Counterfactuals and Contingency in WTO Dispute Settlement History

Abstract With the benefit of hindsight, much scholarship across political science, law, and economics has told the story of the international trade regime as if it had been pulled all along by a definite aim. By contrast, this article emphasizes the contingent aspects of the trade regime's development, looking especially to its dispute settlement mechanism. The very creation of the Appellate Body had by no means a certain outcome, and once created, the tribunal's evolution was largely unanticipated by states. An often-overlooked actor played a key role in that development: the WTO Secretariat. Drawing on recent findings, this article lays out the full extent of the Secretariat's role in dispute settlement, which remains largely hidden from view, and deliberately so. From appointing adjudicators and managing their remuneration, to providing them with legal arguments and drafting final rulings, the Secretariat of the WTO looms larger than in any comparable tribunal. Making its influence more transparent, I argue, would go a long way to returning the system to the shape it was designed to have at its outset.


Introduction
The study of international economic law has a distinct teleological bent. With the benefit of hindsight, much scholarship across political science, law, and economics has told the story of the international trade regime as if it had been pulled all along by a definite aim. The resulting impression is that what started as an informal fallback arrangement among two dozen states in 1947 was meant to become, through half a century of successive trade rounds, a highly legalized multilateral system encompassing nearly all of world trade, amounting to a quasiconstitutional international order.
Economists thus predicted that less productive import-competing interests would contract as they were exposed to global competition, while more productive, export-oriented interests would prosper, leading to a natural shift away from protectionism, and towards openness. Political scientists expected that the growing ambitious reciprocal commitments that states exchanged would require stricter enforcement mechanisms to be made credible. Diplomats would make way for lawyers. Loopholes would appear in one period and be closed in the next, as the degree of 'precision, obligation, and delegation' to international adjudicatory bodies rose. 1 The incomplete contract would gradually grow more complete. Reflecting similar ideas, trade negotiators invariably invoked the 'bicycle theory', according to which the trade regime, like a bicycle, would fall unless it was given constant forward motion. 2 In its most bullish version, such continuous liberalization was seen as inevitable. As a popular undergraduate political science textbook from 2015 expressed it, 'such a trend, moreover, is self-reinforcing and hence unlikely to be reversed'. 3 Recent events have shown that reversals in global governance are not so unlikely after all. The front wheel has been taken off the bicycle. As the trade regime's original champion has gone about disabling its vaunted dispute settlement mechanism by blocking further appointments to the Appellate Body, the field of international economic law has been forced to reassess the arc of history, and the extent to which it bends towards a definite point. This special issue of the World Trade Review is one part of that reassessment.
In the same spirit, this article emphasizes the contingent aspects of the trade regime's development, looking especially to its dispute settlement mechanism. I insist on how unexpected that development was, and how the key actors themselves did not anticipate the emergence of a fullyfledged court-like body. An often-overlooked actor played a key role in that development: the Secretariat. The extent of the WTO Secretariat's role in dispute settlement remains little knownthe figure of the judge usually captures most of the attention paid to tribunals, both domestic and international. Yet in the case of the WTO, focusing on adjudicators misses most of the action.
Indeed, the Secretariat's influence at every step of the WTO's dispute settlement process is now more significant than for any comparable international tribunal. It is WTO staff who help appoint the panelists. It is they who first summarize the facts of the dispute and lay out a proposed solution for adjudicators, before the latter ever meet to discuss a case. It is they who draft the questions that adjudicators ask the litigating parties. And as recent empirical findings demonstrate, it is the Secretariat staff who 'hold the pen' when drafting the final ruling: in fact, text analyses using authorship detecting tools suggest that they play a more significant role in that drafting than the adjudicators themselves. 4 The Secretariat did not always wield such influence. The change came as a result of a decision by Member States in the early 1980s. The appointment of Secretariat staff as 'assistants' to adjudicators was how governments attained a particular equilibrium: it was a way of preventing the outright mistakes of law that had begun creeping up in rulings as disputes dealt with increasingly complex issues of law, while continuing to rely on ad hoc panelists drawn from the ranks of trade officials who would be more deferential to state interests. Its nominal role as an 'assistant' notwithstanding, the Secretariat was thus conceived from the outset as an agent of states, rather than adjudicators. This, by itself, distinguishes the WTO Secretariat staff from the permanent staff and clerks working within most other tribunals, domestic or international. Adjudicators are usually the ones to appoint their assistants; in the case of the WTO, the reverse is true.
