1. Introduction
Two World Trade Organization (WTO) agreements – the Agreement on the Application of Sanitary and Phytosanitary Measures (SPS)Footnote 1 and the Technical Barriers to Trade (TBT) AgreementFootnote 2 – are particularly pertinent to the development of food regulations. As food safety has been a leitmotif of international food policy in recent decades,Footnote 3 the SPS Agreement with its objective of ensuring human, plant, and animal health is naturally the focal point for assessing WTO compatibility with domestic regulations. Since the WTO's first SPS ruling against the EU's prohibition of meat containing growth hormones,Footnote 4 the detrimental impact of these rules on democratic policy-making has been vigorously expounded.Footnote 5 Many of the most controversial measures in this sensitive domain – hormones in meat, genetically modified organisms, cloned foods – invoke broader societal concerns, in addition to questions of food safety. These are often considered to be inadequately addressed by the SPS Agreement's science-based framework.Footnote 6 A conventional viewFootnote 7 is that the TBT Agreement, which does not impose the same rigorous standards as the SPS Agreement, offers greater ‘regulatory autonomy’Footnote 8 to WTO Members. Such reflections raise the intriguing prospect that regulators may seek to ‘forum shop’, that is, contrive to reconceive sensitive food regulations as TBT measures in order to circumvent the stricter SPS rules.Footnote 9 In light of recent case law that has enriched understanding of the TBT regime,Footnote 10 this article reassesses these assumptions, analysing the extent to which regulatory autonomy may be curtailed under both agreements. In so doing, it evaluates the possible merits for domestic regulators of ‘forum shopping’.
The article will consider three features of technical measures which shape the contours of regulatory autonomy under the WTO's technical agreements: the legitimacy of a regulatory objective (Section 3), the level of protection sought (Section 4), and the evidentiary burden for demonstrating the rationality of the chosen measures (Section 5). To help situate this discussion in a policy-making context, the article will make illustrative use of the dispute between the EU and US on antimicrobial treatment of poultry.Footnote 11 Section 2 will provide a brief background to this case, following some introductory comments on ‘forum shopping’ among WTO technical rules.
2. TBT as the easy option and the opportunities for ‘forum shopping’
The conventional view of SPS disciplines has been that they are ‘stricter’,Footnote 12 ‘rather rigid requirements’Footnote 13 compared to their ‘less stringent’Footnote 14 and ‘less rigorous’Footnote 15 TBT counterparts, the former offering ‘more limited flexibility’Footnote 16 to national regulators.Footnote 17 This perception is founded on two lines of argumentation. Firstly, the ‘legitimate objective’ that is required to justify any trade-restrictive TBT measure is not closely defined, thus ‘provid[ing] greater grounds for state regulatory intervention than does the SPS Agreement’.Footnote 18 If many policy objectives are legitimate, the argument would appear to run, a complainant WTO Member will face a stiffer task of demonstrating the non-legitimate nature of a respondent's measures. The second, and more frequently advanced, explanation is that the SPS Agreement establishes a greater evidentiary burden; food safety measures must be underpinned by ‘sufficient scientific evidence’Footnote 19 and ‘based on an assessment … of the risks’.Footnote 20 Defence of an SPS measure is thus more onerous than the ‘more liberal’ TBT Agreement that ‘permits a number of factors to be considered in the risk assessment’.Footnote 21 Commentators have therefore considered respondent states to have a fundamentally more difficult task in justifying measures under the SPS Agreement.Footnote 22 For example, the regulation of genetically modified foods ‘based on non-SPS objectives would have stronger grounds for being upheld as consistent with the EU's WTO obligations’.Footnote 23 Moreover, the divergence between TBT and SPS rules
not only imposes a higher justificatory burden on WTO Members wishing to adopt measures aimed at protecting themselves from health risks but, conversely, it also makes it easier for potential complainants to challenge them.Footnote 24
The perceived disparity between the constraints imposed by the two agreements naturally leads to strategic considerations. As Gruszczynski explains:
As a consequence, it may be tempting for an individual WTO Member to formulate its regulation in a way which will trigger the applicability of one or another agreement. From the perspective of the WTO, such forum shopping between agreements is highly undesirable as it would allow WTO Members to evade their obligations.Footnote 25
Before analysing whether the TBT Agreement does indeed provide a more promising path for those WTO Members looking to safeguard their preferred food measures, some preliminary observations on ‘forum shopping’Footnote 26 may be helpful.
Firstly, it is worth clarifying the situations in which ‘forum shopping’ could realistically occur. It is widely recognized that regulatory measures may reflect multiple policy objectives.Footnote 27 Where each goal can be clearly associated with individual parts of the measure, the viability of ‘forum shopping’ would seem limited. For example, where a food regulation establishes both safety limits for contaminants and ethical labelling requirements, a panel can assess each objective independently under respective SPS and TBT provisions.Footnote 28 The safety limits could not easily, or at least convincingly, be presented as non-safety measures falling outside the scope of the SPS Agreement. There are therefore inherent limits to the extent that ‘forum shopping’ could occur.
