Handbook of Commercial Policy

By Elsevier Science

Handbook of Commercial Policy explores three main topics that permeate the study of commercial policy. The first section presents a broad set of basic empirical facts regarding the pattern and evolution of commercial policy, with the second section investigating the crosscutting legal issues relating to the purpose and design of agreements. Final sections cover key issues of commercial policy in the modern global economy.

Every chapter in the book provides coverage from the perspectives of multilateral, and where appropriate, preferential trade agreements. While most other volumes are policy-oriented, this comprehensive guide explores the ways that intellectual thinking and rigor organize research, further making frontier-level synthesis and current theoretical, and empirical, research accessible to all.

• Covers the research areas that are critical for understanding how the world of commercial policy has changed, especially over the last 20 years
• Presents the way in which research on the topic has evolved
• Scrutinizes the economic modeling of bargaining and legal issues
• Useful for examining the theory and empirics of commercial policy

• Language: English
• Publisher: Elsevier Science
• Release date: Nov 2, 2016
• ISBN: 9780444639264

Book preview

Handbook of Commercial Policy, Volume 1B

First Edition

Kyle Bagwell

Stanford University, Stanford, CA, USA

Robert W. Staiger

Dartmouth College, Hanover, NH, USA

Table of Contents

Cover image

Title page

Copyright

Introduction to the Series

Preface

Trade Agreements: Issue Areas

Contributors

Chapter 1: Enforcement and Dispute Settlement

Abstract

1 Introduction

2 A Simple Model of Trade

3 Self-Enforcing Trade Agreements and Dispute Settlement

4 Incomplete Contract and Dispute Settlement

5 Empirical Studies

6 Concluding Remarks

Acknowledgments

Chapter 2: The Escape Clause in Trade Agreements

Abstract

1 Introduction

2 The Escape Clause in Existing Trade Agreements

3 The Motivation for Escape Clauses

4 Escape Clauses and Optimal Contracts

5 Other Flexibility Mechanisms

6 Does the Escape Clause Undermine or Promote Trade Liberalization?

7 Empirical Analysis

8 Conclusions

Acknowledgments

Chapter 3: Dumping and Antidumping Duties

Abstract

1 Introduction

2 A Brief Primer on the History of AD Laws and Basics on Implementation

3 Facts and Figures on AD Use

4 Key Issues Traditionally Addressed in the Economics and Law Literature

5 Recent Research Developments

6 Concluding Comments

Acknowledgments

Chapter 4: Subsidies and Countervailing Duties

Abstract

1 Introduction

2 Domestic Subsidies and Shallow Integration

3 Domestic Subsidies and Deep Integration

4 Prohibition of Export Subsidies

5 Conclusions

Acknowledgments

Chapter 5: Nontariff Measures and the World Trading System

Abstract

1 Introduction

2 NTMs: Definitions, Data, and Trade Effects

3 Treatment of NTMs in Trade Agreements: Theory

4 Treatment of NTMs in the WTO

5 Other Approaches to International Coordination Over NTMs

6 Conclusion

Acknowledgments

Appendix

Chapter 6: Preferential Trade Agreements

Abstract

1 Introduction

2 Stylized Facts and a Taxonomy

3 Trade and Welfare Effects on Members

4 Economic Effects of Deeper PTAs

5 PTA Formation and Policies: Motives and Determinants

6 Agreement and Policy Interdependence

7 Lessons and Future Research

Acknowledgments

Chapter 7: Special and Differential Treatment for Developing Countries

Abstract

1 Introduction

2 The Institutional Setting

3 Theoretical Analyses

4 Empirical Evidence on the Effects of SDT

5 Concluding Remarks and Future Research

Acknowledgments

Appendix Programs of Nonreciprocal Preferences

Chapter 8: Trade, Intellectual Property Rights, and the World Trade Organization

Abstract

1 Introduction

2 TRIPS and the Global Distribution of Patents

3 Economics of International Patent Protection

4 North-South Models of Technology Transfer

5 Exhaustion of IPRs

6 Compulsory Licensing Under TRIPS

7 Empirical Evidence on IPR Protection

8 Conclusion: Lessons Learnt and What Next?

Acknowledgments

Chapter 9: Issue Linkage

Abstract

1 Introduction

2 Where Do We See Issue Linkage in Reality?

3 A Unifying Framework

4 Gains and Losses From Enforcement Linkage

5 Gains and Losses From Negotiation Linkage

6 Gains and Losses From Participation Linkage

7 Taking Stock

8 Transaction-Cost Arguments Against Linkage

9 Empirical Work on Issue Linkage

10 Coercive Trade Sanctions

11 Conclusion

Acknowledgments

Index

Copyright

Introduction to the Series

Kenneth J. Arrow

Michael D. Intriligator

The aim of the Handbooks in Economics series is to produce Handbooks for various branches of economics, each of which is a definitive source, reference, and teaching supplement for use by professional researchers and advanced graduate students. Each Handbook provides self-contained surveys of the current state of a branch of economics in the form of chapters prepared by leading specialists on various aspects of this branch of economics. These surveys summarize not only received results but also newer developments, from recent journal articles and discussion papers. Some original material is also included, but the main goal is to provide comprehensive and accessible surveys.

The Handbooks are intended to provide not only useful reference volumes for professional collections but also possible supplementary readings for advanced courses for graduate students in economics.

Preface

Kyle Bagwell; Robert W. Staiger

The Handbook of Commercial Policy surveys recent developments in the study of commercial policy. Economic research on commercial policy has flourished in the past two decades. Using new theoretical models, empirical methodologies, and data sources, important recent research offers valuable new insights regarding the determinants and effects of unilateral trade policies. In turn, this work feeds into a primary focus of modern commercial policy research that concerns the role of trade-agreement rules in shaping trade-policy conduct. From this general perspective and motivated by the success of the General Agreement on Tariffs and Trade (GATT) and its successor organization, the World Trade Organization (WTO), a vibrant area of recent research studies the purpose and design of trade agreements. The GATT/WTO has in effect served as the constitution of the postwar international trading system, and an important objective of recent research is to interpret and evaluate the key GATT/WTO design features. Another central feature of the modern trading environment is the rising prominence of preferential trading agreements (PTAs). A large set of recent research also offers new theoretical and empirical insights regarding the implications of PTAs for member and nonmember countries and for the multilateral trading system more generally.

