Towards an Effective Appellate Mechanism for ISDS Tribunals

Abstract This Article identifies the problems of an Appellate Mechanism for ISDS Tribunals in relation with its possible benefits. We propose the inclusion of certain design features to improve the working of an eventual Appellate Mechanism and help mitigate problems related to procedural, conflict resolution, and substantive concerns. We finish by identifying the most central problems with a possible Appellate Mechanism, which helps to narrow down options within the ongoing reform process at UNCITRAL. Overall, we illustrate how institutional choice is always contextual and that all institutional options are imperfect and subject to important trade-offs.


Introduction
Pressure to reform 'Investor-State Dispute Settlement (ISDS)' mounted in recent years as investors, governments, and advocacy groups expressed concerns about the existing system. 1 One proposal of note is the creation of an 'Appellate Mechanism (AM)', which would add an additional layer of review of arbitral decisions.Proponents of the AM argue that it will bolster the rule of international investment law.Yet, the benefits are far from certain.We ask: What are the pros and cons associated with establishing an AM for ISDS?And, are there possible solutions to such risks?In answering these questions, this article provides an overview of an AM's costs and benefits.It also identifies key priorities for the reform efforts currently in progress at the 'United Nations Commission on International Trade Law (UNCITRAL)'.
For many legal scholars, the main problem with the current system of international investment law enforcement is that it is based on a model of international commercial arbitration. 2It relies on ad hoc tribunals of party-appointed arbitrators to resolve one-off disputes, even though the disputes may involve public law and policy. 3The tribunals interpret vague treaty rules such as provisions demanding 'fair and equitable treatment' and prohibitions against 'measures tantamount to expropriation'.At the same time, the system currently lacks an appeal process other than a narrow annulment proceeding that has been routinely criticized. 4As a result, conflicting decisions, sometimes involving the same facts, can raise rule-of-law, consistency, and coherence concerns. 5These issues have important policy implications.Given the potential for large damage awards, the threat of litigation under the uncertainty created by ISDS, it is contended, can chill regulation. 6o resolve many of these core problems with international investment law, an AM has been prescribed as an important (or even necessary) innovation of ISDS. 7By adding a layer of review, many believe that an AM will reduce these problems and enhance the structural limitation (i.e., an ad hoc system with party appointment arbitrators without review) resulting from the arbitration process. 8However, the benefits of an AM are rarely assessed in light of its multiple risks.One contribution of this Article is to describe the costs of an AM in relation to the widely anticipated benefits of such a mechanism. 9oreover, the lack of a framework for assessing the range of institutional alternatives considering their practical issues can be problematic in this legal context. 10Hence, the Article provides a way to think about the relative trade-offs of an AM.Our claim is basic but important: all institutional alternatives are imperfect, but imperfect in different ways.Without considering the tradeoffs and the imperfections, a reform process might be of little help and can make for bad public policy.In this sense, we identify and cluster the anticipated costs of an AM in three distinct categories: procedural, conflict resolution, and substantive.In each area, there are good reasons to be concerned about the downsides associated with an AM.
We do not argue that an AM is necessarily a bad idea.Instead, we propose features that might help mitigate the costs of an AM.We believe that such a contribution could help narrow down the options within the ongoing UNCITRAL process of ISDS reform with a better sense of the trade-offs.We think our intervention is timely as Working Group III of such a UN body will continue deliberations in the coming years on the basis of draft provisions on 'an appellate mechanism, including issues regarding the enforcement of decisions that would be rendered through a standing mechanism '. 11 To be sure, since the proposed AM is part of a much broader reform process, the AM's effects are contingent on variables that have not yet been determined.For example, the scope of the AM's mandate remains to be defined, as does the selection process for its members.These, along with a wide variety of other moving parts, will affect any cost-benefit analysis of the eventual AM.As a result, this Article is not intended to be an exhaustive legal or policy discussion of the risks associated with establishing any particular AM.Instead, this Article provides a broad survey, identifying the most salient plausible effectsand how to address them.For a reader seeking a deeper dive on particular features of the AM, other resources exist. 12he Article is organized as follows: Section 2 highlights the perceived benefits of an AM; Section 3 discusses the potential problems; Section 4 considers design features that could mitigate such problems; Section 5 narrows down the issues and identifies key questions for the UNCITRAL process; and Section 6 concludes.

Perceived Benefits of an Appeals Mechanism
The three conventional goals of investment protection reflected in ISDSfairness, efficiency, and peaceare linked both analytically and consequentially to a broader principle: the rule of law. 13n the words of ICJ Judge James Crawford, one of the main roles 'of international law is to reinforce, and on occasions to institute, the rule of law internally'. 14The rule of law provides the guiding principle for international investment law.The concept resonates with traditional justifications for investment law, such as the obligation not to 'deny justice', contemporary arbitral jurisprudence regarding the 'minimum standard of treatment' and 'fair and equitable treatment', and the preamble and other provisions of investment treaties. 15hile we believe that, in general, an AM could enhance the rule of law, it is often not clear how exactly an appeals process might do so.Therefore, before addressing the AM's costs, this section identifies its perceived benefits.We divide our analysis into two broad categories.Tangible benefits of the AM are corrections to problems in ISDS decision-making. 16Intangible benefits are corrections to the problems of perceived bias and a lack of legitimacy.There is some overlap in these concepts, but the core difference is between the legal decision itself and the impact (or perception) of the system by a larger legal field, governments, and investors.We take this approach because the role of dispute settlement systems is not only to provide tangible benefits such as security and predictability so that individuals and firms can plan their pursuits.Ultimately, for the rule of law to become effective, it must be legitimized as part of a broader institutional culture.Hence, we believe that a more legitimate dispute settlement process can enable intangible benefits that may improve legal practice. 17

Help Clarify Excessively Broad (or Overly Narrow) Interpretations
There is a widespread perception that 'incorrect' interpretations of investment treaties are common in ISDS decisions. 18Recent findings show that tribunals' rulings can often conflict with the treaty text by taking an excessively broad (or overly narrow) view of the rules or exceptions. 19ne classic example is the Metalclad tribunal's rather strenuous interpretation of the agreement's phrase 'measure tantamount to … expropriation'. 20As explained by a Canadian court reviewing the award, the definition adopted by the tribunal was too broad and risked sanctioning 'legitimate rezoning of property by a municipality or other zoning authority'. 21In other words, the interpretation imposed a burdensome reading on the State that is especially vulnerable to this danger.At the same time, the interpretation arguably conflicted with the treaty's text. 22he AM's proponents argue that an appeals system may help correct faulty interpretations by adding a layer of review.In this way, the AM serves the purpose of a traditional appellate body: to confirm, modify, or overturn the ruling of the lower body.An AM in ISDS may also help clarify the proper methods of interpretation to better reflect the extent of the treaty parties' obligations.

