Who Guards the “Guardians of the System”? The Role of the Secretariat in WTO Dispute Settlement

Abstract

For all the attention paid to the panelists and Appellate Body of the World Trade Organization (WTO), the Secretariat plays an overlooked and increasingly important role in the dispute settlement mechanism (DSM), including in selecting panelists, writing “issue papers” for adjudicators, providing economic expert advice, participating in internal deliberations, and drafting actual rulings. This Article argues that, given these functions, the DSM is better understood to be a sui generis administrative review process than it is a “World Trade Court.”

I. Introduction

From the moment of its inception in 1995, the World Trade Organization's (WTO) Dispute Settlement Understanding (DSU) has functioned as a first-rate scholarly specimen, serving the purposes of the various academic tribes studying it. Legal scholars held it up as the consummation of the ambitions of public international law: an independent tribunal with compulsory jurisdiction, tasked with adjudicating legal disagreements between sovereign states, and generating its own expansive jurisprudence.¹ Political scientists looking for cases of successful self-binding by democracies saw in it the rational delegation of power by sovereigns to an international body: a means of managing the demands of domestic interest groups, while defusing interstate power politics.² Economists recognized it as a means of dealing with a necessarily incomplete contract, and rendering credible the threat of retaliation in a way that increased the odds of cooperation between trade partners.³

In this Article, we present the WTO's “crown jewel” in a new light, by looking to the largely overlooked role played by its Secretariat, as compared to that of the appointed adjudicators themselves. Drawing on archival and anecdotal evidence, and on a set of recent empirical findings, we argue that the permanent staff of the WTO's Secretariat has played a more significant role within the organization's development than the staff of any comparable state-to-state tribunal. Looking to the history of the institution, we show how this was partly by design: following a series of problematic rulings in the 1980s, member states construed the Secretariat's Legal Affairs Division as a body tasked with holding “rogue” panels in check. Rather than an agent of adjudicators, akin to the role staff traditionally play in other tribunals, the WTO Secretariat was an agent of governments from its beginnings, designed to impart legal accuracy and institutional memory on diplomat-adjudicators, and guard the system against the risk of egregious rulings. Yet as we suggest, the agent may have outgrown that initial mandate, owing to a number of unanticipated structural features which led, over time, to an increasing shift of power and responsibilities from adjudicators to Secretariat staff, in a way that governments could never have anticipated when they first established a Legal Affairs Division staffed by three lawyers back in the 1980s. Now counting more than ninety staff, of whom as many as a dozen may be assigned to a given dispute, the Secretariat's experience dwarfs that of the adjudicators they are tasked with “assisting.” As we explain, this gap in expertise only keeps increasing with the accrual of WTO jurisprudence: permanent staff in situ in Geneva accumulate knowledge that adjudicators employed part time, and who most often fly in for hearings, cannot hope to match.

Whether intended or unanticipated, the WTO Secretariat's expansive role holds considerable implications both for practice and theory. We argue that it has left the organization open to concerns over (external) legitimacy and a lack of transparency and accountability. It has also affected legal outcomes in several respects, from increasing reliance on precedent and obiter dicta to elongating proceedings and expanding the scope of rulings. These effects are especially notable, since they coincide with the chief criticisms recently directed at the dispute settlement system by the United States and other members. In this sense, the institution's current impasse can be linked back, in part, to the Secretariat's role in dispute settlement. While there has been scarcely any explicit criticism of the Secretariat by member states—an indication that internal legitimacy may be largely intact—the remarkable irony is that “the guardians of the system” may have contributed to the system's demise by the expansion of their influence.

As one prominent WTO practitioner recently put it: “[i]t would be wrong to blame the Appellate Body (AB) crisis on the Secretariat but to ignore its role in the functioning of the WTO dispute settlement system would not be correct either . . . . The role of the WTO Secretariat in WTO disputes cannot be underestimated and merits more attention than has been given to it.”⁴ In this Article, we give the Secretariat the attention it deserves. The result is a wholesale reevaluation of what has often been described as the most successful international tribunal in existence, and of its partial demise.

Indeed, the broadest implication of our examination of the Secretariat's overlooked influence is to reassess the very nature of WTO dispute settlement. Rather than the “World Trade Court” that some of its earliest observers portrayed it as and actively pushed for—an independent international judicial body⁵—our account recasts the DSU as something else entirely: a sui generis process of international administrative review that (1) rather than operating independently, as in separation of powers judicial review, remains part of the WTO as a members-driven, diplomatic organization and (2) where decisions are not so much made by individual judges or an independent court, but institutionally crafted by both diplomat-adjudicators and in-house Secretariat staff and experts. The result is as much a technocratic exercise as a judicial one. This recasting of the very nature of WTO dispute settlement largely accounts for the influential role of Secretariat staff. Even so, adjustments may be appropriate. Among these, we list the possibility of term limits for Secretariat staff, and an increase in accountability and transparency surrounding their functions and output.

II. The WTO Secretariat: Where We Are, and How We Got Here

The long-time director of the AB Secretariat, who held his post from 2006 until recently, has been described as “arguably the most powerful international civil servant that nobody has ever heard of.”⁶ In the last days of the AB’s existence, in December 2019, the then AB chair openly called for the staff director’s resignation, arguing that he “sought to guide appellate body members toward rulings that coincide with his views.”⁷ Elsewhere, it was reported that the director of the AB Secretariat “personally oversees every dispute . . .[,] meaning he reads and revises all issue papers—identifying the issues at question based on disputing members’ submissions—and draft reports before they are sent to the Appellate Body members. He is involved in nearly all the discussions during the appeals.”⁸ Shortly after the AB became dysfunctional following the blocking of new AB appointments by the United States, the AB Secretariat was also dismantled, and all budget to the AB and its Secretariat frozen. The AB staff director was moved and put in charge of a newly created Division for Knowledge and Information Management.

The above offers a glimpse of how things ended for the AB and its Secretariat. But where did they start? What explains the Secretariat’s expansive influence over WTO dispute settlement? We work backward through time to address this question, from the current status quo, back to the Secretariat’s historical origins.

WTO rules bestow a broadly defined role on its staff. DSU Article 27.1 provides that “[t]he Secretariat shall have the responsibility of assisting panels, especially on the legal, historical and procedural aspects of the matters dealt with, and of providing secretarial and technical support.” Notice how this provision explicitly includes assistance on the “legal . . . aspects” of a case. In practice, WTO panels are assisted by an increasing number of various WTO staff: lawyers from the Legal Affairs Division (LAD), or the Rules Division when the dispute concerns so-called trade remedies;⁹ experts in the substantive area at issue sourced from operational divisions, such as specialists in agriculture, sanitary measures or customs valuation; and economists from the Research Division, who play an advisory role on economic questions and may provide econometric expertise. In each dispute, one or more junior lawyers report to the lead lawyer who, in turn, reports to the director of the relevant division. As several divisions are involved, multiple directors supervise staff work and need to coordinate among themselves. Four deputy director-generals (DDGs) are each in charge of three to five divisions. Traditionally, the same DDG oversees both the LAD and Rules Division. At the head of the Secretariat is the WTO director-general (DG).¹⁰ The DG and DDGs do not normally get involved in specific disputes, with the exception of panel appointments, which we discuss below.

As pertains to the AB Secretariat, DSU Article 17.7 provides that “[t]he Appellate Body shall be provided with appropriate administrative and legal support as it requires.” Here, too, the staff’s mandate explicitly includes “legal support.” The AB thus has its own Secretariat, composed exclusively of lawyers and administrative support staff, which is separate and independent from the broader WTO Secretariat. It reports only to the DG for non-dispute-related administrative matters. Yet, all staff, including that of the AB, are appointed (and promoted) by the DG, without any formal role for WTO members, panelists, or the AB itself.¹¹ An AB division of three AB members, hearing a specific dispute, is assisted by a team of junior staff lawyers led by a senior lawyer, who in turn reports to the director of the AB Secretariat. When a dispute involves complex economic arguments or econometric evidence, the AB has also sought advice from economists in the Research Division. As a result, any given WTO panel or AB division is assisted by anywhere from four to more than ten WTO staff members, with one or more directors in supervisory roles.

