Reshaping the Agrifood Sector for Healthier Diets Exploring the Links between Agrifood Public Support and Diet Quality Knowledge Note © 2025 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington, DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy, completeness, or currency of the data included in this work and does not assume responsibility for any errors, omissions, or discrepancies in the information, or liability with respect to the use of or failure to use the information, methods, processes, or conclusions set forth. The boundaries, colors, denominations, links/footnotes, and other information shown 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. The citation of works authored by others does not mean the World Bank endorses the views expressed by those authors or the content of their works. Nothing herein shall constitute or be construed or considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Rights and Permissions This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) http:// creativecommons.org/licenses/by/3.0/igo. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: Attribution—Please cite the work as follows: World Bank. 2025. Reshaping the Agrifood Sector for Healthier Diets: Exploring the Links between Agrifood Public Support and Diet Quality. World Bank, Washington, DC. License: Creative Commons Attribution CC BY 3.0 IGO Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content—The World Bank does not necessarily own each component of the content contained within the work. The World Bank therefore does not warrant that the use of any third- party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to re-use a component of the work, it is your responsibility to determine whether permission is needed for that re-use and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. All queries on rights and licenses should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; e-mail: pubrights@worldbank.org. Cover design: Veronica Gadea, GCS, World Bank Group; Template and accessibility: Will Kemp, GCS, World Bank Group Table of Contents Acknowledgments  vii Abbreviations and Acronyms viii Executive Summary x 1 Failing to Tackle Malnutrition 1 1.1. The growing triple burden of malnutrition 1 1.2. The high costs of malnutrition  6 1.3. Poor diets are a crucial part of malnutrition 9 References17 2 An Opportunity to Leverage Agrifood Public Support 27 2.1. The problem with current agrifood support 27 2.2. An opportunity to repurpose agrifood support 36 Notes44 References45 3 Global-Level Analysis: Agrifood Support and Food Consumption 49 3.1. Existing research on agrifood support and food consumption 49 3.2. Different types of agrifood public support 52 3.3. Different types of food commodities 55 3.4. Methodology: Cross-country estimations  55 3.5. Results: Support to the agrifood sector 56 3.6. Results: Support to specific agrifood commodities 59 Notes68 References68 4 Country-Level Cases: Comparing Different Interventions 71 4.1. Input subsidies, roads, and cash transfers in Bangladesh 71 4.2. Input subsidies, irrigation, and cash transfers in Malawi 81 Data and methods 83 Notes92 References93 5 Future Directions 99 Appendix A: OECD Database on Agrifood Support 101 Appendix B: Global Dietary Database on Food Consumption 105 Appendix C: Cross-Country Regression Estimations 107 Tables Table 3.1 Top Producers and Consumers of Grains 60 Table 3.2 Top Producers and Consumers of Meats 63 Table 3.3 Top Producers and Consumers of Sugar 65 Table 4.1 Basic Descriptive Statistics of Selected Variables 73 Table 4.2 Basic Descriptive Statistics of Different Food Security Variables 74 Table 4.3 Average Value of Different Food Security Variables for the Beneficiaries and Nonbeneficiaries of the OAA Social Security Program 75 Table 4.4 Average Impact (Treatment Effect) of Social Protection Program on the Food Security of Households  76 Table 4.5 Average of Food Security Variables by the Presence of Improved Roads 77 Table 4.6 Average Impact (Treatment Effect) of Improved Rural Infrastructure on the Food Security of Households 78 Table 4.7 Average of Food Security Variables by Input Subsidy Card Status 79 Table 4.8 Average Impact (Treatment Effect) of Input Subsidy on the Food Security of Households 80 Table 4.9 Food and Nutrition Variables and Program Participation 84 Table 4.10 Pairwise Correlation of Program Participation 85 iv Reshaping the Agrifood Sector for Healthier Diets Table 4.11 Mean Differences in Food and Nutrition Security Measures by Program 86 Table 4.12 Descriptive Statistics for Pooled Sample 86 Table 4.13 Effects of Interventions on Household Dietary Diversity 87 Table 4.14 Effects of Interventions on Food Consumption Score 88 Table 4.15 Effects of Interventions on Coping with Food Insecurity 89 Table B.1 List of Countries in the Analysis (N = 44) 106 Table C.1 Specifications on the Choice of Support Variables 107 Table C.2 Descriptive Statistics: Overall Support Variables 108 Table C.3 Descriptive Statistics: Grain Consumption and Single Commodity Support Variables 109 Table C.4 Descriptive Statistics: Meat Consumption and Single Commodity Support Variables 110 Table C.5 Descriptive Statistics: Sugar Consumption and Single Commodity Support Variables 111 Table C.6 Regression: Grain Consumption and Overall Support 112 Table C.7 Regression: Grain Consumption and Single Commodity Support 113 Table C.8 Regression: Meat Consumption and Overall Support 114 Table C.9 Regression: Meat Consumption and Single Commodity Support 115 Table C.10 Regression: Sugar Consumption and Overall Support 116 Table C.11 Regression: Sugar Consumption and Single Commodity Support 117 Figures Figure 1.1 Trends in the Prevalence of Stunting, Wasting, and Overweight in Children under Five 2 Figure 1.2 Global Trends in Adult BMI 4 Figure 1.3 Regional Trends in Obesity 5 Figure 1.4 Diet-Related NCDs 5 Figure 1.5 Composition of Cause of Death, WHO Regions and Global, 2000-2048  6 Figure 1.6 Hidden Costs of Food and Land Uses 7 Figure 1.7 Global Food Intake (grams/day per adult) and Global Dietary Recommendations, 2018 10 Figure 1.8 Global and Regional Mean Absolute Differences in Alternative Healthy Eating Index (AHEI) Component Scores in Adults between 1990 and 2018 11 Figure 1.9 Drivers of Undernutrition 13 Figure 1.10 Ultra-Processed Foods and Adverse Health Outcomes 14 Figure B1.2.1 Pathways by Which Sugar Impacts Health 15 Figure 2.1 Public Support to Agrifood Sector, by Type of Support 28 Figure B2.1.1 Change in Productivity Due to Use of Nitrogen Fertilizer 29 Figure 2.2 Food Groups Based on Per Capita Consumption Relative to Dietary Guidelines 30 Table of Contents v Figure 2.3 Agrifood Support by Food Commodity 31 Figure B2.2.1 Differences between Recommended and Current Intake, by Food Group in Africa 33 Figure 2.4 Sugar Consumption, Taxes on SSBs, and Support to Sugar Producers 34 Figure 2.5 Greenhouse Gas Emissions of Food Commodities across the Supply Chain (per kg of food product), 2018 37 Figure 2.6 Repurposing Simulations at the Global Level: Potential Win-Win Scenarios 39 Figure 3.1 Composition of Agrifood Public Support 53 Figure 3.2 PSE vs CSE, 2020–22 58 Figure 3.3 MPS, Non-MPS, GSSE, and CSE, 1986–2022 58 Figure 3.4 Composition of Support to Grains, and PSCT and CSCT by Country, 2020–22 61 Figure 3.5 Grains: Daily Intake, MPS, and non-MPS, 1990–2018 62 Figure 3.6 Composition of Support to Meats and PSCT and CSCT by Country, 2020–22 64 Figure 3.7 Meats: Daily Intake, MPS, and non-MPS, 1990–2018 65 Figure 3.8 Composition of Support to Sugar, and PSCT and CSCT by Country, 2020–22 66 Figure 3.9 Sugar: Daily intake, MPS, and non-MPS, 1990–2020 67 Maps Map B1.3.1 Geographical Density of the Number of Dietary Surveys in the GDD 2017 by Country (A), including publicly available surveys (B), and non-public surveys submitted by data owners (C) GGD, Global Dietary Database 16 Map B2.3.1 Taxes on SSBs, as of August 2023  35 Map 3.1 Total Support and PSE across Countries 57 Boxes Box 1.1 Negative Externality Costs of the Food System 7 Box 1.2 Guidelines on the Consumption of Sugar, Sodium, Fats, and Alcohol 15 Box 2.1 A Closer Look at Fertilizer Use 29 Box 2.2 Differences between Recommended Intake and Current Intake, by Food Groups in Africa 33 Box 2.3 Global Implementation of Sugar-Sweetened Beverage Taxes  34 Box 2.4 Lessons in Transitioning Farmers away from Tobacco Farming  41 Box 3.1 Trade Policy and Obesity  49 Box 3.2 Indicator Definitions for Public Support Variables 53 Box A.1 Producer Support Estimate Categories and Subcategories 101 vi Reshaping the Agrifood Sector for Healthier Diets Acknowledgments This Knowledge Note was prepared by a team led by Felipe Dizon (Senior Economist, Agriculture and Food, World Bank). Chapter 1 was co-authored by Jenna Golan (Cornell University) and Elizabeth Ludwig-Borycz (University of Michigan). Chapter 2 was co-authored by Nikita Makarenko and Felipe Dizon (Agriculture and Food, World Bank), with inputs on tobacco farming from Danielle Bloom and Evan Blecher (Global Tax Program, World Bank) and inputs on sugar-sweetened beverage taxes from Kate Mandeville and Libby Hattersley (Health, Nutrition, and Population, World Bank). Chapter 3 was co-authored by Jisang Yu and Jiyeon Kim (Kansas State University) and Felipe Dizon and Mykyta Makarenko (World Bank), with inputs from Daniel Jefte Garcia Gomar (World Bank). For chapter 4, the case study on Bangladesh was led by Sakil Ahmmed (University of Dhaka) with support from Md Mansur Ahmed (Agriculture and Food, World Bank), while the case study on Malawi was led by Mirriam Matita (Lilongwe University of Agriculture and Natural Resources) and David Zingwe (University of Malawi). Substantial inputs were provided by Hina Sherwani (Agriculture and Food, World Bank) on earlier versions of the note. This note also greatly benefited from the support of Jubilee Nkechinyere Ahazie, Mariam Haidary, and Teguest Demissie Bekele (World Bank). Critical guidance in the work for this Knowledge Note was provided by Julian Lampietti (Practice Manager, Agriculture and Food, World Bank), Adama Toure (Global Lead for Food and Nutrition Security, World Bank), Geeta Sethi (Global Lead for Healthy and Sustainable Diets, World Bank), and Kyoko Shibata Okamura (Senior Nutrition Specialist, World Bank). The note greatly benefited from peer reviewers Sergiy Zorya (Global Lead for Agrifood Public Policies and Expenditures, World Bank), Meera Shekar (Global Lead for Nutrition, World Bank), and Valentina Pernechele (Economist, Food and Agriculture Organization of the United Nations, or FAO). The note was edited by Hope Steele and the GCS Print and Publishing Unit and was designed by Veronica Gadea. This work was made possible by the financial support of Food Systems 2030 and the Nutrition Multi-Donor Trust Fund. Acknowledgments vii Abbreviations and Acronyms Acronyms Description 1M5R One Must Five Reductions AHEI Alternative Healthy Eating Index AIP Affordable Inputs Program (Malawi) AMR anti-microbial resistance ASF animal source foods ATT Average Treatment Effect on the Treated AWD alternate wetting and drying BDT Bangladeshi taka (currency) BIHS Bangladesh Integrated Household Survey BMI body mass index BOT budgetary and other transfers BTs budgetary transfers CNPC Consumer Nominal Protection Coefficient CO2eq carbon dioxide equivalent CRD chronic respiratory disease CSCT Consumer Single Commodity Transfers CSE Consumer Support Estimate CVD cardiovascular disease DALYs disability-adjusted life years EFC excess feed costs FAO Food and Agriculture Organization of the United Nations FBDGs food-based dietary guidelines FCS Food Consumption Score FIES Food Insecurity Experience Scale FISP Farm Input Subsidy Program (Malawi) FOLU Food and Land Use Coalition GDD Global Dietary Database GDP gross domestic product GFR gross farm receipts GHG greenhouse gas GIFT FAO/WHO Global Individual Food Consumption Data Tool GLOPAN Global Panel on Agriculture and Food Systems for Nutrition GMM generalized method of moments GSS General Services Support GSSE General Services Support Estimate HDDS household dietary diversity score HFCS high-fructose corn syrup viii Reshaping the Agrifood Sector for Healthier Diets Acronyms Description HIV human immunodeficiency virus IAT inclusive agriculture transformation IDB Inter-American Development Bank IDD iodine-deficiency disorders IFAD International Fund for Agricultural Development IFPRI International Food Policy Research Institute IV Poisson instrumental variable Poisson LMICs low- and middle-income countries LSMS Living Standards and Measurement Studies LV price levies MK Malawian kwacha (currency) MPS Market Price Support MVAC Malawi Vulnerability Assessment Committee NAFTA North American Free Trade Agreement NCDs noncommunicable diseases NRA nominal rate of assistance NPP net primary productivity OECD Organisation for Economic Co-operation and Development OOA Old Age Allowance (Bangladesh) OR odds ratio PPP purchasing power parity PSCT Producer Single Commodity Transfers PSE Producer Support Estimate PSM propensity score matching SCTP Social Cash Transfer Program (Malawi) SDD Sustainable Development Goals SOFI State of Food Security and Nutrition in the World SSB sugar-sweetened beverages TCT Transfers to Consumers from Taxpayers TSE Total Support Estimate UNDP United Nations Development Programme UNEP United Nations Environment Programme VAD vitamin A deficiency WASH water sanitation and hygiene WFP World Food Programme WHO World Health Organization WRA women of reproductive age All dollar amounts are US dollars unless otherwise indicated. Abbreviations and Acronyms ix Executive Summary The world is off track in meeting Sustainable deficiencies, including those of iron, vitamin A, and Development Goal 2 (SDG2)— progress in iodine. Across all age groups, females have nearly curbing the various forms of malnutrition has twice the prevalence of anemia (31.2 percent) as slowed. Undernutrition in children, including males. Pregnant women and children under five stunting and wasting, has multifaceted causes. are at a higher risk of vitamin A deficiency. During The health and economic consequences of pregnancy, iodine-deficiency disorders can result undernutrition are also complex and can have in stillbirth, spontaneous abortions, and congenital intergenerational consequences that result in abnormalities such as cretinism. undernutrition and poverty in future generations. In recent years, the decline in undernutrition has The prevalence of overweight and obesity is plateaued. Most children under five affected by steadily increasing among children and adults. stunting and wasting live in Asia (52 percent and Overweight and obesity are a main risk factor 70 percent, respectively) or Africa (43 percent for many noncommunicable diseases (NCDs). and 27 percent, respectively). Globally, the The increase in NCDs has significant economic prevalence of stunting and wasting in children effects because of their high health care costs under five has decreased from 33.0 and 8.7 and associated loss of income. Once considered percent in 2000 to 22.3 and 6.8 percent in 2022. a high-income country problem, overweight and This decrease is inadequate to meet the 2030 obesity are on the rise in low- and middle-income SDG2 target of halving the number of children countries (LMICs). Obesity has negative health and affected by stunting; only about one-third of all economic consequences throughout the lifespan. countries are expected to meet this goal. The The largest proportion of children affected by prevalence of underweight in adolescent girls overweight live in Asia (48 percent) or Africa (28 (ages 10 to 19 years) has remained the same, percent). It is projected that, by 2030, 18 percent of at 8 percent, since 2000. In 1990, underweight all adults will be living with obesity. Dietary factors was more prevalent than obesity, when 14.5 account for 6 of the top 11 risk factors for NCDs. percent of females and 13.7 percent of males were underweight. This prevalence has steadily Failing to tackle malnutrition poses huge costs declined over the past three decades. In 2022, 7.0 to society. These health costs are the largest percent of females and 6.2 percent of males were hidden costs that come from our food systems. underweight. Underweight and thinness among The global food industry accounts for more than adults are associated with increased morbidity and 10 percent of global gross domestic product (GDP), mortality from infectious diseases and lower work with an estimated market value of $10 trillion. productivity and wages. In pregnant women, a low However, the current food system has significant body mass index is associated with intrauterine hidden health, environmental, and economic costs growth restriction and stunting and wasting in that exceed its market value by an estimated $2 early childhood. trillion. The hidden health costs, which include obesity and undernutrition, are the highest among Women and children continue to be the most these costs. The cost that the food systems impose vulnerable to micronutrient deficiencies. on health is $6.6 trillion dollars; on the environment Women of reproductive age (WRA) are at a high it is $3.1 trillion; and on the global economy it risk of undernutrition, including underweight and is $2.1 trillion. Globally, stunting is estimated to thinness in pregnant women. This undernutrition result in an annual per capita income penalty of 5 can result in adverse birth outcomes. Children percent. The global cost of overweight and obesity and WRA are at a higher risk of micronutrient was estimated to be $1.96 trillion in 2020. x Reshaping the Agrifood Sector for Healthier Diets Unhealthy diets are a core driver of this numerous adverse health outcomes, including malnutrition problem. On average, diets are cardiometabolic and mental disorders as well as currently poor. People are overconsuming mortality. In response to the increasing burden sugar, red meat, and processed meat while of diet-related diseases, the Food and Agriculture underconsuming whole grains, fruits, and Organization of the United Nations (FAO) and the vegetables. Improvements in diets have been slow, World Health Organization (WHO) have created in part because of the rise of the consumption dietary guidelines—also known as food-based of less healthy food such as sugar, red meat, dietary guidelines—that set recommendations and processed meat. Poor diet quality and for which foods to limit and which to consume in high food prices drive malnutrition—including greater quantities. These guidelines recommend undernutrition, stunting, wasting, micronutrient eliminating or significantly limiting the deficiencies, overweight, obesity, and the resulting consumption of free sugar, sodium (salt), trans rise in NCDs such as diabetes, hypertension, fats, and alcohol. In addition to limitations, they cardiovascular disease, kidney disease, cancer, recommend increasing the intake of healthy foods osteoarthritis, depression, asthma, and Alzheimer’s such as fruits, vegetables, legumes, seeds and disease. In many LMICs, consumption of animal nuts, and whole grains. These guidelines serve sourced foods (ASFs) is limited and those who as crucial goalposts of what does and does not need their nutritional benefits (for example, high- constitute a healthy diet. quality proteins and bioavailable micronutrients) the most (that is, WRA and children) have the Public support to the agrifood sector is large, least access. This contrasts with many people distortionary, and inefficient. Globally, support in high-income countries, where ASFs are over- to the agrifood sector comes to $854 billion per consumed thereby increasing the risk of NCDs. A year. Most of this support is focused on producers country’s unique situation needs to be taken into and comes largely in the form of trade and market account when making recommendations about policies that distort prices. Out of the total support ASF consumption. Such recommendations can to agrifood, 74 percent was targeted toward include adjusting prices to include environmental producers; most of these measures are market costs and public awareness campaigns in distortive. Over 50 percent of producer support high-income countries as well as making ASFs is in the form of trade and market policies, which more affordable in LMICs by increasing farm impact the market prices of agrifood commodities, productivity, improving market efficiency, and or Market Price Support (MPS). Smaller shares of raising household incomes. support are allocated to General Services Support (GSS, 12 percent). These are investments in private Food system shifts have precipitated a or public services—such as institutions and major dietary transition toward increasingly infrastructure. And even smaller shares are also ultra-processed foods and away from more allocated to consumer subsidies (13 percent)— traditional dietary patterns. These changes which can improve the consumption of nutritious have occurred over the past half-century, first in food if targeted effectively. Support is also often high-income countries and then in LMICs such as regressive, benefiting wealthier farmers, who use countries in Sub-Saharan Africa, South and East more inputs and produce more output. Notably, Asia, and Latin America. Ultra-processed foods are farmer subsidies are leading to excessive use of high in added sugar, sodium, and saturated fats, fertilizers, particularly in East Asia and the Pacific with sugar-sweetened beverages (SSBs) being and in South Asia. the primary source of added sugar consumed in most countries. Food system shifts in tandem Public support to the agrifood sector is with increased consumption of ultra-processed not geared toward promoting healthier foods have worsened diets globally, especially diets, making it potentially misaligned with in LMICs; these shifts are associated with public health policies. Public support to the Executive Summary xi agrifood sector is focused much more on food reductions. Because agrifood support is biased commodities that are already high in consumption toward less healthy and nutritious commodities, such as grains and meats, and much less on it potentially fuels the high costs of malnutrition. underconsumed, healthier food commodities such There is an imperative to rethink and repurpose as fruits, vegetables, and dairy products. Between agrifood support in a way that better promotes 2020 and 2022, in terms of the dollar value of healthy diets. While there are often tradeoffs to support, the most supported commodities were repurposing, global simulation exercises have maize, rice, pork, poultry meat, and beef and veal. demonstrated that there are some windows of In terms of the share of this support to gross farm opportunity for multiple wins across different receipts, the most supported commodities were outcomes—such as climate, nutrition, and poverty sugar, with around 24 percent of farmers’ incomes alleviation. In a World Bank–International Food from sugar deriving from public support, followed Policy Research Institute (IFPRI) report, one by maize, rice, and poultry meat and beef and veal. simulation scenario from 2020 to 2040 shows that For this reason, the current set of agrifood policies redirecting public support toward research and is probably not aligned with policies in other development and other technological investments sectors, such as in public health. For example, could lead to a 1.6 percent increase in real national some countries tax the consumption of SSBs while income, a 1 percent reduction in extreme poverty, at the same time they provide significant support an 18 percent reduction in the cost of healthy to domestic sugar production. In several of these diet, a 16 percent increase in crop production, countries with seemingly inconsistent policies an 11 percent increase in livestock, and a 41 around sugar consumption and production, percent reduction in emissions from agriculture levels of sugar consumption exceed the WHO and land use. In the 2022 State of Food Security recommendations. and Nutrition in the World (SOFI) Report, one simulation scenario from 2017 to 2030 shows that However, the scale and the composition of repurposing agrifood support (fiscal subsidies) agrifood public support varies largely across from producers to consumers and for healthier regions and across countries. In Africa, for foods can reduce extreme poverty by 0.06 percent example, support to the agrifood sector is globally (and by 0.22 percent among low-income generally low and support to this sector would countries), reduce greenhouse gas emissions by require simultaneously increasing and rebalancing 0.18 percent, and decrease the cost of a healthy support. In 2015, countries in Sub-Saharan Africa diet by 3.34 percent globally. provided only $680 million in agriculture subsidies. This number is extremely low, considering that But repurposing agrifood support requires a set agriculture constitutes 23 percent of GDP in Sub- of complementary interventions across various Saharan Africa and employs 60 percent of people sectors, including agrifood systems, health in the region. Some countries in Sub-Saharan systems, social protection, and the environment. Africa allocate only 6 percent of their expenditure It also requires a careful consideration of the to food and agriculture, which is well below political economy in terms of what is feasible and the African Union’s target of 10 percent. Many how to overcome political economy constraints. countries in Sub-Saharan Africa prioritize inputs Moreover, assessing options for repurposing (for example, fertilizers), allocating 88 percent of requires leveraging better data and techniques the total agriculture support to inputs. to assess the links between agrifood policy and support, food consumption, and healthy diets; it Recent simulation analyses show that there will require a rich set of country-level deep dives are windows of opportunity to repurpose that investigate policy reform simulations and their public support to the agrifood sector for impacts on health and environmental outcomes. healthier diets and better nutrition, while supporting poverty alleviation and emissions xii Reshaping the Agrifood Sector for Healthier Diets This Knowledge Note presents two sets of food consumption. The level of total non–market analytical work to further our understanding price support (non-MPS) has been growing quickly of the links between public support in the over time, while growth in total GSS has been agrifood sector and healthy diets. While the slow. In contrast, it is precisely GSS that is found analytical work here is mostly suggestive, it paves to increase overall productivity and translate into the way for more in-depth research to unpack increased consumption of all food commodities. the relationship of agrifood support with healthy There were no clear impacts of total non-MPS and diets. The first set of analytics uses cross-country MPS on consumption. But across all types of non– estimations to assess the potential correlation commodity-specific support, the impact of GSS of agrifood public support with healthy diets for is found to be positive on the consumption of all an average country. The second set of analytics commodities assessed. A 10 percent increase in GSS uses within-country estimations to assess the leads to a 0.14 percent increase in consumption of correlation of agrifood public support with healthy grains, a 0.22 percent increase in the consumption diets in the context of a particular country. The of meats, and a 0.35 percent increase in the overall principles that underpin these analyses consumption of sugars. Looking at commodity- are: (1) that not all types of agrifood public specific support, an increase in commodity-specific support are the same, (2) that not all types of food MPS is seen to decrease the consumption of grains commodities are the same, (3) that the impacts of and sugar, which are often more easily traded a given type of support can vary depending on the than perishable meat products, for example. type of food commodity, and (4) that the impacts A 10 percent increase in grain-specific MPS reduces of various policy options on healthy diets will vary consumption of grains by 0.35 percent and a across countries. These analytics help to foster 10 percent increase in sugar-specific MPS reduces more dialogue and a deeper exploration of the consumption of sugar by 0.60 percent. These links between agrifood support and healthy diets. distortionary commodity-specific MPS measures increase domestic prices and reduce consumption. The cross-country analysis finds that GSS (that Thus, policies to remove commodity-specific MPS is, public goods) increases the consumption must be accompanied by complementary measures of food commodities, while distortionary MPS to curb consumption of less healthy and already- reduces consumption, particularly for more overconsumed commodities. easily traded and less perishable commodities such as grains and sugar. The global cross- Two within-country case studies look at the country analysis makes use of a database that impacts of various public interventions on merges detailed agrifood public support data healthy diets, one in Bangladesh and one in with consumption of various commodities Malawi. For both countries, the work compares across years and countries. The data are being the impacts of input subsidies, rural agriculture made publicly available to encourage further infrastructure, and social protection programs. research.1 The analysis explores the correlation of The cross-country analysis indicates that while the levels of different types of agrifood support there are correlations between MPS policies and with the consumption of different types of food consumption, there were no clear correlations commodities. Building on the global simulation between agrifood subsidies and consumption, on work on repurposing, this work unpacks some key average. This could suggest that country context assumptions of how repurposing public support varies significantly and that the impact of various could impact healthy diets by estimating the policy and intervention options are largely country- actual correlations between public support and specific. The goal of these case studies is to look at the country-level impacts of various types of agrifood sector support (that is, input subsidies, 1 The data are available here: https://datacatalog.worldbank.org/ search/dataset/0066597/agrifood_public_support_and_food_ irrigation, infrastructure, and social protection) on consumption. measures of food security and healthy diets. Executive Summary xiii In the case of Bangladesh, contrary to the On the other hand, subsidies are mainly used impact of social protection programs and for growing maize, which increases monotony farmer input subsidies, there is a positive rather than diversity of production. In other impact of rural infrastructure development words, the input subsidy program increases the on various measures of healthy diets. Looking quantity and not the quality of food. In contrast, into the impact of the social protection program food and cash transfers increase the likelihood as proxied by the country’s Old Age Allowance of undertaking negative coping strategies, but program, there are no significant impacts of the they lead to improved diet diversity and a higher program on the food consumption score, dietary food consumption score. Households that receive diversity score, the food insecurity experience food and cash transfers undertake negative scale, or the household hunger scale. Previous coping strategies for food insecurity between 22 literature confirms that the impact depends on the and 26 percent more on average, which could size of the benefits as well as the duration of the be a manifestation of the characteristics of the support. The benefits of the Old Age Allowance beneficiaries who are often poorer. However, program are very small, and they might fall short receiving food and cash transfers increases of diversifying the food basket of the beneficiary household dietary diversity by at least two food households. The impact of input subsidies on groups and raises the food consumption score food security and nutrition is also insignificant, above acceptable levels of food security. This consistent with other studies. Although the input implies that food and cash transfers support subsidy helps farmers increase production, the market purchases and consumption of a wider size of the benefits as well as their duration variety of foods, but it does not support the might be inadequate to improve food security. In quantity of food consumption. Cash transfers can contrast, households having access to improved be used to purchase food and non-food items, in roads have a significantly higher per capita food contrast to input subsidies, which are only used expenditure and diversified food basket than those to produce food crops. Finally, in the Malawi case having no access to such roads. In this analysis, study, irrigation infrastructure does not impact direct transfer programs, such as social protection food and nutrition security, potentially because of and input subsidies, are less effective than indirect its low coverage and capacity. But other studies support programs such as rural infrastructure suggest positive effects of improved irrigation. development. There are important efforts to Large-scale irrigation investments, such as rethink the effectiveness of fertilizer subsidies the Shire Valley Transformation Project, could in Bangladesh, including piloting an e-voucher transform food systems as well as enhance food system with support from the World Bank. and nutrition security and household resilience in Malawi. In the case of Malawi, farmer input subsidies and food and cash transfers have different Looking at the various interventions together impacts on food and nutrition security, while in Malawi, there are no positive impacts of access to irrigation schemes does not have an a combination of input subsidies, food and impact on diets. Input subsidies, which typically cash transfers, and irrigation infrastructure. support maize production, reduce the likelihood Nevertheless, households receiving food and of undertaking negative coping strategies, but do cash transfers alone showed improvements in not improve diet diversity or the food consumption food consumption and dietary diversity scores; score. Households that receive input subsidies and households receiving input subsidies alone undertake negative coping strategies for food helped reduce negative coping strategies with insecurity between 13 and 16 percent less on food insecurity, but an over-reliance on agriculture average. This implies that food availability through input subsidies leads to reduced variety in own-food production makes it less likely for consumption. The findings suggest that the households to use negative coping mechanisms. current standalone implementation strategy may xiv Reshaping the Agrifood Sector for Healthier Diets not offer opportunities to build synergies between investments in irrigation infrastructure and social assistance. A coordinated strategy across these programs may be necessary to support both food quantity and food quality. Like in Bangladesh, the World Bank is supporting efforts to rethink input subsidies in Malawi, by promoting efficient usage of fertilizer subsidies and incentivizing diversification of crop production. It is important that we continue to build on this knowledge agenda to demonstrate the value of repurposing agrifood public policies and support for healthy diets. Cross-country analysis globally or regionally can shed light on trade policy for healthy diets. Moreover, a deeper dive into political economy analysis can elucidate lessons from successful cases. Within-country analyses can include better articulations of the hidden health and nutrition costs of food systems and the impact of various policy options to mitigate these costs. Similarly, within-country analyses can also assess the climate and nutrition tradeoffs and win-wins across various policy options. This Note is organized as follows. Chapter 1 establishes that the costs of malnutrition are high, that progress toward curbing malnutrition has slowed, and that the promotion of healthy diets is a cornerstone to tackling malnutrition. Chapter 2 demonstrates that current public support to the agrifood sector does not promote healthy diets, and that there are windows of opportunity to repurpose public support to the agrifood sector in a way that better supports diets. Chapter 3 presents the cross-country analytical work, while chapter 4 presents the case studies for Bangladesh and Malawi. Chapter 5 provides some future directions. Appendix A defines the terms of the Organisation for Economic Co-operation and Development (OECD)’s database on agrifood support with examples, appendix B provides information on the Global Dietary Database (GDD) on Food Consumption, and appendix C provides cross-country regression estimations. Executive Summary xv xvi Reshaping the Agrifood Sector for Healthier Diets 1 Failing to Tackle Malnutrition the prevalence of stunting in children under five 1.1. The growing triple burden has decreased from 33.0 percent (204.2 million of malnutrition cases) in 2000 to 22.3 percent (148.1 million cases) in 2022 (figure 1.1). This decrease is inadequate Undernutrition in children, including stunting to meet the 2030 Sustainable Development Goal and wasting, have multifaceted causes. Its (SDG) target of halving the number of children health and economic consequences can have affected by stunting; only about one-third of all countries are expected to meet this goal. In Asia, intergenerational significance, resulting the prevalence of stunting decreased from 28.2 in undernutrition and poverty in future percent (106.8 million) in 2012 to 21.3 percent (76.6 generations. In recent years, the decline in million) in 2022. In Africa, while the prevalence undernutrition has plateaued. of stunting has decreased, the total number of children affected by stunting has increased from The lifelong consequences of stunting in childhood 61.3 million (34.4 percent) in 2012 to 63.1 million increase the risk of stunting in the next generation, (30.0 percent) in 2022 (UNICEF/WHO/World Bank contributing to intergenerational cycles of poverty Group 2023). and malnutrition. Stunting, a measure of chronic undernutrition for children, has serious short- and Wasting causes developmental delays, disease, long-term consequences. It can begin in utero and death (UNICEF 2021). Wasting is characterized due to poor maternal nutrition and/or recurrent by low weight-for-height as a measure of acute infections and can continue through childhood malnutrition. Children affected by wasting have (Christian, Afful-Dadzie, and Marquis 2023; de Onis weakened immune systems, increasing their risk and Branca 2016). Stunting can occur concurrently of infection and worsening their nutritional status with wasting and overweight/obesity (Steyn and (UNICEF/WHO/World Bank Group 2023). Over time, Nel 2022; Thurstans et al. 2022). In childhood, wasting can lead to stunting if acute malnutrition stunting is associated with increased morbidity and persists (Thurstans et al. 2022). mortality from infection. In school-age children, stunting is associated with lower educational The majority of children affected by wasting live in attainment (including both lower grades and fewer Asia (70 percent) and Africa (27 percent). Globally, years of educational attainment) (Hoddinott et al. the prevalence of wasting has decreased from 2013; Martorell et al. 2010). In adulthood, stunting 8.7 percent (54.1 million cases) in 2000 to 6.8 during childhood is associated with economic percent (45.0 million cases) in 2022 (figure 1.1). The losses through a loss of physical growth potential, prevalence of wasting is highest in Southern Asia, cognitive impairments, and a higher risk of chronic where 14.3 percent of children are affected. Five diseases (Hoddinott et al. 2013; Huxley, Shiell, and countries—India, South Sudan, Sri Lanka, Sudan, Law 2000; Whincup et al. 2008). and Yemen—had a prevalence of wasting that Most children under five affected by stunting live in exceeded 15.0 percent (UNICEF/WHO/World Bank Asia (52 percent) and Africa (43 percent). Globally, Group 2023) Failing to Tackle Malnutrition 1 Figure 1.1  Trends in the Prevalence of Stunting, Wasting, and Overweight in Children under Five Percentage Number (millions) 40 220 204.2 197.1 200 35 184.8 180 169.1 30 33.0 154.3 160 25 140 148.1 120 20 22.3 100 15 80 8.7 60 54.1 52.9 51.2 49.5 10 6.8 46.4 45.0 40 5 35.5 36.2 38.0 37.8 37.0 33.0 5.3 5.6 20 0 0 00 05 10 15 20 22 00 05 10 15 20 22 20 20 20 20 20 20 20 20 20 20 20 20 Percentage of children under 5 Number (millions) of children under 5 affected by stunting, wasting and affected by stunting, wasting and overweight, global, 2000–22 overweight, global, 2000–22 Stunting Wasting Overweight 95% confidence interval Source: UNICEF/WHO/World Bank Group 2023. underweight was more prevalent than obesity Women and children are the most vulnerable. (NCD-RisC 2024), when 14.5 percent of females Women of reproductive age are at high risk and 13.7 percent of males were underweight. The of undernutrition, including underweight prevalence has steadily declined over the past and thinness in pregnant women, and three decades. In 2022, 7.0 percent of females and 6.2 percent of males were underweight undernutrition can result in adverse (NCD-RisC 2023; NCD-RisC 2024). Underweight birth outcomes. Children and women of and thinness among adults are associated with reproductive age are at a higher risk of increased morbidity and mortality from infectious micronutrient deficiencies, including those of diseases and lower work productivity and wages. iron, vitamin A, and iodine. In pregnant women, a low body mass index (BMI) is associated with intrauterine growth restriction The prevalence of underweight in adolescent girls and stunting and wasting in early childhood (ages 10 to 19 years) has remained the same, (UNICEF 2023). at 8 percent, since 2000 (UNICEF 2023). In 1990, 2 Reshaping the Agrifood Sector for Healthier Diets Short height is associated with a higher risk of highest prevalence in South Asia and Sub-Saharan adverse birth outcomes, including preterm birth Africa. and small-for-gestational-age infants (Kozuki et al. 2015; Özaltin, Hill, and Subramanian 2010; During pregnancy, iodine-deficiency disorders Victora et al. 2021). Short height is an indicator of (IDD) can result in stillbirth, spontaneous intergenerational chronic undernutrition. It is both abortions, and congenital abnormalities such a cause and consequence of childhood stunting. In as cretinism. During infancy, childhood, and 2020, in South Asia, 69 percent of adolescent girls adolescence, IDD can impair skeletal and central and women were shorter than 155 centimeters, nervous system development and can lead to 35 percent were shorter than 150 centimeters, hypothyroidism or hyperthyroidism (Zimmermann and 11 percent were shorter than 145 centimeters and Andersson 2021). Since 2003, the number (UNICEF 2023). of countries with adequate iodine intake has nearly doubled, reaching 118 in 2020. However, Across all age groups, females have nearly 21 countries remained iodine deficient, including twice the prevalence of anemia (31.2 percent) as Burkina Faso, Burundi, Cambodia, Finland, males (GBD 2021 Anaemia Collaborators 2023). Germany, Haiti, Israel, Iraq, the Republic of Since 1990, the global prevalence of anemia has Korea, Lebanon, Madagascar, Mali, Mozambique, decreased by only about 4 percent (24.3 percent in Nicaragua, Norway, the Russian Federation, 2021) and in children under five, the prevalence of Samoa, South Sudan, Tajikistan, Vanuatu, and anemia was highest in Africa (60.2 percent) (WHO Viet Nam (Zimmermann and Andersson 2021). GHO 2024a). Inadequate dietary iron intake is a leading cause of anemia (WHO 2023a). Children The prevalence of overweight and obesity has and women of reproductive age are at higher continued to increase in children and adults. risk of anemia. Anemia has been associated with increased morbidity and mortality across the Overweight and obesity are a main risk factor life course (Chaparro and Suchdev 2019), during for many noncommunicable diseases. The pregnancy with poor birth outcomes (Haider et al. increase in noncommunicable diseases has 2013), during childhood with impaired cognitive significant economic effects as a result of the and behavioral development (Walker et al. 2007), associated high health care costs and loss of and during adulthood with decreased work income. productivity and lower income (Haas and Brownlie IV 2001). Once considered a high-income country problem, overweight and obesity are on the rise in low- and Pregnant women and children under five are at middle-income countries (figure 1.2). It is one a higher risk of vitamin A deficiency (VAD). VAD of the few indicators of malnutrition that has can cause vision problems and is associated with increased over the past 20 years. Overweight a higher risk of infectious diseases, including and obesity occur when there is an imbalance measles, diarrhea, respiratory diseases, and of energy intake and energy requirements. In death. In 2019, the global prevalence of VAD most cases, it is a multifactorial disease due to was 6.8 percent; among children one through psychosocial factors, genetics, and obesogenic four years old, the prevalence was more than environments (WHO 2024a), which promote high twice that (16.0 percent) (Hess et al. 2022), with the energy intake and sedentary behavior (WHO 2016). Failing to Tackle Malnutrition 3 Figure 1.2  Global Trends in Adult BMI Proportion of the population (%) Indicator 50 Overweight 25 Obesity Underweight 0 1990 2000 2010 2020 Year Source: Original figure for this publication based on data from WHO GHO 2024b. Note: BMI = body mass index. Obesity has negative health and economic live in the Americas (66.9 percent) and the lowest consequences throughout the lifespan. During in Southeast Asia (30.2 percent) and Africa (30.7 childhood and adolescence, overweight and percent). Low- and middle-income countries are obesity are associated with a greater risk and catching up to the high prevalence of overweight earlier onset of several NCDs, including type 2 in high-income countries, but its prevalence is still diabetes and cardiovascular diseases along with rising in high-income countries. In high-income, adverse psychosocial consequences, which affect upper-middle-income, lower-middle-income, and school performance and quality of life. Children low-income countries the prevalence of overweight with obesity are most likely to continue to live with increased from 1990 to 2022 by 42.4 percent to obesity in adulthood (WHO 2024a), which further 56.2 percent, 23.6 percent to 45.9 percent, 13.7 increases the risk of NCDs, including cardiovascular percent to 35.4 percent, and 10.5 percent to 26.4 diseases, diabetes, cancers, neurological disorders, percent, respectively (WHO GHO 2024b). chronic respiratory diseases, and digestive disorders (GBD 2019 Risk Factors Collaborators It is projected that, by 2030, 18 percent of all adults 2020). It also has significant economic impacts will be living with obesity (Lobstein et al. 2023). through higher health care costs and decreased Globally, the prevalence of obesity in adults more income (Okunogbe et al. 2022). than doubled from 6.6 percent in 1990 to 15.8 percent in 2022 (WHO GHO 2024b). The bulk of this The largest proportion of children affected by growth occurred in high-income countries (16.9 overweight live in Asia (48 percent) and Africa (28 percent in 2000 to 25.9 percent in 2022); however, percent). In children and adolescents, overweight obesity prevalence increased in low- and middle- rose from 8 percent in 1990 to 20 percent in 2022 income countries too, ranging from 3.6 percent in and from 25.2 percent to 42.8 percent among 2000 to 15.8 percent in 2022 (figure 1.3), with the adults during the same time. Among adults who Americas seeing the largest increase—from 12.9 are living with overweight, the highest percentage percent to 33.8 percent (WHO GHO 2024b). 