1. Introduction
Consumer preferences and Article III
This paper focuses on how consumer preferences have been interpreted and applied in deliberation of WTO disputes under the GATT National Treatment Principle (‘Article III’). The entry point for the analysis is the concept of ‘like’ products. This concept, which recurs throughout the GATT and WTO Agreements, is an influential aspect of WTO law in general (Choi, Reference Choi2003). The consideration of what constitutes ‘like’ products has differed both between, and within, relevant Articles, and the slightly varied phrasing of these Articles also influences their interpretation. However, the basic concept is that less favourable treatment should not be awarded to ‘like’ products, regardless of their origin.
Article III generally stipulates that, with regard to regulations, imported products must be treated the same way as domestic products. The majority of Article III disputes fall under either Article III(2), which specifically addresses taxation, or Article III(4), which addresses other forms of regulation. The first step in determining whether a national tax or regulation discriminates against imported products, thus constituting an Article III violation, is to determine whether two products, domestic (favoured) and imported (unfavoured), are ‘like’. If they are, the dispute settlement bodies, which consist of the Panel and the Appellate Body (‘AB’), then consider whether the measure is discriminatory. The interpretation of the term ‘like’ thus outlines the boundary of Article III's jurisdiction.
When establishing whether products are ‘like’, consumer preferences are an important consideration, for two reasons. First, as will be further discussed, Article III jurisprudence demonstrates that ‘likeness’ is determined in the marketplace, based upon consumers' perceptions of whether goods are in a competitive relationship with one another. Second, ‘consumer tastes and habits’ form one of four traditional criteria, drawn from a 1970 GATT Border Tax Adjustment Working Party report (‘BTA criteria’), utilized to determine whether products are ‘like’. These include: ‘the product's end-uses in a given market; consumers’ tastes and habits, which change from country to country; the product's properties, nature and quality' (GATT, 1970: 3) [Emphasis added].Footnote 1 This paper uses the term ‘consumer preferences’, rather than simply the GATT criterion's ‘consumer tastes and habits’, to avoid confusion, as the influence of consumer preferences goes beyond the application of this criterion, and includes a broader consideration of product competitiveness.
The BTA criteria are meant to define what ‘like’ means, and product competitiveness is an essential component of ‘likeness’. However, it has not been clearly established that the BTA criteria are intended to add somehow up to a definition of competitiveness. The two enquiries seem to operate in parallel, sometimes overlapping.Footnote 2 Overall, the approach varies widely, and there is not much consistency from dispute to dispute, or generalized guidance.
In the GATT Agreement, as well as the other WTO Agreements, there are very few explicit mentions of consumers, and none in Article III.Footnote 3 Though it is not a mandatory aspect of Article III deliberation, of the 20 disputes that have fallen under Article III since the WTO's formation in 1995, 14 cite the BTA criteria,Footnote 4 and more than half utilize consumer preferences in coming to their conclusions, often through analysis of whether products are competitive with one another.Footnote 5
Recently, it has been suggested that consumer preferences could play a special role in WTO disputes that have to do with public policy regulations, specifically with regard to emerging regulations that respond to climate change (Vranes, Reference Vranes, Herrman and Terhechte2010: 11). This is because products otherwise ‘like’ from a WTO perspective might seem distinct from the perspective of consumers. For example, consumers might prefer goods that have been produced with low carbon emissions, and see them as different from conventional goods. In much the same way, consumer preferences would also be relevant in a potential dispute on biofuels sustainability criteria, a controversial set of EC regulations discussed in this paper. Yet, in spite of ongoing controversy about the treatment of environmental regulation in the WTO, and scrutiny of the Appellate Body approach to such disputes, not much targeted analysis has been undertaken of how consumer preferences have influenced past disputes on ethical public policy regulations, and how they should be interpreted in the future. Indeed, much of the existing literature has analysed, from an economic perspective, the best methodologies for determining ‘like’ products (cf. Berg, Reference Berg1996; Choi, Reference Choi2003). An economic approach to ‘like’ products faces particular challenges when confronted by such public policy regulations. This paper will examine these challenges (from a legal, rather than economic, perspective) by contrasting the economic approach taken in Japan–Alcohol with the approach taken in EC–Asbestos, a dispute that raised more difficult ethical questions.
While the focus here is on Article III, whose disputes include an abundance of jurisprudence about consumer preferences, a parallel analysis might be undertaken of disputes under other WTO Agreements, which also contain the concept of ‘like products’, for example within the Agreement on Technical Barriers to Trade (‘TBT Agreement’), or the Most Favoured Nation Principle (‘MFN Principle’).
Approaches to consumer preferences
This paper focuses on two Article III disputes: Japan–Taxes on Alcoholic Beverages Footnote 6 and European Communities–Measures Affecting Asbestos and Asbestos-Containing Products Footnote 7 Others have also included substantial analysis of consumer preferences. However, the selected disputes, as well as providing important benchmarks in Article III jurisprudence, also well illustrate a spectrum of interpretive possibility.
In Japan–Alcohol, the dispute settlement bodies utilized what this paper calls an objective approach to consumer preference.Footnote 8 This approach emphasized the importance of ‘objective’ BTA criteria, such as how products appeared (‘the product's properties, nature and quality’) and were used (‘the product's end-uses in a given market’), as opposed to ‘subjective’ consumer preferences, such as traditional patterns of consumption.
Further, the dispute emphasized the importance of a competitive market relationship between products, from a consumer perspective, as a way to determine their ‘likeness’ (Bronckers and McNelis, Reference Bronckers, McNelis and Cottier2000: 347). To establish competitiveness, the objective analysis of consumer preferences was approached economically. For example, the parties utilized elasticity of substitution studies to measure whether consumers would switch to one disputed alcoholic beverage if the price of the other rose.
In EC–Asbestos, the dispute settlement bodies faced a very different set of circumstances. None of the disputing parties had submitted any evidence, market-based or otherwise, about consumer preferences. In part, this reflected the fact that, rather than many individuals, asbestos was purchased by few, industrial consumers, which made their preferences more difficult to quantify. Further, the measure in question was a public health regulation, which raised ethical concerns for consumers. The Panel applied the same ‘objective’ approach, concluding that consumer preference, as it could not be measured, was not a relevant consideration.
