LEGISLATED PROTECTION AND THE WORLD TRADE ORGANIZATION

• Author: T. Renee Bowen
• Published date: 01 November 2015
• DOI: http://doi.org/10.1111/iere.12140
• Date: 01 November 2015

INTERNATIONAL ECONOMIC REVIEW

Vol. 56, No. 4, November 2015

LEGISLATED PROTECTION AND THE WORLD TRADE ORGANIZATION∗

BYT. RENEE BOWEN1

Stanford Graduate School of Business and the Hoover Institution, U.S.A.

Tariff bindings and administered protection are two characteristics of the World Trade Organization (WTO) that are

little understood. Tariff bindings place a ceiling on tariffs that is not always reached, whereas administered protection

provides all industries with some minimum import protection, effectively creating a floor for protection. How do these

policies affect applied most favored nation (MFN) tariff rates that are enacted through the legislature? I model tariffs

determined by a dynamic legislative process and show that low applied MFN tariffs are less likely with tariff bindings

and more likely with administered protection than under purely legislated protection.

1. INTRODUCTION

How does the World Trade Organization (WTO) affect tariffs that are enacted by a domestic

legislature? The main objective of the WTO is to reduce tariffs among its member countries.

It does this by facilitating the negotiation of tariff limits (or bindings) among members and

allowing domestic governments the freedom to set their applied most favored nation (MFN)

tariff rates within those limits.2Since a high tariff in one industry benefits producers in that

industry at the expense of all consumers, a broad-based tariff reduction can only come about

through a compromise among industries. What effect does the WTO have on the ability of

legislatures to reach the political compromise necessary to enact lower applied tariffs?

To answer this question, I first present a formal model of tariff determination through the

legislative process and ask under what circumstances will the legislative process result in low

applied MFN tariffs. I develop a dynamic model of a small open economy based on the static

economy of Grossman and Helpman (1994). Within the economy, there are legislative districts

that specialize in different industries. Individuals in each district are identical, so preferences

of elected representatives reflect the preferences of all members of the district. These members

have preferences for high tariffs on the good produced in their district and negative tariffs

(import subsidies) on all other goods. In this model, such preferences lead to dead-weight

losses because of losses in consumer surplus, whereas free trade is the utilitarian optimum.

Each period, trade policy is determined through the legislative process as a game among

locally elected representatives. A trade policy vector is proposed by a randomly selected

legislator and is passed by a majority vote. This stylized legislative process is common in the

literature on legislative bargaining. It was introduced by Baron and Ferejohn (1989), who

argued that with a large number of legislators, each seeking to put forward his own policy, a

legislative process that does not favor a particular legislator will result in a randomly selected

∗Manuscript received February 2012; revised August 2013.

1I wish to thank Roger Lagunoff, Rodney Ludema, Kyle Bagwell, Emily Blanchard, Marc Busch, and participants

at the Stanford University International Workshop, West Coast Trade Workshop, The Conference on the Political

Economy of International Organizations, International Trade Commission seminar, the Midwest International Trade

and Theory meetings, the Midwest Economics Association, and the Georgetown University Trade Group for invaluable

comments and suggestions. Thanks are also due to Yingtian He and Jackie Chan for helpful research assistance, Ravi

Pallai for technical assistance, and anonymous referees for insightful comments and suggestions. The author is solely

responsible for any errors or omissions. Please address correspondence to: T. Renee Bowen, Stanford Graduate School

of Business, 655 Knight Way, Stanford, CA 94305. Phone: (650)721-1299. E-mail: trbowen@stanford.edu.

2Applied MFN tariffs are those generally applied to all members of the WTO by another WTO member.

1349

C

(2015) by the Economics Department of the University of Pennsylvania and the Osaka University Institute of Social

and Economic Research Association

1350 BOWEN

proposer each period. This is appropriate in the context of trade policy, as legislators are

constantly vying for protection for their industry. By modeling each district with a single

industry, I provide the starkest possible representation of trade policy conflict.

The status quo trade policy is modeled as endogenous. If the current period’s proposal fails

to achieve a majority vote, the previous period’s tariff vector remains effective. In trade policy,

a reversion to the status quo tariff reflects the fact that trade policies remain effective until

amendments are passed by the legislature.

I show that an equilibrium exists in which low applied tariffs—defined as an outcome where

all districts, except one, maximize their joint stage utility, resulting in low positive tariffs for all

but a single industry—is a possible outcome of the legislative process. However this equilibrium

is dependent on initial conditions. For initial conditions that closely resemble free trade, the

outcome will be low applied tariffs, whereas for any other set of conditions the outcome will be

abiased outcome—where each period a single industry receives high protection and all other

industries receive negative protection.3

Given this model of legislated trade policy, I consider the impact on the equilibrium outcome

of two characteristics of the WTO. The first is tariff bindings when they are set above applied

rates. This applies mainly to developing countries who negotiate tariff bindings well above

applied rates to comply with WTO requirements. Negotiations over tariff bindings are usually

conducted by an executive branch of the government, whereas applied MFN tariffs are set

by a domestic legislature that views tariff bindings as an exogenous ceiling. The second

characteristic considered is administered protection, for example, antidumping duties. This

ensures that all industries have access to some minimum import protection, effectively creating

a floor for applied tariffs.4I show that tariff bindings decrease the set of initial conditions that

result in low applied tariffs, but low levels of administered protection expands the set of initial

conditions that results in low applied tariffs.

The intuition for the result is as follows: To sustain an equilibrium in which low applied rates

are possible, there must be a threat of spiralling toward the biased outcome. The biased outcome

therefore acts as a “punishment,” because it results in a lower long-run payoff than low applied

tariffs.5Tariff bindings essentially impose a ceiling on protection allowed to all industries,

thereby making the biased outcome less biased. This increases the expected payoff to the biased

outcome and hence increases the incentive to enact the biased tariffs versus low applied tariffs.