4 Reshaping the Agrifood Sector for Healthier Diets Figure 1.3  Regional Trends in Obesity Proportion of the population (%) 50 Region Americas Eastern Mediterranean 25 Europe Africa Western Pacific Southeast Asia 0 1990 2000 2010 2020 Year Source: Original figure for this publication based on data from WHO GHO 2024b and World Obesity Federation 2023. Dietary factors account for 6 of the top 11 risk in 2000. Four NCDs—cardiovascular disease (CVD), factors for NCDs (figure 1.4) (WHO 2023c). The cancer, chronic respiratory disease (CRD), and major diet-related risk factors for NCDs are alcohol diabetes—killed about 33.3 million people in 2019, consumption, obesity, hypertension, child and a 28 percent increase from 2000 (WHO 2023c). It is maternal malnutrition, high cholesterol, and high expected that NCDs will account for 86 percent of fasting plasma glucose. NCDs caused 74 percent of global deaths by 2048 (figure 1.5). global deaths in 2019, an increase from 61 percent Figure 1.4  Diet-Related NCDs Six of the top 11 risk factors driving the global burden of disease related to diets Dietary risks 240 High sytolic blood pressure 210 Child and maternal malnutrition 180 Air pollution 145 Tobacco smoke 145 High BMI 130 Alcohol and drug use 120 High fasting plasma gulucose 110 Unsafe water 80 Unsafe sex 70 High cholestrol 60 0 50 100 150 200 250 300 Global all-age disability-adjusted life year (in millions, 2019) Source: IHME 2019. Note: NCDs = noncommunicable diseases. Failing to Tackle Malnutrition 5 Figure 1.5  Composition of Cause of Death, WHO Regions and Global, 2000-2048 Percentage (%) African region Region of the Americas Eastern Mediterranean region European region 100 80 60 40 20 0 00 10 20 30 40 20 8 00 10 20 30 40 20 8 00 10 20 30 40 20 8 00 10 20 30 40 48 4 4 4 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 South-East Asia region Western Pacific Global 100 80 60 40 20 0 00 10 20 30 40 20 8 00 10 20 30 40 20 8 00 10 20 30 40 48 4 4 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Communicable, maternal, perinatal and nutritional conditions Injuries NCDs Source: WHO 2023c. Note: NCDs = noncommunicable diseases. see figure 1.6). Human capital losses have been 1.2. The high costs of quantified through the Human Capital Index which malnutrition is a measure of both the expected education and health of a child born today (World Bank 2018). The The hidden costs of food systems exceed its index has three components: (1) survival (under- market value by an estimated $2 trillion in five mortality rate), (2) expected years of school, negative externalities; hidden health costs are and (3) health (the rate of stunting of children the highest of these. The global food industry under age five and the adult survival rate, defined accounts for more than 10 percent of global gross as the proportion of fifteen-year-olds who will domestic product (GDP), with an estimated market survive until age sixty). Those most affected by the value of $10 trillion. However, the current food hidden costs to health live in Asia and Sub-Saharan system has significant negative externalities of $12 Africa (Burgess et al. 2019; FOLU 2019a). Negative trillion—such as hidden health, environmental, externalities for the hidden health costs of the and economic costs—exceeding its market food system can be quantified for malnutrition, value. The highest cost is to health at $6.6 trillion including undernutrition, overweight, and obesity dollars, then to the environment at $3.1 trillion, (see box 1.1). and economic costs at $2.1 trillion (FOLU 2019a; 6 Reshaping the Agrifood Sector for Healthier Diets Figure 1.6  Hidden Costs of Food and Land Uses Trillions USD, 2018 prices 10.0 10 9 2.7 8 7 1.8 6 5 2.1 4 3 6.6 1.5 2 1 1.7 0 3.1 0.8 –1 1.3 –1.9 –2 –3 2.1 Market Value of Health Environment Economic Food System Value Global Food System Net of Hidden Costs Obesity Greenhouse Gas Emissions Rural Welfare Undernutrition Natural Capital Costs Food Loss & Waste & Fertiliser Leakage Pollution, Pesticides & Anti-Microbial Resistance Source: FOLU 2019a. Note: FOLU = Food and Land Use Coalition. Box 1.1 Negative Externality Costs of the Food System The hidden health costs for the food system ozone pollution and caused by pollution from were estimated by multiplying the gross household solid cooking fuels. The cost of domestic product (GDP) per capita by relevant air pollution also considered the total global disability-adjusted life years (DALYs). The greenhouse gas (GHG) emissions from food GDP per capita, purchasing power parity (2018 and land use systems and the proportion of international $), was the average output per solid cooking fuels from biomass. Pesticide capita for different regions of the world in exposure included DALYs per kilogram of 2018 international dollars. DALYs are the sum insecticide, herbicide, fungicide, and bactericide of years of life lost due to premature mortality multiplied by the metric tons of annual and the years lived with a disability. For obesity, pesticides applied. Anti-microbial resistance DALYs related to high body mass index (BMI) (AMR) did not use GDP per capita. Instead, this risk factors were included. Undernutrition was estimated by multiplying the percentage included DALYs related to child growth of AMR related to food systems by the total failure, including child stunting, wasting, and annual GDP loss attributed to AMR associated underweight. Air pollution included DALYS with human immunodeficiency virus (HIV), related to ambient particulate matter and tuberculosis, malaria, E. coli, S. aureus, and K. Failing to Tackle Malnutrition 7 pneumoniae. It is likely that this methodology are unsustainable or at risk of becoming missed additional hidden costs that the food unsustainable. Biodiversity loss was estimated and land use system have on health. using the economic value of ecosystem services from tropical forests and mangroves and the The costs of the food and land use system annual rate of deforestation and mangrove loss were estimated through GHG and natural caused by agriculture and aquaculture. The cost capital costs, including land degradation, of overexploitation was estimated using the water scarcity, biodiversity loss, and value of crop production reliant on pollinator overexploitation of land. The cost of GHG services, the yield reduction from the loss of emissions was estimated using the total pollinators, and the economic cost of over- annual GHG emissions from food and land use fishing beyond the maximum sustainable yield. systems, the production of nitrogen fertilizer, the GHG emissions from the production of The hidden economic costs include rural nitrogen fertilizer, and the marginal abatement welfare, food loss and waste, and fertilizer costs for GHG emissions. The cost of land leakage. The costs from rural welfare were degradation was estimated using the total area estimated using the respective region’s poverty of degraded land, the annual value of crop line, the average rural poverty gap, and the production per hectare of cropland, the annual total population living below the poverty line. value of livestock production per hectare of Food loss and waste were estimated using the grassland, the economic value of soil ecosystem share of the total food production that is lost services per hectare, and the percent loss of or wasted and the annual value of agricultural soil biodiversity from land degradation. Water production. Fertilizer leakage was estimated scarcity was estimated by the total annual using the average leakage rate, total annual freshwater withdrawals for agriculture, the application, and global average price nutrients scarcity cost of water, and the percent of for nitrate and phosphate fertilizers. freshwater withdrawal for agriculture that Source: FOLU 2019b. Malnutrition—including undernutrition, considered the loss of income due to education, overweight, and obesity—has huge negative height, and cognition as a result of stunting in externality costs through high health care childhood (Galasso and Wagstaff 2019). Additional economic impacts due to health care costs costs and loss of income. These costs are throughout the lifespan have not been estimated. incurred throughout the lifespan and have (Note: The hidden costs of stunting have been national consequences beyond the individual. estimated quite differently than the estimates of the hidden costs of negative externalities from the Globally, stunting is estimated to result in an food system.) annual per capita income penalty of 5 percent. The highest impact was in South Asia, which was The global cost of overweight and obesity in estimated to have a 9 percent decrease in per 2020 is estimated to be $1.96 trillion. This is capita income. Sub-Saharan Africa was estimated expected to rise to $4.32 trillion in 2035 (Lobstein to have a 6 percent decrease in per capita et al. 2023; Okunogbe et al. 2022). Obesity income. North America had the lowest decrease has significant health and economic costs. in per capita income at 2 percent. The estimates 8 Reshaping the Agrifood Sector for Healthier Diets Overweight and obesity are significant risk factors 1.3. Poor diets are a crucial part of NCDs, which increase health care costs. In of malnutrition addition to increasing mortality, obesity reduces productivity, increases disabilities, increases health care costs, results in early retirement, Diets are currently poor, with people and reduces the length of disability-free healthy overconsuming sugar, red meat, and living across the life cycle, which impacts processed meat while underconsuming whole human capital outcomes (Shekar and Popkin 2020). The estimated health care cost of obesity grains, fruits, and vegetables. Improvements includes the cost of 38 disease conditions that in diets have been slow in part because of the are associated with a high BMI (GBD 2016 Risk rise in the consumption of unhealthy foods Factors Collaborators 2017; Lobstein et al. 2023). such as sugar, red meat, and processed meat. There are likely to be additional health care costs from additional comorbidities, including mental Globally, diets fall short of meeting dietary health and neurological conditions, endocrine recommendations. The World Health Organization disorders, respiratory conditions, and dental (WHO) recommendations and food-based dietary caries. The economic costs included the impact guidelines (FBDGs) show that adults’ dietary of high BMI on economic productivity, including intake falls far short of meeting the recommended absenteeism, reduced productivity at work, and consumption of fruits, vegetables, and whole premature retirement or death. There may be grains while exceeding recommendations for additional economic costs from lower educational sugar, mainly in the form of sugar-sweetened attainment, unemployment, long-term disability, beverages (SSBs), and both processed and and early retirement (Lobstein et al. 2023). (Note: unprocessed meats (figure 1.7) (FAO et al. 2023). The hidden costs of overweight and obesity have The WHO recommends a minimum intake of 200 been estimated quite differently than those of the grams per day for fruits and vegetables while negative externalities from the food system.) the FBSGs recommend just shy of 300 grams per day of fruit and more than 350 grams per day of Estimating the diet-related hidden health costs vegetables. In 2018, adults globally ate less than of the food systems at the country level is an 100 grams of fruit per day and about 150 grams emerging and important piece of work. Such of vegetables per day. Adults consumed only analysis enables better policy dialogue framed around 50 grams of whole grains in comparison to around a more accurate view of the magnitude the more than 300 grams per day recommended of the problem and an outline of possible food by the FBSGs. The WHO recommends less than systems solutions, tailored to a specific country. 50 grams of sugar consumed per day and global A new method jointly developed by FAO and the consumption on SSBs alone far exceeds the 50 World Bank was tested in the Philippines and grams per day (FAO et al. 2023). Ethiopia. That work found that unhealthy diets contribute significantly to child stunting and Improvements to diets have been slow over adult NCDs. For instance, child stunting costs time in part because of the rise in unhealthy in the Philippines and Ethiopia amounted to 1.5 food consumption such as sugar, red meat, and percent and 16.5 percent of GDP, respectively. processed meat (figure 1.8). The global overall Moreover, the economic costs (including treatment diet quality has improved modestly since 1990, and premature mortality) of diet-related NCDs as measured by the Alternative Healthy Eating amounted to 4.4 percent of GDP in the Philippines Index (AHEI) in figure 1.8 (Miller et al. 2022). and 1.9 percent of GDP in Ethiopia (FAO and World Some of these gains were due to increased fruit Bank, forthcoming). consumption, by about 5.3 grams per day, and increased nut and seed consumption, by about Failing to Tackle Malnutrition 9 2.3 grams per day. The gains were countered by et al. 2015). The AHEI did not improve in every losses from decreased consumption of whole region, though; in South Asia it stayed roughly the grains by 8.5 grams per day along with increased same and in Sub-Saharan Africa it declined, largely red meat and SSB consumption by 1.5 grams per because of increased SSB consumption (Miller et day and 0.37 servings (8 ounces = 248 grams) per al. 2022). week, respectively (Lara-Castor et al. 2023; Micha Figure 1.7  Global Food Intake (grams/day per adult) and Global Dietary Recommendations, 2018 Fruits Vegetables Beans & Legumes Nuts & Seeds Whole Grain Milk Fish Unprocessed Meats Processed Meats SSBs Fruits Juices 0 50 100 150 200 250 300 350 400 WHO min recommendation for Fruits & Vegs intake (>200 g/day each) WHO max recommendation for sugar intake (<50 g/day) FBDGs recommendation based on Healthy Diets Basket (g/day) WRC/IARC’s recommendations for meats: 50–71 g/day of unprocessed meat; and 0 g/day of processed meat Global adult consumption in grams/day Source: Original figure for this publication based on World Bank data. Data for global food intake are from the Global Dietary Database (GDD) 2023. 10 Reshaping the Agrifood Sector for Healthier Diets Figure 1.8  Global and Regional Mean Absolute Differences in Alternative Healthy Eating Index (AHEI) Component Scores in Adults between 1990 and 2018 Absolute difference (2018–1990) 5.0 2.5 0 –2.5 –5.0 World Southeast/ Central/ High-income Latin Middle East/ South Asia Sub- East Asia Eastern Europe countries America/ North Africa Saharan and Central Asia Caribbean Africa AHEI score Fruit Vegetables Whole grains Legumes/nuts Red/processed meat SSBs PUFA Seafood omega-3 Sodium Source: Miller et al. 2022. http://creativecommons.org/licenses/by/4.0/ and add to the burden of wasting and stunting Poor diet quality and high food prices drive (Skoufias, Vinha, and Sato 2019). A scoping poor malnutrition—including undernutrition, literature review on the health and nutrition costs stunting, wasting, micronutrient deficiencies, of poor diets in low- and middle-income countries overweight, obesity, and the consequent was conducted in 2024 and helped quantify the noncommunicable diseases. Food system association between poor diets and stunting (Siekmans et al., 2024). A meta-analysis found shifts are part of the problem because they that children five- to eighteen-years old living have precipitated a major dietary transition in low- and middle-income countries who had a toward increasingly ultra-processed foods low diet diversity score were associated with an away from more traditional dietary patterns. increased risk of stunting, with a stunting odds ratio (OR) = 1.43 (n = 14 studies) and wasting Poor diets drive undernutrition, such as stunting, OR = 2.18 (n = 2 studies) (Zeinalabedini et al. wasting, and micronutrient deficiencies. Poor diets 2023). The results are similar for children under that lack diversity, quality, and sufficient quantity five (Abdulahi et al. 2017). This research helps are associated with micronutrient deficiencies underscore the integral nature poor diets play in (such as deficiencies of iodine, vitamin A, and iron) contributing to the burden of stunting globally, Failing to Tackle Malnutrition 11 along with highlighting healthy diets as a critical middle-income countries. Their research found component in mitigation strategies. In 2019, that a 5 percent increase in the real price of food Skoufias, Vinha, and Sato conducted an analysis of increased the risk of wasting by 9 percent and 33 Demographic Health Surveys from Sub-Saharan severe wasting by 14 percent. The risk of moderate Africa to quantify the marginal effects on the and severe stunting for children two through five probability of stunting based on children’s access years old was increased if food inflation occurred to the three determinants of nutrition: (1) food and during pregnancy by 1.6 percent and 2.4 percent care; (2) water, sanitation and hygiene (WASH); and and if it occurred during the first year after birth (3) health. Their study found that a child’s access to by 1.8 percent and 3.4 percent, respectively. These all three determinants of nutrition was associated effects were more severe for children from poor with a 5.4 percent decrease in the probability of and rural households. Wasting, an acute measure stunting, while the independent effect of a child’s of malnutrition, was more severely impacted access to food and care decreased the probability following a 5 percent increase in food prices. of stunning by 0.3 percent (Skoufias, Vinha, and However, increased food prices went beyond Sato 2019). increasing acute malnutrition to affect chronic malnutrition as well, by increasing the risk of In many LMICs consumption of animal sourced stunting (Headey and Ruel 2023). It is evident that foods (ASFs) is limited and those who need their poor diets are exacerbated as food prices increase, nutritional benefits (high-quality proteins and further driving increased undernutrition and bioavailable micronutrients) the most (women stunting (figure 1.9 shows a conceptual framework). of reproductive age and children) have the Increased food prices make nutritious foods less least access (International Food Policy Research accessible to households by making them less Institute 2024). This contrasts with many people in affordable (Skoufias, Vinha, and Sato 2019). HICs, where ASFs are over-consumed increasing the risk of NCDs. The country’s unique situation Food system shifts have precipitated a major needs to be taken into account when making dietary transition toward increasingly ultra- recommendations about ASF consumption processed foods and away from more traditional and can include adjusting prices to include dietary patterns. These changes have occurred environmental costs and public awareness over the past half-century, first in high-income campaigns in HICs and making ASFs more countries and then in low-and middle-income affordable in LMICs through increased farm countries such as Sub-Saharan Africa, South and productivity, improved market efficiency, and East Asia, and Latin America (Gill et al. 2023; Lane raising household incomes. et al. 2024). Ultra-processed foods are high in added sugar, sodium, and saturated fats, with High food prices are also known to drive SSBs being the primary source of added sugar undernutrition. Food prices are integral to the consumed in most countries (Baker and Friel burden of nutrition-related diseases as they 2014; Baker et al. 2020; Monteiro et al. 2013, 2018; impact the accessibility of a diverse, high-quality, Popkin and Hawkes 2016; Popkin and Reardon and sufficient quantity of food. This is especially 2018; Reardon et al. 2021). Food system shifts in true for people of low socioeconomic status tandem with increased ultra-processed foods have in low- and middle-income countries, as these worsened diets globally, especially in low- and people spend a larger proportion of their budget middle-income countries, and are associated with on food than those of higher socioeconomic numerous adverse health outcomes including status (Skoufias, Vinha, and Sato 2019). Headey cardiometabolic and mental disorders as well as and Ruel recently assessed the impact of food mortality (figure 1.10). inflation on wasting and stunting in 44 low- and 12 Reshaping the Agrifood Sector for Healthier Diets Figure 1.9  Drivers of Undernutrition Long-term consequences: Adult height, cognitive ability, economic Short-term consequences: productivity, reproductive performance, Mortality, morbidity, disability overweight and obesity, metabolic and cardiovascular diseases Maternal and child undernutrition Immediate Inadequate dietary intake Diseases causes Underlying Household food Inadequate care and Unhealthy household Inadequate health causes insecurity feeding practices environment services Household access to adequate quantity and quality of resources: Land, education, employment, income, and technology Basic Inadequate financial, human, physical, and social capital causes Social, cultural, economic, and political context Source: Skoufias, Vinha, and Sato 2019. Adapted from UNICEF 1990. http://creativecommons.org/licenses/by/3.0/igo Poor diets drive overweight and obesity, along were associated with reduced risk of overweight, with other NCDs (see figure 1.10) (Siekmans et obesity, diabetes, and hypertension. Conversely, al., 2024). Overweight and obesity further drive risk factors for overweight, obesity, diabetes, NCDs such as asthma, Alzheimer’s disease, cancer, and hypertension were greater consumption of cardiovascular disease, depression, diabetes, fried foods, SSBs, ultra-processed foods, and red hypertension, kidney disease, and osteoarthritis and processed meats (Siekmans et al., 2024). In (Siekmans et al., 2024; WHO 2024b). A scoping 2022, a Lancet study showed that one in eight literature review of the health and nutrition people are living with obesity (1 billion) (NCD-RisC costs of unhealthy diets in low- and middle- 2024), and every year more than 5 million lives income countries conducted in 2024 found are lost from overweight- and obesity-attributable that a healthy dietary pattern, vegetarian and diseases (GBD Collaborative Network 2021; Kontis vegan diets, higher consumption of fruits and et al. 2014). vegetables, and higher consumption of dairy Failing to Tackle Malnutrition 13 Figure 1.10  Ultra-Processed Foods and Adverse Health Outcomes Credibility: I Convincing II Highly suggestive III Suggestive IV Weak V No evidence Grade: Moderate Low Very low Respiratory health Mortality V Asthma All cause mortality III II II Wheezing Cancer related mortality V Cardiovascular disease IV I Cardiovascular health related mortality III III Cardiovascular disease events Heart disease related mortality V II combined (morbidity + mortality) III III Cardiovascular disease morbidity Cancer III Hypertension Breast cancer V V V Hypertriglyceridaemia Cancer overall III IV Low high density lipoprotein Central nervous system V cholesterol levels tumours Gastrointestinal health Chronic lymphocytic V IV Crohn’s disease leukaemia V Ulcerative colitis Colorectal cancer IV III Pancreatic cancer V Metabolic health Prostate cancer V V III III Abdominal obesity V Hyperglycaemia Mental health IV Metabolic syndrome Adverse sleep related outcomes II IV Non-alcoholic fatty liver disease Anxiety outcomes I III II Obesity Combined common mental I III III Overweight disorder outcomes IV IV Overweight + obesity Depression outcomes II I II Type 2 diabetes Source: Lane et al. 2024. http://creativecommons.org/licenses/by/4.0/ guidelines, also known as food-based dietary In response to the increasing burden of guidelines, that set recommendations for which diet-related diseases, the WHO and the FAO foods to limit and which to consume in greater have created dietary guidelines, also known quantities (FAO 2021; World Cancer Research Fund as food-based dietary guidelines, that set International 2015; WHO 1998, 2003, 2015). They recommendations for which foods to limit and recommended eliminating or significantly limiting which to consume in greater quantities. They the consumption of free sugar, sodium (salt), trans recommended eliminating or significantly fats, and alcohol (Nishida et al. 2004). In addition to all of the limitations, the WHO also recommends limiting the consumption of free sugar, sodium increasing the intake of healthy foods such as (salt), trans fats, and alcohol. In addition to the fruits, vegetables, legumes, seeds and nuts, and limitations, they recommend increasing intake whole grains. To meet these FBDGs and mitigate of healthy foods such as fruits, vegetables, obesity and its associated NCDs, the WHO legumes, seeds and nuts, and whole grains. recommends a suite of interventions that includes fiscal policies to promote healthy diets, nutrition Global recommendations by the WHO and the labeling policies, regulations on harmful marketing FAO encourage limiting the consumption of sugar, of food and beverages to children, breastfeeding sodium, fats, and alcohol (box 1.2) and increasing promotion and support, and school food and fruits, vegetables, legumes, seeds and nuts, and nutrition policies that include regulations on the whole grains (see Figure B1.2.2). In response to sale of products close to schools that are high in the increasing burden of diet-related diseases, fats, sugars, and salt (WHO 2024a). the WHO and the FAO have created dietary 14 Reshaping the Agrifood Sector for Healthier Diets Box 1.2 Guidelines on the Consumption of Sugar, Sodium, Fats, and Alcohol Free sugar has been strongly associated with Organization of the United Nations (FAO), increased weight gain, obesity, diabetes, UNICEF, the World Bank, the World Cancer hypertension, fatty liver disease, kidney Research Fund International, and the World dysfunction, cancers, and dental caries (figure Health Organization (WHO) all strongly advise B1.2.1) (Huang et al. 2023; WHO 2015; World against free sugar consumption (UN-Nutrition Cancer Research Fund International 2015); 2022; World Bank 2020; World Cancer research has shown that these effects are Research Fund International 2015). The WHO especially pronounced for free sugar in sugar- recommends consuming not more than 5 sweetened beverages (SSBs) (Singh et al. percent of total kilocalories per day from free 2015). As a result, the Food and Agriculture sugar (World Bank 2015). Figure B1.2.1  Pathways by Which Sugar Impacts Health 4 3 2 1 Energy in Displacement of Alteration of taste High glycemic load Sugar liquid form more satiating foods preferences Passive calorie Increased Primary Postprandial Increased intake of overconsumption hunger hyperinsulinemia hyerglycemia sugary foods when drinking to decreased intake of satisfy thirst vegetables and fruits Insulin Oxidative Lower intake of fiber resistance stress micronutrients, antioxidants, and Increased Fat other phytochemicals energy deposition intake Beta-cell Metabolic syndrome dysfunction (low HDLC, high triglyceride, hypertension, coagulopathy. chronic inflammation) Obesity and possibly Coronary stunting Diabetes Dental caries Heart disease Source: Popkin and Hawkes 2016. Increased sodium consumption is a risk factor 2018). The WHO recommends limiting the for hypertension, cardiovascular disease, consumption of sodium to less than 2 grams stroke, and other NCDs (Afshin et al. 2019; per day (5 grams per day of salt) for adults. Bibbins-Domingo et al. 2010; WHF 2021; WHO In tandem, the WHO advises processed food Failing to Tackle Malnutrition 15 companies to eliminate as much sodium Amico et al. 2021; WHO 2018; Willett and as possible from their recipes and consider Ascherio 1994). potassium salt as a substitute for sodium salt (World Bank 2022). Heavy alcohol consumption is a major cause of preventable death and injuries, violence, Trans fats should make up no more than 1 and abuse. More than 3 million deaths and percent of kilocalories per day and the WHO more than 120 disability-adjusted life years recommends eliminating them as much as (DALYs) lost are due to heavy drinking (Afshin possible to governments and policy makers et al. 2019; Cao et al. 2015; Kloner and Rezkalla (World Bank 2022). Trans fats are linked to a 2007; O’Keefe et al. 2014; Stampfer et al. 1988; large increased risk of death primarily through USDA 2015). Recommendations on how much coronary heart disease, with high consumption alcohol can be consumed safely vary widely increasing the risk of hospitalization and death by country—the WHO suggests consuming no by more than 50 percent (Afshin et al. 2019; more than 14 drinks per week spread out over three or more days (World Bank 2022). Map B1.3.1  Geographical Density of the Number of Dietary Surveys in the GDD 2017 by Country (A), including publicly available surveys (B), and non-public surveys submitted by data owners (C) GGD, Global Dietary Database Source: Miller et al. 2021. https://creativecommons.org/licenses/by/4.0/. Slight changes have been made to the original image. 16 Reshaping the Agrifood Sector for Healthier Diets References Abdulahi, A., S. Shab-Bidar, S. Rezaei, and K. Djafarian. 2017. “Nutritional Status of Under Five Children in Ethiopia: A Systematic Review and Meta-Analysis.” Ethiopian Journal of Health Sciences 27 (2): 175. https://doi.org/10.4314/ejhs​.v27i2.10. Afshin, A., P. J. Sur, K. A. Fay, L. Cornaby, G. Ferrara, J. S. Salama, E. C. Mullany, K. H. Abate, C. Abbafati, Z. Abebe, et al. 2019. “Health Effects of Dietary Risks in 195 countries, 1990–2017: A Systematic Analysis for the Global Burden of Disease Study 2017.” The Lancet 393 (10184): 1958–72. https://doi.org/10.1016/S0140-6736(19)30041-8. Amico, A., M. G. Wootan, M. F. Jacobson, C. Leung, and A. W. Willett. 2021. “The Demise of Artificial Trans Fat: A History of a Public Health Achievement.” The Milbank Quarterly 99 (3): 746–70. https://doi.org/10.1111/1468​-0009.12515. Baker, P. and S. Friel. 2014. “Processed Foods and the Nutrition Transition: Evidence from Asia.” Obesity Reviews 15 (7): 564–77. https://doi.org/10.1111/obr.12174. Baker, P., P. Machado, T. Santos, K. Sievert, K. Backholer, M. Hadjikakou, C. Russell, O. Huse, C. Bell, G. Scrinis, A. Worsley, S. Friel, and M. Lawrence. 2020. “Ultra‐Processed Foods and the Nutrition Transition: Global, Regional and National Trends, Food Systems Transformations and Political Economy Drivers.” Obesity Reviews 21 (12): e13126. https://doi.org/10.1111/obr.13126. Bibbins-Domingo, K., G. M. Chertow, P. G. Coxson, A. Moran, J. M. Lightwood, M. J. Pletcher, and L. Goldman. 2010. “Projected Effect of Dietary Salt Reductions on Future Cardiovascular Disease.” New England Journal of Medicine 362 (7): 590–99. https://doi.org/10.1056/NEJMoa0907355. Burgess, P., J. Harris, A. Graves, and L. Deeks. 2019. Regenerative Agriculture: Identifying the impact; enabling the potential. Report for SYSTEMIQ. Bedfordshire, UK: Cranfield University. https://farmpep.net/index.php​ /node/164. Cao, Y., W. C. Willett, E. B. Rimm, M. J. Stampfer, and E. L. Giovannucci. 2015. “Light to Moderate Intake of Alcohol, Drinking Patterns, and Risk of Cancer: Results from Two Prospective US Cohort Studies.” BMJ 351: h4238. https://doi​.org/10.1136/bmj.h4238. Chaparro, C. M. and P. S. Suchdev. 2019. “Anemia Epidemiology, Pathophysiology, and Etiology in Low- and Middle-Income Countries.” Annals of the New York Academy of Sciences 1450 (1): 15–31. https://doi.org/10.1111/nyas.14092. Christian, A. K., E. Afful-Dadzie, and G. S. Marquis. 2023. “Infant and Young Child Feeding Practices Are Associated with Childhood Anaemia and Stunting in Sub-Saharan Africa.” BMC Nutrition 9 (1): 1–13. https://doi.org/10.1186​/S40795-022-00667-9. de Onis, M. and F. Branca. 2016. “Childhood Stunting: A Global Perspective.” Maternal & Child Nutrition 12 (Suppl. 1): 12–26. https://doi.org/10.1111​/mcn.12231. Failing to Tackle Malnutrition 17 FAO (Food and Agriculture Organization of the United Nations). 2021. Food-Based Dietary Guidelines. https://www.fao.org/nutrition/education/food-based​-dietary-guidelines. FAO, IFAD, UNICEF, WFP, and WHO (Food and Agriculture Organization of the United Nations, International Fund for Agricultural Development, UNICEF, World Food Programme, and World Health Organization). 2023. The State of Food Security and Nutrition in the World 2023: Urbanization, Agrifood Systems Transformation and Healthy Diets across the Rural–Urban Continuum. Rome: FAO, IFAD, UNICEF, WFP, and WHO. FOLU (Food and Land Use Coalition). 2019a. Growing Better: Ten Critical Transitions to Transform Food and Land Use. The Global Consultation Report of the Food and Land Use Coalition (FOLU) September 2019. https://www​.foodandlandusecoalition.org/global-report/. FOLU (Food and Land Use Coalition). 2019b. SystemIQ. Technical Annex. https://www.foodandlandusecoalition.org/wp-content/uploads/2019/09​ /FOLU-GrowingBetter-TechnicalAnnex.pdf. Galasso, E. and A. Wagstaff. 2019. “The Aggregate Income Losses from Childhood Stunting and the Returns to a Nutrition Intervention Aimed at Reducing Stunting.” Economics & Human Biology 34: 225–38. https://doi.org/10.1016/j​.ehb.2019.01.010. GBD 2016 Risk Factors Collaborators. 2017. “Global, Regional, and National Comparative Risk Assessment of 84 Behavioural, Environmental and Occupational, and Metabolic Risks or Clusters of Risks, 1990–2016: A Systematic Analysis for the Global Burden of Disease Study 2016.” The Lancet 390 (10100): 1345–1422. https://doi.org/10.1016/S0140​-6736(17)32366-8. GBD 2019 Risk Factors Collaborators. 2020. “Global Burden of 87 Risk Factors in 204 Countries and Territories, 1990–2019: A Systematic Analysis for the Global Burden of Disease Study 2019.” The Lancet 396 (10258): 1223–49. https://doi.org/10.1016/S0140-6736(20)30752-2. GBD 2021 Anaemia Collaborators. 2023. “Prevalence, Years Lived with Disability, and Trends in Anaemia Burden by Severity and Cause, 1990–2021: Findings from the Global Burden of Disease Study 2021.” The Lancet Haematology 10 (9): E713–E734. https://doi.org/10.1016/S2352-3026(23)00160-6. GBD Collaborative Network (Global Burden of Disease Collaborative Network). 2021. Global Burden of Disease Study 2019 (GBD 2019) Results. Institute for Health Metrics and Evaluation. https://vizhub.healthdata.org/gbd-results/. Gill, J., G. Elabed, S. Guerrero Escobar, V. Pernechele, J. J. E. Yerovi, F. P Fontes, and A. P. Mas Aparisi. 2023. “Repurposing Agricultural Support Policies for Sustainable Food Systems : Toolkit.” https://documents.worldbank.org/en​ /publication/documents-reports/documentdetail/099121823174517138​ /p1736580e6f98e05088070e95fabf1175c. 18 Reshaping the Agrifood Sector for Healthier Diets Haas, J. D. and T. Brownlie IV. 2001. “Iron Deficiency and Reduced Work Capacity: A Critical Review of the Research to Determine a Causal Relationship.” Journal of Nutrition 131 (2 Suppl. 2): 676S–688S. https://doi.org/10.1093​/jn/131.2.676s. Haider, B. A., I. Olofin, M. Wang, D. Spiegelman, M. Ezzati, and W. W. Fawzi. 2013. “Anaemia, Prenatal Iron Use, and Risk of Adverse Pregnancy Outcomes: Systematic Review and Meta-Analysis. BMJ 347 (7916). https://doi.org​/10.1136/bmj.f3443. Headey, D. and M. Ruel. 2023. “Food Inflation and Child Undernutrition in Low and Middle Income Countries.” Nature Communications 14 (1). https://doi.org/10.1038/s41467-023-41543-9. Hess, S. Y., A. C. McLain, H. Lescinsky, K. H. Brown, A. Afshin, R. Atkin, and S. J. M. Osendarp. 2022. “Basis for Changes in the Disease Burden Estimates Related to Vitamin A and Zinc Deficiencies in the 2017 and 2019 Global Burden of Disease Studies.” Public Health Nutrition 25 (8). https://doi.org/10.1017/S1368980021004821. Hoddinott, J., H. Alderman, J. R. Behrman, L. Haddad, and S. Horton. 2013. “The Economic Rationale for Investing in Stunting Reduction.” Maternal and Child Nutrition, 9 (S2): 69–82. https://doi.org/10.1111/mcn.12080. Huang Y., Z. Chen, B. Chen, J. Li, X. Yuan, J. Li, et al. 2023. “Dietary Sugar Consumption and Health: Umbrella Review.” BMJ 381: e071609. https://doi.org/10.1136/bmj-2022-071609. Huxley, R. R., A. Shiell, and C. M. Law. 2000. “The Role of Size at Birth and Postnatal Catch-Up Growth in Determining Systolic Blood Pressure: A Systematic Review of the Literature.” Journal of Hypertension 18 (7): 815–31. https://doi.org/10.1097/00004872-200018070-00002. International Food Policy Research Institute. 2024. 2024 Global Food Policy Report: Food Systems for Healthy Diets and Nutrition. Washington, DC: International Food Policy Research Institute. https://hdl.handle.net/10568/141760 IHME (Institute for Health Metrics and Evaluation). 