmosphere, or too rapid a rise in CO2, could have the same kind of catastrophic effect. But we don't know the thresholds beyond which such a catastrophe could occur. In the face of such igno- rance, it is prudent to buy insurance--that is, to try to keep CO2 levels low and rising slowly. Gitz, hourcade, and Ciais (2006) show that for- est carbon could be a crucial, cost-effective part of a 199 AT L O G G E r h E A d s ? long-term climate change mitigation program. In their model, inexpensive forest carbon offers insurance over the next few decades--after which the world may be better able to assess the risk of catastrophe. If a dan- gerous threshold is then imminent, the world could continue to rely on forests as a carbon sink, or step up investments in geological carbon sequestration. · Temporary sequestration could be a bridge to a clean energy future. Under Kyoto rules, industrial countries need to meet limits on total carbon emissions. They can park their carbon in trees temporarily, but when their storage contracts are up, they need to put that carbon someplace else--or reduce emissions some- place else. This strategy will work nicely if, at the end of the contract term, there are new, cheaper opportuni- ties for storing carbon or reducing emissions. Translated from the project to global scale, this sug- gests that a temporary, renewable decision to protect forests could buy time for technological advance. The strategy would be to protect threatened forests with low opportunity costs. Over time those costs might rise if there is pressure for agricultural expansion. develop- ment of emissions-reducing technologies would then allow the option of substituting emission reductions for continued forest maintenance. (But, as the next section suggests, forestholders at that time might choose not to exercise that option.) For the global community it makes sense to approach climate change mitigation through a pro- gram that uses not-necessarily-permanent avoidance of deforestation as one way to buy time for more effec- tive investments in energy research and development. There is no need to tie the two approaches at a project level, but rather to move toward simultaneous global implementation of avoided deforestation and more vig- orous research and development. The faster that cheap emissions-reducing technologies are developed, the less time has to be bought through temporary seques- tration--potentially allowing forest owners to exercise their option of forest conversion. 200 M O B I L I z I N G G L O B A L I N T E r E s T s F O r F O r E s T C O N s E rvAT I O N · Temporary sequestration could become permanent. however, the history of the forest transition suggests that "temporary" sequestration could bridge the trough of the transition and end up being permanent. Many places face temporary pressures to convert forests for small gains. A 20- to 40-year effort to halt deforestation would not involve large opportunity costs, so equitable compensation could be arranged. At the end of that period, rising wages and appreciation of biodiversity values could prompt a reevaluation of the desirability of forest conversion. The forest owner and the host country may not want to exer- cise their option for conversion at that time. Thus tem- porary efforts to avoid deforestation provide a valuable climatic service and may end up being permanent. "If You Protect One Forest, Someone Will Just Cut Down Another" Problem: does it really do any good to protect a forest plot from con- version to agriculture, or to reforest a working pasture? Won't mar- ket pressures just push someone else to deforest some other plot, to meet demands for food and employment? This problem is called leakage or slippage, and it occurs in many contexts where a project acts locally but has distant repercussions. It's a concern in policies that seek to retire farmland to in order to prevent erosion or shore up commodity prices--do the farmers retire one field and open another? It also occurs in projects intended to reduce energy use and associated carbon emissions: switching a city from coal to wind power nudges down the price of coal slightly. Elsewhere, millions of people respond by increasing their coal con- sumption a bit. such effects can add up to a large proportional dimi- nution of the putative gains at the project site. Solution: Leakage from forest protection isn't necessarily hect- are for hectare (Chomitz 2002), as a naïve view would suggest. sup- pose that a forested property is about to be converted to pasture, but is protected instead. The immediate effect is to drive up the price of beef a scintilla and to send a small amount of capital and a smaller amount of labor looking for other opportunities. One possibility is that the cowboys and ranchers move to an adjacent forest plot and set up a ranch there. But it is also possible that another ranch, pos- sibly a distant one, intensifies a bit, adding a few animals and farm- 201 AT L O G G E r h E A d s ? hands. This is especially likely if the protected forest would have been used for low-intensity grazing. In addition, the slight upward pressure on beef prices may nudge some consumers toward chicken. In sum, leakage will be smaller if other parts of the economy can intensify production and absorb the freed capital and labor; and if consumers are sensitive to the price of beef (or whatever commod- ity is affected by the forest project). Leakage can be moderate even without any effort to control it. The U.s. Conservation reserve Program pays farmers to revegetate erosion-prone land. Wu (2000) found leakage of about 20 percent in terms of area and 9­14 percent for erosion prevention. In other words, for every 5 hectares of land put into the program, 1 hectare of forest was converted to agriculture outside it. But the newly con- verted land was less erosion-prone than the protected land. Murray, McCarl, and Lee (2004) simulate the impacts of a hypo- thetical U.s. program that would protect forestland from agricultural conversion, putting it under sustainable timber management instead. depending on where the program was implemented, carbon leakage ranged from ­4 percent (implying a gain in carbon sequestration outside the program) to 73 percent. The different outcomes could be due to differences in whether the system responds through extensifi- cation (land conversion) or intensification (higher productivity). The solution to leakage, then, is to neutralize it by encouraging sustainable agricultural intensification in nonforest areas--intensi- fication that soaks up the workers, commodity supply, and capital diverted by forest protection. And of course it is important to seek intensified systems that do not produce environmental burdens such as agrotoxic or nitrogen emissions. "It's Too Expensive to Monitor Carbon" Problem: It takes a fair amount of effort to measure the amount of carbon in a tree, let alone on a farm. Measuring changes over time makes things even more complex. Is it affordable to gauge the impact of carbon sequestration efforts? Solution: Measuring forest carbon, in a district or a nation, involves two steps (to oversimplify a bit). The first is estimating how much carbon there is in a tree of a given size, based on its volume and characteristics. The second involves counting the num- ber of trees of different sizes and multiplying by the amount of car- bon in each tree. The second step could be done by tallying every tree--difficult even in a small forest. But technology is making this 202 M O B I L I z I N G G L O B A L I N T E r E s T s F O r F O r E s T C O N s E rvAT I O N approach cheaper. For instance, it is possible to take aerial pictures and have computers recognize trees and estimate their volumes. still, the cost per tree or hectare is significant, as the airplane must cover the countryside in many low-altitude swaths. statistical techniques offer potentially huge economies of scale in carbon measurement (Chomitz 2002). statistics can be used to esti- mate the number or volume of trees based on a sample. And statistical methods have a remarkable property, familiar from household surveys: the accuracy of an estimate depends on the size and representative- ness of the sample, not the size of the population being sampled. With 2,000 interviews it is possible to accurately assess mean household income--for a city, province, or nation. hence there are huge econo- mies of scale, in costs per ton, of measuring changes in carbon stocks at a national rather than project level. Although the statistical issues in drawing appropriate samples can get complicated, the principle is clear: enlisting a few statisticians can drastically reduce the number of fieldworkers or aircraft needed to measure carbon. Implementing Incentives for Avoided Deforestation The solutions to these concerns about forest carbon are mutually supportive. They strongly suggest working at a national level, to incorporate leakage-neutralizing policies and drastically cut the costs of carbon monitoring. Potential steps toward avoiding defores- tation in developing countries include: · Agreement by some industrial countries to provide incentive payments for developing countries to reduce deforestation. · development of national systems for forest and carbon monitoring, including win-win steps to reduce exces- sive deforestation. · Elaboration of the forest carbon infrastructure into national programs for deforestation avoidance. These programs would use the international incentive pay- ments to fund deforestation reduction activities. · Emphasis on neutralizing emissions though sustainable agricultural development. · Incentive payments would be "pay as you go"--based on annual reductions against a reference level. 203 AT L O G G E r h E A d s ? Fostering Sources of Global Finance for Avoided Deforestation and Supporting Research and Development on Emissions Programs for avoided deforestation in developing countries will require global finance. There are different ways to provide it. One is within a Kyoto framework: some countries accept a cap on emis- sions but can meet that cap by purchasing emission reductions abroad, including from averted deforestation. This approach could lead to a market for emission reductions based on forest carbon, with pricing based on supply and demand. Developing National Forest Carbon Infrastructure To manage and use incentive payments, investments must first be made in building capacity and creating needed physical and institu- tional infrastructure. These investments, supported by donor financ- ing, would include win-win investments that reduce deforestation pressure in any case. At the core of the system are institutions and hardware for monitoring forest cover, forest and land fires, and carbon. Initially the system could track land cover--providing rapid, indicative mea- sures of change. Later it could become a more comprehensive and accurate carbon monitoring system, combining new remote sensing technologies (such as MOdIs) with ground-based observations. such a system can do far more than provide the carbon read- ings needed for incentive payments. It could aid in land use plan- ning, forest fire prevention, and forest law enforcement. To facilitate this, the monitoring system would map the boundaries of protected areas, forest concessions, indigenous areas, and large private prop- erties. Authorities could then use this information to help enforce forest laws and improve management of public forestlands. Public disclosure of these data would raise awareness of the issues and might help build constituencies for enforcement of laws against ille- gal forest conversion and logging. Creating National Programs to Reduce Deforestation The next step is to develop a blueprint for a program of domestic institutions, policies, and initiatives to reduce emissions from defor- estation and, probably, increase carbon storage in agricultural and forest landscapes. This program would translate international incen- tive payments for reduced deforestation into incentives for forest owners to contribute to the achievement of these reductions. 204 M O B I L I z I N G G L O B A L I N T E r E s T s F O r F O r E s T C O N s E rvAT I O N One approach would be through direct pass-through of incen- tive payments to individual property owners. But this approach has disadvantages. It doesn't address illegal deforestation or deforesta- tion on public land. It doesn't facilitate government policies that can affect entire landscapes. It fails to recognize the contribution of agricultural intensification in reducing leakage and facilitating emis- sion reductions. And measurement, monitoring, and transactions costs are prohibitively high at the property level, especially for small properties. An alternative is to delink incentives to the nation from incen- tives to individuals and firms. The national government can use incentive payments to fund diverse interventions in different sectors and locations. These interventions might cover the range of options discussed earlier in this report, such as: · Paying communities for reduced deforestation or natu- ral regeneration. · Funding fire prevention programs. · Improving tenure security. · Enforcing regulations against illegal deforestation and logging. · setting up taxation of large-scale land clearance. · Promoting off-farm employment. · Intensifying agriculture in favorable areas to attract or divert workers from marginal lands at the forest fringe. · Implementing strategic planning of road improvements. · supporting community forestry where it deters conver- sion to agriculture. These programs might then be certified for participation in a globally financed incentive program. Certification could facilitate grants or loans from donors or international financial institutions to invest in the programs. To be certified, programs would have to meet some basic criteria. For instance, the monitoring system would have to eliminate any perverse incentives to replace natural forests with planted ones. There are strong benefits to combining forest protection pro- grams with agricultural and silvicultural intensification programs. 205 AT L O G G E r h E A d s ? First, the latter neutralize leakage. For instance, road improvements in less-forested areas can lead to agricultural intensification and increased demand for labor, reducing migration to the forest fron- tier. second, the combination creates a broad constituency of benefi- ciaries who can support program implementation. How Should National Incentives Be Set? Funders and recipients of incentive payments will be keenly inter- ested in how prices and quantities are set. The framework envisions a negotiated reference level (RL) of emissions or net emissions. Incentives would be offered for reductions below that level. (This is different from offering an incentive based on total forest area.) The key terms to be negotiated are how to set the reference level, how much to reward reductions below it, and whether and how to dis- courage emissions above it. First, if the reference level is set above the unobservable base- line (business as usual) of emissions, the country will receive rents--pure transfers unrelated to emission reductions. If these rents are too large, funders may decline to participate. But if the refer- ence level is set too far below the baseline, deforesting countries may decline to participate. reference levels could be set at historical emission levels, but these may be difficult to assess if data are lack- ing, or may reflect market conditions that no longer exist. setting reference levels at current emission levels would introduce moral hazard because countries might be tempted to increase deforesta- tion to obtain a higher target. An alternative is to compute a normative reference level. The normative estimate would be based on a standardized estimate of the rate of increase in agricultural production, adjusted for an esti- mate of the rate of increase in agricultural productivity as well as the mean carbon content of forestland at the agricultural margin. separate estimates could be made for logging-related emissions and the rate of abandonment of current lands. A normative reference level would tend to reward countries already trying to reduce defor- estation, without introducing perverse incentives to increase defor- estation to get more credits. second, what is the reward for reducing emissions below the reference level, and should the temporary nature of the reductions be dealt with? For simplicity, suppose that the reductions result from national incentives and are not tied to the Kyoto Protocol's carbon market or a successor. In that case funders and recipients could 206 M O B I L I z I N G G L O B A L I N T E r E s T s F O r F O r E s T C O N s E rvAT I O N negotiate a payment amount per period per ton-year. The calcula- tions would be made as follows. reductions in year t would be cal- culated as: Rt = RL ­ Et, where Et is measured net emissions. R could be positive (meaning a reduction relative to the reference level) or negative (implying emis- sions above the reference level). The payment at time T would be: T max(0,P Rt), t=1 where P is the price per ton-year and t = 1 marks the beginning of the program. This formula is cumulative because it focuses atten- tion on carbon storage. Each year, the country is rewarded if its actual carbon storage is greater than the baseline implied by the emissions scenario. The price P can be thought of as a storage fee, paid per ton per year. suppose the country protects two hectares from deforestation in year 1, and an additional hectare in year 2, and suppose that each hectare would release 100 tons of CO2 if deforested. Then R1 = 200, R2 = 100; the country would receive 200P in year 1, and 300P in year 2. suppose however, that in year 2, no additional hectares were protected, and in fact one of the previ- ous year's protected hectares was deforested, so that R2 = ­100. Then only 100P would be paid in year 2. how should the price P be set? Ton-years have value because delaying emissions is valuable and because there is a significant chance of unintended permanence. But these values, while real, are difficult to compute on the basis of available information. so P would have to be decided by negotiation. A natural reference point would be the current interest rate times the market or normative price of a carbon allowance. This is the rental value of a permanent allowance. For example, under the EU Emissions Trading scheme (ETs) a ton of CO2 abatement is currently priced at about $20. At 6 percent annual interest, this implies a payment of $1.20 a ton per year. At that rate, averting deforestation of a hectare of moist rain- forest might return a few hundred dollars a year. But even a price based on 6 percent of $3 per ton of CO2 might still return an amount comparable to the annual payment rate in Costa rica's payment for environmental services program (see chapter 6). This approach could be made Kyoto-compatible by setting an exchange rate between ton-years and permanent tons. There is a 207 AT L O G G E r h E A d s ? long, inconclusive history of discussion on the proper exchange rate for ton-years. Again, setting it at, say, 6 percent of a permanent reduction might be a starting point for discussion. Though the obstacles to setting up avoided deforestation pro- grams are considerable, so are the potential benefits. Moreover, solu- tions to those obstacles might be self-reinforcing (table 7.1). The policies discussed here require fairly sophisticated insti- tutional capabilities and so may not be immediately applicable to all forested countries. But countries could proceed in steps--start- ing by creating forest carbon infrastructure and proceeding to pilot tests of national-to-individual incentives. These early stages might be rewarding to participating countries and beneficial to the global climate, while providing information that would improve the design of international incentives. Table 7.1 Policies to Reward Avoided Deforestation Can Have Synergistic Effects Permanence/ contribution to long-run climate Acceptability to Policy mitigation host country Leakage Monitoring Bundle with X commitment to research and development Invest in X X X agricultural intensification Set national X X baselines Secure temporary X commitments from host country Focus on X X marginal areas with ephemeral pressure or risk of irreversibility Catalyze technology X X diffusions 208 M O B I L I z I N G G L O B A L I N T E r E s T s F O r F O r E s T C O N s E rvAT I O N Related Opportunities for Biodiversity Conservation The emergence of global carbon markets may blaze a trail for new approaches to global biodiversity finance. A precondition would be mobilization of significant funding. This might be achieved through biodiversity offset requirements (ten Kate, Bishop, and Bayon 2004). For instance, large mines might be encouraged or required to offset any nonmitigatable habitat destruction by buying offsets--protec- tion of areas elsewhere of equivalent or greater biodiversity or envi- ronmental importance. Offsets could be done by putting an equivalent domestic area under protection or by contributing to a fund for conservation of globally significant biodiversity. such funds could solicit contracts with landholding individuals, communities, and local or national governments. Would-be participants could submit bids specifying the services to be provided and the compensation requested, and the bids could be ranked by cost-effectiveness--as in the U.s. Con- servation reserve Program and Australia's BushTender. Geographic criteria might include the combination of temporary threat and likely irreversible degradation in the absence of action. A side benefit of this approach is that it might stimulate investments in better survey- ing the condition and geographic distribution of biodiversity, using new technologies such as genetic bar coding. That in turn could catalyze renewed, more sophisticated, and more productive markets for bioprospecting from which developing countries and communi- ties could benefit. Summary Carbon storage and biodiversity conservation are forest services that benefit large but diffuse global constituencies. Mobilizing those con- stituencies to finance forest maintenance and negotiating and imple- menting agreements with forestholders pose institutional challenges of planetary magnitude. still, the potential gains to all parties are big enough to motivate such efforts. Endnote 1. Based on pointcarbon.com, August 31, 2006. 