The Question of Like Product Is Surprisingly Controversial

The question of ‘like product’ is surprisingly controversial given the scant number of cases that have come before the WTO’s dispute resolution panel. Although there are numerous provisions within the GATT that rely upon this term, this paper will only focus upon its use within art.III GATT.

The common starting position for such a discussion is the idea that the concept of a ‘like product’ can only be distinguished upon a ‘case-by-case’ basis. This suggests that the term ‘like product’ is not capable of a meaningful definition or the application of firm parameters. Pg.1

Perhaps the best place to start is a clear understanding of the function of art.III GATT. Art.III GATT contains the National Treatment principle upon which the WTO is based. The purpose of this provision is to prevent differential treatment of products which will lead to market distortions. The concept of ‘like products’ appears within art. III in paragraph 2 and 4 and its inclusion is clear from paragraph 1. The rule that internal measures should not give less favourable treatment to ‘like products’ will be successful when they are defined as competitive foreign products. Pg.4 However the two separate inclusions of ‘like product’ terms within art.III has produced two separate meanings of the term. Both a flexible reading and a literal reading can be attributed to the provision and the woolly case law has exasperated this uncertainty. Pg.1

The confusion of past decisions has led to a restricted test for what constitutes a ‘like product’. As such the appellate body has been quick to rely upon the comparison of the physical characteristics of goods mostly. But if any physical differences are identified it then becomes necessary to apply other criteria to determine whether the goods are ‘like products’ or not Pg.3.. Hence there is an urgent need for a clear list of other criteria that can be taken into account when determining the ‘likeness’ of two products to help bring stability and certainty to the WTO dispute settlement body. Such a list is a necessity within the WTO structure to help all member states to clearly identify rule breakers to prohibit all preferential treatment and ensure that the true spirit of art.III is not restricted. On the other hand a clear test for ‘likeness’ will also prevent an excessively draconian application of art. III. Pg.1

DEFINITION IN art.III (2)

To help create a meaningful definition of ‘likeness’ any framework would have to build upon the purpose behind art. III. This purpose can be gleaned from the choice of phrase itself. In paragraph 2 the Ad Note highlights that a ‘like product’ is a far more specific term than ‘directly competitive’ products, as it refers to less products. Therefore it is not hard to agree that the Appellate Body is correct in focusing upon the mere physical characteristics of products they wish to compare. Pg.5 However this should only be a starting point, as such an approach cannot successfully embody the anti-protectionist aim of art. III GATT.

Clearly the term ‘like product’ is utilised in different ways within art. III GATT. In paragraph 2 the drafters felt the need to employ a two sentence structure clarifying that ‘likeness’ is simply a section of goods that are competitive and that require automatic protection. However the use in paragraph 4 does not follow this same structure. Therefore it could be argued that the same rationale was not at work i.e. the term ‘likeness’ has a different meaning in paragraph 4. As a result the two paragraphs require different definitions of the term ‘like products’. pg.6.

INHERENT CONFLICT BETWEEN (2) AND (4)

These definitions of ‘like products’ within paragraph 2 and 4 of art. III GATT may be inherently different but this does not affect the ability to create a cohesive test for ‘like products’. Indeed WTO panel rulings have exemplified this in their rulings by according different definitions to each provision. In the Japan – Taxes on Alcoholic Beverages case, the Appellate Body in its reasoning, made clear their position. ‘We believe that, in Article III:2, first sentence of the GATT 1994, the accordion of “likeness” is meant to be narrowly squeezed.’ Japan case at page 21. Presumably then the Appellate Body must not apply the same definition to all uses of the term ‘like products’ within art. III GATT. Moreover this suggests that the default definition for ‘like products’ must be a broad one, encompassing the economic factor of directly competitive products. This could help explain the inherent conflict between paragraph 2 and 4. In addition it would also solve the individual two sentence structure of art.III:2 if the second sentence of paragraph 2 was taken to refer to dissimilar goods that share no physical characteristics but do share a competitive relationship in the market place.

