To

Recent debates on the operation of the WTO’s dispute resolution mechanism have focused primarily on the Appellate Body (AB). We argue that this neglects the first-order issue confronting the rules-based trading system: sustaining the principle of de-politicized conflict resolution that is reflected in the negative consensus rule for adoption of dispute settlement findings. Improving the quality of the work of panels by appointing a roster of full-time professional adjudicators, complemented by reforms to WTO working practices that reduce incentives to resort to formal dispute settlement, can resolve the main issues that led to the AB crisis. Effective, coherent, and consistent WTO dispute resolution need not include an AB. An appropriately redesigned single-stage process can serve just as well, if not better.


Introduction *
The quest by Uruguay Round negotiators when they crafted what became the WTO DSU (Dispute Settlement Understanding) was to design an effective system to resolve trade disputes. The mechanism they crafted, the means to achieve this end, had two fundamental components: (i) a two-instance adjudication regime; and (ii) a negative consensus decision-making rule that made it impossible for any of the parties to a dispute to block the formation of a dispute settlement panel or the adoption of a ruling by the adjudicators. These two elements, prima facie at least, constituted a paradigm shift, not just for the multilateral trading system, but more broadly for adjudication in the realm of international relations. Of the two features, negative consensus was the real innovation. Going from a single instance to a twoinstance dispute settlement system is simply immaterial, if it is possible to block the adoption of the final reporthowever generated. Removing the possibility to block adoption of a dispute settlement report implies that findings issued by foreign experts must be implemented at home. 1 In this paper, we question whether the debate among WTO members on the operation of the secondinstance court of the world trading system, the WTO Appellate Body (AB), addresses the first-order issue confronting the trading system. We argue that the goals and interests of WTO members, as reflected in the substance of the rules negotiated in the Uruguay Round and the institutional arrangements put in place to implement the various agreements and monitor trade policies, can only be served if negative consensus for adoption of rulings is maintained. Whether there is an AB is not material to the realization of the objectives WTO members established for themselves in 1995. The WTO membership may lose little, if anything, from a decision not to bring the AB back. Instead, it could benefit from opting for a few simple institutional amendments to strengthen the panel stage of dispute settlement.
The plan of the paper is as follows. In Section 1, we characterize the key features of the WTO dispute adjudication regime. Section 2 revisits the arguments in favour of having an AB in the first place, drawing on both arguments advanced during the Uruguay Round negotiations and the rationale for appeals courts in domestic jurisdictions. In Section 3, we discuss several amendments to the DSU that would be required, if the WTO were to proceed without an AB, conditioned on the premise that WTO members continue to support a de-politicized dispute adjudication system, i.e., remain committed to the negative consensus rule. In our view, this is criticalmoving away from negative consensus is likely to spell the end of the post-war trade regime. In the current geo-economic context, in which an ascending power (China) confronts a declining power (US) trying to hold on to its privileges, and both players have little in common when it comes to perceptions about adjudication, it is impossible to think of positive consensus-based decision-making to resolve trade disputes. An implication is that "going back to the GATT" is not a feasible second-best solution to the AB dispute. Section 4 places the role and operation of formal dispute settlement in the broader context of WTO reform, highlighting the importance of bolstering other mechanismsall of which already existto prevent invocation of formal dispute resolution. Section 5 concludes.

Setting the scene: historical context
The two innovations embodied in the DSUtwo -instance adjudication (the creation of the AB) and negative consensuswere, from a negotiating perspective, intertwined. As noted in Hoekman and * We would like to thank Bill Davey, Claus-Dieter Ehlermann, Rodd Michael Izadnia, Alejandro Jara, Patrick A. Low, Doug Nelson, David Palmeter, Alan O. Sykes, David Walker and Robert Wolfe for helpful discussions and comments on previous drafts.

