The following inquiry into the practice of GATT and WTO dispute settlement intends to determine which approach has been prevalent in the judicial decisions of panels and the Appellate Body.
The review of dispute settlement reports of panels established under the GATT 1947 confirms that explicit discrimination on the basis of origin was a major concern in the early history of the GATT.31 It also becomes apparent that allegations of de facto discrimination nevertheless surfaced very early. Those disputes show that contracting parties brought such complaints only against measures resulting in significant asymmetry. Due to an occasionally narrow interpretation of "likeness", the panels did not arrive at the stage where they could have applied the diagonal or the asymmetry approach.
The dispute about the Australian Subsidy on Ammonium Sulfate 32 dealt with the introduction of a less advantageous treatment of sodium nitrate fertilizers compared to ammonium sulfate. Chile had based its violation claim on GATT Article I:1, since its fertilizer exports consisted of sodium nitrate. In other words, the facially neutral differentiation placed Chile de facto at a disadvantage in relation to those other countries which did not predominantly export sodium nitrate. The working party dismissed this violation claim because it did not find the two types of fertilizers to be "like products". Hence, the working party did not have to examine the alleged denial of most-favored-nation treatment. Under the assumption of the two fertilizers being "like", the working party could have found a violation of Article I:1 both under the diagonal and the asymmetric impact approach given the factual asymmetry.
In the Sardines dispute 33 Norway's violation claim (GATT Articles I:1 and XIII) was based on the different tariff rates, internal taxes and quantitative restrictions Germany applied to the Portuguese and Scandinavian types of canned sardines. For procedural reasons, the panel did not establish whether the products were like. Thus, no conclusions can be drawn from this case, but the factual distribution of the differently treated types of imports between Portugal and Norway was clearly asymmetrical.
In 1956, the U.S.A. brought a complaint against the graduated French tax on automobiles. 34 The violation claim related to GATT Article III:2 and the fact that the French formula resulted in a considerably higher tax on the larger, mainly U.S. automobiles. On the same grounds, the U.S.A. could have relied on Article I:1, 35 comparing the effective tax burden on U.S. automobiles with other car imports. France defended the tax as a sort of luxury tax, but the complaint did not result in a ruling. This case again involved factual asymmetry.
The United States complaint against the Chilean auto taxes 36 also remained without a ruling. The sharply graduated tax imposed a disproportionate burden on U.S. automobiles compared with the cheaper Chilean models. Since the U.S. claim did not cite GATT Article I or III, but relied on impaired tariff concessions, it has been suggested that the claim was non-violation nullification or impairment. 37 However, the U.S.A. could have relied on a violation of these two provisions, given that the facts met the conditions of both approaches, diagonal and asymmetric impact.
The European regulation at issue in the transatlantic dispute on animal feed proteins 38 included, among other things, a formally origin-neutral distinction between types of feed proteins commonly produced in the European Communities and those commonly produced in other countries. The panel's ruling denied the alleged violations of GATT Articles I:1 and III:4 on the basis that the different types of feed were not "like products". A violation of Article III:5 and the principle of Article III:1 resulted from the absence of substantial domestic production of several of the disadvantaged products.
The U.S. complaint against Spain in the Soyabean dispute related to internal restrictions on the sale of soybean oil. 39 Those limits did not apply to other vegetable oils and were intended to protect the more expensive domestic olive oil. The panel adopted a very restrictive interpretation of "like" as "more or less the same product". It therefore did not find the internal quantitative restriction to be a violation of GATT Article III:4, ruling that the other vegetable oils treated more favorably were not "like products".
Another such unsuccessful complaint, again in the context of GATT Article I:1, was the dispute between Canada and Japan over the different import duties applied to dimension lumber. Japan had a zero tariff on dimension lumber made of hemlock-fir, the kind typically imported from the U.S.A., but maintained a custom duty of 8 % on the sort typically exported by Canada. Canada argued that the tariffs were designed so that a considerable part of its lumber exports fell under the less favorable category. The panel concluded that the different kinds of dimension lumber were not "like products". 40
Of higher probative, though not dispositive value are dispute settlement reports which scrutinized measures with asymmetric impact and resulted in violation findings. Panels and the Appellate Body relied on the asymmetry to a varying extent, be it in connection with border measures, taxation or regulation.
An early example in the context of GATT Article I:1 is the Spanish Coffee case. Departing from the practice of most countries, Spain had introduced different tariff rates on different kinds of unroasted, non-decaffeinated coffee beans. The panel's first conclusion in assessing most-favored-nation treatment was that the various types should be regarded as "like products". The panel then noted that Brazil mainly exported to Spain the types falling within the higher duty category and concluded that the new Spanish tariff scheme discriminated against unroasted coffee originating in Brazil. 41 This last statement is noteworthy because the panel did not merely conclude that Spain failed to extend the more advantageous treatment to the more highly burdened sub-categories within the like products. Instead, the panel established a link between the disfavored types and their predominant presence among goods originating in Brazil. On that basis, the panel labeled the treatment of the entire group of unroasted coffee beans as discriminatory. This is a clear example of the application of the asymmetric impact test.
