A. WTO Law Does Not Deal in All Things Considered Reasons

WTO law is a subset of public international law. Specifically, at least for the purposes of discussing WTO adjudication, it is the law deriving from the ‘Covered Agreements’, the WTO Agreement and its various annexes.Footnote ¹⁹ The task of the WTO Dispute Settlement System is to resolve disputes ‘in accordance with the rights and obligations under … the covered agreements’.Footnote ²⁰ The governing law in WTO dispute settlement is thus WTO law, rather than public international law as a whole.Footnote ²¹ As such, the questions that Panels and the Appellate Body (AB) must answer are not all-things-considered legal questions.Footnote ²² Where they find a violation of a particular provision, this need not mean that a member has acted contrary to ‘international law’, as a comprehensive scheme of rights and obligations. Equally, where no breach is found this does not mean a member has complied with international law generally. The question of all-things-considered legal obligation is logically downstream from WTO adjudication.

Why might this matter? Any normatively plausible positivist view of law will recognize a distinction between legal obligation (what the law says we should do) and moral obligation (what we in fact should do).Footnote ²³ This allows the positivist to characterize an unjust law as law, while acknowledging there may be no all-things-considered obligation to obey it.Footnote ²⁴ Identifying our legal obligations is simply one step in identifying our all-things-considered obligations.Footnote ²⁵

To the extent this strategy works for law in general, it applies a fortiori to any particular subset of law. To see why, consider the following example. We might think that a (domestic) legal system that avoidably placed individuals in extreme poverty was ipso facto unjust. This might lead us to think that relevant laws should be amended, or that we were justified in disobeying them, or that they should be (re-)interpreted to take account of the moral imperative to eliminate avoidable poverty. However, narrowing our focus to contract law, we might find nothing obviously wrong with this particular subfield, nor any obvious reason to reinterpret settled rules to, for example, relieve the needy of debts lawfully incurred. We could agree that the law should address avoidable poverty, and that its failure to do so was relevant to its moral force, but think some other subfield (tax, welfare or insolvency laws, for example) was better placed to do this. The upshot is that, if we ask the kind of partial question commonly asked in law schools (‘What is the correct legal analysis as a matter of contract law?’) we might feel comfortable bracketing such concerns entirely. Similarly, if the question ‘what are Agraria's all-things-considered legal obligations?’ is distinct from the question ‘what are Agraria's obligations, as a matter of WTO law?’, then, this argument suggests, there is nothing wrong with excluding values in answering the latter question. We can leave addressing these values to other subfields (environmental law, or human rights, or labour law).

However there are at least two reasons why this move won't work. First, as a matter of practice, WTO law and dispute settlement enjoy disproportionately greater influence than other areas of public international law and venues of international adjudication.Footnote ²⁶ And there is no mechanism whereby the conclusions of WTO dispute settlement are integrated with other international legal obligations (or—a fortiori—substantive values) to determine the demands States make of each other's conduct: a successful complainant can demand compliance with the terms of a dispute settlement report, and withdraw concessions where there is continuing non-compliance, regardless of whether there are good non-WTO legal reasons for that non-compliance. The result is that, while formally characterized as merely stating rights and obligations under the WTO agreements, in practice dispute settlement reports are treated as stating the all things considered obligations of members. In these circumstances, it is as statements of all-things-considered, rather than partial, obligations that they must be judged.Footnote ²⁷

Second, even if we allow that many competing values can be addressed at subsequent stages, when propositions of WTO law are integrated with others, there may still be values that specifically relate to practices addressed by WTO law, and so require to be integrated here. Recalling again the example of contract law, imagine a law that regularly enforces agreements procured by force or fraud. If we assume that the function of contract law is to enforce promises’Footnote ²⁸ or to protect legitimate expectations,Footnote ²⁹ or to give effect to welfare enhancing exchanges,Footnote ³⁰ then a contract law that regularly enforces agreements in these circumstances is defective as contract law. It is not sufficient that concerns about force and fraud are addressed elsewhere, in criminal or tort law for instance, because these are not defects of the legal system as a whole. They are defects of that part of the system that enforces promises/protects expectations/enables exchanges. Similarly, to the extent there are values specific to trade regulation, we cannot bracket these by reference to the partial nature of WTO law. This does not tell us whether there are such values. This will turn on our substantive views about trade regulation and the functions of WTO law, which are considered in Section IV.Footnote ³¹ However, it at least means that arguments for value-neutrality in WTO law omit a crucial step.

B. WTO Law Is Exclusively Contractual

A second possible argument highlights the ostensibly contractual quality of WTO law. The WTO Agreements, on this view, are a set of bargained commitments given by the members exercising their respective sovereignties. The only obligations attaching to them are those expressly accepted, running to (at most) an obligation to comply with those commitments.Footnote ³² There is thus no basis on which other values could play a role in their interpretation.Footnote ³³

This argument might be understood in various ways. As an argument about the moral obligations of States generally, it is obviously question-begging. Various scholars have written at length on the substantive political morality of trade regulation, which is considered further in Section IV.Footnote ³⁴ For the present, suffice to note that if we assume that States do not owe any natural duties of economic justice to those beyond their borders, this view seems plausible; if we accept the existence of such duties, its appeal diminishes.Footnote ³⁵

We might instead read this argument as specifically addressing the moral obligations attaching to membership of the World Trade Organization. We might accept that States had certain moral obligations in relation to their trade policies. Further, we might accept that these substantially overlap with WTO legal obligations. Nonetheless, we might deny that their legal obligations as WTO members were expressive of, or fell to be interpreted by reference to, those moral obligations: the two may simply be separate issues, with legal obligation constituting an additional, voluntarily accepted, layer, sitting apart from general moral obligation, and including only what is positively agreed.Footnote ³⁶

To the extent that this argument, so understood, is tenable, it depends on two controversial premises: first, a highly idealized view of the voluntary nature of agreement; and second, a quasi-libertarian account of the force of agreement as deriving from the will of self-owning individuals.Footnote ³⁷ As to the first, whatever its applicability to interpersonal contracting given an appropriate background of social protections within domestic societies, it is wholly inadequate for a multilateral scheme like the WTO, negotiated among countries of widely varying power without effective protections, and where membership is effectively non-voluntary for many States, to the extent they seek to access international markets.Footnote ³⁸ As to the second, it is significantly weakened by the move from interpersonal contracts, between persons plausibly possessed of the kind of autonomous wills that it presupposes, to interstate treaties, between collective agents whose choices are themselves political acts, deriving from complex social institutions and expressed through the exercise of non-voluntary political authority over individuals subject thereto. These points do not tell us what the binding force of interstate agreements is; but they suggest that a simple analogy to interpersonal contracting is inadequate to exclude values from their interpretation.

Finally, perhaps most significantly, the claim that values can be excluded because WTO law is exclusively contractual, howsoever understood, depends for its force on the assumption that we can in fact interpret contracts without recourse to values. However, as any contract lawyer will attest, in practice the interpretation of contracts relies heavily on judges’ views about what is fair and reasonable between the parties, albeit clothed in the language of objective intent.Footnote ³⁹ When identifying what contracting parties have agreed, judges have implicit recourse to what they think those parties, in their particular circumstances, should have agreed. The problem here is analogous to that of teleological treaty interpretation, discussed further in Section III, and the objection is itself one of possibility rather than desirability. Therefore, while it is mentioned here because it has particular relevance for those who regard contractual interpretation as an alternative to moralized interpretation, further discussion will be deferred until the Section III. Suffice to note that the international lawyer's utopia of amoral, quasi-contractual treaty interpretation does not survive contact with the actual practices of contract lawyers and judges.

