Exports and welfare: why agricultural liberalization does not maximize trade in Brazil’s bilateral strategies

Exportações e bem-estar: por que a liberalização agrícola não maximiza o comércio nas estratégias bilaterais do Brasil

Abstract

Abstract  This paper combines a computable general equilibrium model (PAEG), calibrated to GTAP 11 data for 2017, with a game-theoretic framework to examine how changes in domestic agricultural subsidies and import tariffs in the European Union, the United States, and China affect Brazilian export competitiveness and welfare. Five policy strategies are treated as strategic choices in three bilateral games, generating 75 scenarios. When exports are the payoff, Total Liberalization (TL) is strictly dominant and the Nash equilibrium is (TL, TL), with Brazilian export gains of USD 13.35 to 14.03 billion. When welfare is the payoff, measured as the equivalent variation of household consumption, Agricultural Liberalization (AL) becomes dominant and the equilibrium shifts to (AL, AL), with positive welfare for Brazil in all three games (USD 1.15 to 1.44 billion). Under (TL, TL), welfare turns negative because fiscal revenue losses and adverse price movements outweigh efficiency gains. A sensitivity analysis varying Armington elasticities by plus and minus 50% confirms the robustness of these equilibria. The results support negotiating packages that discipline partners’ subsidies while preserving part of Brazil’s tariff margin. The approach is designed for medium-term structural analysis of trade policy and does not capture short-run adjustment dynamics or distributional effects within Brazil.

Keywords:
agricultural subsidies; tariffs; export competitiveness; game theory

Resumo

Resumo  Este estudo combina um modelo de equilíbrio geral computável (PAEG), calibrado com dados do GTAP 11 para 2017, com teoria dos jogos para examinar como mudanças em subsídios agrícolas domésticos e tarifas de importação na União Europeia, nos Estados Unidos e na China afetam a competitividade exportadora e o bem-estar do Brasil. Cinco estratégias de política são tratadas como escolhas estratégicas em três jogos bilaterais, gerando 75 cenários. Quando as exportações são o payoff, a Liberalização Total (TL) é dominante e o equilíbrio de Nash é (TL, TL), com ganhos de exportação entre USD 13,35 e 14,03 bilhões. Quando o bem-estar é o payoff, medido pela variação equivalente do consumo, a Liberalização Agrícola (AL) torna-se dominante e o equilíbrio muda para (AL, AL), com bem-estar positivo nos três jogos (USD 1,15 a 1,44 bilhão). Sob (TL, TL), o bem-estar é negativo porque perdas de receita fiscal e movimentos de preços desfavoráveis superam os ganhos de eficiência. Uma análise de sensibilidade variando as elasticidades de Armington em mais e menos 50% confirma a robustez desses equilíbrios. Os resultados indicam que pacotes que disciplinem subsídios dos parceiros e preservem parte da margem tarifária brasileira são preferíveis à eliminação tarifária total. A abordagem é apropriada para análise estrutural de médio prazo da política comercial e não captura dinâmicas de ajustamento de curto prazo nem efeitos distributivos no Brasil.

Palavras-chave:
subsídios agrícolas; tarifas; competitividade exportadora; teoria dos jogos

1 Introduction

Brazil, as one of the world’s largest agricultural exporters, with USD 164.4 billion in exports in 2024, representing 49% of total Brazilian exports, operates under an international trading environment shaped by pervasive government intervention in competing economies (Brasil, 2025). Producer subsidies averaged USD 842 billion per year across 54 economies in 2021–2023, concentrated in three economies: China (37% of total positive support), the United States (15%), and the European Union (13%) (Organisation for Economic Co-operation and Development, 2024). Brazil, despite generating nearly 5% of the combined agricultural value of production of those economies, received less than 1% of total positive support in the same period, with a Producer Support Estimate of 3.3% of gross farm receipts against a 54-country average of 9% (Organisation for Economic Co-operation and Development, 2024; Viana Martins et al., 2024). This asymmetry compresses profit margins for Brazilian producers, distorts world price signals, and underpins the competitive environment within which Brazil negotiates at the World Trade Organization (WTO) Committee on Agriculture.

Brazil’s response to this environment has been active and litigious. In the DS267 cotton dispute, a WTO panel concluded that U.S. subsidies violated trade rules and distorted markets; in the sugar case involving India, the panel found export incentive schemes to be prohibited subsidies and recommended their withdrawal (World Trade Organization, 2014, 2021). Brazil currently presses for disclosure from the EU on Common Agricultural Policy payments, from the United States on Farm Bill programs, from China on corn and soybean subsidies, and from Canada, Japan, and India on their respective support regimes (World Trade Organization, 2024). These actions reflect a broader negotiating posture: Brazil uses multilateral norms to challenge the dominance of subsidized producers, but the payoff of different reform packages remains quantitatively uncertain.

The economic literature presents a broad body of work on the application of Computable General Equilibrium (CGE) models to Brazilian agricultural trade. Megiato et al. (2016) analyze bilateral trade between Brazil and the European Union using GTAP to identify the sectors that benefited the most from 2002 to 2012. Buchmann et al. (2021) examine bilateral agreements via GTAP to simulate their impacts on Brazil’s international trade and welfare. Martinez (2023) evaluates the impact of the Mercosur-EU agreement on GDP and welfare using GTAP. While these contributions demonstrate the consolidation of CGE modeling as the standard analytical tool for assessing trade policy scenarios involving Brazil, they generally treat policy choices as exogenous shocks rather than as outcomes of interaction among governments.

Studies that analytically integrate CGE models with game theory remain comparatively scarce. Sampaio & Sampaio (2009) use GTAP and game theory to analyze the effects of different trade policies on Brazil’s GDP across four trade agreement scenarios. Sanguinet et al. (2017) employ the same two frameworks to analyze the soybean market, focusing on four countries with payoffs based on exports. Taken together, these studies confirm that trade policy involves interdependent decisions among governments, yet none of them examines how the Nash equilibrium might shift when the payoff criterion changes from exports to welfare, nor do they extend the analysis to the full agricultural sector across Brazil’s three primary trading partners simultaneously. This gap has become consequential in the post-2020 context, characterized by heightened geopolitical tensions, the restructuring of global value chains, the ongoing Mercosur-EU negotiations, and the expansion of agricultural subsidies in China and the United States, factors that amplify the distortions motivating this study.

Against this backdrop, this study estimates the effects of changes in domestic agricultural subsidies and import tariffs in the European Union, the United States, and China on Brazilian export competitiveness, output, and welfare, integrating the PAEG model with a game-theoretic framework in three simultaneous bilateral games (Brazil-EU, Brazil-U.S., and Brazil-China) across 75 policy combinations. The PAEG model is calibrated to GTAP 11 data for 2017 and combined with a game structure in which five policy packages are treated as choices available to each player, and the analysis tests whether the Nash equilibrium shifts when the payoff criterion changes from exports to welfare, deriving the implications of this distinction for Brazil's negotiating posture at the WTO.

The bilateral game structure adopted here is an analytical delimitation that allows the payoff matrices to be read directly as strategic choices between Brazil and each major trading partner. Multilateral interactions, coalition formation, and cross-retaliation dynamics are not modeled, and are identified in Section 5 as natural extensions of the present framework.

Export volumes and household welfare can move in opposite directions because tariff elimination forgoes fiscal revenue and may worsen the terms of trade (Bagwell & Staiger, 1999, 2002; Ossa, 2014; Bekkers & Keck, 2024), yet no study has tested whether the equilibrium itself shifts when the payoff changes. The coexistence of two distinct equilibria, one pointing toward Total Liberalization when exports are the payoff and another toward Agricultural Liberalization when welfare is the payoff, offers a political economy explanation for why broad liberalization does not emerge in actual WTO negotiations: governments optimize household welfare under domestic political constraints, not export volumes.

The remainder of the article is organized as follows. Section 2 presents the theoretical framework covering the terms-of-trade rationale, the political economy of agricultural protection, and the game structure. Section 3 describes the PAEG model, the aggregation choices, the policy instruments, and the scenarios. Section 4 reports the results, identifies the Nash equilibria under alternative payoffs, and discusses the WTO implications. Section 5 concludes.

2 Theoretical Foundation

Trade policy choices in agriculture involve strategic interdependence: the subsidy or tariff that one government adopts alters prices, market shares, and real incomes in all trading partners. Bagwell & Staiger (1999, 2002) show that this terms-of-trade externality is the problem that trade agreements are designed to solve, because reciprocal liberalization commitments internalize the externality and allow all parties to reach outcomes that no country would choose unilaterally. Ossa (2014) provides quantitative support, estimating that terms-of-trade-driven tariff wars would raise average tariffs well above current applied rates and generate welfare losses for most countries. A direct implication is that export volumes and household welfare need not move together: when a country eliminates tariffs, trade expands but fiscal revenue falls and import prices may deteriorate, so welfare can decline even as exports rise. Bekkers & Keck (2024) confirm this in CGE simulations of multilateral tariff rules, showing that the regions with the largest export increases do not always obtain the largest welfare gains because terms-of-trade losses offset part of the efficiency improvements.

Agricultural protection persists because the political economy of trade policy systematically favors concentrated producer interests over diffuse consumer benefits. Olson (1965) argued that small, organized groups can secure policy rents at the expense of diffuse publics because the per-capita cost of protection is low for consumers but the per-capita benefit is high for producers. Grossman & Helpman (1994) formalized this logic in a model where lobbies offer political contributions in exchange for protection, generating tariffs and subsidies that persist even when they reduce aggregate welfare. In agriculture, these forces are reinforced by the political salience of food prices, rural employment, and land values. Swinnen (2021) documents how successive waves of agricultural policy reform in both developed and developing countries have been shaped by this dynamic: even when governments commit to reducing support, new instruments emerge and transfers remain high relative to what efficiency would prescribe. Anderson (2022) adds that uncertainty about reciprocity makes each government reluctant to liberalize unilaterally, because the welfare cost of opening depends on whether trading partners follow. The combination of concentrated producer interests and reciprocity uncertainty creates a setting where partial reform packages are more likely to emerge than full liberalization, which is precisely the pattern that the game-theoretic analysis in this paper formalizes.

