JANUARY 2022 REPURPOSING AGRICULTURAL POLICIES AND SUPPORT Options to Transform Agriculture and Food Systems to Better Serve the Health of People, Economies, and the Planet i   Madhur Gautam, David Laborde, Abdullah Mamun, Will Martin, Valeria Piñeiro, Rob Vos © The World Bank and IFPRI REPURPOSING AGRICULTURAL 1818 H Street NW, Washington DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved POLICIES This work is a product of the staff of The World Bank and the International Food Policy Research Institute (IFPRI). The Food and Agriculture Organization of the United Nations (FAO) facilitated the funding for this study but does AND SUPPORT not carry responsibility for the findings or views expressed in this report. Likewise, the findings, interpretations, and conclusions expressed in this Options to Transform Agriculture and work do not necessarily reflect the views of the Executive Directors of The World Bank or of the governments they represent, or of IFPRI. The World Food Systems to Better Serve the Health Bank and IFPRI do not guarantee the accuracy of the data included in this of People, Economies, and the Planet work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory, or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank and IFPRI encourage dissemination of their knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Attribution Please cite the work as follows: Gautam, M., Laborde, D., Mamun, A., Martin, Madhur Gautam1, David Laborde2, Abdullah Mamun2, W., Piñeiro, V. and Vos, R. 2022. Repurposing Agricultural Policies and Will Martin2, Valeria Piñeiro2, Rob Vos2 Support: Options to Transform Agriculture and Food Systems to Better Serve the Health of People, Economies, and the Planet © The World Bank and IFPRI.” 1 Agriculture and Food Global Practice, Sustainable Development 2 International Food Policy Research Institute All queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@ worldbank.org. Cover photo: © 2019 Tavarius/Shutterstock Report design: Spaeth Hill JANUARY 2022 i i   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T Appendix E. Political Economy Challenges of TABLE OF CONTENTS Repurposing Agricultural Policies and Support.............................. 79 References........................................................................................................ 82 Foreword..............................................................................................................vi Acknowledgments..........................................................................................viii FIGURES, TABLES, BOXES Overview............................................................................................................... x Figures 1. Introduction......................................................................................................1 Figure O.1:  Growth and Volatility Trends in Food Production Per 2. Background and Context............................................................................9 Capita, 1980–2000...................................................................................................... xii 3. Patterns of Agricultural Support and Emissions .............................. 13 Figure O.2: Impact of Climate Change on Productivity, 1960–2020................... xiii Figure O.3:  Total Annual Support to Agriculture Provided by 79 The Scale and Nature of Agricultural Producer Support.......................... 14 3.1  Countries, 2016–18 (in billions of current dollars and 3.2 The Database for Emissions.................................................................................... 24 percentage share) ..................................................................................................... xiv Figure O.4: Baseline Projections, 2020-2040.................................................................... xvi 4. Simulating Policy Options........................................................................ 29 Figure O.5:  Global Implications of Repurposing Domestic Support The Modeling Framework for Assessing Policy Trade-Offs.................. 30 4.1  (Percentage Change Relative to Baseline Projections for 2040)....... xix 4.2 Continuing With Business As Usual..................................................................... 31 Figure 1.1:  Increase in Food Production Per Capita, Population, and 4.2.1 Scenario 0: Baseline Trends..................................................................................32 Agricultural Land, 1961–2018...................................................................................... 2 Figure 1.2:  Food Production Per Capita: Growth and Volatility Trends, 4.3 Removal of Current Support Measures............................................................ 35 1980–2000 ........................................................................................................................ 3 4.3.1 Scenario 1a: Remove All Domestic Support............................................... 35 Figure 1.3:  Global, Regional, and Country Level Impacts of Anthropogenic Climate Change................................................................................................................. 4 4.3.2 Scenario 1b: Remove Both Domestic Support and Trade Barriers .................................................................................................................41 Figure 1.4:  Impact of Climate Change on Agricultural Productivity, 1960–2020 ......................................................................................................................... 4 4.4 Realign Agricultural Policies and Support for Figure 3.1:  Total Annual Support to Agriculture Provided by 79 Countries, Better Outcomes....................................................................................................................43 2016–18 (in billions of current dollars and percentage share) .........15 4.4.1 Scenario 2: Shift to Less Distorting Forms of Support and Figure 3.2:  Agricultural Support Across Main Countries and Country Lower-Emitting Activities...................................................................................................51 Groupings, 2016–18 ..................................................................................................... 17 Figure 3.3: Nominal Rate of Assistance, by Major Component (%) .......................18 4.4.2 Scenario 3: Condition Support on Environmental Services ...........52 Figure 3.4:  Changes in Levels of GHG Emissions by Main Economic 4.4.3 Scenario 4: Repurpose Support to Target Emission Sector (megatons of CO2eq)............................................................................... 24 Reduction and Productivity Enhancement .......................................................... 54 Figure 3.5:  Shares in Global Gross GHG Emissions by Main 4.4.4 The Impact of Individual Country Actions ............................................... 58 Economic Sector (%)................................................................................................. 25 Figure 4.1:  Baseline Projections of Key Economic and Environmental 5. Avenues for Further Policy Analysis: Implementation ................... 61 Outcomes, 2017–2040 (average annual growth rates 6. Conclusions.................................................................................................. 65 in percent)......................................................................................................................... 33 Figure 4.2: Key Features of Baseline Projections.............................................................. 34 Appendix A. Trends in Support................................................................... 69 Figure 4.3:  Contributions to Growth in Emissions from Agriculture Appendix B. Methods for Deriving the Database, and Agricultural Land-Use Change, 2020-2040 (%)........................... 34 and the Emissions Modeling Framework................................................. 70 Figure 4.4:  Global Implications of Removing All Current Domestic Appendix C. The Modeling Framework..................................................72 Support (Percentage of Change Relative to Baseline Projections for 2040)................................................................................................37 Appendix D. Detailed Simulation Results.................................................75 Figure 4.5:  Impact on GHG Emissions of Removing Different Types of Support........................................................................................................................ 39 i i   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T iii  4. Simulating Policy Options 29 ACRONYMS AND Figure 4.6:  Impact on GHG Emissions and Production of Removing All Support (% change from baseline in 2040).................................................41 ABBREVIATIONS Figure 4.7: G  lobal Implications of Repurposing Domestic Support (Percentage Change Relative to Baseline Projections for 2040)............. 50 Figure 4.8:  Impacts of Country-Specific Repurposing Scenarios (Percentage Change Relative to Baseline Projections for 2040)........... 59 AEZ Agroecological zone IFA International Fertilizer Association Figure A.1: Trends in the Nominal Rate of Protection by Income Level............... 69 Figure B.1:  Creation of GHG Agricultural Emissions Database by Source, ALU Agricultural production and land use IFPRI  International Food Policy Research Institute Location, Commodity, Production Stage, and Technology................... 71 BCA Border carbon adjustment Figure C.1: Linking Emissions to Production in MIRAGRODEP....................................73 IMF International Monetary Fund CAP  Common agricultural policy (European Union) IPCC  Intergovernmental Panel on Climate Change (United Nations) Tables CES Constant elasticity of substitution IO  International Organizations Table 3.1: P  ositive Protection Rates for Key Commodities by Region and CET Constant elasticity of transformation Consortium Country, 2016–18 (%)......................................................................................................19 CFS Committee on World Food Security LES Linear expenditure system Table 3.2: N  egative Protection Rate for Key Commodities by Region and Country, 2016–18 (%)....................................................................................................22 CGE Computable general equilibrium LUC Land-use change Table 3.3: D  istribution of Domestic Support by Product, Instrument, and CO2eq Carbon dioxide (CO2) equivalent MAFAP  Monitoring and Analyzing Food Country Grouping, 2016–2018 (percentage shares)............................... 23 COP26  Twenty-Sixth Conference of the and Agriculture Policies (FAO) Table 3.4: R  ate of Domestic Support by Instrument, Sector, and Country Parties of the UNFCCC (2021) MToE Million tons of energy use Grouping, 2016–18 (%)................................................................................................ 23 CoSAI  Commission on Sustainable NDC  Nationally determined Table 3.5: S  hares of Emissions from Agricultural Production by Agriculture Intensification contributions (UNFCCC) Commodity and Source, 2017 (%)...................................................................... 26 CSA Climate-smart agriculture NRA Nominal rate of assistance Table 3.6: E  mission Intensities for Key Products and Regions, 2017 (kg CO2eq/kg of product)......................................................................................... 26 EC Emission coefficient NRP Nominal rate of protection Table 4.1: Scenarios Considered................................................................................................. 49 EMDE  Emerging market and developing OECD  Organisation for Economic economies Co-operation and Development Table 4.2: Impact of Labor Mobility on Real Farm Wages........................................... 56 Table D.1:  Global Impacts of Removing Components of Agricultural FAO  Food and Agriculture Organization of PPP Purchasing power parity Support (% change in each indicator by 2040 with the United Nations R&D Research and Development respect to baseline).......................................................................................................75 FAOSTAT  Food and Agriculture Organization TFP Total factor productivity Table D.2: R  esults by Selected Countries for a Scenario of Corporate Statistical Database Abolition of All Subsidies (% change by 2040 in each UNCAS UN Climate Action Summit FOLU Food and Land Use Coalition indicator with respect to the baseline)............................................................76 UNDESA  United Nations Department of FSIN Food Security Information Network Economic and Social Affairs Table D.3: G  lobal Impacts of Repurposing Simulations (% change in each indicator with respect to baseline).........................77 GDP Gross domestic product UNDP  United Nations Development Table D.4: I mpacts of Country-Specific Repurposing Scenarios: GFR Gross farm receipts Programme Productivity-Enhancing and Emission-Reducing Farm UNEP  United Nations Environment GHG Greenhouse gas Practices in Individual Countries (% change in each Programme indicator by 2040 with respect to baseline) ..............................................78 GI Green innovation UNFCCC  United Nations Framework GNI Gross national income Convention on Climate Change Boxes Gt Gigatons UNFSS  United Nations Food Systems Box O.1: Methodology.......................................................................................................................... xv HLPE  High Level Panel of Experts Summit Box 4.1: Baseline Simulations......................................................................................................... 32 (of the CFS) WDI World Development Indicators IDB Inter-American Development Bank WTO World Trade Organization Note: $ refers to US dollars. i v   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T v  Acronyms and Abbreviations FOREWORD At a time when farmers bear the brunt of worsening climate change impacts, volatile food prices, rising input costs, and shifting consumer demand, government support is much needed and could be much better targeted. The report uncovers that for every budgetary dollar spent under Providing nutritious and affordable food for a growing global population current farm policies, only 35 cents end up with farmers. In rethinking while protecting the vital natural systems that sustain life is one of the agricultural policies, governments must be mindful of farmers’ bottom lines critical challenges of our times. Current agricultural practices have yielded and the particularities of country and even local contexts. Indeed, farmers’ impressive productivity gains, but are increasingly associated with high support for policy changes, incremental or otherwise, will be key to the greenhouse gas emissions, biodiversity loss, and chronic disease, while success of reform efforts. leaving many rural people who depend on farming in poverty. We hope readers will find that this report makes a useful contribution How can agricultural support policies be repurposed to make the food to a growing literature on how to repurpose current agricultural policies system deliver better outcomes? This was the broad question the World and drive reform, as the World Bank and IFPRI, together with other Bank and the International Food Policy Research Institute (IFPRI) sought to partners, including FAO, work with policymakers to reexamine their answer in this study. The report finds that there are important current and support programs and chart ways forward for food systems that better projected trade-offs for policymakers to consider as they work to deliver benefit people, the planet, and the world’s economies on the promise of food systems for sustainable development. All solutions are not equal when it comes to rethinking agricultural public policies and support. The report finds that greenhouse gas emissions would increase substantially in the future if current policies are untouched. Simply rearranging or even removing current support would not bring about the changes needed for sustainability. Nor would applying environmental conditionality to the support provided while relying solely on currently available technologies: While it could help reduce emissions Martien van Nieuwkoop Johan Swinnen in the short term, lower yields could induce farmers to expand land use Global Director, Agriculture and Director General, IFPRI and Global Director for agricultural production. Both changes in incentives and investments in Food Global Practice, World Bank for Systems Transformation, CGIAR innovations that simultaneously pursue productivity enhancements and greenhouse gas emission reductions are needed in order to deliver broad and long-standing wins. The report finds that repurposing a portion of government spending on agriculture each year to develop and disseminate more emission-efficient technologies for crops and livestock could reduce overall emissions from agriculture by more than 40 percent. Meanwhile, millions of hectares of land could be restored to natural habitats. The economic payoffs to this type of repurposing would be large. Redirecting about $70 billion a year, equivalent to one percent of global agricultural output, would yield a net benefit of over $2 trillion in 20 years. Most importantly, repurposing would deliver large benefits to people. It would raise rural incomes, contributing to improved food security. It would substantially reduce the cost of healthy diets, contributing to better nutritional outcomes. And it would accelerate poverty reduction. v i   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T vii  Foreword ACKNOWLEDGMENTS This study was undertaken by a joint World Bank and International Food Policy Research Institute team comprised of Madhur Gautam, David Laborde, Abdullah Mamun, Will Martin, Valeria Piñeiro, and Rob Vos. Special thanks to Martien Van Nieuwkoop(Global Director), Louise Scura (former Practice Manager for Global Engagements) and Julian Lampietti (current Practice Manager for Global Engagements) at the Agriculture and Food Global Practice of the World Bank for their guidance and support from the conceptualization to the implementation of this study. This study contributes to the aims and objectives of Food Systems 2030, the World Bank’s vision for transforming food systems to deliver healthier people, a healthy planet, and healthy economies. The Food and Agriculture Organization of the United Nations (FAO) gracefully facilitated World Bank-sourced funding that made this study possible. The authors are also grateful to FAO staff, Marco V. Sánchez and Valentina Pernechele, for helpful comments to an early draft of the study. The authors further gratefully acknowledge the valuable inputs and feedback received at various stages of this study from Hanane Ahmed, Tobias Bedaeker, Eva Tortella Canellas, Raffaello Cervigni, Richard Damania, Ghada Elabed, Marianne Fay, Joshua Gill, Charlotte Hebebrand, Sebastian Heinz, Christine Heumesser, Stephen Ling, Michael Morris, Clare Jessica Murphy- McGreevey, Flore Martinant de Preneuf, Robert Townsend, Michael Toman, Dina Umali-Deininger, and Sergiy Zorya. The authors also thank the participants at the following seminars, workshops, and conferences for helpful comments on earlier drafts of this paper: an IFPRI policy seminar (April 21, 2020); the National Bureau of Economic Research (NBER) Spring 2020 Conference on “Agricultural Markets and Trade Policy” (April 30, 2020); the 23rd Global Trade Analysis Project (GTAP) Annual Conference on Global Economic Analysis (June 17, 2020); the UN Food Systems Summit Pre-Summit Parallel Event on “Repurposing Public Support to Food and Agriculture: A Just Rural Transition to Sustainable Food Systems” (July 26, 2021); the UN Food Systems Pre-Summit affiliated session “Rebalancing Public Agricultural Support for Health, for Prosperity and for the Planet – Policy Realities” (July 28, 2021); the Agricultural and Applied Economics Association Annual Meetings (August 3, 2021); the 31st International Conference of Agricultural Economists (August 28, 2021); the 49th Session of the Committee on World Food Security (CFS) side event on “Prioritizing Climate Resilience: Building a New Policy Consensus with Smallholder Farmers” (October 11, 2021); the Global Landscapes Forum (November 7, 2021); the 10th Asian Society of Agricultural Economists International Conference (December 7, 2021); and the Academy of Global Food Economics and Policy (December 9, 2021). The publication and editorial services for this report were provided by the Translation and Interpretation Unit in the Global Corporate Solutions Department (GCSTI) of the World Bank. v i i i   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T ix  Foreword KEY MESSAGES • Current governmental support for agriculture provides incentives for unsustainable patterns of production and consumption, with agriculture and land-use change responsible for 22 percent of global greenhouse gas emissions (GHG).  iven a “business-as-usual” scenario of unchanged support, GHG o G emissions from agriculture would increase by 58 percent, and 56 million hectares would be converted to agricultural land between now and 2040. • Current support for agriculture delivers low value for money as a way of helping farmers; for every dollar of public support, the return to farmers is just 35 cents. • Simple reductions in or rearrangement of current support will not yield game-changing reductions in global emissions. • Policy conditionality tying support to the adoption of environment- friendly but lower-yielding farm practices could potentially reduce emissions, but would entail tradeoffs for people, nature, and economic prosperity with lower agricultural production, higher poverty, higher agricultural land use and an increase in the cost of healthy diets. • Concerted efforts to repurpose a part of current domestic support as incentives to develop and adopt green innovations that reduce both emissions and costs could potentially deliver substantial gains for the planet, the economy, and people. o Simulation results suggest that investments in innovations designed OVERVIEW to lower emissions and raise productivity by 30 percent could reduce emissions from agriculture and land use by more than 40 percent, returning 105 million hectares of agricultural land to natural habitats, while delivering substantial gains in poverty reduction, nutrition, and the overall economy. • There is a strong case for policymakers to scrutinize and rethink their current domestic policies. The biggest gains would accrue through a coordinated effort of all countries to reset their policies to address the global threat of climate change, and to better meet nutritional and social needs. Securing affordable access to a healthy, nutritious, and safe diet for the growing world population in the face of climate change and wide- spread resource degradation is a major global challenge. Demand for food is expected to increase rapidly between now and 2050. The world’s population is projected to reach almost 10 billion by 2050, and per capita R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T xi  Overview incomes are rising rapidly. Agricultural performance in meeting the chal- FIGURE O.2: Impact of Climate Change on Productivity, 1960–2020 lenge of feeding the world over the past 60 years has been impressive, as food production has substantially outpaced population growth. However, continuing to meet global food needs successfully and sustainably is becoming increasingly difficult. Global hunger has been on the rise since 2015, and the growth in food output per capita has both decelerated and become more volatile (Figure O.1). FIGURE O.1: Growth and Volatility Trends in Food Production Per Capita, 1980–2000 1.6 1.8 Standard deviation in annual growth, percant 1.4 1.6 1.2 Trend growth, percent 1 1.4 0.8 1.2 0.6 Source: Ortiz-Bobea et al. 2021. 0.4 1 0.2 Building better food systems requires a fundamental change in incen- 0.8 tives. This study finds that if countries continue on a “business-as-usual” 0 path by keeping current policies in place, emissions from agricultural 1981-90 1982-91 1983-92 1984-93 1985-94 1986-95 1987-96 1988-97 1989-98 1990-99 1991-00 1992-01 1993-02 1994-03 1995-04 1996-05 1997-06 1998-07 1999-08 2000-09 2001-10 2002-11 2003-12 2004-13 2005-14 2006-15 2007-16 2008-17 2009-18 2010-19 -0.2 0.6 production would double by 2040, and an additional 56 million hectares Growth (10 year trend growth rate, left axis) of new land would be converted to agriculture between 2020 and 2040. Volatility (10 year Std. Dev. of annual growth, right axis) These outcomes reflect the patterns of production and consumption that have emerged, influenced in part by incentives created through longstand- Source: FAOSTAT ing governmental measures taken to support agriculture. In 2016–18, the Climate change is not a distant threat—it is already adversely affecting governments of the 79 countries for which data are available supported agriculture. Recent analysis indicates that since 1960 climate change has agricultural production and food consumption with measures that gen- slowed productivity growth by 21 percent globally, and by as much as 40 erated net transfers of $638 billion per year (Figure O.3). More than 70 percent in parts of Africa and other tropical zones. More worryingly, as percent of this total support, about $456 billion, consisted of support for shown in Figure O.2, this adverse impact appears to be intensifying, push- agricultural producers, of which 82 percent was provided through mea- ing the world more quickly toward a “tipping point” where climate change sures that the Organisation for Economic Co-operation and Development impacts will offset all productivity growth, and beyond which the economic (OECD) refers to as “potentially most distorting.” These include subsidies and social consequences could be devastating. linked to outputs, inputs, or production factors like land area (referred to as domestic support in this study) as well as market price supports provided While agriculture is highly vulnerable to climate change, it is also a major through trade restrictions such as import tariffs and other border mea- contributor to the problem. The agri-food system contributes about a sures (referred to as trade barriers in this study). About 11 percent of the third of the world’s total anthropogenic GHG emissions. About two-thirds total support was provided to poor consumers, for instance through public of these, or about 22 percent of the total, are generated on farms, from food assistance or food distribution programs. Of the remainder, about agricultural production and land-use change; the rest come from pre- and 17 percent was for public goods and services like research and irrigation, post-production activities in the broader agri-food system. Agriculture and and another 5 percent was “green” subsidies, that is, subsidies to support food systems also generate other major negative externalities, including the better environmental outcomes. Governments have been providing these loss of biodiversity, the degradation of natural resources, and the adverse broad types and levels of support to agriculture and food systems for effects on human health of costly nutrition-adequate diets. x i i   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T xiii  Overview a long time. This public support has helped to raise productivity and lower 4. Repurposing: In this scenario, a portion of current domestic support the price of food, especially of basic staples such as cereals; but it has also would be repurposed for increased spending on green innovations; promoted the unsustainable patterns of production and unhealthy diets that that is, the development, diffusion, and adoption of new technologies characterize today’s food systems. that both reduce emissions and raise productivity. The remainder would be returned to taxpayers and would be potentially available FIGURE O.3: Total Annual Support to Agriculture Provided by 79 Countries, 2016–18 to deliver as nondistorting transfers to producers and other stake- (in billions of current dollars and percentage share) holders. This could be used to compensate them for potential losses due to this reform, and to spend on rural infrastructure and other Input subsidies Output Subsidies $86.3b (14%) essential public goods and services that are fostering agricultural and $73.3b (11%) rural development. Public Goods and Services, $108.2b (17%) BOX O.1: METHODOLOGY Other Green Subsidy, Using the International Food Policy Research Institute’s (IFPRI’s) $207.5b (33%) $28.9b (5%) global general equilibrium model, MIRAGRODEP, this study analyzes Consumer the likely impacts of several different policy options on the planet Market Price Support $211.7b (33%) Support, (that is, on GHG emissions and land use); the economy (national $70.4b (11%) income); and people (poverty, food security, and the cost of a healthy diet). These scenarios assess the potential effects of Decoupled transfers $59.6b (9%) removing, restructuring, attaching conditionality to, and/or repurposing current domestic producer support. Source: Authors, using data from AgIncentives International Organizations Consortium. Note: b=billion Our analysis assumes a phased implementation of reforms and Could the current support to producers be repurposed to deliver better focuses on longer-term outcomes rather than immediate impacts. outcomes? Given the scale and structure of the support to agricultural In all of the scenarios, reforms are assumed to be implemented producers globally, this study assesses several options for repurposing current gradually over a five-year period (2020–2025), and impacts agricultural policies and support to achieve better economic, environmental, measured against a projected baseline for 2020–2040. This would social, nutritional, and climate outcomes. The scenarios analyzed are: allow the investment and consumption responses to changes in income resulting from the reforms to be fully incorporated when 0. Baseline: A business-as-usual (or “zero”) scenario simulates a “no policy considering outcomes. change” option that assumes current policies and patterns of producer support will continue unchanged. 