A. Overlooked by Design

The view that WTO insiders have of the Secretariat's role in dispute settlement is aptly summarized by Thomas Cottier's recent assessment: “[I]t is obvious that . . . Panel reports cannot be the sole responsibility of three part-time panelists. It is shared with the WTO Secretariat . . . . The professionalism of the Secretariat and its institutional memory are the main capital in WTO dispute settlement. . . . There is absolutely nothing wrong with the role of the Secretariat, albeit many (including the Secretariat itself) prefer to hide such involvement.”¹² This preference for opacity matches our own assessment: whereas the WTO has opened itself up to the public in most respects, it remains (and in some respects has actually grown more) circumspect about the role played by the Secretariat, especially as pertains to dispute settlement.¹³ The intent, as per Thomas Cottier himself, who served as a panelist and chair on a large number of WTO disputes, is to maintain the “fiction” of rulings being delivered from on high by independent adjudicators, whereas in fact, a large team of permanent staff is devoted to handling increasingly complex disputes.

The sheer extent of this involvement is thus known to Geneva insiders, but comes as a surprise even to scholars who closely study the trade regime.¹⁴ It is in every way an open secret. Tommaso Soave (2019) has described it as an instance of “collective denial.” Below, we outline the extent of this involvement, before examining its far-reaching implications.

B. What Does the WTO Secretariat Actually Do?

Drawing on a review of formal and informal working procedures, written testimonies from the relevant actors, and conversations with both former and current Secretariat members and adjudicators at both the panel and the AB level, our analysis reveals that WTO staff play a substantive role in WTO panel and AB proceedings in eight major respects. These include: (1) the appointment of panelists; (2) control over the ad hoc financial compensation of panelists and AB members; (3) setting timetable and working procedures; (4) writing the “issues paper” that adjudicators see before they ever meet to discuss the case; (5) drafting the questions adjudicators ask the parties; (6) providing expert advice on non-legal issues (e.g., econometric evidence); (7) actively participating in hearings and internal deliberations; and (8) drafting of reports.¹⁵ Some differences exist between the Secretariat's role during panel and AB proceedings. Most importantly, adjudicators on panels are appointed for a given case; AB members for four years (renewable once). One might thus expect that ad hoc panelists would be guided more by staff lawyers—most of whom were already employed under the General Agreement on Tariffs and Trade (GATT)—while AB members, given their longer terms, would take more control, especially since the AB Secretariat was created at the same time as the AB itself. Yet, these differences have eroded over time. As time progressed, and AB members came and went, while AB staff stayed and accumulated experience, asymmetries between AB members and their staff increased and eventually became as important as those between ad hoc panelists and the staff lawyers assisting panels.¹⁶ Indeed, while a subgroup of panelists became increasingly expert (collecting frequent reappointments) thereby increasing their weight as against long-time GATT/WTO staff, AB members, in contrast, remained more often than not former trade negotiators, not law professionals,¹⁷ whose legal expertise was increasingly dwarfed by AB staff, especially senior staff that served the AB since its inception in 1995. With the exception of adjudicator appointment, discussed below, today, it can be said that “[t]he work of the Appellate Body Secretariat has traditionally been very similar to that of LAD [Legal Affairs Division] and the Rules Division [advising panels].”¹⁸ If anything, today, the role of the Secretariat in AB proceedings may be more prominent (and it certainly has been more controversial) than that of staff advising panels: the AB staff is uniformly controlled by one single director and tends to consider itself as the main guardian of AB consistency. Staff assisting panels, by contrast, are dispersed over several divisions, with multiple directors, and may be more at ease with divergences between panel reports, as the AB remains above them to keep things in check.¹⁹ We outline the core Secretariat functions below.

Appointment of Panelists and Financial Control. First-instance panelists must be agreed upon by both disputing parties. Yet, it is the WTO Secretariat (the director and one or two senior lawyers in the LAD or Rules Division) that initially proposes names to the parties.²⁰ Moreover, if no party agreement can be found, which, today, happens in more than two-thirds of disputes, the DG is tasked to appoint panelists.²¹ The DG makes her decision on the recommendation of, or at least in close consultation with, senior LAD/Rules Division lawyers. This is the opposite of, for example, the U.S. Supreme Court, where the judges themselves appoint their clerks. In the WTO, it is the Secretariat staff that propose and, in many cases, appoint panelists, taking into account the parties’ preferences, instructions, and red lines.²² In principle, panelists are supposed to control the WTO staff assisting them. Yet panelists are aware that they owe their appointment at least in part to the WTO staff now helping them. The same staff also receives and must approve the number of days a panelist claims to have worked and the compensation and expense reimbursement that comes with it. Upon appointment, panelists receive an indicative general guideline of expected panel work for each stage of the proceeding (e.g., two to four days to prepare for the first substantive meeting). If the circumstances of a dispute require work in excess of the general guideline, such extra work and fees must be justified and approved by the director of LAD or the Rules Division, who must sign the claim form.²³ This staff influence over appointments and financial compensation blurs the traditional principal-agent relation between adjudicator and staff present in other tribunals. A panelist with an eye to reappointment and/or compensation for extra days worked thus has an incentive to avoid ruffling the WTO staff's feathers, and to follow the guidance received in Secretariat proposals and drafts.

AB members, in contrast, are appointed by consensus of all WTO members for a once-renewable term of four years.²⁴ The selection committee making proposals to the Membership is composed of committee chairs as well as the DG. AB staff play no formal role in the appointment of AB members, yet the Director of the AB Secretariat does attend the hearings of the selection committee. Like staff assisting panels, AB staff are not appointed by AB members themselves, but rather by internal selection procedures overseen by the DG. While AB staff play no formal role in the selection of ABMs, staff does hold the strings over AB members’ payment and reimbursement requests. AB members receive a relatively small monthly retainer. Actual days worked, and related expenses, must be claimed and approved by the director of the AB Secretariat. In sum, neither panel nor AB adjudicators appoint their own staff, nor can they fire them or ask that they be replaced. Conversely, however, levers are in place whereby staff exercise control over the adjudicators they supposedly serve.

Setting Timetable and Working Procedures. The first task of WTO panels is an organizational meeting where timetable and working procedures are fixed. The DSU provides a template,²⁵ but this template has proven very flexible.²⁶ Over the years, additional working procedures have been developed through trial and error.²⁷ Until recently, most of these adjustments were not made public.²⁸ As a result, the WTO Secretariat serves as the repository of best practices, and panelists—who are appointed to decide only one given dispute—rely heavily on its guidance. AB working procedures are publicly available and fixed ex ante by the AB itself,²⁹ with fewer case-specific adjustments.³⁰ Since many AB staff have been with the WTO for much longer than AB members themselves, here too, Secretariat staff played a crucial role in the original design of AB procedures in 1995 and continue to exert influence as institutional memory and conveyor belt of best practices.³¹ The AB Secretariat also holds considerable influence over the setting of timetables, as it needs to coordinate with other ongoing cases and distribute available staff resources. The time limits for panel and AB proceedings provided in the DSU are most often exceeded. In another illustration of the essential role of the Secretariat during adjudication, the availability of the Secretariat staff is often invoked as one of the core causes of these delays.³² As adjudicators work and get paid part-time, setting hearing dates and oversight over the number of days allocated to each case also ends up determining the number of days that adjudicators devote to each dispute, and thus how much they get paid. Indeed, whereas WTO staff are paid a fixed annual salary with no oversight by adjudicators, payment requests by adjudicators are scrutinized by and need approval from WTO staff.

Writing the “Issues Paper.” The parties in dispute submit voluminous filings and exhibits to both panels and the AB. Adjudicators are expected to digest all of these documents, yet the sheer volume of submissions makes this difficult in practice. To assist them, WTO Secretariat staff write an “issues paper,” which summarizes the facts and arguments of the parties and, crucially, identifies the issues to be decided. It also offers what WTO staff consider to be the applicable WTO treaty rules and past panel or AB rulings that may be on point. Importantly, adjudicators receive this report before they ever meet to discuss the facts of the case. In other words, it is not the adjudicators who deliberate among themselves, and then instruct staff; the opposite is true, especially in the early stages of a dispute.

Nothing in the issues paper is in any way binding. Yet since it effectively sets the agenda, identifying the possible ways in which the case may be resolved, its influence is significant. As one Geneva source reportedly put it, in respect of the AB Secretariat's director: his oversight of issues papers “allows him to control the flow of information to Appellate Body members by prioritizing certain issues or not addressing others. He is able to direct a large portion of the work before the members even get involved.”³³ This is especially likely for the many adjudicators whose native language is not English. Though crucially important in setting the agenda for deliberations, issues papers are never shared with the parties or the public.

Drafting the Panel/AB Questions to the Parties. Besides written submissions, panels and the AB also hold one or more hearings with the parties. Adjudicators can ask questions of the parties and third parties both before and after these hearings. These questions play an important role in identifying the issues and focusing the parties’ attention to certain matters that the adjudicators consider to be important. WTO staff, who often have the deepest knowledge of the parties’ submissions, and who wrote the issues paper, play a leading role in drafting questions. Adjudicators themselves are the ones who actually read out the questions, and they are free to add their own. Yet here too, the fact that the initial list of questions is drafted by the Secretariat offers it considerable agenda-setting power.

