TRADE, INVESTMENT AND COMPETITIVENESS TRADE, INVESTMENT AND COMPETITIVENESS EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT Unfair Advantage: Distortive Subsidies and Their Effects on Global Trade © 2023 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org 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, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. 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 The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. Photo credits: Shaunl, iStock >>> Contents Acknowledgements 6 Abbreviations 7 Executive Summary 9 1. Introduction 15 2. Scope and Defining Characteristics of Subsidies 20 3. The Extent of Subsidy Measures 22 3.1 A New Compilation of Subsidy Measures 25 3.2 Global Landscape of Measures 26 Regional Patterns 27 Sectoral Patterns 28 Types of Measures and Objectives 29 3.3 Effects of Subsidies on Trade 32 Econometric Estimates 33 Implications for Developing Countries 37 4. International Disciplines and Gaps on Subsidies 40 4.1 Existing Rules 41 World Trade Organization 41 EU State Aid Law 42 Select Free Trade Agreements 42 >>> Contents 4.2 Gaps and Limitations in the Current Subsidy Governance 44 Framework Limitations Related to the Scope of the Current Disciplines 44 Limitations Related to Current Substantive Rules 45 Limitations Related to Governance and Other Systemic 46 Concerns 5. Moving Ahead on Subsidies and Trade 47 Appendix A. Antidumping Measures 49 Appendix B. Subsidy Measures Classification 51 Appendix C. Subsidy Objectives (Broad Categorization) 54 Appendix D. Summary Description of Subsidy Values 56 Appendix E. Additional Information on Econometric Analysis 57 Appendix F. Effects of Reforming Trade-Distorting Subsidies 60 References 63 >>> Boxes Box 1.1. Several main insights from this work 19 Box 3.1. Subsidy Rationales and Their Possibly Distortive Effects 31 >>> Figures Figure 1.1 Subsidies have been escalating as global trade has slowed down over 16 the last decade Figure 1.2 Most subsidies are implemented by large economies 17 Figure 3.1. Subsidies have been the most frequent form of intervention 22 Figure 3.2. New unilateral barriers to subsidized imports have accelerated 23 Figure 3.3. China has emerged as the principal target of countervailing measures 24 Figure 3.4. The top three trading blocs account for most subsidies 27 Figure 3.5. Subcentral government subsidies account for the largest share of 27 measures Figure 3.6. Support to manufacturing predominates 28 Figure 3.7. Subsidies take various forms depending on the economy 30 Figure 3.8. Support to improve sectoral competitiveness is a principal objective 30 Figure 3.9. Trade effects of subsidies are particularly significant for agriculture 35 Figure 3.10. Certain agricultural and manufacturing sectors exhibit significant 36 trade effects Figure 3.11. Low-income countries are particularly exposed to distorted 38 agricultural trade Figure 3.12. Middle-income countries are particularly exposed to distorted 39 manufacturing trade >>> Tables Table 3.1. The scope of products subject to countervailing measures is 24 diversifying Table 3.2. Economies Selected for the Database 26 Table 3.3. Top affected sectors tend to vary for each main user 29 Table 3.4. Econometric estimates of bilateral trade and the effects of subsidies 34 by broad sectors >>> Acknowledgments This study was prepared by a team led by José Signoret (Task Team Leader) and Martin Molinuevo (co-Task Team Leader), and composed of Milla Cieszkowsky, Lillyana Daza Jaller, Giulia Jonetzko, Jesse Kreier, Panagiotis Kyriakou, Mario Larch, Fernando Nieto, Maria Seara e Pereira, Jan Remy, Serge Shikher, Tommaso Soave, Marinos Tsigas, and Yoto Yotov. The team acknowledges comments and inputs provided at different stages by Chad Bown, Chiara Bronchi, Ndiame Diop, Simon Evenett, Caroline Freund, Jennifer Hillman, Robert Koopman, Hania Kronfol, Martha Licetti, and Georgiana Pop. Chris Wellisz provided editorial review. The team benefited from helpful comments received during an internal workshop and an internal seminar at the World Bank and during a trade seminar at the US International Trade Commission. The team acknowledges the support of the World Bank management team, including Mona Haddad (Global Director, Trade, Investment and Competitiveness) and Antonio Nucifora and Sebastien Dessus (Practice Managers, Global Trade and Regional Integration) throughout the process of preparation. This activity was funded by the multidonor Umbrella Facility for Trade. >>> Abbreviations AoA Agreement on Agriculture AP Asia-Pacific ASCM Agreement on Subsidies and Countervailing Measures AVE ad valorem equivalent CES constant elasticity of substitution CGE computable general equilibrium The Comprehensive and Progressive Agreement for Trans-Pacific CPTPP Partnership EU European Union FDI foreign direct investment FTA free trade agreement GATS General Agreement on Trade in Services GATT General Agreement on Tariffs and Trade GDP gross domestic product GTA Global Trade Alert GTAP Global Trade Analysis Project IMF International Monetary Fund International Standard Industrial Classification of All Economic ISIC Activities ITPD-E International Trade and Production Database for Estimation HS harmonized system LA Latin America LAC Latin America and the Caribbean MENA Middle East and North Africa >>> Abbreviations cont. n.e.c. not elsewhere classified NA North America NAFTA North American Free Trade Agreement NTM nontariff measure OECD Organisation for Economic Co-operation and Development prod. products PSE producer-support equivalent PPP purchasing power parity R&D research and development ROW rest of world RTAs regional trade agreements SCM Subsidies and Countervailing Measures SMEs small and medium enterprises SOEs state-owned enterprises SSA Sub-Saharan Africa TFEU Treaty on the Functioning of the European Union TRIMs Agreement on Trade-Related Investment Measures UK United Kingdom UNCTAD United Nations Conference on Trade and Development USA United States USMCA United States-Mexico-Canada Agreement WTO World Trade Organization >>> Executive Summary Government subsidies have emerged as a flashpoint in the geopolitical tensions that threaten to undermine the global rules-based trading system. Subsidies are among the tools deployed by the world’s major economies to tackle climate change, gain an edge in frontier technologies, and revamp supply chains disrupted by the COVID-19 pandemic and the war in Ukraine. Yet, even when deployed in pursuit of legitimate goals, subsidies can harm trading partners, fuel tensions, and provoke countermeasures. They can nullify the benefits of global trade and investment by distorting international prices and limiting market access, as in the case of local content requirements, and they can create inefficiencies in global value chains. Because the international trading system is ill-equipped to discipline the use of subsidies, governments increasingly are responding with countervailing tariffs or subsidies of their own when they can afford it. Subsidy programs—which have been rising since the 2008 global financial crisis—can create substantial distortions to trade: • Subsidies are concentrated among the world’s biggest trading economies, including China, the European Union (EU), and the United States; • Manufacturing has emerged as the principal beneficiary, surpassing agriculture, in terms of number of measures; • Subsidies can be more distortive to trade than tariffs—ad valorem equivalents of export support average 15 percent for agriculture and 8 percent for manufacturing. This is double and quadruple average tariffs (at 8 percent and 2 percent), respectively. These developments risk triggering a tit-for-tat cycle of rising barriers and trade- distorting subsidies that pose a threat to global economic prosperity. The economic toll has yet to be assessed, but preliminary data on global trade and investment flows suggest that induced efficiency costs could be significant. Trade has been a critical channel for the diffusion of technology, which in turn drives productivity and economic growth. Higher barriers to trade mean less competition and therefore lower efficiency and slower growth. More rigid and localized supply chains would be more vulnerable to shocks, potentially creating shortages and fueling inflation. Geopolitical tensions could lead to restrictions on flows of credit and investment, creating financial stability risks. 1. See, for example, Irwin (2019). EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 9 Developing countries are most at risk from a loss of ones need to be addressed. The amount of support and the confidence in global trade rules. To attract much-needed conditions under which a subsidy is granted—for example, investment and establish or strengthen links with global value whether it is linked to production levels—can be important chains, developing countries need the certainty provided by a determinants of their effects across borders. Furthermore, credible and coherent system of global trade rules. They need due to the paucity of subsidy data, the magnitude of their international trade to drive growth, reduce poverty, diversify trade effects remains understudied. their economies, and respond to rising challenges such as climate change. Perhaps most importantly, they need to be able to compete at fair terms, a capacity that trade-distortive Subsidies have proliferated subsidies and protectionist policies will hurt. With limited fiscal and institutional capacity, they often cannot respond with their Since the global financial crisis of 2008, subsidies have own subsidies or trade defense actions. far surpassed tariffs and nontariff measures as the most frequent form of policy intervention in trade. The rising Information on the extent and economic effects of number of measures has coincided with a marked slowdown subsidies is hard to come by, making it difficult to regulate of global trade from its pre-crisis pace. Countries principally their use under international trade rules. Governments involved in subsidy interventions include large economies lack the information they need to have informed discussions in the Group of 20, which have the fiscal and administrative and decision making on subsidies that may distort trade. capacity to increase the use of subsidies. The magnitudes are Because subsidies serve various objectives and take many significant. For instance, the support provided to agriculture forms—including direct financial grants, tax breaks, and by governments across a wide range of countries exceeded in-kind contributions—it may not be easy to identify which US$600 billion per year from 2016 to 2018. >>> Figure ES.1. Subsidies are concentrated in the top trading blocs and predominate for manufacturing in terms of number of measures (Distribution by number of measures, 2018) 35 30 25 30 20 Percent 25 15 10 20 Percent 5 15 0 Manufacturing technical activities Professional, scientific, and fishing Agriculture, forestry, and communication Information and Electricity and gas 10 Other sectors 5 0 EU United China Canada Australia Others States Source: Compilation of subsidy measures. Note: EU = European Union (EU28); “Others” combines other 20 major trading economies. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 10 This analysis builds a framework to identify trade- More than 20 percent of subsidy programs appear to be distorting subsidies, monitor them, and offer potential aimed at supporting the competitiveness of a particular solutions. It develops a working definition and classification sector. A second major category consists of incentives to help of subsidy measures and builds a global database covering generate or transfer technology, such as capital and rewards more than 2,000 programs in 50 countries accounting for for innovation; knowledge-sharing and technology-transfer more than 70 percent of global trade. This database provides platforms; financial and business support for start-ups; and a picture of subsidies as of 2018 in terms of intended policy capacity building. Subsidies with an environmental goal are objectives, types of measures, and regional and sectoral less common. They often include rewards for conserving patterns. The analysis breaks new ground by quantifying the energy and reducing pollution by the use of coal filters, for trade-distortive effects of subsidies. example, or by substituting natural gas for coal. Nevertheless, fossil fuel subsidies still outnumber those for clean energy. These new data show that subsidies are not a single- Certain interventions have emerged recently with stated country or single-region problem. China, the EU, and the climate objective that remains to be assessed as they unfold, United States—big economies with the potential to influence but that have already raised concerns for their possible trade global markets—account for about 75 percent of programs of effects. These include large subsidies to promote clean energy the documented number of measures. Australia and Canada in the United States and the EU, under the Inflation Reduction are also among the world’s top subsidizers. Most government Act and the Green Deal Industrial Plan, respectively. programs are implemented at the sub-national level. In China and the United States, sub-national subsidies account for 90 percent and 82 percent of programs, respectively. Thus, all Subsidies can distort trade levels of government need to be aware of and subscribe to established disciplines on subsidy provisions and acknowledge Subsidies create trade-distorting effects for both the need for full transparency. agriculture and manufacturing exports. The sectoral inventory of measures developed here provides useful Subsidies predominate in manufacturing. They are also information on sectors countries consider strategic. Based common in certain services and in agriculture. Subsidies on their stated objectives, these interventions are often benefiting manufacturing account for about 25 percent of put in place with the likely intention of benefiting sectoral programs. They are prevalent in industries such as electronics, competitiveness, and less clearly with the goal of addressing vehicles, machinery, ships, chemicals, food and beverages, a market failure or a legitimate social concern. This is an and metals and metal products. The second largest category important distinction, not only because of the uneven playing encompasses services related to professional, scientific, and field that certain subsidies can generate but also because of technical activities (primarily R&D), followed by agriculture their often-large opportunity costs. New empirical estimates— and fishing in third place. Together, those three top sectors derived from a structural, general equilibrium gravity account for about 60 percent of subsidies. Other sectors, model—show that subsidies can skew trade and production with less concentrated programs, include electricity and significantly for both agriculture and manufacturing. In gas, mining and quarrying, and various services such as particular, the introduction of a subsidy program can boost information and communication, construction, and financial exports by, on average, between 3 and 4 percent. and insurance activities. Subsidies can be more distortive to trade flow than Measures vary considerably by type and stated objective. existing tariffs barriers. The distortionary effect of subsidies In the EU, about six in 10 take the form of financial grants. In the on trade, expressed in ad valorem equivalents, is estimated United States, tax incentives predominate. In both economies, at 15 percent for agriculture and 8 percent for manufacturing. about 20 percent of measures are loans, guarantees, or These estimates suggest that subsidies, while not necessarily capital injections. In China, most programs provide grants. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 11 intended to influence trade, can have more pronounced comparison to tariff barriers are even larger. Furthermore, effects on trade flows than border measures applied directly these effects differ by region. The biggest distortions in to trade flows, such as tariffs barriers. On average, subsidies agriculture originate in the advanced economies of the EU in agriculture can be almost twice as distortive to trade as and North America. Subsides in Asia disproportionately affect agricultural tariffs. For manufacturing, the distortions in trade in manufactures. >>> Figure ES.2. Subsidies can distort trade more than tariff barriers (Estimated ad valorem equivalents for tariffs and subsidies) 16 14 12 10 Percent 8 6 4 2 0 agriculture ma nufacturing Tariffs Subsidies Source: UNCTAD TRAINS Database and original estimates of the trade effects of subsidies. Trade-distorting subsidies can displace trade and Developing countries are especially vulnerable to the production in other trading partners, with important trade-distortive effects of subsidies. Their production and repercussions for developing countries. Beyond possibly trade tend to be concentrated in agriculture, where distortions being distortive and inefficient for domestic objectives, caused by subsidies are greatest. In much of Sub-Saharan subsidies can have negative spillovers across borders. The Africa, products for which subsidies are found to be particularly uneven playing field they create can affect the commercial trade distortive—such as fresh fruits and vegetables, pulses interests of countries that compete with subsidized products and legumes, nuts, cotton, and meat—represent more than in foreign markets. This does not need to be only in direct 20 percent of goods exports. In manufacturing, subsidies competition in the same market, as a disproportionally large cover a wide range of products and significantly affect a large number of subsidy programs are implemented by large share of international trade, with particularly pronounced trading countries with the potential to influence world prices. effects in certain types of machinery and equipment, parts Subsidies by other countries may also affect domestic sales and components, and miscellaneous manufactures. These in industries that cannot compete on fair terms with such distortive subsidies affect middle-income countries—including subsidized imports. These channels of uneven competition Costa Rica, Mexico, and the Philippines—involved in can be significant for developing countries. manufacturing value chains. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 12 >>> Figure ES.3. Low- and middle-income countries are particularly exposed to distorted agriculture and manufacturing trade (Shares for distorted agricultural trade) (Shares for distorted manufacturing trade) 100 100 90 90 GNB 80 80 PHL MEX 70 COM 70 CRI CHN 60 AFG 60 THA BWA Percent Percent CIV VNM MMR 50 50 MWI ETH KEN TON GMB UGA CMR ZWE ECU WSM TUN SRB 40 GRD 40 CAF LBN 30 CAF DJI GTM 30 MAR ZAF IND MKD MDG DOM BIH BDI LBR BFA BEN RWA TLS GHA COL NRUBRB PSE IDN BGR TZA TKM 20 HND IRN 20 10 10 0 - 6.0 7.0 8.0 9.0 10.0 11.0 12.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 Log GDP per capita, PPP Log GDP per capita, PPP Tra de Production Tra de Production Source: Original calculations for this study. Note: GDP = gross domestic product; PPP = purchasing power parity. Eliminating trade-distorting subsidies in agriculture could multilateral level, a multiplicity of interests, deep differences enhance agricultural output in most parts of the world. In the of views among countries, and the broader gridlock in the short and medium term, agricultural output would increase in all World Trade Organization (WTO) negotiating mechanism regions, except for North America and the EU, between 1 and 3 have barred any meaningful changes to WTO rules. As a percent. Global agricultural supply would fall by just 0.1 percent. result, the governance system for subsidies is a patchwork Similarly, agricultural exports also would increase in all regions of overlapping instruments, with the most far-reaching reform except for North America and the EU. The largest expansions efforts taking place at plurilateral or bilateral levels, particularly would be for the Asia Pacific region followed by Sub-Saharan in the context of free-trade agreements. Africa, where agricultural exports would increase by 8 and 4 percent, respectively. For manufacturing, eliminating subsidies One glaring gap in the existing subsidies rules is the virtual would reduce production and exports in Asia. Production would absence of services—the fastest-growing portion of trade. increase in Sub-Saharan Africa (3 percent), Middle East and Investment-distorting subsidies, which could lead to trade North Africa (2 percent), and Latin America and the Caribbean distortions, are also not covered. Moreover, WTO rules do not (1 percent) to partially cover the lower manufacturing trade. explicitly apply to enterprises owned or controlled by the state, Given the initial levels, the largest expansion in dollar terms which are important economic actors in many countries. The would be in Latin America and the Caribbean. rules are deemed to cover economic transfers by government entities that are vested with authority to exercise governmental functions. They are also deemed to cover private bodies that Global rules are ill-equipped to deal have been entrusted or directed by a government to make a financial contribution. Demonstrating such direction by the state, with subsidies however, may be particularly challenging, especially where measures are either unwritten or not formally communicated. Existing subsidy disciplines have not kept pace with the rising challenges of global economic governance. At the EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 13 The paucity of information on subsidies is another enforced, and remedies could be strengthened to include impediment to action. While governments are required to the recovery of subsidies or even monetary compensation. notify the WTO of subsidy programs, notifications are often However, the dispute settlement system is under severe late, incomplete, or missing entirely. State and provincial pressure, given the paralysis of the Appellate Body, and governments, which account for most subsidies in many consensus even on broad reform directions is lacking. A shift countries, are often poorly informed about or unaware of to more prescriptive rules, based on prohibitions or rebuttable global trade rules. Where data are available, they are often not presumptions for certain types of subsidies, would appear to comparable—a shortcoming that this study helps to address. be a more viable alternative. Moreover, the WTO dispute settlement system has Strengthening transparency and analysis on subsidies limitations in effectively addressing challenges to extended by major global trade players will be critical. subsidies. Outright prohibitions apply only to subsidies More data and analysis will be needed to understand the contingent on the volume of exports or the use of local trade-distorting effects of subsidies and to inform the debate content. Other types of subsidies can be challenged, but this on reforms. Understanding which subsidy measures are rarely happens. A member must wait for the harm, possibly inefficient or ineffective in achieving their intended goals and irreparable, to materialize before bringing a challenge. In their potential effects on other countries will require more addition, there may be scant incentive to do so. Gathering and detailed information on the design, implementation, and analyzing evidence to prove adverse effects is difficult and beneficiaries of such interventions. This may require enhanced costly, and remedies do not reverse the subsidy measure or cooperation in fulfilling notification requirements, as well as offer monetary compensation for damage inflicted. other actions by country peers. International organizations can play a facilitating role and provide guidance as to how best to There are ways to strengthen the dispute settlement consolidate this information and monitor subsidies within and system. For example, deadlines could be shortened or strictly across countries. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 14 1. >>> Introduction The expansion in trade and global value chains (GVCs) over the past 30 years contributed to higher growth and an unprecedented reduction in poverty.1 From 1989 to 2018, GDP growth in low- and middle-income countries averaged 4.4 percent a year; their trade as a share of GDP increased from less than 30 percent in 1987 to between 50 and 60 percent in the early 2000s. Meanwhile, the proportion of people in extreme poverty ($2.15 a day, PPP terms) declined from 43.6 percent in 1981 to 8.4 percent in 2019 The benefits of trade operate through various channels. Trade integration can boost productivity by shifting production toward sectors and firms with greater comparative advantage and higher efficiency, expanding markets and creating opportunities for overall output and employment. Increased trade also makes available a wider range of intermediate inputs, lowering costs and increasing firm competitiveness, and enhancing investment, technological spillovers, innovation, and other dynamic effects that can cumulate over time. On the consumer side, greater trade integration can also lead to lower prices and a wider and increasing variety of goods and services, benefiting consumers through higher real incomes and a greater choice of consumption goods and services. Despite these gains, long simmering skepticism about the benefits of free trade is intensifying. Critics blame trade and GVCs for the loss of manufacturing jobs in advanced economies, environmental degradation, and disruptions to supplies of vital goods like vaccines. These concerns, combined with geopolitical tensions, are prompting major players to raise barriers to trade and investment and to subsidize the domestic production of goods deemed essential and strategic. With shades of economic nationalism, this setback to multilateral integration could become a source of unwarranted inefficiencies and a drag on the global economy. Developing countries have the most to lose from these increased trade tensions and protectionist policies. To attract investment, they critically need the certainty provided by a credible and coherent system of global trade rules. Smaller developing economies cannot be 1. See, for example, Irwin (2019). EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 15 self-sufficient and need to export to import. They critically global value chains. In many cases, it is difficult to categorize need to be able to compete at fair terms, a capacity that ex ante a subsidy measure as trade distortive. In some trade distortive subsidies and protectionist policies will hurt. cases, however, subsidies can be intended explicitly to Furthermore, many developing countries lack the fiscal skew competitive opportunities in favor of certain groups resources to counter steps taken by advanced countries to of economic operators to the detriment of others. These subsidize domestic production. measures are of particular concern because of their potential beggar-thy-neighbor effects. Ultimately it is an empirical While recent subsidy programs by major economies have question whether the receipt of subsidies has a statistically exacerbated international trade frictions, government and economically significant trade-distorting effect. And this subsidies often have legitimate economic and social is a key question addressed in the current report. goals. They can be used to reduce market failures. If properly designed and implemented, subsidies may lead to favorable Subsidies have proliferated as a key policy lever over the outcomes, such as creating employment in marginalized past decade, as the growth of global trade has slowed areas, ensuring adequate food or medical supplies in a crisis, down. Data on government interventions starting in the 2008 or achieving environmental targets. More broadly, they can global financial crisis show that subsidies have been principal be used to align market equilibriums with appropriate social interventions during the period. The stock of subsidy measures goals. Thus subsidies, like taxes, can be an important policy in effect per year has been increasing continuously since the tool for governments and should not be banned ex ante. collapse of world trade in 2008. This coincides with the period over which global trade has slowed down, especially relative Yet even when put in place on the most solid to the rapid trade expansion prior to the global financial crisis socioeconomic or environmental grounds, subsidies (figure 1.1). Furthermore, these subsidies, while documented can still harm trading partners. They can nullify the benefits in a wide range of countries, are implemented for the most of global trade and investment by distorting international part by large economies. The G20 countries, for example, prices and limiting market access, as in the case of local consistently represent about 90 percent of all global subsidies content requirements, and they can create inefficiencies in by count (figure 1.2). >>> Figure 1.1 Subsidies have been escalating as global trade has slowed down over the last decade 30 9,000 8,000 25 7,000 Number in effect 20 6,000 Trillion USD 5,000 15 4,000 10 3,000 2,000 5 1,000 0 - 99 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 19 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Subsidy measures (right axis) Wo rld exports Source: World Bank World Development Indicators and Global Trade Alert, World Bank staff calculations. Note: Collection of subsidy interventions starts in 2009. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 16 >>> Figure 1.2 Most subsidies are implemented by large economies 9,000 8,000 7,000 Number in effect 6,000 5,000 4,000 3,000 2,000 1,000 - 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 G20 economies Other Source: Global Trade Alert, World Bank staff calculations. Subsidies and their likely trade-distorting effects are Rising tensions and counteractions point to the need an area of growing multilateral friction, which has been to address subsidies at the multilateral level. For the exacerbated by major programs introduced by large, global trade community to come together, it is crucial to find advanced economies. Indeed, subsidies are the largest a way forward on addressing subsidies in goods as well as in category of intervention, far exceeding tariff and trade remedy services, and in related policy areas such as competition policy. measures.2 Moreover, countries are increasingly resorting to The potential for subsidies to distort the level playing field and unilateral trade remedies in reaction to subsidies put in place to heighten the perceived inadequacy of existing systems by others. Updated data on trade defense barriers show that that govern these interventions add pressure to the current interventions in the form of antidumping and countervailing global trading system, with potentially dire consequences for actions have grown significantly over the years.3 These duties, prosperity, especially in the developing world. while highly concentrated by sector, are being applied more often and by a larger number of countries, affecting a growing Yet subsidies remain a difficult issue in the multilateral share of global trade. arena. They can take many forms and serve various objectives, so it may not be easy to identify which ones Developing countries are most vulnerable to the trade- need to be addressed. They can include grants or in-kind distorting effects of subsidies and other forms of contributions, tax or tariff incentives, preferential financing, intervention in the exchange of goods and services provision of services at lower rates, bailouts, or support across borders. Developing countries rely on trade to drive prices. How they take place depends in part on country- economic growth, reduce poverty, diversify their economies, specific fiscal or administrative capacities. It is therefore and respond to rising challenges such as digitalization and important to establish definitions (or categories) for various climate change. And any fracturing of the global, rules-based types of subsidies for purposes of identification, monitoring, trading system would be especially harmful to developing impact evaluation, and negotiation. countries, which rely on the certainty that the system provides to attract investment. 2. Global Trade Alert database. See figure 3.1 for the relative frequency of subsidies versus other documented government interventions during the period. 3. See World Bank, “Temporary Trade Barriers Database,” https://www.worldbank.org/en/data/interactive/2021/03/02/temporary-trade-barriers-database, for a data visual- ization of these temporary trade barriers across countries and over time. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 17 A subsidy, in a broad economic sense, is an action taken multilateral system, with two-way accusations and diverging by a government to make a particular good or service views and goals as to how to discipline them. Transparency more abundant at a lower price or raise the income of a issues on the use of subsidies still predominate, thus little is producer. However, governments deploy a variety of policy known about their effects on global trade and the economy. instruments to implement subsidies, and each of these has different direct and indirect effects. Because markets are Using both quantitative and qualitative analyses, this study interconnected, the same policy objective may be achieved aims to expand the evidence base and point to potential by paying consumers or producers and may be implemented policy recommendations. Several insights that emerge are using various instruments at different points in the supply highlighted in box 1.1. Among the study’s goals are: chain (IMF, OECD, World Bank, and WTO 2022). • To develop a broader working definition and classification Beyond definitions, lack of transparency on the extent of or categorization of various subsidy measures and to government subsidies poses a major hurdle in moving make this nomenclature publicly available to researchers ahead with negotiations. Because there is lack of clarity as and data suppliers; to what may constitute subsidy measures by a government, • To build a new database on subsidies across many there is also a general lack of transparency about their countries collected in a uniform way; occurrence. Data on subsidies are scarce and much more • To describe the global landscape of identified measures incomplete than information on other policy instruments such in terms of utilization by countries, sectors, and types of as tariffs and certain nontariff measures. Filling these data subsidy measures and their intended goals to provide a gaps needs to be a priority for the international community. broad picture of what it is known and still unknown on the Better data on subsidies would support analysis and inform subject matter; the discussion on the international spillovers of subsidies and • To develop a new analytical framework, building on the need or desire to design new rules. Furthermore, a lack of the new data, to estimate the trade-distortive effects of transparency significantly hinders parties from establishing a subsidies across countries and sectors using a structural, negotiating baseline. general equilibrium gravity model; and • To assess the international regulatory environment on This study aims to lay the groundwork and make significant the use of subsidies interfering with trade, identifying progress in describing different interventions that can important gaps. be considered as subsidies and analyzing their trade- distorting effects. Subsidies are a matter of concern in the EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 18 BOX 1.1. SEVERAL MAIN INSIGHTS FROM THIS WORK Whereas distinctions and differences exist on what should be counted as a subsidy measure or disciplined under existing or new rules, some observed patterns can be helpful in discussions of subsidies at the multilateral level. • Transparency on subsidies remains an issue around the world, with significant data gaps on the provision of this support. For example, interventions provided by nongovernmental entities, such as state-owned enterprises (SOEs), can be important and generally difficult to identify. Furthermore, information on the beneficiaries of subsidies and on the depth of these interventions is often unclear or unavailable. • Newly collected data suggest that the incidence of subsidies is highly concentrated in a relatively small number of countries or trading blocs. The European Union (EU), the United States, and China together account for three out of every four measures globally. Given their disproportionally large share of global trade, subsidies by these top trading partners can possibly have significant cross-border effects. • Subsidies are applied very broadly. They go beyond agriculture, where their use poses a clear problem, to include manufacturing and services. Manufacturing is the sector in which the most measures are applied. But support in agriculture is prevalent, and subsidies to innovation services are also common. • Most subsidies are implemented by subcentral government entities, such as states and provinces. It highlights not only the importance of enhancing transparency in government support measures at subnational levels, but also the importance of ensuring that the applications of these adhere to prevailing international rules. • Subsidies in in most cases are used as an industrial policy tool to support potentially strategic sectors. Many of these subsidy programs are intended to provide financial support for the acquisition of capital goods and materials and for business operations. Subsidies with an environmental objective are relatively less common, and many are oriented toward “cleaner” fossil fuels, such as natural gas. • New estimates in this study show that subsidies can disproportionately affect international trade. Subsidies for both agriculture and manufacturing can displace trade and production in developing countries and limit their participation in global value chains. The largest trade-distorting effects are in agriculture, with an estimated ad valorem equivalent of about 15 percent. Subsidies for agriculture distort the playing field particularly against low- income countries. The advanced economies of the EU and North America account for the largest distortive effects in agriculture. • Subsidies to manufacturing are also distortive to trade, with an estimated ad valorem equivalent of 8 percent. Asia is the largest source of trade-distorting subsidies for manufacturing. • The prevalence of subsidies and their significant trade effects contrast with deficiencies in the multilateral framework governing them. The current arrangement to apply trade defense ex post, after harmful effects are done and determined, seems inadequate to solve the root causes, because it does not remove the distortions in global markets. Many importers, such as most countries in Sub-Saharan Africa, do not have the legal or institutional infrastructure to respond with trade remedies. • Many actions related to subsidies are also not covered, or clearly covered, by prevailing rules. Disciplines on subsidies related to cross-border services and foreign direct investment (FDI) are virtually nonexistent. And subsidies to and from nongovernmental entities, including SOEs, also fall outside most international regulations. • Increased transparency and international cooperation will become even more important and should be strengthened. This is especially the case after disruptions to supply chains caused by the COVID-19 pandemic put pressure on governments to support certain critical industries. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 19 2. >>> Scope and Defining Characteristics of Subsidies Understanding what may constitute a subsidy, or at least recognizing that different definitions apply in different circumstances, is a precondition to analyzing data on subsidies. Similarly, a definition and classification of the types of government interventions that may act as a subsidy to trade and investment are needed to determine how those measures are regulated by international trade rules. For example, the definition of a subsidy under the World Trade Organization (WTO) Agreement on Subsidies and Countervailing Measures (ASCM or SCM Agreement) does not capture several government interventions, such as those that are not specific, that can be considered subsidies in economic terms.4 One important objective of this study is to offer a working definition and classification of subsidy actions for the purpose of tracking them and building a database. Meant to provide the conceptual basis for the collection and classification of subsidies and subsidy-like measures that have the potential to distort trade and investment, the definition is broad and based on the nature and function of the policy intervention rather than its legal or economic form. In deciding on the limits of this definition a number of existing and proposed frameworks on subsidies were examined.5 4. Article 1.1 of the WTO ASCM defines subsidies as government measures containing three basic elements: (a) a financial contribution (b) by a government or any public body within the territory of a Member (c) which confers a benefit; in addition, the measures must be “specific” to an enterprise or industry to be covered by the agreement. The agreement does not apply to measures affecting trade in services or incentives to foreign direct investment (FDI), which are covered under other WTO agreements (notably, the General Agreement on Trade in Services, GATS), under different concepts and rules. 5. Consulted disciplines include but are not limited to: the ASCM and the General Agreement on Trade in Services (GATS), which form part of the agreements covered bythe WTO; the state aid system established under the relevant agreements of the European Union; and select regional trade agreements such as the United States-Mexi- co-Canada Agreement and the Canada-EU Comprehensive Economic and Trade Agreement. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 20 At the most basic level, subsidies share certain core may involve furnishing or buying goods, services, real features: property, natural resources, or intellectual property—assets that may not be immediately convertible into monetary terms 1. They must be directed at or affect economic actors, such but have economic value. as producers of goods or suppliers of services, and have the potential to distort cross-border trade or investment. The transfer of value in a subsidy may emanate from 2. They must entail the transfer of economic value or a public authority, whether at the central or subcentral nonfinancial assets. level, but potentially it could also involve actions by 3. They must be provided by a granting authority, which is enterprises owned or controlled by the state. In the latter a public authority or entity acting on behalf of or at the case it may not be easy to ascertain whether an entity can behest of a public authority. be identified as a granting authority (feature 3). On the one 4. They must confer a benefit or advantage on the recipient, hand, state ownership or control does not necessarily mean in the sense that they are provided so conditions are more that these entities are not commercial. Left to their own favorable than those prevailing in the marketplace. devices, they could be profit-maximizing companies driven 5. They must be limited, whether explicitly or implicitly, to a by the market. On the other hand, when the state can control subset of all possible recipients on the market. A detailed the behavior of such entities, the potential exists that they discussion of elements is included in appendix B. may behave in a noncommercial manner to advance public- policy objectives. Such entities include state-owned and While no definition effort is infallible, wholly exhaustive, state-invested enterprises and chartered or concessionaire or completely neutral, the defined scope aims to abstract companies tasked with providing public goods or services from actions that are of potentially less analytical interest, (for example, municipal transport companies). such as the provision of social services, or measures, such as a currency devaluation, that are horizontal in To qualify as a subsidy, a government action must offer nature. In that way, feature 1 is meant to focus on measures terms that cannot be obtained in the open market. For applied to, or directed at, actors or activities of an economic, example, financing by a state bank may not meet the definition productive, or commercial nature (collectively referred to as of a subsidy if the terms are no better than those offered by “economic undertakings”), because it is more likely that this private banks (feature 4). Similarly, the government purchase type of public spending affects cross-border activity and of goods or services is not per se a form of subsidy, but it can potentially distorts competition. Similarly, feature 5 is meant become one when the transaction is part of a price-support to identify measures that benefit specific actors, enterprise program. The transfer of value at a zero price, such as the activities, sectors, or regions, whether de jure or de facto grant of a financial asset or intangible asset such intellectual limited to a subset of economic undertakings. property rights, is clearly a subsidy. For other types of value transfer, however, determining whether it confers a benefit The transfer of economic value can involve different requires a careful examination of the terms and conditions modalities beyond the transfer of government money to of the transaction and comparing them to those available in a private entity (for example, a negative tax). Subsides in the market.6 feature 2 include direct transfers of funds, such as grants, loans, equity infusions, and debt-to-equity swaps. Guarantees, insurance, or lines of credit may also be offered to satisfy future financial claims or be contingent on the occurrence of a certain event. Other types of subsidies may involve forgoing government revenue, as in a wide range of tax or duty breaks (such as exemptions, deductions, remissions, and so forth), which are not costless to the government. Finally, subsidies 6. A hypothetical benchmark may be needed for purposes of comparison if no market for the asset exists or if the market is significantly distorted, for example, by the government’s presence. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 21 3. >>> The Extent of Subsidy Measures Subsidy measures have proliferated in recent years. Data collected after the global financial crisis of 2008 show that subsidies have been the principal interventions during the period, surpassing tariffs and other nontariff measures. Subsidies account for close to 45 percent of the interventions documented by the Global Trade Alert project (figure 3.1), followed by tariffs at about 20 percent and trade remedies at about 10 percent. It is not always clear what counts as a subsidy. Local content requirements, the most common form of government procurement measure in the period, can be seen as subsidy-like measures. >>> Figure 3.1. Subsidies have been the most frequent form of intervention (Shares of documented measures, 2009–20) Percent 50 45 40 35 30 25 20 15 10 5 0 Government Quantity Subsidies Tariffs Tra de Other procurement controls reme dies me asures Source: Global Trade Alert, World Bank staff calculations. Note: Shares refer to counts of government interventions in the GTA; bars sum to 100 percent. 7. Global Trade Alert (GTA) has been tracking potential trade-distorting and trade-liberalizing measures of countries in the G-20 and around the world since the end of 2008. The data from GTA correspond to the full download of the database accessed in October 2021. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 22 Unilateral actions are the dominant response to Countries are increasingly resorting to countervailing contentious subsidies. Under WTO rules, countries can duties and other trade remedy measures. According to the initiate consultations on subsidy issues as part of the dispute World Bank Temporary Trade Barriers Database, the average settlement mechanism, but this happens rarely. For example, number of countervailing procedures more than doubled in 2018, just six requests for formal consultations involved from 19 new cases per year in 2007–09 to 43 in 2017–19 subsidies. Of those, four were more about the application of (figure 3.2).9 Because these measures remain in place for unilateral remedy actions by another WTO member. Subsidies several years, the stock of existing global measures almost that are specific, per the WTO ASCM, can be addressed quadrupled from 2007 to 2019. through the dispute settlement or by unilateral actions if they cause injury. Countervailing subsidies unilaterally via tariffs is the dominant course of action.8 >>> Figure 3.2. New unilateral barriers to subsidized imports have accelerated (Initiated countervailing actions per year) 50 45 40 2017-19 35 average Initiated actions 30 25 20 15 2007-09 10 average 5 0 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Source: World Bank Temporary Trade Barriers Database. Global measures against subsidies are highly in part reflects that, before 2007, countervailing actions concentrated. The United States, Canada, and the EU are were not taken against China, as a nonmarket economy, the top three users, accounting for more than 80 percent of under US trade remedy procedures. However, the continued global countervailing measures. The United States alone increase in actions against China over the following decade accounts for more than half. Countervailing actions are also suggests a sustained concern about domestic injury caused concentrated by the target country. Since 2010, China has by subsidized imports from China. Other key targets include been by far the largest target of such actions (figure 3.3). Türkiye, Republic of Korea, and Vietnam, but at significantly Prior to that, India was the principal target and is now the lower levels. second largest. The rapid increase in actions against China 8. Most unilateral countervailing actions on subsidies involve imposing tariffs, meant to counter any advantage, on the imports of goods that a country deems to be subsi- dized and harmful to its domestic industries. 9. Chad P. Bown, Milla Cieszkowsky, Aksel Erbahar, and Jose Signoret. “Temporary Trade Barriers Database,” World Bank, 2020, www.worldbank.org/en/data/interac- tive/2021/03/02/temporary-trade-barriers-database. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 23 >>> Figure 3.3. China has emerged as the principal target of countervailing measures 120 100 80 Measures 60 40 20 0 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 China India Türkiye Korea, Rep. Vietnam Source: World Bank Temporary Trade Barriers Database. Similarly, the product mix of unilateral actions against with the share increasing as iron and steel from China were subsidies is highly skewed toward some sectors. Since targeted. Together iron and steel and their products and paper 2007, by far the largest number of countervailing actions are and paperboard have accounted for the overwhelming majority found in the iron and steel sector, which includes products of global countervailing measures. Measures have diminished in the harmonized system (HS) chapter 72 (table 3.1). in sectors such as electronics, edible fats, and sugar, while Downstream iron and steel products, contained in HS chapter they have risen in sectors such as aluminum and plastics and 73, account for the second-largest number of measures, rubber and their products. >>> Table 3.1. The scope of products subject to countervailing measures is diversifying (Top sectors targeted by countervailing actions) 2007 2019 Chapter Measures % Chapter Measures % Iron and steel 72 257 80.802 Iron and steel 72 630 59.66 Iron/steel products 73 10 3.14 Iron/steel products 73 127 12.03 Paper and paperboard 48 6 1.89 Paper and paperboard 48 39 3.69 Aluminum and Electronics and parts 85 6 1.89 76 36 3.41 products Edible fats 15 5 1.57 Plastics and products 39 25 2.37 Sugar/sugar products 17 4 1.26 Rubber and products 40 23 2.18 Source: World Bank Temporary Trade Barriers Database. Note: The count reflects the number of source countries and HS 6-digit products affected. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 24 Antidumping measures are used jointly with These data supplement available information on countervailing actions. While practically all countervailing subsidies that are fragmented. Information on subsidies is cases are also brought up as antidumping cases, the reverse very incomplete and not always collected consistently across is not the case. The data on antidumping measures suggest sectors, nor does it follow similar definitions or methodologies, that some countries may be using them to defend against as reviewed by a recent study by the IMF, OECD, World less-than-normal-value imports in general.10 Antidumping Bank, and WTO (2022). Some data are narrowly focused measures have been increasing continuously over time, over on specific sectors or industries, such as energy, while other a wider range of countries, and in sectors similar to those data may focus on specific types of measures, such as tax targeted by countervailing measures. This description of the incentives.11 A source of data that aims to be broad in sectoral antidumping data is in appendix A. scope is the Global Trade Alert, which collects information on government interventions since 2009. This source is based Trade remedies are likely skewed toward sectors where on announcements of new interventions, so that it captures industry interests are relatively better organized. To the flow of measures, but it could miss the totality of applied implement countervailing actions, countries must determine measures (the stock). that the subject imports are subsidized or sold for less than the normal value and that these imports have caused (or threaten A multidimensional categorization is used to record these to cause) material injury to a domestic industry. This process actions and characterize the nature and extent of identified can involve significant time and effort. Yet in some countries subsides. This includes the following principal fields: and sectors, there may not be an import-competing industry, or the import-competing industry is too dispersed to seek • How the subsidy is applied (for example, a law or a protection. Some countries lack the legal and institutional directive) capacity to undertake these trade-defense actions (as in • The stated objective most of Sub-Saharan Africa, for example). Thus, while the • Whether it is applied to a good or a service information on trade remedies is indicative of certain sectors • The type of subsidy (for example, grant, loan, and so forth) where there may be high levels of tension, it provides a narrow • The granting authority (for example, central or subcentral picture of the extent of subsidies. level) • The recipient and/or beneficiary • Selection criteria (for example, firms or activities) 3.1. A New Compilation of Subsidy Subcategories to these headings provide a detailed list of Measures features to describe the data (appendix B).12 This study collects data on subsidies in a wide range of To focus efforts, data were limited to major exporters countries from available trade-policy documents. These likely to affect global markets. The data on subsidies include notifications by governments on subsidy programs cover 31 jurisdictions that account for over 70 percent of implemented in their countries; assessments of the trade global goods trade. These include G-20 countries plus system as documented by the Trade Policy Review process; other major trading countries, ranked by their export values. and concerns raised by trading partners as reflected in Supranational measures implemented at the EU28 level (i.e., documents such as the US National Trade Estimate Report including at the time the United Kingdom) are recorded and and the EU Market Access database. A similar compilation denoted as EU subsidies. Subsidies by six principal traders for nontariff measures (NTMs) is the CoRe NTM database in the EU (Belgium, France, Germany, Italy, the Netherlands, (Martinez, Mora, and Signoret 2009). These sources can be and the United Kingdom) were also categorized by country used to identify the products or sectors that are more frequently (table 3.2). In the interest of covering more countries outside targeted by subsidies as well as the types of subsidies, and to of the EU, national subsidies in the EU-wide region were not benchmark the prevalence of subsidies by country. fully recorded.13 10. Technically a foreign firm could be dumping its products in some other country to the extent that it can price-discriminate across destination markets. 11. See, for example, OECD (2018) and OECD and IEA (2021) for support measures for fossil fuel. See Redonda, von Haldenwang, and Aliu (2021) for a cross-country database on tax expenditures. 12. This expands in various ways other current nomenclatures, such as the latest classification of nontariff measures by the United Nations Conference on Trade and Development (UNCTAD). 13. For purpose of a simple count, all subsidy programs reported by the EU-28 members can be calculated, as in figure 3.4. But the data of EU members not explicitly targeted for the database are not further categorized, for example, in terms of sectors, type of subsidy used, or intended purpose. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 25 >>> Table 3.2. Economies Selected for the Database Argentinaa Indonesiaa Saudi Arabiaa Australiaa Israel Singapore Belgiumb Italyab South Africaa Brazila Japana Switzerland Canadaa Korea, Rep.a Türkiyea Chile Malaysia United Arab Emirates Chinaa Mexicoa United Kingdomab European Uniona Netherlandsb United Statesa Franceab Norway Vietnam Germanyab Philippines Indiaa Russian Federationa Notes: For completeness, all EU-28 countries’ counts of subsidies are included in the data; but the dataset does not include all variables for the European Union countries not listed. a. G-20 member. b. European Union (EU28) member. The data provide a cross section of over 2,000 subsidy populated as completely as possible from the original source programs in a broad range of sectors. This count reflects with some follow-up research to supplement the data source. programs that were in effect through all or part of the period 2017–18. Countries’ notifications to the WTO, the most granular source of information, include much of the information needed 3.2. Global Landscape of Measures to populate most dimensions of the database. However, assigning sectors was not always straightforward when This new database is used to assess global subsidies assigning an industrial classification such as ISIC to a vague in terms of their intended policy objectives, types of textual description (such as “children’s goods”).14 Information measures, and regional and sectoral patterns. The on a specific recipient or beneficiary was generally omitted to database offers a snapshot of the global economy, covering preserve confidentiality. Subsidy amounts (while not broken over 2,000 programs among the nations that are likely down by individual recipient) were sometimes available at the to shape trade trends in significant ways through their program level, but values were missing for a large number disproportional effect on global markets. The data represent of programs. The missing information sometimes depends on a cross-sectional inventory of subsidies in 2018,15 including the form of subsidy (for example, tax breaks where estimates legacy measures enacted prior to that year. Using 2018 as of the revenue forgone were not available in some countries) a baseline avoids atypical years when data can be distorted or on the level of the implementing jurisdiction (for example, by extraordinary shocks such as the global financial crisis or subnational programs for which the monetary values of the the COVID-19 pandemic.16 measures are not documented). Fields in the database were 14. ISIC stands for International Standard Industrial Classification of All Economic Activities. ISIC Revision 4 was used to classify the subsidy programs per affected econom- ic sector. 15. The data track subsidy programs in place in a notional year of 2018, for which most intended targeted countries were available. For only a small number of countries this may refer to 2017 or earlier. 16. Recent events such as the COVID-19 pandemic, the Ukraine-Russia war, and disruptions in global supply chains are adding demand for subsidies. One consideration is whether in mitigating these shocks the trends and patterns described in this study are likely to continue going forward. That subsidies are more prevalent in major trading partners is a trend that seems robust over time (in crisis and noncrisis times). EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 26 R E G I O N A L PAT T E R N S (figure 3.4) and are the top global traders. These figures are consistent with other broad databases on subsidies. For The EU, the United States, China, Canada, and Australia example, the OECD found that these three economies are the are top global subsidizers by number of programs. main subsidizers in terms of budgetary support to agriculture The top two alone—the EU (inclusive of measures in EU28 (OECD 2021). Similarly, a database on the flow of measures member countries) and the United States—account for over in the last decade or so provides evidence that most new half of measures and, combined with China, the three account interventions refer to the EU, the United States, and China.17 for about 75 percent of measures included in the database >>> Figure 3.4. The top three trading blocs account for most subsidies (Distribution by number of measures, 2018) 35 30 25 20 Percent 15 10 5 0 EU United States China Canada Australia Others Source: Compilation of subsidy measures. >>> Figure 3.5. Subcentral government subsidies account for the largest share of measures (Distribution by number of measures in each region, 2018) 100 90 80 70 60 Percent 50 40 30 20 10 0 EU United States China Canada Australia Central Subcentral Source: Compilation of subsidy measures. Note: EU = European Union. 17. IMF, OECD, World Bank, and WTO (2022) calculation based on GTA data. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 27 Most subsidy programs are implemented by government A wide range of manufacturing activities are supported by entities at the subcentral level (figure 3.5). For the EU, treated subsidies. The largest share of programs is directed broadly as a single market, subcentral measures include those taken by to manufacturers. Programs are prevalent in the electronics, EU member states at the national and subnational levels. For vehicles, machinery, ships, chemicals, food and beverages, all top users of subsidies (counting the EU as a region or bloc), and metals and metal products industries, among others. the vast majority of government subsidies relate to support About 25 percent of programs by broad sectors in the database measures at this subregional level. In the United States and affect manufacturing industries (figure 3.6). The second broad China, subcentral subsidies account for 90 percent and 82 category refers to professional, scientific, and technical activities percent of programs, respectively.18 These findings highlight (primarily R&D), followed by agriculture and fishing. Together, the importance of ensuring that all levels of government are those three top broad sectors account for about 60 percent of aware of and subscribe to established disciplines on subsidy subsidies at that level of industrial or sectoral aggregation. Other provisions and acknowledge the need for full transparency. service sectors can be important beneficiaries of subsidies; for example, information and communications and construction, but their shares of subsidy measures are relatively low. S E C T O R A L PAT T E R N S Primary sectoral targets of subsidies vary by country and Scientific research and development (R&D), several region. Table 3.3 shows the share of subsidy interventions by manufacturing industries, agriculture, and energy are the the number of measures in the top implementing economies. principal recipient sectors for subsidies. The database uses For the EU, the largest share of subsidies goes to crop and international industry classification codes to categorize subsidy animal production (ISIC 01), accounting for about 9 percent of programs.19 Depending on the program, the sectoral association the total, followed by food products and scientific R&D (ISIC can be made at the two-, three-, or four-digit industry code level. 10 and 72), each accounting for about 5 percent. For the About 8 percent of cases are categorized as “horizontal” either United States, the three most frequently subsidized sectors because measures seemed applicable to any sector, or because are machinery and equipment; computer, electrics, and optical the textual description was too vague or broad to allow for a products; and chemicals (ISIC 28, 26, and 20). For China, the sectoral classification. The broader categorization at the two-digit largest shares are for sectors related to scientific R&D, motor level shows that the recorded subsidies benefit upward of 80 vehicles, and fisheries (ISIC 72, 29, and 03). industries or sectors. However, about half of the observations by program and sector fall among the top 15 sectors. >>> Figure 3.6. Support to manufacturing predominates (Top broad sectors by number of measures) 30 25 20 Percent 15 10 5 0 Ma nufacturing Professional, Agriculture, forestry, Information and Electricity and gas Other sectors scientific, and and fishing commu nication technical activities Source: Compilation of subsidy measures. Note: Breakdown by 1-digit (section) code of International Standard Industrial Classification of All Economic Activities (ISIC), rev. 4. 18. Some evidence in the academic literature shows subsidy competition among US states. Ossa (2015), using a quantitative economic geography model calibrated to US states, finds that states have strong incentives to engage in subsidies to outcompete other states in attracting economic activity, with potentially large costs from an escalation of subsidy competition. 