It is becoming clear then that a single definition will not satisfy the reasoning of both paragraphs of art. III GATT. This was acknowledged by the Appellate Body itself in the Japan Alcoholic Beverages Case. Japan–Alcoholic Beverages, Appellate Body Report at page.21. The statement that ‘QUOTE’ is a perfect explanation of the current situation. It also explains why it is necessary for there to be two separate definitions of ‘like products’ within art. III. The two paragraphs require different tests to be implemented in order for the provisions to achieve their aim Pg.9. It is clear that this must be the case. However the difficulty comes in how these tests are carried out and what criteria are advanced for each separate paragraph. It may be inherent in the drafting that these two provisions should result in two individual definitions of the term ‘like products.’ What is not apparent is why they should not employ a joint test with a defined list of criteria? The longer this issue is ignored the greater the divergence between the two terms will be, particularly as the existing WTO decisions have not offered any clear criteria. To formulate a cohesive test it is important to explore the current approach by the Appellate Body and the decided case law.

CASES UNDER ART.III (2)

Pg.18 – D. The Article III:2 cases

Although it is possible to set aside Article III:2 definitions of “like product” as sui generis

due to the two-sentence architecture of Article III: 2, the tendency to cross cite all manner of “like product” cases makes it relevant to examine the contribution of such cases to the “like product” literature. The cases of particular importance are (1) the two Japan Alcoholic Beverages cases — the 1987 GATT panel report in the first case,32 and the 1996 Appellate Body report in the second case33 — and (2) the recent Appellate Body report in the Periodicals case.34

Japan Alcoholic Beverages I and II. In the first, 1987 Japan Alcoholic Beverages case, the

panel made an interesting series of “like product” rulings. It ruled that standard distilled spirit

classifications were “like products” — gin, vodka, whiskey, grape brandy — as well as classic liqueurs, still wine and sparkling wine. The functional effect of the holding was to rule that Japan could not impose different internal taxes on different quality grades within these product categories – – in other words, that it was impossible to subdivide these product categories according to quality. Arguably the rejection of distinctions based on quality offered a broader and more rigorous definition of “like product” than one would expect to find applied to product distinctions made in tariffs, where similar distinctions appear routinely. The panel did nothing, however, to suggest that Article III requires a more stringent test. The panel said that its “likeness” conclusions had relied upon the criteria listed in the 1970 working party report, plus a variety of tariff and statistical reporting categories that subdivided alcoholic beverages into these categories without further breakdown — the same kind of reasoning involved in all other “like product” cases. The panel also

relied on the broader definition applied in the Coffee case, citing that decision for the proposition that minor differences in taste, color and other properties would not affect the “likeness” of products.

Pg.19 – The panel in the first Japan Alcoholic Beverages case indicated that it would not have had a problem with tax differences based on alcohol content, which it viewed as an “objective” basis for product distinctions. The remark indicates that the determining issue in its “like product” analysis, whether consciously or not, may have been the question of protective purpose. The quality distinctions involved in that Japanese tax law apparently struck the panel as protective in purpose, partly because their arbitrary and subjective character, and partly because of their manifestly protective results.

In the second Japan Alcoholic Beverages case, the main “like product” issue involved an

internal tax on shochu that was lower than the comparable tax on several Western varieties of

distilled spirits. Under the first sentence of Article III:2, the question was whether any of Westerntype distilled spirits were “like” shochu. The initial panel decision had followed the first panel’s “like product.” classification in which only vodka, which is almost identical to shochu in all respects except filtration, was found to be “like” shochu. In distinguishing other types of distilled spirits, the panel in the second Japan Alcoholic Beverages case claimed to be applying market principles, but in fact the panel did do so. Instead of relying on market-based characteristics such as consumer preferences, the panel relied on physical characteristics reminiscent of those employed to make very fine tariff distinctions — color, raw materials, presence of additives. The Appellate Body confirmed the narrowness of the criteria applied by the panel, stating that the juxtaposition of “like” products in the first sentence of Article III:2 with “directly competitive” products in the second sentence of III:2 requires giving “like product” a very narrow reading in this particular context.