Providing Expert Advice on Non-legal Issues. WTO adjudicators also receive technical help from non-legal WTO staff on substantive questions in dispute. This may range from input on negotiating history and recent committee practice, to help with understanding technical terms such as “risk management” (under the SPS Agreement) or modern telecom regulations (under the General Agreement on Trade in Services (GATS)),³⁴ and the increasingly complex econometric evidence submitted by the parties. The staff providing this assistance come not from the legal divisions (LAD, Rules, or the AB Secretariat) but from operational divisions (where staffers also assist WTO members in ongoing committee work and negotiations or training and capacity building) and economists from the Research Division. Input from staff economists has proven particularly controversial, raising legal concerns of its own.³⁵ In its very last report, the AB in Australia – Plain Packaging found that the underlying panel had violated due process by (1) raising highly technical objections in an Appendix of over 150 pages to the evidence presented by claimants, even though the defendant had not raised those objections and (2) by failing to provide the parties with an opportunity to respond to those objections.³⁶ As Honduras, one of the claimants, put it in its Notice of Appeal, instead of relying on its authority to appoint outside, technical/economic experts, the panel relied on a “ghost expert,” apparently from within the WTO Secretariat.³⁷

Participation at Hearings and in Internal Deliberations. WTO staff attend all the hearings with the parties and third parties. In most cases, they sit right next to the panelists or AB members themselves, on the main podium, to enable the passing of notes and other communications. Crucially, WTO staff also attend and play an active role in all of the internal and confidential panel and AB deliberations.³⁸ Here too, they literally sit at the main table, not behind it. In fact, WTO staff normally take the first step in these deliberations, by presenting the issues paper at an internal meeting right before the first hearing with the parties. After the hearing, the panel or AB hold further deliberations to discuss the issues, prepare rulings, and go over drafts. WTO staff are present and take a leading role at every stage.³⁹ Weiler, himself a former WTO panelist, concludes that “the legal deliberation will often have taken place between legal secretary and other members of the Secretariat and not, in any meaningful sense within the Panel.”⁴⁰

Drafting of the Panel/AB Report. The DSU provides that the reports of panels and the Appellate Body “shall be drafted without the presence of the parties to the dispute.”⁴¹ No mention is made of WTO support staff. Yet, it is an “open secret” in Geneva trade circles that in most, if not all cases, it is the Secretariat that writes not only the issues paper, but also the first draft of the actual ruling.⁴² As Weiler puts it, “[t]here is an unintended honesty in the style of most Panel Reports: they are written in Third Party Reporting style: ‘The Panel considered, the Panel deliberated et cetera.’ It is honest since the writer is typically the legal secretary to the Panel, himself or herself a member of the Secretariat reporting to his or her supervisor.”⁴³ Indeed, the information note received by panelists upon their appointment does not mention any drafting of the panel report among the list of tasks it outlines for panelists; only “[r]eview of panel report drafts,” for which a mere 3–5 days is counted.⁴⁴ More generally, with the number of days a panelist is supposed to work at each stage so low (a total of 34.5 days for an entire panel proceeding), given the current length and complexity of proceedings and reports,⁴⁵ the clear implication is that much of the work will be handled not by panelists, but by the Secretariat.⁴⁶ Recent empirical work has taken up this question explicitly. Relying on two different text analysis approaches for authorship attribution, Pauwelyn and Pelc show that the WTO Secretariat exerts significantly more influence over the drafting of WTO panel reports than the panelists themselves.⁴⁷ There is variation in the extent of this influence depending on the case, the specific legal issues raised, and the adjudicators involved. Yet on average, looking at the text of the final ruling, it is the Secretariat staff's fingerprints that are most discernible.

To be sure, drafting a ruling is not the same as deciding the case. Adjudicators may provide instructions before drafting, they may revise and edit drafts, and they must, ultimately, approve the draft as their own, final ruling. It is the three panelists or AB members who sign the report, since as we note above, the Secretariat staff's input is no longer mentioned anywhere. Yet by drafting the text of the ruling itself, the Secretariat exerts considerable influence over the way in which legal interpretations and rulings are worded and argued.

Writing in the context of Registry staff assisting International Court of Justice (ICJ) judges, Hugh Thirlway draws an analogous distinction “between the decision on the issues in a case, and the expression of that decision in the best words possible.”⁴⁸ In his view, “[t]he first task is for the judges alone; but the Registry staff can and do help in the performance of the second.”⁴⁹ Yet he immediately cautions that “[l]aw is a matter of words; and it may be said that the choice of words to convey a legal point is in itself the decision of, or a decision on, that point.”⁵⁰ Partasides puts it similarly in the context of international arbitration: “[t]he act of writing is the ultimate safeguard of intellectual control. An arbitrator should be reluctant to relinquish it.”⁵¹ This concern is often explicitly inscribed into treaty texts: in arbitration proceedings under the EU-Ukraine Association Agreement, for example, the rules of procedure explicitly provide that “[t]he drafting of any ruling shall remain the exclusive responsibility of the arbitration panel and shall not be delegated.”⁵²

In a setting like the WTO, operating under a form of de facto precedent, where earlier reports are almost religiously quoted, dissected, and most often followed, the exact words and reasoning used in reaching the decision matter as much as the direction of the ruling itself.⁵³ As things stand today, panel and AB decisions not only offer a resolution of the dispute for the parties; they also interpret or give meaning to terms in the WTO treaty that will impact future cases.⁵⁴ These interpretations, in turn, have distributional consequences: they favor some countries’ interests over others. In sum, who “holds the pen” matters, nowhere more so than in the current WTO system.

C. What Explains the Central Role of the Secretariat?

In reflection of its role within WTO dispute settlement, the size of the Secretariat has grown significantly over time. At the end of 1999, about five years into WTO dispute settlement, a total of thirty-seven staff worked in the LAD, Rules, and AB secretariat divisions combined.⁵⁵ By the end of 2018, this number had increased to ninety, counting only permanent positions.⁵⁶ Staff in the three legal divisions also represented a growing share of the overall staff numbers: 7 percent in 1999, increasing to 14 percent of total staff in 2018. The AB budget has increased in parallel, from 2.3 million CHF (1.8 percent of total) in 2000, to 7.6 million CHF (3.9 percent of total) in 2019.⁵⁷ What these numbers suggest is that more WTO staff are needed today to service a number of cases that has in turn remained relatively stable over time.⁵⁸

The contrast with the days of the GATT could not be starker. The ambivalence, if not distrust, of GATT negotiators and diplomats toward lawyers, from the 1940s to the 1980s, is well documented.⁵⁹ Until the mid 1980s, the GATT Secretariat did not even have a formal legal division. Pragmatism was the leitmotif.⁶⁰ As one of the first “legal officers” ever hired by the GATT Secretariat put it, “[d]uring my first contacts inside the GATT Secretariat [in the early 1980s], most colleagues . . . told me that—in their view—the Secretariat should never have an Office of Legal Affairs; lawyers should not participate in GATT dispute settlement proceedings so as to avoid undue ‘legalisation’; and . . . the GATT should never become a tribunal.”⁶¹ It would be an understatement to say that things have changed.

The change has a precise origin. In the early 1980s, a number of GATT panel reports, composed mainly of diplomats and non-lawyers, made what many considered “legally unsustainable” rulings.⁶² Hudec described these as “embarrassingly poor decisions.”⁶³ In particular, an errant 1981 panel ruling on Spain's discriminatory treatment of imported soybean oil, containing plain errors of law, proved decisive. In 1982, GATT parties tasked the GATT Secretariat with “assisting” panels including on the “legal . . . aspects of the matter dealt with.”⁶⁴ As noted earlier, in 1994, this language was copied almost verbatim into Article 27.1 of the DSU.⁶⁵ In other words, the Secretariat's role in dispute settlement resulted from an attempt by members to exert additional oversight on the proceedings. In this sense, the Secretariat was, from the start, designed as an agent of member states, rather than adjudicators. Its purpose was to serve as a check on adjudicators. Weiler describes the division of labor between diplomat-panelists and staff in the GATT days as follows: “[c]rafting outcomes which would command the consent of both parties and thus be adoptable was the principal task of the Panellists. Custodianship over the Law of the GATT was far from both the minds, and let us be frank, the ability of many Panellists. Both the drafting of Reports and the legal ‘mumbo-jumbo’ were left in the hands of the secretariat.”⁶⁶

A subsequent event reinforces this view. Shortly before the inception of the WTO, the United States pushed to create a separate secretarial body to deal uniquely with trade remedies cases, the Rules Division. The United States was, and remains, a frequent user of these remedies. In effect, the United States thus created a specialized advisory division of permanent staff to exercise oversight over the area of law that it cared about most deeply—the same area of law that is arguably at the heart of the current impasse in the institution.⁶⁷ Here too, the intent was plainly to exercise political oversight, rather than to provide additional assistance to adjudicators.