The AB disagreed, asserting that consumers would not prefer asbestos. This conclusion was not based in evidence, but common sense, grounded in international standards. Though their reasoning represented a departure from the objective approach described above, they disingenuously presented it as objective. They did so by basing their argument on likely quantitative impacts in the market (a negative competition effect), and the objective BTA criterion of physical characteristics of the products in question, which then affected consumer preferences.
EC–Asbestos brings up a fundamental question: should this approach be generalized, so that the AB can interpret what constitutes consumer preference when dealing with normative regulations that limit, or eliminate, the ability of consumers to choose, and thus have their preferences measured? If this approach is not permissible, should consumer preferences just be disregarded, despite their relevance to such an important public policy decision?
Preliminary analysis
Considering the different strategies employed in these two disputes, it is possible to simply assert that the interpretation of consumer preferences has been inconsistent, and even illogical. However, it is perhaps more useful to examine the causes behind these discrepancies, and what they reveal about the challenges of capturing consumer preferences more generally. In EC–Asbestos, the AB evaluated consumer preferences, and product ‘likeness’, without reference to evidence of competitiveness between the products. For this to occur, the public policy at issue needed to represent a fairly unambiguous consensual societal norm, such as human health, as well as a situation where objective evidence about consumer preferences was not available, or difficult to obtain. In such a situation, it became politically difficult to keep consumer preference, and the associated public policy goal of supporting human health, out of Article III, due to the potential for public backlash.
The dispute settlement bodies' approach to consumer preferences reflects oft-discussed strategies of Article III interpretation employed more broadly in each dispute. Thus, considering the influence of consumer preferences involves grappling with some of the same controversies that they raised. Principal among these is the question of whether, under Article III, product ‘likeness’ is determined solely by product competitiveness. On the other hand, can public policy goals that inform how the product is regulated make one product unlike another? In other words, to what extent should the types of concerns addressed under Article XX, the General Exception, be considered within the ‘likeness’ test of Article III?
In this sense, consumer preferences raise some of the same questions as the ‘aim-and-effect’ test, an approach to interpreting product ‘likeness’ whose rejection forms an interrelated theme of the disputes in question. This approach was rejected on the basis that it unduly empowered a government's unprovable claim: that its aim in setting a regulation was not to discriminate. Consumer preferences would seem not to have much in common with this approach, as they represent the collective behaviours of individuals, rather than the discretion of a single government. Also, consumer preferences are an objectively measured entity, rather than an unprovable claim.
Or are they? In EC–Asbestos, consumer preferences had two important features in common with ‘aim-and-effect’. First, they provided a tool (though used rather disingenuously) for broadening the scope of what constituted ‘like’ products, to include public policy considerations. Second, they stood in for a discretionary approach to determining the validity of a regulation, though employed by the AB, rather than a national government.
Perhaps one reason this occurred is that, while ‘aim and effect’ has been rejected, consumer preferences still play an important role in Article III dispute deliberation. Future public policy disputes that fall under Article III will likely provoke the same problems and concerns, even if not so strongly as in EC–Asbestos. The potential for correlation between products with import bans or restrictions, products that generate strong ethical preferences from consumers, and products that come under scrutiny for violating WTO rules such as Article III, suggests that the measurement of consumer preferences forms an important area of enquiry. To explore further the problems that consumer preferences raise, the paper then considers the illustrative example of a hypothetical dispute on EC sustainability standards for biofuels.
The EC has developed sustainability standards to regulate how imported and domestic biofuels are produced. A biofuels-exporting country might argue that sustainable and non-sustainable biofuels are ‘like’ products, and that the regulation discriminates unfairly against imports, thereby violating Article III(4). As in EC–Asbestos, the biofuels supply chain would likely make it difficult to quantify consumer preferences. To complicate things further, a common sense determination is probably more difficult than in the example of carcinogenic asbestos. Thus, an approach based on measurement might be difficult; an approach based on interpretation would leave a huge amount of discretion to the AB to decide the importance of sustainability to consumers. ‘EC–Biofuels’ is useful for illustrating the difficulties that arise regarding consumer preferences in public policy disputes, when ‘common sense’ becomes less common.
2. Article III: a brief overview
The majority of Article III complaints cite violations of either paragraph 2 or paragraph 4. Both are interpreted with reference to the same chapeau. Both paragraphs focus on internal regulatory measures and whether they favour domestic products; Article III(2) deals with taxation, and Article III(4) regulation.
Article III(1), the chapeau, reads:
The Members recognize that internal taxes and other internal charges, and laws, regulations and requirements affecting the internal sale, offering for sale, purchase, transportation, distribution or use of products, and internal quantitative regulations requiring the mixture, processing or use of products in specified amounts or proportions, should not be applied to imported or domestic products so as to afford protection to domestic production.
Article III(2) reads:
The products of the territory of any contracting party imported into the territory of any other contracting party shall not be subject, directly or indirectly, to internal taxes or other internal charges of any kind in excess of those applied, directly or indirectly, to like domestic products. Moreover, no contracting party shall otherwise apply internal taxes or other internal charges to imported or domestic products in a manner contrary to the principles set forth in paragraph 1.
The second sentence has an interpretive note, which reads:
A tax conforming to the requirements of the first sentence of paragraph 2 would be considered to be inconsistent with the provisions of the second sentence only in cases where competition was involved between, on the one hand, the taxed product, and, on the other hand, a directly competitive or substitutable product which was not similarly taxed.
In the first sentence of Article III(2), there are two key questions. First, are the domestic and imported products ‘like’? Second, if they are ‘like’, is the imported product taxed ‘in excess of’ the domestic product? In the second sentence of Article III(2), combined with the Interpretive Note and Article III(1), to which it explicitly refers, the first key question is: are the imported and domestic product ‘directly competitive or substitutable’. If so, are the two products ‘not similarly taxed’, and, finally, does dissimilar taxation operate in a manner ‘so as to afford protection’ to domestic production?
Article III(4) reads:
The products of the territory of any contracting party imported into the territory of any other contracting party shall be accorded treatment no less favourable than that accorded to like products of national origin in respect of all laws, regulations and requirements affecting their internal sale, offering for sale, purchase, transportation, distribution or use. The provisions of this paragraph shall not prevent the application of differential internal transportation charges which are based exclusively on the economic operation of the means of transport and not on the nationality of the product.
The key questions are: first, are the domestic and imported products ‘like’? Second, if they are, is the imported product receiving treatment ‘no less favourable’ than that accorded to the domestic product? This paragraph does not include the concept of ‘directly competitive or substitutable’.
