Barriers to Panel Composition in RTA Dispute Settlement: Evaluating Solutions to a Perennial Problem

Abstract

In the wake of the demise of the WTO's Appellate Body, there has been a growing trend of states resorting to the dispute settlement mechanisms under their regional trade agreements (RTAs) to resolve international trade disputes. While the vast majority of these mechanisms have never been used, many contain defective procedural provisions that are likely to slow down or completely derail the dispute settlement process should those provisions be invoked. This is particularly true of mechanisms that effectively permit a respondent to block or delay the composition of a panel to hear a dispute. This article examines the issues of ‘panel blocking’ and panel composition delay tactics in RTA dispute settlement with reference to both past and present practice, and provides a textual analysis of a cross-section of existing RTAs to identify procedural defects and prescribe solutions for ensuring timely panel composition in future disputes.

1. Introduction

Following the demise of the WTO's Appellate Body in December 2019 due to the US’ persistent objections to the appointment of new Appellate Body Members, the prospect of having international trade disputes satisfactorily resolved by a binding third-party dispute settlement mechanism (TPDSM) at the WTO has been less certain than at any point since the organization's founding. While WTO Members have brought and continue to bring disputes against other Members under the WTO's Dispute Settlement Understanding (DSU) since the Appellate Body ceased functioning,¹ in 14 disputes² where a panel report has been issued, these reports have been appealed ‘into the void’, effectively blocking the enforcement of those panel reports until a reconstituted Appellate Body can consider the appeal.³ Encouragingly, efforts by WTO Members to keep appellate review alive through the establishment of the multi-party interim appeal mechanism (MPIA) has found support among some crucial Members, however so far the MPIA has not attracted the participation of other important players such as the US, India, and South Korea.⁴

Yet somewhat paradoxically, outside the WTO system, many states have recently sought to strengthen and legitimize binding third-party adjudication in state-to-state trade disputes. Since the Appellate Body ceased functioning, several large regional trade agreements (RTAs) – including, among others, the United States–Mexico–Canada Agreement (USMCA), Regional Comprehensive Economic Partnership (RCEP), and EU–UK Trade and Cooperation Agreement (TCA) – have been concluded between major trading states, all of which contain TPDSMs.

Additionally, whereas states had previously not found it necessary to resort to RTA TPDSMs, there are clear indications of a renewed focus on these mechanisms for resolving state-to-state trade disputes. Since January 2019, the EU has initiated dispute settlement proceedings under four of its RTAs. Of these four disputes, three have proceeded to the panel stage and two have produced panel reports.⁵ Two disputes between Canada and the US and one dispute involving all three USMCA parties have been brought under that agreement's TPDSM since May 2021,⁶ while New Zealand initiated the first bilateral trade dispute under the Comprehensive and Progressive Agreement for Trans-Pacific Partnership (CPTPP) in May 2022 challenging Canada's allocation of tariff rate quotas on dairy products.⁷ Though the reasons motivating the use of RTA TPDSMs to resolve these disputes are likely many and varied, given the uncertain situation at the WTO, cases that otherwise would be adjudicated under that forum may instead be resolved increasingly through recourse to RTA TPDSMs.

Given this trend, states seeking to rely on RTA TPDSMs for resolving disputes should assess whether they contain exploitable procedural flaws allowing one party to delay or indefinitely block the dispute settlement process. One such procedural issue that historically has posed a problem for complaining parties in international trade disputes is the capacity for one party to block or delay the composition of a panel to hear a dispute. Indeed, a chief criticism of the TPDSM under the General Agreement on Tariffs and Trade 1947 (GATT) was the respondent's ability to unilaterally delay or block the dispute settlement process at various stages, including the panel composition stage.⁸ While the adoption of the ‘negative consensus’ rule during the Uruguay Round eliminated this concern under WTO dispute settlement,⁹ the ability for one party to delay or block panel composition is still imbedded in many RTA TPDSMs, jeopardizing both the efficiency and effectiveness of third-party adjudication under such agreements.

More than a mere annoyance, an aggrieved party can incur significant economic penalties if faced with prolonged delays in the panel composition process. As the well-known adage goes, justice delayed is justice denied. This is particularly true of international trade disputes, in which the emphasis is on bringing the offending party into compliance with the underlying agreement rather than compensating the aggrieved party for damages suffered. As a result, the longer it takes to compose a panel to rule on an impugned measure, the longer that measure is likely to remain in force and continue causing harmful economic effects.

This article assesses the historical and contemporary instances of ‘panel blocking’ and delays to panel composition, the mechanisms that enable their exploitation in inter-state trade disputes, and considers the practical methods for removing these defects under existing and future RTA TPDSMs. In doing so, this article proceeds by: examining how parties in practice have exploited procedures that allow for panel blocking and panel composition delays in order to avoid third-party dispute settlement in international trade disputes (Section 2); conducting a textual analysis of a cross-section of RTAs and identifying common varieties of procedural defects that may allow a party to block or delay panel composition (Section 3); outlining and assessing remedial approaches to combating panel blocking employed in recent RTAs (Sections 4 and 5, respectively); and offering concrete proposals for eliminating the possibility of a party to delay or block panel composition in current and future RTA TPDSMs (Section 6).

2. Panel Blocking and Panel Composition Delays in Practice

2.1 GATT Practice

The dispute settlement system under the GATT started its life as an informal process, founded upon what Petersmann has characterized as ‘very sketchy legal and institutional underpinnings’, but over time the GATT dispute settlement process became increasingly ‘legalized’ and ‘codified’.¹⁰ Formally, GATT Article XXV:4 provides that unless otherwise stated, decisions by the Contracting Parties under the GATT – including those with respect to dispute settlement proceedings under Article XXIII – are to be taken by a majority of the votes cast.¹¹ However, since 1968 and continuing until the creation of the WTO, the unwritten practice of the Contracting Parties was to take decisions by consensus, including in approving requests to establish a panel.¹² This ‘tradition of decision making by consensus’ effectively afforded any Contracting Party, including the responding party in a GATT dispute, the power to block the composition of a panel and adoption of a panel report simply by voting against that action.¹³

Despite this tradition of consensus, Hudec emphasizes that for the majority of the GATT's existence, it was ‘established practice that complaining parties are entitled to a panel adjudication if they want one’.¹⁴ By the late-1980s, however, this practice had changed. While it was more typical for parties to allow a panel to be composed and then block the adoption of the panel report, it became increasingly common for the responding party to delay or block the composition of a panel in order to extract concessions or derail the dispute settlement process entirely. Between May 1985 and April 1989, eight complainants in GATT disputes requested the establishment of a panel, which the respondents either blocked or delayed.¹⁵ Reasons for delaying the establishment of a panel differed depending on the circumstances of each case; for example, in Nicaragua v US – Trade Measures Affecting Nicaragua, the US delayed the establishment of a panel until it was agreed that such panel would be precluded from considering the US’ GATT Article XXI defence,¹⁶ while in Finland v US – Procurement of Antarctic Research Vessel, the US blocked the establishment of a panel and called for a return to consultations, which Finland agreed to and which ultimately led to Finland abandoning its panel request.¹⁷

Hudec highlights two cases in particular where US retaliatory measures were clearly GATT-inconsistent and yet the US was able to resist compliance by delaying or blocking the establishment of panels to consider the impugned measures.¹⁸ In US v EC – Animal Hormones Directive, the US requested the establishment of a technical panel under the Standards Code to consider a proposed ban on the sale of beef from hormone-treated cattle.¹⁹ The European Communities (EC) objected to this process and instead favoured a mixed legal-technical panel.²⁰ The US refused the appointment of a mixed panel, and after breaking off the dispute settlement process under the Standards Code, retaliated by placing a 100% tariff on EC food products.²¹ The EC subsequently requested the establishment of a panel to consider the US’ retaliatory measures, however, the US blocked the panel's establishment and asserted it would only allow a panel to be established if the EC agreed to establish a technical panel in the underlying dispute.²² The EC never agreed to the US’ proposal and as a result, the panel was never established.