The primary objective of this handbook is to provide a comprehensive survey of a large body of frontier economic research on commercial policy. The handbook will be an important resource for research economists already working in commercial policy, and it will also provide an essential guide to state-of-the-art research in commercial policy for economists and graduate students who are new to the field. While the handbook does not aim to provide a comprehensive treatment of the legal design of trade agreements, it also provides some legal background on trade agreements that will be useful for research economists.

The Handbook of Commercial Policy is organized into three parts that are contained in two volumes. The first volume holds the first two parts of the handbook. The first part has four separate chapters covering economic research that describes the broad set of basic empirical facts regarding the pattern and evolution of commercial policy, the unilateral determinants of commercial policy in developed and in developing countries, the effects of commercial policy intervention on economic magnitudes of interest, and quantitative models of commercial policy. The second part contains four separate chapters providing the relevant legal background for economic research on multilateral and preferential trade agreements and covering economic research on cross-cutting issues relating to the purpose and design of trade agreements. The second volume contains the third part of the handbook, which entails nine separate chapters covering economic research on key issue areas relevant to trade agreements in the modern global economy. The specific issue areas covered here are: enforcement and dispute settlement, escape clauses, dumping and antidumping duties, subsidies and countervailing duties, nontariff measures, preferential trade agreements, special and differential treatment for developing countries, intellectual property rights and the WTO, and issue linkages.

We are deeply grateful to the authors of these chapters for their hard work and exceptional contributions. As a result of their efforts and expertise, this handbook will play a valuable role for economic research on commercial policy by organizing a large body of existing research, identifying key open questions, and stimulating important new research.

We also gratefully acknowledge the work of those associated with the Elsevier Handbooks in Economics series, including Founding Editors Kenneth J. Arrow and Michael D. Intriligator, current Editors Julio Rotemberg and Michael Woodford, Editorial Project Manager Hannah Colford, Acquisitions Editor Kirsten Shankland, and Project Manager Radhakrishnan Lakshmanan. We are especially grateful to Scott Bentley for his encouraging response to this project during an initial meeting at Stanford and for his guidance of the project throughout the early stages. We also thank Dartmouth College, which provided generous financial support and facilities for a conference in which initial drafts of chapters were presented. For their valuable comments, we also thank the chapter discussants—Manuel Amador, Mostafa Beshkar, Emily Blanchard, Bruce Blonigen, Eric W. Bond, Chad P. Bown, Lorenzo Caliendo, Paola Conconi, Arnaud Costinot, Meredith Crowley, Avinash Dixit, Caroline Freund, Penny Goldberg, Bernard Hoekman, Doug Irwin, Brad Jensen, Pravin Krishna, Nuno Limao, Rodney Ludema, Giovanni Maggi, Petros C. Mavroidis, Anna Maria Mayda, Phil McCalman, Peter Neary, Marcelo Olarreaga, Emanuel Ornelas, Jee-Hyeong Park, Nina Pavcnik, Michele Ruta, Kamal Saggi, Peter Schott, Joel Watson, Alan Winters, and Mark Wu—and other conference participants. Finally, we are grateful to Christianne Wohlforth, whose organizational expertise was instrumental in the success of the handbook conference.

Trade Agreements: Issue Areas

Contributors

Chapter 1

Enforcement and Dispute Settlement

J.-H. Park    Seoul National University, Seoul, Republic of Korea

Abstract

This chapter reviews the literature on the enforcement and dispute settlement of international trade agreements. It organizes the review of theoretical developments according to the following questions, of which relevant studies provide new insights: How can governments enforce trade agreements and settle disputes in the presence of incomplete information about trading environments or imperfect information about protection measures? Can a multilateral trading environment and asymmetry among trading countries affect the enforcement and dispute settlement of trade agreements? How can governments enforce trade agreements when renegotiation is possible? What is the role of a dispute settlement system that involves a third-party ruling when trade agreements are largely incomplete contracts because of various transaction costs, such as contracting costs, costly transfers, and asymmetric information? The review of empirical studies discusses the evidence for self-enforcing policy coordination implied by theoretical models, and then examines the empirical studies on trade disputes of the GATT/WTO.

Keywords

Trade agreements; Enforcement; Dispute settlement; GATT; WTO

JEL Classification Codes

F02; F13

1 Introduction

1.1 Basic Questions

This chapter reviews the recent developments in the literature on the enforcement and dispute settlement of trade agreements.a Any international trade agreement poses two main issues for participating governments, namely, negotiating an agreement and enforcing the negotiated agreement. Although these two issues are closely related as known in the literature, this chapter only reviews those studies that primarily focus on enforcement and dispute settlement. This chapter also excludes important studies on topics that are highly related to enforcement and dispute settlement, including escape clause, nontariff measures, preferential trade agreements, and issue linkage, because these topics have been extensively covered in other parts of this handbook.

Although the recent theoretical developments in this research area can be organized in many ways, I classify these developments according to the theoretical frameworks that are commonly employed in the literature. Specifically, the review of theoretical works is divided into self-enforcing trade agreements and dispute settlement (Section 3) utilizing repeated game frameworks, and incomplete contract and dispute settlement (Section 4) typically utilizing nonrepeated-game frameworks that largely abstract away from self-enforcement-related issues. This classification sorts the studies that employ a common analytical framework into closely related topic areas, thereby facilitating the review of these studies.

To comprehend the distinctive issues that are addressed in the reviewed studies, one must recognize the existence of closely related areas of research and understand how the main issues of the reviewed literature differ from those of studies from related areas. The first related study pertains to the analysis of collusion using a repeated game framework in the industrial organization literature. In contrast to collusion under which firms coordinate their actions typically without any explicit agreement or institution, self-enforcing trade agreements and dispute settlement involve both explicit agreements and institutional arrangements, such as the dispute settlement system of the World Trade Organization (WTO) as discussed in the following subsection. Most of the reviewed studies in Section 3 analyze the role of such an institution in facilitating or hindering the policy coordination of governments in a repeated trade relationship. The second related study pertains to the analysis of incomplete contracts in the law and economics literature. Although studies on incomplete contract and dispute settlement as reviewed in Section 4 and the corresponding studies in law and economics both analyze the contractual environments that prevent interested parties from signing a fully-contingent and efficient contract, trade agreements still have certain distinctive characteristics, such as absence of cash transfers between governments as a way of compensation for breaching a contract. The reviewed studies in Section 4 mostly examine such special aspects of trade agreements to characterize the optimal contractual form for trade agreements.