1. Removal: Two scenarios consider the removal of two distinct forms of producer support: 0. Baseline. A “business-as-usual” (or zero) scenario with unchanged a. Remove the current domestic support provided to producers. policies projects a substantial increase in agricultural emissions by b. Remove both domestic support and trade barriers or market price supports. 2040. Figure O.4 shows the projections for key outcomes. From 2020 to 2040, in line with past trends, agricultural value added would increase 2. Restructuring: Two forms of restructuring domestic support that would by about 3 percent per year, and emissions from agricultural production rely on currently available technologies and practices are analyzed: would double. In this business-as-usual scenario, agricultural land use a. Replace the current pattern of support, which targets certain agricultural is projected to increase by 1 percent, equivalent to drawing 56 million products, with a uniform rate of support for all agricultural products. hectares of new land into agriculture from 2020–2040. This expansion b. Target current domestic support to only low-emission intensity products. of agricultural land would increase losses in biodiversity and ecosystem 3. Conditionality: In this scenario, producer support would be conditional services; increase emissions as a result of forest conversion to farmland; on farmers adopting emission-reducing practices, using currently avail- and reduce carbon sequestration capacity by 7 percent. able technologies. x i v   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T xv  Overview FIGURE O.4: Baseline Projections, 2020-2040 • This scenario also reveals that the vast public resources spent to benefit farmers is delivering very little “value for money.” Agri Value Added ($Trillion) Poverty (PPP$1.90, %) Domestic support to producers costs around 14 percent of agri- 7.36 cultural value added but yields an increase in real value added of 8.20 5.47 only 5 percent. If farm support is thought of as providing transfers 3.92 to farmers, its implied transfer efficiency is very low, at only about 7.22 7.15 35 percent. In contrast, lump-sum transfers (that is, payments to 2020 2030 2040 2020 2030 2040 producers that are not linked to inputs or outputs) would almost Agric Land (Bill. ha) Net Agric. Emmissions, Gt triple the gains to farmers, while avoiding the distortions created by Agric Production LULUC current forms of support. 4.87 • Removing trade barriers as well as domestic support would 4.81 4.84 7.24 9.12 yield somewhat greater income gains but would limit the 5.76 reduction in emissions. Trade barriers in the form of import tariffs 2020 2030 2040 -2.29 -2.18 -2.13 support production but tax consumption in protecting countries. 2020 2030 2040 Their removal (Scenario 1b) would thus have partially offsetting Source: Authors’ baseline scenario projections. effects on supply and demand. Economic efficiency gains would be larger if both trade barriers and domestic support were reduced 1. Removal: What is current agricultural support “buying”? This question (which would be about $135 billion, or 0.09 percent in 2040), and is addressed by the first set of complementary scenarios (1a and 1b), which global poverty would fall slightly. With a more muted decline in assume the removal of domestic support and of all producer support, global agricultural output as compared to removing only direct including market price support (Figure O.5). support, however, this more comprehensive reform would limit • A simple removal of domestic producer support would involve the reduction in global GHG emissions induced by the removal of important trade-offs. Removing domestic support (Scenario 1a) domestic support to about 39 megatons of CO2eq, or 0.55 percent would have small but favorable impacts on the climate and on of total agricultural emissions in the baseline. This muted impact nature by reducing agricultural GHG emissions by the equivalent is explained in part by the effect of removing protection on food of about 103 megatons of CO2 (CO2eq), or 1.5 percent of total prices, which would fall in protecting countries, thereby increasing agricultural emissions in the baseline, as well as reducing the global demand for food and offsetting some of the decline in global territorial footprint of agriculture, saving 27 million hectares, or production from the removal of domestic support. about 49 percent of the projected conversion of land to agricul- ture. However, these environmental gains are far short of what is Approaches that specifically aim to reduce emissions can be game needed to appreciably curb agriculture’s contribution to climate changers for agriculture’s impact on climate change; but they require change. Moreover, the economic outcomes would be mixed. On careful consideration of current and projected trade-offs. The options the one hand, removing distortionary domestic support would for maintaining but redirecting domestic support to agriculture considered generate some efficiency gains, reflected in a small increase in real in this study are representative of a broad range of specific policy options world income of $74 billion (0.05 percent) per year relative to the that are conceptually similar but that need to be tailored to individual baseline projections for 2040. On the other hand, major political country contexts. The impacts of selected repurposing options are shown economy challenges would be likely to emerge as farm output in Figure O.5 and compared with those of the previous scenarios involving and real farm income per worker would decline, reinforcing policy- the removal of current supports. These scenarios assume an international makers’ concerns about food security and the welfare of farmers. consensus, under which all governments would repurpose support toward The current farm-support regimes were not designed to reduce common global objectives. poverty or to improve diets, but their abolition would likely increase food prices, contributing to more poverty (albeit marginally) and raising the cost of healthy diets. x v i   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T xvii  Overview 2. Restructuring. Maintaining support for agriculture at the current FIGURE O.5: G  lobal Implications of Repurposing Domestic Support (Percentage Change Relative to Baseline Projections for 2040) levels but restructuring it either by moving to uniform rates of assis- tance for all products, or by favoring low-emission products would ECONOMIC FARM SECTOR yield surprisingly small economic, social, and environmental gains. (Real national income, % change 2040) (Agricultural production volume, % change 2040) Replacing the current highly variable system of agricultural support with a Removal of Dom. Support uniform rate (Scenario 2a) mimics a shift toward decoupled transfers and Removal of all support would remove the present bias toward certain products. However, moving Uniform subsidy support away from high-emission to low-emission intensity products Target CO2 efficient products (Scenario 2b) would have surprisingly little impact on emissions. Para- Env. Conditionality CROPS LIVESTOCK doxically, transferring all subsidies to low-emission crop cultivation would Repurposing for GI actually increase global emissions by increasing demand for cropland and -3.2 -2.2 -1.2 -0.2 0.8 1.8 -20.0 -10.0 0.0 10.0 stimulating land-use change from forests, even though some pastureland would be retired as livestock production fell. These outcomes suggest that SOCIAL DIETS while this scenario is appealing at face value, merely shifting subsidies (Poverty at PPP$1.90, % change 2040) (Healthy food prices, % change 2040) away from emissions-intensive commodities would do little in terms of Removal of Dom. Support overall emission reduction. Removal of all support 3. Conditionality. Making support “conditional” on reducing emissions Uniform subsidy would be positive for planetary health but could entail trade-offs for Target CO2 efficient products people and economic prosperity. Promotion of production methods and Env. Conditionality practices that improve environmental outcomes but reduce the produc- Repurposing for GI tivity of land (Scenario 3) could potentially deliver important reductions -1.3 -1.1 -0.9 -0.7 -0.5-0.3 -0.1 0.1 0.3 0.5 0.7 -20 -15 -10 -5 0 5 10 in GHG emissions; but it might also come with economic and social costs. CLIMATE NATURE Drawing on the literature on emission reductions and cost increases (Reduction in emissions from agriculture and (Agricultural land, % change 2040) associated with existing policy proposals for this type of conditionality, land use, % change 2040) an illustrative simulation makes farm support conditional on production Removal of Dom. Support techniques that reduce emission intensities by 10 percent, while raising Removal of all support costs by the same amount. This would reduce global GHG emissions from Uniform subsidy agricultural production by 19 percent through the reduction in emissions Target CO2 efficient products per unit of output, and a decline in global output. But this gain would be Env. Conditionality offset by increases in emissions from land-use change, because additional Repurposing for GI land would need to be brought into agriculture. The net reduction in emis- -40 -30 -20 -10 0 10 20 -2.5 -1.5 -0.5 0.5 1.5 sions from agriculture and land-use change would be 15 percent. This gain would come at cost of a 0.8 percent decline in global income, and a drop Source: Authors, using model simulation results. Note: Brown bars indicate movement toward, and teal bars indicate movement away from of more than 5 percent in agricultural production, while poverty and the achieving the related SDG(s). GI= Green Innovation. cost of a healthy diet would both increase. Decreased biodiversity would incur additional losses since an increase in the use of land for agriculture 4. Repurposing for green innovation. The repurposing option, which would result in the loss of forest habitat. would redirect a part of domestic support toward targeted investments in technologies that are both productivity-enhancing and emis- sions-reducing, appears to hold the potential to deliver “triple wins” for a healthy planet, economy, and people. The key point of departure in the final option considered (Scenario 4) is the focus on green innovation; x v i i i   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T xix  Overview that is, technologies and practices that would reduce emissions while • Productivity-driven growth reduces poverty and makes nutri- increasing productivity. Recognizing that achieving this is not without tionally adequate diets more affordable. In this scenario, global cost, the focus of this scenario is on redirecting some of the domestic extreme poverty would fall by 1 percent, while the cost of a healthy support currently provided to agriculture toward more public spending diet would drop by a substantial 18 percent. on research and development (R&D), and incentives for the development • Incomes of farm workers would increase, while farm employ- and adoption of green innovations. Some such innovations already exist or ment would fall as part of structural economic transformation are emerging. Based on an examination of the literature on the potential of over the long term—between now and 2040. The repurposing of recent innovations to raise productivity and reduce agricultural emissions, current agricultural support could facilitate farm labor moving into this illustrative scenario assumes a 30 percent increase in production other parts of the economy, because some of this money could and a 30 percent reduction in emissions per unit of output. The literature be spent instead on human capital and skills development, as well on past agricultural productivity growth suggests that the cost of raising as on rural financing and infrastructure. Through structural trans- agricultural productivity by 30 percent on a sustainable basis would be formation, farm labor could become more productive both within roughly equivalent to one percent of the value of farm output. This sce- agriculture and in nonfarm work if governments invested more in the human capital of rural people. nario considers repurposing the equivalent of one percent of the value of farm output from the current domestic support for agriculture to invest in Notwithstanding the substantial potential gains for people, the planet, R&D, under the assumption that with reoriented R&D priorities, this level of and the economy that could result from the repurposing options research intensity would also apply to the generation of green innovations. discussed in this study, current agricultural support measures need to The remaining domestic support would amount to a saving for taxpayers be carefully scrutinized in various country contexts. A key insight from and would be potentially available to deliver as nondistorting transfers to this study is that current agricultural support is a very blunt instrument for producers and other stakeholders to compensate them for any losses they fighting climate change and for addressing the challenges of global food might incur due to this reform, and for spending on extension services, security and nutrition. There appears to be great potential for achieving rural infrastructure, and other essential public goods and services that major gains on these fronts by repurposing support toward public invest- are fostering agricultural and rural development. The importance of green ments that facilitate the widespread adoption of productivity-enhancing innovations in delivering these wins is clear from Figure O.5, which shows and emission-reducing technologies for agri-food systems. Further, these the results of the repurposing scenario. policies are likely to have strongly positive international spillovers. Innova- tions that reduce environmental impacts and raise productivity are likely • Global real income would be higher, reflecting large economic efficiency gains. In 2040, the projected world income would be 1.6 to either be rapidly adapted in other countries, or to provide a basis for percent higher than the business-as-usual projection. developing technologies for other agroecological environments. • Adoption of these improved technologies would deliver huge Nevertheless, even the best-designed policy reforms will face political benefits for the climate and nature. Between 2020 and 2040, hurdles. Agricultural support policies are the prerogative of national overall emissions from agriculture would fall by more than 40 governments. Overcoming national resistance to agricultural policy reform percent, or nearly 2.8 Gt CO2eq—avoiding nearly 80 percent of from affected stakeholders will be a huge challenge. National farm and the incremental emissions expected under the baseline (busi- agricultural policies have a long history in most countries and have devel- ness-as-usual) scenario. Productivity growth would also release oped well-established entitlements and vested interests. Recognition of production factors (for a given level of demand), including land. the major private and societal gains to be achieved, and multistakeholder About 2.2 percent less agricultural land would be needed in this engagement to discuss the potential trade-offs associated with policy scenario, releasing about 105 million hectares of agricultural land options and to devise acceptable strategies should help to earn political for restoration to natural habitats, with potentially substantial support for smart repurposing of existing support at the national level. biodiversity benefits. This approach would spare not only the additional 56 million hectares of land that would be transferred to For reforms to foster sustainable global development, effective policy agriculture between 2020 and 2040 under the baseline scenario coordination and technological innovations that are attractive to but would also release another 48 million hectares currently being both individual producers and governments are needed. At present, used for agriculture that could be restored as natural habitats. x x   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T xxi  Overview agricultural support is distributed unevenly across nations. Poorer nations have less fiscal space with which to provide agricultural support. Also, their national agricultural research systems generally have weaker resource capacity for developing high-productivity and sustainable farm technol- ogies and practices relevant to the local context, and their farmers and other food producers face bigger obstacles in adapting those practices. Hence, to be most effective at the global level, a more even-handed diffusion of both technologies and financial resources is needed in order to allow countries to reap the benefits of agricultural policy reform and contribute most effectively to solving global challenges. International coordination is vitally important to achieve the needed reductions in global emissions from agriculture. Climate change and environmental sustainability are global challenges that transcend borders, and national policies have strong international spillover effects. Policy- makers are well-placed to scrutinize and rethink domestic policies – but ultimately all countries need to act together to effectively address the global threat of climate change to our food systems. x x i i   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T xxiii  Overview 1 Remarkable progress has been made in increasing global food produc- tion over the past 60 years. The increase in the production of calorie-rich staples helped ward off the widespread hunger and famine forecast across much of the developing world in the 1960s and 1970s (Fuglie et al. 2020). Driven by an overriding focus on food availability, intense public support has helped cereal production more than triple since the early 1960s, out- pacing population growth, which has increased about two and a half times since 1961 (Figure 1.1). Much of this increase is credited to productivity growth, with expanded area under agriculture contributing relatively less. Despite this progress, the world is off-track for meeting its Sustainable Development Goal (SDG2) targets, with hunger and food insecurity on the rise since 2014 (FAO et al. 2021).1 FIGURE 1.1:  Increase in Food Production Per Capita, Population, and Agricultural Land, 1961–2018 Index: 1961=1 Food Production: 3.66 Population: 2.49 Land: 1.07 1961 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015 2017 2019 INTRODUCTION Source: FAOSTAT Looking forward, the world faces an even more daunting and complex task than it did in the 1960s: that is, ensuring that a growing population has access to affordable, healthy, and safe food in the face of climate change and the rapid degradation of natural resources. The world’s population is projected to reach 9.7 billion by 2050, which together with rapid urbanization and rising incomes will increase the demand for food, especially for more resource-intensive animal-source foods, and fruits and vegetables. Efforts to successfully and sustainably meet this challenge are already encountering stiff headwinds. Growth in food production 1  FAO et al. (2021) estimates that as many as 811 million people faced hunger and nearly 2.4 billion people (or one in three people worldwide) were without access to adequate food in 2020, with the COVID-19 pandemic worsening trends that had been already deteriorating since 2014. Healthy diets were out of reach for 3 billion people in 2020 due to the high cost of healthy foods. Over 149 million children under 5 are estimated to be stunted, 45 million suffer from wasting, and 39 million are overweight. R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 2  Introduction per capita appears to have been slowing since about 2010. At the same FIGURE 1.3:  Global, Regional, and Country Level Impacts of Anthropogenic Climate Change time, the long-term decline in the volatility of food production per capita appears to have reversed and has risen since the mid-2000s (Figure 1.2). FIGURE 1.2: Food Production Per Capita: Growth and Volatility Trends, 1980–2000 IMPACT (%): 1.6 1.8 >0 -5 TO 0 Standard deviation in annual growth, percant 1.4 -10 TO -5 1.6 -15 TO -10 1.2 -20 TO -15 Trend growth, percent -25 TO -20 1 -30 TO -25 1.4 -35 TO -30 0.8 -40 TO -35 <-40 1.2 N/A 0.6 0.4 1 0.2 0.8 0 Source: Ortiz-Bobea et al. (2021).  This map was produced by the Cartography Unit of the World 1981-90 1982-91 1983-92 1984-93 1985-94 1986-95 1987-96 1988-97 1989-98 1990-99 1991-00 1992-01 1993-02 1994-03 1995-04 1996-05 1997-06 1998-07 1999-08 2000-09 2001-10 2002-11 2003-12 2004-13 2005-14 2006-15 2007-16 2008-17 2009-18 2010-19 -0.2 Bank Group. The boundaries, colors, denominations and any other information shown on this map do 0.6 not imply, on the part of the World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. Growth (10 year trend growth rate, left axis) Volatility (10 year Std. Dev. of annual growth, right axis) The impacts of climate change are also accelerating (Figure 1.4). If it is Source: FAOSTAT unchecked, it will be increasingly difficult to maintain the impressive rate of productivity growth seen over the past 60 years. Positive technological gains Climate change is no longer a distant threat; it is already adversely could be overwhelmed by worsening climate change, leading to an expansion affecting agriculture. The trends shown in Figure 1.2 are indicative of this, in the amount of agricultural area needed to feed the growing population, and and a rigorous new study establishes a concrete link between climate pushing the world more quickly toward an eventual “tipping point,” with poten- change and productivity (Ortiz-Bobea et al. 2021). This study estimates tially enormous economic and social consequences (Johnson et al. 2021). that climate change has reduced global agricultural productivity growth, as measured by the growth in total factor productivity (TFP), by 21 percent FIGURE 1.4: Impact of Climate Change on Agricultural Productivity, 1960–2020 since 1961. This is the equivalent of wiping out seven years of productivity gains globally; in other words, neutralizing technologically driven productiv- ity growth since 2013 (Ortiz-Bobea et al. 2021). Importantly, the impacts are more severe in tropical agriculture, with productivity growth falling by as much as 40 percent or more in parts of Africa and other areas (Figure 1.3). These impacts are already being felt through rising levels of hunger (FAO et al. 2021), and acute food insecurity across large swathes of Africa, Central America, and parts of South Asia and the Middle East that overlap with the areas most affected by climate change (FSIN 2021). Source: Ortiz-Bobea et al. 2021. 3   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 4  Introduction While agriculture is highly vulnerable to climate change, the way that in productivity within agriculture, and a shifting of labor out of agriculture much of it is practiced now is also a major contributor to climate and into other sectors—as well as progress on poverty reduction.2 change, and to degradation of the natural resource base on which it Incentives are key to viable solutions to the enormous challenge of relies. The current agri-food system is associated with substantial “hidden making agriculture more productive, sustainable, and nutrition-sensi- costs” that are becoming increasingly apparent (WRI 2019; FOLU 2019). It tive. A key question to consider is whether current agricultural policies are contributes about a third of total anthropogenic greenhouse gas (GHG) creating incentives that will help producers make appropriate decisions emissions (Tubiello et al. 2021; Crippa et al. 2021; IPCC 2020). Clark et for achieving the desired goals. The amount of support currently provided al. (2020) conclude that reducing GHG emissions from agriculture will to agriculture by governments around the world is substantial—the 79 be essential to meeting the IPCC targets of holding global temperature countries for which data are available provided net transfers of $638 billion increases to 1.5° or 2°C. Agricultural production and additional land being annually between 2016 and 2018, through a combination of explicit trans- brought into agricultural production have an outsized environmental fers funded by public expenditures and implicit transfers through policies footprint—they accounted for 22 percent of the total emissions in 2018— that alter the prices producers receive for their products.3 that is, two-thirds of agri-food emissions, with the remaining coming from pre- and post-production activities--but only around 4 percent of global This support is beneficial to farmers, at least in the short run, but the GDP. About 31 percent of the on-farm emissions are attributed to the societal outcomes of this support are disappointing. These include conversion of land for agricultural purposes. the outcomes for environmental sustainability, the resilience of agri-food systems, poverty reduction, and food security and nutrition. In many Even though the historical contribution of land expansion to increased countries, the bulk of the support for agriculture is delivered in a manner food production may appear to be relatively small, it has an enormous that is both regressive and highly distortionary, creating incentives that environmental impact. Over the past 60 years, the area of agricultural frequently drive producer decisions in favor of targeted commodities and land has increased by only 7 percent, with land under crops growing by encouraging resource-intensive and polluting practices. At the same time, 15 percent and pasture by only 2 percent. It has nevertheless pulled a the large draw on public resources constrains the provision of core public substantial 309 million hectares into agriculture (205 million hectares goods such as agricultural research, and advisory and extension services. into crop production, and 104 million hectares into pastures for livestock The resulting market distortions also often disincentivize private invest- production). This conversion has come at the expense of natural habitats, ment in research and innovation as well as in value chain development for particularly forests, which are dense stores of carbon. As a result, the less-favored commodities, which are often healthier foods. conversion of forests for agricultural use has historically been a major source of GHG emissions, accounting for about 11 percent of global emis- Could the current level of support be repurposed to deliver better sions over the years 2007–2016 (IPCC 2020). The remaining 12 percent of economic, environmental, social, nutritional, and climate outcomes? global emissions comes from crop and livestock production. More recent Given the current level of global support provided to agricultural produc- estimates by Tubiello et al. (2021) suggest that in 2018, of the 22 percent of ers, an outstanding question concerns the scope for repurposing these global emissions accounted for by agriculture, about 15 percent were from funds in a form that promotes more desirable outcomes for productivity, on-farm production processes, and 7 percent from land-use change. Current agricultural practices and the conversion of land from natural habitats have other large negative externalities. Agriculture is the biggest Several recent global reports provide evidence of the complex and multifaceted problems that were facing the 2  driver of biodiversity loss, and it generates enormous economic costs due agriculture and food system even before COVID-19. With the majority of the remaining poor in rural areas, the slowing progress on poverty reduction—with a rising total number of poor in Sub-Saharan Africa—is a stark to lost ecosystem services (Johnson et al. 2021). Beyond the effect on the reminder of the need for continued attention to rural incomes (World Bank 2018). Progress on SDG target 2.1 (Eradicating Hunger and Malnutrition) was already off-track, with the number of undernourished rising environment, current production patterns encourage unhealthy diets with from 615 million in 2014 to as many as 811 million in 2020—a reversal of a decades-long declining trend (FAO large human capital and health costs. Furthermore, current practices are et al. 2021). Beyond undernourishment, the world was also off-track for meeting the targets for malnutrition (stunting and low birthweight), while at the same time child and adult obesity were rising. Furthermore, 2 undermining both current and future economic growth as key resources — billion people were food insecure in 2019, meaning that they lacked regular access to safe, nutritious, and land, labor, water, and energy — are either misallocated or degraded. This sufficient food. The FOLU Global Report (2019) provides estimates of the huge environmental, health, and socioeconomic costs associated with the current food system. constrains the pace of structural transformation--which entails increases Among these 79 countries, 11 countries implicitly taxed their farmers by about $74 billion (in the form of 3  negative market price support), implying total positive transfers of more than $714 billion. 