Abstract
Scholars have increasingly theorized, and debated, the decision by states to create and delegate authority to international courts, as well as the subsequent autonomy and behavior of those courts, with principal–agent and trusteeship models disagreeing on the nature and extent of states’ influence on international judges. This article formulates and tests a set of principal–agent hypotheses about the ways in which, and the conditions under which, member states are able use their powers of judicial nomination and appointment to influence the endogenous preferences of international judges. The empirical analysis surveys the record of all judicial appointments to the Appellate Body of the World Trade Organization (WTO) over a 15-year period. We present a view of an Appellate Body appointment process that, far from representing a pure search for expertise, is deeply politicized and offers member-state principals opportunities to influence Appellate Body members ex ante and possibly ex post. We further demonstrate that the Appellate Body nomination process has become progressively more politicized over time as member states, responding to earlier and controversial Appellate Body decisions, became far more concerned about judicial activism and more interested in the substantive opinions of Appellate Body candidates, systematically championing candidates whose views on key issues most closely approached their own, and opposing candidates perceived to be activist or biased against their substantive preferences. Although our empirical study is specific to the WTO, our theory and findings have implications for the judicial politics of a large variety of global and regional international courts and tribunals.
Introduction
One of the most significant developments of the past several decades has been the legalization, or more precisely the judicialization, of international politics, with a growing number of international courts enjoying regional or global jurisdiction over issues such as trade, human rights, and international peace and security (Abbott et al., 2000). With this development has come an academic debate over the most useful way to theorize the decision by states to create and delegate judicial authority, as well as the subsequent autonomy and behavior of international courts.
One approach, that of principal–agent (PA) analysis, draws from rational-choice theories of domestic and international politics, arguing that instrumentally rational actors (voters or legislators at the domestic level, states at the international level) delegate powers to executive and judicial agents systematically in order to lower the transaction costs of policy making, and that in doing so they tailor the discretion of their agents, again systematically, as a function of several factors, including the demand for credible commitments, the demand for policy-relevant information, and the expected gap between the preferences of the principals and the agents. Following the initial act of delegation, some PA accounts go on to make predictions about the autonomy and influence of executive and judicial agents, which are typically theorized to vary as a function of the administrative and oversight mechanisms available to the principals. As such, the PA approach has been a limited but fruitful middle-range theory, allowing scholars to problematize and generate testable hypotheses about delegation and agency in a variety of empirical contexts (Elsig, 2011; Hawkins et al., 2006; Pollack, 2003). With respect to international courts and tribunals, PA analysts suggest, states are motivated primarily by a desire to establish an impartial third-party dispute resolution system, and in so doing typically create judicial bodies that are largely insulated from day-to-day political pressures, enjoying considerable (but not perfect) independence as a result.
Despite its purported benefits, the PA approach to judicial delegation has come under sustained criticism from scholars who suggest that delegation to international courts is different not simply in degree but also in kind from other acts of delegation. In this view, delegation to international courts is guided by a distinct ‘logic of delegation,’ with international jurists theorized not as agents but as ‘trustees’ largely or entirely beyond the influence of the member-state principals that appointed them. For adherents of this trusteeship approach, the various control mechanisms explored by PA analysts are either unavailable or ineffective, rendering member-state principals, qua principals, essentially irrelevant to international judicial behavior, which is instead guided primarily by rhetorical and legitimacy politics (Alter, 2008; Grant and Keohane, 2005; Majone, 2001).
Such trusteeship arguments have made a useful contribution, by pointing to the difficulty for principals of controlling international courts, the multiplicity of interlocutors and sources of influence on such courts, and the distinctive nature of legal argumentation as a decision-making mode that places limits on the naked exercise of state power. Nevertheless, we argue, the designation of international courts as trustees that are by definition beyond the control of its member-state principals is counter productive, rejecting by assumption hypotheses that should instead be subject to systematic empirical testing.
Against this backdrop, we seek in this article to formulate and test a set of hypotheses about the ways in which, and the conditions under which, member-state principals use their powers of judicial nomination and appointment to influence the endogenous preferences of international judges, and thereby judicial behavior. Unlike the trusteeship view, which holds that principals select international judges solely on the basis of expertise and personal reputation, we hypothesize that individual member-state principals will seek to influence international courts by nominating and appointing international judges whose nationality, judicial philosophy, and views on specific issues most closely approximate their own. This effort by principals to shape ex ante the preferences of judges is well established in the literature on domestic judicial appointments, and we hypothesize that the same effort takes place, in modified form, with respect to international courts.
Our empirical focus is on the Appellate Body (AB) of the World Trade Organization (WTO), which closely approximates the ideal type of a trustee-court. The trustee and PA approaches therefore generate competing predictions in this case, with the trusteeship view predicting that member states will refrain from attempting to use their limited instruments of control over their trustees, while the PA view leads us to expect member states to rely on judicial appointment as a rare source of leverage over a largely autonomous court. Evidence that member-state principals systematically employ nomination and appointment procedures as a means to influence the preferences of AB members, therefore, would lend support to the PA view, and raise the prospect that other courts may be subject to the same sorts of efforts.
The literature on WTO dispute settlement has paid scant attention to the selection of the dispute settlement bodies. 1 We add to the literature by presenting a detailed empirical analysis of the nomination and appointment procedures for members of the WTO AB during the first decade and a half of its existence, from the initial appointment of seven AB members in 1995 through the most recent appointments in 2009, that is, the entire universe of such cases to date.
We present a view of an AB appointment process that, far from representing a pure search for expertise, is deeply politicized. WTO member states, we demonstrate, can and do influence the endogenous preferences of the members of the AB through a rigorous screening process that pays close attention to the views of candidates on specific issues of interest to those states. We further demonstrate that the AB nomination process has become progressively more politicized over its first decade and a half. During the earliest nominations to the AB, WTO member states attempted to secure the appointment of members of the AB from their own states, but otherwise opted for extensive legal and policy experience. With the passage of time, however, as the AB adopted controversial decisions with profound implications for states’ sovereignty and economic interests, WTO member governments became far more concerned about judicial activism and more interested in the substantive opinions of AB candidates, systematically championing candidates whose views on key issues most closely approached their own, and opposing candidates perceived to be activist or biased against their substantive preferences.