Each subsection in Section 3 is related to the following fundamental questions of enforcement and dispute settlement that recent studies have tried to answer: How can governments enforce trade agreements and settle disputes in the presence of incomplete information about trading environments (Section 3.2.1) or imperfect information about protection measures (Section 3.2.2)? Can a multilateral trading environment and asymmetry among trading countries affect how governments enforce trade agreements and settle disputes (Section 3.3)? How can governments enforce trade agreements when renegotiation is possible (Section 3.4)? The reviewed studies in Section 4 mostly focus on the following questions: What should be the role of a dispute settlement system that involves a third-party ruling, such as that of the WTO, when trade agreements are incomplete contracts because of various kinds of transaction costs? How is the answer to this question related to what we observe in reality? To answer these questions, Section 4.2 examines the contracting costs and potentially erroneous third-party rulings as primary factors of contractual incompleteness. Given the imperfection of third-party rulings, the studies in Section 4.3 analyze how the costly transfers among governments, which are associated with possible renegotiation, affect the optimal trade agreement when random shocks to economies necessitate a contingent protection, of which governments may disagree over the legitimacy.

Section 5 is divided into two subsections. Section 5.1 reviews the empirical studies that provide evidence on self-enforcing policy coordination that is implied by theoretical models of trade agreements. Given that Section 3 mostly focuses on those studies that emphasize informational issues related to enforcement and dispute settlement, Section 5.1 extensively discusses theoretical studies that emphasize observable economic/political shocks as determinants of protection levels, prior to reviewing the empirical studies that test predictions of such models. Section 5.2 reviews the empirical studies on the trade disputes of the General Agreement of Tariffs and Trade (GATT) and the WTO. Section 5.2 discusses notable facts on the WTO dispute settlement data (Section 5.2.1), reviews the empirical studies on selection issues associated with GATT/WTO disputes (Section 5.2.2), surveys studies that investigate the determinants of dispute outcomes, such as early settlement and trade volume outcomes (Section 5.2.3), and reviews the empirical studies on trade disputes related to the Most Favored Nation (MFN) requirement (Section 5.2.4).

To facilitate the review of theoretical studies, Section 2 presents and develops a simple bilateral trade model into two kinds of trade-policy models, one of which has political economy influences (Section 2.2.1) and the other involves economic uncertainty and concealed trade barriers (Section 2.2.2). Section 3 presents a basic repeated-game framework that is commonly used in the literature before reviewing those studies on self-enforcing trade agreements and dispute settlement. Section 4 reviews the recent studies that analyze trade agreements as incomplete contracts. Given that some of these analyses employ a trade model that is distinctive from the ones specified in Section 2, Section 4 explains such a model in detail. Section 5 reviews the empirical studies and links their findings with predictions of the theoretical models that are reviewed in the preceding sections. Section 6 concludes this chapter.

1.2 Institutional Background: A Minimal Sketch

Prior to reviewing studies on the enforcement and dispute settlement of international trade agreements, this subsection briefly describes the institutional background of this issue by referring to detailed discussions in the other parts of this handbook.b Given the central role of WTO in the world trading system, the rules and procedures of this organization for enforcement and dispute settlement must be reviewed to understand the model setups of the theoretical and empirical studies that are reviewed in this chapter. The Dispute Settlement Understanding (DSU) of WTO specifies the rules and procedures of the organization in settling disputes. Any member government (excluding private entities) may file a petition to the WTO's Dispute Settlement Body (DSB) against the measures of its trading partner that is suspect of violating the anticipated commitments under the trade agreements. While DSB comprises all WTO member countries in the sense that the outcome of such petition can only be rejected through a unanimous vote among all members, in practice, the active participants of a dispute settlement procedure include a complainant government, a respondent government, a third-party panel that comprises a few specialists who are supposed to be neutral with regard to the disputed issues and make a ruling on the petitioned case, the Appellate Body that judges the consistency of the WTO rulings when the ruling of a third-party panel is appealed, and other governments who list themselves as interested third party and may share the potential benefit/cost of a ruling. If the dispute cannot be resolved through mandatory consultations with the respondent resulting from a formal dispute petition, the complainant government can trigger a formal legal process in which the legal representatives of the two sides present arguments and evidence to a third-party panel. Approximately two-fifths of the 502 dispute cases that are filed to the WTO from its inception in 1995 to January 2016 have resulted in formal legal rulings by third-party panels. More than half of these cases with formal legal rulings have been appealed to the Appellate Body.

Although DSU and its workings, as discussed in the preceding paragraphs, demonstrate an elaborate multilateral system of dispute settlement that resembles a domestic court system, the current system was developed in response to aggressive unilateralism that had gained its power under the GATT regime, especially during the 1980s. As discussed in a book edited by Bhagwati and Patrick (1990), major trading countries, including the United States, were not satisfied with the workings of enforcement and dispute settlement under GATT. These countries tried to solve trade disputes and implement trade-policy commitments in their interests through the threat of unilaterally invoking a tariff war that is based on their own domestic trade laws, such as Section 301 of the US trade law. Such discontent of GATT member countries has driven efforts to refine the multilateral dispute settlement system under the Uruguay round negotiation, thereby creating the aforementioned dispute settlement system. Although DSU prohibits unilateral imposition of retaliatory measures against the potential violations of trade agreements, DSU allows a complainant to use retaliatory measures of withdrawing concessions that will inflict a damage equal to that caused by the illegal measure(s) that a respondent refuses to change in compliance with the ruling of a panel. This process reflects the special nature of an international dispute settlement system that lacks an external coercive enforcement authority, thereby ultimately relying on the actions of interested parties against deviations.