Forum shopping is rather confined to a particular situation where an individual measure is ‘indivisible’Footnote 29 into subparts, that is, with two alternative rationales, one SPS-based and one TBT-based. In Biotech, the Panel rejected complainants' arguments that a measure falling within the definition of Annex A (1) should be considered an SPS measure alone. Instead, it shared the EC's view that any additional non-SPS regulatory goal should be assessed under another relevant WTO Agreement. Thus, provided a measure has an ‘autonomous raison d’être’,Footnote 30 subsequent analyses of SPS and TBT objectives would be possible.Footnote 31 Where found in breach of the SPS Agreement, a WTO Member could presumably nevertheless maintain the substance of the measure, provided all references to the SPS raison d’être are removed. The TBT justification could therefore conceivably provide a safety net in cases where a WTO Member has reasons to doubt the WTO proof nature of its scientific evidence.Footnote 32
Secondly, the term ‘forum shopping’ has negative connotations, invoking a WTO Member seeking to avoid its obligations.Footnote 33 Yet it should not be assumed that parallel SPS and TBT rationales for measures that do not neatly conform to the distinct categories established by each agreement necessarily connote regulatory mischief. If we accept that the perceived distinctions in SPS rules and TBT rules reflect the deliberate intentions of negotiators,Footnote 34 we can have no objections as such to a multifaceted measure being maintained in accordance with the TBT Agreement, even where an SPS justification has been rejected and removed.Footnote 35 On the contrary, if WTO rules help root out illegitimate rationales, dual SPS/TBT scrutiny of a measure could arguably enhance regulatory transparency and coherency to the benefit of domestic stakeholders.Footnote 36 The criticism implicit to forum shopping must be reserved to those more exceptional circumstances where Members are deliberately seeking to dissemble questionable SPS measures.
Finally, while the potential for forum shopping merits exploration, the extent to which WTO considerations resonate in national policy-making, and therefore the likelihood of future shopping sprees, should not be overstated. While regulators are sensitive to potential complainants by third countries,Footnote 37 they typically need to accommodate numerous domestic interests that would severely complicate the task of developing a forum shopping strategy in the way contemplated.
2.1 The potential for forum shopping: antimicrobial treatment of poultry
To illustrate the issue at hand more clearly, consider the EU's measures relating to pathogen reduction treatments on poultry which have been brought before the WTO.Footnote 38 Applying antimicrobial rinses to carcasses is a well-established practice in US poultry processing as an effective way of reducing microbes in meat. This treatment, banned since 1997 in the EU,Footnote 39 is perceived by some to undermine hygiene practices, and deliver inferior food.Footnote 40 The European Parliament has railed against the US treatment of poultry viewed to be at odds with the EU's ‘total food chain approach’.Footnote 41 Notwithstanding the well-documented threat to public health of the transfer of antimicrobial resistance from animals to humans, the European Food Safety Authority (EFSA) failed to establish risks associated with specific antimicrobial substances used in the treatment of US poultry.Footnote 42 In June 2008, the European Commission consequently presented a limited legislative proposal to permit the use of four antimicrobial substances ‘according to very strict conditions and requirements’.Footnote 43 The Council rejected the substances, citing insufficient knowledge and the precautionary principle as the basis for its decision.Footnote 44
The Council elected for a food safety rationale of its measure, or in SPS terms, an Article 5.7 justification. Without entering into a detailed analysis of the scientific issues at hand, it is reasonable to contend that EFSA's favourable assessment of the four substances and therefore limited scientific justification for their prohibition exposed the EU to a potential WTO SPS challenge. An alternative Council rationale could have been ‘the prevention of deceptive practices' as defined in Article 2.2 of the TBT Agreement. After all, the European Parliament had emphasized the potential for consumers to be misled through chlorination which ‘might alter the appearance of meat to make it look fresher than it is’.Footnote 45 Notwithstanding the particularities of the case, if the general expectations of legal commentators are correct, this alternative TBT argumentation could leave the EU less vulnerable to a successful WTO challenge than its chosen SPS-based justification.Footnote 46 The remainder of this article explores these expectations, evaluating the respective approaches of the TBT and SPS Agreements to the various elements that determine a Member's regulatory autonomy.
3. The legitimacy of a regulatory objective
As noted above, it has been proposed that WTO Members are offered considerable flexibility in terms of defining their TBT objectives.Footnote 47 If the EU were to elaborate on an alternative TBT-based justification for its prohibition of antimicrobial treatment of poultry, to what level of scrutiny would this be submitted when compared with the SPS regime? There are two potential challenges for the EU: demonstrating that its public policy goal constitutes a ‘legitimate objective’ within the meaning of TBT Article 2.2 and persuading a panel that the claimed objective reflects the real purpose of its measure.