A full appreciation of the Secretariat's functions throws into question the very nature of WTO dispute settlement. Instead of the 'World Trade Court', as many scholars, and even some of the founding AB members, sought to portray it, 5 WTO dispute settlement was designed from the 2 As James Bacchus put it: 'We all know about this bicycle. We all talk about it every time we talk about the future of the WTO. We all ride it every time we try to make the case for the future of the WTO' (Bacchus, 2003). The bicycle theory is one of perpetual liberalization. Attached to bicycle theory is a theory of judicial development. Jurisprudence would accumulate over time, as it tackled new issues that were not foreseen by the original drafters. According to a colleague, Bacchus himself would routinely refer to AB members as 'judges', 'much to the amusement and raised eyebrows of the participants' (Ganesan, 2015, fn 9). 3 It happens to be the textbook through which McGill undergraduates are introduced to political science (Mintz et al., 2015, 440). 4 See Pauwelyn and Pelc (2022a). This article draws and expands on the findings of this and two recent related articles. See Pauwelyn and Pelc (2022b,c). 5 The World Trade Court descriptor was often invoked, especially in the WTO's early days, including by some AB members. See, for instance, C.-D. Ehlermann (2002), Van den Bossche (2005). 'The Appellate Body is now, in all but name, the World Trade Court.' See also, below, Weiler (2004), calling for an updated nomenclature reflecting the de facto judicial nature outset as something far more modest: a process of review conducted by an international agency staffed by a group of experts. Its decisions were always intended to be as much technocratic outcomes as judicial ones. As the institution strayed from that intended model, the irritation of members such as the US grew.
As is often the case with an agent imbued with vaguely circumscribed power by its principals, the WTO Secretariat outgrew its initial mandate. Its numbers swelled, from the three lawyers appointed to the Legal Affairs Division in the 1980s to the more than 90 permanent staff, devoted to dispute settlement, by 2022. Independent, and largely insulated from Member oversight, this legal bureaucracy developed their own distinctive preferences and practices. These practices had effects of their own, contributing to a lengthening of rulings and proceedings; a tendency towards expansive reports that went beyond what is needed to settle the dispute at hand; a practice of extensively invoking past precedents, and a stifling of dissenting opinions.
These effects lead to the article's conclusion, which draws out one striking implication of the argument, which is that the Secretariat's expansive influence may have unwittingly contributed to the very trends Member states, such as the US, have criticized in obstructing the system. 6 Ambassador Katherine Tai, the United States Trade Representative, recently traced the historical arc of the institution in these terms: It started as a quasi-diplomatic, quasi-legal proceeding for presenting arguments over differing interpretations of WTO rules. A typical panel or Appellate Body report in the early days was 20 or 30 pages. Twenty years later, reports for some of the largest cases have exceeded 1,000 pages. They symbolize what the system has become: unwieldy and bureaucratic. 7 What such criticisms overlook is that the bureaucracy in question emerged as a result of the efforts of governmentsthe US chief among themto retain political oversight in the process. Members delegated power to a bureaucracy to check the efforts of ad hoc adjudicators who were often out of their depth in the face of the growing complexity of disputes. The demise of the 'World Trade Court', in this sense, is not a tragedy, where the main actors are inexorably drawn to their downfall through hubristic ambition; it is a comedy, one where the actors miscalculate and stumble and try to right themselves, all the while complaining about the unanticipated chain of events their own actions have led to.