To be clear, we do not seek to replace the trusteeship view with a crude PA view depicting the WTO AB as the puppet of the US or other powerful states. However, we argue that a trusteeship view, which assumes the ineffectiveness of member-state control mechanisms by definition, ends up directing our attention away from one of the most important developments of the past 15 years, namely, the increasing politicization of the AB selection process. Rather than bluntly dichotomizing obedient ‘agents’ and autonomous ‘trustees,’ we argue for a more subtle, discriminating approach to international courts, which draws theoretical lessons from the domestic law-and-politics literature, and which problematizes and studies empirically the specific mechanisms whereby states are — or are not — able to influence the preferences and the behavior of the international judges they appoint.
The article is organized in four parts. Following this introduction, the second section reviews and draws lessons from the scholarly literature on domestic judicial appointments, and derives four hypotheses about the nature of the international judicial appointment process and its evolution over time. The third section presents our empirical analysis of judicial nomination and appointment to the WTO AB, demonstrating the extensive and growing use of judicial selection as a means whereby member-state principals attempt to influence the preferences of AB members. The fourth section concludes, summarizing our findings with respect to the WTO and locating our WTO case within the broader universe of international courts.
The politics of international judicial appointments
Domestic judicial appointments
Although the study of international judicial appointments is in its infancy, such appointments have a close analogue in the appointment, by national executives and legislatures, of domestic judges. More specifically, our understanding of the judicial appointment process owes much to the study of US judicial politics, where a community of scholars has examined and established clear empirical findings about the process of appointing federal judges to the Supreme Court and to lower federal courts. US federal judges, as is well known, are appointed for life terms, subject to removal only for impeachable offenses, and in this sense closely approximate the ideal of judges as trustees. And yet the American judicial politics literature points clearly to a judicial appointment process that is profoundly ‘political,’ in the sense that political actors with the power to nominate and confirm judicial candidates are guided in large part (although not exclusively) by partisan and ideological concerns (Epstein and Segal, 2005; Moraski and Shipan, 1999). Put more formally, political actors seek to influence public policy through their judicial appointments by appointing candidates whose judicial philosophies and ideological views are likely to move a court in the direction of their own preferences (Moraski and Shipan, 1999: 1073). Furthermore, there is substantial evidence that the ideological preferences of judicial candidates at the time of their nomination really do shape the subsequent behavior of those judges, suggesting that judicial appointment is an effective means whereby political actors shape judicial outcomes (Epstein and Segal, 2005: 117–141; Segal, 2008: 24–28). Although a complete review of this literature is beyond the scope of this article, its essential outlines can be summarized for our purposes in terms of the ‘four I’s’ — interests, interactions, institutions, and information — emphasized in rational-choice theories common to American, comparative, and international politics (Frieden et al., 2009; Lake, 2008).
First, with respect to interests, the key players in the US judicial appointment process are the President and members of the Senate, all of whom have complex preferences. Regarding the President, judicial scholars have demonstrated that nominating behavior is shaped by a variety of motivations, which include not only a desire to nominate well-qualified candidates, but also efforts to appease political allies, interest groups, and the electorate; to provide a source of patronage for friends and allies; and to shape the ideological balance of courts by nominating candidates who share the liberal or conservative views of the presidents who nominate them. Senators have similarly complex preferences, combining concerns about patronage with ideological, policy, electoral, and interest-group considerations. Despite these complexities, empirical evidence points to partisan and ideological factors as being among the most important predictors of both Presidential and Senatorial choices (Epstein and Segal, 2005: 47–116), and, for this reason, theoretical models of judicial appointments typically assume that both the President and Senators use judicial appointments to move the court toward their own policy preferences (Moraski and Shipan, 1999).
Second, however, examining the sincere preferences of Presidents and Senators is insufficient. We must also understand the strategic interactions among Presidents and Senators in the two-stage ‘nomination game,’ with the President having the constitutional right to nominate judges, subject to the advice and consent of the Senate (Moraski and Shipan, 1999: 1071). Presidents, in the US system, have the right to nominate judicial candidates, which gives the President formal agenda-setting power vis-a-vis the Senate, given only a straight up-or-down vote on the President’s nominee. However, the President must anticipate the likely reception of his candidate in the Senate, and should select the candidate closest to his preferences who is likely to be confirmed by the Senate. This two-stage process means that no single actor can dominate the process, and judicial appointments always represent the outcome of a strategic interaction between a President with agenda-setting power and a Senate with collective veto power.
Third, the outcome of this strategic interaction will be shaped in large part by the institutional rules imposed by the Constitution and by Senatorial rules and traditions. The most important of these rules for our purposes is that the candidate must receive the votes of a simple majority of 51 Senators, although the possibility of a filibuster raises the effective threshold to 60 votes. This is an imposing institutional obstacle, which becomes particularly so when the Senate is controlled by the party in opposition to the President (Epstein and Segal, 2005: 85–116; Moraski and Shipan, 1999).
Fourth, information — its availability and its distribution among the various actors — plays an important role in both the nomination and confirmation processes. As in any act of delegation, the principals possess only imperfect information about the true preferences of any judicial candidate. This situation provides the candidate with an incentive to utilize information strategically, revealing information that is appealing to the President and to a majority of Senators, while concealing information likely to elicit opposition (Kagan, 1995). Conversely, both the President and Senators should exert considerable efforts to collect independent sources of information through various ‘vetting’ procedures. Furthermore, the domestic literature points to a likelihood of increasing politicization of judicial appointments over time, as information about the stakes of judicial appointment increases. When a new court is first established, political actors are uncertain about the nature of future cases and the behavior of judges on the court. With the passage of time, however, principals will become aware of the nature and the policy implications of judges’ behavior, and adjust their expectations accordingly (Russell, 2006: 421).
We should not, of course, make the mistake of assuming that judicial appointment processes in all political systems take the same form as US federal appointments, and indeed there is evidence that appointment procedures vary substantially across political systems, both at the state level within the US, and among national common-law and civil-law systems. 2 This understudied institutional variation in judicial appointment procedures should make us wary of simply ‘scaling up’ the US model, or indeed any domestic model, uncritically to international courts, and we therefore modify the ‘off-the-shelf’ domestic model to international politics in the next section. Nevertheless, our point here is that even in the US, where federal judges closely approximate the ideal type of a trustee, there is overwhelming evidence that political factors play a crucial role in judicial appointments and behavior. Furthermore, as we shall see presently, the two-stage model of judicial nomination and confirmation corresponds surprisingly closely to the appointment process to most international courts, albeit with different institutional rules and different actors in the central roles.