2019. Global Burden of Disease Database. https://www.healthdata.org/research-analysis/gbd. Kloner, R. A. and S. H. Rezkalla. 2007. “To Drink or Not to Drink? That Is the Question.” Circulation 116 (11): 1306– 17. https://doi.org/10.1161​/CIRCULATIONAHA.106.678375. Kontis, V., C. D. Mathers, J. Rehm, G. A. Stevens, K. D. Shield, R. Bonita, L. M. Riley, V. Poznyak, R. Beaglehole, and M. Ezzati. 2014. “Contribution of Six Risk Factors to Achieving the 25×25 Non-Communicable Disease Mortality Reduction Target: A Modelling Study. The Lancet 384 (9941): 427–37. https://doi​.org/10.1016/S0140-6736(14)60616-4. Failing to Tackle Malnutrition 19 Kozuki, N., J. Katz, A. C. C. Lee, J. P. Vogel, M. F. Silveira, A. Sania, G. A. Stevens, S. Cousens, L. E. Caulfield, P. Christian, et al. 2015. “Short Maternal Stature Increases Risk of Small-for-Gestational-Age and Preterm Births in Low- and Middle-Income Countries: Individual Participant Data Meta-Analysis and Population Attributable Fraction. The Journal of Nutrition 145 (11): 2542–50. https://doi.org/10.3945/JN.115.216374. Lane, M. M., E. Gamag, S. Du, D. N. Ashtree, A. J. McGuinness, S. Gauci, et al. 2024. “Ultra-Processed Food Exposure and Adverse Health Outcomes: Umbrella Review of Epidemiological Meta-Analyses” BMJ 384: e077310. https://doi​.org/10.1136/bmj-2023-077310. Lara-Castor, L., R. Micha, F. Cudhea, V. Miller, P. Shi, J. Zhang, J. R. Sharib, J. Erndt-Marino, S. B. Cash, D. Mozaffarian, M. Bas, J. H. Ali, S. Abumweis, A. Krishnan, P. Misra, N. C. Hwalla, C. Janakiram, N. I. Liputo, A. Musaiger, et al. 2023. “Sugar-Sweetened Beverage Intakes among Adults between 1990 and 2018 in 185 Countries.” Nature Communications 14 (1): 5957. https://doi.org/10.1038/s41467-023-41269-8. Lobstein, T., R. Jackson-Leach, J. Powis, H. Brinsden, and M. Gray. 2023. World Obesity Atlas 2023. London: World Obesity Federation. https://s3-eu-west-1​ .amazonaws.com/wof-files/World_Obesity_Atlas_2023_Report.pdf. Martorell, R., B. L. Horta, L. S. Adair, A. D. Stein, L. Richter, C. H. D. Fall, S. K. Bhargava, S. K. D. Biswas, L. Perez, F. C. Barros, and C. G. Victora. 2010. “Weight Gain in the First Two Years of life Is an Important Predictor of Schooling Outcomes in Pooled Analyses from Five Birth Cohorts from Low- and Middle-Income Countries.” Journal of Nutrition 140 (2): 348–54. https://doi.org/10.3945​/jn.109.112300. Micha, R., S. Khatibzadeh, P. Shi, K. G. Andrews, R. E. Engell, and D. Mozaffarian. 2015. “Global, Regional and National Consumption of Major Food Groups in 1990 and 2010: A Systematic Analysis Including 266 Country- Specific Nutrition Surveys Worldwide.” BMJ Open 5 (9): e008705. https://doi.org​/10.1136/bmjopen-2015-008705. Miller, V., G. M. Singh, J. Onopa, J. Reedy, P. Shi, J. Zhang, A. Tahira, M. L. Shulkin Morris, D. P. Marsden, S. Kranz, S. Stoyell, P. Webb, R. Micha, and D. Mozaffarian. 2021. “Global Dietary Database 2017: Data Availability and Gaps on 54 Major Foods, Beverages and Nutrients among 5.6 Million Children and Adults from 1220 Surveys Worldwide.” BMJ Global Health 6 (2): e003585. https://doi.org/10.1136/bmjgh-2020-003585. Miller, V., P. Webb, F. Cudhea, P. Shi, J. Zhang, J. Reedy, J. Erndt-Marino, J. Coates, D. Mozaffarian, M. Bas, J. H. Ali, S. Abumweis, A. Krishnan, P. Misra, N. C. Hwalla, C. Janakiram, N. I. Liputo, A. Musaiger, F. Pourfarzi, et al. 2022. “Global Dietary Quality in 185 Countries from 1990 to 2018 Show Wide Differences by Nation, Age, Education, and Urbanicity.” Nature Food 3 (9): 694–702. https://doi.org/10.1038/s43016-022-00594-9. 20 Reshaping the Agrifood Sector for Healthier Diets Monteiro, C. A., G. Cannon, J.-C. Moubarac, R. B. Levy, M. L. C. Louzada, and P. C. Jaime. 2018. “The UN Decade of Nutrition, the NOVA Food Classification and the Trouble with Ultra-Processing.” Public Health Nutrition 21 (1): 5–17. https://doi.org/10.1017/S1368980017000234. Monteiro, C. A., J.-C. Moubarac, G. Cannon, S. W. Ng, and B. Popkin. 2013. “Ultra‐Processed Products Are Becoming Dominant in the Global Food System.” Obesity Reviews 14 (S2): 21–28. https://doi.org/10.1111/obr.12107. NCD-RisC (NCD Risk Factor Collaboration). 2023. NCD Risk Factor Collaboration Home Page. https://ncdrisc.org/index.html. NCD-RisC (NCD Risk Factor Collaboration). 2024. “Worldwide Trends in Underweight and Obesity from 1990 to 2022: A Pooled Analysis of 3663 Population-Representative Studies with 222 Million Children, Adolescents, and Adults.” The Lancet 403 (10431): 1027–50. https://doi.org/10.1016/S0140​-6736(23)02750-2. Nishida, C., R. Uauy, S. Kumanyika, and P. Shetty. 2004. “The Joint WHO/FAO Expert Consultation on Diet, Nutrition and the Prevention of Chronic Diseases: Process, Product and Policy Implications.” Public Health Nutrition 7 (1a): 245–50. https://doi.org/10.1079/PHN2003592. O’Keefe, J. H., S. K. Bhatti, A. Bajwa, J. J. DiNicolantonio, and C. J. Lavie. 2014. “Alcohol and Cardiovascular Health: The Dose Makes the Poison…or the Remedy.” Mayo Clinic Proceedings 89 (3): 382–93. https://doi.org/10.1016/j​.mayocp.2013.11.005. Okunogbe, A., R. Nugent, G. Spencer, J. Powis, J. Ralston, and J. Wilding. 2022. “Economic Impacts of Overweight and Obesity: Current and Future Estimates for 161 Countries.” BMJ Global Health 7 (9): e009773. https://doi.org/10.1136​/BMJGH-2022-009773. Özaltin, E., K. Hill, and S. V. Subramanian. 2010. “Association of Maternal Stature with Offspring Mortality, Underweight, and Stunting in Low- to Middle-Income Countries.” JAMA 303 (15): 1507–16. https://doi.org/10.1001/jama.2010.450. Popkin, B. M. and C. Hawkes. 2016. “Sweetening of the Global Diet, Particularly Beverages: Patterns, Trends, and Policy Responses.” The Lancet Diabetes & Endocrinology 4 (2): 174–86. https://doi.org/10.1016/S2213-8587(15)00419-2. Popkin, B. M. and T. Reardon. 2018. “Obesity and the Food System Transformation in Latin America.” Obesity Reviews 19 (8): 1028–64. https://doi.org/10.1111​/obr.12694. Reardon, T., D. Tschirley, L. S. O. Liverpool-Tasie, T. Awokuse, J. Fanzo, B. Minten, R. Vos, M. Dolislager, C. Sauer, R. Dhar, C. Vargas, A. Lartey, A. Raza, and B. M. Popkin. 2021. “The Processed Food Revolution in African Food Systems and the Double Burden of Malnutrition.” Global Food Security 28: 100466. https://doi.org/10.1016/j.gfs.2020.100466. Shekar, M. and B. Popkin, eds. 2020. Obesity: Health and Economic Consequences of an Impending Global Challenge. Human Development Perspectives Series. Washington, DC: World Bank. https://doi.org/10.1596/978-1-4648-1491-4. Failing to Tackle Malnutrition 21 Siekmans, K., P. Fracass, T. Kato, T. K. Seow, D. Carter, S. Horton, F. Dizon, and K. S. Okamura. 2024. Understanding the Links between Diet Quality, Malnutrition, and Economic Costs: An Evidence Review for LMICs. Policy Research Working Paper 10747, PEOPLE. Washington, DC: World Bank Group. Singh, G. M., R. Micha, S. Khatibzadeh, S. Lim, M. Ezzati, and D. Mozaffarian. 2015. “Estimated Global, Regional, and National Disease Burdens Related to Sugar- Sweetened Beverage Consumption in 2010.” Circulation 132 (8): 639–66. https://doi.org/10.1161/CIRCULATIONAHA.114.010636. Skoufias, E., K. Vinha, and R. Sato. 2019. All Hands on Deck: Reducing Stunting through Multisectoral Efforts in Sub-Saharan Africa. Africa Development Forum. Washington, DC: World Bank and Agence française de développement. https://doi.org/10.1596/978-1-4648-1396-2. Stampfer, M. J., G. A. Colditz, W. C. Willett, F. E. Speizer, and C. H. Hennekens. 1988. “A Prospective Study of Moderate Alcohol Consumption and the Risk of Coronary Disease and Stroke in Women.” New England Journal of Medicine 319 (5): 267–73. https://doi.org/10.1056/NEJM198808043190503. Steyn, N. P. and J. H. Nel. 2022. “Prevalence and Determinants of the Double Burden of Malnutrition with a Focus on Concurrent Stunting and Overweight/ Obesity in Children and Adolescents.” Current Nutrition Reports 11: 437–56. https://doi.org/10.1007/s13668-022-00417-0. Thurstans, S., N. Sessions, C. Dolan, K. Sadler, B. Cichon, S. Isanaka, D. Roberfroid, H. Stobaugh, P. Webb, and T. Khara. 2022. “The Relationship between Wasting and Stunting in Young Children: A Systematic Review.” Maternal & Child Nutrition 18 (1): e13246. https://doi.org/10.1111/mcn.13246. UNICEF (United Nations Children’s Fund). 1990. Strategy for Improved Nutrition of Women and Children in Developing Countries: A UNICEF Policy Review. New York: UNICEF. UNICEF. 2021. Nutrition and Care for Children with Wasting: Treating Children with the Most Life-Threatening Form of Malnutrition. https://www.unicef.org​/nutrition/child-wasting. UNICEF. 2023. Undernourished and Overlooked: A Global Nutrition Crisis in Adolescent Girls and Women. https://doi.org/10.18356/9789213626764. UNICEF/WHO/World Bank Group. 2023. Levels and Trends in Child Malnutrition. Joint Child Malnutrition Estimates of 2023. https://www.who.int/teams/nutrition​ -and-food-safety/monitoring-nutritional-status-and-food-safety-and-events​ /joint-child-malnutrition-estimates. UN-Nutrition. 2022. Transforming Nutrition. UN-Nutrition Journal Volume 1. Rome: FAO. https://doi.org/10.4060/cc2805en. USDA (US Department of Agriculture). 2015. Scientific Report of the 2015 Dietary Guidelines Advisory Committee: Advisory Report to the Secretary of Health and Human Services and the Secretary of Agriculture. US Department of Agriculture, Agricultural Research Service, Washington, DC. 22 Reshaping the Agrifood Sector for Healthier Diets Victora, C. G., P. Christian, L. P. Vidaletti, G. Gatica-Domínguez, P. Menon, and R. E. Black. 2021. “Revisiting Maternal and Child Undernutrition in Low-Income and Middle-Income Countries: Variable Progress towards an Unfinished Agenda.” The Lancet 397 (10282): 1388–99. https://doi.org​/10.1016/S0140-6736(21)00394-9. Walker, S. P., T. D. Wachs, J. Meeks Gardner, B. Lozoff, G. A. Wasserman, E. Pollittere, and J. A. Carter. 2007. “Child Development: Risk Factors for Adverse Outcomes in Developing Countries.” The Lancet 369 (9556): 145–57. https://doi.org/10.1016/S0140-6736(07)60076-2. WHF (World Heart Federation). 2021. “WHF Partners with LoSalt® in Bid to Tackle Hypertension,” May 18, 2021. WHF partners with LoSalt® in bid to tackle hypertension 2021. World Heart Federation. https://world-heart-federation​ .org/news/whf-partners-with-losalt-in-bid-to-tackle-hypertension/. Whincup, P. H., S. J. Kaye, C. G. Owen, R. Huxley, D. G. Cook, S. Anazawa, E. Barrett-Connor, S. K. Bhargava, B. E. Birgisdottir, S. Carlsson, S, R, De Rooij, R. F. Dyck, J. G. Eriksson, B. Falkner, et al. 2008. “Birth Weight and Risk of Type 2 Diabetes: A Systematic Review.” JAMA 300 (24): 2886–97. https://doi​.org/10.1001/jama.2008.886. WHO (World Health Organization). 1998. Preparation and Use of Food-Based Dietary Guidelines: Report of a Joint FAO/WHO Consultation. WHO Technical Report Series 880. Geneva: WHO. https://iris.who.int/bitstream/handle​ /10665/42051/WHO_TRS_880.pdf?sequence=1 WHO (World Health Organization). 2003. Report of the Joint WHO/FAO Expert Consultation on Diet, Nutrition and the Prevention of Chronic Diseases: Report of a Joint WHO/FAO Expert Consultation. WHO Technical Report Series 916. Geneva: WHO. https://iris.who.int/bitstream/handle/10665/42665/WHO​ _TRS_916.pdf?sequence=1. WHO (World Health Organization). 2015. Guideline: Sugars Intake for Adults and Children. Geneva: WHO. https://iris.who.int/bitstream/handle/10665/149782​ /9789241549028_eng.pdf?sequence=1. WHO (World Health Organization). 2017. Report of the Commission on Ending Childhood Obesity. Implementation Plan: Executive Summary, 105(9). WHO (World Health Organization). 2018. Replace Trans Fat: Policies to Eliminate Industrially-Produced Trans Fat Consumption. Information Sheet. https://www.who.int/docs/default-source/documents/replace-transfats​ /replace-act-information-sheet.pdf. WHO (World Health Organization). 2023a. Anaemia. Fact Sheet, May 31, 2023. https://www.who.int/news-room/fact-sheets/detail/anaemia. WHO (World Health Organization). 2023b. World Health Statistics 2023: A Visual Summary. https://www.who.int/data/stories/world-health-statistics-2023-a​ -visual-summary/. Failing to Tackle Malnutrition 23 WHO (World Health Organization). 2023c. World Health Statistics 2023: Monitoring Health for the SDGs, Sustainable Development Goals. Geneva: World Health Organization. https://www.who.int/publications/i/item/9789240074323. WHO (World Health Organization). 2024a. Obesity and Overweight. Fact Sheet, March 1, 2024. World Health Organization. https://www.who.int/news-room​/fact-sheets/detail/obesity-and-overweight. WHO (World Health Organization). 2024b. “One in Eight People Are Now Living with Obesity.” Press Release, March 1, 2024. https://www.who.int/news/item/01​ -03-2024-one-in-eight-people-are-now-living-with-obesity. WHO GHO (World Health Organization Global Health Observatory). 2024a. Anaemia in Women and Children. https://www.who.int/data/gho/data/themes/topics​ /anaemia_in_women_and_children. WHO GHO (World Health Organization Global Health Observatory). 2024b. GHO Indicators: Prevalence of Obesity among Adults. https://www.who.int/data​ /gho/data/indicators/indicator-details/GHO/prevalence-of-obesity-among​ -adults-bmi--30-(age-standardized-estimate)-(-). Willett, W. C., M. J. Stampfer, J. E. Manson, et al. 1993. “Intake of Trans Fatty Acids and Risk of Coronary Heart Disease among Women.” The Lancet 341 (8845): 581–85. https://doi.org/10.1016/0140-6736(93)90350-P. Willett, W. C. and A. Ascherio. 1994. “Trans Fatty Acids: Are the Effects Only Marginal?” American Journal of Public Health 84 (5): 722–24. https://doi​.org/10.2105/AJPH.84.5.722. World Bank. 2018. The Human Capital Project. https://www.worldbank.org/en​/publication/human-capital. World Bank. 2020. Taxes on Sugar-Sweetened Beverages: Summary of International Evidence and Experiences. Washington, DC: World Bank. https://thedocs​ .worldbank.org/en/doc/d9612c480991c5408edca33d54e2028a-0390062021​ /original/World-Bank-2020-SSB-Taxes-Evidence-and-Experiences.pdf. World Bank. 2022. Repurposing agrifood policies which support unhealthy food ingredients. Washington, DC: World Bank. World Bank and FAO. Forthcoming. Uncovering the Economic Costs of Unhealthy Diets: A new methodology and country case studies in Ethiopia and the Philippines. Washington DC: World Bank. World Cancer Research Fund International. 2015. Curbing Global Sugar Consumption: Effective Food Policy Actions to Help Promote Healthy Diets and Tackle Obesity. https://www.wcrf.org/policy/our-publications/curbing​ -global-sugar-consumption World Obesity Federation. 2023. World Obesity Atlas 2023. London: World Obesity Federation. https://www.worldobesity.org/resources/resource-library/world​ -obesity-atlas-2023. 24 Reshaping the Agrifood Sector for Healthier Diets Zeinalabedini, M., B. Zamani, E. Nasli-Esfahani, and L. Azadbakht. 2023. “A Systematic Review and Meta-Analysis of the Association of Dietary Diversity with Undernutrition in School-Aged Children.” BMC Pediatrics 23 (1): 269. https://doi.org/10.1186/s12887-023-04032-y. Zimmermann, M. B. and M. Andersson. 2021. “Global Perspectives in Endocrinology: Coverage of Iodized Salt Programs and Iodine Status in 2020.” European Journal of Endocrinology 185 (1): R13–R21. https://doi.org/10.1530/EJE-21-0171. Failing to Tackle Malnutrition 25 26 Reshaping the Agrifood Sector for Healthier Diets 2 An Opportunity to Leverage Agrifood Public Support make up well over 75 percent of the total support 2.1. The problem with current (OECD 2023a). Furthermore, agrifood support agrifood support measures in low- and middle-income countries often lead to the most distorting outcomes, since they tend Public support to the agrifood sector is to prioritize coupled subsidies—unlike in high-income large, distortionary, and inefficient. Globally, countries, where subsidies are commonly uncoupled support to the agrifood sector comes to and have a less distortive impact. In addition, agrifood subsidies in high-income countries focus more on at least $854 billion per year. Most of this research and infrastructure development than those support is focused on producers, not on in low- and middle-income countries, which leads to consumers, and comes largely in the form of more market harm in the latter (Damania et al. 2023). trade and market policies that distort prices. Public support to the sector is also often Agrifood support is also producer-focused and regressive, benefiting wealthier farmers. regressive. Between 2020 and 2022, agrifood consumers and GSS together received $221 billion, Agrifood support is sizeable, but it is also largely while producers received $633 billion on average distortionary and inefficient, particularly in low- per year. Aggregate support to producers in all 54 and middle-income countries. Total support to the countries represented 9.8 percent of gross farm agrifood sector averaged $854 billion from 2020 to receipts (GFR) on average in 2020–22. This is 13.7 2022 (see figure 2.1) in the 54 countries combined percent of GFR in positive support and 3.9 percent in that include Organisation of Economic Co-operation negative Market Price Support (MPS) (OECD 2023a). and Development (OECD) members and other major Large producer support, however, does not aid all agrifood producers.1 This number is an increase from recipients equally. The OECD notes that producer the total support to agrifood of $714 billion annually support reforms have stagnated over the last 10 for the years between 2014 and 2016 (OECD 2023a). years. Agrifood subsidies that governments allocate This considerable growth in the sectoral support to their agriculture producers often regressively to agrifood, despite increasing global agrifood benefit wealthier farmers, since they, compared production, is also highly distortive to the markets. to poorer farmers, “use more inputs and produce Out of the total support to agrifood, 74 percent was more outputs,” and the subsidies rarely have a targeted toward producers; most of these measures positive effect on the efficiency of the agricultural are market distortive. Over 50 percent of producer production. For example, subsidies are leading to support is in the form of trade/market policies, which excessive use of fertilizers, particularly in East Asia impact the market prices of agrifood commodities. and the Pacific and in South Asia (see box 2.1). Even Smaller shares of support are allocated to General targeting subsidies to the poorer farmers often fails Services Support (GSS, 12 percent). These are to address the regressivity of a policy and, ultimately, investments in private or public services—such as those with stronger political or market connections institutions and infrastructure. Smaller shares are end up having better access to subsidies (Damania also allocated to consumer subsidies (13 percent)— et al. 2023). The issue, however, is not only how which can improve the consumption of nutritious governments support their agrifood sectors, but also food if targeted effectively. Agrifood public support what commodities these funds are aimed at. Chapter is concentrated in a few jurisdictions: China, India, 4 also discusses the fertilizer subsidies in more detail the United States, and the European Union together in the cases of Bangladesh and Malawi. An Opportunity to Leverage Agrifood Public Support 27 Figure 2.1  Public Support to Agrifood Sector, by Type of Support a. Composition of support to agriculture, 2014–16 Total support to agriculture for OECD+ countries adds up to USD 714 billion per year Distorting support, US$302 billion General Services Support Estimate 42% (GSSE), US$108 billion 15% Producer Support Estimate (PSE), 75% US$534 billion 10% 32% Transfers to Consumers (TCT), US$73 billion Other producer support, US$231 billion b. Composition of support to agriculture, 2020–22 Total support to agriculture for OECD+ countries adds up to USD 854 billion per year Distorting support, General Services US$334 billion Support Estimate (GSSE), US$106 billion 39% 12% Producer Support Estimate (PSE), 74% US$633 billion 13% Transfers to Consumers (TCT), 35% US$115 billion Other producer support, Source: Original figure for this publication based on data from OECD 2023b. US$299 billion Note: Total support to agriculture indicates positive transfers to the agricultural sector, computed as the sum of transfers to consumers (that is, Transfers to Consumers from Taxpayers) (TCT), General Services Support Estimates (GSSE), and Producer Support Estimates (PSE). To consider implicit taxation of producers in some emerging economies, PSE is combined with transfers away from producers (negative Market Price Support, or negative MPS). Distorting support includes MPS, transfers away from producers (negative MPS), payments based on output, and payments based on unconstrained variable input use. This covers 54 OECD+ countries, which refers to the 38 OECD countries, the 5 non-OECD EU Member States, and 11 emerging economies. 28 Reshaping the Agrifood Sector for Healthier Diets Box 2.1 A Closer Look at Fertilizer Use According to a 2023 World Bank report, in Strikingly, around 50 percent of global caloric East Asia and Pacific, excessive fertilizer use production occurs in areas that also overuse is contributing to a decline in agricultural fertilizer, making the reduction of fertilizer productivity, with South Asia expected to face use there an opportunity that is unlikely to the same challenge in the near future. In these negatively impact agriculture yields. The two regions, nitrogen fertilizer use, promoted overconsumption of fertilizers is also leading by agriculture subsidies, is reaching or has to the deterioration of water quality, which is reached the diminishing returns to net primary yet another reason to bring the use down to productivity (NPP) (Damania et al. 2023; see optimal levels (Damania et al. 2023). Since the figure B2.1.1). On the other hand, countries price of fertilizers has risen in recent years, in Sub-Saharan Africa and some in Europe reducing its use in overconsumed areas might and Central Asia exhibit a lack of fertilizer positively support the increase in fertilizer use application, which is preventing the growth in underconsuming geographies as a result of in agriculture productivity, measured by NPP. the lower global price of those inputs. Figure B2.1.1  Change in Productivity Due to Use of Nitrogen Fertilizer % change in NPP relative to Q1:[0–3) 30 Middle East and East Asia Sub-Saharan Africa North Africa South Asia and Pacific 25 Europe and Global and Latin America Central Asia and the Caribbean 20 15 10 5 0 Q2:[3–6) Q3:[6–16) Q4:[16–32) Q5:[32–51) Q6:[51–66) Q7:[66–82) Q8:[82–154) Q9:> 154 Quantile of fertilizer use (kilograms per hectare) Source: Damania et al. 2023. Note: NPP = net primary productivity. An Opportunity to Leverage Agrifood Public Support 29 Public support to the agrifood sector is classified as high priority. Products whose actual focused much more on food commodities level of consumption exceeded 120 percent of that are already high in consumption such the recommended level of consumption, were characterized as low priority. And products with as grains and meats, and much less on a per capita consumption of 80 and 120 percent underconsumed, healthier food commodities of the recommended level, were characterized such as fruits, vegetables, and dairy. as medium priority” (Glauber and Laborde 2023, p. 17). Fruits and vegetables are among the least Many regions fall short of the recommended supported agricultural commodities and have the consumption of fruits, vegetables, and dairy; highest priority. In 95 percent of regions, fruits very few regions fall short in rice, wheat, and and vegetables are identified as either high- or maize. Glauber and Laborde (2023) have classified medium-priority products, pointing to their agriculture commodities based on their level of overwhelming underconsumption (Glauber and per capita consumption and the recommended Laborde 2023). On the other hand, wheat, maize, dietary guidance for that geography, which derives and rice are generally overconsumed, since in from the Food and Agriculture Organization of fewer than 10 percent of regions globally these the United Nations (FAO)’s food-based dietary commodities are identified as high priority. It guidelines (see figure 2.2). “Products whose is worth noting that sugar is not present in this average actual consumption level was less than exercise. 80 percent of the recommended level, were Figure 2.2  Food Groups Based on Per Capita Consumption Relative to Dietary Guidelines Fruits and vegetables Dairy products Raw milk Fishery products Vegetable oils Poultry and pork (cut) Poultry and pork (raw) Cattle meat, cuts Cattle meat, raw Oilseeds Rice (paddy) Rice (processed) Wheat Maize and other grains 0 10 20 30 40 50 60 70 80 90 100 Percent of Regions High priority Medium priority Low priority Source: Glauber and Laborde 2023. Agrifood support is imbalanced—it is higher Between 2020 and 2022, in terms of the dollar for agrifood commodities that are already value, the most-supported commodity among sufficiently consumed, while lower for much the 54 countries studied by the OECD was maize, healthier, underconsumed foods (see figure with over $56 billion in total support. The second 2.3). Commodity-specific support to the agrifood most supported commodity in dollar terms was sector is concentrated among a few products. rice, with $32.2 billion; third was pork, with $31.5 30 Reshaping the Agrifood Sector for Healthier Diets billion; fourth and fifth were poultry meat as well 2023a). Alternatively, vegetables, dairy, and fruit as beef and veal, which received $26.6 billion are located to the far right of figure 2.3 because and $24.4 billion in agrifood support. Figure 2.3 of their very low (or negative) levels of producer also ranks the commodities, based on another support. As shown in chapter 1, globally, fruit support measurement, which is the Producer and vegetable consumption fall short of World Single Commodity Transfers (PSCT) as a share Health Organization (WHO) or food-based dietary of the total Gross Farm Receipts, meaning how guideline (FBDG) recommendations, while meat much income a single-commodity farmer gets consumption exceeds International Agency for from PSCT as a share of their total income. The Research on Cancer recommendations and sugar- most-supported commodity as a share of GFR was sweetened beverage (SSB) consumption exceeds sugar, with around 24 percent of farmers’ incomes WHO recommendations. There is thus a disconnect from sugar deriving from public support. The between which agrifood commodities receive second most-supported was maize with around public support and which ones are healthier, 20 percent, third was rice with approximately 15 indicating a need to rethink agrifood support in a percent, and fourth was poultry meat with 11 way to ensure that it is better guided by health and percent of GFR coming from public support (OECD nutrition outcomes. Figure 2.3  Agrifood Support by Food Commodity Size of bubble indicates amount of Producer Single Commodity Transfers Total Producer Support (as a share of Gross Farm (in US$, billions), total support, Receipts for Commodity) (%), average of 2020–22 average of 2020–22. 30 50 25 Poultry 11.0% (US$26.6 20 billion) 15 20 Other meats 8.7% (US$6.9 billion) 10 Sugar 5 23.7% 10 (US$15.5 billion) Other grains –0.3% Maize 19.2% Rice 15.4% –(US$0.2 billion) (US$56.7 billion) (US$ 32.2 billion) Beef 0 8.8% Pork 8.5% Soybeans Vegetables (US$24.4 (US$31.5 3.4% 2.5% Wheat billion) billion) (US$5.6 (US$0.9 –1.5% –10 billion) billion) –(US$2.9 billion) Milk –9.7% –(US$32.1 billion) –20 Fruits –21.3% –(US$17.6 billion) –30 Agriculture commodities Source: Original figure for this publication based on data from OECD 2023b. Note: The OECD indicator database includes a total of 60 food groups. A full list of commodities can be found at https://www.oecd.org​ /agriculture/topics/agricultural-policy-monitoring-and-evaluation/documents/producer-support-estimates-manual.pdf. The full list of the OECD database includes 54 countries, including the 38 OECD countries, five non-OECD EU Member States, and 11 emerging economies. The measure of total public support, percentage of Producer Single Commodity Transfers (%PSCT) of the gross receipts for the single commodity (GR), is defined as the value of gross transfers from consumers and taxpayers to agricultural producers, measured at the farm gate level, arising from policies linked to the production of a single commodity. Other meats include sheep meat. Other grains include alfalfa, oats, sorghum, and barley. An Opportunity to Leverage Agrifood Public Support 31 The nature of agrifood public support varies percent of the total agriculture support to inputs, across regions and across countries. In Sub- 4 percent was given to farms directly, 3 percent in Saharan Africa, support to the agrifood sector a form of coupled support; the rest was directed at producers in other ways. Low levels of agriculture is generally low. Agriculture subsidies in the support in Africa and the importance of the sector region amounted to only $680 million per for the livelihoods of people on the continent have year, with public spending in the sector at already pushed members of the African Union, 6 percent of national budgetary resources, which includes all countries in Africa, to commit to well below the 10 percent target from the increase their spending on rural development and Maputo Declaration. Thus, there is a need agriculture to 10 percent of budgets back in 2003 to simultaneously increase agrifood public as part of the Maputo Declaration (GLOPAN 2022). Further complicating the issue are differences both support and rebalance the support away from in diets across regions and between recommended a heavy focus on input subsidies and toward and current intake of various food groups more research and development. (box 2.2). Not all regions are the same. In Africa, support The current set of agrifood policies can be to the agrifood sector is low. Globally, agriculture disconnected from public policies in other receives significant public support, which (as noted earlier) is often regressive, distortionary, leads sectors, such as those in the health sector. For to inefficiencies, and supports low-priority foods. example, some countries tax the consumption These trends, however, are not universal. The of sugar-sweetened beverages while providing African continent is spending considerably less significant support to domestic sugar than other parts of the world. According to the production. For several of these countries Global Panel on Agriculture and Food Systems for with seemingly inconsistent policies around Nutrition (GLOPAN) 2022 report, in 2015, countries sugar consumption and production, the in Sub-Saharan Africa provided only $680 million in agriculture subsidies. This number is extremely levels of sugar consumption exceed the WHO low, considering that agriculture constitutes 23 recommendations. percent of gross domestic product (GDP) in Sub- Saharan Africa and employs 60 percent of people The nature of agrifood support can point to a lack in the region. A recent report by the FAO, the of policy coherence. In certain cases, agrifood United Nations Development Programme (UNDP), public policy and support might be at odds with and United Nations Environment Programme public policy in other sectors, such as in the health (UNEP) (FAO et al. 2022) shows that the trend of sector. For example, the production of sugar is low agriculture expenditure in Africa continues, one of the most publicly supported. However, particularly citing the negative agriculture many governments around the world are also producer support, which suppresses prices, taxing the consumption of SSBs in recognition of favoring consumers for whom food affordability the harmful effects of a final product with high plays a vital role. In addition, Pernechele et al. sugar content (see box 2.3). This demonstrates (2021) affirm that some countries in Sub-Saharan potential inconsistencies in the commodities Africa allocate only 6 percent of their expenditure that governments support production of and the to food and agriculture, which is well below the commodities that they curb consumption of. Figure African Union’s target of 10 percent. Per GLOPAN, 2.4 presents a scatterplot of various countries, with when it comes to the components of public grams of sugar consumption per day in 2020 in support, countries in Sub-Saharan Africa prioritize the y-axis, and public support to sugar producers inputs (for example, fertilizers), allocating 88 as a share of gross receipts from sugar averaged 32 Reshaping the Agrifood Sector for Healthier Diets between 2020 and 2022 in the x-axis. Countries have provided relatively significant support to their marked as green triangles have implemented some sugar producers but have also imposed an SSB taxes form of SSB tax, while countries marked as red have on their consumers. These countries are close to or not. Countries such as Costa Rica, India, Mexico, have exceeded the WHO maximum threshold of 50 the Philippines, South Africa, and the United States grams/day of sugar consumption. Box 2.2 Differences between Recommended Intake and Current Intake, by Food Groups in Africa Diets differ across regions, making the percent more whole grain, 32 percent more recommended intakes of foods vary throughout milk, 55 percent more fish, 54 percent more the globe. In Africa, the consumption of high- fruits and vegetables, and 19 percent less meat priority agricultural commodities remains an to bring the consumption levels to meet the issue. Figure B2.2.1 shows the deviation of FBDGs. Meeting FBDGs has been shown to consumption in selected African countries from have the ability to reduce premature mortality the recommended local food-based dietary by 15 percent while also reducing greenhouse guidelines (FBDGs). The positive percentage gas (GHG) emissions by 13 percent on average indicates by how much the consumption of (GLOPAN 2022). To meet the FBDGs or at least selected food groups needs to increase to initiate a positive switch in diets in Africa, meet the FBDGs. For example, people in Africa, repurposing current subsidies could be an in order to meet the dietary guidelines, need effective tool to impact people’s consumption to consume 240 percent more legumes, 113 habits. Figure B2.2.1  Differences between Recommended and Current Intake, by Food Group in Africa Legumes +240% Whole grain +113% Milk +32% Fish +55% Nuts and seeds +29% Fruit and vegetables +54% (Fruits +50%; Vegetables +58%) Eggs +20% Sugar –2% Meat –19% (Poultry-18%; Red meat -15%; Processed meat -46%) Energy intake +7% –50 0 50 100 150 200 250 Source: This figure has been amended from Springmann et al. 2020. Note: The data presented in the figure relate to a subset of African countries that have implemented food-based dietary guidelines. These are Benin, Kenya, Namibia, Nigeria, Seychelles, Sierra Leone, and South Africa. An Opportunity to Leverage Agrifood Public Support 33 Figure 2.4  Sugar Consumption, Taxes on SSBs, and Support to Sugar Producers Sugar consumption (grams per day) 50 IND 100 MEX CHL RUS COL CRI USA BRA GBR PHL UKR IDN 50 VNM ZAF AUS TUR CHE JPN CHL 0 0 20 40 60 80 Support to sugar producers (as share of GR) No SSB tax SSB tax Source: Authors’ calculations using data from the OECD (2023). OECD Agriculture statistics (database), IDB (2023). IDB Agrimonitor - PSE Agricultural Policy Monitoring System (data), The Global Nutrition and Policy Consortium (2022). Global Dietary Database, and The World Bank (2023). Global SSB Tax Database, August 2023. Note: The measure of support to sugar producer—the percentage of Producer Single Commodity Transfers (%PSCT) of the gross farm receipts (GR) for sugar—is defined as the total gross transfers from consumers and taxpayers to agricultural producers, measured at the farm gate level, arising from policies linked to the production of sugar. Countries are indicated by their three-letter ISO codes, available at https://www.iban.com/country-codes./. SSB = sugar-sweetened beverages. Box 2.3 Global Implementation of Sugar-Sweetened Beverage Taxes Sugar-sweetened beverage (SSB) taxation is a analysis that considers costs and benefits in fiscal policy tool primarily aimed at reducing the longer term suggests that SSB taxes can the health burden associated with consuming have a progressive impact, with lower-income unhealthy drinks,1 which can also raise households expected to benefit from a additional revenue. SSBs are a key contributor disproportionate share of improved health to excess sugar and energy intakes around outcomes, reduced health care costs, extended the world; they are strongly linked to long- working lives, and reduced years of life lost term weight gain, obesity, and a range of (Fuchs, Mandeville, and Alonso-Soria 2020). noncommunicable diseases (NCDs) including As SSBs are generally cheaper than other type 2 diabetes, dental caries, cardiovascular unhealthy goods such as cigarettes and diseases, and at least 12 cancers (World tobacco, their revenue potential is smaller— Bank 2020). Taxes on SSBs should raise retail an average of 0.07 percent of gross domestic prices, which can be viewed as increasing product in countries for which data are the tax burden on poorer groups. However, available (World Bank 2023b). 34 Reshaping the Agrifood Sector for Healthier Diets Evidence demonstrating the effectiveness of and Hernandez-Cortes 2022; Hernández-F, SSB taxes is expanding rapidly (Andreyeva et Cantoral, and Arantxa Colchero 2021) and the al. 2022). There is strong evidence that SSB United Kingdom (Rogers, Conway et al. 2023; taxes raise prices and reduce sales of taxed Rogers, Cummins et al. 2023). products and can encourage substitution with healthier alternatives. Certain tax designs SSB taxation is not a novel or untested policy can also create supply-side incentives to tool, although only a few existing taxes are reduce the volume of sugar sold through optimized for health outcomes. More than 100 SSBs. Given that many SSB taxes have been economies applied national-level taxes on SSBs recently implemented, evidence of impacts as of 2023, covering more than half (57 percent) on population health outcomes remains of the world’s population and more than four more limited, though early improvements in in five people (82 percent) in low- and lower- oral health and obesity indicators have been middle-income economies (map B2.3.1). observed in Mexico (Gracner, Marquesz-Padilla, Map B2.3.1  Taxes on SSBs, as of August 2023 Source: World Bank 2023a. Note: Targeted excise tax: Unsweetened bottled water is exempt or taxed at a lower rate than SSBs. Untargeted excise tax: Unsweetened bottled water is taxed at the same or higher rate as SSBs. However, one-third of low- and middle-income represent a missed opportunity to reduce sugar economies that tax SSBs also tax unsweetened consumption by shifting demand. bottled water—a key healthy substitute—at the same or higher rate. These untargeted excise As with other health taxes, tax design is crucial taxes on nonalcoholic beverages more broadly to achieving health and equity objectives. Fewer than one in five SSB taxes worldwide An Opportunity to Leverage Agrifood Public Support 35 (19 percent) target sugar content (either half of SSB taxes worldwide cover sweetened including a component based on absolute milk-based drinks (42 percent) and only one sugar content or with sugar content tiers), in three (36 percent) cover 100 percent juices, with these taxes concentrated in high-income undermining potential health gains. economies (World Bank 2023a).2 Fewer than Notes 1. SSBs are all beverages containing free sugars, including caloric carbonated soft drinks (sodas), energy drinks, concentrates, and syrups that are diluted to make drinks, was well as sweetened and unsweetened juices, sweetened (flavored) milk-based drinks, sweetened (flavored) waters, sports drinks, and sweetened ready-to-drink teas and coffees. Free sugars include monosaccharides (such as glucose and fructose) and disaccharides (such as sucrose or table sugar) that are added to foods and beverages by the manufacturer, cook, or consumer, as well as those naturally present in honey, syrups, fruit juices, and fruit juice concentrates. 2. See the World Bank Global SSB Tax Database, August 2023 version (World Bank 2023a). 