3.1 Regulatory goal as a legitimate objective
As the SPS Agreement elaborates on the public health exceptions established under GATT Article XX (b), the inherent legitimacy of trade-restrictive SPS measures that genuinely strive to meet these goals is not in question. TBT Article 2.2 also permits Members to restrict trade through technical regulations provided that they do so to ‘fulfil a legitimate objective’. TBT Article 2.2 sets out a non-exhaustive list of legitimate objectives which broadly echo some of the national policy areas outlined in GATT Article XX, including the prevention of deceptive practices and protection of human health.Footnote 48 In US–Tuna II, the AB clarified that TBT Article 2.2 is not a ‘closed list’, but rather ‘a reference point for which other objectives may be considered to be legitimate’.Footnote 49 Further guidance is offered by the recitals of the preamble of the Agreement and other covered WTO agreements.Footnote 50 In US–COOL, Canada argued that the Panel had erred in interpreting the US objective of providing consumer information on origin as a legitimate objective, claiming that any non-listed objective must be demonstrably linked to those listed.Footnote 51 Rejecting Canada's argument, the AB found that the US's objective ‘bears some relation to the objective of prevention of deceptive practices’.Footnote 52 The low threshold implied by ‘some relation’, combined with the non-exhaustive nature of the list as clarified in US–Tuna II, lends general support to the argument that WTO Members have considerable scope to establish their own legitimate objectives. Moreover, as the burden of proof lies on the complainant,Footnote 53 demonstrating the illegitimacy of an objective will be challenging.Footnote 54 Nevertheless, as Andenas and Zleptnig have underlined ‘the notion of legitimate objectives … implies that there are illegitimate objectives as well’.Footnote 55 By uniquely drawing on other WTO texts as guidance for legitimacy, the AB in US–Tuna II and US–COOL has arguably circumscribed the scope of the non-closed list for future cases. Notwithstanding the broad interpretation of legitimate objectives that emerges from recent jurisprudence,Footnote 56 a residual risk therefore remains that a presented TBT policy goal will be adjudged ‘illegitimate’ under Article 2.2.
3.1.1 Acceptability of characterization of measure
A greater challenge for the would-be WTO forum shopper is to credibly establish both the SPS and TBT objectives for a given measure. While there would seem little merit in challenging the characterization as an SPS measure,Footnote 57 a complainant may claim that the pronounced TBT objective is fabricated or mischaracterized. Importantly, in recent TBT disputes, the AB has been unwilling to accept the respondent's proposed objective at face value.Footnote 58 Rather, it ‘may take into account the texts of statutes, legislative history, and other evidence regarding the structure and operation of the measure’.Footnote 59 The unpredictable nature of this exercise was illustrated in US–COOL, where the Panel considered the US measure to be a dual one (aimed at provision of information and avoiding consumer confusion), only for the AB to largely discount the ‘additional, and secondary objective’.Footnote 60 This does not necessarily mean that a satisfactory TBT rationale for a vulnerable SPS measure cannot be developed.Footnote 61 Simply, given the scrutiny to which any measure's characterization would be subjected, and the number of elements that may be construed, any strategic intentions of this kind by national regulators would have to be carefully constructed from the outset of the legislative process.
In the light of the above considerations, it seems somewhat puzzling that commentators view the indeterminacy of the ‘legitimate objective’ under the TBT Agreement – the flexibility to choose public policy justifications – as a potential advantage for the implementing WTO Members. Let us consider the case of antimicrobial treatment of chicken. Certainly, the risk of consumer information and perception objectives being found illegitimate as such would seem marginal given the explicit recognition of ‘the prevention of deceptive practices’ in Article 2.2 and the ‘some relation’ standard established in US–COOL. Yet, while the legitimacy of measures seeking to ensure consumer protection from SPS risks is undisputed, alternative TBT rationales would, at the very least, be subjected to scrutiny. For example, a post-hoc TBT-oriented defence of EU measures based on preventing ‘deceptive practices’ could be undermined by the non-articulation of these concerns in the initial legislative proposal.
4. The level of protection sought
An EU measure that prohibits antimicrobial substances ensures a maximum level of protection, effectively zero risk both in terms of food safety and consumer deception. A second consideration when contemplating the strategic merits of forum shopping is whether the TBT and SPS Agreements offer comparable scope for a WTO Member to defend its chosen ‘level of protection’.Footnote 62 The AB has strongly expressed in the context of the GATT a ‘right to regulate’ as an ‘undisputed’Footnote 63 and a ‘fundamental principle’.Footnote 64 By extension, the right to choose one's own level of protection has, in the context of the recent EC–Seals dispute, been claimed by the EU to be a ‘well-established’Footnote 65 prerogative and was referred to in a submitted Amici Curiae Brief as ‘a fundamental component of WTO law’.Footnote 66 If WTO treatment of the level of protection is indeed as settled and homogenous as implied, forum shopping in this respect would not create any advantage. However, such a forthright articulation of the WTO's respect of domestic regulatory autonomy contrasts with an earlier analysis by Du. This author noted considerable inconsistency in the AB's deference towards Members’ regulatory sovereignty, characterized by a pragmatic weighing up of the importance of the value at stake against the impact on trade.Footnote 67 Before turning to this question of potential trade-offs between trade interests and regulatory autonomy in Section 4.2, let us first compare how the level of protection has been treated conceptually under the two agreements.
