The GATT and WTO Before the EU Courts: Judicial Avoidance Techniques or a Case Apart?

IV
The GATT and WTO Before the EU Courts: Judicial Avoidance Techniques or a Case Apart?

1. Introduction

This chapter assesses the EU Courts’ case law in relation to the legal effects accorded GATT and WTO norms in the EU legal order. Strikingly, particularly during the GATT era, this has been amongst the most criticized case law to have emanated from the Luxembourg Courts. This line of jurisprudence commenced with a seminal case on the 1947 GATT in which the ECJ first advanced its principled stance against reviewing EU measures vis-à-vis the GATT. And it has proceeded to reach a similar outcome with respect to its successor, the WTO and the Agreements therein.

The chapter is divided into two main sections. The first assesses the GATT-era case law and in doing so provides a corrective to the largely one-sided and revisionist accounts that have emerged; accounts which in particular fail to show a sufficiently nuanced understanding of the nature of the GATT or the implications of general domestic legal review in relation to its norms. The second main section provides an assessment of the WTO-era case law which has also witnessed the emergence of a principled stance against review. And it too has generated a predominantly critical academic response viewing this largely as nothing more than unprincipled recourse to avoidance techniques to immunize EU action from review in relation to legally binding WTO norms. The argumentation advanced herein will put forth an alternative account of the judicial stance and in doing so will also highlight other manifestations of judicially sanctioned legal impact that WTO norms are having in the EU legal order.

2. GATT Agreements Before the EU Courts

This section assesses the case law concerning GATT Agreements that arose before the EU Courts via two main subsections which consider, first, the principled judicial stance adopted vis-à-vis GATT Agreements and, secondly, the remaining manifestations of EU judicial application of GATT norms.

2.1 The reasoning in International Fruit: lasting first impressions

The ECJ first addressed itself expressly to the legal effect of GATT norms in a seminal judgment in 1972. The bulk of this section provides an assessment of the brief reasoning enunciated therein which was reiterated throughout the GATT-era case law.

2.1.1 The International Fruit judgment and its GATT-era progeny

In 1970 the Commission adopted Regulations protecting the EU market from certain agricultural imports which were challenged in relation to the GATT by importers in a direct action and a separate action in a Dutch court. In the direct action, the GATT arguments were not addressed, nor were any reasons for not doing so provided.¹ In any event, the alleged GATT incompatibility of the Regulations was soon before the ECJ again for the Dutch court put forth two questions: whether the validity of EU measures also refers, within the meaning of Article 267, to their validity under international law and, if so, whether the relevant Regulations were invalid as contrary to Article XI GATT.²

On the first question, the starting point was that preliminary rulings jurisdiction extends to all grounds capable of invalidating acts of EU institutions including ‘whether … they are contrary to a rule of international law’. Two criteria for validity review were announced: first, the EU must be bound by the relevant international law provision; and, secondly, for invalidity to be relied upon before a national court the said provision must also be capable of conferring rights. On being bound, the ECJ reasoned that by concluding the Treaty of Rome the Member States could not thereby ‘withdraw from their obligations to third countries’, indeed their desire to observe the GATT was made clear by Member State declarations when the EC Treaty was presented to the GATT’s Contracting Parties under Article XXIV GATT and also by Treaty provisions (particularly what became Arts 206 and 351(1) TFEU). Furthermore, the gradual transfer of trade policy powers to the EU showed the Member States’ wish to bind the EU by their GATT obligations, a transfer of powers put into concrete form in the GATT and recognized by the other Contracting Parties. In short, GATT bound the EU insofar as it assumed the powers previously exercised by the Member States in areas governed by the Agreement.

The International Fruit reasoning was reiterated, occasionally in condensed form, in a number of judgments from 1973 to 1995 whether in challenges to EU or domestic action, regardless of the GATT provision invoked,³ and even in a much-anticipated Member State annulment action which saw the GATT’s features held to preclude it from being taken into account in considering the legality of a Regulation.⁴ There were some exceptions to this reiteration of the International Fruit reasoning,⁵ the most interesting of which was arguably the Dürbeck ruling where the alleged GATT incompatibility of a Commission Regulation was rejected via reliance on information supplied by the Commission that a GATT panel had ‘criticized’ the EU, but found the measures not to infringe Article I or II of the GATT.⁶

2.1.2 Reappraising the International Fruit logic

The Court’s reasoning sinks below criticism.