As part of the legislative matrix available under Article III, in the event that the measure in dispute has been adopted in support of a public policy goal, such as protection of human health or exhaustible natural resources, or the prevention of deceptive practices, inter alia, a Member State can defend their tax or regulatory measure under Article XX, the General Exception. A measure found to violate Article III may be justified under Article XX if, in the language of the chapeau, it is ‘not applied in a manner which would constitute a means of arbitrary or unjustifiable discrimination between countries where the same conditions prevail, or a disguised restriction on international trade’.
3. Case study 1: Japan-Alcohol
The dispute
In 1996, the newly formed WTO faced its first Article III dispute, Japan–Alcohol, which followed from a 1987 GATT dispute, and had many similar elements.Footnote 9 In the GATT dispute, the EEC argued that, although Japan's taxation policy with regard to alcoholic beverages was origin-neutral, it violated Article III(2). They claimed that the policy was both protective and discriminatory, as it levied higher taxes on imported ‘like’ products, such as whisky and vodka, than domestic products, such as Japanese shochu. Japan argued that its intent was non-discriminatory. Rather, liquor tax was levied according to the tax-bearing ability. Japan also claimed that traditional shochu was not ‘like’ vodka and other competitors. The Panel dismissed Japan's argument about the non-discriminatory intent of its measure, concluding that a wide range of the beverages in question was ‘like’, and that the taxation system discriminated against imported alcoholic beverages.
Japan–Alcohol, in 1996, followed from the GATT Japan–Alcohol. The US, EC, and Canada charged that there was still de facto discrimination against imported alcohol under Article III(2), even after the 1987 GATT ruling. Japan had responded to the 1987 dispute by changing its laws, but put in place an excise tax reduction for small producers of shochu. Also, Japan continued to tax ‘white’ and ‘brown’ liquor differently, which the complainants charged was a discriminatory practice; in particular, they argued that (brown) whisky was ‘directly competitive or substitutable’ to (white) Japanese shochu (Panel Report, section III: paras. 3.1–3.4).
One of the core debates of both disputes concerned consumer preferences: were the differences consumers perceived between shochu, Japan's traditional alcoholic beverage, and imported alcoholic spirits significant enough to determine that the products were not ‘like’ or ‘directly competitive or substitutable’? The dispute settlement bodies' response to this question demonstrated what this paper is calling the ‘objective’ approach to consumer preferences.
The objective approach to consumer preferences
This approach explicitly embraced aspects of consumer preferences that were observable, measurable, and quantifiable. The 1987 GATT Panel Report illustrated some components of the objective approach. In coming to its decision, the 1996 Panel Report also quoted this section of the 1987 Report:
[Thus, even] though the Panel was of the view that the ‘likeness’ of products must be examined taking into account not only objective criteria (such as composition and manufacturing processes of products) but also the more subjective consumers' viewpoint (such as consumption and use by consumers) the Panel agreed … that Japanese shochu (Group A) and vodka could be considered as ‘like’ products in terms of Article III:2 because they were both white/clean spirits, made of similar raw materials, and their end-uses were virtually identical. (para. 5.7) [Emphasis added]
The Panel established a hierarchy of value between subjective and objective components of the ‘like’ products determination. They asserted that physical characteristics (which can be observed) and product end-uses (which examine not what consumers feel about a product, but what they can do with it) were more decisive BTA criteria than ‘subjective consumers’ viewpoint'. This paragraph suggested that these, here defined as ‘consumption and use by consumers’, were not influential at all. (In EC–Asbestos we will see that this hierarchy was still in operation, but in a more nested way: consumer preferences were made ‘objective’ by being measured using these ‘objective’ BTA criteria.)
In another section quoted by the 1996 Report, the 1987 Panel affirmed that since, as ‘the aim of Article III:2 … could not be achieved if differential taxes could be used to crystallize consumer preference for traditional domestic products, the Panel found that the traditional Japanese consumer habits with regard to shochu provided no reason for not considering vodka to be a “like” product’ (para. 5.7). In other words, taxation measures should not interfere with consumer preferences, whose basis should be the free market, rather than the influence of government taxation. The final aspect of the ‘objective’ approach was the quantification of consumer preferences, to establish the market competitiveness of products in dispute.
The objective approach and the rejection of ‘aim and effect’
By examining the GATT Report, we can deduce that ‘subjective consumers’ viewpoints' referred to two of Japan's arguments: first, shochu was a traditional beverage and thus consumers perceived it differently. Second, consumers perceived imported alcoholic beverages as luxury products.Footnote 10 Japan called upon this second argument to justify its intent, which, it stated, was not to discriminate against imports, but rather to tax based upon tax-bearing ability.Footnote 11
Japan presented similar arguments on both points in 1996, arguing that the products were differentiated by the ways in which they were consumed. For example, shochu was drunk before dinner; vodka after dinner. Many shochu drinkers added hot water to the drink, which they did not do to other spirits (Panel Report, para. 4.54). These differences resulted from the fact that shochu was a traditional beverage, and the others were not. Further, Japan claimed that its taxation system was based upon the principle of ‘horizontal equity’, so that tax/price ratios were consistent between all alcoholic beverages (Panel Report, para. 4.19; 4.71). In other words, wealthier consumers' preference for luxury imported alcohol justified the tax differential. In adopting this argument, Japan specifically advocated the application of an ‘aim-and-effect’ test. Under this test, the dispute settlement bodies' consideration of whether products were ‘like’ under Article III would have included an examination of whether the intent or ‘aim’ of the tax was to protect domestic products.
Critical analyses have characterized the dispute settlement bodies' approach to Japan–Alcohol as ‘objective’ (Horn and Weiler, Reference Horn, Weiler and Horn2003: 31), ‘formalist’ (Trebilcock and Howse, Reference Trebilcock and Howse2005: 90), or ‘end-use’ (Choi, Reference Choi2003: 21). Though these analyses differ, they all note that the dispute settlement bodies dismissed not only Japan's argument that its intent was non-discriminatory, but also the justification of a measure based on intent: the ‘aim-and-effect’ test. This dismissal became a consistent strategy in Article III.