2.2. An opportunity to outcomes, there are also repurposing scenarios that provide windows of opportunity for multiple repurpose agrifood support wins. Repurposing agrifood support for healthy diets should consider regional differences, Given that agrifood support is biased toward country contexts, and political economy less healthy and nutritious commodities, this constraints; repurposing must be accompanied by support potentially fuels the high costs of complementary interventions. malnutrition. There is an imperative to rethink The most supported commodities are not only less and repurpose agrifood support in a way that healthy, but they are also typically worse for the better promotes healthy diets. While there are climate. However, repurposing agrifood support often tradeoffs to repurposing, there are some requires careful thinking about how to achieve better windows of opportunity for multiple wins climate and nutrition outcomes together. There across different outcomes—such as climate, are often tradeoffs between these intertwined but nutrition, and poverty alleviation. In Sub- different objectives. Commodities that receive the Saharan Africa, some scenarios indicate that most support—such as beef, pig, poultry, rice, and simultaneously increasing and rebalancing sugar—have much higher greenhouse gas (GHG) emissions and freshwater use than other healthier agrifood support can yield important benefits. commodities that receive less support, such as fruits and vegetables (see figure 2.5). Thinking of We know that the costs of malnutrition are high, climate and nutrition together would be important that progress to curb malnutrition has slowed, in repurposing agrifood support. However, targeting and that the promotion of healthy diets is a climate outcomes does not necessarily lead to cornerstone to regaining the moment to address better nutrition outcomes, or vice versa. The EAT– malnutrition. At the same time, public support Lancet Commission’s planetary health diet is an to the agrifood sector is not geared toward important benchmark diet that considers both promoting healthier diets and better nutrition. human health and the environment—including There is, however, a growing body of work both GHG emissions, water use, and biodiversity loss across and within countries, largely leveraging (Willet et al. 2019). But while this benchmark simulation methods, which explores the options proposes a global standard, there is huge variation for repurposing agrifood support. While there in the environmental impacts of dietary changes are often tradeoffs between climate and nutrition for different types of environmentally friendly 36 Reshaping the Agrifood Sector for Healthier Diets adjustments (Kim et al. 2020) and for various by 10 percent if consumption were adjusted to the countries (Semba et al. 2020). For example, if all national dietary guidelines (which consider only countries adopt the planetary health diet, overall health outcomes), and, counterintuitively, by more global GHG emissions would go down but in some, than 20 percent if adjusted to the planetary health primarily low- and middle-income countries, they guidelines (which consider both environment would increase. It is crucial then to assess dietary and health outcomes) (Mehra et al. 2022). Similar pattern, nutritional value, and environmental impact country-level tradeoffs have been carefully explored at the country level. In Bangladesh, for example, in Indonesia (de Pee et al. 2021, among others). evidence shows that GHG emissions would increase Figure 2.5  Greenhouse Gas Emissions of Food Commodities across the Supply Chain (per kg of food product), 2018 a. Meat and cereals Greenhouse gas emissions per kilogram 99.5 kg 100 35 33.3 kg 80 25 60 40 15 12.3 kg 9.9 kg 20 6.3 kg 5 4.5 kg 3.2 kg 3.2 kg 1.0 kg 1.8 kg 1.7 kg 1.6 kg 0 Beef (beef herd) –5 Beef Milk Soymilk Soybean Cane Beet (dairy Oil Sugar Sugar herd) Milk Pig Poultry Soybeans Rice Sugar Maize Wheat & Meat Meat Rye b. Fruits and vegetables 4 3.2 kg 2.1 kg 2 1.8 kg 1.5 kg 0.4 1.1 kg 0.9 kg 1.0 kg kg 0.4 kg 0.4 kg 0.5 kg 0.5 kg 0.5 kg 0.4 kg 0.5 kg 0 –2 –4 s pe & t s t s es as s s bl r es bl t ek & es ui ui ta oo ta the ut na ut oe ca G ies Le ns Pe ls to s Po s O s s Fr Fr pl N dn si ge R at na e e Pu ge O ta o Ap as rr er us m un ni Ba ra Be er Br th To tr ro th Ci O G Ve Ve O Fruits Vegetables Land use change Farm Animal feed Processing Transport Retail Packaging Losses Total Source: Poore and Nemecek, 2018, with data from Our World in Data. Note: Greenhouse gas emissions are measured in kilograms of carbon dioxide-equivalents (CO₂eq) per kilogram of food. An Opportunity to Leverage Agrifood Public Support 37 Global simulation exercises that look into various investments, aimed at increasing productivity repurposing scenarios demonstrate important and decreasing emissions (Gautam et al. 2022); tradeoffs across a number of outcomes. Multiple this scenario is called “Repurposing for green modeling exercises have been conducted to innovation.” The modeling in this case shows analyze the potential of repurposing agriculture that redirecting public support to research and public support. Several scenarios of the joint World development and other technological investments Bank and International Food Policy Research could lead to “triple wins” for a healthy planet, Institute (IFPRI) report outline some crucial economy, and people. With the baseline year of tradeoffs (Gautam et al. 2022). For instance, in 2020, the study finds that, by 2040, repurposing for one scenario, eliminating the domestic support green innovation can lead to a 1.6 percent increase to producers would reduce agriculture-related in real national income, a 1 percent reduction in GHG emissions by around 1.5 percent of total extreme poverty, and an 18 percent reduction agricultural emissions in the baseline and lower in the cost of healthy diet globally compared the land area used by the sector. These results, to the business-as-usual scenario. Agricultural however, come at the cost of lower farm output productivity also increases in this repurposing and income per farmer as well as increased scenario, with the volume of crop production rising food insecurity, higher cost of healthy diets, and by 16 percent and of livestock by 11.5 percent. increased poverty. In another scenario, making In addition, repurposing for green innovation producer support conditional on the adoption can vitally lead to a 40.5 percent reduction in of emission-reducing practices would reduce emissions from agriculture and land use, which agriculture related GHG emissions by 15 percent, partially derives from a 2.15 percent decrease in yet this number is offset by an increase in agricultural land. On the negative side, however, agricultural land use, a decline in global income the World Bank–IFPRI modeling finds that the and agriculture production, and an increase in the above-mentioned scenario can also result in an 8 cost of a healthy diet and in poverty. percent decrease in real farm income per worker, a 10.5 percent reduction in farm employment, and a The 2022 State of Food Security and Nutrition in the 27.5 percent increase in the consumption of sugar, World (SOFI) global report—which is produced which is already an overconsumed commodity by the FAO jointly with the International Fund (Gautam et al. 2022). for Agricultural Development (IFAD), UNICEF, the World Food Programme (WFP), and the WHO— At the same time, these simulation exercises presents multiple repurposing scenarios and also suggest that there are also opportunities for sheds light on these tradeoffs. In one scenario, win-wins when targeting for better health eliminating fiscal subsidies to producers reduces outcomes. The 2022 SOFI report’s scenario 6 GHG emissions but leads to food and nutrition looks at repurposing agrifood support (fiscal insecurities (FAO et al. 2022). On the other hand, subsidies) from producers to consumers for repurposing fiscal subsidies to producers improves priority foods to support healthy diets (figure the affordability of a healthy diet and lowers global 2.6). This scenario targets high-priority foods extreme poverty; however, in this scenario, GHG with a 10 times increase in average consumer emissions from agriculture increase by 1.5 percent subsidy, keeps the average level of consumer as a result of higher production. subsidy for medium-priority foods, and keeps only one-tenth of the average level of support However, these same global simulation exercises for low-priority commodities. With the baseline suggest that there are opportunities for multiple year of 2017, the study finds that, by 2030, this wins when targeting for better climate outcomes repurposing scenario can reduce extreme poverty (figure 2.6). Scenario 4 of the joint World Bank by 0.06 percent, reduce GHG emissions by 0.18 and IFPRI report looks at repurposing some percent, and, importantly, decrease the cost of of domestic support to specific technological a healthy diet by 3.34 percent globally. Extreme 38 Reshaping the Agrifood Sector for Healthier Diets poverty falls the most—by 0.22 percent in low- impacts high-income countries; low-income income countries. Since repurposing for healthy countries experience an increase in farm income diets considers the reduction in the support to of 1.61 percent and agricultural production of 0.36 low-priority foods, the consumption of sugar and percent (FAO et al. 2022). In practice, repurposing sweeteners can be lowered by 0.04 percent while would require transitioning farmers away from at the same time increasing the consumption of their current livelihoods. While often politically dairy as well as vegetables and fruits. A substantial difficult, there are lessons that can be learned from tradeoff in this scenario is the reduction in farm similar experiences, for example for transitioning income of 3.74 percent and agricultural production away from tobacco farming (see box 2.4). of 0.2 percent. However, this outcome significantly Figure 2.6  Repurposing Simulations at the Global Level: Potential Win-Win Scenarios Impact of repurposing on economies, people, and planet Impact of scenarios on per capita consumption Scenario 1 (baseline year – 2017; end Scenario 2 (baseline year – 2020; end year – 2030): Repurposing agrifood year – 2040): Repurposing to Invest in fiscal subsidies from producers to R&D and incentives for adoption of Sugar and sweeteners consumers for priority foods to practices that both raise productivity support healthy diets and reduce emissions. –0.04% 27.5% (FAO et al. 2022) (Gautam et al. 2022) Fats and oils Percentage 25.3% 8.7% 20 14.0 10 n/a 1.6 n/a Dairy 0 –0.2 –0.06 –1.0 –3.3 –0.18 –2.15 –10 2.95% 16.4% –20 –18.0 –30 Fruits and vegetables –40 0.41% 12.0% –40.5 –50 ECONOMIC FARM SECTOR SOCIAL DIETS ENVIRONMENT NATURE National Farm Extreme Cost of a GHG Land Use Income Production Poverty Healthy Diet Emissions Repurposing for Healthy diets Repurposing for Green innovation Sources: Original figure for this publication based on data from FAO et al. 2022 and Gautam et al. 2022. Note: The scenario on the left is listed as Scenario 6 in FAO et al. (2022), shifting fiscal subsidies from producers to consumers in support of healthy diets. In this new scenario the fiscal subsidies initially allocated to producers no longer stay within the agrifood sector, although they remain within the agrifood system. The scenario on the right is listed as Scenario 4 in Gautam et al. (2022). In this scenario, a portion of current domestic support would be repurposed for increased spending on green innovations; that is, the development, diffusion, and adoption of new technologies that both reduce emissions and raise productivity. “Farm Production” is the average change in the production volume of crops and livestock. In Africa, simulation exercises indicate that it to producers to cultivate high-priority crops such simultaneously increasing and rebalancing as fruits, vegetables, legumes, and nuts. With the agrifood support can yield important benefits. baseline year of 2011, the study finds that this Additional Africa-focused modeling studies repurposing scenario, in 12 years, could have have been conducted to look at the repurposing dramatically increased the production of high- possibilities on the continent. Particular attention priority horticulture by 18.269 million metric tons. was paid to how increasing agriculture spending to However, other commodities would experience up to 10 percent of national budgets and allocating a production decline; for instance, wheat and An Opportunity to Leverage Agrifood Public Support 39 other grain production would fall by over 1.100 Repurposing for better diet and health outcomes million metric tons and the production of sugar requires a set of complementary interventions. would be cut by 522,000 metric tons. Despite Such interventions are vital to ensure that there some of the losses, the overall agrifood sector are incentives (disincentives) that work together would gain almost $5.2 billion in sectoral income with repurposing scenarios to push for the and export revenue. Excessive or insufficient changes in food supply chains and environments consumption of various agricultural commodities and shift consumer behavior toward healthier creates diet-related risk factors, which can even options (FAO et al. 2022). A suite of complementary lead to death.2 Hence, deaths can be averted by measures is essential to mitigate and protect either increasing the consumption of high-priority against unintended consequences stemming commodities or reducing the consumption of from reallocating support, particularly if such low-priority ones. This repurposing scenario also transitions result in challenges to accessing models the avoidance of thousands of deaths nutritious foods and healthy diets for vulnerable and leads to a reduction in GHG emissions of and disadvantaged populations. Agrifood systems around 0.7 percent. Among the avoided deaths, transformation must consider the roles of private this scenario tackles the underconsumption of enterprises, such as agribusinesses, that impact legumes, vegetables, nuts and seeds, and fruits the availability, access, utilization, and stability of and lessens the impact of the underweight risk food systems. factor. On the other hand, obesity and overweight deaths rise, but far less than the number of Complementary interventions to repurposing prevented deaths (GLOPAN 2022). This result is include relevant agrifood systems policies, among particularly important for children, since poor others. These policies include implementing diets can lead to lasting impacts on physical limits or targets for food reformulation to health and cognitive development, such as improve nutritional quality and availability; stunting, which can influence overall well-being enhancing nutritional content via fortification and future productivity. Lastly, the tradeoff of and biofortification; implementing laws regarding this action would be an increase in the use of the promotion of food and beverages, along land and water by around 2.5 percent and 3.8 with enforcing policies on nutrition labeling; percent, respectively. Thus, expanding agriculture implementing taxes on energy-dense foods, rich spending in Africa while targeting high-priority in fats, sugars, and/or salt; integrating land-use commodities is a viable option for the continent policies with other complementary measures to improve its competitiveness in global markets to tackle food deserts and swamps; enforcing and to increase the nutritional value of diets in policies for healthy public food procurement and the region. In addition, newer evidence indicates services (for example, having nutrition-guided that there is dynamism in the fruits and vegetable meals and beverages in public settings, such as supply in Africa owing to investments in wholesale schools). Other policies include social protection markets, roads, other infrastructure such as system policies (for example, expanding existing electrification, and agricultural research/extension social protection programs, such as social (Reardon et al. 2024). insurance and labor market interventions; building new shock-responsive systems to safeguard Repurposing agrifood support requires a vulnerable populations against potential tradeoffs set of complementary interventions across of public support repurposing); environmental various sectors, including agrifood systems, and climate-related policies and incentives (for health systems, social protection, and the example, supporting adaptation and mitigations environment. Central to repurposing is a measures, promoting intercropping, and reducing careful consideration of the political economy chemical fertilizer use); health system policies (for example, focusing on vulnerable populations to determine what is feasible and how to by strengthening mother and child nutritional overcome political economy constraints. 40 Reshaping the Agrifood Sector for Healthier Diets services, enhancing the One Health approach,3 highlight some important lessons around political hence the communication between agrifood, economy. These include the significance of the environment and health systems); and other power of ideas around both the links between food system policies (for example, boosting efficiencies security and food availability (without considering in transportation and energy to reduce the cost diversification) and the contrast between self- of the repurposing transition, ensuring the sufficiency and multilateral cooperation; the availability of cold chains to prevent food losses of importance of institutional commitments; and the perishable commodities). impetus provided by the failure to sustain high or distorted market support prices. Ultimately, reform Moreover, a careful consideration of the political is hindered without strong commitment from economy around repurposing is crucial. The political political leaders and an effort toward multilateral economy constraints to policy reform are significant collaboration (Vos, Martin, and Resnick 2022). On because existing policies are often politically the topic of crops that promote unhealthy practices, popular and serve well-entrenched interests. A other lessons on reform can be learned from review of recent policy reform experiences in China, transitioning away from tobacco crop (see box 2.4). the European Union, India, and the United States Box 2.4 Lessons in Transitioning Farmers away from Tobacco Farming Due in large part to global tobacco control given that farmers and leaf-buyers typically measures including tobacco taxes, cigarette face low tax rates and often benefit from consumption and, in turn, demand for tax exemptions. Most tobacco-producing tobacco leaf have been declining for nearly countries—particularly those with the largest two decades. This has coincided with a shift relative economic contribution—are typically in production from developed to developing net exporters of tobacco leaf, making the countries, often with government support and connection between domestic excise taxes, economic incentives, and stagnant or declining consumption, and tobacco growing marginal. prices that farmers are paid for tobacco leaf. Tobacco-growing countries have scope to This generates concerns for the livelihoods of increase domestic tobacco taxes with limited farmers and economic development in tobacco- impact on their leaf-growing markets. growing countries, and pressure to seek alternatives for tobacco growing. While tobacco growing is generally a small contributor to economies, it can have a Unprocessed tobacco leaf is rarely subject to significant impact on poverty and farmer tobacco excise taxes. Tobacco excise taxes welfare. In the 90 countries that grow tobacco, levied on manufactured products such as the crop contributes less than 0.2 percent of cigarettes are among the most effective tools gross domestic product (GDP) in 70 of them; at reducing the negative externalities and however, it has a considerable economic impact internalities associated with tobacco use. in a very small number of countries. In many Unprocessed tobacco leaf production is not countries, tobacco is also grown by smallholder a large source of tax revenue on its own, farmers, with these farmers having higher An Opportunity to Leverage Agrifood Public Support 41 rates of poverty than their peers. They are likely paid work in nearby towns, entrepreneurial to be in a debt cycle with tobacco companies, endeavors, and small businesses. Farmers who with few safeguards in place to protect spend less time tending their crops have more farmers. They experience food insecurity and time to commit to these other endeavors, and consistently use household or child labor as data show that these farmers are consistently a low-cost alternative to hiring adult workers. making a better overall livelihood through Tobacco cultivation can have negative health alternative economic activities. For example, effects for farmers and is one of the most in Indonesia, average total household income environmentally devastating crops. of former tobacco farmers was found to be 30 percent higher than that of current tobacco As global demand is expected to continue farmers, reflecting widespread poverty among to decline, many tobacco-growing countries Indonesian tobacco farmers. Some farmers are particularly vulnerable since they have now move in and out of tobacco cultivation little effect on global markets and are price based on tobacco yields, prices, and often takers, highlighting the urgency of considering crude weather predictions. This transition transitioning to alternatives. Alternative may be more feasible in countries that livelihoods programs for tobacco farmers are undergoing a broader shift of income- have been scarce, and the few that have generating opportunities out of agriculture emerged have suffered from well-documented into manufacturing and services, supported by challenges. For example, a set of programs in a business-friendly investment environment Kenya in the late 2000s introduced bamboo as well as investment in human capital and as an alternative crop. However, the programs economic infrastructure. There are several failed to establish a strong market mechanism actions that governments can take to support for the farmers to sell their crop, resulting in successful shifting away from tobacco growing: long-term negative impacts when farmers struggled to sell the bamboo. A new program • Invest in and support marketplaces to connect with a focus on beans is aiming to address this farmers to buyers of their crops. Research core issue by ensuring that markets exist for suggests that many tobacco farmers grow farmers to sell their crops. tobacco because they feel relatively assured of its economic viability and a guaranteed A considerable number of farmers are shifting market for the crop. The probability of gradually away from tobacco where the successfully shifting is greatly enhanced conditions are favorable. Research from major where there are accessible markets to sell tobacco-growing regions suggests that tobacco other crops. farmers have dynamic on-farm economic lives, • Improve supply and value chains for other often growing other crops including staple and products to incentivize farmers to shift. For other food crops, many in surplus to sell on example, in Migori County, a major tobacco- the open market (for example, groundnuts in growing region in Kenya, a sweet potato Malawi, wheat in North Macedonia, maize in processing plant has been established South Africa, and green vegetables in Zambia). by the government and many farmers Many farmers in tobacco-growing regions also there have begun to grow sweet potato, pursue off-farm economic activities including appearing to make more money doing so. 42 Reshaping the Agrifood Sector for Healthier Diets • Expand extension services. Many farmers • Ensure access to information and related report that they grow tobacco because it is technological infrastructure. Farmers often all they know how to do well and that they lack up-to-date information that is key to receive extension services directly from making better decisions, including planting the tobacco companies. To address this decisions (for example, information about issue, governments can consider expanding prices, buyers, weather, and climate). extension services for farmers to learn how Improving access to and the quality of to better grow other locally suitable crops. information results in better outcomes for farmers. For example, in Kenya, farmers • Support rural credit schemes to provide more with access to mobile phones that have options to farmers beyond tobacco. Farmers reliable coverage were found to be more consistently cite lack of credit as a reason likely to shift away from tobacco, have they choose to grow tobacco, and contracts more diverse economic portfolios, and with leaf-buying companies provide a have better overall economic livelihoods. form of credit. There is a widespread lack Governments can help farmers by ensuring of access to credit in rural settings, and reliable mobile phone coverage and by governments can actively support rural developing local platforms on which credit schemes to provide more options to information is shared. farmers beyond tobacco contracts. Source: Drope et al. 2023. reducing poverty (Sanchez and Cicowiez 2022). Assessing options for repurposing requires Focusing solely on reducing the cost of diets leveraging better data and techniques to would somewhat slow down progress toward carefully assess the links between agrifood IAT. However, focusing solely on IAT objectives policy and support, food consumption, and ensures faster progress toward increased output, healthy diets. Assessing options will also more jobs, and lower poverty while at the same time reducing the cost of diets, suggesting a require a rich set of country-level deep dives window of opportunity for win-wins. In Uzbekistan, that investigate policy reform simulations and the government is implementing programs to their impacts on health and environmental support modernization in agriculture, including outcomes. drip irrigation adoption and investments in viniculture (World Bank 2021). Irrigation plays Country-level simulation exercises in the context a pivotal role in the country’s agriculture of public expenditure reviews provide a good sector, particularly the production of cotton. A understanding of tradeoffs and win-wins around modeling exercise shows that utilizing effective healthy diets and climate outcomes; such studies drip irrigation systems could decrease water can be used as a blueprint for other country-level consumption and expenses related to fertilizers, deep dives. In Ethiopia, modeling work focused fuel, and machinery, meanwhile increasing cotton on healthy diets has assessed ways in which the yields. Rising demand for irrigation services government could rework its budget to both then urges the private sector to enhance their reduce the cost of diets while achieving inclusive investments in these systems. Furthermore, drip agriculture transformation (IAT), which consists irrigation adoption is also expected to lower of increasing agrifood output, creating jobs, and the use of fertilizers, fuel, and machinery and, An Opportunity to Leverage Agrifood Public Support 43 ultimately, almost double the profits of farmers to the largest reductions in GHG emissions. over the course of five years. However, this Combining a national carbon tax with technical will also increase labor costs (Rudenko 2019; solutions is expected to yield a much higher GHG World Bank 2021). In Viet Nam, two technically reduction. The simulations here are valuable and economically sound strategies have been exercises, but there is no-one-size fits all analysis piloted that can preserve or increase rice yields on repurposing for diets at the country level. A and farmer incomes while decreasing GHG variety of analytical work at the country level can emissions. The first one is alternate wetting help elucidate policy options. Chapter 3 presents and drying (AWD), which promotes water some global-level analysis on the links between management; the second one is One Must Five public agrifood support and food consumption, Reductions (1M5R) techniques, which focus on while chapter 4 presents some country-level case optimal input use (World Bank 2022). Simply studies on the impacts of various programs on adopting these technical solutions will not lead healthy diets. Notes 1. The term public support refers to the total amount of monetary transfers to agriculture producers, consumers, and general services. Public support includes budget subsidies and Market Price Support that derive from trade measures. Chapter 3 provides a detailed explanation of public support and how it is calculated. 2. Risk factors include low consumption of fruits, vegetables, legumes, and nuts and seeds; high consumption of red meat; and being overweight or obese. 3. Details about the WHO’s One Health approach can be found at https://www.who.int/health-topics/one-health#tab=tab_1. 44 Reshaping the Agrifood Sector for Healthier Diets References Andreyeva, T., K. Marple, S. Marinello, T. E. Moore, and L. M. Powell. 2022. “Outcomes Following Taxation of Sugar-Sweetened Beverages: A Systematic Review and Meta-Analysis.” JAMA Network Open 5 (5): e2215276-e. https://doi.org​/10.1001/jamanetworkopen.2022.15276. Damania, R., E. Balseca, C. de Fontaubert, J. Gill, K. Kim, J. Rentschler, J. Russ, and E. Zaveri. 2023. Detox Development: Repurposing Environmentally Harmful Subsidies. Washington, DC: World Bank. https://www.worldbank.org/en​ /topic/climatechange/publication/detox-development. de Pee, S., R. Hardinsyah, F. Jalal, B. F. Kim, R. D. Semba, A. Deptford, J. C. Fanzo, R. Ramsing, K. E. Nachman, S. McKenzie, and M. W. Bloem. 2021. “Balancing a Sustained Pursuit of Nutrition, Health, Affordability, and Climate Goals: Exploring the Case of Indonesia.” The American Journal of Clinical Nutrition 114 (5): 1686–97. https://doi.org/10.1093/ajcn/nqab258, Drope, J., C. Lane, D. Prinz, E. Blecher, C. Ozer, and D. Bloom. 2023. Tobacco Excise Taxes and Tobacco Leaf Farming—Key Considerations. Global Tax Program Health Taxes Knowledge Note Series. World Bank. https://thedocs​ .worldbank.org/en/doc/a95559bf187f64dba0346d973d16f991-0350012023​ /original/KN3-Health-tax-Farming.pdf. FAO, IFAD, UNICEF, WFP, and WHO (Food and Agriculture Organization of the United Nations, International Fund for Agricultural Development, UNICEF, World Food Programme, and World Health Organization). 2022. The State of Food Security and Nutrition in the World 2022: Repurposing Food and Agricultural Policies to Make Healthy Diets More Affordable. Rome: FAO. https://doi.org​/10.4060/cc0639en. Fuchs, A., K. Mandeville, and A. C. Alonso-Soria. 2020. “Health and Distributional Effects Taxing Sugar-Sweetened Beverages: The Case of Kazakhstan.” World Bank Poverty & Equity Notes June 2020, Number 24. https://documents1.worldbank.org/curated/en/166111591327025410/pdf/ Health​-and-Distributional-Effects-Taxing-Sugar-Sweetened-Beverages-the- case-of​-Kazakhstan.pdf. Gautam, M., D. Laborde, A. Mamun, W. Martin, V. Piñeiro, and R. Vos. 2022. Repurposing Agricultural Policies and Support: Options to Transform Agriculture and Food Systems to Better Serve the Health of People, Economies, and the Planet. Washington, DC: World Bank and IFPRI. http://hdl.handle.net/10986​/36875. Glauber, J. and D. Laborde. 2023. “Repurposing Food and Agricultural Policies to Deliver Affordable Healthy Diets, Sustainably and Inclusively: What Is at Stake?” Background paper, The State of Food Security and Nutrition in the World 2022. FAO Agricultural Development Economics Working Papers, No. 22-05. Rome: FAO. An Opportunity to Leverage Agrifood Public Support 45 GLOPAN (Global Panel on Agriculture and Food Systems for Nutrition). 2022. Exploring Potential Benefits of Repurposing Agricultural Subsidies in Sub-Saharan Africa. Technical Brief No. 4. September 2022. https://www​.glopan.org/subsidies/. Gracner, T., F. Marquez-Padilla, and D. Hernandez-Cortes. 2022. “Changes in Weight-Related Outcomes among Adolescents Following Consumer Price Increases of Taxed Sugar-Sweetened Beverages.” JAMA Pediatrics 176 (2): 150–58. https://doi.org/10.1001/jamapediatrics.2021.5044. Hernández-F, M., A. Cantoral, and M. Arantxa Colchero. 2021. “Taxes to Unhealthy Food and Beverages and Oral Health in Mexico: An Observational Study.” Caries Research 55 (3): 183–92. https://doi.org/10.1159/000515223. IDB (Inter-American Development Bank). 2023. Agrimonitor Database. https://agrimonitor.iadb.org/en/. Kim, B. F., R. E. Santo, A. P. Scatterday, J. P. Fry, C. M. Synk, S. R. Cebron, M. M. Mekonnen, A. Y. Hoekstra, S. de Pee, M. W. Bloem, R. A. Neff, and K. E. Nachman. 2020. “Country-Specific Dietary Shifts to Mitigate Climate and. Water Crises.” Global Environmental Change 62 (101926). https://doi.org/10.1016/j.gloenvcha.2019.05.010. Mehra, D., J. Ton, F. Dizon, and S. de Pee. 2022. “Healthy and Sustainable Diets in Bangladesh.” Policy Research Working Paper 10160. Washington, DC: World Bank Group. https://documents1.worldbank.org/curated​ /en/099250009012254667/pdf/IDU06764ba7f056d80448e0a10409985​ 341922cd.pdf. OECD (Organisation for Economic Co-operation and Development). 2023a. Agricultural Policy Monitoring and Evaluation 2023: Adapting Agriculture to Climate Change. Paris: OECD. https://doi.org/10.1787/b14de474-en. OECD (Organisation for Economic Co-operation and Development). 2023b. OECD Agriculture Statistics, “Producer and Consumer Support Estimates.” Database. https://doi.org/10.1787/agr-pcse-data-en. Pernechele, V., F. Fontes, R. Baborska, J. Nkuingoua, X. Pan, and C. Tuyishime. 2021. Public Expenditure on Food and Agriculture in Sub-Saharan Africa: Trends, Challenges and Priorities. Rome: FAO. https://doi.org/10.4060/cb4492en. Poore, J. and T. Nemecek. 2018. “Reducing Food’s Environmental Impacts through Producers and Consumers.” Science 360 (6392): 987-92. https://doi.org​/10.1126/science.aaq0216. Reardon, T., L.S.O. Liverpool-Tasie, B. Belton, M. Dolislager, B. Minten, B. Popkin, and R. Vos. 2024. “African Domestic Supply Booms in Value Chains of Fruits, Vegetables, and Animal Products Fueled by Spontaneous Clusters of SMEs.” Applied Economic Perspectives and Policy. Published “early view” April. http://doi.org/10.1002/aepp.13436. Rogers, N. T., D. I. Conway, O. Mytton, C. H. Roberts, H. Rutter, A. Sherriff, M. White, and J. Adams. 2023. “Estimated Impact of the UK Soft Drinks Industry Levy on Childhood Hospital Admissions for Carious Tooth Extractions: 46 Reshaping the Agrifood Sector for Healthier Diets Interrupted Time Series Analysis.” BMJ Nutrition, Prevention & Health e000714. https://nutrition.bmj.com/content/early/2023/10/31/bmjnph-2023-000714. Rogers, N. T., S. Cummins, H. Forde, C. P. Jones, O. Mytton, H. Rutter, S. J. Sharp, D. Theis, M. White, and J. Adams. 2023. “Associations between Trajectories of Obesity Prevalence in English Primary School Children and the UK Soft Drinks Industry Levy: An Interrupted Time Series Analysis of Surveillance Data.” PLoS Medicine 20 (1): e1004160. https://doi.org/10.1371/journal.pmed​.1004160. Sánchez, M. V. and M. Cicowiez. 2022. “Repurposing Agriculture’s Public Budget to Align Healthy Diets Affordability and Agricultural Transformation Objectives in Ethiopia.” Background paper, The State of Food Security and Nutrition in the World 2022. FAO Agricultural Development Economics Working Paper 22-04. Rome: FAO. https://doi.org/10.4060/cc1174en. Semba, R. D., S. de Pee, B. Kim, S. McKenzie, S. Nachman, and M. W. Bloem. 2020. “Adoption of the ‘Planetary Health Diet’ Has Different Impacts on Countries’ Greenhouse Gas Emissions.” Nature Food 1: 481–84. https://doi.org/10.1038​/s43016-020-0128-4. Springmann, M., L. Spajic, M. A. Clark, J. Poore, A. Herforth, P. Webb, M. Rayner, and P. Scarborough. 2020. “The Healthiness and Sustainability of National and Global Food Based Dietary Guidelines: Modelling Study.” BMJ 2020; 370: m2322. https://doi.org/10.1136/bmj.m2322. Vos, R., W. Martin, and D. Resnick. 2022. “The Political Economy of Reforming Agricultural Support Policies.” IFPRI Discussion Paper 2163. Washington, DC.: International Food Policy Research Institute (IFPRI). https://doi.org​/10.2499/p15738coll2.136545. Willett, W., J. Rockström, B. Loken, M. Springmann, T. Lang, S. Vermeulen, et al. 2019. “Food in the Anthropocene: The EAT–Lancet Commission on Healthy Diets from Sustainable Food Systems.” The Lancet 393 (10170): 447–92. https://doi.org/10.1016/S0140-6736(18)31788-4. World Bank. 2020. Taxes on Sugar-Sweetened Beverages: International Evidence and Experiences. Washington, DC: World Bank. World Bank. 2021. “Uzbekistan: Second Agricultural Public Expenditure Review.” World Bank, Washington, DC. http://hdl.handle.net/10986/36561. World Bank. 2022. Spearheading Vietnam’s Green Agricultural Transformation: Moving to Low-Carbon Rice. Washington, DC: World Bank. http://hdl.handle​.net/10986/38074. World Bank. 2023a. Global SSB Tax Database, August 2023 version. https://ssbtax​.worldbank.org/. World Bank. 2023b. Unpacking the Empirics behind Health Tax Revenue. Global Tax Program Health Tax Project, Knowledge Note 4. Washington, DC: World Bank Group. An Opportunity to Leverage Agrifood Public Support 47 48 Reshaping the Agrifood Sector for Healthier Diets 3 Global-Level Analysis: Agrifood Support and Food Consumption of agricultural policies. The study documents 3.1. Existing research on heterogenous effects of supports and protections agrifood support and food and finds that these measures affect the food consumption affordability for only a subset of countries. This chapter analyzes the cross-country relationship Another related cross-country study is that of between different types of agrifood support and Takeshima (2024), which examines the relationship consumption of different types of food commodities. between sectoral public expenditure and health Cross-country empirical studies such as those of outcomes and finds that public expenditure Magrini et al. (2017) and Miller and Coble (2008) on agriculture reduces poverty, child stunting, document the heterogeneous effects of agrifood overweight, and malnutrition as well as food policies on food affordability. Magrini et al. prices. A recent study by Gilbert et al. (2024) (2017) investigate the comprehensive impacts of provides a value-chain approach to explain how agricultural policies on food security for 64 countries trade policies influence the cost of a healthy diet. over the 1990–2010 period using the nominal rate The study indicates that most food retail items of assistance (NRA). The findings suggest that both in African countries are based on some imports excessive disincentives and incentives for agriculture and emphasizes that food imports need to be may discourage food security, while moderate viewed as inputs in the food value chain. Their support is likely to enhance food availability, access, assessments imply that the share of cost of a and utilization. Miller and Coble (2008) estimate the healthy diet that is attributed to trade measures impact of government agricultural support on the is small (that is, 0.67 to 2.45 percent globally). affordability of food in 10 Organisation for Economic Newer work from Hsiao, Moscona, and Sastry Co-operation and Development (OECD) countries— (2024) shows that governments intervene in Australia, Canada, Iceland, Japan, Mexico, New agriculture markets in response to heat shocks Zealand, Norway, Switzerland, the United States, primarily through border protection measures, and Türkiye—over the 1986–2004 period using the while producer assistance increases with foreign Producer Support Estimate (PSE) and the Consumer production shocks. A body of literature also Nominal Protection Coefficient (CNPC) as measures documents positive links between trade policy and obesity (see box 3.1). Box 3.1 Trade Policy and Obesity Globally integrated economies can be one of between globalization and body weight in the key explanations for global obesity trends. 