Correct Rulings that Lack a Textual Basis
Investment treaties, along with the Vienna Convention on the Law of Treaties (Vienna Convention), afford tribunal's interpretative authority to analyze the text, especially in the presence of ambiguity or in situations not properly addressed by the instrument.However, in situations where there is a lack of textual basis (or legal authority) for rulings, arbitrators may overreach, essentially deciding issues beyond their jurisdiction.Accusations of overreach are prevalent in other areas of international economic law, most notably trade dispute at the 'World Trade Organization (WTO)', where such claims form a basis for US opposition to appellate body decisions. 23ISDS has not been immune to similar worries. 24The decision of the tribunal to rely on a Most-Favoured-Nation Treatment clause with specific language to expand the claimant's international dispute settlement options in Impregilo could be an example of this type of overreach. 25he AM's benefit would be avoiding similar situations by reversing decisions that are not grounded firmly within the treaty limits, providing an effective check on tribunals that overstep their mandates.For example, an AM may review a finding that is illogical, that failed to adopt a sound interpretative methodology, or that substantially deviates from the treaty text and reverse that finding.That option is not available under the existing review mechanisms that tend to 'address the integrity and fairness of the process rather than the consistency, coherence, or correctness of the outcomes'. 26

Discourage Impractical Interpretations
Ambiguous, vague treaty obligations create an additional problem that an AM may help solve.Arbitral decisions assigning concrete meaning to vague treaty provisions may lead to rulings that are politically infeasible and all but impossible to implement.In Metalclad, the tribunal found that the Minimum Standard of Treatment imposed a transparency obligation.Interpreted that way, the obligation required that States inform investors of all potential legal and bureaucratic requirements.Such a standard is unrealistically burdensome, particularly given the difficulties identifying when an investor ought to be notified. 27n response, beyond simply reversing such decisions outright, an AM may help discourage impractical decisions that do not sufficiently contextualize a legal dispute.The AM may propose alternative analyses that more accurately reflect the challenges and sensitivities of policymaking.

Avoid Imprecise Applications of Law
Tribunals are also expected to interpret investment obligations within the larger context of international law.Yet, in practice, some tribunals have shown their limited understanding of that context, leading to errors in the interpretation or application of treaty provisions.For example, in Railroad Development Corp., the tribunal conceded that its finding regarding the interpretation of the Minimum Standard of Treatment was based on an outdated interpretation of NAFTA despite differences between the NAFTA and the treaty text in question (CAFTA-DR). 28n AM, by casting a shadow of potential review, may invite tribunals to more precisely identify key aspects in legal agreements and systematically analyze such features within a larger context of international law when making decisions.Failing more careful interpretation by arbitrators, the AM may itself appreciate that context when issuing decisions.

Intangible Benefits
The above benefits are all about how an AM may address concrete, tangible problems with current ISDS decisions themselves.However, an AM may also help address less tangible, but equally contentious concerns.One is the perceived lack of independence and impartiality of arbitrators.The other is doubt over the system's external and relational legitimacy.Proponents of establishing an AM argue that an additional level of review can ameliorate these criticisms and rehabilitate public and institutional perceptions of international investment law.We believe it is important to take this potential benefits seriously.From practical perspective, the ultimate challenge for the rule of law is its implementation, which is mediated by social institutions and legal culture.Thus, the AM can, potentially, bring added benefits to the practice already embedded in ISDS and help to legitimize this contested legal field.

Reduce Bias in Appointment System
Some accusations of bias and partiality stem mostly from the current system for appointing arbitrators, especially party appointments. 29For example, there are few limits against repeat appointments, or of the prejudging of legal issues.At the same time, arbitrators are perceived to have a vested interest in the ad hoc model because they are remunerated for their services on an hourly basis. 30Arbitrators' reliance on reappointments may encourage them to cater to the parties' interests, thereby producing biased interpretations that favor a particular judicial philosophy, even if decision-making is not correct.Four types of biases or effects are especially relevant in ISDS: 31 • Selection effect: Because parties nominate the arbitrators, they can choose arbitrators with the maximum predisposition toward their case.• Affiliation effect: Party-appointed arbitrators may be affected by implicit bias, tending to side with the nominating party even if they attempt to maintain neutrality and independence.Proponents of an AM argue that a standing body would redress bias in various forms introduced by the ad hoc nature of arbitral tribunals and the method of appointment.For example, by having a standing body of members selected for a periodno matter the leanings of the decisionsselection, affiliation, and compensation effects may be reduced.And, by having a more diverse pool of AM members, the epistemic/cultural effect could be mitigated.

Narrow the Legitimacy Gap
Several traits of the current process can undermine the legitimacy of ISDS as a general practice of international law.First, arbitrators understand there are few grounds upon which parties may challenge arbitral awards.Thus, there is little incentive for arbitrators to follow any cohesive doctrine from case to case.In the absence of formal practice of stare decisis, individual decisions may appear inconsistent with the total body of ISDS case law or the body of general international law more broadly.Second, as previously mentioned, arbitrators are appointed on an ad hoc basis and are allowed to represent clients in other arbitrations (unless the BIT limits this practice).As a result, arbitrators face incentives to decide cases in a manner favoring the party that appointed them or the World Trade Review position of their clients in other cases.This contributes to perceptions of partiality in outcomes, which has drawn criticism from governments, firms, and advocacy groups. 32nstituting an AM, with tenured members insulated from arbitrators and other actors (e.g., law firms) interests, and free from incentives related to possible reappointments, may help remedy these issues and boost the external and relational legitimacy of ISDS.More generally, a permanent AM is also expected to rehabilitate legitimacy by adding 'consistency to investment law, as ad hoc tribunals have a natural tendency to diverge more than … standing tribunals with an in-built element of tradition and continuous collegiality'. 33That consistency may reduce perceptions of partiality by establishing a more predictable set of interpretations that would appear less dependent on the appointment process or the identity of the parties or their counsel 34 (see Table 1).

