Several arguments for and against the diagonal approach rely on the plain language of the provisions at issue. GATT Articles III:4 and III:2 refer to imported "products", and "like domestic products" in the plural, which has led to a textual argument in favor of comparing the entire respective groups, 168 as the Appellate Body indicates in paragraph 100 of the Asbestos report. The designations of origin "imported" versus "domestic" could also support a requirement of discrimination related to that origin. 169 Article III:1, relevant directly in Article III:2, second sentence, and informing the entire Article, does not end on "a", "some" or "any domestic production", but on "domestic production" as a general term without specification.
However, these arguments are not compelling given that the Note Ad Article III uses "product" in singular as does Article III:3. 170 Even if these provisions are not directly relevant in all Article III disputes, 171 they substantially weaken the argument that the plural determines the approach for the comparison. GATT Article I:1 does not only express "product" in the singular. It also refers to "any advantage" granted to "any product", but on the other hand refers to "the like product" from other countries. 172 In addition to this inconclusiveness, singulars often appear in legal provisions as pars pro toto and plurals in order to convey general validity. It might also be exaggerated to give dispositive weight to specific words. 173
The text of the provisions, therefore, arguably allows both the diagonal and asymmetry tests. All the diagonal test does is select a sub-group of like imports to be compared with a sub-group of like domestic goods. However, equally selective would be the comparison within the regulatory sub-categories. 174 By definition, origin-neutral rules then would not be able to give rise to less favorable treatment, unless the imposed burden is qualitatively heavier for imported goods (asymmetric intensity 175). The concept of de facto discrimination would become widely redundant because none of the discussed cases could have resulted in a violation finding. Such an approach would be legally unacceptable, because it relies on excluding a part of the like products from the comparison. 176 This leads to the question whether the diagonal approach relies on a similarly selective exclusion of a part of the like products from the picture. The answer depends upon whether one accepts or rejects the premise that an origin-neutral measure may treat no single import less favorably than any like domestic good.
The titles "Most-Favored-Nation Treatment" and "National Treatment" may be indicative of the basic meaning of the obligations. 177 Rather than stipulating general equal treatment for imports, these titles suggest that the subsequent provisions prohibit WTO Members from discriminating against imports on the basis of their origin. An approach resulting in the prohibition of origin-neutral rules, even when they predominantly favor imports over like domestic goods, goes beyond preventing disadvantages related to foreign origin. "National Treatment" ("Most-Favored-Nation Treatment") may suggest that a Member complies with the obligation, if it treats imports exactly the same way as its own products (imports from one Member the same as those from other countries), unless the facially origin-neutral criteria it employs result de facto in worse treatment in a majority of cases, i.e. an asymmetry.
In support of this suggestion, one may also refer to the context of Articles I:1, III:2 and III:4. The exceptions and special rules in the other paragraphs of GATT Articles I and III mostly refer to origin-related discrimination. The rule on mixing requirements in GATT Article III:5 presupposes that mixing requirements are not as such illegal, although they necessarily favor the use of one product over another. For the second sentence of Article III:5, the Ad Note provides that substantial production of all products in question shields from a violation.
Whether the purpose of GATT Articles I and III and of the WTO Agreement in general 178 is to prevent origin-related discrimination and not also discrimination in the wider sense of the diagonal approach, 179 is precisely the question. As the Appellate Body and panels have repeatedly noted, the purpose of GATT Article III is to provide equal competitive conditions for imported products in relation to domestic products 180 once they have cleared through customs. 181 In such general terms, an origin-neutral measure will not affect the competitive opportunities for imports as a whole if its burden falls proportionally on imported and domestic goods. Obviously, the competitive opportunities of some imports will always suffer.
One original purpose of the national treatment obligation is to prevent circumventions and frustrations of tariff concessions, 182 given that discriminatory internal taxes could neutralize the effect of tariff reductions. In the relationship between imports and domestic goods, the tariff is origin-based by definition so that only an openly discriminatory tax or one with appreciably asymmetric impact can achieve the same or similar effect. The diagonal approach, therefore, goes beyond protecting the benefits of tariff reductions.
The purpose of GATT Article I is to prevent discrimination between WTO Members. Where an origin-neutral measure has an asymmetric impact, it produces a disadvantage for one Member in relation to other Members. In non-asymmetric cases, it merely favors one sector over another, and the different Members to similar extents partly enjoy the benefit, and partly receive the disadvantage resulting from the differentiation. The diagonal approach thus provides for equality between different kinds of like products. 183
It would certainly be at odds with the object and purpose of the non-discrimination rules to reduce them to inutility. However, the asymmetric impact test does not render the national and most-favored-nation obligations void of meaning. In particular, an asymmetry does not only exist in situations in which there is hardly any or no domestic production of the disfavored product, 184 but wherever there is an imbalance between imports and like domestic goods in their distribution across the favored and the disfavored regulatory sub-categories.