In Brazil v US – Import Restrictions on Certain Products from Brazil, Brazil challenged the imposition by the US of a 100% tariff on Brazilian products, which the US claimed was done in retaliation for Brazil's failure to afford protection for pharmaceutical patents owned by US firms.²³ Brazil requested the establishment of a panel in August 1988, however the US blocked the request until February 1989 after more than 60 countries intervened to express their support for the panel's establishment.²⁴ Disagreement over the terms of reference delayed proceedings even further, and the composition of the panel was not announced until September 1989.²⁵ Ultimately, the US withdrew the measure in July 1990 after Brazil committed to implementing legislation on patent protection, leading Brazil to withdraw its complaint.²⁶

By 1988, the majority view among the Contracting Parties was that restrictions should be placed on the ability of disputants to block the establishment of a panel, though the Contracting Parties in general still favoured retaining the consensus approach for adopting panel reports.²⁷ In April 1989, the Contracting Parties adopted the Improvements to the GATT Dispute Settlement Rules and Procedures which, while providing for ‘several automatic, non-consensual procedural steps that would keep the process moving even if one of the parties does not agree’, offered only an ‘ambiguous answer’ to the issue of panel blocking without actually closing the loophole.²⁸ It would not be until the creation of the WTO that panel blocking was formally eliminated.²⁹

2.2 RTA Practice

It bears mentioning at the outset that there are limitations to assessing the practice of panel blocking under RTAs since relatively few disputes have been brought under RTAs, particularly when compared to those brought under the GATT and WTO. Indeed, states have consistently shown a preference for WTO dispute settlement in circumstances when the dispute could have been brought under an RTA.³⁰ Moreover, as Porges notes, the list of known RTA disputes is likely unrepresentative of the actual number that have been brought, since the ‘vast majority of RTA dispute settlement procedures do not require public disclosure of consultation requests, panel requests, or even panel reports’.³¹ Any assessment of panel blocking in RTA practice, therefore, is limited to the handful of cases where the dispute's procedural history is known.

A well-known historic example of RTA panel blocking occurred in 2000, when Mexico attempted to compose a panel under Chapter 20 of the North American Free Trade Agreement (NAFTA) to challenge the legality of US quotas on Mexican sugar imports (the Sugar dispute).³² Although the NAFTA required that the Parties establish and maintain ‘a roster of up to 30 individuals who are willing and able to serve as panelists’, at the time Mexico initiated dispute settlement, no such roster existed.³³ Taking advantage of the absence of an established roster and procedural rules under the NAFTA allowing a party to exercise a peremptory challenge against any candidate not selected from the roster, the US repeatedly rebuffed Mexico's efforts to appoint panelists³⁴ over the course of 16 months.³⁵ Frustrated by its inability to resolve the dispute under Chapter 20, Mexico imposed retaliatory measures,³⁶ which were then successfully challenged by the US before a WTO panel and the Appellate Body,³⁷ and in investor–state proceedings under NAFTA Chapter 11.³⁸

More recently, a pending dispute where panel blocking tactics could be used – if they have not already – is the bilateral dispute under the EU–Algeria Association Agreement (AA). Since June 2015, the EU has consistently complained of a host of trade restrictive measures adopted by Algeria that allegedly contravene the EU–Algeria AA.³⁹ After efforts to resolve the dispute amicably failed, in March 2021 the EU initiated arbitration, notifying Algeria of the appointment of its panelist and inviting Algeria to appoint a panelist of its own and cooperate on the appointment of a chairperson.⁴⁰ Article 100(4) of the AA states that the responding party ‘must then appoint a second arbitrator within two months’, though the agreement provides no safeguard procedure for ensuring panel composition if the respondent fails to appoint its panelist.⁴¹ According to the EU, since March 2021 the parties have intensified ‘technical consultations for agreeing on an amicable solution’,⁴² but as of October 2022, Algeria has not appointed its panelist and there are no procedures under the AA for composing the panel absent Algeria's cooperation. As such, if the parties fail to reach a solution through technical consultations, Algeria can evade any attempts by the EU to have the issue resolved through binding dispute settlement by continuing to hold out on appointing its panelist.

Even in cases where panels were eventually formed under RTA TPDSMs, frequently, disputing parties have faced significant delays throughout the panel composition process. In all three NAFTA Chapter 20 disputes that were successfully concluded, the panel composition process greatly exceeded the 30-day timeline envisioned under the agreement. While the panels in the Canada – Agricultural Products and US – Broomcorn Brooms cases were composed in a relatively speedy six months, it took the parties almost 17 months to compose the panel in US – Cross-Border Trucking Services.⁴³ Without an established roster from which panelists could be selected, panel composition faced delays, as the parties had to agree on the appointment of each panelist.⁴⁴

Similarly, in a 2005 dispute brought under the Mexico–Northern Triangle Free Trade Agreement (FTA),⁴⁵ panel composition was delayed due to the lack of a roster. The FTA called for the establishment of two rosters: one listing nationals of the treaty parties who could be selected to serve as panelists, and another listing non-nationals of the treaty parties who could serve as chair.⁴⁶ While the former roster had been established by the time the dispute arose, the treaty parties had failed to agree on a roster of chairs.⁴⁷ In the absence of a roster of chairs, the FTA provides that the chair can only be appointed by mutual agreement of the disputing parties.⁴⁸ While the details of the panel composition process in this case are not public, the fact that it took the parties almost 15 months to compose the panel – rather than the maximum 20-day timeline contemplated under the FTA – indicates that it is likely the absence of a roster of chairs delayed the panel composition process.⁴⁹

In three recent disputes brought under EU RTAs – the EU–Southern African Development Community (SADC) Economic Partnership Agreement (EPA), the EU–Ukraine AA, and the EU–South Korea FTA – the panel composition process also took much longer than anticipated under the respective agreements. The EU–South Korea FTA and EU–Ukraine AA envision a panel being composed within 10 to 15 days after the submission of a request to establish a panel, and yet in the recent disputes brought under these agreements, the process took approximately six to seven months to complete.⁵⁰ These delays, however, pale in comparison to that experienced in the EU–South African Customs Union (SACU) dispute under the EU–SADC EPA concerning poultry safeguards. Under the EPA, the panel should have been composed within a maximum of 25 days from the request to establish the panel,⁵¹ but in SACU – Poultry Safeguards the process took nearly 20 months.⁵²

According to the EU, the delays in SACU – Poultry Safeguards were the result of unforeseen events coupled with stalling tactics by the SACU. Following the breakdown of consultations, the EU requested the establishment of a panel on 21 April 2020, but the next day agreed to suspend the panel composition process at the SACU's request in light of the emerging COVID-19 crisis. By November 2020, the EU considered it appropriate to restart the panel composition process, however the SACU disagreed and refrained from nominating its panelist. In response, the EU invoked Article 80 of the EU–SADC EPA and requested the chairperson of the Trade and Development Committee (TDC) to appoint the SACU's panelist and the panel chair. This move seems to have motivated the SACU to act, as three days later it appointed its panelist. Thereafter, the EU withdrew its Article 80 request and the parties agreed on criteria to follow for selecting the panel chair.⁵³

The EU contends that by the end of January 2021, the party-appointed panelists had identified a suitable and available candidate to serve as chair, but the SACU refused to accept the proposed candidate and instead put forth additional names and arguments, ultimately delaying the panel composition process further. The EU invoked Article 80 again, and a week later, the chair of the TDC selected the panel chair by lot. Though the three panelists were identified by early March 2021, the panel was not formally composed until late-November 2021.⁵⁴

2.3 Conclusions

While the sample size is limited, a review of past practice reveals two important lessons. First, in some cases, disputing parties are willing to take advantage of procedural rules that allow them to block or delay panel composition when they perceive it to be within their interest to do so. States may not always resort to such actions, and indeed the vast majority of GATT cases led to a positive outcome despite the inherent ability of the disputing parties to block the dispute settlement process at various stages. Nevertheless, the above cases demonstrate that when permitted under TPDSMs, disputants have been willing to exploit procedural rules in order to avoid binding dispute settlement.