Periodicals. The narrow view of “like product” taken in the second Japan Alcoholic

Beverages case was implicitly confirmed by the outcome in the subsequent Periodicals case. The initial issue was whether certain foreign periodicals were “like” more favorably treated domestic periodicals. Although the two kinds of periodicals were clearly competitive, the panel apparently had great difficulty in finding a plausible rationale to explain why such periodicals were “like,” in the face of an argument that the content of the two groups of periodicals was an important distinction. The panel eventually issued a difficult-to-follow rationale that the Appellate Body could not accept. But rather than try its hand a better “like product” rationale — the usual Appellate Body practice in such cases –, the Appellate Body proclaimed the record inadequate to make such a judgment, and so decided the case under the “directly competitive and substitutable” standard of the second sentence of Article III:2 — an issue as to which there was even less information in the record. The tortured course of the “like product” issue strongly suggests that both panel and Appellate Body were operating within a quite narrow view of “like product” in which different characteristics carried more weight than competitiveness.

Pg.20 – Given the tendency of GATT/WTO tribunals to treat all “like product” decisions as fungible, these two most recent Appellate Body decisions on the “like product” issue will probably direct subsequent decisions toward a narrower rather a broader interpretation of “like product,” policy context notwithstanding. If so, it may well be a long time before the policy distinctions argued for in this paper are accepted by WTO legal decisions. And that is where the issue now stands.

Pg 8 – For instance, GATT Panels ruled that ‘likeness’ does not exist between three types of sardines (pilchard, herring and sprat) or between dimension lumber produced from different tree species (SPF and hemlock-fir lumber) for purposes of Article III:2 GATT. Another GATT Panel explicitly refused to consider the competitive relationship between ammonium sulphate fertilizer and nitrate fertilizer and found the two products to be ‘unlike’.

Pg.5 – However, it is difficult to accept the reasoning given by the panel in the Japan–Alcoholic Beverages case. The point is that the Panel missed one of the pertinent points in its line of reasoning. The missing point is that not all tax based on different consumer habits crystallises consumer habits. Hence the panel should have examined whether the different tax imposed on schochu and vodka would be likely to crystallise consumer habits in Japan, just as any panel examining the likeness between dog and pork meat should do. In other words, if the force of consideration comes from the possibility of crystallising consumer habits, one should at least pay due attention to “ possibility” .

Pg.5 – No consideration was given to this point by the Panel; it simply assumed that all taxes imposed differently according to different consumer habits would result in such crystallisation because “ all habits are variable” .33

Pg.5 – This is not to argue that all consumer tastes or habits are relevant in determining “ likeness” . The argument is that certain consumer habits are intertwined with market conditions so strongly and steadily that the former can’t be conveniently separated from the latter. Taking account of all these “ obvious” elements doesn’t make the GATT likeness test unstable.

Pg.6 – Recently, with respect to the interpretation of “ directly competitive or substitutable” in the second sentence of art.III: 2, the AB in Japan–Alcoholic Beverages put forward a new approach, which explicitly endorsed the need to look at the marketplace.34 According to the EC–Asbestos case, an examination of market-based evidence relating to consumers’ taste and habits is an “ indispensable” .35 This approach is being followed by the WTO tribunals in Korea–Alcoholic Beverages and Chile–Alcoholic Beverages. 36 As a result of this trend, mere references to similar physical characteristics, general end-uses and/or tariff classification are not enough; a balance consideration should be given to the marketplace.

Pg.6 – Korea–Alcoholic Beverages deepened the concept of market-based end-use. It added an additional dimension of “ potentiality” to the concept of end-use. Not only current end-use is relevant, according to the Panel; whether there is “ increasing end use substitutability” is important.37 In the end, the WTO tribunals seemed to be equipped with an accordion in each hand–one being the traditional “ accordion of provision-by provision” 38 and other the brand new “ accordion of potentiality” .39

Pg.6 – In short, although the market-based approach suggested by the WTO Panels and the AB improved likeness by restoring the balance between the objective and market-based elements and by providing subsequent panels or member with some methodological guidance. However, a more serious problem is that questions of how and why the definitions of these terms should vary from clause to clause and how much different interpretation of identical terms may be possible within a clause are left wholly unanswered; nevertheless this part is taken care of in the subsequent portion of this odyssey.40