The Secretariat was thus a strategic creation of governments, intended to serve as their agent, with a view to acting as a check on adjudicators. Yet it bears asking how much of the current institutional power of the Secretariat results from this original decision by member states, and how much has been an outgrowth of its initial mandate. In other words, how much of the Secretariat's expansion was intended?

D. How Much of the Secretariat's Expansion Was (Un)Intended?

A series of “rogue” GATT rulings containing egregious mistakes of law, and a resulting need to increase members’ oversight of panels were the initial reasons for the appointment of WTO staff lawyers to disputes. Yet a number of unforeseen factors have likely contributed to the Secretariat's rise in influence beyond what any member state had initially anticipated.

The first of these is a growing gap in sheer legal expertise. Governments continue to frequently nominate diplomats or non-lawyers as WTO panelists.⁶⁸ Many AB members also lack experience in a judicial function or legal practice; some have a distinguished career in diplomacy or civil service, but no law degree.⁶⁹ WTO staff lawyers, in contrast, must have a law degree, and are competitively selected among increasingly large pools of highly qualified and experienced applicants. Despite the increasingly “legalized” nature of WTO dispute settlement, Members continue to rely on ad hoc panelists who are often pulled from the ranks of trade policy officials. As Robert Hudec pointed out as early as 1999, the criticism of the Secretariat having “no mandate to perform this quasi-decision-making role . . . is well-founded,” but he went on, it is unlikely to change given panelists’ frequent lack of legal expertise, which means that “the Secretariat has been, and will remain, the only available source of legal expertise” in most disputes.⁷⁰ Little has changed in the intervening years.

This difference in experience would only have grown over time, given the many structural asymmetries between WTO staff and adjudicators. Consider the contrast: WTO staff work full-time, live in or around Geneva, and receive a relatively generous monthly salary and full UN-style benefits. WTO panelists and AB members, on the other hand, are devoted to their task on a part-time basis; they fly into Geneva for hearing days from different time-zones; and they are paid relatively low amounts effectively as consultants, without benefits such as pensions.⁷¹ This leads to an asymmetry of experience and time allocation. On the one hand, in situ staff paid full-time, with years of experience and detailed knowledge of procedure and past rulings. On the other, adjudicators often new on the job, who have (better-paid) competing demands on their time, and who need to travel to Geneva to attend hearings.⁷² As a result, it is WTO staff who increasingly operate as the “institutional memory” of the organization and prime experts on WTO treaty rules and jurisprudence.

The de facto recognition of precedent over the DSU's history has given further weight to the comparative advantage of WTO staff over adjudicators, and may have further added to the delegation from adjudicator to expert staff. As Thomas Graham, former AB chair, put it: “adherence to precedent . . . made it more important to know the past than to think anew. It empowered those who best knew the past, such as staffs.”⁷³ In this way, the practice of buttressing current reasoning with citations to past cases has elevated the status of the Secretariat staff, who have more familiarity with WTO case law than anyone. Given the cumulative nature of jurisprudence, the relative advantage of Secretariat staff only increased over time: as WTO jurisprudence grew and became more diverse and complex, encompassing tens of thousands of pages of legal reasoning, so would the adjudicators’ inclination to rely on, and defer to, the staff's guidance.

In sum, the appointment of Secretariat staff as agents of governments to oversee the dispute settlement process was initially a rational response to the growing complexity of cases and lessons learned in the early 1980s. Yet in installing a second agent to oversee the work of the first, member states put in place no means of oversight over that second agent—likely because these would have seemed unnecessary when the Legal Affairs Division was made up of a total of three lawyers, all new to the task. Yet over time, the Secretariat effectively became a free agent. Under the guise of its much-vaunted “independence,” it gradually became accountable to no one.⁷⁴ As noted earlier, in the WTO, no adjudicator appoints its own staff. Staff is appointed on a standing basis by the DG. Junior staff first and foremost report to their hierarchy and director within the division.⁷⁵ Division directors report to the DG but the DG is extremely hesitant to meddle in the specifics of dispute settlement, for fear of undermining its independence. The lack of control mechanisms over staff and especially staff directors was never as clear as when complaints arose against the director of the AB Secretariat, and it became apparent that no one—neither member states nor the director-general—had the combined authority and willingness to act.

This is not to say that the risk of a bureaucracy growing in influence was entirely unforeseen. In fact, some of those warnings came from the United States itself, which had championed the creation of both the Legal Affairs Division and the Rules Division. As early as 1995, within months of the WTO's creation, Alan Wolff, who was later to become the WTO deputy director-general, warned the U.S. Senate against a “[p]owerful, ensconced staff” in downright prophetic terms:

Another troubling procedural aspect of the system is the effective authority of the WTO secretariat. The secretariat advisors may remain the same from year to year while the panelists serve only infrequently. The secretariat, therefore, is in a position where it may exert a substantial influence on panel decisions. With ad hoc panelists and very limited time, the authority of the secretariat—unelected officials appointed without effective review by WTO member representatives—is likely to grow. The impact of this type of influence on the WTO Appellate Body is also unknown.⁷⁶

The warning was strikingly prescient. As we argue above, it is exactly these structural asymmetries, combined with a growing gulf in its experience vs. that of the average panelist or AB member, that explain why the Secretariat evolved beyond anything governments had initially imagined.

E. The WTO Secretariat Compared to Staff Assisting Other International Tribunals

As a growing literature attests to, the influence of legal bureaucracies and judicial assistants has been on the rise across international tribunals over the last decades. This is the result of a number of common factors: swelling caseloads, as states have been bringing matters pertaining to a wider range of policy domains for tribunal review; the growing complexity and fragmentation of the relevant bodies of law; and the rising stakes of disputes.⁷⁷ The ICJ's Registry has thus seen its various departments and technical divisions grow ever more specialized.⁷⁸ The number of référendaires in the Court of Justice of the European Union (CJEU) has steadily grown, to four times their initial number.⁷⁹ Issue-area experts have been increasingly called upon both in European supranational courts and investment arbitration, as disputes now often turn on questions of science or technical economic data.⁸⁰

Yet even amidst this general trend toward greater delegation of power to legal bureaucracies, the WTO remains exceptional: all things considered, its Secretariat exerts more influence over dispute settlement proceedings than the staff of any comparable state-to-state tribunal. This is the result of a combination of deliberate decisions taken early on by governments and unanticipated factors that arose subsequently. We arrive at this conclusion on the basis of three factors: (1) the combined tasks that WTO staff perform; (2) the asymmetries between WTO staff and adjudicators; and (3) the control by WTO staff over adjudicators, a reversal of the usual relation.

ICJ staff, like the staff of most international tribunals, may play an important role in the drafting of rulings.⁸¹ Compared to WTO staff, however, they play a far less active role in the internal deliberations of ICJ judges, if they are present at all. The asymmetry in expertise between ICJ staff and judges is also smaller: whereas (permanent) lawyers in the Department of Legal Matters of the ICJ's Registry may equal the seniority and built-up expertise of WTO staff, ICJ judges are overall more expert in the law than WTO adjudicators. The influence of Registry lawyers (advising the Court in general, with overall coherence in mind) is also in competition with, and balanced out by, law clerks picked by and assisting individual ICJ judges. In the WTO, Secretariat staff fulfill both roles and represent and generally advise the collective; WTO adjudicators are not supported by individual clerks. ICJ judges are appointed for nine-year terms (renewable without limit) compared to ad hoc panelists and AB members appointed for four years (renewable only once). ICJ staff also play no role in the appointment of ICJ judges. Unlike WTO adjudicators, ICJ judges are employed full-time and receive fixed salaries and benefits. This means that ICJ staff, unlike WTO staff, exercise little or no financial control over ICJ judges. Finally, whereas WTO staff are hired by and report to an internal hierarchy, and ultimately the DG (who is traditionally wary of intervening in dispute settlement), ICJ judges themselves choose and control their individual law clerks, and more generally control Registry lawyers, as the ICJ registry “is accountable to the Court alone.”⁸²