What Did Governments Envision?
The sheer attraction of functionalism likely accounts for the prevalent tendency among students of the WTO to attribute its successes to strategic decisions by governments. In this way, the creation of a quasi-judicial institution delivering binding rulings on sovereign states is perceived, after the fact, as having been the intent all along. Optimistic accounts of farsighted governments suggested that the drafters of the treaty knowingly tied the hands of an acquiescent US Congress, so that it might not fall prone to the siren song of domestic protectionist interest groups (e.g. Thompson, 2007). According to this view, the creation of the Appellate Body is the natural consummation of a widespread move to international legalization: an independent court-like body, steeped in the Vienna Convention's rules on treaty interpretation, reflecting at length on its own of the dispute settlement body: 'My preference would be for an official namethe International Court of Economic Justiceand a diminutivethe World Trade Court.' 6 For a summary of these critiques, see Davey (2022), in this volume. 7 Ambassador Katherine Tai's Remarks as Prepared for Delivery on the World Trade Organization, https://ustr.gov/aboutus/policy-offices/press-office/speeches-and-remarks/2021/october/ambassador-katherine-tais-remarks-prepared-delivery-worldtrade-organization jurisprudence through its binding rulings, whose findings are all but irreversible. 8 These accounts have obvious appeal to scholars who saw in the WTO's dispute settlement system an instantiation of the Pareto improvements envisioned by institutional theory.
Yet a closer look at the rules themselves, and the actors at the heart of the regime, suggests that the emergence of the dispute settlement system in its current form was far from obvious. It also goes a long way to explaining the dissatisfaction with the status quo that the US has been voicing. The US Congress has always been, and remains, particularly resistant to international courts second-guessing US policy. It insisted all along that the AB should be invoked only exceptionally, a feeling that was shared by many Members. As the US representative asked back in 1988 during negotiations over the design of the dispute settlement mechanism, 'How might we ensure that the review process is used only in extraordinary cases … ?' 9 Even Canada, credited for first proposing creating an Appellate Body, kept arguing that it was an option that should only be invoked in 'rare cases'. 10 In fact, Canada explicitly insisted that '[t]he intent would not be to have appellate review become a quasi-automatic step in the dispute settlement process'. It was reserved for cases where a party considered that the ruling was 'so fundamentally flawed that it should not be accepted '. 11 In a recent overview, Cottier (2021) points out the many design features of the Appellate Body that hint at how a full-fledged judicial branch was never the intent of drafters. Among these, AB members are employed part time, more akin to consultants than a sitting bench. They are not expected to reside in Geneva, and often fly in for deliberations (meanwhile, Secretariat staff are permanent employees with pensions who live in Geneva and devote themselves full time to WTO dispute settlement, which also accounts for the growing gulf in experience between the two sets of actors). Tellingly, the clock for appellate review was set at three months, suggesting how narrow the appeals were expected to beduring negotiations, the proposed time-limit was initially even shorter, at 60 days. 12 One might add to these design features the amount of discretion that AB members were given over the actual 'judicial' process: in the months after the founding bench of the AB members was selected, these seven individuals had to come up with their own working procedures. These covered such key matters as how AB members would be appointed to a given case, how dissenting opinions would be treated, how internal deliberations would be conductedall of which had been left open by negotiators. 