The combination of CGE models with game-theoretic frameworks has antecedents in the trade-policy literature beyond the Brazilian studies cited in the introduction. Yilmaz (2006) uses a CGE model of the global cocoa market to derive Nash and Stackelberg optimum export taxes, showing that the welfare-maximizing tax rate depends on each country’s market share and supply elasticity, and that the choice between simultaneous and sequential game structures alters the equilibrium outcome. Beckman et al. (2023) employ a CGE model within a Nash-in-Nash bargaining framework to evaluate bilateral trade agreements among Trans-Pacific Partnership countries, concluding that welfare under a multilateral agreement exceeds the welfare achievable through any combination of bilateral deals, because bilateral negotiations tend to exclude sensitive agricultural sectors.

Cui et al. (2019) combine the GTAP model with a non-cooperative static game to analyze the trilateral FTA among China, Japan, and South Korea, proposing a compromise scheme with agricultural protection that reduces the resistance of Japan and South Korea to liberalization. Conforti & Salvatici (2004) simulate alternative liberalization scenarios in the Doha Round using a CGE framework and analyze the strategic interaction between developed and developing country negotiating groups. In common with these studies, the present analysis uses CGE-generated payoffs as inputs to a game-theoretic structure; the distinguishing feature is that the same set of 75 simulations is evaluated under two payoff criteria, which makes possible a direct test of whether the Nash equilibrium shifts when the objective changes from exports to welfare.

2.1 Game Structure and Equilibrium Concepts

Game theory analyzes decision-making problems involving multiple economic agents whose strategies influence the outcomes for all participants. Applications range from oligopolistic competition and auctions to monetary policy interactions between central banks and, relevant here, international trade negotiations in which countries compete or cooperate over tariffs and subsidies (Gibbons, 1992).

In a simultaneous game, all players choose their strategies at the same time without knowing what others will do. To define such a game, three elements must be specified: the players, who are all agents involved in the strategic interaction; the strategies, which are the actions available to each player; and the payoffs, which represent each player’s gain given the strategies adopted by all participants (Fiani, 2015).

To formalize, consider a simultaneous game with players, where each player has a finite set of available strategies. The structure adopted in this study consists of 2 players, each with two strategies, where $n,i\in \left\{\mathrm{1,2}\right\}$. Thus, player $n$receives payoff $a_{ij}$when player 1 adopts strategy $i$ and player 2 adopts strategy $j$. Table 1 represents this structure.

Table 1
Structure of a simultaneous game with two players, each with two strategies.

		Player 2
		$s_1$	$s_2$
Player 1	$s_1$	${\pi }_1^{11}$, ${\pi }_2^{11}$	${\pi }_1^{12}$, ${\pi }_2^{12}$
	$s_2$	${\pi }_1^{21}$, ${\pi }_2^{21}$	${\pi }_1^{22}$, ${\pi }_2^{22}$

• Source: Authors’ elaboration.

This study analyzes three bilateral games in simultaneous format: Brazil versus the USA, Brazil versus China, and Brazil versus the EU. Player payoffs depend on the full strategy profile and are measured through exports, real GDP, and welfare, where welfare is the equivalent variation, defined as the change in household purchasing power required to make consumers indifferent between the initial and the counterfactual equilibrium.

Nash equilibria are identified under two alternative payoff criteria: exports and welfare. A strategy is strictly dominant for a player when it always yields a higher payoff than any alternative, regardless of what the other player does. When no strictly dominant strategy exists, the Nash equilibrium, the combination of strategies where each player’s choice is the best response to the other’s, provides the solution concept. An equilibrium in dominant strategies is always a Nash equilibrium, but the converse is not necessarily true (Fiani, 2015; Bierman & Fernandez, 1998). Comparing Nash equilibria under both payoff criteria reveals whether the strategy that maximizes trade volumes also maximizes welfare, and whether the equilibrium itself shifts when the payoff changes. An equilibrium is Pareto optimal if no other strategy profile could improve one player’s payoff without worsening the other’s; a Nash equilibrium satisfies best-response conditions but does not guarantee Pareto optimality (Fiani, 2015; Bierman & Fernandez, 1998).

The connection between the CGE model and the game-theoretic framework in this study is external: the PAEG model generates the payoff values for each strategy profile, and the Nash equilibrium analysis is then applied to those values outside the model’s structural equations, rather than being solved simultaneously within the CGE optimization. This approach is standard in the literature that combines CGE with game theory (Sanguinet et al., 2017; Beckman et al., 2023; Yilmaz, 2006) and allows the use of a fully specified general equilibrium model to compute payoffs while preserving the analytical tractability of the game structure.

3 Methodology

3.1 The Computable General Equilibrium Model

The analysis employs the Projeto de Análise de Equilíbrio Geral da Economia Brasileira (PAEG), a static multi-regional CGE model implemented in MPSGE within GAMS, calibrated to a social accounting matrix from the GTAP 11 Data Base for 2017 (Aguiar et al., 2022; Gurgel et al., 2010). The database is aggregated into 21 regions, including the five Brazilian macroregions, the main trading partners, and composite rest-of-world regions; 19 sectors, of which seven are primary agricultural activities; and four primary factors: capital, labor, land, and natural resources.

On the supply side, firms operate under perfect competition and constant returns to scale, with a nested CES-Leontief production structure. Aggregate intermediate inputs and value added enter a Leontief function at the top level; within the value-added nest, factors substitute according to CES functions with sector-specific elasticities. Land is allocated across agricultural activities through a CET function, natural resources are sector-specific, and intermediate demand follows the Armington assumption, distinguishing domestic and imported varieties by region of origin. Primary factors are mobile across sectors within each region but immobile across Brazilian macroregions and foreign regions, so adjustment to policy shocks occurs through changes in wages and rental rates rather than unemployment or excess capacity.

On the demand side, each region is represented by a single representative household that allocates factor income and net tax revenues to private consumption, government consumption, and savings through a CES expenditure system. The macroeconomic closure fixes the marginal propensity to save in each region and holds regional current account balances at benchmark levels. Taxes and subsidies enter explicitly as ad valorem wedges between border, producer, and consumer prices, covering production subsidies, import tariffs, export taxes, and consumption taxes, all of which can be shocked in the policy experiments.

The simulations are comparative static: each experiment compares the 2017 benchmark equilibrium with a counterfactual obtained under alternative subsidy and tariff configurations, with fixed factor endowments and technology. Results describe how the sectoral structure of production, trade, and factor remuneration adjusts once the economy reaches the new equilibrium (Nazareth et al., 2019).

Because the Armington elasticities that govern substitution between domestic and imported goods and among import origins are known to influence the magnitude and sometimes the direction of CGE trade-policy results (Hertel et al., 2007), a sensitivity analysis was conducted in which both the domestic-import elasticity (esubd) and the intra-import elasticity (esubm) were varied by plus and minus 50% relative to their GTAP 11 default values. The PAEG model was recompiled under each alternative configuration and all 75 scenarios were re-simulated. The full payoff matrices and Nash equilibrium identification under these configurations are reported in Appendix B.

3.2 Regional and Sectoral Aggregation

The regional aggregation includes Brazil’s five macroregions treated as separate economies, the four major subsidizing economies (EU, United States, China, and Japan), other countries relevant to agricultural trade (Canada, India, Argentina, and Mercosur partners), and aggregate rest-of-world regions, as shown in Box 1. Disaggregating Brazil into five macroregions is required by the model’s factor market structure: capital and labor are mobile across sectors within each macroregion but immobile across macroregions, so aggregating Brazil into a single region would overstate the economy’s adjustment capacity in response to policy shocks. Although the analysis focuses on national-level results, this regional disaggregation ensures that the aggregate outcomes correctly reflect the constraints on interregional factor reallocation.

Box 1. Regional aggregation

Region (R)	R code	Description
North-BR	NOR	Brazilian states: Acre, Amapá, Amazonas, Pará, Rondônia, Roraima, Tocantins
Northeast - BR	NDE	Brazilian states: Alagoas, Bahia, Ceará, Maranhão, Paraíba, Pernambuco, Piauí, Rio Grande do Norte, Sergipe
Central-West - BR	COE	Brazilian states: Distrito Federal, Goiás, Mato Grosso, Mato Grosso do Sul
Southeast - BR	SDE	Brazilian states: Espírito Santo, Minas Gerais, Rio de Janeiro, São Paulo
South - BR	SUL	Brazilian states: Paraná, Rio Grande do Sul, Santa Catarina
Rest of Mercosur	RMS	Argentina, Paraguay, Uruguay
USA	USA	United States of America
Canada	CAN	Canada
Mexico	MEX	Mexico
Rest of the Americas	ROA	Other American countries excluding Brazil, USA, Canada, Mexico and Mercosur partners
European Union	EUR	27 EU member states
Rest of Europe	REU	European countries outside the EU (e.g. United Kingdom, Switzerland, Norway and others)
Japan	JPN	Japan
Russia	RUS	Russia
China	CHN	China
India	IND	India
Australia and New Zealand	ANZ	Australia and New Zealand
Fast development Asia	ASI	Emerging East and Southeast Asian economies (e.g. South Korea, Indonesia, Malaysia, Thailand, Viet Nam, Singapore)
Africa	AFR	African countries
Middle East	MES	Middle Eastern countries
Rest of Asia	RAS	Remaining Asian economies not in CHN, IND, ASI, JPN or MES

Source: Authors’ elaboration based on GTAP 11 database and PAEG framework.

The 65 GTAP sectors are grouped into 19, with detail preserved where agricultural support is concentrated: seven primary agricultural activities (paddy rice, cereal grains, oilseeds, sugar cane, animal products, raw milk, and other agricultural products), one processed food sector, and the remaining sectors covering manufacturing, utilities, construction, trade, transport, and services, as listed in Box 2.

Table 2
Effects on exports for the Brazil vs. USA game, in USD billion (2017 prices).