19. In particular, this database uses the International Standard Industrial Classification of All Economic Activities (ISIC), Revision 4. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 28 >>> Table 3.3. Top affected sectors tend to vary for each main user (Percentage of the count of programs for top sectors, by top regions, 2018) EU USA China Crop/animal prod 8.9 Machinery/equip. 4.4 Scientific R&D 16.3 Food products 5.2 Computer/elect./optical 4.3 Motor vehicles 6.1 Scientific R&D 5.2 Chemicals 4.1 Fishing and aquaculture 5.4 Electricity, gas 3.5 Scientific R&D 4.0 Computer/elect./optical 4.5 Business services 3.5 Transport equipment 3.8 Computer programming 3.1 Land transport 3.1 Electricity, gas 3.6 Transport equipment 2.9 Pharmaceuticals 2.9 Food products 3.3 Civil engineering 2.9 Transport equip. 2.9 Motor vehicles 3.2 Crop/animal prod 2.7 Civil engineering 2.9 Basic metals 3.1 Basic metals 2.7 Beverages 2.3 Electrical equip. 3.1 Electrical equip. 2.5 Source: Compilation of subsidy measures. Note: Sectors designated by 2-digit (division) code of International Standard Industrial Classification of All Economic Activities (ISIC), rev. 4. Elect = electrical; equip. = equipment; EU = European Union; prod = products; R&D = research and development; USA = United States. While subsidy measures are generally uniform through based on a count of programs because information on their sectors in the United States, they are more skewed toward value is often missing. agriculture in the EU and technology in China. For example, the top and fifth sectors for the United States (machinery and However, the types of subsidy measures can vary transport equipment, respectively) both account for about 4 significantly by implementing region (figure 3.7). For the percent of US programs. Meanwhile, in the EU, the top sector EU, for example, about six in 10 programs take the form of (crop and animal production) accounts for more than double financial grants. This is consistent with subsidies provided the number of subsidy programs as the fifth sector (business by the EU in agriculture, where support to producers takes services). The difference is even more pronounced in China, the form of direct payments. For the United States, on the where the top sector (scientific R&D) receives more than five other hand, a similar proportion of measures takes the form times the number of programs than the fifth sector (computer of tax incentives. For both the EU and the United States, programming). about 20 percent of measures refer to financial instruments such as loans, guarantees, or capital injections. For China, most programs refer to grants, while financial interventions TYPE OF MEASURES AND OBJECTIVES are apparently rare. This is at odds with some accounts of subsidies to industrial sectors in China in the form of Financial grants account for slightly over half of subsidy below-market financing; for example, in cases where measures in the database, followed by tax breaks such as aluminum products from China benefit from loans granted rebates, base or rate reductions, and exemptions. Other on a preferential, noncommercial basis.20 Yet these are interventions take the form of soft loans, loan guarantees, provided through state-owned or state-directed banks that and financial interventions such as capital injections and are inherently difficult to identify. purchases of equity stakes. Again, this characterization is 20. See, for example, the 2017 dispute at the WTO by the United States on primary aluminum from China, in consultations; US countervailing duties on aluminum foil from China of 2021. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 29 >>> Figure 3.7. Subsidies take various forms depending on the economy (Count shares of subsidy forms by top users, 2018) 100 90 80 70 60 Percent 50 40 30 20 10 0 EU USA China Grants Tax expenditures, other rev forgone Loans, guarantees, capital injections Source: Compilation of subsidy measures. Note: rev = revenue. Identified subsidies can be classified as aiming to achieve areas (see appendix C), of which the top 15 are listed in a wide variety of stated objectives. Deciphering the intent figure 3.8.21 Primary areas are intended to support strategic of subsidy measures is not straightforward. Nevertheless, sectors and the acquisition of knowledge; other goals include subsidies could be classified based on the objectives or goals the environment, support to fuel and energy production, of the programs described in the source measures. Based infrastructure development, and social objectives. on this information, programs were catalogued in 30 broad >>> Figure 3.8. Support to improve sectoral competitiveness is a principal objective (Top stated objectives of programs, 2018) Sectoral R&D/Innovation Environmental Preservation Economic/Regional Development SMEs & Entrepreneu rship Industry Restructuring Fu el/Energy Provision Clean Energy Social International Trade Infrastru ctu re Employment Capacity Bu ilding Disaster/Climate Change Preparedness .. Cultural Enrichment 0 5 10 15 20 25 Percent Source: Compilation of subsidy measures. Note: rev = revenue. 21. Cognizant that a particular measure may serve more than one objective, two variables were allowed per measure to record those multiple concerns and are counted separately using those. For example, measures intended to improve the competitiveness of a particular sector, as a primary objective, and also to advance issues of regional development, if noted as such, are included with both goals in the calculation. These characterizations, of course, are limited by the relative richness of the program descriptions. Also, as an overarching caveat, intended goals per the descriptions may not necessarily reflect the real political economy objectives pursued by these interventions. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 30 Most recorded measures appear related to a sectoral illustrates the stated subsidy rationales and their ex ante objective. This covers general sectoral financial support possibly distortive effects to identify which should be subject for the acquisition of capital goods, materials, and business to particular scrutiny. operations. This form of sectoral objective includes slightly more than 20 percent of cases, suggesting that subsidies are Environmental subsidies are less common. These, in used in pursuit of industrial policy. A second large category principle, include measures to preserve certain species or consists of incentives to help generate or transfer technology, ecosystems; reduce earth, air, or water pollution; and conserve such as capital and rewards for innovation, knowledge-sharing natural resources through tax credits or other incentives for and technology-transfer platforms, financial and business the adoption of more environmentally friendly or sustainable support for start-ups, and capacity building. production, processes, transportation, and so forth. These include rewards for conservating energy and reducing pollution Support of trade competitiveness is a common goal. through the use of coal filters, for example, or by substituting Although relatively fewer subsidy measures are directly natural gas for coal. Subsidies for clean energy, categorized described as a way to advance trade (10th in the objectives separately, are even further down the list and below subsidies category in figure 3.8), many nevertheless represent industrial for other fuels and energy, showing how fossil fuels continue policy efforts to increase competitiveness, as in the sectoral to receive more support than renewables. category, which is the most common. Box 3.1 further BOX 3.1. SUBSIDY RATIONALES AND THEIR POSSIBLY DISTORTIVE EFFECTS It is difficult to categorize a particular subsidy measure as trade distortive ex ante, because it often depends on how it was designed and how it is implemented. In principle, subsidies can be the first-best policies in the presence of market failures, where they can be seen as corrective interventions, for example, to capture positive externalities. For instance, economists generally recognize that private agents may underinvest in research and development because their rate of return in such innovation investments is less than the social return. Subsidies can also be often justified in the pursuit of social objectives or in the context of an economic, health, or environmental crisis. However, from mere descriptions of a subsidy’s intentions, it is often difficult to identify which market failure, if any, particular subsidies are trying to address. And depending on the industry benefiting, often a tax, rather than a subsidy, may be a more economically justified action. One example of this is the still significant levels of subsidies directed to fossil fuels that are an important area of debate (IMF, OECD, World Bank, and WTO 2022). Even in more ambiguous cases, an important consideration should be whether the intended objective can be achieved with fewer subsidies or with subsidy actions that could have a less trade-distorting effect. These are important considerations, not only because of the uneven playing field that certain subsidies can generate but also because of the opportunity cost of subsidies. In the context of limited resources, the presence of subsidies means that some other goods or services are underproduced. A good number of subsidies could be reviewed for their distortive effects. In view of their intended objectives, where available (in about two-thirds of collected subsidy information), subsidies could be classified on the basis of whether the goal suggests a higher potential for being distortive in favor of the beneficiaries or whether the subsidy could be seen as having either a more ambiguous or lesser/more indirect effect on distorting trade and competition. Objectives are divided along these lines in table B3.1.1. Roughly a third of programs in the database fall within each category, with 35 percent of subsidies intending to support sectoral goals, industry restructuring, state-owned enterprises, and so on. In terms of reforming subsidies, initial scrutiny could focus on the interventions with the potential of being more distortive. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 31 >>> Table B3.1.1. Possible Distortive Effect by Objective Possibly distortive Possibly ambiguous Possibly less distortive Sectoral Economic/regional development R&D/innovation Industry restructuring SMEs and entrepreneurship Public services SOEs Transportation Infrastructure Fuel/energy provision Clean energy Disaster/climate change readiness/ International trade Financial markets participation response FDI Business services Public health/safety PPPs Capacity building Social Regulatory compliance, certification Employment Environmental preservation National traditions/heritage Support for ethnic minorities Support for people with disabilities Gender-related support Cultural enrichment Children/adult education Water conservancy/treatment Source: Original figure for this publication. Note: FDI = foreign direct investment; R&D = research and development; PPP = public-private partnership; SMEs = small and medium enterprises; SOE = state- owned enterprise. 3.3 Effects of Subsidies on Trade of the good) and the presence of a subsidy. Second, in the absence of frictions to trade, a divergence in the prices of Subsidies can have cross-border spillovers for other identical goods on international markets should be eliminated trading partners. A disproportionally large number of programs by arbitrage. Thus, part of the observed price gap depends on are implemented by major trading countries that have the a range of trade costs (for example, distance, the presence of economic heft to distort global markets for goods and services. borders, and so forth) that need to be controlled for as best as This section assesses the potentially distortive effects of possible, especially if the measure of the subsidy is a residual subsidies on global trade, a primary goal of this inquiry. term. Third, even if the extent or value of subsidies can be measured with reasonable confidence, the question of how The literature on the effects of subsidies on trade is trade is distorted remains unanswered. sparse, mainly because of the paucity of data. Some available estimates of the value of subsidies, such as those Government support measures in agriculture are among for the energy sector, primarily depend on estimates of the best documented. That is the result of over 30 years of differences between the observed, subsidized price of fuel analysis at the Organisation for Economic Co-operation and and an undistorted reference price. While this approach Development (OECD) on support for producers and consumers is useful, it is insufficient for an understanding of the trade- of agricultural products. For example, the producer-support distorting effects of subsidies for at least three reasons. First, equivalent (PSE) captures the production support in dollar prices can partly reflect qualitative characteristics of goods, terms or as a percentage of gross farm receipts. This includes so it is difficult to distinguish between the relative impact on estimates for 54 countries. The largest producer support prices of quality differences (for example, based on the source numbers by far refer to the EU, followed by China. The United EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 32 States, in third place for agricultural subsidies, provides much E C O N O M E T R I C E S T I M AT E S of this assistance through consumer support; the same is true for India (OECD 2021). Outside of these four economies, The methodological approach followed here estimates agricultural subsidies on a value basis are relatively small. As the effect of subsidies on trade. It builds a multicountry a percentage of gross farm receipts, agricultural subsidies and multisector general equilibrium framework that captures average around 12 percent for all countries and 17 percent for the impact of subsidies on production and trade. Using this OECD countries. model, a set of theory-based estimating equations can be derived to obtain a structural gravity equation. Augmented By contrast, estimates of the monetary value of subsidies with tariff and subsidy information, the model makes it in industrial sectors are generally lacking. There is the possible to estimate the direct effects of subsidies on trade sense that subsidies are common in a wide range of industries and their ad valorem equivalents.23 such as electronics, vehicles, other transport equipment (for example, aircraft, ships, rail), pharmaceuticals, other This empirical approach explicitly incorporates subsidies chemicals, and machinery. There is also the sense that they in various sectors and countries. It takes advantage of the are more prominent in the larger economies of China, the EU, inventory of subsidy programs described in this study and and the United States, as described in Section 3.2.22 But how enters it directly as an explanatory variable in the estimating distortive these programs may be and how consequential they equation, building on its variation for sectoral identification. In may be for the commercial interests of other trading partners these data, only a small fraction of identified programs are is not well understood. Similar observations can be made for horizontal. While many subsidy programs with identified sectors subsidies related to services, which are also common. fall within manufacturing, scientific R&D, and agriculture, there are nuances by countries in terms of industries. For example, Even for agricultural subsidies, it is difficult to ascertain within transport equipment, some countries’ programs may be trade-distorting effects. Subsidies can be implemented in directed to aircraft, others to motor vehicles or railways. ways that can have more or less distortionary effects. For example, some measures may be “coupled” to the recipient’s The analysis exploits the available signal in terms of level of output, while others may be decoupled (such as direct sectors of interest targeted for support. While the inventory payment to farmers), which may be less distortionary. In any of measures is not exhaustive, it is nevertheless likely to case, it is necessary to take the estimated support equivalents provide useful information as to strategic sectors by country to some type of model to translate the effect of the subsidy (or by combining available information on the incidence of subsidy its removal) on trade. For agriculture this type of analysis has measures with statistical inference from econometrics. The been done in the context of computable general equilibrium approach followed in the main set of results refers to using not (CGE) model simulations (see, Anderson, Valenzuela, and only the presence of subsidies in a country-sector, but also van der Mensbrugghe 2010). Such modeling builds on added the intensive margin from the multiplicity of subsidies provided information in the Global Trade Analysis Project (GTAP) to sectors (such as in receiving a variety of subsidies). To database over the years on agricultural support equivalent estimate this impact of subsidies on trade, the information on (Dimaranan and McDougall 2005; Anderson and Valenzuela subsidy programs is mapped to 170 sectors for which trade 2008). However, this limited information on agricultural and production data across countries are brought into the subsidies exaggerates the relative level of subsidization analysis,24 along with other country and sectoral covariates in agriculture and precludes a comprehensive analysis on of trade, such as those used in the empirical gravity literature. subsidy reform (i.e., both agricultural and non-agricultural). However, the data analysis is conducted separately for 22. Evenett and Fritz (2021) also consider China, the EU, and the United States major players in the subsidy field and focus attention on subsidies in these economies. The authors build an inventory of subsidy interventions in China, the EU, and the United States since 2008. As in the GTA database, this information is collected from government information made available online on subsidies granted, complemented by information on subsidies receipts by publicly listed Chinese companies (from the Win.d database). The information on subsidy receipts to Chinese companies shows that not only the number of subsidies has increased over time, but also the annual amounts received, which have been growing at a fast pace during the period (at an annual growth rate of 14.6 percent). No monetary values of subsidy receipts are reported for the EU or the United States. 23. Appendix E presents some basic assumptions, the structure, and the estimating approach of the structural model. Additional technical details can be found in Larch et al. (2021). In general, the assumptions employed are relatively mild and common to the many structural gravity frameworks. 24. These data are from the recent International Trade and Production Database for Estimation (ITPD-E) database (Borchert et al., 2021), using the average of the last three years in that database to smooth for across years. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 33 agriculture and manufacturing industries, excluding the flows (for example, RTAs can increase trade by 16 percent for relatively less tradable sectors of mining and services.25 manufactured goods and by up to 43 percent for agriculture). Several subgroups within both agriculture and manufacturing On the opposite side, the coefficient for an indicator variable are formed to estimate the effect of subsidies on trade in other for trade sanctions shows large negative effects on trade. sectoral subgroups of interest, as described further below. The variable for domestic market, an indicator variable for trade that refers to domestic sales, aims to capture the The estimated trade equation shows plausible estimates extent of home bias; while the interaction with gross domestic and some comparable effects to the empirical gravity product (GDP) per capita allows for this to vary by the level literature. The estimated coefficients for all terms show the of development of the producing country. The term for tariffs expected signs and are in general statistically significant (table captures the variable cost of tariff barriers on bilateral trade.26 3.4). Distance lowers trade, with a slightly larger coefficient for Its estimated coefficient is used to recover the trade elasticity agricultural goods. Other gravity measures such as contiguous and is estimated at around 3 percent for agriculture and 8 borders, common language, and an indicator variable for percent for manufacturing, which are plausible magnitudes colonial relationship tend to increase bilateral trade flows. given other estimates in the literature.27 This trade elasticity, Policy variables include indicator variables for membership in in turn, is used, together with the estimated coefficient for regional trade agreements (RTAs) and the EU single market, the term for subsidies, to calculate the ad valorem equivalent both of which show significant positive effects on bilateral trade (AVE) of subsidy measures. >>> Table 3.4. Econometric estimates of bilateral trade and the effects of subsidies by broad sectors (1) (2) Agriculture Manufacturing −0.891 −0.741 Log (distance) (0.103)** (0.034)** 0.651 0.354 Contiguous (0.136)** (0.048)** 0.294 0.366 Colony (0.202) (0.077)** 0.676 0.150 RTA (0.269)* (0.066)* 1.271 0.191 EU (0.281)** (0.096)* −1.747 −1.776 Sanctions (0.410)** (0.325)** 10.968 6.889 Domestic (1.650)** (0.704)** Log (GDP per capita) −0.683 −0.405 *Domestic (0.163)** (0.068)** 0.025 0.037 Subsidy measures (0.008)** (0.013)** −3.428 −8.467 Tariff (1.172)** (0.924)** 25. The agricultural sector is defined as the first 26 industries in ITPD-E data; while manufacturing covers over 100 sectors from industry 34 to 153. 26. The information on tariffs at the product level is from the International Trade Center’s Market Access Map database (MacMap) and reflects preferential rates in effect among trading partners. 