2
Robert Schuman Centre for Advanced Studies Working Papers Mavroidis (2020), two-instance adjudication was an insurance policy against the loss of sovereignty resulting from the passage to negative consensus, even though the ultimate statutory expression of this function (as reflected in Article 17 of DSU) does not make this explicit. In his monumental work on the GATT, Hudec (1993) showed that the passage to negative consensus associated with the creation of the WTO was a matter of evolution, not revolution: over time the GATT had come to operate on a de facto quasi-negative consensus basis. Nonetheless the formalization of negative consensus for dispute settlement was a major innovation by making explicit what had gradually come to be the common practice.
The AB crisis puts some and perhaps all of this into question. The officially expressed critique of the AB by the United States centres on the way the AB has executed its mandate. The US argument is that AB rulings imply an impermissible usurpation of national sovereignty, and that this justified forcing through single-handedly the demise of the AB in December 2019. This has left the WTO with only the first instance panel-stage of the DSU. In April 2020, a partial substitute version of the AB was put in place by an EU-led coalition of WTO members, the Multi-Party Interim Appeal Arbitration Arrangement (MPIA). 2 At the time of writing nineteen WTO members have signed the MPIA. They represent 14 percent of the membership, counting the European Union (EU) as one. Signatories include several WTO members that are heavy users of the dispute settlement system (e.g., Australia, Brazil, Canada, China, the EU and Mexico). The MPIA illustrates that some important members of the WTO are committed to the maintaining the possibility of appealing panel findings on an interim basis until a permanent resolution of the AB conflict can be found by the WTO membership. Disputes between MPIA signatories accounted for about a quarter of the total DSU case load during 1995-2019, suggesting the MPIA will cover a substantial share of future potential disputes. If more WTO members also decide to join, this 'plan B' will provide a half-way house in allowing many disputes to continue to be resolved through a two-instance system. However, the MPIA does not and will not encompass the United States, which was involved as a defendant in 18 percent of post-1995 WTO disputes. Any longer-term solution must include the US and the more than 100 other WTO members that have not signed the MPIA and that collectively accounted for about half of the formal dispute settlement case load during 1995-2019.
Although the United States has not formulated clear proposals in favour of undoing negative consensus, it has left the door open to this possibility. 3 There are legitimate concerns to the effect that the current US administration effectively wants to go back to the GATT days, when formation of panels and adoption of panel reports could be blocked. Alternatively, the US goal may be limited to going 'back to 1995', that is, the US accepts the principle that the DSU should be a two-instance system, but it seeks to ensure that the AB does not overstep the mandate it was given when the WTO entered into force. In this scenario the presumptive goal of the US is to ensure this will not happen again. 4 Going back to the GATT (i.e., giving up negative consensus) would seriously undermine the rulesbased trading system. It is already a setback that the AB has been put out of business, as it suggests an erosion of the commitment to pursue a de-politicized dispute resolution system. Irrespective of the eventual assessment of the record of the AB during its 25 years of operation, the very existence of the AB was a concrete signal of the acceptance of the WTO membership to de-politicize the workings of trade dispute settlement. Removing the AB from the institutional edifice established during the Uruguay This position can be inferred from the many interventions of the US before the DSB in 2019. See e.g., Hoekman and Mavroidis (2019).

To AB or Not to AB? Dispute Settlement in WTO Reform
European University Institute 3 round may cast doubt on the continued commitment to this goal. We will argue in what follows that the demise of the AB is not sufficient, in and of itself, for this concern to become a reality. Hudec (1993) shows that de-politicization of dispute settlement was an incremental process. The legitimacy that early panellists enjoyed, the quick resolution of disputes, and the pragmatic remedies (an insult to the purist, but a welcome feature for the trading community) contributed towards an increase in trust in "GATT panels", the body called to perform the task of adjudication. 5 Hudec also points to a reversal in the march towards de-politicization. During and after the Tokyo round-era (the years following 1973) several reports issued by GATT panels remained un-adopted. Technically, this meant that the report would not become a decision of the GATT CONTRACTING PARTIES. 6 As a result, on occasion some technically impeccable panel reports never became the law of the "trade land". By refusing to consent to adoption of the panel reports, as they had the legal if not a moral right to do, losing parties reduced the value of such reports to almost nothing.
This was not because the dispute extended to areas where it was unclear whether existing disciplines addressed the matter raised in a dispute. There were examples of such cases, notably disputes concerning the use of farm subsidies, where panel reports had been adopted. Instances of non-adoption most likely reflected two factors: i.
Those losing before panels would rather legislate the matter head-on in the context of a trade round, and receive a quid pro quo for whatever liberalization they might agree upon; and/or ii. Political economy factors: non-adopted cases often concerned agricultural support policies and at the time the farm lobby, especially in the European Community, was very powerful.
The refusal to adopt panel reports exacerbated the intensity of voices in the US in favour of unilateralism, with the ensuing surge in Section 301 cases in the 1980s. The transatlantic partners resolved their differences with respect to farm trade in the WTO Agreement on Agriculture. Judging by the number of cases against the EU, the prime culprit in the realm of farm trade during the GATT years, this 'legislative solution' embodied in the Agreement on Agriculture proved a stable equilibrium. According to the WTO webpage, 84 cases were raised between 1995 and the end of March 2020, where at least one claim involved an allegation that the Agreement on Agriculture had been violated. On 13 occasions, the EU was defendant, a small number by any reasonable account, were one to view it in its historical perspective. Only 2 of these 13 cases involved subsidization (DS283 and DS286, both dealing with export subsidies on sugar). The remaining cases concerned either public health-related measures (e.g., domestic regulation of genetically modified organisms or hormone-treated beef), or administrative measures that covered farm goods as well as other types of products. Importantly, the EU implemented all adverse reports, including, albeit with long delay, the report on hormone-treated beef through a mutually agreed solution with the US.
The foregoing is not meant to suggest that the reason for Section 301 cases was the EU refusal to adopt reports condemning its farm policies. This was an exacerbating aspect, but this is all it was. Many cases were initiated against other trading partners, including Brazil, India, and Japan. Quite often the subject-matter of Section 301 complaints concerned policy areas not covered by the GATT mandate, such as investment, services, and intellectual property. 7 It is against this background that the Uruguay round was negotiated. In part, an agreement to extend the coverage of the world trading system to new areas was the quid pro quo for the US to stop pushing 5 Reputed scholars like Davey (1987) and Schwartz and Sykes (2002), who have analyzed GATT dispute settlement in detail, share the basic view and arguments developed by Hudec. 6 The convention during the GATT years was to capitalize all letters of this term to denote instances when the highest organ in the GATT edificeall signatories of the GATT acting in concertengaged in deliberations or took decisions. Robert Schuman Centre for Advanced Studies Working Papers for this extension unilaterally. The passage to negative consensus and imposition of tight deadlines to enforce the assumed obligations was the quid pro quo for the US to accept to limit the use of Section 301 to an instrument of pure diplomatic protection. Since only states have access to WTO courts, Section 301 would become the means for private parties to alert the US government about illegalities abroad. The US government in turn, would evaluate the merits of pursuing a complaint at the WTO, observing the agreed rules as embedded in the new DSU.
The DSU was one element of a resounding rejection of unilateral enforcement of national trade law by large economies against trading partners, and the associated trade tensions and tit-for-tat retaliation by targeted nations. The negotiating history is discussed in greater detail in Mavroidis (2016a). The agreed objective was a conflict resolution system with embedded strict deadlines, where decisions could be consistently enforced. This is shorthand for de-politicization. The AB was a last-minute addition that was meant to serve this quintessential feature of the agreed system of adjudication. The key to getting there was the willingness to engage in a negotiation to establish the substantive rules of the game to be enforcedthe successful conclusion of which involved a complex package deal that reflected numerous domestic political economy trade-offs in all major countries. 8