209 Juan Pablo Moreiras / Fauna & Flora International / Comisión Centroamericana de Ambiente y Desarrollo photo archive. CHAPTER 8 Conclusions and Recommendations A re development and the environment at loggerheads in tropi- cal forests? This report has shown that trade-offs sometimes exist between the two. Poverty and deforestation are not closely linked at the local level, so we shouldn't expect fixing one problem to automatically solve the other. Indeed, some deforestation contributes a lot to development, poverty alleviation, or both. And when poor people deforest for paltry gains, it's often because the alternatives--including forest maintenance--are less attractive. The vast environmental benefits of forests have been difficult to tap in ways that motivate forest conservation while improving livelihoods. Moreover, forest poverty and deforestation are difficult prob- lems to solve. They revolve around the allocation and enforcement of rights. Strong, equitable institutions are needed to resolve these problems--but such institutions are in short supply in many devel- oping countries. Still, there are grounds for hope. Technological and institutional innovations create possibilities for catalyzing change and for mini- mizing or transcending trade-offs. But problems must be properly diagnosed. Challenges for poverty, equity, and the environment are systematically different in areas in and beyond the agricultural frontier, and different kinds of management institutions are needed at the international, national, and local levels. Various policy and institutional interventions could help reduce poverty, ease environ- mental damage, and make allocations of wealth more equitable (box 8.1). 211 AT L o g g e R h e A d S ? Box 8.1 This Report's Recommendations International level to agriculture where doing so offers high, · Mobilize carbon finance to reduce defores- sustainable returns and does not threaten tation and promote sustainable agriculture. irreplaceable environmental assets. · Mobilize finance for conservation of glob- · Plan for rational, regulated expansion of ally significant biodiversity. road networks--including designation of · Finance national and global efforts to moni- roadless areas. tor forests and evaluate the impacts of forest · experiment with new ways of providing projects and policies--including devolution services and infrastructure to low-density of forest control. populations. · Foster the development of national-level research and evaluation organizations Frontier areas through twinning with established foreign · equitably assign and enforce property partners. rights. · Plan and control road network expansion. National level · discourage conversion in areas with hydro- · Create systems for monitoring forest con- logical hazards, or encourage community ditions and forest dwellers' welfare, make management of these watersheds. land and forest allocations and regulations · Use remote sensing, enhanced communica- more transparent, and support civil society tion networks, and independent observers organizations that monitor regulatory com- to monitor logging concessionaires and pro- pliance by government, landholders, and tect forestholders against encroachers. forest concessionaires. The prospect of car- · Consider using carbon finance to sup- bon finance can help motivate these efforts. port government and community efforts to · Make forest and land use regulations more assign and enforce property rights. efficient, reformulating them to minimize · encourage markets for environmental ser- monitoring, enforcement, and compliance vices in community-owned forests. costs. economic instruments can help. Disputed areas Areas beyond the frontier · Where forest control is transferred to local · Avert disruptive races for property rights by communities, build local institutions with equitably assigning ownership, use rights, upward and downward accountability. and stewardship of these lands. · Where community rights are secure and · options for forest conservation include markets are feasible, provide technical assis- combinations of indigenous and community tance for community forestry. rights, protected areas, and forest conces- · Make landholder rights more secure in "for- sions. Still, some forest may be converted ests without trees." 212 C o n C L U S I o n S A n d R e C o M M e n dAT I o n S Box 8.1 (continued) · When forest tenure is secure, use carbon systems--through research and develop- markets to promote forest regeneration and ment, extension efforts, community organi- maintenance. zation, and reform of agriculture and forest regulations. Mosaiclands · develop a wide range of markets for envi- · Reform regulations so that they don't penal- ronmental services--carbon, biodiversity, ize tree growing. water regulation, recreation, pest control-- · Promote greener agriculture--such as inte- to support more productive, sustainable grated pest management and silvopastoral land management. International Level Two areas stand out for international cooperation. Financing Forest Environmental Services Internationally financed incentives for avoiding deforestation and stimulating forest regrowth could catalyze global forest conservation and agricultural development while cutting the costs of mitigating climate change. Any serious effort to grapple with global climate change must place an explicit or implicit charge on carbon dioxide (Co2) emissions. And at any reasonable charge on Co2 emissions, there are huge dividends to engaging in more intensive, labor- absorbing agriculture on degraded lands--instead of deforestation that yields trifling, ephemeral gains. That substitution, which would also provide domestic benefits, won't happen automatically because private incentives work against it. But sharing the global dividends of Co2 reduction could provide the funds and motivation for needed national-level efforts. The Un Framework Convention on Climate Change, whose 189 signatories have agreed to the goal of stabilizing greenhouse gas concentrations in the atmosphere, provides a natural venue for negotiating financ- ing for international incentives. These efforts would have to be coordinated with research, development, and diffusion of environmentally friendly agricul- tural technologies and practices. Some approaches to agricultural intensification can create or exacerbate environmental problems, 213 AT L o g g e R h e A d S ? including eutrophication, pesticide pollution, and emergence of new pathogens related to animal husbandry. new technologies--such as integrated pest management and other landscape management tech- niques--could mitigate the potential side effects of intensification. The international community could also provide direct incen- tives for global biodiversity conservation. Financing sources might include industries dependent on biodiversity, such as agriculture based on plants with wild relatives in forests. Financing could be funneled through national programs for environmental service payments. Addressing Monitoring and Information Needs The international community could undertake monitoring efforts that would have immediate payoffs while contributing to these long-term financing goals. A priority is to fund and coordinate basic monitoring on the rate, location, and causes of global deforestation and forest poverty and the impacts of project and policy interven- tions. Without this information, policy makers are flying blind, and interest groups lack a solid basis for dialogue. For instance: · despite intense concerns about deforestation, among developing countries only Brazil and India report for- est cover on a regular basis. (Indonesia is developing a system.) For Africa, estimates of deforestation vary by a factor of 10. · despite billions of dollars spent setting up and run- ning protected areas, there has been little analysis of their conservation and development impacts and of how these are related to their funding, management strategies, and context. Similarly, there has been little monitoring and analysis of the impacts of the massive transfers in forest tenure of recent years. · despite hopes for securing support for forest conser- vation based on local environmental services such as hydrological benefits, there are few studies and tools for quantifying those benefits and relating them to specific interventions in specific places. And measure- ments of weather and rivers, the foundation for such studies, are increasingly difficult because meteorologi- cal and gauging stations are being abandoned. 214 C o n C L U S I o n S A n d R e C o M M e n dAT I o n S · despite concerns about forest poverty, information is spotty on the numbers and conditions of people in two distinct poverty situations: deep forest (highly forest- dependent people in remote areas) and mosaiclands (who get a small but significant share of their income from forests). · despite concerns about global biodiversity loss, infor- mation on the global distribution of biodiversity is inadequate. Conservation scientists have made great strides in organizing available data, but systematic sampling is lacking. With the possible exception of the last, these information gaps are relatively easy to remedy. new remote sensing technologies make it feasible and affordable to identify hotspots of deforestation. Rid- der (2006) estimates that it would cost $12 million to create a global network for forest monitoring that could produce annual, medium- resolution estimates of deforestation. That price tag includes support for building local capacity to acquire and interpret remote sensing data. A World Bank­WWF survey tool for managing protected areas (Stolton and others 2003), already in use by the global environ- ment Facility (geF), could be applied more widely and integrated with quantitative indicators of biodiversity status. household survey instruments could be adjusted to better account for forest income. In addition, collaborative research institutes--twinning staff from industrial and developing countries--could conduct monitoring and evaluation studies from a policy perspective, as well as research and development on land and forest management. Such efforts would build capacity and generate analytic and research results and could nurture the development of policy entrepreneurs as described in Steinberg (2001). National Level At the national level, strengthening the voice and influence of dif- fuse interests--for environment and for poverty--is essential to institutional reforms. environmental councils might be one way to mobilize people affected by forest externalities. There may be a virtuous circle between mobilizing these interests and generat- ing better information on forest conditions: interest groups demand 215 AT L o g g e R h e A d S ? information, which empowers them to negotiate better outcomes. The critical factor is local demand for forest regulation, which exter- nal sources of finance (such as for carbon) could help spark. national monitoring of forests is increasingly easy thanks to improvements in remote sensing and communications. A national forest monitoring program--combined with mapping of indigenous areas, protected areas, forest concessions, and other tenure zones-- can form the basis for better forest law enforcement and fire control, and for certification or rating of large logging firms and landowners. Public disclosure of this information is important for encouraging enforcement. Internationally financed incentives for forest carbon could be a powerful inducement for countries to set up national monitoring. Within countries, this report's policy recommendations are dif- ferentiated by three forest regions: beyond the frontier, at the fron- tier and in disputed areas, and within the frontier. Areas beyond the Frontier A few countries have large, remote forests beyond the range of most agriculture or timber extraction. Pressures to exploit those forest resources will likely eventually arise--in some places chainsaws are already almost within earshot. In other, more remote forests such pressures may be decades off and will result from technological changes. For instance, the development of productive, forest-com- peting biofuel crops could greatly increase global land demands and pressures for forest conversion. now--before those pressures arise and stakeholders are entrenched--is the time to think about how to manage those for- ests, accounting for their unique ecological values. This process entails recognizing and enforcing indigenous land claims where applicable and demarcating and institutionalizing protected areas. Indigenous and protected areas have been relatively successful in deterring deforestation, though the determinants of their effective- ness and their impacts on livelihoods are poorly understood. now is also the time to think about long-term planning of road networks in areas beyond the frontier. Strong technical methodologies exist for planning roads and protected areas, but must be exercised in a context that mediates between competing interests. Poverty is often high in transfrontier areas, as is the proportion of indigenous people. Protecting rights and building capacity can help empower these isolated people. In addition, innovative tech- 216 C o n C L U S I o n S A n d R e C o M M e n dAT I o n S nologies--such as satellite communications--can lower the cost of providing services such as health and education to low-density populations. Frontier and Disputed Areas governance at the frontier and in disputed areas requires equitable, secure assignment of land and forest rights. There is no easy pre- scription for achieving that, since forests are subject to elite capture as they take on value. Although models of participatory zoning and conflict resolution exist for small areas, challenges remain in scaling them up to legitimately disentangle forest claims at the provincial and national levels. A special challenge is rationalizing tenure in "forests without trees"--areas where agriculture is legally prohib- ited but forest has been severely degraded. Many forests are nominally owned by governments, but poorly administered. devolving nationally owned forests to local owner- ship and control might result in a more equitable allocation of forest rents and better management. But local ownership and management are no panacea, because communities face their own institutional issues, including elite capture. Where markets are accessible, build- ing capacity, providing marketing assistance, and streamlining regu- lations could help community forestry. In some contexts, putting forests under regulated logging con- cessions could provide income to the forest owner (whether state or community) while maintaining environmental values. new tools-- including auctions and independent monitoring--are available for ensuring that concession awards and operations contribute to the public good. Shifting the balance from forest degradation toward for- est maintenance could require technical assistance in production, management, and marketing. But neither concessionaires nor local communities have strong incentives to manage forests for national or global environmental benefits. So regulations or incentives (or both) will be needed to secure those benefits. efficient regulation focuses on the flexible achievement of clear environmental goals, uses criteria that are cheap and easy to monitor, does not lend itself to petty rent seeking by enforcers, and imposes low opportunity costs on those who comply. Because roads are an important driver of deforestation in fron- tier areas, some control should be exercised over the expansion of road networks--especially discouraging road extensions into areas 217 AT L o g g e R h e A d S ? with low agricultural potential. In other places clarification of forest tenure should precede road extension. Mosaiclands Within the frontier, high-density mixes of people and forest frag- ments generally lead to both stronger pressures for deforestation and greater incidence of environmental benefits. But this is a hetero- geneous region, presenting a range of challenges. Where marginal uplands are near agriculturally favorable lowlands, intensification of the latter might ease pressure on hillside forests. elsewhere, balancing environmental externalities against pressures for forest conversion or exploitation will require a variety of interven- tions. There is scope for exploring technological and institutional inno- vations for greener management of agriculture and forest mosaiclands. Again, efficient regulation is needed, placing on landholders only restrictions and responsibilities that provide clear external benefits, are easy to monitor and enforce, and do not burden poor people. Areas within the frontier have the greatest scope for developing payment systems for environmental services. But if they are to be effective in delivering the promised services, these systems must be tightly focused on efficiency. International finance for forest carbon or threatened biodiversity could be important in these areas, which are likely to be home to threatened species. Accelerating the Forest Transition As development progresses at the national level, rising wages attract farmers to urban employment and away from low-return farming at the forest fringe (see box 2.1). In some countries a demographic transition--with shrinking youth populations--will intensify this trend, driving up wages and reducing the number of people willing to live hard lives at the frontier. Many developing countries are at the cusp of another demographic transition, with their rural popula- tions poised to decline (figure 8.1).1 In much of Sub-Saharan Africa, however, this transition is distant. The prospect of a forest transition isn't cause for complacency. True, industrial countries have seen a remarkable rebound in forest cover. For instance, despite its high population and economic densi- ties, Japan is about two-thirds forest. But the quality and nature of regenerated forest may be quite different from the original. This risk is profound in tropical forests, 218 C o n C L U S I o n S A n d R e C o M M e n dAT I o n S Figure 8.1 Some Forested Countries Will See Shrinking Rural Populations 160 140 100)= 120 2005 100 (Index: 80 60 Brazil Peru Madagascar Cameroon population 40 India Indonesia Vietnam Mexico Rural 20 Congo, Dem. Malaysia Republic of 0 1950 1960 1970 1980 1990 2000 2010 2020 2030 Source: UN Population Division 2004. where ecologies are far richer and more complex than in temperate regions, and where soils are poorer and more degradable. The dan- ger is that, for ecological reasons, the pulse of tropical forest clear- ance over the next few decades will often yield paltry benefits and leave behind not a renaissance of the original forest but a degraded landscape where biodiversity and carbon storage have been perma- nently impaired. At the global level, pro-poor growth, the creation of sustain- able cities, and the development of agricultural technologies that are intensive, labor-absorbing, and environmentally benign can help accelerate the forest transition. Incentives for carbon storage and biodiversity conservation can help countries maintain these assets, bridging the trough of the transition. It is in this important sense that poverty alleviation, development, and forest conservation are fully aligned. Endnote 1. however, the figure shows a cautionary lesson from Brazil. There, market forces and road building drove deforestation even as the rural population declined. 