	Effects on exports	U.S.	U.S.	U.S.	U.S.	U.S.
		SQ	DM	AL	SL	TL
Brazil	SQ	(0; 0)	(0.07; 0.35)	(0.07; 0.35)	(0.07; 0.35)	(0.14; 3.47)
Brazil	DM	(6.05; 0.06)	(6.12; 0.41)	(6.13; 0.41)	(6.12; 0.41)	(6.21; 3.55)
Brazil	AL	(6.14; 0.07)	(6.22; 0.42)	(6.22; 0.42)	(6.22; 0.42)	(6.31; 3.56)
Brazil	SL	(5.96; 0.04)	(6.04; 0.39)	(6.04; 0.39)	(6.03; 0.39)	(6.12; 3.53)
Brazil	TL	(13.16; 0.21)	(13.24; 0.61)	(13.24; 0.61)	(13.24; 0.61)	(13.35; 3.80)

• Source: Authors’ elaboration. Note: SQ: Status Quo (maintain subsidies and tariffs); DM: Doha Moderate (cut subsidies and tariffs by 50%); AL: Agricultural Liberalization (eliminate subsidies, cut tariffs by 50%); SL: Selective Liberalization (maintain subsidies, cut tariffs by 50%); TL: Total Liberalization (eliminate subsidies and tariffs). Nash equilibrium cells highlighted in bold.

Box 2. Sectors and factor aggregation

Sectors (S)	S code	Agricultural classification	Production factors (F)	F code
Paddy rice	PDR	Primary agriculture	Capital	CAP
Cereal grains	GRO	Primary agriculture	Labor	LAB
Oilseeds	OSD	Primary agriculture	Land	LND
Sugar cane	C_B	Primary agriculture	Natural resources	RES
Animal products nec	OAP	Primary agriculture
Raw milk	RMK	Primary agriculture
Other agricultural products	AGR	Primary agriculture
Food products	FOO	Processed food
Textiles	TEX	Nonagricultural
Wearing apparel and leather products	WAP	Nonagricultural
Wood products	LUM	Nonagricultural
Paper products and publishing	PPP	Nonagricultural
Chemical, rubber, and plastic products	CRP	Nonagricultural
Other manufacturing	MAN	Nonagricultural
Electricity, gas, and water	SIU	Nonagricultural
Construction	CNS	Nonagricultural
Trade	TRD	Nonagricultural
Transport	OTP	Nonagricultural
Services	SER	Nonagricultural

Source: Authors’ elaboration based on GTAP 11 database and PAEG framework.

This aggregation strategy concentrates analytical resolution on the activities and regions where agricultural support and tariff protection generate the strongest competitive effects on Brazilian trade, which is the focus of the policy experiments described in the remainder of this section.

3.3 Representation of Policy Instruments

Agricultural subsidies in the PAEG model follow the OECD Producer Support Estimate classification, inherited from the GTAP Data Base, where output subsidies, input subsidies, and Market Price Support are calibrated from OECD PSE data and incorporated as ad valorem wedges in production, intermediate input use, and market prices (Gurgel et al., 2010; Aguiar et al., 2022; Wang & Aguiar, 2023). Import tariffs are modeled as ad valorem taxes on bilateral trade flows, differentiated by source country and commodity, calibrated from bilateral tariff schedules in GTAP 11. Export subsidies appear in the 2017 benchmark only as residual wedges following the WTO Nairobi Ministerial Decision on Export Competition (World Trade Organization, 2015) and are held constant across all scenarios. The policy experiments therefore concentrate on domestic production subsidies and import tariffs, which the literature identifies as the main sources of distortion affecting Brazilian export competitiveness (Bouët & Laborde, 2010; Aguiar et al., 2022).

3.4 Policy Strategies and Scenario Design

Each player chooses from five policy strategies that reflect positions observed in WTO agricultural negotiations, defined along two dimensions: domestic production subsidies (maintain, reduce 50%, or eliminate) and import tariffs (maintain, reduce 50%, or eliminate). The 50% reduction approximates the Amber Box cuts proposed in the Doha Round modalities (World Trade Organization, 2008). Box 3 maps each strategy to its subsidy and tariff parameters.

Table 3
Effects on exports for the Brazil vs. China game, in USD billion (2017 prices).

	Effects on exports	China	China	China	China	China
		SQ	DM	AL	SL	TL
Brazil	SQ	(0; 0)	(0.25; 0.92)	(0.25; 0.79)	(0.25; 1.05)	(0.55; 2.11)
Brazil	DM	(6.05; -0.04)	(6.32; 0.88)	(6.32; 0.74)	(6.32; 1.01)	(6.63; 2.07)
Brazil	AL	(6.14; -0.18)	(6.42; 0.73)	(6.42; 0.59)	(6.42; 0.86)	(6.73; 1.90)
Brazil	SL	(5.96; 0.10)	(6.23; 1.03)	(6.23; 0.89)	(6.23; 1.16)	(6.54; 2.23)
Brazil	TL	(13.16; 0.19)	(13.45; 1.19)	(13.45; 1.05)	(13.45; 1.32)	(13.79; 2.47)

• Source: Authors’ elaboration. Note: SQ: Status Quo (maintain subsidies and tariffs); DM: Doha Moderate (cut subsidies and tariffs by 50%); AL: Agricultural Liberalization (eliminate subsidies, cut tariffs by 50%); SL: Selective Liberalization (maintain subsidies, cut tariffs by 50%); TL: Total Liberalization (eliminate subsidies and tariffs). Nash equilibrium cells highlighted in bold.

Box 3. Policy strategy definitions

Strategy	Domestic Subsidies	Import Tariffs	Abbreviation
Status Quo	Maintain	Maintain	SQ
Doha Moderate	Reduce 50%	Reduce 50%	DM
Agricultural Liberalization	Eliminate	Reduce 50%	AL
Selective Liberalization	Maintain	Reduce 50%	SL
Total Liberalization	Eliminate	Eliminate	TL

Source: Authors’ elaboration.

The Status Quo (SQ) maintains domestic subsidies and import tariffs at 2017 GTAP benchmark levels, serving as the baseline against which all other strategies are compared. The Doha Moderate strategy (DM) implements a 50% reduction in both instruments, a level that falls within the range of tariff and subsidy cuts discussed in the Doha Round revised modalities (World Trade Organization, 2008), which proposed tiered formulas with deeper reductions for higher tariffs and for countries with higher levels of trade-distorting support. Agricultural Liberalization (AL) eliminates domestic production subsidies entirely while reducing tariffs by 50%, reflecting the negotiating position historically advanced by the Cairns Group, which has pressed for the elimination of trade-distorting domestic support as a condition for market access concessions (Anderson et al., 2006; Hopewell, 2016). Selective Liberalization (SL) maintains subsidies at baseline levels while reducing tariffs by 50%, capturing the asymmetric outcome frequently observed in trade negotiations, where countries accept market access concessions while resisting disciplines on domestic support, a pattern documented in the post-Uruguay Round period (Anderson & Martin, 2005). Total Liberalization (TL) eliminates both subsidies and tariffs, representing the theoretical free-trade benchmark and providing an upper bound on the welfare gains available from comprehensive agricultural reform (Anderson et al., 2006). The 5×5 matrix generates 25 scenarios per bilateral game and 75 combinations across the three games.

4 Results and Discussion

The PAEG simulations report results in absolute values (USD billion at 2017 prices). Each payoff cell contains two values in parentheses: Brazil’s outcome and the trading partner’s outcome. Nash equilibrium cells are highlighted in bold. The section presents export payoff matrices first, followed by welfare payoff matrices, where welfare is the equivalent variation of household consumption, reflecting changes in real income from resource reallocation, price movements, and fiscal adjustments.

Across Tables 2 to 4, TL is strictly dominant for both players under the export payoff: Brazil’s export gain is highest under TL regardless of the partner’s strategy, and the same holds for each partner. The unique Nash equilibrium is therefore (TL, TL) in all three games. In the Brazil-U.S. game (Table 2), Brazil’s export gain under (TL, TL) is USD 13.35 billion, while U.S. gains reach USD 3.80 billion. In the Brazil-China game (Table 3), the corresponding values are USD 13.79 billion for Brazil and USD 2.47 billion for China. In the Brazil-EU game (Table 4), Brazil gains USD 14.03 billion and the EU gains USD 5.28 billion. In percentage terms, Brazil’s gains range from 1.68% to 1.76% of baseline exports, while partner gains range from 0.07% to 0.17%, implying asymmetric payoff ratios of roughly 10:1 to 25:1 across the three games.

Table 4
Effects on exports for the Brazil vs. EU, in USD billion (2017 prices).

	Effects on exports	European Union	European Union	European Union	European Union	European Union
		SQ	DM	AL	SL	TL
Brazil	SQ	(0; 0)	(0.29; 1.45)	(0.29; 1.22)	(0.29; 1.67)	(0.73; 4.59)
Brazil	DM	(6.05; -0.13)	(6.36; 1.33)	(6.36; 1.10)	(6.36; 1.55)	(6.82; 4.49)
Brazil	AL	(6.14; -0.28)	(6.45; 1.17)	(6.45; 0.94)	(6.45; 1.39)	(6.92; 4.31)
Brazil	SL	(5.96; 0.01)	(6.27; 1.48)	(6.27; 1.25)	(6.27; 1.71)	(6.73; 4.66)
Brazil	TL	(13.16; 0.11)	(13.51; 1.77)	(13.51; 1.54)	(13.51; 2.00)	(14.03; 5.28)

• Source: Authors’ elaboration. Note: SQ: Status Quo (maintain subsidies and tariffs); DM: Doha Moderate (cut subsidies and tariffs by 50%); AL: Agricultural Liberalization (eliminate subsidies, cut tariffs by 50%); SL: Selective Liberalization (maintain subsidies, cut tariffs by 50%); TL: Total Liberalization (eliminate subsidies and tariffs). Nash equilibrium cells highlighted in bold.

The asymmetry in proportional gains is consistent with simulation evidence that competitive suppliers capture larger relative benefits when distortions are reduced in heavily subsidized economies, even when aggregate world effects look modest in percentage terms (Anderson et al., 2006; Elbehri & Leetmaa, 2001). The PAEG numbers quantify that logic in a bilateral payoff table: Brazil, as an efficient exporter operating under relatively low domestic support, gains disproportionately more from reform than the subsidizing partners, whose export performance depends in part on policy-induced cost advantages.