The Panel's dismissal of aim and effect rested on the logic that it redoubled the purpose of Article XX, which contains a ‘definitive list of grounds justifying departure from obligations that are otherwise incorporated in Article III’ (Panel Report, para. 6.17). If, as the Panel asserted, the ‘aim’ of the measure does include the type of issues addressed under Article XX, then its rejection has another important implication, fundamental to the objective approach: it narrows the scope of Article III. Rejecting the ‘aim-and-effect’ test implies that extra-market issues will be considered as exceptions, rather than being a part of the initial determination of whether products are ‘like’. This rejection echoed the 1987 GATT Panel's assertion that taxation should not crystallize consumer preferences. The Panel affirmed the market basis of the Article III ‘like’ products determination.
In their dismissal of aim and effect, the dispute settlement bodies also argued that Japan was utilizing subjective arguments, such as analysis of traditional patterns of consumption, to support an unprovable claim of its ‘intent’. Japan had explained its ‘aim’, and justified this aim as non-discriminatory with reference to its analysis of consumer preferences. The dispute settlement bodies equated ‘aim’ with a subjective approach to consumer preferences, and the problem of subjectivity in general: how can you prove an ‘aim’? By the extension of their own logic, ethical concerns dealt with under Article XX might be categorized as subjective. Given the hierarchy of importance between subjective and objective, this would be a cause for concern.
An econometric approach
Thus far, this analysis has focused on the criteria for determining consumer preferences as part of the consideration of ‘like’ products, and the rejection of aim and effect in this context. On this topic, the Panel suggested one principal danger of a subjective approach to consumer preferences: a government might use factors, such as traditional patterns of consumption, to support an unprovable claim of its intent. However, we have not discussed the efficacy of its alternative: a proof based, objective approach.
A positive ‘objective’ approach to consumer preferences is best demonstrated in the dispute settlement bodies' examination of ‘directly competitive or substitutable’ products. After determining that vodka and shochu were ‘like’, they turned their attention to the remaining alcoholic beverages in dispute, and whether they fit into this broader category of comparison.Footnote 12
Japan–Alcohol is noteworthy for having developed perhaps the most econometric approach of any Article III dispute (Horn and Mavroidis, Reference Horn and Mavroidis2004: 62), which reflected an overall emphasis on market competitiveness. The Panel Report stated that ‘in the Panel's view, the wording makes it clear that the appropriate test to define whether two products are “like”’ or “directly competitive or substitutable products” is the marketplace. The Panel recalled … the words used in the Interpretative Note ad Article III, paragraph 2, namely “where competition exists”: competition exists by definition in markets’ (Panel Report, para. 6.22). The AB broadly agreed with this interpretation (AB Report, para. 25).
The dispute settlement bodies used econometric methods to measure competition. Both the Panel and the AB had asserted the importance of elasticity of substitution as ‘the decisive criterion’ in determining whether the alcoholic beverages were ‘directly competitive or substitutable’ (Panel Report, para. 6.22; AB Report, Section 2(a) para. 55).
One of the main products in dispute was whisky. Japan used elasticity of substitution studies to argue that an increase in the price of shochu would not cause significant numbers of consumers to switch to whisky, or vice-versa. Japan cited a study conducted by its Institute for Social Studies, which found that, in either case, only 10% of consumers would switch. Also, the study found no significant impact on whisky consumption based on the price of shochu, or vice versa (Panel Report, para. 4.85). The complainants found fault with Japan's study, pointing out procedural errors, such as the way that the time-series model was calculated, and the need for more variables which might alter consumer behaviour. They introduced separate surveys, which indicated that there had been a negative competitiveness effect of high taxes of whisky, as compared with shochu (Panel Report, para. 4.82).
The Panel had to interpret these competing claims. In the end, they rejected Japan's 10% figure, as it was obtained in the context of the current pricing regime, the subject of dispute. Here, a tax seen as distorting the market was cause to invalidate Japan's evidence regarding consumer preferences. This rationale recalls the GATT Japan–Alcohol Panel Report statement, that a tax measure should not crystallize consumer preferences. The Panel then stated that, even so, 10% elasticity was sufficient to prove direct competitiveness or substitutability (para. 6.31). The AB agreed that the beverages were directly competitive or substitutable (section 2(a) para. 56).
In this example, econometric analysis was no ‘objective’ silver bullet; it demanded discretion on the part of the dispute settlement bodies in order to answer questions such as: how much cross-price elasticity was necessary? Which econometric techniques and studies were most reliable? Certainly, given the imperfection of the econometric tools, the dichotomy between subjective and objective should not be too neatly drawn.
However, despite these imperfections, an econometric approach such as this seems to represent the best possible attempt to capture consumer preferences in a way that conforms to the objective approach the dispute settlement bodies intended. The approach was consistent with a market-based interpretation, translating consumer preferences into competitiveness indices. It was also made possible by the fact that consumers were choosing between products on the shelf, and could express their preferences clearly in a quantifiable way. Consumer preferences could be captured through an explicitly market-based approach. In this sense, Japan–Alcohol forms an illustrative contrast with EC–Asbestos.
4. Case study 2: EC–Asbestos
The dispute
Under Article III(4), Canada challenged an import ban that France had imposed on chrysotile asbestos, as it was a known carcinogen. Rather than imposing a ban, Canada argued France could have instituted special installation and maintenance procedures that would lessen the carcinogenic effect. This would be a less trade-restrictive approach. France defended its ban as a public health protection regulation.
Canada argued that two types of products, its asbestos and France's substitute, were ‘like’ with respect to all of the BTA criteria (Panel Report, para. 3.419). With regard to ‘consumer tastes and habits’, they asserted that as the products were not for mass, but instead industrial, consumption, normal consumers found them to be virtually identical. Therefore, there was no reason to differentiate between the two (Panel Report, para. 3.422). The EC disagreed that the products were ‘like’. With respect to physical characteristics, they pointed out that the smaller diameter of asbestos fibres, a differing physical property, made them carcinogenic (Panel Report, para. 3.431). When it came to ‘consumer tastes and habits’, however, they dismissed the criterion altogether, as it was only industrial consumers who purchased the product.