27. For example, Fontagné, Guimbard, and Orefice (2022) estimate trade elasticities for all possible HS six-digit codes and find that the estimated parameter has a median of −5.6 EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 34 >>> Table 3.4. Econometric estimates of bilateral trade and the effects of subsidies by broad sectors 11.733 12.462 Constant (0.912)** (0.319)** N 46228 479227 R2 0.982 0.972 14.529 8.443 Subsidy AVE (6.743)* (2.843)** Note: Standard errors in parentheses. Contiguous = contiguous borders; language = a common language; colony = a colonial relationship; RTA = membership in regional trade agreements; EU = membership in the EU single market; domestic = domestic sales; GDP per capita allows for this to vary by the level of development of the produc- ing country; tariff = the variable cost of tariff barriers on bilateral trade. AVE = ad valorem equivalent. Regressions include exporter and importer fixed effects. +p < 0.10, *p < .05, **p < .01 Subsidies promote a country’s exports disproportionately. The effects of subsidies on trade also differ by region, This impact is estimated by the effect of the subsidy variable in with subsidies favoring manufacturing exports being the trade equation that captures the effects of subsidies above statistically significant for Asia-Pacific. Estimating the and beyond their effects on the domestic economy. As the econometric model and allowing for interaction effects with subsidy variable is the count of subsidies for the sector, the country regions shows this differentiated effect. Estimated estimated coefficients for agriculture and for manufacturing coefficients for most model variables are comparable to suggest that the introduction of a subsidy program increases those presented in table 3.4, and the comparable table is exports by about 3 to 4 percent, relative to domestic sales. included in appendix E. Figure 3.9 shows the resulting ad Given the recovered trade elasticities for the tariff term, the valorem equivalents of subsidy measures by region (as effects of the subsidies on trade can be restated in ad valorem estimated at one standard deviation) for those estimates for equivalents of about 15 percent for agriculture and 8 percent which the ad valorem equivalents are statistically significant for manufacturing.28 These are new empirical estimates in favor of exports.29 Only the ad valorem estimate for that show statistically significant, trade-distorting effects of manufacturing in the Asia-Pacific shows a positive (and subsidies for both agriculture and manufacturing exports. statistically significant) value.30 >>> Figure 3.9. Trade effects of subsidies are particularly significant for agriculture (Estimated export-support equivalents, ad valorem, by sector and region) 18 16 14 12 Percent 10 8 6 4 2 0 Asia-Pacific EU LA North America Agriculture Ma nufacturing Source: Original calculations for this publication based on estimates from a gravity estimation. Note: Ad valorem estimates of subsidies that statistically favor exports. EU = European Union; LA = Latin America. 28. These AVEs are a nonlinear function of the coefficients for subsidies and tariffs, the latter being the trade elasticity. The point estimate for the AVEs is calculated as follows: [exp(S*α/β) -1]*100, where α is the estimated coefficient for the subsidy term and β is the estimated elasticity. The subsidy variable S is evaluated at one standard deviation. 29. With a significance level of at least 10 percent. 30. The Asia-Pacific region includes the following countries in the subsidy database: Australia, China, India, Indonesia, Japan, Republic of Korea, Malaysia, the Philippines, and Vietnam. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 35 Subsidies in the advanced economies of the EU and At a more disaggregated level, subsidies for agricultural North America disproportionally affect their agricultural products, including meat and several horticultural exports. The estimated ad valorem equivalent for both products, are found to be particularly trade-distorting. regions is in the range of 14 to 16 percent (with overlapping Beyond estimating the trade effects of subsidies in the broad confidence intervals).31 Agricultural subsidies in the EU and agriculture and manufacturing categories, additional estimates the United States are known to be large (OECD 2021), which by subsectors aim to highlight groups of products for the may be consistent with these results. But according to the distorting effect of subsidies on trade that can be established same source, agricultural supports in China and India are with statistical significance. With this in mind, all goods sectors also estimated to be large, although this econometric analysis (the more than 100 sectors with trade and production data) are does not suggest disproportional trade effects in agriculture in pooled into 22 product groups. For agriculture, this includes six those regions.32 Conversely, agricultural subsidies in certain subgroups.35 Of these, particularly distortive effects are found Latin American countries are found to disproportionally affect for the sector that contains meats, livestock products, and live trade.33 This effect in Latin America, however, is significantly animals; as well as for the sector that includes fresh fruits and lower, at about 3 percent. Egas Yerovi and de Salvo (2018) vegetables, pulses and legumes, nuts, cotton, tobacco leaf, review agricultural support policies in Latin America and find and spices, among other agricultural products not elsewhere that countries had extended agricultural supports during classified (n.e.c.). For these subgroups, the estimated effect the period 2009–16 that amounted to 3.29 percent of gross of an additional subsidy measure is to promote exports by 3 to agricultural receipts on average. This estimate, in turn, would 4 percent. An estimated ad valorem equivalent of subsidies for be higher at 6.29 if Argentina (which during that period taxed plant products would be about 13 percent.36 the agricultural sector) were excluded. This same study estimated agricultural support across the Latin American and the Caribbean region at 10 percent of agricultural GDP.34 >>> Figure 3.10. Certain agricultural and manufacturing sectors exhibit significant trade effects (Estimated marginal effects of subsidy measures, by sector) 0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 Plant products Me at Rubber and plastic Ma chinery Other manufacturing Source: Original calculations for this publication based on estimates from a gravity estimation. Note: Estimated coefficients of subsidy variable that statistically favor exports. 31. The region of North America includes Canada and the United States. 32. Specifically, the estimated point estimate for the agricultural ad valorem equivalent in Asia-Pacific is 5.4 percent, with a 90 percent confidence interval that covers the values from zero up to about 11 percent. 33. The Latin America grouping refers to the following countries included in the subsidy database: Argentina, Brazil, Chile, and Mexico. 34. Subsidized inputs, for example with respect to energy and fertilizers, are important in Brazil, Chile, and Mexico (Egas Yerovi and de Salvo, 2018). 35. For agriculture, these groups are cereals; oilseeds; sugars; plant products; meat; and animal products. These descriptions are broad, and each covers one or more of the 26 agricultural categories in the ITPD-E data, which in turn group a number of agricultural products per the FAOSTAT Commodity List. 36. Figure 3.10 focuses on the estimated coefficient of the subsidy measures for product subgroups for which the parameter is at a significance level of at least 10 percent. Because the estimated trade elasticities on these more granular regressions are not always significant, ad valorem equivalents are generally not reported. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 36 Similarly, for manufacturing, trade-distorting effects double digits (on average 15 percent), helping their producers of subsidies are found for a wide range of capital, undercut developing-country competitors that might otherwise intermediate, and consumer goods. For manufacturing, be competitive in such products. estimation is done separately for 16 subgroups.37 Among these groups, subsidies have particularly distortive effects For many developing and especially low-income for rubber and plastics products, for which each subsidy countries, a significant share of trade and production is measure expands exports by about 7 percent. Another concentrated in agricultural activities subject to trade- subgroup for which subsidies promote exports includes distorting subsidies. Figure 3.11 shows the percentage a wide range of products such as machinery; electrical of goods trade and production in each country where trade- equipment (electric motors, generators, batteries, and so distorting subsidies are found per figure 3.10. On a value forth); electronic components and boards; communications basis, exports of those agricultural goods represent less than equipment; consumer electronics; medical and optical 5 percent of global trade.38 However, for poor countries, these instruments; and motor vehicles and parts. Lastly, an industry agricultural activities are significant, representing over 20 group with significant subsidy effects includes miscellaneous percent of goods exports for countries such as Afghanistan, manufactures such as furniture, jewelry, musical instruments, Burkina Faso, Burundi, Central African Republic, Ethiopia, sports goods, games and toys, and other manufacturing n.e.c. The Gambia, Guinea-Bissau, Liberia, Madagascar, Malawi, Rwanda, and Uganda. Indeed, there is a strong negative correlation between the relative importance of these I M P L I C AT I O N S F O R D E V E L O P I N G agricultural exports and income levels, as suggested by the COUNTRIES fitted line in figure 3.11 for both exports and production, so that employment and livelihood are more dependent on these Trade-distorting subsidies can displace trade and activities among the poorest countries. To a lesser extent, production in other trading partners, with important these sectors also represent significant export activities repercussions for developing countries. Domestic in middle-income countries such as Colombia, Ecuador, subsidies, beyond possibly being distortive and inefficient Grenada, Guatemala, and Turkmenistan. for domestic objectives, can have negative spillovers across borders. The uneven playing field created by subsidies can affect the commercial interests of countries that compete with subsidized products in foreign markets. This does not need to be only in direct competition in the same market, as a disproportionally large number of subsidy programs are implemented by large trading countries with the potential to affect world prices. Furthermore, subsidies by other countries may also affect domestic sales in import- competing industries that cannot compete on fair terms with such subsidized products. Both of these channels, displacing exports and displacing domestic sales, can be important for developing countries. Subsidies for both agriculture and manufacturing can be important for different reasons. For agriculture, the empirical evidence presented suggests that the trade-distortive effects of subsidies are likely higher. For example, the ad valorem equivalents of subsidies by advanced economies are in the 37. For manufacturing products, these groups are manufactured food; drinks and tobacco; textiles; apparel; wood; pulp and paper; publishing and printing; petroleum and fuel; chemical products; rubber and plastic; minerals; steel; metal products; machinery; other transportation; and other manufacturing. 38. Note that in the industry classification used here, manufacture of food products is under the broad manufacturing sector. If the HS classification were used, global exports under chapters 1 through 24, which also include prepared foodstuffs, beverages, spirits, and tobacco products, account for less than 9 percent of merchandised trade. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 37 >>> Figure 3.11. Low-income countries are particularly exposed to distorted agricultural trade (Share of trade and production for distorted agricultural trade) 100 90 GNB 80 70 COM 60 AFG Percent CIV 50 MWI ETH KEN TON GMB UGA ZWE ECU 40 CMR GRD 30 CAF DJI GTM BDI MDG BFA BEN GHA TLS COL LBR RWA TZA TKM 20 HND IRN 10 0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 Log GDP per capita, PPP Tra de Production Source: Original calculations for this publication. Note: GDP = gross domestic product; PPP = purchasing power parity. Manufactured goods represent the vast majority of in total merchandise exports and income levels (figure 3.12). global trade, so that subsidies in industrial goods can At the same time, these manufacturing exports are significant be widespread.39 Focusing only on those manufacturing for a notable number of middle-income countries, including activities with trade-distorting subsidies per figure 3.10 China, Costa Rica, Mexico, the Philippines, Thailand, Tunisia, involves slightly less than half of global exports. Thus, and Vietnam. These are countries that have integrated into together with the percentage of trade where there are trade- manufacturing value chains. Low-income countries, on the distorting subsidies in agriculture, as discussed above, other hand, are not significant exporters of these manufacturing about half of global trade is affected—either favored by or products.40 A possible factor could be that subsidized exports competing with distortive subsidies. of industrial goods, including parts and components, prevent developing countries from entering manufacturing value Manufacturing exports that are subject to trade-distorting chains; this may especially be the case as they lack the subsidies can be a stepping-stone to higher value-added resources to counter the effects of other countries’ subsidies. trade. Manufacturing exports are more prevalent in high- This in turn can limit the growth potential that trade offers low- income countries, such as the Czech Republic, Hungary, and middle-income countries, as participation in manufacturing Malaysia, and the Slovak Republic, where they can account value chains is typically associated with higher investment for as much as 70 percent of goods exports. Indeed, there is a and technological spillovers. large positive correlation between the share of those products 39. About 80 percent of global trade is contained in HS chapters 28 and above, which cover merchandised trade excluding food and agriculture, minerals, and fuels. 40. The Central African Republic is an outlier, where the high percentage of manufacturing exports reflects exports from sawmilling and planing of wood, and to a lesser extent vehicle parts. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 38 >>> Figure 3.12. Middle-income countries are particularly exposed to distorted manufacturing trade (Share of trade and production for distorted manufacturing trade) 100 90 80 PHL MEX 70 CHN CRI 60 THA BWA Percent 50 MMR VNM WSM 40 CAF TUN SRB LBN 30 IND MAR MKD DOM ZAF BIH BRB BGR PSE IDN NRU 20 10 - 6.0 7.0 8.0 9.0 10.0 11.0 12.0 Log GDP per capita, PPP Tra de Production Source: Original calculations for this publication. Note: GDP = gross domestic product; PPP = purchasing power parity. Reforming subsidies in agriculture and manufacturing in the rest of the world combined, with the global supply of can affect regional trade and production patterns and manufactures largely unchanged (less than 0.1 percent lower). economic welfare. These economic effects can be analyzed, Furthermore, as subsidies can cause global production and based on economic modeling, under the scenarios where all trade to deviate from their otherwise optimal levels, they can trade-distorting subsidies on agriculture and manufacturing create misallocations or allocative inefficiencies that result in are eliminated.41 These indicative simulations show that both deadweight losses to the global economy. exports and production would partially shift from regions with more subsidies to others with less. For example, conservatively in the medium term, export of agricultural products can expand in Sub-Saharan Africa by close to 5 percent. Under this scenario, agricultural output would fall in North America and the European Union (by about 4 percent combined), while it would increase in all other regions by between 1 and 3 percent (by about 2 percent combined), as (more) subsidized agricultural products are substituted by their own products. Global agriculture supply, after removing trade-distorting subsidies, would remain largely unchanged (0.1 percent lower). Similarly, regional production patterns would shift in the case of removing trade-distorting manufacturing subsidies, decreasing in Asia and increasing 41. These scenarios are simulated separately for agriculture and manufacturing. Analysis details are in Appendix F. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 39 4. >>> International Disciplines and Gaps on Subsidies Disciplines on subsidies exist in a wide range of international instruments comprising multilateral, regional/plurilateral, and bilateral treaties and arrangements.42 These disciplines vary widely in approach and scope. They may be subsidy-specific or may regulate subsidies within the context of a broader economic or noneconomic governance system. Moreover, they may take the form of treaty law or nontreaty commitments. The enforcement and implementation of such disciplines may rely on a dispute-settlement mechanism or a nonbinding mechanism that relies on peer pressure. The current governance system for subsidies is a patchy framework with few binding, substantial rules, limited sectoral coverage, and weak institutional settings. While recent international agreements, particularly free trade agreements (FTAs), introduce small innovations focused on expanding the system, current international disciplines remain subject to major gaps that undermine their effectiveness. At the global level, the virtual absence of rules on subsidies for cross-border services and foreign investment limits the governance framework for subsidies. WTO Subsidies and Countervailing Measures (SCM) rules apply only to goods, and GATS rules provide little guidance on services. Rules related to trade in goods are further hampered by a narrow definition of granting authorities, which limits the potential to address subsidies offered by SOEs and other nongovernment entities. Broad and unclear language on the regulation, particularly regarding the “specificity” or “selectivity” of the measures to be captured as subsidies, also reduces the effectiveness of the framework. Substantive rules on subsidies in goods trade are more comprehensive but remain hampered by strict requirements on how to apply them. In particular, the SCM requirement that actionable subsidies effectively must have an “adverse effect” on other members severely limits the scope of the regime. Free-trade agreements have expanded the list of subsidies that 42. This section draws on the work of Kreier, Remy, and Kyriakou (2022). EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 40 need not demonstrate adverse effects (that is, “prohibited” Only economic transfers by certain “granting authorities” subsidies), but the SCM rules limit that framework to export to beneficiaries in a WTO member country fall under subsidies. the SCM agreement. Granting authorities may be central governments and all subnational government authorities, The multilateral institutional setting of the WTO SCM, “public bodies” vested with elements of “governmental focused on increasing transparency and providing ex authority,” and private entities “entrusted” or “directed” by a post remedies, offers only superficial rules and weak government to make a financial contribution. Under these incentives for compliance. Rules on the notification of state rules and existing WTO jurisprudence, state ownership and aid measures are marred by shortcomings that hinder their control of entities such as SOEs are not sufficient to classify intended transparency benefits. The WTO offers a strong those entities as “public bodies.”43 rule-based dispute settlement mechanism, but disputes on the use of subsidies fit poorly in its procedural rules and, The SCM Agreement focuses on manufacturing.44 importantly, provide prospective-only remedies. Agricultural subsidies that conform to specific requirements under the WTO AoA may be protected under the SCM Agreement. Also, subsidies that benefit the service sector 4.1 Existing Rules are excluded from the scope of the SCM agreement and are, instead, to be further negotiated under Article XV of GATS, which has thus far failed to yield services-specific disciplines. There are three major sources of international disciplines on subsidies: (a) the rules contained in the “covered WTO rules apply to a specific set of “financial agreements” of the WTO, which are nearly universal in contributions” that confer a “benefit.” Economic transfers geographical scope; (b) the state aid system of the EU; and covered by the SCM Agreement are enumerated as direct (c) subsidy-related provisions in FTAs among some of the transfers of funds; potential direct transfers of funds or liabilities; world’s major trading powers, including the Comprehensive the forgoing of revenue that is otherwise due; the provision and Progressive Agreement for Trans-Pacific Partnership of goods and services other than general infrastructure; (CPTPP), the United States-Mexico-Canada Agreement the purchase of goods (although not of services); and price (USMCA), and EU-led agreements. The WTO offers the only or income support. Notably, the SCM Agreement does not global framework and remains a benchmark for disciplines cover regulatory actions that do not involve a listed financial on subsidies, especially those related to trade in goods; the contribution, including export restrictions that can provide an EU and, to a smaller degree, FTAs, have elaborated on those implicit subsidy to downstream activities. Importantly, these rules, offering relevant guidance on ways to improve the measures must confer a “benefit” to its recipient to constitute multilateral framework. a subsidy, understood as a financial contribution being made. Finally, financial contributions must be “specific” to an W O R L D T R A D E O R G A N I Z AT I O N enterprise or industry to fall under SCM rules. A subsidy is de jure specific to an enterprise or industry when there is The WTO Agreement on Subsidies and Countervailing an express regulatory limitation on access. It may also be Measures features the most comprehensive global de facto specific, depending on whether it is predominantly rules on subsidies, but they focus only on goods trade. or disproportionately used by certain enterprises, and on how Narrower disciplines are contained in the General Agreement much discretion the granting authority exercised in deciding on Tariffs and Trade (GATT) 1994, the General Agreement to grant the subsidy. A subsidy is also considered specific on Trade in Services (GATS), the Agreement on Agriculture when limited to a certain region in a designated geographical (AoA), the Agreement on Trade-Related Investment Measures area. Finally, subsidies that are contingent on export or the (TRIMs), and several WTO Accession Protocols. This analysis use of domestic content (that is, prohibited subsidies under focuses predominantly on the rules of the SCM Agreement. 