219 Appendix A: Tables Table A.1 Findings of Studies Assessing How Road Proximity Affects Deforestation Location, square Variable capturing Study kilometers, year(s) road effect Impact on deforestation Andersen and Reis Brazilian Amazônia, Road length +*** (1997) ~5 million, 1970 to 1985 Andersen and Brazilian Amazônia, Unpaved road density ­ (at least **) others (2002) ~5 million, 1980 to 1985 Unpaved road × + (at least **) cleared land Paved road density n.s. 1985 to 1995 Paved road density + (at least **) Unpaved roads + (at least **) density Paved roads × cleared ­ (at least **) land Unpaved road × + n.s. cleared land Bray and others Mexico, 7,300, 1976 to Distance to roads n.s. (1976­84) (2004) 1984 and 1984 to 2000 ­*** (1984­2000) Chomitz and Gray Belize, 11,712, 1989­92 Distance to market ­*** (1996) Chomitz and Brazilian Amazônia, Proportion of land +*** Thomas (2003) 4.86 million, 1970 to within 50 km from 1985 main federal roads Distance to cities with ­*** populations > 25,000 Distance to cities ­ (n.r.) with populations > 100,000 Cropper, Griffiths, Thailand, 514,000, 1976 Road density +*** and Mani (1999) to 1989 220 A p p e n d i x A Control for spatial Control for Control variables (impact on deforestation) auto-correlation? endogeneity? Soil quality Slope Altitude Yes No n.i. n.i. n.i. No Yes n.i. n.i. n.i. No Yes n.i. n.i. n.i. Yes Yes n.i. n.i. n.i. Yes Yes +*** ± n.i. No Yes n.r. n.i. n.i. n.a. Yes + ­*** n.a. (continued) 221 AT L o g g e r h e A d s ? Table A.1 (continued) Location, square Variable capturing Study kilometers, year(s) road effect Impact on deforestation Cropper, Puri, and Northern Thailand, area Cost to nearest ­*** Griffiths (2001) not reported, 1986 market Deininger and Mexico, 160,000, Distance to nearest ­*** Minten (2002) 1980­90 paved road Etter and others Colombia, 1.1 million, Distance to roads, ­*** (2006) 1998 town Geoghegan and Mexico, 22,000, 1988 to Distance to roads ­*** (1988­92) others (2001) 1992 and 1992 to 1995 ­*** (1992­95) Distance to market +** (1988­92) +*** (1992­95) Distance to village +*** (1988­92) ­*** (1992­95) Kirby and others Brazilian Amazônia, 5 Distance to roads ­** (2006) million, 1999 McConnell, Madagascar, 940, 1957 Distance from village ­ n.r. Sweeney, and to 2000 Mulley (2004) Mertens and Pará, Brazil, 56,300, 1986 Distance to main road ­*** (planned colonization, 1) others (2002) to 1992 +*** (small-scale colonization, 2) ­*** (medium-size colonization, 3) ­*** (large fazendas, 4) Distance to secondary ­** (1) road ­*** (2) +* (3) + n.s. (4) Distance to village +** (1) ­*** (2) +*** (3, 4) 1992 to 1999 Distance to main road ­ *** (1) +*** (2, 3) + n.s. (4) Distance to secondary ­*** (1, 2, 3, 4) road Distance to village ­*** (1, 2, 3, 4) 222 A p p e n d i x A Control for spatial Control for Control variables (impact on deforestation) auto-correlation? endogeneity? Soil quality Slope Altitude Yes No +*** ­*** ­*** No No +* ­*** ­*** No No +*** ± n.i. No No +*** +*** ­*** Yes No n.s. n.i. n.i. Yes No n.i. ­ n.r. ­ n.r. Yes Yes n.i. n.i. ± Yes Yes n.i. n.i. ± (continued) 223 AT L o g g e r h e A d s ? Table A.1 (continued) Location, square Variable capturing Study kilometers, year(s) road effect Impact on deforestation Mertens and Bolivia, 364,000, <1989 Distance to roads ­*** others (2004) and 1989 to 1994 and to Santa Cruz Müller and Munroe Vietnam, ~1,390, 2000 Distance to long- + n.s. (sample includes only less- (2005) established road remote areas) network Müller and Zeller Vietnam, ~2,390, 1975 to Distance to nearest ­*** (1975­92) (2002) 1992 and 1992 to 2000 all-year road ­*** (1992­2000) Distance to district +*** (1992­2000) capital n.i. (1975­92) Travel time to all-year +*** (1992­2000; n.s. for road paddy) +*** (1975­2000; n.s. for mixed agriculture) Munroe, Honduras, 1,015, 1987 to Distance to nearest ­*** Southworth, and 1996 village Tucker (2004) Distance out of region ­ n.s. Naidoo and Paraguay, 2,920, Distance to unpaved ­* (smallholders) Adamowicz (2006) 1991 to 2004 roads ­ n.s. (ranchers) + n.s. (soybeans) Distance to paved ­***(smallholders) roads ­ n.s. (ranchers, soybeans) Distance to towns ­ n.s. (all groups) Nelson and Mexico, area not reported, Cost to nearest road ­** Hellerstein (1997) 1973 or village Cost to large + n.s. population center Nelson, Harris, Panama, 15,995, Cost to border ­ (n.s.) and Stone (2001) 1987 to 1997 Cost to port +*** Cost to village ­*** Cost to nearest town ­*** 224 A p p e n d i x A Control for spatial Control for Control variables (impact on deforestation) auto-correlation? endogeneity? Soil quality Slope Altitude Yes Yes ± n.i. n.i. Yes Yes +*** ± ± Yes Yes +*** ­*** ­*** Yes No n.i. +*** ­*** Yes No +*** ­*** ± Yes No n.s. ­*** ­*** Yes Yes +*** ­*** ­*** (continued) 225 AT L o g g e r h e A d s ? Table A.1 (continued) Location, square Variable capturing Study kilometers, year(s) road effect Impact on deforestation Nelson and others Panama, 16,100, Cost of transporting ­ n.r. (2004) 1987 to 1997 wood to market (by road or river) Osgood (1994) Indonesia, area not Extension of roads + n.s. reported, 1972 to 1988 Panayotou and Northeast Thailand, Rural roads extension +* Sungsuwan (1994) 169,000, 1973 to 1982 Distance to Bangkok ­*** Pender and others Uganda, area not reported, Change in distance ­** (2004) 1990 to 1999 to tarmac roads Change in distance ­ n.s. to market Pendleton and Bolivia, area not reported, Walking time to ­** (primary forest) Howe (2002) clearance during 1995 roads + n.s. (secondary forest) Walking time to +*** (primary forest) closest market +** (secondary forest) Pfaff (1999) Brazilian Amazônia, Density of unpaved +*** area not reported, roads 1975 to 1988 Density of paved ­ n.s. roads Pichón (1997) Ecuador, ~70,000, 1990 Distance to roads, ­*** nearest market Reis and Guzmán Brazilian Amazônia, Extension of unpaved +** (1994) 5 million, 1983­87 roads Extension of paved + n.s. roads Distance to state ­ n.s. capital 226 A p p e n d i x A Control for spatial Control for Control variables (impact on deforestation) auto-correlation? endogeneity? Soil quality Slope Altitude Yes Yes n.i. n.i. n.i. n.a. No n.i. n.i. n.i. n.a. No n.i. n.i. n.i. No Yes n.i. n.i. n.i. No No n.i. n.i. n.i. Yes Yes +*** n.i. n.i. n.a. Yes +*** ­*** n.i. Yes Yes n.i. n.i. n.i. (continued) 227 AT L o g g e r h e A d s ? Table A.1 (continued) Location, square Variable capturing Study kilometers, year(s) road effect Impact on deforestation Serneels and Kenya, 10,694, 1975 to Mechanized agriculture Lambin (2001) 1985 and 1985 to 1995 Squared distance to ­*** (1975­85) roads ­*** (1985­95) Distance to roads +*** (1975­85) +*** (1985­95) Distance to village ­*** (1975­85) +*** (1985­95) Distance to Narok ­*** (1975­85) (district seat) ­*** (1985­95) Smallholders (1975­85 model only): Distance to roads ­*** (log) Distance to village ­*** (log) Distance to Narok +*** Southworth and Honduras, 1,015, Distance to roads and ­** others (2004) 1987 to 2000 regional market Tucker and others Guatemala, 1,053, Distance to nearest ­** (2005) 1987 to 1996 town/local market Distance out of +** also increases probability of region (capital city or forest regrowth regional market) Vance and Yucatan, 22,000, Distance to market ­*** Geoghegan (2002) 1984­87 and 1994­97 Wilson and others Chile, 42,000, 1995­96 Distance to roads and ­*** (2005) town Note: Years x­y indicates a single cross-section analysis based on composite forest cover data for the period x­y; years x and y indicates separate cross-section analyses for years x and y; years x to y indicates an analysis of forest cover change between years x and y. n.a. Not applicable; n.i. Variable not included; n.s. Not significant; n.r. Significance not reported; ± Effect differs for different land uses. *, **, *** represent 10 percent, 5 percent, and 1 percent significance levels, respectively. 228 A p p e n d i x A Control for spatial Control for Control variables (impact on deforestation) auto-correlation? endogeneity? Soil quality Slope Altitude Yes No +*** n.i. ­*** Yes No +*** n.i. ­*** No No n.i. +* ­** Yes No n.i. ­** +** No No +*** ­*** ­*** No No +** ­*** ­*** 229 AT L o g g e r h e A d s ? Table A.2 Findings of Studies on How Roads Affect Development Control for Study Location, year(s), and data level endogeneity? Binswanger, Khandker, and India, 1960­82, district level Yes Rosenzweig (1993) Buys, Deichmann, and Wheeler Sub-Saharan Africa, 1999­2004, country level Not applicable (2006) De Castro (2002) Brazil, 1970­96, municipal level No De Vreyer, Herrera, and Mesple- Peru, 1997­2001, household survey, geographic Yes Somps (2002) variables at district level Dewi, Belcher, and Puntodewo (2005) East Kalimantan, Indonesia, 1992­97, spatial No data at village level Escobal and Ponce (2002) Peru, 1994­2000, household survey Yes Fan, Nyange, and Rao (2005) Tanzania, 2000­01, household level Yes Fan and Chan-Kang (2004) Uganda, 1999, household level Yes Fan, Hazell, and Haque (2000) India, 1970­94, district level Yes Fan, Zhang, and Zhang. (2004) China, 1978­2000, province level Yes Fan and Zhang (2004) China, 1996­97, province level Yes Gibson and Rozelle (2003) Papua New Guinea, 1996, household survey Yes 230 A p p e n d i x A Dependent variable Road impact on dependent variable Agricultural output Elasticity (road length): 0.20*** Trade Implementation of a $35 billion regional road improvement project would increase trade by $250 billion in 15 years Agricultural output Elasticity with respect to road density: 0.33*** Consumption growth Road density and percentage of paved roads: n.s. Elasticity with respect to distance to provincial capital: ­0.024* Economic diversity index (heterogeneity Density of provincial and district road: +*** of income sources) Income Road improvements increased average income by 35% in villages with motorized roads Poverty Marginal effect of distance to public transportation facilities on probability of being poor: 0.0022­0.0033 per km depending on region. The coefficient was not significant in Lake Victoria and the southern coast. 30 people lifted out of poverty for each $1,000 invested in roads Agricultural output Returns to government investment in feeder roads: 600% (center--richest) 870% (east) 490% (north--poorest) 920% (west) Poverty reduction 16 (center), 81 (east), 109 (north), and 46 (west) people lifted out of poverty for each $1,000 invested in feeder roads Agricultural output Elasticity (road density): 0.18** (irrigated). Of 13 rainfed zones, 6 display a negative elasticity but only 1 (most productive land) is significant: ­0.28**. The other 7 rainfed zones display positive and significant elasticities, with the least productive land showing the highest magnitude: 0.082** to 1.38** Poverty reduction 0.25 (irrigated areas) and 0.03­5.18 (rainfed areas) people lifted out of poverty for each $1,000 invested in roads Agricultural output Elasticity (road density): 0.099* Poverty reduction 2.2 (coastal region), 6.9 (central), and 8.3 (western) people lifted out of poverty for each $1,000 invested in roads Agricultural output Elasticity (road density): 0.032** Ln (region-specific poverty line) Marginal effect of travel distance (hours) to nearest road: ­0.04**. Expanding roads to be within a two-hour walk of everyone would reduce the number of poor people by 6­12% (continued) 231 AT L o g g e r h e A d s ? Table A.2 (continued) Control for Study Location, year(s), and data level endogeneity? Guimaraes and Uhl (1997) Para, Brazil, 1994, household survey No Hettige (2006) Indonesia, Philippines, and Sri Lanka, No 1993­2001, household level Instituto Cuanto (2005) Peru, 1994­2004, household survey Yes Jacoby (2000) Nepal, 1995­96, household level Yes Jalan and Ravallion (2002) Four Chinese provinces, 1985­90, household, Yes village, and county level Lofgren, Thurlow, and Robinson Zambia, 2001, household level Yes (2004) Minten (1999) Madagascar, 2000­01, commune census data No Pender and others (2004) Uganda, 1999­2000, community level Yes Renkow, Hallstrom, and Karanja Kenya, 1999, household and village level Not applicable (2004) 232 A p p e n d i x A Dependent variable Road impact on dependent variable Transport costs Road improvement: ­ After road rehabilitation projects: Travel time Generally at least 50% shorter than in control villages Access to electricity 17% more households covered than in control villages Increase in nonfarm income in the past 9% more households than in control villages five years Changes due to road (R) and track (T) rehabilitation (medium- to long-term effects): Travel time ­61.8% (R and T) Freight transport costs ­5.7% (trucks) to ­46.4% (minibuses) Passenger transport costs ­8.8% (trucks) to ­40.6% (minibuses) Students registered ­0.4% (R), 14.1% (T) Visits to health centers 45.6% (R), 25.4% (T) Male wages 20.6% (R), 6.5% (T) Female wages 2.4% (R), ­10.6% (T) Children wages ­21.3% (R), ­9.0% (T) Poverty ­4.1% (R), ­5.7% (T) Quantity of agricultural land 15.8% (R), ­39.0% (T); geographically, increase in farmed lands following road rehabilitation occurred mostly in the southern mountainous region and in the forest Land value Elasticity of land value with respect to "time to market center": ­0.26*** Wage rate Elasticity of wage with respect to "time to market center": ­0.048*** Consumption growth Elasticity with respect to of road density: 0.015*** Agricultural output 10% increase in feeder roads leads to 0.1 percentage point increase in agricultural GDP growth rate Poverty 10% increase in roads leads to 3­4% decrease in rural poverty Producer price of rice Increase of $1.20 to $17.00 per ton per hour reduction in access time to paved road. (Mean price $389/ton) Increase in nonfarm activities Elasticity with respect to reduction in distance to tarmac road: 0.089*** Fixed transactions costs in the market Effect of distance to nearest village by truck (road): +*** for maize (continued) 233 AT L o g g e r h e A d s ? Table A.2 (continued) Control for Study Location, year(s), and data level endogeneity? Warr (2005) Lao PDR, 1997­2003, household survey and Yes district level World Bank (2001) Peru, 1994­2000, household survey No Zeller, Diagne, and Mataya (1998) Malawi, 1993­95, household level No Zhang and Fan (2001) India, 1971­94 Yes district level n.s. = Statistically not significant Note: Where only the sign of the effect is reported, it either was not possible to give a meaningful interpretation of the coefficients in the study or no coefficient was reported. 234 A p p e n d i x A Dependent variable Road impact on dependent variable Real per capita expenditure District built road during 1997­2002 dummy: 0.188*. Providing all- weather road access to everyone would reduce by 7% the number of Lao PDR's rural poor (5.6% of its total population) Changes due to road (R) and track (T) rehabilitation (short-term effects): Travel time ­33.3% (R and T) Freight transport costs ­7.9% (trucks), ­13.6% (buses) Passenger transport costs ­14.3% (trucks), ­41.1% (minibuses) Students registered 0.2 (R), 6.9% (T) Student dropouts ­9.0% (R), 24.7% (T) Visits to health centers 4.1% (R), ­2.9% (T) Share of area cropped under new Elasticity with respect to travel costs to agricultural market: technologies (hybrid maize) ­0.276** Agricultural productivity Elasticity with respect to road density: 0.043** 235 AT L o g g e r h e A d s ? Table A.3 Forest Management and Tenure (thousands of hectares unless otherwise indicated) Tropical forest Tropical Permanent forest estate area (millions of closed natural Natural Planted Continent, country hectares) forest production production Protection Africa 208,581 70,461 825 39,271 Cameroon 13.3­23.8 19,985 8,840 17 3,900 Central African Rep. 22.9­29.3 4,826 3,500 3 300 Congo, Dem. Rep. 128.0­135.0 126,236 20,500 55 27,000 Congo, Rep. 20.3­22.1 22,000 18,400 72 2,860 Côte d'Ivoire 7.1­11.7 3,248 3,400 167 734 Gabon 25.8 21,800 10,600 25 2,700 Ghana 2.7­6.3 1,634 1,150 97 353 Liberia 3.5­5.7 4,124 1,310 n.d. 101 Nigeria 9.7­13.5 4,456 2,720 375 1,010 Togo 0.5­1.1 272 41 14 313 Asia 216,791 100,522 39,669 76,900 Cambodia 9.3­11.1 5,500 3,460 17 4,620 Fiji 0.8­0.9 747 0 113 241 India 64.1­76.8 22,500 13,500 32,600 25,600 Indonesia 105.0­120.0 100,382 46,000 2,500 22,500 Malaysia 19.3­19.5 19,148 11,200 183 3,210 Myanmar 34.4 32,700 9,700 710 3,300 Papua New Guinea 30.6 30,150 8,700 80 1,700 Philippines 5.4­7.2 5,288 4,700 274 1,540 Thailand 13.0­14.8 10,127 0 1,870 8,260 Vanuatu 0.9 442 117 2.1 8.37 Vietnam 19.0 12,307 3,145 1,320 5,921 236 A p p e n d i x A Production forest management Tenure Natural Planted Community Public Private License or Management Sustainably Management owned or forest forest concession plan Certified managed plan Certified managed (%) (%) 44,049 10,016 1,480 4,303 488 0 n.d. n.d. n.d. 4,950 1,760 0 500 n.d. 0 M 100 0 2,920 650 0 186 n.d. 0 L n.d. n.d. 15,500 1,080 0 284 40 0 L 100 0 8,440 1,300 0 1,300 45 0 L 100 0 1,870 1,110 0 277 120 0 L 100 0 6,923 2,310 1,480 1,480 10 0 L 100 0 1,035 1,150 0 270 97 0 L 100 0 1,310 0 0 0 0 0 L n.d. n.d. 1,060 650 0 n.d. 175 0 L 100 0 41 5.5 0 5.5 1.2 0 M 27 73 75,045 55,060 4,914 14,397 11,456 184 n.d. n.d. n.d. 3,370 150 0 0 7 0 L 100 0 n.d. n.d. n.d. n.d. 90 0 H n.d. n.d. 13,500 9,720 0 4,800 8,150 0 M 98 2 43,200 18,400 275 2,940 2,500 0.152 L 100 0 6,790 11,200 4,620 4,790 183 183 L 93 7 n.a. 9,700 0 291 0 0 L 100 0 5,600 4,980 19 1,500 n.d. 0 H n.d. <3 n.d. 910 0 76 274 0 H 89 11 n.d. n.d. n.d. n.d. 250 1 L 87 13 n.d. 0 0 0 2.1 0 H 0 0 5,955 n.d. n.d. n.d. n.d. n.d. M 56 18 (continued) 237 AT L o g g e r h e A d s ? Table A.3 (continued) Tropical forest Tropical Permanent forest estate area (millions of closed natural Natural Planted Continent, country hectares) forest production production Protection Latin America and Caribbean 788,008 184,727 5,604 351,249 Bolivia 52.2­59.5 47,999 17,000 60 14,700 Brazil 444.0­544.0 489,515 98,100 3,810 271,000 Colombia 49.6­65.6 51,437 5,500 148 8,860 Ecuador 8.4­11.4 10,854 3,100 164 4,300 Guatemala 2.9­4.3 2,824 1,140 71 1,240 Guyana 16.9 16,916 5,450 12 980 Honduras 5.4 3,811 1,590 48 1,600 Mexico 55.2­64.0 33,120 7,880 100 5,600 Panama 2.9­3.5 3,052 350 56 1,580 Peru 65.2­86.4 64,204 24,600 250 16,300 Suriname 13.6­14.8 14,100 6,890 7 4,430 Trinidad and Tobago 0.2­0.3 250 127 15.4 59.1 R.B. de Venezuela 49.9­55.0 49,926 13,000 863 20,600 Total 1,213,380 355,710 46,098 467,420 n.d = No data. Note: Data in italics are tropical and nontropical forests combined. L, M, and H indicate that community forests constitute a low, medium, or high proportion of the forest (based on authors' interpretations). Public and private proportions of forest do not necessarily add to 100%; the remaining share may include community forests, or areas of undefined tenure. Source: FAO 2005; ITTO 2006; Nguyen 2006. 238 A p p e n d i x A Production forest management Tenure Natural Planted Community Public Private License or Management Sustainably Management owned or forest forest concession plan Certified managed plan Certified managed (%) (%) 34,651 31,174 4,150 6,468 2,371 1,589 n.d. n.d. n.d. 5,470 5,470 2,210 2,210 n.d. 0 M 85 10 n.d. 5,250 1,160 1,360 1,350 1,350 M n.d. n.d. 2,150 n.d. 0 200 80 58 M n.d. n.d. n.d. 65 0 101 65 21.3 M 77 0 540 697 520 672 25 7.57 M 42 53 3,800 3,730 0 520 0 0 M 66 0 1,070 671 37 187 28 0 M 75 25 8,600 8,600 163 163 34 0 H 59 0 86 63 0 0 32 12.2 M 10 90 8,000 5,000 59.5 560 8 0 M 83 15 1,740 73 0 0 7 0 L 0 75 75 0 15 15.4 0 L 75 25 3,120 1,480 0 480 727 140 L 90 n.d. 153,745 96,250 10,544 25,168 14,315 1,773 n.d. n.d. n.d. 239 Appendix B: data and Methods This appendix briefly describes data and methods used for unpub- lished analyses in this report. Global Data and Analyses (Chapter 1 and Figure 2.3) Gridding All mapped data were converted to a pantropical grid of 1 x 1 kilo- meter cells. Tables and figures in chapter 1 and figure 2.3 are based on this grid. data scope included the tropical forest and savanna biomes, excluding Australia, Japan, and the United states. Accessibility to Major Market This index, constructed by Andrew nelson, represents the notional time to travel to the nearest city of 100,000 or more people. it was constructed using standard geographic information system (gis) methods, road information from the digital Chart of the World (dCW), and assumed travel speeds for different road classes. Because the index is based on assumed rather than measured travel times, and because dCW maps are inconsistent and out of date, the imputed times should be considered only rough indexes. This is especially the case for Africa, where travel in forest areas of the democratic republic Congo is probably much slower than represented by the maps and assumptions used for calculations. elsewhere, some areas classified as very remote may in fact be rela- tively accessible due to recently constructed roads or rapidly grow- ing towns. Accessibility measures could not be computed for some areas, mostly in Asia. These include island portions of indonesia and the philippines, some island nations of the southwest pacific, and for Taiwan, China. populations and areas of these islands were included in table 1.3 based on judgments about remoteness but excluded from figures 1.3, 1.4, and 2.3. 240 A p p e n d i x B Land Cover The global Land Cover 2000 (gLC2000) database (eCJrC 2003) was used for information about land use and forest cover throughout the tropics. This database is based on 1-kilometer resolution data from the spoT-4 satellite. dates for the data range from 1 november 1999 to 31 december 2000. The 23 land cover classifications used by the gLC2000 are shown in appendix table B.1, along with this report's 7-class aggregation of these classifications. Table B.1 GLC2000 Land Cover Categories Aggregated land cover Code Type of land cover class 1 Tree cover, broadleaved, evergreen Forest 2 Tree cover, broadleaved, deciduous, closed Forest 3 Tree cover, broadleaved, deciduous, open Forest 4 Tree cover, needle-leaved, evergreen Forest 5 Tree cover, needle-leaved, deciduous Forest 6 Tree cover, mixed leaf type Forest 7 Tree cover, regularly flooded, fresh water Forest 8 Tree cover, regularly flooded, saline water Forest 9 Mosaic of tree cover and other natural vegetation Forest 10 Tree cover, burnt Forest 11 Shrub cover, closed-open, evergreen Bush 12 Shrub cover, closed-open, deciduous Bush 13 Herbaceous cover, closed-open Bush 14 Sparse herbaceous or sparse shrub cover Bush 15 Regularly flooded shrub and/or herbaceous cover Bush 16 Cultivated and managed areas Agriculture 17 Mosaic of cropland, tree cover, and other natural vegetation Mosaic 18 Mosaic of cropland, shrub, and/or grass cover Mosaic 19 Bare areas Bare 20 Water bodies Water/missing 21 Snow and ice Water/missing 22 Artificial surfaces and associated areas Artificial 23 No data Water/missing Source: ECJRC 2003; authors' aggregations. 241 AT L o g g e r h e A d s ? Forest Situation Typology The report defines human-affected rural gridcells as being in "agri- culture" or "mosaic" classes (using the 7-class aggregation). it then measures the distance from these collections of cells to areas that are forest or bush in the aggregated classification. The nearest 6 kilo- meters are called forest edge or savanna edge depending on biome. Forest or bush cells more than 6 kilometers from the nearest human- affected cells are designated as forest core or savanna core. But a special rule is applied to small patches of forest and bush cells--those less than 8 square kilometers--that are completely surrounded by agriculture and mosaic cells. These are designated embedded forests. Mosaic forests consist of embedded forests and mosaic cells. Mosaiclands consist of mosaic forests and agricultural cells. Rural Population Density The report uses population density figures calculated from the grUMp (alpha version) population count grid (Ciesin and others 2004b) and the grUMp area grid (Ciesin and others 2004a). These figures are based on census data reported at a local administrative level--usually the equivalent of a county or municipio, or smaller. Within the administrative unit, grUMp identifies the popula- tion living in cities, towns, and villages of about 2,500 people or more (Ciesin and others 2004c). The rural remainder is assumed to be evenly distributed across the rest of the administrative unit. The assumption of even distribution likely overstates the population density of forested and remote areas of the unit and understates the density of agricultural and mosaic areas. so the forest population densities reported here should be taken with caution--as with all global, spatially explicit population datasets. Forest Cover Change The report team is grateful to the Food and Agriculture organiza- tion (FAo) for sharing data from the Forest resources Assessment remote sensing survey (FrA-rss) (FAo 2001a; b). The remote sens- ing survey examined a stratified random sample of 10 percent of the world's tropical forest, using Landsat scenes as a sampling frame. high-resolution (30-meter) images were used to identify nine types of land cover, sampled at 2-kilometer intervals. Change was detected by direct comparison of earlier and later scenes. 