CASES UNDER ART.III (4)

Pg.17. – The “aim and effects” cases. The two other “like product” cases involving Article III:4 were the two 1990s cases in which GATT panels suggested abandoning the traditional definitions of “like product” and replacing them with what became called the “aim and effects” test – Malt Beverages and Auto Taxes. These cases represented an effort, inter alia, to establish a definition of like product that would require analysis of market effects in every case. These two decisions were rejected by the WTO Appellate Body as an incorrect application of the “like product” concept under Article III:2. Shortly thereafter, the aim and effects test was rather summarily rejected as an incorrect application of the “like product” test under Article III:4 as well.31

Pg.18. – In sum, except for the two now-repudiated “aim and effects” cases, the scanty body of panel decisions interpreting the “like product” concept with regard to internal measures covered by Articles I:1 and III:4 shows little sign of adopting an interpretation that would focus on the competitive relationship between the products. Indeed, there is little evidence of any interest in distinguishing in any way between cases involving internal measures from those involving tariffs.

Oddly enough, the broadest definition of “like product” to date appears to be the Coffee decision – a case involving tariff distinctions. However good the policy reasons for a broader “like product” test with regard to product distinctions in internal measures, GATT and WTO legal decisions have not so far accepted it.

Pg 8 – Finally, the GATT Panel in EEC – Animal Feed Proteins held that different products used for the purpose of adding protein to animal feeds are ‘unlike’ under Articles I and III:4 GATT. All these reports relied heavily on different tariff classifications as well as physical differences between the products as criteria of the ‘likeness’ analysis.16

Pg.10 – The Appellate Body Report on EC – Asbestos nicely illustrates this change in jurisprudence with regard to the assessment of discriminatory regulations (Article III:4 GATT).22

WORKING GROUP AND CASE CONCLUSIONS

The sparse number of rulings by the WTO upon the meaning of ‘like product’ have created more confusion than clarification. Consequently two conclusions arise from the line of cases. Either we attempt to coerce the legal rulings into a meaningful legal dialogue, or we disregard the cases as a whole. In either case it is necessary to go back to the original comment upon ‘like or similar products’ within the Working Party on Border Taxes’ 1970 report. the working party concluded that the term caused uncertainty, needed improvement. They want on to conclude that the concept of ‘like products’ would need to be decided on a case-by-case basis and this has been whole heartedly persued by the WTO panel rulings. Moreover the working group suggested that there were some criteria that could be taken into account when assessing the ‘similarity’ of products such as consumers’ tastes and habits, which change from country to country; the product’s properties, nature and

Quality. The wording of the report makes it clear that these criteria are not a recommendation but examples of what may be considered under a test for ‘like products’. pg.12 However there is clearly scope for these criteria to become a definite test upon which the WTO panel can rely. Some academics suggest that the Working Party’s suggestions have already become a framework adopted by the WTO. PG.8 but in reality the only set rule it has developed is the case-by-case approach Pg.12. The Working Party’s suggestions represent a real missed opportunity to develop a clear test based upon the Working Party’s list of both physical characteristics and competitive factors. find

ECONOMIC CONSIDERATIONS

Despite the apparent lack of a firm application by the WTO panels of the economic factors offered by the Working Group, there is still a wide spread use of them to some degree. What is now necessary is a clear list of economic considerations that must be taken into account when considering ‘like products’. Paramount of which is the concept of competitiveness.

Pg.4. – Many of the criteria of likeness that have been offered in GATT legal discussions of the “like product” concept can be viewed as overlapping variations on the idea of competitiveness. . First, there is substitutability — the extent to which consumers perceive two products as functionally equivalent, measured by the consumer’s willingness to substitute one for the other, a willingness which in turn is usually measured by the extent to which relatively small changes in price affect consumer preferences for one or the other. Next, there is concept of functional likeness, the extent to which the two products do in fact perform the same function, like sweeping dirt. Finally, although the producer-oriented provisions sometimes do employ “likeness” criteria that do not, strictly speaking, relate to the competitiveness of the goods in question — e.g., the extent to which two products are made from the same raw materials, in the same establishments, by the same capital goods, or by the same workers — the competitiveness criteria are still the first and most important factor in the “like product” decisions in those areas as well.4

Pg.5. – “Competitiveness” in this sense is best measured by the substitutability of the foreign product – the extent to which consumers are willing to choose the foreign product in substitution for the domestic product.