(Conforti (1993: 30))

The reasoning first employed in International Fruit was the subject of frequent and widespread opprobrium.⁷ The focus of this criticism has been on the application of the twin-pronged test for review. Before turning to its application, it is useful to commence with a few comments as to the test itself.

2.1.2.1 The twin-prong test for review

Although in International Fruit the ECJ had been faced with an international agreement, it responded, as the Advocate General had before it, and in line with the question referred, in general terms, at least in the context of a preliminary ruling, as to the relationship between international law and EU law. It established that the EU’s secondary legislation can be reviewed vis-à-vis a rule of international law. However, the Court took the conclusion it reached for granted, and unlike its Advocate General, did not invoke any detailed argumentation in support. The persuasiveness of reasoning that the absence of any express limitations on grounds of validity under Article 267 amounts to an obligation to examine validity in relation to international law was unsurprisingly questioned as soon as the ink was dry on the judgment.⁸

The first prong has a seemingly innate logic. One need only pose the question why validity review should be possible in relation to provisions of international law by which the EU is not bound, to sense, at least prima facie, the elementary logic of this requirement. One could seek a domestic analogy and consider how controversial it would be for a domestic court to embark on direct validity review of a domestic measure in relation to international law provisions that are not binding on the State. Matters are, however, considerably more complex in the EU context because it is composed of States which have their own internationally binding obligations which can both pre-exist and post-date the Treaty of Rome. The crucial question becomes when is the EU so bound. Article 351(1) TFEU specifically speaks to the situation of pre-existing treaties concluded by the Member States with third parties. It provides that the rights and obligations arising from such agreements between one or more Member States and one or more third countries shall not be affected by the provisions of the Treaty. One might argue that it follows that the EU is bound by all prior agreements of the Member States.⁹ Alternatively, it can be argued that had the drafters intended all such agreements to be binding on the EU they could have stated it in the express terminology of Article 216(2). That Article 351(1) was not determinative of the matter was made apparent in the International Fruit decision; were it otherwise, the Court could simply have invoked Article 351 to support the contention that the GATT binds the EU rather than also underlining the Treaty-based transfer of competences to the EU.¹⁰

However, it is the second hurdle, requiring the international provision to be capable of conferring rights, that has generated greater comment.¹¹ In International Fruit the ECJ articulated no justification for this requirement. One justification would be to ensure a measure of symmetry between the approach to the relationship between EU law and the domestic legal order, and that between EU law and international norms.¹² Put simply, if a hurdle is in place for the enforceability of EU norms in the domestic legal order, such a hurdle should also apply to international norms. This logic took hold after International Fruit with respect to EU Agreements as Chapter II illustrated, and inevitably once such symmetry takes hold there, its applicability cannot be coherently contested for treaties not even concluded by the EU.

Nevertheless, a critique emerged. Schermers suggested that one would be inclined to reject this condition all together.¹³ The reasoning is essentially one of normative prescription, a vision of domestic courts—including the ECJ—faithfully applying all binding norms of international law. In the words of Schermers: ‘What is the meaning of a binding rule of international law if there is no international court which has (compulsory) jurisdiction and if national (and [EU]) courts refuse to apply it in the vast majority of cases? If ever we want to establish some form of an international legal order all courts should apply its binding legal rules.’¹⁴ In keeping with this vision, he expressed a preference for the ECJ ‘to apply all rules of international law binding upon the Communities and not merely the self-executing rules’. Clearly, simply as a matter of empirical observation, there is a great deal of meaning to binding rules of international law even absent strict judicial enforceability by international and national courts. International norms operate as constraints on the behaviour of nations, and crucially their legislative and administrative output, notwithstanding that domestic courts may not apply them.¹⁵ To argue that greater respect for particular international norms would result if there were an international court with compulsory jurisdiction and if national courts routinely apply such norms would appear axiomatic.¹⁶ In contrast, to imply that the meaning of an international norm is suspect when it is not going to be judicially applied is wholly unpersuasive.

As for the actual type of criterion being used, the Court looked to ‘the spirit, the general scheme and the terms’ of the GATT in its conferring rights determination. Criticism was forthcoming for applying the test employed in the relationship between EU and domestic law rather than the intent of the parties test.¹⁷ The concern appeared to be that a Van Gend en Loos-type test would result in fewer agreements qualifying as conferring rights when contrasted with a strict intent of the parties test.¹⁸ Were the ECJ to look for a parallel to the EU Treaty in all international agreements as a condition precedent to their capacity to be used to review EU or national measures, then few if any would not be found wanting. Other treaties lack even the limited textual evidence in favour of direct applicability marshalled in Van Gend en Loos. An international lawyer concerned with domestic judicial enforceability would inevitably be anxious that if this kind of evidence were looked for it would scupper the chances for greater enforcement of treaties through the medium of EU law. Such concerns proved misplaced as the ECJ’s jurisprudence on EU Agreements was to illustrate.