Legal assistants and secretaries to ad hoc international arbitrations (be they state-to-state such as trade disputes under free trade agreements, or investor-state under bilateral investment treaties) may exert influence in the specific case at hand, but they are not permanent staff and, therefore, do not accumulate the same institutional knowledge and impact over time the way WTO staff does. Staff assisting International Centre for Settlement of Investment Disputes (ICSID) and Permanent Court of Arbitration (PCA) arbitrations may be (partly) long-term/permanent and those Secretariats can also play a role in arbitral appointments. That said, ICSID/PCA arbitrators are generally more expert in the law and procedure than WTO adjudicators (thereby reducing the staff/adjudicator asymmetry) and, substantively, the role played by ICSID and PCA staff and assistants to international arbitrations generally is overall less important and more administrative:⁸³ there are generally no “issues papers” with proposed solutions, nor active participation in internal deliberations. The WTO is an outlier in the extent to which support staff actively participate in a panel's confidential, internal deliberations. In international commercial arbitration, active participation of this kind in internal deliberations by Tribunal secretaries could be seen as a reason to set aside or annul the arbitral award. The 2016 UN Commission on International Trade Law (UNCITRAL) Notes on Organizing Arbitral Proceedings draw a clear line in this respect: “it is recognized that secretaries are not involved and do not participate in the decision-making of the arbitral tribunal,” except in certain rare, specialized types of arbitration.⁸⁴ Similarly, the recent Agreement between the United States of America, the United Mexican States, and Canada (USMCA) explicitly provides that “[o]nly panelists may take part in the deliberations of the panel. Assistants, Secretariat personnel, interpreters, or translators may be present if the panel determines they are necessary.”⁸⁵ The EU-Ukraine Association Agreement provides for similar limitations and also distinguishes between being “present,” and actually “taking part” or participating in, deliberations. Only the former is permitted if the tribunal so decides.⁸⁶

One needs to turn to staff lawyers advising the General Court and the Court of Justice of the EU (with its Registry and référendaires) and, in particular, the European Court of Human Rights (ECtHR) (with Registry rapporteurs, case lawyers, as well as the Jurisconsult⁸⁷) to find anything approaching the function played by the WTO Secretariat. As it happens, these actors have also drawn growing critical attention as “unseen actors,” with the associated concerns over legitimacy.⁸⁸ The regional character and sheer size of the caseloads of these two European courts, numbered in the thousands, and the fact that proceedings in these courts are generally not state-to-state, further complicate the comparison. Yet even here, the role of the référendaires is considerably more limited: they serve “at the member's will,” and are deemed “contractual agents” from the standpoint of EU staff rules.⁸⁹ In neither the ECtHR nor the CJEU do staff play a role in the appointment of judges. Similarly, even in the case of these large legal bureaucracies, there is no equivalent to the financial control by WTO Secretariat staff over adjudicators. On the other hand, in both European courts, as with the ICJ, judges themselves have a more permanent function: they get a fixed compensation package and are appointed for longer terms than AB members: non-renewable nine-year terms for ECtHR judges; six-year, renewable terms for CJEU judges.

III. How the Role of the Secretariat Matters

As we argue next, the WTO Secretariat's expansive role holds considerable implications for the institution itself, as well as for its study. We place these effects under three banners: institutional legitimacy, accountability, and legal outcomes. More broadly, we argue that the sheer extent of the Secretariat's role within the WTO should lead us to reassess the established view of what the WTO's dispute settlement mechanism actually is. Rather than an independent judiciary operating at the international level, a full appreciation of the Secretariat's role recasts WTO dispute settlement as something more akin to a process of administrative or agency review.

A. Effects on Internal vs. External Legitimacy

International tribunals stand or fall by their perceived legitimacy. Lacking any enforcement power of their own, unable to compel sovereign states to support their interpretations and comply with their rulings other than through consent, they depend on both internal and external audiences granting them a “right to rule.”⁹⁰ The “internal” audience of international tribunals, in this sense, is constituted of the member delegations and their trade ministries at home, while the “external” audience is made up of citizens, multinational corporations, NGOs, and national media, all of whom member states are to some degree answerable to.⁹¹ Such attention to the perceptions of different constituencies speaks to what is most often called “sociological legitimacy,” which relates to subjective beliefs about a tribunal’s authority, and which is the facet of legitimacy that most interests us here.⁹²

One might argue that as far as the WTO’s legitimacy is concerned, the empowerment of a Secretariat widely lauded for its expertise and high levels of professionalism and independence should be all to the good.⁹³ After all, the presence of “guardians of the system” suggests that the long-term interests of the institution are duly protected, and reflected in rulings. Staff may improve the adjudicative process and keep “rogue” adjudicators in check, as intended. In a recent look at the civil servants of the ECtHR’s Registry, Creamer and Godzimirska argue that it ought to have a net positive effect on trust in the court. Tellingly, they focus on the Registry’s effect on the perception of member states, and allow that this effect could prove different for the court’s other constituencies. This is the distinction we insist on here. The high degree of opacity under which the WTO Secretariat operates, and which exists by design, masks an unresolved clash with the formal setup in WTO treaties, one that is condoned by members.

The ways in which this highly competent, empowered, and mostly unseen Secretariat impacts the institution’s legitimacy are thus not straightforward: the WTO’s various constituencies are likely to perceive these effects differently. In this section, we unpack the trade-off that results: in brief, members bought a measure of internal legitimacy at a high price to external legitimacy. With the growing attention now being paid to the unseen actors of international tribunals, the gap between outward appearances and the legal process as it actually plays out may become a growing burden on the institution.

In a recent survey of the determinants of the perceived legitimacy of various international tribunals, Follesdal observes that “risks arise from the often opaque roles of many ‘unseen actors’ who nominate judges and help administer and advise cases.”⁹⁴ Follesdal is building on the work of Baetens, who has devoted an entire edited volume to teasing out the various ways unseen actors affect the perceived legitimacy of the tribunals in which they operate. She argues that “an assessment of [unseen actors’] legitimacy should include an examination of their origins, function (including their day-to-day procedure), and performance (that is, the outcome and effects of its function).”⁹⁵ This is the course we take up here.

As we argue in Sections II.C and II.D above, the Secretariat's Legal Affairs and Rules Divisions were both created with a particular purpose in mind: they were means of preserving the diplomatic character of dispute settlement, while ensuring that the quality of legal reasoning would not suffer unduly. By insisting on retaining ad hoc panelists and part-time AB members with limited time devoted to case proceedings, but also creating a highly empowered legal bureaucracy to appoint panelists and guide adjudicators at both the panel and appellate stage, members achieved a provisional balance: adjudicators would remain sensitive to government interests, being themselves most often pulled from the ranks of diplomats and trade officials, and a legal bureaucracy would oversee the legal soundness of rulings, even as members were unable to exercise this oversight function themselves (rulings are automatically adopted unless there is a consensus against adoption; authoritative interpretations by WTO members require a three-fourths majority and, in practice, consensus).⁹⁶ The result, as Weiler saw early on, “put a premium on settlement and acceptability, and in a finely tuned process which often combined the diplomatic skill and reflex of the panel with the legal expertise of the Secretariat.”⁹⁷

This particular compromise required the maintenance of a fiction, according to which the selected adjudicators single-handedly rule on complex issues, similarly to judges in a municipal court. As Weiler went on to say, the result “shades the truth in that the legal deliberation will often have taken place between legal secretary and other members of the Secretariat and not, in any meaningful sense within the Panel.”⁹⁸ Most legal observers predicted that this imbalance would gradually be corrected, and the need for this fiction would be reduced. The most widely advocated solution, by Weiler among others, was a shift toward more professional adjudicators, who would be less dependent on the Secretariat's expertise. This did not come to pass. In a way that speaks to the limits of delegation to law that sovereign states are willing to countenance, amid the institution's much-vaunted “judicialization,” governments insisted on retaining the diplomatic character of dispute settlement. Continued reliance on diplomats over jurists was “reassuring to the internal players,” who were loath to give it up.⁹⁹ As Debra Steger plainly puts it, “[t]he continuing selection of panelists on an ad hoc basis from the pool of Geneva-based government officials contributes to the ‘internal’ legitimacy of the dispute settlement system.”¹⁰⁰ Yet this compromise has required a highly empowered Secretariat, the full function of which is kept hidden from view, and whose influence over time has only grown as a result. As Weiler concluded, “the persistence of diplomatic practices and habits in the context of a juridical framework might end up undermining the very rule of law and some of the benefits that the new DSU was meant to produce.”¹⁰¹

Internal legitimacy stands for the acceptance by governments of an adjudicator's right to rule; but governments have just as much of a stake in external legitimacy: without sufficient buy-in by their domestic constituencies, from NGOs to voters themselves, governments risk domestic pushback when they try to act on the adjudicator's recommendations. Indeed, given the United States’ move against the AB, scholars of the WTO have understandably focused on how the institution secures buy-in from member states. Yet it is worth recalling that other international tribunals, from the investment protection regime to European supranational courts, have also faced growing backlash, which has for the most part originated in skepticism from domestic audiences, rather than governments.¹⁰² Writing about the ECtHR, for instance, the rulings of which became a fixture on the front page of British newspapers in the years preceding the Brexit vote, Voeten argues that it is the lack of public support that encouraged non-compliance.¹⁰³ In a later article, Voeten shows how such popular skepticism has made international courts especially tempting targets for populist political candidates. But there too, the prime mover was a sense of generalized skepticism and mistrust of international courts, which was often linked with the perceived non-transparency or inaccessibility of these bodies.