What is the AB Good For?
Our discussion so far points to the conclusion that the prime objective of the DSU was de-politicization of conflict resolution. The pragmatic evolution of GATT dispute settlement, as Hudec (1993) has demonstrated, contributes to this perspective. The enactment of the DSU formalized it. In what follows, we explore whether the WTO would lose this quintessential feature, were the membership to decide to proceed without the AB. We do so under the assumption that not reconstituting the AB does not also lead to ditching the rest of the DSU, i.e., our premise is that all WTO members continue to want to have access to an operational multilateral dispute settlement system. Crucially, we will also assume that the demise of the AB does not imply that WTO members ditch negative consensus as well. Given these assumptions, we will argue that if the WTO membership were to strengthen the panel stage of dispute settlement by establishing a body of full time professional panellists, the loss of the AB would not matter for the realization of the purpose of the DSU: effective and consistent enforcement of the substantive trade policy commitments made by WTO members.
With this in mind, we turn to the main question that is the subject of this Section, namely, why was the AB part of the DSU package? What is its purpose? Understanding this is a necessary condition for considering how a DSU without an AB might be designed to achieve the same goal.

Institutional Reasons for the AB
The original conception of GATT dispute adjudication comprised two phases:  A diplomatic phase, where contracting parties would attempt to resolve their disputes through bilateral consultations;  A "court-like" phase, where, in case consultations had failed to resolve the dispute, a panel of three impartial and independent experts would be asked to do so.
This latter phase, originally at least, was closer to a second diplomatic bite at the apple as opposed to the courts that operate in domestic settings. Eventually, especially after the 1962 Uruguayan complaints 8 Thus, for example, the Agreement on Agriculture reflected not only the goal of US farmers to improve access to the European Community (EC) and to reduce EC export competition on world markets, but the interest of finance ministers in European and other OECD member countries to reduce the magnitude and the inefficiency of agricultural support programs that had come to constitute a significant fiscal burden as well as imposing large costs on domestic taxpayers and consumers. See e.g., Hoekman and Kostecki (2009). against various practices by developed nations, the process and the reasoning adopted by panels became more and more court-like. As is well known however, GATT panels continued to demonstrate substantial pragmatism in addressing the issues before them (Hudec, 1993). They were also pragmatic in recommending 'soft' remedies, with the exception of a few cases during the 1980s in which panels recommended revocation of antidumping and countervailing duties, and reimbursement of all duties found to be illegal. 9 This was a short-lived parenthesis that was soon closed.
Pragmatism also pertained to the selection of the panellists. Panel members were part of the trading community, individuals well versed in the negotiation of the underlying contract. Dana Wilgress, a senior GATT negotiator for Canada, was the most frequent panellist during the first years (Mavroidis 2016a). The situation did not change in the later years. In the GATT period, the panellists were trade officials: GATT negotiators and/or members of trade delegations. Even when the roster of individuals not affiliated with governments was established, 10 it comprised overwhelmingly former government officials. The GATT dispute settlement ethos reveals a preference for individuals that have participated in the negotiation, that have inside knowledge regarding what is at stake.
When the GATT moved to bring in outside experts to assist in adjudicating disputes involving technical issues associated with measures that did not fall within the standard ambit of trade policy, as in the dispute on the US domestic international sales corporation (DISC), there was an immediate, almost institutional, backlash. The dispute was brought by what was then the European Community and concerned an element of the US fiscal regime through which US-based producers could benefit from lower income tax rates on income derived from exports. The US responded with a case against the EC, alleging that European tax regimes also acted to promote exports through lower taxation of income derived from export activities.
The panellists chosen were not the usual suspects. They were tax experts, academics, who reportedly often met in closed rooms without assistance by the GATT Secretariat. They owned the process, and drafted the final reports. Both complainants prevailed, and the two reports were submitted to the CONTRACTING PARTIES for adoption in 1976. In a dramatic, and probably unexpected reversal, the GATT CONTRACTING PARTIES, when considering adoption of the submitted panel reports, decided that only the US was in violation of its commitments. The choice of panellists, the process from A to Z, and the eventual dramatic U-turn by the CONTRACTING PARTIES, all provoked the interest of legal scholars. The leading trade law experts of the time, in separate accounts [Davey (1987); Hudec (1988); Jackson (1978) and (1990)] agreed that this was probably a dispute too far for the GATT system. They also agreed that it was a turning point for the dispute settlement process.
The DISC case had at least two consequences of interest to the subject of this paper: i.