13 The AB's Working Procedures hint at the role of contingency over historical necessity. The process was not the reflection of the strategic thinking and bargaining power of governments defending their commercial interests, it was instead left to the whims of a half dozen individuals. In Pauwelyn and Pelc (2022b), we look at how the AB sought to use these Working Procedures to insulate itself from political oversight, by committing themselves to unanimous opinions (while the texts explicitly allowed dissents) and to a practice of 'collegiality', whereby each division of three would consult with the remaining fourmaking it harder for governments to censure specific AB members. If one were to indulge in some counterfactual history, one might speculate that had states specified the AB's working procedures in the treaty texts, adjudicators would have been more exposed to political control, might have been less bold in their rulings as a result, and the recent political backlash might not have occurred. 8 Member States did retain the legislatively overriding AB rulings through 'authoritative interpretations'. This option has never been used. See Creamer and Godzimirska (2016); see also, Lamp (2021). 9 GATT doc MTN.GNG/NG13/W/40, pp. 5-6. 10 GATT doc MTN.GNG/NG13/W41: 'In rare cases, where a party to a dispute considered, despite the review by the panel, that a report was so fundamentally flawed that it should not be accepted, the GATT dispute settlement system should provide for a means of correcting errors. The addition of an appellate mechanism would serve that purpose.' 11 Ibid. 12 See GATT Doc MTN.GNG/NG13/W/41. 13 The Working Procedures were then amended six times after 1995. For the most recent version, see 'Working procedures for appellate review', https://www.wto.org/english/tratop_e/dispu_e/ab_e.htm In sum, the very creation of the AB itself was by no means a certain outcome, and once created, its evolution was not anticipated by states: it was never meant to become today's quasi-automatic, wholesale review of panel rulings. 14 Elsig (2017) quotes negotiators themselves as being taken aback by the very outcome of the Uruguay Round: 'I couldn't believe it myself. Why had the US accepted the creation of the Appellate Body?' Another participant claimed to be 'very surprised to see agreement on issues such as automaticity in accepting rulings'. In his own interviews, Weiler (2004) describes the 'incredulity' of some country delegations that had arrived at the sudden realization, 'we have created a court'. Van den Bossche (2005) aptly sums up the transformation that then took place in the first years of the WTO: the AB, which came very close to never seeing the day, went 'from afterthought to centerpiece'.
Scholars have been complicit in this latter transformation. Political scientists latched onto the WTO and its tribunal as the epitome of a widespread trend of legalization and judicialization in international affairs. 15 Students of constitutional law were quick to see in the dispute settlement mechanism the emergence of a constitutional legal order at the international level. In fact, legal observers like Weiler (2004) explicitly pushed for the de jure recognition of the shift. Governments needed to come to terms with what they had wrought, and call things by their names. Weiler derided the fiction maintained for the sake of members like the US, whereby a plainly judicial process was referred to as 'dispute settlement,', rulings were 'reports', and judges were 'panelists' and 'AB members.' While recognizing its purpose -'the nomenclature is convenient for some domestic constituencies, notably the US. Congress'he favored calling the mechanism by a name that reflected its de facto function: 'the International Court of Economic Justice'. In the last five years, we have been reminded of the political import of conventions, and the associated risk of ripping off masks in this way.
Weiler's sentiment was widely shared among students of the institution. The World Trade Court would emerge, whether government liked it or not, since the push towards greater integration required it. In securing the gains from trade, countries would necessarily have to match the ambition of the concessions they tradedwhich now ran to behind-the-border measures that increasingly encroached on realm of traditional sovereign controlwith the ambition of increasingly legalized enforcement mechanism. This also had a sociological dimension, as international relations grew ever more 'lawyerized', and the practices and assumptions of the legal profession gradually replaced those of professional diplomats. 16 The outcome was overdetermined: global commerce would henceforth play out on the stage of public international law. The danger of confounding contingency and necessity, as events have shown, is that it blinds scholars to the possibility of reversals.