In Canada Automobiles, the WTO witnessed another violation of GATT Article I:1 through a formally origin-neutral measure. Canada accorded duty-free treatment to motor vehicles imported by certain manufacturers producing automobiles in Canada. The list of those companies was closed as of 1989. The main beneficiaries were the principal United States auto manufacturers, and in practice they imported only their own automobiles duty-free. This case is particular in that the legal differentiation was not based on product characteristics and the list of beneficiaries was closed. Nevertheless, it is interesting that the panel inquired into the source countries of the imports, which in practice benefited from the duty exemption. It found that the regime favored products of certain origins and concluded that Canada did not accord the advantage on equal terms to like products of different origin. 42 The panel based its decision on the discriminatory effects of the measure with respect to origin and did not merely look for some like product from another country.43
The Appellate Body supported this finding on the basis of both the text and the practical operation of the facially origin-neutral measure. It concluded that Canada had granted an advantage "to some products from some Members" and not "to `like' products `originating in or destined for the territories of all other Members.'" 44
In one of the few GATS cases to date, the Bananas III dispute, both the panel and the Appellate Body based their reasoning on the asymmetric impact of the measures adopted by the European Community (EC). 45 In relation to the most-favored-nation clause of GATS Article II:1, one violation consisted in formally origin-neutral rules differentiating between operator categories. The criteria for the subdivision of wholesalers were based on past trade patterns and resulted in most wholesale service suppliers from African, Caribbean and Pacific countries being classified in the more favorable operator category. In contrast, most operators of the complainant's origin fell within the less advantageous group.46
As regards the national treatment obligation of GATS Article XVII, the panel in EC - Bananas III found that various aspects of the EC's banana regime modified the conditions of competition between domestic and foreign service suppliers to the detriment of the latter. It similarly based this finding on a thorough quantitative analysis of the share of both EC and third-country service suppliers within the formally origin-neutral categories which received different treatment under the regime. 47 The Appellate Body confirmed this approach. 48
The first Japanese Alcohol dispute 49 was the earliest case of an origin-neutral tax resulting in a finding of national treatment violation. The law at issue imposed different tax rates according to types and sub-types of alcoholic beverages. Exactly the same rates applied to both imported and Japanese products and all categories actually included goods of both origins. However, the majority of domestic drinks fell into categories with low taxes, and many imported alcoholic beverages disproportionately often came within categories with high rates.
Strictly read, the panel assessed the national treatment conformity of the national tax scheme in a somewhat ambiguous manner as regards the question of diagonal versus asymmetry test. In respect of GATT Article III:2, first sentence, the panel found in a diagonal manner that special and first grade whiskies and brandies imported from the European Economic Community (EEC) were taxed in excess of the like domestic first and second grade whiskies and brandies. However, just before, the panel had noted that almost all the imported whiskies and brandies were subject to the higher rates, whereas the majority of like domestic products enjoyed the lower rates. 50
Similarly, the panel condemned the mixed system of specific (up to a certain selling price) and ad valorem taxes (above the threshold) without explicit inquiry into the predominant origin of the goods actually falling within these tax categories. However, it also observed that Article III:2 does not prohibit differentiated taxes as such. 51
As to the higher rates on beverages with high raw material content, the panel assessed the excess taxation in a diagonal manner. It also referred to the exceptions in Articles II:2 and XX in support of its statement that different tax treatment of like products was not always GATT incompatible. On the other hand, the panel noted the EEC's argument that almost all its liqueurs faced the higher rates whilst some Japanese liqueurs benefited from lower rates. 52
Against this background, it is questionable whether this part of the report truly suggests a diagonal product comparison under Article III:2, first sentence. 53 On the contrary, Professor Hudec has deduced from the case that a finding of excess taxation requires that most imports within the "like product" group fall into the high-tax category and that the low-tax category contains most of the domestic goods. 54
As regards the second sentence of Article III:2, the panel again expressly relied on the fact that the more favorably treated shochu was almost exclusively produced in Japan in order to establish that Japan afforded protection to domestic production, "rather than to the production of a product produced in many countries ... in relation to another product". 55 Thus, the panel applied an unequivocal asymmetric impact test. 56
Japan replaced the GATT incompatible regime with a new tax system, but again designed it with different beverage categories in such a manner that imported drinks were predominantly subject to higher specific tax rates than like domestic products. Whether these higher specific rates truly resulted in a disadvantage in competitive opportunities expressed in the price ratio of the different drinks, remained somewhat doubtful. This is, however, immaterial for the purpose of exploring in what manner the sub-categories were compared in the Article III:2 assessment.
In the new dispute, the panel confined itself to establishing that vodka was obviously taxed in excess of shochu, without even adding the attributes imported/domestic. 57 In respect of Article III:2, second sentence, the panel similarly established that the tax rates applicable to the different directly substitutable drinks at issue were dissimilar by more than de minimis. Hence, Japan afforded protection to shochu in contradiction to the GATT. Again, the panel did not add the attributes of origin contained in GATT Article III:2 and, more importantly, the comparison did not take place between the entire groups of domestic and imported substitutable goods. Only after making these findings did the panel refer to the fact that the favored "shochu is essentially a Japanese product." 58
The reasoning, therefore, does not reflect the asymmetric impact approach, 59 although the facts certainly allowed it 60 and the European Communities had expressly relied on it and accordingly furnished market share numbers. 61 Interestingly, the U.S.A. who won at the panel stage appealed and referred precisely to the dilemma of the diagonal test, that tax distinctions among like products automatically result in a violation of Article III:2, first sentence. 62 Rather than promoting an asymmetric impact test, the appeal in vain pushed for the adoption of the "aims and effects" approach to remedy the dilemma.