It has been argued that the diagonal approach follows directly from the emphasis on competitive opportunities as the criterion used in the GATT/WTO jurisprudence to identify discrimination. Conversely, according to this argument, the asymmetry approach amounts to the "trade effects" test that has been rejected in that jurisprudence.185 It is, however, important not to confuse these two quite different questions, since the asymmetry test is perfectly in line with the object and purpose of Article III, which is to protect effective equality of competitive opportunities between imports and domestic goods. The asymmetry approach does not dismiss a complaint because (seemingly) no actual trade effect can be shown.186 It examines the configuration of imports and domestic products at one point in time in order to explore whether the measure at issue modifies the competitive relationship between imports and like domestic goods, both taken as a whole, to the detriment of the former. It does not attempt to measure the significance of actual trade effects by comparing the levels of trade volumes at two points, or in fact any point in time. Neither does it exclusively take account of actual trade flows; on the contrary, potential imports are of paramount importance. Only if the measure does not distort the overall competitive relationship between (potential) imports and like domestic goods, because the competitive opportunities of both groups are equally affected, would a complaint be dismissed. Production and import data or market shares serve only as a more precise and differentiated indicator of how the respective competitive opportunities are affected. In the realm of de facto discrimination, it is anything but surprising that such facts matter, and not only the explicit language of the laws at issue. The real life (de facto) situation to which the facially neutral rule applies is the marketplace in its particular composition. If there were any connection with a "trade effects" test, it would be odd for the Appellate Body to refer to the objective of Article III as providing for "equality of competitive conditions" and then to rely on the significant asymmetry in the market.187
Economically, both most-favored-nation and national treatment intend to enhance the benefits of comparative advantage. An origin-neutral distinction generating an asymmetric impact impairs the comparative advantage of imports in relation to domestic goods, because it limits the capacity of the disadvantaged part of the like imports to take over a higher market share. The effect is similar to that of a tariff.
Without asymmetry, that distinction alters only the competitive relationship between the regulatory sub-categories, not between imports and domestic supplies. The destination country does not forego potential benefits accruing from comparative advantage, i.e. from importing a product instead of producing it, because imports and domestic goods compete without distortion in both sub-categories. There remains, of course, a distortion between the sub-sectors and, as a result, a potential efficiency loss, which however does not impair the beneficial effects of the different comparative advantage of different Members. A symmetric distribution, therefore, suggests that the domestic production is similarly competitive in both sub-sectors. Whenever other countries in general or a particular country enjoys a comparative advantage in one sub-category as opposed to the other, an asymmetry in the distribution of potential imports should reflect this: if that superiority is due to the efficiency of that foreign sub-sector, it should by tendency enjoy a larger market share abroad among like products than in the market in question. If it is due to the domestic inefficiency of the other sector, then without the measure that sector would have a smaller production share domestically than abroad. The differentiating measure should not prevent imports from reducing the amount and share of domestic production in that sub-sector.
If the measure on the whole favors more imports than domestic goods, 188 the relative scarcity of domestic products of the favored sub-category may well be due to that sub-sector being the comparative advantage of other countries. The national measure having this reverse asymmetric effect to some extent operates like an import subsidy, and potentially undermines the domestic comparative advantage in the other sub-category which the measure places at a disadvantage. However, the GATT is generally not concerned with reverse discrimination.
Where the purpose of the national measure is to address a market failure, the national efficiency effect of outlawing the differentiation should ipso facto be negative. The asymmetric approach then offers the advantage that it gives wider margin to the government to maintain distinctions.
Where there is no market failure, the effect of the differentiation on the competitive relationship will be distortive. Another negative potential effect of the asymmetry requirement is that it often allows the application of regulatory distinctions to imports, which could result in the exportation of economic distortions: in other countries, the intended effect might be less desirable or even harmful. Operators in other countries will, however, only orient their behavior according to the differentiation, if the destination market is sufficiently important. 189
The negotiating history suggests that the drafters recognized possible conflicts of origin-neutral rules with the national treatment obligation. At the Geneva preparatory conference in 1947, the passage now in Article III:2, second sentence, was explained with an example of a country producing apples, but no oranges, and imposing a high tax on oranges. 190 Accordingly, the 1947 Geneva draft version commenced with the words: "In cases in which there is no substantial domestic production of like products". 191 The 1948 Havana Conference removed that passage from Article 18 of the Havana Charter, but the Sub-Committee stressed as the only substantial change another modification, the immediate prohibition of such discriminatory taxes. 192 It maintained that Article 18 would not preclude regulations and taxes "assisting the production of a particular domestic product" that "are directed as much against the domestic production of another product ... of which there was a substantial domestic production as they are against imports".193 Except for the difference between "substantial quantities" and "equivalent ratios", these statements reflect the asymmetry requirement. It appears that the drafters intended to prohibit, in addition to formal discriminations against imports, such measures which indirectly achieve a similar result because they disfavor predominantly imports.