Along these lines, since the very inception of the WTO, there have been sustained calls for a more professionalized corps of panelists. Hudec (1999) claimed that "[i]f panels were composed of members with such legal expertise, one would no longer need to worry about the undue decision-making power of the panel's legal staff." Later, Weiler (2004) explicitly called for more expert panelists: 'The professional backgrounds of panelists should be commensurate with the gravity and profundity of the issues decided in a globalized world. This has conspicuously not been the case in some of the most important instances.' Some of my own work added empirical support to these claims: in joint work with Marc Busch, we demonstrated how panelist experience mattered: in particular, rulings presided over by panel chairs with lower experience were more likely to be appealed, and overturned on appeal, than those presided over by chairs with more experience (Busch and Pelc, 2009). Observers continue calling for more 14 See Davey (2022) on how narrow the role of the AB was initially designed to be, limited to reviewing '"bad" panel reports'. 15 See the special journal issues devoted to these trends, all of which held up the WTO's DSU as prime example (e.g. . For a later special issue contemplating the possible end of judicialization, see Abebe and Ginsburg (2019). 16 On the lawyerization of foreign affairs, see, e.g. Nuñez-Mietz (2016). professional panelists to this day. Earlier this year, Wauters (2021) argued that '[t]he problem is not so much the Secretariat but the ad hoc nature of the panel system and … panelists whose professional qualifications and part-time participation are simply ill-suited for writing the kind of legal reports that WTO dispute settlement panels are expected to issue'. Of course, Member-States have been well aware of these proposals. Yet they have consistently rejected them: governments continue to appoint ad hoc panelists. The reasons are political. They view such individuals, often pulled from the ranks of former trade officials, as more likely to be sensitive to their policy concerns.
Of course, claiming that the views of academics played much of a role in the evolution of the postwar trade regime might itself be derided as scholarly conceit. The point should not be overstated. It is merely to say that scholars had reasons to see the regime as something it was not, and that states never intended it to become. This may have kept them from realizing how adamant Member States were to retain the non-legal aspect of the dispute settlement mechanism, as in their insistence on ad hoc panelists. WTO scholars would gain from recognizing the intensity of these state preferences, which, as I argue next, largely account for the Secretariat's outsize influence over dispute settlement.

The Secretariat as the Solution to a Problem of Institutional Design
The idea of there being two fundamental judicial objectives in tension with one another is a point long made by the literature on courts. As Ferejohn and Kramer (2002) observed in the US context, 'We are told we cannot have it both ways. We can have a bench that is independent or a bench that is accountable, but we must accept a trade-off that sacrifices one or the other of these goals to some yet to be defined extent.' The same holds, to an even greater extent, at the international level. There, state commitments rely on the credibility of an enforcement body; yet sovereign nations are even more loath to delegate power to judges in international courts than executives are to delegate powers of review to domestic judicial branches. Domestic audiences have recently been especially averse to the idea of unelected foreign magistrates ruling over questions falling under traditional sovereign purview, as the British public's recriminations against the European Court of Human Rights make clear (Voeten, 2020). Dunoff and Pollack (2017) have proposed that non-identifiability of judges is one means of handling this dilemma. By obscuring the identity of individual adjudicators, they argue, international tribunals might maintain both a high level of transparency and a high level of political oversight. 17 In the 1980s GATT era, states settled on another. Following a series of legally problematic rulings, culminating with the 1981 panel report in Spain-Soyabean Oil, the Member States created the Legal Affairs Division, a dedicated part of the Secretariat which would be in charge of guiding panelists, and ensuring that they did not commit egregious mistakes of law.
They did so reluctantly. As Porges (2015) puts it, 'Governments that did not want to have a group of lawyers in the Geneva headquarters telling them what they could do … had run up against the limits of what was possible without lawyers.' In the face of growingly complex cases, appointing a body of legal experts to supervise the work of often legally underqualified panelists had become a necessity. The creation of the Legal Affairs Division was thus the solution to a tricky problem of institutional design. There was a prevalent concern over the quality of rulings. Yet governments were loath to give up too much control by appointing professional judges. Appointing one agent to check the other was the compromise that states settled on. Later, the US pushed for the creation of the Rules Division, a distinct part of the Secretariat devoted specifically to the aspect of trade rules the US cared most about: trade remedies. 17 It is worth noting that transparency is not, in and of itself, an inherent value of tribunals, in contrast to independence and accountability. The identifiability of judicial opinions has well studied costs and benefits, and a tribunal lacking in identifiability need not suffer in its essential functions. In this way, tribunals often deliver per curiam decisions, reached as a bench, with no individual judicial voices identified, and they suffer no loss in legitimacy or effectiveness for it. This origin story underscores the difference between WTO Secretariat staff and the clerks or assistants in other tribunals. Its nominal role as an 'assistant' notwithstanding, the Secretariat was always conceived of as the agent of states, rather than that of adjudicators.