Public opinion about courts thus matters for governments. The outward appearance of WTO dispute settlement as a strictly judicial process has served member states who must justify reforming domestic policies in response to the “recommendations” of panels and AB members. Christina Davis has demonstrated at length how adjudication allows states to achieve settlements not attainable through simple negotiations, by making outcomes more palatable to domestic audiences through the legitimating invocation of international law.¹⁰⁴ Much of this persuasive authority is tied to the image of independent judge-like figures impartially applying international rules to deliver rulings from on high. As Hannah Arendt famously put it: “If authority is to be defined at all, then, it must be in contradistinction to both coercion by force and persuasion through arguments.”¹⁰⁵ In other words, WTO adjudicators’ ability to gain acceptance for their rulings relies not merely on the legal validity of their reasoning, but in their very status as adjudicators: highly visible individuals, who place their imprimatur on the ruling document by literally signing it “in the original”;¹⁰⁶ who are appointed according to an agreed-upon process; and to whom authority to rule has been delegated by sovereign states, alongside feedback mechanisms to revoke that authority if these individuals overstep their mandate. Governments rely on this perception of authority to increase their bargaining power vis-à-vis domestic import-competing industries, who incur the costs of compliance most directly, and are thus most staunchly opposed to it.¹⁰⁷ The figure of the judge has a unique claim to what Thomas Franck called law's “compliance pull,” in a way that a collective body of unidentified staff applying the same law to the same cases may lack.¹⁰⁸ A tribunal's authority in the eyes of external observers thus rests in part on the assumption that the appointed adjudicators are the authors of the rulings handed down by the tribunal, instead of permanent staff whose names are no longer found in the final report.

In sum, member states have tried to have it both ways. They have put in place a system that draws all the authority of a judicial process, while maintaining the deference to members’ political sensibilities thanks to ad hoc panelists pulled from the ranks of trade officials. An active Secretariat has been key to achieving this balance, but its full influence has been deliberately hidden from view. Internal legitimacy has been bought at the price of external legitimacy, yet this balance is increasingly untenable. As the full role of the Secretariat increasingly comes to light, our claim is that the WTO will become highly vulnerable to claims of illegitimacy from external observers. It may undo years of efforts by member states to portray the process as transparent, legalistic, and drawing the full authority of a judicial system. Indeed, as Shaffer, et al. have documented, member states have tried to secure greater buy-in from civil society for the WTO by agreeing to making dispute hearings open, and publishing their submissions as a matter of policy. “This practice,” they write, “makes the proceedings appear to be more transparent and legalistic and thus (potentially) less objectionable.”¹⁰⁹ Such an image is likely to be disrupted as more is revealed about the heretofore hidden influence of a permanent body of largely unaccountable staff over the dispute settlement process.

B. Effects on Accountability

Member states empowered one agent (the Secretariat) to oversee the other (adjudicators). But in increasing the accountability of adjudicators, members have had to relinquish their oversight of the Secretariat, rendering the latter, in effect, accountable to no one. This effect on accountability, one of the essential components of procedural legitimacy, is the one we examine next.¹¹⁰

The WTO treaties painstakingly define the qualifications, independence, geographical distribution, and appointment process and terms of the adjudicators.¹¹¹ Meanwhile, the Secretariat staff, which as we have shown, exercise considerable influence over the adjudicators themselves at every stage, from appointment to their financial remuneration, which provides the adjudicators with issues papers setting the agenda and framing the terms of the debate, and which drafts the actual final rulings—receives comparatively little attention. No analogous vetting of Secretariat members exists in the run-up to disputes. The names of the relevant staff have ceased appearing on the final report, just as their influence over the proceedings has grown. The staff are mandated to “assist” and “support” adjudicators, but the line between such “support” and involvement in final decision making is ambiguous and unenforceable, precisely since much of the process takes place behind closed doors.

Similarly, no feedback mechanism exists to review the inputs from Secretariat staff. Beyond feedback on specific staff-produced documents in a given case, staff do not answer to panelists and AB members; they report to their team leader or senior lawyer who, in turn, follows orders of the staff director. If they are dissatisfied with staff in a particular proceeding, neither panelists nor AB members have the power to ask for a replacement, let alone to remove staff from their functions. The WTO staff are paid exclusively out of the overall WTO Secretariat budget, and it is instead adjudicators who must get their payment requests approved by WTO staff. More striking still, members themselves also exercise no real oversight of staff. The accountability of WTO staff is limited to the internal hierarchy within the Secretariat.¹¹² While panelists or AB members that are seen as having “overreached” can be blacklisted or see their reappointment blocked, most WTO staff have permanent contracts. Neither adjudicators nor WTO members have direct control over them, let alone the power to terminate their contracts. This was vividly illustrated in December 2019, when the AB chair wanted to fire the director of the AB Secretariat, but was unable to do so.

Whereas the Secretariat's formal role “assisting” panelists and AB members might at first paint it an agent of the adjudicators—the traditional role of staff in other tribunals—these considerations explain why it is more accurately seen as an agent of WTO members. Here is a case of one agent checking another.¹¹³ This is consistent with the institutional history we outlined above. The balancing act that members have sought to achieve, where the outward appearance of strictly “judicial” proceedings conceals a more managed system behind-the-scenes, has required members to delegate considerable power to the Secretariat. Yet such delegation brings with it an inevitable loss of control. As time has passed, staff have grown in number, settled into permanent contracts, and expanded their role. Once appointed, staff are accountable neither to adjudicators nor WTO member states. Even the DG lost effective control over AB staff—who are technically outside of the WTO Secretariat—as any intervention by the DG could be seen as “political meddling.” As a result, WTO staff have fallen between the cracks of any real accountability mechanism. They have effectively become a free agent. This explains how the Secretariat has adopted a set of practices with effects that were unanticipated even by the system's most powerful members.

C. Legal Effects

Beyond concerns over external legitimacy and accountability, the Secretariat's expansive role also has a concrete impact on outcomes. We outline four such effects, relating to: (1) the status of precedent; (2) the length of reports and proceedings; (3) the expansive scope and ambition of rulings; and (4) the suppression of dissenting opinions. None of these are inherently desirable or undesirable, but what is notable, as we lay out in Table 1, is that they happen to coincide to a large degree with the “systemic concerns” the United States raised around WTO dispute settlement. This underscores a key aspect of principal-agent relations: delegating power to an agent necessarily brings with it a loss of control, even for the most powerful entities.

Table 1. The Role of the Secretariat in WTO Dispute Settlement: Causes and Effects

Status of Legal Precedent. Although they are not formally legally binding on future cases, it has become standard practice for panel and AB reports to pay close attention to past rulings, giving rise to a form of de facto stare decisis. Rulings spend pages summarizing and referring back to past reports and further interpreting, distinguishing, and developing the precise sentences and words previously employed. The AB, in particular, has stated that “absent cogent reasons” its rulings are to be followed by panels and future AB divisions alike.¹¹⁴ The current practice is likely a result of a series of factors, including the very creation of a second-level AB, the AB's principle of collegiality and the way parties and litigators have argued cases. Yet the increased role of WTO staff in the process is another likely factor.