2.1.2.2 Applying the twin-pronged test to the GATT

The most frequently heard, and prima facie compelling, critique of the International Fruit reasoning was that it constituted a misrepresentation of the GATT. In the cases cited, the ECJ generally avoided engaging with the specific provisions with which EU or national measures were allegedly incompatible, it apparently sufficed to contend that the Agreement in which the provisions were housed, characterized as it is by ‘great flexibility’, detracts from the notion that the particular provision invoked could possibly confer rights. In effect, one could say that the appearance of provisions that are clear, precise, and unconditional¹⁹ is belied by the spirit, general scheme, and terms of the Agreement. But was the GATT really characterized by the great flexibility of its provisions? Many commentators over the years have taken issue with these observations. Essentially, the argument runs that this was a caricature of the GATT, the reality being that it was injected with a much greater degree of legalism and accordingly much less flexibility.²⁰ The assumption, not always expressly stated, is that indeed the GATT should in principle meet the criteria enunciated for review. The argument advanced, contrary to received wisdom,²¹ is that the ECJ was largely on firm ground in what is unquestionably a superficial assessment of the GATT.

The safeguard clause Some commentators intimate that there was something amiss in relying on the safeguard clause. Kuilwijk argued that it could only ‘be invoked under strictly defined circumstances’ and that ‘The clause itself is not flexible at all’.²² He continued: ‘By saying that the GATT clause is too flexible, the Court confuses GATT law with GATT practice.’²³ If substantiated, such assertions constitute a chink in reading Article XIX as an exemplar of the GATT’s flexibility. However, the ECJ did not actually assert that the clause was too flexible, but rather that the Agreement was characterized by the great flexibility of its provisions. The context of the argument being that this was one of the provisions evincing the flexibility inherent in the Agreement itself, which is difficult to contest given that the objective of a general safeguard clause is, indeed, to inject a measure of flexibility. In this respect, Kuilwijk’s attempt to contest the Court’s position misses its target. More importantly, it is rather telling that Kuilwijk was content to leave the reader with the impression that there was nothing flexible about Article XIX without any mention of the criteria for its applicability. A brief foray into the criteria provides a rather different picture.

To ‘escape’ GATT obligations, the language of Article XIX:1(a) required, somewhat oversimplified, imports in increased quantities causing or threatening serious injury and resulting from both unforeseen developments and GATT obligations. From the very earliest of days, these requirements were interpreted loosely.²⁴ A 1948 Working Party report concluded that the increased quantities of imports included a relative increase.²⁵ A few years later, a Working Party report took a very generous approach to the ‘unforeseen developments’ and to the causation requirement such that the author of the acknowledged GATT law ‘bible’,²⁶ queried: ‘If the nature of the development can be foreseen and yet the degree of its impact on imports is that which fulfils the “unforeseen development” prerequisite, can this not be the case in every substantial increase in imports?’²⁷ That Working Party’s determination of whether serious injury had been sustained or threatened, concluded that the Contracting Party invoking Article XIX was ‘entitled to the benefit of any reasonable doubt’ and that the complaining party had ‘failed to establish that no serious injury has been sustained or threatened’.²⁸ Jackson contended that the net result was ‘to render tariff concessions and other GATT obligations less stable’.²⁹ That the doyen of GATT lawyers put it in such stark terms renders it surprising that Kuilwijk referred to the strictly defined circumstances and the absence of flexibility of the safeguard clause.