4.1 Conceptualization of chosen level of protection
In the context of the SPS Agreement, regulatory ambition finds expression through the frequently recurring ‘appropriate level of protection’.Footnote 68 In Australia–Salmon, the AB explained that the determination of the appropriate level of protection precedes and is independent of the establishment of the SPS measure.Footnote 69 Thus, even where a measure fails to achieve its pre-determined level in practice, this neither undermines the ‘prerogative’Footnote 70 of the member to determine its own level of protection nor lessens the challenging member's obligation to meet that intended level.Footnote 71 In other words, the SPS level of protection is an abstraction of regulatory intention separated from, and uncorrupted by, the actual performance of any given measure.Footnote 72 Only if the WTO Member does not fulfil its implicit obligation to be precise about the level of protection it considers appropriate,Footnote 73 could a panel be permitted to infer the level from the measure applied.Footnote 74 This approach has therefore justifiably been characterized as one that ‘demonstrated sensitivity and deference’Footnote 75 to a Member's regulatory intent.
In the TBT Agreement, the notion of level of protection is considerably less prominent than in SPS provisions, with its articulation only found explicitly in the sixth recital of the preamble, according to which a Member can establish measures ‘at the levels it considers appropriate’. In US–COOL, the parties reflected on the relationship between a TBT measure's ‘objective’ and ‘level of protection’. The EU not unreasonably suggested that the ‘level’ referred to in the sixth recital is analogous to the ‘appropriate level of protection’ identified in the SPS Agreement.Footnote 76 To bring clarity, the US, in basic agreement with the EU, proposed that this level of protection be distinguished by the term ‘“level of fulfilment (of the objective)”’.Footnote 77 The AB in US–Tuna appeared to follow this conceptualization. It held that the requirement that a regulation ‘fulfil an objective’ under Article 2.2 is a question of degree and not an expectation that a Member fully achieves its goals, thus setting apart the regulatory aim from the effects of the measure in place.Footnote 78 Having done so, and with the task of assessing both the contribution of the measure to the legitimate objectiveFootnote 79 and potentially the acceptability of alternative measures, it would logically seem incumbent upon the AB to identify with some precision the level of protection sought by the Member. This is recognized as a feasible exercise for a panel to undertake as ‘by preparing, adopting, and applying a measure in order to pursue a legitimate objective, [a Member] articulates either implicitly or explicitly the level at which it seeks to pursue that particular legitimate objective’.Footnote 80 The methodology outlined in US–Tuna could reasonably be read to parallel the established understanding of the SPS ‘appropriate level of protection’.Footnote 81
Nevertheless, in US–COOL a divergent TBT treatment of the level of protection emerges. Here, the panel distinguished the objective (provision of information) and the USA's regulatory ambition (‘as much clear and accurate origin information as possible’).Footnote 82 However, this SPS-style approach was condemned by the AB:
Neither Article 2.2 in particular, nor the TBT Agreement in general, requires that, in its examination of the objective pursued, a panel must discern or identify, in the abstract, the level at which a responding Member wishes or aims to achieve the objective. Rather, what a panel is required to do, under Article 2.2, is to assess the degree to which a Member's technical regulation … contributes to the legitimate objective pursued by the Member.Footnote 83
By blurring the distinction between objective and level of protection, the member's specific articulation of its level of protection, treated with such deference under the SPS Agreement, loses importance.Footnote 84
The distinctive treatment in identifying the level of protection has implications for a member contemplating forum shopping. Firstly, in the absence of any abstractly defined level of protection, the actual level of protection achieved by a TBT measure takes on far greater weight. While a Member's SPS level of protection is predefined and unaffected by the realities of implementation, the TBT level of protection can be inferred from the measure and its application.Footnote 85 Accordingly, in EC–Seals, the Panel drew its own conclusions about the level of protection of public morals from the EU Regulation (taking into account the perceived disregard for public morals in the exceptions to the banFootnote 86) rather than accepting the EU's claimed ‘high level of protection’.Footnote 87 This approach would controversially imply that, through the prism of TBT law, the more ineffective the Member's regulatory choices, the more constrained its right to regulatory autonomy. Secondly, by playing down the importance of the level of protection as an independent element of the analysis, the AB indirectly increases the significance of the precise characterization of the measure's objective. Yet, as noted above,Footnote 88 ‘what a Member seeks to achieve’Footnote 89 can be legitimately scrutinized and redefined by the panel. Thirdly, with limited weight accorded to a Member's stated level of protection, the information with which a Panel may discern not only a Member's objective, but its level of protection (to the extent this distinction is now relevant) and the contribution of the measure to the objective – namely ‘design, structure, and operation of the technical regulation, as well as from evidence relating to its application’Footnote 90 – is essentially the same.Footnote 91 This seemingly circular analysis would appear to give a panel considerable leeway in construing both the aims and effectiveness of TBT measures. In this context, a WTO Member can have no real confidence that specific regulatory intentions of its TBT measure will be correctly identified, let alone respected.
4.2 Balancing trade and non-trade objectives
Having identified the different conceptual approaches to the level of protection, let us return to the question of whether a WTO Member may, under either agreement, have to compromise its regulatory autonomy as a result of a measure's impact on trade.