C. WTO Law Does Not Claim Authority

A third potential argument against importing values into WTO adjudication queries whether WTO law purports to impose obligations on members at all. This may seem an odd suggestion, particularly to those persuaded by Joseph Raz's argument that claiming authority is a conceptual necessity of law.Footnote ⁴⁰ However a number of scholars have queried whether WTO law claims authority in the sense in which Raz uses that term.Footnote ⁴¹ Rather than expressing obligations that attach to members, this view suggests, WTO law simply records a standing political bargain, from which States are free to depart, but at the cost of a consequent rebalancing by other affected States. This may be through the withdrawal of concessions following dispute settlement, the modification or withdrawal of concessions under GATT Art XXVIII, rebalancing in response to safeguards under GATT Article XI and the Safeguards Agreement, or otherwise. However, in each case, it is simply a rebalancing, as opposed to an expression of fault, or blame, or punishment, of a kind that might be appropriate where an obligation has been breached.

This view of WTO law has been heavily criticized.Footnote ⁴² However, assuming that it is plausible, might it support an argument against admitting a role for values in adjudication? The claim might be that, in other contexts, the linking of legal and moral reasoning reflects the fact that both law and morality are understood as normative. Both claim to guide action, answering the same questions, and so unavoidably interact.Footnote ⁴³ If WTO law is not normative in this sense, then it might be more plausible to keep it insulated from moral reasoning.

This, however, ignores the extent to which, even if only conceived as a trigger for rebalancing, the questions WTO law answers are appropriate to moral evaluation. Where a member withdraws concessions, this can have significant impacts on the material well-being of individuals, including (at least) producers in the country against which concessions are withdrawn, and consumers in the retaliating country.Footnote ⁴⁴ Without prejudging the answer, intentionally causing such impacts is clearly a potential object of moral evaluation.Footnote ⁴⁵ To the extent that withdrawal of concessions is justified as ensuring compliance with, or responding to violation of, the WTO agreements, the moral evaluation of those agreements in turn seems required. The specific question asked may be slightly different: the morality of endorsing or being bound by a rule is not identical to the morality of enforcing it against others who themselves are not morally bound.Footnote ⁴⁶ However, simply denying the authority of WTO law doesn't take us very far.

D. Admitting Values in WTO Law Will Be Self-Defeating

A fourth set of arguments against engaging with values in WTO law and adjudication emphasizes the compliance mechanisms through which law affects State policies. If we assume (which seems reasonable) that the point of engaging values in WTO law depends significantly on affecting State behaviour, then we might ask how likely this is, given plausible views about how and why compliance with international law happens.Footnote ⁴⁷ If moral law will not motivate moral behaviour, and especially if it risks undermining the less ambitious but still significant goods of stability, predictability and mutually beneficial cooperation, this would provide a significant argument against such an approach.Footnote ⁴⁸

Consider three alternative stories, drawing on three prominent schools of international relations theory. First, realists characterize compliance as driven by pressure from powerful affected States.Footnote ⁴⁹ The prominence of retaliation in WTO law suggests that this plays an important role. At first glance, this might seem to empower adjudicators by putting State power at their disposal.Footnote ⁵⁰ However, this is true only in so far as interpretations proposed by adjudicators accord with those demanded by relevant powerful States. Where interpretations do not meet the expectations of powerful States, these States can be expected to instead act on their own judgment. The United States’ recent enthusiasm for unilateral retaliation against perceived ‘cheating’ by China highlights this point.Footnote ⁵¹ The worry here is that engaging with values in WTO adjudication risks leading the Appellate Body to diverge from powerful States’ expectations to such an extent that decisions cease to be credible triggers for retaliation.Footnote ⁵²

Second, institutionalists characterize compliance as driven by reputational concerns.Footnote ⁵³ Under anarchy, the lack of centralized enforcement makes reputation a key determinant of States’ capacity to enter into mutually beneficial agreements. This gives them reasons to value and preserve their reputations for good faith compliance with those agreements. Various features of the WTO system, including multilateral surveillance, suggest these mechanisms play an important role. The function of dispute settlement in this model is to increase the reputational costs of non-compliance. However, this will only work if decisions are seen as tracking States’ compliance with commitments. If decisions are perceived as going beyond the commitments States saw themselves as making, then they cease to provide relevant information for assessing counterparties’ compliance, and respondents no longer suffer reputational costs for non-compliance. So again, it appears that engaging with values risks undermining the effectiveness of dispute settlement.Footnote ⁵⁴

Finally, we might understand WTO compliance in constructivist terms. States, on this view, comply because they see this as appropriate behaviour, having regard to their identities as liberal/law-compliant/fair etc.Footnote ⁵⁵ Compliance reflects not an instrumental calculus of enlightened self-interest, but an intrinsic judgment of right action in relevant circumstances. Given that, on this view, ideational, including moral, factors play an immediate role in determining State behaviour, we might imagine that it was most compatible with admitting values in WTO adjudication. However, this depends on a careful reading of the relevant social norms, identities and understandings. An identity as law-compliant gives greater influence to those authorized to interpret the law. However, it is also exposed to criticisms that adjudicators have gone beyond interpretation to impose their own values. The WTO Appellate Body has a plausible claim to be an authoritative legal interpreter, but is only one amongst many potential contributors to debates about relevant values. And if its interpretations move too far from the mainstream, it will lose even that position, at least for some relevant stakeholders.Footnote ⁵⁶

Is there a response to these objections? As regards the first two objections (realist and institutionalist), there remains significant leeway within the constraints imposed by the triggering and reputation functions that each attributes to dispute settlement. Constructing institutions, whether as tools for imposing one's will, or managing cooperation, is difficult and expensive. Acting through ostensibly impartial institutions provides benefits that are lost if these are too readily set aside, or are seen to be too responsive to the demands of more powerful States.Footnote ⁵⁷ As such, States will tolerate some deviation from their perceived optimal outcomes before they seek to directly challenge, and ultimately undermine, existing institutions.Footnote ⁵⁸ This at least provides a range within which adjudicators have agency to accommodate values in adjudication.

As to the third objection (constructivist), while the Appellate Body is not an authoritative interpreter of the morality of trade, it is plausibly a privileged contributor to discussions on the subject, in much the same way that constitutional courts enjoy a privileged place in domestic political debate. There is a limit to how far any individual voice can move a debate, and equally there is a limit to how far any contributor can move from the centre of the debate while retaining influence.Footnote ⁵⁹ But within these constraints, incorporating values in adjudication can potentially move political debates about appropriate conduct in trade regulation, and the trade regime as a whole, in the right direction.