Much was certainly happening outside the context of Article XIX inconsistent with the spirit, if not necessarily the letter, of the GATT, notably, the Voluntary Export Restraints (VERs) that had come into vogue by the early 1970s. In this sense, the safeguard clause could be viewed as too exacting in that were it less so recourse to VERs outside the strictures of Article XIX may not have been needed.³⁰ One can interpret Kuilwijk’s discussion as hinting in this direction. This would be the only way in which he can be interpreted to have marshalled support, if only implicitly, for the strictly defined circumstances and inflexibility of the safeguard clause. There is nothing inappropriate in pointing out that the consultation requirement of Article XIX, and the compensation or retaliatory responses to which it can give rise, have constrained its role as a general escape clause and that it had failed to satisfy the needs of the post-war trading regime. Kuilwijk, however, seized on the VERs as support for the notion that the Court made the most elementary of blunders by confusing GATT law with GATT practice. This allegation is immediately followed by a discussion of VERs, with the implication being that this may have been what the ECJ had in mind in terms of flexibility in that Article XIX was not providing an effective discipline because it was largely being ignored. But there is simply no opening in the mere seven sentences of paragraph 26 of International Fruit to suggest that any such thing was intended. Indeed, the ECJ confined itself to merely paraphrasing Article XIX, accurately when compared with the text.³¹ It is, thus, impossible to see from where any inkling that the ECJ confused Article XIX with the practice of employing VERs could be derived. And yet in the summary and conclusion to Kuilwijk’s work the point is put in even starker terms: ‘[the ECJ]… took the rather peculiar view that this neglect [of Article XIX] was a sign of GATT’s institutional weakness’.³² The Court, however, told us nothing of any neglect of Article XIX nor did it need to in order to make the rather axiomatic point that the capacity to resort to unilateral safeguard measures is evidence of the flexibility of the GATT.³³

This brings us to a related though more veiled criticism of reliance on the safeguard argument alluded to by Kuilwijk when he points out that unilateral safeguard measures are also allowed under the EC Treaty in certain circumstances.³⁴ The implication is that there was something amiss in invoking the Article XIX safeguard measure as exemplifying flexibility, given that the EU Treaty itself allowed for unilateral safeguard measures. The point was made more explicitly in a piece shortly after International Fruit to which Kuilwijk refers in which Waelbroeck essentially argues that although the EU system is more stringent, notably because of a posteriori control by the Commission and potentially also the Council, it is only a difference of degree not substance.³⁵ But as both acknowledged, the old Article 109(1)TEC was concerned with balance of payments difficulties and accordingly did not play the role of a general safeguard clause as did Article XIX. This makes any comparison misplaced because insofar as balance of payments problems were being addressed it would be via Article XII and Article XVIII:B and not the general safeguard regime of Article XIX.³⁶ Indeed, the more appropriate comparator would have been the old Article 226 which did perform the role of a general safeguard clause comparable to Article XIX. That this general safeguard clause lapsed following the culmination of the transitional period³⁷ is itself testament to one of the differences between the EU and GATT regimes. But even if it had remained in operation, it would have been of little use to those implying there was some sophistry in invoking the GATT’s general safeguard clause. To the contrary, it exemplifies a further sharp difference between the two, namely, that Article 226 did not allow for unilateral safeguard measures.³⁸ This was clearly, then, a qualitatively different regime from that put in place by Article XIX of GATT,³⁹ one which constituted a difference of substance and not merely of degree. Given the rigidity inherent in the old Article 226 procedure, policed as it was by the Commission, Council, and ECJ, it is unsurprising that the Court considered the GATT counterpart to be an indicator of its great flexibility.

Despite the conspicuous shortcomings in Kuilwijk’s attempt to counter the invocation of the safeguard clause, the conclusion reached was that ‘The alleged flexibility of the safeguard clause never should have prevented the Court from granting direct effect to precise and unconditional GATT rules.’⁴⁰ At the risk of pedantry, the way this and related sentences are formulated could mislead the reader unacquainted with the International Fruit reasoning.⁴¹ It bears repeating that the safeguard argument was not in and of itself the reason for the judicial unwillingness to use GATT provisions as a review criterion. Rather it formed part and parcel of a broader argument premised on the GATT being characterized by great flexibility such that the ECJ was in principle unwilling to employ it in this fashion. The capacity unilaterally to suspend GATT obligations, especially when read alongside the other arguments, does suggest that the ‘precise’ and ‘unconditional’ GATT rules to which Kuilwijk, amongst others,⁴² referred may not be quite so precise and unconditional after all. This was the unambiguous message of the Court.

Provisions conferring the possibility of derogation This brings us to the related reference to provisions conferring the possibility of derogation. It is related because Article XIX would also qualify as such a provision,⁴³ and because the same objective is served by referring to the derogation provisions and Article XIX, namely, to invoke evidence in support of the ‘great flexibility’ characterizing the GATT. However, one searches in vain for elaboration of the derogation provisions that support the flexibility allegation. It is not that a detailed consideration of all the specific derogations was called for, but merely to leave the issue in this fashion opened the decision to criticism. Accordingly, it is worth briefly touching upon some of the major clauses which might have been invoked in support.