127 countries from 1980 to 2008. Miljkovic Vogli et al. (2014) examine how economic et al. (2015) develop a theoretical model that globalization and body weight outcome are describes how globalization exacerbates connected and find a positive association obesity in the importing countries under free Global-Level Analysis: Agrifood Support and Food Consumption 49 trade agreements. Costa-Font and Mas (2016) in 31 low- and middle-income countries. Their empirically test whether globalization impacts findings validate the association of trade and obesity using panel data on health outcomes fiscal policies on unhealthy foods and body and various measures of globalization for 26 weight outcomes and indicate heterogeneity countries in the period 1989–2005. They find the in degrees of exposure to these policies across effects of globalization to be an increase of both the level of wealth in households. obesity and caloric intake, both of which are driven by different factors of social globalization, In line with prior research examining the link such as changes in lifestyles. between trade policy and body weight on a global scale, findings from Mexico similarly Reduction in tariffs on unhealthy foods can suggest that a decrease in tariffs could provoke obesity or overweight by reducing contribute to obesity. Giuntella, Rieger, and the relative price of unhealthy foods. Using Rotunno (2020) examine whether Mexico’s the aggregate measure that captures tariff greater exposure to foods and beverages from and nontariff barriers for 9 countries, Cutler, the United States affects obesity through tariff Glaeser, and Shapiro (2003) show that more reductions of the North American Free Trade tariff and nontariff barriers to agriculture Agreement (NAFTA). Using anthropometric data are likely to decrease obesity. Boysen et al. for adult women with a shift-share approach, (2019) analyze how higher import tariffs on they show that increasing US food exports to highly processed foods influence body weight Mexico aggravated the prevalence of obesity outcomes in Sub-Saharan Africa countries. across Mexican states between 1988 and 2012 They find tariff differences between highly and find that most impacts of food imports on processed foods and less processed foods obesity come from unhealthy food imports. increase obesity and mitigate underweight and Using annual food supply data from the the effects are differential in income level of the FAOSTAT, Barlow et al. (2017) analyze the impact country, genders, and regions. They conclude of NAFTA on the supply of high-fructose corn that, while imposing sales taxes can mitigate syrup (HFCS) that experienced tariff removal obesity and aggravate underweight and vice once NAFTA was signed. Their findings show versa, an integrated policy approach is needed that tariff reductions led to more consumption to treat these two issues simultaneously. of caloric sweeteners, including HFCS, in Abay, Ibrahim, and Breisinger (2022) argue Canada. These findings underscore concerns that food trade policies can influence the regarding the potential adverse health food consumption patterns leading the rise of implications of forthcoming trade agreements overweight and obesity, and they study how despite partial offsets from the targeted public tariff rates on unhealthy foods and government health policies. subsidies affect individuals’ health outcomes A recent report from the World Bank (2024) on These disruptions have worsened food insecurity trade and food security highlights the critical and have led to lower diet quality, especially for importance of international trade and integration poorer populations. Inefficient policies like price in international markets for food and nutrition insulation have often amplified these problems by security. Over the past five decades, from 1970 to creating greater market instability. Ensuring global 2023, the global food system has been repeatedly food security requires comprehensive reforms strained by external shocks such as financial in trade policies, climate-resilient agricultural crises, geopolitical conflicts, and climate change. practices, and more efficient logistics systems 50 Reshaping the Agrifood Sector for Healthier Diets to address both price volatility and dietary • Strengthening regional markets and enhancing challenges: agrifood logistics can ensure food security and dietary diversity. In times of global disruption, • Reducing the use of price insulation measures, regional trade agreements and robust agrifood such as export restrictions and tariffs, will logistics systems serve as essential buffers, help governments protect consumers from enabling countries to maintain access to food price shocks while maintaining access diverse food sources and stabilizing local diets. to diverse, nutritious diets. Price insulation, Data from Chile and Colombia illustrates how designed to shield domestic markets from countries that relied on regional imports were global price volatility, frequently achieves the able to mitigate the effects of global price opposite result, increasing domestic price shocks (World Bank 2024). For instance, during swings and restricting access to essential a period of heightened global maize prices, foodstuffs. Only about 50% of global price Chile increased its share of maize imports from changes in key staples, such as rice and wheat, regional partners to nearly 100%, which limited are passed through to domestic markets price hikes. Similarly, Colombia’s reliance on when price insulation policies are in place. regional maize imports rose from 22% to 52%, For instance, countries like India and Japan, allowing the country to avoid more severe which applied heavy insulation policies, saw price increases that would have occurred if significant price volatility in their domestic it had depended solely on global markets. markets despite their efforts to stabilize prices. Regional trade plays an instrumental role in In Japan, insulated markets for rice often smoothing out supply disruptions and helping displayed the same or even higher levels of to maintain a consistent flow of essential food price volatility as global markets. This not items. only affects market stability but also disrupts • Enhancing agrifood logistics, particularly household food consumption patterns, leading shipping routes and transportation to poorer diet quality. Conversely, governments infrastructure, is equally critical in safeguarding that focus on providing targeted safety nets food security. The logistical challenges created instead of relying on generalized subsidies by the COVID-19 pandemic and the the invasion or trade barriers can better stabilize prices have exposed the vulnerability of global food while promoting access to higher-quality diets, supply chains. Rising freight rates, coupled with especially during crises. Case studies from delays at key shipping points like the Suez and Egypt and Morocco show how entrenched food Panama Canals, have further exacerbated food subsidies for staples like bread have imposed price instability. Investing in resilient logistics significant fiscal burdens, making it difficult for networks is important—such as upgrading governments to reform these policies. In Egypt, infrastructure for storage, reducing post- for instance, over 70% of the population relies harvest losses, and using digital technologies on subsidized bread, a system that became to streamline food supply chains. For instance, politically entrenched and unsustainable during using blockchain technology to coordinate global food crises. When global prices rose logistics between smallholder farmers and sharply after Russia’s invasion of Ukraine, Egypt food vendors, as implemented in Kenya, has was forced to increase its already enormous reduced food spoilage and increased efficiency. subsidies, exacerbating the financial strain. By improving logistics and reducing food waste, Shifting toward more adaptable and cost- countries can ensure more stable access to a effective policies, such as targeted safety nets, variety of foods, thereby maintaining both food will help address food security challenges more security and dietary diversity in times of global efficiently, ensuring that vulnerable populations crisis (World Bank 2024). receive timely assistance while maintaining fiscal sustainability (World Bank 2024). Global-Level Analysis: Agrifood Support and Food Consumption 51 • Climate-resilient agricultural practices are The OECD support estimates aim to quantify the essential for securing future food production gross transfers to agriculture that originate both and consumption patterns. Climate change from consumers and producers of agricultural will have a profound impact on agricultural commodities and from taxpayers. These transfers productivity and food systems, necessitating are the result of diverse government support proactive investment in sustainable, climate- policies, which may include direct payments, smart agricultural practices to safeguard future trade barriers, and other forms. It is crucial diets. Artuc et al. (2023) highlights that climate to understand that these indicators primarily change is expected to alter growing conditions measure the level of governmental support effort, in many regions, with some areas benefiting rather than the direct impact of this support. from warmer temperatures and extended The OECD methodology encompasses several growing seasons, while others face devastating indicators. These indicators, collectively, offer declines in productivity. For example, regions a comprehensive view of government support in Sub-Saharan Africa and South Asia are in agriculture, aiding in the understanding and particularly vulnerable to reduced yields analysis of policy impacts on the sector. Figure 3.1 of staple crops like wheat, maize, and rice. displays the sector and commodity levels of this Without significant investment in climate- study’s analysis, based on OECD data for years resilient agricultural infrastructure, poorer between 2020 and 2022 (average, per year). households in these regions will face even more severe food insecurity and reduced access At the agrifood sector level, there are different to nutritious foods. Through the adoption of types of public support. At the sector level, sustainable agricultural technologies—such as agriculture support is comprised of three improved irrigation methods, drought-resistant components that, added together, make up the crop varieties, and soil health management Total Support Estimate (TSE). These three elements practices—countries can mitigate some of the are Producer Support Estimate (PSE), Consumer negative impacts of climate change on food Support Estimate (CSE), and General Services Support production. Real income losses due to climate- Estimate (GSSE). PSE is the largest recipient of related shocks could be as high as 63% in agrifood public support, with 74 percent; CSE the most vulnerable areas (Artuc et al. 2023). is second, with only 13.5 percent; and GSSE is Investments in climate-resilient infrastructure third, with around 12.5 percent of public support will be critical not only to protect local food (OECD 2023). It is important to note that PSE can production but also to ensure that international be negative if government interventions lead to trade in climate-resilient crops continues to negative transfers, essentially taxing producers supply diverse and nutritious foods to regions rather than supporting them. at risk of climate-driven food shortages. Besides the sectoral level, agriculture support 3.2. Different types of agrifood can further be separated into commodity-specific indicators. Both PSE and CSE can be broken public support down into a combination of commodity-related indicators, unlike GSSE, which remains only at the The overall principles that underpin the sectoral level. Producer Single Commodity Transfers analyses proposed here are that not all types (PSCT) and Consumer Single Commodity Transfers of agrifood public support are the same, that (CSCT) are key components of TSE and allow for not all types of food commodities are the commodity-specific analysis. A large portion of same, and that the impacts of a given type PSCT and PSE is Market Price Support (MPS), which of support can vary depending on the type of is a distortive measure that impacts the price of a commodity. Definitions of the above-mentioned food commodity. indicators are provided in box 3.2. 52 Reshaping the Agrifood Sector for Healthier Diets Figure 3.1  Composition of Agrifood Public Support Total Support Estimate (TSE) Sector level US$851 billion Consumer Support General Services Support Producer Support Estimate (PSE) Estimate (CSE) Estimate (GSSE) US$630 billion – 74% US$115 billion – 13.5% US$106 billion – 12.5% Group All Other Sector wide commodity commodities transfers Single commodity transfers transfers transfers transfers to producers Producer support Consumer support Total support to a single commodity Commodity level Consumers Single Producer Single Commodity Commodity Transfers Transfers (PSCT) (CSCT) Price incentives Fiscal support 3a. Market 3b. Non-MPS: 3c. Fiscal subsidies to Price Support fiscal subsidies to (intermediary and final) (MPS) producers consumers US$333 billion – 39% US$297 – 35% Market/trade Budgetary transfers/public expenditure interventions Source: Original figure for this publication based on data from OECD 2023. Note: Producer Support, Consumer Support, Total Support to a Single Commodity, Price Incentives, Fiscal Support, Market/Trade Interventions, and Budgetary Transfers / Public Expenditure are all characteristics of different agrifood public support indicators and are not official OECD terms. 3B and 3C are also not official OECD terms but are a simplified version of the OECD methodology for our analysis and will be introduced below. Box 3.2 Indicator Definitions for Public Support Variables Total Support Estimate (TSE): The annual monetary value of all gross transfers from taxpayers and consumers arising from policies that support agriculture, net of the associated budgetary receipts, regardless of their objectives and impacts on farm production and income, or consumption of farm products. Producer Support Estimate (PSE): The annual monetary value of gross transfers from consumers and taxpayers to agricultural producers, measured at the farm gate level, arising from policies that support agriculture, regardless of their nature, objectives or impacts on farm production or income. Consumer Support Estimate (CSE): The annual monetary value of gross transfers from (to) consumers of agricultural commodities, measured at the farm gate level, arising from policy measures that support agriculture, regardless of their nature, objectives or impacts on consumption of farm products. General Services Support Estimate (GSSE): The annual monetary value of public expenditures aimed at providing public goods and services to the agriculture sector, such as research and development, infrastructure, extension services, and marketing support. This indicator captures the government’s investment in supporting the overall agricultural sector rather than targeting specific commodities or producers. Global-Level Analysis: Agrifood Support and Food Consumption 53 Producer Single Commodity Transfers (PSCT): the annual monetary value of gross transfers from consumers and taxpayers to agricultural producers, measured at the farm gate level, arising from policies linked to the production of a single commodity such that the producer must produce the designated commodity in order to receive the transfer. Consumers Single Commodity Transfers (CSCT): The annual monetary value of gross transfers from (to) consumers of agricultural commodities, measured at the farm gate level, arising from policy measures that support agriculture, regardless of their nature, objectives or impacts on consumption of farm products. Market Price Support (MPS): The annual monetary value of gross transfers from consumers and taxpayers to agricultural producers arising from policy measures that create a gap between domestic market prices and border prices of a specific agricultural commodity, measured at the farm gate level. Usually, expressed as percentage of GFR. This Knowledge Note, similar to existing literature, also refers to MPS’s “distortionary support” or equivalent. Non-MPS Fiscal Supporta (aka Σ BOT, which is aggregate budgetary and other transfers to producers from policies): The annual monetary value of gross transfers from taxpayers to agricultural producers arising from policy measures that do not create a gap between domestic market prices and border prices of a specific agricultural commodity, measured at the farm gate level. Usually, expressed as a percentage of GFR Source: OECD 2018. Note: a. OECD methodology does not include the term non-MPS Fiscal Support. This category was created by taking the OECD’s classification of BOT (budgetary and other transfers). Since Σ BOT includes a lot of different categories and for the analysis, they were combined into one category, which is PSE - MPS = non-MPS Fiscal Support. Public subsidies to producers from national budgets are generally the largest contributor to non-MPS. The definition for Non-MPS Fiscal Support here is the same as MPS, yet the gap between domestic market and border prices is not created. GFR = gross farm receipts. The data used come from the OECD and the Inter- • The OECD collects and reports agriculture American Development Bank (IDB), which contains support data for 54 countries—specifically the information on public support to agricultural 38 OECD countries, 11 emerging economies,2 commodities for all OECD members, other key and 5 non-OECD EU Member States. The agricultural producers, and IDB countries. The OECD allows for an opportunity to compare two sources provide data from 1986 until 2021, agriculture support data across these which is also the data range the analysis follows. countries, as the numbers are also converted This data set is matched with the Global Dietary into US dollars. OECD data are aggregated Database (GDD), which contains information on and used for Agricultural Policy Monitoring and the mean daily intake of different food groups Evaluation reports (OECD, various years); the across 187 countries since 1990 (described below). statistics are available at stats.oecd.org. Depending on the commodity in the analysis, the • The IDB Agrimonitor collects and reports number of countries in the sample changes, since agriculture support data for countries in Latin not all countries in the data sample produce each America and the Caribbean (26 countries);3 it given commodity. For instance, the sample includes also has data for the European Union, Canada, 44 countries that are in both the GDD and the and the United States.4 The Agrimonitor agrifood policy data set based on OECD and IDB’s follows the OECD methodology, separating Agrimonitor.1 There are more observations for the public support into PSE, CSE, and GSSE, which sugar analysis as there is one more year (2020) in makes the comparison and matching between the GDD for sugar. The data allow for a comparative the two sources possible. analysis across different types of agrifood support at both the sector and commodity levels. 54 Reshaping the Agrifood Sector for Healthier Diets 3.3. Different types of food through its detailed stratification, categorizing data by 16 age groups, urban and rural settings, commodities and gender. This expansive database is informed by more than 300 nationally representative The analysis in this note covers three commodities dietary surveys, covering the dietary preferences with large public support: grains, meats, and of approximately 1.75 million individuals. This sugar. The analysis assesses the relationship represents about 89 percent of the global adult between public support to the agrifood sector and population (Muhammad et al. 2017). The GDD’s the consumption of grains, meats, and sugar.5 extensive scope includes 14 food groups, 7 Consumption data come from the GDD, which is beverage categories, 15 macronutrients, 19 based at the Friedman School of Nutrition Science micronutrients, and 2 indexes of carbohydrate and Policy at Tufts University. Further information quality, making it one of the most comprehensive about the commodities in the analysis is provided dietary databases globally. below. • Grains: The GDD reports refined and whole 3.4. Methodology: Cross- grains separately, measured in grams per day. The analysis uses the sum of the two country estimations categories. The cross-country analysis estimates correlations • Meats: The GDD reports processed and of different types of agrifood support with food unprocessed meats separately, measured in consumption of different food items using cross- grams per day. The analysis uses the sum of country panel regressions. The regression analyses the two categories. leverage the panel of 44 countries in the years • Sugar: The GDD reports both sugar-sweetened 1990, 1995, 2000, 2005, 2010, 2015, 2018 (and for beverages (SSB) in grams per day and added sugar, 2020 is also covered) and employ two-way sugars in percentage of total kilocalories per fixed effects to estimate correlations between day. The analysis used added sugars and public support for agrifood commodities and converted percentage of total kilocalories per consumption. The primary specification is: day to grams per day. The data used come from the GDD. Despite extensive nutrition research, the complex effects where yit is the outcome—that is, the consumption of various dietary factors on health remain largely of a food group (in grams per day) for country ambiguous. This gap largely stems from a lack i in year t; Xit is a vector of the agrifood support of systematic and comprehensive efforts to variables; and ut and vi are year- and country- consolidate evidence from controlled interventions specific fixed effects. Equation (1) is estimated by and observational studies. Historically, global each food group separately. nutritional research has predominantly focused on individual micronutrients, neglecting the For the specification of the vector, Xit, two broader spectrum of dietary components including approaches are considered. The first is an estimate macronutrients, foods, and overall dietary patterns of how different types of sectorwide (that is, that might have equally significant or greater non–commodity-specific) agrifood support (TSE, impacts on health. To address this deficiency, the PSE, CSE, and GSSE), and correlate that with GDD has been developed as a pivotal resource. the consumption of each food group. For this The GDD serves as a compendium of dietary approach, five specifications of Xit are considered data, encapsulating 55 dietary factors across (TSE, PSE, CSE, GSSE, and [PSE, CSE, GSSE]). 185 countries (GDD 2018).6 It distinguishes itself The second focuses on the commodity-specific Global-Level Analysis: Agrifood Support and Food Consumption 55 agrifood support. Here, five specifications of Xit The level of support varies across countries. Map (PSCT, MPS, Non-MPS, [Non-MPS, Non-MPS2], 3.1 presents total agrifood support and PSE across and [MPS, Non-MPS, Non-MPS2]) are considered, countries (average of dollar values from 2020 and as well as the specification without and with 2022). China strongly supported both the agrifood controlling for CSCT. sector and its producers, with $310 billion in total support and $271 billion in PSE. India was the second country that strongly supported agrifood, 3.5. Results: Support to the with $124 billion, and the United States was the third, with $122 billion in total support. In terms agrifood sector of PSE, the European Union was the second most supported agricultural producer, with $93 billion, At the agrifood sector level, the level of and the United States was the third, with $44 non-Market Price Support has been growing billion in PSE. Argentina, Australia, Brazil, Canada, quickly over time, while growth in General and Russia show overall low total support and low Services Support Estimate has been slow. level of PSE. But it is precisely General Services Support Estimate that is found to increase overall Most countries have positive PSE and negative CSE. Figure 2.3 presents PSE as a share of gross farm productivity and translate into increased receipts (GFR) and CSE as share of consumption consumption of all food commodities expenditures. Countries with the highest PSE as a assessed. There were no clear impacts of share of GFR include Norway, Iceland, Switzerland, sector wide non-Market Price Support and the Republic of Korea, Japan, and the Philippines. Market Price Support on food consumption. Countries with the highest CSE as a share of However, across all types of sectorwide consumption expenditure are India, the United States, Argentina, and Viet Nam. Among those support, the impact of General Services four, the United States is the only country also Support Estimate is found to be positive with positive PSE. Public support to agriculture in on the consumption of all commodities most countries provides transfers to agricultural assessed. A 10 percent increase in General producers and burdens consumers through taxes Services Support Estimate leads to a 0.14 and other measures. percent increase in consumption of grains, a 0.22 percent increase in the consumption The level of total non-MPS has been the fastest and of meats, and a 0.35 percent increase in the most steadily growing type of support over time, so that it now surpasses the level of total MPS. consumption of sugar. Figure 3.3 presents MPS, non-MPS, GSSE, and CSE over time. The level of non-MPS support has been This section presents descriptives of the measures increasing since 1980s; until recently, GSSE has of total agrifood support (PSE, CSE, and GSSE) also been increasing over time, but much more and estimations of their relationship with the slowly than the increase in non-MPS and has lower consumption of grains, meats, and sugar, as these amount. Therefore, the gap between non-MPS and are the three commodities with the most public GSSE has widened over time while both have been support. Section 3.5 presents descriptives of the increasing steadily. In contrast, despite various measures of commodity-specific support (PSCT, MPS, fluctuations, MPS and CSE have largely stayed at and non-MPS) and estimations of their relationship the same level over time. The largest swings for with the consumption of grains, meats, and sugar, MPS and CSE occurred around 2008, 2011, and respectively. 2021, when MPS dropped and CSE increased. 56 Reshaping the Agrifood Sector for Healthier Diets Map 3.1  Total Support and PSE across Countries a. Total support, 2020–22 b. PSE, 2020–22 Source: Original map for this publication based on data from OECD 2023. Notes: The measure of total support includes Market Price Support (MPS), transfers away from producers (negative MPS), Transfers to Consumers from Taxpayers (TCT), and General Services Support Estimate (GSSE). Producer Support Estimate (PSE) is defined as the gross transfers from consumers and taxpayers to agricultural producers, measured at the farm gate level, arising from policies that support agriculture. Global-Level Analysis: Agrifood Support and Food Consumption 57 Figure 3.2  PSE vs CSE, 2020–22 CSE % 40 IND 20 USA ARG KAZ EU 0 UKR CAV MEX VNM TUR KBR –20 ISR CHN NOR ISL PHL JPN CHE –40 KOR –60 –20 –10 0 10 20 30 40 50 60 PSE % Negative PSE%; Positive CSE: Countries taxing producers Positive PSE%; Positive CSE: Countries taxing producers and consumers Positive PSE%; Negative or zero CSE: Countries taxing consumers Source: Original figure for this publication based on data from OECD 2023. Note: The percentage of Producer Support Estimate (%PSE) measures transfers to producers as a share of gross farm receipts. The percentage of consumer support estimate (%CSE) measures transfers to consumers as a share of consumption expenditure measured at the farm gate. The PSE and CSE are averaged across 54 OECD+ countries, which include 38 OECD countries, 5 non-OECD EU Member States, and 11 emerging economies. Untitled data points represent Australia, Brazil, Costa Rica, Colombia, Chile, and Brazil, Indonesia, New Zealand, Russia, and South Africa. Countries are indicated by their three-letter ISO codes, available at https://www.iban.com/country​ -codes./ CSE = Consumer Support Estimate; EU = European Union; OECD = Organisation for Economic Co-operation and Development; PSE = Producer Support Estimate. Figure 3.3  MPS, Non-MPS, GSSE, and CSE, 1986–2022 Billion USD 300 200 100 0 –100 –200 1985 1990 1995 2000 2005 2010 2015 2020 2025 Distorting support (MPS) Other producer support (Non-MPS) General Services Support Estimate (GSSE) Consumer Support Estimate (CSE) Source: Original figure for this publication based on data from OECD 2023. Note: Distorting support indicates Market Price Support (MPS), defined as the gross transfers to agricultural producers arising from policy measures that distort domestic market prices of a specific commodity. Other producer support indicates Non-MPS Support, defined as the gross transfers to agricultural producers arising from policy measures that do not distort domestic market prices (calculated as PSE – MPS). General Services Support Estimate (GSSE) is defined as the government’s investment to support the overall agrifood sector. Consumer Support Estimate (CSE) is defined as the gross transfers from (to) consumers of agricultural commodities. This covers 54 OECD+ countries, which refers to the 38 OECD countries, the 5 non-OECD EU Member States, and 11 emerging economies. EU = European Union; OECD = Organisation for Economic Co-operation and Development. 58 Reshaping the Agrifood Sector for Healthier Diets At the overall sectorwide level, there are no clear (table C.10, column 5), which translated to a 0.35 impacts of TSE seen on the consumption of grains, percent increase in consumption as a response to meat, and sugar. But there are impacts of GSSE a 10 percent increase in GSSE seen on the total consumption of all commodities assessed—grains, meat, and sugar. This suggests that there are overall impacts of GSSE on 3.6. Results: Support to specific productivity, which translate into consumption agrifood commodities gains across food commodities: An increase in commodity-specific Market • For grains, no statistically significant effects are found for TSE, PSE, CSE, and GSSE when they Price Support decreases the consumption of each enter in regression equations standalone grains and sugar, which are often more easily (table C.6 in appendix C). GSSE becomes traded than meat. A 10 percent increase in significant and positive when PSE and CSE grain-specific Market Price Support reduces are controlled for (table C.6, column 5). The consumption of grains by 0.35 percent, and a estimated results imply that for every increase 10 percent increase in sugar-specific Market of $1 million in GSSE, the consumption of Price Support reduces consumption of sugar grains increased by 0.00258 grams per day (p-value < 0.1) when controlling for PSE and by 0.60 percent. These distortionary Market CSE (table C.6, column 5), which translated to Price Support measures increase domestic a 0.14 percent increase in consumption as a prices and reduce consumption. Hence, policy response to a 10 percent increase in GSSE. reform toward removing market-distorting • For meats, a positive and significant effect of commodity-specific Market Price Support TSE, PSE, and GSSE is found (table C.8, columns must be accompanied by complementary 1, 2, and 4). While estimates are noisy, it measures to curb the consumption of seems to be the case that the positive effect commodities that are less healthy and already of TSE is driven by GSSE (table C.8, column 5). overconsumed. The estimated results imply that $1 million increases in TSE, PSE, and GSSE led to an increase in consumption of meat ranging from Support to grains 0.000206 grams per day to 0.00108 grams per day. In other words, a 10 percent increase The United States, China, and the European in TSE would increase consumption by 0.22 Union are the largest producer of grains, while percent and for PSE and GSSE, those are 0.18 Kazakhstan, Korea, and India are the largest and 0.22 percent increases, respectively. consumers of grains per capita per day (see table • For sugar, no statistically significant effects of 3.1). Top producers such as the United States, TSE, PSE, CSE, and GSSE are found when they Brazil, Russia, Argentina, Ukraine, and Canada are each enter in regression equations standalone not part of top consumers, as many of their grain (table C.10). GSSE becomes significant and productions are for feed or export. Asian countries, positive when controlling for PSE and CSE (table with rice-based diets, are located as top grain- C.10, column 5). The estimated results imply that consuming countries. Only China and Indonesia a $1 million increase in GSSE leads to a 0.000986 are also in the list of top producers. grams per day increase in sugar consumption (p-value < 0.05) when controlling for PSE and CSE Global-Level Analysis: Agrifood Support and Food Consumption 59 Table 3.1  Top Producers and Consumers of Grains Top producers in 2018 Top consumers in 2018 Production Daily Production Daily (million metric consumption (million consumption Rank Country tons) (grams per day) Country metric tons) (grams per day) 1 United States 606.6 108.4 Kazakhstan 19.3 543.9 2 China 548.8 402.6 Republic of Korea 4.1 432.3 3 European Union 277.1 — India 261.1 424.2 4 India 261.1 424.2 Japan 9.4 415.5 5 Brazil 221.5 216.6 China 548.8 402.6 6 Russia 105.3 130.6 Philippines 20.2 369.9 7 Argentina 99.8 185.1 Viet Nam 48.9 348.5 8 Indonesia 81.5 302.1 Indonesia 81.5 302.1 9 Ukraine 68.2 195.5 South Africa 13.8 295.3 10 Canada 65.5 169.2 Chile 2.4 294.4 Source: Original table based on production data from OECD 2023 and consumption data from the Global Nutrition and Policy Consortium 2022. Note: This covers 54 OECD+ countries, which refers to the 38 OECD countries, the 5 non-OECD EU Member States, and 11 emerging economies. Top producers and consumers are ranked in descending order based on the production quantity and consumption in 2018. Data on the daily consumption for the European Union are not available. EU = European Union; OECD = Organisation for Economic Co-operation and Development; — = not available. Of the $91.4 billion in PSCT to grains, close to Mean daily consumption of grains was 245 $70 billion is considered distortionary support grams per day in 1990 (figure 3.5). It dropped to and $22.3 billion is non-distortionary support. The 222 grams in 1995 and steadily increased since monetary value of transfers from consumers to then until 2010 (265 grams). In 2015 and 2018, grain producers is $74.2 billion. Japan, Korea, and the consumption slightly decreased to 261 and Croatia are the countries with the highest PSCT 259 grams. Mean daily consumption of grains for grains as a percentage of gross receipts from remained at or above 260 grams per day between grains, while India, Argentina, and Russia are the 2005 and 2018, increasing slightly from the levels countries with the highest CSCT for grains as a of consumption during 1990–2000. Levels of MPS share of gross receipts from grains. Overall, the and non-MPS were higher in more recent years. PSCT and the CSCT have a negative correlation Levels of MPS were highest in 2015 and 2018, (see figure 3.4). while levels of non-MPS were highest in 2018. MPS has been generally substantially higher than non-MPS. 60 Reshaping the Agrifood Sector for Healthier Diets Figure 3.4  Composition of Support to Grains, and PSCT and CSCT by Country, 2020–22 a. Composition of support Distorting support, US$69.1 billion -(Consumers Single Producer Suport Commodity Single Commodity Transfers Transfers (PSCT), (CSCT)), US$91.4 billion US$74.2 billion Other producer support, US$22.3 billion b. PSCT and CSCT (% of GR) CSCT % 50 IND 25 ARG KAZ 0 RUS VNM UKR NOR –25 CHN PHL JPN –50 –75 KOR –100 CRI –20 –10 0 10 20 30 40 50 60 PSCT % Negative PSCT%; Positive CSCT: Countries taxing producers Positive PSCT%; Positive CSCT: Countries taxing producers and consumers Positive PSCT%; Negative or zero CSCT: Countries taxing consumers Source: Original figure for this publication based on data from OECD 2023. Note: The measure of public support, the Producer Single Commodity Transfers (PSCT), is defined as the total gross transfers from consumers and taxpayers to agricultural producers, measured at the farm gate level, arising from policies linked to the production of a single commodity. The percentage of Producer Single Commodity Transfers (%PSCT) is defined as the single commodity transfers as a share of gross farm receipts for the specific commodity. The percentage of Consumer Single Commodity Transfers (%CSCT) is defined as transfers from (to) consumers of a single commodity as a share of gross farm receipts (GR) for the specific commodity. Panel A covers 54 OECD+ countries, which refers to the 38 OECD countries, the 5 non-OECD EU Member States, and 11 emerging economies. Panel B covers OECD+ countries except Chile, New Zealand, Iceland, and Australia, where PSCT and CSCT are zero. Untitled data points represent Israel, Brazil, United Kingdom, Mexico, United States, Türkiye, the European Union, Switzerland, Colombia, South Africa, Indonesia, and Canada. Countries are indicated by their three-letter ISO codes, available at https://www.iban.com/country-codes./ EU = European Union; OECD = Organisation for Economic Co-operation and Development. Global-Level Analysis: Agrifood Support and Food Consumption 61 Figure 3.5  Grains: Daily Intake, MPS, and non-MPS, 1990–2018 40 150 270 264.7 259.2 261.0 258.9 Mean daily intake (grams per day) 35 120 250 245.1 245.5 Non-MPS (US$, billions) 30 90 MPS (US$, billions) 25 230 60 222.1 20 30 15 210 0 10 190 5 –30 0 –60 170 1985 1990 1995 2000 2005 2010 2015 2020 Mean daily intake of grains (grams per day) MPS Non-MPS Source: Original figure for this publication based on data from the Global Nutrition and Policy Consortium 2022; IDB 2023; OECD 2023. Note: Market Price Support (MPS) is defined as the gross transfers to agricultural producers arising from policy measures that distort domestic market prices of a specific commodity. Non-MPS is defined as the gross transfers to agricultural producers arising from policy measures that do not distort domestic market prices. This covers 44 countries in the samples for the analysis, which refers to the 27 OECD+ countries and the 17 countries from Latin America and the Caribbean. The consumption, MPS, and non-MPS amounts are averages across countries by year. The data are not balanced, so not all countries possess complete information on the gross transfers for the entire period. OECD = Organisation for Economic Co-operation and Development. Commodity-specific support to grains in the United States, India, Mexico, Australia, and Canada form of MPS reduces grain consumption, but are top producers that are not in the list of the top the effects are quite small. Commodity-specific consumers. Most of the top consumers, except single-commodity support for producers (PSCT China and Brazil, are not top producers, indicating and MPS) has negative and statistically significant that they are sourcing their meat consumption via effects (p-value < 0.01) on consumption when import. controlling for CSCT (table C.7, columns 6 and 7 in appendix C). Estimations indicate that a 10 percent Of the $89 billion in PSCT to meats, $82 billion increase in the level of PSCT leads to a 0.34 percent is considered distortionary support. About $113 decrease in consumption of grains. Similarly, a 10 billion in consumer transfers to meat producers percent increase in MPS leads to a 0.35 percent is estimated to have occurred between 2020 and decrease in consumption of grains. 2022. Switzerland, Iceland, Norway, Korea, and Japan are the countries with the highest PSCT Support to meats for meats as a percentage of gross receipts from meats, while Viet Nam and Argentina are the The United States, China, and the European Union countries with the highest CSCT for meats as a are the largest producers of meats, while Russia, share of gross receipts from meats. Similar to the Israel, and Colombia are the largest consumers case of grains, the negative correlation between of meats per capita per day (see table 3.2). The CSCT and PSCT is evident (see figure 3.6). 62 Reshaping the Agrifood Sector for Healthier Diets Table 3.2  Top Producers and Consumers of Meats Top producers in 2018 Top consumers in 2018 Production Daily Production Daily consumption (million consumption Rank Country (million metric tons) (grams per day) Country metric tons) (grams per day) 1 China 85.2 115.7 Russia 9.3 240.3 2 European Union 48.5 — Israel 0.5 221.5 3 United States 45.8 58.6 Colombia 2.8 175.3 4 Brazil 26.5 114.3 South Africa 3.2 164.6 5 India 9.7 6.3 Kazakhstan 0.9 142.1 6 Russia 9.3 240.3 Costa Rica 0.3 122.2 7 Mexico 6.8 65.5 China 85.2 115.7 8 Australia 6.0 74.0 Brazil 26.5 114.3 9 Argentina 5.8 112.1 Norway 0.4 112.6 10 Canada 5.6 68.2 Argentina 5.8 112.1 Source: Original table for this publication based on production data from OECD 2023 and consumption data from the Global Nutrition and Policy Consortium 2022. Note: This covers 54 OECD+ countries, which refers to the 38 OECD countries, the 5 non-OECD EU Member States, and 11 emerging economies. Top producers and consumers are ranked in descending order based on their production quantity and consumption in 2018. Data on the daily consumption for the European Union are not available. EU = European Union; OECD = Organisation for Economic Co-operation and Development; — = not available. Mean daily consumption of meats was 91 grams Support to sugar per capita per day in 1990 and declined to 83 grams in 2000, but it has been increasing over time Brazil, Mexico, and India are the largest producers to reach 102 grams per capita per day in 2018, of sugar, while Iceland, India, and Korea are the surpassing the levels of consumption in 1990 (see largest consumers of sugar per capita per day (see figure 3.7). Levels of MPS for meat have increased table 3.3). Mexico, India, Russia, and the United substantially after 2005, while levels of non-MPS States are the only countries that are in both top for meat have declined significantly after 2010 and producers and top consumers lists. South Africa, have remained negative after 2015. The increase in China, Australia, and Türkiye are top producers MPS from 2005 to 2010 coincides with the increase that do not consume sugar heavily. in consumption (from 87 to 96 grams between 2005 and 2010). Again, similar to grains, non-MPS Commodity-specific support to meats has no clear is overall substantially lower than MPS. impact on meat consumption. Consistent effects of commodity-specific single commodity support for Commodity-specific support to meats has no clear producers are not found as the signs and statistical impact on meat consumption. Consistent effects of significances are noisy or mixed across different commodity-specific single commodity support for specifications (table C.9 in appendix C). Positive producers are not found as the signs and statistical and significant effects were found for PSCT and significances are noisy or mixed across different MPS when other support variables were not specifications (table C.9 in appendix C). Positive controlled for. and significant effects were found for PSCT and MPS when other support variables were not controlled for. Global-Level Analysis: Agrifood Support and Food Consumption 63 Figure 3.6  Composition of Support to Meats and PSCT and CSCT by Country, 2020–22 a. Composition of support Distorting support, US$82.3 billion -(Consumers Single Producer Suport Commodity Single Commodity Transfers Transfers (PSCT), (CSCT)), US$89.4 billion US$113.2 billion Other producer support, US$7.1 billion CSCT % b. PSCT and CSCT (% of GR) 20 VNM 10 ARG KAZ 0 IND EU CRI BRA –10 TUR ISL –20 UDN CHN GBR RUS ISR NOR –30 PHL –40 KOR –50 CHE –60 JPN –20 –10 0 10 20 30 40 50 60 PSCT % Negative PSCT%; Positive CSCT: Countries taxing producers Negative PSCT%; Negative CSCT: Positive PSCT%; Negative or zero CSCT: Countries taxing consumers Source: Original figure for this publication based on data from OECD 2023. Notes: The measure of public support—the Producer Single Commodity Transfers (PSCT)—is defined as the total gross transfers from consumers and taxpayers to agricultural producers, measured at the farm gate level, arising from policies linked to the production of a single commodity. The percentage of Producer Single Commodity Transfers (%PSCT) is defined as the single commodity transfers as a share of gross farm receipts for the specific commodity. The percentage of Consumer Single Commodity Transfers (%CSCT) is defined as transfers from (to) consumers of a single commodity as a share of gross farm receipts for the specific commodity. Panel A covers 54 OECD+ countries, which refers to the 38 OECD countries, the 5 non-OECD EU Member States, and 11 emerging economies. Panel B covers OECD+ countries except Australia and South Africa, where PSCT and CSCT are zero. Untitled data points represent Mexico, United States, Chile, Ukraine, New Zealand, Colombia, and Canada. Countries are indicated by their three-letter ISO codes, available at https://www.iban.com​ /country-codes./ EU = European Union; OECD = Organisation for Economic Co-operation and Development. Support to sugar table 3.3). Mexico, India, Russia, and the United States are the only countries that are in both top Brazil, Mexico, and India are the largest producers producers and top consumers lists. South Africa, of sugar, while Iceland, India, and Korea are the China, Australia, and Türkiye are top producers largest consumers of sugar per capita per day (see that do not consume sugar heavily. 