International judicial appointments
Despite the similarities between the domestic and international judicial appointment processes, advocates of the trustee approach suggest that international courts are different in fundamental ways from domestic courts, rendering the appointment process almost insignificant as a control mechanism. As Karen Alter argues: In the domestic realm, the appointment and promotion process is often a potent tool to influence the judiciary, wielded by whoever has dominant control of the executive and legislative branches of government. The international constitutional order established after World War II, however, was tailor-made to ensure that any ‘international’ decision requires the political support of multiple states. Thus in contrast to the domestic process where political branches can control the nomination process, in the international realm each country chooses which individuals it nominates for international positions. Selecting from among international judicial nominees is certainly politicized; the larger point is that the overall nomination and appointment process cannot be controlled by any one state or organized group of states. While states did choose to keep the international judiciary beyond the control of the most powerful states, the probably unintended result is that international judges are institutionally less subject to appointment politics than their domestic counterparts. (2008: 46)
Alter is correct about the differences distinguishing international judicial appointments from their domestic counterparts, including the large number of state actors involved in the nomination and selection of most international judges. Nevertheless, we disagree with Alter on two fundamental points. First, we note that Alter focuses almost entirely on the use of reappointment as an instrument of ex post influence over judicial decisions. Judicial appointment procedures, however, are not only or even primarily useful as an ex post sanction for judges; they are also and especially useful as a means for shaping ex ante the endogenous preferences of judges. Second, Alter assumes that the judicial appointment procedure is radically decentralized, with any given state able to influence only its own nomination. This is indeed the case for the European Court of Justice (ECJ), but in other international courts state nominations are only the beginning of a winnowing process whereby a larger intergovernmental body selects from among judicial nominees.
Put simply, we find important analytic similarities between the domestic and international judicial appointment processes across all of the ‘four I’s,’ that is: (a) the complex, and at least partly ‘political,’ interests of the key actors; (b) the nature of the appointment process as a two-stage strategic interaction between agenda-setting parties that nominate judicial candidates and an electorate that confirms them; (c) the importance of institutional rules that shape the process and the power of the nominators and the electorate; and (d) the strategic use of information by both state principals and judicial candidates.
Until very recently, both legal scholars and political scientists had neglected the study of international judicial appointments, largely because of the empirical difficulties of studying what has been (and generally remains) a largely opaque process (Terris et al., 2007: 15). Nevertheless, a number of studies have appeared over the past several years, offering important insights into the interests of state actors, the strategic interactions of states and the importance of institutions in the appointment process, and the strategic use of information by both states and judicial candidates (Mackenzie et al., 2010; Terris et al., 2007; Voeten, 2007, 2008, 2009; Wood, 2007).
First, with respect to the interests or preferences of states, Erik Voeten usefully distinguishes multiple motivations for states in their choice of judicial candidates, which include a search for the best qualified candidate, concerns to maximize the credibility of commitments, and concerns to limit the sovereignty costs of judicial decisions (Voeten, 2009: 389). 3 We agree with Voeten that states have complex preferences, but we argue that such complexity is not unique to the international sphere. Our claim here is not that member-state preferences are simple or one-dimensional, but rather that just as US presidents manage to nominate like-minded judicial candidates despite complex preferences, so states will tend, ceteris paribus, to nominate and support candidates with substantive preferences similar to their own.
Second, in terms of the strategic interactions among the players, international judicial appointment procedures vary considerably across the various courts, but nearly all of them follow the same basic two-stage model in which individual member states nominate candidates in the first stage, with the formal appointment of judges, typically by a plenary of all member states, in the second stage (Mackenzie et al., 2010: 2). Different states arrange their domestic nomination process in various ways, but in most cases the process of selecting national nominees is an informal one, undertaken within the executive branch by senior civil servants who identify and screen candidates through analyses of their writings, interviews, and background checks (Terris et al., 2007: 25–27). Our expectation, spelled out further below, is that states will use this control to press for the appointment of judges whose substantive preferences are close to their own.
Third, the agenda-setting power of any given nominating state depends fundamentally on institutions, that is, the formal and informal rules of the international court in question. Here, we can distinguish three fundamental institutional rules that shape the power of nominating states and the ultimate choice of international judges. First, following Mackenzie et al., we distinguish between: full representation courts, where each state has a judge of its own nationality on the court permanently, and ‘selective representation’ courts, where there are fewer seats than the number of states that are parties to the court’s statute. In the latter type of court, a choice has to be made between candidates from different states, thus giving rise to a greater degree of competition in which political influences, among other factors, can and do hold sway. (2010: 7)
In full representation courts, such as the ECJ, each member state is guaranteed a seat on the court, and the various members typically rubber-stamp each other’s nominees, giving the nominating states near-total agenda-setting power and making the confirmation stage a formality. 4 By contrast, in selective representation courts, the nomination of candidates is only the first stage in a longer process in which nominees are weighed and a selection made from a potentially large number of candidates.
In such cases, judicial appointments will be shaped by a second institutional rule, namely, the voting rule governing the election of judges. The statutes of the various international courts specify varying voting rules for the final appointment of judicial nominees. Appointment of judges to the International Tribunal on the Law of the Sea, for example, requires a two-thirds vote among the states-parties, while nominations to the WTO AB call for a consensus among the membership, which sets a much higher threshold and grants member states an effective veto over other states’ nominees. The institutional voting threshold, therefore, plays a crucial role in shaping the international ‘nomination game,’ with lower thresholds favoring nominating states and higher thresholds favoring the electorate.
Not all member states will enjoy equal influence in such settings, and a number of international courts feature formal or informal rules relating to the geographic distribution of seats among the member states, including the possibility of guaranteed or permanent seats for certain countries or regions (third institutional rule). In such cases, countries with guaranteed seats are in a particularly strong agenda-setting position, with their candidates virtually guaranteed acceptance, while other members must consider the likely reception of their nomination among the electorate. In the ensuing elections, member states typically engage in extensive lobbying, bargaining, and ‘horse-trading’ on behalf of their national candidates, and final appointments are often the result of messy compromises among the members (Terris et al., 2007: 34–36). For our purposes here, the important point is that in selective representation courts, the judicial appointment process is not as radically decentralized as Alter suggests, since a collective decision determines the ultimate choice of judges on the court.
