The General Agreement on Tariffs and Trade (GATT) is an important institution. Established in 1947 to encourage the reduction of trade barriers among its 23 member countries, GATT (and now its successor, the World Trade Organization -- WTO) has grown in membership to a roster that currently exceeds 125 countries. The expanding GATT membership reflects the success that this organization has had in facilitating multilateral tariff liberalization. As a consequence of the eight negotiating rounds that GATT has sponsored, the average ad valorem tariff on industrial goods has fallen from over 40 percent to below 4 percent. In light of its significance for the world economy, an important question for economists is whether a theoretical interpretation of GATT and its main features can be provided. Much of my recent research has been an attempt to answer this question, by seeking answers to four sets of related questions.

First, a most basic question: Why do governments bother to negotiate trade agreements? After all, major policy decisions are routinely made by governments without consulting their trading partners. If a government wishes to reduce its trade barriers, why must it wait for a trading partner to reach the same conclusion? Second, given that governments do choose to negotiate trade agreements with one another, what do they hope to accomplish in agreeing to a prior set of rules by which negotiations are to proceed and outcomes abide? These rules, which are contained in the articles of GATT, remain largely a mystery to economists. How do these rules affect trade bargaining outcomes, and why would governments choose to adopt them? Third, how are these agreements enforced, and how do the limits of enforcement shape the features of GATT? And finally, can the same set of principles on which postwar multilateral liberalization has been based be applied to the host of "new" trade policy issues (for example, labor and environmental standards) currently before the WTO? This article describes my research on each of these questions.

Why Negotiate Trade Agreements?

What do governments achieve when they negotiate a trade agreement? Because trade negotiations are voluntary, each government should gain as a result of them, and this is possible only if there exists an inefficiency (relative to the governments' own preferences) that the negotiations can correct. What, then, is the source of the inefficiency that a trade agreement can correct? My research points to two possibilities.

First, governments may face credibility problems, and view trade agreements as a means by which to make policy commitments relative to their own private sectors. In joint work, Guido Tabellini and I have explored the inefficiencies that can arise if the unilateral trade liberalization announced by a government is not deemed credible by its private sector. We provide some empirical evidence in support of the general position that commitments made in a trade agreement can bolster the credibility of trade policy decisions. Our empirical investigation studies U.S. tariff choices, and exploits the fact that trade policy in the United States is implemented in a variety of institutional environments. One environment that we examine is the Tokyo Round of GATT negotiations, and in particular the decision by the United States to exclude certain industries from the across-the-board tariff cuts negotiated in that round. The other is the decision by the United States to grant protection to certain industries through GATT's "escape clause" (under which a government may temporarily "escape" from its GATT commitments). We observe that in each environment the U.S. government had to decide whether to address distributional concerns by granting costly tariff increases (or forgoing beneficial tariff reductions), but that only in the former environment did GATT rules and procedures place explicit constraints on the ability of the U.S. government to subsequently reverse its decision. Therefore, only in the former environment could GATT rules serve as a potential commitment device relative to the private sector. Comparing decisions made across these two environments, we find that GATT rules did help the U.S. government make domestic trade policy commitments, for example by allowing the decision to protect to be more responsive to the cost of the production distortions that would arise as a consequence of that decision. This supports the view that GATT's rules indeed can enhance the credibility of trade policy decisions, and it suggests one possible answer to the question posed earlier: trade agreements can help governments make commitments to reduce their own trade barriers.

But this question also has a common sense answer that every trade policy practitioner knows: governments negotiate trade agreements not because they wish to reduce their own trade barriers but because they seek to reduce the trade barriers imposed by their trading partners, and they are willing to "pay" - with market access "concessions" of their own - for the enhanced access to foreign markets that lower foreign barriers would bring. According to this view, the real question for economists is, Must this behavior be interpreted as reflecting an irrational mercantilist distraction, or can it be understood with sound economic principles? The answer to this question has serious implications for the study of GATT as an institution: if the behavior of trade negotiators makes no sense on any level, then there can be no internal logic to GATT, since this behavior is woven into the very fabric of GATT law and practice.

In joint work, Kyle Bagwell and I have shown that there is a deeper economic logic behind this seemingly irrational behavior, and that it is tied to the terms-of-trade motivations familiar from the classic work of Harry G. Johnson. Moreover, this logic is very general; it applies, for example, to each of the leading political economy models of trade policy (and is in fact the only reason that governments might wish to negotiate a trade agreement in these models). The logic itself is very simple, and can be described in intuitive terms, once it is observed that the terms-of-trade effects of a government's policy choices refer simply to its ability to shift the costs of its policies onto trading partners. This cost-shifting will occur if some of the incidences of a government's policies are borne by foreign exporters. Thus, for example, when a domestic government offers protection to an import-competing industry, some of the costs of that protection will be shifted abroad if foreign exporters accept lower export prices for their sales in the domestic market. When such cost-shifting occurs, governments will likely make distorted policy choices, as they do not bear the whole cost of their decisions. Consequently, from the perspective of cost-shifting, terms-of-trade effects represent a natural source of inefficiency associated with unilateral policy decisions. At the same time, these effects can help to provide an economic explanation for the mercantilist orientation of actual negotiations; they imply that each government is right to pin its hopes for a beneficial outcome of negotiations on its ability to gain enhanced access for its exporters to the markets of its trading partners. Moreover, as we discover in subsequent work, viewing the central "problem that trade agreements can solve" as a terms-of-trade-driven prisoners' dilemma reveals a simple logic to a number of GATT's key principles. These are discussed in the next section.