From an SPS perspective, the answer appears relatively clear. Article 5.6 obliges Members to ‘ensure that … measures are not more trade restrictive than required to achieve their appropriate level of protection’. It intends to reduce to a minimum the obstacles created by the measure to trade, but ‘does not require any reductions in protection, no matter how costly in trade terms’.Footnote 92 This particular standard for determining the appropriate relationship between trade and non-trade objectives has been described as the ‘least trade-restrictive means’ test.Footnote 93 Does the TBT Agreement establish a similar standard or potentially one that gives greater sway to trade interests?
TBT Article 2.2 provides that ‘technical regulations shall not be more trade restrictive than necessary to fulfil a legitimate objective, taking account of the risks non-fulfilment would create’. Rather than preserving the Member's regulatory choices, the final sub-clause seemingly invites an evaluation of the implications of failing to respect them. Indeed, if non-fulfilment of a legitimate objective had not been considered an alternative and acceptable scenario by the drafters of the TBT Agreement, the clause would be redundant. On an ordinary reading of Article 2.2,Footnote 94 the TBT Agreement would rather appear to invite a proportionality test stricto sensu, that is, ‘a true weighing and balancing of competing objectives’.Footnote 95
In recent TBT cases, a number of WTO Members have advanced a proportionality stricto sensu reading of Article 2.2. In US–COOL, Canada considered that ‘a regulation that restricts trade will be more or less justifiable, depending in part on the risk of harm resulting from a failure to fulfil the objectives it pursues’.Footnote 96 In other words, a measure could be found to be non-compliant with WTO rules where its impact on trade outweighs the value of the regulatory goals the measure seeks to obtain. For example, in US–Cloves, under a proportionality test, the Panel would have been expected to weigh up the costs to health of Indonesia's proposed alternativesFootnote 97 to the US's ban of clove cigarettes against the projected trade gains. However, in its report, the Panel chose to interpret the word ‘necessary’ by drawing on the GATT Article XX necessity analysis performed by the AB in Brazil–Retreaded Tyres.Footnote 98 Here, the AB – citing earlier jurisprudence that emphasized the Member's right to achieve its desired level of protection – required the complainant to demonstrate that an alternative ‘would make an “equivalent” contribution to the achievement of the objective at the level sought’.Footnote 99 A similar approach was taken in US–Tuna, where the Panel determined that a less trade-restrictive measure would not be valid where it ‘would entail greater “risks of non-fulfilment”’.Footnote 100 In short, the Panels interpreted Article 2.2 to require a ‘least trade-restrictive means’ rather than a proportionality test stricto sensu.Footnote 101
As the notion of an unelected trade body weighing up and recalibrating the relative policy preferences of sovereign governments has long proved problematic to many commentators,Footnote 102 the Panel's ‘least trade-restrictive means’ approach has unsurprisingly found support.Footnote 103 Perhaps contentment at the reassertion of regulatory autonomy, heavily (and arguably overlyFootnote 104) reliant on the reference to ‘level of protection’ in the TBT preamble, has deflected from the shaky legal grounding for this finding. For to justify its eschewal of any judgement on the goals in question, the Panel in both US–Tuna II and US–Clove Cigarettes effectively had to rewrite Article 2.2. In their analyses, they substituted the phrase ‘the risk that non-fulfilment would create’ with ‘risks of non-fulfilment’,Footnote 105 thus transforming an evaluation of the necessity of the challenged measure into a narrower assessment of the effectiveness of alternatives to that measure.Footnote 106 Yet, this interpretation makes no sense if, as the AB subsequently determined, a measure can be deemed to breach Article 2.2 even where no alternative is proposed.Footnote 107 Such unexplained tampering with the text of 2.2Footnote 108 was arguably unnecessary. Given that the importance of the objectives pursued – health and protection of the environment – have solid support in WTO jurisprudence,Footnote 109 under a proportionality stricto sensu test, the Panel could have satisfactorily argued that even the smallest of shortfalls would have been unacceptable. However, the Panel's chosen methodology also precludes serious scrutiny of regulatory goals whose importance may be more tendentious. The AB did not directly address in either case the precise significance of the non-fulfilment provision.Footnote 110 Echoing the GATT Article XX cases, the AB did state in US–Tuna II that the burden is on the complainant to put forward an alternative measure that ‘makes an equivalent contribution to the relevant objective’Footnote 111 an approach subsequently reiterated in US–COOL.Footnote 112 The inference one could draw is that the AB too is consciously distancing itself from the proportionality stricto sensu standard. This standpoint would effectively align the SPS and TBT Agreements’ deference to regulatory intent.