64 Reshaping the Agrifood Sector for Healthier Diets Figure 3.7  Meats: Daily Intake, MPS, and non-MPS, 1990–2018 5 80 99.2 102.5 95.6 Mean daily intake (grams per day) 100 90.9 70 85.6 87.0 4 82.6 Non-MPS (US$, billions) 60 80 MPS (US$, billions) 3 50 60 40 2 30 40 1 20 20 10 0 0 0 1985 1990 1995 2000 2005 2010 2015 2020 Mean daily intake of meat (grams per day) MPS Non-MPS Source: Original figure for this publication based on data from the Global Nutrition and Policy Consortium 2022; IDB 2023; OECD 2023. Note: Market Price Support (MPS) is defined as the gross transfers to agricultural producers arising from policy measures that distort domestic market prices of a specific commodity. Non-MPS is defined as the gross transfers to agricultural producers arising from policy measures that do not distort domestic market prices. This covers 44 countries in the samples for the analysis, which refers to the 27 OECD+ countries and the 17 countries from Latin America and the Caribbean. The consumption, MPS, and non-MPS are averages across countries by year. The data are not balanced, so not all countries possess complete information on the gross transfers for the entire period. OECD = Organisation for Economic Co-operation and Development. Table 3.3  Top Producers and Consumers of Sugar Top producers in 2020 Top consumers in 2020 Daily Production Daily Production consumption (million consumption Rank Country (million metric tons) (grams per day) Country metric tons) (grams per day) 1 Brazil 654.5 62.1 Iceland — 146.5 2 Mexico 53.5 81.7 India 31.1 115.3 3 India 31.1 115.3 Korea — 93.0 4 South Africa 19.3 36.8 Mexico 53.5 81.7 5 European Union 15.5 — Chile 0.2 77.0 6 China 14.4 3.0 Russia 5.8 74.4 7 United States 7.1 67.9 New Zealand 5.2 73.1 8 Russia 5.8 74.4 Colombia 2.2 70.5 9 Australia 4.2 46.6 Costa Rica 0.4 69.6 10 Türkiye 3.0 31.9 United States 7.1 67.9 Source: Original table for this publication based on production data from OECD 2023 and consumption data from Global Nutrition and Policy Consortium 2022. Notes: This covers 54 OECD+ countries, which refers to the 38 OECD countries, the 5 non-OECD EU Member States, and 11 emerging economies. Top producers and consumers are ranked in descending order based on their production quantity and consumption in 2020. Data on the daily consumption for the European Union and data on the production for Iceland and Korea are not available. EU = European Union; OECD = Organisation for Economic Co-operation and Development; — = not available. Global-Level Analysis: Agrifood Support and Food Consumption 65 Of the $15 billion in PSCT to sugar, $14.4 billion the countries with the highest PSCT for sugar as is considered distortionary support. About $16 a percentage of gross receipts from sugar, while billion in consumer transfers to sugar producers Brazil, Türkiye, Chile, Colombia, and the European is estimated between 2020 and 2022. While PSCT Union have the lowest. A noticeable difference to sugar is smaller than that for grains and meat from the earlier two commodities is that there is as noted above, PSCT as a share of GFR is highest no country with a positive CSCT or negative PSCT. for sugar (see chapter 2). The Philippines, Ukraine, Overall, the correlation between PSCT and CSCT China, Japan, the United States, and Indonesia are appears to be negative (see figure 3.8). Figure 3.8  Composition of Support to Sugar, and PSCT and CSCT by Country, 2020–22 a. Composition of support Distorting support, US$14.4 billion -(Consumers Single Producer Suport Commodity Single Commodity Transfers Transfers (PSCT), (CSCT)), US$15.0 billion US$16.2 billion Other producer support, US$1.2 billion CSCT % b. PSCT and CSCT (% of GR) BRA THR EU CHE 0 CRI COL RUS –25 IND ZAR MEX CHL VNM PHL –50 CHN UKR GBR –75 USA IDN –100 –125 JPN –150 0 10 20 30 40 50 60 70 80 PSCT % Positive PSCT%; Negative or zero CSCT: Countries taxing consumers Source: Original figure for this publication based on data from OECD 2023. Notes: The measure of public support—the Producer Single Commodity Transfers (PSCT)—is defined as the total gross transfers from consumers and taxpayers to agricultural producers, measured at the farm gate level, arising from policies linked to the production of a single commodity. The percentage of Producer Single Commodity Transfers (%PSCT) is defined as the single commodity transfers as a share of gross farm receipts for the specific commodity. The percentage of Consumer Single Commodity Transfers (%CSCT) is defined as transfers from (to) consumers of a single commodity as a share of gross farm receipts for the specific commodity. Panel A covers 54 OECD+ countries, which refers to the 38 OECD countries, the 5 non-OECD EU Member States, and 11 emerging economies. Panel B covers OECD+ countries except Korea, Israel, New Zealand, Iceland, Norway, Argentina, Australia, Kazakhstan, and Canada, where PSCT and CSCT are zero. OECD = Organisation for Economic Co-operation and Development. 66 Reshaping the Agrifood Sector for Healthier Diets Mean daily consumption of sugar has been around effects (p-value < 0.1) on consumption when 60 grams per capita per day in the period 2015–20 controlling for CSCT (table C.11, columns 6 and 7). (see figure 3.9). The consumption decreased to 54 Estimations indicate that a 10 percent increase in grams in 2010 from a peak of 60 grams in 2000. the level of PSCT leads to a 0.62 percent decrease The increase in consumption from 2010 to 2015 in consumption of sugar. Similarly, a 10 percent coincides with the increase in the levels of MPS increase in MPS leads to a 0.60 percent decrease for sugar, with nearly $17 billion of distortionary in consumption of sugar. The magnitudes of the support in 2018. Levels of non-MPS for sugar have estimates are unsurprising given that the PSCT for generally followed an upward trend from 1990 sugar mostly consists of MPS. to 2020, except during 2015–18, which were the periods with the lowest non-MPS for sugar. Similar Using the existing cross-country database to the two earlier commodities, the levels of non- developed for this analysis, there are opportunities MPS are generally lower than those of MPS. to revisit the cross-country regression specifications, such as by looking at cross-elasticities. Further Commodity-specific support to sugar in the analytical work could also expand on the range form of MPS reduces sugar consumption, but of commodities assessed (for example, including the effects are quite small. Commodity-specific fruits, vegetables, and milk), and by looking at single commodity support for producers (PSCT subcategories of broader commodity groups (for and MPS) has negative and statistically significant example, different types of grains and meats). Figure 3.9  Sugar: Daily intake, MPS, and non-MPS, 1990–2020 59.9 61.5 59.5 59.4 2.0 20 60 56.5 56.9 56.0 53.7 Mean daily intake (grams per day) 50 Non-MPS (US$, billions) 1.5 15 MPS (US$, billions) 1.0 40 10 0.5 30 5 0 20 0 10 –0.5 –1.0 –5 0 1985 1990 1995 2000 2005 2010 2015 2020 2025 Mean daily intake of sugar (grams per day) MPS Non-MPS Source: Original figure for this publication based on data from the Global Nutrition and Policy Consortium 2022; IDB 2023; OECD 2023. Notes: Market Price Support (MPS) is defined as the gross transfers to agricultural producers arising from policy measures that distort domestic market prices of a specific commodity. Non-MPS is defined as the gross transfers to agricultural producers arising from policy measures that do not distort domestic market prices. This covers 44 countries in the samples for the analysis, which refers to the 27 OECD+ countries and the 17 countries from Latin America and the Caribbean. The consumption, MPS, and non-MPS are averages across countries by year. The data are not balanced, so not all countries possess complete information on the gross transfers for the entire period. OECD = Organisation for Economic Co-operation and Development. Global-Level Analysis: Agrifood Support and Food Consumption 67 Notes 1. The 44 countries in the analysis include Argentina, Australia, Barbados, Belize, Bolivia, Brazil, Canada, Chile, China, Colombia, Costa Rica, the Dominican Republic, Ecuador, El Salvador, the European Union, Guatemala, Guyana, Haiti, Honduras, India, Indonesia, Israel, Jamaica, Japan, Kazakhstan, Mexico, New Zealand, Nicaragua, Norway, Panama, Paraguay, Peru, Philippines, Russian Federation, South Africa, Suriname, Switzerland, Trinidad and Tobago, Türkiye, Ukraine, the United Kingdom, the United States, Uruguay, and Viet Nam. 2. Argentina, Brazil, China, India, Indonesia, Kazakhstan, the Philippines, the Russian Federation, South Africa, Ukraine, and Viet Nam. 3. Argentina, The Bahamas, Barbados, Belize, Bolivia, Brazil, Chile, Colombia, Costa Rica, Dominican Republic, Ecuador, El Salvador, Guatemala, Guyana, Haiti, Honduras, Jamaica, Mexico, Nicaragua, Panama, Paraguay, Peru, Suriname, Trinidad and Tobago, Uruguay, and Venezuela. 4. Data can be downloaded at https://mydata.iadb.org​ /Indicator-Catalog/IDB-Agrimonitor-PSE-Agricultural​ -Policy​-Monitoring/2dqw-u35p/about_data. 5. Other commodities that are accessible in the final database include fruits, vegetables, and dairy/milk. 6. The GDD 2018 is available at https://globaldietarydatabase​.org/available-gdd-2018-estimates-datafiles. References Abay, K. A., H. Ibrahim, and C. Breisinger. 2022. “Food Policies and Obesity in Low- and Middle-Income Countries.” World Development 151: 105775. https://doi​.org/10.1016/j.worlddev.2021.105775. Artuc, E., G. Porto, and B. Rijkers. 2023. “Crops, Conflict and Climate Change.” World Bank Mimeo. https://www.colorado.edu/economics/sites/default/files​ /attached-files/artuc.pdf. Barlow, P., M. McKee, S. Basu, and D. Stuckler. 2017. “Impact of the North American Free Trade Agreement on High-Fructose Corn Syrup Supply in Canada: A Natural Experiment Using Synthetic Control Methods.” Canadian Medical Association Journal 189 (26): E881–E887. https://doi.org/10.1503/cmaj.161152. Boysen, O., K. Boysen-Urban, H. Bradford, and J. Balié. 2019. “Taxing Highly Processed Foods: What Could Be the Impacts on Obesity and Underweight in Sub-Saharan Africa?” World Development 119: 55–67. https://doi.org​/10.1016/j.worlddev.2019.03.006. Costa-Font, J. and N. Mas. 2016. “ ‘Globesity’? The Effects of Globalization on Obesity and Caloric Intake.” Food Policy 64: 121–32. https://doi.org/10.1016/j.foodpol​.2016.10.001. Cutler, D. M., E. L. Glaeser, and J. M. Shapiro. 2003. “Why Have Americans Become More Obese?” Journal of Economic Perspectives 17 (3): 93–118. https://doi.org​/10.1257/089533003769204371. Gilbert, R., L. Costlow, J. Matteson, J. Rauschendorfer, E. Krivonos, S. A. Block, and W. A. Masters. 2024. “Trade Policy Reform, Retail Food Prices and Access to Healthy Diets Worldwide.” World Development 177: 106535. https://doi​.org/10.1016/j.worlddev.2024.106535. 68 Reshaping the Agrifood Sector for Healthier Diets Giuntella, O., M. Rieger, and L. Rotunno. 2020. “Weight Gains from Trade in Foods: Evidence from Mexico.” Journal of International Economics 122: 103277. https://doi.org/10.1016/j.jinteco.2019.103277. Global Nutrition and Policy Consortium. 2022. Global Dietary Database. https://globaldietarydatabase.org/. Hsiao, A., J. Moscona, and K. Sastry. 2024. “Food Policy in a Warming World.” National Bureau of Economic Research Working Paper 32539. https://www​.nber.org/papers/w32539 IDB (Inter-American Development Bank). 2023. Agrimonitor Database. https://agrimonitor.iadb.org/en/. Magrini, E., P. Montalbano, S. Nenci, and L. Salvatici. 2017. “Agricultural (Dis) Incentives and Food Security: Is There a Link?” American Journal of Agricultural Economics 99 (4): 847–71. https://doi.org/10.1093/ajae/aaw103. Miljkovic, D., S. Shaik, S. Miranda, N. Barabanov, and A. Liogier. 2015. “Globalisation and Obesity.” The World Economy 38 (8): 1278–94. https://doi.org/10.1111/twec.12260. Miller, J. C. and K. H. Coble. 2008. “An International Comparison of the Effects of Government Agricultural Support on Food Budget Shares.” Journal of Agricultural and Applied Economics 40 (2): 551–58. https://doi.org/10.1017​/S107407080002383X. Muhammad, A., A. D’Souza, B. Meade, R. Micha, and D. Mozaffarian. 2017. “How Income and Food Prices Influence Global Dietary Intakes by Age and Sex: Evidence from 164 Countries.” BMJ Global Health 2: e000184. https://gh.bmj.com/content/2/3/e000184. OECD (Organisation for Economic Co-operation and Development). 2018. Agricultural Policy Monitoring and Evaluation. Paris: OECD. https://doi.org/10.1787/agr​_pol-2018-en. OECD (Organisation for Economic Co-operation and Development). 2023. Agriculture Statistics. Database. https://doi.org/10.1787/agr-data-en. OECD (Organisation for Economic Co-operation and Development). Various years. Agricultural Policy Monitoring and Evaluation. Annual ISSN: 22217371 (online) https://doi.org/10.1787/22217371. Takeshima, H. 2024. “Public Expenditure’s Role in Reducing Poverty and Improving Food and Nutrition Security: Cross-Country Evidence from SPEED Data.” The European Journal of Development Research 1–29. https://doi.org/10.1057/s41287-023-00623-8. Vogli, R. D., A. Kouvonen, M. Elovainio, and M. Marmot. 2014. “Economic Globalization, Inequality and Body Mass Index: A Cross-National Analysis of 127 Countries.” Critical Public Health 24 (1): 7–21. https://doi.org/10.1080/09581596.2013.768331. World Bank. 2024. Trade Policy and Food Security in an Era of Climate Change. Washington, DC: World Bank. http://documents.worldbank.org/curated/en​ /099061424110025183/P18103613dfff3011a9d9184325f1f9fc1. Global-Level Analysis: Agrifood Support and Food Consumption 69 70 Reshaping the Agrifood Sector for Healthier Diets 4 Country-Level Cases: Comparing Different Interventions Chapter 3 assessed the relationship between 4.1. Input subsidies, roads, different types of agrifood support and the consumption of different types of commodities. and cash transfers in However, while there are correlations between Bangladesh Market Price Support policies and consumption, there are no clear correlations between agrifood In the case of Bangladesh, contrary to the subsidies and consumption, on average. This suggests that country context varies significantly, impact of social protection programs and and that the impact of various policy and input subsidies, there is a positive impact intervention options are largely country specific. of rural infrastructure development, an This chapter presents two country cases: one for important public good, on various measures Bangladesh and one for Malawi, both of which are of healthy diets. receiving support from the FoodSystems2030 Trust Fund to repurpose agrifood support for better Public investment in the agriculture sector in outcomes.1 The goal of these case studies is to Bangladesh has mainly been through input look at the country-level impacts of various types subsidies, which constitute around 43 percent of agrifood sector support (that is, input subsidies, of the country’s Ministry of Agriculture budget. irrigation, infrastructure, and social protection) on The major component of the input subsidy measures of food security and healthy diets. in Bangladesh is the fertilizer subsidy, which is based on the twin policy objectives of the For the Bangladesh case study, with a grant of government: keeping the food grain price low $18 million and $500 million leveraged,2 farmers enough to make it affordable for the poorer will receive direct support through an innovative section of the society and keeping the price inputs subsidy pilot utilizing an e-voucher system high enough to incentivize the farmers to to experiment with various options, with the produce crops. Although the existing studies aim to eventually inform gradual repurposing show a significant impact of input subsidies on of the current Fertilizer Subsidy Program of enhancing the production and income of farming over $2 billion and expand coverage nationally. households, little is known about its impact In Malawi, there is a grant of $20 million with on food and nutrition security in Bangladesh $250 million leveraged,3 which will allow farmers (Nasrin, Bauer, and Arman 2018; Sarma and to gain more technical support for efficient Hossian 2020; Sarma and Rahman 2020; Uddin usage of fertilizer subsidies and are provided and Dhar 2018). However, alternative agriculture incentives to diversify crop production, improve interventions—such as irrigation infrastructure/ soil health restoration practices, and assist with community infrastructure as well as research paying for ecosystems services. The aim is for and agricultural education—are also important the government to repurpose the existing inputs to facilitate positive outcomes and complement program by testing other options and scaling up input subsidies. Improved rural infrastructure such the most promising interventions. as roads can create opportunities for economic growth and poverty reduction through a range of mechanisms (Khandker, Bakht, and Koolwal 2009). Similarly, social cash transfers targeting the ultra-poor or vulnerable groups in Bangladesh Country-Level Cases: Comparing Different Interventions 71 could complement input subsidies, hence calling more importantly, the same unseen traits (more for harmonization of the different assistance essential because they are more difficult to control programs from the government. However, existing for) when a randomized design evaluation is studies show a mixed impact of different social done effectively. This eliminates selectivity issues, protection programs on the food and nutrition establishes a reliable foundation for comparison, security of Bangladeshi households (Ahmed et al. and establishes the direction of causality. A 2009; Badhan et al. 2019; Mamun 2019; Rahman randomized design also has the benefit of making 2012). In this context, this Bangladesh case study program impact simple to calculate, which makes evaluates the relative effectiveness of different it easier to comprehend and discuss. interventions for rural farming households to ensure their food and nutrition security. However, because the three different types of programs (that is, input subsidy, social Data and methodology protection, and rural agricultural infrastructure) in Bangladesh had already been put into action This country case study utilizes the Bangladesh prior to the evaluation, a randomized approach Integrated Household Survey (BIHS) 2018-2019 was not practical for the assessment. As a result, developed by the International Food Policy a nonrandomized approach was used for impact Research Institute (IFPRI). This was the third- assessment in this work. round survey IFPRI collected, after the first round in 2011–12 and the second round in 2015. It is a In this study, the propensity score matching nationally representative survey that collects data (PSM) estimator is used to estimate the effect of at both household and community levels. It mainly different programs on the household spending collects data on the rural areas of Bangladesh, and behavior. The PSM method consists of generating thus it is representative of the agricultural sector a single propensity score based on the observable of rural Bangladesh. Here it needs to be noted characteristics to construct a counterfactual that, to make the result of different interventions group. In this method, each treated observation comparable, this study restricts the sample to only is matched to several control observations based farming households. on the propensity score, which would help to construct a counterfactual to examine what would It is essential to create a counterfactual happen to the food and nutrition security of assessment of what may have happened in the program beneficiary household if it did not receive absence of the program in order to compare the benefits of the program. Thus, this method beneficiary results to what those outcomes would ensures the robustness of the effect of treatment have been had the program not been introduced. variable on the outcome variable. This is how program impact is measured. The most effective method for creating a legitimate In order to construct the propensity score, a counterfactual is to choose recipients at random participation equation needs to be estimated from a group of candidates who are all equally where the dependent variable will be the deserving. Every person (or community, school, treatment variables. In this case study, the and so on) has an equal chance of getting chosen treatment variables for the three different for the program if the assignment is made at interventions are different. For example, for the random. The average results for individuals who input subsidy the treatment variables will be 1 for were not chosen at random ought to offer a fair those receiving the subsidy and 0 for those not assessment of what the beneficiaries would have receiving the subsidy. For the social protection gone through in the absence of the program. program, the treatment variable will be 1 for those enjoying the benefit of the program and Beneficiaries and nonbeneficiaries will, on 0 for those not enjoying the benefit. Because average, have the same observed features and, the treatment variable is a dichotomous or 72 Reshaping the Agrifood Sector for Healthier Diets binary variable, the propensity score is usually To evaluate the relative effect of different rural calculated using the probit model. The propensity support schemes in Bangladesh on food and score, which can be expressed as P (D = 1|X), nutrition security, five different indicators of represents the probability of enjoying the benefit food security based on the BIHS 2018-2019 of a particular intervention based on observed were constructed. The five indicators include the different characteristics of the households and household per capita food expenditure, household communities. The basic descriptive statistics of dietary diversity score (HDDS), Household Hunger the variables used for estimating the participation Scale (score), the Food Consumption Score (FCS), equation are presented in table 4.1. and the Food Insecurity Experience Scale (FIES) (table 4.2). Table 4.1  Basic Descriptive Statistics of Selected Variables Standard Variable Observations Mean deviation Minimum Maximum Social protection program beneficiary 2,675 0.059 0.236 0 1 Beneficiary of improved infrastructure 2,675 0.597 0.491 0 1 Beneficiary of input subsidy 2,675 0.033 0.177 0 1 Household wealth index 2,675 0.0912 0.2879 0 1 Household total land (hectares) 2,674 118.49 111.065 7 1,311.5 Primary level of education 2,675 0.267 0.442 0 1 Secondary level of education 2,675 0.227 0.419 0 1 Tertiary level of education 2,675 0.043 0.204 0 1 Household size 2,675 5.946 2.342 1 23 Proportion of household members (age < 15) 2,675 0.238 0.186 0 0.778 Proportion of household members (age 15–64) 2,675 0.509 0.207 0 1 Proportion of household members (age ≥ 65) 2,675 0.254 0.228 0 1 Community has a government clinic 2,675 0.326 0.469 0 1 Number of educational institutions in the 2,675 2.884 1.929 0 10 community Community has a government primary school 2,675 0.769 0.422 0 1 Barisal (division) 2,675 0.068 0.252 0 1 Rangpur (division) 2,675 0.136 0.343 0 1 Sylhet (division) 2,675 0.122 0.327 0 1 Chattogram (division) 2,675 0.12 0.325 0 1 Rajshahi (division) 2,675 0.131 0.338 0 1 Khulna (division) 2,675 0.116 0.32 0 1 Source: Original table for this publication based on data from BIHS 2018-2019 (IFPRI 2020). Note: BIHS = Bangladesh Integrated Household Survey. Country-Level Cases: Comparing Different Interventions 73 Table 4.2  Basic Descriptive Statistics of Different Food Security Variables Standard Food security variable Observations Mean deviation Minimum Maximum Per capita food expenditure (BDT) 2,664 403.45 337.77 5.36 3,107.14 Food consumption score 2,675 71.113 17.05 25.5 112 Household dietary diversity score (HDDS) 2,675 10.36 1.314 6 12 Food Insecurity Experience Scale (FIES) score 2,675 1.327 1.813 0 8 Household Hunger Scale (score) 2,675 0.051 0.293 0 5 Source: Original table for this publication based on data from BIHS 2018-2019 (IFPRI 2020). Note: BDT = Bangladeshi taka (currency); BIHS = Bangladesh Integrated Household Survey. The per capita food expenditure is an economic the benefits of any social protection programs are gauge that reflects whether households spend considered control groups. However, there can adequately on food and dietary intake. Food be major problems associated with using such consumption expenditure is calculated based on wide-ranging treatment and control groups. The money spent on food items. Total food-related problem with using the beneficiaries of all the consumption expenditure per month was then programs as the treatment group is that each of divided by the number of household members the programs has different target groups (mainly to calculate per capita food consumption poor) based on some selection criteria; this makes expenditure. Dietary diversity is measured as the it difficult to find control groups with similar total food groups consumed in the last 7 days characteristics. Moreover, the data used in this as per the BIHS. The Food Consumption Score case study have not been designed to conduct can be measured as weighted frequency of the this kind of outcome research and, therefore, number of food groups consumed in the previous they contain all kinds of households (that is, 7 days. The household hunger score is calculated rich and poor). In this context, including all the based on three questions related to availability households not receiving the benefits of social of food and hunger of the household members. protection program might lead to a biased result Finally, the FIES consists of a total of eight binary as the treatment and control groups would not be questions pertaining to the household’s behaviors comparable. and experiences that are associated with the accessibility of the food. A score ranging between Against this backdrop, this case study first 0 and 8 is developed based on the responses of considers the beneficiaries of a particular social these eight questions. protection program as the treatment group. Second, to make the treatment and control groups Social protection and food and comparable, it considers only the households that nutrition security belong to the poorest groups where the poorest is defined as the households belonging the first In evaluating the effect of social protection quintile of household total expenditure. Finally, programs on food and nutrition security, the to make the results comparable with the impact first step is to identify the treatment and control of other interventions such as input subsidy groups. In this regard, the usual practice is to and rural agricultural infrastructure, this case consider the households receiving the benefits study considers only farming households while of at least one social protection program as the evaluating the effect of social protection program treatment group while households not receiving on food and nutrition security. 74 Reshaping the Agrifood Sector for Healthier Diets The social protection program considered in this 65 years or older in the households. Based on study is the Old Age Allowance (OAA), which is the propensity score, the treatment and control a cash transfer program. The OAA program was groups have been matched with different introduced in 1998 to provide a monthly cash matching techniques. payment to older people to help reduce their vulnerabilities and income insecurity. Currently, The result of the treatment effect estimated based it is one of the largest social protection programs on PSM shows that social protection programs in Bangladesh. The program is not universal: have a positive impact on the household per beneficiaries are prioritized on the basis of age capita food expenditure (table 4.4). The result is and socioeconomic status, so that the poorest are consistent across different matching techniques. selected to become recipients. For example, the ultra-poor households receiving OAA have, on average, 70–75 more BDT per The average value of different food capita for food expenditure than those who do security indicators for the beneficiaries and not receive the benefits. However, for another nonbeneficiaries of different OAA is presented in two indicators of food and nutrition security— table 4.3. Here the ultra-poor households receiving the food consumption score and the household the benefits of OAA have a higher per capita dietary diversity score—the effect is negative food expenditure than those not receiving the and statistically insignificant. This implies that benefits of the OAA program. However, there is no receiving the benefits of the OAA program does significant difference in the average value of other not lead to a significant impact on diversifying food security indicators between the two groups. the food basket of the beneficiary households. Moreover, in terms of the FIES and Household The impact of OAA on the different food and Hunger Scale, the beneficiary households lag the nutrition security indicators estimated with a nonbeneficiary households, which contradicts our PSM technique is presented in table 4.4. Here it expectation. There might be several reasons for needs to be noted that, to obtain the average this contradictory result. First, the Food Insecurity treatment effect based on PSM, first it is necessary Experience Scale (FIES) and Household Hunger to calculate the propensity score, which reflects Scale scores reflect the psychological aspect of the probability of being treated or the probability food security of the households; households of receiving the benefits of OAA. The propensity report their past experience, when they might score has been estimated including different not have availed themselves of the program and, variables such as the wealth index and household therefore, experienced food insecurity. Second, the land holding, as well as the presence of people apparent insignificant impact of social protection Table 4.3  Average Value of Different Food Security Variables for the Beneficiaries and Nonbeneficiaries of the OAA Social Security Program Food security variable OAA nonbeneficiary OAA beneficiary Per capita food expenditure (BDT) 233.97 387.65 Food consumption score 59.29 58.14 Food Insecurity Experience Scale (FIES) score 2.28 2.71 Household Hunger Scale (score) 0.25 0.44 Household dietary diversity score (HDDS) 9.43 9.28 Source: Original table for this publication based on data from BIHS 2018-2019 (IFPRI 2020). Note: BDT = Bangladeshi taka (currency); BIHS = Bangladesh Integrated Household Survey; OOA = Old Age Allowance. Country-Level Cases: Comparing Different Interventions 75 Table 4.4  Average Impact (Treatment Effect) of Social Protection Program on the Food Security of Households Average Treatment N Effect Matching (treatment N on the Standard Outcome variable technique group) (control) Treated Control Treated error T-statistic Per capita food Nearest 999 110 387.65 312.63 75.01 58.44 1.28 expenditure (BDT) neighbor (3) Nearest 999 110 387.65 309.96 77.68 57.49 1.35 neighbor (5) Kernel 999 110 387.65 316.85 70.79 27.97 5.49 Food consumption Nearest 999 110 57.99 58.68 –1.4 1.48 –0.95 score neighbor (3) Nearest 999 110 57.99 58.68 –0.68 1.96 –0.35 neighbor (5) Kernel 999 110 57.99 58.94 –0.95 1.63 –0.58 Food Insecurity Nearest 999 110 2.73 2.45 0.27 0.3 0.91 Experience Scale neighbor (3) (FIES) score Nearest 999 110 2.73 2.42 0.3 0.29 1.03 neighbor (5) Kernel 999 110 2.73 2.37 0.36 0.26 1.38 Household Nearest 999 110 0.445 0.33 0.11 0.11 1.05 Hunger Scale neighbor (3) (score) Nearest 999 110 0.445 0.32 0.127 0.102 1.24 neighbor (5) Kernel 999 110 0.445 0.309 0.136 0.091 1.5 Household dietary Nearest 999 110 9.28 9.23 0.04 1.9 0.22 diversity score neighbor (3) (HDDS) Nearest 999 110 9.28 9.27 0.009 0.185 0.05 neighbor (5) Kernel 999 110 9.28 9.3 –0.02 0.167 –0.12 Source: Original table for this publication based on data from BIHS 2018-2019 (IFPRI 2020). Note: Nearest neighbor and Kernel are matching techniques that have been used to estimate the impact of the programs. The numbers indicate the variants of nearest neighbor matching. Number 3 indicates that matching has been done with the three nearest neighbors. Number 5 indicates that matching has been done with the five nearest neighbors. BDT = Bangladeshi taka (currency); BIHS = Bangladesh Integrated Household Survey. 76 Reshaping the Agrifood Sector for Healthier Diets program on dietary diversity might be because Here the improved rural roads significantly these households are extremely vulnerable and contribute to ensure the food security of the therefore often take fewer calories than required. farming households. Households enjoying the Therefore, although they might have spent the benefits of improved rural roads have a higher per allowance for food to fill the gap, they might fail to capita food expenditure than those not enjoying spend the amount for a more diversified food. the benefits. Moreover, these households have a higher food consumption score and dietary Rural infrastructure and food and diversity score. Finally, although the household nutrition security with improved rural roads have significantly lower scores in terms of the Household Hunger Scale In evaluating the effect of rural agricultural than those without improved infrastructure, the infrastructure on food and nutrition security, this difference is not significant for the Food Insecurity case study considers the presence of roads in the Experience Scale. village passable by bus or truck during 12 months of the year as an indicator of rural infrastructure. The next step is to estimate the impact of rural The households residing in such villages are road infrastructure based on the PSM technique. considered to be the treatment group. Households In this regard, a participation equation has been residing in villages having roads not passable by estimated to match the treatment and control a bus or truck for even a single month in the year groups based on the propensity score. The are considered to be the control group. To make participation equation reflects the probability of the result comparable, the analysis is restricted to the households enjoying the benefits of improved only farming households. roads, which has been estimated based on the different household- and community-level The average value of different food security variables such as location of the households indicators for the households enjoying the benefits or community, the presence of government of improved roads and those not enjoying the hospital or primary schools in the community, the benefits have been presented in table 4.5 education level of the heads of households, and the wealth of the households. Table 4.5  Average of Food Security Variables by the Presence of Improved Roads Food security variable Without improved road With improved road Per capita food expenditure (BDT) 189.256 222.791 Food consumption score 57.995 60.517 Household dietary diversity score (HDDS) 9.233 9.598 Household Hunger Scale (score) 0.107 0.073 Food Insecurity Experience Scale (FIES) score 1.864 1.863 Source: Original table for this publication based on data from BIHS 2018-2019 (IPFRI 2020). Note: BDT = Bangladeshi taka (currency); BIHS = Bangladesh Integrated Household Survey. Country-Level Cases: Comparing Different Interventions 77 The impact of improved rural road infrastructure capita food expenditure than households on the different indicators of food and nutrition not enjoying the benefits of the improved security is shown in table 4.6. The result shows infrastructure. that rural improved roads have positive impact on per capita food expenditure of the farming The impact of improved rural road infrastructure households in Bangladesh. The impact is highly on the other two indicators of food and nutrition statistically significant and consistent across the security—the food consumption score and dietary three matching methods. The Average Treatment diversity score—is also positive and statistically Effect on the Treated (ATT) ranges between significant. The results show that the treatment BDT 99 and BDT 107, which indicates that, for group has a significantly more diversified food each matched group, households enjoying the basket than the control group. In other words, the benefits of improved rural road infrastructure improved infrastructure not only leads to a higher on average have a BDT 99–107 higher per per capita food expenditure, it also improves Table 4.6  Average Impact (Treatment Effect) of Improved Rural Infrastructure on the Food Security of Households Average N Treatment Outcome (treatment N Effect on Standard variable Matching technique group) (control) Treated Control the Treated error T-statistic Per capita food Nearest neighbor (3) 1,071 1,552 441.84 335.38 106.5 18.46 5.76 expenditure Nearest neighbor (5) 1,071 1,552 441.84 340.1 101.7 17.86 5.7 (BDT) Kernel 1,071 1,552 441.84 343.04 98.8 15.56 6.35 Food Nearest neighbor (3) 1,071 1,552 72.22 70.588 1.633 1.033 1.58 consumption Nearest neighbor (5) 1,071 1,552 72.22 70.76 1.453 1.001 1.45 score Kernel 1,071 1,552 72.22 70.85 1.36 0.857 1.6 Household Nearest neighbor (3) 1,071 1,552 10.48 10.24 0.238 0.079 2.99 dietary Nearest neighbor (5) 1,071 1,552 10.481 10.255 0.226 0.077 2.92 diversity score (HDDS) Kernel 1,071 1,552 10.481 10.27 0.204 0.066 3.08 Food Nearest neighbor (3) 1,071 1,552 1.259 1.37 –0.111 0.112 –0.99 Insecurity Nearest neighbor (5) 1,071 1,552 1.259 1.366 –0.106 0.11 –0.97 Experience Scale (FIES) Kernel 1,071 1,552 1.259 1.385 –0.126 0.092 –1.36 score Household Nearest neighbor (3) 1,071 1,552 0.0464 0.043 0.003 0.0175 0.15 Hunger Scale Nearest neighbor (5) 1,071 1,552 0.046 0.042 0.004 0.017 0.25 (score) Kernel 1,071 1,552 0.046 0.045 5E–04 0.015 0.04 Source: Original table for this publication based on data from BIHS 2018-2019 (IFPRI 2020). Note: Nearest neighbor and Kernel are matching techniques that have been used to estimate the impact of the programs. The numbers indicate the variants of nearest neighbor matching. Number 3 indicates that matching has been done with the three nearest neighbors. Number 5 indicates that matching has been done with the five nearest neighbors. BDT = Bangladeshi taka (currency); BIHS = Bangladesh Integrated Household Survey. 78 Reshaping the Agrifood Sector for Healthier Diets the quality of the food basket of the beneficiary Against this backdrop, this case study follows an households. alternative approach to find the impact of input subsidy on the food and nutrition security of the Finally, the improved rural infrastructure also helps household. The BIHS data include information to reduce the food insecurity of the rural farming about the input subsidy card, which is not households as measured by the Household Hunger universal. Only cardholder farmers can get the Scale and Food Insecurity Experience Scale. The subsidized input for free or at a lower price. score for the treatment group is lower than the Cardholder households can be considered to be score for the control, which confirms the impact the treatment group while the non-cardholder of the intervention on reducing food insecurity. households can be considered to be the control However, the effect is not statistically significant group. To make the result comparable with the across the different matching techniques. impact of other interventions, only the rural farming households receiving no input subsidy Input subsidy, productivity, and food and card are considered as the control group for nutrition security evaluating the impact of input subsidy. As discussed earlier, in evaluating the impact of a The average value of different food security program, a treatment group enjoying the benefits indicators by the input subsidy card status has been of the program is needed, as well as a control group presented in table 4.7. Here the households having with similar characteristics that has not enjoyed an input subsidy card have a lower per capita food the benefits of the program. The most prevalent expenditure than the households without the input input subsidy in Bangladesh is fertilizer subsidy. The subsidy card. This indicates a negative effect of the major problem with evaluating the effect of input input subsidy card on the household per capita food subsidy is that it is universal in Bangladesh and, expenditure, which is contradictory to expectations. therefore, there is no control group for the program Looking into the other indicators of food and with similar characteristics. The government sets nutrition security shows mixed results. Table 4.7 a ceiling price for different fertilizers that is much shows that households having an input subsidy lower than the market price. The government card have a higher food consumption score and pays the difference between the market price and dietary diversity score compared to the households the subsidized price. Anyone can purchase the not having the card. However, the average value of subsidized fertilizers at the subsidized price and, Household Hunger Scale, and the Food Insecurity therefore, can enjoy the benefits of the subsidy. Experience Scale is higher for the beneficiaries of the input subsidy card than for their nonbeneficiary Table 4.7  Average of Food Security Variables by Input Subsidy Card Status Households without input Households with input Food security variable subsidy card subsidy card Per capita food expenditure (BDT) 207.98 190.69 Food consumption score 59.18 64.03 Household dietary diversity score (HDDS) 9.42 9.64 Household Hunger Scale (score) 0.085 0.214 Food Insecurity Experience Scale (FIES) score 1.847 2.357 Source: Original table for this publication based on data from BIHS 2018-2019 (IFPRI 2020). Note: BDT = Bangladeshi taka (currency); BIHS = Bangladesh Integrated Household Survey. Country-Level Cases: Comparing Different Interventions 79 counterparts. This indicates that households wealth index, education of the household head, and receiving the input subsidy card are relatively more other socioeconomic characteristics. food insecure. Table 4.8 shows the impact of input subsidy card The next step estimates the impact of input subsidy on the different indicators of food and nutrition on the household food and nutrition security security. The result shows that input subsidies based on the PSM technique. In this regard, a have a negative effect on the per capita food participation equation has been estimated to expenditure of the rural farming households, match the treatment and control groups based on which is counter to expectations. This surprising the propensity score. The participation equation result might be because these households are reflects the probability of the households of having consuming the domestically produced food more the input subsidy card; this probability has been and therefore have to spend less on food. estimated based on the different variables such as location of the households or community, the Table 4.8  Average Impact (Treatment Effect) of Input Subsidy on the Food Security of Households Average Treatment N Effect Outcome Matching (treatment N on the Standard variable technique group) (control) Treated Control Treated error T-statistic Per capita food Nearest neighbor (3) 87 2,576 366.29 431.17 –64.88 37.48 –1.73 expenditure Nearest neighbor (5) 87 2,576 366.29 416.37 –50.08 34.98 –1.43 (BDT) Kernel 87 2,576 366.29 431.85 –65.56 29.87 –2.19 Food Nearest neighbor (3) 87 2,576 75.24 73.72 1.51 2.204 0.69 consumption Nearest neighbor (5) 87 2,576 75.25 73.36 1.88 2.14 0.88 score Kernel 87 2,576 75.25 74.58 0.661 1.938 0.34 Household Nearest neighbor (3) 87 2,576 10.471 10.475 –0.003 0.167 –0.02 dietary Nearest neighbor (5) 87 2,576 10.471 10.448 0.023 0.161 0.14 diversity score (HDDS) Kernel 87 2,576 10.471 10.474 –0.003 0.143 –0.02 Food Insecurity Nearest neighbor (3) 87 2,576 1.25 1.386 –0.134 0.243 –0.55 Experience Nearest neighbor (5) 87 2,576 1.252 1.488 –0.235 0.236 –0.99 Scale (FIES) score Kernel 87 2,576 1.252 1.512 –0.259 0.206 –1.25 Household Nearest neighbor (3) 87 2,576 0.0804 0.0344 0.046 0.047 0.97 Hunger Scale Nearest neighbor (5) 87 2,576 0.0804 0.0344 0.046 0.046 0.046 (score) Kernel 87 2,576 1.252 1.512 –0.259 0.206 –1.25 Source: Original table for this publication based on data from BIHS 2018–2019 (IFPRI 2020). Note: Nearest neighbor and Kernel are matching techniques that have been used to estimate the impact of the programs. The numbers indicate the variants of nearest neighbor matching. Number 3 indicates that matching has been done with the three nearest neighbors. Number 5 indicates that matching has been done with the five nearest neighbors. BDT = Bangladeshi taka (currency); BIHS = Bangladesh Integrated Household Survey. 