242 A p p e n d i x B in the most forested class we grouped closed canopy, open canopy, and long fallow, followed by a second class consisting of fragmented forest and a third class of agriculture, short fallow, and shrubs (see table 2-4 in FAo 2001a or table 46-1 in FAo 2001b). We defined as degradation any shift from a more to a less densely for- ested cover. degradation rates were computed by dividing the num- ber of cells where degradation was observed from one time period to the next by the number of cells that were degradable--that is, nei- ther in the third class nor covered by water or clouds in the earlier period. The changes reported are based on comparison of remote sensing images from around 1990 and 2000. Because of cloud cover, the actual image dates may vary. We did not adjust for the variation in observation period. in calculating degradation rates, we used a weighting scheme based on the sampling scheme provided by the FAo to accompany the FrA-rss. The weights used were calculated by dividing the land area computed from all Landsat scenes in the subregion and forest cover stratum by the land area of the sampled Landsat scenes from the respective subregions and strata. Biomes WWF (2001) distinguishes 13 biomes. This report's "forests" com- prise three WWF tropical and subtropical biomes: moist broadleaf forest, dry broadleaf forest, and coniferous forest. This report's "savannas" correspond to WWF tropical and subtropical grasslands, savannas, and shrublands. Agricultural Suitability This measure corresponds to plate 46, "suitability for rain-fed crops- maximizing technology mix," from the global Agro-ecological Zones dataset (FAo and iiAsA 2000). This is a gridded dataset with 5 arc- minute resolution. it takes into consideration slope, soil fertility, soil depth, drainage, soil chemical, soil texture, and climate constraints. Threatened Amphibians These data are from the global Amphibian Assessment, which describes the "extent of occurrence" and threat status of almost all known amphibian species (iUCn, Conservation international, and natureserve 2005). The extent of occurrence is a rough depiction of the known geographic range of the species, based on recorded 243 AT L o g g e r h e A d s ? observations, and may include areas of habitat unsuitable for the species. Figure 1.4 shows the proportion of all gridcells, for each distance category, containing the extent of occurrence of at least one endangered or critically endangered amphibian species, using the classification of the World Conservation Union (iUCn) redlist (http://www.redlist.org/info/categories_criteria2001.html). Imminent Extinctions data on imminent extinctions are updated from ricketts and oth- ers (2005) using data from the Alliance for Zero extinction (www .zeroextinction.org, dataset v2.1). Locations of the imminent extinc- tion sites were mapped into the pantropical forest (nonsavanna) gridcells. Figure 1.5 shows, for each distance category, the propor- tion of gridcells containing an imminent extinction site (multiplied by 100,000). National Poverty, Forest, and Deforestation Data (Chapter 3 and Figure 2.1) Brazil The farmgate price imputation is from an unpublished analysis by iMAZon (instituto do homem e Meio Ambiente da Amazônia). rainfall (annual mean) is 1-kilometer resolution data from hijmans and others (2004). deforestation data are from 1:250,000 scale digital maps of incremental clearing based on remote sensing, and cover primary forests (excluding savannas) covering the officially-defined Amazô- nian region of Brazil (inpe 2004). Literacy and population data at the census tract level are from the Brazilian demographic Census 2000 (iBge 2003). Municipal income and education data are from Undp (2004) and derived from the 2000 census. India data on the proportion of forest cover are from india's State of For- est Report 2003 (Forest survey of india 2005). Literacy data are from india's 2001 population census (government of india 2001). Units of observation are at the district level. 244 A p p e n d i x B Indonesia data on forest cover in 2000 are from the Forest Watch indonesia land cover map, derived from the Ministry of Forests forest cover data for 2003. Forest cover for 1990 was generated from the 1:250,000 land cover map of the national Forest inventory 1993. poverty map data for 2000 were constructed by the Central Bureau of statistics (Bps 2005) using the method of elbers, Lan- jouw, and Lanjouw (2003), which imputes consumption to census households based on regressions estimated using a separate house- hold survey. There is a tradeoff between spatial precision and preci- sion of the imputed mean consumption level. The kecamatan-level estimates shown here have relatively high standard errors and thus are useful for illustrating, for example, relationships between forest cover and poverty, rather than for providing poverty counts for a particular kecamatan. Madagascar Forest cover and deforestation data are from satellite images inter- preted and analyzed by Conservation international for 1990 and 2000 (harper n.d.; steininger and others 2004), and from the World Bank poverty map for Madagascar, based on welfare measures computed from the 1993 census of population and housing (Bureau Central du recensement), as well as from a household survey (Mistiaen, razaf- imanantena, and razafindravonona 2002) again using the method of elbers, Lanjouw and Lanjouw (2003). data points were computed at the firaisana level. The size of the bubbles is based on the popu- lation of the firaisana in 1993. Nicaragua extreme rural poverty rates are from Gobierno de Nicaragua (2001) and based on the 1995 census and imputed consumption using a 1998 survey. imputed access time to Managua was computed using gis methods from a late 1990s road map and assumptions about mean travel speed on four classes of roads. rural population density was computed as rural population divided by total municipio area-- implicitly assuming that urban areas occupy a negligible portion of municipio land area. Tabulations excluded island municipios and a few that "fissioned" between 1995 and 2001. 245 AT L o g g e r h e A d s ? Chapter 5 data on Brazilian municipal environmental councils and problem perceptions are from iBge (2002). Chapter 6 The dataset for parks is the 2005 national iUCn point dataset from WdpA Consortium (2005). Figures 6.1 and 6.2 were produced using it. The analysis was limited to tropical forest and tropical savanna biomes in developing countries, and further excluded marine parks, those without information on establishment date, and those for which accessibility could not be computed. 246 references Achard, Frederic, hugh d. eva, philippe Mayaux, hans-Jürgen stibig, and Alan Belward. 2004. "improved estimates of net Carbon emissions from Land Cover Change in the Tropics for the 1990s." Global Biogeochemical Cycles 18. Achard, F., J. p. Malingreau, T. phulpin, g. saint, B. saugier, B. seguin, and V. Vidal-Madjar. 1994. "A Mission for global Monitoring of the Continen- tal Biosphere." VegeTATion international Users Committee secretariat, Joint research Centre, ispra, italy. Agrawal, Arun. 2001. "Common property institutions and sustainable gov- ernance of resources." World Development 29 (10): 1649­72. Agrawal, Arun, and Kent redford. 2006. "poverty, development, and Biodi- versity Conservation: shooting in the dark?" Working paper 26. Wildlife Conservation society, new York. Akella, Anita sundari, and James B. Cannon. 2004. Strengthening the Weak- est Links: Strategies for Improving the Enforcement of Environmental Laws Globally. Washington, d.C.: Center for Conservation and govern- ment [Conservation international]. Alix-garcia, Jennifer, Alain de Janvry, elisabeth sadoulet, Juan Manuel Torres, Josefina Braña, and Maria Zorilla ramos. 2005. "An Assessment of Mexico's payment for environmental services program." Food and Agriculture organization, rome. Alston, Lee J., gary d. Lipecab, and Bernardo Mueller. 2000. "Land reform policies, the sources of Violent Conflict, and implication in deforestation in Brazilian Amazon." Journal of Environmental Economics and Manage- ment 39 (2): 162­88. Andersen, L. e., and e. J. reis. 1997. "deforestation, development and government policy in the Brazilian Amazon: An econometric Analysis." Working paper 513. institudo de pesquisa econômica Aplicada (ipeA), rio de Janeiro. Andersson, Krister. 2003. "What Motivates Municipal governments? Uncov- ering the institutional incentives for Municipal governance of Forest resources in Bolivia." Journal of Environment and Development 12 (1): 5­27. Angelsen, Arild. 1995. "shifting Cultivation and `deforestation': A study from indonesia." World Development 23 (10): 1713­29. ------. 2006. "A stylized Model of incentives to Convert, Maintain, or estab- lish Forests." Background paper for this report. 247 AT L o g g e r h e A d s ? Angelsen, Arild, and david Kaimowitz. 2001. Agricultural Technologies and Tropical Deforestation. Wallingford, U.K.: CABi publishing. Antinori, Camille, and david Barton Bray. 2005. "Community Forest enter- prises as entrepreneurial Firms: economic and institutional perspectives from Mexico." World Development 33 (9): 1529­43. Antona, M. e. Motte. 2002. "property right Transfer in Madagascar's Biodi- versity policies." paper presented at the Bioecon meeting, rome. Araújo, s. Alves de. 2006. "sistema de Licenciamento Ambiental em pro- priedades rurais do estado do Mato grosso." presentation at the Bra- zilian Ministry of the environment workshop "sLApr na Amazônia Legal--estudo de Caso das experiências de Mato grosso, rondônia, et Tocantins," February 8­10, palmas. Arima, eugenio Y., paulo Barreto, and Marky Brito. 2006. Cattle Ranching in the Amazon: Trends and Implications for Environmental Conservation. Belém, Brazil: imazon. Arima, eugenio Y., robert T. Walker, stephen g. perz, and Marcellus Caldas. 2005. "Loggers and Forest Fragmentation: Behavioral Models of road Building in the Amazon Basin." Annals of the Association of American Geographers 95 (3): 525­41. Arnold, J. e. M. 2001. "Forestry, poverty and Aid." occasional paper 33. Center for international Forestry research, Bogor, indonesia. Arnold, J. e. M., gunnar Köhlin, and reidar persson. 2006. "Woodfuels, Livelihoods, and policy interventions: Changing perspectives." World Development 34 (3): 596­611. Asner, gregory p., david e. Knapp, eben n. Broadbent, paulo J. C. oliveira, Michael Keller, and Jose n. silva. 2005. "selective Logging in the Brazil- ian Amazon." Science 310 (5747): 480­82. Aylward, Bruce. 2002. "strategic Framework." report to the World Bank under the program for sustainable Management of rural Areas in the panama Canal Watershed. ------. 2005. "Land-Use, hydrological Function and economic Valuation." in M. Bonell and L. A. Bruijnzeel, eds., Forest, Water and People in the Humid Tropics. Cambridge, U.K.: Cambridge University press. Baillie, J. e. M., C. hilton-Taylor, and s. n. stuart, eds. 2004. 2004 IUCN Red List of Threatened Species: A Global Species Assessment. gland, switzer- land and Cambridge, U.K.: World Conservation Union (iUCn). Baland, Jean-Marie, and Jean-philippe platteau. 1996. Halting Degradation of Natural Resources: Is There a Role for Rural Communities? oxford: Clarendon press. Balk, deborah, F. pozzi, Uwe deichmann, g. Yetman, and A. nelson. 2004. "The distribution of people and the dimension of place: Methodologies to improve global estimation of Urban extents." Columbia University, Center for international earth science information network, palisades, new York. 248 r e F e r e n C e s Bandyopadhyay, s., p. shyamsundar, and A. Baccini. 2006. "Forests, Bio- mass Use and poverty in Malawi." draft background paper for the Malawi poverty Assessment. World Bank, environment department, policy and economic Team, Washington, d.C. Barbier, e. B., and Mark Cox. 2004. "An economic Analysis of shrimp Farm expansion and Mangrove Conversion in Thailand." Land Economics 80 (3): 389­407. Barbosa, g. n. 2006. "A experiência de Mato grosso: uma visão critica do Ministerio publico estadual do Mato grosso." presentation at the Bra- zilian Ministry of the environment workshop "sLApr na Amazônia Legal--estudo de Caso das experiências de Mato grosso, rondônia, et Tocantins," February 8­10, palmas. Barnes, douglas F., Kerry Krutilla, and William F. hyde. 2005. The Urban Household Energy Transition. Washington, d.C.: resources for the Future. Bartholomé, e., and A. s. Belward. 2005. "gLC2000: A new Approach to global Land Cover Mapping from earth observation data." International Journal of Remote Sensing 26 (May 9, 10): 1959­77. Batagoda, B. M. s., Kerry r. Turner, robert Tinch, and Katrina Brown. 2000. "Towards policy relevant ecosystem services and natural Capital Values: rainforest non-Timber products." Cserge Working paper geC 2000-06. University of east Anglia, norwich, U.K. Baulch, Bob, Truong Thi Kim Chuyen, dominique haughton, and Jonathan haughton. 2004. "ethnic Minority development in Vietnam and the potential for Targeting." in paul glewwe, nisha Agrawal, and david dol- lar, eds., Economic Growth, Poverty, and Household Welfare in Vietnam. Washington, d.C.: World Bank. Bennett, Michael T. 2005. "China's sloping Land Conversion program: insti- tutional innovation or Business as Usual?" revised version of a paper presented at the ZeF-CiFor workshop on "payments for environmental services: Methods and design in developing and developed Countries," Titisee, germany, June 15­18, 2005. Binswanger, hans p., shahidur r. Khandker, and Mark r. rosenzweig. 1993. "how infrastructure and Financial institutions Affect Agricultural output and investment in india." Journal of Development Economics 41 (2): 337­66. Boccucci, M., K. d. Muliastra, and g. dore. 2005. "poverty Analysis of indo- nesia's Forest Land." draft report. World Bank, Jakarta, indonesia. Boltz, Frederick, douglas r. Carter, Thomas p. holmes, and rodrigo Jr. pereiara. 2001. "Financial returns under Uncertainty for Conventional and reduced impact Logging in permanent production Forests of the Brazilian Amazon." Ecological Economics (39): 387­98. Bonell, M., and L. A. Bruijnzeel, eds. 2005. Forests, Water and People in the Humid Tropics. Cambridge, U.K.: Cambridge University press. 249 AT L o g g e r h e A d s ? Boscolo, Marco, and Jeffrey r. Vincent. 2000. "promoting Better Logging practices in Tropical Forests: A simulation Analysis of Alternative regu- lations." Land Economics 76 (1): 1­14. Bps (Biro pusat statistik). 2005. Peta Penduduk Miskin Indonesia, 2000. Brandon, Katrina, Larry J. gorenflo, Ana s. L. rodrigues, and robert W. Waller. 2005. "reconciling Biodiversity Conservation, people, protected Areas, and Agricultural suitability in Mexico." World Development 33 (9): 1403­18. Bray, david Barton, Leticia Merino-perez, patricia negreros-Castillo, gerardo segura-Warnholtz, Juan Manuel Torres-rojo, and henricus F. M. Vester. 2003. "Mexico's Community-Managed Forests as a global Model for sus- tainable Landscapes." Conservation Biology 17 (3): 672­77. Brito, Brenda, paulo Barreto, and John rothman. 2005. "Brazil's new envi- ronmental Crimes Law: An Analysis of its effectiveness in protecting the Amazonian Forests." Belém, imazon. Brooks, Thomas M., stuart L. pimm, and Joseph o. oyugi. 1999. "Time Lag between deforestation and Bird extinction in Tropical Forest Fragments." Conservation Biology 13 (5): 1140­50. Brown, david. 2004. "review of independent Forest Monitoring." odi, London. Bruijnzeel, L. A. 2004. "hydrological Functions of Tropical Forests: not see- ing the soil for the Trees?" Agriculture, Ecosystems & Environment 104: 185­228. Bruijnzeel, L. A., M. Bonell, d. A. gilmour, and d. Lamb. 2005. "Forest, Water and people in the humid Tropics: An emerging View." in M. Bonell and L. A. Bruijnzeel, eds., Forest, Water and People in the Humid Tropics. Cambridge, U.K.: Cambridge University press. Bruinsma, J. 2003. World Agriculture: Towards 2015/2030--An FAO Perspec- tive. London: earthscan. Bruner, Aaron g., raymond e. gullison, richard e. rice, and gustavo A. B. da Fonseca. 2001. "effectiveness of parks in protecting Tropical Biodiver- "effectiv pr er- sity." Science 291: 125­28. Bureau Central du recensement. 1993. Recensement général de la popula- tion et de l'habitat. Antananarivo, Madagascar: direction de La démo- graphie et des statistiques socials. Burgess, neil d., Jennifer d'Amico hales, Taylor h. ricketts, and eric diner- stein. 2006. "Factoring species, non-species Values and Threats into Bio- diversity prioritisation across the ecoregions of Africa and its islands." Biological Conservation 127 (4): 383­401. Buys, p., U. deichmann, and d. Wheeler. 2006. "road network Upgrad- ing and overland Trade expansion in sub-saharan Africa." World Bank, Washington, d.C. Calder, ian r. 2005. The Blue Revolution: Land Use and Integrated Water Resources Management. London: earthscan. 250 r e F e r e n C e s Cassman, Kenneth g., and stanley Wood. 2005. "Cultivated systems." in rashid hassan, robert scholes, and neville Ash, eds., Ecosystems and Human Well-Being: Current State and Trends. Washington, d.C.: island press. Castro, newton de. 2002. "Custos de transporte e produção agrígola no Bra- sil, 1970­1996." Agricultura em São Paulo 49 (2): 87­109. Ceballos, gerardo, paul r. ehrlich, Jorge soberón, irma salazar, and John p. Fay. 2005. "global Mammal Conservation: What Must We Manage?" Science 309: 603­07. Cernea, Michael M., and Kai schmidt-soltau. 2003. "The end of Forcible displacement? Conservation Must not impoverish people." Policy Mat- ters 12: 42­51. Chazdon, robin L. 2003. "Tropical Forest recovery: Legacies of human impact and natural disturbances." Perspectives in Plant Ecology, Evolu- tion and Systematics 6 (1.2): 51­71. Chomitz, Kenneth M. 2002. "Baseline, Leakage, and Measurement issues: how do Forestry and energy projects Compare?" Climate Policy 2 (1): 35­49. ------. 2004. "nicaragua economic geography: A snapshot." in Nicaragua: Drivers of Sustainable Rural Growth and Poverty Reduction in Central America, Vol. II. Washington, d.C.: World Bank. Chomitz, K. M. and d. gray. 1996. "roads, Land Use, and deforestation: A spatial Model Applied to Belize." The World Bank Economic Review 10: 487­512. Chomitz, K. M., and K. Kumari. 1998. "The domestic Benefits of Tropical Forest preservation: A Critical review emphasizing hydrological Func- tions." The World Bank Research Observer 13 (1): 13­35. Chomitz, Kenneth M., and Timothy s. Thomas. 2003. "determinants of Land Use in Amazônia: A Fine-scale spatial Analysis." American Journal of Agricultural Economics 85: 1016­28. Chomitz, Kenneth M., and sheila Wertz-Kanounnikoff. 2005. "Measuring the initial impacts on deforestation of Mato grosso's program for envi- ronmental Control." policy research Working paper 3762. World Bank, Washington, d.C. Chomitz, K. M, e. Brenes, and L. Constantino. 1999. "Financing environ- mental services: The Costa rican experience and its implications." Sci- ence of the Total Environment 240: 157­69. Chomitz, K. M., Timothy s. Thomas, and Antônio salazar p. Brandao. 2005. "The economic and environmental impact of Trade in Forest reserve obligations: A simulation Analysis of options for dealing with habitat heterogeneity." Revista Economia e Sociologia Rural 43 (4): 657­82. Chomitz, K. M, daniel da Mata, Alexandre Ywata de Carvalho, and João Carlos Magalhaes. 2005a. "spatial dynamics of Brazilian Labor Markets." "spatial dynamics policy research Working paper 3752. World Bank, Washington, d.C. 251 AT L o g g e r h e A d s ? Chomitz, K. M., K. Alger, T. s. Thomas, h. orlando, and p. Vila nova. 2005b. "opportunity Costs of Conservation in a Biodiversity hotspot: The Case of southern Bahia." Environment and Development Economics 10 (3): 293­312. Chomitz, Kenneth M., gunawan setiadi, Azrul Azwar, nusye ismail, and Widiyarti. 1998. "What do doctors Want? Two empirical Analyses of indonesian physicians' preferences regarding service in rural and remote Areas." policy Working paper series 1888. World Bank, Wash- ington, d.C. Chomitz, Kenneth M., g. A. B. Fonseca, d. M. stoms, M. honzák, e. C. Landau, T. s. Thomas, d. Thomas, and A. davis. 2006. "Viable reserve networks Arise from individual Landholder responses to Conservation incentives." Unpublished paper. Ciesin (Center For international earth science information network), iFpri (international Food policy research institute), World Bank, and CiAT (Centro internacional de Agricultura Tropical). 2004a. "global rural-Urban Mapping project (grUMp), Alpha Version: Land Area grids." Columbia University, socioeconomic data and Applications Cen- ter (sedAC), palisades, n.Y. [http://sedac.ciesin.columbia.edu/gpw]. Accessed december 22, 2005. ------. 2004b. "global rural-Urban Mapping project (grUMp), Alpha Ver- sion: population grids." Columbia University, socioeconomic data and Applications Center (sedAC), palisades, n.Y. [http://sedac.ciesin .columbia.edu/gpw]. Accessed February 7, 2006. ------. 