Second, even when disputing parties have acted in good faith, delays have arisen in cases where the TPDSM does not provide adequate safeguards for continuing the panel composition process in the face of an impasse between the parties, or where the parties, in practice, have chosen not to invoke the safeguards under the RTA.⁵⁵ In numerous cases, these delays have been significant and prolonged the deleterious effects of the impugned measures on the aggrieved party.

While an empirical study of past cases reveals instances in which panel blocking and composition delays have frustrated proceedings, what such a review cannot capture are situations where a state chose not to initiate a dispute because it concluded that any attempt to establish a panel would ultimately be blocked or delayed by the opposing party.⁵⁶ Thus, it is difficult to determine the effect caused by the ability of a party to block or delay panel composition just by reviewing cases actually brought, as it fails to account for disputes that otherwise would have been initiated had the procedure not allowed for panel blocking in the first place.

The use of panel blocking tactics in the Sugar dispute likely contributed to the NAFTA Parties’ eventual abandonment of Chapter 20 dispute settlement in favour of WTO dispute settlement.⁵⁷ Between 1996 and 2001, panel reports were issued in three Chapter 20 disputes; however, after the US blocked establishment of a panel in the Sugar dispute, the NAFTA Parties stopped resorting to Chapter 20 altogether while continuing to litigate cases between them at the WTO.⁵⁸ According to Vidigal, between 2006 and 2017, the NAFTA Parties initiated five WTO disputes that could have been brought under NAFTA Chapter 20, while in fact no Chapter 20 disputes were initiated during this period.⁵⁹ Notably, the Chapter 20 roster was recomposed in 2006,⁶⁰ and its existence would have precluded the types of panel blocking tactics employed in the Sugar dispute. Nevertheless, this new roster lapsed in 2009 and was never re-established, indicating a lack of interest among the NAFTA Parties in resorting to Chapter 20 for future disputes.⁶¹ The active use of Chapter 31 dispute settlement under the USMCA, which provides a ‘fix’ to the panel blocking issues under the NAFTA, provides further support for the notion that the ability to unilaterally block or delay panel proceedings under Chapter 20 led to its abandonment.

3. Types of Problematic Panel Composition Procedures under RTA DSMs

Of the nearly 400 RTAs currently in force, over 250 provide for some sort of ad hoc adjudicative process.⁶² The aim of such mechanisms is to provide for effective, binding dispute settlement in the event that a dispute arises between the RTA parties. This model is also intended to be a more certain means of dispute resolution than other mechanisms that are diplomatic in nature.⁶³ Whereas a diplomatic process favours negotiated agreements, the adjudicative model injects legalism into dispute settlement and is reflective of the trend toward binding third-party dispute settlement that has become an increasingly common feature of RTAs.⁶⁴ These TPDSMs attempt to create an automatic procedure for settling disputes, but in many cases, they contain procedures that can be exploited by an RTA party to undermine that intended automaticity.

3.1 Absence of Any Default Mechanism in Case of Non-Participation

One obstacle to panel composition arises where a TPDSM does not provide any safeguard provisions in the event of non-participation in the panel composition procedure. This defect is commonly found in older EU RTAs, such as the Euro-Mediterranean (EUROMED) Agreements entered into between the EU and southern Mediterranean states soon after the WTO's establishment.⁶⁵ Indeed, a number of EUROMED Agreements still in force provide for recourse to arbitration while making it easy for one party to block the process unilaterally. These RTAs typically provide that a party ‘may notify the other of the appointment of an arbitrator’, and that ‘the other Party must then appoint a second arbitrator within two months’.⁶⁶ The process can easily be frustrated if the respondent fails to appoint the second arbitrator, as there are no provisions detailing the process for composing the panel if the respondent refuses to participate.

The same issue also arises in more recently concluded RTAs. Under Turkey's RTAs with Chile and Malaysia, safeguard procedures are in place to ensure the appointment of the third arbitrator in the event that the parties cannot come to an agreement; however, there are no procedures to ensure the panel is composed if a party fails to appoint its own arbitrator.⁶⁷ Likewise, the US–Oman FTA and US–Bahrain FTA provide that each party shall appoint one panelist, and the parties shall endeavour to agree on a chair, failing which the party chosen by lot shall select the chair.⁶⁸ Again, these TPDSMs provide a safeguard in the event that the parties cannot agree on a chair, but do not address how the panel will be composed if a party refuses to appoint its panelist.

The presence of such fundamental flaws in these RTA TPDSMs raises the question of whether they were intentionally designed to allow the parties to avoid binding dispute settlement. These systems are more reminiscent of a diplomatic model, with consultations and negotiations being the key focus of dispute settlement as recourse to arbitration essentially requires the consent of both parties.⁶⁹ Certainly, the EUROMED Agreements were intentionally designed in this manner, as all pre-2000 EU RTAs favoured the diplomatic approach to dispute settlement.⁷⁰ What is less clear is whether the same approach is intentionally incorporated into more recent Turkish and American RTAs, especially given that they include safeguards for appointing the panel chair, but fail to do the same for the appointment of panelists.

3.2 Barriers to Roster Composition

A much more common and subtler defect is found in RTA TPDSMs that establish a standing ‘roster’ of individuals, usually appointed by the parties soon after the RTA enters into force, who are to serve as panelists if called upon by the disputing parties.⁷¹ While rosters may help speed up panel composition by providing a ready-made list of potential panelists acceptable to the RTA parties, rosters may also serve as a barrier to panel composition, particularly when the roster is difficult to establish or maintain.

As discussed in Section 2 above, the inability to compose a panel due to the inexistence of a roster has been dubbed the ‘fatal flaw’ of the state-to-state TPDSM under NAFTA Chapter 20.⁷² On a cursory reading of the provisions under NAFTA Chapter 20 regarding the panel composition procedure, there is nothing that stands out as being intrinsically problematic. The disputing parties are to ‘endeavor to agree’ on a panel chair within 15 days of the delivery of the request to establish a panel, failing which the disputing party chosen by lot is to select within five days an individual who is not a citizen of that party to serve as the panel chair.⁷³ Within 15 days of the selection of the chair, each party is to select two panelists who are citizens of the other disputing party.⁷⁴ If a party fails to appoint its panelist within the allotted period, a citizen of the other disputing party is to be selected by lot from the roster to serve as the remaining panelist.⁷⁵

A critical problem arises in the selection of the remaining panelist. While the NAFTA Chapter 20 procedure anticipates the possibility that a disputing party will fail to appoint its panelist within the prescribed period, the contingency requires selection of the remaining panelist from among the roster established under Article 2009.⁷⁶ In accordance with Article 2009(1), the NAFTA Parties must ‘establish by January 1, 1994 and maintain a roster of up to 30 individuals who are willing and able to serve as panelists’.⁷⁷ Roster members serve for a term of three years, with the possibility of reappointment, but they must be appointed by consensus of the NAFTA Parties.⁷⁸

Despite the obligation to establish a roster soon after the NAFTA's entry into force, the status of the Chapter 20 roster at various points of time is ambiguous. Gantz suggests that the NAFTA Parties informally agreed on a roster in the 1990s, but it was never formally adopted or publicly disclosed.⁷⁹ The tribunal in Cargill v. Mexico, adjudicating an investor–state dispute under NAFTA Chapter 11, recalled that a Chapter 20 roster existed in 1998 but had lapsed at some point thereafter, and a roster was not successfully re-established until December 2006.⁸⁰ What is clear, however, is that the roster was not continuously maintained.⁸¹

