A Political Economy Model of GATT/WTO Trade Complaints

The paper seeks to explain patterns of formal trade complaints brought to the GATT/WTO by member governments. Formal complaints are significant because they have distributional implications both at the national level, among domestic producers, and at the international level, between nations. Domestically, a successfully litigated complaint can produce highly excludable benefits for a particular sector. Internationally, complaints are the primary method of enforcing GATT/WTO law, as they provide a stimulus for targeted states to bring non-conforming trade policies into consistency with the rules of an agreement. But not all infractions of rules by states become the subject of a complaint. The selection of complaints for litigation out of a universe possible complaints is subject to a complex political process of demand by private producers and supply by governments. The upshot of this process is that the rules of multilateral trade agreements are unevenly enforced. The largest countries in world trade may be de facto more constrained by international trade law across sectors than the smaller and especially the smallest countries, contrary to the expectations of most theories of international relations.

An earlier version of this essay was presented at the annual meeting of the American Political Science Association, San Francisco, California, 29 August - 1 September 1996. I wish to thank James Alt, Robert Keohane, Pieter Jan Kuijper, Lisa Martin, Andy Moravcsik, John Odell, and Joseph Weiler for their helpful comments and suggestions. I am also grateful to the trade representatives in Washington, D.C., Brussels, and Geneva and to the WTO Secretariat officials who were so generous with their time and insight in anonymous interviews conducted between August 1995 and June 1996. The views expressed in this paper are those of the author.

Since international regimes themselves generally do not provide for hierarchical enforcement of the rules and principles which they embody, enforcement in the international system is decentralized to its constituent states.[1] The General Agreement on Tariffs and Trade (GATT), and its successor institution, the World Trade Organization (WTO), are no exceptions. The WTO secretariat, consisting of a body of independent international civil servants, has no standing to bring proceedings against countries which infringe the rules of the agreements, and no independent sanctioning mechanism. Rather, it is governments who enforce GATT/WTO law, primarily through bringing complaints against other governments in formal dispute settlement.

Multilateral trade complaints are significant because they have distributional consequences both domestically and internationally. At the domestic level, a successfully litigated complaint can produce highly excludable benefits for a particular industry or sector. For example, U.S. leather exporters obtained improved market access to Japan, when that country removed quantitative restrictions following a GATT panel ruling in 1983 supporting a U.S. complaint.[2]

At the multilateral level complaints perform two main functions. Firstly, a country's compliance with the agreements is only at issue when a trade-related complaint is raised by another contracting party to the agreement. Complaints are thus a signaling device which indicate both the identity of the alleged transgressor (i.e. the state who is the defendant), and the nature of the transgression. They are a readily observable, albeit probably biased, estimator of countries' compliance with the standing rules of an agreement.

Secondly, complaints perform an enforcement function for the multilateral trading system, by providing a stimulus for targeted states to bring non-conforming policies into consistency with the rules of an agreement. For example, the U.S. had to amend its customs legislation in 1990 following a complaint by the European Community and Canada that the legislation violated GATT rules.[3] In this capacity, complaints have important distributional implications regarding the burden of compliance with international trade agreements ex post, since they determine which of the signatories are required to adjust their policies in light of specific interpretations of written rules.

In this essay I will focus on the multilateral function of complaints. I seek to explain and predict variation in the type, frequency and distribution of GATT/WTO complaints brought by governments against one another during the implementation stage of the agreement. I argue that decentralized enforcement of treaty obligations by self-interested states has important and systematic consequences for signatories' adherence to the rule-based trading system. Not all infractions of rules by states become the subject of a complaint. Rather, the selection of complaints for litigation out of a universe possible complaints is subject to a complex political process of demand by private producers and supply by governments. The upshot of this process is that the rules of international trade agreements are unevenly enforced. I argue that the largest countries in world trade may be de facto more constrained by their GATT/WTO obligations across sectors than the smaller and especially the smallest countries, contrary to the expectations of most theories of international relations.

I will begin with a brief discussion of complaints and the role they play in dispute settlement under the provisions of the old GATT and the new WTO. The discussion that follows will reveal why the frequency and distribution of complaints is a biased indicator of countries' true compliance with GATT/WTO obligations, and will suggest a typology of complaints. A "two-level" political economy model for the discovering and raising complaints in the multilateral forum will be presented, and some hypotheses drawn from the model will be tested against a data set of approximately 300 multilateral trade complaints, initiated between 1948 and 1996.

Why complaints are important

A brief review of GATT/WTO dispute settlement

The primary method of enforcing GATT/WTO law is the dispute settlement mechanism, triggered by a formal complaint from a contracting party. As one seasoned observer of the GATT has argued, "it is often the case that a GATT rule which is not being observed voluntarily cannot be enforced unless there is a formal complaint from a member country."[4] Another commentator concurs that "enforcement in GATT circles is primarily equated with dispute settlement."[5]

The dispute settlement mechanism is based on Articles XXII and XXIII of the General Agreement. A request for consultations is the first step in the dispute-settlement process. A party can request consultations with any other party "with respect to any matter affecting the operation" of the Agreement.[6] If bilateral consultations fail to resolve a dispute, the aggrieved party may lodge a formal complaint if it believes that a benefit accruing to it is being nullified or impaired as a result of

a) the failure of another contracting party to carry out is obligations under the Agreement, or

b) the application by another contracting party of any measure, whether or not if conflicts with the provisions of the Agreement, or

c) the existence of any other situation.[7]

The vast majority of formal complaints allege a legal violation (LV) of the Agreement (paragraph (a)), although some important cases have also involved non-violation nullification and impairment (NV N&I) complaints (paragraph (b)). Article XXIII:2 provides that the contracting parties as a collective body promptly investigate the matter, and make appropriate recommendations and rulings. Article XXIII does not specify a procedure for the consideration of complaints, but the GATT has traditionally relied upon a panel of arbitrators to issue a report which the contracting parties could then accept or reject by consensus.