The Court could, for example, have referred to the waiver provision in Article XXV(5) which certainly constitutes a derogation clause, indeed, by far the most powerful of them. It enables the Contracting Parties to take a decision waiving a GATT obligation of a Contracting Party. The procedural thresholds are high for it must be approved by a two-thirds majority and the majority must constitute more than half the Contracting Parties. The objective of such a provision is to inject flexibility and is thus perfectly consistent with the logic of the Court’s allegation, reiterating the Advocate General who explicitly invoked the waiver provision, as to flexibility. By the time of International Fruit some 67 waivers had been granted under Article XXV(5),⁴⁴ most controversially a wide-ranging waiver accorded the US in the agricultural domain which the Advocate General had been eager to bring to the Court’s attention.⁴⁵ Curiously, Kuilwijk asserted that ‘Article XXV:5 never has been a good example to prove the flexibility of GATT’.⁴⁶ The drafters, the negotiators, and the Contracting Parties who had been granted waivers would surely disagree. Neither is support to be found for such a proposition in the relevant scholarly literature. The Jackson treatise contended that ‘The legal technicalities of GATT provide a considerable measure of flexibility allowing “pragmatic” solutions to real concrete problems. The waiver provision … as well as the escape clause, are illustrations of this.’⁴⁷ Whilst Hudec argued that ‘Even in its most legalistic days the GATT was quite generous with waivers …’.⁴⁸

Another core provision attesting to flexibility was Article XXVIII which allows for the renegotiation of tariff concessions. This provision provided ‘a periodic right to withdraw concessions for any reason whatsoever’.⁴⁹ Here we thus find another provision providing clear testimony to the GATT’s flexibility.⁵⁰ There are also permissible derogations that apply specifically to the provision invoked in International Fruit itself (Art XI), notably the balance of payments exceptions in Article XII (and in Art XVIII:B with respect to developing countries) and an agricultural exception in Article XI:2(c). The latter was cited by both the Advocate General and the Commission in International Fruit; the Court was unlikely to be keen on making this provision an explicit tenet of its reasoning due to its case-specific applicability, that is, quantitative restrictions in the agricultural domain.⁵¹

An overall sense of why the ECJ was on safe ground in underlining the derogation provisions was provided by the leading GATT legal treatises at the time of International Fruit. The leading French treatise was published in 1968 and argued that: ‘Dans la technique du G.A.T.T., non seulement l’exception confirme la règle mais, bien plus, elle constitue la règle’.⁵² Flory’s account clearly supported the invocation of the derogation provisions as exemplifying the great flexibility of the GATT. The same can be said of the seminal GATT treatise published the following year. It is telling that Part III of this treatise was headed ‘Exceptions to GATT Obligations’ and contained well over 200 pages, whereas Part II headed ‘Law of GATT Obligations’, which also included some of what were considered to be the ‘minor exceptions’ such as Article XI:2, contained less than 350 pages. Jackson provided detailed consideration of the manifold derogations that can apply and rightly refers to them as having ‘provided the necessary flexibility without which the General Agreement might never have endured in the face of the pressures that have buffeted it.’⁵³ The key phrase of ‘flexibility’ again makes an appearance. The other major treatise was published in 1970 by an author adopting a markedly anti-legalist approach famously remarking that ‘ “Illegality” is an uncertain and ambiguous concept when applied to the General Agreement’.⁵⁴ It is thus submitted that the ECJ was on solid ground in underlining that the derogations were both flexible and conferred great flexibility on the system. The practice as it then existed, recognized in the work of the leading scholars in the field, was unmistakable testimony to this.

The dispute settlement system If the ECJ was on secure ground in pointing to the derogations and the safeguard clause as exemplars of the flexibility of the GATT system, what of the reference to the dispute settlement system? The Court, following the Advocate General, limited itself to paraphrasing Articles XXII and XXIII and while there was strictly nothing inaccurate in doing so, the system had moved much further than one could have anticipated on the basis of that bare text. The provisions did not enshrine a system of third party adjudication, an embryonic form of which took hold by the 1950s.⁵⁵ It is accordingly not surprising that the ECJ was criticized for apparently wilful misrepresentation of the dispute settlement system. On the other hand, by the early 1970s the dispute settlement system had been deep in a process of decline,⁵⁶ with the US and the EU being the main proponents of an anti-legalistic and non-confrontational GATT agenda.⁵⁷ Given it was not possible to foresee what the future would hold, to have painted a picture of the system as it had existed in the late 1950s, rudimentary as it may have been, would also have misrepresented the extant reality. More importantly, were a more nuanced picture of the dispute settlement process presented this would only have bolstered the International Fruit reasoning rather than detract from it. In a seminal 1971 article, the commentator most closely acquainted with the dispute settlement system discussed its flexibility:

proceedings have generally been conducted with as much room as possible for negotiated solutions. Formal legal rulings have seldom been sought before the parties are in deadlock. The rulings themselves have usually been expressed in terms that are soft and tentative and have almost invariably been accompanied by an invitation to continue negotiations. And, in fact, the parties have very often gone back to the bargaining table after a ruling and worked out a bilateral settlement.⁵⁸

Hudec brought out clearly the large discretion inherent in all stages of the process giving ample room for (frequently taken) negotiated settlements. The Jackson treatise also underlined that throughout the panel procedure an attempt was made to achieve conciliation.⁵⁹ The system thus encouraged the parties to the dispute to produce as far as possible a result palatable to them given the economic and political exigencies of the time. If indeed judicial resolution of disputes can ever be seen simply as the mechanical application of a legal rule to a particular factual scenario, one can confidently state that such a vision did not even come close to accurately representing GATT reality at the time of International Fruit. Indeed, panels drawn from diplomats rather than lawyers, hence Hudec’s coinage of ‘a diplomat’s jurisprudence’,⁶⁰ was still dominant in the 1980s. It was not until the early 1980s that an Office of Legal Affairs was created for the GATT which generated a marked improvement in the legal quality of panel reports,⁶¹ but the willingness of the losing party to block reports—the system operated with a consensus requirement at most stages of the process including adoption⁶²– became increasingly common with the growth in panel activity since the 1980s.⁶³ The remarkable evolution of the GATT dispute settlement system since the initial International Fruit judgment cannot be gainsaid and perhaps many would have found the initial International Fruit reasoning, which merely paraphrased GATT provisions, more palatable were it to have come with recognition of the increasingly legalized nature of the dispute settlement process. Ultimately, however, the critics were not merely seeking judicial recognition of the evolution of the dispute settlement system, but a different substantive outcome, one in which the GATT was accepted as a general review criterion.

The preceding subsection was not intended to suggest that the seven paragraphs in International Fruit devoted to assessing the GATT, much of which is mere selective paraphrasing and was articulated yet more concisely in later judgments, actually sheds light on the complex system that was the GATT. International Fruit and its offspring were far from a model of clarity and carefully reasoned analysis. A finer grasp of the issues involved can be gleaned from the Advocates General Opinions and the reports for the hearings of the relevant cases. But much reading between the lines is needed. ECJ judgments, as has rightly been noted, ‘may well be satisfactory but readers must often supply for themselves the reasons, political or juridical, which grounded the decision.’⁶⁴ International Fruit, and its progeny, exemplified this proposition. Whilst we do get a strong sense of the reasoning which grounded the decision, fundamentally that we are not dealing with hard and fast legal rules, as attested to by the assortment of applicable derogations and the nature of the dispute procedures, it is nonetheless essential for this to be supplemented by an understanding of some of the more openly political considerations lurking in the background.

The most significant of such considerations, that has rarely received open recognition in the largely one-sided critiques, is the place of agriculture in the GATT. Whilst it formally always applied to agricultural trade, the reality was that this sector had largely escaped GATT discipline.⁶⁵ The US was the first Contracting Party to make it clear that it was not willing to see its agricultural policies constrained.⁶⁶ A 1951 congressional amendment obliged the President to implement ad valorem fees up to 50 per cent or such quantitative limitations as necessary in order to prevent agricultural imports materially interfering with various domestic farm programmes. This led in the mid-1950s to the US receiving a GATT waiver, unlimited in time, from its obligations under the most-favoured-nation clause of Article I and the quantitative restrictions interdiction in Article XI.⁶⁷ Essentially, this was forced through the GATT machinery, all parties fully aware that the granting of a waiver was not going to determine whether the dictates of the amendment to an Act of Congress would be followed.⁶⁸