Fourth and finally, information plays a significant role in the judicial appointment process at the international level, albeit in slightly different ways from the domestic level. By contrast with the domestic level, both the national nomination process and the final election of judges are non-transparent, taking place behind closed doors (Mackenzie et al., 2010; Terris et al., 2007). For this reason, public opinion and interest-group pressures are less constraining on international court appointments. That said, however, the essential logic of judicial appointments, in which candidates attempt to reveal only favorable information while their principals attempt to determine their true preferences, should operate at the international level as at the domestic level. Furthermore, the informational logic whereby political principals gain access to greater information about the stakes and substance of judicial decisions over time, and increase their attention to judicial appointments, should also operate at the international level.
Four hypotheses
The foregoing discussion of interests, interactions, institutions, and information allows us to generate four hypotheses about the behavior of both states and judicial candidates in the international judicial appointment process.
First, with respect to the interests of member-state principals, we hypothesize that member governments, in nominating their own judicial candidates and in evaluating the candidates of other countries, will attempt to secure the appointment of judges whose preferences on substantive issues likely to come before the court are as close as possible to their own. Of course, even powerful states cannot hope to exert more than a marginal influence on the court through their own nominations or through their votes on the nominations of others; however, we would expect states to use their limited power over judicial appointment to move the court, however modestly, toward their own ideal points (H1).
Second, with respect to strategic interactions, member-state principals must operate within the constraints of the ‘nomination game.’ Nominating states, therefore, are unlikely to act upon their sincere preferences in their choice of judicial candidates, but will instead behave strategically, anticipating the likely reaction of other states to their nominations and putting forward candidates closest to their preferences who are also likely to receive the endorsement of the electorate. By the same logic, member states at the second stage should attempt to use such influence as they have to secure the confirmation of candidates who are most likely to move the court toward their substantive preferences, and reject candidates with conflicting preferences (H2).
Third, the relative power of states in both stages of the game will be shaped by three institutional features: (a) the full or selective representation of member states on the court; (b) the decision rule governing the selection of judges; and (c) the formal and informal rules about (geographic) distribution of seats on the court. More precisely, we predict that member states in full representation courts such as the ECJ will enjoy nearly complete agenda-setting power vis-a-vis their own nominees but little or no such power over the nominees of other states. By contrast, selective representation courts will, ceteris paribus, decrease the agenda-setting power of nominating states, while increasing the power of states to select among candidates from other states (H3a). Within selective representation courts, the decision rule for the collective appointment will also influence the respective power of states at the nominating and selection stages, with more demanding supermajority or consensus rules acting, ceteris paribus, to decrease agenda-setting power and increase the power of states at the selection stage (H3b). Despite these general tendencies, rules about geographic distribution of seats can increase or decrease the agenda-setting powers of specific states, particularly great powers that are granted the right to a permanent seat (H3c).
These institutional hypotheses (H3a-H3c) allow us to make specific predictions about judicial appointments across international courts, as a function of these three specific institutional rules. With respect to the ECJ, for example, they predict that the full-representation nature of the court will empower nominating states vis-à-vis the collective membership, giving each state significant influence over its own choice but radically decentralizing that choice, as Alter (2008) suggests. With respect to the WTO AB, by contrast, the combination of selective representation and a consensus decision rule should generally increase the power of the electorate at the expense of nominating states, who need to be highly strategic in their nominations to secure consensus support; by the same token, however, we would expect the US, the EU, and Japan to have greater agenda-setting power than other members, given their informal right to a permanent seat on the AB. 5
Our final hypotheses concern the role of information in the appointment process. Following PA models, we expect candidates to international judgeships to provide selective information to governments, designed to convince member states that the candidates are not activists and share their core substantive preferences (H4a). We also hypothesize that member-state principals will undertake increasingly careful screening of judicial candidates over time, as the court’s developing case law makes clear the substantive stakes of judicial appointments (H4b).
Put simply, then, the dependent variable of our study is the behavior of principals (and, for H4a, of judicial candidates) in the nomination and appointment of international judges, and our central expectation is that Members of the WTO will attempt to use their powers as principals to shape the endogenous preferences of the judiciary, rather than basing their decisions solely on the expertise of judicial candidates. We do not, by contrast, purport to test the hypothesis that nomination and appointment decisions actually influence judicial behavior, an effort that would require us to establish a correlation between the ex ante views of judicial appointees and the ex post behavior of those same judges (Voeten, 2008). Nor do we attempt to explain variation in institutional design and reform of appointment processes across international courts, although we concur with Epstein et al. (2002) that such an analysis would be highly desirable.
Judicial appointments to the World Trade Organization Appellate Body
In this section, we examine the politics of nominations and appointments to the WTO AB, seeking to test our four hypotheses. Given that the study covers AB selections since the origins of the new dispute settlement system, we exclude potential threats to our research design due to selection bias. We first discuss the appointment of the original seven members. Then we outline the changing context and focus in more detail on a number of individual selections using process-tracing tools that elucidate the preferences of the key players, the strategic interactions among nominating states, judicial candidates, and the collective member-state electorate in Geneva. We rely on primary and secondary sources, including extensive use of official WTO sources and specialized press reports, supplemented by 33 semi-structured interviews with involved stakeholders. 6
The WTO dispute settlement system consists of two stages. First, disputes can be brought by WTO Members to expert groups composed of three individuals (panels). These individuals are chosen by the disputing parties; if parties to the dispute cannot agree, the Director-General (DG) of the WTO appoints panel members. Second, parties have the possibility of turning to another institution to demand a review of a panel decision by appealing panel findings: the AB. This institution, created in 1995, is composed of seven members, three of whom serve on appeals defined by a rotating principle. Recommendations by the AB cannot be overturned easily. WTO Members would have to reject the AB report by consensus, or WTO Members (by a three-quarters majority) would have to agree on an ‘authoritative interpretation’ of the provisions of a covered agreement, thereby challenging the interpretation of the AB.
The designers’ intentions and the evolving profile of Appellate Body Members
In the old GATT times, dispute settlement developed as a quasi-diplomatic tool based on the panel system. There was no right to a panel (a panel was only constituted if both parties agreed), and panel recommendations had to be accepted by both parties (and by the entire membership) to be binding. During the Uruguay Round negotiations, a consensus emerged on abandoning the practice of blocking the establishment of panels and allowing the automatic adoption of reports (Bernauer et al., 2012). This increased delegation of power to panels, however, led to concerns. One proposal was the creation of an additional institution that would hear appeals: the AB. Negotiators were aware of, and concerned about, the potential importance of this body, limiting its mandate and to some extent its discretion through a number of specific design features.