With exports as the payoff, TL is individually attractive for both players regardless of what the other does, so the game has the structure of a coordination problem rather than a Prisoner’s Dilemma. In a Prisoner’s Dilemma, mutual defection occurs because cooperation is individually costly even when collectively superior. Here, cooperation is individually attractive when measured in exports, which means both players would willingly move to (TL, TL) if export gains were the only consideration. The obstacle to reaching (TL, TL) lies not in the export calculus but in the welfare calculus: governments optimize household real income under domestic political constraints, and welfare does not move in the same direction as export volumes (Bagwell & Staiger, 1999, 2002). Eliminating subsidies and tariffs in a single negotiating move creates concentrated losses for organized producer groups and diffuse gains for consumers and taxpayers, a configuration that systematically slows reform regardless of the aggregate gains on offer (Olson, 1965). This pattern is also consistent with quantitative assessments showing that packages focused on domestic support alone deliver smaller price and welfare effects than packages that also cut agricultural tariffs, because border protection continues to constrain market access and export incentives even after domestic subsidies are removed (Diao et al., 2001; Hoekman et al., 2004).

An additional feature appears in the China and EU games: when Brazil liberalizes unilaterally while the partner remains at the status quo, Chinese and European exports fall, whereas the same pattern does not occur in the U.S. game. The magnitudes are modest in absolute terms, but the direction is informative. Brazil’s unilateral opening increases competitive pressure in sectors where these partners rely on support-induced cost advantages, which reduces their export performance even without any policy change on their side. This is consistent with simulation evidence that partner-country subsidies affect Brazilian export performance not only in bilateral trade but also in third markets, as Costa et al. (2012) demonstrate for orange juice exports. In game-theoretic terms, this feature also helps explain why countries that anticipate losing export share may resist reforms even when aggregate world welfare gains from liberalization are positive (Khurana, 2022; Mughwai, 2020).

The export game provides a useful benchmark by showing what countries leave on the table when they maintain current policy distortions. Compared with (TL, TL) as a free-trade reference, the status quo implies forgone export gains on the order of USD 13 to 14 billion for Brazil, depending on the partner, converting a qualitative claim about the cost of protection into a metric that can be weighed against adjustment costs and negotiation constraints. The export game does not, however, explain why (TL, TL) fails to emerge in actual negotiations. That answer lies in the welfare payoffs.

Tariffs are not only trade barriers; they also generate fiscal revenue and influence import prices, so that when Brazil eliminates tariffs under TL it forgoes that revenue and may face less advantageous price movements relative to trading partners. Real GDP can rise as production becomes more efficient and trade expands, yet welfare can fall if revenue losses and adverse price movements outweigh the efficiency gains from trade expansion (Bagwell & Staiger, 1999, 2002; Ossa, 2014). Bekkers & Keck (2024) confirm this pattern in CGE simulations of multilateral tariff liberalization: export gains are not a reliable indicator of welfare gains when terms-of-trade effects are large. This divergence between trade volumes and welfare is precisely why export payoffs alone do not capture what governments optimize when they set trade policy. Tables 5 to 7 report the welfare matrices for the three bilateral games.

Table 5
Effects on welfare for the Brazil vs. USA game, in USD billion (2017 prices)

	Effects on welfare	U.S.	U.S.	U.S.	U.S.	U.S.
		SQ	DM	AL	SL	TL
Brazil	SQ	(0; 0)	(0.10; 0.01)	(0.10; 0.02)	(0.10; 0.01)	(0.27; -2.77)
Brazil	DM	(0.61; 0.06)	(0.71; 0.08)	(0.71; 0.08)	(0.71; 0.07)	(0.88; -2.71)
Brazil	AL	(1.05; 0.08)	(1.15; 0.09)	(1.15; 0.09)	(1.15; 0.08)	(1.32; -2.69)
Brazil	SL	(0.10; 0.05)	(0.20; 0.06)	(0.20; 0.07)	(0.20; 0.06)	(0.37; -2.72)
Brazil	TL	(-0.92; 0.24)	(-0.84; 0.26)	(-0.84; 0.26)	(-0.84; 0.25)	(-0.68; -2.53)

• Source: Authors’ elaboration. Note: SQ: Status Quo (maintain subsidies and tariffs); DM: Doha Moderate (cut subsidies and tariffs by 50%); AL: Agricultural Liberalization (eliminate subsidies, cut tariffs by 50%); SL: Selective Liberalization (maintain subsidies, cut tariffs by 50%); TL: Total Liberalization (eliminate subsidies and tariffs). Nash equilibrium cells highlighted in bold.

Table 6
Effects on welfare for the Brazil vs. China game, in USD billion (2017 prices)

	Effects on welfare	China	China	China	China	China
		SQ	DM	AL	SL	TL
Brazil	SQ	(0; 0)	(0.28; 0.11)	(0.28; 0.27)	(0.28; -0.06)	(0.61; 0.11)
Brazil	DM	(0.61; -0.16)	(0.90; -0.05)	(0.90; 0.12)	(0.90; -0.22)	(1.23; -0.05)
Brazil	AL	(1.05; -0.52)	(1.34; -0.40)	(1.34; -0.24)	(1.34; -0.57)	(1.68; -0.40)
Brazil	SL	(0.10; 0.19)	(0.38; 0.30)	(0.38; 0.46)	(0.38; 0.13)	(0.71; 0.29)
Brazil	TL	(-0.92; -0.11)	(-0.66; -0.01)	(-0.66; 0.16)	(-0.65; -0.17)	(-0.35; -0.02)

• Source: Authors’ elaboration. Note: SQ: Status Quo (maintain subsidies and tariffs); DM: Doha Moderate (cut subsidies and tariffs by 50%); AL: Agricultural Liberalization (eliminate subsidies, cut tariffs by 50%); SL: Selective Liberalization (maintain subsidies, cut tariffs by 50%); TL: Total Liberalization (eliminate subsidies and tariffs). Nash equilibrium cells highlighted in bold.

Table 7
Effects on welfare for the Brazil vs. EU, in USD billion (2017 prices)

	Effects on welfare	European Union	European Union	European Union	European Union	European Union
		SQ	DM	AL	SL	TL
Brazil	SQ	(0; 0)	(0.39; 0.22)	(0.39; 0.25)	(0.39; 0.20)	(0.99; 0.09)
Brazil	DM	(0.61; 0.02)	(1.00; 0.24)	(1.00; 0.27)	(1.00; 0.22)	(1.59; 0.11)
Brazil	AL	(1.05; -0.16)	(1.44; 0.06)	(1.44; 0.09)	(1.43; 0.04)	(2.03; -0.07)
Brazil	SL	(0.10; 0.19)	(0.49; 0.42)	(0.49; 0.45)	(0.48; 0.39)	(1.08; 0.28)
Brazil	TL	(-0.92; 0.47)	(-0.61; 0.72)	(-0.61; 0.74)	(-0.61; 0.69)	(-0.14; 0.55)

• Source: Authors’ elaboration. Note: SQ: Status Quo (maintain subsidies and tariffs); DM: Doha Moderate (cut subsidies and tariffs by 50%); AL: Agricultural Liberalization (eliminate subsidies, cut tariffs by 50%); SL: Selective Liberalization (maintain subsidies, cut tariffs by 50%); TL: Total Liberalization (eliminate subsidies and tariffs). Nash equilibrium cells highlighted in bold.

The welfare matrices reverse the equilibrium. Agricultural Liberalization (AL) is strictly dominant for both players in all three games because it eliminates domestic production subsidies while reducing tariffs by only 50%, preserving part of Brazil’s tariff margin and avoiding the fiscal and terms-of-trade costs of full liberalization. The Nash equilibrium shifts from (TL, TL) to (AL, AL), where Brazil’s welfare gains are positive in all three games: USD 1.15 billion against the United States, USD 1.34 billion against China, and USD 1.44 billion against the European Union. Under (TL, TL), welfare is negative in all three games (USD -0.68, -0.35, and -0.14 billion), because full tariff elimination reduces real income through fiscal revenue losses and adverse price movements that outweigh efficiency gains.

The Brazil-EU game illustrates this divergence most clearly. Under (TL, TL), exports reach USD 14.03 billion and real GDP rises by USD 8.63 billion, yet welfare is negative at USD -0.14 billion. Under (AL, AL), exports and GDP are lower (USD 6.45 and 5.64 billion), but welfare turns positive at USD 1.44 billion. The result formalizes the terms-of-trade rationale of Bagwell & Staiger (1999, 2002): by retaining partial tariff protection, Brazil avoids the revenue losses and price deterioration that make (TL, TL) welfare-reducing, and the welfare difference between the two equilibria measures how much the fiscal and terms-of-trade channels matter relative to the pure efficiency effect.

The dominance of AL is consistent with the composition effects documented in Diao et al. (2001), Hoekman et al. (2004), and Buchmann et al. (2021): welfare gains depend on the structure of the reform package, not on trade volume alone. The PAEG results also show that DM outperforms SL from Brazil’s perspective, because tariff cuts that leave subsidized output intact limit the export response they would otherwise generate (Anderson et al., 2006).

The (AL, AL) equilibrium aligns with Brazil’s historical Cairns Group posture of disciplining partner subsidies while resisting full border liberalization (Barral, 2007; Hopewell, 2016), and is reinforced by the political economy logic in which organized producer groups resist concentrated losses even when aggregate gains are positive (Olson, 1965; Grossman & Helpman, 1994; Swinnen, 2021). Real GDP matrices in Appendix A follow the same dominance pattern as exports, with (TL, TL) delivering the largest GDP gains (USD 8.10 to 8.63 billion), and do not alter the equilibrium identified in the export game.

The coexistence of the two equilibria is the central finding. The export game establishes an upper bound of USD 13 to 14 billion in Brazilian gains; the welfare game reveals why countries stop short of that bound, settling at (AL, AL), where Brazil captures USD 6 to 6.5 billion in export gains while maintaining positive welfare. Because AL is dominant for both players, (AL, AL) is self-enforcing and therefore a plausible focal point for WTO-style bargaining, where reciprocity and verifiable commitment are institutional requirements (Conforti & Salvatici, 2004). The evidence from Viana Martins et al. (2024) reinforces the case for partner subsidy reform as the channel through which Brazil gains the most, while the welfare simulations show that preserving part of Brazil’s tariff margin is the mechanism that converts those export gains into net welfare improvements.