The Appellate Body 63 also did not significantly rely on the factual asymmetry in the tax treatment of imports and domestic goods as a whole. 64 The report neither provides support for the diagonal test, except for the fact that the Appellate Body upheld the panel's reasoning, 65 which however should not be overestimated given the facts of the case. 66
In Korea - Alcoholic Beverages, the panel heavily relied on the factual asymmetry in establishing protection to domestic production. The panel pointed out that there was "virtually no imported soju" so that the beneficiaries of the tax structure were "almost exclusively domestic producers" whilst less than 1 % of domestic production faced the higher tax. 67 The Appellate Body equally relied on the fact that "the tax operates in such a way that the lower tax brackets cover almost exclusively domestic production, whereas the higher tax brackets embrace almost exclusively imported products". 68
The pisco dispute might be the classic example in which the asymmetric impact of the measure under scrutiny played the most significant role in the reasoning of both the panel and the Appellate Body. In assessing whether the Chilean tax system was "so as to afford protection" the panel asked whether imports or domestic products received the benefit of the dissimilar taxation. In view of the distribution of domestic and imported products across the tax categories, namely the fact that pisco was exclusively produced domestically, the panel found that the beneficiary under the transitional system was the domestic industry. 69
The panel further established that the new system introduced unfavorable
treatment for products which happened to be primarily imports. Its design,
structure and architecture were such that 70-80 % of the Chilean
production enjoyed the lowest tax rate whereas "almost 95 % of current
imports will be taxed at the highest rate". On that basis the panel concluded
that the new law was a de facto discriminatory system. 70
The Appellate Body applied the asymmetry test already under "not similarly taxed" 71 in a "comprehensive examination" of the respective tax burdens on "all of the directly competitive or substitutable domestic and imported products". 72 In determining that the law was applied "so as to afford protection", the Appellate Body insisted on equal competitive relationships for "all directly competitive or substitutable imported products in relation to domestic products", and not simply for "those imported products within a particular fiscal category". Accordingly, the Appellate Body noted that the highest tax bracket contained approximately 95 % of the imports, whereas 75 % of all domestic production faced the lowest tax rate. 73
While neither the panel nor the Appellate Body expressly mandated an asymmetric impact test, both heavily relied on the discriminatory overall effect of the Chilean legislation. 74
An example of another WTO rule on non-discrimination where the asymmetry question surfaced is Article 5.5 of the Agreement on the Application of Sanitary and Phytosanitary Measures (SPS Agreement). 75 In the EC - Hormones dispute, the panel established that the European ban on meat treated with growth hormones de facto discriminated against U.S. meat. The basis for that conclusion was that, not only during but already prior to the ban, the percentage of animals treated with such hormones "was significantly lower in the European Communities than in the United States." 76
Several aspects of the United States - Malt Beverages dispute raised the issue of discrimination through formally origin-neutral rules. The panel report is among the few atypical ones to have applied the "aims and effects" test to origin-neutral measures. This means that the GATT Article III test was modified and the likeness of differently treated products also depended on the presence of some kind of protective effect and aim. 77
One of the tax measures under scrutiny was the Mississippi wine tax which favored wine made from a grape type growing only in the Southeastern U.S.A. and the Mediterranean. The panel pointed to the absence of this grape type in other regions and considered this as an implied geographical distinction. It consequently found the tax to afford protection to domestic production. 78 This implied or indirect geographical differentiation is a very clear case of asymmetric impact. Wines imported from other countries usually do not contain the Mississippi grape.
In the absence of reports explicitly mandating an asymmetry requirement, only the dismissal of a complaint could serve as compelling evidence for such a condition. It appears that only two non-representative decisions relied, to some extent, on lacking asymmetry in order to conclude that the measure at issue did not violate GATT Article III.
The first such decision is the Malt Beverages report mentioned above. The dispute also concerned the different treatment of beer according to alcohol content for purposes of sale and distribution in various states of the U.S.A. (Article III:4). Assessing whether the measures were applied so as to afford protection to domestic production, the panel remarked that both imported and domestic goods contained both high and low alcohol beer. "The burdens resulting from these regulations thus do not fall more heavily on Canadian than on United States producers." 79 In rejecting the Article III:4 violation claim this denial of asymmetry, however, was only one factor together with remarks on the legislative purpose. Based on this "aims and effects" approach, the discussion of "so as to afford protection" was part of the "likeness" test. Even if this discussion of protection to some extent relied on an asymmetry requirement, the panel simultaneously assumed that once products are like, any product differentiation becomes inconsistent with Article III. 80 Taken in isolation, this obiter dictum is reminiscent of the diagonal approach.
Two of the three taxes under consideration in United States - Taxes on Automobiles were, from a formal point of view, entirely origin-neutral: the graduated Gas Guzzler tax on vehicles achieving less than 22.5 miles per gallon and the luxury tax on cars sold above 30,000 USD. It was alleged that, in the first year after the introduction of the luxury tax, 80 % of the cars subject to it were imports. Despite those alleged facts, the panel was not convinced that the regime favored U.S. cars, since many imported cars sold below the threshold and an imbalance could have been due to marketing, production or consumers' decisions. The panel finally declined to make a finding of protection for domestic automobiles, because a price above the threshold was not "inherent" to imported cars. Foreign manufacturers had the capabilities to produce smaller cars and did so for other markets. 81 In relation to the fuel efficiency tax, the panel stated that, irrespective of trade flow data, the low fuel economy was not "inherent" to foreign automobiles. It also stated that the exempt light trucks, a production dominated by domestic manufacturers, were not "inherently" of domestic origin. 82
This panel report, which has not been adopted, is as atypical as the report in Malt Beverages in that it followed the "aims and effects" approach and, therefore, examined "afforded protection" as part of "like product". In that examination, however, the panel applied a very strict test of "inherent" asymmetry. 83
A problematic finding of the Malt Beverages panel concerned Minnesota's tax advantage for small beer producers. For the panel it did not matter whether the tax credit applied to foreign beer, 84 which, although contentious, was arguably the case. Applying a diagonal test, 85 the panel confined itself to holding that imported beer from large producers was like beer from small domestic producers and that the former was taxed in excess of the latter. 86 Although the panel did not rely on the factual distribution of domestic and imported goods across the tax categories, one may defend an asymmetry in this context. Small beer producers usually have their main market in the region. The share of imports among beer from small brewers should, therefore, be smaller than among beer from large manufacturers. 87 An additional implication of this part of the case is that the distinction was not based on product characteristics.
Another case, which seems to disprove the hypothesis that asymmetric impact is a necessary requirement for a GATT Article III:2 violation, is the Canada - Periodicals dispute. The measure at issue was a prohibitive Canadian tax on advertising revenues from the Canadian edition of so called "split-run" magazines, i.e. magazines issued with similar editorial content in other countries. The intention of this rule applying both to Canadian and to foreign magazines was to protect Canadian periodicals against competition from United States publications containing advertisements specifically targeting Canadian consumers.