First, Article 3.2 of the Understanding on Rules and Procedures Governing the Settlement of Disputes (DSU) includes a general provision that ‘recommendations and rulings of the Council for Dispute Settlement (DSB) cannot add to or diminish the rights and obligations provided in the covered agreements.’ This passage reflects a concern among contracting parties that panels or the AB would interpret agreements in ways that impact on principals’ expected benefits agreed through negotiations.
Second, the treaty states that, ‘an appeal shall be limited to issues of law covered in the panel report and legal interpretations developed by the panel’ (Art. 17.6 DSU). The objective of this condition was that the AB abstains from reviewing facts or introducing new issues of law that had not been addressed at the panel stage.
Third, the mandate of the individual Appellate Body Member (ABM) is restricted to four years, with the possibility of reappointment once (Art. 17.2 DSU). This short and renewable term for ABMs suggests a concern for judicial accountability. Furthermore, while details of appointment procedures were not clearly defined in the treaties, final appointment was to be made by consensus. This rule meant that each individual member state could, albeit at a diplomatic cost, exercise a potential veto over the appointment of individual judges (Steinberg, 2004).
Fourth, we also witness attempts at discursive control. The text of the DSU refers to its highest instance as a body (and not a court), which is composed not of judges but of members. The AB, furthermore, writes reports (not judgments) and it makes recommendations to the DSB (comprising Ambassadors) that accepts (or rejects) the report.
On the issue of ‘profile,’ the text remained quite general: the AB ‘shall comprise persons of recognized authority, with demonstrated expertise in law, international trade and the subject matter of the covered agreements generally.’ The negotiated text also proposed that ‘they shall be unaffiliated with any government’ (Art. 17.3 DSU).
In Figure 1 we briefly map the change of profile of ABMs over time. The data suggest that WTO Members increasingly select candidates with extensive trade policy experience and who have a familiarity with the WTO system and its particularities, gained through negotiation and panel activities, to the disadvantage of other key characteristics (e.g. public international law background, court experience). In 2011, all ABMs, prior to their appointment, have gathered experience in the field of trade law (in 1995: five); and five are former trade negotiators (in 1995: three). None of the seven has prior experience as a judge (in 1995: one; from 2001 to 2006: two) and the number of former panelists has risen to three (in 1995, only one ABM had previous GATT panel experience). In terms of overall expertise, the average age declined from 65.4 years in 1996 to 58 years in 2010.
Composition of AB over time.
Nomination and appointment over time
The DSU did not lay down in detail the procedures to appoint members of the AB. In addition to ‘expertise in law, international trade and the subject matter of the covered agreements,’ the original designers insisted that the AB ‘shall be broadly representative of membership’ (Art. 17.3 DSU). In 1994, a committee was mandated to provide further guidance for the nomination process. On the notion of expertise, the committee stated that it should allow ABMs to resolve ‘issues of law covered in the panel reports and legal interpretations developed by the panel.’ 7 On the question of what was meant by ‘broadly representative,’ it suggested that ‘factors such as different geographical areas, levels of development, and legal systems shall be duly taken into account.’ 8 On the issue of how to organize the process, it was proposed that a group of six comprising the Director-General (DG) and the chairpersons of five important bodies (General Council, DSB), Goods, Services, and Trade-Related Intellectual Property Rights (TRIPS)) should elaborate jointly a proposal after ‘appropriate consultation’ with Members.
The first seven
The chairman of the DSB (Ambassador Donald Kenyon from Australia) organized the first round of appointments in Geneva in February 1995. 9 Members were encouraged to propose candidates and the Committee of the Six (C6 hereafter) would recommend to the DSB the seven candidates for acceptance. Members agreed on the process and emphasized the importance of finding the ‘right balance.’ There were 32 candidates overall from 23 countries (13 alone from EU states). A total of 54 separate delegations provided their views to the C6. However, demands by the EU and the US derailed the process. The EU signaled that two seats should go to nationals from EU countries to reflect its economic importance. Similarly, the US argued that it wanted two seats given its role in world trade. Most delegations opposed such a partition. Over time, the US showed willingness to accept one seat if the EU did the same. 10 In November 1995, agreement was reached not to block the suggested list prepared by the C6 and the first group of ABMs was finally approved by the DSB. 11 Many delegations voiced disappointment, in particular the EU, which argued that the list did not sufficiently take into account EU interests (five seats went to candidates from APEC countries and one member was from a close US ally). In addition, India and Brazil were disappointed that their candidates were not among the original seven.
As to the context of the first selection process, the key objective of many Members was to have one of their nationals elected. Therefore, politicization occurred along ‘national dimensions.’ 12 For one thing, it was generally accepted that the EU, the US, and Japan would each get a seat (informal permanent seat). EU states put forward candidates, many of whom had an academic, trade law, or general law background. The selected EU candidate (Ehlermann) gained widespread support. It helped that he was from an important EU member state and was highly respected in Brussels. 13 The US did not nominate any trade law practitioner or an academic working on trade law. The Office of the US Trade Representative (USTR) suggested to the membership two candidates with quite different profiles: a former Congressman (Bacchus) nominated by the White House for his help ‘getting NAFTA through Congress,’ and a professor of international economic law not specialized in trade law. 14
The C6 realized that it was important to have an excellent group of eminent persons on the AB. A member of the original C6 summarized the choice: ‘It was clear that three seats were taken. The rest of the field was open. We needed a balance in regional representation, legal systems, experience, and developed and developing countries.’
15
In terms of individual candidates he continued: Bacchus did very well in the discussions. Ehlermann was an excellent candidate among a number of very good candidates from Europe. Japan at that time was economically strong and was the target of many disputes. We considered the Japanese candidate Matsushita strong. Then we needed someone from a developing country in Asia. Feliciano from the Philippines was very good — much better than the Indian candidate who was a former Ambassador. Feliciano was the only judge … and we wanted that element. El-Naggar was the best of a group of weaker candidates (but we wanted to have Africa represented). As to South America, Brazil had two candidates, both trade policy experts, Brazil put a lot of pressure on us, but Lacarte had impressed us and we wanted someone to be there who knew the GATT system so well … and Beeby represented Oceania.