The 'Costs' of an Appeals Mechanism
Having outlined the main goals behind creating an AM, we turn to the potential problems.This section covers issues related to: (1) procedure; (2) conflict resolution; and (3) the substance of rulings.Our conceptual distinction follows the view that 'substance' and 'procedure' can be separated clearly. 35To this, we add a third category, which refers to the impact, or 'downstream effects' of the ISDS process.Ultimately, the system's principal goal is the peaceful and efficient resolution of transnational conflicts.Yet, creating an AM may, as we explain, weaken enforcement and dampen compliance with decisions, thus impacting on this core goal.
As noted at the outset, the problems or issues with the AM are, at present, theoretical.It is hard to forecast the effects of the AM in isolation, in part because there are many different changes to the ISDS system proposed under the current reform process.However, we can draw from recent history and other international legal or dispute settlement settings to identify the possible difficulties below.
Beyond our more technical discussion, the argument here is simple: in critiquing (or advocating for) a particular institutional choice, one should not focus on the potential benefits only.We should instead assess institutions from multiple perspectives, including one that avoids ideal characterizations and takes into account institutional pathologies and tradeoffs.We believe   Greater consistency in rulings may also have the benefit of promotion common understandings of the lawthat is, it may help clarify frequently disputed concepts such as Fair and Equitable Treatment.that much international legal scholarship related to the AM fails to analyze institutional alternatives, especially legal processes, accounting for their real-world complexity. 36

Increase in Duration
Adding an appeals option inevitably extends the time to final dispute resolution, prolonging the period of uncertainty during which parties await a decision.The average length of arbitration proceedings under current rules is 40 months, or 3.3 years until a decision.If a party files an annulment application, cases may be tied up an additional 1.8 yearsa 48% increase in total dispute duration. 37otential delays will depend on the AM's timeframes, functions, and standard of review, among other features.For instance, an AM that reviews errors of fact would likely cause longer delays than if it reviewed only errors of law.That is because reviewing facts often requires a reevaluation of the evidence, not just a consideration of the legal argument.Some questions may require a new hearing (for example, the assessment of the functioning of a legal system) and, as a result, extend the proceedings.At the same time, unless the bases for appeal and annulment (or set-aside) are combined, an AM runs the risk of adding a third instance of ISDS proceedings, further lengthening the amount of time before a final decision.

Increase in Cost
Introducing an AM will likely raise the costs associated with ISDS, especially legal fees.The costs of investment arbitrations are composed of: (1) tribunal fees, such as arbitrator fees and institution maintenance; and (2) legal fees, such as counsel and expert costs.Legal fees account for approximately 90% of the total costs.The 'International Centre for Settlement of Investment Disputes (ICSID)' reports average annual expenditures of $127,000 per case.At an average length of 3.3 years, it totals $420,000 in expenditures. 38Annulment proceedings add another $230,000 to the total cost, on average.An appeals process, especially if it exists in addition to the annulment process, will increase these costs much further.
Another cause of higher costs is the likelihood that an AM may result in a greater number of challenges to arbitral decisionsat least in the short run.Even under the currently narrow grounds, 51% of the ICSID Convention awards resulted in annulment applications. 39This number could increase markedly.For example, at the WTO, panel reports are appealed 71% of the time. 40There are important differences between WTO disputes and ISDS and the use of the Calculations are based on financial information from ICSID Annual Reports.We divide the total annual 'expenses related to arbitration/conciliation proceedings' by the total number of disputes active in a given year (at any stage of the legal process).This method of calculation produces a simple average that understates the expenditures on some disputes while overstating others.Cf.WTO's practice might be far-fetched. 41However, it is possible that parties will seek a review more often under an ISDS system where the AM has a broad scope.
Quite apart from the costs of litigation, additional costs may also arise from maintaining the AM's bureaucracy.In 2020, the existing system's annual administrative expenses hit all-time highs of $14.3 million, representing a 12% increase over 2019.Depending on the cost structure, some States may end up facing even greater expenses establishing and maintaining a permanent AM facility (i.e., an office) and staff.This could include indirect costs such as AM members' pensions and other benefits.

Increase in Bias
Increases in duration and costs have an additional implication: privileging wealthier litigants over poorer ones.Wealthy, well-resourced States and deep-pocketed investors could reap the benefits of the lengthier, more expensive process.This may disadvantage smaller investors and developing States who face tighter resource constraints.
There is already a wide disparity under current rules.Most cases target poorer States, which tend to have more limited legal capacity.In fact, 75% of all cases involve respondents whose economies are half the size (or less) of the investor's home country.Unsurprisingly, poorer respondents are also 30% more likely to settle, perhaps because their limited resources prevent them from seeing the legal process through to its end.By contrast, richer States, with greater legal capacity, are targeted less frequentlyand they settle less often. 42ere, too, the history of trade litigation may be instructive.Only 52 of the WTO's 164 members have ever filed a complaintand almost all of those are upper-or middle-income countries.Only two African countries have filed complaints and studies show that capacity shortages fundamentally shape access to the legal system that, in theory, should treat members equally. 43It is no coincidence that the GATT/WTO system is sometimes called a 'country club' in which the largest marketsincluding those who can better defend themselves in trade disputesare seen to enjoy larger benefits from membership. 44An appeals stage added to ISDS could aggravate these issues, especially if no efforts are made to mitigate imbalances resulting from the cost structure.

Increase in Litigiousness
The AM may also work against one of international dispute settlement's main goalsnamely, reducing the incidence of conflict.As mentioned, an appeals process may result in more litigation because decisions might become more impactful, especially for States.This may lead to an issue that scholars have identified, and long struggled to measure: litigiousness.
'Litigiousness' refers to the strategic use of more robust legal systems to extort settlements from the threat of legal action or the use of legal proceedings to create or manipulate legal precedent.For litigious firms, 'such behavior can be beneficial because the filing of many cases creates a strong public signal of resolve, [also] litigious firms may become better at understanding the odds of success, at selecting cases that best meet their goals, at dissecting successful arguments and ultimately at generating useful precedent'. 45We see this in other areas of international economic law.For example, research on the WTO shows that complainants design legal arguments with the express purpose of establishing favorable precedent. 46They also pay careful attention to previous rulings when deciding whether to settle. 47In both situations, the idea is that litigation is strategic, that past rulings matter, and that parties engage in disputes partly to shape the law.Incentives to do so, especially when an AM may more effectively establish precedent, will generate rather than dampenincentives to litigate.