It should come as no surprise that WTO staff, formally tasked with providing “legal support,” and which has been “advising” past panels or AB divisions—including involvement in internal deliberations and drafting of past rulings, in some cases, for over a decade—are more inclined to refer back to past decisions. They may be especially likely to refer to those past decisions they were themselves part of, or for which they wrote the crucially important issues paper. Recent empirical work supports this expectation. As Pauwelyn and Pelc demonstrate using text analysis tools, those disputes where the Secretariat appears to have most input in drafting rulings are also those that are subsequently most highly cited.¹¹⁵

As permanent staff, paid and employed by the WTO, Secretariat lawyers may also be predisposed to defend the WTO and its “jurisprudence” as an institutional value and, in that context, pursue “consistency” and respect for “precedent” (and, at times, refuse to admit or correct mistakes) as an independent goal. This may be the case especially for AB staff, which tends to consider itself as the main guardian of WTO jurisprudence.¹¹⁶ Blustein, referring to the director of the AB secretariat, reports the following:

his arguments are generally perceived as stemming from a passion to safeguard institutional respectability—in particular, ensuring that new rulings follow principles set forth in prior cases—rather than pursuing some political agenda. His overriding goal, in other words, is that the Appellate Body should be consistent.¹¹⁷

Even defenders of the AB Secretariat director recognized that his “aims are consistency in Appellate Body rulings and protecting the institution's integrity.”¹¹⁸ Another source stated that “[t]he secretariat has placed an emphasis on collegiality and consistency, but overshot the target when the Appellate Body's culture no longer allowed it to openly reconsider its jurisprudence.”¹¹⁹

A striking demonstration of panel staff pursuing “consistency” is a 2005 panel report that copies almost verbatim (but without reference) the findings of a panel report circulated only two weeks earlier, on the same issue of “entrustment or direction” under the SCM Agreement.¹²⁰ As the panelists were different in the two cases, it is highly unlikely that the findings of the first panel were publicly available when the second panel drafted its report (drafting occurs weeks before the report is publicly circulated). The implication is that it was most likely Secretariat staff that cross-checked these panels’ findings, to ensure coherence.¹²¹

Reference to past rulings, in turn, increases the role and influence of WTO staff: since they are intimately involved in the history, internal deliberation and drafting of past decisions, they become the organization's “institutional memory.” Adjudicators themselves, who often lack experience or detailed knowledge of case law, may feel they have no choice but to rely on the staff's guidance. In this sense, the WTO staff's enhanced role and the elevation of precedent are mutually reinforcing.¹²²

Length of Reports and Proceedings. Panel and AB reports have grown longer, as has the time required to deliver them.¹²³ One reason is the aforementioned practice of making detailed reference back to past decisions. When WTO staff write the first draft—already a product of compromise and correction between junior and senior staff lawyers—and then various adjudicators add to it, with the overall goal of consensus reports (i.e., avoiding dissents), rulings inevitably become longer, with unnecessary overlap and repetition. As one outgoing AB member put it in his farewell speech, “an excessive striving for consensus decisions coupled with a discouragement of dissents . . . led to excessively long and unclear compromise reports. It also encouraged over-reach, gap filling, and advisory opinions, as a way of accommodating conflicting views.”¹²⁴

Paradoxically, this implies that increasing the role of WTO staff may, at times, lengthen and complicate the process and rulings, rather than make the process faster and more efficient. The fact that both panels and the AB now consistently exceed DSU prescribed timeframes may be partly due to a mutually self-reinforcing effect: the Secretariat's involvement in rulings results in a more bureaucratic style where “no stone is left unturned,” and hypertrophic reports replete with references to earlier rulings. These, in turn, invite longer and more articulate party submissions, which, in turn, call for more detailed and time-consuming examination by panels and the AB, all of which requires yet more staff input.

Expansive Scope and Ambition of Rulings. The formal objective of WTO dispute settlement is to resolve the particular matter at issue in the dispute at hand. A mutually agreed settlement is explicitly put forward as the preferred outcome (DSU Article 3.7). Experienced adjudicators have reason to focus their reasoning on certain elements, while relegating others to the margins of the analysis, exercising a type of “selection prerogative.” WTO staff, by contrast, have more incentives to consider a ruling's impact on the broader system, rather than merely resolving the specific case before them. They have more reason to care about coherence with past rulings, with the other eye on future cases. Staff may thus see greater value in “completing” the incomplete contract than rotating panelists and AB members. The central role of WTO staff in writing the issues paper, asking questions of the parties, actively participating in deliberations, and drafting the actual ruling all provide ample opportunity to act on these temptations.

Suppression of Dissents. For the same reason that high staff involvement enhances the likelihood of reference to past rulings, WTO staff, assigned to a panel or AB division as a whole (rather than individual adjudicators) and in many ways representing the “institutional memory” and “consistency” of the WTO as an institution, has reason to deter dissents by individual adjudicators, and support consensus outcomes.¹²⁵ The dissent rate in the WTO is unusually low, occurring in less than 10 percent of AB rulings.¹²⁶ The role of WTO staff and its conscious, or unconscious, defense of the system and of its coherence may be among the factors behind this low rate of dissents.

In the DSU's early years, especially, collegiality and consensus were seen as means of ensuring the perceived credibility and predictability of the AB. In this context, the Secretariat's ability to provide a common starting point through its write-up of the issues paper, and the way it served as a “neutral” advisor and natural focal point, made its guidance a good means of arriving at consensus opinions.

These legal effects are notable in and of themselves, yet they appear especially so in the light of the United States’ recent “systemic concerns,” which have led to the system's current paralysis. Among these concerns, the United States has pointed to the misguided invocation of precedent, overreliance on obiter dicta, a tendency to lock in past mistakes, and overly lengthy rulings and proceedings, focusing on litigation rather than negotiations and settlement.¹²⁷ Echoing many of these concerns,¹²⁸ the last U.S. AB member denounced “an excessive striving for consensus decisions coupled with a discouragement of dissents,” along with a “prevailing ethos” “created and maintained over many years by the Appellate Body staff leadership” according to which the AB was “a self-anointed international court.”¹²⁹

The Secretariat has never been formally singled out for criticism by the United States, which was instrumental in expanding its role in dispute settlement. Yet it is hard to overlook how many of the U.S. criticisms are related to trends that can be linked back to that expanded role, especially that of the AB Secretariat. The implication is noteworthy: the Secretariat's empowerment has had a number of effects that were unanticipated even by the members who pushed for that empowerment. As a result, the “guardians of the system” may have contributed to several trends that ultimately led to backlash against the system. Whether normatively desirable or not, the unanticipated aspect of the Secretariat's expansion appears to have had system-wide consequences.

D. A New Lens On WTO Dispute Settlement

Our examination of the role of the Secretariat leads us to recast the WTO dispute settlement system as altogether different from what legal scholars and political scientists have long portrayed it as. In part, these hopes were self-serving. Political scientists saw in the institution the exemplar of a broader theorized trend toward legalization and judicialization in international affairs.¹³⁰ Legal scholars viewed the WTO as a judicial organ that imbued the trade regime with a constitutional logic. Some even called on governments to recognize what they had created, by rechristening the dispute settlement mechanism as the “World Trade Court” or even “the International Court of Economic Justice.”¹³¹

Instead, a full appreciation for the Secretariat's outsized influence, and the power that this collective body amassed over time, relative to adjudicators, should lead us to fundamentally revise this idealized picture of an international court exercising judicial review. The common understanding of “judicial review” relates to the power of an independent judiciary, or court of law, to determine whether the acts of other components of the government or organization in question are in accordance with the constitution or other relevant legal instrument. In contrast, “administrative review” is understood as an internal process that is not fully independent from the actor or actors it is reviewing, and where compliance decisions are made, not by independent judges, but a technocratic or administrative process.¹³² Applying this distinction to WTO dispute settlement, we see two main reasons to reclassify it as a sui generis process of “administrative review.” First, rather than independent from the other components of the organization, following the doctrine of separation of powers, which is inherent in the concept of judicial review, WTO dispute settlement remains closely embedded within the broader political and diplomatic operation of the WTO.¹³³ As one of us has written in a different context:

Panels and the AB meet in the same building in Geneva where trade negotiations and monitoring (trade policy review) meetings are held. Many of the panelists are trade diplomats who, both before and after their panel appointments, walk the very same halls but wearing a different hat—that of country negotiator. Likewise, many AB members have been trade diplomats or were stationed in Geneva in prior careers. More broadly (in terms of institutional context), WTO panel and AB reports have no legal value unless and until they are adopted by the WTO's Dispute Settlement Body, on which all . . . WTO members have a seat. Although DSB adoption is virtually automatic (one single vote in favor suffices), DSB meetings are religiously attended, members provide formal feedback on reports, and the DSB plays a prominent role in monitoring implementation.¹³⁴

WTO dispute settlement is also closely linked to, rather than independent from, the WTO Secretariat: It is the Secretariat that proposes names of panelists and actually appoints panels in the majority of cases. AB members, in turn, are (politically) appointed by consensus of all WTO Members.