Under each of these three perspectives, then, there is scope for admitting substantive values in adjudication. They do, however, sound an important warning: if engaging with values moves adjudicators too far from the commitments particular members—and particularly more powerful members—see themselves as making, or their respective and shared understandings of right action, this must undermine the effectiveness of WTO law. These limits must therefore be borne in mind in identifying the ‘best’ interpretation in any particular case.Footnote ⁶⁰ At the same time, the diversity of such subjective commitments and understandings, which vary across members, and as regards a particular member across relevant persons and groups, and across time, mean that simply seeking to track those understandings cannot yield consistent answers. (This point is developed further in Sections IIIC and IIID below.) Further, making political acceptability the sole focus of, as opposed to a background constraint on, interpretation risks undermining the social legitimacy of law and dispute settlement, and indeed the system's essential nature as legal. Consider: most observers of the WTO will recognize that, if the United States, or the European Union, or China, loses every dispute, then they can be expected to disengage, with consequent damage to the system as a whole; yet an interpretive approach that explicitly sought to accommodate the demands, however unreasonable, of one or more of these powers, would thereby give up any pretence of being a system of reasoned decision-making. It would thus concede its authority both over the leading powers (who know that what matters is the force of their political demands, not the strength of their arguments) and over less influential States (who see that success depends on political acceptability to and relationships with the great powers, rather than legal rights and principles). What is required is to strike a difficult balance, recognizing and accommodating to political reality, without abdicating the distinctive function of law.Footnote ⁶¹

The institutionalist and constructivist stories have a second important implication. Each highlights the importance, not simply of outcomes in adjudication, but also of the reasoning that supports those outcomes. The same conclusion, they suggest, might be effective or ineffective in moderating State behaviour, or might reinforce or undermine the regime as a whole, depending on the quality of the reasoning supporting it. Decisions that appear arbitrary or unmotivated are less likely to attract compliance and support than those that are supported by convincing reasoning. It is therefore necessary to consider whether, and to what extent, different ways of thinking about the role of values in adjudication might serve to enhance, or alternatively to undermine, the force of that reasoning.

A. Two Positivisms and the Role of Discretion in Adjudication

Before proceeding, it is necessary briefly to recite some familiar ideas from analytical jurisprudence. While this will take us a little way from WTO dispute settlement, the detour will be short, and will yield results that can be brought directly to bear on our object of inquiry.

Consider, first, the tradition that legal theorists label analytical positivism. As a tradition, it comprises a range of distinct views.Footnote ⁶² However, it is perhaps most readily identified with some version of what Joseph Raz labels the Social Thesis: ‘that what is law and what is not is a matter of social fact’.Footnote ⁶³ Legal propositions, on this view, are valid exclusively in virtue of social facts, including most prominently facts about the past actions of political institutions. Whether a given proposition is or is not law is answered exclusively by reference to such facts, which we label sources. In the domestic context, these may include primary legislation, administrative acts, judicial decisions and so forth. Internationally, they will include treaties and the actions of both States and jurists that together constitute customary international law.

A key implication of legal positivism, so understood, is that the law is found without recourse to values or moral reasoning. It may be that in particular instances, the law will direct adjudicators to enquire into matters of value, or to apply moral standards, but the question of the law's content is answered exclusively by reference to non-moral facts.Footnote ⁶⁴ Gardner offers the following formulation for this idea, which he characterizes as the (minimal) shared core of analytical legal positivism:

It is not my purpose here to engage with the substantive merits of legal positivism. Rather, the intention is to highlight one implication that everyone—positivists and anti-positivists—accepts that this view implies. This is the fact that the law, so understood, is necessarily incomplete. If we understand law as exclusively a matter of sources, then it will inevitably be the case that, in at least some situations, there will be no legal ‘fact of the matter’.Footnote ⁶⁶

This may arise for one of a number of fairly uncontroversial reasons. First, it reflects the inherent limits of linguistic determinacy. Hart refers to this as the ‘open textured’ nature of language, which means that, for any given term, proposition or precedent, there will be some cases that fall squarely within it, which he labels the ‘core’, and others that might or might not fall under it, which he labels the ‘penumbra’.Footnote ⁶⁷ Second, it reflects the impossibility of legislators anticipating in advance all possible circumstances that may arise in the future, which means that it falls to those applying the law to decide whether unforeseen situations that arise fit under a given rule or not. This is at once necessary (given the cognitive limits of human legislators), efficient (given the transaction costs of seeking to anticipate everything) and desirable (given the better position that judges are in to answer questions as they arise, instead of seeking to do so in the abstract in advance).Footnote ⁶⁸ Third, it reflects the political realities of both domestic and international law-making, where plural and often conflicting interests and perspectives are only partly integrated. This leads to gaps, ambiguities and contradictions, as lawmakers seek to paper over disagreement through ambiguous language, effectively delegating controversial decisions to subsequent interpreters.Footnote ⁶⁹

The upshot is that, while positivists characterize law's content as exclusively a matter of social facts, they do not claim that adjudication is, or can be, exclusively a matter of law, so understood.Footnote ⁷⁰ Rather, adjudicators necessarily reach beyond law to resolve disputes in the penumbra, to fit rules to unforeseen circumstances, and to answer questions that legislators left unresolved.Footnote ⁷¹ These are commonly characterized as cases where adjudicators exercise ‘discretion’ in reaching decisions that take them beyond existing law.Footnote ⁷² Where their decisions have formal or informal precedential value, we may also characterize them as delegated judicial legislation.Footnote ⁷³

Positivism, in this analytical sense, is related to, but distinct from, positivism as that term is often understood by international lawyers.

For the international lawyer, positivism is primarily a claim about the relation between international law and State consent.Footnote ⁷⁴ International law is law because, and only because, it has been consented to as such by States who are subject to it.Footnote ⁷⁵ Treaty and custom are each understood as expressing that consent. Analytical legal positivism is a view about what kinds of facts (social facts) constitute law, whereas international legal positivism is a view about which social facts (State consent) count as sources. An obvious case where the two come apart is the acts of international organizations, such as General Assembly Resolutions. These are clearly social facts, of a kind that analytical positivists can readily admit might make law, whereas international legal positivists will deny their legal status, unless they can be treated as directly or indirectly expressing State consent.Footnote ⁷⁶ Equally, an international legal positivist might readily admit (analytical) anti-positivist elements, including in particular a concern with object and purpose, in the interpretation of consent-based sources.Footnote ⁷⁷ So analytical positivists need not be international law positivists, nor must international law positivists be analytical positivists.Footnote ⁷⁸

B. Can WTO Adjudicators Be (Analytical) Positivists?

Analytical positivism thus implies a particular view about the role of values in adjudication.Footnote ⁷⁹ Law, understood exclusively in terms of social facts, answers many questions. Where the law is silent, however, adjudicators committed to deciding cases have little choice but to look beyond law, including in particular to relevant values, to fill the gaps and provide answers in hard cases.Footnote ⁸⁰ This, it is suggested, renders analytical positivism incompatible with the political and doctrinal context in which the Appellate Body exists. The Appellate Body publicly endorses international law positivism; but the same reasons that lead it to that position preclude it from accepting the consequences of analytical positivism. The Appellate Body's self-presentation, formal mandate and political context are sketched below, showing how each points towards international legal positivism, but away from analytical positivism.