Decrease in Settlements
Establishing an AM may negate the promptness and finality of resolution.Awaiting an appeals decision would prolong legal proceedings and potentially delay the settlement process.Currently, settlements tend to occur within the first two years of a dispute.However, given that governments may be required to exhaust legal options, by law or policy, before offering settlement terms, an AM may effectively delay the resolution of the dispute.
Effects on settlements may have consequences for foreign investments.If firms lack any assurance of recovery, in whole or in part, they may be less willing to invest in the future. 48Moreover, since an AM may dislocate the ISDS system from the general system of international arbitrationand from enforcement of awards more generallyan AM can also encourage less transparent investment planning decisions (with concomitant effects on taxation).
Finally, adopting an AM to provide a substantive review will open debates on the scope of the AM's power to interpret provisions, including under the ICSID Convention.
the development of a strong norm of de facto precedent, as seen in many areas of international law.Developing and leaning on precedent comes with risks, however.An AM that develops a strong norm of precedent may increase dissatisfaction with a decision and reduce the willingness of States to comply.This effect is seen at the WTO Appellate Body, where heavy reliance on precedent is associated with a reduction in the rate of timely compliance and with longer process delays. 51The United States, for example, complied with nearly 70% of adverse WTO rulings prior to the decision in US -Stainless Steel (Mexico) to require a panel to follow a prior Appellate Body interpretation unless the panel has 'cogent reasons' for departing from that inter-Since that decision, US compliance rates have fallen to under 40%. 52he analogy to the WTO is imperfect because the GATT/WTO is one set of texts rather than the network of BITs implicated in ISDS.However, any efforts to develop a coherent interpretation of common terms, such as Fair and Equitable Treatment, may result in a body of case law that ossifies interpretations of investment treaties in ways with which the Parties are dissatisfied.Of course, compliance may also vary by the clarity of rulings.States may delay compliance if the AM issues ambiguous and/or imprecise decisionsin part, because there is more at stake at the appeals stage.

Decrease in Enforcement
Enforcement of current ISDS awards depends, mainly, on mechanisms of the ICSID Convention and the 'New York Convention on the Recognition and Enforcement of Foreign Arbitral Awards (New York Convention)' as well as domestic set-aside proceedings. 53We identify three issues in which an AM can complicate the enforcement process (although many more may emerge given the strategies of able lawyers).
First, to be enforceable within the current system, AM decisions will need to be considered final (i.e., the 'award').Because of this, either the ICSID Convention will need to be amended or countries need to adopt an inter se modification of the Convention accounting for the AM in defining an enforceable award.In that latter scenario, the AM still could decrease enforcement by dislocating the process from the general systems of enforcement of arbitration awards.This may weaken, rather than strengthen, enforcement because the new regime would apply only to the contracting Parties to the new system.Therefore, unlike the ICSID or the New York Conventions, which have more than 150 contracting Parties each, a new enforcement regime would suffer from weakness until a plurality of States adopt it.
Second, and related, some ICSID members will inevitably not join the proposed AM.As a result, three (or more) different ISDS enforcement mechanisms may co-exist.This fragmentation can cause confusionand, as per above, lead to additional litigation.For example, if an ICSID award is appealed, and a third State who is not a member to the new AM is asked to enforce the award, it may not be bound to the ICSID Convention provisions as it relates to that award. 54his discrepancy in enforcement mechanisms could be avoided if ICSID members passed an amendment under Article 66, but that would be unprecedented.
Third, adopting an AM could also lead to uncertainty over the timing of a potential (re)espousal of a claim by the home State of the investor.Under Article 27 of the ICSID Convention, States can only provide diplomatic protection if the other contracting State fails to pay an award pursuant to Article 53.The question of when an award is final may also affect when a State could be considered recalcitrant.Questions such as this are relevant to the general conflict resolution goals of ISDS. 55

Loss of Precision
One of the AM's benefits is supposed to be greater coherence and consistency in the body of law.Yet, given the sheer number of BITs, emphasizing consistency across treaties could lead to the loss of treaty specificity.For example, AM members may be tempted to apply reasoning from a previous dispute under a different treaty.Applying de facto precedent in this way may be inappropriate given potential differences in text, intent, context, relevant facts, and arguments.Even very similar treaties may demand different interpretations depending on the contracting Parties' subsequent practice and other elements relevant under the Vienna Convention.Therefore, any generalizing by the AM runs the risk of conflicting with specific treaty provisions.
Systemically, this could have a second-level effect.Recognizing that the current system is far from ideal, a desire for legal consistency could dampen innovation.States could be disincentivized from revising treaty texts if States perceive that investment treaties are misinterpreted consistently. 56

Ossification of Precedent
Related, prioritizing legal consistency may ossify decisions that States see as incorrect or problematic.Compared to other areas of international dispute settlement, there seems to be currently a relatively weak(er) norm of precedent in ISDS.However, an AM with a strong interest in legal coherence may strengthen the norm of precedent in investment disputes.
The WTO Appellate Body illustrates the danger of ossification, even where there is no formal stare decisis.Appellate Body rulings almost always adhere to precedent on at least one of the disputed issues.More controversially, the Appellate Body sometimes goes further, extending past rulings in no less than 10% of its decisionsi.e., applying a prior to reading in a way that deepens States' obligations. 57As mentioned, the AB's reliance on a strong norm of precedent fueled accusations of overreach from States dissatisfied with past interpretations and eventually contributed to the gridlock in WTO dispute settlement. 58The creation of an AM for ISDS runs a similar risk because legal bodies often have a vested interest in promoting coherence and consistency within the regime.
In fact, this danger of ossification is even greater in ISDS because only investors bring claims (at least for now).Decisions that establish lasting precedent may reduce the policy leeway States enjoy under the law.As a result, States may worry that previous decisions, when applied to future disputes, may fundamentally alter, and harden their obligations under BITs.
Overall, we do not consider the further legalization or 'juridification' of investment disputes that may result from a more robust legal setting as a cost per se.However, a less supportive view of international adjudication may see such juridification as a general problem.World Trade Review 573

Conflicting Outcomes
As mentioned, it is improbable that all States would adopt an AM.BITs generally afford investors the choice between different arbitration rules and systems.If an AM is added to ICSID Convention awards, investors could avoid the AM by opting for a different arbitration process.Hence, there is a possibility of related-party disputes ending up on different arbitration 'tracks.'For these reasons, an AM creates the potential for conflicting, contradictory outcomes in cases involving related parties.The existing system is not immune to this problem. 59However, the creation of an AM may exacerbate it (see Table 2).