43. See WTO Appellate Body Report, DS379: United States—Definitive Anti-Dumping and Countervailing Duties on Certain Products from China, 2012, paras. 318 and 353. This ruling regarding the meaning of a “public body” has provoked substantial legal and policy debate. 44. Investment and services subsidies are indirectly governed by the nondiscrimination provisions of GATS and TRIMs. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 41 WTO rules) are deemed to be specific. On the other hand, 107(1) of the TFEU also includes a “material selectivity” or a subsidies based on “objective criteria,” such as the number “regional selectivity” threshold for the state aid to be subject to of employees or the size of the enterprise, are not considered its disciplines—an element that connects with the “specificity” “specific” and hence fall out of the SCM coverage. requirement in WTO rules. Subsidies may be “prohibited” or “actionable.” The SCM The EU state aid framework contains horizontal and Agreement contains two principal sets of rules: prohibitions sectoral disciplines. At the horizontal level the Council of the for a limited set of subsidies that are considered most likely European Union and the European Commission have devised to distort trade, and a cause of action for all other specific special rules applicable to regional aid, R&D, environmental subsidies that can be demonstrated to have “adverse effects” aid, and other forms of non-sector-specific aid. At the sectoral on the interests of another member. Subsidies contingent level, separate EU regulations govern areas such as finance, on export or local content—either de jure or de facto—are transport, and fisheries. prohibited outright. Actionable subsidies—subsidies that are not local content- or export-contingent but cause adverse State aid under EU rules must first be notified and effects (that is, serious prejudice, material injury, or nullification approved, with fines imposed for noncompliance. The and impairment)—can be challenged through WTO dispute European Commission plays an active role in the EU state aid settlement with a view to seeking their withdrawal. rules, examining notifications and granting approval, as well as imposing fines and defending its decisions before the Court All specific subsidies, whether prohibited or actionable, of Justice of the European Union. The remedy for unlawful may be subject to countervailing duties. To impose state aid is recovery from the beneficiary. countervailing duties, the importing member must show that the subsidized imports are causing or threatening to cause injury to its domestic industry, or that they retard the SELECT FREE TRADE AGREEMENTS establishment of such an industry. Subsidies are also subject to transparency and notification requirements. Traditionally, subsidy-related disciplines contained in FTAs simply reaffirmed member states’ obligations under the SCM Agreement. Recently, certain WTO E U S TAT E A I D L AW members have included special subsidy rules in their FTAs, whether as part of subsidy-specific chapters or The EU state aid framework is arguably the broadest and under chapters dedicated to the regulation of SOEs. The most developed international framework for subsidies. CPTPP, the USMCA, and a range of FTAs concluded by the This framework is laid out in Articles 107 and 108 of the Treaty EU are notable examples of an emerging trend to develop on the Functioning of the European Union (TFEU) and its specific disciplines on subsidies and SOEs in FTAs. In many subsequent regulations. cases, these new provisions reflect a determination to plug perceived gaps in WTO subsidy disciplines. The EU state aid framework applies to all economic sectors and to the central and subnational governments of any EU member state, as well as other public bodies and SOEs to THE COMPREHENSIVE AND PROGRESSIVE the extent that the economic transfer involves the use of AGREEMENT FOR TRANS-PACIFIC PARTNERSHIP “state resources” and may be attributed to the state. The framework requires that the state aid, which can take any form, The CPTPP’s most significant innovations in subsidies including direct grants, loans, direct investment in corporate relate to noncommercial assistance provided both by capital, in-kind benefits, guarantees, and revenue foregone, and to SOEs. The CPTPP is intended to regulate SOEs, entail a direct or indirect “advantage” to the recipient. Article both as potential subsidizers and as recipients of subsidies. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 42 While the concept of noncommercial assistance is built on • Noncommercial assistance provided by a party or SOE the types of financial contributions provided for under the to an SOE, in circumstances where the recipient is SCM Agreement,45 CPTPP covers SOEs operating in both insolvent or on the brink of insolvency, without a credible the goods and service sectors. Importantly, an enterprise is restructuring plan to return the SOE within a reasonable an SOE under the CPTPP if the government has majority period to long-term viability; and ownership, controls a majority of voting shares, or has the • Conversion by a party or SOE of the outstanding debt power to appoint a majority of the management board. of an SOE to equity, in circumstances where this would be inconsistent with the usual investment practice of a The main provision mandates that no party cause adverse private investor. effects or injury to the interests of another party by using noncommercial assistance provided to or by SOEs. Notably, these disciplines apply where a party is subsidizing the EUROPEAN UNION FREE TRADE AGREEMENTS activities of its SOEs abroad (transnational subsidies), arguably reaching beyond WTO rules. In addition to establishing a state- The EU has developed a significant body of subsidy rules to-state dispute settlement mechanism, the CPTPP creates under its FTAs. Several contain specific chapters on subsidies, new transparency requirements on the operation of SOEs and in addition to chapters regulating SOEs, while others do not the provision of noncommercial assistance. contain specific chapters but still prescribe rules that impose meaningful constraints on subsidies. The EU-South Korea Besides SOE-related rules, the CPTPP contains FTA, for instance, covers subsidies in its competition chapter provisions addressing subsidies in specific sectors. For which, like the USMCA, prohibits unlimited guarantees to example, Article 20.16.5 prohibits subsidies that negatively cover a company’s debts or liabilities and subsidies for ailing affect fish stocks in areas that are overfished, as well as companies without a credible restructuring plan, insofar as subsidies to any fishing vessel engaged in illegal, unreported, they adversely affect international trade (that is, domestic or and unregulated fishing. export markets). FTAs such as those with Replubic of Korea and Singapore include subsidies to services sectors. UNITED STATES-MEXICO-CANADA AGREEMENT The EU-Japan FTA further includes an obligation to ensure that enterprises use subsidies only for the specific purpose The United States-Mexico-Canada Agreement (USMCA) for which they are granted. The EU-Japan FTA permits the includes some of the most far-reaching disciplines on parties to rely on Article XX of GATT 1994 to justify subsidies state aid featured in a free trade agreement. While most granted in pursuit of a closed list of legitimate policy objectives. of the USMCA framework mirrors the scope, content, and governance features of the subsidy-related disciplines set out in the CPTPP, its definition of SOEs extends even beyond the OTHER AGREEMENTS definition given under the CPTPP, encapsulating situations where the government has the power to control an enterprise Energy and environment-related agreements or through an indirect or minority interest. arrangements may include disciplines that limit the use of subsidies in areas that may have a negative impact The USMCA prohibits three categories of noncommercial on the environment. The Paris Agreement on climate assistance provided to or by SOEs of a party. These are change does not contain rules specific to subsidies but applicable to goods, without the need to demonstrate adverse requires the parties to commit to the decarbonization of effects and injury. the global economy. This implies in practice an overhaul • Loans or loan guarantees provided by an SOE to an of the energy subsidy framework, as made explicit through uncreditworthy SOE; the goal to make “finance flows consistent with a pathway 45. Although it notably excludes revenue forgone (presumably because SOEs lack taxing powers) and the purchase of goods. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 43 towards low greenhouse gas emissions and climate- GRANTING AUTHORITIES resilient development” (Article 2.1(c)). The G-20 Fossil Fuel Subsidies Agreement established an overall goal to eliminate Under which circumstances the actions of nongovernmental “inefficient” fossil-fuel subsidies and create a mechanism for entities should be attributable to the state has become country peer reviews of fossil-fuel subsidies. particularly contentious in the context of subsidies granted to and by SOEs. Under WTO rules, the term “public body” in the SCM Agreement has been construed to mean that relevant 4.2 Gaps and Limitations in the Current entities must be “vested with governmental authority” to be subject to the agreement’s disciplines. As a result, the ability Subsidy Governance Framework of a government to control an entity is not sufficient to establish attribution, meaning that even subsidies provided by SOEs that It is widely accepted that existing subsidy disciplines have are wholly- or majority-owned by the state may not be subject not fully kept pace with the recent developments and rising to WTO rules. Given that SOEs in many countries are important challenges of global economic governance. This holds economic actors, particularly as suppliers of inputs and financial particularly true at the multilateral level, where the multiplicity services, this may be seen as a major loophole. Moreover, of the interests at stake, deep differences of views among demonstrating “entrustment” or “direction” of private bodies countries, and the broader gridlock in the WTO negotiating by the state may be particularly challenging, especially where mechanism have barred any meaningful amendment to the measures are either unwritten or not formally communicated. rules. As a result, the governance systems for dealing with subsidies are a patchwork of overlapping instruments, with the most far-reaching reform efforts taking place at plurilateral or TYPES OF ECONOMIC TRANSFERS bilateral levels, particularly in the context of FTAs. While the SCM Agreement contains a list of covered “financial contributions,” certain categories of economic L I M I TAT I O N S R E L AT E D T O T H E S C O P E transfers, particularly those related to regulations, may OF THE CURRENT DISCIPLINES not fit within the list. For example, export restrictions and differential export taxation on inputs, which may confer an SECTORAL COVERAGE advantage to downstream user industries, are not included under the area of “financial contribution” (Hoekman and Nelson The main gap in subsidies rules is the virtual absence of 2020). While there are widely varying views about whether the services sector from its coverage. As the SCM covers these types of government behavior should be addressed the goods sector exclusively, the main substantial GATS rules as subsidies, through other disciplines, or not at all, it may on subsidies in the services sector are its nondiscrimination be useful to evaluate whether and in what circumstances the obligations. There may be several factors playing into this current range of financial contributions could be adjusted. gap, including the difficulty of collecting data; the profound sensitivity of some critical services sectors including health, education, transportation, and banking; and the complex TYPES OF ECONOMIC ADVANTAGES interaction of commercial and noncommercial activity in these services areas. Even so, this lack of disciplines on services In WTO rules, “benefit” is established in relation to the subsidies can be perceived as a major weakness, especially market, which is likely to require a complex, case-specific considering the rapidly rising share of services in global trade. analysis to identify an appropriate benchmark. The Meaningful outcomes in this sector would likely need to be situation becomes particularly difficult where no market exists negotiated through treaty amendments in the WTO, which because the government or public body fully occupies the area appears unlikely to happen in the short term.46 of economic activity, or where the heavy presence of state 46. Potential reforms include adopting a scheduling approach to disciplines, enabling members to select those sectors where they have more flexibility to accept disciplines; blocking exemptions for particularly sensitive sectors or types of subsidies; and grandfathering (through scheduling) existing support systems in specific sectors. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 44 actors distorts the market. While WTO law recognizes that in L I M I TAT I O N S R E L AT E D T O C U R R E N T certain circumstances a resort to proxies may be required, S U B S TA N T I V E R U L E S the circumstances in which this is allowed, and the way such proxies are constructed, remain unclear. Given that this is a “PROHIBITED” SUBSIDIES highly technical issue it could be addressed through technical discussion in an expert group to develop recommendations. While current SCM Agreement rules are designed to Examples of these include the groups used by the Tokyo address trade-distorting subsidies, the list of subsidies that Round Antidumping and SCM Committees, or by the WTO are deemed to distort trade and are thus a priori prohibited SCM Committee. is very short. In effect, only subsidies that are contingent upon exportation (export subsidies) or upon the use of domestic over imported goods (local content subsidies) are prohibited. Some SPECIFICITY AND SELECTIVITY WTO members argue that the SCM Agreement would be more effective if it prohibited other categories of subsidies that, while The current rules do not provide concrete guidance on not conditioned upon the achievement of trade objectives, are how targeted (that is, specific or selective) a subsidy must the most likely to distort trade because of their magnitude, be to be subject to discipline. For example, it is unclear their role in creating economic actors or productive capacity whether subsidies to the entirety of a country’s agriculture (or maintaining such actors’ capacity when not economically sector can be deemed specific. Also, the application of the viable), or their involvement in granting access to inputs on a specificity test may exclude from discipline certain subsidies discriminatory basis.48 with substantial trade-distorting effects, such as the low- cost provision of energy or other inputs economy-wide. This area too could benefit from convening a group of experts to DEMONSTRATION OF EX POST ADVERSE TRADE consider whether the existing rules regarding specificity would EFFECTS benefit from clarification and, more generally, whether such rules are useful and appropriate in all circumstances. As discussed earlier, subsidies that do not fall under the “prohibited” category are only covered under the SCM Agreement when they cause “adverse effects” to the TRANSNATIONAL SUBSIDIZATION interests of another member. As a result, members must wait until they have suffered (possibly irreversible) economic Issues related to subsidies provided by one member that harm before they can challenge the measure in WTO dispute affect economic activity in another member have become settlement. This is particularly problematic when coupled with more common in recent years. This is the case both in the the prospective nature of WTO remedies, which can neither context of countervailing duty investigations involving imports undo the harm suffered nor offer monetary compensation. There from a third country, and in situations where a member argues is thus little incentive for members to avoid causing adverse that subsidies provided by an external party to entities within effects through their subsidy measures. Significant evidentiary its territory are distorting internal competition (European challenges make it difficult to demonstrate adverse effects. Commission 2020). There is legal uncertainty as to the extent to which such subsidies are, or should be, within the scope Strengthening the dispute-settlement system could of the SCM Agreement.47 Clarity regarding these issues in principle address the weakness in WTO subsidy would be desirable before any comprehensive treatment of disciplines. Deadlines could be shortened or strictly enforced, transnational subsidies is entertained. and remedies could be strengthened to include recouping subsidies or even monetary compensation. However, the dispute settlement system is currently under severe pressure, considering the inoperability of the Appellate Body, and 47. Crochet and Hegde (2020), for instance, conclude that transnational subsidies can be covered under Article 3, which does not contain the limitation of “within the jurisdiction” stipulated under Article 2.1. 48. See, for example, the approach under the USMCA and a number of EU FTAs. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 45 consensus even on broad reform areas or directions is lacking. a two-year cycle. In practice, this means that even timely A shift to more prescriptive rules, based on prohibitions or notifications of subsidies may be two or three years old before rebuttable presumptions for certain subsidies, would therefore they are subject to review. Moreover, since notifications appear to be a more viable alternative. rely on self-identification, WTO members often fail to send notifications of subsidies entirely. This may happen out of a lack of understanding on whether a measure is covered by DISTORTIONS IN INVESTMENT FLOWS the SCM agreement, or often, to avoid “self-incrimination,” especially regarding the subsidies that are the most likely to be Investment distortions can harm the economic welfare of challenged: those most trade distortive. Lack of notification is affected businesses and workers just as much as trade particularly common for subsidies adopted at the subnational distortions.49 Furthermore, investment distortions may be seen level, largely due to low awareness of WTO disciplines. as precursors to trade distortions, as shifts in the distribution of production may ultimately affect trade flows. While multilateral rules on investment have proved highly controversial in the DISPUTE SETTLEMENT AND REMEDIES past, possible disciplines related to investment-distorting subsidies could strengthen the ability to mitigate the adverse Subsidy-related disputes have proven particularly effects of mobile capital of multinational corporations. Any new challenging. First, subsidy-related disputes are heavily rules would require changes to the SCM Agreement. evidence-based. Second, since the principal subsidy disciplines are effects-based, the process of gathering evidence and conducting intensive economic analysis50 is NONTRADE EFFECTS OF CERTAIN SUBSIDIES difficult and extremely time-consuming for WTO dispute settlement. Critically, economic analytical exercises rely on A comprehensive reform of the international governance part-time, ad hoc panels that lack the subpoena and other system for subsidies should aim to address the nontrade investigative powers of national judicial systems. Third, effects. The current multilateral framework of the WTO neither WTO dispute settlement is inherently prospective in nature. protects subsidies that seek to advance or protect legitimate While a prospective remedy may be reasonably effective policy objectives (for example, the environment, R&D, health, and in addressing recurring subsidies, it is notably ineffective in welfare) from challenge nor seeks to balance the trade-distorting dealing with large, one-off subsidies that create or maintain effects of a subsidy against the impact of achieving the policy economic activity (for example, the creation of new production objectives. Potential ways to address this include the resurrection capacity) and whose effects may continue for decades. Under of a nonactionable category of subsidies, the creation of a set WTO law there are no obvious ways to remove the adverse of general exceptions analogous to GATT Article XX, or the effects of subsidies, which is aggravated by the fact that in development of a facilitated mechanism to review and approve certain cases the complainant must wait until negative effects waivers from subsidy disciplines in appropriate circumstances. emerge before being able to bring the claim. L I M I TAT I O N S R E L AT E D T O DATA COLLECTION GOVERNANCE AND OTHER SYSTEMIC CONCERNS One of the main challenges for the international governance of subsidies is the poor quality of comparable statistical TRANSPARENCY AND NOTIFICATION data. e of the most frequently expressed concerns relates to the limitations of data collection on the extent of subsidization The current WTO SCM framework relies on notification provided by governments around the world (See Hoekman requirements to increase transparency on the use of 2015; Horlick and Clarke 2017). Currently, there is no global subsidies, but these are often late and incomplete. WTO information hub, and importantly, there is no agreed-upon rules provide for the ex post notification of subsidies on common methodology for the collection of data on subsidies. 