Before the panel, Canada pointed out that the tax on split-runs also covered Canadian magazines and that in fact one such magazine had ceased production of its edition for the United States. On that basis, Canada strongly argued in favor of comparing domestic and like imported goods as classes, thus an asymmetry requirement, lest distinctions within like products should automatically run afoul of Article III:2, first sentence, even if imports on the whole benefited. 88 The United States did not principally oppose this approach, but disputed the absence of asymmetric impact. It considered the additional (GATT Article XI inconsistent) import ban on split-run magazines to be responsible for the fact that imports might not be preponderant among split-runs. 89
The panel, having established the likeness of both split-run and non-split-run periodicals, determined that Canada evidently imposed on imported split-runs taxes in excess of those applied to domestic non-split-run periodicals. On the grounds of this diagonal comparison the panel found the tax to breach GATT Article III:2, first sentence. 90
Literally read, the Appellate Body to some extent followed this comparison "between imported split-run periodicals and domestic non-split-run periodicals" when assessing under the second sentence of Article III:2 whether the dissimilar tax was so as to afford protection to domestic production. However, its final conclusion was that Canada afforded protection to domestic periodicals, not that it protected domestic non-split-run periodicals. 91
Strictly speaking, such a conclusion in relation to domestic periodicals, as a whole, presupposes a certain form of asymmetric impact on imported substitutable magazines. Admittedly, it remained unclear in the decision whether imported magazines would indeed more frequently be split-runs than domestic editions, 92 given that the report only refers to one United States and one Canadian magazine which both ceased their split-run edition. 93 However, looking at the peculiarities of the North American market, there are good reasons to assume, as the U.S.A. suggested, 94 that in the absence of both ban and tax, split-run editions on the Canadian market would disproportionately often be U.S. magazines. Irrespective of this speculation, the tax on split-run editions is noteworthy in that it does not distinguish according to physical characteristics of the product, but on the basis of the destination of a part of the production. Canada imposed the tax if the product of whatever origin was also sold abroad. The tax, therefore, did not prevent imports together with the sale of domestic products, as is typically the case with origin-neutral legislation, but imports together with exports. In other words, lifting the tax (and the ban) would not result in easier imports together with easier domestic production of the previously restricted kind for the domestic market. Thus, Canadian magazines would face additional competition from U.S. split-runs, but there would not be additional (potential) competition from domestic editions. This might not be accurate in absolute terms because, without the restrictions, new editors might establish in Canada with a view to selling split-run editions in both countries, and this effect is greater than across the border. Nonetheless, the assumption should still hold as a tendency, so that there would not be an equivalent amount of additional domestic competition. Following that reasoning, the tax on split-run editions did have an asymmetric impact on imported magazines. 95
The recent Asbestos dispute is highly interesting given the contrast between the reports of the panel and the Appellate Body, not only in relation to "likeness", but also as regards "less favorable treatment". The panel, after finding Asbestos products to be like the domestic substitutes (PVA, cellulose or glass fiber and fibro-cement), established less favorable treatment in a bluntly diagonal manner. It contented itself with the trivial observation that the French asbestos ban de jure treated imported asbestos products less favorably than the like domestic substitutes. The distribution of imported and domestic goods across the two regulatory categories, banned and allowed, in other words, the relative incidence of the ban, received no attention whatsoever. The panel was the first to draw the logical conclusion from the diagonal approach: that the origin-neutral distinction amounted to de jure discrimination. 96 Usually, only the concept of de facto discrimination occurs in the context of formally origin-neutral rules. However, on the basis of the diagonal approach, the distinction in the law itself triggers the discrimination finding, not the factual characteristics of the goods of different origin. 97
Considering the undisputed health risks of asbestos products, the Appellate Body denied their likeness compared to the safer substitute products and in this respect reversed the panel's findings. The question of less favorable treatment, therefore, was no longer relevant in this appeal. Nevertheless, the Appellate Body dedicated the following obiter dictum to this condition:
100. We recognize that, by interpreting the term "like products" in Article III:4 in this way, we give that provision a relatively broad product scope - although no broader than the product scope of Article III:2. In so doing, we observe that there is a second element that must be established before a measure can be held to be inconsistent with Article III:4. Thus, even if two products are "like", that does not mean that a measure is inconsistent with Article III:4. A complaining Member must still establish that the measure accords to the group of "like" imported products "less favourable treatment" than it accords to the group of "like" domestic products. The term "less favourable treatment" expresses the general principle, in Article III:1, that internal regulations "should not be applied ... so as to afford protection to domestic production". If there is "less favourable treatment" of the group of "like" imported products, there is, conversely, "protection" of the group of "like" domestic products. However, a Member may draw distinctions between products which have been found to be "like", without, for this reason alone, according to the group of "like" imported products "less favourable treatment" than that accorded to the group of "like" domestic products. In this case, we do not examine further the interpretation of the term "treatment no less favourable" in Article III:4, as the Panel's findings on this issue have not been appealed or, indeed, argued before us. 98
The penultimate sentence of this paragraph, the statement "... without, for this reason alone ...", perfectly matches the question whether the diagonal test is accurate: under the diagonal approach applied by the panel, a distinction between like products is a sufficient condition for less favorable treatment. There only needs to exist some national goods in the sub-group receiving the better treatment and some (potential) like imports in the sub-group receiving the worse treatment. Instead of diagonally comparing sub-groups, the Appellate Body kept comparing the treatment accorded to the "group of `like' imported products" with the treatment of the "group of `like' domestic products". The group in singular implies that there is only one group on each side which must, therefore, include all of the like products, whether favored or disfavored, under the origin-neutral measure. 99 Applying such an aggregate comparison, it makes perfect sense that a distinction between like products does not necessarily result in less favorable treatment of the entire group of imports, although some like imported products (necessarily) do receive worse treatment than some like domestic goods. For instance, there is no less favorable treatment, when the quantitative ratio between favored and disfavored products is equivalent on both sides and the measure imposes no burden which is qualitatively heavier for foreign goods (asymmetric intensity 100). The use of italics to designate and emphasize the origin of goods further supports this reading that the burden must specifically affect imports, thus the discriminatory effects must indirectly be related to origin. 101
This interpretation accords well with the semantic equivalence the Appellate Body gives to "less favorable treatment" and "protection to domestic production", the paradigms of GATT Article III:4 and Article III:1 respectively. In particular, the Appellate Body again speaks of "`protection' of the group of `like' domestic products", i.e. the entire group of all foreign goods meeting the requirements of likeness.