Arguments against the diagonal approach have also been made a maiore ad minus: if under GATT Article III:4 a country may outlaw the sale of all foreign and domestic wine, it should also be able to restrict all foreign and domestic wine containing pesticides. 194 Conceptually, this is a valid point, provided that the total prohibition really covers all like products, i.e. none of the spared other beverages are "like" wine. Under that condition, even the diagonal approach would not succeed in showing that the total prohibition violates national treatment.
168 Ole K. Fauchald, supra note 9, at 220, 221; see also Canada's claim in Canada - Periodicals, Panel Report, supra note 88, para. 3.109.
169 William J. Davey & Joost Pauwelyn, supra note 8, at 39-40.
170 Counterargument of the U.S.A. in Canada - Periodicals, see Panel Report, supra note 88, para. 3.107.
171 Ole K. Fauchald, supra note 9, at 220, note 15, therefore finds the U.S. argument unconvincing.
172 The wording of Article I is understood as less supportive of an asymmetry requirement than that of Article III by William J. Davey & Joost Pauwelyn, supra note 8, at 40; as indication against such a requirement by Ole K. Fauchald, supra note 9, at 251.
173 Critical of such an approach in another context: Robert E. Hudec, supra note 31, at 641.
174 See supra section III.A.4, text accompanying note 14, and the dotted arrows in 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.
175 See supra section III.A.2, note 10.
176 Appellate Body Report, Chile - Alcoholic Beverages, supra note 72, paras. 50-52. See also supra section III.A.4.
177 The Appellate Body has sought guidance from titles as part of the text and of the context. See e.g. Appellate Body Report, U.S. - Section 211 Appropriations Act, supra note 103, paras. 154 and 155 (text); and Appellate Body Report, Argentina - Safeguard Measures on Imports of Footwear, WT/DS121/AB/R, adopted 12 January 2000, para. 93 (context).
178 See the 3rd para. of the Preamble of the WTO Agreement.
179 This is the position of William J. Davey & Joost Pauwelyn, supra note 8, at 40, 41; Ole K. Fauchald, supra note 9, at 221, 231.
180 Appellate Body Report, Japan - Alcoholic Beverages, supra note 63, at 109-110 (section F); Appellate Body Report, Korea - Taxes on Alcoholic Beverages, supra note 68, para. 119.
181 Panel Report, Italian Discrimination Against Imported Agricultural Machinery, adopted 23 October 1958, BISD 7S/60, para. 11.
182 Panel Report, U.S. - Malt Beverages, supra note 6, para. 5.30; Panel Report, 1987 Japan - Alcoholic Beverages, supra note 49, para. 5.5 at b).
183 William J. Davey & Joost Pauwelyn, supra note 8, at 41.
184 See the argument of Kenneth W. Dam, supra note 144, at 118-119, and text accompanying note 144.
185 See e.g., as regards "trade effects", Kazumochi Kometani, Trade and Environment: How Should WTO Panels Review Environmental Regulations Under GATT Articles III and XX?, 16 Nw. J. Int'l L. & Bus. 441, 463 (1996); and, despite an opposite conclusion with respect to "competitive opportunities", Frieder Roessler, supra note 146, at 25.
186 Relying for instance on a low total amount of imports or in fact increasing imports, defendants have often contended that their (discriminatory) measure produces no actual detrimental effects. Not only in WTO law, such a submission is indeed in contradiction to basic principles of international economic law. Of course, without infringement on those principles, a "demonstrated actual protective effect" can serve as evidence for the protective effect of a measure. See, Appellate Body Report, Canada - Periodicals, supra note 71, at 476 (section VI.B.3).
187 Appellate Body Report, Chile - Alcoholic Beverages, supra note 72, para. 67; supra section IV.B.7, text accompanying notes 71-73.
188 See the scenario illustrated supra in section III.A.3.c).
189 See Paul Farmer & Richard Lyal, supra note 123, at 76.
190 United Nations Economic and Social Council, Second Session of the Preparatory Committee of the United Nations Conference on Trade and Employment, U.N. Document E/PC/T/A/PV/9, at p. 7.
191 See Report of the Second Session of the Preparatory Committee of the United Nations Conference on Trade and Employment, U.N. Document E/PC/T/186 Corr. 1, at p. 18.
192 Reports of Committees and Principal Sub-Committees, United Nations Conference on Trade and Employment, ICITO 1/8, p. 61, para. 36 (1948).
193 Id., p. 64, para. 54.
194 Steve Charnovitz et al., supra note 159, at 305 note 36.