49. Johnson and Toledano (2013). 50. Some observers also believe that the WTO dispute settlement has been reluctant to fully embrace economic analysis. See Mavroidis and Neven (2017). EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 46 5. >>> Moving Ahead on Subsidies and Trade Closing data and analytical gaps should remain a priority area in the discussion of subsidies. Subsidies remain a delicate and poorly documented matter in global trade policy. Understanding which subsidy measures are inefficient or ineffective in achieving their intended goals and their potential effects on other countries by distorting trade would require detailed information on the design, implementation, and beneficiaries of such interventions. This may require enhanced cooperation in fulfilling notification requirements, as well as other actions by peers (see a short discussion below on some possible mechanisms). International organizations can play a facilitating role and provide guidance as to how best to consolidate this information and monitor subsidies (within and across countries).51 More knowledge is needed on subsidies and their international implications. There is room for greater collaboration to improve the collection, vetting, and dissemination of data on subsidies. Enhanced information on subsidies, including indirect subsidies by nongovernmental entities such as SOEs, could naturally feed into a research workstream to better assess the role of subsidies on trade. This work could be done in collaboration with international agencies with global coverage such as the World Bank, the International Monetary Fund (IMF), WTO, UNCTAD, and the OECD, as well as external experts on subsidies. Technical experts and government entities could be encouraged to participate in collecting and vetting data.52 This effort could initially focus on a limited set of countries that, as shown in the current analysis, disproportionally employ subsidy measures. Nonbinding policy guidelines issued by an independent body could have a high impact, given the prevalence of subsidy interventions and the need to inform multilateral reform discussions. Limited transparency should not preclude (or be used to preclude) progress on reforms where agreement can form. Shortfalls of information about the source, scale, and extent of subsidies can deter reform as much as political gridlock. This is particularly true in contexts where it is virtually impossible to know how governments are assisting the private sector. 51. IMF, OECD, World Bank, and WTO (2022). 52. To ensure success and broad-based participation, such an exercise needs to be (a) delinked from any ongoing efforts in rulemaking; (b) conducted without prejudice to any consequences under existing binding or nonbinding regimes; and (c) able to manage confidential data. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 47 A comprehensive global framework for subsidies is ideal to subsidy rules could be expanded to include transnational but unlikely. Any sensible proposal for governance reform subsidies as well as export duties and taxes. must distinguish between the desirable and the feasible. At the • Increase transparency through peers’ actions. multilateral level, the current gridlock in broader WTO reforms Shortcomings in the notification process could be addressed limits any undertaking from being adopted. The sensitivity through (a) peer pressure in the SCM Committee; (b) the of subsidies governance adds another layer of caution. involvement of successive SCM Committee chairs; (c) While most countries would agree that better coordination the publication of “status of notification” documents; (d) in formulating subsidy rules may be necessary, there is no the use of counternotifications; (e) technical cooperation consensus on the design of these rules or what role the WTO— by the Secretariat; and (f) the work of the WTO’s Trade or other bodies—should have in developing or enforcing them, Policy Review Body. Moreover, institutional or legal much less on which substantive areas should be prioritized. consequences could be attached to non-notification, and Thus, it is unlikely to expect any major multilateral effort toward transparency could be improved by creating domestic reform of subsidy rules in the foreseeable future. More likely bodies with the mandate and legal authority to collect is a continuation of the divergent approaches to, and levels information from national and subnational government of ambition in, subsidy rules, as reflected in recent bilateral agencies and report to the WTO. and plurilateral FTAs. Sectoral disciplines under nontrade • Use soft law options to fill loopholes. Instead of instruments may also attract more regulatory activity. improving subsidy rules through treaty amendments (which in the current climate may be unfeasible), the suggested The nature of subsidies requires the agreement of all recommendations could be implemented through major trading partners. Unlike tariffs and other market alternative means such as authoritative interpretations, access deals, whose benefits can be limited to participating waivers, or expert group recommendations for more parties in regional trade agreements, subsidy disciplines are technical issues (including the calculation of benefit or the unlikely to succeed unless all the major subsidizers come determination of “specificity”). to an agreement, mainly due to challenges surrounding the free-rider problem. This means that solutions ultimately must be negotiated multilaterally, especially in the WTO or in other forums that include all the major economies. In this context, some priority steps include the following: • The need for subsidy disciplines in cross-border services and investment. While reform on this area may be beyond the current appetite, the importance of the issue requires keeping it on the table. Increased transparency on the use of subsidies in the services sector, including on investment in services, could be negotiated as “additional commitments” to the GATS. • Support multilateral rules with bilateral interpretations. Even where disciplines exist under the SCM Agreement, there is a need to adapt its rules to reflect current jurisprudence and realities. For instance, interpreting the meaning of a “public body” could be guided by definitions used in FTAs, and the types of government action subject EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 48 >>> Appendix A: Antidumping Measures The application of antidumping measures has been increasing continuously, albeit at a slower pace than countervailing measures. Antidumping measures around the world increased by 46 percent from 2007 to 2019, with growth averaging 3.2 percent per year (figure A.1).53 In 2019, about 2,000 antidumping measures were in place, or about 10 times more than countervailing measures. That compares with 2007, when antidumping measures were more than 20 times as numerous as countervailing duties, and 2013, when antidumping measures were 15 times more common than countervailing measures. >>> Figure A.1. Applied Antidumping Measures Per Year, 2007–2019 2,500 2,000 Number of measures 1,500 1,000 500 0 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Source: World Bank Temporary Trade Barriers Database. A diverse group of rapidly emerging economies has become top users of antidumping measures. Starting from essentially zero, India rapidly started implementing antidumping duties at the end of the 1990s. In 2003 it passed the EU as the second-largest user of global antidumping measures after the United States. Brazil and Türkiye have also increased their use of antidumping measures very rapidly, Türkiye in the early 2000s and Brazil after 2012. Together, the United States, India, Brazil, and Türkiye account for slightly over half of the global total (figure A.2). China is by far the largest target of antidumping measures, accounting for slightly more than one third of the global total. China is the principal target for all top 10 users of antidumping measures, as shown in figure A.2 (excluding China itself). Meanwhile, most of China’s antidumping measures are aimed at the United States, Japan, the EU, and Republic of Korea (figure A.3). 53. This is based on the World Bank Temporary Trade Barriers Database. Chad P. Bown, Milla Cieszkowsky, Aksel Erbahar, and Jose Signoret. 2020. Temporary Trade Barriers Database. Washington, DC: World Bank. www.worldbank.org/en/data/interactive/2021/03/02/temporary-trade-barriers-database. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 49 >>> Figure A.2. Share of Global Antidumping Measures by Top Users, 2007–2019 90 80 70 60 50 Percent 40 30 20 10 0 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 USA IND BRA TUR CHN ARG EUN CAN ME X AUS Source: World Bank Temporary Trade Barriers Database. Note: ARG = Argentina; AUS = Australia; BRA = Brazil; CAN = Canada; CHN = China; EUN = European Union; IND = India; MEX = Mexico; TUR = Türkiye; USA = United States. >>> Figure A.3. Share of Applied Antidumping Measures by Target Economies, 2007–2019 80 70 60 50 Percent 40 30 20 10 0 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 CHN KOR TWN IND THA USA JPN IDN RUS VNM Source: World Bank Temporary Trade Barriers Database. Note: CHN = China; IDN = Indonesia; IND = India; JPN= Japan; KOR= Korea, Rep.; RUS= Russian Federation; THA = Thailand; TWN = Taiwan, China; USA= United States; VNM = Vietnam. Antidumping measures are highly concentrated in the same sectors as countervailing actions (figure A.4). Iron and steel together with iron and steel products (under HS chapters 72 and 73 respectively) accounted for about 6 of every 10 targets of antidumping duties in 2019. Similarly, plastics and plastic products and paper and paperboard are commonly targeted by trade remedy actions. Meanwhile, chemicals and synthetic textiles are common in antidumping cases but not in countervailing cases. This in part reflects the application of antidumping measures in these sectors in countries such as India, Mexico, and Türkiye, which rarely apply countervailing duties. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 50 >>> Appendix B: Subsidy Measures Classification The following provides a schematic guide to the type of information to be collected and classified for an analytical database on subsidy measures.54 1. Type of measure through which subsidy is provided* • Law/legislation/regulation • Administrative practice/direction • Omission • Other (state) 2. Objectives or objectives of measure (Use direct language where explicitly stated.) • Research, innovation, technology • Environment, social impact • Financial bail-out/relief • Disaster relief • Other (please state) 3. Good/service/sector in which economic activity takes place* • Goods ◦ Manufacturing (specify) ◦ Agriculture (specify) ◦ Fishing (specify) • Services ◦ Business (including professional and computer services) ◦ Communication (including postal, telecoms, and courier) ◦ Construction and related ◦ Distribution ◦ Educational ◦ Environmental ◦ Financial ◦ Health and social ◦ Tourism and travel ◦ Recreational, cultural, and sporting ◦ Transport ◦ Other (The Harmonized System can be consulted to categorize good at the chapter, heading, or subheading level as possible. For services see WTO Services Sectoral Classification List: MTN.GNS/W/120. A sectoral classification following the ISIC nomenclature is also recommended.) 54. An asterisk (*) in the headings below designates mutually exclusive criteria. Provided options under heading are for initial guidance and can be regrouped as necessary. General considerations include: i) a single measure may include different subsidies; each subsidy must be separately assessed; ii) the form of a measure is not necessarily dispositive of its ultimate categorization and classification: A complete review of the measure’s design, structure, and operation is needed; iii) it is preferable to err on the side of being overly inclusive; iv) where a single measure can fall into different categories, a determination should be made as to the most appropriate category, and if not possible, both categorizations could be included, with relevant indications to avoid double counting; v) and references/citations should be provided as possible to ensure the work is easily verifiable. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 51 4. Category of transfer of economic value or (nonfinancial) assets* • Direct transfers of funds ◦ Grant ◦ Grant ◦ Loan ◦ Equity infusion ◦ Joint venture ◦ Debt forgiveness ◦ Transfer of liability ◦ Transfer of shareholding ◦ Other • Potential direct transfers of funds or liabilities ◦ Guarantee ◦ Insurance ◦ Issuance of line of credit ◦ Other • Foregoing of government revenue otherwise due ◦ Indicate the type of revenue forgone: ▪ Income-related taxes (e.g., on wages, profits, interest, rents, royalties, and other income) ▪ Product-related internal taxes (such as sales, excise, turnover, and value-added taxes) ▪ Real property taxes ▪ Import and export charges (such as tariffs and duties on imports and exports) ▪ Social welfare charges ▪ Other ◦ Indicate the manner in which revenue forgone is forgone ▪ Exemption ▪ Remission ▪ Deferral ▪ Other • Provision of goods, services, real property, or other nonfinancial assets: ◦ Provision of goods/services ◦ Provision of real property right/intellectual property right ◦ Other • Purchase of goods, services, real property, or other nonfinancial assets: ◦ Purchase of goods/services ◦ Purchase of real property/intellectual property right ◦ Other • Other EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 52 5. Granting authority (indicate, where possible, the nationality and name of the granting authority) • National governments • Subnational authorities • Inter-, supra- and transnational institutions • Parastatal entities (specify how control is exercised (e.g., through ownership/specific transactions) • Private actors/ bodies acting under the instruction of a public authority (specify the public authority) 6. Recipient and/or beneficiary of transfer • Recipient of economic/asset transfer (specify whether first, second, third, and so forth) • Ultimate beneficiary of subsidy (if applicable, indicate how is different from prior recipients) • Identify market (if one does not exist, indicate and describe the constructed market) • Amount/value of benefit 7. Selectivity criteria (Indicate basis for selectivity and whether de jure or de facto)* • Enterprise • Sectoral • Territorial • Trade-related: export or domestic content contingent EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 53 >>> Appendix C: Subsidy Objectives (Broad Categorization) >>> Table C.1. Subsidy Objectives Objectives Scope General sectoral financial support; for the acquisition of capital goods, materials, 1 Sectoral business operations or not earmarked Reorganization, upgrades, technical transformation, and other measures to 2 Industry restructuring enhance competitiveness 2.1 Overcapacity reduction Research and development/ Financial support for research and development of new technologies and 3 innovation innovations in various fields/sectors Income supplement for struggling firms or individuals; measures of social inclusion/ 4 Social integration such as leisure and social activities Measures related to expanding employment options and/or opportunities for 5 Employment individuals Small and medium enterprises Financial support for the creation or growth of small and medium enterprises, seed 6 and entrepreneurship capital, venture capital, and so forth Training and instructional sessions (or the financing of such programs) geared at increasing the knowledge of businesses, employees, municipalities, and 7 Capacity building so forth, on topics such as environmental protection, business management, entrepreneurship, industry-related workshops, and so forth The provision of business-related services or the financing thereof; for example, 8 Business support/services marketing, accounting, management, and taxation Measures geared at facilitating the creation or attraction of new businesses, 9 Economic/regional development economic diversification, and job creation in a given region Measures aimed at the creation or enjoyment of works in the arts, film, music, and 10 Cultural enrichment literature industries, among others Financial support to maintain and/or pass on national traditions to a new generation, such as specific hunting/fishing techniques, arts and crafts skills and 11 National traditions/heritage styles, as well as the maintenance of historical, geographic, natural, or artificial landmarks or monuments Support for ethnic minorities through income supplement, business support, 12 Support for ethnic minorities capacity building, employment, and so forth Support for people with Support for people with disabilities through income supplement, business support, 13 disabilities capacity building, employment, and so forth Support for gender minorities through income supplement, business support, 14 Gender-related support capacity building, employment, and so forth EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 54 Objectives Scope Efforts to preserve certain species or ecosystems, reduce earth/air/water pollution and conserve resources through tax credits or other measures to incentivize 15 Environmental preservation the adoption of more environmentally friendly and/or sustainable production, processes, transportation, and so forth—except clean energy Measures aimed at increasing the production or substitution of renewable, more environmentally friendly energy for traditional ones. Note: Hydroelectric power, for 16 Clean energy the purpose of this classification, though cleaner than the energy produced with fossil fuels, is not listed as clean energy due to the severe environmental disruption it causes. Measures to encourage the prospecting, exploration, or management of fossil fuel sources and hydroelectric power. This classification also includes measures to 17 Fuel/energy provision offset taxes on or fuel prices (regardless of their source) as well as tax credits and other financial rewards for energy/fuel conservation. Construction or installation of networks necessary to improve transportation, 18 Infrastructure communication, the provision of certain public services, and so forth Financial support for government agencies, municipalities and/or private firms performing public services such as sewage treatment and trash collection and 19 Public services disposal, among others (except construction of infrastructure and provision of transportation) Financial aid to facilitate the management, modernization, and/or repairs of 20 Transportation public transport vehicles, as well as the provision of transportation services for passengers and businesses Financial support for regions/industries that are susceptible to or affected by Disaster/climate change 21 disasters and other harmful events (whether natural or artificial) that cause capital preparedness and response losses (such as agricultural pests, droughts, wildfires, and meteorological disasters) Prevention/reduction of accidents and disease; medical services, incentives for a 22 Public health/safety healthier lifestyle Subsidy to increase/incentivize initial public offerings (IPOs) and other participation 23 Financial markets participation in financial markets 24 International trade Subsidies to finance or encourage imports/exports 25 FDI Foreign direct investment 26 Water conservation/treatment 27 Children/adult education Grade-school and above education Financial aid for the payment of governmental or professional certification fees and Regulatory compliance, 28 inspections, for the maintenance or improvement of conditions to maintain or obtain certification acquisition such certifications, or to meet regulatory or industry standards 29 State-owned enterprises Financial aid to government-owned (or partially owned) firms and farms 30 PPPs Public-private partnerships Source: Authors EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 55 >>> Appendix D: Summary Description of Subsidy Values For slightly over half of collected subsidy programs in the database, subsidy values were reported and entered. While the missing information precludes a robust treatment of this information, a brief description of some patterns is provided here, subject to the data caveat that this information can be missing in important ways both across countries and for certain types of programs within a country. The largest providers of subsidies across economies are the EU, China, and the United States, with the largest value for the EU in the range of US$250 billion. The fact that these are the largest values in part reflects that these are the largest economies. In fact, there is a positive correlation between (log) GDP and total subsidies, either in absolute value or as percent of GDP. Among the largest subsidy programs are payments to farmers in the United States (of over US$64 billion) and in the EU (under the First Pillar of the EU Common Agricultural Policy with close to US$63 billion). Also substantial is the European Regional Development Fund, with a budget of about US$52 billion. Other large programs are China’s Fund for Development of Agriculture (with US$30 billion) and Türkiye’s Eximbank Export Credits for the manufacturing industry, with close to US$40 billion. Table D.1 shows at a high level, which sectors may be the primary beneficiaries by country. >>> Table D.1. Primary Sectors Receiving Subsidy Resources by Economy Country Sector Country Sector Belgium g R&D Norway g Agriculture France g R&D Mexico g Agriculture Israel g R&D Indonesia g Agriculture Türkiye g Manufacturing Chile g Agriculture India g Manufacturing Argentina g Agriculture European Union g Agriculture Russia g Transport equipment China g Agriculture Canada g Renewable energy United States g Agriculture Netherlands g Renewable energy Japan g Agriculture Italy g Multiple sectors Switzerland g Agriculture United Kingdom g Multiple sectors Brazil g Agriculture Germany g Multiple sectors Korea, Rep. g Agriculture Australia g Multiple Sectors Note: R&D = research and development. 