80 Reshaping the Agrifood Sector for Healthier Diets Although the input subsidy has a positive effect protection program depends on the size of the on the food consumption score of the households, benefits, as well as the duration of the support. The the effect is not statistically significant. Moreover, benefit under the OAA program is very small. This while the impact of input subsidy on dietary negligible support program helps poor households diversity varies across different matching to increase their per capita food expenditure. But techniques, the impact is statistically insignificant. the benefit might fall short of diversifying the food It implies that input subsidy does not have any basket of the beneficiary households. significant impact on diversifying the food basket of the farming households. Finally, the impact The impact of input subsidies on food and is also statistically insignificant for the Food nutrition has also been found to be insignificant. Insecurity Experience Scale and Household Hunger Although the result is inconsistent with Scale. The insignificant impact of the input subsidy expectations, other studies also confirm a similar on different indicators of food security might be result. For example, Uddin and Dhar (2018) due to the very poor size of the support provided confirmed that the beneficiaries of an input through the input subsidy card. Moreover, the subsidy card continued to have a lower calorie transmission mechanism of the impact of input intake than the minimum calorie requirement. subsidy may also be weak, such that it fails to Although the input subsidy is helpful for farmers influence the food security of the households. to increase production, the size of the benefits as well as their duration might be inadequate to Discussion improve their food security. This country case study evaluates the relative Contrary to the impact of social protection effectiveness of different interventions for rural programs and input subsidy, this study finds a farming households to ensure their food and significant positive impact of rural infrastructure nutrition security. The interventions include the development on the different indicators of food social protection program, the input subsidy and nutrition security. Households having access program and rural infrastructure development. The to improved roads have a significantly higher per PSM technique has been applied to find the impact capita food expenditure and a more diversified of the different interventions. While looking into food basket than those having no access to such the impact of social protection program as proxied roads. Therefore, a direct transfer program such by the OAA program, this study finds that such a as social protection and input subsidy works less program leads to a significant increase in the per effectively than an indirect support program such capita food expenditure of the poor beneficiary as rural infrastructure development. households. However, no significant impact of the program was found on other variables of food and nutrition security such as food consumption 4.2. Input subsidies, irrigation, score, dietary diversity score, the Food Insecurity and cash transfers in Experience Scale score, or the Household Hunger Scale score. The impact of social protection Malawi programs on food and nutrition security in Bangladesh is mixed in the extant literature. Ahmed In the case of Malawi, input subsidies and et al. (2009) and Uraguchi (2011) found a similar result of the impact of cash transfer program on food and cash transfers have different food security. However, there are other studies impacts on food and nutrition security. (Ahmed et al. 2009; Badhan et al. 2019; Mamun Input subsidies, which typically support 2019) showing positive effects of the program. maize production, reduce the likelihood of The literature confirms that the impact of a social undertaking negative coping strategies but Country-Level Cases: Comparing Different Interventions 81 do not improve diet diversity or the food also been found to support food and nutrition consumption score. In contrast, although the security (Baird, McIntosh, and Özler 2019; Nkhata, likelihood of undertaking negative coping Jumbe, and Mwabumba 2014). However, there is a dearth of studies evaluating the impacts strategies increase with food and cash of participating in multiple programs (Tirivayi, transfers, the transfers lead to improved Knowles, and Davis 2016). This country case study diet diversity and food consumption score. compares and assesses the combined effects Irrigation infrastructure is found to have of social protection and irrigation investments no effects on food and nutrition security. using integrated household panel surveys for Moreover, there are no positive effects of three years, extending the literature on food and a combination of input subsidies, food and nutrition security. cash transfers, and irrigation infrastructure, Input subsidy programs have evolved in Malawi suggesting that production- and transfer- (Nkhoma 2018). Malawi’s prominent Farm Input based programs are not designed to leverage Subsidy Program (FISP), a 15-year program, synergies was replaced by the Affordable Inputs Program (AIP) in 2020. This program provides inorganic Global priorities for sustainable development fertilizers and improved maize, sorghum, or rice to include food and nutrition security. However, productive poor farmers. The Social Cash Transfer climate change, pandemics, economic shocks, Program (SCTP) provides monthly cash transfer of and conflicts pose threats to food systems (Poole, Malawian kwacha (MK) 14,919 (equivalent to $9) Donovan, and Erenstein 2021). Malawi faces to ultra-poor households with labor constraints. malnutrition, stunting, obesity, and hunger, The program also responds to disasters such with 37 percent of children under the age of five as floods, food inadequacy, and the COVID-19 stunted (NSO and ICF 2017) and one-third of the pandemic. Food aid is another direct welfare 20 million people living in the country at risk of transfer from government used to abate shock- hunger (IPC 2022). Among strategies to boost induced hunger and food insecurity. Households incomes and productivity of resource-constrained receive cereals (maize), pulses (beans), cooking smallholder farmers in Malawi and other African oil, and nutritious food (corn soya blend) through countries are agricultural input subsidy programs the Malawi Vulnerability Assessment Committee (Jayne et al. 2018). However, there are several gaps (MVAC) response and supplementary feeding with the model, including tight fiscal space for the programs. Malawi’s government irrigation program and other public investments such as projects, managed by cooperatives, have benefited agricultural extension (Holden 2019). Furthermore, 348,572 smallholder farmers, but only 36 percent the greater emphasis on maize and inorganic of potential 407,862 hectares are irrigated. The fertilizers is linked to reduced diversification, farmers cultivate a range of crops including maize, higher consumption of energy-dense diets, and rice, and vegetables such as tomatoes, onions, soil health issues (Ickowitz et al. 2019). The use of cabbages, and potatoes. Government irrigation inorganic fertilizers in maize production in Malawi projects with credit financing include the Shire has also been found to be largely unprofitable at Valley Transformation Program and the Linga current market prices, under the assumption that Irrigation Scheme. This study considers the paired farmers face positive transaction costs associated effects of these social protection and irrigation with fertilizer use (Darko, Ricker-Gilbert, and investment programs on food and nutrition Kilic 2024). Additionally, input subsidies’ impact security. on productivity, welfare, and food and nutrition security shows mixed results (Hemming et al. A recent study assessed whether Malawi’s input 2018). Other interventions—such as food aid, subsidy programs should target non-poor farmers cash transfers, and irrigation investments—have instead of poor farmers by estimating the net 82 Reshaping the Agrifood Sector for Healthier Diets gain in maize yield from targeting non-poor the individual/standalone effects of participation farmers (Darko 2024). The analysis accounted for in farm input subsidies program, social differences in inorganic fertilizer use efficiency assistance program, or public irrigation scheme. and differences in crowding out of commercial The combined effects of farm input subsidies fertilizer by subsidized fertilizer in both groups. with social protection or irrigation investment The findings show that non-poor farmers are on food and nutrition security are assessed significantly more efficient in their use of inorganic using the interaction term (Interv1it × Interv2it). fertilizer, but experience much higher levels of Control variables, including household crowding out compared to poor farmers. However, socioeconomic characteristics and location-specific the study also indicates that targeting non-poor fixed effects, are denoted by Xit, whereas εit is the farmers would yield an overall increase of 3.14 to idiosyncratic error term and β represent the slope 4.33 kilograms of maize per kilogram of nitrogen coefficients. distributed through the subsidy program, even after accounting for the crowding out effect. This This case study uses several measures of food opens the question of whether focusing on non- and nutrition security including household dietary poor farmers, rather than poor farmers, might diversity score (HDDS), food consumption score, better achieve the goal of enhancing productivity and coping strategy index (Maxwell, Vaitla, and through Malawi’s input subsidy program (Darko, Coates 2014). The main independent variables are Ricker-Gilber, and Kilic 2024). production and cash transfer–based investments. Production policies under consideration include the input subsidies and irrigation investment Data and methods measures. The case study also considers participation in social assistance programs—in This case study uses Living Standards particular, receipt of cash transfer and/or food Measurement Studies (LSMS) data for Malawi, aid. Because of the modest sample sizes for panel data for the years 2013, 2016, and 2019. The each component, participation in the two social 2013 survey was conducted between April and assistance programs was pooled to facilitate October 2013, the 2016 survey field work occurred meaningful analysis. Apart from these variables between April 2016 and April 2017, and the 2019 of interest, the study accounts for household and survey was conducted between April 2019 and location-specific characteristics. March 2020. The surveys included individual and community questionnaires collecting data on To estimate dietary diversity, the case study agriculture, employment, and consumption assets, employs instrumental variable Poisson (IV Poisson) among others, to explain the living conditions of regression via the generalized method of Malawians. In Malawi, the surveys are conducted moments (GMM). The instrumental variable is by the National Statistical Office with support from employed because there is potential selection the World Bank. bias for the variables of interest resulting in endogeneity. Several instrumental variables are To estimate the impact of social protection and used, depending on the endogenous variable irrigation investment on food and nutrition under consideration. The instrumental variables’ security, a model of the following form is used: validity is checked using the Instrumental Variable Two Stage Least Squares (IV-2SLS) estimation for panel-data models since the IV Poisson method does not produce post-estimation tests. The combined effects of the interventions are assessed where FNSit represents the food and nutrition by including interaction terms in the models. security measure for an individual household i in the period of survey t. Interv1it and Interv2it capture Country-Level Cases: Comparing Different Interventions 83 Models estimating program effects on food consumed food from 8 food groups and the consumption scores are estimated using single- maximum was 12 food groups. The consumption equation instrumental-variables regression via of 8 food groups compares well with other the GMM to take advantage of pooled data. national-level studies (Jones, Shrinivas, and Additional analysis considers food and nutrition Bezner-Kerr 2014; Matita et al. 2022). The mean security as measured by coping strategies using food consumption score was 49, which indicates tobit estimation. A comparison of the panel tobit that food and nutritional insecurity is not severe model with the pooled tobit model using the but visible. The proportion of households with a likelihood-ratio test shows that there were panel- borderline and poor food security situation was level effects in the relationship, hence results 20.4 and 2.5 percent of the sample, respectively. based on panel data analysis are reported. Close to 77.1 percent reported acceptable levels of food security. About coping with food shortages, at Descriptive statistics least 61 percent of the sample used some coping strategy in the past week of the survey to deal with Table 4.9 presents the descriptive statistics for food insecurity. The average coping strategy index the pooled sample. On average, households score was 6 with a maximum of 56 – the higher the Table 4.9  Food and Nutrition Variables and Program Participation Overall Standard Variable mean deviation Minimum Maximum Household dietary diversity 8.058 2.017 1 12 Food consumption score 49.28 18.53 8 126 Coping strategy index score 5.664 7.603 0 56 Received subsidized inputs (1/0) 0.280 0.449 0 1 Received food/cash transfer (1/0) 0.115 0.319 0 1 Community has irrigation scheme (1/0) 0.133 0.340 0 1 HH uses irrigation scheme (1/0) 0.032 0.177 0 1 Subsidized inputs × irrigation scheme presence (1/0) 0.037 0.188 0 1 Subsidized inputs × use of irrigation scheme (1/0) 0.012 0.107 0 1 Subsidized inputs × share of HH in irrigation scheme 0.001 0.007 0 0.09 Subsidized inputs × food/cash transfer (1/0) 0.042 0.200 0 1 Food/cash transfer × irrigation scheme presence (1/0) 0.015 0.120 0 1 Food/cash transfer × share of HH in irrigation scheme 0.001 0.013 0 0.30 Food/cash transfer × HH use an irrigation scheme (1/0) 0.004 0.066 0 1 Received cash transfer 0.028 0.164 0 1 Received food aid 0.087 0.282 0 1 Number of observations 3,009 Source: Original table for this publication based on an analysis of Living Standards Measurement Studies (LSMS) data for Malawi; data panel for 2013, 2016, and 2019 (LSMS-ISA). Note: HH = household. (1/0) indicates dichotomous variable equal to 1 for the included category, otherwise equal to 0 for the base category. 84 Reshaping the Agrifood Sector for Healthier Diets coping strategy index score the greater the use of while stakeholders support multiple program negative coping mechanisms with food insecurity. participation to enhance possibility of graduation, The finding of a lower coping strategy index score informal rules of exclusion are promoted by and a higher food consumption score and dietary chiefs and communities in the targeting process diversity is in line with the consensus that there is (Chirwa et al. 2016). an inverse relationship between coping strategy index and other related scores (Maxwell, Caldwell, The study’s estimates show that participation in and Langworthy 2008). both cash transfer and input subsidy programs is limited to only 4 percent of the sample with no About 28 percent of the sample received significant correlation. There is, however, strong subsidized farm inputs. Government transfers positive correlation between receiving food aid were accessible to few households; only 3 percent and input subsides (p < 0.01), suggesting that received cash transfers and 9 percent obtained households might concurrently benefit from the food aid. In the econometric analysis, receipt of programs. either cash transfers and/or food aid representing 12 percent of the sample were considered. The Table 4.11 presents mean differences in food and presence of an irrigation scheme in a community nutrition measures by program participation. It was reported by 13 percent of the sample and is observed that, although households receiving only 3 percent have access to use the schemes. input subsidies cope less with food insecurity, Joint program participation is minimal in the their FCS is significantly lower. A significantly sample. For example, only 4 percent of households lower FCS and dietary diversity as well as an reporting the presence of an irrigation scheme increase in usage of coping strategies to address in their community received input subsidies. food insecurity is linked to presence and use of According to pairwise correlation in table 4.10, an irrigation scheme. Furthermore, receiving receiving input subsidies is weakly associated food/cash transfers is associated with higher with the use of an irrigation scheme (p < 0.10). dietary diversity; but such households cope with The proportion using the irrigation schemes and significantly more food insecurity. benefiting from the subsidy program is far less: only 1.2 percent. A qualitative assessment about The modeling approach also controlled for several the eligibility of SCTP beneficiaries for other variables whose descriptive statistics are reported social assistance programs demonstrates that, in table 4.12. The majority of the household heads Table 4.10  Pairwise Correlation of Program Participation Received HH use of an Irrigation subsidized inputs irrigation scheme scheme presence Program (1/0) (1/0) (1/0) HH use an irrigation scheme (1/0) 0.0323* n.a. n.a. Irrigation scheme presence (1/0) –0.0029 n.a. n.a. Food aid and/or cash transfer (1/0) 0.0688 *** 0.0115 –0.0035 Food aid 0.0795*** 0.0237 –0.0052 Cash transfer –0.0060 –0.0193 0.0005 Source: Original table for this publication based on an analysis of Living Standards Measurement Studies (LSMS) data (LSMS-ISA). Note: HH = household; n.a. = not applicable. (1/0) indicates dichotomous variable equal to 1 for the included category, otherwise equal to 0 for the base category. Reported values are correlation coefficients. * p < 0.10, ** p < 0.05, *** p < 0.01. Country-Level Cases: Comparing Different Interventions 85 Table 4.11  Mean Differences in Food and Nutrition Security Measures by Program Household Food consumption Coping strategy Number of Program dietary diversity score index observations Received subsidized inputs (yes) 8.144 48.27 5.24 846 Received subsidized inputs (no) 8.024 49.81** 5.83** 2,163 HH use an irrigation scheme (yes) 7.89 45.83 6.16 97 HH use an irrigation scheme (no) 8.06 49.50 ** 5.65 2,912 Irrigation scheme presence (yes) 7.76 47.52 7.09 396 Irrigation scheme presence (no) 8.10 *** 49.66 5.45 *** 2,613 Received food/cash transfer (yes) 8.23 50.82 6.71 343 Received food/cash transfer (no) 8.04 * 49.19 5.53 *** 2,666 Food aid (yes) 8.58 52.29 5.79 262 Food aid (no) 8.01 *** 49.10 ** 5.65 2,747 Cash transfer (yes) 7.12 45.99 9.61 83 Cash transfer (no) 8.08*** 49.47* 5.55*** 2,926 Source: Original table for this publication based on an analysis of Living Standards Measurement Studies (LSMS) data (LSMS-ISA). Note: (1/0) indicates dichotomous variable equal to 1 for the included category, otherwise equal to 0 for the base category. * p < 0.10, ** p < 0.05, *** p < 0.01. Table 4.12  Descriptive Statistics for Pooled Sample Variable Mean Standard deviation Minimum Maximum Age of household head (years) 44.06 15.98 10 104 Male headed household 1/0) 0.735 0.441 0 1 Household size 4.995 2.195 1 16 Years of education for the head 7.152 3.725 0 23 Number of household members 0–5 years 0.872 0.841 0 5 Number of household members 6–14 years 1.168 1.175 0 8 Number of household members 15–55 years 2.410 1.298 0 9 Number of household members ≥ 56 years 0.412 0.689 0 4 Wage employment (1/0) 0.130 0.336 0 1 Business employment (1/0) 0.107 0.310 0 1 Agriculture employment (1/0) 0.591 0.492 0 1 Asset index 0.794 4.263 –0.042 23.22 Distance to nearest weekly market 7.605 5.745 0 37 Number of crops cultivated 2.632 1.397 0 9 Received any extension service (1/0) 0.680 0.466 0 1 Received credit (1/0) 0.246 0.431 0 1 Farmland size (acreage) 1.780 1.604 0 18 Southern region (1/0) 0.494 0.500 0 1 Central region (1/0) 0.387 0.487 0 1 Number of observations 3,021 Source: Original table for this publication based on an analysis of Living Standards Measurement Studies (LSMS) data for Malawi (LSMS-ISA). Note: (1/0) indicates dichotomous variable equal to 1 for the included category, otherwise equal to 0 for the base category. 86 Reshaping the Agrifood Sector for Healthier Diets (74 percent), who are on average forty-four years diversity, food consumption scores, and coping old and have seven years of schooling, are men. strategy index—in tables 4.13, 4.14, and 4.15, On average, a household with five members respectively. The models presented are all jointly cultivates 1.9 acres of land. Only 25 percent of statistically significant at the 1 percent level as the sample accessed credit, although over half indicated by the obtained log-likelihood ratio received extension services (68 percent). chi-squared statistic. The included instrumental variables were valid and the test of endogeneity Regression results was significant. Two endogenous variables are considered in each paired analysis. All models Summarized results are presented for a set of include control variables such as household food and nutrition measures—household dietary Table 4.13  Effects of Interventions on Household Dietary Diversity Model 1 Incidence rate Standard Dependent variable: Household dietary diversity ratio errors Panel A Received subsidized farm inputs (1/0) 1.348 (0.317) HH uses an irrigation scheme (1/0) 1.509 (0.925) Subsidized inputs × use of irrigation scheme (1/0) 0.550 (0.397) Panel B Received subsidized farm inputs (1/0) 1.348 (0.277) Community has an irrigation scheme (1/0) 1.183 (0.438) Subsidized inputs × community has irrigation scheme (1/0) 0.690 (0.299) Panel C Received subsidized farm inputs (1/0) 0.808 (0.321) Received food/cash transfer (1/0) 2.072 *** (0.528) Subsidized inputs × food/cash transfer (1/0) 0.660 (0.237) Panel D Community has an irrigation scheme (1/0) 0.706 (0.375) Received food/cash transfer (1/0) 1.738*** (0.287) Irrigation scheme × food/cash transfer (1/0) 0.858 (0.447) Panel E Received food/cash transfer (1/0) 1.810*** (0.300) HH uses an irrigation scheme (1/0) 1.401 (0.773) HH uses an irrigation scheme × food/cash transfer (1/0) 0.443 (0.263) Source: Original table for this publication based on an analysis of Living Standards Measurement Studies (LSMS) data for Malawi (LSMS-ISA). Note: All models include control variables such as household characteristics and location fixed effects. Exponentiated coefficients. Standard errors are in parentheses. (1/0) indicates dichotomous variable equal to 1 for the included category, otherwise equal to 0 for the base category. HH = household. * p < 0.10, ** p < 0.05, *** p < 0.01. Country-Level Cases: Comparing Different Interventions 87 socioeconomic characteristics and location fixed an additional two food groups compared to effects. those without such transfers (panels C, D, and F). Nonetheless, this significant standalone effect Table 4.13 reports the effects of various driven by food aid does not persist when food/ interventions on household dietary diversity cash transfers are considered in combination with estimated using IV Poisson regression. No other interventions. For instance, the results show statistically significant joint effects on dietary that households receiving food/cash transfers in diversity from participation in any of the programs communities where an irrigation scheme exists do under consideration were found. However, not experience dietary diversity that is significantly receiving food/cash transfers improves food different from those in communities without consumption (p < 0.01). The results suggest an irrigation scheme and who are not receiving households receiving food/cash transfers consume government social transfers. Table 4.14  Effects of Interventions on Food Consumption Score Model 1 Standard Dependent variable: Food consumption score Coefficient errors Panel A Received subsidized farm inputs (1/0) 29.146 (20.216) HH uses an irrigation scheme (1/0) 72.026 (67.444) Subsidized inputs × use of irrigation scheme (1/0) –90.862 (77.749) Panel B Received subsidized farm inputs (1/0) 31.290 (20.468) Community has an irrigation scheme (1/0) 47.295 (36.394) Subsidized inputs × community has irrigation scheme (1/0) –66.622 (44.156) Panel C Received subsidized farm inputs (1/0) –12.241 (21.464) Received food/cash transfer (1/0) 72.839** (35.939) Subsidized inputs × food/cash transfer (1/0) –55.317* (31.705) Panel D Community has an irrigation scheme (1/0) 2.341 (23.620) Received food/cash transfer (1/0) 45.362*** (16.032) Irrigation scheme × food/cash transfer (1/0) –40.182 (26.488) Panel E Received food/cash transfer (1/0) 47.961*** (17.949) HH uses an irrigation scheme (1/0) 45.316 (50.633) HH uses an irrigation scheme × food/cash transfer (1/0) –86.161 (57.079) Source: Original table for this publication based on an analysis of Living Standards Measurement Studies (LSMS) data for Malawi (LSMS-ISA). Note: All models include control variables such as household characteristics and location-fixed effects. Obtained coefficients are the same as conditional marginal effects (dy/dx). Standard errors in parentheses. (1/0) indicates dichotomous variable equal to 1 for the included category, otherwise equal to 0 for the base category. * p < 0.10, ** p < 0.05, *** p < 0.01. 88 Reshaping the Agrifood Sector for Healthier Diets Table 4.14 shows the effects of different with food/cash transfers (panel C in the table). interventions on food consumption scores Households receiving input subsidies as well as generated from the IV Poisson regression. Positive food/cash transfers experience reduced food joint effects on food consumption scores are consumption scores by a margin of 55 points found to be lacking. In other words, there are no (p < 0.10). However, receiving food/cash transfers significant combined effects on food consumption alone has significant favorable benefits on food scores from investing in irrigation and receiving consumption scores (panels C, D, and E), implying input subsidies. Additionally, the findings that households receiving government food/ demonstrate that households are significantly cash transfers experience greater balance in food disadvantaged in food consumption scores when consumption. they combine participation in input subsidies Table 4.15  Effects of Interventions on Coping with Food Insecurity Model 2 Standard Dependent variable: Coping strategy index (CSI) Coefficient errors Panel A Received subsidized farm inputs (1/0) –0.135* (0.074) HH uses an irrigation scheme (1/0) –0.318 (0.217) Subsidized inputs × use of irrigation scheme (1/0) 0.267 (0.360) Panel B Received subsidized farm inputs (1/0) –0.115 (0.077) Community has an irrigation scheme (1/0) –0.001 (0.108) Subsidized inputs × community has an irrigation scheme (1/0) –0.094 (0.203) Panel C Received subsidized farm inputs (1/0) –0.158** (0.078) Received food/cash transfer (1/0) 0.144 (0.120) Subsidized inputs × transfer (1/0) 0.182 (0.200) Panel D Received food/cash transfer (1/0) 0.257** (0.104) Community has an irrigation scheme (1/0) 0.026 (0.098) Irrigation scheme × food/cash transfer (1/0) –0.412 (0.283) Panel E Received food/cash transfer (1/0) 0.224** (0.099) HH uses an irrigation scheme (1/0) –0.159 (0.187) HH uses an irrigation scheme × food/cash transfer (1/0) –0.548 (0.503) Source: Original table for this publication based on an analysis of Living Standards Measurement Studies (LSMS) data (LSMS-ISA). Note: All models include control variables such as household characteristics and location fixed effects. Because the dependent variable is log transformed, the obtained coefficients can be interpreted in terms of percentage change. Standard errors in parentheses. (1/0) indicates dichotomous variable equal to 1 for the included category, otherwise equal to 0 for the base category. * p < 0.10, ** p < 0.05, *** p < 0.01. Country-Level Cases: Comparing Different Interventions 89 Table 4.15 presents panel tobit estimates of investments using integrated household panel the effects of social assistance and irrigation surveys for three years. investments on food and nutrition security as measured by coping strategy index. The results Overall, the findings suggest weak and indicate no significant combined effects from insignificant joint effects from participating in participation in multiple programs. However, several programs. In particular, the availability of receiving input subsidies alone has a significant irrigation infrastructure in a community and its effect on a household’s ability to cope with use by households that receive food/cash transfers food insecurity (panels A and C). For example, or input subsidies has no influence on food and households receiving input subsidies cope nutrition security outcomes. This is contrary to between 13–16 percent less with food poverty evidence demonstrating standalone positive on average, suggesting a better food security effects of irrigation infrastructure on production, situation, insofar as they engage less in negative livestock, and livelihoods (Fan, Gulati, and Thorat coping strategies. Conversely, receiving food/ 2008; Nkhata, Jumbe, and Mwabumba 2014), but cash transfers is linked to a significant use of these are localized studies that did not focus on coping mechanisms (panel D and E). On average, food and nutrition security measures. It is also beneficiaries of food/cash transfers cope 21–26 likely that the low coverage of the programs may percent more with food and nutrition insecurity be driving the counterintuitive results. (p < 0.05), thus implying greater use of negative measures such as reducing food portions and the This case study has also revealed that receiving number of meals, among others, to deal with food input subsidies has no effect on dietary diversity, insecurity. consistent with other studies suggesting that the effects are indirect—that is, through the sale of Discussion maize, the growing of nutrient-rich crops, food expenditures, or filling the maize basket (Karamba A growing number of Malawian households 2013; Matita et al. 2022; Smale, Thériault, and are experiencing food and nutrition insecurity Mason 2020; Snapp and Fisher 2015). The finding because of the effects of COVID-19 pandemic; about the lack of input subsidies’ effect on food supply chain disruptions caused by Russia’s consumption scores, however, differs from that invasion of Ukraine; and climate shocks that of Harou (2018), who reported that, depending result in the loss of produce, livestock, and on the number of vouchers received between livelihoods. Furthermore, the recent devaluation 2008 and 2013, the effects are positive in Malawi. of the Malawi kwacha relative to the US dollar Nevertheless, a recent study by Chakrabarti et al. has made the macroeconomic environment (2024) corroborates that the impact of input even more dire, making it difficult to afford subsidies on food and nutrition are largely indirect nutritious foods. Public programs to achieve food and come through improvements in productivity and nutrition security include the provision of and income. Therefore, the lack of joint impacts input subsidies, the establishment of irrigation may not necessarily reflect a lack of relationship infrastructure, and direct welfare transfers in the but rather that the association between food form of food or cash. Prior evaluation of these security and the programs is indirect. programs found positive association with food and nutrition security; there is, however, a dearth This case study indicated that the combined of studies evaluating the impacts of participating effects of food/cash transfers and input subsidies in multiple programs (Tirivayi, Knowles, and on FCS are negative, which is contrary to study Davis 2016). This country case study contributes expectations. Especially in light of research to the literature by comparing and assessing the showing that input subsidies raise household food combined effects of social protection and irrigation production capacity while cash transfers address 90 Reshaping the Agrifood Sector for Healthier Diets liquidity constraints, which—when combined— the quantity of maize. In other words, subsidizing have positive incremental effects (Pace et al. farm inputs for the growing of maize increases 2017; Thome, Taylor, and Filipski 2014). Even then, food monotony rather than diversification, as the lack of influence of the two programs could also demonstrated by Chakrabarti et al. (2024). be a manifestation of the characteristics of the Conversely, receipt of food/cash transfers is beneficiaries. Since typically social cash transfer associated with increased use of negative food programs targets labor-constrained households coping mechanisms and positively impacts on both in Malawi, households receiving food assistance dietary diversity and food consumption score— or cash transfers are already constrained. They both measures of diversity of dietary intake. This lack the labor to use the subsidized inputs in exemplifies how cash/food transfers may not be their production. Furthermore, those receiving sufficient to promote other dimensions of food food/cash transfers are more vulnerable to food and nutrition security in Malawi. Providing food/ insecurity, and sourcing food for consumption cash transfers supports market purchases and would take precedence over making efficient use consumption of a wider variety of foods at home of the subsidized inputs. There is also literature but not the quantity of food consumption. indicating that households that are food insecure in terms of meeting their calorie requirements There are very few collaborative programs in are unlikely to respond to incentives to grow Malawi that encompasses both production- and nutrient-rich crops or to participate in output transfer-based instruments. This results from, marketing to earn income, both of which are key among other things, the structural or inherent to improving the diversity of foods consumed distinctions in the program designs, where the (Aberman and Roopnaraine 2020; Carletto, Corral, targeting of transfer-based programs focuses on a and Guelfi 2017; Chirwa and Matita 2012; Matita criterion linked to the characteristics of individuals et al. 2022). Additionally, this is in keeping with the or households. For instance, transfer-based findings of Thome, Taylor, and Filipski (2014) that instruments are primarily designed for ultra- households with larger land holdings benefit more poor and labor-constrained households, whereas from the input subsidies and the finding of Pace production-based instruments are purposed et al. (2017) that positive synergies exist for labor- to enhance crop productivity among resource- unconstrained households receiving cash transfers constrained smallholder farmers but with land and input subsidies. Therefore, the combined (Harou 2018). This implies that, in as much as effects could be heterogenous. joint programs show significant impact elsewhere (Sulaiman 2016), there are no deliberate efforts Furthermore, looking across the indicators, input to consolidate these instruments into a unified subsidies are found to have no effect on dietary approach in Malawi. diversity and food consumption scores and are only useful in lowering the use of negative coping This study concludes that the combined effects of strategies, in keeping with Zingwe, Manja, and input subsidies, food/cash transfers, and irrigation Chirwa (2021). It is likely that food availability investments on food and nutrition security are assured through own-food production makes it minimal. Therefore, there might not be many less likely for households to use negative coping opportunities to build synergies between social mechanisms, a finding corroborated by Tirivayi, assistance and irrigation investments through Knowles, and Davis (2016) in a review of the links the existing stand-alone implementation model. between social protection and agriculture. The Nevertheless, households simply receiving food/ notion that the coping strategy index captures cash transfers showed improvements in food more food quantity than quality helps to further consumption and dietary diversity scores. In explain this (Vaitla et al. 2017). The coping addition, input subsidies helped households to strategy index is therefore more sensitive to input use fewer coping strategies to manage their food subsidies, as subsidies in Malawi aim to increase insecurity even if the program also reduced dietary Country-Level Cases: Comparing Different Interventions 91 diversity and food consumption scores, suggesting The results of this study should be interpreted that overreliance on agriculture input subsidies with caution. First, the various program impacts leads to reduced variety in consumption. Policy on food and nutrition security are largely indirect pursuits for more explicit links between programs, via crop production, productivity, and likely income while commendable, might require addressing increases, pathways that have not been addressed in parallel the need to diversify and rebalance in this study. Therefore, the weak joint program public spending intended to reduce food and effects may be resulting from the indirect relation nutrition insecurity. This might include rebalancing between the food security and the programs, spending toward food/cash transfers from and not necessarily a lack of relationship. Second, high spending on agricultural input subsidies, low coverage and under-sampling of programs’ sequencing interventions as well as targeting participants might have contributed to the weak agricultural input subsidies to poor households effects estimated. Future studies could address receiving food/cash assistance that might make these limitations better use of these inputs. Notes 1. Details about the World Bank’s FoodSystems 2030 Umbrella Multi-Donor Trust Fund can be found at https://www.worldbank.org/en​/programs/food-systems-2030. 2. This grant comes from the Program on Agricultural and Rural Transformation for Nutrition, Entrepreneurship, and Resilience (PARTNER) in Bangladesh is to promote diversification, food safety, entrepreneurship, and climate resilience in the agri-food systems of Bangladesh. See https://projects.worldbank.org/en/projects-operations/project-detail/P176374 for details. 3. This grant comes from the Food Systems Resilience Program for Eastern and Southern Africa Project, which aims to increase the resilience of food systems and preparedness for food insecurity in the participating countries. Details about this project are available at https://projects.worldbank.org/en/projects-operations/project-detail/P178566. 92 Reshaping the Agrifood Sector for Healthier Diets References Bangladesh Ahmed, A. U., A. R. Quisumbing, M. Nasreen, J. Hoddinott, and E. Bryan. 2009. Comparing Food and Cash Transfers to the Ultra Poor in Bangladesh. IFPRI Research Monograph 163. Washington, DC: IFPRI. Badhan, S. A., S. Haque, M. Akteruzzaman, N. Zaman, K. Nahar, and F. Yeasmin. 2019. “Role of Social Safety Net Programmes for Ensuring Food Security and Reducing Poverty in Char Area of Jamalpur District in Bangladesh.” Progressive Agriculture 30 (1): 75–85. IFPRI (International Food Policy Research Institute). 2020. “Bangladesh Integrated Household Survey (BIHS) 2018-2019.” https://doi.org/10.7910/DVN/NXKLZJ. Khandker, S. R., Z. Bakht, and G. B. Koolwal. 2009. “The Poverty Impact of Rural Roads: Evidence from Bangladesh.” Economic Development and Cultural Change 57 4): 685–722. Mamun, M. 2019. “The Impact of Social Safety Net Programmes on Poverty Reduction in Bangladesh: An Evidence from Old Age Allowance.” Bangladesh Journal of Public Administration 27 (1): 63–78. Nasrin, M., S. Bauer, and M. Arman. 2018. “Assessing the Impact of Fertilizer Subsidy on Farming Efficiency: A Case of Bangladeshi Farmers.” Open Agriculture 3 (1): 567–77. Rahman, M. M. 2012. “Estimating the Effects of Social Safety Net Programmes in Bangladesh on Calorie Consumption of Poor Households.” Bangladesh Development Studies 35: 67–85. Sarma, P. K. and A. Hossain. 2020. “Impact of Agricultural Input Subsidy Assistance Card Program on Farm Production in Rajbari District of Bangladesh.” Journal of the Bangladesh Agricultural University 18 (3): 688–99. Sarma, P. and M. M. Rahman. 