Why Adopt a "Rules-Based" Approach to Negotiations?

If GATT were simply the codification of a set of negotiated tariffs, would governments have been able to achieve the same degree of success in liberalizing trade barriers over the postwar period? This is a counterfactual situation that we can never observe, but economic theory - together with a knowledge of the institution - can help to provide an answer. Of course GATT is not an agreement to operate at a particular point on the efficiency frontier. Rather, it is a negotiating forum, membership in which carries an obligation to abide by a set of rules under which future negotiations can occur and future conduct will be judged. While these rules are laid out in a series of GATT Articles, the pillars of GATT are the principles of reciprocity and nondiscrimination (most favored nation, or MFN), and that enforcement mechanisms form the heart of the GATT system. I will discuss reciprocity and MFN here, leaving issues of enforcement for the following section.

The principle of reciprocity is a GATT norm under which one country agrees to reduce its level of protection in return for a reciprocal concession from its trading partner. At the broadest level, this principle refers to the "ideal" of mutual changes in trade policy that bring about equal changes in import volumes across trading partners. The principle of nondiscrimination is a separate norm, under which a member government agrees that any tariff applied to the exports of a given product from one trading partner will apply equally to the exports of that product from all other trading partners. Do the principles of reciprocity and nondiscrimination serve governments as simple rules of negotiation that promote efficiency, by "undoing" the terms-of-trade-driven inefficiency that arises in the absence of an agreement? My joint work with Bagwell suggests that they do.

Specifically, our work identifies at least two roles that these tandem principles can play in aiding governments as they seek to implement efficient trade agreements. First, these principles can help direct bargaining outcomes toward the tariffs that each government would have chosen had it ignored its ability to shift the costs of its protection onto trading partners through terms-of-trade effects (for example, if governments sought to maximize national income with their tariff choices, this would correspond to multilateral free trade). This feature in turn can encourage weaker countries to participate in GATT negotiations without fear of exploitation by their stronger, bigger trading partners. Second, these principles can help to protect the value of market access concessions won by a government in a current negotiation from being eroded in a future bilateral negotiation to which it is not a party. This feature can serve to facilitate the efficient exchange of market access concessions between countries, by reducing or eliminating the potential for opportunistic bilateral agreements in the future. Each of these roles arises from basic properties of reciprocity and MFN - namely, that mutual changes in trade policy conforming to reciprocity will stabilize the terms of trade, and that MFN tariffs will ensure that all countries can trade on the same terms (face the same set of exporter prices). Together, these properties serve to neutralize the terms-of-trade motives for trade policy intervention. Consequently, as we establish in our papers, when the fundamental inefficiencies that governments seek to correct with a trade agreement arise as a result of incentives to shift costs through terms-of-trade movements, these properties create a negotiating environment with the desirable features noted earlier. These papers point out that preferential agreements - which are by their nature discriminatory - can interfere with these desirable features, to the detriment of the multilateral trading system.

How Are Trade Agreements Enforced?

As there is no "world jail," an international agreement must be self-enforcing if it is to be credible, and GATT is no exception. As a result, GATT must (and does) attempt to specify credible retaliatory measures against any country that places additional restraints on trade in a way that violates the agreement. This amounts to maintaining a balance between the short-term temptation to deviate unilaterally from an agreed-on trade policy and the long-term penalty of a consequent future loss of cooperation (that is, the cost of a future retaliatory "trade war"). Viewed in this way, it is evident that any event that alters the current temptation to cheat or the value of maintaining cooperation into the future can alter this balance, and consequently the enforceable level of cooperation may change through time. In a number of papers Bagwell and I have built on this observation to evaluate and interpret several of GATT's features, including its escape-clause provisions and the impacts of preferential agreements on sustainable multilateral tariff cooperation.

In one paper I argue that enforcement concerns in a nonstationary world may help to explain the degree to which GATT's liberalization process has spread over time. I show that an initial round of liberalization can set in motion changes in an economy's resource allocation that eventually lead to a relaxation of enforcement constraints, creating the possibility of sustaining a further round of liberalization as the process continues toward the efficiency frontier. It is encouraging for the prospects of reciprocal trade liberalization in a world of limited international enforcement power that the liberalization process can gather momentum and sow the seeds of further liberalization. It suggests that such limits need not keep the world permanently away from the efficiency frontier. But there is a more ominous side to this observation: anything that interrupts the expected future progress of the liberalization process may negatively affect the ability to sustain the liberalization that has been achieved already. This notion embodies the core feature of what has come to be known informally as the "bicycle" theory of GATT liberalization ("If you don't keep pedaling, you will fall off"). It emerges naturally when GATT's liberalization is considered from the perspective of a self-enforcing agreement.