Stakeholder voices in favour of de-politicization of dispute settlement gathered pace. The US was particularly unhappy with the outcome, for understandable reasons; ii. The then Director-General of the GATT, Arthur Dunkel, succumbed to the many voices arguing that the GATT Secretariat's resources in terms of legal expertise should be strengthened. As a result, a GATT Legal Office was created in 1982. 11 The decision by the GATT CONTRACTING PARTIES in the DISC case was not only unprecedented, it was never repeated. But there was a parallel reaction. Both the membership and the GATT bureaucracy came to the view that adjudication could not be totally outsourced. This belief was shared for both good and bad reasons. Some wanted to increase their "ownership" over the outcome of disputes. Others pointed to the risk of incoherent jurisprudence if some centralization of the process could not be established. The establishment of a GATT Legal Office was meant to provide in-house expertise on past 9 This is discussed at greater length in Mavroidis (2016). 10 GATT Doc. L/5752, published in GATT Doc. BISD 31S/9 et seq. Robert Schuman Centre for Advanced Studies Working Papers case law. Its small staff of Secretariat members were to be keepers of the institutional memory, and to contribute towards guaranteeing coherent jurisprudence across disputes.
In a way, without having the power to decide cases themselves, staff assigned to the Legal Office were mandated to exercise a function akin to that of the AB. In principle, whenever requested, the members of the Legal Office were supposed to provide their understanding of where GATT case-law was on a specific issue. Of course, this is an exercise which involves an element of subjectivity. The challenge was to reduce this element when explaining the status of extant interpretations of GATT agreements and commitments. But the upshot was that by introducing de facto precedent into dispute adjudication, other, political elements, such as bargaining power, the temporary context etc., would become less prevalent. The trading community would then be taking another step towards depoliticization.
Viewed from this perspective, the AB is a further stage in the quest for de-politicization. By appointing seven members for a fixed term, adding language in both the DSU and the Working Procedures for the AB underscoring, besides expertise, the need for independence (from the rest of the WTO), and impartiality (from parties to the dispute) for its members, GATT/WTO adjudicators were provided for the first time with institutional guarantees enabling them to serve their function without heeding to concerns of a political nature. Indeed, Article 17.3 of DSU, states inter alia, that the AB members "… shall be unaffiliated with any government. "It's Working Procedures, and especially Annex II (entitled "Rules of Conduct for the Understanding on Rules and Procedures Governing the Settlement of Disputes") prohibits AB members to accept employment conflicting with their function, further strengthening this point. The application of this provision relies on self-disclosure, the details of which are included in Annexes 2 and 3 to Annex II of the AB Working Procedures. 12

Guarantee Coherence in Jurisprudence
The statutory language of the DSU does not address explicitly the function of the AB, other than by explaining its powers. In its case law, the AB held that its jurisprudence should be observed by subsequent panels dealing with the same issue. It made the link between coherence in jurisprudence and Article 3.2 of DSU, which reads: The dispute settlement system of the WTO is a central element in providing security and predictability to the multilateral trading system. Coherent case law is thus, in the AB understanding, the means to guarantee security and predictability. First, in US-Stainless Steel (Mexico) in § §158-162, and then in US-Continued Zeroing ( §362), the AB underscored this point. The de facto precedential value has become a bone of contention, and a salient feature in the US critique of the AB. A benign reading of its position would suggest the US is not putting into question the content of Article 3.2 of DSU. Indeed, its whole legal system is built around this premise. The US is unhappy with the rote repetition of what it considers the "wrong" precedent. No repetition in other words, only for the sake of observing precedent. After all, the WTO regime does not know of binding precedent (stare decisis). We will proceed on this basis in what follows.
In principle, the framers of WTO could have inserted a paragraph acknowledging stare decisis. They did not. They could have also crystallized case law into a legal document, a GATT decision. Indeed, as Davey (1987) notes, this is exactly what the GATT members did in 1979 when they re-visited panel practice up to that time, and agreed on its description. This did not happen either. But, in Article 17.6 of the DSU, they pointed in the direction of de facto precedential value of AB reports, when stating that: An appeal shall be limited to issues of law covered in the panel report and legal interpretations developed by the panel.
Issues of law, by definition, apply to a wide set of facts. The same law, in other words, should apply to the same facts, irrespective of the identity of the disputing parties. 13 This should entail that, absent distinguishing factors, one should expect adherence to the precedent. The AB reports on US-Stainless Steel (Mexico) ( § §158-162) and US-Continued Zeroing ( §362) cited above, are, viewed from this perspective, but a short walk from the statutory language embedded in Article 17.6 of the DSU. With few notable exceptions, like the zeroing cases, WTO panels have consistently followed prior AB rulings.
Is an AB needed to assure security and predictability? We do not think so. In domestic adjudication, jurisdiction may also be a function of geographic location. For example, depending on whether a contract has been signed in Nordrhein-Westfalen or Bavaria it could be that the courts of Dusseldorf or Munich are competent to adjudicate disputes. Appellate courts serve also the purpose of homogenizing case law produced by courts in different jurisdictions in the country. Formal WTO disputes, however, can only be brought before WTO panels. 14 Yes, by changing the identity of panellists, there is still a risk for incoherent case law. Different panellists might see the same issue in different ways, and, as stated above, the WTO Secretariat has no mandate to compel coherence. It can only suggest it. However, an AB is not the only solution to this potential problem. An alternative is to bolster the panel process so as to improve the quality and coherence of the reasoning of panels, thereby making a second stage redundant. We discuss this in more detail in Section 3. Suffice to state for now that the AB is not passage obligé for coherence to be achieved. There are feasible alternatives.