Without a formally established roster, the panel composition procedure is vulnerable to indefinite blockage. Article 2011(3) allows a disputing party to ‘exercise a peremptory challenge against any individual not on the roster who is proposed as a panelist’.⁸² In the absence of a roster, any individual proposed by a disputing party is automatically subject to a peremptory challenge, as they are not members of an established roster under Article 2009.⁸³ Consequently, a party may indefinitely delay the dispute settlement process and effectively prevent the composition of a panel by continuously vetoing any panelist proposed by the other party.⁸⁴

Indeed, the US exploited this loophole in the Sugar dispute, revealing the ineffectiveness of Chapter 20 dispute settlement absent the cooperation of the NAFTA Parties in maintaining the roster or appointing panelists. The safeguards built into Article 2011 intended to ensure automaticity in panel composition only work if a formal roster exists at the time of the dispute. Even if a roster had been established in 1994, as required under Article 2009(1), the NAFTA Parties were apparently incapable of maintaining it. While the exact reasons for this failure are unclear, the text of Article 2009 makes maintaining a roster difficult absent the sustained cooperation of the NAFTA Parties.

First, the term limit for each roster member is capped at three years.⁸⁵ As a result, the NAFTA Parties are essentially required to enter into negotiations every three years to reconstitute the Chapter 20 roster. Though roster members can be reappointed, they must be appointed by consensus.⁸⁶ By choosing not to participate in the process of reconstituting the roster or by refusing to agree to the appointment roster members, a NAFTA Party could unilaterally frustrate the operation of Chapter 20. These textual limitations likely explain the inability of the NAFTA Parties to maintain an active roster. According to the Cargill tribunal, a 15-member roster existed in 1998 but lapsed sometime thereafter. Between 1997 and 2004, the NAFTA Parties engaged in at least six separate sets of discussions on the composition of the Chapter 20 roster before finally agreeing on a full 30-member roster in 2006,⁸⁷ however by 2009, the roster had lapsed again and was never reconstituted.⁸⁸

The potential for panel blocking arises under other RTAs that similarly envision recourse to a roster where the parties fail to agree on or fail to appoint roster members. This is particularly true of RTAs that entered into force prior to the late-2000s, which incorporate many of the limitations found under NAFTA Chapter 20. For example, the US–Singapore FTA and Canada–Costa Rica FTA also limit the term of roster members to three years, require consensus for appointment to the roster, and afford parties the right to a peremptory challenge against candidates who are not drawn from the roster.⁸⁹ Accordingly, if the RTA parties fail to agree on a new roster every three years, the roster will lapse and the respondent to a dispute under the agreement could block panel composition by using the same tactics employed in the Sugar dispute.

Newer US RTAs such as the US–Australia FTA (AUSFTA) do away with explicit term limits for roster members, however appointment to the roster still requires the agreement of both parties and is therefore susceptible to unilateral blockage.⁹⁰ Moreover, RTAs that eliminate the peremptory challenge right for non-roster appointees may still be problematic if the fail-safe provisions for ensuring panel composition in the event a disputing party refuses to participate require selection from a roster. For instance, the US–Singapore FTA does not include a right of peremptory challenge, but states that ‘a panelist shall be selected by lot from the contingent list’ if a party fails to appoint its panelist within the specified period of time.⁹¹ The fail-safe procedure is thus frustrated in the absence of a roster. Considering the US–Singapore FTA, like other problematic roster systems, requires roster members to be appointed by consensus and limits their terms to three years, there is the possibility that even if established, the roster will subsequently lapse and nullify the effect of the fail-safe provision.⁹²

Surveying the range of roster systems that could be described as procedurally problematic, some provide a greater opportunity for exploitation than others. Under NAFTA Chapter 20, the respondent is capable of avoiding dispute settlement in the absence of a set roster by vetoing non-roster appointees. Given the consensus requirement for roster appointees and the automatic expiry of the roster every three years, establishing and maintaining a roster requires ongoing cooperation between the NAFTA Parties. In contrast, RTAs such as the EU–Korea FTA that limit the consensus requirement, lengthen term limits for roster members, and eliminate the right to peremptorily challenge non-roster appointees do not require the same level of cooperation.⁹³ While there may be some slight barriers to establishing the roster, once composed it is easier to maintain. Indeed, rosters have been established under the EU–Korea FTA, the Dominican Republic–Central America FTA (CAFTA–DR), and the EU–SADC EPA, among others, despite those mechanisms relying, at least to some degree, on the good faith participation of the RTA parties in composing the roster.⁹⁴ Nevertheless, there are a number of RTAs currently in force that employ roster systems that are vulnerable to abuse by one party, which may ultimately delay or derail the dispute settlement process.

4. Approaches to Ensuring Timely Panel Composition in RTA DSMs

States have adopted a variety of methods to combat the potential of panel blocking and composition delays in RTA TPDSMs, with some mechanisms being more effective at resolving the issue than others. Four broad approaches can be gleaned from assessing RTAs currently in force: (i) RTAs with provisions safeguarding panel composition despite the absence of a roster; (ii) RTAs with provisions safeguarding roster composition; (iii) RTAs with panel composition procedures that do not envision use of a roster; and (iv) RTAs that provide for recourse to an appointing authority. This Section describes each approach, while the next Section assesses their effectiveness and efficiency.

4.1 Provisions Safeguarding Panel Composition Absent a Functioning Roster

One approach to resolving the issue that arose in the Sugar dispute is to include provisions in the TPDSM that safeguard panel composition even if the RTA parties fail to establish a roster. Under the CPTPP, if the respondent fails to appoint its panelist within 20 days of the request for the establishment of a panel, the complainant is to select a panelist from among the individuals appointed to the respondent's indicative list of panelists.⁹⁵ If the respondent has not appointed anyone, the complainant gets to select from among the individuals appointed to the roster of panel chairs.⁹⁶ Finally, if no roster of panel chairs has been established, the remaining panelist is to be randomly chosen from among three candidates proposed by the complainant.⁹⁷

The CPTPP procedure effectively closes the panel blocking loophole by ensuring that a panel can be composed despite the non-participation of the respondent and the lack of an established roster. The final stage of the fail-safe procedure – where the remaining panelist is chosen from among candidates proposed by the complainant – is dependent only upon the participation of the complainant, and thereby circumvents the need to resort to a roster. Notably, however, the CPTPP Parties managed to compose the roster of panel chairs within a year of the CPTPP coming into force.⁹⁸ Accordingly, a situation in which the process progresses to the third stage of the fail-safe procedure is unlikely to arise. Rather, in the event that a CPTPP Party fails to participate in the process, the most likely outcome is that the panel will be composed of the complainant's chosen panelist and two individuals appointed to the roster of panel chairs.

A similar approach has been incorporated into an assortment of modern EU RTAs. The EU–Canada Comprehensive Economic and Trade Agreement (CETA) TPDSM requires selection from a roster in the event that the parties cannot agree on the composition of the panel; however, Article 29.7(6) provides that if no roster has been established or if it does not contain a sufficient number of roster members, ‘three arbitrators shall be drawn by lot from the arbitrators who have been proposed by one or both of the Parties’.⁹⁹ This same sort of provision is also found under the EU–Vietnam FTA and EU–Ukraine FTA.¹⁰⁰ Other EU RTAs include similar provisions, but are more detailed and contain specific procedures for selecting the panel chair. The EU–Singapore FTA, for example, provides that in the absence of a complete roster of panel chairs, the chair ‘shall be selected by lot from among former Members of the WTO Appellate Body’.¹⁰¹ Under these types of procedures, provided the complainant proposes individuals for appointment to the roster by the time the panel is to be composed, the lack of a formalized roster will not stymie panel composition.