The sensitivity analysis reported in Appendix B confirms that the Nash equilibria identified above are robust to variations in the Armington elasticities. Under the high-elasticity configuration, (AL, AL) remains the welfare equilibrium in all three games, with Brazilian welfare gains of USD 1.34 to 1.61 billion. Under the low-elasticity configuration, AL remains strictly dominant for Brazil, though the partner’s dominant strategy shifts to Status Quo in the U.S. and EU games. The export equilibrium remains at (TL, TL) across all configurations (13).

5 Conclusions

Brazil’s choice between advocating full agricultural liberalization at the WTO and pursuing intermediate packages depends on what governments optimize when they negotiate. When the payoff criterion is exports, Total Liberalization is strictly dominant for both players in all three games, with Brazilian gains of USD 13.35 to 14.03 billion. When the payoff is welfare, measured as the equivalent variation of household consumption, Agricultural Liberalization becomes dominant and the equilibrium shifts to (AL, AL), with welfare gains of USD 1.15 to 1.44 billion. Under Total Liberalization, welfare turns negative because fiscal revenue losses and adverse price movements outweigh the efficiency gains from trade expansion. The two equilibria coexist because they answer different questions, and identifying which one a negotiator is optimizing changes the policy package that follows.

This explains why broad liberalization does not emerge in actual negotiations even when export volumes recommend it. Real-income calculations include fiscal revenue, price movements, and the political economy of reform, none of which is captured by an export payoff. Intermediate packages such as Brazil’s AL paired with a partner’s DM deliver export gains of USD 6.22 to 6.45 billion, a large share of the Total Liberalization benchmark, while generating positive welfare and avoiding the concentrated distributive costs of full tariff elimination. The pattern reflects the terms-of-trade rationale for trade agreements (Bagwell & Staiger, 1999, 2002) and the political economy logic in which concentrated losses on organized producer groups slow reform regardless of the aggregate gains on offer.

The results should be read within the bounds of what the framework can and cannot do. The simulations are comparative static and calibrated to 2017, so they describe medium-term structural responses rather than transition dynamics, adjustment costs, or the geopolitical and value-chain reconfigurations that have shaped agricultural trade since 2020. As discussed in Section 3.1, the coupling between the PAEG model and the game-theoretic structure is external rather than simultaneous, which precludes feedback between the players’ policy choices and general equilibrium adjustments within a single optimization, a feature that endogenous formulations of the game could incorporate in future work.

The bilateral structure abstracts from coalition formation and issue linkage in multilateral settings, where Brazil’s negotiating position might differ when multiple partners are at the table simultaneously. The PAEG allows disaggregation of the representative household into multiple income classes for Brazilian regions, but combining 10 household classes across 5 macroregions with 75 scenarios in 3 bilateral games would expand the dimensionality of the results beyond what the comparative analysis of Nash equilibria requires. The welfare measure therefore captures changes in aggregate real income without resolving distributional effects across income groups, and the extension to heterogeneous households remains open for future studies.

On the parametric side, the sensitivity analysis in Appendix B addresses uncertainty in the Armington elasticities but does not extend to factor substitution elasticities or the elasticity of transformation governing land reallocation across crops, which in Brazil are shaped by ongoing structural change in land use and may affect the sectoral adjustment paths. The political economy of reform resistance is invoked in the discussion to interpret the welfare equilibrium, but the weight of organized producer lobbies is not formally incorporated into the payoff structure, since the model treats governments as maximizing aggregate welfare or aggregate exports rather than as responding to pressure from interest groups. Incorporating lobby weights endogenously would require extending the game beyond what the current CGE-based payoff approach supports.

Even with these caveats, the exercise yields a clear contribution: within the structural relationships captured by the 2017 calibration, Brazilian welfare is higher under packages that discipline partner subsidies while retaining partial tariff protection than under full liberalization. The pattern holds across all three bilateral games and is robust to variations in the Armington elasticities. The result does not imply that full liberalization would always reduce welfare under any configuration of the global economy, but it does indicate that the welfare cost of full tariff elimination, through fiscal revenue losses and adverse price movements, is a quantitative factor that negotiators should weigh against the export gains that broad liberalization would deliver. For Brazilian trade policy at the WTO Committee on Agriculture, this translates into a negotiating posture that pushes for the disciplining of trade-distorting domestic support in partner countries while preserving enough tariff margin to convert any export gains into net welfare improvements.

Appendix A Real GDP payoff matrices

Table A1
Effects on real GDP for the Brazil vs. USA game, in USD billion (2017 prices)

	Effects on real GDP	U.S.	U.S.	U.S.	U.S.	U.S.
		SQ	DM	AL	SL	TL
Brazil	SQ	(0; 0)	(0.08; 0.21)	(0.08; 0.21)	(0.08; 0.21)	(0.18; 12.46)
Brazil	DM	(4.78; 0.03)	(4.87; 0.24)	(4.87; 0.24)	(4.87; 0.24)	(4.97; 12.49)
Brazil	AL	(5.29; -0.02)	(5.38; 0.20)	(5.38; 0.20)	(5.38; 0.20)	(5.49; 12.45)
Brazil	SL	(4.21; 0.07)	(4.29; 0.28)	(4.29; 0.28)	(4.29; 0.28)	(4.40; 12.53)
Brazil	TL	(7.92; 0.20)	(7.99; 0.44)	(7.99; 0.44)	(7.99; 0.44)	(8.10; 12.73)

• Source: Authors’ elaboration. Note: SQ: Status Quo (maintain subsidies and tariffs); DM: Doha Moderate (cut subsidies and tariffs by 50%); AL: Agricultural Liberalization (eliminate subsidies, cut tariffs by 50%); SL: Selective Liberalization (maintain subsidies, cut tariffs by 50%); TL: Total Liberalization (eliminate subsidies and tariffs). Nash equilibrium cells highlighted in bold.

Table A2
Effects on real GDP for the Brazil vs. China game, in USD billion (2017 prices)

	Effects on real GDP	China	China	China	China	China
		SQ	DM	AL	SL	TL
Brazil	SQ	(0; 0)	(0.24; 0.62)	(0.24; -0.17)	(0.24; 1.40)	(0.52; 1.55)
Brazil	DM	(4.78; -0.42)	(5.03; 0.20)	(5.03; -0.59)	(5.03; 0.97)	(5.32; 1.13)
Brazil	AL	(5.29; -1.28)	(5.55; -0.68)	(5.55; -1.46)	(5.55; 0.10)	(5.85; 0.24)
Brazil	SL	(4.21; 0.42)	(4.45; 1.05)	(4.45; 0.26)	(4.45; 1.83)	(4.73; 1.99)
Brazil	TL	(7.92; -0.37)	(8.16; 0.31)	(8.16; -0.48)	(8.16; 1.08)	(8.44; 1.31)

• Source: Authors’ elaboration. Note: SQ: Status Quo (maintain subsidies and tariffs); DM: Doha Moderate (cut subsidies and tariffs by 50%); AL: Agricultural Liberalization (eliminate subsidies, cut tariffs by 50%); SL: Selective Liberalization (maintain subsidies, cut tariffs by 50%); TL: Total Liberalization (eliminate subsidies and tariffs). Nash equilibrium cells highlighted in bold.

Table A3
Effects on real GDP for the Brazil vs. EU, in USD billion (2017 prices)

	Effects on real GDP	European Union	European Union	European Union	European Union	European Union
		SQ	DM	AL	SL	TL
Brazil	SQ	(0; 0)	(0.34; 0.74)	(0.34; 0.29)	(0.34; 1.18)	(0.86; 2.98)
Brazil	DM	(4.78; -0.09)	(5.13; 0.67)	(5.13; 0.21)	(5.12; 1.11)	(5.66; 2.92)
Brazil	AL	(5.29; -0.36)	(5.64; 0.38)	(5.64; -0.07)	(5.64; 0.83)	(6.18; 2.62)
Brazil	SL	(4.21; 0.18)	(4.55; 0.94)	(4.55; 0.49)	(4.55; 1.38)	(5.08; 3.20)
Brazil	TL	(7.92; 0.34)	(8.20; 1.24)	(8.20; 0.79)	(8.20; 1.68)	(8.63; 3.67)

• Source: Authors’ elaboration. Note: SQ: Status Quo (maintain subsidies and tariffs); DM: Doha Moderate (cut subsidies and tariffs by 50%); AL: Agricultural Liberalization (eliminate subsidies, cut tariffs by 50%); SL: Selective Liberalization (maintain subsidies, cut tariffs by 50%); TL: Total Liberalization (eliminate subsidies and tariffs). Nash equilibrium cells highlighted in bold.

Appendix B Sensitivity Analysis: Armington Elasticities

This appendix reports the payoff matrices under two alternative configurations of Armington elasticities: a low-elasticity configuration, in which the elasticities of substitution between domestic and imported goods (esubd) and among import origins (esubm) are reduced by 50% relative to the GTAP 11 default values, and a high-elasticity configuration, in which both parameters are increased by 50%. The PAEG model was recompiled under each configuration and all 75 scenarios were re-simulated. The procedure follows the standard practice in the GTAP literature for systematic sensitivity analysis (Hertel et al., 2007).

1 3 report the welfare payoff matrices under the low-elasticity configuration. 4 6 report the corresponding export payoff matrices. 7 12 report the welfare and export matrices under the high-elasticity configuration. 13 summarizes the Nash equilibria across all configurations.

Table B1
Effects on welfare for the Brazil vs. USA game under low Armington elasticities (−50%), in USD billion (2017 prices)

	Effects on welfare	U.S.	U.S.	U.S.	U.S.	U.S.
		SQ	DM	AL	SL	TL
Brazil	SQ	(0.00; 0.00)	(0.11; -0.03)	(0.12; -0.02)	(0.11; -0.04)	(0.32; -3.68)
Brazil	DM	(0.38; 0.09)	(0.49; 0.06)	(0.49; 0.07)	(0.49; 0.06)	(0.69; -3.59)
Brazil	AL	(0.93; 0.15)	(1.04; 0.12)	(1.04; 0.12)	(1.04; 0.11)	(1.24; -3.54)
Brazil	SL	(-0.22; 0.04)	(-0.11; 0.01)	(-0.11; 0.02)	(-0.11; 0.01)	(0.09; -3.64)
Brazil	TL	(-1.09; 0.26)	(-0.98; 0.22)	(-0.98; 0.23)	(-0.98; 0.22)	(-0.79; -3.43)

• Source: Authors’ elaboration. Note: SQ: Status Quo (maintain subsidies and tariffs); DM: Doha Moderate (cut subsidies and tariffs by 50%); AL: Agricultural Liberalization (eliminate subsidies, cut tariffs by 50%); SL: Selective Liberalization (maintain subsidies, cut tariffs by 50%); TL: Total Liberalization (eliminate subsidies and tariffs). Nash equilibrium cells highlighted in bold.