5   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 6  Introduction sustainability, and nutrition. This study was conceived to fill this knowledge gap by assessing potential options for repurposing current agricultural policies and support. Such reforms are expected to have wide-ranging impacts on various outcomes of interest. Accordingly, this analysis con- siders the potential of a range of options for achieving better economic, environmental, social, nutritional, and climate outcomes. It is important to also note that the focus of this study on repurposing support does not consider all of the potential strategies that could be used to transform the food system. These include the large array of potential measures that might influence consumer demand, such as consumption taxes on particular foods or measures to facilitate agri- cultural development, such as investments in infrastructure, value chain efficiencies, nutrition supplements, or biofortification of foods. This report also does not constitute an impact assessment of the strategies discussed as such; the modelling scope does not include all of the measures (for example, food waste reduction targets, dietary shifts, and organic action plans) that could alter the impacts reported. In other words, not all policies that would affect the transition are captured by this model. Other analyt- ical approaches and tools will be necessary to arrive at a more complete picture of the potential impacts of this transition. This study is timely in view of the rising global attention to repurposing public support to agriculture to transform the agriculture and food systems in the interest of realizing better health for people, economies, and the planet. Since the launch of the Just Rural Transition4 at the UN Climate Action Summit (UNCAS) in 2019, growing global momentum has led to recognition of the potential of repurposing public support to agriculture as a potential “game-changing” solution cluster under the action track for boosting nature-positive production that was discussed at the UN Food Systems Summit (UNFSS) in 2021.5 Repurposing public support and incen- tives is also identified as one of five core “imperatives” in the new Food Finance Architecture proposed by the Summit’s Finance Lever of Change.6 The study is also timely in light of the UN Climate Change Conference (COP26), held in November 2021, and the Nutrition for Growth Summit in December 2021: these were two additional venues for promoting reforms with a view to achieving the Sustainable Development Goals by 2030. Just Rural Transition, https://justruraltransition.org/ 4  “A Just Transition to Sustainable Agriculture through Policy Reform and Public Support: Meeting the Triple 5  Challenge of Food and Nutrition Security, Climate and Biodiversity,” https://foodsystems.community/ repurposing-public-support-to-food-and-agriculture-2/ https://www.worldbank.org/en/events/2021/07/02/un-food-systems-summit-public-finance-forum 6  7   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 8  Introduction 2 When considering reforms of agricultural support policies, it is vitally important to understand the complex and multifaceted links between the support provided and the outcomes achieved. This requires identifying measurable outcomes and then exploring the relationships between the support instruments and the policy goals. Reducing emissions of GHGs in agriculture is critical to environmental sustainability.7 This is because agriculture is both an important contributor to global warming and is strongly affected by the impacts of climate change and variability. But policies cannot focus solely on the impacts of reforms on GHG emis- sions—policymakers are also deeply concerned about impacts on poverty, nutrition, and the natural environment. Thus, the question becomes whether it is possible to identify policy reforms that help—or at least do not hinder—achievement of those goals, while also achieving reductions in GHG emissions. The analysis for this study was conducted in two phases. Findings from the first phase were presented in Laborde et al. (2020). That phase used the existing Organisation for Economic Co-operation and Development (OECD) database for agricultural support in 51 countries (OECD 2020), adjusting the agricultural support for border measures and domestic support that influ- ence output and input decisions (coupled subsidies). It then augmented the modeling framework to estimate the changes in GHG emissions associated with shifts in output and inputs resulting from the modeled policy shifts. The BACKGROUND Laborde et al. study examined the impacts of removing agricultural support for agricultural production but did not consider the impacts of agricultural land-use change. Therefore, the second phase of this study, the results of which are reported here, expanded the agricultural support database to AND CONTEXT include an additional 28 developing countries, as discussed in Section 3. Importantly, it also includes the impacts of policy shifts on land-use change and the associated changes in emissions. The first phase of the analysis provided important insights into the degree and channels of influence of agricultural support on production and GHG emissions. A key insight from Laborde et al. (2020, 2021) is that these impacts differ depending on whether producer support is provided Agriculture is also the lead contributor to biodiversity loss, through the conversion of natural habitats to 7  agricultural land, and degradation of the natural resource base, including land and water. The analysis in this study, however, is focused on GHG emissions because incorporating all the other dimensions explicitly is enormously complex: this remains a task for future research. Nevertheless, some of these externalities are implicitly subsumed in the estimation of GHG emissions (for example, GHG emissions associated with land- use change, and emissions directly from soils, crop burning, fertilizer and other chemical uses) even though the longer-term and “hidden” costs associated with these externalities, such as loss of ecosystem services and their potential impacts on future productive potential, are not adequately accounted for. As such, the estimates of the economic impacts in this study may be considered as lower-bound estimates of the true cost associated with policies that influence the decisions of agricultural producers. R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 10  Background and Context through domestic support, or market price support. Transfers that are the assessment from the first phase of this study (Laborde et al. 2021). tied to the production of specific outputs or the use of certain inputs and This matters for approaches that reduce emissions through conditionality, provided through public expenditures are referred to as direct domestic which may reduce productivity and require the conversion of forests for support. Support provided through market prices typically arises from cropland. It also matters when funds can be repurposed from providing trade measures that seek to alter the price producers receive for their direct support into supporting R&D, which can both reduce emissions and outputs. Indirect transfers are effectively transfers from consumers to increase productivity. In the earlier analysis (Laborde et al. 2021), some producers that are generated through higher prices and are not financed of the benefits of such R&D were found to be offset by a rebound effect, through public expenditures. with higher productivity lowering prices, increasing demand and output, and thus offsetting much of the initial reduction in emissions. The current, This study also builds on a more recent FAO-UNDP-UNEP study. The more complete analysis, with land-use change incorporated, finds that this FAO-UNDP-UNEP (2021) study uses the expanded database and the aug- rebound effect is outweighed by the benefits of reductions in land use, mented modeling framework described above to also look at the impacts and hence in emissions from land-use change. of removal of agricultural support. It provides a qualitative discussion of the outcomes that can be expected from repurposing current support, and The empirical analysis for this study was undertaken in three steps. lays out a six-step process for developing a potential repurposing strategy. The first step was to enhance the global modeling framework in order to include the expanded database of agricultural support measures; update The main insight from these studies, confirmed by the analysis pre- the baseline estimates of emissions from both agricultural production and sented in this report, is that the removal of support involves important land-use change; and include model specifications for the links between trade-offs. The current pattern of direct support provided to producers support measures, agricultural production, land-use change, and GHG through various forms of transfers induces higher levels of global agricul- emissions. The second step was to perform experiments to study the tural production and GHG emissions than is seen in the absence of such impacts of existing support by creating counterfactuals for production, support. In contrast, with market price support created by border mea- emissions, incomes, prices, and so on, in the absence of various kinds of sures, the stimulus to global output (and emissions) provided by higher support. The third step was to conduct experiments that would examine prices in the protected markets is offset by a contraction in global demand changes in the use of support, including the refocusing of domestic resulting from higher consumer prices in those markets. support away from products with high emission intensities; conditionality The present study expands the analysis in an important way. It explores designed to reduce the emission intensity of production; and greater specific approaches to repurposing support for better environmental, investments in R&D both to lower emissions and increase productivity, economic, and social outcomes. And it looks at the specific implications along with incentives to foster the adoption of such “green innovations.” of alternative approaches to repurposing current agricultural support — This report is organized as follows. Section 3 describes the analytical such as the use of conditionality to require the use of emission-reducing tools needed for this study, particularly the modeling framework, the technologies, and investing in innovations that reduce emissions and emissions database, the measures of agricultural distortions, and the raise productivity. household models used to assess impacts on poverty. Section 4 examines The inclusion of emissions from land-use change and the design of the results from a range of simulations. Section 5 discusses the challenges potential repurposing measures are hugely important. This is partly of implementing a repurposing agenda, and Section 6 presents a short because, globally, gross GHG emissions associated with land-use change8 summary and conclusions. are of the same order of magnitude as those from agricultural production. It is also important to account for other potential negative externalities associated with land-use change, such as biodiversity loss (Johnson et al. 2021). Accounting for land-use change is likely to affect the results of “Gross” in this context refers to emissions from land-use change, not accounting for change in GHG 8  sequestration by soils and forests. 1 1   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 12  Background and Context 3 This section briefly describes the databases developed for this study. The main building blocks are an expanded database on agricultural support; and an enhanced database on emissions from agriculture and land-use change.  HE SCALE AND NATURE OF AGRICULTURAL 3.1 T PRODUCER SUPPORT Total support to agricultural producers flows through multiple channels. These include market price support, which is provided using many differ- ent instruments, including measures such as tariffs, licenses, tariff-rate quotas, quotas, and trade bans. While in most countries this support is positive, some countries effectively (either implicitly or explicitly) tax producers using measures such as export taxes or restrictions on export volumes (including quotas or outright bans). The combined impacts of these measures are calculated using comparisons between domestic and world prices for the same product. OECD (2021) computes these measures for 54 countries (including all 38 OECD members, 5 non-OECD European Union (EU) member states, and 11 emerging and developing economies). These measures are complemented by data from the Inter-American Development Bank’s Agrimonitor program and the FAO’s Monitoring and Analyzing Food and Agriculture Policies (MAFAP) program, and curated by the International Food Policy Research Institute (IFPRI) in the Ag-Incentives PATTERNS OF database for the International Organizations (IO) Consortium (www. ag-incentives.org). This database allows the tariff equivalent, or nominal rate of protection (NRP), to be calculated as a summary measure capturing the effects of all prevailing border measures. The coverage of countries in AGRICULTURAL the resulting database varies since some countries do not have data for all years since 2005. At its peak in 2012 the database included 88 countries, accounting for 88 percent of global agricultural production. The coverage declined to 73 countries with the data for 2017, but it still accounts for 83 SUPPORT AND percent of the value of world production, and the pattern of protection remains consistent. The nominal rate of assistance (NRA) is a comprehensive measure of EMISSIONS support provided to agricultural producers. This includes both market price support and support provided through budgetary transfers from governments to agricultural producers. For modeling purposes, this type of support may be broadly categorized into tied transfers (or subsidies) to produce certain outputs, inputs, or factors of production, such as land, labor, or machinery. For this study, the database combines all of the avail- able measures of distortions in a way that allows the impacts of changes in these measures on output and production to be modeled. R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 14  Patterns of Agricultural Support and Emissions The scale of total support provided to agriculture is quite large. Looking About 5 percent, or nearly $29 billion, is provided as “green” subsidies, at only the 79 countries for which data are available for 2016–18, the aver- or subsidies to support environmental outcomes. This support is chan- age annual positive support (as transfers from the government or between neled through various instruments across countries—as input subsidies to consumers and producers through market price support) is estimated to promote less-polluting inputs, or to encourage the production of outputs be $713 billion. This was offset by implicit taxation of producers through with fewer negative externalities, or as payments for resource conservation negative market price support, which on average between 2016 and 2018 or land set-asides. Support through green subsidies has increased in amounted to about $74.3 billion across 11 countries, leaving $638.3 billion recent years but it remains limited both in volume and in the number of in net support, as shown in Figure 3.1. Of the total, about 11 percent was 9 countries providing such support. provided through measures to support poor consumers: for example, In contrast, the bulk of transfers to producers was provided through public food coupons or food distribution programs. measures that the OECD refers to as “potentially most distorting.” FIGURE 3.1: T  otal Annual Support to Agriculture Provided by 79 Countries, 2016–18 These include subsidies linked to outputs, inputs, or factors of production (in billions of current dollars and percentage share) such as land area (referred to as domestic support in this study), as Input subsidies well as market price support provided through trade restrictions such as Output Subsidies $86.3b (14%) import tariffs and other border measures (referred to as trade barriers $73.3b (11%) in this study) (OECD 2020). These measures account for 82 percent of the $456 billion provided annually between 2016-18 as producer support Public Goods (that is, total support less expenditures on public goods and services and and Services, $108.2b (17%) for consumer support). Of the remaining 19 percent of producer support, 13 percent is in the form of relatively less-distorting decoupled income Other Green Subsidy, transfers, and 6 percent in the form of “green” subsidies. $207.5b (33%) $28.9b (5%) Behind these aggregate global numbers, the level and nature of support Consumer varies significantly across countries. Importantly, market price support Market Price Support Support, $211.7b (33%) remains the dominant form of distortionary support for most countries $70.4b (11%) (Figure 3.2). Several emerging and developing countries continue to implic- itly tax their producers by keeping domestic prices for key commodities Decoupled transfers $59.6b (9%) below the world market (or reference) prices. In most OECD countries, positive market price support through trade measures remain the most Source: Authors, using data from AgIncentives International Organizations Consortium. b=billion popular form of support that governments provide to producers. As a group, the emerging and developing countries provide the largest share Public goods and services account for only 17 percent of the total of their direct public support for agricultural public goods and services. support. Of this, about 31 percent is for R&D; 42 percent for infrastructure Green subsidies are emerging, but the evidence shows that except in (most prominently irrigation development); and the remaining 27 percent China, these are largely offered in the developed countries. for other public services. In other words, only about 5.3 percent of the total support provided for agriculture is devoted to R&D spending, even though it is identified as a core driver of productivity, and a key instrument for addressing the challenge of resilience in the face of climate change. More recent data (the average of the years 2018–20) are available for the 54 countries monitored by OECD 9  (OECD 2021). These data show that these 54 countries provided $720 billion per year as positive transfers to producers, which were counteracted by $104 billion in implicit taxation of farmers through negative market price support in some countries, resulting in a net global transfer to producers of about $616 billion. 1 5   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 16  Patterns of Agricultural Support and Emissions FIGURE 3.2: A  gricultural Support Across Main Countries and Country Groupings, 2016–18 FIGURE 3.3: Nominal Rate of Assistance, by Major Component (%) 25% Other Inputs Outputs NRP Nominal Rate of Assistance (NRA) 20% 15% 10% 5% 0% 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 Year Source: Authors’ database of agricultural support. Source: Authors, using data from AgIncentives IO Consortium (IFPRI, OECD, FAO, IDB, and the World Bank). Border protection measures accounted, on average, for roughly half of Note: EMDE = emerging market and developing economies. the total assistance provided (8.1 percent of GFR on average for 2016–18). Of the tied transfers or domestic support provided directly by govern- Focusing on the elements of support that have the potential to alter the ments, most are not based on production inputs or outputs, but on factor incentives that influence producer decisions, it is useful to assess their use, such as amount of area planted or numbers of animals (4.3 percent size relative to the value of agricultural production. Two key measures that of GFR). While roughly 45 percent of this support is not linked to current capture the major forms of support are the NRP and the NRA. The NRP reflects production decisions (OECD 2020, 102), most of it nevertheless indirectly market price support, or support provided by border measures, as a proportion influences production through conditions such as the restrictions on of gross farm receipts (GFR) valued at world market prices. Adding to market payments to active farmers (Abbott 2020; EU 2015), or through percep- price support, the direct transfers provided as output and input subsidies and tions that payment bases are likely to be updated (Bhaskar and Beghin other forms of support, including support decoupled from production, yields 2010). In order of magnitude, support based on factor use is followed by the NRA, which is the total support relative to the GFR. The trends in global support coupled to input use (such as water, power, and fertilizer), at about agricultural producer support, using data for countries for which consistent 2.7 percent of GFR. Direct transfers coupled to output are the smallest, time series are available are presented in Figure 3.3. averaging 0.4 percent of GFR. The NRA averaged 15.4 percent of gross farm receipts for the three Market price support has been rather volatile, particularly in periods of years 2016–18. Figure 3.3 indicates the predominance of the support sharply rising commodity prices, such as in 2008 and 2011, when many provided by protection relative to support provided in the form of public governments reduced protection of or increased taxation on farmers expenditures—through transfers to outputs or inputs, or support that is (for example, through export restrictions), as they sought to reduce the conditional on other variables. As noted above, support varies significantly impacts of increases in world prices on consumers. Unfortunately, this led across countries. To capture this, a breakdown of NRAs by groups of coun- to a serious collective action problem, since containing domestic prices tries classified by income levels is provided in Appendix A (Figure A1.1). It increased the demand for food and thus exacerbated the increases in shows the minimal average support provided by low-income countries, with world prices (Anderson, Ivanic, and Martin 2014). During periods of stability some countries implicitly increasing taxes for farmers as a way of keeping in world market prices, such as 2016–18, market price support has been consumer prices in check. The level of publicly funded support also tends relatively stable. to be low (about 5 percent on average) across middle-income countries, but is close to 15 percent in high- income countries. 1 7   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 18  Patterns of Agricultural Support and Emissions The incidence of market price support varies widely across countries. PORK/ REGION/COUNTRY BEEF MILK POULTRY RICE SUGAR WHEAT The database used in this analysis provides estimates of the amount of support provided, not only for major producer countries but also for many China 15.5 72.0 16.9 31.3 103.1 55.0 Colombia 3.5 40.5 8.3 113.9 13.0 countries where agriculture accounts for a substantial share of the national Costa Rica 0.0 0.6 43.8 130.0 29.1 economy. Table 3.1 provides 2016–2018 data on positive market price support in a range of economies and key regions for products that were Dominican 29.4 230.8 45.1 58.9 Republic important both for the economy and for GHG emissions. Almost all of the Ecuador 0.0 19.0 36.1 64.2 support shown in this table was provided by import protection, with just El Salvador 14.4 22.4 148.2 22.4 40.8 a few cases where protection was provided to export-oriented industries Ethiopia 15.7 through implicit export subsidies. European Union 27.0 0.1 9.0 23.9 3.1 0.8 For countries with positive market price support, the differences Guatemala 67.9 6.2 81.5 between developed and developing countries are relatively small. Table Honduras 26.1 22.1 28.9 7.2 27.5 3.1 highlights the relatively small differences in rates of protection between Iceland 37.2 92.7 210.2 developed and developing countries, with two notable exceptions. For rice, India 25.2 25.8 0.4 the protection rate is nearly three times higher in developed countries. For Israel 4.5 50.4 43.1 16.9 wheat, many developed countries that export the crop keep protection Japan 38.5 134.8 94.0 228.3 34.3 0.0 rates relatively low (0.7 percent), while developing countries maintain Kazakhstan 3.5 0.0 0.0 4.2 higher levels of protection (24.4 percent on average). Overall, while many Kenya 30.6 80.8 18.3 developing countries have minimal protection for many commodities, they Korea, Republic of 43.4 133.8 136.5 102.3 tend to apply relatively high rates to those they do protect. The highest Mali 13.9 34.7 rate of protection in the table, for instance, is for milk production in the Mexico 0.0 0.2 0.3 0.0 42.5 0.0 Dominican Republic (at 231 percent). Mozambique 9.8 4.3 New Zealand 0.0 0.0 7.5 0.0 TABLE 3.1: P  ositive Protection Rates for Key Commodities by Region and Country, 2016–18 (%) Nicaragua 0.0 0.0 83.3 75.2 0.0 Norway 84.0 102.2 111.6 89.1 PORK/ 0.0 0.0 0.0 0.0 0.0 REGION/COUNTRY BEEF MILK POULTRY RICE SUGAR WHEAT Paraguay Philippines 10.0 37.3 136.6 59.7 World 11.5 17.5 14.1 46.1 28.5 12.6 10.7 15.8 11.8 123.8 25.8 0.7 Russian 22.1 38.8 11.5 46.3 0.2 Developed Federation Developing 12.5 22.6 15.4 34.4 29.1 24.4 Rwanda 80.7 104.6 Africa 0.7 0.9 0.5 38.8 59.1 12.3 Senegal 20.0 Asia 25.0 56.2 20.8 48.3 44.4 27.3 South Africa 0.0 0.9 0.5 56.1 0.4 Latin America 0.9 6.6 3.6 38.9 9.7 2.0 71.4 29.0 138.9 2.1 32.5 Switzerland & Caribbean Turkey 120.3 0.5 39.4 2.8 0.0 Argentina 0.0 0.0 2.9 0.0 Uganda 74.9 48.3 Australia 0.0 0.0 0.0 0.0 0.0 0.0 Ukraine 0.0 4.5 2.7 0.0 Benin 61.1 United States 0.0 27.6 0.0 0.0 81.6 0.0 Brazil 0.8 4.4 0.0 21.5 0.0 6.3 Uruguay 0.0 0.0 43.0 0.0 0.0 Burkina Faso 60.0 Vietnam 21.2 0.0 9.2 98.6 Burundi 48.0 Canada 0.0 66.9 0.9 0.0 Source: Authors’ calculations for countries/commodities with positive protection. Chile 0.0 0.0 0.0 3.5 0.0 1 9   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 20  Patterns of Agricultural Support and Emissions China currently provides substantial positive support for key agricul- TABLE 3.2: N  egative Protection Rate for Key Commodities by Region and Country, 2016–18 (%) tural products. The current situation is a complete reversal from China’s overwhelming taxation of the sector in the 1980s and 1990s (Huang et al. PORK/ 2009). The European Union (EU), by contrast, provides relatively modest REGION/COUNTRY BEEF MILK POULTRY RICE SUGAR WHEAT market price support for the key commodities in Table 3.1, while high rates World -2.0 -12.1 -0.9 -8.4 -1.1 -1.0 of market price support in the United States are applied only on milk and Developed 0.0 0.0 0.0 0.0 0.0 -1.8 sugar. Japan continues to have relatively high protection on rice and milk. Developing -6.0 -18.7 -3.1 -8.8 -1.3 -0.3 A set of relatively land-scarce high-income countries including Iceland, Africa -2.6 -56.7 -42.4 -20.3 the Republic of Korea, Norway, and Switzerland also have high rates of Asia -26.2 -20.7 -34.3 -7.4 -12.6 protection. But high rates of protection are also seen in many developing Latin America -18.7 -31.8 -35.6 -3.0 -0.3 countries, with rates above 100 percent on rice in Colombia, Costa Rica, & Caribbean and the Philippines. Argentina -18.3 -33.5 -39.2 -0.3 Burundi -0.1 A very different pattern is evident across countries with negative Dominican -12.4 -3.0 protection. Negative protection is generally the result of the explicit or Republic implicit taxation of exports, and occasionally of import subsidies that are Ghana -63.1 used to keep domestic prices below world prices. Such negative support Guatemala -22.2 -9.2 is almost nonexistent in the developed countries and is much less widely Honduras -2.2 used than positive import protection in developing countries (Table 3.2). India -26.2 -20.7 -6.2 Developing countries that do apply export taxes or import subsidies, Kazakhstan -52.4 -12.6 however, tend to do so at quite high rates, ranging up to 26 percent for Kenya -49.2 beef, 57 percent for milk, and 39 percent for poultry. One important case Malawi -20.3 is India, where domestic prices for bovine meat and milk are substantially Mali -2.6 below world prices, with important implications for global production and Mozambique -25.8 consumption levels.10 Argentina is another important outlier, particularly at Russian -6.5 its income level, with domestic prices of beef, milk, and pork all substan- Federation tially below world prices. Rwanda -56.7 Tanzania -37.4 From earlier analysis, it is known that domestic support increases Uganda -45.9 GHG emissions more than market price support does. As Laborde et al. Ukraine -21.1 (2020, 2021) point out, this is partly because domestic support increases Vietnam -34.3 -12.9 output without the offsetting impact on global demand associated with Source: Authors’ calculations for commodities/commodities with negative protection. market price support, and partly because this support is often in the form of direct support, or subsidies tied to the use of inputs such as chemical One key question is whether support is being provided at high rates fertilizer or pesticides that directly affect emissions. or in large volumes to the products that are responsible for the largest contribution to agricultural emissions. As will become clear in Section 4, by far the most agricultural emissions arise from beef and dairy pro- duction, and from rice. But these products receive only around a quarter of total support, as shown in Table 3.3. This table also shows that most current support, about 72 percent of total domestic support, accrues to crops, and only 28 percent to livestock products. Table 3.4 shows that the rates of domestic support on those products are about or below the Bovine meat exports from India are composed entirely of carabeef, or the meat of water buffalo. Trading of the 10  average for all products. meat of cows, oxen, and calves is prohibited in India. 