2004c. "global rural-Urban Mapping project (grUMp), Alpha Version: Urban extents." Columbia University, socioeconomic data and Applications Center (sedAC), palisades, n.Y. [http://sedac.ciesin .columbia.edu/gpw]. Accessed december 22, 2005. CiFor (Center for international Forestry research) and FAo (Food and Agri- culture organization). 2005. "Forests and Floods: drowning in Fiction or Thriving on Facts?" rAp publication 2005/03. CiFor and FAo regional office for Asia and the pacific, Bogor, indonesia. Cochrane, Mark A., Ane Alencar, Mark d. schulze, Carlos M. souza Jr., daniel C. nepstad, paul Lefebvre, and eric A. davidson. 1999. "positive Feedbacks in the Fire dynamic of Closed Canopy Tropical Forests." Sci- ence 284: 1832­35. Colchester, Marcus. 2006. "Justice in the Forest: rural Livelihoods and For- est Law enforcement." Center for international Forestry research, Bogor, indonesia. Colfer, Carol J. pierce, and Ana doris Capistrano. 2005. The Politics of Decen- tralization: Forest, Power and People. London: earthscan. Contreras-hermosilla, Arnoldo, and elisa peter. 2006. "Best practices for improving Law Compliance in the Forestry sector." FAo and iTTo, rome. 252 r e F e r e n C e s Contreras-hermosilla, Arnoldo, and Chip Fay. 2005. "strengthening For- est Management in indonesia through Land Tenure reform: issues and Framework for Action." Forest Trends, Washington, d.C. Costa, Marcos heil, Aurelie Botta, and Jeffrey A. Cardille. 2003. "effects of Large-scale Changes in Land Cover on the discharge of the Tocantins river, southeastern Amazonia." Journal of Hydrology 283 (1­4): 206­17. Cowles, paul d., soava rakotoarisoa, haingolalao rasolonirinamanana, and Vololona rasoaromanana. 2001. "Facilitation, participation, and Learning in an ecoregion-Based planning process: The Case of AgerAs in Toliara, Madagascar." in Louise e. Buck, Charles g. geisler, John schelhas, and eva Wollenberg, eds., Biological Diversity: Balancing Interests through Adaptive Collaborative Management. Boca raton, Fla.: Crs press. Cowling, r. M., r. L. pressey, M. rouget, and A. T. Lombard. 2003. "A Conservation plan for a global Biodiversity hotspot--The Cape Floristic region, south Africa." Biological Conservation 112 (1­2): 191­216. Coxhead, ian, and Bayou demeke. 2004. "panel data evidence on Upland Agricultural Land Use in the philippines: Can economic policy reforms reduce environmental damages?" American Journal of Agricultural Eco- nomics 86 (5): 1354­60. Cropper, Maureen, Charles griffiths, and Muthukumara Mani. 1999. "roads, population pressures, and deforestation in Thailand, 1976­1989." Land Economics 75 (1): 58­73. Cropper, Maureen, Jyotsna puri, and Charles griffiths. 2001. "predicting the Location of deforestation: The role of roads and protected Areas in north Thailand." Land Economics 77 (2): 172­86. Cruz, rex Victor o., and Maricel A. Tapia. 2006. "A review of the Multi-sec- toral Forest protection Committees in the philippines." overseas devel- opment institute, London. Curran, L. M., s. n. Trigg, A. K. Mcdonald, d. Astiani, Y. M. hardiono, p. siregar, i. Caniago, and e. Kasischke. 2004. "Lowland Forest Loss in pro- tected Areas of indonesian Borneo." Science 303 (5660): 1000­03. davies, p., and p. Abelson. 1996. "Value soils in the Tropical Lowlands of eastern Bolivia." in p. Abelson, ed., Project Appraisal and Valuation of the Environment, General Principles and Six Case Studies in Developing Countries. London: Macmillan. davis, M. 2005. "Forests and Conflict in Cambodia." International Forestry Review 7 (2): 161­64. defries, ruth s., richard A. houghton, Matthew C. hansen, Christopher B. Field, david skole, and John Townshend. 2002. "Carbon emissions from Tropical deforestation and regrowth Based on satellite observations for the 1980s and 1990s." Proceedings of the National Academy of Sciences 99 (22): 14256­61. 253 AT L o g g e r h e A d s ? deininger, Klaus. 2003. Land Policies for Growth and Poverty Reduction. A policy research report. new York: oxford University press for the World Bank. deininger, Klaus, and Bart Minten. 1999. "poverty, politics, and deforesta- tion: The Case of Mexico." Economic Development and Cultural Change 47 (2): 313­44. ------. 2002. "determinants of deforestation and the economics of protec- tion: An Application to Mexico." American Journal of Agricultural Eco- nomics 84 (4): 943­60. de Merode, emmanuel, Katherine homewood, and guy Cowlishaw. 2004. "The Value of Bushmeat and other Wild Foods to rural households Liv- ing in extreme poverty in democratic republic of Congo." Biological Conservation 118 (5): 573­81. dennis, rona A., Judith Mayer, grahame Applegate, Unna Chokkalingam, Carol J. pierce Colfer, iwan Kurniawan, henry Lachowski, paul Maus, rizki pandu permana, Yayat ruchiat, Fred stolle, suyanto, and Thomas p. Tomich. 2005. "Fire, people and pixels: Linking social science and remote sensing to Underlying Causes and impacts of Fires in indone- sia." Human Ecology 33 (4): 465­504. deschamps, Vince, and paul hartman. 2005. "Trends in Forest ownership, Forest resources Tenure and institutional Arrangements: Are They Con- tributing to Better Forest Management and poverty reduction? Case studies from indonesia." paper prepared for a Food and Agriculture organization regional workshop, october 17­21, Bangkok, Thailand. de Vreyer, philippe, Javier herrera, and sandrine Mesple-somps. 2002. "Consumption growth and spatial poverty Traps: An Analysis of the effects of social services and Community infrastructures on Living stan- dards in rural peru." document de travail 2002/17. diAL, paris. dewi, sonya, Brian Belcher, and Atie puntodewo. 2005. "Village economic opportunity, Forest dependence, and rural Livelihoods in east Kaliman- tan, indonesia." World Development 33 (9): 1419­34. dilley, Maxx, robert s. Chen, Arthur L. Lerner-Lam, Margaret Arnold, Jona- than Agwe, piet Buys, oddvar Kjekstad, Bradfield Lyon, and gregory Yet- man. 2005. "natural disaster hotspots: A global risk Analysis." World Bank, Washington, d.C. dorosh, paul, and Bart Minten. 2005. "rice price stabilization in Madagas- car: price and Welfare implications of Variable Tariffs." Working paper. U.s. Agency for international development, sAgA, Washington, d.C. douglas, ellen, stanley Wood, Kate sebastian, Charles J. Vorosmarty, Ken- neth M. Chomitz, and Thomas p. Tomich. Forthcoming. "policy implica- tions of a pan-Tropic Assessment of the simultaneous hydrological and Biodiversity impacts of deforestation." Water Resources Management. douglas, ellen M., Kate sebastian, K. M Chomitz, Charles J. Vorosmarty, and stanley Wood. 2005. "The role of Tropical Forests in supporting 254 r e F e r e n C e s Biodiversity and hydrological integrity." policy research Working paper 3635. World Bank, Washington, d.C. dudley, nigel, and sue stolton. 2003. Running Pure. Washington, d.C.: World Bank and World Wildlife Fund. dudley, nigel, Kalemani Jo Mulongoy, sheldon Cohen, sue stolton, Vic- tor Charles Barber, and sarat Babu gidda. 2005. "Towards effective pro- tected Area systems. An Action guide to implement the Convention on Biological diversity programme of Work on protected Areas." Technical series 18. secretariat of the Convention on Biological diversity, Mon- treal, Canada. eCJrC (european Commission Joint research Centre). 2003. "global Land Cover 2000 database." [http://www.gvm.jrc.it/glc2000]. economist intelligence Unit. 2005. "Country profile 2005: Brazil." London. environmental Law institute. 2003. "Legal Tools and incentives for private Lands Conservation in Latin America: Building Models for success." Washington, d.C. escobal, Javier, and Carmen ponce. 2002. "The Benefit of rural roads: enhancing income opportunities for the rural poor." grAde, Lima, peru. esri (environmental systems research institute). 1993. "digital Chart of the World: digitized data, Based on defense Mapping Agency Maps at 1:1,000,000." [http://www.maproom.psu.edu/dcw]. etter, Andres, Clive Mcalpine, Kerrie Wilson, stuart phinn, and hugh pos- singham. 2006. "regional patterns of Agricultural Land Use and defor- estation in Colombia." Agriculture, Ecosystems & Environment 114 (2­4): 369­86. Fa, John e., dominic Currie, and Jessica Meeuwig. 2003. "Bushmeat and Food security in the Congo Basin: Linkages between Wildlife and peo- ple's Future." Environmental Conservation 30 (1): 71­78. Faith, d. p., p. A. Walker, and C. r Margules. 2001. "some Future prospects for systematic Biodiversity planning in papua new guinea--and for Bio- diversity planning in general." Pacific Conservation Biology 6: 325­43. Faith, d. p., g. Carter, g. Cassis, s. Ferrier, and L. Wilkie. 2003. "Comple- mentarity, Biodiversity Viability Analysis, and policy-Based Algorithms for Conservation." Environmental Science and Policy 6: 311­28. Fan, shenngen, and Connie Chan-Kang. 2004. "returns to investment in Less-Favored Areas in developing Countries: A synthesis of evidence and implications for Africa." Food Policy 29 (4): 431­44. Fan, shenggen, and xiaobo Zhang. 2004. "infrastructure and regional eco- nomic development in rural China." China Economic Review 15 (2): 203­14. Fan, shenngen, peter hazell, and T. haque. 2000. "Targeting public invest- ments by Agro-ecological Zone to Achieve growth and poverty Allevia- tion goals in rural india." Food Policy 25 (4): 411­28. 255 AT L o g g e r h e A d s ? Fan, shenggen, david nyange, and neetha rao. 2005. "public investment and poverty reduction in Tanzania: evidence from household survey data." discussion paper 18. international Food policy research institute, development strategy and government division, Washington, d.C. Fan, shenggen, xiaobo Zhang, and Linxiu Zhang. 2004. "reform, invest- ment, and poverty in rural China." Economic Development and Cultural Change 52 (2): 395­421. FAo (Food and Agriculture organization). 2001a "FrA 2000: pan-Tropical survey of Forest Cover Changes 1980­2000." FrA Working paper 49. rome. ------. 2001b. "global Forest resources Assessment 2000: Main report." Forestry paper 140. rome. ------. 2005. "global Forest resources Assessment 2005: progress towards sustainable Forest Management." Forestry paper 147. rome. FAo (Food and Agriculture organization) and iiAsA (international insti- tute For Applied systems Analysis). 2000. "The global AeZ Cd-roM." [http://www.iiasa.ac.at/research/luc/gaez/index.htm]. Accessed octo- ber 13, 2005. Fay, Chip, and g. Michon. 2005. "redressing Forestry hegemony: When a Forestry regulatory Framework is Best replaced by an Agrarian one." Agroforestry Systems 15: 193­209. Fearnside, philip M. 1995. "global Warming response options in Brazil's Forest sector: Comparison of project-level Costs and Benefits." Biomass and Bioenergy 8 (5): 309­22. ------. 2003. "deforestation Control in Mato grosso: A new Model for slow- ing the Loss of Brazil's Amazon Forest." Ambio 35 (5): 343­45. Feddema, Johannes J., Keith W. oleson, gordon B. Bonan, Linda o. Mearns, Lawrence e. Buja, gerald A. Meehl, and Warren M. Washington. 2005. "The importance of Land-Cover Change in simulating Future Climates." Science 310 (5754): 1674­78. Fernandez, Tania, and Carlos Munoz. 2006. "Correlations between poverty and deforestation in Mexico." ine-dgipeA draft report eA-0602. Mexico, ine-dgipeA eA d. F. Ferraro, p. J. 2002. "The Local Costs of establishing protected Areas in Low- establishing pr income nations: ranomafana national park, Madagascar." Ecological Economics 43: 261­75. Ferraro, paul J., and Agnes Kiss. 2002. "direct payments to Conserve Biodi- versity." Science 298: 1718­19. Finan, Frederico, elisabeth sadoulet, and Alain de Janvry. 2005. "Measur- ing the poverty reduction potential of Land in rural Mexico." Journal of Development Economics 77: 27­51. Fisher, r. J., stewart Maginnis, W. J. Jackson, edmund Barrow, and sally Jeanrenaud. 2005. "poverty and Conservation: Landscape, people and power." Landscapes and Livelihoods series 2. World Conservation Union (iUCn), gland, switzerland. 256 r e F e r e n C e s Fnp Consultoría & Agroinformativos. n.d. "Valor da Terra." Foley, Jonathan A., Michael T. Coe, Marten scheffer, and guiling Wang. 2003. "regime shifts in the sahara and sahel: interactions between ecological and Climatic systems in northern Africa." Ecosystems 6 (6): 524­32. Forest survey of india. 2005. State of Forest Report 2003. dehradun, india: Ministry of environment and Forests. Forest Trends. 2006. "Logging, Legality and Livelihoods in png: synthe- sis of official Assessments of the Large-scale Logging industry Forest Trends." Washington, d.C. Forman, richard T. T. 1995. Land Mosaics: The Ecology of Landscapes and Regions. Cambridge, U.K.: Cambridge University press. Foster, Andrew d., and Mark r. rosenzweig. 2003. "economic growth and the rise of Forests." Quarterly Journal of Economics 118 (2): 601­37. Fundação estadual do Meio Ambiente. 2001. "environmental Control sys- tem on rural properties." governo do estado de Mato grosso, Cuiabá. go estado gr ------. 2002. "relatório dos resultados alcançados na implementação do sistema de controle ambiental de propriedades rurais no estado de Mato grosso." governo do estado de Mato grosso, Cuiabá. go estado gr Fundacão getulio Vargas. data series 22716 and 22759. [http://fgvdados .fgv.br]. Accessed June 12, 2005. FWi (Forest Watch indonesia) and gFW (global Forest Watch). 2002. "The state of the Forest: indonesia." Bogor, indonesia. garrity, d. p., M. soekardi, M. noordwijk, r. Cruz, p. s. pathak, h. p. M. gunasena, n. so, g. huijun, and n. M. Majid. 1996. "The Imperata grasslands of Tropical Asia: Area, distribution, and Typology." Agrofor- estry Systems 36 (1­3): 3­29. gautam, A. p., g. p. shivakoti, and e. L. Webb. 2004. "Forest Cover Change, physiography, Local economy, and institutions in a Mountain Watershed in nepal." Environmental Management 33 (1): 48­61. gbetnkom, daniel. 2005. "deforestation in Cameroon: immediate Causes and Consequences." Environment and Development Economics 10: 557­72. geF (global environment Facility). 2006. "The role of Local Benefits in global environmental programs." evaluation report 30. geF evaluation office, Washington, d.C. geisler, Charles, and ragendra de sousa. 2001. "From refuge to refu- gee: The African Case." Public Administration and Development 21: 159­70. geist, helmut J., and eric F. Lambin. 2001. "What drives Tropical defores- tation? A Meta-analysis of proximate and Underlying Causes of defores- tation Based on subnational Case study evidence." LUCC report series 4. ihdp and igBp, Louvain-la-neuve, Belgium. geoghegan, Jacqueline, sergio Cortina Villar, peter Klepeis, pedro Macario Mendoza, Yelena ogneva-himmelberger, rinku roy Chowdhury, i. i. 257 AT L o g g e r h e A d s ? Turner, and Colin Vance. 2001. "Modeling Tropical deforestation in the southern Yucatan peninsular region: Comparing survey and satellite data." Agriculture, Ecosystems & Environment 85 (1­3): 25­46. ghimire, Krishna B., and Michel p. pimbert. 1997. Social Change & Conser- vation. London: earthscan. gibson, Clark C., John T. Williams, and elinor ostrom. 2005. "Local enforce- ment and Better Forests." World Development 33 (2): 273­84. gibson, John, and scott rozelle. 2003. "poverty and Access to roads in papua new guinea." Economic Development and Cultural Change 52 (1): 159­85. gilmour, don, Y. B. Malla, and Mike nurse. 2004. "Linkages between Com- munity Forestry and poverty." reCoFTC, Bangkok, Thailand. gitz, Vincent, Jean-Charles hourcade, and philippe Ciais. 2006. "The Tim- ing of the Biological Carbon sequestration and Carbon Abatement in the energy sector under optimal strategies against Climate risks." The Energy Journal 27. global Forest Watch. 2005. "interactive Forestry Atlas of Cameroon (Version 1.0): An overview." Wri­MineF, Washington, d.C. glomsrod, solveig, Maria dolores Monge, and haakon Vennemo. 1998. "structural Adjustment and deforestation in nicaragua." Environment and Development Economics 4: 19­43. gobierno de nicaragua. 2001. Mapa de pobreza extrema de nicaragua, Censo 1995­eMnV 1998. gockowski, James, Mathurin Tchatat, Jean-paul dondjang, gisele hietet, and Terese Fouda. 2006. "The Value of Biodiversity in the Beti Cocoa Agroforests of southern Cameroon." Journal of Sustainable Forestry. gorenflo, L., C. Corson, K. M Chomitz, g. harper, M. honzák, and B. ozler. 2006. "poverty, population, and deforestation in Madagascar." Conserva- tion international, Washington, d.C. government of india. 2001. Census of India 2001. Literacy rates by district available at http://educationforallinindia.com/page157.html. grimes, Alicia, sally Loomis, paul Jahnige, Margo Burnham, Karen onthank, rocio Alercón, Walter palacios Cuenca, Carlos Cerón Martinez, david neill, Michael Balick, Brad Bennet, and robert Mendelsohn. 1994. "Valu- ing the rain Forest: The economic Value of nontimber Forest products in ecuador." Ambio 23 (7): 405­10. guimaraes, Andrè Loubet, and Christopher Uhl. 1997. "rural Transport in eastern Amazonia: Limitations, options, and opportunities." Journal of Rural Studies 13 (4): 429­40. hall, gillette, and h. A. patrinos. 2005. Indigenous People, Poverty and Human Development in Latin America. United Kingdom: palgrave Macmillan. hamilton, L. s., and p. n. King. 1983. "Tropical Forested Watersheds: hydrologic and soils response to Major Uses and Conversions." West- view press, Boulder, Colorado. 258 r e F e r e n C e s hansen, M., and r. deFries. 2004. "detecting Long Term Forest Change Using Continuous Fields of Tree Cover Maps from 8km AVhrr data for the Years 1982­1999." Ecosystems 7: 695­716. harper, g. n.d. "Madagascar: Forest Cover and deforestation, 1990­2000." Conservation international, Center for Applied Biodiversity science, Washington, d.C. harvey, C. A., C. Villanueva, J. Villacís, M. Chacón, d. Muñoz, M. López, M. ibrahim, r. gómez, r. Taylor, J. Martinez, A. navas, J. saenz, d. sán- chez, A. Medina, s. Vilchez, B. hernández, A. perez, F. ruiz, F. López, i. Lang, and F. L. sinclair. 2005. "Contribution of Live Fences to the eco- logical integrity of Agricultural Landscapes." Agriculture, Ecosystems & Environment 111: 200­30. hettige, hemamala. 2006. "When do rural roads Benefit the poor and how?" Asian development Bank, Manila, philippines. hijmans, r. J., s. e. Cameron, J. L. parra, p. g. Jones, and A. Jarvis. 2004. "The WorldClim interpolated global Terrestrial Climate surfaces, Version 1.3." [http://biogeo.berkeley.edu/]. hoare, Alison e. 2006. "divided Forests: Towards Fairer Zoning of Forest Lands." rainforest Foundation, London, U.K. holden, stein. 2001. "A Century of Technological Change and deforestation in the Miombo Woodlands of northern Zambia." in Arild Angelsen and david Kaimowitz, eds., Agricultural Technologies and Tropical Deforesta- tion. Wallingford, U.K.: CABi publishing. house, J., V. Brovkin, r. Betts, r. Costanza, M. Assunçao silva dias, e. hol- land, C. Le Quéré, n. Kim phat, U. riebesell, M. scholes, A. Arneth, d. Barratt, K. Cassman, T. Christensen, s. Cornell, J. Foley, L. ganzeveld, T. hickler, s. houweling, M. scholze, F. Joos, K. Kohfeld, M. Manizza, d. ojima, i. C. prentice, C. schaaf, B. smith, i. Tegen, K. Thonicke, and n. Warwick. 2006. "Climate and Air Quality." in Millennium Ecosystem Assessment 2005--Current State and Trends: Findings of the Condition and Trends Working Group. Ecosystems and Human Well-being. Washing- ton, d.C.: island press. howard, Andrew F., and Juvenal Valerio. 1996. "Financial returns from sus- tainable Forest Management and selected Agricultural Land-use options in Costa rica." Forest Ecology and Management 81: 35­49. hyde, William F. Forthcoming. The Global Economics of Forestry. ------. 2005. "institutions, sustainability, and natural resources: institu- tions for sustainable Forest Management." in s. Kant and r. Berry, eds., The Limitations to SFM: An Economic Perspective. dordrecht, nether- lands: springer. hyde, William F., M. dalmacio, e. guiang, and B. harker. 1997. "Forest Charges and Trusts: shared Benefits with a Clear definition of responsi- bilities." Journal of Philippine Development xxiV (2): 56­223. iBge (instituto Brasileiro de geografia e estadística). 1998. "Censo Agrope- cuário 1995­1996--no. 24 Mato grosso." rio de Janeiro. 259 AT L o g g e r h e A d s ? ------. 2003. "Censo demográfico 2000--agregado por setores censitários dos resultados do universo, 2nd edition." rio de Janeiro. ------. 2006. "produção agrícola municipal." [http://www.sidra.ibge.gov .br]. iBge (instituto Brasileiro de geografia e estatistica) 2005 "perfil dos muni- cipios Brasileiros--meio ambiente." rio de Janeiro. inpe 2004. "projeto prodes." [http://www.obt.inpe.br/prodes/]. Accessed January 2005. ------. 2006. "Mean rate gross deforestation (km2/year) from 1978 to 2000, projeto prodes." [http://www.obt.inpe.br/prodes/prodes_1988_ 2003.htm]. insan hitawasana sejahtera. 2003. "indonesia Forest and Media (inForM) Campaign Monitoring and evaluation: results of the Baseline survey" Jakarta. instituto Cuanto. 2005. "evaluacion economica, social, ambiental e institu- cional del provias rural fase i." Lima, peru. instituto socioambiental. 2005. "Mato grosso, Amazônia (i) legal--desma- tamentos de florestas em propriedades rurais integradas ao sistema de Licenciamento Ambiental rural entre 2001 e 2004." Brasilia. iTTC (international Tropical Timber Council). 2005. "Achieving the iTTo objective 2000 and sustainable Forest Management in Mexico." Yoko- hama, Japan. iTTo (international Tropical Timber organization). 2006. "status of Tropical Forest Management 2005." Technical series 24. Yokohama, Japan. iUCn (World Conservation Union). 1994. "guidelines for protected Area Management Categories." gland, switzerland. [http://www.iucn.org/ themes/wcpa/pubs/pdfs/iucncategories.pdf]. iUCn (World Conservation Union), Conservation international, and natureserve. 2005. "global Amphibian Assessment." [http://www .globalamphibians.org]. Jacoby, h. g. 2000. "Access to Markets and the Benefits of rural roads." The Economic Journal 110: 713­37. Jalan, Jyotsna, and Martin ravallion. 