16
As to the screening of candidates, few bilateral meetings with delegations took place. The key interviews were conducted by the C6. 17 Discussions did not focus on positions on trade policy issues or WTO law. 18 ‘The questions were very general, not like … would you vote against the US in an antidumping case, on case law, or on filling gaps.’ 19 Candidates were asked about their view of how the dispute settlement system ‘should’ work in practice. A candidate remembered: ‘We were asked about the approach the AB should take.… I told them that the AB was a slender tender plant that should be protected from too strong winds, the AB should also act cautiously … I think the Ambassadors probably liked that.’ 20
In sum, the selection of the first seven members of the AB demonstrates a concern for the eminence and expertise of the candidates, although these considerations interacted with concerns about national representation and geographic diversity. Politicization of the process was present primarily in terms of Members pressing for appointment of ‘their’ candidates, but there appears to have been relatively little screening of candidates with respect to their views on specific issues.
Second ‘generation’ of ABMs
The first reappointment process began in summer 1997. Article 17.2 of the DSU stipulated that the first term for three of the seven members would end after two years, thus producing a staggered nomination process thereafter. Three of the seven ABM were chosen by lot and were reappointed without notable discussion for their second four-year term. 21 Then, in 1999, Members were called upon to reappoint the remaining four of the original group for a second term. Their first term was to expire by December 1999. Yet, the membership did not have a clear process for reappointment in place and in addition, and something that came as a surprise, two of the original members (El-Naggar and Matsushita) did not seek a second term for personal reasons. The chair suggested renewing the terms of office for ABMs Bacchus and Beeby and starting a process of selecting two new members. 22
After agreeing to renew the mandate of two continuing ABMs, the Geneva process was launched to appoint two new ABMs. In total there were seven candidates for the two seats. 23 During the selection process, ABM Beeby passed away. Therefore, a new selection process for a single vacancy started, which overlapped with the ongoing process. In place of the two original candidates, the DSB selected in April 2000 Georges Abi-Saab from Egypt and A.V. Ganesan from India. 24 For the third vacancy, there were only two official candidates. The Japanese candidate Yasuhei Taniguchi was finally accepted by the DSB. 25
The first reappointments and new selections were characterized by Members taking decisions under time pressure. Overall, the original regional balance shifted slightly (the seat from Oceania went to India). In addition, the political context started to change. While Members would continue to lobby for their nationals, there was increasing concern in Geneva that the original seven were showing some judicial activism. Along with the increased case load came a number of contentious decisions taken by the AB. However, it was largely over a procedural issue that a large majority of WTO Members started to organize opposition. The matter was whether panels or the AB had the right to accept or reject unsolicited briefs from non-governmental organizations (so-called amicus curiae briefs). A panel held that unsolicited communications should be disregarded. 26 The AB overruled this view and argued that the DSU allows for an interpretation that would give panels the option to ‘receive’ such briefs. 27 Many Members was furious about this interpretation. In reaction to mounting criticism, the AB engaged in a number of clarifications (including concessions to address the criticism) (Mavroidis, 2001). In November 2000, Members convened an extraordinary meeting of the General Council, the most important decision-making body in Geneva. 28 The main criticism suggested that the AB decision changed the balance of rights and obligations by allowing special rights to private actors and organizations and therefore ‘trespassing its own mandate and becoming a legislator’ (Mavroidis, 2001: 9). Only the US took the floor and supported the practice. While Members were unable to overturn the AB’s decision, they gave a strong warning to the AB. This debate continued, and influenced the next AB appointments.
In 2001, the end of the terms of office for the ABMs from the EU, Uruguay, and the Philippines prompted another selection exercise. 29 In this next round, screening intensified and bilateral meetings with delegations in Geneva increased. Candidates had to answer more detailed questions. At the end of the selection process, Brazil finally got a seat at the table as well as Australia. The European seat was uncontested. 30 In September 2001, the chair of the DSB announced that, of 12 initial candidates, they would propose the following three: Luiz Olavo Baptista (Brazil), John Lockhart (Australia), and Giorgio Sacerdoti (EU). 31 Let us briefly focus on the EU seat, which illustrates the growing scrutiny to which candidates were subjected.
In comparison to the original appointment procedure in 1995, the EU Members put forward fewer candidates. The Commission was also more involved. Those candidates nominated by the member states went to Brussels for meetings with Commission officials ‘over a cup of coffee not taking more than 30 minutes.’ 32 While the Commission developed some preference for certain candidates, it abstained from influencing other Ambassadors in Geneva about their priorities. 33 The individual profiles of the three main contenders were quite different. The successful candidate, Giorgio Sacerdoti, was an outsider to the WTO world. 34 He was a professor with general expertise in international economic law. He also had experience of working with the Organization for Economic Co-operation and Development (OECD) on anti-bribery issues and had been an arbitrator in investment disputes. The other candidates had worked for the Commission, were experienced trade law experts and panelists, and had written widely, as academics, on the WTO. Both faced considerable criticism, in particular from the US. One candidate recalled that he was involved in a number of panels on trade remedies, ‘where the decision went against the US … I also had some firm opinions about reform issues and judicialization of the system, whether the selection process was transparent enough, whether we needed the participation of parliaments … and I liked to write and talk about that. I can imagine that didn’t fly very well.’ 35 In the end, the two trade law experts were effectively vetoed, while a third candidate with no clear positions on existing WTO case law was more acceptable.
The last appointment of this second group of ABMs concerned the US seat in 2003. The US for the first time established an internal screening through pre-selection interviews in Washington, DC. Interviews addressed more specific questions related to case-law development. One candidate recalled being asked about ‘the role of international law, the standard of review, interpretation of agreements, and my views on various reform issues.’ In addition, party affiliation mattered. 36 Finally, the US nominated two former USTR negotiators close to the Republican Party for the position: Merit Janow, a law professor from Columbia University, and a second candidate representing the interests of US import-competing industries who had publicly criticized the AB on various occasions. Given his views, many Members voiced concerns and opted for Janow. 37 The US nomination was a reflection of growing criticism in the US over AB opinions (Barfield, 2001), but also strategic in the sense that USTR could anticipate which of the two US nominees the Members would select.