2 1   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 22  Patterns of Agricultural Support and Emissions TABLE 3.3: D  istribution of Domestic Support by Product, Instrument, and Country Grouping, 2016–2018 3.2 THE DATABASE FOR EMISSIONS (percentage shares) GHG emissions from agricultural production and land-use change HIGH-INCOME COUNTRIES LOW- AND MIDDLE-INCOME COUNTRIES remain important components of total emissions. Figure 3.4 shows key Output Input Factor Output Input Factor World components of global GHG emissions in megatons of CO2eq for 1990 and SECTOR subsidies subsidies subsidies Subtotal subsidies subsidies subsidies Subtotal total 2018. Emissions from agricultural production increased 17 percent during Cattle 0.1 1.3 4.1 5.5 0.1 1.2 1.1 2.4 7.9 that period. In contrast, gross emissions from land-use change declined Dairy 0.4 1.2 4.4 6 0.2 1.3 0.5 2 8 by 25 percent because of a reduction in the rate of deforestation and Poultry & Pigs 0.3 1.8 5.1 7.2 0.2 2 2.3 4.5 11.7 in emissions from organic soils. However, the annual quantity of CO2eq Livestock sequestered by forests also declined by 24 percent, partly because of 0.8 4.3 13.6 18.7 0.5 4.5 3.9 8.9 27.6 subtotal declines in overall forest cover and forest health. But the increases in Fibers 0.1 0.1 0.6 0.8 6.1 0.9 1.2 8.2 9 emissions associated with agriculture were small relative to the increases Maize 0.4 0.9 2.8 4.1 0.1 2.1 7.2 9.4 13.5 observed for other sectors, especially energy/electricity (81 percent), Oilseeds 2.4 0.7 1.8 4.9 0.1 1.8 3.6 5.5 10.4 transport (79 percent), and industry (82 percent). Other crops 0 0.9 3.5 4.4 0.3 3.3 1.1 4.7 9.1 Longer-term trends show a decline in agricultural and land-use Rice 0.1 0.1 0.8 1 0.1 3.7 2.1 5.9 6.9 emissions, but agriculture remains a significant contributor to total Sugar crops 0.3 0.2 0.5 1 0.1 0.8 0.2 1.1 2.1 emissions. Figure 3.5 shows that the share of agriculture and land-use Vegetables 0 2.1 5.5 7.6 0.5 4.8 2.5 7.8 15.4 change in gross emissions (excluding sequestration), which averaged 23 & Fruits percent over the period, fell from 28 percent in 1990 to 18 percent in 2018. Wheat 0.8 0.4 1.9 3.1 0.2 1.6 1.1 2.9 6 This leaves the share of agriculture and land-use change roughly on par Crops 4.1 5.4 17.4 26.9 7.5 19 19 45.5 72.4 with the shares for transport, industry, and “other,” which includes fugitive subtotal emissions from energy production and waste. Despite the decline in All products 4.9 9.7 31 45.6 8 23.5 22.9 54.4 100 agriculture’s share, achieving global climate goals will likely not be possible Source: Authors’ calculations. without major efforts to reduce emissions from agricultural production and related land-use change. TABLE 3.4: R  ate of Domestic Support by Instrument, Sector, and Country Grouping, 2016–18 (%) FIGURE 3.4: C  hanges in Levels of GHG Emissions by Main Economic Sector   HIGH-INCOME COUNTRIES LOW- AND MIDDLE-INCOME COUNTRIES (megatons of CO2eq) Output Input Factor Output Input Factor 20000 SECTOR subsidies subsidies subsidies Total subsidies subsidies subsidies Total Cattle 0.3 1.8 4.5 6.6 0.4 1.2 1.7 3.3 Dairy 0.7 2.2 5.7 8.6 0.7 1.5 1 3.2 15000 Poultry 0.3 1.7 4.2 6.2 0.7 1.2 1.9 3.8 & Pigs 10000 Fibers 2.1 2.5 8 12.6 34.6 3.2 3.3 41.1 Maize 0.8 2.6 4.2 7.6 0.5 1.8 5.1 7.4 5000 Oilseeds 7.7 2.4 4.7 14.8 0.4 1.7 3 5.1 Other crops 0 2.7 12.4 15.1 1.5 3.1 1.5 6.1 0 Rice 0.1 1.3 9.3 10.7 0.7 3.2 2.4 6.3 Sugar crops 6 1.8 4.3 12.1 0.4 2.5 0.9 3.8 -5000 Agriculture Land use Sequestration Electricity Transport Ind & Manuf Other Vegetables Change 0 2 5.4 7.4 1 2 1.5 4.5 & Fruits 1990 2018 Wheat 5 2.3 6 13.3 0.9 3.2 2.6 6.7 Source: FAOSTAT for agricultural and land-use change emissions (faostat.org, extracted April Source: Authors’ calculations. 17, 2021). Other categories of emissions from Climate Watch data (www.climatewatchdata.org, extracted April 17, 2021). 2 3   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 24  Patterns of Agricultural Support and Emissions FIGURE 3.5: Shares in Global Gross GHG Emissions by Main Economic Sector (%) TABLE 3.5: S  hares of Emissions from Agricultural Production by Commodity and Source, 2017 (%) 100   OTHER RUMINANT PIG, POULTRY 90 RICE CEREALS MILK MEAT MEAT, AND EGGS TOTAL 80 Burning of 0.2 0.5 0.0 0.0 0.0 0.7 crop residues 70 Crop residues 1.3 3.2 0.0 0.0 0.0 4.5 60 Enteric 0.0 0.0 28.0 21.2 0.5 49.7 50 fermentation 40 Manure 0.0 0.0 2.2 1.9 3.4 7.5 management 30 Manure left 0.0 0.0 6.4 10.0 1.2 17.7 20 on pasture 10 Manure applied 0.0 0.0 1.1 1.0 2.0 4.1 to soils 0 Pesticides 0.0 0.3 0.0 0.3 0.0 0.6 2010 2011 2012 2013 2014 2015 2016 2017 2018 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Rice cultivation 10.7 0.0 0.0 0.0 0.0 10.7 Agriculture LUC Electricity Transport Industry Other Synthetic 0.3 1.9 0.0 2.2 0.0 4.5 fertilizers Source: FAOSTAT for agricultural and land-use change emissions (faostat.org extracted April 17, 2021). Other categories of emissions from Climate Watch data (www.climatewatchdata.org, TOTAL 12.5 5.9 37.6 36.8 7.1 100.0 extracted April 17, 2021). LUC = land-use change. Source: FAOSTAT. For this study, detailed databases of emissions from agricultural Another potentially important distinction is between emission inten- production and from land-use change were created. FAOSTAT presents sities across products and countries. At the world level, the emission data on emissions by type and by commodity for each country, but a intensity of ruminant meat is roughly 40 times that of chicken and 17 full matrix of emissions by type, commodity, and source is needed in times that of pork. The emission intensity of rice is more than four times order to consider changes in emissions by type in the production of each that of other cereals. There are also large differences between countries commodity, such as reductions in emissions from enteric fermentation in and regions, with the emission intensity for beef much lower in the beef production. The approach this study uses to develop this database United States than in countries like Australia that primarily use grass-fed is described in detail in Appendix B, along with the modeling approaches production methods, while emission intensities are particularly high in India used to capture the impacts of policy changes. (Table 3.6). For both milk and ruminant meat, the emission intensities in the Emissions by commodity and source show the predominance of industrial countries are much lower than in developing countries. livestock products in total emissions from agricultural production. TABLE 3.6: E  mission Intensities for Key Products and Regions, 2017 (kg CO2eq/kg Table 3.5 highlights the extraordinary importance of emissions from milk of product) and ruminant meat in overall emissions. Enteric fermentation associated with the production of these commodities accounts for almost half of COUNTRY/ CEREALS BOVINE REGION EXCL. RICE EGGS MEAT CHICKEN PIG MEAT MILK RICE total emissions, while manure accounts for another 22.6 percent. The Australia 0.2 0.4 24.5 0.2 2.5 0.6 0.7 main source of emissions from crop production is rice, accounting for 12.5 Brazil 0.2 0.8 34.6 0.3 2.4 1.1 0.5 percent of total agricultural emissions, while crop residues account for 5.2 European Union 0.2 0.7 15.5 0.3 1.5 0.5 3.1 percent, and chemical fertilizers 4.5 percent. India 0.3 0.5 108.3 0.4 5.0 1.0 0.7 United States 0.2 0.5 11.9 0.3 2.0 0.4 1.1 Developed 0.2 0.6 15.0 0.3 1.7 0.5 1.1 Developing 0.2 0.7 31.8 0.7 1.4 1.3 0.9 World 0.2 0.6 25.5 0.6 1.5 0.9 0.9 Source: FAOSTAT. 2 5   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 26  Patterns of Agricultural Support and Emissions Emissions from land use and land-use change were estimated by a detailed tracking of the carbon stock adjustments. Beginning with an inventory of land in each region mapped to the category “Cropland, Pasture, Forest, and Other,” the stocks of carbon associated with land use and land-use change were then tracked using procedures consistent with FAOSTAT and IPCC (2003). Carbon stock accumulation in croplands and grasslands and sequestration in forests was tracked, as well as conversion between cropland and forests. 2 7   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 28  Patterns of Agricultural Support and Emissions 4 This section considers the modeling framework used to assess the impacts of changes in support policies to global GHG emissions. The global modeling framework used for this study is briefly introduced first. The scenarios used for the analysis are then presented, followed by their key results.  HE MODELING FRAMEWORK FOR ASSESSING POLICY 4.1 T TRADE-OFFS This study uses a global dynamic general equilibrium model to simulate the outcomes to 2040 based on a series of policy shifts. IFPRI’s global computable general equilibrium (CGE) model, MIRAGRODEP, provides the core of the modeling framework. This is an extension of the widely used MIRAGE multisector, recursive dynamic CGE model of the global economy (Decreux and Valin 2007; Laborde, Robichaud, and Tokgoz 2013), which allows for a detailed and consistent representation of the economic and trade relations between countries. Appendix C provides a detailed description of the model, including an explanation of how the global model framework is linked to the large set of household data and models needed to assess the impacts of agricultural policy reforms on poverty. As with any ex-ante modeling analysis, the findings discussed in this study are intended to provide strategic guidance on trade-offs and SIMULATING potential outcomes. The simulations are not intended to provide precise predictions or quantitative assessments of impacts, but rather insights into the outcomes associated with policy options of the broad type con- sidered. The magnitudes of the shocks applied were chosen to be relevant POLICY to recent policy proposals and/or assessments of the impacts of potential reforms, and to provide a basis for understanding the qualitative effects of reforms that might involve larger or smaller shocks than those imple- mented. If, for instance, a reform involves a reduction in emissions per unit OPTIONS of land but lower yields, does the reduction in emissions from production outweigh the increase in emissions from land-use change as additional land is brought into agricultural use? The results from the baseline simulation are presented first. This “zero” scenario simulates the business-as-usual, or “no policy change” option; that is, it assumes that current policies and patterns of producer support will continue unchanged, and projects global economy-wide outcomes in such a case, from 2020 to 2040. It provides the benchmark against which agricultural policy changes can be examined. This approach allows the analysis to reflect the anticipated changes in the structure of the world economy, and particularly the changes in the share of developing economies over this period. Simulating relative to this dynamic benchmark R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 30  Simulating PolicyOptions also allows the effects of changes in policies to cumulate over time. In countries from 2021–2040, as are the rates of growth for population, the baseline scenario, the rates of protection and assistance provided by agricultural total factor productivity (TFP), and agricultural emissions import and export barriers and subsidies were held constant. All of the (Figure 4.1). scenarios yield simulated projections of impacts for the period 2020–2040. The first reform scenario considers removing current producer support. Two complementary simulations (1a and 1b) examine removal of the two BOX 4.1: BASELINE SIMULATIONS distinct forms of producer support–domestic support provided to produc- The basic ingredients for the baseline simulations are a set of ers, and both domestic support and trade barriers, or market price support. economic projections that provide output targets and a set of These simulations help to shed light on the potential trade-offs associated demographic projections for the evolution of the labor force. with a blanket removal of current support and establish the “value for These are treated as exogenous parameters, with the economy- money” for the substantial public resources spent on wide rate of productivity growth that would be consistent domestic support. with these outcomes determined within the model. With this The next three sets of scenarios simulate the outcomes associated with information, the model solves for spending and saving levels various options for redirecting or repurposing current domestic support in each year and calculates the opening stock of capital for to producers. These scenarios correspond to three broad categories of the next year. It then solves repeatedly to create a projection options. Scenarios 2a and 2b consider restructuring the current pattern of to 2040. Because of the particular importance of agricultural domestic support, relying on currently available technologies and practices, productivity growth in this study, it was specified separately either to make support uniform or to focus it only on low-emission prod- from nonagricultural productivity growth, with a slower rate ucts. Scenario 3 makes domestic support conditional on environmental reflecting the depressing effect of climate change on agricultural outcomes, using currently available technologies and practices. Scenario productivity. For subsequent policy simulations, the roles of 4 simulates repurposing a part of the current domestic support to target GDP and productivity growth are reversed, making productivity investments in green innovations; that is, technologies that reduce emis- growth exogenous, and allowing the model to determine the level sions while also enhancing productivity. More details on the simulations of GDP in the policy simulations. This allows the model to assess analyzed are provided below, with detailed results presented in Appendix D. the impacts of changes in policies on the full range of variables determined within the model. 4.2 CONTINUING WITH BUSINESS AS USUAL The potential impacts of policy changes are estimated as deviations 4.2.1 Scenario 0: Baseline Trends from the baseline projection of outcomes with unchanged policies. It is therefore important to have a good understanding of the baseline trends The “business-as-usual” (or zero) scenario provides projections of and their underlying assumptions. The key assumptions are summarized probable outcomes with unchanged policies. The projections for key in Box 4.1. To focus on the core issues at hand—the impacts of policy economic indicators show important differences in the outlook for devel- reforms—and to avoid what are likely, in retrospect, to be extraordinary and oped and developing countries. Figure 4.1 shows substantial differences uncertain adjustments to the baseline due to COVID-19 shocks, the last in GDP growth rates between developed and developing countries. In pre-pandemic set of economic forecasts from the World Economic Outlook addition, much higher TFP growth rates are projected in developing coun- were used—that is, those from October 2019 (IMF 2019). These provide tries for both agricultural and nonagricultural products, a projection that is historical data up to 2018, and then forecasts to 2024. The GDP forecasts consistent with continuing income convergence (Martin 2019; Startz 2020). to 2024 were used to capture adjustment dynamics during that period. The Agricultural TFP was adjusted relative to nonagricultural TFP, taking into growth rates for 2024 were then used to provide a benchmark growth rate account information on rates of yield growth, the desirability of avoiding for subsequent years. In line with the trends of recent decades and most excessive changes in real agricultural prices during the projection baseline, long-term projections for the world economy, the average rate of income and the adverse impacts of climate change on yield growth going forward growth in developing countries is assumed to be higher than in developed (Schlenker 2021; Ortiz-Bobea et al. 2021). 3 1   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 32  Simulating PolicyOptions FIGURE 4.1: B  aseline Projections of Key Economic and Environmental Outcomes, FIGURE 4.2: Key Features of Baseline Projections 2017–2040 (average annual growth rates in percent) Agri Value Added ($Trillion) Poverty (PPP$1.90, %) Developed Developing 4 7.36 8.20 3.5 5.47 3 3.92 7.22 7.15 2.5 2020 2030 2040 2020 2030 2040 2 Agric Land (Bill. ha) Net Agric. Emmissions, Gt 1.5 Agric Production LULUC 1 0.5 4.87 4.81 4.84 9.12 0 5.76 7.24 Real GDP Non-Ag TFP Agric TFP Ag Emissions Source: Authors’ baseline scenario. 2020 2030 2040 -2.29 -2.18 -2.13 2020 2030 2040 Key projected outcomes relevant to the agriculture sector are shown Source: Authors’ baseline scenario projections. in Figure 4.2. Real agricultural value added in 2017 in US dollars rises by 88 percent, from $3.92 trillion to $7.36 trillion. This growth includes an The largest increase in emissions is expected from livestock production. increase of 87 percent in crop production and 48 percent in livestock pro- Figure 4.3 shows the increase in agricultural emissions by source. Dairy duction between 2020 and 2040. Global poverty headcount rates fall from alone accounts for 46 percent of the incremental emissions to 2040, and 8.2 percent to 7.2 percent. The slow rate of decline in poverty is strongly dairy, beef, and pork production together account for 77 percent. Fourteen influenced by the relatively low rate of growth in agricultural productivity percent of the total growth in emissions will be from crop production, part (Ivanic and Martin 2018). Agricultural land use rises by 23 million hectares of which reflects increased demand for livestock feed. However, the prima- between 2020 and 2030—an increase similar to the 28 million hectare ry contributor to these increased emissions will be synthetic fertilizers (55 expansion projected by the World Bank (Johnson et al. 2021, 27)—and then percent of crop emissions and 9 percent of all total production emissions). is projected to increase further by 33 million hectares by 2040, implying an FIGURE 4.3: C  ontributions to Growth in Emissions from Agriculture and Agricultural expansion in agricultural land use of 56 million hectares (an increase of 1.2 Land-Use Change, 2020-2040 (%) percent) over the entire period 2020–2040. 9% 3% 3% The baseline scenario projects a substantial increase in the level 4% of agricultural emissions in coming decades. Continuing with busi- 4% Rice ness-as-usual, GHG emissions from production would increase from 5.8 Wheat gigatons CO2 equivalent (Gt CO2eq) to 9.12 Gt CO2eq between 2020 and Maize 2040, an increase of 58 percent (Figure 4.2). Net annualized emissions from land use and agricultural land-use change are negative, because Other emissions from land-use change are offset by sequestration. However, the 27% Beef size of these net benefits declines by 7 percent because of the dual effect Pork/Poultry of higher emissions from forest conversion and reduced capacity 46% Dairy for sequestration. LUC 4% Source: Authors’ baseline scenario projections. 3 3   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 34  Simulating PolicyOptions 4.3 REMOVAL OF CURRENT SUPPORT MEASURES as the country reducing emissions increases its imports or reduces its exports, leading to carbon “leakage.” The key results from the first experi- Before turning to options for repurposing existing support, it is import- ment, which assumes that all domestic support provided through transfers ant to understand the impacts of the current support measures. This to producers would be eliminated, are shown in Figure 4.4. Given the major can be assessed by looking at what would happen if the current support differences in the nature and form of support provided by developed and provided to agriculture were removed. Answering this question provides developing countries, this figure shows impacts both for the world as a insight into the likely trade-offs, if any, that a change in current policies whole and for developed and developing countries.11 might entail. One issue of particular interest to policymakers is the likely impact on production, which is often equated to food security. It is equally One important message emerging from these simulations is that a important to assess the “value for money” for the hundreds of billions of blanket removal of all domestic support would entail important trade- dollars that are currently spent globally on agricultural support. Domestic offs. Tracking the impact of such reforms through to outcomes related to support measures are of particular interest, as these rest on re-allocable income, GHG emissions, land-use change, poverty, and nutrition reveals fiscal resources within limited budgets; earlier work has found that this the complexity of these effects. type of support tends to have a greater impact on global GHG emissions Removal of domestic support would have favorable, but small, impacts than do market access barriers (Laborde et al. 2021). on the climate and nature. Abolishing domestic support would reduce A key question is how current support measures are affecting devel- agricultural GHG emissions by about 103 megatons of CO2eq. These opment outcomes. These outcomes include food production, national reductions would be induced by a decline in the use of agricultural inputs income, poverty, the cost of healthy diets, the level of GHG emissions from and the factors of production that are currently supported; the larger fall agriculture and land use, and the demand for agricultural land (along with in crops than in livestock reflects the relatively higher current support for its corollary impact on forest habitat). To delineate the potentially distinct crops. The environmental gains would also vary across countries, reflecting impacts of domestic support from those of trade barriers, the analysis first the level of support provided by individual countries—hence the reduction simulates the impact of removing only domestic support measures, and in GHG emissions would be larger for developed than for developing coun- then considers the removal of both domestic support and trade barriers tries. The removal of support would also reduce the territorial footprint of simultaneously. All of the policy reforms considered are implemented agriculture, reducing the amount of land under agriculture by a substantial progressively between 2020 and 2025, with the 2025 policy position held 27 million hectares by 2040—and preventing nearly 49 percent of the constant during the projections to 2040 to allow the longer-term impacts potential conversion of land to agriculture that is projected over the next of the policy changes to be identified. Most of the impacts are reported as 20 years under the current policy and support regime. This land savings deviations from the benchmark (or baseline) outcomes in 2040; that is, the would directly contribute to an increase in forest habitat, with important outcomes projected assuming there were no change in policies from the positive contributions to reducing GHG emissions through sequestration “business-as-usual” scenario. and protecting biodiversity. The patterns observed are consistent with the nature and structure of domestic support across countries. 4.3.1 Scenario 1a: Remove All Domestic Support The first policy experiment (Scenario 1a) simulates the effect of all countries eliminating all domestic support simultaneously. While achieving such a global consensus would be challenging in practice, this set of simulations helps to quantify the influence of current forms of support on global outcomes of interest, including climate outcomes: a truly global public good. It should be noted that collective action is vitally important in achieving progress on climate outcomes. While reducing GHG emissions requires actions at the country level, the gains at the individual Additional experiments were also carried out to assess the impact of eliminating various components of 11  country level may be offset by increases in production in other countries the coupled support differentiated by form of payment, and whether to crops or livestock, or for developed countries only. Detailed results from these experiments are presented in Table D.1. 3 5   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 36  Simulating PolicyOptions FIGURE 4.4:  Global Implications of Removing All Current Domestic Support baseline.12 On the other hand, such reform would face a major political (Percentage of Change Relative to Baseline Projections for 2040) economy challenge because it would lead to a decrease in farm output, ECONOMIC FARM SECTOR reinforcing some policymakers’ concerns about food security. Crop (Real national income, % change 2040) (Agricultural production volume, % change 2040) production would fall by 2.6 percent in developed countries and 1 percent in developing countries, and livestock production would fall by 1 percent in Developing the developed countries and close to zero in developing. These reductions Developing in output would drive up global prices, but despite the price increases, Developed Developed real farm incomes per worker would decline by about 4.5 percent globally, LIVESTOCK CROPS with more dramatic declines in developed countries (11.4 percent) than in World World developing countries (2.7 percent). -0.05 0.00 0.05 -1.0 -0.5 0.0 Impacts on poverty and nutrition would also be adverse. Another prob- lem is associated with impacts on the poorest, and on diets at the national SOCIAL DIETS level. While current farm support regimes do not appear to have been (Poverty at PPP$1.90, % change 2040) (Healthy food prices, % change 2040) designed to reduce poverty or to improve diets, their abolition would likely both increase poverty (albeit marginally) and make healthy diets more Developing costly. The rise in prices would make food more expensive for the poor, Developing constraining progress on poverty reduction. Rising prices would also raise Developed Developed the prices of nutrient-dense foods such as dairy products, vegetables, and fruit, thus reducing their consumption; at the same time, sugar consump- World World tion would also fall. -0.01 0.00 0.01 0.01 -0.2 0.8 1.8 One important insight from this analysis is that the type of support matters. Different types of support have heterogenous impacts, making it CLIMATE NATURE (Reduction in emissions from (Agricultural land, % change 2040) difficult to generalize across different interventions (Figure 4.5). Different agriculture and land use change, %) support measures and policy instruments show broadly similar effects on some outcomes but notably different effects on others (see Table D.1). This Developing Developing points to the need for a carefully considered and nuanced strategy for reforming current agricultural support. For example, among transfers, direct Developed Developed input subsidies have the largest impact on production emissions. This is not surprising, since most input subsidies are targeted at crop inputs, World World specifically fertilizers; but they account for less than half the reduction in emissions when all domestic support to crop production is removed. -6.0 -3.0 0.0 -0.2 -0.1 0.0 Eliminating all support to crops would reduce emissions the most—equiv- alent to a substantial 136 megatons of CO2eq, or a 24 percent reduction in Source: Authors, using model simulation results. Note: Brown bars indicate movement toward, and teal bars indicate movement away from the estimated incremental emissions from crops between 2020 and 2040 achieving the related SDG(s). when compared to the baseline scenario. This is a sizable impact consid- ering that production would fall by only 1.3 percent. While most crops have However, these environmental gains would come with mixed economic low direct emission intensities, this result also partly reflects the fact that outcomes. On the one hand, the removal of economic distortions created crop output is also a major input (as feed) into the livestock sector. by distortionary domestic support would generate efficiency gains, raising real world income in 2040 by about $74 billion per year relative to the Economic efficiency gains were calculated using the projected estimate of global real GDP of $149.8 trillion in 12  2040 (an increase of 82.1 percent from the 2017 real GDP of $81.7 trillion). 