Improving the Design of the Appeals Mechanism
The AM's proponents are aware of many if not most of the above concerns, and they offer various design suggestions to address these problems.Based on our analysis of these suggestions, we now offer ideas on how to mitigate the identified concerns.

Establish Timelines
The imposition of short, but realistic timelines for each stage of the appeal proceeding could reduce duration and, by implication, higher costs.The time frames must reflect the complexity of ISDS disputes and the resources available to an AM.However, measures to reduce the length and number of written submissions, to dismiss issues summarily, or to incentivize the AM to decide in a timely fashion (i.e., reduction of fees for delays), could effectively lower the length and cost of proceedings.For example, while the WTO's Appellate Body receives heavy criticism for often missing its 90-day deadline, appeals lengthen the average trade dispute by only about 22%.By contrast, as mentioned above, annulment proceedings extend ICSID Convention investment arbitrations by almost 50%.Therefore, while deadlines at the WTO are frequently missed, having some default rules may help constrain duration and costs. 60

Provide Technical Assistance
Even with efforts to expedite proceedings, introducing an AM may impose significant financial burdens on some litigants, especially smaller States.This can exacerbate systemic biases wherein smaller States are less able to defend their interests under the system.Establishing a funding mechanism to help alleviate ISDS burdens is crucial and could play a key role in dispute prevention (or, given a dispute, additional resources may help poor litigants assert their rights).This could also include mechanisms to increase the legal capacity to deal with investment disputes, similar to the advisory services provided at the WTO.The establishment of an advisory body is already on the agenda and could have substantive benefits beyond alleviating procedural concerns.

Clarify Functions of Body
One of the most direct ways to address time and cost inefficiencies is establishing clear functions of the AM.It may help to limit appeals to 'errors in the application or interpretation of applicable law' and 'manifest errors in the appreciation of facts' for final decisions only.This would reduce slippage in the AM's mandate and, by extension, forestall concerns about legal overreach.Of course, to prevent uncertainty it is crucial that the ambiguities in these terms are clearly defined.
Moreover, enabling an AM to confirm, reverse, modify, or annul the decisions of the first-tier body may significantly reduce the duration of proceedings because it will prevent a remand to the arbitration tribunal.And, most obviously, structuring an AM so as to be final (that is, not subject to annulment) could drastically reduce time-cost inefficiencies and other problems as we explain below.

Built-in Conflict Management
Conflict resolution and management techniques could increase the likelihood of settlements prior to, or during the adjudicatory process.For example, parties may benefit from subjecting themselves to an additional layer of mediation prior to initiating an appeal.Even if the parties are unable to reach a negotiated settlement on all claims at issue, mediation may help to narrow or streamline the dispute.The ample literature on conflict resolution system design could be useful to inform the options as well as the tradeoffs of the different decisions. 61

Curtail Abuse of Process
Rules like cost-shifting or establishing a mechanism to impose sanctions against parties or, in very limited situations, against their counsels that bring unmeritorious or frivolous claims may lead to more efficient proceedings, depending on the standards for considering abusive behavior.Some rules like the ICSID Arbitration Rules already contain a textual basis for ICSID tribunals to dismiss spurious claims or to dispose the cases more efficiently. 62Overall, improving the timecost efficiency of proceedings requires clear and transparent guidance about the principles upon which an AM will shift costs or sanction litigants or their counsel.

60
On the issue of timeframes, see, J. Hillman (2018) 'Three Approaches to Fixing the World Trade Organization's Appellate Body: The Good, the Bad and the Ugly?', Institute of International Economic Law 8-9.

Clarify Standard and Scope of Review
A clear standard and scope of review are crucial to addressing issues of dispute settlement and compliance with awards.Clarity on this issue would promote consistency in the ISDS process, but also prevent parties from appealing decisions to delay an undesirable outcome.
For example, UNCITRAL's Working Group III has considered whether an AM should provide for a review of issues de novo or whether it should accord deference to the findings of the body of first instance.The most recent draft suggests consensus that should an AM be implemented, instances of appeal should be limited to 'error in the application or interpretation of the law' and 'manifest error in the assessment of the facts', thereby affording some deference to the findings of first instance adjudicators. 63This is, we believe, a good start.Consistent with the interpretation of other arbitration provisions, a 'manifest' error should be patently obvious and not require the AM to conduct a very complex analysis to conclude that such an error exists. 64This, in turn, preserves the legitimacy of the first-tier body fact-finding process, and it takes into account efficient management of costs and time.If support continues to grow for permitting an AM to review for errors of fact, differentiating what constitutes an error of law from an error of fact can help minimize any complications that might arise when determining whether issues are subject to appellate review.

Address Redundancy
To avoid the danger that an AM becomes a third stage of adjudication, negotiators could merge the legal bases of appeal with the bases of annulment.Generally, an appeals system is broader than annulment, and may encompass the (current) narrow scope of ICSID's annulment.This would reduce complexity in the proceedings, thereby helping address the aforementioned potential loss of finality.

Adapt the New York Convention
While unlikely, it may be useful to revise the New York Convention.The goals would be to increase the enforcement of ISDS awards, to establish a presumption of validity of AM decisions, and discourage the intervention of courts before the AM has issued a final decision.If the Convention is not reformed, we believe, parties to proceedings may try to use States' national courts to enforce or vacate arbitral awards pending an appellate proceeding. 65

Substantive Concerns
Among the AM's core purposes are reviewing the integrity of the process leading to the decision as well as the substantive correctness of the decision. 66However, some inconsistencies and inaccuracies would be inevitable, especially at the early stages of the AM's operations.As such, the 63 UNCITRAL, Draft Note by the Secretariat, Possible Reform of Investor-State Dispute Settlement (ISDS): Appellate Mechanism, para.17 (accessed 16 December 2022).following mechanisms of control may address the concerns of States, investors, and nondisputing Parties.Some of these features are already common practice in some BITs.