Brief Re-Cap
Our discussion and analysis so far, supports the following conclusions:  The negotiation of the WTO DSU was a decisive step towards de-politicization;  The AB is part and parcel of this process; and  The purpose of the WTO DSU is to promote security and predictability in international trade relations.

De-politicization with a Single Stage DSU
If one accepts the argument that the primary institutional rationale for the AB is the de-politicization of dispute settlement, the question of interest is whether achievement of this goal can be achieved if WTO members limit adjudication to a one-instance system. As noted above, our presumption is that the trading community wants to honour the Uruguay Round commitment to de-politicization of the adjudicating process. While perhaps unrealistic, we think it is important to undertake this thought experiment because it leads to the conclusion that in principle de-politization does not require an AB. It can be achieved by making changes to the panel stage of the dispute settlement process that: (i) improve the quality of adjudication, thereby reducing the prospects of legal errors (the main task of the AB) and, as important, (ii) address weaknesses in the performance of panels that are due to inadequate analysis and/or inconsistent treatment of facts across similar cases over time.
In our view, the core of any serious effort to bolster the panel process must centre on the establishment of a standing body of full-time panellists. While observing collegiality, panellists would be tasked with adjudicating disputes, supported by legal clerks and professional economists tasked with 13 Whether the facts are the same is an important matter in practice-one where panels could do a better job and one on which the AB was precluded from opining. More on this below. 14 As we discuss in Section 4 below, the DSU is just one mechanism to resolve disputes. Depending on the type of issue involved other WTO mechanisms may be more appropriate and effective in addressing problems associated with the (non-) implementation of WTO commitments by a member. Greater support for and use of such mechanisms would reduce the burden placed on formal dispute resolution.

8
Robert Schuman Centre for Advanced Studies Working Papers establishing the facts and providing analysis of the empirical aspects of disputes. A key benefit of having permanent panellists would be the need to look for distinguishing factors every time they deviate from their reasoning and rulings in previous cases. What would be implied in terms of changes to the DSU? All that is required is to delete one provision of the DSU (Article 17) and amend several others:  Article 8 would have to be amended so as to introduce a procedure for permanent panellists;  Article 16.4 would have to be amended as well o Negative consensus will continue to be the decision-making mode; but o There will be no room for appeals anymore  All references to the Appellate Body and the possibility for appeal (for example, in Articles 18, 19) would have to be removed as well.
The new provisions should, in our view, reflect the following elements.

A Small Number of Adjudicators Working in Chambers
The membership should agree on twelve to fifteen panellists. Judging from the number of cases submitted to the WTO since 1995 15 and the workload involved per case, as calculated in Nordström (2005), this number seems more than adequate to deal with the workload. Organizing work in chambers of three or five members, depending on the disputed value of trade affected, would expand the ability of the panel body to handle a multiplicity of cases. Disputes above a certain value threshold or in areas where no case law exists, would be submitted to the plenum.

Criteria for Eligibility
All permanent panellists must have demonstrable expertise in trade law and/or economics. Prior experience at the bench, is an asset but should not be a requirement. Panellists should not be affiliated with governments and be constrained in their ability to take up government positions when their terms expire. The aim should be to take seriously potential bias and conflicts of interest. Ideally criteria for appointment should include a prohibition on appointment of officials who are on leave from their governmental function and a requirement that panellists cannot go (back) to government for five years after their employment as panellists, or work as lawyer for governments submitting disputes to the WTO.

Selection Process
Each WTO member chooses from another's list. For example, the US would choose an individual from the EU list, the EU from the Indian list, and so forth. The process would aim at generating a shortlist of forty potential panellists comprising at least one person from a WTO member that was one of the twenty most frequent users of WTO dispute settlement between 1995-2020. In parallel, a committee of eleven people, comprising senior academics, judges from the highest national courts, and experienced past panellists/AB members would be selected by the WTO membership. This committee would select fifteen panellists from the shortlist put together by WTO members.
One third of the selected panellists should have non-traditional trade policy expertise, i.e., knowledge of services trade issues, domestic regulation, intellectual property, and investment policies. Another one third, which could overlap with the prior category, should have demonstrable economics expertise. The WTO is a contract where knowledge of economics is more than warranted, it is necessary. The factual dimensions of WTO cases increasingly revolve around determining both the behavioural incentives created by trade-related policies and the size of the associated effects on economic variables of interest.
These are matters that require knowledge of economics so that panels can assess the type of analysis that should be performed by their clerks and the secretariat, and effectively use the results of such analysis.