The Appellate Body publicly presents itself as an exponent of international legal positivism.Footnote ⁸¹ Thus, in one of its earliest decisions, it observed as follows:

This concern with consent as the root of legal obligation among sovereign States has featured in the Appellate Body's reasoning in a number of subsequent cases.Footnote ⁸³ More generally, its emphasis on the priority of the text in its interpretive approach seems to reflect a concern with State consent as the source of its authority.Footnote ⁸⁴

This professed enthusiasm for consent can be understood as a response to the Appellate Body's specific mandate, as set out in Article 3.2 of the Dispute Settlement Understanding, specifically its duty to ‘preserve the rights and obligation of Members under the covered agreements’, and prohibition on ‘add[ing] to or diminish[ing] the rights and obligations provided in the covered agreements’. While there is a lot squeezed into Article 3.2, a key message the Appellate Body has taken from it is that its task is squarely focussed on the positive law found in the WTO Agreements.Footnote ⁸⁵ Those agreements are to be central to its reasoning, and constitute, in the first instance, the sources from which it is expected to work.Footnote ⁸⁶ The denial of precedential force to Panel/AB decisions further emphasizes the priority of the texts to which the members have consented.

Emphasizing the text of agreements and State consent is also understandable as a response to the Appellate Body's particular political and institutional vulnerability, especially in its early years. Like all international adjudicators, the Appellate Body is reliant on the continuing support of States to give effect to its decisions. However, three features render it especially vulnerable.

First, it has compulsory jurisdiction and is tasked with resolving disputes of immediate economic significance for both complainant and respondent States. This distinguishes it from human rights courts, where complainants are either individuals, or other States whose interest is often principled rather than practical. It also distinguishes it from the ICJ, where judgments on matters of high politics are expected to contribute to, but rarely to settle, ongoing political disputes, and where the need to establish jurisdiction limits risks from non-compliance. The AB is therefore more exposed than other international tribunals to non-compliance undermining its social legitimacy.

Second, as the current crisis illustrates, the role of consensus in WTO decision-making provides members—and especially those more powerful members capable of standing alone against pressure from others—with a powerful, but dangerously blunt, tool for disciplining adjudicators who fail adequately to legitimate their decisions and reasoning in the eyes of those members. Article IX.2 of the WTO Agreement provides a mechanism for precisely targeted political control of Appellate Body jurisprudence, through the adoption of authoritative interpretations. However, the requirement of a three-fourths majority in that provision means that it does not provide an effective remedy for members dissatisfied with the way the agreements have been interpreted. This leaves procedural obstruction as the only effective mechanism of control. Yet the dramatic (because it requires making an existential threat against the system) and blunt (because it necessarily attacks the system as a whole) nature of that mechanism in turn places a significant onus on the Appellate Body to pre-empt its use by legitimating its conclusions in terms of the expressed consent of States.

Third, as Howse argues, the Marrakesh Agreement marks a high-water mark for neo-liberal globalization. Within a few years of the WTO's creation, anti-globalization protest movements were challenging the legitimacy of its project. This led the Appellate Body in its early years to adopt a cautious, technical, judicial minimalism, insulating it from challenges to the WTO as a political institution.Footnote ⁸⁷

However, those same factors that push the Appellate Body towards international legal positivism constitute a barrier to its adopting analytical positivism. Recall, the analytical positivist commitment to understanding law as a matter of social facts brings with it a recognition that adjudication cannot be simply a matter of applying law; rather, on this view, adjudication involves a significant degree of judicial discretion, and decisions will frequently turn on extra-legal considerations.

Yet it seems clear from Article 3.2 DSU that the members understood the role of the Appellate Body as a strictly legal one. It is not the purpose of the dispute settlement system, at least as described there, to go beyond the law in deciding cases, or to create new law through extra-legal discretion. It is the membership, under Article IX of the Marrakesh Agreement that is authorized to adopt authoritative interpretations, exercising the political discretion to fill any gaps in those agreements.Footnote ⁸⁸

Further, given the political context outlined above, any explicit recognition by the Appellate Body of the kind of discretion positivists identify with adjudication could be expected to evoke opposition from both member representatives within the ‘political’ WTO, and from wider political legitimacy constituencies. More than bare judicial fiat is required to legitimate interpretation.

There are certainly rational-choice models of adjudication under which ‘we have decided’ would be sufficient for these purposes, in the sense of motivating the parties thereto to comply. If we model adjudication as an iterated game of indefinite duration for relatively low stakes, then provided others can also be expected to do so, the losing party to a dispute will be motivated to comply by the expectation of future successes.Footnote ⁸⁹ However, given the diversity of constituencies and interests involved, such mechanisms cannot plausibly stabilize a regime of 164 members: what are small stakes for some will be large stakes for others, decisions in one dispute have consequences for others, and no set of decisions can be expected to ensure every member continues to have a narrowly rational interest in compliance. Further, once we acknowledge the role of wider compliance constituencies beyond the ‘trade insider’ community, simply tracking the rational self-interest of members, as understood by relevant constituencies within their trade bureaucracies, ceases to be sufficient: decisions on economic, political, environmental and ethical matters of concern to wider populations require to be legitimated to those populations.Footnote ⁹⁰ Simply asserting that the Appellate Body has so decided, without more, can be expected to carry little weight in this context.

We can here distinguish the position of the Appellate Body from certain other tribunals, both domestic and international, which might enjoy sufficient social legitimacy to demand respect for decisions simply in virtue of their origin. Domestic courts, at least in stable democracies, are embedded within widely publicly endorsed social institutions. Further, they usually enjoy a degree of—admittedly mediated—democratic legitimacy, in virtue of their appointment processes, as well as epistemic legitimacy deriving from the professional reputation of individual judges and the social acceptance of their procedures. Even where—as in the United States—higher courts are intensely politicized, their decisions are accorded significant respect by other political institutions and by the wider population, simply in virtue of the fact that they are court decisions. Legal professionals, who themselves constitute an influential class within official bureaucracies and wider civil societies, are socialized to respect those decisions, whether or not they appear correct or well-reasoned on their merits.Footnote ⁹¹ This gives domestic courts a greater capacity to decide ‘off their own bat’ when faced with legal lacunae.

The international tribunal that has enjoyed perhaps the greatest success in establishing its own social legitimacy within its judicial domain, the European Court of Justice, has done so precisely by recognizing this fact, and building implicit alliances with these influential domestic professional and judicial communities.Footnote ⁹² But the key tools through which the ECJ forged those alliances—preliminary reference and individual rights of action—are not available to the Appellate Body.Footnote ⁹³ Other tribunals that have achieved significant popular legitimacy, such as the European Court of Human Rights, have relied on the support of civil society actors who in turn enjoy significant popular sympathy, while also—as noted above—addressing disputes where non-compliance is less likely to undermine long-term social legitimacy. The upshot is that—unlike many other tribunals—the Appellate Body lacks the degree of deep social legitimacy required to effectively exercise the kind of extra-legal discretion implied by analytical positivism.

Further, the Appellate Body has itself been careful to avoid acknowledging any such discretion in its role. This is perhaps clearest in its approach to the Anti-Dumping Agreement, where Article 17.6(ii) specifically anticipates legal indeterminacy, providing:

From an analytical positivist perspective, this provision is perfectly intelligible. Where the law settles a particular question, domestic authorities should be required to follow it. Where a point is indeterminate (ie there is no legal fact of the matter), any permissible interpretation should be allowed, and the Panel/AB should not exercise its (extra-legal) discretion to determine the matter.