4.2 Provisions Safeguarding Roster Composition

Another less common method of combating the panel blocking issue is to include provisions within the TPDSM that guarantee a roster will be composed, regardless of whether all RTA parties participate in composing the roster. This is the approach employed under the USMCA. Just as the panel composition process under NAFTA Chapter 20 is dependent on the existence of a functioning roster, so too is the process under Chapter 31 of the USMCA.¹⁰² Yet, whereas establishing and maintaining the roster under NAFTA Chapter 20 was a considerable challenge absent the good faith participation of all NAFTA Parties, the USMCA allows for the creation and maintenance of a roster without the active participation of all three USMCA Parties.

In appointing members to the roster, each USMCA Party is allowed to designate up to 10 individuals.¹⁰³ Article 31.8(1) states that the Parties ‘shall endeavor to achieve consensus on the appointments’, but unlike under the NAFTA, consensus is not required.¹⁰⁴ If each of the USMCA Parties nominates its 10 individuals, and there is no agreement on all 30 proposed individuals within 30 days, the individuals nominated by each Party are appointed to the roster.¹⁰⁵ However, if for example the US failed to nominate its 10 individuals while Canada and Mexico each nominated their prospective roster members, the 20 individuals proposed by Canada and Mexico would form the roster. As a result, provided one USMCA Party proposes individuals to the roster, the roster will be formed within 30 days of the agreement entering into force.¹⁰⁶ The USMCA also eliminates the three-year term limit for roster members found under NAFTA Chapter 20.¹⁰⁷ With respect to reappointments and new appointments, the Parties are to seek consensus, however if there is no agreement within 30 days, the individual proposed is added to the roster.¹⁰⁸

In practice, this procedure has worked remarkably well. The USMCA entered into force on 1 July 2020, and the roster of panelists for Chapter 31 disputes was released only a few days later.¹⁰⁹ The fact that there are 30 individuals appointed to the roster indicates that there was active participation by the US, Canada, and Mexico in appointing individuals to the roster.¹¹⁰ As a final safeguard, the Rules of Procedure for Chapter 31 disputes outline the procedure to be followed in the event the failure of a Party to make its roster appointments ‘impedes the composition of a panel … due to an insufficient number of individuals on the roster’.¹¹¹ Yet, recourse to this procedure is not necessary given how quickly the Parties moved in fully composing the roster.

4.3 Ad hoc Panel Composition Procedures

Given past challenges in composing and maintaining rosters and the potential for panel blocking that may arise as a result, many RTAs forego the roster system entirely and employ an ‘ad hoc’ approach. One type of panel composition procedure commonly found in RTAs involves each disputing party appointing its arbitrator, and proposing a list of three or four individuals to serve as chair.¹¹² If a disputing party fails to appoint a panelist within the specified period of time, the remaining panelist is either chosen by the other disputing party, or selected by lot from among the four individuals proposed by that party to serve as chair.¹¹³ The same safeguard applies for appointing the chair, whereby the chair is appointed from among the individuals proposed by the complainant if the respondent fails to participate, or if the parties otherwise cannot agree on a chair.¹¹⁴ This simple procedure therefore affords the complainant the ability to compose a panel unilaterally without the participation of the respondent and without having to resort to an established roster to appoint the remaining panelists.

4.4 Recourse to an Appointing Authority

In the event that the disputing parties are unable to agree on the appointment of panelists, several RTAs resort to a designated appointing authority to overcome the impasse.¹¹⁵ Under this approach, a political body or representative steps in and appoints the remaining panelists on behalf of the parties if they cannot come to an agreement.¹¹⁶ A number of RTAs designate a representative of a body internal to the RTA system as the appointing authority. The Protocol of Olivos, which governs dispute settlement for the Southern Common Market (MERCOSUR), designates the MERCOSUR Administrative Secretariat as the residual appointing authority.¹¹⁷ Likewise, under the Enhanced Dispute Settlement Mechanism (EDSM) of the Association of East Asian Nations (ASEAN), if the disputing parties fail to agree upon the appointment of panelists, it falls to the ASEAN Secretary-General to appoint the remaining panelists.¹¹⁸

EU RTAs frequently feature both roster and internal appointing authority systems, where the appointing authority steps in to appoint a panelist or chair from the roster by lot in the event that the disputing parties cannot agree. Under the CETA, if the parties cannot agree on the composition of the panel, a party can request that the panelists be drawn by lot from the roster by the Chair of the CETA Joint Committee or their delegate.¹¹⁹ Similar provisions are found in both the EU–SADC EPA and the EU–Japan EPA.¹²⁰

Other RTAs make recourse to an external appointing authority in the event that the disputing parties cannot agree on the composition of the panel. A host of Asian–Pacific RTAs, including the China–New Zealand FTA and ASEAN–Japan Comprehensive Economic Partnership Agreement (CEPA), designate the WTO Director-General as the appointing authority.¹²¹ The Secretary-General of the Permanent Court of Arbitration (PCA) is also frequently designated as the appointing authority in RTAs,¹²² and the President of the International Court of Justice (ICJ) has been designated as the appointing authority under a few agreements.¹²³ The Regional Comprehensive Economic Partnership (RCEP) is unique in that it provides a fail-safe appointing authority; it designates the WTO Director-General as the default appointing authority, but in the event that the WTO Director-General fails to act, a party may request that the PCA Secretary-General make the necessary appointments.¹²⁴

5. Assessment for Approaches to Ensuring Timely Panel Composition in RTA DSMs

In many instances, states have reduced the likelihood of panel blocking and panel composition delays in their more recent RTAs, but have not employed sufficient fail-safe mechanisms that would guarantee, in all circumstances, timely panel composition upon a party's request. Accordingly, there are still opportunities for parties to undermine dispute settlement procedures by failing or refusing to participate in panel or roster composition. As the previous Section demonstrates, states have employed various methods to respond to the panel blocking problem and ensure automaticity in roster or panel composition. This Section offers a critical analysis of the effectiveness and efficiency of each of those approaches,¹²⁵ but first identifies a common theme that appears in the TPDSMs most successful in eliminating panel blocking.

5.1 A Common Approach to Eliminating Panel Blocking: Redistributing Procedural Control and Incentivizing Participation

What defective RTA panel composition procedures tend to have in common is that they do not allow the complainant to compose a panel unilaterally without cooperation from the other RTA parties or a third party. Under many roster-style RTAs, the consensus requirement for appointments to the roster necessitates the active participation of the RTA parties in composing the roster, and without a composed roster, the fail-safe provisions designed to ensure automaticity are inoperable. Other RTAs provide no safeguards in the event the respondent fails to appoint its panelists. And under RTAs that designate an appointing authority, the complainant cedes control of the process to the appointing authority, which may refuse to act (see Subsection (4) below). In these cases, the complainant cannot push ahead with composing the panel and is either left at the mercy of the respondent, or some other third party, which may result in unintended delays.

It follows that RTA TPDSMs that either vest control over the dispute settlement process with the complainant, or create strong incentives for RTA parties to participate in the process, or ideally, do both, should effectively ensure timely panel composition. This is a common element of effective panel composition procedures regardless of whether the TPDSM employs a roster, ad hoc, or appointing authority approach. Accordingly, panel blocking can be eliminated from each type of panel composition procedure, provided the TPDSM either vests control over the process with the complainant or creates strong incentives for the RTA parties to actively participate.

5.2 Efficiently Eliminating Panel Blocking in Roster Systems

Many modern RTAs that employ a roster system effectively deal with the panel blocking issue, but do so using different methods. The USMCA approach guarantees roster composition and maintenance without the participation of all the RTA parties. This in turn incentivizes each party to participate in composing the roster, because by failing to participate a party may lose out on appointing its preferred candidates to the roster.