This discrepancy between the WTO and all other comparable international tribunals becomes clearer still when one considers the tasks the Secretariat began taking on in the WTO era. Today, Secretariat staff are present from the start of the dispute settlement process through to its end. They are responsible for proposing panelists for specific cases. 18 They then play a crucial agendasetting function, writing what is called an 'issues paper', which sets out the facts and possible resolutions of the case, and draws out the most relevant jurisprudence. Crucially, adjudicators receive the issues paper before they ever meet to discuss the case. It is the first word on the direction the dispute might take. Staff are also the most active actor throughout the deliberations: they sit in the room during meetings with the parties. They even draft questions for the adjudicators to ask the parties. Finally, they are the ones who 'hold the pen' when it comes to drafting the actual ruling.
This last claim has often been made anecdotally, without being demonstrated empirically. In joint work with Joost Pauwelyn, we do just that. We review all panel rulings for which we can identify the specific Secretariat staff who were assigned to the dispute. We are able to do so for all disputes up to DS302, at which point the identity of staff stopped being publicized. This is one of the few instances of the WTO becoming less transparent over time; tellingly, it reflects a decision on the part of the Secretariat itself, rather than Member States. We then search for outside texts written by the staff and the panelists, and we build stylometric profiles for each individual using these texts. Comparing these profiles to the texts of final rulings, we are able to determine with high probability who the most likely authors are, or alternatively, who are the authors with the greatest likely influence over the drafting of the texts. In the supermajority of cases, the Secretariat staff's apparent influence is greater than that of the adjudicators themselves.
This comports with a great deal of anecdotal evidence. 'Many panelists have told me that they feel they cannot meaningfully challenge the legal secretary on points of law,' writes Weiler (2004). As he goes on, 'this is no secret. The only novelty, if any, is putting it on paper so explicitly'. By analogy, the novelty of our findings has been to demonstrate empirically what Geneva insiders have long been aware of. It also allows us to show evidence for more detailed claims on how this institutional power manifests itself: for instance, we find that the Secretariat's fingerprints are more apparent for disputes of systemic importance. 19 And those disputes where those fingerprints are most apparent also go on to become more influential in jurisprudence, as measured by subsequent citations. The Secretariat's influence over jurisprudence, in other words, is self-reinforcing.
In sum, although they were initially conceived of as 'assistants' of the tribunal, the Secretariat eventually became what insiders now refer to as the 'guardians of the system'. They are 'the repository of institutional memory, of horizontal and temporal coherence, of long-term hermeneutic strategy'all functions that are the traditional purview of adjudicators themselves (Weiler, 2004).