In the GATT's early history, the panel procedure was quite informal. Disputes were simply referred to the chair of the plenary meeting of the contracting parties, who issued a ruling to the parties. As the volume of complaints grew, the practice of referring a dispute to a "working party" composed of the disputants and other interested parties developed. From 1952 onward the procedure became more formalized, with disputes being referred to a panel of independent experts acting in their own capacities, and not as representatives of the member states. These panels issued rulings based on a legal interpretation and examination of evidence submitted by the disputants, and could make policy recommendations about what steps, if any, should be taken to bring the offending policies into compliance with GATT rules, or what compensatory measures would be sufficient to make restitution to the complainant. Under the old GATT system, any state, including the defendant, could exercise a veto power over the decision to establish a panel to hear a case as well as the decision to adopt a panel report.

The 1994 Dispute Settlement Understanding in the WTO Agreement establishes procedures for dispute settlement under a new entity, the Dispute Settlement Body (DSB) which is composed of representatives of all the member states. In contrast to the previous system, strict time limits for consultations have been set,[8] and all consultations must be notified to the DSB.[9] When consultations fail to resolve a dispute, a complaining party has a right to the establishment of a panel to adjudicate its case, barring a consensus by the DSB not to establish it[10]. Unless a party to the dispute formally notifies the DSB that it wishes to appeal, adoption of panel reports is virtually automatic, since a "reverse" consensus by the DSB to not adopt the report is required for rejection.[11] In the case of an appeal, a standing Appellate Body will review the case and issue its own report, which is to be adopted by the DSB unless, once again, there is a consensus not to adopt.[12]

If a panel report finds a complaint to be justified, it will generally recommend that the measure in violation be withdrawn.[13] After a report is adopted, the DSB is charged with surveilling the implementation of its recommendations and rulings.[14] The Understanding further provides that compensation from the losing party or suspension of concessions by the plaintiff are "temporary measures available in the event that the recommendations and rulings are not implemented within a reasonable period of time", but specifies that "neither compensation nor the suspension of concessions or other obligations is preferred to full implementation of a recommendation to bring a measure into conformity with the covered agreements."[15] All of these reforms are intended to create a more legalistic and binding dispute settlement mechanism in the WTO, as compared to its predecessors under the GATT.

Empirical work on dispute settlement

GATT dispute settlement has been the subject of an enormous amount of international legal scholarship, but with few attempts to subject theoretically-driven hypotheses to empirical testing. Many of these studies provide descriptions of cases which have been litigated, legal analysis and interpretation of panel rulings.[16] Explanations of observed patterns of complaints, if any, are generally post hoc, and prediction of variation is not the goal of these studies. Nevertheless, legal scholarship in this area has yielded some very important empirical findings which justify further social scientific inquiry into the process of complaint-raising by governments.

In particular, Robert Hudec has compiled a comprehensive data set of formal GATT complaints which has been underexploited by political scientists. Hudec's evaluation of the outcomes of dispute settlement cases under the GATT demonstrates that legal systems can have a constraining effect on the actions of governments.[17] Dispute settlement under the GATT is generally noted more for its shortcomings than its strengths: as mentioned above, the defendant could block the establishment of a panel or veto the adoption of panel reports by the contracting parties. Even under the previous system, however, being the subject of an unfavorable panel ruling very often meant policy adjustment by the losing state in a manner favorable to the complainant. Of the 207 complaints launched between the years 1948 and 1989, Hudec finds that among cases where there was a legal ruling of violation against the defendant, 89% of all cases (65 out of 74) resulted in "full satisfaction" or "partial satisfaction" of the legal claim. Full satisfaction is defined as "cases where the legal claim has been fully (or almost fully) vindicated, usually by removing a measure found to be in violation of GATT."[18] Partial satisfaction consists of "cases in which some significant remedial action was taken by the defendant, short of complete removal but sufficient to justify treating the legal claim as having been successful."[19] Moreover, Hudec counts 64 additional cases of complaint without a panel ruling but where the defendant settled or conceded the validity of the legal claim. 100% of these cases resulted in full or partial satisfaction. In only 12% of cases does Hudec find a negative outcome, defined as "cases in which the legal system has failed to enforce a valid claim", either due to no remedial action taken at all by the defendant or the imposition of a trade restriction.[20]

Furthermore, the relatively high incidence of policy adjustment by the defendant cannot simply be attributed to endogeneity of treaty selection or endogeneity of complaints in favor of "easy" cases -- that is to say, strategic actors do not only litigate cases where change is unproblematic or non-costly to the defendant.[21] On the contrary, many of these cases are plausibly ones where remedial action would not have been taken by governments in the absence of a complaint. Examples of such cases which resulted in high or partial policy change by the defendant include: the removal of QRs in sensitive products by France;[22] termination of an exchange rate guarantee for Airbus Industrie by Germany;[23] amendment of Superfund legislation by the U.S. to remove a discriminatory tax on foreign petroleum imports;[24] removal of discriminatory taxes on foreign alcoholic beverages by Japan;[25] adjustment of the calculation of the Value Added Tax to lower the threshold for foreign bidding on government contracts in the European Community[26]; Canada's repeal of the Foreign Investment Review Act, which required foreign investors to favor domestic suppliers over imports.[27]

Even among the more infamous cases of the 1980s where a GATT panel ruling was blocked by the defendant, some policy change favorable to the plaintiff was eventually achieved. The EC, for example, reduced the amount of subsidy to producers of pasta and processed fruit subsequent to complaints launched by the U.S. in 1982.[28] In these cases, the act of blocking itself indicates that the defendant would have been unlikely to change the status quo policy in the counterfactual situation of no complaint. Although it is probable that the threat or act of retaliation (large tariff increases by the United States on certain EC goods) was the main inducement for the defendant to change its behavior in these cases, transgressions must first be identified before sanctions can be deployed. Awareness of rule violations via complaints must be prior to any determination of adequate punishment levels for defection.