The CPTPP approach, in contrast, does not guarantee that a roster will be established, but allows the complainant to appoint individuals not on the roster to the panel in the event of the respondent's non-participation and without the respondent being able to block those appointments. This procedure also encourages the respondent's participation in composing the panel, as the complainant can appoint two panelists if the respondent does not participate.

Finally, the approach adopted under recent EU RTAs permits the disputing parties to select as panelists individuals proposed to the roster, even if the roster has not been formally established or does not contain a sufficient number of roster members required to preside over a dispute. While ensuring that a panel can still be composed in the absence of a full roster, this procedure also encourages the parties to make their roster appointments. A complainant may bring a dispute and at the same time propose members for the roster to ensure there are prospective panelists even if the roster is not formalized prior to the panel being composed. The complainant's proposals would then motivate the respondent to make its own proposals for the roster to minimize the likelihood that the panel is composed entirely of individuals nominated by the complainant. Each of these approaches therefore prevents another party from unilaterally frustrating the dispute settlement process, while also providing incentives for the RTA parties to participate in the panel or roster composition process.

While the USMCA, CPTPP, and EU RTA approaches effectively deal with the panel blocking issue, all three have the potential to be inefficient. Under the CPTPP TPDSM, the respondent can still delay the panel composition process by choosing not to participate. In particular, the procedure for selecting the panel chair is vulnerable to extended delays that are outside the complainant's control, especially where the respondent refuses to participate.¹²⁶ If the disputing parties fail to agree on a chair, which would be inevitable if the respondent refused to participate, it falls to the two appointed panelists to select an individual from the roster of panel chairs.¹²⁷ If the panelists cannot agree on a chair, the CPTPP provides that ‘the two panelists shall appoint the chair with the agreement of the disputing Parties’.¹²⁸ Again though, if one of the parties refuses to agree, the procedure is further delayed, with the next step being that the disputing parties ‘select the chair by random selection’ from the roster of panel chairs.¹²⁹ Unlike the USMCA, the CPTPP does not provide guidance on how to proceed if a disputing party fails to participate in choosing the chair via ‘random selection’.¹³⁰ In any event, it can take up to 60 days to reach this point, and the complainant cannot accelerate the process if the panelists cannot agree on appointing a chair and if the respondent refuses to participate.¹³¹ Accordingly, while effective in eliminating panel blocking, the CPTPP approach is likely to result in panel composition delays in the absence of the participation of the respondent.

In contrast, under the USMCA, the complainant has control over the process in the absence of the respondent's participation. At each stage, the USMCA envisions that the parties will attempt to come to a consensus on appointments, but provides that in the event of the respondent's non-participation, the complainant shall make the appointment.¹³² In total, the panel composition procedure is intended to take a maximum of 40 days, allowing the complainant to compose the panel quickly and unilaterally if the respondent refuses to participate.¹³³

Importantly though, the USMCA retains the right of parties to exercise a peremptory challenge against non-roster appointees, which is vulnerable to abuse by respondents. In selecting panelists under the USMCA, disputing parties can either appoint from the roster or choose someone not on the roster, though in the latter case there is a risk that the appointee will be subject to a peremptory challenge by the opposing party. For complainants, there is a strong incentive then to select from the roster, as panel composition could be delayed if the respondent challenges the complainant's non-roster appointee. For respondents, however, the risk of a peremptory challenge is less consequential as resulting delays in the panel composition process simply means a delayed ruling, which does not prejudice the respondent in the same way it prejudices the complainant. In fact, there is an incentive for respondents to go ‘off the roster’, as it forces the complainant to choose between capitulating and accepting the respondent's preferred appointee or delaying the process further by exercising a peremptory challenge.

Indeed, the USMCA experience thus far suggests that the peremptory challenge right cancels out any incentive for the respondent to appoint panelists from the roster. In the three cases brought under Chapter 31, the respondent in two of those cases did not select its panelists from the roster.¹³⁴ In both cases, the panel composition process exceeded the timeline established under the USMCA.¹³⁵ By contrast, in Canada – Dairy TRQs, all three panelists were drawn from the roster and the panel composition process still exceeded the established 35-day timeline, but only by six days.¹³⁶ These cases suggests that while the peremptory challenge right may not necessarily lengthen the panel composition process, it negates any incentive for the respondent to appoint panelists from the roster, which could lead to panel composition delays unless the complainant capitulates on the respondent's appointee.

Examining the EU RTA approach to eliminating panel blocking, delays may arise under those RTAs that employ a hybrid roster/appointing authority approach (discussed further in Section 5.4 below). Setting aside potential issues with the appointing authority for now, the EU RTA approach seems to provide an efficient means of eliminating panel blocking. Under the CETA, for example, the disputing parties are to consult with a view to agreeing on the panel's composition within 10 days of the request for the establishment of the panel.¹³⁷ If they cannot agree, it falls to the chair of the CETA Joint Committee to appoint the remaining panelists within five working days, drawing by lot from the panelist and panel chair rosters.¹³⁸ Again, if any of the rosters have not been formally established at the time the party requests the chair's assistance, the chair will draw from among the individuals proposed for inclusion on the rosters by one or both of the parties.¹³⁹ In theory, the process for appointing panelists should not exceed 15 days, and the non-participation of the respondent in the process does not risk delaying that timeline further.

In practice, however, long delays in panel composition have been common in recent disputes involving these types of hybrid procedures.¹⁴⁰ The main reason for delay seems to be parties’ reluctance to submit a request to the appointing authority to make the outstanding appointments, even if doing so would expedite the process. Though panel composition in SACU – Poultry Safeguards took nearly 20 months to complete, it also revealed the effectiveness of the hybrid roster/appointing authority approach when properly invoked. After the SACU failed to appoint its panelist, the EU requested the chair of the TDC to appoint the SACU's panelist, ultimately motivating the SACU to appoint its panelist three days later to avoid appointment by the TDC chair. When the parties could not agree on a panel chair, the EU requested the chair of the TDC to appoint the chair, which he carried out five days later. While disputing parties may have good reason to hold off on moving forward unilaterally with panel composition by invoking fail-safe provisions,¹⁴¹ practice suggests that the hybrid roster/appointing authority approach provides for efficient panel composition when such provisions are invoked in accordance with the timeline established under the TPDSM.

5.3 The Roster vs. Ad Hoc Approaches to Panel Composition

The ad hoc panel composition approach may be effective in preventing the respondent from delaying or blocking panel composition, and also in encouraging the respondent's participation. As detailed in Section 4.3, under a well-designed ad hoc system, the complainant retains control over composing the panel in the event that the respondent does not participate. Moreover, the respondent has a strong interest in participating in the panel composition process under such systems since if it fails to appoint anyone, the panel is composed entirely of the complainant's proposed panelists.