Filters Prior to Appeal
Treaty provisions can effectively limit the issues that parties bring to appeal.Provisions may limit the advancement of aggressive interpretations to be assessed by the AM, reducing the potential for errors.For example, a sort of 'political filter' could require that the investor's home State agrees with the investors' interpretation prior to the submission of the appealin general or on some treaty provisions.The veto of expropriation claims resulting from tax measures pioneered in the NAFTA can serve as a model. 67This provision supplies an initial screening process by the home State to determine when taxation constitutes a form of expropriation.Of course, this and other control mechanisms can be more effective if treaty Parties include annexes with clear substantive guidelines to adjudicators to interpret provisions in accordance with such directives.

Consultation Processes with Treaty Parties
The ability of the AM (or, perhaps, any of its members) to refer questions of interpretation or application of the treaty to the Parties could serve to control the authority of the body.For example, the USMCA established a referral procedure (from judicial or administrative proceedings) by which a domestic body can request clarity in matters of interpretation or application of the treaty with a commission formed of representatives of Parties to the treaty. 68ther similar mechanismsalbeit less formalinclude requiring the AM to circulate a draft copy of the decision and allowing for written comments to review the proposed decision.This practice has been included in US BITs and could be useful as a way to roll back incorrect interpretations or that result in unintended policy outcomes. 69

Binding Treaty Interpretations
Provisions in the AM's design could also establish an obligation to consider (or to apply) agreed upon interpretations.This agreement could happen in different ways, including: (a) by a joint statement of a body involving the treaty Parties with the authority to dispatch such function; (b) by a separate articulation of a position by each of the treaty Parties in prior submissions; (c) by a separate articulation of a position by all treaty Parties in prior statements; or (d) by any other method considered by international law to reflect subsequent agreements.Any requirement to follow agreed upon interpretations could then be complemented by a mandate that the body explains its reasons to deviate from following an interpretation when a State argues for the consideration of agreed treaty interpretations.

Include Sunset Provisions for Body
Provisions that create a temporary mandate or that routinely review and revise aspects of the functioning of an international body could be useful to control the interpretative authority (and potential drift) of the AM.These mechanisms are often referred to as 'sunset provisions' and are already included in treaty practice, albeit discreetly. 70Such clauses anticipate changes in political constituencies and permit fine-tuning of the system as well as limiting the effects of incorrect or unworkable decisions.

Establish Qualifications for Body Members
Specific qualifications can help ensure that members of the AM are not perceived as biased and are better equipped to apply and interpret specific investment agreements, for example: (a) requiring particular background or expertise of all members, the chair, or some of the members of the AM that will hear an appeal; 71 (b) introducing random selection of seating members from a larger pool; (c) mandating that certain appeals (e.g., appeals under US BITs/FTAs) can only be heard by body members from a particular pool or by members with required qualifications; or (d) instituting a right to veto members based on specifically defined situations (e.g., conflict of issue).

Revise Support Architecture of Body
The institution assisting the body provides a de facto control mechanism for international tribunals.The Secretariat's involvement in the process could foster more coherence between decisions and help navigate politically sensitive issues.However, there have been accusations of undue influence and bias leveled against some Secretariats perceived to be too influential. 72While strong institutional involvement by a Secretariat could be beneficial, it could backfire as at the WTO.Therefore, calibrating the exact level of desired influence of the Secretariat is very important (see Table 3).

Insights for the Uncitral Process
We have now identified the AM's possible benefits, its possible costs, and some of the design decisions that may help address those costs.Clearly, there are a large number of factors to consider, and we cannot claim to cover every issue.Our effort is, rather, to first identify the list of relevant concerns.At the same time, our discussion so far raises important policy questions: Which problems are more important than others?And which design solutions promise the largest benefit?In reply, we offer an approach to identify the most central problems with the AM, which can help narrow down the options within the UNCITRAL process.
To identify the key priorities, we start by recognizing that many of the issues we enumerate are linked closely together.This applies to items within categories as well as across them.In terms of costs, for example, we note above that a longer process will inevitably be more expensive.At the same time, a more expensive process may increase bias (by tipping access/success to the system toward wealthier litigants).In the same way, some of the design solutions that have been offered in anticipation of these costs can help address more than one issue.For example, providing litigants with technical assistance may help address those issues of cost as well as bias.
With these connections in mind, we map out the web of associations between issues.We focus on where we see opportunities for design solutions to address the key costs we identify.Hence, the tradeoffs and balances to which we refer.Some issues may play a more central role in the 'network' of the AM's costs.To get a sense of relative importance (or centrality), we map out the connections between these issues below (see Figure 2).
Despite our simple approach, our central claim remains valid: an AM, like all institutional alternatives, is highly imperfect because of the dynamics of participation, and thus criticism of one institutional alternative without comparatively assessing the imperfections of real-life alternatives is of little help.Therefore, adopting an institution like the AM without understanding the potential impacts can make for bad public policy too.That is why we provide a roadmap to engage in that debate with a better sense of where to focus the negotiating efforts.

Approaching the Debate within Uncitral
Section 4 laid out a wide variety of AM design options.Some of those options are (in theory) very straightforward solutions, such as imposing strict timelines on the appeals process.Others are far more complicated, and perhaps infeasible, such as revising the New York Convention.
Feasibility aside, some of the design options are also more (or less) impactful than others.For example, adapting the New York Convention would be incredibly difficult and, perhaps, not worth the effort or outside of the current mandate of UNCITRAL.We made a careful assessment of how many problems each design solution may address.Table 4 reports those relationships.It offers a simple indication of whether a particular design option may help address one of the several costs or risks we identify.This coding offers a way to identify core prioritiesi.e., the design features that are more or less important to the overall cost of the AM.
The most central design issue, perhaps unsurprisingly, is clarifying the functions of the AM.As stated above, enabling an AM to confirm, reverse, modify, or annul the decisions of the first-tier body may significantly reduce the duration of proceedings because it will prevent a remand to the arbitration tribunal.Clarifying the AM's functions may also help prevent increases in litigiousness, increases in bias, and the loss of precisionall of which are important concerns for those skeptical about an AM.For these reasons, the conversation around adding an AM to ISDS must naturally focus on delineating the boundaries of the AM's powers and purpose.
Perhaps less obvious is the importance, in our view, of establishing clear qualifications for members as well as demarcating the relationship between the AM and the Secretariat.These Curtail abuse/apply sanctions