On this reading of recent TBT jurisprudence, the forum-shopping regulator may feel relatively confident that, regardless of the severity of a TBT measure's trade impact, no compromise of her regulatory aims is required. A few nagging doubts may nevertheless remain. Firstly, views diverge as to whether GATT Article XX jurisprudence does exclude a proportionality stricto sensu test.Footnote 113 As has been extensively analysed elsewhere,Footnote 114 there are unresolved contradictions in the jurisprudence that descend from the first AB analysis of GATT Article XX necessity in Korea–Beef.Footnote 115 On the one hand, the AB establishes an analytical framework that announces a proportionality stricto sensu test.Footnote 116 On the other hand, the AB systematically fails to apply this methodology to the facts of the relevant cases, but instead, reiterates the WTO Members’ right to set and maintain its level of protection.Footnote 117 As the latter implies a ‘least trade-restrictive means’ test, the AB appears to be presenting two irreconcilable methods of assessing necessity. Even accepting that the trend in recent case law is deliberately deferential to regulatory preferences, it has been achieved in a way that has ‘done much violence to text, to precedent, and to legal logic’.Footnote 118 Given the ‘internally incoherent’Footnote 119 interpretation of Article XX and continued reference to, if not application of, the proportionality stricto sensu test, it cannot be excluded that future Panels treating Article 2.2 apply more vigorously and systematically the various elements of the test identified in Korea–Beef.Footnote 120
Secondly, accepting that the AB has, with regard to GATT Article XX, taken a clear and irreversible course away from proportionality stricto sensu and towards a ‘least trade-restrictive means’ test, the relevance of this jurisprudence is not settled. There are very evident textual and contextual divergences between GATT Article XX and TBT Article 2.2. As has been noted by Howse and Levy, the structure of Article 2.2 (an obligation not an exception justifying a violation), additional criterion (‘the risks non-fulfilment would create’), and different burden of proof (on the complainant not respondent) clearly set the two apart.Footnote 121 Indeed, as the US dryly pointed out in US–COOL, ‘the only similarity between the two texts is the use of the term “necessary”’.Footnote 122 Most crucially, while a measure that is indispensable to secure a regulatory goal is ‘certainly’ ‘necessary’Footnote 123 under GATT Article XX, with the basic legitimacy of a TBT objective open to question, such certainty cannot be ascribed to Article 2.2. This has not escaped the attention of the AB in the relevant TBT cases, where it has been notably circumspect in its reference to GATT Article XX. While ostensibly supporting a ‘least-trade restrictive means’ test, and recognizing the similarities between Article XX and Article 2.2,Footnote 124 the AB acknowledges that the ‘risks non-fulfilment would create’ provision ‘suggests a further element of weighing and balancing in the determination of whether the trade-restrictiveness of technical regulation is “necessary”…’.Footnote 125 In addition, the AB reads Article 2.2 to require a ‘relational analysis’Footnote 126 or ‘comparative analysis’Footnote 127 which could be read to imply a more critical and direct evaluation of the various factors concerned than the impressionistic ‘weighing and balancing’ associated with Article XX. As a result, if clear conclusions could be drawn from the necessity test emerging from GATT jurisprudence, these insights may prove limited in interpreting future TBT cases.
Thirdly, in US–COOL, the case where the legitimate objective (consumer information) was most contested, the AB did not exclude the possibility of a proportionality stricto sensu test. On the contrary, the AB appeared to take a significant step in precisely this direction, addressing all the factors relevant to such an analysis. In characterizing the US measure as providing ‘some contribution’ to its objective, identifying ‘a considerable degree of trade-restrictiveness’ and determining that the consequences of the non-fulfilment of the legitimate objective ‘would not be particularly grave’, the AB was creating a framework for establishing a potentially less effective, but less trade-restrictive alternative measure.Footnote 128 The Panel's limited findings with regard to alternatives did not permit completion of this analysis, but were such an outcome deemed impossible, the AB's specific identification of the above elements would have been irrelevant.
On the basis of recent jurisprudence, our hypothetical forum shopper may in general terms feel comfortable that consumer deception arguments related to chemically treated chicken would withstand a challenge under TBT Article 2.2.Footnote 129 Nevertheless, the above analysis exposes certain challenges for a member seeking to defend its chosen level of protection for a TBT measure. A WTO complainant who confronts an SPS antimicrobial resistance measure reflecting an explicit zero risk approach would have to demonstrate that any proposed less trade-restrictive alternative meets this stated level of protection.Footnote 130 By contrast, the AB's treatment of the level of protection in TBT cases opens up opportunities for a complaining Member to probe the defendant's regulatory aims. In the case of antimicrobial resistance, even an explicit claim of a zero risk approach to consumer deception is open to contestation and dependent on the panel's judgement of the measure as developed and applied. A complainant may thus potentially successfully propose an alternative measure providing protection lower than the claimed zero risk. Moreover, it may have to offer no alternative at all if the disproportionality of the effect on trade can be demonstrated. As a result, a promising TBT litigation strategy – casting doubt upon the EU's zero deception risk claim whilst documenting trade disruption – is available to WTO members, which simply would have no traction under the SPS Agreement.