Table B2
Effects on welfare for the Brazil vs. China game under low Armington elasticities (−50%), in USD billion (2017 prices)

	Effects on welfare	China	China	China	China	China
		SQ	DM	AL	SL	TL
Brazil	SQ	(0.00; 0.00)	(0.33; -0.01)	(0.33; 0.13)	(0.34; -0.15)	(0.70; -0.09)
Brazil	DM	(0.38; -0.20)	(0.71; -0.21)	(0.71; -0.07)	(0.71; -0.35)	(1.08; -0.29)
Brazil	AL	(0.93; -0.60)	(1.27; -0.61)	(1.27; -0.47)	(1.27; -0.75)	(1.64; -0.69)
Brazil	SL	(-0.22; 0.19)	(0.11; 0.18)	(0.11; 0.32)	(0.11; 0.04)	(0.47; 0.10)
Brazil	TL	(-1.09; -0.24)	(-0.76; -0.26)	(-0.76; -0.12)	(-0.76; -0.40)	(-0.40; -0.34)

• Source: Authors’ elaboration. Note: SQ: Status Quo (maintain subsidies and tariffs); DM: Doha Moderate (cut subsidies and tariffs by 50%); AL: Agricultural Liberalization (eliminate subsidies, cut tariffs by 50%); SL: Selective Liberalization (maintain subsidies, cut tariffs by 50%); TL: Total Liberalization (eliminate subsidies and tariffs). Nash equilibrium cells highlighted in bold.

Table B3
Effects on welfare for the Brazil vs. EU game under low Armington elasticities (−50%), in USD billion (2017 prices)

	Effects on welfare	European Union	European Union	European Union	European Union	European Union
		SQ	DM	AL	SL	TL
Brazil	SQ	(0.00; 0.00)	(0.45; -0.09)	(0.45; -0.07)	(0.44; -0.11)	(1.02; -0.34)
Brazil	DM	(0.38; 0.01)	(0.82; -0.08)	(0.82; -0.06)	(0.82; -0.10)	(1.39; -0.33)
Brazil	AL	(0.93; -0.15)	(1.37; -0.24)	(1.37; -0.22)	(1.37; -0.26)	(1.94; -0.49)
Brazil	SL	(-0.22; 0.17)	(0.22; 0.08)	(0.22; 0.10)	(0.22; 0.06)	(0.79; -0.18)
Brazil	TL	(-1.09; 0.30)	(-0.68; 0.21)	(-0.68; 0.22)	(-0.68; 0.19)	(-0.16; -0.06)

• Source: Authors’ elaboration. Note: SQ: Status Quo (maintain subsidies and tariffs); DM: Doha Moderate (cut subsidies and tariffs by 50%); AL: Agricultural Liberalization (eliminate subsidies, cut tariffs by 50%); SL: Selective Liberalization (maintain subsidies, cut tariffs by 50%); TL: Total Liberalization (eliminate subsidies and tariffs). Nash equilibrium cells highlighted in bold.

Table B4
Effects on exports for the Brazil vs. USA game under low Armington elasticities (−50%), in USD billion (2017 prices)

	Effects on exports	U.S.	U.S.	U.S.	U.S.	U.S.
		SQ	DM	AL	SL	TL
Brazil	SQ	(0.00; 0.00)	(-0.01; 0.20)	(-0.01; 0.20)	(-0.01; 0.20)	(-0.03; 4.53)
Brazil	DM	(3.64; 0.00)	(3.63; 0.20)	(3.63; 0.20)	(3.63; 0.20)	(3.61; 4.53)
Brazil	AL	(4.04; -0.01)	(4.04; 0.19)	(4.04; 0.19)	(4.04; 0.19)	(4.01; 4.52)
Brazil	SL	(3.25; 0.00)	(3.25; 0.21)	(3.25; 0.21)	(3.25; 0.21)	(3.22; 4.54)
Brazil	TL	(7.13; 0.02)	(7.12; 0.23)	(7.12; 0.23)	(7.12; 0.23)	(7.11; 4.58)

• Source: Authors’ elaboration. Note: SQ: Status Quo (maintain subsidies and tariffs); DM: Doha Moderate (cut subsidies and tariffs by 50%); AL: Agricultural Liberalization (eliminate subsidies, cut tariffs by 50%); SL: Selective Liberalization (maintain subsidies, cut tariffs by 50%); TL: Total Liberalization (eliminate subsidies and tariffs). Nash equilibrium cells highlighted in bold.

Table B5
Effects on exports for the Brazil vs. China game under low Armington elasticities (−50%), in USD billion (2017 prices)

	Effects on exports	China	China	China	China	China
		SQ	DM	AL	SL	TL
Brazil	SQ	(0.00; 0.00)	(0.04; 0.58)	(0.04; 0.41)	(0.04; 0.75)	(0.08; 1.29)
Brazil	DM	(3.64; -0.02)	(3.69; 0.56)	(3.69; 0.38)	(3.69; 0.73)	(3.73; 1.27)
Brazil	AL	(4.04; -0.07)	(4.09; 0.50)	(4.09; 0.33)	(4.09; 0.68)	(4.14; 1.21)
Brazil	SL	(3.25; 0.03)	(3.30; 0.61)	(3.30; 0.44)	(3.30; 0.78)	(3.34; 1.32)
Brazil	TL	(7.13; 0.07)	(7.18; 0.66)	(7.18; 0.49)	(7.18; 0.84)	(7.23; 1.39)

• Source: Authors’ elaboration. Note: SQ: Status Quo (maintain subsidies and tariffs); DM: Doha Moderate (cut subsidies and tariffs by 50%); AL: Agricultural Liberalization (eliminate subsidies, cut tariffs by 50%); SL: Selective Liberalization (maintain subsidies, cut tariffs by 50%); TL: Total Liberalization (eliminate subsidies and tariffs). Nash equilibrium cells highlighted in bold.

Table B6
Effects on exports for the Brazil vs. EU game under low Armington elasticities (−50%), in USD billion (2017 prices)

	Effects on exports	European Union	European Union	European Union	European Union	European Union
		SQ	DM	AL	SL	TL
Brazil	SQ	(0.00; 0.00)	(-0.02; 0.92)	(-0.02; 0.65)	(-0.02; 1.19)	(-0.04; 2.92)
Brazil	DM	(3.64; -0.10)	(3.63; 0.82)	(3.63; 0.55)	(3.63; 1.09)	(3.61; 2.82)
Brazil	AL	(4.04; -0.17)	(4.03; 0.75)	(4.03; 0.48)	(4.03; 1.02)	(4.01; 2.75)
Brazil	SL	(3.25; -0.04)	(3.24; 0.89)	(3.24; 0.62)	(3.24; 1.16)	(3.22; 2.90)
Brazil	TL	(7.13; 0.02)	(7.12; 1.00)	(7.11; 0.73)	(7.12; 1.27)	(7.09; 3.07)

• Source: Authors’ elaboration. Note: SQ: Status Quo (maintain subsidies and tariffs); DM: Doha Moderate (cut subsidies and tariffs by 50%); AL: Agricultural Liberalization (eliminate subsidies, cut tariffs by 50%); SL: Selective Liberalization (maintain subsidies, cut tariffs by 50%); TL: Total Liberalization (eliminate subsidies and tariffs). Nash equilibrium cells highlighted in bold.

Table B7
Effects on welfare for the Brazil vs. USA game under high Armington elasticities (+50%), in USD billion (2017 prices)

	Effects on welfare	U.S.	U.S.	U.S.	U.S.	U.S.
		SQ	DM	AL	SL	TL
Brazil	SQ	(0.00; 0.00)	(0.09; 0.05)	(0.09; 0.05)	(0.09; 0.05)	(0.28; -2.72)
Brazil	DM	(0.87; 0.06)	(0.95; 0.11)	(0.95; 0.11)	(0.95; 0.11)	(1.14; -2.67)
Brazil	AL	(1.25; 0.05)	(1.34; 0.10)	(1.34; 0.11)	(1.34; 0.10)	(1.53; -2.67)
Brazil	SL	(0.39; 0.06)	(0.48; 0.12)	(0.48; 0.12)	(0.48; 0.11)	(0.67; -2.67)
Brazil	TL	(-0.80; 0.27)	(-0.75; 0.33)	(-0.75; 0.33)	(-0.75; 0.32)	(-0.58; -2.45)

• Source: Authors’ elaboration. Note: SQ: Status Quo (maintain subsidies and tariffs); DM: Doha Moderate (cut subsidies and tariffs by 50%); AL: Agricultural Liberalization (eliminate subsidies, cut tariffs by 50%); SL: Selective Liberalization (maintain subsidies, cut tariffs by 50%); TL: Total Liberalization (eliminate subsidies and tariffs). Nash equilibrium cells highlighted in bold.

Table B8
Effects on welfare for the Brazil vs. China game under high Armington elasticities (+50%), in USD billion (2017 prices)

	Effects on welfare	China	China	China	China	China
		SQ	DM	AL	SL	TL
Brazil	SQ	(0.00; 0.00)	(0.26; 0.19)	(0.26; 0.37)	(0.26; 0.00)	(0.59; 0.21)
Brazil	DM	(0.87; -0.14)	(1.13; 0.05)	(1.13; 0.23)	(1.13; -0.14)	(1.46; 0.07)
Brazil	AL	(1.25; -0.48)	(1.52; -0.29)	(1.52; -0.11)	(1.52; -0.48)	(1.86; -0.26)
Brazil	SL	(0.39; 0.19)	(0.65; 0.38)	(0.65; 0.56)	(0.65; 0.19)	(0.97; 0.40)
Brazil	TL	(-0.80; -0.02)	(-0.57; 0.18)	(-0.58; 0.36)	(-0.57; -0.01)	(-0.30; 0.19)

• Source: Authors’ elaboration. Note: SQ: Status Quo (maintain subsidies and tariffs); DM: Doha Moderate (cut subsidies and tariffs by 50%); AL: Agricultural Liberalization (eliminate subsidies, cut tariffs by 50%); SL: Selective Liberalization (maintain subsidies, cut tariffs by 50%); TL: Total Liberalization (eliminate subsidies and tariffs). Nash equilibrium cells highlighted in bold.