55. Several countries do not reflect subsidy values at all. For the United States, subsidy amounts for subfederal programs are not reported. For China, subsidy expenditures for tax incentives are generally missing. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 56 >>> Appendix E: Additional Information on Econometric Analysis This Appendix describes in brief form the basis of the modeling frameworks developed to introduce the role of subsidies, including their assumptions and derived equations for estimation, in the principal estimations.56 The model starts from the basic premise that economic agents in each country produce and consume and, across countries, engage in trade. This basic idea leads to some principal equations, following economic theory, for the structures on preferences, technology, and trade interactions. In this framework, the value of production is assumed to follow a standard Cobb-Douglas function, while consumer preferences are described by a constant elasticity of substitution (CES) utility function, nested within the Cobb-Douglas utility function across sectors. Each consumable consists of a variety differentiated by its place of origin (as in Armington, 1969). Meanwhile, trade is assumed to be subject to iceberg costs. In addition to being increased by trade frictions of this nature, delivered prices can be influenced by policies in the destination markets as well as by subsidies by the country of origin. The model can be solved in two stages: First, the solution of the upper level determines the aggregate variables in the model— for example, sectoral output and expenditure. Then, the solution to the lower level gives the bilateral trade flows. The solved model arrives at a structural system of subsidies, production, and trade as described in the following system of equations: . In a nutshell, these equations describe the economic equilibrium conditions for exports (X) of industry l from country i to j as in the first equation, which will be a principal objective for the current estimation. The other three equations describe the outward multilateral resistance (Π), consumer prices (P), and the exporter’s level of production (Y). Other variables and parameters in this system include expenditures (E), iceberg transport cost (t), ad valorem tariffs (τ), export and production subsidies (z and s, with the possibility that only a fraction φ of the latter influences exports), and the elasticity of substitution σ. Additionally, an equilibrium output includes the CES preference parameter β, as well as A, L, K, and γ corresponding to production technology, labor, capital, and output elasticity with respect to capital, respectively. 56. See Larch, Signoret, Shikher, and Yotov (2021) for more details on the model and estimation robustness checks. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 57 This theoretical model is translated into an empirical framework that allows us to identify the direct impact of subsidies on production and trade. With this framework, the first equation on the determinants of bilateral trade between trading partners ij is estimated, with the inclusion of country subsidies in i, which are assumed to be applied irrespective of the partner country j. Given that the subsidy data are for a single period of time, the econometric model to be estimated is set in a cross-section. The trade function is specified as a gravity equation as follows: where the dependent variable denotes nominal trade flows in industry l from source i to destination market j. The estimation of the gravity equation takes advantage of production data by sector to include internal trade flows (that is, domestic sales) when i = j. The first two terms in the exponential function refer to exporter-sector and importer-sector fixed effects that control for the structural multilateral resistances as in Anderson and van Wincoop (2003). These fixed effects also aim to control for other exporter-sector and importer-sector unobserved factors. The vector GRAV includes a number of standard gravity variables that proxy for trade cost. These include natural frictions such as distance, as well as policy instruments such as regional trade agreements. Meanwhile, the vector INTRA includes additional factors for domestic trade costs, including an indicator variable for home bias. The last two terms, Z and S, refer to two possible types of subsidies in the i country—the first one to capture subsidies specifically targeted to export activities (that is, trade subsidies) and the second to capture subsidies that are not specific to trade but which could have, nonetheless, effects on trade flows (production supports). As the vast majority of collected subsidy measures are not specific to exports, the current estimations do not make this distinction, so that only one coefficient is estimated. It should be noted that the effect of subsidies on bilateral trade is measured from the perspective of exports, while it could also be looked at from the import side. In that respect, the average effect of subsidies on exports controls for subsidies in the importing country (via the importer-sector fixed effect) but separating this effect would not be possible with cross-sectional data. The gravity equation, in log-linearized form, is estimated via a Poisson Pseudo Maximum Likelihood estimator (Santos Silva and Tenreyro, 2006, 2011). This provides for consistent estimates and takes advantage of the information contained in the zero trade flows. >>> Table E.1. Estimates of Trade and Subsidies Varying by Country Regions (1) (2) Agriculture Manufacturing −0.882 −0.757 Log (distance) (0.105)** (0.034)** 0.592 0.336 Contiguous (0.123)** (0.046)** 0.424 0.438 Language (0.177)* (0.075)** 0.631 0.463 Colony (0.271)* (0.087)** 0.385 0.140 RTA (0.161)* (0.067)* 1.377 0.178 EU (0.273)** (0.098)+ −1.730 −1.815 Sanctions (0.413)** (0.328)** EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 58 10.674 7.799 Domestic (1.625)** (0.766)** −0.613 −0.513 Log (GDP per capita)*Domestic (0.159)** (0.075)** 0.038 0.040 Subsidies AP (0.020)+ (0.013)** 0.222 −0.095 Subsidies EU (0.063)** (0.028)** 0.267 −0.490 Subsidies LA (0.076)** (0.199)* 0.029 −0.171 Subsidies NA (0.007)** (0.045)** −3.460 −7.805 Tariff (1.223)** (0.913)** 11.331 13.527 Constant (0.919)** (0.355)** N 46228 479227 5.442 2.503 Subsidy AVE AP (3.320) (0.820)** 14.096 −2.479 Subsidy AVE EU (6.725)* (0.817)** 3.030 −2.399 Subsidy AVE LA (1.442)* (1.075)* 16.284 −32.959 Subsidy AVE NA (7.485)* (8.488)** Note: Standard errors in parentheses. Contiguous means contiguous borders; language = a common language; colony = a colonial relationship; RTA = membership in regional trade agreements; EU = membership in the EU single market; domestic = domestic sales; GDP per capita allows for this to vary by the level of development of the producing country; tariff = the variable cost of tariff barriers on bilateral trade. AP = Asia-Pacific; AVE = ad valorem equivalent; LA = Latin America; NA = North America. Regressions include exporter and importer fixed effects. + p < 0.10, * p < .05, ** p < .01 EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 59 >>> Appendix F: Effects of Reforming Trade- Distorting Subsidies The trade effects of subsidies, as estimated by the gravity approach in this study, are further utilized to simulate the economic impact of reforming trade-distorting subsidies.57 In these simulations the impacts of export subsidies are translated into impacts on sectoral trade, production, and other economic variables (for example, consumption, prices, income, and so forth). The computable general equilibrium (CGE) framework employed is a multiregion, multisector model that quantifies the medium-run effects in a comparative static specification. Economic sectors include those highlighted in the econometric analysis and in particular three agricultural aggregates (referred to as certain plant products, livestock products, and rest of agriculture) and four manufacturing aggregates (rubber/plastic, machinery, other manufacturing, and rest of manufacturing) in addition to other sectors of the economy. Regional aggregates correspond to one North America region, one region for the European Union, two groupings for Latin America and the Caribbean (subsidizing and the rest of Latin America and the Caribbean), a regional aggregate for Sub-Saharan Africa, one region for the Middle East and North Africa (MENA), two regions for Asia and the Pacific (subsidizing and the rest of Asia-Pacific), and a rest of the world. Two hypothetical scenarios are simulated to describe the elimination of all trade-distorting subsidies on agriculture and manufacturing, separately. To focus attention, the own sectoral displacement effects are examined first (for example, how agricultural subsidies affect agricultural trade and production across borders), and additional effects are reported in this appendix. To develop a baseline, an initial equilibrium is solved for the base year 2017 with subsidy information in the Global Trade Analysis Project (GTAP) database updated along the lines of the estimated ad valorem equivalent (AVE) of subsidies. Agricultural subsidies, for example, are introduced in North America and the European Union at AVEs of 15 percent, of 3 percent for Latin America and the Caribbean countries in the subsidy database, and of a nominal 1 percent for all other regions. Subsidies on manufacturing goods are introduced in a similar stylized form, based on the econometric estimates. The economic simulations show that eliminating the trade distortions of agricultural subsidies would shift agricultural trade flows by reducing exports from North America and the European Union by about 36 and 37 percent, respectively, and in Latin America and the Caribbean by about 1 percent (figure F.1). However, exports from all other regions would be higher. They would increase by about 8 percent in Asia-Pacific and by about 4 percent in MENA, Sub-Saharan Africa, and the rest of the world. These higher exports would amount to about US$27 billion in Asia-Pacific, about US$2 billion in MENA and in Sub-Saharan Africa, and about US$3 billion in the rest of the world. Agricultural production would fall in North America and the European Union (by about 4 percent combined), while it would increase in all other regions by between 1 and 3 percent (by about 2 percent combined), as (more) subsidized agricultural products are substituted by their own products. In this scenario, agricultural output would increase by US$91 billion in Asia-Pacific, US$14 billion in Latin America and the Caribbean, US$12 billion in Middle East and North Africa, and US$7 billion in Sub-Saharan Africa. Agricultural output worldwide, after removing trade-distorting subsidies, would be just slightly lower than without the removal of the subsidies (0.1 percent lower). 57. The general equilibrium effects discussed are based on the Global Trade Analysis Project (GTAP) computable general equilibrium (CGE) model of international trade (Hertel 1997; Corong et al. 2017). A CGE model uses economic statistics and estimated economic parameters to simulate how markets in the global economy might react to trade policies. The model builds on the GTAP database (version 11). EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 60 >>> Figure F.1. Effects on Agriculture of Eliminating Agricultural Subsidie 10 5 0 -5 -10 Percent -15 -20 -25 -30 -35 -40 NA EU LAC AP ME NA SSA ROW Exports Production Source: Original calculations for this publication. Note: Exports for the EU refer to extra-EU exports. AP = Asia-Pacific; EU = European Union; LAC = Latin America and the Caribbean; MENA = Middle East and North Africa; NA = North America; ROW = rest of world; SSA = Sub-Saharan Africa. The effect of removing trade-distorting subsidies in manufacturing would reduce manufacturing trade beyond Asia-Pacific, where the large subsidy AVEs are (figure F.2). Exports from Asia-Pacific would drop by about 6 percent, while exports from North America and the European Union would also be lower by between 6 and 8 percent. This would correspond to lower manufacturing exports by US$338 billion in Asia-Pacific and by about US$110 billion and US$111 billion in North America and the European Union, respectively. Manufacturing goods from Asia-Pacific fuel manufacturing exports from North America and the European Union, so that in the medium term the reduction of subsidies in Asia-Pacific also translate into lower exports from these regions. Similarly, manufacturing exports from Latin America and the Caribbean would also fall, while more modestly by about US$7 billion (1 percent). Manufacturing production across regions, on the other hand, would not fall except for in Asia-Pacific. In regions such as Latin America and the Caribbean, Middle East and North Africa, and Sub-Saharan Africa manufacturing production would increase to partially cover the lower manufacturing trade. Production would increase by US$15 billion in Sub-Saharan Africa (about 3 percent), US$26 billion in Latin America and the Caribbean (1 percent), and US$29 billion in Middle East and North Africa and in the rest of the world (2 percent). EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 61 >>> Figure F.2. Effects on Manufacturing of Eliminating Manufacturing Subsidies 4.0 2.0 0.0 -2.0 Percent -4.0 -6.0 -8.0 -10.0 NA EU LAC AP ME NA SSA ROW Exports Production Source: Original calculations for this publication. Note: Exports for the EU refer to extra-EU exports. AP = Asia-Pacific; EU = European Union; LAC = Latin America and the Caribbean; MENA = Middle East and North Africa; NA = North America; ROW = rest of world; SSA = Sub-Saharan Africa. In both simulated scenarios global production and trade deviate from their otherwise optimal levels. These distortions expand production of subsidized goods at the expense of other goods and services in the economy, creating misallocations or allocative inefficiencies. These inefficient quantities generate deadweight losses to the global economy. From the model, one could calculate those allocative efficiency effects that reflect the removal of the export subsidies, thus ignoring the effect of all other existing distortions which may remain in place (that is, as if there were no other distortions). These effects are in general positive for all regions and would sum to an annual waste of about US$65 billion, with the majority (US$52 billion) related to manufacturing subsidies. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 62 References Anderson, James E., and Eric van Wincoop. 2003. “Gravity with Gravitas: A Solution to the Border Puzzle.” American Economic Review 93(1): 170–92. Anderson, K., and E. Valenzuela. 2008. “Estimates of Global Distortions to Agricultural Incentives, 1955 to 2007.” Washington, DC: World Bank. Anderson, K., E. Valenzuela, and D. van der Mensbrugghe. 2010. “Global Poverty Effects of Agricultural and Trade Policies Using the Linkage Model.” In Agricultural Distortions, Inequality and Poverty, edited by K. Anderson, J. Cockburn, and W. Martin. Washington, DC: World Bank. Armington, Paul S. 1969. “A Theory of Demand for Products Distinguished by Place of Production.” IMF Staff Papers 16(1): 159–176. Borchert, Ingo, Mario Larch, Serge Shikher, and Yoto Yotov. 2021. “The International Trade and Production Database for Estimation (ITPD-E).” Office of Economics Working Paper 2020–05–C, US International Trade Commission, Washington, DC. Corong, Erwin L., Thomas W. Hertel, Robert McDougal, Marinos E. Tsigas, and Dominique van der Mensbrugghe. 2017. “The Standard GTAP Model, Version 7.” Journal of Global Economic Analysis 2(1). Crochet, Victor, and Vineer Hegde. 2020. “China’s ‘Going Global’ Policy: Transnational Subsidies under the WTO SCM Agreement.” Working Paper No 220, Leuven Center for Global Governance Studies, Leuven, Belgium. Dimaranan, B. V. and R. A. McDougall. 2005. “Global Trade, Assistance, and Production: the GTAP 6 Data Base.” Centre for Global Trade Analysis, Purdue University, West Lafayette, IN. Directorate-General for Competition, European Commission. 2020. “White Paper on Levelling the Playing Field as Regards Foreign Subsidies.” European Commission, Brussels. Egas Yerovi, Juan José, and Carmine Paolo De Salvo. 2018. “Agricultural Support Policies in Latin America and the Caribbean: 2018 Review. Inter-American Development Bank, Washington, DC. Evenett, S., and J. Fritz. 2021. “Subsidies and Market Access: Towards an Inventory of Corporate Subsidies by China, the European Union, and the United States.” The 28th Global Trade Alert Report. London: CEPR Press. Fontagné, Lionel, Houssein Guimbard, and Gianluca Orefice. 2022. “Tariff-Based Product-Level Trade Elasticities.” Journal of International Economics 137(July): 103593. Hertel, T. W., ed. 1997. Global Trade Analysis: Modeling and Applications. Cambridge, UK: Cambridge University Press. Hoekman, B. 2015. “Subsidies and Spillovers in a Value Chain World: New Rules Required?” E15Initiative. Geneva: International Centre for Trade and Sustainable Development (ICTSD) and World Economic Forum. Hoekman, Bernard, and D. Nelson. 2020 “Rethinking International Subsidy Rules.” Global Economic Dynamics Working Paper, Bertelsmann Stiftung, Gütersloh, Germany. https://www.bertelsmann-stiftung.de/fileadmin/files/user_upload/MT_ Rethinking_Subsidy_Rules_2020_EN.pdf. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 63 Horlick, G., and P. A. Clarke. 2017. “Rethinking Subsidy Disciplines for the Future: Policy Options for Reform.” Journal of International Economic Law 20(3): 673–703. Irwin, D.A. 2019. “Does Trade Reform Promote Economic Growth? A Review of Recent Evidence.” NBER Working Paper 25927, Cambridge MA. IMF (International Monetary Fund), OECD (Organisation for Economic Co-operation and Development), World Bank, and WTO (World Trade Organization). 2022. “Subsidies, Trade, and International Cooperation.” Washington, DC: World Bank. Johnson, L., and P. Toledano. 2013. “Background Paper for the Eighth Columbia International Investment Conference on Investment Incentives: The Good, the Bad and the Ugly: Assessing the Costs, Benefits and Options for Policy Reform,” November 2013, New York. Kreier, Jesse, Jan-Yves Remy, and Panagiotis Kyriakou. 2022. “International Governance Frameworks for Subsidies: The Existing Rules, Their Limitations, and Paths to Reform.” Background Paper, World Bank, Washington, DC. Kreier, Jesse, Jan-Yves Remy, and Tommaso Soave. 2022. “Subsidies in International Trade and Investment: Scope and Defining Characteristics.” Background Paper, World Bank, Washington, DC. Larch, Mario, Jose Signoret, Serge Shikher, and Yoto V. Yotov. 2021. “State Aid, Production, and Trade: A Structural Estimation Framework.” Background Paper, World Bank, Washington, DC. Mavroidis, P. C., and D. Neven. 2017. “Land Rich and Cash Poor? The Reluctance of the WTO Dispute Settlement System to Entertain Economics Expertise: An Institutional Analysis.” In The Use of Economics in International Trade and Investment Disputes, edited by T. Carpenter, M. Jansen, and J. Pauwelyn. Cambridge, UK: Cambridge University Press. Martinez, Andrew, Jesse Mora, and José E. Signoret. 2009. “The CoRe NTMs Database: A Compilation of Reported Non-Tariff Measures,” Working Paper No. 2009-12-A, Office of Economics, US International Trade Commission, Washington, DC. OECD (Organisation for Economic Co-operation and Development). 2018. “OECD Companion to the Inventory of Support Measures for Fossil Fuels 2018.” Paris: OECD Publishing. https://doi.org/10.1787/9789264286061-en. OECD (Organisation for Economic Co-operation and Development). 2019a. Measuring Distortions in International Markets: The Aluminum Value Chain. OECD Trade Policy Papers, No. 218, OECD Publishing, Paris. OECD (Organisation for Economic Co-operation and Development). 2019b. Measuring Distortions in International Markets: The Semiconductor Value Chain. OECD Trade Policy Papers, No. 234, OECD Publishing, Paris. OECD (Organisation for Economic Co-operation and Development). 2021. Agricultural Policy Monitoring and Evaluation 2021: Addressing the Challenges Facing Food Systems, OECD Publishing, Paris, https://doi.org/10.1787/2d810e01-en. OECD (Organisation for Economic Co-operation and Development) and IEA (International Energy Agency). 2021. “Update on Recent Progress in Reform of Inefficient Fossil-Fuel Subsidies that Encourage Wasteful Consumption.” Ossa, Ralph. 2015. “A Quantitative Analysis of Subsidy Competition in the US.” Working Paper 20975, National Bureau of Economic Research (NBER), Cambridge MA. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 64 Redonda, A., C. von Haldenwang, and F. Aliu. 2021. The Global Tax Expenditures Database (GTED). Companion Paper, Deutsches Institut für Entwicklungspolitik, Bonn, Germany. https://gted.net/2021/05/the-global-tax-expenditures-database- companion-paper/. Santos Silva, João M. C., and Silvana Tenreyro. 2006. “The Log of Gravity.” Review of Economics and Statistics 88(4): 641–58. Santos Silva, João M. C., and Silvana Tenreyro. 2011. “Further Simulation Evidence on the Performance of the Poisson Pseudo- Maximum Likelihood Estimator.” Economics Letters 112(2): 220–22. EQUITABLE GROWTH, FINANCE & INSTITUTIONS INSIGHT <<< 65