By the end of the 1950s most Contracting Parties were likewise unwilling to see their domestic agricultural policies hemmed in by GATT rules. The GATT certainly granted greater flexibility for agricultural products:⁶⁹ it remained, however, overly ambitious and the developed countries promoting greater self-sufficiency found themselves producing surpluses which not only led them to restrict agricultural imports, but also to ‘dump’ the excess on the international market with additional trade-distorting effects. The arrival of the EEC with its attachment to a burgeoning protectionist agricultural agenda consolidated the de facto demise of the GATT’s rules as they pertained to agriculture.⁷⁰ Agricultural protectionism that flew in the face of the letter of the GATT was yet more rife by the time of the 1994 Bananas ruling (C-280/93), when the Council and Commission strenuously contested the GATT’s capacity to form a general review criterion, than at the time of International Fruit itself. It is telling that these two seminal cases, coming at different points in the GATT’s evolution, concerned challenges to EU agricultural measures. In effect these challenges were seeking to have GATT agricultural rules resurrected vis-à-vis the EU and policed by the EU Courts. This would be a one-sided resurrection of the rules, for the protectionist agenda of other powerful Contracting Parties, most prominently the US, would remain and with no need to fear domestic judicial challenge. The ECJ was certainly aware of this context and in such circumstances the rejection of the GATT as a review criterion could be expected. This, however, may be the very reason why some have criticized the ECJ, seeing it as having ‘needlessly weakened the GATT’.⁷¹ The logic seemingly being that requiring strict adherence to the rules, which the GATT system itself was incapable of generating and, indeed, can be viewed as having de facto accommodated non-compliance, would have bolstered the GATT.⁷² It is difficult to believe, however, that the EU’s political institutions and the Member States would have been willing to countenance such onesided enforcement of the GATT bargain. Perhaps a waiver for the EU’s protectionist agricultural agenda could have been sought which would at least provide some semblance of GATT conformity. In any event, it had long been emphasized that the size and weight of the project could have justified renegotiating the GATT rules to produce those that would have been negotiated in 1947 had the EU existed and that, in fact, this is essentially what happened, albeit not openly and directly.⁷³

The de facto status of GATT discipline in relation to agricultural trade is also testimony to a further distinction between the GATT and the other trade-related EU Agreements, considered in Chapter III, and thus weakening the oft-heard double standards critique.⁷⁴ After all, with the other EU Agreements the EU negotiators were able formally to accommodate the increasingly protectionist Common Agricultural Policy. Even if judicial enforcement of such Agreements were to be one-sided this would not constitute a threat to the EU’s emerging agricultural policy. Independently of this distinction, the other Agreements were not characterized by the same level of flexibility that had come to characterize the GATT. Whilst it can be conceded that satisfactory reasons for the different treatment meted out to different Agreements was not provided,⁷⁵ it should equally be acknowledged that the reasons are not hard to find.⁷⁶

One final significant critique concerns the relevance of the International Fruit reasoning where the challenge is brought by a Member State as many had assumed a different outcome would prevail.⁷⁷ This proved to be mistaken, as illustrated with the Germany v Council ruling where the reasoning was considered equally applicable. This logic generated considerable criticism,⁷⁸ but as Eeckhout has argued, the success of a challenge should not depend on the route by which it is brought before the EU Courts or who is bringing the challenge.⁷⁹ If the clear implication of the International Fruit reasoning was that GATT rules were not sufficiently unconditional for domestic judicial enforcement then the fact that the EU measure is being challenged via a different procedure and with a different challenger should be irrelevant. A closely related point is that the GATT’s dispute settlement procedures would be bypassed regardless of the procedure by which the challenge is brought and regardless of the litigant. The EU, if challenged, might defend a particular position in dispute settlement and this is imperilled if the ECJ reviews EU measures vis-à-vis its own interpretation of GATT rules. Admittedly, in the Bananas case a panel had already found the measures in breach of GATT rules, but the EU vetoed adoption of the ruling.⁸⁰ Some might suggest that it is the ECJ’s place to make up for such shortcomings in the GATT’s dispute settlement system in the interests of ensuring a more rules-based international order. However, that non-adopted panel ruling triggered negotiations within the GATT framework between the EU and various affected Contracting Parties concluded after the Bananas judgment.⁸¹