Table B9
Effects on welfare for the Brazil vs. EU game under high Armington elasticities (+50%), in USD billion (2017 prices)

	Effects on welfare	European Union	European Union	European Union	European Union	European Union
		SQ	DM	AL	SL	TL
Brazil	SQ	(0.00; 0.00)	(0.36; 0.54)	(0.37; 0.58)	(0.36; 0.50)	(1.04; 0.51)
Brazil	DM	(0.87; 0.03)	(1.22; 0.58)	(1.23; 0.62)	(1.22; 0.54)	(1.89; 0.55)
Brazil	AL	(1.25; -0.17)	(1.61; 0.37)	(1.61; 0.41)	(1.61; 0.33)	(2.28; 0.35)
Brazil	SL	(0.39; 0.23)	(0.75; 0.78)	(0.75; 0.82)	(0.75; 0.74)	(1.41; 0.75)
Brazil	TL	(-0.80; 0.65)	(-0.59; 1.31)	(-0.59; 1.35)	(-0.59; 1.27)	(-0.19; 1.24)

• Source: Authors’ elaboration. Note: SQ: Status Quo (maintain subsidies and tariffs); DM: Doha Moderate (cut subsidies and tariffs by 50%); AL: Agricultural Liberalization (eliminate subsidies, cut tariffs by 50%); SL: Selective Liberalization (maintain subsidies, cut tariffs by 50%); TL: Total Liberalization (eliminate subsidies and tariffs). Nash equilibrium cells highlighted in bold.

Table B10
Effects on exports for the Brazil vs. USA game under high Armington elasticities (+50%), in USD billion (2017 prices)

	Effects on exports	U.S.	U.S.	U.S.	U.S.	U.S.
		SQ	DM	AL	SL	TL
Brazil	SQ	(0.00; 0.00)	(0.15; 0.52)	(0.15; 0.52)	(0.15; 0.52)	(0.31; 3.27)
Brazil	DM	(8.52; 0.11)	(8.68; 0.64)	(8.69; 0.63)	(8.68; 0.64)	(8.90; 3.43)
Brazil	AL	(8.34; 0.13)	(8.50; 0.66)	(8.50; 0.66)	(8.50; 0.66)	(8.72; 3.44)
Brazil	SL	(8.70; 0.08)	(8.86; 0.62)	(8.87; 0.62)	(8.86; 0.62)	(9.08; 3.41)
Brazil	TL	(19.65; 0.39)	(19.85; 1.04)	(19.85; 1.04)	(19.85; 1.04)	(20.11; 3.97)

• Source: Authors’ elaboration. Note: SQ: Status Quo (maintain subsidies and tariffs); DM: Doha Moderate (cut subsidies and tariffs by 50%); AL: Agricultural Liberalization (eliminate subsidies, cut tariffs by 50%); SL: Selective Liberalization (maintain subsidies, cut tariffs by 50%); TL: Total Liberalization (eliminate subsidies and tariffs). Nash equilibrium cells highlighted in bold.

Table B11
Effects on exports for the Brazil vs. China game under high Armington elasticities (+50%), in USD billion (2017 prices)

	Effects on exports	China	China	China	China	China
		SQ	DM	AL	SL	TL
Brazil	SQ	(0.00; 0.00)	(0.44; 1.28)	(0.44; 1.18)	(0.44; 1.38)	(0.99; 3.06)
Brazil	DM	(8.52; -0.04)	(8.99; 1.25)	(8.99; 1.15)	(8.99; 1.35)	(9.58; 3.03)
Brazil	AL	(8.34; -0.23)	(8.81; 1.03)	(8.81; 0.93)	(8.81; 1.13)	(9.40; 2.79)
Brazil	SL	(8.70; 0.15)	(9.17; 1.45)	(9.17; 1.35)	(9.17; 1.55)	(9.75; 3.27)
Brazil	TL	(19.65; 0.36)	(20.18; 1.83)	(20.18; 1.73)	(20.18; 1.93)	(20.86; 3.87)

• Source: Authors’ elaboration. Note: SQ: Status Quo (maintain subsidies and tariffs); DM: Doha Moderate (cut subsidies and tariffs by 50%); AL: Agricultural Liberalization (eliminate subsidies, cut tariffs by 50%); SL: Selective Liberalization (maintain subsidies, cut tariffs by 50%); TL: Total Liberalization (eliminate subsidies and tariffs). Nash equilibrium cells highlighted in bold.

Table B12
Effects on exports for the Brazil vs. EU game under high Armington elasticities (+50%), in USD billion (2017 prices)

	Effects on exports	European Union	European Union	European Union	European Union	European Union
		SQ	DM	AL	SL	TL
Brazil	SQ	(0.00; 0.00)	(0.63; 2.05)	(0.64; 1.87)	(0.63; 2.24)	(1.78; 6.94)
Brazil	DM	(8.52; -0.15)	(9.20; 1.94)	(9.20; 1.75)	(9.19; 2.12)	(10.42; 6.89)
Brazil	AL	(8.34; -0.36)	(9.01; 1.71)	(9.02; 1.52)	(9.01; 1.89)	(10.23; 6.62)
Brazil	SL	(8.70; 0.06)	(9.38; 2.16)	(9.38; 1.98)	(9.38; 2.35)	(10.60; 7.15)
Brazil	TL	(19.65; 0.22)	(20.47; 2.77)	(20.48; 2.59)	(20.47; 2.96)	(21.97; 8.64)

• Source: Authors’ elaboration. Note: SQ: Status Quo (maintain subsidies and tariffs); DM: Doha Moderate (cut subsidies and tariffs by 50%); AL: Agricultural Liberalization (eliminate subsidies, cut tariffs by 50%); SL: Selective Liberalization (maintain subsidies, cut tariffs by 50%); TL: Total Liberalization (eliminate subsidies and tariffs). Nash equilibrium cells highlighted in bold.

Table B13
Sensitivity of Nash equilibria to Armington elasticities

Configuration	Game	Nash (exports)	Nash (welfare)	EV Brazil	EV Partner	Exp. Brazil
Base (GTAP)	Brazil-U.S.	(TL, TL)	(AL, AL)	1.15	0.09	13.35
	Brazil-China	(TL, TL)	(AL, AL)	1.34	-0.24	13.79
	Brazil-EU	(TL, TL)	(AL, AL)	1.44	0.09	14.03
Low (-50%)	Brazil-U.S.	(TL, TL)	(AL, SQ)	0.93	0.15	7.11
	Brazil-China	(TL, TL)	(AL, AL)	1.27	-0.47	7.23
	Brazil-EU	(TL, TL)	(AL, SQ)	0.93	-0.15	7.09
High (+50%)	Brazil-U.S.	(TL, TL)	(AL, AL)	1.34	0.11	20.11
	Brazil-China	(TL, TL)	(AL, AL)	1.52	-0.11	20.86
	Brazil-EU	(TL, TL)	(AL, AL)	1.61	0.41	21.97

• Source: Authors’ elaboration. Note: EV and exports in USD billion at Nash equilibrium of the welfare game (columns 5 and 6) and export game (column 7).

Under the low-elasticity configuration, Agricultural Liberalization (AL) remains strictly dominant for Brazil in all three welfare games, confirming that the preference for packages that discipline partner subsidies while preserving part of the tariff margin does not depend on the calibration of trade elasticities. In the Brazil-China game, the Nash equilibrium remains at (AL, AL), with Brazilian welfare of USD 1.27 billion. In the Brazil-U.S. and Brazil-EU games, the partner’s dominant strategy shifts from AL to Status Quo (SQ), so that the welfare Nash equilibrium becomes (AL, SQ), with Brazilian welfare of USD 0.93 billion in both cases. This shift reflects the fact that, when trade flows respond less to policy changes, the benefits of subsidy reform for the subsidizing partner are attenuated, reducing the incentive to liberalize. Under the high-elasticity configuration, (AL, AL) is restored as the welfare Nash equilibrium in all three games, with Brazilian welfare gains ranging from USD 1.34 to 1.61 billion, above the baseline values. The export Nash equilibrium remains at (TL, TL) across all configurations, with Brazilian export gains ranging from USD 7.09 to 7.23 billion under low elasticities and from USD 20.11 to 21.97 billion under high elasticities, compared with USD 13.35 to 14.03 billion in the baseline. The variation in magnitudes is expected, as higher trade elasticities amplify the trade-volume response to tariff removal, while lower elasticities compress the response. The finding that the payoff criterion determines the equilibrium and that AL dominates when welfare is the objective is robust to these parameter variations.

Data availability:

Research data is not available.

• How to cite:
  Martins, M. M. V., & Cruz, A. A. (2026). Exports and welfare: why agricultural liberalization does not maximize trade in Brazil’s bilateral strategies. Revista de Economia e Sociologia Rural, 64, e305516. https://doi.org/10.1590/1806-9479.2026.305516
• Financial support:
  Nothing to declare.
• Ethics approval:
  Not applicable.
• JEL Classification:
  C68; F13; Q18.