2002. "geographic poverty Traps? A Micro Model of Consumption growth in rural China." Journal of Applied Econometrics 17 (4): 329­46. Jensen, henning Tarp, sherman robinson, and Finn Tarp. 2004. "general equilibrium Measures of Agricultural policy Bias in Fifteen developing Countries." discussion paper 04-25. University of Copenhagen, institute of economics, denmark. Joshi, L., g. Wibawa, g. Vincent, d. Boutin, r. Akiefnawati, g. Manurung, M. Van noordwijk, and s. Williams. 2002. "Jungle rubber: A Tradi- tional Agroforestry system Under pressure." World Agroforestry Centre (iCrAF), Bogor, indonesia. 260 r e F e r e n C e s Kaimowitz, david, and Arild Angelsen. 1998. "economic Models of Tropi- cal deforestation: A review." Center for international Forestry research, Bogor, indonesia. Kaimowitz, david, and Joyotee smith. 2001. "soybean Technology and the Loss of natural Vegetation in Brazil and Bolivia." in Arild Angelsen and david Kaimowitz, eds., Agricultural Technologies and Tropical Deforesta- tion. Wallingford, U.K.: CABi publishing. Kaimowitz, david, Benoit Mertens, sven Wunder, and pablo pacheco. 2004. "hamburger Connection Fuels Amazon destruction." Center for interna- tional Forestry research, Bogor, indonesia. Kazianga, harounan, and William A. Masters. 2005. "property rights, pro- duction Technology and deforestation: Cocoa in Cameroon." [http:// www.columbia.edu/~hk2252/KaziangaMasters.pdf]. Khan, Asmeen, paul Jepson, Michael Wells, scott guggenheim, and Wahjudi Wardojo. 1999. "investing in Biodiversity: A review of indonesia's inte- grated Conservation development projects." World Bank, Washington, d.C. Khanna, Madhu, Wanhong Yang, richard Farnsworth, and hayri onal. 2003. "Cost-effective Targeting of Land retirement to improve Water Quality with endogenous sediment deposition Coefficients." American Journal of Agricultural Economics 85 (3): 538­53. Kiersch, B., and s. Tognetti. 2002. "Land-water Linkages in rural Water- sheds." Land Use and Water Resources Research 2. Kihiyo, Vincent B. M. s. 1996. "economic evaluation of rural Woodlots in a developing Country: Tanzania." Journal of Environmental Management 46 (3): 271­79. Kirby, Kathryn r., William F. Laurance, Ana K. Albernaz, gotz schroth, philip M. Fearnside, scott Bergen, eduardo M. Venticinque, and Carlos da Costa. 2006. "The Future of deforestation in the Brazilian Amazon." Futures 38 (4): 432­53. Kishor, M. nalin, and Luis F. Constantino. 1993. "Forest Management and Competing Land Uses: An economic Analysis for Costa rica." dissemi- nation note 7. World Bank, LATen, Washington, d.C. Kloss, dirk. 2006. "Forest Conservation through protected Area establish- ment and Maintenance." Background paper. Köhlin, gunnar, and gregory s. Amacher. 2005. "Welfare implications of Community Forest plantations in developing Countries: The orissa social project." American Journal of Agricultural Economics 87 (4): 855­69. Lamb, david, peter d. erskine, and John A. parrotta. 2005a. "restoration of degraded Tropical Forest Landscapes." Science 310: 1628­32. ------. 2005b. "restoration of degraded Tropical Forest Landscapes-- supporting Material." [http://www.sciencemag.org/cgi/content/ full/310/5754/1628/dC1]. 261 AT L o g g e r h e A d s ? Larson, Anne M. 2004. "democratic decentralisation in the Forestry sector: Lessons Learned from Africa, Asia and Latin America." Center for inter- national Forestry research, Bogor, indonesia. Laurance, William F., Barbara M. Croes, Landry Tchignoumba, sally A. Lahm, Alfonso Alonso, Michelle e. Lee, patrick Campbell, and Claude ondzeano. 2006. "impact of roads and hunting on Central African rain- forest Mammals." Conservation Biology 20 (4): 1251­61. Laurance, W. F., T. e. Lovejoy, h. L. Vasconcelos, e. M. Bruna, r. K. did- ham, p. C. stouffer, C. gascon, r. o. Bierregaard, s. g. Laurance, and e. sampaio. 2002. "ecosystem decay of Amazonian Forest Fragments: A 22-year investigation." Conservation Biology 16: 605­18. Lawton, r. o., U. s. nair, r. A. pielke sr., and r. M. Welch. 2001. "Climatic impact of Tropical Lowland deforestation on nearby Montane Cloud For- ests." Science 294 (5542): 584­87. Ledec, george, and paula J. posas. 2006. "Biodiversity Conservation in road projects: Lessons from World Bank experience in Latin America." Trans- portation Research Record 1 (1819): 198­202. Lescuyer, guillaume, Alexandre emerit, edouard essiane Mendoula, and Joseph Junior seh. 2001. "Community involvement in Forest Manage- ment: A Full-scale experiment in the south Cameroon Forest." rdFn paper 25c. odi, London. Lima, A., C. T. irigaray, J. C. Figueira, s. guimarães, r. T. de silva, and s. Araújo. 2005. "sistema de Licenciamento Ambienal em propriedades rurais do estado do Mato grosso: Análise de lições na sua implementação." Final report to the Brazilian Ministry of the environment, Brasilia. Lofgren, hans, James Thurlow, and sherman robinson. 2004. "prospects for growth and poverty reduction in Zambia, 2001­2015." discussion paper 11. international Food policy research institute, Washington, d.C. Lombardini, C. 1994. "deforestation in Thailand." in K. Brown and d. pearce, eds., The Causes of Tropical Deforestation. London: UCL press. Mace, georgina, hillary Masundire, and Jonathan Baillie. 2005. "Biodiver- sity." in rashid hassan, robert scholes, and neville Ash, eds., Ecosys- tems and Human Well-being: Current State and Trends. Washington, d.C.: island press. MacKinnon, Kathi. 2005. "parks, people and policies: Conflicting Agendas for Forests in southeast Asia." in e. Bermingham, C. W. dick, and C. Moritz, eds., Tropical Rainforests. Past, Present and Future. Chicago: Uni- versity of Chicago press. Mahapatra, Krushna, and shashi Kant. 2005. "Tropical deforestation: A Multinomial Logistic Model and some Country-specific policy prescrip- tions." Forest Policy and Economics 7 (1): 1­24. Mahar, d. J. 2000. "Agro-ecological Zoning in rondônia, Brazil: What Are the Lessons?" in Anthony hall, ed., Amazonia at the Crossroads: The 262 r e F e r e n C e s Challenge of Sustainable Development. University of London, institute of Latin American studies. Malla, Y. B. 2000. "impact of Community Forestry policy on rural Liveli- hoods and Food security in nepal." Unasylva 51 (202): 37­45. Margules, C. r., and r. L. pressey. 2000. "systematic Conservation plan- ning." Tropical Forest research Centre and rainforest Cooperative research Centre, Atherton, Australia. Margulis, sergio. 2004. "Causes of deforestation of the Brazilian Amazon." Working paper 22. World Bank, Washington, d.C. Mather, A. s. 1992. "The Forest Transition." AREA--Institute of British Geog- raphers 24 (4): 367­79. Mayers, James, and stephen Bass. 1999. Policy That Works for Forests and People. London: iied. McAndrew, John p., Mam sambath, hong Kimly, and Ky Bunthai. 2004. "phnong Villagers Adapt to decline in natural resources." Cambodia Development Review 8 (1): 9­12. McConnell, William J., sean p. sweeney, and Bradley Mulley. 2004. "physi- cal and social Access to Land: spatio-temporal patterns of Agricultural expansion in Madagascar." Agriculture, Ecosystems & Environment 101 (2­3): 171­84. McKenney, Bruce, Yim Chea, prom Tola, and Tom evans. 2004. "Focusing on Cambodia's high Value Forests: Livelihoods and Management." Cam- bodian development resource institute and Wildlife Conservation soci- ety, phnom penh. Mcneely, Jeffrey A., and sara J. scherr. 2003. Ecoagriculture: Strategies to Feed the World and Save Wild Biodiversity. Washington, d.C.: island press. Meinzen-dick, ruth s., and Monica di gregorio. 2004. "Collective Action and property rights for sustainable development 2020." Focus 11. CApri, Washington, d.C. Mendelsohn, robert, Ariel dinar, and Apurva sanghi. 2001. "The effect of development on the Climate sensitivity of Agriculture." Environment and Development Economics 6: 85­101. Merry, F. d., p. e. hildebrand, p. pattie, and d. r. Carter. 2002. "An Analysis of Land Conversion from sustainable Forestry to pasture: A Case study in the Bolivian Lowlands." Land Use Policy 19 (3): 207­15. Mertens, B., r. poccard-Chapuis, M.-g. piketty, A.-e. Lacques, and A. Ven- turieri. 2002. "Crossing spatial Analyses and Livestock economics to Understand deforestation processes in the Brazilian Amazon: The Case of sao Felix do xingu in south pará." Agricultural Economics 27 (3): 269­94. Mertens, Benoit, david Kaimowitz, Atie puntodewo, Jerry Vanclay, and patricia Mendez. 2004. "Modeling deforestation at distinct geographic 263 AT L o g g e r h e A d s ? scales and Time periods in santa Cruz, Bolivia." International Regional Science Review 27 (3): 271­96. Millennium ecosystem Asssessment. 2005. Ecosystems and Human Well- being: Synthesis. Washington, d.C.: island press. Ministry of environment and Forests. 2005. State of Forest Report 2003. deh- radun, india. Minten, Bart. 1999. "infrastructure, Market Access, and Agricultural prices: evidence from Madagascar." Mssd discussion paper 26. international Food policy research institute, Washington, d.C. Minten, Bart, and philippe Méral. 2005. "international Trade and environ- mental degradation: A Case study on the Loss of spiny Forest Cover in Madagascar." in Bart Minten, ed., Trade, Liberalization, Rural Poverty, and the Environment: The Case of Madagascar. Washington, d.C.: World Wildlife Fund. Mistiaen, J., B. soler, T. razafimanantena, and J. razafindravonona. 2002. "putting Welfare on the Map in Madagascar." Working paper 34. World Bank, Africa region, Washington, d.C. Molnar, Augusta, sara J. scherr, and Arvind Khare. 2004. "Who Conserves the World's Forests? A new Assessment of Conservation and investment Trends." Forest Trends, Washington, d.C. Monela, g. C., s. A. o. Chamshama, r. Mwaipopo, and d. M. gamassa. 2004. "A study on the social, economic and environmental impacts of Forest Landscape restoration in shinyanga region, Tanzania." Ministry of natural resources and Tourism (Forestry and Beekeeping division) and World Conservation Union (eastern Africa regional office), dar es salaam, Tanzania. Moser, Christine, Christopher B. Barrett, and Bart Minten. 2005. "Missed opportunities and Missing Markets: spatio-temporal Arbitrage of rice in Madagascar." Working paper 180. sAgA. Working paper. 180, U.s. Agency for international development, sAgA, Washington, d.C. Müller, daniel, and darla K. Munroe. 2005. "Tradeoffs between rural devel- opment policies and Forest protection: spatially explicit Modeling in the Central highlands of Vietnam." Land Economics 81 (3): 412­25. Müller, daniel, and Manfred Zeller. 2002. "Land Use dynamics in the Central highlands of Vietnam: A spatial Model Combining Village survey data with satellite imagery interpretation." Agricultural Economics 27 (3): 333­54. Muñoz-piña, Carlos, Alejandro guevara, Juan Manuel Torres, and Josefina Braña. 2005. "paying for the hydrological services of Mexico's Forests: Analysis, negotiations and results." revised version of a paper presented at the ZeF-CiFor workshop on "payments for environmental services: Methods and design in developing and developed Countries," Titisee, germany, June 15­18, 2005. Munroe, darla K., Jane southworth, and Catherine M. Tucker. 2002. "The dynamics of Land-cover Change in Western honduras: exploring spatial and Temporal Complexity." Agricultural Economics 27: 355­69. 264 r e F e r e n C e s ------. 2004. "Modeling spatially and Temporally Complex Land-Cover Change: The Case of Western honduras." The Professional Geographer 56 (4): 544­59. Murali, K. s., indu K. Murthy, and n. h. ravindranath. 2002. "Joint Forest Management in india and its ecological impacts." Environmental Man- agement and Health 13 (5): 512­28. Murray, Brian C., Bruce A. McCarl, and hemg-Chi Lee. 2004. "estimating Leakage from Forest Carbon sequestration programs." Land Economics 80 (1): 109­24. Myers, norman, russell A. Mittermeier, Cristina g. Mittermeier, gustavo A. B. da Fonseca, and Jennifer Kent. 2000. "Biodiversity hotspots for Conservation priorities." Nature 403: 853­58. naidoo, robin, and Wiktor L. Adamowicz. 2005. "economic Benefits of Bio- diversity exceed Costs of Conservation at an African rainforest reserve." Proceedings of the National Academy of Sciences 102 (46): 16712­16. ------. 2006. "Modeling opportunity Costs of Conservation in Transitional Landscapes." Conservation Biology 20 (2): 490­500. national research Council, Water science and Technology Board. 2004. Valuing Ecosystem Services: Towards Better Environmental Decision-Mak- ing. national Academies press, Washington, d.C. ndjanyou, L., and C. h. Majerowitz. 2004. "Actualisation de l'audit de la fiscalité decentralisée du secteur forestier camerounais." institutions et developpement, Boulogne, France. nelson, Andrew, and Kenneth M. Chomitz. 2006. "The Forest-hydrology- Watershed nexus in Central America: An heuristic Analysis." Environ- ment, Development and Sustainability. nelson, g. C., and d. hellerstein. 1997. "do roads Cause deforestation? Using satellite images in econometric Analysis of Land Use." American Journal of Agricultural Economics 79: 80­88. nelson, gerald, Alessandro de pinto, Virginia harris, and steven W. stone. 2004. "Land Use and road improvements: A spatial perspective." Inter- national Regional Science Review 27 (3): 297­325. nelson, gerald C., Virginia harris, and steven W. stone. 2001. "deforesta- tion, Land Use, and property rights: empirical evidence from darien, panama." Land Economics 77 (2): 187­205. nepstad, d. C., g. Carvalho, A. C. Barros, A. Alencar, J. p. Capobianco, J. Bishop, Moutinho, p. Lefebvre, U. Lopes silva Jr., and e. prins. 2001. "road paving, Fire regime Feedbacks, and the Future of Amazon For- ests." Forest Ecology and Management 154: 395­407. nepstad, d. C., s. schwartzmann, B. Bamberger, M. santilli, d. ray, p. schlesinger, p. Lefebvre, A. Alencar, e. prinz, g. Fiske, and A. rolla. 2006. "inhibition of Amazon deforestation and Fire by parks and indig- enous Lands." Conservation Biology 20 (1): 65­73. nguyen, Quang Tan. 2006. "Trends in Forest ownership, Forest resources Tenure and institutional Arrangements: Are They Contributing to Bet- 265 AT L o g g e r h e A d s ? ter Forest Management and poverty reduction? The Case of Viet nam." Food and Agriculture organization, rome. ninan, K. n., and Jyothis sathyapalan. 2005. "The economics of Biodiver- sity Conservation: A study of a Coffee growing region in the Western ghats of india." Ecological Economics 55 (1): 61­72. nittler, John, and henry Tschinkel. 2005. "Community Forest Management in Maya Biosphere reserve of guatemala." submitted to UsAid; steering Committee on nature, Wealth and power (nWp); sustainable Agricul- ture and natural resource Management (sAnreM); and Collaborative research support program (Crsp), University of georgia. norton-griffiths, Michael, and Clive southey. 1995. "The opportunity Costs of Biodiversity Conservation in Kenya." Ecological Economics 12 (2): 125­39. olschewski, roland, and pablo C. Benitez. 2005. "secondary Forests as Temporary Carbon sinks? The economic impact of Accounting Methods on reforestation projects in the Tropics." Ecological Economics 55 (3): 380­94. olson, d. M., and e. dinerstein. 1998. "The global 200: priority ecoregions for global Conservation." Annals of the Missouri Botanical Garden 89: 125­26. olson, david M., eric dinerstein, eric d. Wikramanayake, neil d. Burgess, george V. n. powell, emma C. Underwood, Jennifer A. d'Amico, illanga itoua, holly e. strand, John C. Morrison, Colby J. Loucks, Thomas F. Allnutt, Taylor h. ricketts, Yumiko Kura, John F. Lamoreux, Wesley W. Wettengel, prashant hedao, and Kenneth r. Kassem. 2001. "Terrestrial ecoregions of the World: A new Map of Life on earth." BioScience 51 (11): 933­38. osgood, d. 1994. "government Failure and deforestation in indonesia." in K. Brown and d. pearce, eds., The Causes of Tropical Deforestation. Lon- don: UCL press. ostrom, elinor. 1990. Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge, U.K.: Cambridge University press. ------. 1999. "self-governance and Forest resources." occasional paper 20. Center for international Forestry research, Bagor, indonesia. oyono, p. rene. 2005. "profiling Local-level outcomes of environmental decentralizations: The Case of Cameroon's Forests in the Congo Basin." Journal of Environment and Development 14 (3): 317­37. pacala, s., and r. socolow. 2004. "stabilization Wedges: solving the Climate problem for the next 50 Years with Current Technologies." Science 305. pacheco, pablo. 2002. "deforestation and degradation in Lowland Bolivia." in Charles h. Wood and roberto porro, eds., Deforestation and Land Use in the Amazon. gainesville: University of Florida press. pagiola, stefano. 2005. "payments for environmental services in Costa rica." revised version of a paper presented at the ZeF-CiFor workshop 266 r e F e r e n C e s on payments for environmental services: Methods and design in devel- oping and developed countries, Titisee, germany, June 15­18, 2005. pagiola, stefano, Agustin Arcenas, and gunars platais. 2005. "Can payments for environmental services help reduce poverty? An exploration of the issues and the evidence to date from Latin America." World Develop- ment 33 (2): 237­53. pagiola, stefano, elías ramírez, José gobbi, Cees de haan, Muhammad ibrahim, enrique Murgueitio, and Juan pablo ruíz. 2006. "paying for the environmental services of silvopastoral practices in nicaragua." pre- pared for submission to special issue of Ecological Economics on "eco- system services and Agriculture," edited by s. M. swinton, F. Lupi, s. h. hamilton, and g. p. robertson. pagiola, stefano, paola Agostini, gobbi José, Cees de haan, Muhammad ibrahim, enrique Murgueitio, elías ramírez, Mauricio rosales, and Juan pablo ruíz. 2004. "paying for Biodiversity Conservation services in Agri- cultural Landscapes." environmental economics series 96. World Bank, Washington, d.C. palm, Cheryl, stephen A. Vosti, pedro sanchez, and polly J. ericksen. 2005. Slash-and-Burn Agriculture: The Search for Alternatives. new York: Columbia University press. palmer, Charles, and stefanie engel. 2006. "For Better or for Worse? Local impacts of the decentralization of indonesia's Forest." eTh, Zürich. pan, William K. Y., stephen J. Walsh, richard e. Bilsborrow, Brian g. Frizzelle, Christine M. erlien, and Francis Baquero. 2004. "Farm-level Models of spatial patterns of Land Use and Land Cover dynamics in the ecuadorian Amazon." Agriculture Ecosystems and Environment 101: 117­34. panayotou, T., and s. sungsuwan. 1994. "An econometric Analysis of the Causes of Tropical deforestation: The Case of northeast Thailand." in K. Brown and d. pearce, eds., The Causes of Tropical Deforestation. Lon- don: UCL press. pearce, david, Francis e. putz, and Jerome K. Vanclay. 2003. "sustainable Forestry in the Tropics: panacea or Folly?" Forest Ecology and Manage- ment 172 (2­3): 229­47. pender, John, pamela Jagger, ephraim nkonya, and dick sserunkuuma. 2004. "development pathways and Land Management in Uganda." World Development 32 (5): 767­92. pendleton, Linwood h., and e. Lance howe. 2002. "Market integration, development, and smallholder Forest Clearance." Land Economics 78 (1): 1­19. perotto-Baldiviezo, h. L., T. L. Thurow, C. T. smith, r. F. Fisher, and x. B. Wu. 2004. "gis-based spatial Analysis and Modeling for Landslide haz- ard Assessment in steeplands, southern honduras." Agriculture, Ecosys- tems & Environment 103 (1): 165­76. 267 AT L o g g e r h e A d s ? pfaff, A. s. p. 1999. "What drives deforestation in the Brazilian Amazon? evidence from satellite and socioeconomic data." Journal of Environ- mental Economics and Management 37: 26­43. pichón, Francisco J. 1997. "Colonist Land-Allocation decision, Land Use, and deforestation in the ecuadorian Amazon Frontier." Economic Devel- opment and Cultural Change 45 (4): 707­44. pinedo-Vasquez, Miguel, daniel Zarin, and peter Jipp. 1992. "economic returns from Forest Conversion in the peruvian Amazon." Ecological Economics 6 (2): 163­73. putz, Francis e., geoffrey M. Blate, Kent h. redford, robert Fimbel, and John robinson. 2001. "Tropical Forest Management and Conservation of Biodiversity: An overview." Conservation Biology 15 (1): 7­20. randrianarisoa, J. C. 2003. "Analyse spatial de la production rizicole mal- gache." presented at Cornell University and international Labour orga- nization conference on developpement économique, services sociaux et pauvreté à Madagascar. ravindranath, n. h., and B. s. somashekhar. 1995. "potentials and eco- nomics of Forestry options for Carbon sequestration in india." Biomass and Bioenergy 8 (5): 323­36. reardon, Thomas, Julio Berdegue, and german escobar. 2001. "rural non- farm employment and incomes in Latin America: overview and policy implications." World Development 29 (3): 395­409. redford, Kent, and Michael painter. 2006. "natural Alliances between Con- servationists and indigenous peoples." Working paper 25. Wildlife Con- servation society, new York. reis, e. J., and r. guzmán. 1994. "An econometric Model of Amazon defor- estation." in K. Brown and d. pearce, eds., The Causes of Tropical Defor- estation. London: UCL press. renkow, Mitch, daniel g. hallstrom, and daniel d. Karanja. 2004. "rural infrastructure, Transactions Costs and Market participation in Kenya." Journal of Development Economics 73 (1): 349­67. ribot, Jesse. 2002. "democratic decentralization of natural resources: insti- tutionalizing popular participation." World resources institute, Washing- ton, d.C. ------. 2003. "democratic decentralization of natural resources: institu- tional Choice and discretionary power Transfer in sub-saharan Africa." Public Administration and Development 23: 53­65. ribot, Jesse, and Anne M. Larson. 2005. Democratic Decentralisation through a Natural Resource Lens. London: routledge. ricker, Martin, robert o. Mendelsohn, douglas C. daly, and guillermo Angeles. 1999. "enriching the rainforest with native Fruit Trees: An ecological and economic Analysis in Los Tuxtlas (Veracruz, Mexico)." Ecological Economics 31 (3): 439­48. 