5. Evidentiary burden: demonstrating the rationality of the chosen measures
In general terms, the evidence required by WTO Members in support of their SPS measures is undoubtedly more explicitly demanding than is the case for their TBT counterparts. The SPS Agreement places additional and independent substantive and procedural demands under Articles 2.2 and 5.1: sufficient scientific substantiation of any measure and evidence demonstrated through risk assessment. The burden that these scientific demands impose on domestic SPS regulators has been extensively assessed elsewhere.Footnote 131 Criticisms include the need for regulators to provide sufficiently specific evidence of the risk in questionFootnote 132 and their limited scope to act upon the theoretical uncertainty commonly associated with complex food–health relationships.Footnote 133 In the absence of such specific obligations, a WTO Member pursuing a TBT objective is seemingly granted greater freedom in both the manner and means of justifying regulatory choices.Footnote 134 Understandably, commentators concerned by the shackles placed upon WTO Members by the SPS regime have particularly emphasized this distinction between the two agreements.Footnote 135
Notwithstanding the clear additional evidential requirements set out in the SPS Agreement, there is a danger of overstating the significance of these distinctions and therefore the lure of forum shopping. There are a number of reasons why the TBT regime's seemingly lower evidential burden may not provide the greater regulatory autonomy anticipated. Firstly, TBT Article 2.2 and SPS Articles 2 and 5 essentially establish rationality tests for technical measures which are more similar than they may first appear. Both tests involve the relationship between the three same fundamental elements: the measure, the policy goal (general and specific level of protection), and supporting evidence in defence of the measure. The primary distinction between the two technical regimes lies in the emphasis placed on each of these three different elements. The SPS rationality standard focuses on the measure and evidence to determine whether there is a ‘rational or objective relationship between the SPS measure and the scientific evidence’.Footnote 136 The TBT Article 2.2 rationality standard brings the measure and its goals to the fore in order to ascertain ‘a genuine relationship of ends and means between the objective pursued and the measure at issue’.Footnote 137 The third element, not explicitly articulated in both cases, nevertheless remains important to establishing rationality.Footnote 138 It is these different emphases that create divergent thresholds for establishing rationality. In the case of the SPS Agreement, the stress placed on evidence creates what Scott describes as a ‘bright line’ of rationality, a minimum of scientific evidence required in order to distinguish between a rational real risk that merits a measure and a theoretical risk that does not.Footnote 139 Under the TBT Agreement, with means and ends accentuated, the minimum threshold required is a demonstration that the proposed measure provides some contribution to the identified policy.Footnote 140 In comparison, the TBT minimum rationality threshold is straightforward to meet, thus bolstering claims as to the burdensome nature of the SPS agreement.Footnote 141
Yet, the analysis should not end here. Another significant difference between the SPS and TBT tests lies in the demonstration of rationality acquired above the minimum evidential level. In the case of the SPS regime, as Scott puts it, ‘[a]bove the line anything would seem to go … there is ample room for Members to have regard to other kinds considerations in regulating risk, including public fears relating to the risk’.Footnote 142 Under the TBT Agreement, by contrast, demonstrating the basic rationality of the measure, while simple, may be of limited value if the defending Member cannot supply convincing evidence to justify the specific level of protection sought. The evidentiary burden for the respondent WTO Member is therefore demanding under either technical agreement. Whereas it has been argued that the SPS Agreement requires too much from Members below the ‘bottom-line’ of rationality and is ‘too lenient’ above,Footnote 143 the opposite may be true of the TBT Agreement. Demonstrating that the measure fulfils a TBT objective may be relatively easy, but beyond that minimum threshold, the evidence required to justify the genuine nature of the chosen level of protection may (especially given the contestable nature of the TBT level of protection as noted above) be excessively demanding.
A second possible misconception surrounds the scientific character of the evidence required for SPS measures, generally construed as a burden on domestic regulators. Even abstractly, the difference between the two regimes in this respect may be overstated given that under TBT Article 2.2 ‘available scientific and technical information’, while not mandatory, is pertinent to evaluation of TBT measures.Footnote 144 In practice, this has proved to be the case, most notably in cases concerning public health and the environment, where evaluation of scientific evidence has been extensive, and important to determining the legitimacy of challenged measures.Footnote 145 For measures responding to ethical and religious concerns, scientific justification is more questionable, potentially even ‘incoherent … because the measure is motivated by considerations that are not amenable to scientific analysis’.Footnote 146 Nevertheless, even in such circumstances, WTO Members have sought to justify measures through data collected according to a scientific methodology.Footnote 147 Furthermore, the scientific burden could be considered a double-edged sword; potentially an obstacle to regulatory autonomy, scientific evidence may also provide respondent Members in some situations with a powerful means for defending trade-distorting measures. The initial burden of proof remains on the complainant to demonstrate prima facie breaches of SPS and TBT obligations.Footnote 148 The steep scientific demands of the SPS Agreement are therefore in the first instance, at least theoretically,Footnote 149 a constraint on the challenging WTO Member which has to offer a science-based demonstration of the irrationality of the measure, one that will be discounted in the face of a plausible scientific rationale.Footnote 150 By contrast, in a TBT context, any justification of policy, scientific or otherwise, can be challenged by another Member using a wide range of counter-evidence including opinion polls,Footnote 151 information from websites,Footnote 152 and trade data.Footnote 153 Given the scientific shield that may be wielded in the SPS context, our hypothetical forum shopper may consider that these multiple avenues of argumentation offer the complaining Member greater rather than diminished potential for exposing the non-rationality of its TBT measure.Footnote 154