By the same token, the basic policy of Article III would not be served by a definition of “like

product” which limited that concept to products that had nearly identical physical characteristics. Such a narrow definition of “like product” would allow governments to give less favorable treatment to a foreign product that, although competitive with the relevant domestic product, had some different physical characteristics.

Pg.8 – Early GATT 1947 jurisprudence pertaining to Articles I (MFN) and III (NT) interpreted the concept of ‘like products’ on the basis of purely formal and objective criteria, mostly ignoring or even denying the relevance of competition.15

g.9 – Under the economic standard, the tertium comparationis is defined by economic parameters indicating the extent to which the market actors are in a competitive relationship. This standard was first applied by GATT 1947 panels for the ‘directly competitive or substitutable products’ element in Article III:2 GATT.

A TIGHTER DEFINITION?

Pg.8. – With regard to the internal measures covered by Article III, there really is no reason to

distinguish between the Article I definition of “like product” and the Article III definition of “like product.” The basic policy of Article III is the economic policy to eliminate market distortions caused by internal measures, and that economic policy is just as compelling with respect to National- Treatment-type discrimination as it is to MFN-type discrimination.9

pg.11 – To sum up, it is suggested that the GATT policy toward tariffs and tariff negotiations would justify a quite narrow interpretation of the “like product” concept with reference to claims of tariff discrimination under Article I:1 — an interpretation in which any objective difference between products would justify a separate tariff classification, and thus a different tariff rate. In contrast, we suggest, the “like product” concept should be given a to a broader interpretation — one that would prohibit product distinctions between directly competitive products — when applied to product distinctions made by internal measures. We suggest that such a broader definition be applied both under Article III and under that part of Article I:1 that applies to internal measures which discriminate between foreign countries. Such a broader interpretation would best be achieved by defining “likeness” in terms of the competitive relationship between the products in question, rather than requiring near identity of physical characteristics.

Pg.16. – The thesis of this paper is that product distinctions involving internal measures — internal taxes, subsidies, and various other regulations affecting internal sale — should be judged under a broader standard that focuses on the competitive relationship between the affected products and pays less attention to physical characteristics. As noted earlier, the particular two-sentence architecture of Article III:2 leads to a result consistent with this standard, but because of that architecture the “like product” concept has been relegated to a subordinate role in achieving that policy. Thus, it is only in Articles I:1 and III:4 in which the “like product” concept is employed to define the governing standard, and it is to these cases that one must look for broader interpretations of “like product” needed to carry out this anti-protection policy.

Pg.11 – 3. Subjective standard

The subjective standard of ‘likeness’ has been developed by different adjudicating bodies of

international economic law in order to balance the tension between international obligations

designed to liberalize trade and investment on the one hand, and domestic non-economic policy objectives such as environmental and consumer protection. The doctrinal reasoning of the subjective standard is to argue that the tertium comparationis is defined by the regulatory

purpose of the measure under scrutiny; for instance, if the measure is designed to protect the

environment, then the products are compared on the basis of their environmental impact. GATT 1947 jurisprudence implemented a subjective standard with the so called ‘aim and effects’ test as part of the ‘like products’ analysis.29

pg.12 – Following this approach a GATT Panel ruled that low and high alcohol content beers are not ‘alike’ for the purpose of Article III:4 GATT because the measures restricting points of sale, distribution and labelling were aimed to encourage the consumption of low alcohol beer. Conversely, wines made from different grapes were found to be ‘like products’ mainly because the respondent was unable to provide any valid public policy purpose in support of its differential tax treatment.30 However, subsequently the WTO panel and Appellate Body strongly rejected the ‘aim and effects’ test for purposes of both GATT and GATS.