Secondly, rather than individual judges or an independent court, decisions in WTO dispute settlement are crafted as much by in-house Secretariat staff and experts as by diplomat-adjudicators. The main capital and strength of WTO dispute settlement is not the expertise, standing or exceptional character of individual panelists or AB members; it is the internal (and mostly confidential) institutional process, combining the views and arguments of parties and third parties, with the expertise of panelists, AB members and Secretariat staff.¹³⁵ As Thomas Cottier recently put it, “[t]he crucial and essential involvement of the WTO Secretariat and closeness to the process of diplomacy raise the question whether the quasi-judicial process [of WTO dispute settlement] is not closer to a particular sui generis or hybrid type of fully legal procedures . . . . Panels and their team operate comparably to administrative law judges, being independent but still part of the WTO as an agency.”¹³⁶

In practice, the WTO Secretariat plays a role closer to that of advocate generals in the civil law tradition, or at the Court of Justice of the EU—formal, public law positions, representing the collective interest, with a neutral voice of their own¹³⁷—than that of law clerks assisting domestic high courts or international tribunals like the ICJ, merely tasked with assisting individual judges. What Lasser says of the advocate general (a branch of the ministère public) in French proceedings could equally be said of the WTO Secretariat and WTO adjudicators: it “occupies a privileged, intermediate position between the parties and the court . . . . Institutionally and professionally, the advocates general are truly the judiciary's brethren.”¹³⁸ Similarly, the WTO Secretariat is an actor on par with the adjudicators. One difference is that the advocate general's submissions, their conclusions, are “nominally public,” though in practice they remain difficult to access. The other, more important, distinction is that the advocates general are known: they are themselves magistrates, they are explicitly identified as the authors of their conclusions, and their function is an elevated, not to say a celebrated one. The WTO Secretariat, by contrast, is an unidentified collective, and their role, output and opinions are purposefully kept opaque by the institution.

The comparison to the advocate general function remains useful, insofar as it leads us to recast the very nature of dispute settlement at the WTO. The DSU was never meant to be, nor did it ever become, the World Trade Court. Rather, the process reveals itself to have been something more akin to administrative review performed at the international level, closer to agency (than judicial) review in domestic legal systems.¹³⁹ One where a group of in-house experts on law and economics—a combination of diplomat-adjudicators and staff-experts, interacting as peers with mutual levers of control over each other—craft solutions that attempt to strike a balance between countries’ obligations and their political sensibilities. The result is as much a technocratic exercise as a judicial one.

As noted earlier, Thomas Cottier argues for a similar re-envisioning of WTO panels.¹⁴⁰ Meanwhile, he continues to view the AB as “more like a court of law.” Given what we now know about the AB Secretariat, we would extend Thomas Cottier's description to encompass the AB, and the entire dispute settlement system. Indeed, the structural asymmetries that led to the Legal Affairs and Rules Divisions’ influence are also present in the AB Secretariat—even more so in some instances. As the case of its director demonstrates, the AB Secretariat became just as much of a “free agent” as the staff assigned to panels.

IV. Conclusion

We spent the first part of this Article outlining the expansive role the WTO Secretariat exerts in dispute settlement. Its influence, as we argue, appears greater than that of any other comparable state-to-state tribunal's staff. The WTO Secretariat is involved not only in setting timetables and working procedures, drafting questions to the parties, and appointing panelists, but it also has the key tasks of writing “issues papers,” actively participating in internal deliberations, drafting the eventual ruling, and exercising elements of financial control. New empirical tools have recently provided more clarity into the question of who drafts the rulings of the WTO, confirming how the permanent staff appear to have more influence over the final text of rulings than the adjudicators themselves. We then offered possible explanations for why Secretariat support staff—whose numbers have grown rapidly since the WTO's creation—play such a prominent role today. These have both a deliberate and a less foreseeable aspect. The Secretariat's role arose from an explicit treaty mandate to provide “legal support” following a series of problematic GATT rulings in the early 1980s. But then growing asymmetries between staff and adjudicators in terms of appointment terms and expertise, and the relative independence and limited accountability of staff, have contributed to expanding the role of the Secretariat beyond anything members might have anticipated at its outset.

We then lay out the implications of this unique and expansive role. Crossing the line between adjudicator and staff tasks may threaten the external legitimacy, perceived authority, and compliance pull of WTO rulings. Looking specifically to the question of drafting, the question of who “holds the pen” takes on special significance in a legal regime that pays as close attention to past rulings as the WTO does. Concretely, drawing on empirical findings and evidence from other courts, we argue that a greater role for the Secretariat is likely to contribute to four key outcomes: increased reliance on precedent; the lengthening of reports and proceedings; the expanding scope and ambition of rulings; and the deterrence of dissents. While these effects are not inherently desirable or undesirable, they overlap to a striking extent with the very U.S. concerns that have led to the demise of the AB: the role of precedent and obiter dicta; the expansive interpretations of the AB's mandate; and frequent flouting of the ninety-day period for AB review. From a force to keep “rogue” panels in check, the Secretariat may therefore have contributed to the very factors that have brought about the system's current paralysis.

We end with a far-reaching claim concerning taxonomy. An appreciation for the full role of the WTO Secretariat should lead us to reevaluate what we understand the WTO legal system to be. Neither a bird nor a plane, in light of our examination, the dispute mechanism appears as more of a domesticated terrestrial animal. In its current form, it is best seen as a sui generis process of administrative review, along the lines of what domestic regulatory agencies are involved in, transposed to the international level, where a group of dedicated experts review the measures of member states.

Our discussion suggests a number of possible reforms going forward. Most fundamentally, greater clarity must be achieved as to the true role of the WTO Secretariat: is it supposed to merely assist adjudicators (as law clerks do) or is it designed to represent the systemic interests and coherence of the WTO collective (as advocate generals do in the civil law tradition). Recent proposals at the highest levels suggest increased willingness to reform the institution in the latter direction. Consider Alan Wolff's suggestion of appointing a new independent Office of Legal Counsel, who among other tasks, could file amicus briefs, along the lines of what the AG does in the European supranational system, and would have an independent investigative function, with the power of referring matters directly to the General Council.¹⁴¹ If the WTO Secretariat was meant to play a role closer to that of law clerks, drastic limitations are needed when it comes to its roles and functions, and the chain of command and control should run from adjudicators to staff, and not the other way around. If, on the other hand, WTO staff are to act more like an advocate general, their current role and impact may be justifiable. However, stricter accountability mechanisms would then be called for. Either way, more transparency is needed. The status quo of the Secretariat as de facto advocate general (and much beyond in some respects) without the transparency and control mechanisms expected even for law clerks is unsustainable.

In this respect, modest steps like disclosing the names of staff involved in a proceeding and circulating the “issues papers” and any econometric advice provided by Secretariat staff to the parties to a dispute, and then publishing these memos alongside the final ruling, would prove significant. Such a small change would have an immediate disciplining effect on the content of “issues papers” and the provision of new economic evidence. To further circumscribe the influence of staff, their active involvement in internal deliberations could be further disciplined, or even prohibited. Term limits could be imposed on the number of years that specific individuals can work for the Legal Affairs or Rules Divisions or for the AB Secretariat. In addition, various tasks handled by the Secretariat could be split between different staff members. Chinese walls could be built, for example, between those individual staff involved in panel appointments or in approving the contractual terms and compensation of panelists and AB members, and those providing procedural and substantive support to panels or the AB in a given proceeding. Finally, more robust accountability mechanisms could be both beneficial and compatible with member state incentives. Adjudicators could be given the explicit power to remove staff from a specific case or even get a formal say in the hiring, promotion and firing of staff, be it as full members of the Secretariat or as legal clerks appointed ad hoc for a given case. Like the ICJ, WTO adjudicators could also be assisted by both a permanent Secretariat (much like the ICJ Registry) and law clerks supporting individual adjudicators. Law clerks could offer a counterweight to permanent staff and, more importantly, empower individual adjudicators especially those whose native language is not English or with diplomatic rather than legal-technical skills. Adjudicator empowerment would, in turn, reduce staff-adjudicator asymmetry which has been one of the main sources of the WTO staff's outsized role. These adjustments are needed even if one reads the WTO dispute settlement system as a sui generis process of administrative review; if one views the mechanism as a judicial one, they become indispensable.

Far from robbing the institution of its perceived legitimacy as something less than a fully-fledged court, more transparency about the nature of the Secretariat's role, and more robust mechanisms to ensure its accountability, have become essential in keeping the WTO dispute settlement mechanism afloat.