Administered protection, on the other hand, essentially imposes a floor on tariffs applied to

any industry in equilibrium. In order to achieve a biased outcome, legislators will cherry-pick

minimum winning coalitions and freeze out the remaining legislators by reducing tariffs on their

industries. Placing a floor on tariffs raises the cost of freezing out legislators, and hence decreases

the incentive to go to the biased outcome. It should be noted that if administered protection is

sufficiently large, or tariff bindings sufficiently low, the equilibrium breaks down. This is consis-

tent with the fact that bindings set low enough result in low applied tariffs trivially and vice versa.

Little formal work has been done to examine the equilibrium effects of administered

protection and tariff bindings, and even less has been done to look at protection as an outcome

of the legislative process. Mayer (1984) looks at tariff determination through direct democracy,

where citizens vote directly over the formation of tariff policy instead of having elected

representatives decide it through a dynamic legislative process. He focuses on the effects of

voter eligibility rules and shows how actual tariff policy may reflect the preferences of a small

minority of well-endowed citizens. Anderson (1992) considers the impact of the prospect of

administrative protection on a country’s incentives to export and the protectionist response

of the exporting country. Thus Anderson (1992) argues that administrative protection in the

3By a renormalization this can be interpreted as no protection for this special symmetric case. It may not be true

more generally, so the prediction of negative protection may remain without the assumption of symmetry.

4An alternate interpretation of administered protection is allowing a tariff strictly higher than the maximum tariff

with some exogenous but small probability. The main results are robust to this interpretation.

5Note that this is not a punishment in the traditional “trigger strategy” sense since it is the equilibrium outcome for

some initial status quo tariffs.

LEGISLATED PROTECTION 1351

domestic country may have the adverse effect of encouraging protectionism in the exporting

country. Bagwell and Staiger (1990) develop a model that explains administered protection.

They consider two countries’ governments setting trade taxes to maximize national welfare and

show that when future trade volumes are uncertain, equilibrium tariffs will be high when trade

volumes are high. I do not provide here a model that explains the existence of administered

protection and tariff bindings. I provide a model that determines MFN tariffs as decided

through the legislative process and assess the effect of administered protection and tariff

bindings on applied MFN tariffs. Grossman and Helpman (2005) discuss the protectionist bias

of majoritarian politics, but focus on intraparty incentives to maintain protection. I argue in

this article that a legislative process characterized by a majority voting rule can sustain low

tariff levels and need not be biased toward protectionism. When combined with administered

protection, the legislature may in fact have a greater likelihood of maintaining low tariffs.

Anderson (1992) and Bagwell and Staiger (1990) consider the effect of administered

protection on the noncooperative interaction between two countries whereas Grossman and

Helpman (2005) consider trade policy determination as the result of interaction within political

parties. Willmann (2005) considers tariff determination in a citizen-candidate model `

alaBesley

and Coate (1997). In recent work, Celik et al. (2012, 2014) consider trade policymaking in a

model of sequential legislative bargaining `

a la Baron and Ferejohn (1989). This article is a first

attempt to model trade policy determination as the outcome of a dynamic legislative process,

as in Baron (1996), combined with administered protection and tariff bindings. Further, by

showing that tariffs corresponding to the Pareto frontier in a commonly used model in trade

policy sum to a constant, I provide a useable framework for further analysis of legislative

bargaining over tariff policy using existing results on legislative bargaining.

The model of the legislative process I follow is similar to that in Dixit et al. (2000),

Kalandrakis (2004), Kalandrakis (2010), and Bowen and Zahran (2012). Policies in these

papers are purely distributive, allocating a share of a fixed surplus each period to legislators.

Trade policy, in contrast, is a multidimensional public good. A positive tariff on any good

imposes negative externalities on all industries through losses in consumer surplus, but creates

a benefit to the industry on which the tariff is applied through gains in producer surplus. This

article is therefore the first to show that equilibria exist in an endogenous status quo legislative

bargaining game for a multidimensional public good. Baron (1996) showed the existence of an

equilibrium with a single-dimensional public good in such a dynamic game.

The remainder of the article is organized as follows: Section 2 presents the model of a

dynamic endowment economy and derives preferences of individuals in different legislative

districts over trade policies. Section 3 specifies the legislative process. Section 4 defines the

Markov-perfect equilibrium (MPE), and Section 5 characterizes an MPE of the legislative game

with low tariffs. In Sections 6 and 7, I examine the effects of tariff bindings and administered

protection and present the main propositions. Section 8 presents a discussion of welfare, and

Section 9 discusses testable implications. Section 10 concludes.

2. THE ECONOMY

A small open economy produces K+1 goods, k=0,1,...,Keach period over an infinite

horizon. The number of goods Kis at least three for majority voting to be applicable and is odd

to avoid ties. Each period, the economy is modeled as in Grossman and Helpman (1994). Let

ykbe the total output in sector keach period. The production technology is such that one unit

of each good requires one unit of a sector-specific factor; hence ykis also the total endowment

of the factor used specifically in sector keach period. All goods are traded. Good zero is the

freely traded numeraire with price p0=1. All other goods, k=1,...,K, have world price p∗

k.

These prices are exogenously given and constant each period. The domestic price of each of the

non-numeraire goods is the world price p∗

kplus a specific tariff τt

kso pt

k=p∗

k+τt

k. The vector

of specific tariffs τt=(τt

1,...,τ

t

K) is determined by the legislative process at the beginning of

period t, and once a tariff policy is selected, individuals make consumption decisions.