268 r e F e r e n C e s ricketts, Taylor h., gretchen C. daily, paul r. ehrlich, and Charles d. Michener. 2004. "economic Value of Tropical Forest to Coffee production." Proceedings of the National Academy of Sciences 101 (34): 12579­82. ricketts, T. h., e. dinerstein, Tim Boucher, Thomas Brooks, s. h. M. Butchart, Michael hoffmann, J. F. Lamoreux, J. Morrison, M. parr, J. d. pilgrim, A. s. L. rodrigues, W. sechrest, g. Wallace, K. Berlin, J. Bielby, n. d. Burgess, d. r. Church, n. Cox, d. Knox, C. Loucks, g. W. Luck, L. L. Masterl, r. Moore, robin naidoo, r. ridgely, g. e. schatz, g. shire, h. strand, W. Wettengel, and e. Wikramanayake. 2005. "pinpointing and preventing imminent extinctions." PNAS 102 (51): 18497­501. ridder, ralph. 2006. presentation at University of Montana workshop on Long-term Monitoring of Vegetation Variables Using Moderate resolution satellites, August 9, Missoula, Mont. rodrigues, A. s. L., s. J. Andelman, M. i. Bakarr, L. Boitani, T. M. Brooks, L. d. C. Fishpool, g. A. B. da Fonseca, K. J. gaston, M. hoffmann, J. s. Long, p. A. Marquet, J. d. pilgrim, r. L. pressey, J. schipper, W. sechrest, s. n. stuart, L. g. Underhill, r. W. Waller, M. e. J. Watts, and x. Yan. 2004. "effectiveness of the global protected Areas network in represent- ing species diversity." Nature 428: 640­43. rojas, Manrique, and Bruce Aylward. 2002. "Cooperation between a small private hydropower producer and a Conservation ngo for Forest protec- tion: The Case of La esperanza, Costa rica." Land-Water Linkages in rural Watersheds Case study series. Food and Agriculture organization, rome. roldan ortega, roque. 2004. "Models for recognizing indigenous Land rights in Latin America." environment department paper 99. World Bank, Washington, d.C. romano, Francesca, and dominique reeb. 2006. "Understanding Forest Tenure: What rights and for Whom? secure Forest Tenure for sustain- able Forest Management and poverty Alleviation: The Case of southeast Asia." Food and Agriculture organization, Forestry department, Forest policy and information division, rome. roper, J. Montgomery. 2003. "An Assessment of indigenous participation in Commercial Forestry Markets: The Case of nicaragua's northern Atlantic Autonomous region." Forest Trends, Washington, d.C. rosenzweig, M. L. 2001. "Loss of speciation rate Will impoverish Future diversity." PNAS 98 (10): 5404­10. ------. 2003. Win-Win Ecology: How the Earth's Species Can Survive in the Midst of Human Enterprise. new York: oxford University press. ross, Michael L. 2001. Timber Booms and Institutional Breakdown in South- east Asia. Cambridge, U.K.: Cambridge University press. rudel, Thomas K. 2005. Tropical Forests. new York: Columbia University press. 269 AT L o g g e r h e A d s ? rudel, Thomas K., Marla perez-Lugo, and heather Zichal. 2000. "When Fields revert to Forest: development and spontaneous reforestation in post-War puerto rico." Professional Geographer 52 (3): 386­97. rudel, Thomas K., oliver T. Coomes, emilio Moran, Frederic Achard, Arild Angelsen, Jianchu xu, and eric Lambin. 2005. "Forest Transitions: Towards a global Understanding of Land Use Change." Global Environ- mental Change Part A 15 (1): 23­31. ruf, François. 2001. "Tree Crops as deforestation and reforestation Agents: The Case of Cocoa in Côte d'ivoire and sulawesi." in Arild Angelsen and david Kaimowitz, eds., Agricultural Technologies and Tropical Deforesta- tion. Wallingford, U.K.: CABi publishing. sanchez-Azofeifa, g. Arturo, g. C. daily, A. s. p. pfaff, and Christopher Busch. 2003. "integrity and isolation of Costa rica's national parks and Biological reserves: examining the dynamics of Land-cover Change." Biological Conservation 109: 123­35. sarin, Madhu, Lipika ray, Manju s. raju, Mitali Chattarjee, narayan Baner- jee, and shyamala hiremath. 1998. "Who is gaining? Who is Losing? gender and equity Concerns in Joint Forest Management." society for promotion of Wasteland development, new delhi. sathaye, J., W. Makundi, L. dale, p. Chan, and K. Andrasko. Forthcom- ing. "ghg Mitigation potential, Costs and Benefits in global Forests: A dynamic partial equilibrium Approach." Energy Journal. scherr, sara J., Andy White, and david Kaimowitz. 2003. "A new Agenda for Forest Conservation and poverty reduction: Making Forest Markets Work for Low-income producers." Forest Trends, Washington, d.C. schmidt-soltau, Kai. 2003. "Conservation-related resettlement in Central Africa: environmental and social risks." Development and Change 34 (3): 525­51. schneider, r. r. 1995. "government and economy on the Amazon Fron- tier." environment paper 11. World Bank, Washington, d.C. schneider, s. h. 2004. "Abrupt non-linear Climate Change, irreversibility and surprise." Global Environmental Change 14: 245­58. schweik, Charles M., harini nagendra, and deb ranjan sinha. 2003. "Using satellite imagery to Locate innovative Forest Management practices in nepal." Ambio 32 (4): 312­19. serneels, suzanne, and eric F. Lambin. 2001. "proximate Causes of Land- use Change in narok district, Kenya: A spatial statistical Model." Agri- culture, Ecosystems & Environment 85 (1­3): 65­81. shively, gerald, and elmer Martinez. 2001. "deforestation, irrigation, employment and Cautious optimism in southern palawan, the philip- pines." in Arild Angelsen and david Kaimowitz, eds., Agricultural Tech- nologies and Tropical Deforestation. Wallingford, U.K.: CABi publishing. shively, gerald, and s. n pagiola. 2004. "Agricultural intensification, Local Labor Markets, and deforestation in the philippines." Environment and Development Economics 9: 241­66. 270 r e F e r e n C e s shone, Bryan M., and Jill L. Caviglia-harris. 2005. "Quantifying and Com- paring the Value of non-timber Forest products in the Amazon." Ecologi- cal Economics 58 (2): 249­267. shyamsundar, priya, and s. Bandyopadhyay. 2004. "Fuelwood Consump- tion and participation in Community Forestry in india." policy research Working paper 3331. World Bank, Washington, d.C. shyamsundar, priya, and randall A. Kramer. 1996. "Tropical Forest protec- tion: An empirical Analysis of the Costs Borne by Local people." Journal of Environmental Economics and Management 31: 129­44. shyamsundar, priya, eduardo Araral, and suranjan Weeraratne. 2005. "devolution of resource rights, poverty, and natural resource Manage- ment: A review." environmental economics series, Working paper no. 104, World Bank, Washington, d.C. simpson, r. david, roger A. sedjo, and John W. reid. 1996. "Valuing Bio- diversity for its Use in pharmaceutical research." Journal of Political Economy 104 (1): 163­85. somanathan, e., r. prabhakar, and Bhupendra singh Metha. 2005. "does decentralization Work? Forest Conservation in the himalayas." discus- sion paper in economics 05-04. indian statistical institute, delhi. southgate, douglas, rodrigo sierra, and Lawrence Brown. 1991. "The Causes of Tropical deforestation in ecuador: A statistical Analysis." World Development 19 (9): 1145­51. southworth, Jane, harini nagendra, Laura A. Carlson, and Catherine Tucker. 2004. "Assessing the impact of Celaque national park on For- est Fragmentation in Western honduras." Applied Geography 24 (4): 303­22. steinberg, paul F. 2001. Environmental Leadership in Developing Countries: Transnational Relations and Biodiversity Policy in Costa Rica and Bolivia. Cambridge, Mass.: MiT press. ------. 2005. "From public Concern to policy effectiveness: Civic Conserva- tion in developing Countries." Journal of International Wildlife Law and Policy 8: 341­65. steininger, M. K., g. harper, C. J. Tucker, d. Juhn, and F. hawkins. 2004 "Forest Clearance and Fragmentation in Madagascar." unpublished digi- tal map. Washington, Conservation international. stockholm environment institute. 2002. "Charcoal potential in southern Africa--ChAposA." Final report. stockholm. stocks, A., B. McMahan, and p. Taber. 2006. "Beyond the Map: indigenous and Colonist impacts and Territorial defense in nicaragua's Bosawas reserve." Unpublished. stolle, F., K. M. Chomitz, eric Lambin, and Thomas p. Tomich. 2003. "spa- tial determinants of Vegetation Fires in Jambi province, sumatra, indo- nesia." Forest Ecology and Management 179 (July): 277­92. stolton, sue, Marc hockings, nigel dudley, Kathy MacKinnon, and Tony Whitten. 2003. "reporting progress at protected Area sites: A simple 271 AT L o g g e r h e A d s ? site-level Tracking Tool developed for the World Bank and WWF." World Bank, Washington, d.C. stoms, david M., Kenneth M. Chomitz, and F. W. davis. 2004. "TAMArin: A Landscape Framework for evaluating economic incentives for rainfor- est restoration." Landscape and Urban Planning 68 (1 May): 95­108. stone, steven W. 1998. "Using a geographic information system for Applied policy Analysis: The Case of Logging in the eastern Amazon." Ecological Economics 27 (1): 43­61. stoneham, gary, Vivek Chaudhri, Arthur ha, and Loris strappazzon. 2003. "Auctions for Conservation Contracts: An empirical examination of Vic- toria's BushTender Trial." Australian Journal of Agricultural and Resource Economics 47 (4): 477­500. sunderlin, William d., sonya dewi, and Atie puntodewo. 2006. "Forests, poverty, and poverty Alleviation policies." Background paper. sunderlin, William d., Arild Angelsen, Brian Belcher, paul Burgers, robert nasi, Levania santoso, and sven Wunder. 2005. "Livelihoods, Forests, and Conservation in developing Countries: An overview." World Devel- opment 33 (9): 1383­1402. Tacconi, Luca. 2003. "Fires in indonesia: Causes, Costs, and policy implica- tions." CiFor occasional paper no. 38, Center for international Forestry research, Bogor. ten Kate, K., J. Bishop, and r. Bayon. 2004. "Biodiversity offsets: Views, experience, and the Business Case." iUCn, gland, switzerland and Cam- bridge, UK and insight investment, London, UK. Thanapakpawin, porranee, Jeffrey e. richey, david Thomas, sarah d. rodda, Bruce Campbell, and Miles g. Logsdon. 2006. "effects of Land Use Change on the hydrologic regime of the Mae Chaem river Basin, nW Thailand." submitted for publication. river systems research group, University of Washington, seattle. Tiffen, Mary, and Michael Mortimore. 1994. "environment, population growth and productivity in Kenya: A Case study of Machakos district." international institute for environment and development, London. Tomich, Thomas p., hubert de Foresta, rona dennis, Quirine Ketterings, daniel Murdiyarso, Cheryl palm, Fred stolle, suyanto, and Meine Van noordwijk. 2002. "Carbon offsets for Conservation and development in indonesia?" American Journal of Alternative Agriculture 17 (3): 125­37. Tomich, Thomas p., Andrea Cattaneo, simon Chater, helmut J. geist, James gockowski, david Kaimowitz, eric Lambin, Jessa Lewis, ousseynou ndoye, Cheryl palm, Fred stolle, William d. sunderlin, Judson F. Valen- tim, Meine Van noordwijk, and stephen A. Vosti. 2005. "Balancing Agri- cultural development and environmental objectives: Assessing Tradeoffs in the humid Tropics." in Cheryl palm, stephen A. Vosti, pedro sanchez, and polly J. ericksen, eds., Slash-and-Burn Agriculture: The Search for Alternatives. new York: Columbia University press. 272 r e F e r e n C e s Tucker, Catherine M., darla K. Munroe, harini nagendra, and Jane south- worth. 2005. "Comparative spatial Analyses of Forest Conservation and Change in honduras and guatemala." Conservation and Society 3 (1): 174­200. Undp (United nations development programme). Atlas do Desenvolvi- mento Humano no Brasil. [http://www.pnud.org.br/atlas/]. Unep. 2005. One Planet Many People: Atlas of Our Changing Environment. Unep: nairobi. Un population division. 2004. "World population prospects 2004, Medium Variant." new York. Uphoff, n. 2003. "higher Yields with Fewer external inputs? The system of rice intensification and potential Contributions to Agricultural sustain- ability." International Journal of Agricultural Sustainability 1 (1): 38­50. UsdA (U.s. department of Agriculture). 2006. "oilseeds: World Markets and Trade." Circular series Fop 3-06. Foreign Agricultural service, Wash- ington, d.C. van der Molen, M. K., A. J. dolman, M. J. Waterloo, and L. A. Bruijnzeel. Forthcoming. "Climate is Affected More by Maritime Than by Continen- tal Land Use Change: A Multiple scale Analysis." Global and Planetary Change. Van noordwijk, Meine, F. Agus, B. Verbist, K. hairiah, and Thomas p. Tomich. 2006. "Managing Watershed services in ecoagriculture Land- scapes." in Jeffrey A. Mcneely and sara J. scherr, eds., The State-of-the- Art of Ecoagriculture. Washington, d.C.: island press. van Wilgen, B. W., C. Marais, d. Magadlela, n. Jezile, and d. stevens. 2002. "Win-Win-Win: south Africa's Working for Water programme." in s. M. pierce, r. M. Cowling, T. sandwith, and K. MacKinnon, eds., Main- streaming Biodiversity in Development: Case Studies from South Africa. Washington, d.C.: World Bank. Vance, Colin, and Jacqueline geoghegan. 2002. "Temporal and spatial Mod- elling of Tropical deforestation: A survival Analysis Linking satellite and household survey data." Agricultural Economics 1683: 1­16. Vedeld, paul, Arild Angelsen, espe sjaastad, and gertrude K. Berg. 2004. "Counting on the environment: Forest incomes and the rural poor." environmental economics series, Working paper 98. World Bank, Wash- ington, d.C. Vincent, r. Jeffrey, C. Clark gibson, and Marco Boscolo. 2003. "The politics and economics of Forest reforms in Cameroon." World Bank, Washing- ton, d.C. von Amsberg, J. 1998. "economic parameters of deforestation." The World Bank Economic Review 12: 133­53. von Thünen, Johann heinrich. 1966 Isolated state (an english edition of Der isolierte Staat). Translated by Carla M. Wartenberg. edited with an intro- duction by peter hall, oxford. new York: pergamon press. 273 AT L o g g e r h e A d s ? Vosti, stephen A., James gockowski, and Thomas p. Tomich. 2005. "Land Use systems at the Margins of Tropical Moist Forest: Addressing small- holder Concerns in Cameroon, indonesia, and Brazil." in Cheryl palm, stephen A. Vosti, pedro sanchez, and polly J. ericksen, eds., Slash-and- Burn Agriculture: The Search for Alternatives. new York: Columbia Uni- versity press. Warnken, philip F. 1999. The Development and Growth of Soybean Industry in Brazil. Ames, iowa: iowa state University press. Warr, peter. 2005. roads and poverty in rural Laos. Working paper 4. rspAs, The Australian national University. Watson, robert T. and the Core Writing Team, eds. 2001. Climate Change 2001: Synthesis Report. A Contribution of Working Groups I, II and III to the Third Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, U.K.: Cambridge University press. WdpA Consortium. 2005. "World database on protected Areas 2005." World Conservation Union (iUCn) and United nations environment pro- gramme­World Conservation Monitoring Centre. Wertz-Kanounnikoff, sheila. 2005. "deforestation in Amazônia: An inquiry into the effect of Land Tenure and population from 2000 to 2003." World Bank. White, Andy, and Alejandra Martin. 2002. "Who owns the World's Forests?" Forest Trends, Washington, d.C. Whitten, Tony, sengli J. damanik, Jazanul Anwar, and nazaruddin hisyam. 2000. The Ecology of Sumatra. singapore: periplus edition. Wigmosta, M. s., L. Vail, and d. p. Wittenmaier. 1994. "A distributed hydrology-vegetation Model for Complex Terrain." Water Resources Research 30 (6): 1665­79. Wilson, Kerrie, Adrian newton, Cristian echeverria, Chris Weston, and Mark Burgman. 2005. "A Vulnerability Analysis of the Temperate Forests of south Central Chile." Biological Conservation 122 (1): 9­21. World Bank. 2001. "peru second rural roads project." project Appraisal document 22110. Washington, d.C. ------. 2002. World Development Report 2003: Sustainable Development in a Dynamic World: Transforming Institutions, Growth, and the Quality of Life. new York: oxford University press. ------. 2003. "implementation Completion report (CpL-34440) on a Loan in the Amount of Us$167.0 Million to the Federative republic of Brazil for a rondônia natural resources Management project." report 26080. Washington, d.C. ------. 2004. "sustaining Forests: A development strategy." Washington, d.C. ------. 2005. "india. Unlocking opportunities for Forest-dependent people in india." report 34481-in. Washington, d.C. Wri (World resources institute). 2005. The Wealth of the Poor: Managing Ecosystems to Fight Poverty. Washington, d.C. 274 r e F e r e n C e s Wu, JunJie. 2000. "slippage effects of the Conservation reserve program." American Journal of Agricultural Economics 82 (4): 979­92. Wunder, s. 2000. The Economics of Deforestation. new York: st. Martin's. Wunder, sven, and William d. sunderlin. 2004. "oil, Macroeconomics, and Forests: Assessing the Linkages." The World Bank Research Observer 19 (2): 231­57. WWF (World Wildlife Fund). 2001. "Terrestrial ecoregions gis database." [http://www.worldwildlife.org/science/data/terreco.cfm]. Accessed May 25, 2005. xu, Zhigang, Michael T. Bennett, ran Tao, and Jintao xu. 2004. "China's sloping Land Conversion programme Four Years on: Current situation, pending issues." International Forestry Review 6 (3­4): 317­26. Yaron, gil. 2001. "Forest, plantation Crops or small-scale Agriculture? An economic Analysis of Alternative Land Use options in the Mount Camer- oon Area." Journal of Environmental Planning and Management 44 (1): 85­108. Yin, hongfu, and Changan Li. 2001. "human impact on Floods and Flood disasters on the Yangtze river." Geomorphology 41 (2­3): 105­09. Yohe, gary, natasha Andronova, and Michael schlesinger. 2004. "CLiMATe: To hedge or not Against an Uncertain Climate Future?" Science 306 (5695): 416­17. Zbinden, simon, and david r. Lee. 2004. "paying for environmental ser- vices: An Analysis of participation in Costa rica's psA program." World Development 33 (2): 255­72. Zelek, C. A., and g. e. shively. 2003. "Measuring the opportunity Cost of Carbon sequestration in Tropical Agriculture." Land Economics 79 (3): 342­54. Zeller, Manfred, Aliou diagne, and Charles Mataya. 1998. "Market Access by smallholder Farmers in Malawi: implications for Technology Adop- tion, Agricultural productivity and Crop income." Agricultural Economics 19 (1­2): 219­29. Zhang, x., and s. Fan. 2001. "how productive is infrastructure? new Approach and evidence from rural india." discussion paper 84. inter- national Food policy research institute, Washington, d.C. Ziegler, Alan d., Thomas W. giambelluca, ross A. sutherland, Mike A. nul- let, sanay Yarnasarn, Jitti pinthong, pornchai preechapanya, and sath- aporn Jaiaree, 2004. "Toward Understanding the Cumulative impacts of roads in Upland Agricultural Watersheds of northern Thailand." Agri- culture, Ecosystems & Environment 104 (2004): 145­58. 275 Index Note:Pagenumbersfollowedbylettersb, f, m,n,p,andt biodiversityoffsetrequirements,proposed,209 refertoentriesinboxes,figures,maps,notes,photos,and bioprospecting,112,209 tables,respectively. definition,110 endemicspecies,19,113 Africa environmentalservicepaymentsystemsand,16,186b charcoalextraction,43­44 andforestfragmentation,113,114b,115 deforestationbysmallholders,44 globalbiodiversityfinance,209,214 deforestationrates,43,214 globaldemandforprotection,9 displacementofpopulationsfromprotectedforestareas, greenagricultureand,179­80 92 andhabitatloss,113­14 effectsofroadsonmammalpopulations,69 informationgaps,16­17,215 forestdegradationandsoilsuitability,46­47,46f instrumentalrationaleforconservation,111­12 frontieranddisputedareas,30 intrinsicrationaleforconservation,111,112­13 populationinsavannaandnonsavannaforests,38,40 localbiologicalresources,111 savannabiomes,33m,38 Madagascar,19,21 tropicalforestbiomes,32m niches,113 agriculturalfrontier.Seeareasbeyondtheagriculturalfrontier See alsoextinctionthreats agriculturalprices.Seefarmgateprices biogeography,113­15 agriculturaltechnology biomassscarcity,90­91 effectsondeforestation,61t,66­67 biomes greeneragriculture,18b,179­80,213­14,213b,218 Africanbiomes,32m,33m greenrevolution,66­67,99 Asiantropicalforestbiomes,34m See alsopestcontrol;technologyforenvironmental andfocusofreport,28b­29b monitoring forestandsavannacomparison,27 agroclimate,47,62,126,162 LatinAmericanandCaribbeanbiomes,35m,36m agroecologicalzoning,139,158­59 tropicalforestsandsavannas,28m­29m agroforestry See alsoareasbeyondtheagriculturalfrontier;frontierand Asianforestedges,34,38 disputedareas;mosaiclands mappingof,34,37b bioprospecting,112,209 rubbercultivation,91,128,179 Bolivia AlternativestoSlashandBurn(ASB)program,56,59,126 agroecologicalzoning,158 areasbeyondtheagriculturalfrontier landclearancecosts,56,60 area,39t landvaluesatfrontier,56,57t,58t,59­60 nationalroleinbalancinginterests,11,153­54,193 problemsencounteredinlocalforestmanagement,173, overview,7,30,31t,216 177 population,39t roadexpansion,74 recommendationsofreport,17b­18b,212b,216 Brazil See alsoforestcores abandonmentofpastures,78 Argentina,192 Bahia,56,62,186b asymmetriesofpower,information,andorganization,10,140 biodiversity,56,59,186b Australia,160­61,186,209 cattleandbeefproduction,21­22 cerrado,21­23,21p,56,66 Biodiversity,106-111 currencydevaluation,21­22,191 balancingbiodiversityconservationagainstproduction, deforestationbylargecommercialinterests,44,95­96,95f 137­39,138f deforestationofpubliclands,98,162 biodiversitycorridors,16,114b,186b 277 At L o g g e r h e A d S ? deforestationratesandclearingsizes,95­98,96m,106n1, referencelevels(rLs)ofemissions,206 191 releasesperhectarefromdeforestation,15,195 environmentalcouncilsinmunicipios,143­44 targetforatmosphericconcentration,196 environmentallawenforcement,22­23,147 targetforemissionreductionsbymid-century,198 farmgatepricesofbeef,deforestationand,63,63f,191 See alsoclimatechange FeMA(stateenvironmentalagencyofMatogrosso),22­23 carbonfinance forestreserverequirements,22,23,181­82 argumentaboutleakageorslippagefromforestprotection, foresttransition,94 201­2 goias,21,22,23 argumentthatforestcarbonmakesmitigationtoocheap, landvalues,22,56,62 197­98 Matogrosso,19,21­23,21p,145,158 calculationsofpaymentsforreducedemissions,207 MatogrossodelSul,21 creatingnationalprogramstoreducedeforestation,204­6 NationalInstituteforSpaceresearch(INPe),21,97, financialincentivestoreduceforestCo2emissions,14­16, 