Pg13 – 4. Combination of standards

The objective, economic and subjective standards of ‘likeness’ may be applied individually or in combination. WTO adjudicating bodies combine the economic and objective standard for purposes of certain GATT non-discrimination provisions. For instance, ‘like products’ in terms of Article III:1, first sentence, GATT is interpreted as requiring both physical similarity and a competitive relationship.

pg.3 – Thus the most important consideration in such cases is the “ relevant criteria” for comparison. Conceptually, one of the proposed approaches19 consist of the following steps: first, what non-discrimination requires is that like treatment be extended to all that are similarly situated with respect to the purpose of a given rule. Whether persons or things are similarly situated cannot be determined meaningfully by neutral or “ objective” standards. Instead, it must be determined by reference to the purpose of the particular rule under which a classification is made. Secondly, a criterion for classification is relevant only when it serves to further the goal of the particular rule. If the criterion does not further the goal, it is irrelevant and should not be used. The connection between the criterion and the asserted goal of the rule should at least be reasonable. Thirdly, unless a criterion for classification is indispensable in furthering the goal of the rule, in which case it must be used, whether to allow a relevant criterion for classification is a normative decision of the court. The court may consider norms and values extrinsic to the immediate purpose of the particular rule in deciding whether to permit a useful but not indispensable factor to be the basis for classification. Fourthly, the question of whether the purpose of a given rule is appropriate or legitimate entails a normative decision that can only be made in light of a norm beyond the framework of classification itself.

Pg.4 – Conceptually, two products are “ like” when the both are similarly situated with respect to the purpose of their comparison. As explained earlier the relevant criteria of comparison become all more important.22 Hence, in order to determine whether Japanese shochu and vodka, or a luxury car and an economy car, or beer of 6 per cent alcohol content and beer of 3 per cent alcohol content, are “ like products” , one must first know for what purpose they are compared and then apply criteria that are suitable to further that purpose. The entire issue of what constitute “ like products” , therefore, can be approached by applying the “ similarly situated” analysis explained above.

Pg.4 – With the above exposition of “ directly competitive or substitutable” product, one can approach the concept of like product in another way; this part underlines that way and further analyses the feasibility of such an approach.

Pg.4 – To begin with, all “ like products” should be understood as a subset of “ directly competitive or substitutable products” .23 In this light, all “ like” products are, by definition, “ directly competitive or substitutable” products, whereas the opposite does not always hold true.24 The obvious corollary to this proposition is that the two products in question are like products only when they are substitutable. As already stated, the GATT/WTO tribunals have shied away from clearly defining what a “ like product” is; one could start to posit a definition from the Japan–Alcoholic Beverages case, in which “ like” products were implied, in the context of art.III:2, as being “ products having no substantial noticeable differences in physical characteristics” .25 The “ like” product relationship stretches or squeezes within the boundary of physical similarity. Physical similarity is an attribute of products, whereas substitutability or competitiveness is an attribute of consumers or producers.

Pg.4 – Thus, in order to be “ like” , the two products being compared should have both physical *Int. T.L.R. 55 similarity and “ non-physical” competitiveness or substitutability.

Pg.4 – To rephrase, what is essential in the “ like” product concept is whether the two products in question share physical characteristics close enough to be considered virtually identical or substantially similar (so as not to be considered noticeably different). A caveat here is that differences in variety, price, quality, and freshness don’t make a difference when considering “ likeness” , as many GATT/WTO panels have emphasised.26

Pg.4 – Therefore, depending upon context and purpose of each provision of the WTO Agreements, the “ like” product concept is seen as a “ range of likeness” , from the narrowest to the broadest meaning of the term, stretching from the concept of “ identical” to the concept of “ remotely similar” as explained above.27

Pg.5 – To summarise, in general, like products should share at least one physical characteristic to be considered “ like” products in the broadest sense. If the products in question lose their similarity owing to processing or commercial manipulation, the same should be considered as taking them outside the scope of the “ like product” coverage. At this stage “ directly competitive or substitutable” relationships might be utilised to cover the situation.

Pg.5 – The next stage from here is to identify evidential elements which can be referred to in identifying such physical similarities or functional interchangeability between the products under consideration. One of the common denominators confirmed by the GATT/WTO Panels