This approach is appealing in its simplicity. The parties are neither required to negotiate and agree on a roster for the TPDSM to be operable nor resort to an appointing authority to overcome an impasse. Yet, it is difficult to compare the merits of the roster and ad hoc approaches since, as Trakman rightly points out, there is a lack of compelling empirical evidence, suggesting that one system is clearly preferable to the other.¹⁴² The primary argument for the roster system is that it ensures that if a dispute arises, the parties can appoint panelists from a list of individuals who, in principle, they have already accepted as suitable candidates. A person named to the roster is regarded as having the required qualifications to serve as a panelist and, generally speaking, should only face disqualification for certain forms of misconduct, such as by failing to make disclosure or failing to exercise independence and impartiality.¹⁴³ In composing the roster, RTA parties put candidates through a ‘pre-clearance process’ to check for conflicts of interest and confirm the candidates are properly qualified before they are considered for a specific dispute.¹⁴⁴ Therefore, by drawing panelists from a roster, the argument is that RTA parties can avoid potential delays by reducing the likelihood that the selected panelist will be challenged or decline the appointment due to a conflict of interest.¹⁴⁵

Unfortunately, testing this hypothesis is difficult. Of the few publicly known RTA cases, nearly all involved some form of roster-based TPDSM. Moreover, the details of negotiations (if any) over the selection of panelists are rarely made public. What is evident, however, is that the mere existence of a roster does not always lead to timely panel composition. As discussed in Section 2.2, in three recent bilateral disputes under EU RTAs, it took the parties between six and 20 months to compose the panels for each dispute,¹⁴⁶ despite the fact that the underlying RTAs in all three disputes required the establishment of a roster and rosters had, indeed, been established prior to each dispute.¹⁴⁷ In Korea – Labour Commitments and Ukraine – Wood Export Ban, the original panelists were all selected from the established rosters, yet there were still delays in composing the panels. The USMCA experience thus far suggests that recourse to a roster system could result in timely appointments; however, as noted in the previous Section, this is dependent on the parties actually choosing panelists from the roster. Consequently, a proper comparison of the efficiency and effectiveness of the roster and ad hoc approaches in RTA TPDSMs requires more empirical evidence than is currently available.

5.4 Recourse to an Appointing Authority

Recourse to an appointing authority can provide an effective method of panel composition in the event that the disputing parties cannot agree on the panelists, or where one party refuses to participate.¹⁴⁸ Indeed, parties in international commercial and investment disputes frequently resort to appointing authorities for assistance in panel composition, which have proven to be rather efficient.¹⁴⁹ Nevertheless, there are potential issues with recourse to both internal and external appointing authorities.

Where an internal representative acts as the default appointing authority, problems may arise depending on how that representative is assigned to their position. The Korea–Chile FTA, for example, designates as the appointing authority the ‘chairperson of the [Free Trade] Commission’ when the parties cannot agree on a chair or panelist.¹⁵⁰ However, the agreement provides that the Free Trade Commission (FTC) ‘shall be chaired alternately by each Party’.¹⁵¹ The EU–SADC EPA provides a similar process, whereby the TDC chairperson is tasked with making appointments, but the chair alternates between the Parties each year.¹⁵² These procedures raise an obvious question: what happens if the chairperson is a national of the respondent and simply refuses to appoint any panelists?¹⁵³ Where the appointing authority lacks complete neutrality between the disputing parties, there is an increased risk that the panel composition procedure may be subject to vulnerabilities.

The CETA also provides for recourse to an internal appointing authority in the event that the parties cannot agree on the panel's composition within 10 days, allowing either party to request that the chair of the CETA Joint Committee or their delegate draw by lot arbitrators from the roster.¹⁵⁴ Article 26 states that the CETA Joint Committee shall be co-chaired by government officials of both Canada and the EU, and that decisions and recommendations of the CETA Joint Committee shall be made by mutual consent.¹⁵⁵ It is not clear, therefore, whether both co-chairs must approve of the selection by lot, or whether the co-chair of the complainant state can act unilaterally absent the consent of the other co-chair. Unfortunately, the Rules of Procedure of the CETA Joint Committee do not offer any clarification.¹⁵⁶ If mutual consent is required, panel composition may be susceptible to paralysis if the respondent's co-chair fails to cooperate.

The EU–Japan EPA seems to resolve this issue. If a party fails to make its appointment, then the co-chair of the Joint Committee from the complainant state steps in to appoint a panelist from the roster.¹⁵⁷ The same general procedure applies in the event that the parties cannot agree on a panel chair, with the co-chair of the Joint Committee from the complainant state making the selection by lot from the roster of panel chairs.¹⁵⁸ Like the CETA, lack of a roster under the EU–Japan EPA is not itself a barrier to composing the panel, provided the complainant has formally proposed individuals for the panelist and panel chair rosters at the time the panel is to be composed.¹⁵⁹ The EU–Japan approach is therefore preferable to the CETA approach as it explicitly permits the co-chair from the complainant state to unilaterally compose the panel, allowing the complainant to push forward with panel composition even in the absence of a roster and the active participation of the respondent.

In light of the potential issues with making recourse to an internal appointing authority, an external appointing authority may be preferred. There are multitudes of such options parties could select from, many of which have significant experience in acting as an appointing authority in international disputes.¹⁶⁰ Still, parties should exercise caution in designating an external appointing authority, as the designated body or individual may still refuse to act even if designated. This may be problematic where RTA parties have chosen the WTO Director-General as the appointing authority. While the Director-General is frequently involved in appointing panelists in WTO disputes, no Director-General has ever appointed panelists under an RTA, nor has a Director-General ever taken a public position on the matter.¹⁶¹ As such, it is unclear whether the Director-General would act even if designated under an RTA.

In contrast, the President of the ICJ is known to occasionally act as an appointing authority.¹⁶² This role and its procedure are not formally set out under the Rules of the Court or the Statute of the ICJ, and ICJ publications do not disclose how often the President fulfils this ‘extrajudicial’ function,¹⁶³ however the ICJ recommends that treaty parties consult the President on their willingness to act as an appointing authority.¹⁶⁴ Shaw notes that in acceding to any request to act as appointing authority, an essential consideration for the ICJ President is ‘that the performance of the function in a particular case should not interfere with the performance by the Court of its primary functions’.¹⁶⁵

The PCA Secretary-General will act as an appointing authority if so designated, provided he is satisfied on a prima facie review of the documents submitted by the parties that he is empowered to act.¹⁶⁶ Consequently, as long as the dispute settlement clause explicitly names the PCA Secretary-General as the appointing authority, and assuming all other conditions for submitting a request have been met,¹⁶⁷ the Secretary-General will act. In contrast with the ICJ, appointing authority services fall within the PCA's regular scope of business and are not subsidiary to the PCA's other dispute settlement functions. Having responded to over 800 appointing authority requests since 1981, the PCA has significant experience in providing such services and the Secretary-General will give effect to appointment procedures, including the timelines for appointment, established under the applicable dispute settlement clauses.¹⁶⁸

6. Concrete Proposals for Eliminating Panel Blocking

The RTA TPDSMs that appear most effective in eliminating the possibility of panel blocking are those that inhibit the ability of one party to control the roster or panel composition process, while providing incentives for the parties to participate in that process. However, these approaches differ in their effectiveness in eliminating panel blocking, as well as their overall efficiency in ensuring timely panel composition. This Section offers concrete proposals for states to eliminate panel blocking under their RTA TPDSMs in a manner that maximizes procedural efficiency in panel composition.

6.1 Improvements to the Roster-Based Approach

As discussed above in Section 3, panel blocking only tends to arise under roster-based TPDSMs if no roster has been formally established, or has lapsed. The CPTPP, USMCA, and EU RTAs employ different approaches to resolving this issue, with varying effectiveness and efficiency. For existing RTAs with pathological roster-based TPDSMs, the simplest method for eliminating panel blocking is to adopt the EU RTA approach. For example, since there is no established roster under the AUSFTA or the KORUS, there is a possibility that the panel composition process could be blocked as the safeguard provisions for ensuring automaticity are tied to the existence of a roster.¹⁶⁹ The RTA parties could resolve this issue by amending – or perhaps by issuing an interpretation of – the text to provide that in the event the roster under the agreement is not fully composed, the remaining panelists shall be selected by lot from among the individuals formally proposed by the parties for inclusion on the roster. The addition of this simple provision would not require a substantial redrafting of the panel or roster composition procedures that already exist under the RTA, yet would provide two crucial benefits. First, it would allow the complainant to compose a panel, even in the absence of an established roster, provided the complainant proposes its individuals for inclusion on the roster. Consequently, the respondent would not be able to block the process by failing to participate in composing the panel or roster. Second, this type of provision would encourage the parties to create a roster, since the complainant would have to propose individuals for the roster in order to ensure the panel is composed, and the respondent would likely respond by proposing its own individuals to ensure the panel is not composed entirely of the complainant's preferred candidates. Accordingly, in adding this language to an existing RTA TPDSM without a fully composed roster, the parties could guarantee panel composition in the absence of a roster while also encouraging the creation of a roster.