The Appellate Body included paragraph 100 recognizing the broad scope it gave to the "like product" scope. It appears to wish to balance this broad scope by emphasizing "less favorable treatment" as a second condition in GATT Article III:4. Contrary to the diagonal approach, only the asymmetric impact test truly represents an additional legal requirement and is able to function as a filter. This supports the hypothesis that the Appellate Body intended to reject the diagonal approach for the "less favorable treatment" comparison.
However, other interpretations of the Appellate Body's obiter dictum have been suggested. For instance, the statement that distinctions between like products do not necessarily amount to less favorable treatment of like imports has been interpreted as a mere restatement of an earlier proposition: in U.S. - Section 337 the panel stated that different treatment of imports need not be less favorable, because it could also be more advantageous. 102 However, the origin-neutral asbestos ban at issue certainly did not suggest any possible advantage for the whole group of like imports compared to their domestic counterparts. In order to make that reading plausible, one therefore has to isolate the Appellate Body's statement from the case and also from its context in paragraph 100, namely the paradigm of group comparisons. More importantly, the described reading also reduces the penultimate sentence of the obiter dictum to the trivial statement that "different" need not be worse, since it could also be better or equally good. It is doubtful that the Appellate Body made the effort of an obiter dictum for such a purpose. 103
For the same reasons, paragraph 100 is not a restatement of an Appellate Body finding in Korea - Beef. 104 The measure at issue in that case provided for separate channels of distribution for imported and domestic goods, i.e. butchers selling domestic beef were banned from simultaneously selling imported beef and vice versa. Relying on the finding of the panel in U.S. - Section 337, 105 the Appellate Body stated that a formal distinction between imports and domestic goods is neither necessary nor sufficient for a violation of GATT Article III:4, because the treatment must be "less favorable", i.e. must modify the conditions of competition to the detriment of imports. 106 It was clear that the Korean law, unlike the French asbestos ban, provided for a formal distinction in the treatment of imported and domestic goods. Since a difference in itself is not necessarily a disadvantage, 107 the Appellate Body justified its finding of less favorable treatment with a more elaborate analysis of the distinction's impact on the market. 108 Hence, also in the light of Korea - Beef, paragraph 100 of the Appellate Body Report should not be reduced to the statement that "less favorable treatment" requires, in addition to a distinction of treatment, that the different treatment of imports be less favorable. 109
Other commentators suggest that paragraph 100 supports the "aims and effects" approach. 110 Admittedly, this approach also avoids automatic Article III violations through origin-neutral national regulation. A supporting indication might be the Appellate Body's reference to GATT Article III:1, one of the provisions in the GATT where the "aims and effects" test has been located. This argument, however, is not compelling because, as witnessed in various GATT disputes, "so as to afford protection" can have other meanings than "aims and effects". More importantly, the Appellate Body interprets the meaning of "less favorable treatment" and coming from that concept reaches the conclusion of protection. Proponents of the "aims and effects" test have usually located that test within the requirements "like" and "so as to afford protection". Under the "aims and effects" approach, the products in question are no longer considered to be like, or the measure to afford protection, once the legislative purpose has been accepted as non-protective.
However, once likeness has been established according to the two-stage approach practiced in the WTO jurisprudence, it is not apparent how this prior finding of likeness can be reversed at the subsequent stage of the legal examination, i.e. less favorable treatment, in view of the aim of the measure. In "less favorable treatment", the criterion at issue in paragraph 100, it also seems impossible to consider, in addition to the design and the effect of the measure, whether its aim is legitimate enough to offset a discriminatory effect. Thus, the "less favorable treatment" test operates in an objective manner irrespective of legislative purpose. Its result fundamentally depends on how imported and domestic goods are compared and what kind of discriminatory effects the measure displays. It is precisely the thesis of this paper that and why it matters so essentially how one compares the treatment accorded to imported and like domestic products, namely in a diagonal manner or by comparing the groups of like products as aggregates.
A final comment on the status of paragraph 100 as obiter dictum: the fact that this statement is not part of the ratio makes it in no way irrelevant. Even the ratio of the decisions is not formally binding as precedent in the WTO. The obiter dictum indicates that the Appellate Body voluntarily and deliberately intended to make a point. One may, therefore, expect that the passage will gain significance in the future. 111 The proposed interpretation of this part of the report, that is, the requirement of an asymmetric impact 112 would make it remarkable given the background of the panel report and the ambiguous preceding jurisprudence highlighted above. Only future cases, however, can give final clarity on the Appellate Body's exact meaning.