In addition, the Trade Policy Review Mechanism of WTO mandates each member country to submit regular reports on how its policies adhere to the rules, disciplines, and commitments that are made under the WTO agreements. The reports to be submitted by four major trading countries every 2 years, by 16 countries every 4 years, and by the other member countries every 6 years are thoroughly reviewed by the WTO Secretariat. Thereafter, the report and review are made public to all WTO members. This review mechanism obviously intends to improve the transparency of the trade-related policies and practices of its member countries in enforcing trade agreements.

The contingent protection allowed under WTO should also be described to understand the reviewed studies in this chapter.c Under specified contingencies, the WTO allows its member countries to impose protection measures against the exports of other members that are beyond the levels by which they have committed to bind their protection. The provisions for antidumping (AD) duties against dumping and countervailing duties against subsidies are examples of rules that allow contingent protection. Another example is the safeguard provision for temporarily restricting the imports of a product if the domestic industry is injured or threatened with injury by a surge in imports. These contingent protection measures are often considered in the WTO dispute settlement procedure because some countries may disagree about the legitimacy of such protection.

2 A Simple Model of Trade

2.1 Basic Setup

This section describes a simple model of trade that is often used in the literature on the enforcement and dispute settlement of trade agreements. We suppose that two countries, Home and Foreign, produce and trade two goods, x and y, under perfect competition. We assume that Home (Foreign) is a natural importer of x (y), with the following linear demand and supply functions:

  

(1)

denotes the domestic price of product j∈{x,y} in country i with α, β, and γare required for Home (Foreign) being a natural importer of x (y). Regarding policy instruments, I focus on the case in which Home (Foreign) government chooses its import protection level, which is denoted by τ (τ*).d In the following analysis, τ represent the world price of product j∈{x,y}, then px ττ.

To characterize the equilibrium of this static model of trade, consider the market for product x. The import demand function of Home, Mx , are defined as follows:

  

(2)

and px yields the following:

   (3)

   (4)

as a function of τ*.

The terms-of-trade-driven incentive to raise the protection level τ by Home is shown in (. (3) also demonstrates that Foreign can figure out the protection level of Home even if such level is not publicly visible because Home is using concealed nontariff barriers. If random elements exist in the economy, such as a random positive shock to the domestic suppliers of Home (a larger value for γx), of which Foreign has limited information, the protection level of Home can become its private information, of which Foreign has imperfect information. Section 3.3 will discuss the issue of enforcement and dispute settlement in the presence of imperfect information about protection levels.

2.2 Government Objective Functions and Policy Choices

Using the simple trade model in the preceding subsection, I characterize two types of government objective functions that are often used in the studies on enforcement and dispute settlement. The first type pertains to the preference of the government under political influences, while the second type pertains to the payoff function of the government in the presence of concealed trade barriers and economic uncertainties.

2.2.1 Political Economy Influences

Following Baldwin (1987), I assume that a political economy parameter, λ (λ*), affects how the Home (Foreign) government values its import-competing sector.f Therefore, the preference of the Home government on its import and export products is defined as follows:

  

(5)

and

   (6)

denotes the profit of Home producers of product j∈{x,y} as a function of its domestic price. The Home government's objective function is then computed as the sum of these two preference functions as expressed in the following:

  

(7)

with the preference of the Foreign government being similarly defined.

Each government observes its own political economy parameter before choosing its policy. However, the Home government may not observe the political economy parameter of the Foreign government and vice versa, thereby leading to incomplete information. Given the separability of the government preference over import-competing and exporting sectors, the presence of incomplete information does not affect the unilaterally optimal static policy choice of each government in the aforementioned model. However, the incomplete information on the political economy parameter will affect the enforcement of international trade agreements and the settlement of disputes, which will be discussed later.

Given the government preference specified, the static Nash protection level and the politically efficient protection level are determined in the following way. The static Nash protection level of the Home government is obtained as follows:

   (8)

The first-order condition of the preceding maximization problem of the Home government is as follows:

   (9)

The above condition demonstrates the well-known motivations behind a unilateral choice of protection level, namely, the terms-of-trade motivation and the political motivation. As shown earlier, an increase in τ ). Therefore, the first term in (9) reflects the terms-of-trade motivation. The Home government also considers the effect of τ on its domestic price, which in turn affects the profit and consumer surplus of the import-competing sector as well as its tariff revenue. Given that the political economy parameter affects the preference weight of the government on profit (λπx, the second term in (9) represents the political motivation that is associated with a choice of protection level.g Bagwell and Staiger (1999) introduce the "politically optimal" tariff (of the Home government) concept that results from setting the protection level in the absence of the terms-of-trade motivation by solving the following:h

   (10)

Such a politically optimal protection level is identical to the efficient protection level that maximizes the sum of the objective functions of the Home and Foreign governments by solving the following:i

  

(11)

Denote the static Nash equilibrium protection level and the efficient protection level by τN and τE. Then,

   (12)

and

   (13)

with λ ∈ [1,7/4) guaranteeing that the static Nash protection level is still not prohibitive even when the political parameter reaches its highest level.j

τN is greater than τE, which indicates the terms-of-trade externality of a unilaterally chosen protection level. The gap between τN and τE is reduced as the political economy parameter, λ, continues to increase; the gap eventually disappears when the parameter becomes high enough to entail a prohibitive Nash protection level.k

For notational simplicity, the payoff of each government can be written as a function of protection levels as follows:

  

(14)

where i(≠j) = * or none.

2.2.2 Economic Uncertainties and Concealed Trade Barriers

I now characterize a government payoff function in the presence of concealed trade barriers and economic uncertainties. Following Park (2011), assume that each government chooses its action, a ≡ (τ,e) ∈ A and a*≡ (τ*,e*) ∈ A*, where e and e* denote the explicit import tariff of each government, while A and A* represent sets of nonnegative real numbers. With τi denoting the total level of protection, τi − ei (≥ 0 being assumed) represents the concealed protection level of each country with i .