To resolve the differing claims about the ‘likeness’ of the products' physical characteristics, the Panel determined that the products must not just be viewed in terms of chemical properties, whose differences might be minute, but also in the context of markets. Thus, they affirmed the market-based approach taken in Japan–Alcohol.Footnote 13 The Panel did not think that differences in physical characteristics, or the end-uses of the products, were significant enough to differentiate the products (para. 8.122). The Panel then stated:
We note first of all that the risk of a product for human or animal health has never been used as a factor of comparison by panels entrusted with applying the concept of ‘likeness’ within the meaning of Article III. In addition to the fact that no other panel has probably ever been called upon to examine a question similar to the one before us, in our view the reason is to be found in the economy of the GATT 1994. Its primordial role is to ensure that a certain number of disciplines are applied to domestic trade regulations. Article XX of the GATT, however, recognizes that certain interests may take precedence over the rules governing international trade and authorizes the adoption of trade measures aimed at preserving these interests while at the same time observing certain criteria. (para. 8.129)
This analysis reiterates the Japan–Alcohol rejection of ‘aim and effect’, as it would replicate the purpose of Article XX. Finally, the Panel concluded that, although the two products were ‘like, and therefore the import ban violated Article III(4), it did satisfy the conditions of Article XX(b).Footnote 14 On this basis, they did not recommend that France remove the ban. Though the AB agreed that the ban could remain, they disagreed with the way that the Panel came to this conclusion; a feature of their disagreement was the Panel's treatment of consumer preferences.
How to measure consumer preferences
In Japan–Alcohol, while a good deal of attention was given to consumers, in particular as they related to competition, their influence was defined narrowly, in the sense that they were choosing between ‘like’ products based upon small price differentials in the marketplace. However, in EC–Asbestos, consumers did not have market choice. Asbestos is an industrial product, with industrial consumers, so it would be more difficult to obtain the type of data used in Japan–Alcohol. The Panel felt that it was too difficult to determine precisely the tastes and habits of consumers in France before the asbestos ban. Also, consumer preferences were inevitably varied. Therefore, the criteria would not provide clear results. They refrained from taking a position (para. 8.139).
In fact, based on the impossibility of assessing consumer preferences in an ‘objective’ manner, neither Canada, nor the EC, nor the Panel considered consumer preferences to be a compelling aspect of their argument (paras. 3.419–3.431; 8.139). In a broader sense, this reveals a limitation in approaching consumer preferences through quantitative analysis; it becomes less applicable in certain types of product supply chains.
Though the Panel's finding was consistent with Japan–Alcohol's objective approach to consumers, the AB criticized the Panel's omission of ‘consumer tastes and habits’ as a criterion for determining whether the two products were ‘like’. Again affirming the Japan–Alcohol approach, the AB affirmed that this criterion was particularly important as Article III dealt with competitive relationships. They stated:
We do not wish to speculate on what the evidence regarding these consumers would have indicated; rather, we wish to highlight that consumers' tastes and habits regarding fibres, even in the case of commercial parties, such as manufacturers, are very likely to be shaped by the health risks associated with a product which is known to be highly carcinogenic. A manufacturer cannot, for instance, ignore the preferences of the ultimate consumer of its products. If the risks posed by a particular product are sufficiently great, the ultimate consumer may simply cease to buy that product. This would, undoubtedly, affect a manufacturer's decisions in the marketplace. (paras. 103–104) [Emphasis added]
Here, the AB attempted to include consumer preferences in an analysis of competition. As the words ‘very likely’ demonstrate, it revealed their interpretation of consumer preference, particularly as none of the disputing parties had made any claims about it. The interpretation was based on a strong common-sense argument, grounded in international standards and guidelines, that consumers would not want to purchase a substance with known carcinogenic properties (World Health Organization, 1998, para. 144). Therefore, unlike the Panel, the AB assumed that consumer preferences could be interpreted in a normative fashion.
But presumably, in Canada, industrial consumers did purchase the product. Neither Canada, nor any of the other disputing parties, had submitted studies on consumer preferences and asbestos to substantiate any claim. The Panel's argument that ‘consumer tastes and habits’ are varied seems more plausible than the AB's argument that they are normative. Further, the logic of the Panel is more consistent with Japan–Alcohol, and an objective view of consumer preferences. Using this approach, public health concerns are addressed under Article XX, which deals with them explicitly. As the Panel pointed out, they were never introduced into the ‘like’ products determination itself.
However, there was a problem with the Panel's logic: it created the potential for a legitimacy crisis (Howse and Tuerk, Reference Howse, Tuerk, Burca and Scott2001). By deeming a carcinogenic and non-carcinogenic product ‘like’, the Panel appeared insensitive to public health concerns. Their reasoning demonstrated a public relations risk of jettisoning extra-market considerations to Article XX, when considering regulations with consumer protection objectives.
The AB was forced to abandon an objective approach to consumer preferences because of these limitations, and adapt a discretionary, subjective, interpretation, which this paper calls a common-sense approach. However, they did not acknowledge this. After affirming the importance of ‘consumer tastes and habits’, the AB questioned the Panel's finding that the two products were ‘like’, but because there was not enough evidence that the products had similar physical characteristics. In so doing, they adopted the EC's argument about differing particle diameter. Then, they interpreted consumer preferences through physical characteristics. The different molecular properties that led one product to be carcinogenic were the reason for consumers' distinction between the products, and thereby made the products uncompetitive.
The AB presented their approach as ‘objective’. ‘Consumer tastes and habits’ were based upon market competitiveness between products, and interpreted in the light of other, more objective BTA criteria of physical characteristics. The AB stated that they did not want to speculate about consumer preferences, but that very speculation seemed to form the substantive justification for the differentiation. Further, considering that many ‘like’ products differ on a molecular level, their argument about physical characteristics does not seem that rigorous.
While the AB claimed they were evaluating the products based on their competitive relationship, it could easily be argued that their rationale reflected a subjective aim and effect approach. Under the ‘likeness’ test itself, rather than under Article XX, the AB honoured the fact that the aim of the regulation was to protect consumers, even though its effect was to eliminate trade in this product. They thus introduced an extra-market dimension into the ‘like’ products determination.
They also overstepped an important boundary. Findings of fact, based in interpretation of the evidence of the dispute, are the responsibility of the Panel. The AB, on the other hand, is empowered only to make findings of law, derived directly from legal principles. Perhaps the most problematic aspect of this decision is not that they made a finding based upon no evidence, but that they made a finding of fact at all. However, the AB did not acknowledge this breach. Instead, they finally concluded that, though they would not agree with the Panel that the products were ‘like’, there was not enough evidence to prove they were unlike.