106n1,144 195­96,203,205 overview,29b forestmanagementtoreduceCo2,costs,126­29,129f povertydensity,83,83m globalsourcesoffinanceforavoideddeforestation,204, povertyrates,82m,83,94 205­6 PublicMinistry,147 importanceforstabilizingclimate,196 relationshipbetweenpovertyanddeforestation,95­98 internationalmarketvaluesforCo2abatement,14­15, reservaParticulardoPatrimnioNatural(rPPN),156 (rPPN), 126­29,128f,195,197,207 roadexpansion,74 localandglobalbenefitsforforestsandruraldevelopment, rubberagroforestry,128 196­97 ruralilliteracydensity,97f,98 marketforemissionreductions,198 ruralPropertyenvironmentalLicensingSystem(SLAPr), optionssubmittedbyCostaricaandPapuaNewguinea, 22­23,187­88 197 soybeancultivation,21,21p,56,66,75 pricenegotiations,207­8 tocantinsriverflow,22,122 referencelevels(rLs)ofemissions,206 transferableforestprotectionobligations,181­82 synergisticeffectsofpoliciesforavoideddeforestation, zoningplans,158 208,208t BrazilianNationalInstituteforSpaceresearch(INPe),21,97, temporarynatureofdeforestationabatement,198­201 106n1,144 See alsoenvironmentalservicepaymentsystems; bushmeat,91,112 internationalroleindevelopmentandenvironmental issues;KyotoProtocol Cambodia,92,147 CBerS2Sino-Braziliansatellite,145 Cameroon cerrado,21­23,21p,56,66.See alsoBrazil;savannas economicvolatility,effectondeforestation,192 certificationprogramsforindustry,148­49 landvalues,56,58t China,50n1,70,116,183­84 oilpalmcultivation,56,58t climatechange overexploitationofbushmeat,91 deforestationeffects,1,9,125­29 reformsinforestpolicyandmanagement,159­60, andextinctionrisk,114,115 168b­170b localandregionalclimatechange,112,125,132 regulationofloggingandtimbersales,91,145,147­48, UnitedNationsFrameworkConventiononClimateChange, 168b­170b 196,197 zoningplans,159­60 weatherpatterns,effectsofdeforestation,9,125 carbondioxide(Co2) See alsocarbondioxide(Co2);carbonfinance abatementcosts,14­15,126­29,128f,197,207 communityforestmanagement environmentalimpact,9 benefitsforremoteforestcommunities,88 forestmanagementtoreduceCo2,biologicalaspects,126, challenges,176­78 127b inguatemala,175,177 forestmanagementtoreduceCo2,costs,126­29,129f India'sJointForestManagementprogram,173­74 globalreleasesfromdeforestation,1,14,126 inMexico,174­75 KyotoProtocollimitsonemissions,197­98 inNepal,174 monitoringandmeasurement,202­3,204 overview,13,155,171 optionsforreducingCo2emissions,196 roads,importanceof,88 278 I N d e x trade-offsandmixedresultswithlocalmanagement,171, localvariables,effectsondeforestationandincome,61t, 173 78­79 See alsoindigenouspeople macroeconomicpolicies,191­93 conflictandviolenceinforests,154­55 off-farmwages,effectsondeforestation,65 Confluenceproject,146 profitsfromdeforestation,55­60,79n4 conservationgroups sustainableforestmanagement,54­55,165­67,170­71 asymmetriesofpower,information,andorganization,10 See alsoenvironmentalservicepaymentsystems;farmgate funding,113 prices;incentivesforconservation;landvalues;timber instrumentalrationaleforconservation,111­12 values intrinsicrationaleforconservation,111,112­13 ecotourism,21,141,156,164,164t nongovernmentalorganizations(Ngos),16,113,142,173 ecuador,57t­59t,59,87 See alsospecific groups elites,10,12,84,139,146 Costarica endangeredspecies BraulioCarrilloNationalPark,189 "criticalorendangered"forests,124 environmentalservicepaymentsystem,183,184,187 definitions,47 FoNAFIFo,183 imminentextinctionsites,48m­49m,49f,50 landclearancecosts,60 incidenceinnonremoteareas,47­50,48f pollinators,111 rangesof,47 protectedareas,162­63,189 See alsoextinctionthreats reforestation,78,115­16 endemicspecies,19,113 roadplanning,189 environmentaleffectsofdeforestation suggestedfinancialincentivestoreduceforestCo2 "criticalorendangered"forests,124 emissions,197 externalitiesofdeforestation,overview,109­10,131­33, watershedprotection,118 132t croplandexpansion,overview,1­2,44.See alsolanduse importanceoftropicalforests,1 instrumentalrationaleforconservation,111­12 deforestationrates intrinsicrationaleforconservation,111,112­13 annualratesinAfrica,43,214 overview,1,9­10,109­10 annualratesinLatinAmericaandAsia,42t,43 waterquality,9,120,124­25 Brazil,95­98,96m,106n1,191 weatherpatterns,9,125 degradationrates,fordomains,45t See alsobiodiversity;extinctionthreats;hydrology­forest hotspotsoftropicaldeforestation,20,44m­45m,45­47, relationship 50n1,98 environmentalservicepaymentsystems indigenouspeople,lowerratesfor,12­13,172 asalternativetoprotectedareas,165 institutionalandpolicyresponsesto,6f auction-basedsystems,186b,187 overview,1,6f biodiversityand,16,186b studiesontropicaldeforestation,42­44 inChina,183­84 See alsoincentivesfordeforestation inColombia,184 degradationofforests inCostarica,183,184,187 biomassscarcity,90­91 designissues,184­87 irreversibledegradationofland,78,112,129­31,192,199 financeissues,184­85 rates,45­47,45t internationalfinancingof,213­14 soilsuitabilityand,46­47,46f,62 logisticsandadministrationissues,184,187­88 tragedyofthecommonsand,30,68,89,90­91 inMexico,183,187,188 democraticrepublicofCongo,91,189 inNicaragua,184 dipterocarptrees,62,92 overview,14,182­83 discountrates,55,56,176,178 regionalIntegratedSilvopastoralecosystemManagement domains.Seeforesttypesanddomains;specific domains Project(rISeMP),180,184,187 structuringofpayments,184,185­87 economicsofdeforestation U.S.ConservationreserveProgram,184,185,186b,202, cashandcreditconstraintsonsmallholders,60,62 209 discountrates,55,56,176,178 See alsocarbonfinance;incentivesforconservation farminputprices,62,64,68 erosion,116­17,119­20,183,202 landclearancecosts,56,60 externalitiesofdeforestation,overview,109­10,131­33,132t 279 At L o g g e r h e A d S ? extinctionthreats forestandsavannadegradationrates,45t andclimatechange,114 Madagascar,21p andforestfragmentation,19,113,114b,115 maps,32m,34m,35m andhabitatloss,113­14 population,34,38­40,39t imminentextinctionsites,48m­49m,49f,50 See alsofrontieranddisputedareas increaseinextinctionrates,111 forestpoverty instrumentalrationaleforconservation,111­12 incomefromforestproducts,estimationof,88­89 intrinsicrationaleforconservation,111,112­13 lackofforestmanagement,9 See alsobiodiversity;endangeredspecies;environmental andlimitedaccesstoforestresources,8­9 effectsofdeforestation overviewandcharacteristics,8­9,81­82,104­6 povertydensity,83­88,83m,85t,86f FAo.SeeFoodandAgricultureorganization povertyrates,82­88,85f,85t farmgateprices relationshiptodeforestation,inBrazil,95­98 Brazilianbeef,63,63f,191 relationshiptodeforestation,India,99,106n3 anddeforestation,62­64,63f,191 relationshiptodeforestation,Indonesia,100­101, macroeconomicpolicies,effectsof,191­93 102m­103m inremoteforests,84 relationshiptodeforestation,Madagascar,20,104 roads,effectof,68­69,192 relationshiptodeforestation,Mexico,94 See alsoeconomicsofdeforestation relationshiptodeforestation,overview,6f,93­95 fires remoteness,effectof,8,84­88,85f,85t anthropogenicfires,100,130 See alsopoverty,general;individual countries forestfires,125,143­44 ForestresourcesAssessment(FrA),42,44m­45m Indonesia,125 ForestresourcesAssessmentremotesensingsurvey(FrA- repeatedfires(viciouscircles),130­31 rSS),42­43,44,45t,50n1 technologyformonitoring,144­45,204 forestrightsandlanduseregulation FoodandAgricultureorganization(FAo),2,44,154,158,161 allocationofforestrights,11­12,12t,155­56,156t forest­agriculturemosaiclands.Seemosaiclands policyissues,10,137 Forestcarbonfinance privateownershipand,14,155­56,178­79,178t argumentaboutleakageorslippagefromforestprotection, regulatedloggingconcessions,13,167,170­71 201­2 See alsogovernance;tenure;zoning argumentoftemporarynatureofdeforestationabatement, foresttransition 198­201 accelerationof,218­19 argumentthatforestcarbonmakesmitigationtoocheap, avoideddeforestationasabridgetotransition,201 197­98 definition,8,77b exclusionofavoideddeforestationfromprotocol,197,198 factorsinvolvedinforesttransition,77­78,77b,218 limitsongreenhousegasemissions,197­98 locations,77­78,77b,94,98­99,218,219f marketforemissionreductions,198 routestoreforestation,77­78,201,219 monitoringandmeasurementofcarbon,202­3,204 foresttypesanddomains referencelevels(rLs)ofemissions,206 inAfrica,32m,33m See alsocarbonfinance;UnitedNationsFramework inAsia,34m ConventiononClimateChange characteristics,27,30­31,31t forestcores correspondencebetweentypesanddomains,31­34,32t,37b area,39t inLatinAmericaandtheCaribbean,35m,36m definition,37b See alsoareasbeyondtheagriculturalfrontier;biomes; forestandsavannadegradationrates,45t frontieranddisputedareas;mosaiclands maps,32m,34m,35m FrA(ForestresourcesAssessment),42,44m­45m population,39t FrA-rSS(ForestresourcesAssessmentremotesensing See alsoareasbeyondtheagriculturalfrontier survey),42­43,44,45t,50n1 forestedges frontieranddisputedareas area,39t area,39t Brazil,21p arrivaloffrontier,triggersfor,74­75 definition,34,37b nationalroleinbalancinginterests,11,153­54,193 degradation,45t overview,7,30,31t,217 280 I N d e x population,39t ICdPs(integratedconservation-developmentprojects), recommendationsofreport,18b,212b­213b,217­18 163­65,164t trajectoriesoutofthefrontier,74­78,76t,79 incentivesforconservation See alsoforestedges directincentivestolandholders,overview,180­81 financialincentivestoreduceforestCo2emissions,14­16, gabon,69,192 195­96,203,205 geographicinformationsystems,22,145,160 transferableforestprotectionobligations,181­82 geographicpositioningsystems,145,187 See alsocarbonfinance;environmentalservicepayment geographicscopeofreport,28b­29b,28m­29m systems globalenvironmentFacility(geF),16,161,164,183,215 incentivesfordeforestation,7­8,53­54,55­60,79n4. governance See alsoeconomicsofdeforestation;landuse;land agroecologicalzoning,139,158­59 values asymmetriesofpower,information,andorganization,10, incomefromforestproducts,estimationof,88­89 140 India balanceofinterestsbetweengroups,10,139­40,193 biodiversityinWesternghats,56 balancingproductionandenvironmentalservices,137­39, foresttransition,77,98­99 138f JointForestManagementprogram,173­74 certificationprogramsforindustry,148­49 landvalueinWesternghats,56,58t mobilizingdomesticconstituencies,140­44 landvalueofteakforests,54 moneylaunderinglaws,149­50 povertyandilliteracy,relationshiptoforestcover,99, publicdisclosureprogramsforindustry,148 106n3 rights,enforcementof,10 sustainableforestmanagementofteakforests,54 weakinstitutions,improvementof,146­50 treelessforests,93 See alsoforestrightsandlanduseregulation;nationalrole indigenouspeople inbalancinginterests Australia,160­61 greeneragriculture,18b,179­80,213­14,213b,218 disadvantagedstatus,87 greenhousegases(ghgs).Seecarbondioxide(Co2) guatemala,87 greenrevolution,66­67,99 InternationalLabourorganization'sConvention169on guatemala IndigenousandtribalPeoples,166,172 communityforestmanagement,175,177 lowerdeforestationrates,12­13,172 criticalwatersheds,120­21,121f Nicaragua,85,88,172 indigenouspeople,87 inremoteareas,87 povertyrates,120­21,121f Vietnameseuplandandhighlandminoritymembers,87 Indonesia hydrology­forestrelationship biodiversity,2 Costarica,115­16 deforestationbylargecommercialinterests,2­3,44,100 dryseasonflows,118­19 economicsofagriculturevs.sustainablemanagement,60 erosion,116­17,119­20,183,202 fires,125 evapotranspiration,effectoftrees,117­18 frontieranddisputedareas,30 floodsandflowregularity,3b,116,118,122­24,183 IndonesiaForestandMedia(INForM),141 historicalstudies,122 integratedconservation-developmentprojects(ICdPs), landslides,119­20,121,123m,133n2 163­64 localhazardsinwatersheds,120­21 Kalimantan,101,103m modelsandsimulations,122­24 Kerinci-SeblatNationalPark,2­3 sedimentation,20,115­17,119,124­25,183 logging,43,101,147 SouthAfrica'sWorkingforWaterprogram,118 Papuaprovince,160,189 tocantinsriverflow,22,122 poverty,relationshiptodeforestation,100­101,102m­103m tree"fountain"effect,117­18 povertyrates,100,101,102m­103m tree"sponge"effect,117­18 ProPerprogram,148 waterquality,9,120,124­25 protectedareas,deforestationin,163 watershedmanagement,variationwithscale,116­17 publicattitudesoncontrolofforestexploitation,141,142f Yangtzeriver,116,183 regulationofrubbertreesales,91 See alsoenvironmentaleffectsofdeforestation;watersheds roadexpansion,74,189 281 At L o g g e r h e A d S ? rubberagroforestry,91,128,179 Madagascar Sulawesi,100­101,102m biodiversity,19,21 treelessforests,92­93,158 Co2emissionsandmarketsforreductions,21 zoningplans,158,160 deforestationin,19­20,78,104,105m,192­93 INForM(IndonesiaForestandMedia),141 endemicspecies,19 INPe(BrazilianNationalInstituteforSpaceresearch),21,97, forestedges,21p 106n1,144 irrigatedagriculture,20 institutionalandpolicyresponsestodeforestation lemursasseeddispersers,130 overview,6f maizeandricecultivation,19­20,104,192 instrumentalrationaleforconservation,111­12 povertyin,19­20,104,105m integratedconservation-developmentprojects(ICdPs), protectedforestareas,20,92,162 163­65,164t roads,relationshiptopovertyanddeforestation,20,104, InternationalLabourorganization'sConvention169on 193 IndigenousandtribalPeoples,166,172 triggersforexpansionoffrontier,74 internationalroleindevelopmentandenvironmentalissues MaeChaemwatershed(thailand),123­24 financingofforestenvironmentalservices,213­14 mahogany,74,175,176 globalbiodiversityfinance,209,214 maize monitoringandanalysisofenvironmentalinformation, inMadagascar,19­20,104,192 214­15 inMexico,63­64 UnitedNationsFrameworkConventiononClimateChange, inthePhilippines,64 196,197,213 Malawi,90­91 See alsocarbonfinance;KyotoProtocol Malaysia,43,125,157,170 Internationaltropicaltimberorganization(Itto),155 mangroves,64,65 intrinsicrationaleforconservation,111,112­13 Mexico irreversibledegradationofland,78,112,129­31,192,199 communityforestmanagement,174­75 environmentalservicepaymentsystem,183,187,188 Japan,foresttransition,77b,78,218 maize,63­64 relationshipbetweenpovertyanddeforestation,94 KyotoProtocol,191-192 roads,effectsonpoverty,70 sustainableforestmanagement,54 landuse MillenniumecosystemAssessment,109,111,112,125,197 competingintensiveandextensivelanduses,64 ModIS(ModerateresolutionImagingSpectroradiometer) economicframeworkof,overview,7­8 images,144,145,204 landuseplanning,157­58,160­61 moneylaunderinglaws,149­50 vonthünen'smodeloflanduse,71­74,72f,73f,189 mosaiclands See alsocroplandexpansion;forestrightsandlanduse area,39t regulation;incentivesfordeforestation;zoning forestandsavannadegradationrates,45t,51n2 landvalues,56,57t,58t,59­60 greenagriculture,promotionof,18b,213b,218 definition,79n2 overview,6­7,30,31t,218 Indianteakforests,54 population,39t soilsand,62 recommendationsofreport,18b,213b,218 tenure,effectof,73,93 roadconstruction,190 variationinforestedareas,57t­59t tenurein,30,31t Laos,70 LivingStandardsMeasurementSurveys(LSMS),88 nationalroleinbalancinginterests logging areasbeyondthefrontier,153­54,193 auctionsystemforconcessions,147,167,168b­169b,217 balancingbiodiversityconservationagainstproduction, dipterocarptreesanddeforestation,62,92 137­39,138f economicsofsustainablemanagement,55,165­67,170­71 creatingnationalprogramstoreducedeforestation,204­6 extractionratesinIndonesiaandMalaysia,43 forestmanagement,promotionof,11 regulatedloggingconcessions,13,167,170­71 forestrights,alternateassignmentof,11­12,12t,155­56, roleindeforestation,overview,2,43 156t stringentregulations,costsandbenefitsof,170­71 frontieranddisputedareas,11,153­55,193 See alsotimbervalues mosaiclands,11 282 I N d e x nationalforestmonitoringprograms,216 protectedforestareas overview,11­14,153­54,193,215­16 alternativestoaddingnewareas,165­66 recommendationsofreport,17b,212b,215­18 auctionsystemforloggingconcessions,147,167,168b­ regionaldevelopmentinventories,promotionof,11 169b,217 roadconstruction,planningandregulation,14,189­90,192 costofaddingnewareas,165­66 tenureandlanduseregulation,11,12­14 displacementofpopulations,92 See alsoforestrightsandlanduseregulation;governance effectonlocalincomes,163­65 Nepal,68,174 effectsondeforestation,12,162­63 Nicaragua FoodandAgricultureorganizationestimates,161 environmentalservicepaymentsystem,184 forestrevenues,166­67 forestcoverandremoteness,84,86,86f growthandexpansion,161­66,162f indigenouspeople,85,88,172 landarea,155,161 landslides,119­20 leakageorslippagefromforestprotection,201­2 populationdensityandremoteness,84,86f Madagascar,20 povertydensityandremoteness,84­86 regulatedloggingconcessions,13,167,170­71 povertyratesandremoteness,84,85f remotenessofnewprotectedareas,165,166t sustainableforestmanagement,163,165­67,170­71 oilpalm publicdisclosureprogramsforindustry,148 Cameroon,56,58t Puertorico,65,78,130 anddeforestation,56,60,68,75 Indonesia,60,179 recommendationsofreport,17b­18b,212b­213b,215­18 regionalIntegratedSilvopastoralecosystemManagement PanamaCanalwatershed,3 Project(rISeMP),180,184,187 PapuaNewguinea,70,159,175­76,197 regulationofforestuse.Seeforestrightsandlanduse Peru,70,74,158 regulation pestcontrol reservaParticulardoPatrimnioNatural(rPPN),156 biodiversityand,111­12 rivers.Seewatersheds integratedpestmanagement,180,213b,214 roads inmosaiclands,10,18b,132t construction,pressuresfor,74­75 pesticides,180,214 effectonfarmgateprices,68­69,192 Philippines,64,65,93,143,190 effectsondeforestationpressures,14,68­70,188 pollinationandpollinators,10,110,111,132t effectsonexpansionoffrontier,74 population effectsonincome,14 inAfricansavannaandnonsavannaforests,38,40 effectsonmammalpopulations,69 distributionof,37­40,39t,41f,50 effectsonpovertyinMexico,70 forestcores,39t effectsonruraldevelopment,69­70 forestedges,34,38­40,39t importanceforcommunityforestmanagement,88 frontieranddisputedareas,39t inMadagascar,20,104,193 mosaiclands,39t inMexico,70 inremotetropicalforests,2,37 planning,importanceof,14,189­90,192 ruralpeoplelivingnearforests,2 poverty­environmenttradeoffs,188­90 poverty,general transportcosts,74,88,191 croplandexpansioncausedby,1­2 off-farmwages,effectsonpoverty,65­66 savannas,27,28m­29m,32m,36m,45t. See alsocerrado povertydensity,83­88,83m,85t,86f sedimentation,20,115­17,119,124­25,183 povertyrates,82­88,85f,85t shiftingcultivation,40,44,54,64 relationshipofdeforestationandpoverty,overview,6f, shrimpfarming,64,65 93­95 silvopastoralsystems,18b,180,184,213b unreliablegeneralizationsabout,2­3,3b,104 soils See alsoforestpoverty;individual countries compaction,78,118­19,131 predicatecrimes,149 forestdegradationandsoilsuitability,7,46­47,46f,62,130­31 privatelandownership,userestrictionsandregulation,14, andlanduse,72­73,75,157 155­56,178­79,178t andlandvalues,62,186b propertyrights.Seeforestrightsandlanduseregulation waterinfiltrationandsaturation,117,118,122 283 At L o g g e r h e A d S ? SouthAfrica,118,193n1 UnitedStates,foresttransition,77b,78 stateownershipofforests,137,155,173.See alsoprotected U.S.ConservationreserveProgram,184,185,186b,202,209 forestareas sustainableforestmanagement Venezuela,192 discountrates,55,56,176,178 Vietnam,77b,87,94,160,190 economics,54­55,60,164­67,170­71 vonthünen,Johannheinrich,71 Indianteakforests,54 vonthünen'smodeloflanduse,71­74,72f,73f,189 Mexico,54 protectedforestareas,163,165­67,170­71 watersheds shiftingcultivation,40,44,54,64 criticalzones,definition,120 erosion,116­17,119­20,183,202 tanzania,43,175 floods,116,118,122­24,183 technologyforenvironmentalmonitoring historicalstudies,122 Brazil'sruralPropertyenvironmentalLicensingSystem hydrologicalmodelsandsimulations,122­24 (SLAPr),22­23,187­88 importanceofprotecting,9,118,120 carbonemissions,monitoringandmeasurement,202­3, localhazardsin,120­21 204 MaeChaemwatershed(thailand),123­24 CBerS2Sino-Braziliansatellite,145 NewYorkCity,120 Confluenceproject,146 PanamaCanalwatershed,3 fires,144­45 scale,effectsonwatershedmanagement,116­17,131­32 FrA-rSS(ForestresourcesAssessmentremotesensing sedimentation,20,115­17,119,124­25,183 survey),42­43,44,45t,50n1 sensitivity,definition,120 geographicinformationsystems,22,145,160 tocantinsriverflow,22,122 geographicpositioningsystems,145,187 waterquality,9,120,124­25 INPe(BrazilianNationalInstituteforSpaceresearch),21, wateryield,effectoftrees,117­18 97,106n1,144 Yangtzeriver,116,183 internationalroleinmonitoringandanalysisof See alsohydrology­forestrelationship environmentalinformation,214­15 weatherpatterns,effectsofdeforestation,9,125 ModIS(ModerateresolutionImagingSpectroradiometer) WorldBank images,144,145,204 Cameroonforestpolicyand,168b tropicalecosystemenvironmentobservationbySatellite environmentaladjustmentloans,143 (treeS),42­43 fieldguidesforenvironmentaleducation,141 tenure WorldBank'sForestStrategy,4,5b benefitstolandholders,67­68 zoningplansinBrazil,158 effectsonaccesstocredit,93 WorldConservationUnion,47,48f,163 effectsondeforestation,61t,68 World Development Report 2003: Sustainable Development in effectsonlandvalues,73,93 a Dynamic World(WorldBank),10 forestrightsanddifferenttenuresystems,155­56 WorldValuesSurvey,140 inmosaiclands,30,31t WWF(WorldwideFundforNature/WorldWildlifeFund), tragedyofthecommonsand,30,68,89,90­91 28b,124,133n3,215 See alsoforestrightsandlanduseregulation thailand,64,65,93,119,123­24 Yangtzeriver,116,183 timbervalues,3b,7­8,61t,64­65,88.See alsoeconomicsof deforestation;logging zoning tragedyofthecommons,30,68,89,90­91 agroecologicalzoning,139,158­59 transportcosts,effectof,74,88,191 indicativeplanning,157 tropicalecosystemenvironmentobservationbySatellite participatorylanduseplans,159­61 (treeS),42­43 prescriptiveplanning,157­58 problemswithimplementation,158­59 Uganda,59t,60,70,165 systematicconservationplanning,157 UnitedNationsFrameworkConventiononClimateChange, technicalplanning,157 196,197,213. See alsoforestrightsandlanduseregulation See alsoKyotoProtocol 284 E c o - A u d i t Environmental Benefits Statement the World Bank is committed to Saved: preserving endangered forests and ·49trees natural resources. the office of the Publisher has chosen to print · 2,929lbsofsolid At Loggerheads? on 50% recycled waste paperincluding25%post-consumer · 17,702gallonsof waste, in accordance with the rec- water ommended standards for paper · 5,400 lbs of net usagesetbythegreenPressInitia- greenhousegases tive, a nonprofit program support- ing publishers in using fiber that · 34 million BtUs is not sourced from endangered oftotalenergy forests.Formoreinformation,visit www.greenpressinitiative.org.