Terms of Appointment, Administrative Support and Staffing
Appointed panellists should serve one non-renewable term of eight years. Half of the first cohort would be appointed for only six years to ensure continuity between the first and the second group of panellists. Appointing panellists for an extended non-renewable term achieves two different objectives: the opportunity to profit from the expertise of incumbents, and to ensure that panellists will not use their output as signalling mechanism for re-appointment. 16 Panellists should own their decisions. The best way to guarantee this, is to break totally from the WTO paradigm, where the Secretariat is heavily involved in the process. In (large) part the reason for this is that currently panellists are ad hoc appointments. Moving to permanent panellists entails that Article 8.6 of the DSU falls into desuetude: there is no need any more to select panellists on a case-bycase basis and much less of a rationale for heavy involvement and support by the Secretariat.
The WTO Secretariat suffers from crisis of identity. It is called to assist panels with the drafting of reports, and also to advise delegations on all sorts of legal questions. When doing the second, they may inadvertently prejudge the former. If a delegation first asks for advice on the status of national law, and then submits a dispute where the advice received becomes a key input, this could happen. This problem did not arise for the AB Secretariat, as it was not implicated in the work of the WTO Secretariat. The problem here is that AB members did not appoint the AB Secretariat. Staff therefore are unlikely to have the relationship of trust that say a US Supreme Court Justice enjoys with his/her clerks. Such a relationship is what is needed in a context of where panel adjudication is professionalized. Consequently, panellists should be given the necessary budget to hire their own clerks, who will be serving them, and not the institution.

Complementary changes
There are a several other areas where changes to the DSU could both further improve the quality of dispute settlement outcomes and address key concerns that arose regarding the operation of the AB, very vocally in the case of the US and more quietly by others. 17 First, the statutory deadlines for panels to complete their work are clearly unrealistic. Panels continuously violate the six-and/or nine-month rule, as documented by the data compiled in Johannesson and Mavroidis (2017). Since there will be no need to select panellists anymore, a twelvemonth statutory deadline seems reasonable. Deviations, whenever warranted and upon justification should, of course, be agreed.
Second, the natural consequence of Article 3.2 DSU is that adjudicators are agents and not principals. They can adjudicate but cannot invent law to adjudicate. This should have as natural consequence the acknowledgment of non liquet (Hoekman and Mavroidis, 2020). Panellists should be allowed to throw the ball back to the membership, whenever they feel that they do not possess the necessary tools to 16 Blustein (2019) points a bleak picture in this respect. 17 For example, Korea was ostentatiously unhappy with the AB report on Korea-Various Measures on Beef (DS169), and criticized the approach adopted by the AB. The absence of serious investigation into market conditions was at the heart of the criticism. The AB could have concluded it was not able to complete the analysis because the facts were poorly discussed. It did not and proceeded to find against Korea notwithstanding the facts. The US was unhappy with the findings in US-FSC (DS108), stating that it simply could not be that a measure aiming at levelling the play field would be characterized a prohibited subsidy. Chile did not hide its disapproval of the AB finding that a progressive taxation scheme, like dozens of others around the world, had been found to be GATT-inconsistent in Chile-Alcoholic Beverages (DS110). Robert Schuman Centre for Advanced Studies Working Papers adjudicate a dispute. Surprisingly, this has never happened so far, even though there have been clearcut cases where an argument in favour of non liquet seemed strong. 18 Panels have consistently hidden behind in dubio mitius in such cases. There is a problem here. In dubio mitius means that the panel feels the current legal regime suffices to adjudicate a dispute, but the interpretation that is more favorable to the defendant should be privileged. Non liquet means there is no law to adjudicate a dispute. The membership should go back to the negotiating table and design the law it wants to see applied in similar cases.
Non liquet respects the legislative imperative embedded in Article 3.2 better than the invocation of in dubio mitius does. This principle should find its way into the DSU, possibly as an addition to the current Article 3.2. Such cases should be adjudicated by the plenum and not a chamber, as they amount to an acknowledgement by the adjudicating body that the provisions of existing agreements are insufficient to make a determination regarding grievances that have been put forward. Note that if the DSU was to be limited to a single stage conflict resolution system, the use of non liquet would be more straightforward than one where there is both a panel and an appeals stage. In a two stage system, panels may feel less constrained to declare non liquet because there is a prospect of legal appeal, while the AB may be less inclined to throw cases back to the membership insofar as it sees its role as the final arbiter on the applicable law and its interpretation.
Third, the DSU prohibits the AB from addressing how panels have considered the factual dimensions of each case. Improving performance on this front by definition requires action focused on the panel stagebringing back the AB will not resolve this source of underperformance because the DSU prohibits the AB from (re-) considering the facts of a case and their interpretation. Improving the quality and consistency of the factual analysis undertaken by panels would reduce a source of tension that arose vis-à-vis the AB, as appeals often concerned matters of fact as opposed to matters of law, and thus could not be addressed by the AB.
Take DS449, US-Countervailing and Antidumping Measures on Certain Products from China. 19 The US Department of Commerce (DOC) decided in the 1980s that nonmarket economies would not be subject to countervailing duty investigations. The US Congress changed the law in this regard leading the DOC to impose countervailing duties in all then pending cases involving nonmarket economies. Not surprising, China was unhappy with this change and challenged the shift in approach. One of issues before the WTO adjudicating bodies concerned retroactive application of the legislation (the Congressional decision). The US argued, among other things, that municipal law was a factual determination and as such should escape the scope of review by the AB. The AB could and should have stopped the analysis there and then. Municipal law, as per standing case law in both the GATT era and the WTO, should be considered as a factual issue, over which the AB has no authority to rule, as per the unambiguous wording of DSU Article 17. If the AB took the view that on this occasion, it was not dealing with a factual issue, it should have explained why this was so. It did not, instead spending several pages discussing in detail the US law, a matter it had no mandate to address. 18 For example, panels and the AB could have asked whether a second compliance panel on the same case is possible, as the DSU is silent on this. The DSU further presumes one reasonable period of time during which implementation should occur. Second, compliance panels routinely deal with measures adopted after the expiry of the implementation period (see, for example, DS108). They do so, arguing the second implementing measure has a nexus with the original measure, which is judge-made law without any statutory underpinnings. In a similar vein, panels, and the AB eventually, could have asked whether panels can legitimately decline a request to issue a suggestion. The wording of Article 19 of the DSU is at best ambiguous in this regard. 19 https://www.wto.org/english/tratop_e/dispu_e/cases_e/ds449_e.htm European University Institute 11