It has been suggested that the approach of paragraph 100, i.e. the comparison between the two entire groups of like products, and hence the asymmetric impact test, departs from established GATT jurisprudence on Articles III:4 and III:2. 113 The panel in U.S. - Section 337 understood "the `no less favourable' treatment requirement of Article III:4 [to be] applicable to each individual case of imported products." Since then, panels and the Appellate Body have consistently rejected "any notion of balancing more favourable treatment of some imported products against less favourable treatment of other imported products." 114
In a certain way, the comparison of the entire groups of like products amounts to such balancing, because those two groups contain goods which benefit as well as goods which are disadvantaged by the measure. The requirement of asymmetric impact implies that disadvantages for one sub-group can be offset with advantages for another sub-group, provided that the sub-groups are equally prevalent among imported and domestic goods. However, the Section 337 dispute dealt with a rule applicable only to goods of foreign origin, thus a case of overt discrimination, 115 not one of de facto discrimination. 116 Contrary to facially discriminatory measures, and with the exception of measures imposing a burden that is qualitatively heavier for imported goods, origin-neutral rules by definition never provide for less favorable treatment within the same regulatory sub-group because imports and domestic products receive identical treatment. 117 Hence, the rejection of the "on the whole" comparison in this jurisprudence, i.e. the inadmissibility of "balancing" and the insistence of equally favorable treatment for every products, cannot simply be transferred to origin-neutral rules. 118
Arguably, it is always legitimate to compare directly the treatment of imported and domestic goods within sub-groups of likeness. Simultaneously, this defends the application of the aggregate group comparison to origin-neutral rules, but not to origin-based ones. In addition, no one would argue in favor of balancing origin-based disadvantages with advantages for some of the affected goods flowing from different laws. Why should it be different, if the source of the advantage is another aspect of the same law which fails to offset the less favorable elements in the case of every 119 import? Hence, there is no contradiction between the asymmetric impact requirement and the "no balancing"-jurisprudence on origin-based discrimination. Neither can this proposition serve as a basis for the diagonal approach in origin-neutral contexts. 120
The review of the jurisprudence reveals significant inconsistencies in a relative sense, but displays few of the extremes that are theoretically conceivable under the two approaches. Panels and the Appellate Body hardly ever applied a straightforward diagonal test. 121 The reasoning, although not the facts, in the Asbestos panel report, in Canada - Periodicals and one finding in Malt Beverages probably come the closest. Without being explicitly declared a necessary condition for a violation, asymmetry plays a significant role in most of the relevant decisions. Hinting at a requirement of some asymmetry, the Appellate Body Report in Asbestos could be a further step in that direction.
31 Robert E. Hudec, GATT/WTO Constraints on National Regulation: Requiem for an "Aim and Effects" Test, 32 Int'l Law. 619, 622 and note 8 (1998); Robert Howse & Elisabeth Tuerk, The WTO Impact on Internal Regulations - A Case Study of the Canada - EC Asbestos Dispute, in The EU and the WTO: Legal and Constitutional Aspects 283, 284-285 (Grainne De Burca and Joanne Scott eds., 2001).
32 Working Party Report, The Australian Subsidy on Ammonium Sulphate, adopted 3 April 1950, BISD II/188.
33 Panel Report, Treatment by Germany of Imports of Sardines, adopted 31 October 1952, BISD 1S/53.
34 L/520, 12 September 1956; SR.11/16, 26 November 1956, at 163.
35 John H. Jackson, World Trade and the Law of GATT 258-259 (1969).
36 L/599, 16 November 1956; SR.11/18, 29 November 1956, at 212.
37 Robert E. Hudec, Enforcing International Trade Law: The Evolution of the Modern GATT Legal System 440 (1993).
38 Panel Report, EEC - Measures on Animal Feed Proteins, adopted 14 March 1978, BISD 25S/49.
39 Panel Report, Spain - Measures Concerning Domestic Sale of Soyabean Oil, L/5142, 17 June 1981, unadopted.
40 Panel Report, Canada/Japan - Tariff on Imports of Spruce, Pine, Fir (SPF) Dimension Lumber, adopted 19 July 1989, BISD 36S/167.
41 Panel Report, Spain - Tariff Treatment of Unroasted Coffee, adopted 11 June 1981, BISD 28S/102, paras. 4.5, 4.9 and 4.10.
42 Panel Report, Canada - Certain Measures Affecting the Automotive Industry, WT/DS139/R, circulated 11 February 2000, paras. 10.40-10.48; see also paras. 10.261-10.262 in relation to GATS Article II:1.
43 See also Edmond McGovern, International Trade Regulation 8.13-2 (1995-November 2000).
44 Appellate Body Report, Canada - Certain Measures Affecting the Automotive Industry, WT/DS139/AB/R, WT/DS142/AB/R, adopted 19 June 2000, paras. 80, 81 and 85.
45 Robert E. Hudec, supra note 31, at 639-641; Ole K. Fauchald, supra note 9, at 251, note 97.
46 Panel Report, EC - Regime for the Importation, Sale and Distribution of Bananas (EC - Bananas III), WT/DS27/R/USA, circulated 22 May 1997, DSR 1997:II, 943, paras. 7.349, 7.384, 7.396; and, in abstracto, Appellate Body Report, EC - Bananas III, WT/DS27/AB/R, adopted 25 September 1997, DSR 1997:II, 591, paras. 229-234.
47 Panel Report, EC - Bananas III, supra note 46, paras. 7.332-7.338 and 7.363-7.368 (quota tariff rate import licenses), 7.378-7.380 (exemption of Category B operators from BFA export certificate requirement), 7.392-7.393 (allocation of hurricane licenses).
48 Appellate Body Report, EC - Bananas III, supra note 46, paras. 243-244, 246.
49 Panel Report, Japan - Customs Duties, Taxes and Labelling Practices on Imported Wines and Alcoholic Beverages, adopted 10 November 1987, BISD 34S/83.
50 Id., para. 5.9 lit. a).
51 Id., para. 5.9 lit. b).
52 Id., para. 5.9 lit. d).
53 As Ole K. Fauchald, supra note 9, at 221, suggests by concluding that there is a possibly irrefutable "violation presumption", once the differently treated products are found to be "like", see id., at 221, 223.
54 Robert E. Hudec, supra note 37, at 539. See also Rex J. Zedalis, A Theory of the GATT "Like" Product Common Language Cases, 27 Vand. J. Transnat'l L. 33, 92-93 (1994).
55 Panel Report, 1987 Japan - Alcoholic Beverages, supra note 49, para. 5.11.
56 See also Ole K. Fauchald, supra note 9, at 229, 246.
57 Panel Report, Japan - Taxes on Alcoholic Beverages, WT/DS8/R, WT/DS10/R, WT/DS11/R, circulated 11 July 1996, DSR 1996:I, 125, paras. 6.24, 6.27.