3 7   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 38  Simulating PolicyOptions FIGURE 4.5: Impact on GHG Emissions of Removing Different Types of Support developing countries as groups, and for some selected large agricultural countries, including Brazil, China, the EU, India, and the US). The reduction Megatons, CO2 equivalent in farm output would vary significantly across countries, with larger -103.1 All domestic support declines in both crop and livestock output for developed than for develop- ing countries. Among the countries included in Table D.2, crop production 2.6 Livestock only would fall significantly in the US (-5 percent) and the EU (-4 percent), but also in India (-3 percent) and China (-2 percent). It would rise in Brazil. -136.0 Crops only Livestock production would shift from the EU and India to Brazil, China, and -28.4 Factor payments the US. In general, countries that are light subsidizers and major exporters, such as Brazil, would gain from such a global reform. Emissions would fall -52.3 Input subsidy in most countries, but with significant variation across countries. The US would experience a 32 percent decline driven by emissions from land use -25.2 Output subsidy and land-use change. On the other hand, emissions in China would rise, as they would in Brazil and the EU, driven by an increase in land-use change Source: Authors, using model simulation results. relative to the baseline in China and the EU, and by increased production Removing subsidies to the livestock sector, on the other hand, gives in Brazil. surprisingly different results. Given the overwhelming importance of The second key message from this first set of simulations is that simple livestock—and particularly ruminants—in overall emissions from agricultural reductions in, or even removal of, all domestic support would not be production, the abolition of subsidies to livestock production might be sufficient to “bend the arc” on climate change. The results show that expected to substantially reduce global GHG emissions. But this does not while the overall reduction in global GHG emissions would be substantial appear to be the case. One reason for this result is that only about a quar- as a proportion of the baseline global level of GHG emissions in 2040 (1.5 ter of domestic support is targeted at livestock production (as shown in percent), or as a proportion of growth in emissions from agriculture and Table 3.3), and livestock production benefits from the substantial support land-use change over the period 2020 to 2040 (3 percent), this reduction to crops such as maize that are used for livestock feed, and which were not in emissions is an order of magnitude short of what is needed to stabilize removed in this scenario. In addition, much of the emissions from livestock the climate. This striking result follows from the limited apparent impact of are generated in countries with low levels of farm support: either advanced domestic support on global output and emissions under current produc- economies like Australia and New Zealand, or developing countries with tion technology and practices. large, low-productivity, and relatively high-emission intensity herds (for example, Ethiopia and India). While emissions from production would The third message is that the large amounts of public spending on decline with the removal of livestock subsidies, emissions from land-use domestic support appear to have low “value for money.” One notable change would increase slightly as cropland expanded, creating a marginal conclusion is that the gains from these subsidies in terms of incremental increase in emissions. Also, in this scenario per capita dairy consumption global farm output and farm returns appear to be quite small. Globally, worldwide would fall by 0.7 percent, including in developing countries, the equivalent of 14.4 percent of real farm value added was provided on where the average level of dairy consumption is already considered to be average between 2016 and 2018 as annual domestic support to producers. below the requirements for nutrition-adequate diets. Under the baseline scenario, this level of support would be maintained for the entire simulation period (to 2040). The model results show that Finally, the global aggregates mask shifts in production across when domestic support is removed, farm value added would fall by about countries. A case in point is the modest reduction in global output, which 5 percent. In other words, domestic support equivalent to 14.4 percent of reflects shifts in production across countries to meet global demand and farm value added—a substantial cost in terms of public expenditures— that is expected to continue to grow. Looking at the impacts of removing would “buy” a return of only 5 percent in value added, implying very low global domestic support for some key agricultural economies, the analysis value for money as a means of transferring income to farmers (OECD shows that the gains in economic welfare would be distributed unevenly 2003). If farm support is thought of solely as a means to provide transfers across countries. (See Table D.2, which shows outcomes for developed and 39   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 40  Simulating PolicyOptions to farmers, its implied transfer efficiency would be only about 35 percent. The key message emerging from these findings is that removing trade The primary causes of this outcome are the declines in prices resulting barriers involves stark trade-offs between economic efficiency and from the increase in supply and the increases in production costs resulting emissions. The combined effects of removing trade barriers (a rise from the distortions. While the lower prices would be a benefit to consum- in demand) and of eliminating domestic support (a decline in output) ers, particularly poor consumers, the distortions to production would be results in a smaller decline in output, and a larger increase in world prices a loss to the world. A policy of transferring income directly to producers for agricultural products than occurs when abolishing only domestic could, in principle, provide almost three times the benefit to farmers while support. The removal of trade barriers offsets some of the decline in global avoiding the incentive distortions that current forms of support create, an production volumes for crops and livestock observed in the first scenario, insight of the type that guided the McSharry reforms of the EU’s Common where domestic support is eliminated. The net result is that, with greater Agricultural Policy (OECD 2011). integration of domestic prices with world market prices, the change in farm incomes per worker (-3.5 percent) is smaller in this scenario than in the first (-4.5 percent).13 Economic efficiency gains would also be larger (about 4.3.2 Scenario 1b: Remove Both Domestic Support and Trade Barriers $135 billion), and poverty would decline slightly more in this scenario than The second experiment (Scenario 1b) removes trade barriers such as in the first one. But this more comprehensive reform would also reduce tariffs and quotas in addition to all domestic support. The main results the impact on global GHG emissions when compared with the scenario in are shown in Figure 4.6, and more detailed results are presented in the last which only direct support is removed. This finding is consistent with the column of Table D.1. It might be expected that because trade measures smaller decline in global agricultural output. A corollary to this outcome is are such a large share of total support, their elimination would significantly that the removal of trade barriers alone would deliver positive economic reduce both global agricultural output and GHG emissions. But, as noted efficiency gains, but also higher emissions. This is because as protected by Mamun, Martin, and Tokgoz (2021), this ignores an important distinction markets are liberalized with the removal of trade barriers, consumers between domestic support and trade measures: trade protection raises would demand more of these products, contributing to higher output and domestic prices, which depresses domestic demand for agricultural prod- emissions than would be the case under protected markets. ucts. So, while the protection is typically intended to raise the incomes To summarize, the results in this section provide important insights of farmers, its transfer efficiency at the global level is zero. Any benefit to into the complexities associated with simply removing all domestic farmers in protected regions is offset by the decline in global demand that support and trade barriers. The analysis of the implications of removing reduces world prices, by the losses to farmers in other countries, and by farm support is purely a thought experiment, designed to assess the the inefficiencies created in both production and consumption. implications of changes in farm support. A clear result of the analyses FIGURE 4.6: I mpact on GHG Emissions and Production of Removing All Support (% for domestic support only, and for all support, is that their effects on key change from baseline in 2040) policy goals would be mixed. While the increase in national income and All domestic support Trade Barriers & Dom. Support the reduction in emissions associated with removing support would be favorable, the associated reductions in farm output would make such a reform extremely challenging from a political economy perspective. A somewhat surprising finding is that the reduction in global emissions and -0.35 -0.49 in average net farm output would be similar (in fact, a bit lower) if both -0.55 trade measures and domestic support were abolished than if only domes- tic support were abolished. As discussed above, this reflects complex dynamics on both the demand and supply sides for agricultural products. -1.23 More concretely, it rests on three factors: the tax on consumers imposed -1.31 -1.48 Emissions Production Volume - Crops Production Volume - Livestock 13  The dynamics at the farmgate are much more complex and context-specific, depending on the mix of policies in effect at the time. They require more detailed analysis at the individual country level, which is beyond the Source: Authors, using model simulation results. scope of this global study. 4 1   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 42  Simulating PolicyOptions by market price support; the presence of substantial negative market price support. The analysis here is again at the global level, recognizing that support in a few major producing and consuming countries; and the fact while policy action needs to be taken at the individual country level, that removing trade protection raises world prices, and hence the prices these actions impact global outcomes such as climate change through received by producers. important transboundary effects—both directly, and indirectly through trade. Achieving a shared global goal thus will require an internationally The impacts on the costs of healthy foods shown in Figure 4.4 point concerted and broadly accepted agenda of policy shifts to make to trade-offs between economic, environmental, and nutritional meaningful progress. Repurposing agricultural policies and support also outcomes. These results reflect a mix of nutritional outcomes. If trade involves significant political economy challenges that must be considered barriers were eliminated, increases in the consumption of dairy products as part of a broader strategy for making agriculture more sustainable, as in developing countries would likely contribute to better nutritional out- discussed in Appendix E. comes. Likewise, in all but one of the simulations involving the removal of domestic support, the sugar consumption per person would decline, which Three broad categories of reform options are examined: they all aim to would also contribute to better diets. By contrast, in the simulation that maintain the current level of support for agriculture, but to redirect and includes the removal of trade barriers, sugar prices would fall, and demand deliver it in more beneficial ways. The analytical framework used in this would increase, which would worsen the quality of diets. Nevertheless, study allows for analysis of the impact of a broad range of policy options in all scenarios the average cost of the “healthy diet” food basket would on the triple goals of reducing GHG emissions; making gains in farm effi- increase. As a result, the share of the global population unable to afford ciency and income, and hence poverty; and achieving improved nutritional healthy diets would also increase in all cases. outcomes. The three categories of options considered in this study are representative of a range of specific policy options that are conceptually Overall, the baseline and support removal scenarios yield sobering similar, but that need to be tailored to individual country contexts. results. With sustained growth in demand as population and incomes continue to grow, simply removing all agricultural support, even if a global Scenario 2: Restructuring domestic support within the current subsidy consensus were achieved to do so, would be insufficient to achieve large budget includes two different experiments. In the first, Scenario 2a, the enough reductions in global GHG emissions to appreciably reduce agricul- budget is redistributed uniformly across all products. Under Scenario ture’s impact on the climate, and nature. This is because, despite shifts in 2b all support is transferred to low carbon-intensity products. production across countries and gains in economic efficiency associated Scenario 3: Conditionality makes the availability of domestic support with the removal of incentive distortions, there would be only modest conditional on producers switching to products or production process- changes in global output as world prices adjusted to the removal of es that are less environmentally harmful (for example, less GHG-inten- support. These results, as well as the adverse impacts on farmer incomes, sive), using currently available technologies. poverty, and nutrition, suggest that “simple” policy options like removing all domestic support and border distortions are naïve. The real trade-offs Scenario 4: Repurposing for green innovation redirects a portion of they entail make such actions extremely challenging politically, and likely the public expenditures currently being spent on domestic support to infeasible. Together with the key finding on the low value for money of invest in the development, dissemination, and adoption of new green the current large volume of domestic support provided to agricultural technologies that both reduce emissions and increase productivity. producers, the results point to the imperative of exploring other options The balance of the domestic support goes back to the taxpayers and is for repurposing current policies and support, to identify much-needed potentially available to deliver as nondistorting transfers to producers “win-win” outcomes. and other stakeholders to compensate them for potential losses due to the reform, and for spending on extension services, rural infrastructure, 4.4 REALIGN AGRICULTURAL POLICIES AND SUPPORT FOR and other essential public goods and services that are fostering agri- BETTER OUTCOMES cultural and rural development. Given the trade-offs and the associated political economy dilemmas The restructuring simulations consider moving from the current, highly involved in reducing or removing support—despite its low value for differentiated set of subsidy rates across outputs, inputs, and factors money—this section explores potential options for repurposing current in two less distortionary directions. Scenario 2a moves to a uniform rate 4 3   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 44  Simulating PolicyOptions across all agricultural products, while Scenario 2b moves to a uniform to lower-yielding technologies requiring an increase in agricultural land subsidy rate on non-emission-intensive products. Both simulations are use to meet global food demand. The direct impacts of reducing polluting based on uniform domestic support rates, but they keep the average inputs can be seen relatively easily since they are tracked in the emissions rate of support unchanged from the level of support in 2020. By reducing modeling framework used in this study. The impacts of reductions in the currently very uneven spread of support across commodities, these productivity are much more wide-ranging, involving changes in the alloca- simulations also move in the direction of decoupled transfers: that is, tion of land and in the product mix, and require a global general equilibrium direct income transfers that are not tied to specific commodities or inputs. approach of the type used here if their full impacts are to be accounted for. The available literature points to indicative values for the productivity In the “conditionality” simulation (Scenario 3), support is conditioned impacts of moving to organic agriculture. Ponisio et al. (2014) found a on farmers’ willingness to provide environmental services. There is smaller reduction in yields from moving to organic agriculture than earlier strong evidence that in countries where there is substantial support studies, but still estimated a decline of 19.2 percent, while the survey by to farmers, cross-compliance conditions can increase the adoption of Seurfert, Ramankutty, and Foley (2014) put the associated yield reductions sustainable agricultural practices (Piñeiro et al. 2020). These policies between 13 and 34 percent. If, for illustrative purposes, the productivity of frequently involve reductions in the use of chemical inputs such as fertiliz- organic agriculture is around 20 percent less than that of nonorganic, then ers and pesticides, and sometimes more comprehensive moves to organic the EU’s proposed requirement for 25 percent organic production would agriculture that reduces emissions. If farmers are minimizing costs, as is translate into an average productivity reduction of 5 percent. assumed in the modeling performed for this study, then requiring them to produce using approaches they have previously rejected can be expected The specific conditionality scenario used in this study assumes a to result in higher costs and lower productivity. Consequently, policies of reduction in both productivity and emission intensities of 10 percent. this type are also likely to involve compliance-monitoring challenges akin While any such scenario is inherently arbitrary, this scale of shock to to those seen with organic food certification (Parker 2021). productivity seems broadly consistent with the impacts of the EU’s Farm to Fork proposals (European Commission 2020b). The goal of this scenario An example of this broad approach in industrial countries is the use is to provide insight into the impact of a potentially plausible policy of enhanced conditionality in the European Union’s future Common reform—one that could be scaled for greater impacts on productivity or Agricultural Policy (CAP) proposal (European Commission 2020a). This emission intensities. seeks to achieve reductions in the emissions associated with specific inputs, while compensating farmers for providing environmental services Impacts on emissions and on productivity are uncertain. Any such by adopting technologies that they otherwise might not adopt, or that policy must carefully assess whether that technology really has higher might be less productive than the technologies they currently apply. The private productivity than the ones that producers would otherwise have EU’s Farm-to-Fork proposal (European Commission 2020b) includes the chosen. If the technology on which support is conditioned has lower condition that in exchange for direct payments farmers should strive to productivity, and hence would require expansion of global agricultural land reduce pesticide use by 50 percent, chemical fertilizer use by 20 percent, use, a key question is whether the emissions associated with the resulting and antimicrobials by 50 percent, while increasing the share of organically land-use change will outweigh the lowered emissions. This question has farmed output to 25 percent. The reduced use of chemical inputs is been addressed for individual countries (see, for example, Smith et al. pursued both because of local externalities (risks to land and water quality, 2019), but not, to the best of the authors’ knowledge, on a global scale. and public health), as well as because of global externalities from GHG The repurposing for green innovation simulation (Scenario 4) redirects emissions. The move to organic agriculture is driven by similar motivations, a part of current domestic support to investment in new technologies including to improve soil quality and reduce GHG emissions. that overcome the limitations or trade-offs associated with current Any conditionality scenario of this type can be expected to have two technologies. This option focuses on green innovations, that is technolo- potentially offsetting impacts on GHG emissions. The first is direct gies and practices that reduce emissions while increasing productivity. Cli- reductions in emissions per unit of output as the use of polluting inputs mate-smart agriculture (CSA) is one approach to this goal. This approach, declines. The second is the likely increase in emissions from land-use promoted in many developing countries (Bell et al. 2018), seeks to achieve change with a move away from the current technologies chosen by farmers three objectives: (1) increasing productivity; (2) increasing resilience; and 4 5   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 46  Simulating PolicyOptions (3) mitigating climate change by reducing emissions. De Pinto et al. (2020) and Hodgson 2021), together with emissions from rice, which account for use crop modeling techniques to argue that widespread adoption of CSA an additional 9 percent, they present an extremely important opportunity techniques could sharply increase agricultural productivity. Where it is for mitigating emissions of this potent GHG. Ocko et al. (2021) see the feasible, approaches that can do this are highly desirable because they potential to lower methane emissions from livestock by 30 percent and can help raise productivity while also lowering GHG emissions. rice by 49 percent using currently available methods. They see this as part of a package of methane emission reductions that—alone—could slow the A key challenge, which is addressed in the paper by Bell et al. (2018), is global mean rate of warming by 30 percent by midcentury. The reduction to identify approaches that will better support the objectives of CSA, or in emissions from agriculture is also consistent with the Global Methane other agroecologically sound approaches, rather than the technologies Pledge for rapid reductions in methane emissions that was supported by and practices currently in use. If the current lack of adoption of these more than 30 countries in the lead-up to COP26 in Glasgow (US Depart- technologies is due to a lack of information, high capital costs, or a need ment of State 2021). for adaptive research to meet particular production conditions, then approaches that will alleviate the associated market failures are better than To turn these aspirations to outcomes will require investment in blunt instruments that induce compliance by, for example, conditioning research and development (R&D). The urgency of such action takes on support on the adoption of new technologies, or simply by regulatory fiat. added importance in light of two key findings from a new study from the Commission on Sustainable Agriculture Intensification (CoSAI) (Dalberg The main challenge is creating green innovations that will achieve these Asia 2021). First, while funding for agricultural innovation across the Global outcomes. Some such innovations already exist or are emerging and have South14 has been increasing, it remains low. It is highly concentrated (China, been proven effective in some contexts. Based on an examination of the Brazil, and India account for 60 percent of public funding for R&D); and its evidence provided by recent literature, the specific simulations used in rate of growth is decelerating. The second finding is that currently only a this report assume a 30 percent reduction in emissions per unit of output, small fraction (7 percent) of the current $60 billion spent on innovation is and a 30 percent increase in productivity. These assumptions are within targeted at sustainable intensification.15 the range observed with key new technologies. (See, for example FAO 2016). A 30 percent reduction in emissions is broadly consistent with the Public support for future R&D requires careful design. The final sim- potential that was identified by Mernit (2018) for ruminant feed supple- ulation (Scenario 4) increases investment in research and innovation to ments and analyzed by Laborde et al. (2020). Runkle et al. (2019) found develop technologies that target reducing emission intensities and raising reductions of 65 percent in methane emissions from rice production, with productivity. A key feature of this scenario is that it demonstrates the substantial water savings and no yield loss, using alternate wetting and critical role of innovation in achieving the desired “triple wins.” To highlight drying practices. More recent evidence suggests that the reductions in this, a subsidiary simulation (Scenario 4a), where green innovations emissions and in the costs associated with livestock feed additives may are assumed as “manna from heaven” (that is, costless to taxpayers), be substantially higher than the 30 percent considered here. For example, demonstrates the importance of innovation in driving appreciable gains Kinley et al. (2020) found emission reductions of 40 to 98 percent, and on the “triple wins.” In addition to investing in the development of green weight gain improvements between 42 and 53 percent in cattle. And innovations even where technologies have a strong demonstrated capa- Chang et al. (2021) have highlighted substantial reductions in emission bility to raise productivity and reduce emissions in specific contexts—as, intensities associated with livestock production in the past two decades; for example, in the Kinley et al. (2020) experiment with cattle in Townsville, they conclude that improvements in livestock production efficiency for Australia—considerable investment in adaptive research and dissemination achieving emission reductions show much more promise than efforts to may be needed before they can be adopted more widely. Based on change consumer demand patterns. available studies, a rough indication of the cost needed to achieve such productivity gains might be given by the benefit-cost ratio investments Reducing methane emissions is a high priority if global warming of over for rural R&D of 10 found by Alston, Pardey, and Rao (2020). If the benefits 1.5°C is to be avoided. The Energy Transitions Commission (2021) and Wolf of this type of R&D follow the 50-year distributed-lag they identified, (2021) both highlight the need for much greater emphasis on reducing methane emissions than is provided under the current nationally deter- mined contributions (NDCs). Since emissions from enteric fermentation account for over a third of anthropomorphic methane emissions (Terazono 14  CoSAI uses the World Bank’s definition of the Global South, which includes Asia (excluding Japan, Singapore, and South Korea), Central America, South America, Mexico, Africa, and the Middle East (excluding Israel). 15  Total of current funding by the domestic and development partners, and the private sector. 