Yet the Appellate Body, on each occasion where Article 17.6(ii) has been invoked, has found that the underlying provision admitted only one permissible interpretation, interpreting its mandate to apply ‘customary rules of interpretation of public international law’ so that it will almost never be the case that a provision ‘admits of more than one permissible interpretation’.Footnote ⁹⁵ Rather, the AB's view is that there will (almost) always be one (legally) right answer.Footnote ⁹⁶ Given the limited textual resources underpinning these decisions and regular disagreements between Panels and Appellate Body, this is a difficult conclusion to square with the analytical positivist view of law's limits. However, once we recognize the doctrinal and political impediments to extra-legal discretionary decision-making by the Appellate Body, it begins to make sense. After all, if the interpretation of the Anti-Dumping Agreement does not lead to a single right answer, then why should we expect the other agreements to do so? If there are multiple reasonable interpretations of the Anti-Dumping Agreement, the same will presumably be true of other agreements? But once that point is conceded, then the Appellate Body will struggle to maintain its authority to prescribe one of those reasonable interpretations as correct, and the others as incorrect.

C. Analytical Positivism and the Customary Rules of Interpretation

The Appellate Body's formal mandate, institutional position, and judicial self-presentation are thus difficult to reconcile with the analytical positivist view of the limits of legal determinacy and the necessity of extra-legal discretion. That analytical positivist view is a necessary one: it claims to describe the nature of law, rather than simply the conditions of any particular legal system.Footnote ⁹⁷ If the Appellate Body's doctrinal mandate and political position preclude it exercising the kind of discretion that positivists identify with adjudication, then this poses a difficulty, regardless of any particular features of WTO law. However, the scale of the difficulty will vary, depending on the extent of the indeterminacy in that law. It is therefore worth exploring a little further how the indeterminacy of positivist sources plays out in the context of WTO adjudication, in order to appreciate the extent of the challenge in reconciling the two.Footnote ⁹⁸

Article 3.2 directs adjudicators to ‘customary rules of interpretation’, which the Appellate Body has consistently read as referring, in the first instance, to those principles codified in Articles 31–33 of the Vienna Convention on the Law of Treaties.Footnote ⁹⁹ Article 31.1, in particular, is regularly invoked in justifying the Appellate Body's interpretive approach. It provides as follows:

Three points should be noted about this provision.

First, Article 31.1 identifies not one but three different approaches to interpreting treaties: textual, contextual and teleological. In some cases, these three may point in the same direction, suggesting a high degree of confidence that a particular interpretation is correct. In others, a provision that appears indeterminate by reference to one standard may be clarified by another, again offering relative certainty as to the interpretation to be preferred. However, in yet other cases, different approaches point in different directions. Where this happens, some way must be found to choose between the interpretation suggested by the ordinary meaning (textual) of the terms of a provision, and that suggested by its object and purpose (teleological).Footnote ¹⁰⁰ Yet neither Article 31, nor the prior or subsequent customary practice that it codifies, provide a rule for choosing among these interpretive methods when they conflict.Footnote ¹⁰¹ The Appellate Body's characterization of the interpretive exercise as ‘holistic’, and prominent emphasis on text in the first instance, merely obscure rather than excluding these choices.Footnote ¹⁰² Rather, in the analytical positivist view, these will be situations where adjudicators must exercise extra-legal discretion in deciding the case before them.Footnote ¹⁰³ The more such situations we find, the greater the need for discretion, and hence—if we accept the positivist view—the greater the legitimation problem.

Second, there are significant problems understanding the third of these criteria, object and purpose, in positivist terms.Footnote ¹⁰⁴ Positivist theorists have generally been quite sceptical of legislative intent as a component of legal interpretation. Joseph Raz, for example, argues that the only legislative intent that is relevant is the intent to enact the particular legislation that has been enacted, understood in accordance with its terms and with prevailing legal and linguistic conventions.Footnote ¹⁰⁵ However scepticism about legislative intent is not a necessary feature of positivism. In circumstances where it is possible to identify the relevant legislative intent by reference to the kinds of social facts that positivists emphasize, and where the prevailing conventions of legal interpretation characterize it as relevant, interpreting the law by reference to such positive intent need not necessarily give positivists pause.Footnote ¹⁰⁶

Third, in the context of the World Trade Organisation, the identification of object and purpose is generally a fraught exercise. For public international lawyers, preambles are a standard source of guidance as to the object and purpose of agreements. However, a number of WTO agreements lack any preamble, while others have preambles that are singularly unilluminating, frequently saying little more than ‘the purpose of this agreement is to provide rules on the topics governed by this agreement’. Further, this is not simply a matter of incomplete evidence. Rather, at key points, the WTO agreements constitute difficult compromises between fundamentally opposed views.Footnote ¹⁰⁷ This is perhaps clearest in the context of trade remedies. Thus, the Safeguards Agreement can be understood either as an attempt to discipline unilateral recourse to protectionism; or an attempt to discourage other ‘grey area’ measures, by making recourse to safeguards easier.Footnote ¹⁰⁸ The Subsidies and Countervailing Measures Agreement is about restraining States’ use of subsidies; but also restraining States’ recourse to unilateral measures to respond to subsidies.Footnote ¹⁰⁹ And the Anti-dumping Agreement is about restraining, while also licensing, anti-dumping.Footnote ¹¹⁰ In no case do we get a clear sense of the problem members saw themselves as trying to solve, of what was objectionable about the underlying practices to which these agreements apply, or of the principles underpinning the solutions adopted. The upshot is that, if we understand object and purpose as referring to the actual goals of actual States at the time the agreements were entered into (which we must if we retain the positivist emphasis on social facts), then there simply is no object and purpose behind many of these agreements, and any reference thereto must involve an exercise of judicial discretion.Footnote ¹¹¹ Indeed, these difficulties may explain why the Appellate Body has been, in general, quite slow to rely on object and purpose as a prominent element in its interpretive approach.

Readers might worry that the degree of uncertainty suggested by the previous paragraphs does not reflect our experience of treaty interpretation in general, or WTO treaty interpretation in particular. There have been and continue to be prominent points of interpretive disagreement among WTO lawyers. However, across much of WTO law, there is substantial agreement on many points. We are not lost in a sea of subjectivity or indeterminacy. To the extent this is the case, the above points suggest that there is more going on than the analytical positivist assumes: the very fact that we do agree about so much suggests that our understanding is built on more than just social facts.Footnote ¹¹² Rather, it reflects shared assumptions about the values that the law should be pursuing; and it is only where those shared assumptions break down that interpretive disagreements become both prominent and entrenched.

D. Constructing Object and Purpose

The implication, then, is that in so far as we find WTO treaty interpretation relatively unproblematic, we must be drawing, whether explicitly or implicitly, on anti-positivist elements; and conversely, that if we insist on adopting a strictly positivist account of WTO law, then we risk undermining much of that certainty, replacing legal interpretation with political choice.

To make what is implicit explicit, this section examines in more detail what is involved in identifying the intention behind a treaty text, where this is not expressly stated. The goal is to demonstrate how far that exercise, whether addressed to text, context or object and purpose, requires making substantive value judgments incompatible with the positivist account of law. The point here is not limited to the WTO, applying (at least) to multilateral treaties in general. However, it is particularly evident in this context, and particularly problematic, given the constraints on Appellate Body discretion noted above.