As 'guardians of the system,' the Secretariat staff have likely contributed to a number of trends that have become salient in the midst of the US' pushback against the WTO's dispute settlement system. Chief among these is the status of precedent. Public international law does not recognize stare decisis: past precedent is not considered binding, a fact the US habitually reminds other Members of. Yet as a number of studies have demonstrated, most of the relevant actors nonetheless behave as if precedent were in fact binding: states invest in precedent-setting cases, they 18 For details of the appointment procedure and the Secretariat's role, see Pauwelyn and Pelc (2022c). 19 We proxy for systemic importance by looking to the number of non-litigant members who join as third parties, and how many of them invoke a 'systemic interest' (as opposed to a trade interest) when doing so. On third parties, their role in dispute settlement, and effects on outcomes, see Johns and Pelc (2014). appeal findings to obtain change of legal language, rather than outcome; industries worry about the precedent that specific findings might represent; markets react to findings against foreign that have implications for countries who were not among the litigants. 20 The Secretariat's influence has likely played a role in this rise in status of precedent. Issue papers largely consist of pointing to the most relevant existing jurisprudence. Through this agenda-setting function, past reasoning constrains legal reasoning in the case at hand. Interviews with former staff and AB members show how 'consistency' is the value that Secretariat members prize most highly. In particular, the director of the AB Secretariat was known for paying inordinate attention to consistency: his arguments are generally perceived as stemming from a passion to safeguard institutional respectabilityin particular, ensuring that new rulings follow principles set forth in prior cases … His overriding goal, in other words, is that the Appellate Body should be consistent. 21 For analogous reasons, the Secretariat staff have a systemic outlook; they are less narrowly focused on 'settlement' of the dispute at hand than the adjudicators. As a result, they have greater incentives to complete legal reasoning with an eye to other ongoing or future disputes. This makes them, as per the criticism of one outgoing AB member, more prone to 'over-reach, gap filling, and advisory opinions'. 22 Finally, the Secretariat has been associated with a push for consensus opinions, and the deterrence of dissenting opinions. It follows that as 'guardians of the system', the Secretariat have a greater interest than anyone in rulings presenting a united front, and downplaying discord between adjudicators. Over time, adjudicators themselves have chafed against such pressure, and the resulting 'groupthink' it has brought about. The aforementioned AB member thus faulted the Secretariat for 'an excessive striving for consensus decisions coupled with a discouragement of dissents'. Such insistence, he went on, 'led to excessively long and unclear compromise reports'. 23 These effects, attributed to an empowered legal Secretariat, are especially notable in a moment when the US has been blocking the dispute settlement mechanism, citing just these trends as its justification. As per the quote from USTR Katherine Tai from the introduction, the US has been critical of the expanding length and scope of the rulings, and the 'unwieldy and bureaucratic' nature of the process. The US has also long denounced the reliance of panels and the AB on past disputes as precedent, insisting that the WTO texts do not allow for any binding effect of past decisions. Claiming that that the overt reliance on precedent 'usurp[s] the authority expressly reserved to Members', they have accused WTO adjudicators of engaging in law making. 24 What these criticisms miss is that one factor contributing to all these trendsfrom the reliance on precedent to the expanding scope and length of rulingsis the very bureaucracy that the US and other members put in place back in the 1980s. That bureaucracy has served its purpose: egregious mistakes of law have been avoided. But there have also been unanticipated effects, which Members like the US are now inveigling against. The picture that emerges is not one of farsighted legislators tying their hands through optimally designed international agreements, ir is one of governments reacting fitfully to problems as they arise, and then lashing out when the solutions they come up with later prove to have unforeseen effects. 20 See, respectively, Bhala (1998); Pelc (2014); Kucik and Pelc (2016a); Daku and Pelc (2017). 21 Blustein (2017)

Conclusion
There has been a renewed surge of interest around contingency in international legal history. 25 This latest push has come predominantly from critical legal scholars who are reacting against a prior emphasis on structure and necessity, which they fault for foregrounding the most powerful actors, to the detriment of those marginalized due to status, class, or wealth discrepancies. 26 Trained to pay attention to context, and eager to imagine alternative historical paths, these legal historians have also been especially alert to the distorting role of scholarship itself in international law, when it embarks on projects of self-legitimation. 27 These studies hold useful lessons for a field of international economic law that has held up the WTO's enforcement mechanism as a shining exemplar for their theories, and which may have led some to overlook the many ways in which the institution did not comport with those theoretical expectations.