Connecting to the Broader WTO Reform Agenda
The dispute regarding the operation of the AB that led to its eventual demise in December 2019 is just one, albeit important, element of the challenge confronting the WTO membership. The consequent response by the EU, supported by Brazil, China, Canada, Mexico and a dozen other countries, to create the MPIA, while understandable, risks fragmenting the dispute settlement system without addressing weaknesses in its design and operation noted above and discussed at greater length in the literature (e.g., Mavroidis, 2016b). Even if the alternative proposed here would be considered, there are other, arguably more fundamental challenges confronting the WTO, notably the difficulty of negotiating new agreements and revisiting old ones (Hoekman and Mavroidis, 2020). Revitalizing the WTO as a forum for trade cooperation is critical to sustain the rules-based multilateral trading system. The large trading powers need to agree on rules of the road in areas where the WTO has little to say or where current disciplines are outdated. More narrowly construed, a willingness to negotiate is also critical for the suggestions made above regarding DSU reform. Thus, the use of non liquet presupposes a readiness of the parties to a dispute on a matter that is not covered by the WTO or where the rules are unclearperhaps by designto (re-) engage in negotiation on the rules.
Clearly this is difficult to envisage in the current context of geopolitical rivalry between the two major global trade powers. What this suggests a need is to facilitate and support dialogue that will help establish a basis for multilateral cooperation on trade policy matters. 20 Many WTO members are actively engaged in an effort to identify reforms to WTO working practices that would increase the effectiveness of WTO bodies as fora to oversee the implementation of specific agreements. Some of the proposals and suggestions made by external observers stress the utility of using WTO bodies as venues for dialogue, learning about the causes of non-implementation of commitments and for informal resolution of potential conflicts (dispute prevention). Some of the ideas that have been put forward are not new in the sense that they already have been applied in the context of some agreements. Building on this experience and adopting what has emerged as good practice more broadly could help establish a basis for greater dialogue on contrasted trade policy issues. It could also reduce the burden on the formal dispute settlement system. Wolfe (2020aWolfe ( , 2020b provides an extensive and insightful assessment of procedural innovations that could be pursued across WTO bodies that largely involve building on initiatives that have been taken in specific WTO committees but not adopted across the organization. Many of his suggestions pertain to actions that would prevent disputes by fostering more dialogue through use of mechanisms such as support for raising specific trade concerns (STCs) 21written questions posed by one member regarding (planned) regulatory measures of another member -and 'thematic sessions' that focus on areas of policy relevant to a WTO committee or other body and that is of interest (concern) to members.
Several proposals have been made in the context of WTO reform discussions that would increase the use and effectiveness of such mechanisms as a means to prevent disputes, identify areas where there are gaps in the rules and address implementation problems. 22 One example of such a procedure has been adopted in the WTO Trade Facilitation Agreement (TFA). Article 18 TFA calls on the TFA Committee to establish an Expert Group of five independent trade facilitation professionals to examine the situation and make recommendations to resolve implementation difficulties notified by developing countries after 20 Such cooperation is more likely than not to take the form of plurilateral agreements, as illustrated by the ongoing 'joint statement initiatives' on investment facilitation, e-commerce, domestic regulation of services and micro-and medium-sized enterprises and plurilateral initiatives in 2020 to remove barriers to trade for medical supplies, protective equipment and food products in the context of the COVID-19 pandemic (Hoekman and Sabel, 2020). 21 Karttunen (2020) provides an excellent in-depth analysis of the genesis and development over time of the use of STCs in the WTO. 22 A prominent example is a proposal for guidelines for the operation of all WTO bodies put forward by the EU supported by Robert Schuman Centre for Advanced Studies Working Papers transition periods have expired. No recourse can be made to the DSU until the recommendation the Expert Group has been considered. An implication is less emphasis on recourse to lawyers and invocation of the DSU in favour of cooperation to help governments implement commitments and use of the TFA Committee as a forum for deliberation.