For effectively eliminating panel blocking in future RTAs while ensuring efficient panel composition, states would be prudent to adopt a modified USMCA approach. The advantage of the USMCA is that it guarantees a roster will be quickly composed regardless of whether all RTA parties participate in composing the roster, while simultaneously encouraging all parties to participate by making non-participation highly disadvantageous. Moreover, the complainant retains control over the panel composition process from start to finish, unlike under the CPTPP and EU RTA approaches, where the complainant must resort to a third party to finalize appointments to the panel. However, states should deviate from the USMCA by making selection from the roster mandatory, which would eliminate any risk of delay resulting from challenges to non-roster appointees.

6.2 Improvements to the Appointing Authority Approach

The appointing authority approach loses its efficiency (and may be ineffective in preventing panel blocking and delay tactics) where the complainant cedes control over the process to an internal body that can be stymied by the respondent, or to an external body that may not act in a timely manner, if at all. For states that wish to incorporate an internal appointing authority into current and future RTAs, the best course of action would be to adopt a procedure similar to that provided under the EU–Japan EPA. The procedure under that agreement allows for recourse to an internal appointing authority, but eliminates the opportunity for the respondent to influence the appointing process by allowing the complainant to appeal to their own co-chair of the Joint Committee in selecting the panelists and chair. This procedure guarantees that a panel will be composed as long as the complainant takes steps to try to compose the panel and enlists the assistance of its co-chair if the respondent fails to participate.

Under RTAs such as the CETA and EU–Vietnam RTA that provide for recourse to an internal appointing authority that is co-chaired by representatives of the parties, but do not specify whether a decision on appointments must be taken by consensus, this ambiguity could be resolved by adopting an interpretation of the provision in question, or by writing the clarification into the applicable rules of procedure. Neither course of action would require any significant change to the panel composition process or the structure of the appointing authority, but would effectively eliminate the possibility of delays or panel blocking by allowing the complainant's co-chair of the ministerial body to make the necessary appointments absent the respondent's participation.

If parties prefer an external appointing authority, it is advisable to select an institution with significant experience in acting as an appointing authority in order to avoid any delays in the panel composition process. In this regard, the PCA has extensive experience in acting as an appointing authority in a large number of cases, including state-to-state disputes. Provided the parties make a proper request to the PCA Secretary-General to act as the appointing authority, there is little doubt that he will act in a timely manner and compose the panel in line with the parties’ needs and preferences. The same cannot be said of requests made to other potential appointing authorities that do not typically provide the institutional appointing authority services offered by the PCA.

6.3 Bringing Practice in Line with Treaty Text

Efforts to eliminate panel blocking and delay tactics in the text of an RTA are pointless if, in practice, the treaty parties do not follow the TPDSM's procedures. This means that, where the complainant has the ability to overcome procedural roadblocks in the panel composition process, it is imperative that they invoke these procedures to achieve timely panel composition.

Nevertheless, there may be times when it is within the parties’ interests or reasonable given the circumstances to adopt a flexible approach and to not strictly adhere to the procedure under the RTA. Though the EU alleges that the panel composition delay in SACU – Poultry Safeguards is partly attributable to the SACU's actions, both parties agreed to amend the applicable procedural timelines in light of the COVID-19 pandemic. Flexibility may also be beneficial where the disputing parties are engaged in constructive negotiations that could be jeopardized by one party unilaterally moving ahead with panel composition. Moreover, it may be the case that it is simply not practical or possible to follow the treaty text if certain procedural steps take longer in practice than anticipated under the RTA. Gantz, for example, questions whether the 30-day timeline for panel composition under NAFTA Chapter 20 was achievable even if panelists were selected from a roster given the need to confirm the availability of prospective panelists and conduct conflict checks.¹⁷⁰ Accordingly, it may be reasonable for a complainant to refrain from moving forward unilaterally with the panel composition process if the respondent is making good faith efforts to assist in composing the panel, even if doing so results in some delay.

There are other cases, however, where the disputing parties have departed from the RTA language in composing the panel, for reasons less apparent, and faced prolonged delays as a result. The delays to panel composition under NAFTA Chapter 20 were in part attributable to the lack of an established roster, but also the failure of the NAFTA parties to follow the agreement's procedures for selecting the chairperson. Article 2011(1)(b) provides that the disputing parties shall endeavour to agree on a chair within 15 days; however, if they cannot agree, a party is ‘chosen by lot’ to select an individual not a citizen of either party to serve as chair.¹⁷¹ In practice, the parties never used the selection ‘by lot’ method for appointing the chair; rather the chair was appointed by agreement among the disputing parties.¹⁷² Though the treaty language is unclear as to how and by whom this selection by lot process should be carried out,¹⁷³ a similar procedure was used successfully under the NAFTA's predecessor, the Canada–US FTA.¹⁷⁴ It seems likely that the NAFTA Parties could have employed a similar system that would have been consistent with the text of Article 2011 while also accelerating the chair selection process.

SACU – Poultry Safeguards is another example of panel composition being delayed because the parties deviated from the treaty text. The EU–SADC EPA provides that the two party-appointed panelists are to select the chair within 20 days of the request to establish a panel.¹⁷⁵ In SACU – Poultry Safeguards, the parties agreed on an alternative procedure that required the party-appointed panelists to identify suitable and available candidates to serve as chair and then invite the parties’ comments. Presumably, the intention of implementing this alternative procedure was to give the parties a greater say in the selection of the chair, whereas under the EPA the party-appointed arbitrators usually have sole discretion in appointing a candidate. The EU accepted the panelists’ candidate, while the SACU did not, leading to further delays as the parties sought consensus. Ultimately, the chair was appointed without the input of either party, as the impasse led the EU to invoke Article 80(3) of the EPA and call on the chairperson of the TDC to select the chair by lot. Had the parties chosen to follow the language of the treaty, the chair would have been appointed five weeks earlier.¹⁷⁶

Where states have negotiated workable panel composition procedures, disputing parties should avoid deviating from those procedures unless there are legitimate reasons and it is within the interests of both parties to do so. In any event, should disputing parties conclude that a deviation is reasonable, it should be limited in scope, set out in writing between the parties, and explicitly applicable only to the dispute in question to avoid its use as a precedent for future disputes.

7. Conclusion

The uncertain state of dispute settlement under the WTO system has provided a renewed focus on the use of RTA TPDSMs as an alternative means of resolving state-to-state trade disputes. Certainly, the proliferation of RTA disputes in recent years suggests that states are increasingly relying on these long-neglected mechanisms. However, both past and present experience demonstrates that where a TPDSM contains flawed panel composition procedures, the effectiveness of that mechanism is undermined, if not nullified. Many RTAs currently in force, varied as they are, contain such defects, which can be (and have been) exploited by respondents at the expense of delaying justice for complainants.

While many states have attempted to eliminate panel composition defects in newer RTAs, the results have been mixed. In most cases, these RTAs have eliminated the possibility of panel blocking, but panel composition is still susceptible to delays. Nevertheless, there are steps states can take to rid their current and future RTAs of such defects, regardless of whether those agreements employ a roster, ad hoc, or the appointing authority approach to panel composition. States would be wise to take a serious look at the panel composition procedures under their current RTA TPDSMs, and consider how they put those procedures into practice, to identify barriers to timely panel composition and determine the best course of action for overcoming those barriers to ensure effective and efficient dispute resolution in future state-to-state trade disputes.