Can GATT's Principles Be Applied to "New" Trade Policy Issues?

GATT is certainly not perfect, but most observers agree that its principles have worked remarkably well to liberalize world trade. Can these principles also be applied to the variety of new issues before the WTO? A number of these issues, such as agreements on labor and environmental standards, would extend GATT's reach well beyond traditional trade policy matters, and as such appear to encroach on traditional limits of national sovereignty. This raises fundamental questions about the structure of international economic relations among sovereign states. Therefore, an important analytical question concerns the minimal range of policies over which international negotiations must proceed if global efficiency is to be achieved.

In recent work Bagwell and I have begun to explore such questions by investigating how domestic labor standards might be handled in the GATT/WTO. We consider several approaches to the treatment of domestic labor standards within a trade agreement. First we show that the "benign neglect" of labor standards within a trade agreement will result in inefficient choices for both trade barriers and labor standards, much as labor interests and social activists have claimed. However, we also show that direct negotiations over labor standards are not required to reach efficient outcomes. Instead, we describe GATT rule changes that in principle could allow governments to achieve efficient policy outcomes while continuing to negotiate over tariffs alone, thereby preserving a degree of national sovereignty over traditionally domestic policy choices. The required rule changes would extend the logic of reciprocity as currently contained in GATT to the choice of domestic labor standards. But there is an important distinction between the GATT rule changes we suggest and the changes that have been proposed in recent WTO discussions - namely, the formal inclusion of a "social clause" that would permit restrictions to be placed on imports from countries not complying with a specified list of minimum standards. These proposed changes would allow governments to raise import restrictions in response to the weak labor standards of their trading partners. In contrast, the changes suggested by our analysis would instead allow governments to raise import restrictions in exchange for tightening their own labor standards. This reorientation, linking the permissible level of import protection in GATT to one's own labor standards rather than the labor standards of one's trading partners, is the paper's central message for the application of GATT's principles to the new trade policy issues.

Endnotes

1 R. W. Staiger and G. Tabellini, "Discretionary Trade Policy and Excessive Protection," American Economic Review 77 (5), (December 1987), pp. 832-37; and R. W. Staiger and G. Tabellini, "Do GATT Rules Help Governments Make Domestic Commitments?," Economics and Politics, forthcoming, July 1999.

2 K. Bagwell and R. W. Staiger, "Reciprocal Trade Liberalization," NBER Working Paper No. 5488, March 1996.

3 H. G. Johnson, "Optimum Tariffs and Retaliation," Review of Economic Studies 1 (2), (1953-4), pp. 142-53.

4 Authoritative references on GATT rules and procedures include J. J. Jackson, World Trade and the Law of GATT, New York: Bobbs-Merrill Co., 1969; and K. W. Dam, The GATT: Law and International Economic Organization, Chicago: University of Chicago Press, 1970.

5 K. Bagwell and R. W. Staiger, "An Economic Theory of GATT," NBER Working Paper No. 6049, May 1997, forthcoming in American Economic Review; "Reciprocity, Non-discrimination and Preferential Agreements in the Multilateral Trading System," NBER Working Paper No. 5932, February 1997; and "Multilateral Trade Negotiations, Bilateral Opportunism and the Rules of GATT," unpublished paper, January 1999.

6 K. Bagwell and R. W. Staiger, "A Theory of Managed Trade," American Economic Review 8 (4), (September 1990), pp. 779-95; "Protection and the Business Cycle," NBER Working Paper No. 5168, July 1995; "Multilateral Tariff Cooperation during the Formation of Customs Unions," Journal of International Economics 42 (1-2), (February 1997), pp. 91-123; "Multilateral Tariff Cooperation during the Formation of Free Trade Areas," International Economic Review 38 (2), (May 1997), pp. 291-319; "Regionalism and Multilateral Tariff Cooperation," NBER Working Paper No. 5921, February 1997, published in International Trade Policy and the Pacific Rim, John Piggott and Alan Woodland, eds. London: Macmillan, 1998.

7 R.W. Staiger, "A Theory of Gradual Trade Liberalization," in New Directions in Trade Theory, A. Deardorff, J. Levinsohn, and R. Stern, eds. Ann Arbor: University of Michigan Press, 1995.

8 K. Bagwell and R. W. Staiger, "The Simple Economics of Labor Standards and the GATT," NBER Working Paper No. 6604, June 1998; published in Social Dimensions of U.S. Trade Policies, A. V. Deardorff and R. M. Stern, eds. Ann Arbor: University of Michigan Press, forthcoming.