The aforementioned questions are just a handful of those that could be asked about the potential impact of judicial enforceability of GATT law. It is also essential to take a step back and consider the broader constitutional setting. The foundational court-created concepts of the EU legal order—direct effect and supremacy—were not yet a decade old at the time of International Fruit, and resistance in various national legal orders had already made its presence emphatically felt. At such a critical juncture in the acceptance of the foundational tenets of the EU legal order, it seems at least arguable that to accord the GATT the status of a review criterion, when one could reasonably infer that the structure of the system was unlikely to lead the courts of the automatic incorporation Member States to reach such a conclusion for domestic purposes, would have been a very controversial result indeed.⁸² Such a result could justifiably have been interpreted as binding the Member States beyond that which they had bargained for in acceding to the GATT and potentially to GATT interpretations which would not accord with the results that might be achieved through its dispute settlement procedures. This is accentuated when one considers that the European project resulted in the Member States giving up their independent regulatory capacity in the external commercial policy sphere. For the result would be that, not only had they lost such autonomous regulatory capacity, but its collective use through the EU regulatory framework would be subject to validity review as against ECJ interpretations of the dictates of GATT ‘law’, which the Member States would be unlikely to have contemplated in transferring the powers. In other words, Member States where the direct effect of the GATT had been rejected, or would be likely to be rejected,⁸³ would to all intents and purposes find that it could be judicially enforced vis-à-vis their external commercial policy, now in an EU guise, through the back door and potentially in their own courts. There seems no doubt of the court’s awareness of the implications at stake,⁸⁴ and it is noteworthy that a judge who sat for the string of 1983 judgments rejecting the direct effect of GATT provisions, and EU-concluded GATT tariff protocols, stated extra-judicially that ‘the Court must reflect upon the necessary acceptance of its judgments’.⁸⁵

2.2 Exceptions to the non-judicial applicability of the GATT

That the EU Courts were not in principle willing to allow GATT norms to be used as a direct review criterion, did not mean that such norms were not judicially applicable at all. Three important ways in which they were cognizable before EU Courts, with varying degrees of legal consequence, were recognized.

2.2.1 The consistent interpretation doctrine

GATT norms could be employed as an aid to the interpretation of EU legislation via the doctrine usually known as consistent interpretation. Application of this interpretive principle first appeared shortly before the International Fruit ruling where tariff agreements concluded within the GATT framework were drawn upon in interpreting the Common Customs Tariff.⁸⁶ Reliance or attempted reliance on GATT norms when interpreting EU legislation was repeated in a growing number of rulings,⁸⁷ but it was not until 1996 that some explicit justification for this doctrine was offered.⁸⁸ In that case it was held that the primacy of EU Agreements over EU secondary legislation, the first such express judicial affirmation, meant that the latter was, as far as possible, to be interpreted consistently with the former. The logic is clear and familiar from domestic courts when dealing with international obligations: where possible superior norms are to shape the interpretation of inferior norms in order that breach of the former, and thus international responsibility, can be avoided. It, thus, constitutes the mirror image of the obligation of consistent interpretation upon domestic courts and authorities in the context of Directives, which ultimately has its grounding in the primacy of EU law of which the cooperation duty in Article 4(3) TEU can be viewed as a manifestation.⁸⁹

2.2.2 The Fediol and Nakajima principles

Consistent interpretation illustrated that GATT norms could indirectly influence the interpretation of EU and potentially domestic legislation,⁹⁰ but their capacity to have a more potent legal impact first became apparent with the 1989 Fediol judgment.⁹¹ This case concerned a Regulation allowing EU producers to lodge complaints as to ‘illicit commercial practices’ by third countries.⁹² A trade federation challenged a Commission Decision adopted under that Regulation. The ECJ reiterated the jurisprudence commencing with International Fruit, but held that this did not mean individuals could not rely on the GATT to obtain a ruling on whether the conduct in the lodged complaint constituted an ‘illicit commercial practice’.⁹³ The Regulation defined ‘illicit commercial practices’ by reference to their incompatibility with international law, and referred expressly to the GATT in the preamble, and on this basis the ECJ concluded that GATT provisions formed part of the rules to which the Regulation referred. The Court reiterated the GATT’s broad flexibility whilst holding that this did not prevent it from interpreting its rules with reference to a specific case to establish the compatibility of specific commercial practices.⁹⁴

The obvious question that Fediol raised was whether it could be reconciled with the general rejection of the GATT as a review criterion. Eeckhout appeared to see no inconsistency. For him, judicial review of the Commission’s examination of the legality of third country commercial practices, an examination required by the Regulation, would not be meaningful if GATT provisions could not be looked at.⁹⁵ That is certainly so, but it does still clearly require the judicial interpretation of GATT provisions. And in Fediol itself this included Article XI, the very provision the ECJ had been unwilling to engage with in International Fruit. Clearly, then, there was not some inherent factor precluding EU judicial interpretations of GATT provisions, but then that much was already apparent by virtue of the doctrine of consistent interpretation.