4 7   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 48  Simulating PolicyOptions a sustained increase in output of just over 30 percent16 would require CHANGE IN EMISSION SCENARIO LABEL REGION INSTRUMENT COEFFICIENT an investment equal to one percent of agricultural output. To finance Green Innovation – Developed Agricultural TFP =+30%; additional the needed investments, this scenario considers repurposing part of 4c publicly funded countries only 1% of Ag Output spent on R&D -30% current domestic support to agriculture that, based on past investment Green Innovation – Developing Agricultural TFP =+30%; additional 4d publicly funded countries only 1% of Ag Output spent on R&D -30% returns, would be enough to generate the 30 percent increase in output. Alternatively, green innovations could be financed through additional Note: TFP=Total factor productivity; GI=Green Innovation; ARD=Agriculture and Rural Development; PG=Public Goods. public funding. Another subsidiary simulation (Scenario 4b) shows that FIGURE 4.7:  Global Implications of Repurposing Domestic Support (Percentage Change Relative to this would yield similar but slightly smaller gains on key outcomes. Many Baseline Projections for 2040) countries face fiscal constraints that would complicate the public funding option. This constraint has become even more binding as many economies ECONOMIC FARM SECTOR struggle to recover from the COVID-19 pandemic (Laborde, Martin, and (Real national income, % change 2040) (Agricultural production volume, % change 2040) Vos 2020). Another consideration is the current low intensity of public Removal of Dom. Support spending on research and innovation, which is further declining in the very Removal of all support countries where it is needed the most (Fuglie et al. 2020). Uniform subsidy The following discussion compares results from the restructuring Target CO2 efficient products scenarios (Scenarios 2a and 2b) with a conditionality (Scenario 3) and a Env. Conditionality CROPS LIVESTOCK repurposing scenario (Scenario 4). The key features of these scenarios are set out in the shaded rows in Table 4.1 (Scenarios 2a, 2b, 3, and 4). Results Repurposing for GI for subsidiary simulations (3a, 3b, 4a, 4b, 4c, and 4d) are similar to those for -3.2 -2.2 -1.2 -0.2 0.8 1.8 -20.0 -10.0 0.0 10.0 Scenarios 3 and 4, and are given in Table D.3. Focusing the discussion on the key outcomes of interest: the main results related to the overall economy, SOCIAL DIETS farm production, social outcomes, people’s diets, emissions, and the effect (Poverty at PPP$1.90, % change 2040) (Healthy food prices, % change 2040) on nature are summarized in Figure 4.7. This figure also shows results from Removal of Dom. Support the previous scenarios that simulate the removal of current support, to put the magnitudes of the projected impacts in perspective. Removal of all support Uniform subsidy TABLE 4.1: Scenarios Considered Target CO2 efficient products CHANGE IN EMISSION SCENARIO LABEL REGION Env. Conditionality INSTRUMENT COEFFICIENT 2a Uniform subsidy World To weighted average None Repurposing for GI Uniform on non-CO2 To weighted average for non-CO2 -1.3 -1.1 -0.9 -0.7 -0.5-0.3 -0.1 0.1 0.3 0.5 0.7 -20 -15 -10 -5 0 5 10 2b -intensive products World intensive products, 0 otherwise None 3 Conditionality World Agricultural TFP= -10% -10% CLIMATE NATURE Developed (Reduction in emissions from agriculture and (Agricultural land, % change 2040) 3a Conditionality countries only Agricultural TFP= -10% -10% land use, % change 2040) Developing 3b Conditionality countries only Agricultural TFP =-10% -10% Removal of Dom. Support Agricultural TFP =+30%; repurpose 1% of Ag Output equiv. of domestic Removal of all support 4 Repurposing for GI World support to invest in R&D, with rest -30% available for other ARD PGs Uniform subsidy Green Innovation – as Target CO2 efficient products 4a “manna from heaven” World Agricultural TFP =+30% -30% Green Innovation – Agricultural TFP =+30%; additional Env. Conditionality 4b publicly funded World 1% of Ag Output is spent on R&D -30% Repurposing for GI -40 -30 -20 -10 0 10 20 -2.5 -1.5 -0.5 0.5 1.5 Source: Authors, using model simulation results. Note: Brown bars indicate movement toward, and teal bars indicate movement away from achieving the related 16  At a discount rate of 5 percent. SDG(s). GI= Green Innovation. 16  At a discount rate of 5 percent. 4 9   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 50  Simulating PolicyOptions 4.4.1 Scenario 2: Shift to Less Distorting Forms of Support and commodities without changing current technologies and practices may Lower-Emitting Activities have surprisingly complex results, and might not necessarily help to reduce overall emissions. Maintaining the current level of support, but changing the patterns would offer only small economic, social, and environmental gains. The two restructuring experiments address the question of whether it is the 4.4.2 Scenario 3: Condition Support on Environmental Services pattern of current direct supports, or their level that most affects their Making support conditional on reducing emissions would be positive for impacts on economic, social, and environmental outcomes. In the first planetary health. The conditionality scenario delivers greater environmen- experiment, moving from the current patterns of support to a uniform tal benefits than the earlier scenarios considered, despite an increase in output subsidy with the same budget cost would have only modest the amount of land being used for agriculture. Emissions from agricultural impacts. Surprisingly, real national income would fall, albeit very slightly, a production would fall by 19 percent, driven by the decline in emissions per second-best result consistent with budget support being transferred away unit of output. This reduction in emissions from production would be only from lightly protected commodities to commodities with greater support partially offset by an increase of almost 4 percent in emissions from land- from border measures. Overall, agricultural prices would fall, with the net use change as the sector drew in more land to offset the adverse impact result that farm incomes per worker would also fall slightly. On the plus on productivity. Therefore, there would be a net reduction in emissions of side, reductions in the prices of dairy products would raise consumption 15 percent. levels, modestly reducing the costs of healthy diets. Emissions from agricultural production would increase slightly by 0.5 percent, but this But conditionality might also entail important and surprising trade-offs increase would be more than offset by a decline of 1.1 percent in land-use for people and for economic prosperity. Shifting to production methods emissions, with total emissions from agriculture and land use falling by and practices that improve environmental outcomes but reduce the 0.7 percent. productivity of land does come with economic and social costs. Real gross national income (GNI) would decline by 0.8 percent, or about $1.21 Reallocating support away from the most emission-intensive agricul- trillion in 2040, compared to the baseline projection in 2040, because this tural commodities to other agricultural commodities might not reduce experiment involves a decline in productivity in an important sector. With emissions, since it would encourage increases in agricultural land use. this decline in productivity, agricultural production would fall by more than In this experiment, support is shifted away from high-emission livestock 5 percent. The decline in output would raise world food prices by a sub- production and rice toward other agricultural commodities, mostly crops stantial 12.7 percent, contributing to an increase in the poverty headcount. that have much lower emission intensities. This scenario would also reduce The decline in productivity, and the consequent 10 percent increase in the real national income, but again only slightly. Production of the highly traded simulated cost of a healthy diet food basket would also cause per capita grains and other non-livestock commodities would expand, while livestock consumption of healthier foods to decline—dairy product consumption production would fall slightly. With a reduction in overall prices driven by would fall by 6.4 percent, and vegetable and fruit consumption by more expanding production of non-livestock products, real farm incomes per than 4 percent. The decline in productivity would also draw additional worker would fall, but less so than in the first scenario. Consumption of resources into the sector —agricultural land use would increase to offset dairy products and fats would decline, while vegetable consumption would the decline in productivity, as would farm employment, slowing structural increase slightly. However, the biggest dietary impact by far would be a transformation. Finally, increased use of land for agriculture and the related 14 percent increase in the consumption of sugar, as increased support loss of forest habitat would incur further biodiversity losses. Against the for production interacts with relatively low demand elasticities. On the backdrop of these developments, the simulated increase in farm income favorable side, the cost of a healthy diet dominated by non-livestock associated with a global reduction in agricultural productivity might seem products would fall by almost 2 percent. However, global GHG emissions surprising, but it is a consequence of the relatively low price elasticities of would increase slightly in this scenario because the decline in emissions demand for agricultural products; that is, food prices would rise more than caused by lower agricultural production would be outweighed by increased proportionately with the decline in output. emissions from deforestation, even though pastureland would be retired with the reduction in livestock production. This experiment suggests that, It is important to remember that these results relate to a move to a while appealing, ideas like shifting subsidies away from emissions-intensive presumed lower productivity technology by all countries. Moving to a 5 1   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 52  Simulating PolicyOptions lower productivity technology in just one individual country would have 4.4.3 Scenario 4: Repurpose Support to Target Emission Reduction exactly the opposite effect on farm income, reducing the volume of output and Productivity Enhancement for sale without a strong compensating increase in prices. A country or an The repurposing scenarios illustrate the impacts of green innovations individual farmer that moves to a lower productivity technology while the that have an assumed 30 percent increase in agricultural productivity rest of the world turns to higher productivity options faces the technology along with a 30 percent decline in emission intensity. The final sim- treadmill problem identified by Cochrane (1958). Farm returns go down ulation presented in Figure 4.7 (Scenario 4) refers to the case in which both because the decline in productivity reduces output and technical domestic support is repurposed to invest in productivity increases, with progress elsewhere lowers output prices. resources equivalent to 1 percent of agricultural output (about $70 billion The conditionality experiments for developed and developing countries of the $244 billion provided as domestic support annually from 2016-18), have substantially smaller impacts than conditionality at the global and redirected to invest in the development of productivity-enhancing and level. Weighting the percentage changes in real farm income for each emissions-reducing technologies and practices. The rest of the domestic country group by its income share would suggest a much smaller increase support would be returned to taxpayers and potentially available to in real farm income than is seen with global implementation. (See Columns deliver as non-distorting transfers to producers and other stakeholders, 3a and 3b in Table D.3). This difference arises because when conditionality to compensate them for any losses they might incur; to finance incentives is introduced in both developed and developing countries, its effects on for the widespread adoption of green innovations, or to spend on other world prices cumulate, increasing key impacts such as the rise in real farm underfunded agricultural public goods and services such as agricultural income and the pressure to use more land in agriculture. An important infrastructure; and to foster broader agricultural and rural development.17 difference is the impact of conditionality on poverty in developing coun- The importance of innovations in driving the results in this scenario are tries. Poverty rises much more when conditionality is used in developing highlighted in the subsidiary simulations, the first of which assumes countries than in rich countries because most poor people in developing productivity “shocks” to come as “manna from heaven”: that is, they countries are farmers. These results share many features with recent are assumed to be exogenous and costless to taxpayers (Scenario 4a). analyses of the EU’s Farm to Fork proposals, which indicate that those Additional subsidiary simulations (Scenarios 4b–4d) consider financing proposals would lead to yield reductions and increased agricultural land such innovations through additional public resources. These show similar, use (Barreiro-Hurle et al. 2021; Henning and Witzke 2021). albeit slightly muted results on some outcomes—for example, a reduction in emissions and agricultural land, and gains in real national income and These conditionality scenarios are potentially very interesting thought structural transformation—and are not discussed here. (See Table D.3 for experiments to foster policy dialogue. The reduction in emissions turns the full set of results). out to be more or less proportional to the reduction in agricultural pro- ductivity. This finding highlights the importance of key links that are often Repurposing support toward targeted productivity investments has the overlooked. First, assuming effective enforcement of the conditionality, the potential to deliver large gains through improved economic efficiency, productivity loss would lead to lower agricultural production, compounding reduced environmental impacts, and better health outcomes. The the reduction in emissions per unit of output. However, the decline in broad impacts of targeted productivity investments are evident from the productivity would induce farmers to expand the amount of land used “productivity” and “repurposing” scenarios shown in Figure 4.7. The dis- for agriculture, leading to higher emissions from land-use change. Thus, it cussion here focuses on the repurposing outcomes, since the productivity becomes particularly important to scrutinize proposals for conditionality simulations give very similar results. Aggregate (world) GNI would be higher to assess their potential impacts on productivity. The validity of assump- than the projected baseline scenario GNI for 2040 by around 1.6 percent, tions that any productivity losses would be small, or that productivity implying a substantial payoff—equivalent to $2.4 trillion in 2040.18 would actually increase, therefore need to be carefully assessed. On average between 2016 and 2018, agricultural domestic support is estimated to have been about 6.6 percent 17  of global agricultural GDP. The equivalent of 1 percent of agricultural GDP would imply repurposing about 15 percent of current domestic support toward targeted emissions-reducing public productivity enhancements. Applying the projected growth rate for GNI from the repurposing scenario to the World Bank’s World 18  Development Indicators (WDI) estimate of the average annual global GNI for 2011–18 of $81.5 trillion would result in a 2040 global GNI of $149.4 trillion compared to the baseline 2040 projection of $147.1 trillion (World Bank 2021). 5 3   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 54  Simulating PolicyOptions This reflects a substantial gain in economy-wide efficiency.19 However, the If unskilled farm and nonfarm labor were perfectly substitutable, labor large global productivity shock and the low elasticity of demand for agri- would move seamlessly out of agriculture, with the returns to labor rising in cultural products would drive prices down by 21 percent as the production both sectors. The simulation indeed shows quite rapid structural transfor- of crop and livestock products would rise by 16 percent and 11 percent, mation, with nearly 11 percent of farm workers shifting from farming to other respectively. Because the increase in national income and the “savings” in activities. However, transforming farm labor into nonfarm labor is often public expenditure from the removal of the remaining domestic support quite difficult, in large part because the educational opportunities for rural would be much larger than the fall in farm income, it would be possible to youth tend to be much more limited than for urban youth, and specialized compensate farmers for any potential loss in income associated with lower agricultural skills are often less useful in employment outside agriculture. farmgate prices as a result of the productivity increases. Such “frictions” and other forms of labor market rigidities are well rec- ognized in the literature. (See, for example, Gollin, Lagakos, and Waugh Importantly, these green innovations would deliver huge benefits 2014; Herrendorf and Schoellman 2018; and Hicks et al. 2017). To account for climate and nature. Overall emissions from agriculture would fall for these differences, a constant elasticity of transformation (Powell and by a substantial 40.5 percent, or nearly 2.8 Gt CO2eq—avoiding nearly Gruen 1968) parameter of 0.9 is used by default within the model. This 80 percent of the incremental emissions expected under the baseline is below the 1.32 value used by Ianchovichina and Martin (2004); the 2.2 (business-as-usual) scenario between 2020 and 2040. Emissions from estimated by Sicular and Zhao (2004, 257) for China; and the 3.7 estimated production would fall by 24 percent, as efficiency gains significantly by Wang and Matthews (2011), using more recent data for China. reduced input use. In addition, about 2.2 percent of agricultural land would move from agricultural use back to its natural uses, resulting in a 16 per- The baseline results suggest that limited movement of labor out of cent reduction in emissions from land-use change. The decline of about agriculture in response to the productivity shock would lower real 105 million hectares of land under agriculture would deliver substantial wages in agriculture and returns per worker. This challenge can only ecological benefits through the restoration of natural habitats and reduced be overcome if farm labor can more readily transition into remunerative biodiversity loss. This scenario would not only avoid the need to add 56 nonfarm work. To explore this hypothesis, a range of simulations was million hectares of agricultural land use between 2020 and 2040 as in the performed raising the elasticity from the baseline value of 0.9 to 25 (Table baseline scenario; it would also allow another 48 million hectares of current 4.2). The results confirm that greater mobility indeed offloads the down- agricultural land to be restored to natural habitats. ward pressure on real agricultural wages, and wages would actually rise for elasticities of transformation of 2 and above, well within the degree of labor Productivity-driven growth would also reduce poverty and generate mobility observed in the literature. With very high mobility (at an elasticity nutritional benefits. Poverty measured against a poverty line of $1.90 pur- of 25), real agricultural wages would increase by almost five percent. Real chasing power parity (PPP) declines substantially (about 1 percent) when farm income per worker, however, would decline even though returns to productivity rises in developing countries. The composition of diets would labor would rise. This is because elasticities of demand for food are low, shift substantially as the cost of healthier food declined and incomes rose; so the decline in prices would exceed the increase in the quantity of food the consumption of dairy products would increase by 16 percent and the demanded, pushing down total returns from food production, and hence consumption of vegetables and fruit by 12 percent. Overall, the cost of a the returns to agricultural land healthy diet would fall by a remarkable 18 percent and would be expected to drive large increases in the consumption of nutrient-dense foods. At the TABLE 4.2: Impact of Labor Mobility on Real Farm Wages same time, however, the prices of unhealthy foods would also fall, explain- ing the increase of almost 28 percent in sugar consumption. LABOR ELASTICITY OF TRANSFORMATION 0.9 2 10 25 Overall incomes, including for farm workers, would be expected to rise Real Farm Income per Worker -4.5 -3.4 -2.2 -2.0 as productivity increases accelerated the process of economic trans- Real Farm Wage Rate -3.4 0.1 4.0 4.7 formation; but there would be important transitional challenges. Source: Authors, using model simulation results. The gain of 1.6 percent in real national income is significantly higher than the 1.1 percent implied by a 30 19  percent increase in productivity applied to the share of agriculture in the world economy of 3.5 percent. 5 5   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 56  Simulating PolicyOptions One key message that emerges from this analysis is the need to 4.4.4 The Impact of Individual Country Actions invest in human capital and skill development, and to implement rural The rise in agricultural incomes and in employment when productivity development policies that help to create new and better nonfarm declines, and the decline in incomes when productivity rises, may employment opportunities. Since these simulations are run over close to seem counterintuitive. But these outcomes are a natural consequence of a generation, there is time to make investments in the skills of farm chil- the low income and price elasticities of demand for food, and the global dren so they will have more employment options during their working lives. nature of the experiments reported, as shown in Figure 4.7 and Table D.3. With improved educational opportunities, an enabling business environ- They are consistent with the results Matsuyama (1992) found for the world ment that encourages new businesses and employment opportunities to as a whole and for individual, closed economies—that when agricultural emerge, and other rural development policies, the barriers to mobility out productivity rises, prices fall and employment declines. of agriculture would decline considerably. To evaluate outcomes from individual country actions, which may be The results when only developed countries, or only developing coun- more likely to happen than a concerted global reform, the next set of tries adopt improved technologies show smaller impacts than when experiments simulated the impact of country-specific productivity-in- all countries do it on a concerted basis (See Scenarios 4c and 4d in crease and emission-reduction scenarios. The main results found for Table D.3). However, in contrast with approaches where unilateral action seven major agricultural countries (Brazil, China, Ethiopia, India, Indonesia, is undermined to some degree by “leakage” to nonadopting countries, the the EU, and the US) are summarized in Figure 4.8, with detailed results sum of the gains from individual adoption are greater than the gains from presented in Table D.4. These simulations provide results that are intuitive full adoption. This is because a country that adopts productivity-enhanc- and generally consistent across the major countries. As with the produc- ing practices gains market share from those that do not. tivity experiments in Figure 4.7, national real income would rise in each An outstanding question is why countries continue to underinvest in country; world prices would decline; agricultural production volumes would agricultural R&D and in supporting wider adoption of new technologies increase very substantially, especially for crops; food prices would decline despite high returns on investment. The puzzle of carefully documented and food consumption would rise; poverty would decline everywhere that and consistently high returns to investment yet persistently low allocations measurable poverty remains in 2040, with a particularly large decline in of resources to agricultural R&D and innovation was explored in detail in Ethiopia; overall emissions would decline but by very different amounts per a recent World Bank study (Fuglie et al. 2020). This study identifies the country; and agricultural land use would decline in every country except implementation challenges for raising productivity and lays out a compre- Indonesia. These outcomes are consistent with individual countries having hensive agenda that calls for a combination of revitalizing public research, incentives to adopt productivity and emission-reducing innovations. The spurring private R&D, and promoting the adoption of available technol- fact that each country, and each farmer, has an incentive to adopt the ogies, particularly by smallholders in developing countries. Briefly, these improved technology to gain market share and to raise its farm returns, include simultaneous actions on both the supply and demand sides of the gives this approach an important advantage over alternative approach- innovation puzzle. On the supply side, the priority is to increase investment es—like conditionality or carbon-tax-based approaches—where individual in R&D. This would require (1) increasing publicly supported R&D for countries and individual farmers have an incentive not to adopt. invention, adaptation, and dissemination of new technologies (for example, In large countries, substantial increases in productivity would depress by increasing, stabilizing, and diversifying funding; incentivizing scientists global prices, but not as dramatically as in the case of a global increase and strengthening universities and public research institutions; aligning in productivity. The increases in output would be substantial, ranging priorities with user needs; and partnering with foreign and international from 25 to 35 percent for crops, and 17 to 29 percent for livestock. Output researchers); and (2) mobilizing the private sector to invest in research responses would have been even larger except for the declines in output and innovation through market liberalization; regulatory reform; intellectual prices, with output increasing not only because of the increases in output property rights; and catalytic and complementary public R&D. On the per unit of resources used, but by drawing additional resources into the demand side, the priority is (1) to remove the constraints to smallholder sector (Martin and Alston 1997). adoption of technologies by addressing outstanding issues in the enabling environment and improving advisory services and access to finance and markets; and (2) to invest in human capital and capabilities. 