In the absence of any uncertainty in these trading economies, the amount of imports of each country is a deterministic function of its (total) protection level and terms of trade. This function allows each government to calculate the exact protection level of the other based on publicly observable information even when concealed trade barriers exist. Introducing economic uncertainties into the model to represent the shocks to technology or preference may prevent such calculation, thereby yielding imperfect information on protection levels.

may occur to Home's import sector, thereby changing the slope of its supply curve from γx to γx + θx, with θx ∈ Θ x. In addition, I assume that the Foreign government cannot observe the realized value of the technology shock in the other country but can privately ..n With the realized values of these random variables and a given level of protection, τas functions of τ,θx.o Similarly, one can introduce technology shocks to the production of ypyas functions of τ*,θy.

The random shocks to economies prevent each government from determining the protection level of the other country based on publicly observable information, import level and terms of trade. A lower level of imports (and less favorable terms of trade) may result from a higher concealed protection level, but may also result from a positive supply shock in the import-competing sector.

replace the corresponding variables in (7), the expected payoff of each government, ui, becomes a function of the protection levels of both governments as shown in the following:

  

(15)

where i(≠j) = * or none. λi is suppressed from the preceding expression to indicate that the focus of analysis can be shifted from the issue of enforcing trade agreements under incomplete information on political shocks to the issue of enforcing trade agreements under imperfect information on concealed trade barriers. For simplicity, I consider the case where the efficient protection level that maximizes the total expected payoffs of both governments is equal to zero by assuming that λ = λ* = 1.p

The following section on the imperfect information of concealed trade barriers analyzes the symmetric equilibria of a repeated protection-setting game between symmetric countries. Therefore, I assume that u(τ,τ*) = u*(τ*,τ) for all τ and τ*. Regarding the derivatives of u(τ,τ*) and u*(τ*,τ) with respect to τ and τ*, I assume that the following standard results continue to hold in the presence of random variables: ∂u/∂τ > 0 at τ = 0 (ie, each country has an incentive to raise its protection level above the efficient level), and ∂(u + u*)/∂τ < 0 (ie, raising the protection level reduces the joint expected payoff of the Home and Foreign governments by producing distortional losses). For analytical simplicity, I assume ∂²u/∂τ² < 0 (ie, the marginal gain from protection decreases as the protection level increases). These assumptions guarantee the existence of a unique static optimal protection level for the Home government, which is denoted by τN. Again, τN >τE = 0 because of the terms-of-trade motivation.

3 Self-Enforcing Trade Agreements and Dispute Settlement

3.1 A Basic Analytical Framework: Perfect Public Equilibrium

This section assumes that countries are in a repeated trade relationship in which the per-period bilateral trade described in Section 2 is repeated forever. Governments are assumed to discount their future payoffs at the common rate of δ with 0 ≤ δ ≤ 1. In period t , where i = * or none.

The vector of public information in period t typically includes explicit tariffs, trade volumes, and terms of trade in the previous period t − 1. However, this vector may also include a set of messages that are sent by governments in period t with regard to their own private information as explained in the following subsection. The vector of private information of each government may contain the realized values of the political economy parameters as discussed in Section 2.2.1 as well as some private information on the concealed trade barriers of its own and that of the other government to be discussed in Section 3.2.2.

As an equilibrium concept of this repeated policy-setting game between governments, this section employs the perfect public equilibrium presented by Fudenberg et al. (1994), except for the private trigger strategy of Park (2011) as discussed in Section 3.2.2. The strategy of each government, denoted by σi ∈ Σi, denoted by ρHt ∈ PHt. However, σi , which denote the publication of private information, before taking a policy action. Let the coordination scheme C, denote the rule of actions that governments agree to follow ex ante as a function of public history and their current types (private information). This coordination scheme C from the perspective of the Foreign government for all σ/i ∈ Σi with i = * or none, after any public history of the game).

As discussed in the preceding section, a unilaterally chosen protection level will not internalize the terms-of-trade externality that is associated with such protection, thereby generating the need for trade policy coordination among governments. In the absence of any external enforcement mechanism, an international coordination scheme must entail a punishment scheme against actions that deviate from the agreed-upon action rule as well as a cooperation scheme that specifies the desired actions of governments. Given the focus on a perfect public equilibrium, a punishment phase (defined by a punishment scheme) will be triggered in period t if the history of public information ρHt . Otherwise, a cooperative scheme specifies the actions in a cooperative phase.

In the absence of any informational issue that may arise either from incomplete information on political economy parameters or from imperfect information on concealed trade barriers, governments may enforce efficient protection levels using a simple trigger strategy as shown by Dixit (1987). A cooperative scheme demands each government to set its efficient protection level, whereas a punishment scheme demands each government to set its static Nash equilibrium forever after a punishment phase is triggered. The discounted payoff (of the Home government, for example) that results from following such a cooperative scheme, denoted by VC, is greater than the discounted payoff that results from deviating from such a scheme by setting the static optimal protection and facing the punishment as follows:

  

(16)

as long as the discount factor δ is greater than the following:

where the last inequality comes from W(τE,τ*E) > W(τN,τ*N).q

The presence of incomplete or imperfect information challenges the enforcement of efficient protection levels as shown in the following subsection.

3.2 Informational Issues

3.2.1 Incomplete Information

To discuss the enforcement of trade agreements in the presence of incomplete information, this subsection considers the model with political economy influences in Section 2.2.1. Specifically, each government privately observes the realized value of its random political economy parameter λi ∈ [1,7/4], which can either be low (λi = L) or high (λi = H), with 1 ≤ L < H in period t. The probability of having λi = H .r In addition, I assume that concealed trade barriers do not exist, having ai = (τi = ei,ei). Therefore, τi simply denotes the import tariff in this subsection.

Two types of deviations may arise from a coordination scheme C (ie, a trade agreement), namely, on-schedule and off-schedule deviations. An off-schedule deviation occurs when tariffs differ from those specified by the coordination scheme C are set, thereby constituting a publicly observable deviation. Following the literature, I assume that such deviation triggers a Nash reversion of playing the tariff war of imposing static Nash tariffs, which may possibly last forever. An on-schedule deviation occurs when a government misrepresents its type, thus such deviation is not publicly observable. Similar to (16), one can show that governments can deter off-schedule deviations as long as the discount factor is sufficiently high. To restrain on-schedule deviations, the coordination scheme C needs to provide each government with the incentive to report its type truthfully.