Some may find the discussion here unhelpfully abstract. Certainly, it does not seek to reflect the ways adjudicators describe their reasoning in their decisions, or even the conscious steps by which they proceed. Rather, it describes what we must be doing when we seek the intention behind another's act. It is the explanation that results if we repeatedly ask, of our shorthand ascriptions of intent, ‘yes, but how do you know?’. It is therefore an effort to describe ‘what is really going on’ when we talk about the intention behind a treaty, including its object and purpose. The answer it points us towards is a firmly anti-positivist one.

If we cannot find direct evidence of the object and purpose of the WTO Agreements, in the form of preambular or other statements, then we might instead hope to deduce these from the agreements themselves. Indeed, this is a standard approach adopted not only by the WTO Appellate Body, but by public international lawyers more generally, in seeking the object and purpose of agreements and indeed of individual provisions. We examine the relevant provisions, in their relevant context, and we ask ‘what are these provisions seeking to achieve?’. Once we have an answer to this question, we can in turn apply it in interpreting the relevant provision.Footnote ¹¹³

However, identifying the object and purpose of an agreement from the agreement itself cannot be a wholly positive inquiry. We cannot deduce a unique purpose by simply applying logic to text: there will always be multiple explanations—of admittedly varying plausibility—that can be offered for why negotiators adopted a particular text. It is only by importing assumptions about the kinds of agents whose object and purpose is being sought that we can hope to choose between these—that we can evaluate their plausibility—and thus move from text to purpose.Footnote ¹¹⁴

The point here is very general, and not limited to treaties, or even to legal instruments. To attach a purpose to an action, we need to know something about the kinds of goals and purposes agents of the relevant kind have. If I observe a person scratch their head, I might conclude their purpose was to relieve an itchy scalp. That conclusion follows not simply from the observed behaviour, but also from my knowledge that humans occasionally have itchy scalps, and are inclined to try to relieve them. I identify the purpose by examining the action together with assumptions about the motivations agents of the relevant kind have.

In the case of persons, the assumptions required are not necessarily normative. Rather, they are psychological assumptions, which are at least potentially subject to empirical examination. However, States are not the sorts of agents about whom there are psychological facts. Individual negotiators may have actual intentions, and these may even be shared amongst negotiating teams. However, once we shift our focus to collective agents, including States, it becomes much harder to identify the ‘hard facts’ on which statements about intention supervene. Rather, the attribution of purposes to States is always metaphorical.Footnote ¹¹⁵

Further, even if we could attribute concrete purposes to individual States, interpreting treaties is a step further removed from psychological facts about individuals, requiring that we identify the shared object and purpose of the parties to the relevant agreement, rather than simply the disparate purposes of each.Footnote ¹¹⁶ Where a treaty explicitly states such a purpose, we might read this as being adopted by the members as a whole; but—as noted above—statements of purpose in the WTO agreements are rarely especially illuminating, where they appear at all. Absent such statements, it is hard to imagine what physical, social or psychological facts claims about shared objects and purposes pick out.Footnote ¹¹⁷

Where an object and purpose is not explicitly stated, then, it will always be the case that interpreters are constructing one, based on assumptions or metaphorical ascriptions of particular motivations to the States parties thereto. How should that ascription be made?

One option is to ascribe to the States the actual motivations of the politicians and diplomats responsible for negotiating on their behalf. In the context of trade agreements these will in turn reflect, at least to some degree, the various interest groups lobbying for particular provisions, as well as the wider political and social trends at the time an agreement is reached. They will also include, in the case of the WTO agreements, the distinctive ‘GATT-think’ which various scholars have highlighted as playing an important role in maintaining and advancing the trade regime.Footnote ¹¹⁸ The telos of the covered agreements, on this view, is likely to be read in neo-liberal terms, hostile to trade barriers and market interventions, and relatively unconcerned with issues of regulatory sovereignty or distributive impacts, whether local or global.Footnote ¹¹⁹

Alternatively, we might ascribe motivations reflecting current political trends and preferences in leading members. While not articulated in these terms, this seems to be the approach that Howse attributes to the Appellate Body. In the earliest years of the WTO era, the Appellate Body recognized that the originalist telos, shared among trade ‘insiders’ and political elites in powerful States in the early 1990s, was significantly in retreat across much of the developed world, and probably never enjoyed much support among developing country members. Their response was—following a number of years of cautious, textualist, and in consequence often unconvincing reasoning, to embrace the much older telos of embedded liberalism, policing the more egregious cases of protectionism, but allowing significant autonomy to members to pursue their own domestic regulatory and economic policies.Footnote ¹²⁰

Unfortunately neither alternative, whatever their merits, solves the problem of plural objectives across States. There may be individual negotiators and policymakers whose particular intentions can be attributed to the States they represent. There may even be shared objectives that we can attribute on an ongoing basis to particular polities. However, the intent that we seek is one that can be attributed to all 166 WTO members. Given clear divergences on key points—evidenced inter alia by the fact of recurring interpretive disputes—it seems clear that focusing on actual intentions, whether historical or contemporary, cannot provide the required answers.Footnote ¹²¹

Instead, absent a clear statement in an agreement, any normatively plausible account of object and purpose must rely on a hypothetical account of the interests and motivations of the States involved. Instead of asking ‘what did these particular States seek to achieve through this agreement?’, we ask ‘what would appropriately motivated States, entering into an agreement of the relevant kind, in the relevant circumstances, have intended to achieve by it?’. The important question then becomes how we should understand ‘appropriately motivated’.

We might, for example, adopt a materialist and egoist account of State motivation. States, on this view, are concerned to advance their own interests, understood in terms of the material well-being of their citizens. This fits well with standard economic readings of the WTO Agreements.Footnote ¹²² States enter into these agreements because they hope to make themselves materially better off as a result. Where there is disagreement about the correct interpretation of a particular provision, we should adopt the interpretation that seems best fitted to realize that goal.

If instead we model States as concerned to protect their prerogatives as sovereign and independent, we will attribute a very different object and purpose to the agreements, addressing specific mutually harmful and destabilizing policies while respecting States’ autonomy over regulation and domestic economic policies.Footnote ¹²³ Many contestable interpretive questions, that on the first view we might read as disciplining State action to eliminate inefficiencies, will now be read as permitting a greater degree of freedom to States. It is only the most egregious protectionism and other mutually harmful policies that should be disciplined on this view.

Nor are these the only plausible ways we can model States motivations. If we model States as materialist, but replace egoism with altruism, we get something closer to a global utilitarian interpretation, in which provisions are understood as advancing global, as opposed to national, welfare. If we model States as concerned to be the authors of their own destinies, we will emphasize those features of the agreements that instantiate some form of market fairness.Footnote ¹²⁴ If we model them as security-seeking positionalists, we will emphasize provisions that allow States to preserve relative gains by challenging others’ policies, even at significant cost to themselves.Footnote ¹²⁵

E. Choosing Assumptions

It seems, then, that how we interpret the agreements turns unavoidably on assumptions about the appropriate motivations of parties thereto. This is not just one among many ways to approach the agreements. It is an implicit requirement of any attempt to attribute a purpose to the words therein, and thereby answer interpretive questions. The previous section highlighted how different assumptions will point towards quite different objects and purposes, and in turn quite different interpretations of the agreements themselves. I will not attempt here to identify which assumptions are appropriate. Rather, let me suggest three strategies for selecting amongst them, which we might label the explanatory, the moral and the political. There may be others, but these seem the most plausible. Significantly, for our purposes, none rescues the analytical positivist aspiration to find WTO law by exclusive reference to social facts.