5 7   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 58  Simulating PolicyOptions FIGURE 4.8: Impacts of Country-Specific Repurposing Scenarios (Percentage Change worker would rise for all the economies considered, except for China and Relative to Baseline Projections for 2040) the EU. In all cases, world prices would decline by much less than they ECONOMIC FARM SECTOR do in the productivity and repurposing scenarios shown in Figure 4.8. In (Real national income, % change 2040) (Agricultural production volume, % change 2040) Brazil, Ethiopia, India, and Indonesia, the favorable impact of the increase in output would more than offset the more modest decline in prices, and real United States United States EU EU farm incomes would rise. For China, the EU, and the US, the price declines Indonesia Indonesia would be larger. The net result is that real farm incomes per worker would India India fall slightly in China and the EU, while farmers in the more export-oriented Ethiopia Ethiopia US economy would see a rise in per-worker income. Agricultural land use China China would decline in all of these countries except Indonesia. The impact on Brazil Brazil global emissions would differ substantially across countries. The decline in 0 2 4 6 8 0 5 10 15 20 25 30 35 emissions from production would be particularly large in China, the EU, and the US, and the share resulting from land-use change would be particularly SOCIAL DIETS large in Brazil. (Poverty at PPP$1.90, % change 2040) (Healthy food prices, % change 2040) United States United States EU EU Indonesia Indonesia India India Ethiopia Ethiopia China China Brazil Brazil -0.8 -0.6 -0.4 -0.2 0 -20 -15 -10 -5 0 CLIMATE NATURE (Agricultural emissions, % change 2040) (Agricultural land, % change 2040) United States United States EU EU Indonesia Indonesia India India Ethiopia Ethiopia China China Brazil Brazil -60 -50 -40 -30 -20 -10 0 -2 -1.5 -1 -0.5 0 0.5 Source: Authors using model simulation results. Note: Brown bars indicate movement toward, and teal bars indicate movement away from achieving the related SDG(s). Finally, in addition to the positive and desirable outcomes, an important shift seen in country-specific productivity scenarios is the impact on real farm incomes per worker. As indicated, when productivity rises in unison across all countries, it pushes prices down and reduces farmer returns to productivity growth. With country-specific productivity increases, as shown in Figure 4.8 (and Table D.4), real farm income per 59   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 60  Simulating PolicyOptions 5 The repurposing agenda outlined here may sound simple to implement but it would require a great deal of additional careful analysis and research. This study makes a modest attempt to analyze the potential outcomes associated with alternative options for repurposing agricultural policies and support, but much work remains to be done. While a lot of thought has been given to allocating research resources toward increasing productivity, little attention has historically been paid to approaches that reduce the emission intensity of production. It seems likely that increasing productivity and reducing emissions are strongly complementary research outcomes. Much of the emission generation from agriculture is the result of inefficiency in the production process. If, for instance, methane could be used to produce desired outcomes, rather than emitted into the atmosphere from ruminant digestion or flooding rice fields, productivity could potentially be increased substantially. The striking combination of lower emissions and higher productivity growth reported by Kinley et al. (2020) in their experiments with cattle suggest that this potential complementarity can be harnessed through research focused on these two goals. Another, less direct, indica- tion that productivity growth and emission reductions are complementary comes from the generally lower emission intensities, particularly for beef, observed in developed countries relative to developing countries (see Table 3.6). The longer history of R&D in these countries has resulted in gen- AVENUES FOR erally higher yields for crops and more rapid growth of livestock that have, in turn, reduced emissions from their production. These outcomes have occurred without a strong focus on emission reduction in the agricultural R&D programs of these countries. Going forward, the key is to identify the FURTHER POLICY innovations that are the most effective in both reducing emissions and increasing productivity. More general rethinking by economists and policymakers about the ANALYSIS: toolbox for dealing with collective action problems also seems to be required. The traditional toolbox focuses on internalization of the externalities and pays little attention to identifying technical changes that might mitigate the problem of collective action. Clearly, the set of IMPLEMENTATION technical changes that might contribute to solving the problem is very large. Increasing the efficiency of production can clearly help. Reducing food loss and waste can similarly help, by reducing the resources and inputs needed for production. The traditional environmental economist’s concern that increasing the productivity of food production might increase emissions by lowering the cost of food and increasing demand is not warranted when agricultural productivity growth is broadly based. While the falling cost of food does increase demand, the low demand elasticity R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 62  Avenues for Further Policy Analysis: Implementation means that the demand for agricultural land is also likely to decline, on changing output levels, making it less effective than a carbon tax on bringing about an associated reduction in emissions from land-use change. emissions from the combustion of fossil fuels. If a BCA is to be introduced, Thus, a 30 percent reduction in emission intensity and a 30 percent it would be vitally important to cover export-oriented production as well increase in productivity would result in a roughly 40 percent reduction in as import-competing production. overall emissions. The range of policy approaches considered in this paper on repur- While some technological progress is the result of decisions by prof- posing support could be complemented by other measures, such as it-making entrepreneurs, public research remains critically important. incentives for dietary change, which was emphasized by Springmann et Governments will need to play a more direct and active role in promoting al. (2017). Given the urgency of the need to reduce GHG emissions from R&D that reduces emissions from agriculture. Some of this is currently agriculture (Clark et al. 2020), it seems likely that more than one approach being done by governments in beef, dairy, and sheep-producing countries, will be required. partly to contribute to environmental goals, and perhaps partly out of Finally, when a repurposing agenda is being undertaken, it is important concern that market access may in the future be restricted by policies to be aware that the specific policy needs will likely differ substantially such as border carbon adjustments. among countries. Many countries will need to adapt new and more A focused repurposing agenda will also require greater attention to productive technologies for their own individual contexts before they can the implications of higher agricultural productivity for farm labor. As be successfully adopted. Individual countries adopting more productive observed in the simulations reported in this paper, large-scale increases and lower-emission technologies at a higher-than-average rate are likely in agricultural productivity put downward pressure on farm prices that to have a greater need for support for their farmers in adopting new tech- increase the importance of helping farm workers who are no longer needed nologies and in dealing with the resulting contraction in demand for farm in agriculture to leave the sector. Doing this successfully will require labor. Country deep-dives involving considerable research and analysis are removing, as much as possible, impediments to the movement of labor likely to be needed to ensure that the interventions make their greatest out of agriculture such as those that are frequently included in land tenure contribution to the economy, the environment, and to pressing social contracts. There is also a need for a more positive agenda. Improvement needs as well. of educational opportunities for rural children will become more important as the need for agricultural labor declines and outmigration becomes more likely. While a productivity-focused approach to lowering emissions is beneficial for producers, adopting countries, and the global commons, other approaches, such as the use of carbon taxes or conditionality, which raise production costs, create disincentives for producers. Frequently, these higher costs will result in pressures for policies to reduce the replacement of imports by products that are produced without these disincentives, or to avoid the contraction of export sectors that are being squeezed by higher costs. One frequent proposal for dealing with these problems is to introduce a border carbon adjustment (BCA) that com- pensates import-competing firms for their increased costs (Martin 2022). One challenge for this approach is that most emissions from agricultural production are process emissions, rather than emissions from combustion. While carbon taxes can be finely calibrated—by fuel type and emission content—to create incentives both to change production techniques and to reduce output, this is not the case with the process emission from agriculture. This forces a carbon tax on agriculture to rely solely 6 3   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 64  Avenues for Further Policy Analysis: Implementation 6 The results of this study provide new perspectives on policy questions regarding the repurposing of current agricultural support measures. Clearly, there is vast scope for improving current agricultural support measures, particularly for achieving the SDGs related to reducing poverty, improving resilience, and increasing sustainability. However, it is essential to be strategic about the type of reforms to be pursued if those goals are to be achieved. These findings highlight clear trade-offs among the environmental, economic, nutritional, and social objectives associated with the “simple” option of removing domestic support. The key messages that emerge from this analysis are that, even if it were politically feasible, the abolition of current domestic support would have significant impacts on GHG emissions at the individual country level and even at the sector level: specifically, in the crops sector. But the impact on global GHG emissions will fall far short of what is needed to appreciably curb agriculture’s contribution to climate change. There are several reasons for this—most notably that the support rates have been determined on political economy grounds that are unre- lated to the environmental impacts of support—but also that some support is provided by countries with relatively low emissions per unit of output. In addition, with sustained and growing demand for agricultural products, there would be shifts in the structure of production across countries, but the overall effect on the level of production would be relatively small. As a result, the aggregate impact on GHG emissions would also be relatively small. Finally, the findings suggest very low transfer efficiency of current domestic support. Each dollar of public expenditure for domestic support contributes only 35 cents in value added, making a powerful case for finding better ways CONCLUSIONS to support producers. Abolition of border measures would actually increase emissions very slightly. This is because the predominant positive support measures combine support for output with a disincentive to consumption. If current protection were simply removed, the results suggest that the stimulus to global demand would slightly outweigh the loss of incentives to production in protected markets, raising the incentive to produce emission- intensive goods. When repurposing rather than removing support for agriculture, the results show significant potential for some of the options to deliver large “triple wins.” The results for various repurposing options, however, also caution against simplistic redesign of policies. Simple rearrangements of current domestic support would tend to have quite limited effects on emissions. In particular, replacing the current highly variable set of subsi- dies with a uniform set would have little impact on emissions. And trans- R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 66  Conclusions ferring all subsidies to low-emission cropping activities would, paradoxically, by excluding efficient producers from developing countries. There is great increase emissions due to the global land-use change that would result. potential for achieving major gains on these fronts by repurposing support toward public investments that facilitate the widespread adoption of Making support “conditional” on farmers’ willingness to provide envi- productivity-enhancing and emission-reducing technologies for agri-food ronmental services could be attractive in terms of emissions reduction. systems. Furthermore, these policies are likely to have strongly positive However, some of these options may result in lower productivity. For instance, international spillovers. Innovations that reduce environmental impacts and if farmers were merely asked to reduce chemical input use or shift to organic raise productivity are likely to either be rapidly adapted in other countries agriculture, the productivity loss would imply significant reduction in national or provide them with a basis for developing technologies for their own income and agricultural production, while poverty and the cost of healthy agroecological environments. diets would increase, and agricultural land use for agriculture would increase at the expense of forest habitat. Balancing a reduction in emissions with Nevertheless, even the best design of the proposed policy reforms productivity loss, especially in developing countries, is a major challenge. undoubtedly will face considerable political hurdles. Agricultural support policies are the prerogative of national governments. Overcoming national Repurposing support toward investments that are targeted at productiv- resistance to agricultural policy reform will be a huge challenge. National ity-enhancing and emissions-reducing technologies holds the greatest farm and agricultural policies have a long history in most countries and have potential for delivering “triple wins” for a healthy planet, economy, and developed well-established entitlements and vested interests. Recognition people. Repurposing a relatively small share of the current domestic support of the major private and societal gains to be achieved, and multistakeholder funded from public expenditures (which represents about 1 percent of engagements to discuss the potential trade-offs associated with policy present agricultural value added) toward the development and diffusion options and to devise acceptable strategies should help earn political of emissions-reducing and productivity-enhancing innovations would support for smart repurposing of the existing support at the national level. improve human welfare while substantially reducing global emissions. Such technologies appear to have the greatest potential for reducing poverty, For reforms to foster sustainable global development, a combination lowering the cost of healthy diets, and reducing the amount of land needed of effective policy coordination and technological innovations that are for agriculture. From the macroeconomic perspective, this repurposing also attractive to both individual producers and governments is needed. At has the strongest positive impact on real national income and structural present, agricultural support is distributed unevenly across nations. Poorer transformation; that is, reducing agricultural employment as labor transitions nations have less fiscal space with which to provide agricultural support. to other sectors of the economy. Policies that lead to the development of Also, their national agricultural research systems generally have weaker new technologies with higher private productivity also have the advantage resource capacity for developing high-productivity and sustainable farm of not requiring concerted action. Countries that choose to adopt more technologies and practices that are relevant to the local context, and their productive and lower-emission technologies will tend to gain market share, farmers and other food producers face bigger obstacles in adopting those avoiding the problems of carbon leakage that plague approaches based on practices. Hence, to be most effective at the global level, a more even-hand- the use of current technologies. ed diffusion of both technologies and financial resources is needed so that countries can reap the benefits of agricultural policy reform and contribute Notwithstanding the impressive results from the repurposing options more strongly to solving global challenges. discussed in this global modeling study, current agricultural support measures need to be carefully scrutinized in individual country contexts. International coordination is vitally important for achieving the needed The “triple win” scenario considered in this study is based on investing reductions in global emissions from agriculture. Climate change and only about 29 percent (about $70 billion) of current domestic support for environmental sustainability are global challenges that transcend borders, agriculture. But the volume repurposed need not be limited to this level. and national policies have strong international spillover effects. Policymakers A key insight from this study is that current agricultural support is a very are well-placed to scrutinize and rethink domestic policies. For the health blunt, and largely counterproductive instrument for fighting climate change of people, economies, and the planet, nations and food system actors must and addressing the remaining challenges to global food security and come together behind a concerted strategy for resetting global food system nutrition. Current support for agriculture distributes much of its benefits to incentives to address the existential threats posed by climate change and the relatively well-off and generates substantial inefficiency and inequity unsustainable food systems. 67   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 68  Conclusions APPENDIX A. APPENDIX B. TRENDS IN SUPPORT METHODS FOR DERIVING THE FIGURE A.1: Trends in the Nominal Rate of Protection by Income Level DATABASE, AND THE EMISSIONS 0.45 HIGH INCOME COUNTRIES - ALL PRODUCTS NRP MODELING FRAMEWORK Outputs 0.35 Inputs In this study, wherever possible, a full matrix was derived by reverse Other 0.25 engineering the FAO emissions data to ensure that the total matched the FAOSTAT estimates. Where this was not possible, as in the case of 0.15 emissions from pesticides, a similar IPCC Tier 1 methodology was used to generate comparable estimates. 0.05 Emission sources were identified using 11 FAOSTAT-based categories, plus 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 -0.05 emissions from agricultural pesticides. The first step was to define the activity levels associated with commodity outputs, such as the amount -0.15 of area used for rice cultivation. The second was to calculate the emission -0.25 coefficients (ECs) for CH4, CO2, and N2O by activity level using, wherever possible, the FAOSTAT database. Finally, emissions of N2O and CH4 were 0.45 MIDDLE INCOME COUNTRIES - ALL PRODUCTS NRP converted to CO2 equivalents, using 310 and 21 for N2O and CH4 respectively. Outputs 0.35 Inputs In many cases, the FAOSTAT emissions database provided implied Other emission factors by activity and emission source, such as the amount of 0.25 area harvested in rice cultivation, and the nitrogen content of manure. In 0.15 some cases, it provided the base activity data, such as area of organic soil cultivation, and the number of head of livestock for enteric fermentation 0.05 and manure management. In other cases, such as burning crop residues, only the data on biomass burned are provided, rather than data on which 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 -0.05 crops were burned. In such cases, base activity data were imported from the FAOSTAT crop and livestock production database for the crops whose -0.15 residues are frequently burned—maize, rice, sugar cane, and wheat. -0.25 For synthetic nitrogen fertilizer, the activity data regarding the agricultural 0.35 LOW INCOME COUNTRIES - ALL PRODUCTS NRP use of nitrogen is missing. Fertilizer use data are obtained from two sourc- Outputs es—FAOSTAT, and the International Fertilizer Association (IFA) (www.ifastat. 0.25 org ). FAOSTAT gives the total fertilizer volume for many countries, while Inputs Other IFA’s data regarding fertilizer use by crop provides the nutrient content of 0.15 fertilizer by crop for 54 countries. Fertilizer use data from FAOSTAT were 0.05 scaled to match IFA numbers for all countries; this was done by mapping the characteristics of IFA countries to the countries listed in FAOSTAT. Finally, emissions were estimated by multiplying fertilizer volume by the 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 -0.05 emission coefficients given in the FAOSTAT database. For the final version -0.15 -0.25 69   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 70  Appendix of the database, the base activity (or index) data were retained in order to estimate the average amount of emissions per index type (land, animals, APPENDIX C. THE MODELING FRAMEWORK output, fertilizer, and energy). The process for creating this new database is presented schematically in Figure B.1 FIGURE B.1:  Creation of GHG Agricultural Emissions Database by Source, Location, Commodity, Production Stage, and Technology The International Food Policy Research Institute (IFPRI)’s global computable general equilibrium (CGE) model, MIRAGRODEP, provides the core of the modeling framework used in this study. It is an extension of the widely used SOURCE OF MIRAGE multisector, recursive dynamic CGE model of the global economy EMISSION (12 sources (Decreux and Valin 2007), which allows for a detailed and consistent repre- DECOMPOSE identified) INTO sentation of the economic and trade relations between countries. In each country, a representative consumer maximizes a CES–LES (Con- stant Elasticity of Substitution–Linear Expenditure System) utility function subject to an endogenous budget constraint in order to generate the Scale of AREA QUANTITY HEAD Production allocation of expenditures across goods. This functional form replaces the (Source: FAO) Cobb-Douglas structure of the Stone-Geary function (that is, LES) with a CES structure that retains the ability of the LES system to incorporate different income elasticities of demand (Stone 1954), with those for food being typically lower than those for manufactured goods and services. The Apply Technical Technical coefficients - Tier 1* Coefficients demand system is calibrated on the income and price elasticities estimat- ed by Muhammad et al. (2017). Once the total consumption of each good (Source: IPCC Guidelines) Emissions per unit has been determined, the origin of the goods consumed is determined by N extraction per head N content in residue another CES nested structure, following the Armington (1969) assumption Leaching and DIRECT INDIRECT of imperfect substitutability between imported and domestic products. volatilization rate EMISSION EMISSION Share left on soil etc. On the production side, demands for intermediate goods are determined through a fixed-coefficient (Leontief) production function that specifies AGGREGATE intermediate input demands in fixed proportions to output. Total value CO2EQ Apply conversion added is determined through a CES function of unskilled labor, and a Allocate emission (by country, coefficients composite factor of skilled labor and capital. This specification assumes by commodities source, sector across sources ** alower degree of substitutability between the last two production factors. (N to N2O, CO2eq; CH4 to CO2eq) and year) In agriculture and mining, production also depends on land and natural resources. Source: Laborde et al. 2021. The underlying database used for the analysis is Pre-Release 1 of the GTAP Notes: * Tier I: Default emission factors from IPCC guidelines (2006). ** Using disaggregation space and linkage matrix. v11 database for 2017 (www.gtap.org). This database includes 141 regions/ countries and 65 products. It includes updated social accounting matrices The allocation of emissions from enteric fermentation and manure manage- for all individually specified countries, and updated estimates of agricul- ment between products such as meat, milk, and wool, from the livestock that tural support measures based on measures of average domestic support produce them (such as buffalo, camels, cattle, goats, and sheep) is in line provided by the Organisation for Economic Co-operation and Develop- with the value of their products. The livestock numbers were then linked to ment (OECD), and adjusted to include the impacts on the bilateral protec- emissions using data from the FAOSTAT emissions database. In the final step, tion rates of major trade preferences. A realistic baseline was constructed, emissions data were produced by country, emission source, and commodity. aligned with the United Nations’ demographic projections and updated IMF 7 1   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 72  Appendix C economic growth estimates to bring the base year values (2017) to those of The macroeconomic assumptions used for the analysis were designed to be the actual years of simulation (2021–25) and on to the comparisons between relatively “neutral” to avoid situations in which macroeconomic adjustments reference and simulated outcomes in 2040. such as real exchange rate changes could outweigh the impacts of interest, and to allow focusing on the impacts of agricultural support policies on The data on agricultural support were adjusted in line with the OECD’s (2016) emissions. These assumptions were that: categories, distinguishing, in particular, agricultural border measures and subsidies that influence output or input decisions (coupled subsidies). The 3. The analysis is based on macroeconomic projections to 2040 imple- mented annually in a recursive-dynamic model. model was augmented with a post-solution module based on the new emission database presented in Appendix B, which links GHG emissions to outputs and 4. Investment is savings-driven, and the real exchange rate adjusts to inputs of agricultural activities within the model. These links are presented keep the current account constant relative to the national GDP. schematically in Figure C.1. The combined model was then used to assess the 5. Aggregate real public expenditures are kept constant, and a consump- impacts of policy reform on emissions of CH4, CO2, and N2O, and these results tion tax is adjusted to keep the government budget balance fixed as a were combined to generate changes in emissions in CO2 equivalents. share of GDP. 6. Land-use change varies across agroecological zones as defined for FIGURE C.1: Linking Emissions to Production in MIRAGRODEP each region specified in the model, and follows the procedure outlined in Hertel et al. (2009), where land is reallocated between forest and Production various types of agricultural land in response to changes in returns. (Y-PY) 7. Total employment as a share of the active population is constant. The Leontief active population is defined by the 15-to-60-year-old group in the United Nations Department of Economic and Social Affairs (UNDESA) projections. Value added Intermediate Consumption (VA-PVA) (CNTER-PCNTER) The modeling approach for land builds on the agroecological zone (AEZ) approach of Hertel et al. (2009). Competition for land between forestry Production emissions from: -Fertilizers CES CES and agricultural uses within 16 agroecological zones is represented using -Chemical Pesticides a constant elasticity of transformation (CET) specification. Land is also -Fossil Fuel Unskilled labor Commodity 1 reallocated between agricultural activities in response to changes in relative Role of couples subsidies (L-PL) (IC-PIC) prices. Emissions from land use and land-use change arise from the conver- sion of land from forestry to agricultural uses; transitions between grassland Land Commodity I and cropland; cultivation of organic soils; and CO2 sequestration. The model (TE-PTE) (IC-PIC) considers only land use and land-use change that was created by changes GHG emissions in agricultural incentives, and thus generates estimates of emissions from from crop-specific Natural resources soil use (e.g. methane (RN-PRN) Production emissions from: the conversion of forest to agricultural land that is less than the gross from rice -Fertilizers estimates of land conversion away from forests reported by the Food and -Chemical Pesticides Capital-Skillled -Fossil Fuel Agriculture Organization of the United Nations (FAO). labor bundle (Q-PQ) Role of coupled subsidies The poverty analyses reported in the paper were conducted using the POVANA household modeling framework documented in Laborde, Martin, CES and Vos (2020). To make this relevant to the 2020–2040 projection period GHG emissions for this study, household incomes within the model were projected forward from livestock Capital Skilled labor in line with the trends of economic growth in each country. This reduced the (enteric (KTOT-PK (H-PH) fermentation) poverty rate in the benchmark to 3.5 percent at the traditional World Bank extreme poverty line value of $1.90, and to 10 percent at the $3.20 poverty line. (Both poverty lines are expressed per person per day, and in purchasing Source: Laborde et al. 2021. power parity dollars, PPP$). 