Most recent wave of appointments
The latest wave of appointment procedures started a bit earlier than anticipated when Lockhardt died in 2006, Taniguchi left early in 2007, and the American ABM did not seek a reappointment. 38 There is no conclusive evidence on why Janow did not seek reappointment. Some evidence from interviews suggests that USTR was concerned about cases where she was part of the three persons hearing the appeals in which the AB ruled against the US. 39 In the reports, she did not use the option of sharing an individual (usually dissenting) view. 40 Most importantly, she was involved in an AB recommendation that disagreed with a panel that found US antidumping practices (so-called zeroing methodology) to be permissible. 41 A former USTR put it more generally: ‘We were not happy with US AB members who bend over backwards to show their independence by ruling against the US.’ 42 In 2008 and 2009, the regular reappointments started. In terms of regional representation, Oceania lost its seat as a Chinese candidate was selected, and the South American seat went to Mexico. 43 The Japanese continued to be represented by a former WTO Ambassador (Shotaro Oshima). The European seat went to Peter Van den Bossche (Belgium). 44
Two key issues characterized this third wave of appointments. First, sparked by the discussions on antidumping and the zeroing cases, the US argued that the AB had overlooked the negotiation history when addressing these issues. Some matters were left deliberately ambiguous, in the US view, and the court should not engage in forms of gap-filling. Therefore, the US promoted the use of dissenting views. While there was a strong view in the first group of ABMs that dissenting views should be avoided and consensus should be reached, this position changed a little. 45 The US and other WTO Members argued that the AB was now mature enough to allow this practice (Flett, 2010; Kolsky, 2006). Second, the membership started in 2002 (in parallel to the Doha Round) a reform discussion of the DSU. 46 In this context, it was the US in particular that put forward a number of reform proposals that would give Members more control over the dispute settlement process. These proposals included the suspension of procedures to allow bilateral discussions to resolve the issue, more guidance as to the rules of interpretation, the prevention of ‘gap filling,’ and restrictions on addressing constructive ambiguity. 47 The main concern was that the AB should not engage in making law. 48
In this context, Members sought to determine candidates’ views on DSU reform and on other case-specific questions and limited their support to candidates whose views were not too distant from their own. 49 This increased attention to candidates’ views favored candidates with non-controversial positions and those who had been careful in the past not to make enemies in Geneva. In addition, candidates’ knowledge of the system as well as experience as arbitrators were valued more highly than experience as judges of national courts. Finally, over time, more Members have been actively using the WTO dispute settlement system, therefore creating more interested parties and more potential veto players in the selection process.
As a result of these developments, the screening process at the Geneva level intensified further. 50 Candidates were now encouraged to meet as many Members as possible and envisage visits to Washington, DC and Brussels. A member of a recent Selection Committee described the increased screening: Members ‘quiz candidates on the role of the DSB and WTO jurisprudence, about the role of AB vis-a-vis negotiated agreements, they have some strong concerns about filling gaps and the AB making law, they want to learn whether candidates have experience in dispute settlement and they ask about specific issues, including agricultural policy or zeroing … important for some Members was the issue of dissenting views.’ 51
In reaction to increased screening, agents give selective information to the nominating principals. Interviewees confirmed that potential candidates seek to prepare for and adapt to Members’ questions. A successful candidate confirmed, ‘I worked on my opening statements which got better over time … and you adjust a bit depending on who you talk to.’ 52 Another successful candidate was aware that the Members were looking for a type of candidate who would be sensitive to the membership’s concerns. ‘It’s important to show that the candidate is sensitive for the political set-up of the system … as a candidate you take a neutral position on politically sensitive issues (not in favor, not against).’ 53 They also get assistance from the Geneva-based national delegations. One Ambassador confirmed that ‘we prepared the candidate for the talks; we explained to him the purpose of WTO law, dispute settlement history, and the preoccupation of Members.’ 54 Another successful candidate concluded that ‘if you want to become ABM, I would advise against writing on the subject matter.’ 55 Below, we briefly examine two selection procedures from this most recent round.
Starting in 2008, for the Latin American seat, there were four candidates in the race. All of them had differing profiles. The Brazilian candidate, a highly respected constitutional judge, was an outsider to the system. The other three were better known in Geneva and had trade policy expertise gathered through trade negotiations. The Brazilian candidate was strongly supported by the Brazilian government. The empirical evidence on this selection suggests that a key factor played against the candidate: the perception of her being an ‘activist judge.’ An Ambassador recalled: ‘The candidate signaled that she didn’t know much about WTO, but had experience as a judge and had developed jurisprudence in Brazil.’ 56 A legal advisor to the Ambassador continued: ‘she mentioned that she helped introduce into the Brazilian system more judicialism; that was a red flag for many.’ 57 In addition, as Brazil realized that its candidate was facing powerful opposition, it started an aggressive campaign. A closely involved observer recalled: ‘The campaign backfired and actually helped members to gather support for another Spanish-speaking candidate instead.’ 58 The two candidates from Argentina and Costa Rica were Geneva insiders. Yet both faced opposition (in particular from the EU): ‘one had attacked some time ago the EC Ambassador over Bananas in a Council meeting after the deal was done. You don’t do that; he lost all support from the membership!’ 59 Another candidate had confronted the EU by criticizing EU fishery and agricultural subsidies during post-Uruguay Round negotiations. 60 There were also issues raised with regard to legal expertise. 61 In the end, Members had fewest issues with the Mexican candidate and elected Ricardo Ramirez, a former deputy general counsel for trade negotiations of the Ministry of Economy in Mexico. This case illustrates the strategic interaction between the nomination and the selection stage, and the growing demands of navigating an increasingly politicized and consensus-based election. Mexico was most strategic in its nomination, while Brazil overestimated its diplomatic power to push a candidate through.