Concluding Remarks: DSU+
Dispute settlement is an important part of the WTO reform challenge that has attracted both too much and too little attention. Too much in the sense that the DSU is just one element of enforcing commitments. As has been argued by many observers what is, and can be, done through other WTO bodies and mechanisms is as if not more important in sustaining cooperation. 23 Too little in the sense that the main axis of the response to the unquestionably indefensible methods used by the US to remove the AB from the board has centred on retaining as much as possible the status quo ante (i.e., the initiative by the EU to establish the MPIA), as opposed to a serious effort to consider how to use the crisis as an opportunity to improve dispute settlement.
Putting in place a professional standing body of panellists endowed with the necessary institutional guarantees will permit this body to act in an impartial and independent manner. Impartiality and independence are distinct notions. The first concerns the relation between adjudicators and the disputing parties; the second, the relation between WTO panels and the other WTO functions. Both are equally important for dispute adjudication to be de-politicized, and for panellists to act as agents and not as principals.
More consistent legal analysis by panels, based on better assessment and analysis of the facts would address at least some and perhaps many of the underlying issues that led to the demise of the AB. Whatever the view that is taken of the suggestions made above, we believe that if the WTO membership continues to stick with the status quo panel process, not much would be achieved by reconstituting the AB as it was, or if this proves impossible, extending the number of signatories to the MPIA. The platitude "never waste a good crisis" should be taken to heart. There are design flaws in the DSU "as is" that extend beyond the AB, notably weaknesses associated with the panel stage. Addressing those through the establishment of a standing body of full-time panellists that can rely on the support of a mix of professional lawyers and economists to establish the facts associated with cases and have much stronger incentives to use such resources to ensure consistency in reasoning and findings over time would go a substantial way towards addressing many of the underlying reasons for the AB crisis.
The AB crisis is not strictly confined to design and operation of the DSU: it is a crisis of multilateralism. 24 Past success in taming GATT-inconsistent border protection has presented the WTO with the formidable challenge of taming behind-the-border protection across a very diverse set of countries. The existing agreements and those that hopefully will be signed in the future, will necessarily be incomplete. Moreover, the outcome of adjudication is more palatable and easier to implement when it involves rejection of excessive tariffs than when it pronounces the illegality of measures that reflect pursuit of a regulatory objective that applies irrespective of the origin of the goods/services involved. The use of second-best instruments is routinely predicated on mixed motives. The new DSU should aim to do what Article 3.7 of the current DSU already postulates, but this time seriously: resolve conflicts. In our view, conflict resolution is as much conflict avoidance as it is adjudication.
The good news is that there is no need to re-invent the wheel. Conflict-avoidance can be effectively promoted by widening the use of instruments already in place. We see a lot of potential in two of them. The provisions in the TFA establishing a mechanism through which subject experts evaluate the reasons for non-implementation of commitments is an illustration. This approach could be applied in the wider 23 See, e.g., Kostecki (2009), McDougal (2018), Karttunen (2020). 24 Hoekman and Mavroidis (2020). See more broadly, for example, Wolfe (2015), Baldwin (2016), Maull (2018) and Smith (2018).

To AB or Not to AB? Dispute Settlement in WTO Reform
European University Institute 13 WTO context, helping to both identify reasons for difficulties countries have in implementing specific WTO agreements or doing so consistently, in the process generating knowledge on good practices and informing adaptive change in the specifics of cooperation in a given area.
Second, efforts have already been undertaken to expand the use of STCs across all agreements. It must be encouraged. The success of STCs is largely because it is experts who participate in the discussions and debate on the rationale, appropriateness and the necessity of potential trade effects of proposed and adopted measures. Why cannot experts also participate for example in such deliberations when it comes to services regulations and measures that may be permitted in principle but have detrimental effects of trade that could be avoided without undercutting the realization of the domestic objective?
Active use of such dispute prevention and cooperation promotion mechanisms, other things equal, can reduce, as per Article 3.7 of the DSU, the number of disputes submitted to panels. We will not, consequently, be leaving it to the discretion of WTO members to decide whether to submit or not to submit. We will be encouraging them to use the dispute-avoidance mechanisms before taking it to the next step. Since administrative resources, because of budgetary constraints, are limited, panels would be asked to focus their energy and attention on fewer cases. The quality of their output can only improve as a result.
Professionalized dispute adjudication will also help. For starters, reducing the system to one-instance adjudication will ipso facto remove the US complaint that facts and law are at times mischaracterized by the AB. Panels will have to discuss both facts and lawand have a much stronger incentive to do a serious job on both fronts. Facts must be considered more seriously for findings to become more consistent and thus to provide the predictability needed by traders. A change in the applicable standard of review in many cases is therefore warranted. Very often assumed obligations protect competitive opportunities. This does not mean that trade effects are immaterial. Indeed, it is based on actual trade effects that we can speak of protection of competitive opportunities in the future. Professional panels, employing both law and economics expertise, will hopefully do so.