58 Id., paras. 6.33 and 6.35.
59 As suggested by William J. Davey & Joost Pauwelyn, supra note 8, at 50, note 119 ("to some extent").
60 See also Daniel A. Farber, Environmental Federalism in a Global Economy, 83 Va. L. Rev. 1283, 1299-1300 (1997).
61 Panel Report, Japan - Alcoholic Beverages, supra note 57, para. 4.95.
62 United States' appellant's submission of 23 August 1996, para. 36.
63 See Appellate Body Report, Japan - Taxes on Alcoholic Beverages, WT/DS8/AB/R, WT/DS10/AB/R, WT/DS11/AB/R, adopted 1 November 1996, DSR 1996:I, 97, 115 (section H.1.(b)) and 119-123 (section H.2.(c)).
64 Ole K. Fauchald, supra note 9, at 246, in so far expresses the opposite view that the Appellate Body Report emphasizes an asymmetry test. The panel in Indonesia - Automobiles describes the question decided in Japan - Alcoholic Beverages relying on a diagonal approach, see Panel Report, Indonesia - Certain Measures Affecting the Automobile Industry, WT/DS54/R, WT/DS55/R, WT/DS59/R, WTDS64/R, adopted 23 July 1998, para. 14.112.
65 Appellate Body Report, Japan - Alcoholic Beverages, supra note 63, at 115 (section H.1.(b)). As regards the language the Appellate Body uses, there is a certain inconsistency and tension between the introduction to the application of Article III:2, first sentence, and the subsequent application of this test: "if the taxes applied to the imported products are `in excess of' those applied to the like domestic products ..." - "The only remaining issue under Article III:2, first sentence, is whether the taxes on imported products are `in excess of' those on like domestic products." See id., at 112 (section H.1) (emphasis added) and 115 (section H.1.(b)).
66 See also Ole K. Fauchald, supra note 9, at 223.
67 Panel Report, Korea - Taxes on Alcoholic Beverages, WT/DS75/R, WT/DS84/R, circulated 17 September 1998, para. 10.102 with note 414.
68 Appellate Body Report, Korea - Taxes on Alcoholic Beverages, WT/DS75/AB/R, WT/DS84/AB/R, adopted 17 February 1999, para. 150.
69 Panel Report, Chile - Taxes on Alcoholic Beverages, WT/DS87/R, WT/DS110/R, circulated 15 June 1999, paras. 7.123, 7.128-7.129.
70 Id., paras. 7.150, 7.155; see also paras. 7.158, 7.159.
71 Other reports have declined to do so, but since "dissimilar taxation" and "afforded protection" are cumulative requirements, this question is secondary. Such other reports are: Appellate Body Report, Canada - Certain Measures Concerning Periodicals, WT/DS31/AB/R, adopted 30 July 1997, DSR 1997:I, 449, 474 (section VI.B.2); Appellate Body Report, Japan - Alcoholic Beverages, supra note 63, at 118-119 (section H.2.(b)); Panel Report, Korea - Alcoholic Beverages, supra note 67, para. 10.100. See also Ole K. Fauchald, supra note 9, at 234.
72 Appellate Body Report, Chile - Taxes on Alcoholic Beverages, WT/DS87/AB/R, WT/DS110/AB/R, adopted 12 January 2000, paras. 50-52.
73 Id., paras. 67, 64.
74 Other commentators have gone a step further and described the Appellate Body's reasoning as an apparent application of an asymmetric impact ("discriminatory effect") approach. See Simon Lester & Kara Leitner, Dispute Settlement Commentary, European Communities - Asbestos (Appellate Body Report) 14 (2001), available at <http://www.worldtradelaw.net/dscsamples/index.htm>, visited 4 February 2002.
75 The relevant passage of Article 5.5 of the SPS Agreement, which is part of Annex 1A of the WTO Agreement, supra note 1, states:
[E]ach Member shall avoid arbitrary or unjustifiable distinctions in the levels it considers to be appropriate in different situations, if such distinctions result in discrimination or a disguised restriction on international trade.
76 Panel Report, EC - Measures Concerning Meat and Meat Products (Hormones), WT/DS26/R/USA, circulated 18 August 1997, DSR 1998:III, 699, para. 8.205. (The Appellate Body dealt with para. 8.242 of this report, where the panel repeated this consideration by reference, and reversed the finding on unrelated grounds. Appellate Body Report, EC - Measures Concerning Meat and Meat Products (Hormones), WT/DS26/AB/R, WT/DS48/AB/R, adopted 13 February 1998, DSR 1998:I, 135, paras. 243-246.)
77 Highly critical of this approach to what he calls cumulative legal requirements ("like products" and "afforded production") is Richard A. Westin, Environmental Tax Initiatives and Multilateral Trade Agreements: Dangerous Collisions 128 (1997).
78 Panel Report, U.S. - Malt Beverages, supra note 6, paras. 5.23-5.26.
79 Id., para. 5.73.
80 Id., para. 5.72. See Ole K. Fauchald, supra note 9, at 222, who further understands this obiter dictum as a clear statement in favor of the diagonal approach, which is questionable given the prior statement on lacking protective effect. Similar Ilona Cheyne, Environmental Unilateralism and the WTO/GATT System, 24 Ga. J. Int'l & Comp. L. 433, 447 (1995).
81 Panel Report, United States - Taxes on Automobiles, DS31/R, 11 October 1994, unadopted, paras. 5.13-5.15.
82 Id., paras. 5.25, 5.31, 5.34.
83 For a discussion of the kind of asymmetry test the panel applied see infra section X.B, text accompanying note 207.
84 Paul Demaret & Raoul Stewardson, Border Tax Adjustments under GATT and EC Law and General Implications for Environmental Taxes, 28 J. World Trade, No. 4, at 5, 37 (1994), therefore characterize the finding as obiter dictum, which is questionable given the open factual basis of the reasoning.