With regards to consumer preferences, EC–Asbestos seems to offer a lose–lose situation. If, following the Panel, consumer preferences are interpreted as ‘varied’, and therefore irrelevant, there is no basis for incorporating consumer preferences into an important public policy decision. However, when consumer preferences were defined by the AB's discretion, this involved a transfer of power, presented as an ‘objective’ evaluation of product competition and physical differences. There is a gap between the stated approach and the content of the decision. Though the ends were good, from the perspective that they humanized the decision, the means were bad, in the sense that they opened up the potential (which perhaps will never be realized) for a future overstepping of discretion.
In this dispute, a market-bound determination on ‘likeness’ under Article III had its limitations. Had product ‘likeness’ been evaluated on the basis of health impacts, rather than product competitiveness, the AB's assessment would have been logical. Horn and Weiler term this an ‘alternative comparator’ approach (Horn and Weiler, Reference Horn, Weiler and Horn2003: 31). So far, however, the dispute settlement bodies have resisted measuring ‘likeness’ using any other baseline than competitiveness. Presenting these complex value judgments as an objective evaluation is somewhat disingenuous. Whether it is used to legitimate or to undermine national regulation, allowing this level of AB discretion in interpreting consumer preferences seems an unsatisfying approach.
5. Case study 3: EC–Biofuels
The hypothetical dispute
Though biofuels have not been the subject of a trade dispute, there is the potential for conflict under Article III(4). International trade in biofuels is increasing, and the EC has developed sustainability criteria for their production, which may present an obstacle for some imports (Swinback, Reference Swinback2009). The EC has stated that, in order to count toward a 10% binding target for renewable energy in the transportation sector, biofuels should meet these criteria (Directive 2009/28/EC of the European Parliament and of the Council, 23 April 2009). The criteria will be applied to both domestic and imported biofuels, and provide a basis for EU Member States to distinguish between sustainable and non-sustainable production. Whether these provisions, when applied, will discriminate unfairly against importers is subject to debate. There could be an Article III(4) controversy if exporting countries argue that sustainable and conventional biofuels are ‘like’, and that sustainability regulations discriminate unfairly against imports. In evaluating this argument, consumer preferences are an interesting area to consider. Perhaps a good way to go about this is to review the three approaches documented here, and how they would apply in this dispute: first, objective and quantitative; second, objective and dismissive; and, third, common sense.
It should be noted, of course, that this is a very speculative exercise. Not only is the dispute hypothetical, but the specific details of the submissions by the Parties would necessarily shape the outcomes. However, without attempting to discern an outcome, this exercise intends to outline some basic contours of the challenge with regard to consumer preferences.
The objective approach 1: quantitative measuring
As well as consumer preferences, evaluated econometrically, Japan–Alcohol also emphasized the importance of physical characteristics and end-uses of products when determining whether they are ‘like’. The criterion of product end-uses is directly relevant to consumers. The criterion of physical characteristics, on the other hand, operates to some extent in parallel to the consumer preferences query. However, EC–Asbestos provided an unusual demonstration of how physical characteristics, in this case molecular differences, might directly influence consumer preferences.
In terms of both physical characteristics and end-uses, sustainable and conventional biofuels would be virtually identical. Even on a molecular level, it would be impossible to distinguish any systematic differences between, for example, palm oil whose production has involved the clearing of virgin forest, and palm oil whose production has not. Therefore, physical characteristics could not justify differing consumer perceptions, as they did in EC–Asbestos.
Also, when it comes to their end-uses, sustainable and non-sustainable biofuels are the same. In Japan–Alcohol, Japan argued that consumers drank shochu in a traditional way; this type of argument could not be made for biofuels. Of course, the dispute settlement bodies did not consider this a persuasive distinction on the part of Japan. On the other hand, it is still a valid observation that, regarding end-uses, sustainable and conventional biofuels are even more ‘like’ than, for example, shochu and vodka.
Both of these points play into one overarching issue: that of product distinctions based upon processing and production methods (‘PPMs’) which have no physical impact on the final product, so-called non-product-related, or ‘NPR’ PPMs. Biofuels, produced sustainably versus conventionally, provide an excellent example of a NPR PPM-based regulatory distinction. Given the ‘like’ products test's emphasis on physical characteristics and end-uses, there has been some debate about whether two otherwise identical products may be seen as unlike, solely because they are produced differently.Footnote 15 This is an important question for environmental regulations in general, as many rely upon such distinctions.
Neither of the disputes analysed herein explicitly addressed the role of NPR PPMs in establishing ‘like’ products; however, both emphasized the importance of physical characteristics. This functions as a counter-argument to basing product distinctions on NPR PPMs. On the other hand, it has been argued convincingly that the market-based approach to consumer preference established in Japan–Alcohol, and emulated in EC–Asbestos, resolves some of the environmental shortcomings associated with a reluctance to acknowledge PPM distinctions (Bronckers and McNelis, 2000: 374–378). This is because, under this approach, if consumers clearly distinguish between physically identical products, so that they are no longer competitive, it forms a strong basis for determining that they are not ‘like’. This de-emphasizes aspects of the ‘likeness’ test related to observable product characteristics and end-uses.
Thus, a greater role for consumer preferences represents a useful evolution in the PPMs debate. The problem that remains is methodological: how will these preferences be measured? ‘EC–Biofuels’ illustrates these difficulties. With regard to the econometric approach to consumer preferences, the Japan–Alcohol approach contained, in particular, one requirement missing in ‘EC–Biofuels’: consumer choice. The approach of measuring consumer preferences necessitated data and surveys on purchasing behaviours, which most likely would not exist in this scenario. The supply chain of most biofuels is probably more similar to that of asbestos, in that consumer preferences are not economically explicit. Within Canada, asbestos was purchased by industrial consumers, and installed without end-of-the-line consumers having any say. The same was true for asbestos substitutes in France. In the same way, it would be unlikely that consumers would have much choice between sustainable and conventional biofuels, in either a complaining or a responding country. Small-scale distribution channels might be feasible for some types of biofuels, such as recycled cooking oil. However, given the presumed difficulty of sourcing, transporting, and delivering fuels on a mass scale, while a separate pump with sustainable biofuel is certainly an option, it seems unlikely as an EU-wide practice. For the same reason, it seems unlikely that econometric data could be drawn from a complainant country.