Thus, existing empirical work suggests that complaints are significant because countries which are targeted very often have to adjust their policies in light of particular interpretations of standing GATT/WTO rules. Even in the absence of an authoritative panel ruling, a request for consultations means that the targeted state's policies are subject to at least a bilateral peer review and discussion; reaching a "settlement" prior to the request for a panel often involves some mutual policy accommodation or unilateral adjustment which palliates the complainant. Extending Hudec's work to include cases up to 1994, Reinhardt finds at least "partial cooperation" -- defined as partial removal of the policies in dispute or relaxation of trade barriers by the defendant or by both states -- in 103 of 148 (69.6%) of cases where there is no panel ruling. He argues that "the mere establishment of a panel acts as a signal of complainant resolve and encourages defendants to settle early."[29] The potential advantages to the defendant of early settlement include the avoidance of legal precedent, which other exporting states could invoke to their advantage, as well the fact than an early settlement may be discriminatory, whereas a post-ruling settlement, if any, may not.[30]

Under the new provisions of the WTO, countries have an even greater incentive to reach a mutually agreeable solution prior to reaching the panel stage.[31] The virtual guarantee of access to a panel and adoption of the report makes formal complaints a useful tool for achieving some kind of policy modification in the target state, with uncertainty about the outcome of litigation driving an early settlement. For example, the European Community recently modified its import regime for apples in an agreement with Chile, "essentially to forestall a probable negative outcome of a...Chilean panel procedure against the Community regime on the importation of apples."[32]

In this essay, I argue that who raises complaints against whom and which complaints are selected for litigation out of a potential universe of complaints has great significance for the operation of an agreement ex post. Because the primary objective of dispute settlement is to resolve trade issues which arise in the bilateral relations between states subject to the GATT/WTO code, and the WTO secretariat itself has no standing to bring proceedings against countries which infringe the rules, GATT/WTO rules and norms are only consequentially enforced as a result of formal complaints. The practical implication of this observation is that international trade rules are selectively enforced according to the interests of the countries initiating complaints. In other words, not all possible infractions or policies which are perceived as being at odds with rules are litigated and hence subjected to a corrective process. Given a certain level of non-compliance by all parties to the agreement -- that is, given the existence of trade-related policies whose consistency with the rules could be challenged -- the probability of observing a complaint varies. The probability is greatest for the largest trading nations in terms of world imports and exports, and least for the smallest trading nations, due to the political economic incentives associated with decentralized enforcement.

The Sources of complaints: Police patrols vs. fire-alarm oversight in the GATT/WTO

There are two main forms of surveillance which parties can use to determine whether and in what way states' behavior violates the goals of the treaty: centralized "police patrols" and decentralized "fire alarms".[33] In an analogous manner to Congressional oversight of Executive Branch activity in American politics, signatories to an international agreement may engage in centralized and routinized monitoring of their partners' policies, with the purpose of detecting and addressing existing treaty violations, as well as pre-empting new violations. Alternatively, parties to the agreement can establish a set of rules, procedures, and informal practices which allow decentralized private interests or even other states to alert the governments concerned to possible breaches of obligation.

In the GATT/WTO system, police patrol surveillance takes the form of regular reporting requirements to the General Council of contracting parties as well as standing committees comprised of representatives of the member states and serviced by the WTO Secretariat. These committees meet several times a year for the purpose of affording contracting parties the opportunity to consult on any matters relating to GATT/WTO law.[34] By submitting relevant legislation for examination and meeting other notification requirements, signatories create a centralized source of information which permits them to monitor their partners' policies and be monitored in turn. Any discrepancies found can be aired in the multilateral forum by voicing questions of clarification, concerns, or complaints. With the agreement establishing the WTO in 1995, various changes have occurred with respect to the organizational structure of the committees, notably the creation of overarching councils on goods, services, and TRIPs (see Figure 1). Nevertheless, the basic function of the committee process in facilitating the operation of the agreement remains the same.

Fire alarm oversight at the domestic level is accomplished by creating channels of access for private producers to bring GATT/WTO-related complaints to the attention of their governments. For example, many disputes initiated by the U.S. in the GATT forum began as private Section 301 complaints under U.S. law. Similarly, the EC's new Trade Barriers Instrument, an extension of the New Commercial Policy Instrument of 1984, [35] is designed to enable Community industries and member states to alert the Commission and Council to foreign "illicit practices" in trade and to activate Community bodies for the purpose of WTO dispute settlement.

At the multilateral level, other states may raise fire alarms by alerting a government with a shared interest to a third parties' possible breach of the agreement. For example, Cairns group countries committed to achieving freer trade in agriculture may find it useful to alert each other and build case-specific coalitions against restrictive agricultural practices by another contracting party.

Thus, police patrol and fire alarm oversight at the international and national levels can be complementary strategies for achieving treaty implementation. Both result in the production of complaints -- allegations of violation or nullification and impairment of benefits -- to be addressed in the multilateral forum.