Consider a coordination scheme C that specifies the first-best tariff τE(λ) for all political contingencies λ ∈ [1,7/4]. Such a coordination scheme will induce the Home government to report its type falsely if λ = L with

   (17)

which occurs iff [τE(L) + τE(H)]/2 < τN(L). If L is close enough to H so that [τE(L) + τE(H)]/2 < τN(L), the Home government with λ = L will have an incentive to pretend that it has λ = H under a cooperative scheme in which it is simply supposed to choose an efficient tariff.

3.2.1.1 Efficient Agreements with Transfers

If governments can make direct transfers to each other when applying their tariffs, Bagwell and Staiger (2005) and Martin and Vergote (2008) show a coordination scheme that can implement the first-best outcome of setting τ = τE(λ) for all possible contingencies of λ ∈ [1,7/4].s Consider the cooperative scheme that specifies a tariff-transfer scheme of {τ(μ),TE(μ)} with τ(μ) = τE(μ) and

  

(18)

in a cooperative phase. One can easily check if such a scheme will satisfy the on-schedule incentive compatibility condition. Using a punishment scheme to trigger a permanent static Nash reversion against (off-schedule) deviations, Bagwell and Staiger (2005) prove the existence of δE ∈ (0,1) such that for all δ ∈ [δE,1], the fully efficient tariffs (τE(λ)) can be implemented as a perfect public equilibrium using the tariff-transfer scheme specified in (18), with δE being implicitly defined as follows:

  

(19)

where Eλ(K) represents the expected value of K over λ.t

3.2.1.2 Costly Transfers and Dispute Settlement Procedures

Assume that governments cannot make direct transfers to each other as originally assumed in the basic analytical framework. Beshkar (2010b) shows that reciprocal tariffs may work as a mechanism that induces the truthful reporting of political parameters by curtailing the opportunistic use of higher tariffs. Instead of a tariff-transfer scheme, a tariff–tariff represents an import tariff that the Foreign government can impose in response to the Home government's reporting of its political parameter μ, namely, a reciprocal-tariff constraint, Beshkar shows that the optimal reciprocal-tariff scheme that maximizes the ex ante . With a permanent static Nash reversion against off-schedule deviations, δRT ∈ (0,1) exists such that for all δ ∈ [δRT,1], the aforementioned reciprocal-tariff scheme can be implemented as a perfect public equilibrium.u

Two types of inefficiencies are associated with such an optimal reciprocal-tariff scheme. First, the assigned tariff for a high political parameter is less than the efficient tariff with τ(H) < τE(His also higher than the efficient tariff τE(L) of the Foreign government.is necessary to curtail the incentive of the Home government to misrepresent its political parameter as H. However, such a reciprocal tariff is a distortional way to set the price for reporting H (= τ(H)) below τE(H) will be higher than the marginal distortion loss associated with the corresponding reduction in τ(H) because the latter is equal to zero when τ(H) = τE(H). Therefore, the optimal reciprocal tariff must entail τ(H) < τE(His optimal, which indicates the optimality of a less-than-proportional response to a safeguard action of imposing τ(H)(> τ(L)).

comes not only from the distortional nature of import tariffs as a means of transfer between governments but also from the informational asymmetry that necessitates the use of distortional tariffs along the equilibrium path. With complete information, governments can enforce fully efficient contingent-tariffs with a threat of invoking a static Nash reversion against any opportunistic behaviors as long as δ is sufficiently enough. If a third party can reduce such informational asymmetry between governments, then there must exist a coordination scheme that can outperform the aforementioned simple tariff–tariff scheme.

Beshkar (2010b) characterizes an optimal tariff–tariff scheme in the presence of a third party who can correctly judge whether the report on the political parameter is truthful with a probability of ψ ∈ (1/2,1). He postulates that the DSB of WTO assumes such a role. When the judgment of DSB is available, the optimal tariff–tariff scheme will assign tariffs not only based on the government's reporting on its political parameter μ, but also based on the judgment of DSB (true = T; false = F) about the truthfulness of such reporting, denoted by μDSB = {T,F}.

Beshkar (2010b) focuses on the tariff–DSB . If ψ is greater than a certain critical level ψc ∈ (1/2,1), then the on-schedule incentive constraint does not bind under an optimal tariff–DSB scheme with

  

(20)

, which allows a government reporting H to raise its tariff without a distortional reciprocal increase in the tariff of the other country when DSB judges that the report is true. Therefore, the optimal tariff–DSB scheme will set τ(H,T) > τ(H), which is closer to the efficient tariff τE(H). This is the second source of efficiency gain. However, τ(H,T) > τ(Hτ(H,T) also increases with an increase in τ(H,T). Given the potential loss that is associated with the wrong judgment of DSB, τ(H,T) < τE(H) and τ(H,T) only increases along with ψ. When ψ > ψc, which prevents the on-schedule incentive constraint from binding, Beshkar (2010b) shows that the above aforementioned efficiency gains will cause the ex ante joint payoff of governments under the optimal tariff–DSB scheme to be strictly higher than that under the optimal tariff–tariff scheme.

Having τ(H,T) >τ(Halso increases the incentive of governments to exaggerate their political parameters in expectation of the wrong judgment of DSB. Such an on-schedule deviation incentive is stronger when the judgment of DSB is less accurate. If such judgment is not accurate enough with ψ < ψc, then the on-schedule incentive constraint will bind and the optimal tariff–DSB scheme will differ from the one described in (. As the ex ante joint payoff of governments decreases with a lower value for ψ, another critical value for ψ, is observed, that is the ex ante joint payoff is identical across the tariff–tariff scheme and the tariff–DSB scheme if ψ . The tariff–DSB scheme generates an informational advantage over the tariff–tariff scheme only when the judgement of DSB is accurate enough with ψ .

Even if the DSB–tariff scheme does not generates any informational advantage over the tariff–tariff scheme with ψ , Beshkar (2010b) shows that the DSB–tariff scheme has an advantage over the tariff–tariff scheme in relaxing the off-schedule incentive constraint.