The explanatory strategy works from assumptions that as far as possible reflect the observed behaviours of relevant agents. Following theoretically inclined social scientists, we seek motivational assumptions that can support reliable predictions about agents’ behaviour or interactions. Examples include the realist image of States as security-seeking positionalists, and the economist's image of persons as rational utility maximizers. Neither necessarily claims to describe how agents of the relevant kind should be, or indeed are, motivated. Rather, they claim that assuming these motivations allows us to explain and predict relevant behaviours.Footnote ¹²⁶

The moral strategy works from an account of how agents of the relevant kind should be motivated. While recognizing that States—or rather the individuals who determine State behaviour—are frequently motivated by various base motives, it models their behaviour as though they were appropriately motivated. A wide variety of different moral motivations might be proposed, from the nationally egoistic to the wholly other-regarding.Footnote ¹²⁷ They might include (amongst many others) the promotion of national economic welfare, the advancement of national projects, the preservation of international peace and cooperation, the promotion of global equality, the maximization of global well-being. The motivation we assume depends on the substantive moral theory we endorse.

The political strategy, like the moral one, focuses on the ways States should be motivated, rather than the ways they actually are. However, unlike the moral strategy, it does not answer this question by reference to a comprehensive moral view. Rather, it looks to the shared understandings that are present in the public political culture of the international system. This is Rawls’ move in his later work, in search of a theory that is ‘political not metaphysical’.Footnote ¹²⁸ Like the first two strategies, adopting the political strategy is not the end of the argument: there will be various plausible interpretations of the conception of the State and its motivations expressed in the international public political culture. However, the disagreement on this view is one of cultural interpretation rather than social or moral theory.

Arguments can be made for each of these strategies. For the present, it suffices to note that the choice among them, just like the choice of a particular conception of State motivation, cannot be made exclusively, or even predominantly, by reference to ‘hard’ physical, social or psychological facts. Rather, it will be made and defended by reference to the function that judgments of object and purpose play in legal interpretation and, more generally, to the function of legal interpretation itself. Different answers will yield different understandings of what legal interpretation is, and what values it realizes. The upshot seems to be that, given the constraints on discretionary decision-making by the Appellate Body, giving intention in general, and object and purpose in particular, a role in WTO law renders it an unavoidably anti-positivist enterprise. Or rather—putting the same point in different terms—the only sense that analytical positivism can make of references to intention in the context of WTO law is as exercises in extra-legal judicial discretion, of a kind that the WTO Appellate Body is both doctrinally and politically precluded from making.

A. From Positivist Moral Revision to Anti-Positivist Moral Reading

If law is only sources, as the positivist asserts, then WTO adjudicators must be doing something more than finding law, in ways incompatible with their doctrinal and institutional role. If, on the other hand, law is a more capacious phenomenon, then adjudication can in turn become a wholly legal enterprise, but at the expense of rejecting analytical positivism. Adopting an anti-positivist approach thus makes possible the reconciliation of linguistic uncertainty with the Dispute Settlement Understanding's demands for legal completeness. However, it also imposes quite specific requirements for the kinds of values and principles to which WTO law must have recourse.

Recall, the positivist sees the judge operating across distinct domains of law and non-law, applying distinct modes of reasoning in each. While in the domain of law, Hart's core, the judge applies rules to facts, and reaches conclusions that are prescribed by those rules. In the penumbra, where the law ‘runs out’, the judge adopts a non-legal mode of reasoning, in which she has recourse to moral, prudential or other values to reach a decision. There are reasons (of administrative efficiency, for example) why her extra-legal conclusions should be consistent with those prescribed by settled rules. Particularly where judicial decisions carry precedential weight, today's discretionary choices constitute tomorrow's legal rules. To the extent the judge can give an account of her reasoning and conclusions that renders law and non-law continuous, she provides a guide that can be readily applied by her successors.Footnote ¹³⁰ However, such concerns aside, the goal of the judge when not constrained by law will be simply to produce the ‘best’ decision possible, both as between the disputing parties, and having regard to the wider implications of her decision.Footnote ¹³¹

It is not necessary, in doing this, that the judge endorse or justify those legal rules that constrain her reasoning. The legal rules are fixed points, defining the area of discretion within which she has recourse to the moral and the prudential. If those rules are, or appear to her, irrational, unjust or ill advised, this need not pose a problem: she follows the rules to the extent the rules require, and seeks to realize justice (or whatever value she thinks adjudication should pursue) within the space left to her. A utilitarian judge, applying libertarian laws, might regret that she is in many cases constrained by rules she cannot endorse: but within the area of discretion afforded to her, she can choose the interpretations that best meet her utilitarian standards.Footnote ¹³²

For the anti-positivist, by contrast, law and morality are linked ‘all the way down’. There are few straightforward questions that are resolved through the mechanical application of rules found through scientific examination of social fact sources.Footnote ¹³³ Rather, legal interpretation is a matter of identifying those principles that best fit and justify the existing practice, and working from these principles both forwards, towards answers to hard cases, but also backwards, to potentially re-evaluate what initially seemed like relatively uncontroversial cases falling under established rules. The result is that the values invoked to support anti-positivist interpretation must not only suggest solutions to the novel questions, but also make sense of the mass of existing rules and principles, which the positivist regards as legally fixed points. The utilitarian anti-positivist cannot simply apply utility maximization in a legal system built on libertarian lines, because that approach cannot make sense of the existing practice: it would not constitute a plausible interpretation of the purpose of that practice. Rather, to make progress in resolving hard cases, she will need to relax her utilitarian commitments, and ask whether there is some principle or set of principles that can make sense of that existing practice, but that at the same time is able to show it in its best moral light.

Reasoning about law in the anti-positivist mode is therefore a two-way, iterative process. Rather than simply applying an independently motivated moral view to criticize the existing law and to provide answers to questions in those areas where the law is silent, we seek a set of principles that both fit and justify the existing practice. A view that fits the existing practice closely, but is morally defective, requires to be rejected in favour or one that fits marginally less well, but is morally preferable. But equally a view that is morally sound, but fails to make sense of large parts of the existing practice must be rejected in favour of one that, while otherwise morally inferior, does significantly better at making sense of the existing practice.Footnote ¹³⁴ (The parallels with the discussion of object and purpose above will hopefully be clear.)

The move to anti-positivism thus imposes new constraints on the values that can plausibly inform legal interpretation. Whether approached from a positivist or anti-positivist perspective, values, to be helpful, must suggest answers to questions adjudicators face: if they are indifferent between alternatives they offer no guidance. However, subject to this relatively minimal constraint, many values may be capable of guiding adjudicators who understand their role in positivist terms. For the anti-positivist, by contrast, relevant values must explain existing practice, as well as guiding innovation. The criterion applied here is a flexible one, as we seek to jointly optimize two aspects, fit and justification, which may lead us to accept a relatively lower value on one for a higher value for the other. However, it remains a significant constraint, which will rule out at least some views.

B. Anti-positivism, Economics and International Fairness

Anti-positivism emphasizes the role of values in law, suggesting that we can say very little about the law without engaging in straightforward moral reasoning. However, it also emphasizes the differences between moral reasoning in the presence and the absence of law. The next question is therefore what counts as good moral reasoning in the context of WTO law? Which schemes of values are plausible candidates to inform an anti-positivist reading of WTO law?