73   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 74  Appendix C APPENDIX D. DETAILED SIMULATION RESULTS Global Impacts of Removing Components of Agricultural Support (% change in each indicator by 2040 with respect to baseline) TABLE D.1:  ALL DOMESTIC OUTPUT INPUT FACTOR CROPS LIVESTOCK DEVELOPED TRADE BARRIERS SUPPORT SUBSIDY SUBSIDY PAYMENT ONLY ONLY ONLY & DOM SUPPORT (1A) (1A.1) (1A.2) (1A.3) (1A.4) (1A.5) (1A.6) (1B) Macroeconomic National Real Income 0.05 0.01 0.01 0.02 0.03 0.01 0.02 0.09 Farm Sector Real Farm Income per Worker -4.51 -0.66 -0.59 -3.37 -3.76 -0.78 -2.43 -3.54 World Prices 2.93 0.74 1.05 1.12 2.66 0.30 1.63 4.38 Production Volume – Crops -1.31 -0.40 -0.57 -0.35 -1.30 -0.02 -0.35 -1.23 Production Volume - Livestock -0.49 0.01 -0.28 -0.22 0.12 -0.61 -0.32 -0.35 Social Farm Employment -0.53 -0.15 -0.60 0.22 -0.49 -0.03 0.29 -1.51 2040 Poverty Rate at PPP$1.90 0.01 0.00 0.01 -0.01 0.01 0.00 -0.01 -0.02 2040 Poverty Rate at PPP$3.20 0.05 -0.01 0.07 -0.02 0.05 0.00 -0.03 0.05 Nutrition/Diets Dairy Cons Per Capita -0.42 -0.04 -0.24 -0.14 0.23 -0.66 -0.25 0.55 Fats Cons Per Capita -0.94 -0.54 -0.01 -0.40 -0.95 0.01 -0.91 -2.68 Sugar Cons Per Capita -1.24 -0.17 -0.97 -0.09 -1.29 0.04 -0.39 4.91 Veg & Fruit Cons Per Capita -0.48 0.04 -0.31 -0.21 -0.50 0.02 -0.21 0.02 Healthy Diet Food Prices 1.70 0.08 0.83 0.79 1.40 0.33 0.89 1.15 Climate Energy in Agriculture – MtoE -1.04 -0.18 -0.55 -0.32 -0.75 -0.29 -0.35 -0.91 Emissions from Production, % of ALU -0.59 -0.03 -0.47 -0.09 -0.40 -0.20 -0.11 -0.20 Emissions from Land-Use Ch., % of ALU -0.89 -0.33 -0.28 -0.31 -1.55 0.24 -0.39 -0.35 Total Emissions – Megatons CO2 eq. -103.1 -25.2 -52.3 -28.4 -136.0 2.6 -35.0 -38.5 Total Emissions - % of ALU -1.48 -0.36 -0.75 -0.41 -1.95 0.04 -0.50 -0.55 Nature Agricultural Land -0.06 0.00 -0.08 0.04 0.09 -0.17 0.01 -0.02 Cropland -0.19 -0.06 -0.04 -0.10 -0.38 0.12 -0.10 -0.08 Pasture 0.01 0.03 -0.11 0.10 0.32 -0.32 0.07 0.01 Note: ALU refers to emissions from Agricultural Production and Land Use. MToE = million tons of oil equivalent energy use. 75   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T | Appendix D Results by Selected Countries for a Scenario of Abolition of All Subsidies (% change by 2040 in each indicator with respect to the baseline) TABLE D.2:  WORLD DEVELOPED DEVELOPING BRAZIL CHINA EU INDIA USA Macroeconomic National Real Income 0.05 0.05 0.04 0.26 0.03 0.11 0.03 0.02 Farm Sector Real Farm Income per Worker -4.51 -11.36 -2.70 0.76 -5.03 -23.07 -2.37 -9.36 World Prices 2.93 2.93 2.93 2.93 2.93 2.93 2.93 2.93 Production Volume – Crops -1.31 -2.56 -1.02 0.66 -1.83 -3.97 -3.06 -5.06 Production Volume – Livestock -0.49 -1.10 -0.07 0.81 0.19 -3.00 -0.82 0.15 Social Farm Employment -0.53 0.25 -0.60 1.04 -1.07 -1.01 -2.62 -1.69 2040 Poverty at PPP$1.90 0.01 0.01 0.01 -0.03 - - 0.06 - 2040 Poverty at PPP$3.20 0.05 -0.01 0.06 -0.06 - - 0.29 - Nutrition/Diets Dairy Cons Per Capita -0.42 -0.49 -0.37 -0.17 -0.06 -0.60 -0.35 -0.13 Fats Cons Per Capita -0.94 -1.16 -0.87 -0.98 -1.42 -0.98 0.64 -1.70 Sugar Cons Per Capita -1.24 -0.93 -1.46 0.33 -0.49 -0.07 -3.98 0.15 Veg & Fruits Cons Per Capita -0.48 -0.54 -0.45 -0.64 -0.33 -0.58 -1.23 -0.73 Healthy Diet Food Prices 1.70 2.17 1.44 1.37 1.09 3.19 1.91 2.40 Climate Energy in Agriculture – MtoE -1.04 -1.43 -0.83 0.83 -0.60 -3.07 -2.08 -1.55 Emissions from Production, % of ALU -0.59 -1.52 -0.38 0.74 -0.30 -6.29 -1.21 -2.42 Emissions from Land-Use Ch., % of ALU -0.89 -4.52 -0.07 -0.29 5.67 6.83 -0.02 -29.73 Total Emissions – Megatons CO2 eq. -103.1 -77.8 -25.3 1.81 22.5 1.46 -19.2 -83.8 Total Emissions - % of ALU -1.48 -6.04 -0.44 0.45 5.37 0.53 -1.23 -32.15 Nature Agricultural Land -0.06 -0.15 -0.01 -0.03 0.85 -1.28 -0.01 -0.15 Cropland -0.19 -0.50 -0.06 0.01 0.32 1.39 -0.01 -2.44 Pasture 0.01 0.01 0.01 -0.05 1.05 -6.50 0.00 1.42 76   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T | Appendix D  lobal Impacts of Repurposing Simulations (% change in each indicator with respect to baseline) TABLE D.3: G 2A FOR UNIFORM NON-CO2 INT. CONDI- 3 ONLY FOR 3 ONLY FOR 4. REPURPOS- 4. AS “MANNA GI -PUB 4B ONLY FOR 4B ONLY FOR SUPPORT PRODUCTS TIONALITY DEVELOPED DEVELOPING ING FOR GI FROM HEAVEN” FUNDED DEVELOPED DEVELOPING 2A 2B 3 3A 3B 4 4A 4B 4C 4D Macroeconomic National Real Income -0.01 -0.03 -0.81 -0.16 -0.63 1.61 1.71 1.57 0.33 1.31 Farm Sector Real Farm Income per Worker -2.28 -1.16 2.02 1.20 0.81 -8.39 -4.54 -4.79 -2.77 -2.22 World Prices -0.63 -2.03 12.71 4.47 7.41 -20.85 -23.21 -23.24 -10.72 -16.12 Production Volume - Crops -0.05 1.41 -6.28 -1.12 -5.19 16.06 18.02 17.95 2.99 14.80 Production Volume - Livestock 2.40 -0.69 -4.66 -1.49 -3.19 11.47 12.19 12.14 3.66 8.41 Social Farm Employment 0.25 0.18 4.65 0.98 3.53 -10.50 -9.79 -9.83 -2.49 -7.99 2040 Poverty at PPP$1.90 -0.01 -0.01 0.58 0.00 0.57 -1.00 -1.02 -0.99 -0.02 -1.01 2040 Poverty at PPP$3.20 -0.06 -0.06 0.58 -0.01 0.58 -0.97 -1.05 -1.02 0.04 -1.07 Nutrition/Diets Dairy Cons Per Capita 3.20 -0.74 -6.37 -2.02 -4.32 16.41 17.15 17.07 5.14 12.14 Fats Cons Per Capita -0.65 -0.18 -3.91 -1.00 -2.86 8.65 9.82 9.77 2.77 7.42 Sugar Cons Per Capita 3.58 13.57 -10.20 -3.90 -6.27 27.53 29.37 29.32 11.28 18.33 Veg/Fruit Cons Per Capita 0.09 1.14 -4.40 -1.05 -3.34 11.95 12.79 12.73 2.87 9.98 Healthy Diet Food Prices -0.49 -1.98 10.01 3.39 6.21 -17.63 -19.06 -19.12 -7.60 -13.34 Climate Energy in Agriculture - MToE 0.98 0.72 -4.34 -1.33 -2.92 10.47 11.96 11.90 3.81 8.72 Emissions-Production, % of ALU 0.49 -0.05 -19.17 -3.42 -15.49 -24.14 -23.48 -23.55 -6.72 -17.41 Emissions - Land Use, % of ALU -1.14 0.31 4.59 1.42 3.11 -16.31 -15.09 -15.22 -4.79 -10.47 Total Emissions – Mtons CO2 eq. -45.2 18.3 -1018.8 -139.3 -865.0 -2825.9 -2694.5 -2708.5 -803.4 -1947.9 Total Emissions - % of ALU -0.65 0.26 -14.58 -1.99 -12.38 -40.45 -38.57 -38.77 -11.50 -27.88 Nature Agricultural Land 0.02 0.00 0.62 0.27 0.35 -2.15 -2.04 -2.05 -0.80 -1.21 Cropland -0.22 0.03 0.45 0.10 0.34 -1.72 -1.49 -1.50 -0.44 -1.09 Pasture 0.13 -0.02 0.70 0.35 0.35 -2.36 -2.31 -2.33 -0.98 -1.27 Forest Habitat 0.04 -0.03 -0.23 -0.08 -0.15 0.40 0.39 0.40 0.15 0.32 7 7   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T | Appendix D TABLE D.4:  Impacts of Country-Specific Repurposing Scenarios: Productivity-Enhancing and Emission-Reducing Farm Practices in Individual Countries (% change in each indicator by 2040 with respect to baseline) BRAZIL CHINA ETHIOPIA INDIA INDONESIA EU UNITED STATES Macroeconomic National Real Income 2.5 2.5 7.5 2.5 2.3 0.5 0.4 Farm Sector Real Farm Income per Worker 4.8 -0.3 6.7 1.2 1.6 -1.2 3.0 World Prices -2.7 -4.0 -0.2 -2.3 -0.4 -3.4 -4.1 Production Volume - Crops 34.9 25.2 26.7 29.1 25.9 32.7 34.0 Production Volume - Livestock 28.8 18.0 22.8 23.5 23.4 16.7 17.2 Social Farm Employment 4.1 -5.6 -3.7 -3.0 -3.1 -1.6 -2.6 2040 Poverty at PPP$1.90 -0.2 - -0.7 0.0 -0.3 - - 2040 Poverty at PPP$3.20 -0.2 0.0 -3.0 -0.7 -0.8 - - Nutrition/Diets Dairy Cons Per Capita 14.7 14.5 23.1 22.1 11.5 8.4 7.7 Fats Cons Per Capita 3.6 10.4 -0.5 -3.6 9.6 2.3 3.2 Sugar Cons Per Capita 23.8 21.3 35.7 20.0 21.8 19.7 16.4 Veg & Fruits Cons Per Capita 9.4 12.5 8.7 11.4 5.4 3.5 4.1 Healthy Diet Food Prices -9.9 -16.8 -12.9 -12.3 -12.9 -10.4 -9.8 Climate Energy in Agriculture - MToE 30.2 19.4 21.9 24.9 24.3 20.6 21.9 Emissions-Production, % of ALU -17.4 -44.7 -13.3 -14.6 -7.7 -33.3 -32.0 Emissions - Land Use, % of ALU -11.1 -8.6 -0.5 -0.3 2.8 -7.1 -10.8 Total Emissions – Megatons CO2 eq. -116.1 -223.6 -51.5 -233.7 -25.4 -110.5 -111.6 Total Emissions - % of ALU -28.5 -53.3 -13.9 -15.0 -4.9 -40.4 -42.9 Nature Agricultural Land -1.2 -1.6 -0.5 -0.2 0.7 -1.6 -1.8 Cropland -0.1 -0.3 -0.5 -0.1 0.8 -0.1 0.5 Pasture -1.6 -2.0 -0.6 -0.8 0.3 -4.5 -3.4 7 8   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T | Appendix D biodiversity (Just Rural Transition 2021). A key question is how public support APPENDIX E. for these broader goals may best be channeled into political support that will POLITICAL ECONOMY CHALLENGES bring about concrete political change. One big problem for achieving vitally important goals such as improving sustainability is the public-good nature OF REPURPOSING AGRICULTURAL of these goods. A reduction in global emissions that reduces global warming by one degree would provide benefits to all. (In other words, “my benefiting POLICIES AND SUPPORT from this reduction does not diminish the benefit to others.”) Perhaps more importantly no person and no country can be excluded from the benefit. This creates the well-known problem of the tragedy of the commons, in which Transforming agricultural policies and support is likely to involve many it is in the interest of individuals and countries to overconsume a resource, serious challenges. Given the deeply political nature of these decisions, potentially leading to its collapse (Frischmann 2018). simply identifying the ways in which the current structure of support fails Three approaches to dealing with such collective action problems are to achieve specified economic goals is not likely to be enough to secure typically considered: reforms. Support measures tend to be in place for particular commodities because of asymmetries in political power between those gaining and 1. Agreements that use taxes or subsidies to internalize the associated those losing from these measures (Grossman and Helpman 1988). Most externalities (Pigou 1932) successful proposals for reform are based on an understanding of the 2. Allocating property rights (Coase 1960) political economy forces that gave rise to the existing measures and the 3. Allowing communities to create rules for resource management ways in which reform might contribute to the mitigation of emissions. These (Ostrom 1990). proposals would necessarily have to be context- and country-specific, and would thus require deeper country-level engagements to identify those A fourth approach that may play an important role is technological change. pathways and options that may be feasible to implement. For example, advances in contraception technology seems to have played an important role—along with the desire to invest in children in Some reforms attempt to redesign policies in ways that continue to serve higher-income societies—in resolving the seemingly intractable Malthusian the powerful interests that were supported by the initial policies, while specter of global overpopulation and resource collapse. Geoengineering reducing the adverse impacts on other affected parties whose strength approaches have also been proposed as a potential means of dealing with has increased. Only rarely are policy reformers able to introduce reforms global warming (Royal Society 2009). that withdraw benefits from strong interest groups. Major reforms to support policies, such as the reform of the EU’s Common Agricultural The usual approach to dealing with collective action problems that spill Policy (CAP) (Swinnen 2015) tend to involve changes in the way in which over between countries—whether of international security, product or assistance is provided to powerful economic groups; or changes in either service standards, or market functioning—is to create an international insti- the cost of providing support or recognition of the rising power of other tution or a body of rules, such as the United Nations, the Universal Postal economic groups. In the seminal case of the CAP reform, for instance, the Union, or the World Trade Organization. Attempts to do this for sustainabil- dramatic increase in the cost of providing support when the EU became a ity through agreements such as the Kyoto Protocol have been unsuccessful net exporter of many products—and hence market price support stopped because of a lack of enforcement powers, with this agreement having been generating tariff revenues and required funding export subsidies—was an replaced using the much more flexible architecture of the Paris Agreement. important source of pressure for reform. Another was pressure from trading Attempts to deal with the climate change problem have used all four of the partners—both unilaterally and through the World Trade Organization approaches considered. The Kyoto Protocol created targets by country for (WTO)—against the use of export subsidies. six main greenhouse gases. This used the Coasian approach of allocating The rise of interest in repurposing agricultural support is associated with property rights with a view to limiting emissions. Had these rights been increased concern about environmental problems such as global warming, freely transferable between countries, they would have mimicked the oper- the need to improve nutritional outcomes, and a desire to increase ation of a Pigovian tax, but with the revenues allocated to the governments in line with their allocations of quotas. Emissions trading systems and 79   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 80  Appendix D carbon pricing systems have been used in a number of countries and cities (World Bank 2020), and resource management systems at the local level REFERENCES have been used to avoid deforestation and the associated GHG emissions (Cardenas 2016). Agricultural support policies and policies designed to mitigate GHG Abbott, C. 2020. “USDA Tightens Eligibility Rules for Farm Subsidies.” Successful emissions are very different. However, there are potentially important links Farming, August 24. www.agriculture.com/news/business/usda-tightens-eligi- between the two, through reforms to agricultural support policies that: bility-rules-for-farm-subsidies Alston, J., P. Pardey, and X. Rao. 2020. The Payoff to Investing in CGIAR Research. 1. Reallocate support toward low-emission agricultural activities, and SoAR Foundation. https://tinyurl.com/uya43kvw away from emission-intensive activities. Anderson, K., M. Ivanic, and W. Martin. 2014. “Food Price Spikes, Price Insulation, 2. Make receipt of support conditional on the adoption of low-emission and Poverty.” In The Economics of Food Price Volatility, eds. J-P. Chavas, D. production methods. Hummels, and B. Wright. Chicago: University of Chicago Press for National Bureau of Economic Research. 3. Reallocate some support resources toward the development of pro- Armington, P. 1969. “A Theory of Demand for Products Distinguished by Place of duction methods that have lower emissions and higher productivity. Production” IMF Staff Papers 16: 159-178. Reallocation of support away from high-emission activities runs into the Barreiro-Hurle, J., M. Bogonos, M. Himics, J. Hristov, I. Pérez-Domínguez, A. Sahoo, political economy problem that some activities have much greater ability G. Salputra, F. Weiss, E. Baldoni, and C. Elleby. 2021. “Modelling Environmental and Climate Ambition in the Agricultural Sector with the CAPRI Model. Explor- to generate political support. It is of course important to assess whether ing the Potential Effects of Selected Farm to Fork and Biodiversity Strategies this approach can, in fact, generate substantial reductions in emissions. Targets in the Framework of the 2030 Climate Targets and the Post 2020 Approaches that make the receipt of benefits conditional on low-emission Common Agricultural Policy.” EUR 30317 EN, Luxembourg: Publications Office of the European Union. production methods have clear appeal in some cases. If a much lower Bell, P., N. Namoi, C. Lamanna, C. Corner-Dollof, E. Girvetz, C. Thierfelder, and T. emission technology is available at low cost and the threat of withdrawing Rosenstock. 2018. “A Practical Guide to Climate-Smart Agricultural Technol- support is large enough, then it might be possible to greatly reduce emis- ogies in Africa.” Research Program on Climate Change, Agriculture and Food sions through this approach. However, there are serious challenges with it, Security Working Paper No. 224, Wageningen, the Netherlands: CGIAR. including slippage that reduces the reduction in emissions achieved, and the Bhaskar, A., and J. Beghin. 2010. “Decoupled Farm Payments and the Role of Base Acreage and Yield Updating Under Uncertainty.” American Journal of Agricul- risk of non-additionality, whereby farmers receive payments for doing what tural Economics 92 (3): 849–858. they would have chosen to do anyway (Mamun, Martin, and Tokgoz 2021). Cardenas, J.C. 2016. “Human Behavior and the Use of Experiments to Understand the Agricultural, Resource, and Environmental Challenges of the XXI Century.” Approaches that reallocate some funding from subsidies toward Agricultural Economics 47 (S1): 61–71. approaches that reduce emissions and that, ideally, also increase produc- Chang, J., S. Peng, Y. Yin, P. Ciais, P. Havlik, and M. Herrero. 2021. “The Key Role of tivity appear to have a great deal of potential. Many sources of emissions Production Efficiency Changes in Livestock Methane Emission Mitigation.” from agriculture, such as methane emissions from ruminants, are both AGU Advances 2 (2): e2021AV000391. a source of potent GHG emissions and an abject waste of a potentially Clark, M., N. Domingo, K. Colgan, S. Thakrar, D. Tilman, J. Lynch, I. Azevedo, and J. valuable resource. It is not surprising that innovations designed to reduce Hill. 2020. “Global Food System Emissions Could Preclude Achieving the 1.5° and 2°C Climate Change Targets.” Science 370:705-8, 6 November. these emissions also appear to raise the efficiency of production (Kinley Coase, R. 1960. “The Problem of Social Cost.” Journal of Law and Economics 3: et al. 2019). Some countries with high GHG emission intensities in their 1–44. agriculture sector have begun to support research on ways of reducing Cochrane, W. 1958. Farm Prices: Myth and Reality. Minneapolis: University of Min- these emissions (see, for example, the NZ Agricultural Greenhouse Gas nesota Press. Research Center). Crippa, M., E. Solazzo, D. Guizzardi, F. Monforti-Ferrario, F. N. Tubiello, and A. Leip. 2021. “Food Systems Are Responsible for a Third of Global Anthropoge- nic GHG Emissions.” Nature Food 2: 198–209. https://doi.org/10.1038/s43016- 021-00225-9 Dalberg Asia. 2021. Funding Agricultural Innovation for the Global South: Does It Promote Sustainable Agricultural Intensification? Colombo: Commission on 81   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 82  Appendix D Sustainable Agriculture Intensification (CoSAI). HLPE. 2019. Agroecological and Other Innovative Approaches for Sustainable Agriculture and Food Systems that Enhance Food Security and Nutrition. A Decreux, Y., and H. Valin. 2007. “MIRAGE, Updated Version of the Model for Trade report by the High Level Panel of Experts on Food Security and Nutrition of Policy Analysis Focus on Agriculture and Dynamics.” CEPII Working Paper the Committee on World Food Security. Rome: FAO. 2007-15. Paris: CEPII. www.cepii.fr. Huang, J., Y. Liu, W. Martin, and S. Rozelle. 2009. “Changes in Trade and Domestic De Pinto, A., N. Cenacchi, H-Y Kwon, J. Koo, and S. Dunston. 2020. “Climate Smart Distortions Affecting China’s Agriculture.” Food Policy 34 (5): 407–16. Agriculture and Global Food-Crop Production.” PLoS ONE 15 (4): e0231764. Ianchovichina, E., and W. Martin. 2004. “Impacts of China’s Accession to the European Commission. 2020a. How the Future CAP Will Contribute to the EU World Trade Organization.” World Bank Economic Review 18 (1): 3–28. Green Deal. Factsheet, May 20. Brussels: European Commission. IMF. 2019. World Economic Outlook. October 2019. Washington DC: International European Commission. 2020b. Farm to Fork Strategy: For a Fair, Healthy and En- Monetary Fund. vironmentally Friendly Food System. Brussels: European Commission. IPCC (Intergovernmental Panel on Climate Change). 2003. Good Practice Guid- European Parliament. 2015. Implementation of the First Pillar of the CAP 2014-20 ance for Land Use, Land-Use Change and Forestry, Institute for Global Envi- in the EU Member States. Study IP/B/AGRI/IC/2014_45. Brussels: Directorate ronmental Strategies. Kanagawa, Japan: IPCC. General for Internal Policies. IPCC. 2020. Climate Change and Land: An IPCC Special Report on Climate Frischmann, B. 2018. “The Tragedy of the Commons, Revisited.” Scientific Ameri- Change, Desertification, Land Degradation, Sustainable Land Management, can Observations (blog), November 19. Food Security, and Greenhouse Gas Fluxes in Terrestrial Ecosystems. January FAO 2016. The State of Food and Agriculture 2016. Rome: FAO. 2020 update. https://www.ipcc.ch/srccl/ FAO, IFAD, UNICEF, WFP, and WHO. 2021. The State of Food Security and Nutri- Johnson, J.A., G. Ruta, U. Baldos, R. Cervigni, S. Chonabayashi, E. Corong, O. tion in the World 2021: Transforming Food Systems for Food Security, Im- Gavryliuk, J. Gerber, T. Hertel, C. Nootenboom, and S. Polasky. 2021. The Eco- proved Nutrition, and Affordable Healthy Diets for All. Rome: FAO. https://doi. nomic Case for Nature: A Global Earth-Economy Model to Assess Develop- org/10.4060/cb4474en ment Policy Pathways. Washington, DC: World Bank. FAO, UNDP, and UNEP. 2021. A Multi-Billion Dollar Opportunity: Repurposing Just Rural Transition. 2021. The Case for Repurposing Public Support for Agricul- Agricultural Support to Transform Food Systems. Rome: FAO. https://doi. ture. Policy Brief. Washington, DC. org/10.4060/cb6562en Kinley, R., G. Martinez-Fernandez, M.K. Matthews, R. de Nys, M. Magnusson, and FOLU (Food and Land Use Coalition). 2019. Growing Better: Ten Critical Transi- N.W. Tomkins. 2020. “Mitigating the Carbon Footprint and Improving Produc- tions to Transform Food and Land Use. Global Report. London: FOLU. tivity of Ruminant Livestock Agriculture Using a Red Seaweed.” Journal of Cleaner Production 259: 120836. FSIN (Food Security Information Network). 2021. Global Report on Food Crises. Rome: FSIN and Global Network Against Food Crises. Laborde, D., A. Mamun, W. Martin, V. Pineiro, and R. Vos. 2020. “Modeling the Im- pacts of Agricultural Support Policies on Emissions from Agriculture.” Agricul- Fuglie, K., M. Gautam, A. Goyal, and W.F. Maloney. 2020. Harvesting Prosperity: ture and Food Discussion Paper. World Bank, Washington, D.C. Technology and Productivity Growth in Agriculture. Washington, DC: World Bank. Laborde, D., A. Mamun, W. Martin, V. Pineiro, and R. Vos. 2021. “Agricultural Sub- sidies and Global Greenhouse Gas Emissions.” Nature Communications 12: Gollin, D., C. Hansen, and A. Wingender, A. 2021. “Two Blades of Grass: The Impact 2601. of the Green Revolution.” Journal of Political Economy 129 (8): 2344–2384. Laborde, D., W. Martin, and R. Vos. 2021. “Impacts of COVID-19 on Global Poverty, Grossman, G.M. and E. Helpman 1994. “Protection For Sale” American Economic Food Security and Diets.” Agricultural Economics 52 (3): 375–90. Review 84(4): 833-850. Laborde, D., V. Robichaud, and S. Tokgoz. 2013. “MIRAGRODEP 1.0: Documenta- Henning, C. and P. Witzke 2021. “Economic and Environmental Impacts of the tion, AGRODEP.” Technical Note 20. Washington, DC: International Food Policy Green Deal on the Agricultural Economy: A Simulation Study of the Impact of Research Institute. the F2F-Strategy on Production, Trade, Welfare, and the Environment Based on the CAPRI-Model.” Mimeo. University of Kiel. Mamun, A., W. Martin, and S. Tokgoz, S. 2021. “Reforming Agricultural Support for Improved Environmental Outcomes.” Applied Economic Perspectives and Pol- Herrendorf, B., and T. Schoellman. 2018. “Wages, Human Capital and Barriers to icy (January 10). https://doi.org/10.1002/aepp.13141 Structural Transformation.” American Economic Journal: Macroeconomics 10 (2): 1–23. Martin, W. 2019. “Economic Growth, Convergence and Agricultural Economics.” Agricultural Economics 50 (S1): 7–27. Hertel, T., H. Lee, S. Rose, and B. Sohngen. 2009. “Modeling Land-Use Related Greenhouse Gas Sources and Sinks and Their Mitigation Potential.” In Eco- Martin, W. 2022. “Border Carbon Adjustments: Should Production or Con- nomic Analysis of Land Use in Global Climate Change Policy, eds. T.W. Hertel, sumption be Taxed?” Researchgate. https://www.researchgate.net/publica- S. Rose, and R.S.J. Tol, Ch. 6. New York: Routledge. tion/357172148 Hicks, J.H., M. Kleemans, N.Y. Li, and E. Miguel. 2017. “Reevaluating Agricultural Martin, W., and J. Alston. 1997. “Producer Surplus Without Apology? Evaluating Productivity Gaps with Longitudinal Microdata.” NBER Working Paper 23253. Investments in R&D.” Economic Record 73 (221): 146–58. Cambridge, MA: National Bureau of Economic Research. Martin, W., and A. Mattoo. 2011. Unfinished Business? The WTO’s Doha Develop- 8 3   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 84  Appendix D ment Agenda. Washington, DC: Center for Economic Policy Research and the Seufert, V., N. Ramankutty, and J. Foley. 2012. “Comparing the Yields of Organic World Bank. and Conventional Agriculture.” Nature 485: 229–34, May. doi:10.1038/na- ture11069 Matsuyama, K. 1992. “Agricultural Productivity, Comparative Advantage and Eco- nomic Growth.” Journal of Economic Theory 58 (2): 317–34. Sicular, T., and Y. Zhao. 2004. “Earnings and Labor Mobility in Rural China: Impli- cations for China’s Accession to the WTO.” In China and the WTO: Accession, Mernit, J. 2018. “How Eating Seaweed Can Help Cows to Belch Less Methane.” Policy Reform and Poverty Reduction Strategies, eds. D. Bhattasali, S. Li, and Yale Environment 360: July 2. W. Martin. New York: Oxford University Press. Muhammad, A., A. D’Souza, B. Meade, R. Micha, and D. Mozaffarian. 2017. The In- Smith, L., G. Kirk, P. Jones, and A. Williams. 2019. “The Greenhouse Gas Impacts fluence of Income and Prices on Global Dietary Patterns by Country, Age, and of Converting Food Production in England and Wales to Organic Methods.” Gender. Economic Research Report No. ERR-225. Washington, DC: US Depart- Nature Communications 10: 4641 | https://doi.org/10.1038/s41467-019-12622-7. ment of Agriculture. Springmann, M., D. Mason-D’Croz, S. Robinson, K. Wiebe, C. Godfray, M. Rayner, Ocko, I., T. Sun, D. Shindell, M. Oppenheimer, A. Hristov, S. Pacala, D. Mauzerall, Y. and P. Scarborough. 2017. “Mitigation Potential and Global Health Impacts Xu, and S. Hamburg. 2021. “Acting Rapidly to Deploy Readily Available Methane from Emissions Pricing of Food Commodities.” Nature Climate Change 7: 69- Mitigation Measures by Sector Can Immediately Slow Global Warming.” Envi- 74, January. ronmental Research Letters 16: 054042. Startz, R. 2020. “The Next Hundred Years of Growth and Convergence.” Journal of OECD. 2003. “The Incidence and Income Transfer Efficiency of Farm Support Applied Econometrics 35 (1): 99–113. Measures.” In Farm Household Income: Issues and Policy Responses. Paris: OECD. Stone, R. 1954. “Linear Expenditure Systems and Demand Analysis: An Application to the Pattern of British Demand.” The Economic Journal 64 (255): 511–27. OECD. 2011. Evaluation of Agricultural Policy Reforms in the European Union. Paris: OECD. Swinnen, J. 2015. The Political Economy of the 2014-2020 Common Agricultural Policy: An Imperfect Storm. London: Rowman and Littlefield. OECD. 2020. Agricultural Policy Monitoring and Evaluation 2020. Paris: OECD. Terazono, E., and C. Hodgson. 2021. “Raising Climate Friendly Cows.” Financial OECD. 2021. Agricultural Policy Monitoring and Evaluation 2021: Addressing the Times, https://www.ft.com/content/73e5f1fc-76ac-48b0-871a-7fa4e8bda69b Challenges Facing Food Systems. Paris: OECD. Downloaded October 11. Ortiz-Bobea, A., T. Ault, C. Carrillo, R. Chambers, and D. Lobell. 2021. “Anthropo- Tubiello, F., C. Rosenzweig, G. Conchedda, K. Karl, J. Gütschow, P. Xueyao, G. Ob- genic Climate Change Has Slowed Global Agricultural Productivity Growth.” li-Laryea, N.Wanner, S.Yue Qiu, J. Barros, A. Flammini, E. Mencos-Contreras, L. Nature Climate Change 11: 306–12. Souza, R. Quadrelli, H. Halldórudóttir Heiðarsdóttir, P. Benoit, M. Hayek, and D. Ostrom, E. 1990. Governing the Commons: The Evolution of Institutions for Col- Sandalow. 2021. “Greenhouse Gas Emissions from Food Systems: Building the lective Action. Cambridge: Cambridge University Press. Evidence Base.” Environmental Research Letters 16 (6): 065007. https://iop- science.iop.org/article/10.1088/1748-9326/ac018e Parker, I. 2021. “The Great Organic Food Fraud.” New Yorker. 15 November, https:// www.newyorker.com/magazine/2021/11/15/the-great-organic-food-fraud US Department of State. 2021. “Joint U.S.-EU Statement on the Global Methane Pledge.” Media Note, October 11. https://www.state.gov/joint-u-s-eu-state- Pigou, A.C. 1920. The Economics of Welfare. London: Palgrave Macmillan. ment-on-the-global-methane-pledge/ Piñeiro, V., J. Arias, J. Dürr, P. Elverdin, A. Ibáñez, A. Kinengyere, C. Opazo, N. Owoo, Wang, H., and A. Matthews. 2011. “Estimating the Labour Transformation Elasticity J. Page, S. Prager, and M. Torero. 2020. “A Scoping Review on Incentives for Between Agricultural and Non-Agricultural Sectors in China.” Food Economics Adoption of Sustainable Agricultural Practices and Their Outcomes.” Nature - Acta Agriculturae Scandinavica, Section C 8 (1): 48–58. Sustainability 3: 809–20. Wolf, M. 2021. “What Is the Least We Need from COP26?” Financial Times, Octo- Ponisio, L.C., L.K. M’Gonigle, K.C. Mace, J. Palomino, P. de Valpine, and C. Kremen. ber 26. 2014. “Diversification Practices Reduce Organic to Conventional Yield Gap.” Proceedings of the Royal Society B 282 (1799): 20141396. World Bank. 2018. Poverty and Shared Prosperity 2018: Piecing Together the Pov- erty Puzzle. Washington, DC: World Bank. Powell, A., and F. Gruen. 1968. “The Constant Elasticity of Transformation Produc- tion Frontier and Linear Supply System.” International Economic Review 9 (3): World Bank. 2020. State and Trends of Carbon Pricing 2020. Washington, DC: 315–28. World Bank. Royal Society. 2009. Geoengineering the Climate. London: Royal Society. WRI (World Resources Institute) 2019. Creating A Sustainable Food Future. Final Report. Washington, DC: WRI Runkle, B., K. Suvočarev, M. Reba, C. Reavis, S. Smith, Yin-Lin Chiu, and B. Fong. 2019. “Methane Emissions Reductions from Alternate Wetting and Drying of Rice Fields Detected Using the Eddy Covariance Method.” Environmental Sci- ence and Technology 53 (2): 671–681. https://doi.org/10.1021/acs.est.8b05535 Schlenker, W. 2021. “Environmental Drivers of Agricultural Productivity Growth and Socioeconomic Spillovers.” Paper presented to the Federal Reserve Bank of Kansas City Agricultural Symposium. May 24–25. 8 5   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T 86  Appendix D 87   R E P U R P O S I N G AG R I C U LT U R A L P O L I C I E S A N D S U P P O R T