3.2.1.3 A Discussion

The aforementioned analyses of Bagwell and Staiger (2005) and Martin and Vergote (2008) demonstrate that governments can effectively deal with on-schedule deviation incentives in the presence of incomplete information if they utilize nondistortional transfers; they can enforce a fully efficient and truth-telling coordination scheme that is contingent on realized values of private political parameters.w Although the intergovernmental direct transfers that can work as truth-revealing prices to pay for the right to increase import tariffs, governments hardly use such transfers and no trade agreement explicitly promotes the use of such transfers. We need to improve our understanding of why governments do not use direct transfers as a way of efficiently enforcing trade agreements.x

Given that governments are limited to using import tariffs as their primary trade policy measures, Beshkar (2010a, b) demonstrates that using a reciprocal tariff increase against a higher contingent tariff may work as a truth-telling coordination scheme. However, this reciprocal tariff scheme involves distortional losses that are associated with reciprocal tariff increases as well as less-than-efficient contingent tariff increases. Beshkar (2010b) shows that a coordination scheme that involves the judgment of DSB on the legitimacy of contingent protection can be helpful as long as such judgment is accurate enough. This scheme can reduce the frequency of distortional reciprocal tariff increases against the legitimate use of contingent protection as well as increases the level of contingent protection closer to the efficient level. Based on this analysis, Beshkar (2010b) rationalizes the reform in the GATT escape clause that requires the judgment of DSB on the illegitimacy of contingent protection before applying a reciprocal tariff against such protection.

Martin and Vergote (2008) emphasize a related yet different rationale for governments not to rely on reciprocal tariffs in restraining their opportunistic use of contingent protection. Even in the absence of an informative judgment from DSB, a punishment mechanism that involves an asymmetric punishment scheme can outperform the reciprocal tariff scheme and possibly achieve full efficiency with sufficiently high discount factors.y Martin and Vergote then surmise that the retaliatory use of AD duties against potentially illegitimate uses of contingent protection may approximate such an optimal asymmetric punishment scheme.

By relaxing the independence assumption of the distribution of political parameters across periods (ie, governments can be subject to nontransitory political shocks), Bagwell (2009) generates noteworthy results with regard to the enforcement issues in the presence of incomplete information. When a private political shock persists through time, then a government that draws a weak political shock has an incentive to hide such information because this government is unlikely to retaliate effectively, which Bagwell (2009) calls "an enforcement ratchet effect. He shows that the government with a low political pressure may pool" and apply its tariff at the bound rate, which is inefficiently high for a government with a low political pressure.z Bagwell (2009) also demonstrates that persistent political pressures may lead to gradualism in trade agreements, that is, a government with a low political pressure applies a tariff that is below the bound rate in the initial agreement and then applies a lower tariff at the reduced bound rate in the subsequent periods.

3.2.2 Imperfect Information

To review the studies on enforcing trade agreements in the presence of imperfect information, this subsection utilizes the trade model with economic uncertainties and concealed trade barriers as discussed in Section 2.2.2. This model also abstracts away from issues that may arise from random political shocks. There exist publicly observable variables that are correlated with the hidden protection levels of governments, such as explicit tariffs, trade volumes, and terms of trade, thereby allowing each government to infer the protection levels of one another based on such public information. However, the random shocks to economies, makes such inferences imperfect as discussed in Section 2.2.2. These governments may also use some privately observable variables that can improve their inferences on the hidden protection of other governments. This subsection reviews those studies that explore the possibility of utilizing both imperfect public and private information in sustaining cooperative trade policy choices in a repeated trade relationship.

This section initially focuses on the symmetric perfect public equilibrium, in which governments use ex ante symmetric strategies. Such focus on the symmetric equilibrium allows us to analyze only the incentive constraints of the Home government. Although governments may utilize the entire history of public information to trigger a punishment phase, the studies on enforcement and dispute settlement often focus on a simple stationary trigger strategy of triggering a punishment phase that lasts for T − 1 periods if the current value of public information belongs to a certain stationary triggering set of public information with ρt ∈ PD . Such simple symmetric stationary trigger strategies are denoted as STS.

Although other complex strategies may also outperform STS in term of maximizing the ex ante joint discounted payoff of governments, the focus on STS significantly simplifies the analysis, especially when governments are required to set their static Nash tariffs (τN) in all punishment periods. Focusing on such STS, I can express the ex ante discounted payoff of the Home government from cooperation as follows:

  

(21)

which implies the following:

  

(22)

and τN = τ*N as indicated in the symmetry assumption. Therefore, characterizing the optimal STS solves the following problem:

  

(23)

Most studies on enforcing trade agreements with imperfect information utilize certain variations of the above STS even though the use of private information in triggering a punishment phase generates different issues as discussed later in this subsection.

3.2.2.1 A Brief Review of the Literature Using Public Trigger Strategies

With regard to the issue of enforcing trade agreements in the presence of concealed trade barriers, Riezman (1991) shows that governments can sustain a cooperative protection level that is strictly lower than the static Nash tariff using an import trigger strategy in which the import level that is lower than a certain critical value triggers a punishment.aa The incentive to raise the concealed trade barriers above the cooperative ones is counter balanced by the increased probability of triggering a costly punishment phase, that is, the import tends to decrease below the critical level with a higher protection level. The periodic tariff wars that occur along the equilibrium path are necessary to restrain the use of concealed trade barriers.

Hungerford (1991) models the dispute settlement procedure (DSP) of GATT as an investigation requirement before initiating a punishment phase even when the import level drops below a critical level. Given that governments endogenously determine the likelihood of detecting concealed trade barriers by including more resources into their investigation, Hungerford (1991) emphasizes that the equilibrium concealed protection must entail a positive protection level because there will be no incentive for governments to invest any resources into the investigation in the absence of concealed trade barriers in the equilibrium. He uses this result to explain the prevalent use of concealed trade barriers or nontariff barriers, which in turn he attribute to the requirement for a GATT investigation before triggering punishments against such barriers.