A year earlier, in 2007, the US looked for a replacement for their ABM. The US had two main objectives when the internal nomination process started. The chosen person should share the concerns of the US on key issues and should be a strong personality able to influence others on the AB. ‘It was a very delicate situation,’ the US coordinator of the internal process recalled; ‘there was this controversy in the US on zeroing; there was this concern that the AB had overstepped its mandate.’ 62 The US selection was not influenced by party politics this time, but staffers of the key Congressional Committees were consulted and USTR Susan Schwab was heavily involved. 63 Among ‘the eight or nine candidates, we were looking for someone who had strong understanding of WTO law, ideally worked for USTR and understood our positions, knew the role of the AB and had good persuasive skills to influence the AB decisions;… we needed someone who could sensitize others.’ 64 One of the candidates recalled of the US internal nomination procedures that: ‘They wanted to know my views on the role of precedent, how to interpret public international law, how well the AB works, under what conditions dissenting should be allowed …; of course they asked me on trade remedy cases … and they also wanted to see whether I could persuade others on the merits of arguments.’ 65
The US again chose two candidates to be presented to the membership, both with a similar profile. As to the Geneva process, one candidate recalled having met some 20–25 Ambassadors. 66 In these meetings and during the interaction with the C6, the questions on the role of dispute settlement and various reform issues took center stage. With regard to reform questions, ‘I would usually tell them — which is also my view — that it’s not for me to provide new solutions or prefer one over the other issue in reform debates … it’s up to the members.’ 67 As to positions on legal issues, ‘they checked my judicial temperament.… The consensus–dissent question was discussed a lot. They wanted to know whether I thought that the judicial system was mature enough to allow dissenting view. … They also wanted to see whether I was a team player … and would not be too strongly opinionated, not too hard to move and change views.’ 68 These questions illustrate that other Members anticipated that the US was attempting to push a candidate close to its own preferences. Yet, their possibilities were constrained as they faced two similar candidates. The Members finally picked Jennifer Hillman. 69
While the focus of this article has not been on the potential effects of individual or collective selection on subsequent case law, this last selection is interesting as Hillman later sat in a dispute over antidumping where one of the judges dissented. 70 The separate and very detailed opinion on US zeroing practices 71 emphasized ‘the differences as opposed to the similarities between different types of zeroing’ and did not ‘comport with prior and subsequent panel and Appellate Body jurisprudence’ (Flett, 2010: 300). It seems reasonable to assume that Hillman drafted this statement, which, if correct, would suggest that the US had influenced the endogenous preferences, and the behavior, of an ABM through the selection process.
The PA interaction between the US and Hillman, however, does not end here, since the four-year renewable term of ABMs provides an opportunity for the nominating countries like the US to respond to the behavior of members on the bench. Indeed, at the time of writing, the USTR decided not to re nominate Hillman for a second term. 72 Although the US did not indicate the reasons for its decision, it seems likely that USTR was unhappy with Hillman because she did not more often and forcefully dissent from decisions against the US. This unusual failure to re nominate a member is testament to growing politicization of the WTO judicial appointment process, and a sign that the nomination process can potentially be used not only to shape the preferences of members ex ante, but also as an ex post warning to sitting members about independence from the governments that nominated them. If this is indeed correct, we might expect ABMs in future cases to pay greater attention to the preferences of the states that nominated them, including a possible increase in the number of separate opinions in politically sensitive cases.
Conclusion
This article has put forward and tested a theory of international judicial appointments drawn from PA models of domestic and international politics. By contrast with trusteeship models that depict international courts as beyond the influence of their member-state principals, PA models predict that principals can and will utilize their individual right to nominate and their collective right to appoint in order to influence ex ante the endogenous preferences of international judges. We then derived four hypotheses about the behavior of states and of judicial candidates in the appointment process, which we tested in the context of a particularly powerful court, the WTO AB. The empirical evidence strongly supports our hypotheses.
We have found ample evidence that WTO Members seek to support candidates whom they consider close to their own positions on salient issues (H1 supported). Looking more closely at strategic interaction over the two-stage process, we find that nominating states have behaved as strategic agenda-setters, putting forward candidates whose positions are close to their own but who are likely to secure the support of the membership as a whole (H2 supported).
We have also seen that institutional features affect the influence of various parties in systematic ways. By contrast with full representation courts such as the ECJ, the selective representation and consensus rules of the AB empower the electorate at the expense of nominating states, providing each member with a potential (if diplomatically costly) veto against candidates perceived to be unhelpful to their interests (H3a and H3b supported). Furthermore, as we have seen, the informal rules on geographic distribution of members, and in particular the de facto permanent seats for the EU and US, provide the latter with greater agenda-setting power compared to other nominating states (H3c supported).
Turning to the role of information, we have demonstrated how candidates to the AB have behaved strategically, providing selective information to national delegations about their views on hot-button issues of interest to particular Members (H4a supported). In addition, we have seen evidence of WTO Members actively coaching and preparing their candidates to run the gauntlet of ambassadorial interviews. The result is a system that resembles nothing so much as the process of appointment to the US Supreme Court, where successful candidates are likely to be those who say little or nothing controversial that might alienate the principals who hold the keys to judicial office.
This study has also documented how scrutiny has increased progressively over the 15 years since the establishment of the WTO. Members have reacted to evolving case law and increasingly look for evidence of candidates’ positions on controversial issues likely to come before the AB (H4b supported). Indeed, some characteristics that were acceptable in the selection for the original seven (e.g. views on filling gaps) have become a decisive factor against nomination and appointment in subsequent waves in light of evolving WTO case law.
We should, of course, be cautious about declaring support for our hypotheses based on a relatively brief experience of judicial nomination to a single international court, but the pattern of findings in this case, and the remarkable consistency of views among a wide range of participants, reinforces our view that member-state principals seek to shape the preferences of international jurors in much the same way that political principals within states seek to shape the preferences of domestic judges, and for the same reasons.
Looking beyond the WTO, we expect our hypotheses about states’ use of their judicial nominating power to ‘travel’ to the vast majority of international courts for which member states nominate judges. Put simply, we would expect Hypotheses 1 (state interests in securing nominees with preferences similar to their own), 2 (strategic behavior by states in seeking the appointment of such nominees), and 4 (the significance of information and its strategic use by states and candidates) to apply across the full range of international courts and tribunals.
Hypothesis 3 and its corollaries, by contrast, allow us to make more specific predictions about variations in state and candidate behavior across international courts as a function of each court’s institutional rules, including in particular the distinction between full representation and selective representation courts. In this context, it is striking that a significant number of international courts have selective representation schemes similar to the AB, including several global courts (e.g. the International Court of Justice, International Tribunal for the Law of the Sea), all international criminal courts (International Criminal Tribunals for Yugoslavia and Rwanda, and the International Criminal Court), most regional human rights courts (the Inter-American Court of Human Rights, the African Court of Human Rights, and the ECOWAS Court), and several economic courts (OHADA, COMESA, ECOWAS, and ASEAN). Based on our theory and our findings here, we would expect states in these regimes to use their nomination and selection powers to shape the views of the judges they appoint, although it is possible that the lower numbers and/or the lower stakes of cases before these other courts provide states with less incentive to invest time and effort in the politics of judicial appointment. Further research into these and other cases would allow us to determine how common, or uncommon, are the politicized appointment procedures we have uncovered in the WTO case.