6 References

• Aguiar, A., Chepeliev, M., Corong, E., & van der Mensbrugghe, D. (2022). The Global Trade Analysis Project (GTAP) Data Base: Version 11. Journal of Global Economic Analysis, 7(2), 1-37. https://doi.org/10.21642/JGEA.070201AF
  » https://doi.org/10.21642/JGEA.070201AF
• Anderson, K. (2022). Agriculture in a more uncertain global trade environment. Agricultural Economics, 53(4), 563-579. https://doi.org/10.1111/agec.12726
  » https://doi.org/10.1111/agec.12726
• Anderson, K., & Martin, W. (2005). Agricultural Trade Reform and the Doha Development Agenda. World Economy, 28(9), 1301-1327. https://doi.org/10.1111/j.1467-9701.2005.00735.x
  » https://doi.org/10.1111/j.1467-9701.2005.00735.x
• Anderson, K., Martin, W., & van der Mensbrugghe, D. (2006). Doha merchandise trade reform: What’s at stake for developing countries? The World Bank Economic Review, 20(2), 169-195. https://doi.org/10.1093/wber/lhj009
  » https://doi.org/10.1093/wber/lhj009
• Bagwell, K., & Staiger, R. W. (1999). An economic theory of GATT. The American Economic Review, 89(1), 215-248. https://doi.org/10.1257/aer.89.1.215
  » https://doi.org/10.1257/aer.89.1.215
• Bagwell, K., & Staiger, R. W. (2002). The economics of the world trading system. Cambridge: MIT Press. https://doi.org/10.7551/mitpress/2628.001.0001
  » https://doi.org/10.7551/mitpress/2628.001.0001
• Barral, W. (2007). The Brazilian experience in dispute settlement (Project Document, No. LC/W.147). United Nations Economic Commission for Latin America and the Caribbean.
• Beckman, J., Ivanic, M., & Shaik, S. (2023). How bilateral trade deals get in the way of multilateral agreements: Why WTO is marginalized. Journal of Policy Modeling, 45(5), 877-894. https://doi.org/10.1016/j.jpolmod.2023.05.006
  » https://doi.org/10.1016/j.jpolmod.2023.05.006
• Bekkers, E., & Keck, A. (2024). Tariff spillovers and new rules for multilateral tariff negotiations (WTO Staff Working Paper, No. ERSD-2024-01). Geneva: WTO.
• Bierman, H. S., & Fernandez, L. (1998). Game theory with economic applications (2nd ed.). Reading: Addison-Wesley.
• Bouët, A., & Laborde, D. (2010). Assessing the potential cost of a failed Doha Round. World Trade Review, 9(2), 319-351. https://doi.org/10.1017/S1474745609990267
  » https://doi.org/10.1017/S1474745609990267
• Brasil. Ministério da Agricultura e Pecuária – MAPA. (2025). Ações estratégicas 2025-2026. Brasília.
• Buchmann, J. L., Massuquetti, A., & Azevedo, A. F. Z. (2021). Análise de cenários do agronegócio brasileiro frente à China, aos EUA e à UE, utilizando um modelo de equilíbrio geral computável. Revista de Economia e Sociologia Rural, 59(4), e221493. https://doi.org/10.1590/1806-9479.2021.221493
  » https://doi.org/10.1590/1806-9479.2021.221493
• Conforti, P., & Salvatici, L. (2004). Agricultural trade liberalization in the Doha round: alternative scenarios and strategic interactions between developed and developing countries (Commodity and Trade Policy Research Working Paper, No. 10). Rome: FAO.
• Costa, C. K. F., Maia, S. F., & Sampaio, L. M. B. (2012). Exportações brasileiras de suco de laranja e subsídios americanos: uma análise empírica de estratégias comerciais (1991-2006). Revista de Economia e Sociologia Rural, 50(1), 83-106. https://doi.org/10.1590/S0103-20032012000100005
  » https://doi.org/10.1590/S0103-20032012000100005
• Cui, L., Song, M., & Zhu, L. (2019). Economic evaluation of the trilateral FTA among China, Japan, and South Korea with big data analytics. Computers & Industrial Engineering, 128, 1040-1051. https://doi.org/10.1016/j.cie.2018.04.029
  » https://doi.org/10.1016/j.cie.2018.04.029
• Diao, X., Elbehri, A., Gehlhar, M. J., Gibson, P. R., Leetmaa, S. E., Mitchell, L., Nelson, F. J., Nimon, R. W., Normile, M. A., Roe, T. L., Shapouri, S., Skully, D. W., Smith, M., Somwaru, A., Trueblood, M. A., Tsigas, M., Wainio, J., Whitley, D. B., & Young, C. E. (2001). Agricultural policy reform in the WTO: the road ahead (Agricultural Economic Report, No. 802). U.S. Department of Agriculture, Economic Research Service.
• Elbehri, A., & Leetmaa, S. E. (2001). How significant are export subsidies to agricultural trade? Trade and welfare implications of global reforms American Agricultural Economics Association.
• Fiani, R. (2015). Teoria dos jogos (4. ed.). Rio de Janeiro: Elsevier/GEN Atlas.
• Gibbons, R. (1992). Game theory for applied economists Princeton: Princeton University Press.
• Grossman, G. M., & Helpman, E. (1994). Protection for sale. The American Economic Review, 84(4), 833-850.
• Gurgel, A. C., Pereira, M. W. G., & Teixeira, E. C. (2010). A estrutura do PAEG (PAEG Technical Paper, No. 1). Viçosa: UFV.
• Hertel, T. W., Hummels, D., Ivanic, M., & Keeney, R. (2007). How confident can we be of CGE-based assessments of Free Trade Agreements? Economic Modelling, 24(4), 611-635. https://doi.org/10.1016/j.econmod.2006.12.002
  » https://doi.org/10.1016/j.econmod.2006.12.002
• Hoekman, B. M., Ng, F., & Olarreaga, M. (2004). Agricultural tariffs versus subsidies: What’s more important for developing countries? The World Bank Economic Review, 18(2), 175-204. https://doi.org/10.1093/wber/lhh037
  » https://doi.org/10.1093/wber/lhh037
• Hopewell, K. (2016). Breaking the WTO: how emerging powers disrupted the neoliberal project. Stanford University Press.
• Khurana, C. (2022). Review of game theory applications in international trade. International Journal of English Literature and Social Sciences, 7(1), 196-206. https://doi.org/10.22161/ijels.71.26
  » https://doi.org/10.22161/ijels.71.26
• Martinez, T. S. (2023). Acordo Mercosul-União Europeia e mudança estrutural: considerações a partir de modelos de equilíbrio geral (Carta de Conjuntura, No. 59, pp. 1-18). Rio de Janeiro: IPEA. Retrieved in 2026, April 20, from https://www.ipea.gov.br/cartadeconjuntura/
  » https://www.ipea.gov.br/cartadeconjuntura/
• Megiato, E. I., Massuquetti, A., & de Azevedo, A. F. Z. (2016). Impacts of integration of Brazil with the European Union through a general equilibrium model. EconomiA, 17(1), 126-140. https://doi.org/10.1016/j.econ.2015.10.001
  » https://doi.org/10.1016/j.econ.2015.10.001
• Mughwai, V. L. (2020). Game theory and its application in international trade: Use of strategic games in trade policy. Academic Journal of Economic Studies, 6(2), 96-101.
• Nazareth, M. S., Gurgel, A. C., & Vieira, W. C. (2019). Federalismo fiscal market-preserving: Uma análise de equilíbrio geral computável para o Brasil. Estudos Econômicos, 49(2), 265-304. https://doi.org/10.1590/0101-41614923maw
  » https://doi.org/10.1590/0101-41614923maw
• Olson, M. (1965). The logic of collective action: public goods and the theory of groups. Cambridge: Harvard University Press. https://doi.org/10.4159/9780674041660
  » https://doi.org/10.4159/9780674041660
• Organisation for Economic Co-operation and Development – OECD. (2024). Agricultural policy monitoring and evaluation 2024: innovation for sustainable productivity growth. Paris: OECD Publishing.
• Ossa, R. (2014). Trade wars and trade talks with data. The American Economic Review, 104(12), 4104-4146. https://doi.org/10.1257/aer.104.12.4104
  » https://doi.org/10.1257/aer.104.12.4104
• Sampaio, L. M. B., & Sampaio, Y. (2009). Prioridade brasileira entre acordos de livre comércio utilizando equilíbrio geral aliado à teoria dos jogos. Revista Economia, 10(3), 533-551.
• Sanguinet, E. R., Siqueira, L. V., Coronel, D. A., & Schultz, G. (2017). Práticas intervencionistas e seus efeitos sobre o comércio internacional de soja: uma análise a partir de um modelo de equilíbrio geral computável e da teoria dos jogos. Revista de Economia e Sociologia Rural, 55(4), 641-660. https://doi.org/10.1590/1234-56781806-94790550402
  » https://doi.org/10.1590/1234-56781806-94790550402
• Swinnen, J. (2021). The political economy of agricultural and food policies. In K. Otsuka & S. Fan (Eds.), Agricultural development: new perspectives in a changing world (pp. 471-502). IFPRI.
• Viana Martins, M. M., Cechin, A., Bispo, S. Q. A., Araújo Pedrosa, F., & Braga Nonnenberg, M. J. (2024). Subsídios agrícolas e comércio internacional: quais são as implicações sobre as exportações mundiais e brasileiras? Estudos Econômicos, 41(82), 155-187. https://doi.org/10.52292/j.estudecon.2024.3508
  » https://doi.org/10.52292/j.estudecon.2024.3508
• Wang, L., & Aguiar, A. (2023). Agricultural domestic support (Global Trade Analysis Project, Chap. 11A). West Lafayette: Center for Global Trade Analysis, Purdue University.
• World Trade Organization – WTO. (2008). Revised draft modalities for agriculture (TN/AG/W/4/Rev.4). Geneva: Committee on Agriculture, Special Session.
• World Trade Organization – WTO. (2014). DS267: United States - subsidies on upland cotton. Dispute settlement. Geneva: WTO.
• World Trade Organization – WTO. (2015). Ministerial decision on export competition (WT/MIN(15)/45, WT/L/980). In Proceedings of the 10th WTO Ministerial Conference, Nairobi.
• World Trade Organization – WTO. (2021). DS580: India - measures concerning sugar and sugarcane - summary of key findings. Dispute settlement. Geneva: WTO.
• World Trade Organization – WTO. (2024). Agriculture Information Management System (Ag-IMS): question and answer 108068. Geneva: WTO.
• Yilmaz, K. (2006). How much should primary commodity exports be taxed? Nash and Stackelberg equilibria in the global cocoa market. The Journal of International Trade & Economic Development, 15(1), 1-26. https://doi.org/10.1080/09638190500523360
  » https://doi.org/10.1080/09638190500523360

Edited by

• Associate Editor:
  Daniel Arruda Coronel

Publication Dates

• Publication in this collection
  17 Aug 2026
• Date of issue
  2026

History

• Received
  13 Feb 2026
• Accepted
  29 June 2026