In contrast to this negative view of the role of DSP, Kovenock and Thursby (1992) emphasize that DSP can have the positive role of imposing an extra cost for breaching international obligations on deviations that goes beyond the typical deviation cost of increasing the probability of triggering a costly punishment phase. DSP publicizes and disseminates information on deviations across countries to create such an extra cost. However, using DSP may delay the beginning of a punishment phase during which a respondent government may continue to impose its deviatory protection level, while a complainant government is not allowed to react to such an action during the investigation phase of DSP. According to Kovenock and Thursby (1992), such a delay may lead to an incentive to punish the deviations outside of DSP, even when employing such punishment incurs an international obligation cost. The complainant government will compare the benefits from reacting to a deviatory action of the respondent government with the international obligation cost from pursuing the punishment outside of DSP.

Chisik (2012) analyzes how various types of dispute settlement mechanisms affect the product quality and export promotion choices of firms when such choices limit their future production flexibility. Specifically, he compares two dispute settlement mechanisms. On the one hand, a weak dispute settlement (WDS) has a third party that settles a dispute (declares settlement) with a certain probability. Similar to the DSP in the analysis of Klimenko et al. (2008) as reviewed in Section 3.4, such a dispute is triggered when the export prices are reduced below a certain critical level by the potential deviations of increasing the protection levels above the cooperative ones. On the other hands, a strong dispute settlement (SDS) utilizes the available information to detect potential deviators and recommends asymmetric continuation payoffs in the punishment phase, in which only the government that is detected as a potential deviator will suffer. Chisik (2012) shows that compared with the equilibrium under WDS, SDS induces more efficient product quality and export promotion choices, thereby generating higher volumes of trade and lower protection levels in the equilibrium.

3.2.2.2 Private Trigger Strategies and Dispute Settlement Procedures

In contrast to studies that rely on STS to analyze the enforcement and dispute settlement of trade agreements in the presence of imperfect information, Park (2011) explores the possibility of utilizing the imperfect private information of potential deviations to restrain the use of concealed trade barriers.

Governments in a repeated trade relationship may have certain information on the concealed trade barriers of their trading partners that is not publicly verifiable but is more accurate than other public measures such as import levels. For instance, the trade model in , which belongs to its set of private information (Φi, such that a punishment phase will be triggered if ωi(∈Ωi) belongs to Ωp. Such trigger strategies that utilizes more information on concealed trade barriers may outperform those that only rely on public information (P). Given our focus on simple symmetric stationary trigger strategies, I denote the former as private trigger strategies (PTS) and the latter as STS.

Despite the informational advantage of PTS over STS, the former potentially raises a well-known issue in repeated games with imperfect private monitoring, that is, if some private information triggers a punishment phase, then governments may not realize whether they are in a punishment or a cooperative phase, which in turn destroys the recursive structure of the repeated game and generates serious analytical issues. To coordinate their punishment actions, each government may impose its static Nash tariff (e = e* = τN with τ − e = τ*− e* = 0) when starting a punishment phase based on its private information. The use of private information still imposes a new constraint on the punishment phase, that is, the expected payoff from invoking a punishment phase must be identical to the expected payoff from not invoking such a phase, which in turn limits the length of the punishment phases under PTS. Surprisingly, such a constraint on the length of the punishment phases simplifies the ex ante discounted payoff of the Home government under PTS as follows:

  

(24)

and τN = τ*N by the symmetry assumption. Characterizing the optimal PTS solves the following problem:

  

(25)

To analyze the possible role of a third party, such as DSB, in enforcing trade agreements in the presence of imperfect private information on concealed trade barriers, Park (2011) compares the optimal PTS with the optimal STS that involves a third party. To check the potential benefit of having the third-party STS that is not based on informational advantage, the third party is assumed to have access to the Ωi of each government. If the third party finds ωi ∈ ΩP, each government is authorized to trigger a punishment phase by setting its static Nash tariff. The potential benefit of having the third-party STS must then come from the flexibility of the length of punishment phases and the associated change in the optimal triggering information set ΩP. Given that the information set Ω yields sufficiently accurate information on concealed trade barriers, Park (2011) shows that the optimal third-party STS entails the minimum punishment length by specifying one-period asymmetric punishment against potential deviations. The ΩP under the optimal third-party STS is also expanded compared with that under the optimal PTS, thereby increasing the sensitivity of the triggering set to an increase in concealed trade barriers. The optimal third-party STS entails weaker and less costly punishments than those under the optimal PTS, which in turn helps governments to trigger a punishment phase more actively against potential deviations and achieve a higher level of cooperation on protection levels.

To generalize the third-party STS of Park (2011), Zeng (2014) considers the possibility of differentiating the retaliatory protection from the static Nash tariff. Assuming that the third party can induce each government to implement any retaliatory protection level during punishment phases, which may incur fixed costs for the third party to access the information set of each government, Zeng (2014) shows that limiting the retaliatory protection level below the static Nash can be beneficial. The benefit comes not only from the less distortional punishment phase but also from the optimal expansion of ΩP, which in turn lowers the cooperative protection level.

3.3 Asymmetry and Multilateral Aspects

In contrast to studies that focus on informational issues in a bilateral trade relationship between symmetric countries, recent studies have made important contributions to the literature by analyzing the enforcement issues in multilateral trade relationships among possibly asymmetric countries. For analytical simplicity, these studies do not explicitly model the informational issues and instead employ the subgame perfect equilibrium in a repeated game as the equilibrium concept.ac

3.3.1 Bilateral Imbalances of Power and a Multilateral Punishment Mechanism

Maggi (1999) develops a simple trilateral trade model with each country trading a pair of relation-specific products, on which each government may impose import tariffs. To restrain the terms-of-trade-driven use of import tariffs, Maggi considers a simple trigger strategy in which a permanent static Nash reversion is applied as retaliation against defections from cooperative tariffs. The analysis shows that employing a multilateral punishment facilitates a higher level of cooperation only when "bilateral imbalances of power [exist] in the sense that partners in each bilateral relationship stand to lose different amounts from a tariff war." In the absence of such imbalances, triggering a multilateral tariff war against deviations does not strengthen the punishment power because each government will set its static optimal tariffs against all of its trading partners when deviating; the incentive constraint for setting a cooperative tariff under the bilateral punishment scheme in (16) remains