2020. “Impact of Government Agricultural Input Subsidy Card On Rice Productivity in Rajbari District of Bangladesh: Application of Endogenous Switching Regression Model.” Universal Journal of Agricultural Research 8 (5): 131–45. Uddin, M. T. and A. R. Dhar. 2018. “Government Input Support on Aus Rice Production in Bangladesh: Impact on Farmers’ Food Security and Poverty Situation.” Agriculture & Food Security 7: 1–15. Uraguchi, Z. B. 2011. “Social Protection for Redistributive Justice: Socio-Economic and Political Drivers of Vulnerability to Food Insecurity in Bangladesh and Ethiopia.” Paper prepared for the International Conference “Social Protection for Social Justice,” Institute of Development Studies, UK, April 13–15, 2011. Country-Level Cases: Comparing Different Interventions 93 Malawi Aberman, N. and T. Roopnaraine. 2020. “To Sell or Consume? Gendered Household Decision-Making on Crop Production, Consumption, and Sale in Malawi.” Food Security 12: 433–47. https://doi.org/10.1007/s12571-020-01021-2. Baird, S., C. McIntosh, and B. Özler. 2019. “When the Money Runs Out: Do Cash Transfers Have Sustained Effects on Human Capital Accumulation?” Journal of Development Economics 140 (March): 169–85. https://doi.org/10.1016/j​.jdeveco.2019.04.004. Carletto, C., P. Corral, and A. Guelfi. 2017. “Agricultural Commercialization and Nutrition Revisited: Empirical Evidence from Three African Countries.” Food Policy 67: 106–18. https://doi.org/10.1016/j.foodpol.2016.09.020. Chakrabarti, A., A. P. Harou, J. Fanzo, and C. A. Palm. 2024. “Exploring Agriculture‑Child Nutrition Pathways: Evidence from Malawi’s Farm Input Subsidy Program.” Food Security 16: 201–21. https://doi.org/10.1007/s12571​-023-01416-x. Chirwa, E. W. and M. Matita. 2012. “From Subsistence to Smallholder Commercial Farming in Malawi: A Case of NASFAM Commercialisation Initiatives.” FAC Working Paper No. 37. Brighton: Future Agricultures Consortium. https://opendocs.ids.ac.uk/opendocs/handle/123456789/2268. Darko, F.A., J.E. Ricker-Gilbert, and T. Kilic. 2024. Profitability of Fertilizer Use in Sub-Saharan Africa: Evidence from Malawi (English). Policy Research Working Paper, PROSPERITY, Living Standards Measurement Study. Washington DC: World Bank Group. Fan, S., A. Gulati, and S. Thorat. 2008. “Investment, Subsidies, and Pro- Poor Growth in Rural India.” Agricultural Economics 39 (2): 163–70. https://doi.org/10.1111​/j.1574-0862.2008.00328.x. Harou, A. P. 2018. “Unraveling the Effect of Targeted Input Subsidies on Dietary Diversity in Household Consumption and Child Nutrition: The Case of Malawi.” World Development 106: 124–35. https://doi.org/10.1016/j​.worlddev.2018.01.011. Hemming, D. J., E. W. Chirwa, H. J. Ruffhead, R. Hill, J. Osborn, L. Langer, L. Harman, C. Coffey, A. Dorward, and D. Phillips. 2018. “Agricultural Input Subsidies for Improving Productivity, Farm Income, Consumer Welfare and Wider Growth in Low- and Middle-Income Countries: A Systematic Review. 3ie Systematic Review 41. https://doi.org/10.23846/SR51062. Holden, S. T. 2019. “Economics of Farm Input Subsidies in Africa.” Annual Review of Resource Economics 11: 501–22. https://doi.org/10.1146/annurev​-resource-100518-094002. Ickowitz, A., B. Powell, D. Rowland, A. Jones, and T. Sunderland. 2019. “Agricultural Intensification, Dietary Diversity, and Markets in the Global Food Security Narrative. Global Food Security 20: 9–16. https://doi.org/10.1016/j.gfs​.2018.11.002. 94 Reshaping the Agrifood Sector for Healthier Diets IPC (Integrated Food Security Phase Classification). 2022. Malawi: IPC Acute Food Insecurity Analysis July 2021 - March 2022. Rome, Italy. https://reliefweb.int​ /report/malawi/malawi-ipc-acute-food-insecurity-analysis-july-2021-march​ -2022-issued-august-2021 Jayne, T. S., N. M. Mason, W. J. Burke, and J. Ariga. 2018. “Review: Taking Stock of Africa’s Second-Generation Agricultural Input Subsidy Programs.” Food Policy 75: 1–14. https://doi.org/10.1016/j.foodpol.2018.01.003. Jones, A. D., A. Shrinivas, and R. Bezner-Kerr. 2014. “Farm Production Diversity Is Associated with Greater Household Dietary Diversity in Malawi: Findings from Nationally Representative Data.” Food Policy 46: 1–12. https://doi​.org/10.1016/j.foodpol.2014.02.001. Karamba, R. W. 2013. “Input Subsidies and Their Effect on Cropland Allocation, Agricultural Productivity, and Child Nutrition: Evidence from Malawi.” PhD thesis, American University. LSMS-ISA (Living Standards Measurement Study - Integrated Surveys on Agriculture). Data sets. https://www.worldbank.org/en/programs/lsms​/initiatives/lsms-ISA. Mason, N. M., A. Wineman, and S. T. Tembo. 2020. “Reducing Poverty by ‘Ignoring the Experts’? Evidence on Input Subsidies in Zambia.” Food Security 12 (5): 1157–72. https://doi.org/10.1007/s12571-020-01032-z. Matita, M., L. Chiwaula, E. W. Chirwa, J. Mazalale, and H. Walls. 2022. “Subsidizing Improved Legume Seeds for Increased Household Dietary Diversity: Evidence from Malawi’s Farm Input Subsidy Programme with Implications for Addressing Malnutrition in All Its Forms.” Food Policy 113. https://doi​.org/10.1016/j.foodpol.2022.102309. Maxwell, D., B. Vaitla, and J. Coates. 2014. “How Do Indicators of Household Food Insecurity Measure Up? An Empirical Comparison from Ethiopia.” Food Policy 47: 107–16. https://doi.org/10.1016/j.foodpol.2014.04.003. Maxwell, D., Caldwell, R., & Langworthy, M. (2008). Measuring food insecurity: Can an indicator based on localized coping behaviors be used to compare across contexts?. Food Policy, 33(6), 533-540. https://doi.org/10.1016/j​.foodpol.2008.02.004 Nkhata, R., C. Jumbe, and M. Mwabumba. 2014. “Does Irrigation Have an Impact on Food Security and Poverty: Evidence from Bwanje Valley Irrigation Scheme in Malawi.” IFPRI Working Paper No. 04. Lilongwe, Malawi. http://ebrary.ifpri​.org/utils/getfile/collection/p15738coll2/id/128180​ /filename/128391.pdf. Nkhoma, P. R. 2018. “The Evolution of Agricultural Input Subsidy Programs: Contextualizing Policy Debates in Malawi’s FISP.” World Development Perspectives 9: 12–17. https://doi.org/10.1016/j.wdp.2017.12.002. NSO (National Statistical Office) and ICF. 2017. Malawi Demographic and Health Survey 2015–16: Key Findings. Zomba, Malawi and Rockville, Maryland, USA. Country-Level Cases: Comparing Different Interventions 95 Pace, N., S. Daidone, B. Davis, S. Handa, M. Knowles, and R. Pickmans. 2017. “One Plus One Can Be Greater than Two: Evaluating Synergies of Development Programs in Malawi.” Journal of Development Studies 54 (11): 2023–60. https://doi.org/10.1080/00220388.2017.1380794. Poole, N., J. Donovan, and O. Erenstein. 2021. “Viewpoint: Agri-Nutrition Research: Revisiting the Contribution of Maize and Wheat to Human Nutrition and Health.” Food Policy 100: 101976. https://doi.org/10.1016/j​.foodpol.2020.101976. Smale, M., V. Thériault, and N. M. Mason. 2020. “Does Subsidizing Fertilizer Contribute to the Diet Quality of Farm Women? Evidence from Rural Mali.” Food Security 12: 1407–24. https://doi.org/10.1007/s12571-020-01097-w. Snapp, S. S and M. Fisher. 2015. “ ‘Filling the Maize Basket’ Supports Crop Diversity and Quality of Household Diet in Malawi.” Food Security 7: 83–96. https://doi​.org/10.1007/s12571-014-0410-0. Sulaiman, M. 2016. “Making Sustainable Reductions in Extreme Poverty: A Comparative Meta-Analysis of Livelihood, Cash Transfer and Graduation Approaches.” Research Gate Working Paper. https://doi.org/10.13140​/RG.2.2.18649.93286. Thome, K., E. J. Taylor, and M. Filipski. 2014. Evaluation of the 2013/14 Farm Input Subsidy Programme, Malawi: The Local Economy Impacts of FISP. Washington, DC: International Food Policy Research Institute. Tirivayi, N., M. Knowles, and B. Davis. 2016. “The Interaction between Social Protection and Agriculture: A Review of Evidence.” Global Food Security 10: 52–62. https://doi.org/10.1016/j.gfs.2016.08.004. Vaitla, B., J. Coates, L. Glaeser, C. Hillbruner, P. Biswal, and D. Maxwell. 2017. “The Measurement of Household Food Security: Correlation and Latent Variable Analysis of Alternative Indicators in a Large Multi-Country Dataset.” Food Policy 68: 193–205. https://doi.org/10.1016/j.foodpol.2017.02.006. Zingwe, D. E., L. P. Manja, and E. W. Chirwa. 2021. “The Effects of Engendered Intra-Household Power Dynamics on Household Food Security and Nutrition in Malawi.” Journal of Gender Studies 32 (2). https://doi.org/10.1080/09589236​.2021.1940110. 96 Reshaping the Agrifood Sector for Healthier Diets 98 Reshaping the Agrifood Sector for Healthier Diets 5 Future Directions There are significant opportunities to build on of newer and more holistic measures of diet this important knowledge agenda. quality, particularly the Global Diet Quality Score. Ongoing joint work of the World Bank and the Some areas for future work include a deeper Food and Agriculture Organization of the United dive on trade policy and political economy Nations (FAO) has developed and is piloting this considerations. Further analysis on trade new country-level method in Ethiopia and in the can focus on how a country’s participation in Philippines. This work looks at the overall burden international trade is linked to dietary diversity of unhealthy diets, including the economic costs and food composition within countries. This of undernutrition, as well as the economic costs work can make use of time series country-level (i.e. treatment costs and premature mortality) data capturing the contribution of individual of major diet risk factors such as overweight items to total domestic food supply (including and obesity, high blood pressure, and high the availability of various micronutrients) and fasting blood glucose. The pilot findings reveal combine this with data characterizing a country’s that both Ethiopia and the Philippines had low participation in international trade. Political overall diet quality scores. Urbanization and economy analysis can identify ways to reform higher income are associated with decreased policies towards curbing the growth in the consumption of nutrient-rich vegetables and consumption of unhealthy foods. The analysis can increased consumption of unhealthy foods. The be based on a framework that considers factors initial findings also highlight the high economic that influence policy design, implementation, and costs of unhealthy diets, particularly from child effectiveness. Successful experiences in addressing stunting in Ethiopia and from diet-related NCDs global public health issues—including parallel in the Philippines. This new method can be lessons from tobacco policies—can be further used to inform health and agrifood investment examined. Key stakeholders, and their incentives discussions, especially in low- and middle- and constraints can be identified, along with income countries (LMICs) undergoing a nutrition their roles in supporting or hindering the reform transition. process. A third area for future work is to develop Another area for future work is to develop country-level assessments of climate and country-level analyses which articulate the nutrition win-wins and tradeoffs. Doing so would hidden health costs of food systems. Doing allow for a more holistic assessment of the so would help in advocating the value of and impact of various policy levers across a range the need for repurposing agrifood policies of outcomes that policymakers need to weigh. and support for healthier diets. Healthy diets To better support policy making, there is a need have been characterized as responding to to assess diet-related synergies and tradeoffs four universal principles—nutrient adequacy, between nutrition and the environment, including dietary diversity, macronutrient balance, and through country assessments of the nutrition and moderation. Previous estimates of the economic environmental impacts of diets. Such work can cost of unhealthy diets have used nutritional assess and compare a country’s current starting status as a proxy or have made estimates of the point and targets for dietary pattern, nutritional impact of noncommunicable diseases (NCDs) by adequacy, nutritional impact of adaptation to adding up known risks of individual diet factors. climate change, environmental impact, and cost Both these methods have limitations. A new and affordability of diet scenarios. Ongoing joint proposed method to accurately estimate hidden work of the World Bank and the World Food health costs of food systems takes advantage Programme (WFP) is piloting the use of a new Future Directions 99 tool in Cambodia to assess the potential impact identifies diet combinations that minimize cost, of project-level interventions on the cost and closeness to current diet patterns, and greenhouse affordability of nutrient adequate and healthy gas emissions. Initial results in Cambodia indicate diets, and on environmental outcomes such as that there are trade-offs between greenhouse greenhouse gas emissions, water use, and land gas emissions and current consumption habits. use change. The Environment, Nutrition, and To achieve lower emissions, current consumption Health Analytics for National Consumer and habits must shift, but these shifts are within Emergency Diets (ENHANCE) tool was recently reach. The pilot in Cambodia will provide solutions developed by WFP, Capgemini, Zero Hunger Lab of across a range of food subsectors, including rice, Tilburg University, and Johns Hopkins University. fruits and vegetables, livestock, and alternative This multi-objective linear programming approach proteins. 100 Reshaping the Agrifood Sector for Healthier Diets APPENDIX A: OECD Database on Agrifood Support composed of two main elements (see table A.1 for Total Support Estimate the categories and subcategories): The Total Support Estimate (TSE) is defined as the • Market Price Support (MPS) is defined as the annual monetary value of all gross transfers from “annual monetary value of gross transfers taxpayers and consumers arising from policy from consumers and taxpayers to agricultural measures that support agriculture, net of the producers, arising from policy measures that associated budgetary receipts, regardless of their create a gap between domestic market prices objectives and impacts on farm production and and border prices of a specific agricultural income or the consumption of farm products. It commodity, measured at the farm gate level,”1 can be found by summing transfers to producers, from which we subtract price levies (LV) and consumers, and general services: excess feed costs (EFC).2,3 The MPS can be positive (in which case policy measures benefit producers) or negative (in which case policy measures are essentially an implicit tax on Producer Support Estimate producers). For example, import tariffs, import quotas, levies, licensing systems, minimum The Producer Support Estimate (PSE) encapsulates guaranteed prices, export taxes, are all MPS the annual monetary value of gross transfers policies benefiting producers via prices except from consumers and taxpayers to agricultural for the export tax. producers. This measure, calculated at the • Budgetary transfers (BT) are direct transfers farm-gate level, encompasses a range of policy from the government to farmers, enhancing measures aimed at supporting agriculture, their financial support. irrespective of their specific objectives or impacts on farm production and income. The PSE is Box A.1 Producer Support Estimate Categories and Subcategories PSE categories and subcategories A. Support based on commodity output A.1. Market Price Support (MPS) A.2. Payment based on output (Continued) Appendix A: OECD Database on Agrifood Support 101 Box A.1 Producer Support Estimate Categories and Subcategories (Continued) PSE categories and subcategories B. Payment based on input use B.1. Variable input use B.2. Fixed capital formation B.3. On-farm services C. Payments based on current area/animal number/receipts/income C.1. Based on current receipts/income C.2. Based on current area/animal numbers D. Payments based on non-current area/animal number/receipts/income, production required E. Payments based on non-current area/animal number/receipts/income, production not required E.1. Variable rates E.2. Fixed rates F. Payments based on non-commodity criteria F.1. Long-term resource retirement F.2. A specific non-commodity output F.3. Other non-commodity criteria G. Miscellaneous payments Over 60 percent of the support to producers for inefficient production locations. Consequently, the 2020–22 period was in the form of policies this has an impact on diets by changing the classified as support based on commodity type of food consumed and the quantity of each output (category A) and variable input use food group consumed. For instance, positive (subcategory B.1); these policies are considered MPS causes domestic prices for the supported the most distorting. They encourage farmers commodity to be higher than international prices, to expand production and the use of variable which leads to an increase of food consumption inputs (fertilizers, pesticides, herbicides) beyond expenditure of the supported commodity. This is what they would have produced/used otherwise. especially true for low-income consumers. As a result, these policies are linked to poor environmental outcomes such as freshwater The remaining 40 percent of the support ecosystem pollution. Furthermore, these policies includes payments based on land area, do a poor job of raising farm incomes: it has animal numbers, receipts, or income for been found that for every policy-induced dollar current (category C), non-current but linked transferred through MPS, farmers’ income to production (category D) or not (category E), increased by 25 cents (OECD 2001). Lastly, these or payments not linked to the production of policies impact the global availability of food agricultural commodities (category F), such by favoring inefficient allocation of resources as payments based on historical entitlements. (farmers might have switched crops or exited Despite not being as distortive as the previously the agricultural sector altogether), leading to mentioned policy categories, policies can distort 102 Reshaping the Agrifood Sector for Healthier Diets incentives, leading to environmental degradation on a broad group of commodities and non- and inefficient resource allocation. This is the case commodity criteria influence the decisions of for support policies coupled to current production farmers to produce a specific commodity less and (C), which create incentives to increase land allow them to respond to market signals. Second, area/livestock. This is likely to cause a negative the GCT, ACT, and OTP necessarily involve environmental impact (in the case of livestock) taxpayers who are either sources or recipients of or, at best, to have an ambiguous effect (in the the transfers. As mentioned above, MPS is included case of land area). In addition, payments to land in PSE and constitutes a share of PSCT. Lastly, the area favor a shift away from livestock production. sum of the indicator makes up the PSE: Lastly, payments based on categories D and E are considered fully decoupled. These are among the least environmentally harmful support policies and allow farmers to respond to market signals. Consumer Support Estimate Indicators of support to producers based on (CSE) commodity specificity Like the PSE, the CSE considers market transfers The OECD also uses complementary indicators and budgetary transfers such as food assistance of producer support based on commodity and cash transfer programs. Like the PSE, it can specificity. There are four indicators: Producer also be measured by commodity, in which case Single Commodity Transfers (PSCT), Group we talk about Consumer Single Commodity Commodity Transfers (GCT), All Commodity Transfers (CSCT). When expressed as a percentage Transfers (ACT), and Other Transfers to Producers (%CSE), it measures transfers to consumers as a (OTP). They measure the annual monetary share of consumption expenditure measured at transfers to producers from consumers and the farm gate. If the CSE is negative it measures taxpayers based on the production of a single the burden (implicit tax) on consumers. This can commodity, one or more commodities from a be the result of higher prices through Market designated list, any commodity, and no commodity Price Support to producers that more than at all, respectively. The payments considered for offsets consumer subsidies that lower prices to each indicator are mutually exclusive, meaning consumers. It is important to note that the MPS that the payments considered for one indicator are component of the CSE is the same one as the PSE not included in the others. For example, transfers but has the opposite sign. In that sense, the CSE to sugar producers coming from a program for usually mirrors PSE such that countries heavily which sugar (along with other commodities) is supporting consumers are generally penalizing eligible will be considered as part of the GCT their consumers. and not the PSCT. These indicators can also be expressed as a share of GFR. General Services Support The PSCT can be computed for specific Estimate (GSSE) commodities and groups, respectively. It is interesting to highlight a few aspects of these The GSSE is defined as the annual monetary value indicators. First, PSCT can be understood as the of gross transfers arising from policy measures amount of incentive distorting transfers. This is that create enabling conditions for the primary because only producers of that specific commodity agricultural sector through the development can benefit from the transfers that can make of private or public services, institutions, and producing that specific commodity relatively more infrastructure, regardless of their objectives attractive than producing another. Transfers based and impacts on farm production and income or Appendix A: OECD Database on Agrifood Support 103 the consumption of farm products. It includes as a percentage of GSSE (%GSSE), it measures policies where agriculture is the main beneficiary the share of total government support going to collectively but not those where individual agriculture. producers or consumers receive direct support. It includes spending on research and development, Spending on GSSE should be favored to spending innovation, inspection services, infrastructure on PSE because it allows producers to follow development and maintenance, marketing and market signals, create an enabling environment promotion, and public stockholding. It essentially to strengthen competitiveness, and increase measures spending on public goods or support to productivity in the long term. goods with public characteristics. When expressed Notes 1. It is important to bear in mind that the MPS measures implicit or explicit transfers arising from a policy creating a price differential, it is not a measure of public expenditure. 2. Price levies are production taxes that are imposed on producers as part of MPS. 3. Excess feed cost is a concept accounting for transfers between feed producers and livestock producers as a result of policies impacting the price of feed crops. Reference OECD (Organisation for Economic Co-operation and Development). 2001. Agricultural Policy Monitoring and Evaluation. Paris: OECD. https://doi.org/10.1787/agr_oecd-2001-en. 104 Reshaping the Agrifood Sector for Healthier Diets APPENDIX B: Global Dietary Database on Food Consumption Global Dietary Database (GDD) data are based where countries are nested within regions, allows on a series of various sources, such as food and for a nuanced analysis of global dietary patterns. consumption reports, partner information, the Food This framework accounts for differences across and Agriculture Organization of the United Nations countries and regions after adjustments for (FAO), and Demographic Health Surveys (DHSs) that demographic and geographic stratifications. This date from 1990 to 2018. As of July 2021, the latest approach is beneficial for capturing the complex GDD 2018 iteration includes 1,240 surveys, which nature of dietary patterns globally. The model were then included in the GDD 2018 model. uses cubic splines for age to address nonlinear age effects and includes country- and year-specific GDD data includes the following regions, which are covariates to fine-tune the estimates. then separated into 185 countries and economies: To ensure the reliability and relevance of its data, Asia: East Asia, Southeast Asia, Oceania, and the GDD model undergoes rigorous validation Asia-Pacific high-income countries processes, including cross-validation techniques and the assessment of global heat maps of Former Soviet Union: Central Asia, Central national mean intakes. This continuous validation Europe, Eastern Europe process is crucial for maintaining the accuracy and Latin America and Caribbean: Caribbean, applicability of the database in reflecting changing Andean Latin America, Southern Latin America, global dietary trends. Tropical Latin America, Central Latin America Middle East and North Africa: Western Europe, Middle East, and North Africa Data set construction South Asia: South Asia, Southeast Asia. To construct this comprehensive data set, Sub-Saharan Africa: Central Sub-Saharan the Organisation for Economic Co-operation Africa, Southern Sub-Saharan Africa, East Sub- and Development (OECD)’s Agricultural Policy Saharan Africa, West Sub-Saharan Africa Indicators database is integrated with the Agrimonitor database from the Inter-American Western high-income countries: Development Bank (IDB). (The OECD indicator Australasia, Western Europe, North America is used if a country appears in both databases). high-income countries. Additionally, food intake data from the 2022 Global Dietary Database (GDD) is incorporated. The GDD prediction model After appending and merging from the three sources mentioned above and dropping The GDD model incorporates multiple levels observations with missing variables on any of of analysis, considering country-specific and GDD or support variables, a total of 44 countries region-specific variations. It factors in the impact is found (see table B.1) and 258 observations for of age, sex, education, urbanicity, and other sugar and 230 observations for other four food relevant demographic details using a hierarchical groups. structure.1 The hierarchical nature of the data, Appendix A: OECD Database on Agrifood Support 105 Table B.1  List of Countries in the Analysis (N = 44) List of countries in the analysis Argentina Ecuador Japan Russian Federation Australia El Salvador Kazakhstan South Africa Bahamas, The Guatemala Korea, Rep. Suriname Bolivia Guyana Mexico Switzerland Brazil Haiti New Zealand Trinidad and Tobago Canada Honduras Nicaragua Türkiye Chile Iceland Norway Ukraine China India Panama United Kingdom Colombia Indonesia Paraguay United States Costa Rica Israel Peru Uruguay Dominican Republic Jamaica Philippines Viet Nam Note 1. At the core of the GDD is its sophisticated Bayesian multilevel prediction model. This model synthesizes the mean intake data from various surveys, adjusting for a multitude of relevant covariates. 106 Reshaping the Agrifood Sector for Healthier Diets APPENDIX C: Cross-Country Regression Estimations Table C.1  Specifications on the Choice of Specifications Support Variables The regression analyses leverage the panel of 44 countries spanning 1990, 1995, 2000, 2005, 2010, Support type Support variables 2015, and 2018 (for sugar, 2020 is also covered) Overall (1) TSE and employ two-way fixed effects to estimate supports (A) (2) PSE correlations between public support for agrifood (3) CSE commodities and consumption. The primary specification is: (4) GSSE (5) TSE, PSE, CSE, GSSE Single (1) PSCT commodity where yit is the outcome—that is, the consumption (2) MPS supports (B) of a food group (grams per day) for country i in (3) Non-MPS year t — and Xit is a vector of the agrifood support (4) Non-MPS, Non-MPS2 variables, and ut and vi are year- and country-specific (5) MPS, Non-MPS, Non-MPS2 fixed effects. Equation (1) is estimated by each food group separately for three food groups (grains, (6) PSCT, CSCT meat, sugar). Various specifications for the vector of (7) MPS, CSCT the agrifood support variables, Xit, are considered. (8) Non-MPS, CSCT Table C.1 summarizes the specifications. (9) Non-MPS, Non-MPS2, CSCT The key underlying assumption behind the (10) MPS, Non-MPS, Non-MPS2, CSCT estimation of equation (1) is Cov(Xit,eit) = 0, which, Note: CSCT = Consumer Single Commodity Transfer; CSE = in words, says that the agrifood support variables Consumer Support Estimate; GSSE = General Services Support are uncorrelated with unobservable time-varying Estimate; PSE = Producer Support Estimate; MPS = Market Price country-specific factors that may influence the Support; PSCT = Producer Single Commodity Transfer; PSE = consumption of each food group. As many Producer Support Estimate. policies and support variables are interlinked, some specifications with a single support variable in table C.1 may face biases from the nonzero covariance. Under such logic, specifications Descriptive statistics of the such as A5 and B5-10 (see table C.1.) would be regression sample less prone to the omitted bias due to possible interlinkages across policies and supports. The While table C.1 provides descriptive statistics limitation for those specifications is that the of overall support variables, tables C.2 through estimations are underpowered due to relatively C.5 provide descriptive statistics of individual large number of variables with small sample size. variables. Hence, the results face multicollinearity where the point estimates are unstable and noisy. Appendix A: OECD Database on Agrifood Support 107 Table C.2  Descriptive Statistics: Overall Support Variables 1990 1995 Support variables Mean SD N Mean SD N TSE (US$, millions) 5,344.00 12,643.98 27 7,898.57 19,248.52 27 PSE (US$, millions) 4,091.88 9,369.36 27 5,371.12 13,766.23 27 CSE (US$, millions) −2,965.42 9,355.97 27 −4,394.75 16,819.18 27 GSSE (US$, millions) 707.80 1,944.36 27 1,765.93 4,751.79 27 2000 2005 Support variables Mean SD N Mean SD N TSE (US$, millions) 8,855.05 17,967.46 27 9,202.95 18,296.31 27 PSE (US$, millions) 6,267.09 12,993.41 27 6,017.09 13,161.48 27 CSE (US$, millions) −3,545.55 11,545.84 27 −2,463.04 12,745.61 27 GSSE (US$, millions) 1,737.28 3,434.62 27 1,969.93 3,372.40 27 2010 2015 Support variables Mean SD N Mean SD N TSE (US$, millions) 9,348.00 25,388.73 43 13,101.08 43,238.18 42 PSE (US$, millions) 6,075.39 21,332.22 43 8,932.87 34,881.45 42 CSE (US$, millions) −2,883.56 23,337.53 43 −4,263.09 29,000.39 42 GSSE (US$, millions) 1,815.32 4,050.62 43 2,397.41 7,329.84 42 2018 2020a Support variables Mean SD N Mean SD N TSE (US$, millions) 13,161.76 39,364.63 37 21,810.08 56,279.73 28 PSE (US$, millions) 8,810.95 32,086.77 37 13,773.84 48,444.43 28 CSE (US$, millions) −3,609.85 28,221.69 37 −5,458.69 56,051.81 28 GSSE (US$, millions) 2,584.39 6,552.68 37 3,373.28 7,416.99 28 Note: a 2020 is available only for sugar. 108 Reshaping the Agrifood Sector for Healthier Diets Table C.3  Descriptive Statistics: Grain Consumption and Single Commodity Support Variables 1990 1995 Support variables Mean SD N Mean SD N Grains (grams per day) 245.13 182.15 27 222.12 121.72 27 PSCT (US$, millions) 1,419.96 3,947.95 27 1,753.39 5,711.70 27 MPS (US$, millions) 1,109.42 3,501.26 27 1,601.97 5,491.47 27 Non−MPS (US$, millions) 310.53 1,363.87 27 151.42 456.64 27 CSCT grains (US$, millions) −1,175.43 4,546.60 27 −2,019.19 7,283.19 27 2000 2005 Support variables Mean SD N Mean SD N Grains (grams per day) 245.52 138.48 27 259.23 148.06 27 PSCT (US$, millions) 1,641.72 4,203.39 27 1,028.75 3,732.56 27 MPS (US$, millions) 1,221.81 3636.99 27 708.08 3,303.47 27 Non−MPS (US$, millions) 419.91 1659.39 27 320.68 974.60 27 CSCT grains (US$, millions) −1,245.87 4071.05 27 −827.36 3,376.13 27 2010 2015 Support variables Mean SD N Mean SD N Grains (grams per day) 264.73 101.46 43 260.97 107.71 42 PSCT grains (US$, millions) 1,197.35 4,756.54 43 2,396.03 11,701.77 42 MPS grains (US$, millions) 998.02 4,565.13 43 2,222.82 11,612.54 42 Non−MPS grains (US$, millions) 199.34 682.03 43 173.21 757.72 42 CSCT grains (US$, millions) −1,083.81 4,331.31 43 −1,911.32 9,229.11 42 2018 Support variables Mean SD N Grains (grams per day) 258.87 110.61 37 PSCT grains (US$, millions) 1,962.38 7,077.07 37 MPS grains (US$, millions) 1487.69 6,315.09 37 Non−MPS grains (US$, millions) 474.70 2,073.12 37 CSCT grains (US$, millions) −1,316.91 5,438.37 37 Appendix A: OECD Database on Agrifood Support 109 Table C.4  Descriptive Statistics: Meat Consumption and Single Commodity Support Variables 1990 1995 Support variables Mean SD N Mean SD N Meat (grams per day) 90.91 61.14 27 85.56 48.49 27 PSCT meat (US$, millions) 327.11 736.97 27 371.03 1,415.12 27 MPS meat (US$, millions) 312.96 730.00 27 324.17 1,369.53 27 Non−MPS meat (US$, millions) 14.15 44.65 27 46.87 112.99 27 CSCT meat (US$, millions) −452.58 1,080.34 27 −512.50 2,316.52 27 2000 2005 Support variables Mean SD N Mean SD N Meat (grams per day) 82.57 40.84 27 86.98 41.96 27 PSCT meat (US$, millions) 375.51 808.79 27 646.84 1,176.50 27 MPS meat (US$, millions) 351.43 802.19 27 614.77 1,172.78 27 Non−MPS meat (US$, millions) 24.08 47.60 27 32.07 62.42 27 CSCT meat (US$, millions) −596.72 1,436.56 27 −1,006.09 1,969.48 27 2010 2015 Support variables Mean SD N Mean SD N Meat (grams per day) 95.61 45.55 43 99.15 47.82 42 PSCT meat (US$, millions) 1,352.81 4,202.55 43 1,472.82 5,850.06 42 MPS meat (US$, millions) 1,326.35 4,189.92 43 1,479.85 5,843.64 42 Non−MPS meat (US$, millions) 26.46 160.38 43 −7.03 121.32 42 CSCT meat (US$, millions) −1,679.65 5,137.33 43 −1,809.58 7,108.18 42 2018 Support variables Mean SD N Meat (grams per day) 102.45 48.71 37 PSCT meat (US$, millions) 1,677.40 6,080.75 37 MPS meat (US$, millions) 1,684.29 6,075.98 37 Non−MPS meat (US$, millions) −6.89 142.35 37 CSCT meat (US$, millions) −2,142.65 7,535.56 37 110 Reshaping the Agrifood Sector for Healthier Diets Table C.5  Descriptive Statistics: Sugar Consumption and Single Commodity Support Variables 1990 1995 Support variables Mean SD N Mean SD N Sugar (grams per day) 56.48 32.66 27 56.90 34.20 27 PSCT sugar (US$, millions) 53.12 161.23 27 −194.89 1,475.03 27 MPS sugar (US$, millions) 51.38 158.27 27 −199.21 1,488.49 27 Non−MPS sugar (US$, millions) 1.74 5.71 27 4.31 14.69 27 CSCT sugar (US$, millions) −124.29 419.53 27 33.20 1,167.79 27 2000 2005 Support variables Mean SD N Mean SD N Sugar (grams per day) 59.54 32.06 27 56.01 23.31 27 PSCT sugar (US$, millions) 252.52 418.03 27 255.28 362.45 27 MPS sugar (US$, millions) 238.57 406.68 27 246.67 353.41 27 Non−MPS sugar (US$, millions) 13.94 37.82 27 8.61 27.32 27 CSCT sugar (US$, millions) −280.58 533.71 27 −318.84 482.09 27 2010 2015 Support variables Mean SD N Mean SD N Sugar (grams per day) 53.67 25.68 43 59.87 24.66 42 PSCT sugar (US$, millions) 141.74 434.36 43 283.41 734.22 42 MPS sugar (US$, millions) 120.54 431.07 43 284.69 737.27 42 Non−MPS sugar (US$, millions) 21.21 78.06 43 −1.28 79.50 42 CSCT sugar (US$, millions) −198.51 594.85 43 −323.80 850.88 42 2018 Support variables Mean SD N Sugar (grams per day) 61.47 31.04 37 PSCT sugar (US$, millions) 444.19 1,011.70 37 MPS sugar (US$, millions) 448.90 1,014.23 37 Non−MPS sugar (US$, millions) −4.72 105.50 37 CSCT sugar (US$, millions) −502.83 1,130.57 37 Appendix A: OECD Database on Agrifood Support 111 Regression results Tables C.6 through C.11 provide regressions results. Table C.6  Regression: Grain Consumption and Overall Support (1) (2) (3) (4) (5) Support variables Grains (grams per day) TSE (US$, millions) −0.0000514 (0.0000732) PSE (US$, millions) −0.0000445 −0.00152** (0.0000723) (0.000685) CSE (US$, millions) −0.0000575 −0.00133** (0.0000935) (0.000552) GSSE (US$, millions) −0.0000842 0.00258* (0.000605) (0.00141) Constant 249.5*** 249.5*** 249.3*** 249.4*** 249.4*** (13.31) (13.34) (13.40) (13.38) (13.47) Observations 230 230 230 230 230 Number of countries 44 44 44 44 44 Elasticity −0.00202 −0.00119 0.000788 −0.000638 Note: Standard errors clustered at the country level in parentheses. Country-specific and year-specific fixed effects are included. Elasticity indicates percent change in consumption as a response to 1 percent change in the support variable of interest. Significance level: * = 0.1, ** = 0.05, *** = 0.01. 112 Reshaping the Agrifood Sector for Healthier Diets Table C.7  Regression: Grain Consumption and Single Commodity Support (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) Support variables Grains (grams per day) PSCT grains −0.0000760 −0.00522*** (US$, millions) (0.000332) (0.00182) MPS grains −0.0000604 0.00000556 −0.00649*** −0.00642*** (US$, millions) (0.000348) (0.000389) (0.00203) (0.00206) Non-MPS −0.00118 −0.00489 −0.00492 −0.00148 −0.00689 −0.00358 grains (US$, (0.000893) (0.00545) (0.00635) (0.00115) (0.00675) (0.00498) millions) Non-MPS 0.000000295 0.000000297 0.000000426 0.000000192 grains^2 (0.000000354) (0.000000415) (0.000000438) (0.000000326) Appendix A: OECD Database on Agrifood Support CSCT grains −0.00632*** −0.00767*** −0.000370 −0.000445 −0.00764*** (US$, millions) (0.00201) (0.00228) (0.000508) (0.000564) (0.00232) Constant 249.5*** 249.4*** 249.7*** 250.1*** 250.1*** 249.2*** 247.9*** 249.4*** 250.0*** 248.4*** (13.31) (13.33) (13.44) (13.47) (13.50) (13.20) (13.19) (13.39) (13.46) (13.33) Observations 230 230 230 230 230 230 230 230 230 230 Number of 44 44 44 44 44 44 44 44 44 44 countries Elasticity −0.000500 −0.000329 −0.00134 −0.00535 −0.0343 −0.0354 −0.00167 −0.00753 Source: Note: Standard errors clustered at the country level in parentheses. Country-specific and year-specific fixed effects are included. Elasticity indicates percent change in consumption as a response to 1 percent change in the support variable of interest. Significance level: * = 0.1, ** = 0.05, *** = 0.01. 113 114 Table C.8  Regression: Meat Consumption and Overall Support (1) (2) (3) (4) (5) Support variables Meat (grams per day) TSE (US$, millions) 0.000206*** (0.0000261) PSE (US$, millions) 0.000255*** −0.0000888 (0.0000183) (0.000167) CSE (US$, millions) −0.000286*** −0.000295** (0.0000448) (0.000136) GSSE (US$, millions) 0.00108** 0.000682 (0.000423) (0.000424) Constant 91.44*** 91.39*** 91.32*** 91.78*** 91.39*** (4.218) (4.214) (4.291) (4.233) (4.201) Observations 230 230 230 230 230 Number of countries 44 44 44 44 44 Elasticity 0.0220 0.0184 0.0106 0.0223 Note: Standard errors clustered at the country level in parentheses. Country-specific and year-specific fixed effects are included. Elasticity indicates percent change in consumption as a response to 1 percent change in the support variable of interest. Significance level: * = 0.1, ** = 0.05, *** = 0.01. Reshaping the Agrifood Sector for Healthier Diets Table C.9  Regression: Meat Consumption and Single Commodity Support (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) Support variables Meat (grams per day) PSCT meat 0.00132*** -0.00108 (US$, millions) (0.000154) (0.00257) MPS meat (US$, 0.00132*** 0.00132*** -0.00107 -0.000955 millions) (0.000153) (0.000155) (0.00271) (0.00265) Non-MPS meat -0.00280 0.00190 0.00402 -0.00129 0.00347 0.00305 (US$, millions) (0.0121) (0.0152) (0.0151) (0.0114) (0.0151) (0.0151) Non-MPS -0.0000194 -0.0000220 -0.0000197 -0.0000179 meat^2 (0.0000251) (0.0000245) (0.0000246) (0.0000247) Appendix A: OECD Database on Agrifood Support CSCT meat -0.00202 -0.00202 -0.00113*** -0.00113*** -0.00192 (US$, millions) (0.00217) (0.00224) (0.000132) (0.000133) (0.00229) Constant 91.98*** 91.96*** 91.88*** 92.06*** 92.14*** 91.88*** 91.90*** 91.90*** 92.09*** 92.05*** (4.255) (4.255) (4.512) (4.512) (4.214) (4.251) (4.240) (4.209) (4.183) (4.215) Observations 230 230 230 230 230 230 230 230 230 230 Number of 44 44 44 44 44 44 44 44 44 44 countries Elasticity 0.0141 0.0139 -0.000490 0.000223 -0.0115 -0.0112 -0.000225 0.000496 Source: Note: Standard errors clustered at the country level in parentheses. Country-specific and year-specific fixed effects are included. Elasticity indicates percent change in consumption as a response to 1 percent change in the support variable of interest. Significance level: * = 0.1, ** = 0.05, *** = 0.01. 115 116 Table C.10  Regression: Sugar Consumption and Overall Support (1) (2) (3) (4) (5) Support variables Sugar (grams per day) TSE (US$, millions) 0.0000131 (0.0000300) PSE (US$, millions) −0.0000157 −0.000110 (0.0000335) (0.000205) CSE (US$, millions) 0.0000518 0.0000364 (0.0000778) (0.000200) GSSE (US$, millions) 0.000397 0.000986** (0.000490) (0.000461) Constant 57.86*** 57.93*** 58.01*** 57.82*** 58.04*** (4.167) (4.149) (4.138) (4.181) (4.109) Observations 258 258 258 258 258 Number of countries 44 44 44 44 44 Elasticity 0.00254 –0.00203 –0.00329 0.0142 Source: Note: Standard errors clustered at the country level in parentheses. Country-specific and year-specific fixed effects are included. Elasticity indicates percent change in consumption as a response to 1 percent change in the support variable of interest. Significance level: * = 0.1, ** = 0.05, *** = 0.01. Reshaping the Agrifood Sector for Healthier Diets Table C.11  Regression: Sugar Consumption and Single Commodity Support (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) Support variables Sugar (grams per day) PSCT sugar (US$, −0.00208 −0.0160* millions) (0.00319) (0.00891) MPS sugar (US$, −0.00212 −0.00243 −0.0162* −0.0204** millions) (0.00311) (0.00304) (0.00861) (0.00824) Non-MPS sugar 0.00217 −0.0198 −0.0214 0.00301 −0.0193 −0.0399** (US$, millions) (0.0243) (0.0146) (0.0168) (0.0244) (0.0153) (0.0182) Non-MPS sugar^2 0.0000698*** 0.0000763*** 0.0000713*** 0.000102*** (0.0000176) (0.0000173) (0.0000167) (0.0000263) Appendix A: OECD Database on Agrifood Support CSCT sugar (US$, −0.0158* −0.0159* 0.00124 0.00138 −0.0202** millions) (0.00922) (0.00948) (0.00325) (0.00327) (0.00951) Constant 57.88*** 57.89*** 57.91*** 57.68*** 57.67*** 57.04*** 57.13*** 57.97*** 57.75*** 56.55*** (4.234) (4.242) (4.175) (4.182) (4.237) (4.108) (4.191) (4.237) (4.246) (4.177) Observations 258 258 258 258 258 258 258 258 258 258 Number of 44 44 44 44 44 44 44 44 44 44 countries Elasticity −0.00816 −0.00792 0.000411 −0.00347 −0.0626 −0.0605 0.000572 −0.00338 Source: Note: Standard errors clustered at the country level in parentheses. Country-specific and year-specific fixed effects are included. Elasticity indicates percent change in consumption as a response to 1 percent change in the support variable of interest. Significance level: * = 0.1, ** = 0.05, *** = 0.01. 117