A unified ‘at the pump’ delivery infrastructure requires a degree of regulatory involvement at the point of purchase. Rather than the individual or industrial consumer, the state might be deeply implicated in deciding what kind of biofuels to certify and import. This is a significant departure from many existing labeling schemes, such as Fair Trade and Forest Stewardship Council certification, which involve producing goods with superior ethics, as indicated by a label and/or explicit marketing, aimed at capturing a premium market share of alternative consumers. The willingness of consumers to pay more for a product which supports their environmental and social values underwrites the diversification of market choice. These types of voluntary schemes are much more immune from WTO action, as the element of consumer choice is preserved, and there are no restrictions on imports of uncertified or unlabelled goods. The fact that sustainability criteria are being developed by the EC, to be applied by EU Member States, makes them potentially more problematic, from a WTO perspective.
For this supply-chain related reason of lack of market choice, it would likely be difficult to amass data about consumer preferences using market-based surveys or data like those in Japan–Alcohol. However, perhaps the notion of ‘measuring’ consumer preferences might be incorporated through another strategy, which does not rely upon such economically explicit data. An attempt to measure consumer preferences through surveys probably represents the best attempt to capture them objectively. For example, surveys could be administered which queried consumers for their views on sustainable versus conventional biofuels. The emphasis, of course, would be on determining competitiveness between products.
To obtain this information, a survey would have to undertake a fairly nuanced set of questions. If more than half of consumers stated that they would prefer sustainable to conventional biofuels, this would represent a collective consumer preference toward sustainable biofuels. Yet, if consumers said they would prefer sustainable biofuels, would this mean that they would not purchase conventional ones, if they were available, even if they were cheaper? If only conventional biofuels were available, would consumers go further and boycott these? There is also the problem that this type of data is speculative. Is there a gap between how consumers would respond to such a survey, and how they would act in real life? Also, to use the word of the GATT Japan–Alcohol Panel, what if a regulation supporting the use of sustainable biofuels ‘crystallized’ consumer preference toward these biofuels? Would the presence of the regulation invalidate data gathered about consumer preferences, as it did in Japan–Alcohol? In that case, would the only legitimate data about consumer preferences be gathered from countries which did not offer sustainable biofuels? There is no real guidance within the WTO system on how to deal with such complexities.
The objective approach 2: dismissing consumer preferences
In EC–Asbestos, the Panel concluded that the products in dispute were ‘like’, and then moved on to an analysis of whether the regulation represented less favourable treatment of imported products. Due to lack of evidence, consumer preferences were dismissed as ‘varied’. The public policy justifications behind the regulation were considered in the context of Article XX, where they were not evaluated in terms of consumer preferences, per se.
Of course, under EC–Asbestos, the Parties in dispute made no claims about consumer preferences. Thus, the Panel had no evidence to go on. Had the Parties undertaken surveys, there would have been more of a rigorous basis for their conclusion. Clearly, an evidence-based approach is the best way to deal with consumer preferences, though whether evidence is submitted is beyond the control of the dispute settlement bodies.
In the event of an ‘EC–Biofuels’ dispute, it might prove impossible to measure, or interpret, consumer preferences with any degree of accuracy. In this event, the most rational and consistent approach, in its conformity with the stated jurisprudence reviewed in this paper, would be to follow the Panel in EC–Asbestos, and dismiss consumer preferences. In a sense, dismissal is unsatisfying, because consumers seem to be such an important component of public policy. Undoubtedly, at least some consumers have strong preferences. However, given the similarities between physical characteristics and end-uses of the products, and lack of data on consumer preferences, it would be easy to argue that sustainable and conventional biofuels were ‘like’, and then undertake the other steps of the analysis.
Based on the example of EC–Asbestos, it seems there is another force at work: an intuition as to the degree of ethical controversy the issue will provoke. The emphasis in determining ‘likeness’ has been product competitiveness. Would consumers really refuse to buy conventional biofuels on a mass scale, making them uncompetitive with sustainable ones? Perhaps, in ‘EC–Biofuels’, it would be easier for the dispute settlement bodies to get away with dismissing consumer preferences than in EC–Asbestos, on the argument that it was less likely to be such a controversial issue for them.
The common sense approach
The example of ‘EC–Biofuels’ is instructive because it inspires thought about what might happen if a common sense approach is applied in a more ethically ambiguous situation. International standards support the classification of asbestos as a carcinogen, whose use should be eliminated. This made the interpretation of consumer preferences in EC–Asbestos fairly unambiguous. The controversial nature of the dispute also influenced the AB to include this consideration under Article III itself, rather than Article XX.
On the importance of sustainable production of biofuels, however, targeted international standards are not so well established. For this reason, a common sense interpretation would likely conclude that sustainability is less fundamental than, for example, the preservation of human health. An implicit hierarchy of concerns would likely be in operation, which reflected the extent to which the regulations in question demonstrated broad-based societal norms. If the AB affirmed, for example, that consumers would ‘very likely’ prefer sustainable to conventional biofuels, and that these preferences would impact upon the marketplace, this argument might prompt more criticism than it did in the EC–Asbestos decision.
If the dispute settlement bodies wanted to affirm the legitimacy of sustainability regulations under Article III, and be logically consistent, they would probably need to adopt an ‘alternative comparator’, so that the relevant criteria for evaluating product ‘likeness’ consisted of environmental or social impact. Alternatively, they could explicitly reinstitute the ‘aim-and-effect’ test.
Of course, there is no way of knowing what the dispute settlement bodies would decide. However, if consumer preferences were to be considered relevant to the dispute, and these preferences were interpreted, rather than measured, this interpretation would grant a large degree of discretionary power to the dispute settlement bodies. To resolve this problem, as the conclusion further discusses, it is important to be clear about what ‘consumer preferences’ really means.
6. Conclusion
As well as biofuels sustainability criteria, the passage of climate change-related regulations with potential to violate WTO rules provides a fresh opportunity to establish how, precisely, the influence of consumer preferences should be incorporated, particularly in ethical public policy disputes. The issue is important, in the context of the argument that consumer preferences can be used to distinguish between products that would otherwise be seen as ‘like’, for example sustainable versus unsustainable, or carbon-intensive versus low-carbon, goods (Vranes, Reference Vranes, Herrman and Terhechte2010).