A third consideration is how the evidence presented in support of a measure will be treated under dispute settlement. Formally, the same standard of review applies under both technical agreements: panels are charged with undertaking an ‘objective assessment of the facts’ as set out in Article 11 of the Dispute Settlement Understanding (DSU),Footnote 155 defined in EC–Hormones as a standard between ‘total deference’ and ‘de novo review’.Footnote 156 However, panels in SPS cases have shown very little deference at all for a Member's demonstration of rationality.Footnote 157 For example, in Australia–Apples, the Panel (upheld by the AB) criticized Australia's treatment of probability and thus ‘came very close to saying that it disagreed with the risk assessment methodology [and] would have done the risk assessment differently’.Footnote 158 These intrusions have proved particularly controversial in the SPS domain, widely perceived to be dominated by a technical approach to risk which marginalizes other genuine non-expert understandings.Footnote 159 Although so far less obvious in jurisprudence, such concerns about the treatment of scientific inputs in complex policy areasFootnote 160 are undoubtedly extremely relevant to the TBT realm. Consider, for example, the Panel's response to the EU argument in EC–Seals that its citizens morally valued the protection of Inuit interests above their concerns relating to seal welfare. To substantiate its position, the European Union pointed to opinion polls demonstrating graver public concern for commercial hunts than those aimed at subsistence. However, the Panel rejected this claim finding that the evidence cited pointed to public acceptance of killing ‘animals’ for subsistence in general rather than ‘seals’ in particular.Footnote 161 Moreover, the quality of the evidence provided was deemed insufficient for the purposes of justifying the EU's measure, the Panel concluding that it had ‘limited probative value because the reliability the results of the surveys … has not been clearly demonstrated’.Footnote 162 The appropriateness of the Panel's approach, challenged by the EU, was not directly addressed by the AB.Footnote 163 However, such interpretations are reminiscent of the type of second-guessing of scientific evidence that have so troubled commentators of the SPS regime.Footnote 164 Any expectation of a less intrusive examination of a Member's underlying argumentation of TBT measures may therefore be misplaced. Indeed, if anything, a Member's uncertainty as to the level of scrutiny to which the factual claims will be subjected could be far greater in the latter context. Notwithstanding the criticisms noted above, traditional risk assessment does at least provide implementing Members with an established and accepted framework for constructing an adequate defence of the rationality of its measures.Footnote 165 This has been further supplemented through SPS jurisprudence with a number of reference points, for example, the validity of data ‘not susceptible to quantitative analysis’Footnote 166 or the right to develop a risk assessment approach reflecting the chosen level of protection.Footnote 167 By contrast, the panel in US–COOL in assessing the ‘genuine relationship’ between measures and regulatory objective, emphasized that ‘panels are not required to apply a particular methodology’.Footnote 168 A would-be forum shopper can only conjecture how any defence of its TBT arguments should be constructed.
Returning to our chicken example, the SPS Agreement clearly places an explicit evidentiary burden on the EU beyond the obligations set out in the TBT Agreement. Given the findings of EFSA on the safety of antimicrobial treatments, the EU may have significant doubts as to whether sufficient scientific evidence is available to meet the rationality threshold set by the SPS Agreement. From this perspective, exploring an alternative TBT justification for its measure has its obvious advantages. The EU may be fairly confident that it can demonstrate that prohibiting antimicrobial treatments makes ‘some contribution’ to a legitimate objective such as consumer perception of food. However, it is much less clear whether the EU would be able to establish the genuine nature of a zero risk level of protection or even how it would set about this task. At the very least, the willingness of the Panel in EC–Seals to question both the claimed level of protection and its evidentiary basis suggests that assumptions as to the light evidentiary burden of the TBT regime need to be rethought.
6. Conclusion
As this article has shown, the general view that the SPS Agreement is more stringent than the TBT Agreement, and thus a more difficult forum in which to defend regulatory autonomy, requires further reflection. The SPS Agreement does set minimum evidentiary demands with which Members may struggle to comply. Nevertheless, SPS jurisprudence has shown a level of deference to a Member's regulatory autonomy that is not paralleled by recent TBT case law. The former may therefore in some respects provide a more predictable and secure basis for defending food regulations. By contrast, many aspects of what a Member may aim to achieve through a TBT measure – the overall objective, the chosen level of protection, the deliberate prioritizing of this goal in spite of negative effects to trade – are open to reassessment and possible rejection by a panel. The TBT Agreement therefore gives rise to a different, but nonetheless significant, set of challenges for domestic regulators searching for a WTO-proof regulatory strategy. In this context, the attractiveness of forum shopping is doubtful. Food regulations that respond to genuine non-safety-related consumer concerns could potentially find refuge under the TBT Agreement where science-based arguments have been found to fall short. Yet, creative legal strategies which seek to circumvent SPS rules with elaborate TBT-oriented rationales seem unlikely to withstand the level of scrutiny that has typified recent TBT cases. For regulators acutely sensitive to potential third country litigation of SPS measures, forum shopping remains an option, but one that offers neither a panacea for defenders of regulatory autonomy, nor a substantial threat to SPS disciplines.