The Dependent Variable

Formal legal complaints

Following Robert Hudec, formal legal complaints will be defined as

those events in GATT's history that contained the basic elements of an Article

XXIII

lawsuit, that is (1) a complaint from one or more individual plaintiffs, (2)

directed against actions of a named defendant or defendants, (3) based on a claim

that the defendant's acts affect the GATT legal rights of the complainant, and (4)

looking eventually to a ruling or decision by the Contracting Parties or the GATT

Council indicating those rights. The

same elements constitute adjudication under

the MTN Codes.[36]

Hudec's data set of 207 GATT disputes, involving 47 developed and developing countries and spanning from 1948-1989, includes not only those disputes which resulted in a panel ruling, but also those settled or otherwise dispatched during the formal consultation stage.[37] To this will be added a data set of formal GATT complaints between 1990 and 1994,[38] and complaints issued since the founding of the WTO in January 1995 up through April 1996.[39]

Particularistic and Systemic Complaints

As mentioned earlier, the GATT/WTO system distinguishes between charges of legal violation and non-violation nullification and impairment. Among these it is instructive to further enumerate two categories of complaints: what shall be termed particularistic versus systemic complaints. Particularistic complaints are ones where the trading interests of a particular producer or concentrated group of producers within a country or across countries are damaged or immediately threatened by the practices of a third member state. Particularistic complaints may be either LV or NV N&I. Systemic complaints, by contrast, are ones where the trading interests of a particular producer or group of producers in the complaining country are not significantly damaged or immediately threatened (see Fig. 1).

Click here for Picture

Box I cases are those where a legal violation by a GATT/WTO contracting party produces concentrated damage to either the export or import interests of the complaining country, leading the latter to voice a particularistic complaint. In Box III, an action that is not inconsistent with GATT obligations nevertheless results in injury to trading benefits expected by another country, and a particularistic complaint on behalf of the harmed domestic producers is lodged. One celebrated example of a Box III complaint is the second U.S. Oilseeds case against the European Community, wherein the U.S. claimed that an EC processing subsidy that was not inconsistent with the GATT operated to undermine an earlier tariff concession on oilseeds, thus harming the exports of American soybean producers.[40] Box II cases are ones where an alleged legal violation by a country does not produce concentrated or immediate damage to partners' trading interests, yet a party who observes the infraction is still willing to draw the members' attention to a possible violation of the rules. Systemic complaints may be thought of as "principled" complaints, primarily for the purpose of ensuring that rules are observed independent of immediate, short-term trade interests. However, ensuring that rules are respected in a particular issue-area may coincide with a party's perception of its long-term trade interests, so systemic complaints are also instrumental in the sense that they are not divorced from cost-benefit calculations. Box IV is an empty set, since a claim of nullification and impairment indicates that a particular interest was harmed.

I argue that this distinction between particularistic and systemic complaints is important, as it carries implications for the operation of the GATT/WTO system and the distribution of the burden of compliance among all members. Systemic complaints may be likened to public goods from which all members would benefit and no member could be excluded, as they serve to protect the integrity of the rule-based system. By drawing the attention of parties to possible non-compliance, they raise the costs to the transgressor of continuing the disputed activity and act as a stimulus for policy adjustment. Excluding situations where a violation of international obligations by another country would discriminate in favor of country A, country A should prefer compliance by all partners in all situations to enhance predictability, stability, and to facilitate the expected flow of benefits arising from the agreement. But as the political economy model of complaints set forth below will suggest, systemic complaints in the GATT forum are likely to be underprovided, and subject to free-ridership.

Decentralized enforcement in the GATT/WTO: A two-level political economy model of complaints

The decision to bring or not to bring a complaint to the GATT/WTO forum is one aspect of a nation's trade policy with distributional implications. The international political economy literature has generally fostered a dichotomy between societal and statist approaches to explaining trade policy at the domestic level.[41] A "societal" view argues that government policy mirrors the preferences of domestic lobbies, interest groups, or other actors in society. These preferences arise from the anticipated effect of policies on the economic return to groups or individuals. Governments are responsive to the most vocal, well-organized and well-funded constituents, and hence the state is largely a tool of private interest.[42] By contrast, a "statist" perspective claims that policymakers represent a logic of their own, and that they are relatively insulated and autonomous in making decisions. The multiplicity of demands and complaints issuing from society must be molded into a coherent national policy by officials in positions of political power. Independent state preferences arise from either national or aggregate objectives, such as full employment and economic growth, bureaucratic interests, or responsiveness to the constraints and opportunities of the international system and strategic interactions with other states.[43]

This section argues that societal and statist perspectives are complementary rather than competing approaches to understanding the production of trade policy. Specifically, the political market for GATT/WTO complaints requires both a demand for complaints as well as their supply by governments. On the demand-side, only a subset of the population of potential GATT/WTO complaints are brought to the attention of governments each year. Potential complaints become known through multiple sources, the most important being private producers. Export-oriented firms encountering barriers to their products in markets abroad or import-competing firms or importers of intermediate goods experiencing difficulty due to the practices of foreign governments have several options: appealing to their home governments for assistance, resolving the conflict through private representations or arbitration, or doing nothing. If the trade problem the producer is experiencing is related to imports entering the domestic market, there are multiple possibilities for seeking government intervention that do not involve a GATT/WTO complaint: petitions for countervailing and anti-dumping duties, for example, are significant forms of "administered protection" which can provide relief from foreign predatory practices without requiring foreign actors to change their behavior.[44] Domestic petitioners can exercise much more influence over policy outcomes in the domestic arena than in the international one. Thus, one can predict that damage to trading interests due to import competition will be addressed more often in the domestic arena, and hence is less likely to appear as a GATT/WTO complaint.

Conversely, when foreign practices impede the exportation of domestically-produced goods (i.e., there are barriers to imports in the foreign market or in third markets), firms have little choice but to appeal to their governments for a solution vis-‡-vis foreign governments. This is because behavior change is often required from foreign actors in order to remedy the injury.