Due to their ethical dimensions, disputes about public policy regulations, such as climate change measures, are relevant to consumers in a crucial way. Yet, this paper argues that they can become more difficult to capture in these very circumstances. In EC–Asbestos, consumer preferences were not measured, in part because the supply chain of the product in the complaining country and the import ban in the responding country made gathering data in both places difficult. Yet, they were undeniably relevant. Perhaps because of this conundrum, the AB introduced consumer preference even though it required a contradictory approach. In fact, the AB seemed to employ the same ‘subjective’ approach to consumers that it had so clearly rejected both in Japan–Alcohol and EC–Asbestos itself. This temptation, to use consumer preferences as a stand-in for discretionary action, may recur in disputes that concern public policy regulations.
In such disputes, rather than actual consumers, the interpretation of consumer preferences perhaps should be re-framed: it reflects more about societal norms. Specifically, the more risky and controversial the public threat being forestalled by a tax or regulation is seen to be, the larger the temptation to deal with it under Article III, instead of Article XX, and, in the absence of data, to interpret consumer preferences. In EC–Asbestos, the AB argued that they were evaluating consumer preferences based upon competitiveness impacts. However, it is probably more accurate to conclude that norms about protecting human health were universal enough to intrude upon the primary, market-based purpose of Article III. This clarification is a step toward a more rational approach. Unavoidably, in a dispute such as EC–Asbestos, rather than consumers, the discussion concerns the less market-based concept of ‘the public’.
The difficulty with measuring consumer preferences, as documented in EC–Asbestos, could simply be seen as a methodological failure. Canada and the EC could have made their approach to consumers more ‘objective’ simply by undertaking studies. They might have administered surveys that questioned consumers about their views: did they see asbestos as a serious health threat, even if it was installed using safer procedures? To improve methodologies for capturing consumer preferences, more techniques could be applied, perhaps drawn from other areas, such as Competition Law, which have already developed an arguably more sophisticated approach to measuring consumer preferences (Choi, Reference Choi2003: 17). With evidence to work with, the AB might have used an interpretative strategy as objective as that employed in Japan–Alcohol.
Certainly, better methodologies would aid the AB in coming to a more democratic interpretation of consumer preferences. However, one important question remains. Presumably, these responses would differ in both countries, based upon the presence, or absence, of a ban. Whose views would be ‘objective’? In GATT Japan–Alcohol, the Panel affirmed that a tax should not crystallize consumer preferences. Utilizing the same rationale, the Panel in WTO Japan–Alcohol rejected data Japan had gathered about consumer preferences under the conditions of the taxation policy in dispute. The underlying assumption, that government regulation should not influence consumer preferences, is not surprising, given the free market basis of the WTO. When applied to a public policy tax or regulation, however, the extension of this logic has radical implications. It would mean, for example, that consumer views in France, the country with the ban, would not be considered relevant. It would also follow that governments should only establish regulations whose source is measured consumer preferences. This invalidates the argument that governments should influence consumer preferences in support of public policy goals. It also suggests that it is the responsibility of consumers, rather than government, to be informed of public health threats, such as asbestos.
Despite the lack of homogeneity in the views of consumers, governments routinely make normative regulations. The GATT Panel's insistence that the tax in Japan–Alcohol should not ‘crystallize’ consumer preferences is anathema to the fact that, in some instances, government regulation does affect consumer preferences. Through setting regulation, such as sustainability criteria, the state influences consumers to adapt to shifting or emerging norms. Regulatory decisions reflect national priorities, and there is certainly no guarantee that they benefit the public. Nor are states impartial guards of the global public good. Their regulatory preferences likely reflect their economic priorities, one factor that makes the dispute settlement bodies' task so difficult. Nevertheless, it is important that the WTO system not act as a braking mechanism for more progressive regulations that support emerging norms, such as, for example, regulations that support environmental sustainability, or climate change mitigation.
A greater recognition of the authority of governments to influence consumer preferences seems to call into question the fundamental ideological premise of the WTO: the free market. On the other hand, if we think about the effect on legal reasoning, it can be argued that this recognition does not need to have such a radical impact. There would be factual and legal implications for the Panel and AB, respectively. Factually, the Panel would evaluate evidence obtained within the context of a regulatory regime in dispute. If our hypothesis of government influence on consumer preferences is correct, this would have the impact of providing more WTO legitimacy for these regulations.
Legally, there might be a systematic approach for identifying progressive or ethical regulations that merit the allowance of more government discretion. Presumably, the first step would be to establish criteria that distinguish these regulations from other types. Given the discretionary range of the AB, there would certainly be room for this within the existing Article III, and it might even clear up inconsistencies between approaches to consumer preferences. The process of establishing criteria, however, could be difficult. For example, a government might present its policy of mandating the purchase only of domestic products as an ethical public policy regulation, though it is anathema to Article III. An attempt to create general criteria which successfully addressed these and other problems might be crushed under the weight of their own complexity.
This is particularly true as there is a better established way of approaching essentially the same problem. This is to question whether Article III should accommodate an aim and effect or ‘alternative comparator’ approach that would allow a larger set of concerns to contribute to the ‘like’ products determination, or whether it should maintain its market-based orthodoxy. It may be that EC–Asbestos represents a unique departure on the part of the AB, which was prompted by an extraordinary set of circumstances: the lack of ambiguity in the public policy value at stake (human health), and public pressure to avoid appearing insensitive to such concerns. These factors forced the AB to incorporate non-competitiveness criteria into the ‘like’ products determination.
On the other hand, shunting such concerns to the category of ‘exceptions’ under Article XX also might prompt controversy in other situations. If the dispute settlement bodies wanted to adjust their norms, concluding that sustainably produced and conventional biofuels were not, in fact, ‘like’ for example, they would need to depart from the approach of Japan–Alcohol and EC–Asbestos. These disputes take as their foundation the notion that ‘like’ products are competitive products. When product ‘likeness’ is measured through competitiveness, it undermines the ability of the dispute settlement bodies to recognize extra-market, ethical considerations as a legitimate basis for distinguishing products under Article III, or to allow Member States to do so. If the dispute settlement bodies considered whether sustainability standards for biofuels formed a violation of Article III, an approach which considered only the competitiveness between conventional and sustainably produced biofuels would likely fail to justify such regulation. If, however, an ‘alternative comparator’ approach were applied, biofuels could be differentiated because of their differing environmental impact. This determination would not rest on an illogical, discretionary interpretation of consumer preferences. Thus, it is a better, more straightforward approach to incorporating non-trade societal norms in Article III.