Amongst candidates that might otherwise be attractive, globally egalitarian views are the most obvious casualties of the requirement that principles fit existing practices. While there are egalitarian elements within the trade regime, including most obviously provision for Special and Differential Treatment for Developing Countries, these represent a relatively limited exception within a regime that otherwise places little emphasis on global economic equality.Footnote ¹³⁵ While arguments can be (and are) made for the pro-development potential of international trade liberalization, WTO law accepts too much international inequality, and too many of the institutional structures that maintain it, for this to plausibly constitute the organizing principle of the trade regime.Footnote ¹³⁶

Perhaps more surprisingly, international egoist views also struggle to inform an anti-positivist reading. By egoist here I mean views that see States as primarily self-interested and deny they owe natural duties of economic justice to those outside their borders.Footnote ¹³⁷ Economic duties to outsiders, on this view, are limited to those affirmatively accepted through agreement.Footnote ¹³⁸ Without more, such views are of little use in either positivist or anti-positivist terms. By emphasizing agreement as the exclusive source of obligation, they become circular—and in consequence mute—when invoked to guide interpretation of agreements. Beyond motivating anti-avoidance provisions,Footnote ¹³⁹ we do not help an interpreter by telling them to interpret an agreement by reference to the obligations that States have taken on under that agreement.Footnote ¹⁴⁰

Egoist views become more useful interpretive guides when they are complemented with an account of the interests of relevant agents. Homo economicus is an example of such an expanded egoist view, assuming that agents seek to advance their interests, and that those interests track either the aggregate economic welfare of national populations, or that of politically influential groups. Both provide plausible explanations of many of the key liberalization and anti-discrimination disciplines in the trade regime.Footnote ¹⁴¹ National welfare accounts also provide a plausible justification of those disciplines as advancing the material well-being of persons.

However, economically informed egoist views struggle to explain large sections of the trade regime. Most prominently, the trade remedies disciplines, which constitute some of the most litigated WTO agreements, make little sense in economic terms.Footnote ¹⁴² The subsidies rules discipline an overbroad category of measures, many of which are likely to advance both national and global welfare, while permitting recourse to countervailing duties which harm national welfare of both the country imposing them, and the country whose exports are targeted. The existence of safeguards rules is explicable in political economy terms, as facilitating a limited return to protectionism where political opposition would otherwise undermine the trade liberalization bargain; but nothing in this explanation makes sense of the specific rules governing when such measures are and are not permissible. And anti-dumping, despite its increasing use by members, is universally criticized by economists as incoherent ‘ordinary protection, albeit with a good public relations program’.Footnote ¹⁴³ As noted above, the requirement of fit is a flexible one, so the fact that a view cannot explain some elements of the trade regime need not be fatal. However, the significance of the trade remedies rules in WTO practice, and their political prominence in contemporary challenges to that regime, mean a view that cannot make sense of them is significantly undermined.

This imperfect fit between economic approaches and positive law is reflected in the varying extent to which the Appellate Body has felt able to engage in constructive interpretation across different areas of the trade regime.

The Appellate Body's clearest invocations of a substantive object and purpose to guide interpretation have been in the context of border liberalization and non-discrimination, where economic views fit best.Footnote ¹⁴⁴ Here, even absent clear textual support, it has been willing to identify the purpose of relevant provisions, and to interpret them having regard to that purpose, even to the point of disregarding significant differences in text.Footnote ¹⁴⁵ In the context of development provisions, object and purpose has also played a role, albeit in managing the tension between identified pro-development and anti-discrimination goals.Footnote ¹⁴⁶ The key point, for our purposes, is that these are areas where the relevant rules make economic sense, with the upshot that economic reasoning can help us understand what the rules are for, how they should be interpreted, and what they should admit or exclude. They are cases where the Appellate Body's reasoning seems most clearly to exhibit a direct engagement with underlying values, but where that reasoning has also generally been broadly accepted by members.

By contrast, in the context of trade remedies the invocation of object and purpose has been largely limited to effectiveness and anti-avoidance interpretations, excluding reference to the underlying goals of the relevant provisions.Footnote ¹⁴⁷ Indeed, in a number of cases the Appellate Body has criticized panels for engaging in such teleological interpretation in the context of trade remedies cases.Footnote ¹⁴⁸ While economically-informed constructive interpretation appears possible elsewhere in the trade regime, trade remedies interpretation has remained stubbornly formalist.Footnote ¹⁴⁹

The problem is one of fit rather than justification. Economic approaches emphasize a value, national economic welfare, that has obvious attractions from various perspectives. However, they cannot make sense of the most prominent features of the trade remedies rules, and so cannot guide or legitimize interpretation of those rules. The upshot has been that some of the most controversial and heavily criticized AB decisions have been in this area. These have included decisions requiring that members demonstrate both unforeseen circumstances, and an increase in imports that is ‘recent enough, sudden enough, sharp enough and significant enough’, before applying safeguards, making recourse to such measure significantly more difficult.Footnote ¹⁵⁰ They have also included a long line of cases on zeroing, a price comparison methodology in anti-dumping cases that has been consistently rejected by the Appellate Body over significant resistance from the United States in particular.Footnote ¹⁵¹ In the latter context, they have included cases disputing the requirement in Article 2.4 that members make a fair comparison between prices, as clear an invitation to normative reasoning as could be imagined.Footnote ¹⁵² In each case, economic approaches can justify imposing the relevant requirements and/or rejecting the criticized practices. However, the Appellate Body has not been able to invoke these, as those same approaches would condemn many other features of the relevant rules, making them implausible as standards for guiding and legitimizing interpretation. Instead, in the cases, we find a rigid textualism in the Appellate Body's self-presentation, which has in turn inspired significant criticism from unsuccessful parties.Footnote ¹⁵³

This points towards the need for an alternative approach that both fits and justifies the existing rules. If those rules are not explicable in exclusively economic terms, then the required approach must similarly go beyond narrow economic rationality. In particular, it will need to take seriously the idea of fairness that is a recurrent feature of discourse around trade remedies in particular, and trade regulation more generally. I have in other work sought to do just this, elaborating in some detail an account of global distributive justice that, I argue, can be applied to international trade regulation, and that generates a set of prescriptions sufficiently familiar to trade lawyers that they can serve to explain, justify and critique the positive law of the WTO. By beginning with the ideas of persons and peoples as free and equal, rational and reasonable, and emphasizing the need to justify coercion to those over whom it is exercised, I derive principles for the justification of various categories of measures, and show how these can be applied to make sense of familiar disciplines on border measures, discrimination, trade remedies and domestic regulation, and to improve in various ways on the explanations generated by standard economic approaches.Footnote ¹⁵⁴ Whether that particular view is endorsed or not, giving space to fairness requires going beyond narrowly egoistic views, to acknowledge and accommodate the possibility of natural duties of economic justice that are owed across borders. These may, but need not, be concerned with economic inequality: views focussed on the competitive process may in fact fit better than those concerned with outcomes.Footnote ¹⁵⁵ However, even such competitive fairness views, in so far as they acknowledge the claims of outsiders to particular conditions of competition, or exclude particular market interventions, not on grounds of efficiency but because they constitute wrongs to competitors, involve a significant ‘thickening’ of the moral backdrop against which international economic law is interpreted.