Export-oriented firms also make a cost-benefit calculation of pursuing alternative strategies to redress their situation, and I hypothesize that private complaints brought to the attention of governments will target those areas where the greatest amount of trade is at stake in terms of value and volume, such that the potential benefits of government-contacting outweigh the costs to the industry or firm of organizing, gathering relevant information, legal fees, and so forth. Although it is certainly true that small country markets may constitute a large proportion of an industry or firm's trade, on average the largest countries in terms of exports and imports will constitute the most important markets across sectors. By contrast, legal violations by small countries (mostly less-developed) are likely to be underreported, since the trade damage associated with an infraction of GATT rules by these countries is likely to be less.

There are also supply-side factors which reinforce the tendency to observe more complaints against the large than the small. At the domestic level, governments should prefer a "fire-alarm" method of oversight of partner's trade policies -- that is, creating rules and procedures which allow decentralized interests to alert them to possible breaches of rules -- as opposed to police patrol monitoring, which is centralized and routinized monitoring of partners' policies with the purpose of detecting and addressing rule violations. This is because elected representatives and executive officials have a self-interest -- in terms of votes, contributions, information, organizational support, etc. -- in serving vocal constituencies. By addressing concrete violations which potential supporters have complained about, the government can take credit for redressing grievances and thus build support for its policies. A second reason why governments should favor a more decentralized fire alarm policy of GATT/WTO oversight is that domestic firms and associations bear the costs of discovering complaints, rather than legislative or executive agencies.

By contrast, under a police patrol method governments spend time remedying GATT violations which harm no domestic supporters and thus yield fewer political benefits. Furthermore, where governments do engage in police patrol oversight, they must be pragmatic about the allocation of scarce bureaucratic and financial resources devoted to surveilling partners' trade policies. Given finite resources and the declining marginal utility of monitoring where bilateral trade as a proportion of total trade is less, developing countries and smaller trading partners will receive less scrutiny. Even for countries with relatively large bureaucratic resources such as the U.S., a unit of time spent monitoring the EC or Japan is likely to yield more political benefits in terms of the trade value of a change in policy than a unit of time spent monitoring Costa Rica or Nigeria. Thus, violations by small states will go undiscovered, and the number of complaints which might otherwise be incurred will be understated.

Governments' preferred method of oversight, I argue, also has consequences for the type of GATT/WTO complaint pursued. I have defined particularistic complaints as ones where the trading interests of a specific producer or group of producers within a country are damaged or immediately threatened by a the practices of another member state. Systemic complaints, on the other hand, are ones where specific producer interests in the complaining country are not immediately threatened. As McCubbins and Schwartz have observed, "fire-alarm oversight tends to be particularistic...it arguably emphasizes the interests of individuals and interest groups more than the public at large [emphasis added]."[45] If governments do indeed prefer a fire-alarm method of oversight which is sensitive to domestic concerns, an observable implication of this would be more particularistic relative to systemic GATT/WTO complaints.

However, the actual supply of complaints in the multilateral forum is not simply a mirror of societal demands, as the potential number of complaints is vast in comparison with their actual production. In the first forty-odd years of the GATT's operation, Hudec identifies only 207 formal disputes. Since only states have access to the multilateral forum and dispute settlement procedures, they can exercise a considerable degree of discretion in voicing complaints which come to their attention. What other incentives, then, do governments face in bringing complaints to the GATT/WTO?

Firstly, societal actors are not the only source of information about potential complaints to which governments respond. Other sources of information about complaints are transnational or international in character, to which policymakers are uniquely sensitive. Embassies and missions keeping track of political and economic developments in the foreign country of residence may become aware of activities which infringe multilateral obligations and inform the home government.[46] Print and other media may also alert governments to inconsistent practices by their partners.[47] GATT/WTO-level police patrol monitoring, discussed above, is a key fount of potential complaints. Finally, the GATT/WTO Secretariat or other governments may call attention to an infringement through informal channels or contacts.[48]

Bringing complaints up in the GATT/WTO forum is a multi-stage process of state-to state consultations and discussions outside and inside the GATT forum, with progressively increasing costs at each stage. Costs include administrative resources of the trade ministry, potential political costs by opposing interests at home or abroad, as well as opportunity costs of pursuing other complaints. Governments are acutely aware of these costs and hence are selective in the complaints which they pursue. Policymakers are also familiar with their own country's policies and implementation with respect to their trade obligations, and may be reluctant to raise complaints with partners on issues where their own government's practices are vulnerable to challenge.[49] Hence, governments act as an important filter of societal demands for GATT/WTO complaints.

Executives are also uniquely sensitive to diplomatic and strategic considerations, and thus may refrain from bringing complaints when there are overriding political considerations. I argue that large countries were historically reluctant to bring complaints against smaller developing countries or exert significant pressure on them to liberalize their markets, due to security concerns about political stability in developing regions in the context of the Cold War. This is not to say that bilateral pressure was not used occasionally to open up specific outlets, and it is certainly plausible that pressure applied in consultations prior to the dispute settlement stage (for example, the threatened withdrawal of trade concessions or a cut in aid) would preempt the need for a formal GATT complaint against a small developing country. But supporting evidence, which I present in the next section, suggests that the industrialized countries refrained from using coercive tactics prior to the dispute-settlement stage in a systematic fashion to hold LDCs to the GATT obligations which they did undertake. Indeed, during the postwar era, the developing countries of Asia, Africa, and Latin America had a wide degree of latitude to engage in highly trade-restrictive import substitution policies.