Conflict Resolution
Clarify standard and scope Binding treaty interpretations Jeff Kucik and Sergio Puig two issues may not seem as high-profile, or fundamental to the process, but we are clear that they deserve careful attention in reform discussions moving forward.
To depict these priorities in another way, Figure 1 is a simple bar chart based on that coding, identifying the features that may deserve the most time and attention.The chart shows that clear functions relate to the largest number of problems we identify.Conversely, revising the New York Convention is related to the smallest number of the problems we identify.Given the costs associated with adapting it, Table 4 suggests that, on balance, negotiating resources may be directed better toward more realistic, achievable goals.
Of course, we recognize that a simple count of these relationships is not perfect.Issues like providing technical assistance score low on our list, but they are nevertheless important.Assisting lower-capacity or resource-strapped governments to better defend themselves is a worthy goal.Our effort here is not to diminish that issue.Instead, it is an effort to identify, on balance, the negotiating priorities that speak to the largest number of AM costs and risks.
Figure 2 recasts these relationships in a network.The nodes are the various costs we list, and the edges between them are weighted by the number of design features that speak to both issues.For example, three design issues relate to both increases in duration and bias.However, nine of them relate to both duration and costs; hence, the thicker edge between duration and costs in Figure 2.
Our effort is to identify key priorities.However, none of these design decisions guarantees a well-functioning AM.Clarifying the AM's power sounds like an obvious first step, but other legal settings show that it is complicated in practice.For example, WTO disputes frequently include debate over the Dispute Settlement Body's jurisdiction and authority.But that is all the more reason to focus resources in this area.

Enhancing ISDS and Institutional Alternatives
A final clarification is in order.Creating an AM is not inevitable.States disinclined to accept the potential costs imposed by a new AM may find ways to enhance ISDSfor example, by improving the system of appointment and accountability of arbitrators.Or, as suggested previously by one of us, States might be more interested in increasing the complementarity of international investment dispute settlement with domestic judicial and administrative institutions. 73Therefore, reforms could be directed to improve decision-making and limit bias (and, indirectly, improve legitimacy) by adopting some concrete actions in a trajectory of improvements.That is of course an entirely different issue and would require a separate paper to do it justice.We just note that other options to improve ISDS remain available, and the cost/benefit of that should also be explored.
What is more relevant here is to understand that institutional choice is always contextual and thus difficult.Therefore, institutional analysis informed by alternatives is essential because all institutional options are imperfect and subject to trade-offs.
We do not claim to have found reasons to reject an AM.Adopting a legal realist and interdisciplinary perspective, 74 our primary goal is to present and apply an analytic framework that helps to clarify the trade-offs of an AM.The option chosen (if any) will be a function of the preference of States that are in different situations and the complex negotiations.The tailoring of institutional choice can be met by what Roberts describes as ongoing pluralism and flexibility in the architecture of institutional mechanisms. 75Yet, looking forward, our analysis reveals that the international investment regime should not simply rely on an AM to improve ISDS unless such a system 73 Shaffer and Puig, 'Imperfect Alternatives: Institutional Choice and the Reform of Investment Law', supra n. 2.  provides for flexibility, such as through opt-out and opt-in provisions.Maintaining flexibility will be keya flexibility that permits States facing different challenges to select in light of the trade-offs.
An AM might be in the horizon of ISDS.However, given that all institutional processes are imperfect, the AM may also be distrusted but in different ways than today's ISDS.Yet, to some degree, the entire system of foreign investment protection is built around mistrust.For example, the creation of investment rights implies distrust of national government and regulators.BITs imply a distrust of domestic law.The turn to international arbitration implies distrust of domestic courts.The use of an AM implies a distrust of ISDS tribunals.In turn, the proposal for a multilateral investment court system implies distrust of ISDS even with an AM.This parade of institutional distrust is not surprising, since each institutional alternative is imperfect.For any meaningful policy analysis, however, their imperfections need to be identified, compared, and contrasted.Our framework provides ways to perform such assessment of the AM and different alternatives to ameliorate possible concerns with its implementation.

Conclusion
Although little can be said regarding the actual effects of an AM for ISDS tribunals, the literature and prior experience in other international dispute settlement processes indicate the existence of three types of potential issues or 'costs': procedural, conflict resolution, and substantive.As we noted, procedural costs refer to potential increases in costs and expenses, extended duration, increase in bias, and increase in litigiousness.Conflict resolution costs refer to decreasing instances of settlement, voluntary compliance, and difficulties in enforcement.Substantive costs involve the loss of precision of treaty provisions, ossification of bad decisions, and conflicting outcomes.This framework could be helpful to understand the risks (in relation to its rewards or goals) as more details emerge on the specifics of the proposed AM.
In this Article, we also proposed the inclusion of certain design features to improve the working of the AM and help mitigate the problems identified.Procedural concerns can be addressed through establishing clear timeframes and language, as well as including conflict management tools to reduce cost, time, and bias.Conflict resolution concerns can be addressed through clear language and treaty provisions that help integrate the AM in the existing systems of arbitration.Substantive concerns can be addressed through a series of ex ante and ex post ways that may tighten control of the AM's interpretative authority.
For States and negotiators involved in the UNCITRAL process, we have identified the most central problems with the AM to narrow down the imperfect alternatives.To be sure, we do not claim that an AM should not be pursued.However, adopting a legal realist and interdisciplinary perspectives, we presented and applied an analytic framework that helps to clarify the trade-offs towards an effective AM for ISDS tribunals.

11
United Nations Commission on International Trade Law, 'Possible Reform of Investor-State Dispute Settlement (ISDS) Ensuring Independence and Impartiality on the Part of Arbitrators and Decision Makers in ISDS', UN Doc A/CN 9/WG III/ WP151, 30 August 2018, para.31.12 See Papers on the Academic Forum on ISDS.http://bit.ly/isds-academic-forum(accessed 12 March 2023).13Shaffer and Puig, 'Imperfect Alternatives: Institutional Choice and the Reform of Investment Law', supra n. 2. 14 J. Crawford (2003) 'International Law and the Rule of Law', Adelaide Law Review 24, 3, 8. 15 See, e.g.CETA, at preamble ('[Recognizing] the importance of … the rule of law for the development of international trade and economic cooperation.').For a discussion, see, Shaffer and Puig, 'Imperfect Alternatives: Institutional Choice and the Reform of Investment Law', supra n. 2. 16 The categories we use in this section are identified by and well-developed in A. de Luca et al. (2020) 'Responding to Incorrect Decision-Making in Investor-State Dispute Settlement: Policy Options', Journal of World Investment & Trade 21, 374, 395.