85 This is equally the interpretation of Ole K. Fauchald, supra note 9, at 221-222.
86 Panel Report, U.S. - Malt Beverages, supra note 6, para. 5.19.
87 Ole K. Fauchald, supra note 9, at 218.
88 Panel Report, Canada - Certain Measures Concerning Periodicals, WT/DS31/R, circulated 14 March 1997, DSR 1997:I, 481, paras. 3.101, 3.102, 3.103.
89 Id., para. 3.119.
90 Id., paras. 5.28, 5.30.
91 Appellate Body Report, Canada - Periodicals, supra note 71, at 475-476 (section VI.B.3).
92 Ole K. Fauchald, supra note 9, at 246.
93 Appellate Body Report, Canada - Periodicals, supra note 71, at 476 (section VI.B.3).
94 Panel Report, Canada - Periodicals, supra note 88, para. 3.119.
95 See also Sydney M. Cone, III, The Appellate Body and Harrowsmith Country Life, 32 J. World Trade, No.2, at 103 (1998), according to whom the typical split-run edition is the Canadian regional edition of a U.S. magazine.
96 Panel Report, European Communities - Measures Affecting Asbestos and Asbestos-Containing Products, WT/DS/135/R, circulated 18 September 2000, paras. 8.155, 8.156.
97 One could oppose that even the diagonal approach requires some facts, namely that at least one (potential) import suffers and one actual domestic product benefits from the distinction. This is, however, a largely theoretical obstacle having hardly any significance in practice.
98 Appellate Body Report, EC - Asbestos, supra note 5, para. 100.
99 Equally interpreting paragraph 100 as supporting the comparison between the groups as a whole: Robert Howse & Elisabeth Tuerk, supra note 31, at 297-298 and note 55; Simon Lester & Kara Leitner, supra note 74, at 15; Joel P. Trachtman, Decisions of the Appellate Body of the World Trade Organization, European Communities - Asbestos, available at <http://www.ejil.org/journal/curdevs/sr13-02.html>, visited 4 February 2002.
100 See supra section III.A.2, note 10.
101 Similar Joel P. Trachtman, supra note 99; Simon Lester & Kara Leitner, supra note 74, at 15.
102 Panel Report, U.S. - Section 337, supra note 10, para. 5.11.
103 See also Robert Howse & Elisabeth Tuerk, supra note 31, at 297-298, note 55, who criticize that such a reading would, because of the paradigm of the comparison between the entire groups of like products, reverse another proposition established since the Section 337 case: that instances of less favorable treatment cannot be balanced with instances of more favorable treatment. See Panel Report, U.S. - Section 337, supra note 10, para. 5.14 and infra section IV.F. The continuing validity of this proposition has been confirmed in the Appellate Body Report, United States - Section 211 Omnibus Appropriations Act of 1998, WT/DS176/AB/R, adopted 1 February 2002, paras. 262 and 264.
104 See, however, Gabrielle Marceau, L'affaire « CE - Amiante » et la nouvelle jurisprudence de l'Organe D'appel de l'OMC concernant les risques à la santé, 39 Annuaire canadien de Droit international 213, 217-218 (2001).
105 Panel Report, U.S. - Section 337, supra note 10, para. 5.11.
106 Appellate Body Report, Korea - Measures Affecting Imports of Fresh, Chilled and Frozen Beef, WT/DS161/AB/R, WT/DS169/AB/R, adopted 10 January 2001, paras. 135-137.
107 See also supra note 102 and accompanying text.
108 Appellate Body Report, Korea - Beef, supra note 106, paras. 143-148.
109 This is the conclusion put forward by Gabrielle Marceau, supra note 104, at 218 (2001).
110 Robert Howse & Elisabeth Tuerk, supra note 31, at 299.
111 See also Robert Howse & Elisabeth Tuerk, supra note 31, at 289, 297, 298.
112 In this direction see also Simon Lester & Kara Leitner, supra note 74, at 15.
113 Simon Lester & Kara Leitner, supra note 74, at 15.
114 Panel Report, U.S. - Section 337, supra note 10, para. 5.14. See also Panel Report, U.S. - Taxes on Automobiles, supra note 81, paras. 5.47, 5.48; Panel Report, United States - Standards for Reformulated and Conventional Gasoline, WT/DS2/R, circulated 29 January 1996, DSR 1996:I, 29, para. 6.15.
115 The same is true for the most recent confirmation of the inadmissibility of "balancing" in the Appellate Body Report, U.S. - Section 211 Appropriations Act, supra note 103, paras. 262 and 264. See id., paras. 264, 268, 279-280.
116 Also John H. Jackson et al., Legal Problems of International Economic Relations 522 (3rd ed. 1995); Robert Howse & Elisabeth Tuerk, supra note 31, at 297-298, note 55; Robert E. Hudec, supra note 31, at 622. The opposite suggestion in relation to Section 337 in the Decision by the Arbitrators (under Article 22.6 of the Dispute Settlement Understanding), EC - Bananas III, WT/DS27/ARB, circulated 9 April 1999, para. 10.86, note 39, must be based on a different concept of de facto discrimination.
117 See supra section III.A.4, and the first illustration supra in section III.A.3.a). The illustration reflects with different rows the two regulatory sub-categories of like products resulting from the legislative differentiation. The dotted arrows represent the comparison within those sub-categories. By definition, this comparison shows equal treatment. In contrast, in the case of a disadvantageous origin-based differentiation, this comparison would show less favorable treatment.
118 Unlike other reports, and only in the context of "not similarly taxed", this is what the Appellate Body did in its Report in Canada - Periodicals, supra note 71, at 474 (section VI.B.2).
119 See Panel Report, U.S. - Section 337, supra note 10, paras. 5.16, 5.17.
120 See however Simon Lester & Kara Leitner, supra note 74, at 14-15, for the opposite opinion.
121 It is therefore open to question whether the "traditional interpretation" truly reflects a diagonal test. See supra note 8.