If large industrialized countries historically refrained from prosecuting LDCs, due to the relatively small size of individual country markets coupled with overriding security concerns, it follows logically that formal complaints launched by large states in the GATT/WTO would mostly target other large states. In the first place, this is because more private complaints will be demanded against the countries which account for the greatest proportion of world trade, as argued above. On the supply side, large developed states generally have highly competent bureaucracies and the legal expertise necessary to implement trade obligations, and hence multilateral police patrol efforts are less likely to reveal inadvertent bureaucratic snafus or simple legislative transposition errors which can account for non-compliance and are easily remediable. Rather, trade barriers are more likely to be cleverly designed measures which "bend the rules without the breaking them" and hence serve important political constituencies while nominally obeying the letter of the GATT/WTO, if not the spirit. Thus, large-against-large state disputes are more likely to surface as formal GATT/WTO complaints.

In sum, I draw the following hypotheses from the foregoing two-level political economy model of complaint-raising:

On the proportion of particularistic to systemic complaints:

Hypothesis 1: One should observe a greater number of particularistic relative to systemic complaints in the GATT/WTO forum.

On the proportion of export-related to import-related complaints:

Hypothesis 2: From the point of view of the complaining country, one should observe more GATT/WTO complaints related to export trading interests (market access or competition in third markets) than import trading interests (primarily injury due to competition in the domestic market). Or in other words, more complaints should target the import policies of the defendant country rather than its export policies.

Corollary 2 a): Complaints will target the export policies of the defendant when

such

policies have significant enough impact in third markets to adversely

affect the export interests of the complaining country.

On the proportion of complaints targeting large countries vs. medium and small-size

countries:

Hypothesis 3: One should observe more GATT/WTO complaints against countries accounting for the greatest proportion of international trade (from all parties), and conversely, fewer complaints against smaller countries (from all parties).

Hypothesis 4: Formal complaints by large states will mostly target other large states.

These are the four hypotheses about formal GATT/WTO complaints which will be empirically evaluated in the following section. At this point it may be helpful to visualize the preceding discussion of the progressive filtering of complaints as a "pyramid model," with the population of potential complaints forming the base and formal disputes with panel rulings issued forming the top (see Fig. 2)

Fig. 2. Pyramid model of GATT/WTO complaints

Inside GATT/WTO Forum

-------------------------------------------------------------------------------------------------------------

Outside GATT/WTO Forum

The Results: Formal Complaints in the GATT/WTO, 1948-1996

The results are set forth in tabular form and are broken down by decade, where relevant, to permit the analysis of trends over time.[50] Each bilateral trade dispute is counted as a single observation, so, for example, if five countries bring a suit against Country A for a particular policy or bundle of policies, that is counted as five disputes. In this way the frequency of complaints and the identity of disputants are more fully representative, since many developing countries tend to bring suits as a group rather than individually. I also present separately complaints brought under the GATT system, from 1948 to 1994, and those litigated under the new WTO from 1 January 1995 - 1 April 1996.

Tables 1(a) and 1(b) indicate the number of systemic relative to particularistic complaints over time. Hypothesis 1 predicted that one should observe a greater number of particularistic relative to systemic complaints in the GATT/WTO forum, due to governments' preference for a fire-alarm oversight policy whereby domestic constituents contact their representatives about sector-specific short-term trade difficulties, and indeed this is the case. Particularistic legal violation and non-violation nullification and impairment complaints together account for 94% of complaints in the 1950s, increasing each decade to reach 100% of complaints by the 1990s and under the new WTO.

Table 1(a). Types of GATT complaints, 1948-1994

Table 1(b). Types of WTO complaints, January 1995 - April 1996

I count only 6 systemic complaints in the entire data set. There are 3 cases in the 1950s, all involving U.S. complaints against France for statistical and customs stamp taxes applying to imports from all countries, not just particular sectors or products. The U.S. claimed that the taxes, ranging from 0.1% to 0.4% ad valorem, were a legal violation of France's GATT Article II tariff bindings. France conceded the violation, and the taxes were eventually abolished. Hudec writes that the complaints involved "an issue of principle more than commercial impact",[51] since the charges were minimal and distributed across all sectors.

A similar case in 1969 involved a U.S. complaint against Italy for administrative and statistical fees. The service fees applied to all import entries, and the U.S. argued that the charges constituted a legal violation of Article II tariff bindings and the Article III:2 requirement of national treatment for imports. Hudec notes that as the fee applied to all products, no particular trade interest was evident. Italy soon enacted legislation withdrawing the fees.[52]

In 1970 the U.S. brought a complaint against Jamaica for exceeding margins of preference permitted by Article I:4 in its preferential tariff for Commonwealth countries. According to case history, the complaint "appeared to be motivated primarily by concern to establish legal principle against increasing margins of preference",[53] rather than serving particular trading interests. And finally, in 1981 the U.S. lodged a formal complaint against the EC for export subsidies maintained by Greece which allegedly violated Article 9 of the Subsidies Code. The complaint had no specific focus and appears to have been triggered by the examination of trade policies that took place upon Greek accession to the EC. However, no American industries were supporting the complaint and the U.S. eventually abandoned the case due to the lack of specific industry interest.[54]

The fact that the U.S. was the sole provider of systemic complaints and that the proportion of systemic complaints declined over time, from 5% of all complaints in the 50s and 60s to 3% in the 70s, 1% in the 80s, and most recently 0% in the 1990s and under the new WTO, is consistent with the predictions of hegemonic stability theory. As applied to international political economy, this theory predicts that the largest state in the system, i.e. the one with a preponderance of resources, such as a large market for imports, will have a self-interest in providing collective goods which tend to be underprovided due to free-ridership. The making and enforcing of formal and informal trade rules are among these collective goods. Yet as the resources available to the hegemon decline-- due to the dissipation of technological advantage, the shifting of competitive advantage to low-cost producers in the periphery, and so forth-- the hegemon is less willing and able to act in the common interest. And indeed, the willingness of the United States to provide formal systemic complaints to enforce multilateral trade rules has declined over the postwar period.[55]