55ICSID
Convention Articles 27(2) and 53.To be sure, similar issues of enforcement remain subject to debate under the current system.See e.g.Letter from the US Department of State Office of the Legal Advisor (1 May 2008) in ICSID Case No. ARB/02/8, http://italaw.com/documents/Siemens-USsubmission.pdf(accessed 12 March 2023) ('Article 54 does not supersede or condition a Contracting State party's obligation under Article 53 in any way.Rather, Article 54 only applies after the losing State fails to pay an award pursuant to Article 53.').
61 L.B. Amsler et al. (2020) Dispute System Design: Preventing, Managing, and Resolving Conflict.Stanford University Press.62 See A. Antonietti (2006) 'The 2006 Amendments to the ICSID Rules and Regulations and the Additional Facility Rules', ICSID Review Foreign Investment Law Journal 21, 427, 430 for a detailed review of the amendments.

64SeeICSID
Rules of Procedure for Arbitration Proceedings (2006) Rule 41(5) and (2022) Rule 41(1) creating a special procedure to dismiss claims 'manifestly without legal merit'.For application of the standard, see e.g., Trans-Global Petroleum, Inc v Jordan (ICSID Case No. ARB/7/25, Decision on the Respondent's Objection under Rule 41(5) of the ICSID Arbitration Rules of 12 May 2008), para.92.65 To be sure, the Court of Arbitration for Sport is a good example of an international arbitral system that implemented an appeal process within the current functioning of the New York Convention.See L. Reilly (2012) 'An Introduction to the Court of Arbitration for Sport (CAS) & the Role of National Courts in International Sports Disputes', Journal of Dispute Resolution 63, 77.66 G. Kaufmann-Kohler and M. Potestà (2016) 'Can the Mauritius Convention Serve as a Model for the Reform of Investor-State Arbitration in Connection with the Introduction of a Permanent Investment Tribunal or an Appeal Mechanism?',Geneva Center for International Dispute Settlement, 3 June 2016 44-45.

67
Under the agreements, if an expropriation claim implicates taxation measures, the competent fiscal authorities of host and investor countries may block arbitration.See NAFTA article 2103(6), 32 ILM 605, 700(1993).68 Agreement between the United States of America, the United Mexican States, and Canada (USMCA), article 31.20, 30 November 2018.69 2012 US Model BIT article 28(9)(a) provides a process to consider comments on a proposed decision or award by the disputing parties.The practice is also used at the WTOalbeit with limited success.See Understanding on Rules and Procedures Governing the Settlement of Disputes [15 April 1994], Marrakesh Agreement Establishing the World Trade organization article 15.

Figure 2 .
Figure 2. Relationships between AM costs 74G.Shaffer (2015) 'The New Legal Realist Approach to International Law', Leiden Journal of International Law 28, 189.
75 A. Roberts (2018) 'Incremental, Systemic, and Paradigmatic Reform of Investor-State Arbitration', American Journal of International Law 112, 410 (noting the possibility of 'open architectural approaches that permit differently situated states to sign up for new multilateral approaches or institutional mechanisms').

36
For a critique of this approach, see, R.Coase (1988)The Firm, the Market and the Law.University of Chicago Press, 28.Langford et al. (2019) 'Empirical Perspectives on Investment Arbitration: What Do We Know?Does It Matter?', ISDS Academic Forum Working Group 7 Paper, 15 March 2019, 16-17.For additional discussion on the complexity of the reform, see United Nations Commission on International Trade Law, Working Group III: Investor-State Dispute Settlement Reform, UN Doc A/CN9/WGIII/WP159/Add1 Submission from the European Union and Its Member States: Possible Reform of Investor-State Dispute Settlement (ISDS) [37th Session, April 2019], https://undocs.org/A/CN.9/WG.III/WP.159/Add.1 (accessed 12 March 2023). 37M.
39 2018 Annual Comment, 'International Centre for Settlement of Investment Disputes (ICSID)' (2018) 36, https://icsid.worldbank.org/en/Documents/resources/2018ICSIDAnnualComment.ENG.pdf(accessed 12 March 2023).40 Appellate Body Secretariat WTO Doc WT/AB/29 at 142 Annual Comment for 2018 [March 2019] Some scholars also argue that the ECHR became 'inundated' with cases when review was introduced.D. Kurban (2016) 'Forsaking Individual The Implications of the European Court of Human Rights' Pilot Judgment Procedure for Victims of Gross and Systematic Violations', Human Rights Law Review 16, 745.41 The WTO litigation and settlement dynamics are different to ISDS.They generally involve upper-or middle-income countries.The remedies are also different.For discussion, see Y. Ngangjoh-Hodu and C.C. Ajibo (2015) 'ICSID Annulment Procedure and the WTO Appellate System: The Case for an Appellate System for Investment Arbitration', Journal of International Dispute Settlement 6(2), 308-331, 310; For other examples, see M.A. Carreteiro (2016) 'Appellate Arbitral Rules in International Commercial Arbitration', Journal of International Arbitration 33(2), 185-216, 200.42 Calculations are based on information from ICSID Annual Reports.See also, J. Calvert, C. Rommerskirchen, and A. van der Heide (2022) 'Does Ownership Matter?Claimant Characteristics and Case Outcomes in Investor-State Arbitration', New Political Economy 27(5).43 M.L. Busch, E. Reinhardt, and G.C. Shaffer (2009) 'Does Legal Capacity Matter?A Survey of WTO Members', World Trade Review 8, 4.For a discussion of these issues in ISDS, see Submission by European Federation for Investment Law and Arbitration to the UNCITRAL Working Group III on ISDS Reform, 'Contributed, 'Ensuring Equitable Access to all Stakeholders: Critical Suggestions for the MIC', UNCITRAL' (15 July 2019) https://uncitral.un.org/sites/uncitral.un.org/ files/media-documents/uncitral/en/wgiii_efila.pdf(accessed 12 March 2023) ('UNCITRAL Working Group III must address the risk that private parties will not see the MIC as a viable forum for enforcing the treaty protections granted to them by States'). Justice:

Table 2 .
Potential costs of AM Summary: Possible Costs of Appellate Mechanism

Table 3 .
Design solutions to address costs of AM