Yet a similar embrace of contingency has long been present in positivist social science, too. Warning against a functionalist tendency to interpret the effects of institutions as the reasons for their creation, scholars like Elster (2000) have long insisted on the distinction between essential constraints, which are designed ex ante for their anticipated effects, and incidental constraints, which arise in an unplanned way. The first emerges by design; the second might be retained ex post, once they reveal themselves to be beneficial. In a common form of this fallacy, scholars mistake incidental benefits arising from the creation and development of international institutions as having been designed all along with those benefits in mind. Devices meant to bind others become self-binding devices; enforcement bodies with limited mandates become calculated investments in public international law.
One of the first facts students learn about the GATT is that it was born of failure, as the US Congress refused to ratify the agreement of the International Trade Organization. As I have emphasized in this article, such haphazardness remained over the next half-century. Three decades in, one fix to the growing tension between legal autonomy and political oversight yielded an empowered Secretariat staff. Yet this body was itself largely insulated from government oversight. It eventually took over all aspects of dispute settlement, from appointing panelists to providing them with summaries of cases and questions for the litigants, and even drafting their final reports. A closer look at the WTO Secretariat, and its role in the development of the WTO's dispute settlement body, demonstrates just how contingent its development was. Insofar as it has come to bind states through a highly developed jurisprudence, it constitutes an incidental constraint far more than an essential one. When that jurisprudence began encroaching on the stated intentions of Member States, some of them began lashing out.
The expansive role of the Secretariat has had observable effects. The fingerprints of its staff are more discernible in the final rulings than those of adjudicators themselves. The rulings in which staff, as opposed to adjudicators, appear to have played an especially large role are also those that go on to have the biggest impact on subsequent jurisprudence. What is more, a number of the practices that the Secretariat staff developed over time as 'guardians of the system' may be to blame for the trends that WTO members like the US have recently taken to denouncing. Protracted proceedings; highly elaborate rulings that go beyond the resolution of the case at hand; a high investment in precedent, and an exceptionally low rate of dissent: all of these are trends that have been associated with an empowered Secretariat. They are also the trends that the US has been citing as going against the original intent of the institution's drafters. This is not to say that the Secretariat's role is deleterious on the whole. In fact, the accumulated experience of the Secretariat is the main reason offered by some observers for a longstanding puzzle, which is that WTO members consistently send their regional disputes to the WTO, rather 25 For a recent examination of contingency across a range of international legal settings, see the edited volume Venzke and Heller, 2021, on 'the possibility of different legal histories'. 26 For an especially cogent take on the various forces pulling towards contingency vs. necessity, see Renard Painter (2021). 27 See Koskenniemi (2001) onwards.
than to the dispute resolution mechanisms contained in their own preferential agreementseven as these have most often been copied on the WTO's DSU (Cottier, 2021). Governments, in other words, continue to see value in the oversight role played by the Secretariat.
How might one retain the many benefits of a highly experienced Secretariat, while avoiding some of the unanticipated trends an empowered bureaucracy might bring about? One solution might be the opposite of that advocated by Weiler back in 2004. Rather than updating the nomenclature to match the de facto function of the ABand triumphantly referring to the process as the 'International Court of Economic Justice'it might be wise to be true once more to the original nomenclature. Rather than the absolute rulings of an independent international tribunal, it may be worth recalling that these were meant to be reports and recommendations: technocratic outcomes as much as judicial ones.
Making the role of the Secretariat more transparent would go a long way to returning the system to the shape it was meant to have had at the outset. In this respect, one might require at a minimum that the Secretariat's issues papers be circulated among the litigants and third parties to a given dispute, who would have an opportunity to respond, and then be made publicly available together with final panel and AB reports. 28 This, by itself, would exert a disciplining effect on the content of these opinions, making any form of bias less likely. 29 After all, the Secretariat is more invested in the good functioning of the system than the average AB member. Just as importantly, such transparency would make plain the true nature of the process: not so much an international constitutional court, as something more akin to a process of administrative legal review. One that has been admirably successful at its professed goal: the settlement of interstate disputes.