However, this does not mean that systemic enforcement no longer takes place in the multilateral system. Formal dispute settlement is not the only arena for systemic complaining available to the contracting parties to the GATT/WTO. As suggested above, lodging a formal complaint and then going through the panel process is a costly and time-consuming prospect. Countries with an interest in principled enforcement have less costly options for bringing pressure to bear on parties that deviate from the rules. Venues such as the GATT/WTO's Trade Policy Review Mechanism (TPRM), whereby each member's trade policies are reviewed at regular intervals, and routine committee meetings are appropriate fora for countries to raise systemic informal complaints. Small export-oriented developed countries who depend on a rule-based trading system, such as Australia and the Nordic states, have developed a reputation for performing this type of enforcement role in the GATT/WTO.[56]

Another hypothesis about selection of complaints according to type has to do with the relative mix of import-related to export-related complaints. Hypothesis 2 predicted that from the point of view of the complaining country, one should observe more GATT/WTO complaints related to export trading interests (market access abroad) than import trading interests (primarily injury due to competition in the domestic market). Or in other words, more complaints should target the import policies of the defendant country rather than its export policies. This is due to the fact that private petitioners and governments have a choice of forums for resolving difficulties encountered by domestic producers. If the issue is injurious competition with domestic products, administered protection is a remedy which does not require foreign actors to change their behavior, so many of these complaints will not show up in the GATT/WTO fora.. But if injury is experienced due to market access barriers, the only recourse is an appeal to foreign governments to find a solution.

What we observe in Tables 2(a) and 2(b) supports these expectations. Over the four decades of the GATT's operation, 84% of all complaints (249 total) have targeted the import policies of the defendant, whereas 14% (40 total) were generated by foreign export policies. Under the WTO, we see a similar distribution, with 6 of 42 complaints or 14% targeting export policies.

Table 2(a). GATT import-related complaints (against defendant) vs. export-related complaints (against defendant), 1948-1994

*imports and exports are targeted in the same complaint

Table 2(b). WTO import-related complaints (against defendant) vs. export-related complaints (against defendant), January 1995-April 1996

This hypothesis is only interesting if one can show that there was a potential demand for GATT/WTO complaints against the export-related practices of foreign governments, but that these were diverted to an alternative forum more efficacious for resolving industry complaints. Evidence for this is found in the number of CVD actions taken by GATT members over time. CVDs are intended to counter the subsidy practices of governments by levying a duty equal to the net amount of the subsidy granted on the product being exported. From the period 1980 to 1993, 162 definitive countervailing duties were levied by parties to the GATT, with the U.S. accounting for 80% (130) of the total, and the rest divided among Australia (13),Canada (11), Chile (4), the EC (3), and Brazil (1).[57] Note that this number is quite similar to the total number of import-related complaints (153) brought to the GATT in the 1980s and 1990s (from Table 2(a)), suggesting that the underlying demand for import- and export-related complaints is about equal. Though it is difficult to say how many of these CVDs indicate foreign subsidy practices which might have been the basis for GATT litigation, either as a legal violation or a non-violation nullification and impairment case under Article XVI of the General Agreement or the Subsidies Agreement, it does indicate the existence of private sector demand for action against foreign government policies affecting exports, particularly in the U.S. Rather than taking a complaint to the multilateral forum, imposing a CVD at the border can often be the quicker and easier route when increased domestic competition is the primary issue.

Table 2(c) looks at the breakdown of cases focusing on the export policies of the defendant. As predicted by Corollary 2(a), the main reason for export-related complaints

against the defendant was increased competition in third-country markets and subsequent loss of market share -- something which cannot be easily redressed without the

cooperation of foreign governments.[58] These cases accounted for 65%, or 26 of the export complaints between 1948 and 1995. Export restrictions on intermediate or consumer goods account for 18% of cases, and only 2 cases (5%) were due to competition in the domestic market. In both of the latter two complaints, the domestic forum was considered as an alternative to the GATT forum to address producers' concerns. One case involved an EC export subsidy on boneless manufacturing beef. Canadian producers were complaining about the depressive effect on domestic prices and lodged a legal violation complaint under the Subsidies Code in 1984. However, instead of continuing to attack the export subsidy in GATT, Canada launched a CVD

Table 2(c). Reasons for export-related complaints (against defendant), 1948-1994

*cases with no apparent commercial interest

investigation and levied duties in 1986.[59] A second GATT case began as a complaint by U.S. pasta producers about subsidized exports of EC pasta on the domestic market. A multilateral dispute was apparently launched only after it was determined that CVD action was not possible, as the volume of EC imports did not satisfy the CVD law's requirement of "material injury."[60] These cases constitute additional supporting evidence of the assertion that many potential export-related complaints are diverted from the GATT/WTO forum by the availability of domestic remedies.

Now we turn to the identity of the defendants. The frequency and distribution of complaints indicates whose compliance with standing multilateral rules was most often formally contested. However, the frequency of contestation may be a systematically biased estimator of the degree of parties' actual compliance relative to each other, discussed further below. Tables 3(a) and 3(b) present GATT complaints and WTO complaints against all parties, divided roughly into three groupings of countries according to size and level of development. Subtotal A includes the four leading exporters and importers in world merchandise trade: the European Union, the United States, Japan, and Canada. In 1995, these 4 parties together accounted for 52.5% of exports and 51.9% of imports globally (excluding European Union intra-trade) and are collectively referred to as "the Quads" in GATT parlance.[61] Subtotal B groups other OECD countries, state-trading countries, and Asian newly-industrialized countries (from 1994 onwards)

together, and Subtotal C includes all developing countries which have been the target of GATT/WTO trade complaints.[62]

Table 3(a). GATT complaints against all parties, 1948-1994

Hypothesis 3 predicted that one should observe more GATT/WTO complaints against countries accounting for the greatest proportion of international trade (from all parties), and conversely, fewer complaints against smaller countries (from all parties). This is because private producers which complain to their governments about barriers to market access are expected to target those countries where the greatest amount of trade is at stake in terms of value and volume. On average the largest countries in terms of exports and imports will constitute the most important markets across sectors. And indeed, we do observe that the large states account for 85% of all complaints over the four decades of the GATT's operation. The EC and its member-states ranked first in the number of appearances as a defendant, with 43% or 127 of the 295 complaints, followed by the U.S. with 28% or 83 complaints, and Japan and Canada at 7% (22 cases) and 6% (18 cases) respectively. Complaints against the latter two countries began surfacing mostly in the 1970s and 80s, whereas the US and EC figures fluctuate somewhat over time. Both in absolute and percentage terms none of the smaller developed countries in category B or the developing countries in category C come close in the frequency of complaints. Norway leads the other developed countries with 5 complaints against it, followed by Australia with 4. In the developing country group Brazil ranks first with 5 complaints and is seconded by Korea with 4 (Korea appeared as a defendant for the first time in the 1980s and seems to be becoming a regular party to dispute settlement in the WTO, see Table 3(b)).

Table 3(b). WTO complaints against all parties, January 1995-April 1996

Looking at the first 16 months of the WTO's operation, there is a notable shift in the distribution of complaints. The Quads still account for the lion's share of cases, but this percentage has dropped significantly to 60%. Most of the shift has occurred towards category B countries, which now include the emerging market economies of Hungary and Poland. They account for 29% of complaints. As these countries, and NICs such as Korea, become more significant as importers and exporters, we can expect the number of complaints against them to rise proportionately, if our expectations about the demand-side incentives for complaints are correct. The number of developing country appearances as defendant increased only slightly to 12%, and is still below the high of 17% witnessed during the 1950s. Thus, the prediction that most complaints should target the largest trading countries still holds, albeit less strongly.

This hypothesis is not necessarily obvious or self-evident. One could have alternatively predicted that small countries would more frequently be the object of lawsuits, since they can be more easily pressured into fulfilling legal claims. Another reason why one might have expected more complaints against category C states in particular -- the developing countries -- has to do with actual levels of compliance. Theoretically, one might expect that small LDCs which lack efficient bureaucracies and sophisticated legal expertise would have more trouble properly implementing and observing GATT/WTO law. Corroborating evidence for this assertion is found in the GATT's early creation of a Technical Cooperation Division designed to assist and advise developing countries in interpreting and implementing their multilateral trade obligations. And empirically, it is well-known that a large number of LDCs had pursued very illiberal, trade-restrictive policies since the early 1950s. Most of Latin America and many nations of southeast Asia and Africa pursued strategies of import substitution aimed at replacing manufactured products that had previously been imported, and required subsidies, high protective tariffs, and exchange controls to encourage domestic industries.[63] It is only since the mid-to-late-1980s that significant strides towards liberalizing domestic markets have been made in the developing countries.[64] Even though LDCs have been subject to fewer legal constraints under the GATT code than the developed countries[65], from the point of view of strict compliance, LDCs very likely did less well in meeting the obligations which they did have than the number of observable complaints suggest.

This is why the number of observable formal complaints against all parties, most of which allege legal violations of the GATT/WTO code, is argued to be a biased estimator of true relative compliance by those countries. On the demand-side, legal violations by the large are more likely to come to governments' attention due to fire alarm oversight which emphasizes private particularistic interests with significant trade value at stake. By contrast, legal violations by small countries (mostly less-developed) are likely to be underreported by private industry due to the smaller trade values involved. Furthermore, private producers exporting to developing country markets, at least in the past, may have expected to encounter some bureaucratic difficulties in a non-legalistic contracting environment, and found other ways to resolve their problems. Bringing a complaint to one's government to exert pressure on relevant officials might be counterproductive for doing business in the future in such an environment, especially since one's firm can usually be identified by the nature and specifics of the complaint.[66]

There are also supply-side factors which reinforce the tendency to observe more complaints against the large than the small. Large countries have historically been the most frequent suppliers of complaints for dispute settlement (about 60% of all cases under the GATT), which is not surprising given their bureaucratic and legal expertise and significant trading interests. Hypothesis 4 predicted that large countries would tend to prosecute other large countries, and refrain from bringing small countries to formal dispute settlement.

Table 4. Complaints by the four largest trading parties against all other parties, 1948-1994

Table 4, depicting complaints by the four largest trading parties against all other parties under the GATT is consistent with this expectation. Between 1948 and 1994, 84% of complaints by large countries were about other large countries (for Canada and Japan, 100% of complaints were against other Quad nations). 5% of complaints were directed at smaller industrialized countries, and 11% of cases were brought against LDCs, with the U.S. somewhat more likely than the EC to prosecute the latter.

A sensible interpretation of this proclivity for "great power" trade conflict is that trade disputes between the large, particularly where important domestic interests are at stake, are less likely to be resolved at earlier stages of the consultation process. Parties will require stronger methods such as panel rulings or even the implicit or explicit threat of retaliation to enforce their claims. Conversely, one could argue that lower levels of pressure are sufficient to obtain policy adjustment when the power asymmetry is in favor of the complainant, thus preempting the need for large-small complaints.

However, there is strong anecdotal evidence which suggests that under the old GATT at least, very little enforcement of obligations vis-‡-vis LDCs occurred at all. As early as 1969, John Jackson, a noted authority on the GATT, observed that

...a spirit see