Regulatory purpose and "Like products" in article III:4 of the GATT(with additional remarks on article III:2)

Journal of World Trade

Volumn 36, Issue 3, 2002, Pages 443-478

  Regan, Donald H ^a  

^a UNIVERSITY OF MICHIGAN   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

GATT; REGULATORY APPROACH;

EID: 23044533061     PISSN: 10116702     EISSN: None     Source Type: Journal    
DOI: 10.1023/A:1016312511993     Document Type: Article

References (99)

1. European Communities - Measures Affecting Asbestos and Asbestos-Containing Products, WT/DS135/AB/R (12 March 2001) [hereafter Asbestos or EC - Asbestos], paras 93, 98.
2. Chile - Taxes on Alcoholic Beverages, WT/DS87 & DS110/AB/R (13 December 1999) [hereafter Chile -Alcohol], para. 62.
3. United States - Measures Affecting Alcoholic and Malt Beverages, BISD 39S/206 (adopted 19 June 1992) [hereafter US - Malt Bevrages].
4. Following the lead of US - Malt Beverages, the Panel in United States - Taxation of Automobiles, DS31 (11 October 1994)
5. originated the famous "aims and effects" language, but the report was not adopted. Thereafter the Appellate Body rejected the "aims and effects" test, and purported to reject all consideration of legislative purpose, in Japan - Taxes on Alcoholic Beverages, WT/DS8 & DS10 & DS11/AB/R (4 October 1996) [hereafter Japan - Alcohol].
6. Japan - Alcohol was decided under Article III:2, but in European Communities - Regime for the Importation, Sale and Distribution of Bananas, WT/DS27/AB/R (9 September 1997) [hereafter Bananas III], para. 216, the Appellate Body seemingly extended the ban on considering regulatory purpose to Article III:4, saying Article III:4 cases should be decided without explicit reference to the Article III:1 policy against protectionism. Chile - Alcohol commences the rehabilitation of regulatory purpose analysis, under Article III:2, second sentence, and Asbestos makes it clear that the consequences of Chile - Alcohol are much broader still. Obviously, there are complexities in the development that are beyond this footnote but that will be revealed as we proceed. This was just to explain roughly how we have "come full circle".
7. As it happens, I think the same is true of "like products" in Article III:2, First sentence. But since that is an even more controversial claim, I postpone it to section III.B below.
8. Asbestos, para. 100, discussed in section IV.A below.
9. For further discussion, see section IV.A. One advantage, from my perspective, of considering regulatory purpose as part of "less favourable treatment" is that that would allow us to uphold process-based measures that embody a non-protectionist regulatory purpose, even if the products made with the different processes are physically identical and are held on that ground to be "like" (mistakenly in my view, see Robert Howse and Donald Regan, The Product/Process Distinction - An Illusory Basis for Disciplining "Unilateralism" in Trade Policy, 11 EJIL 2 (2000), pp. 249-289).
10. Robert Howse and Donald Regan, The Product/Process Distinction - An Illusory Basis for Disciplining "Unilateralism" in Trade Policy, 11 EJIL 2 (2000), pp. 249-289) Id.
11. Asbestos, paras 96, 99.
12. This is what the Asbestos Appellate Body says in para. 88, deftly setting aside any possible relevance of Bananas III (of which more below) on the ground that in Bananas III they were not called upon to address the meaning of "like products". Para. 88, n. 57.
13. But see the discussion in section IV.A below of a relevant ambiguity in the Appellate Body's argument.
14. Uniquely in the annals of the Appellate Body, there is a "concurring statement" of six paragraphs by one Member of the Division interpolated into the report, Asbestos, paras 149-154.
15. Asbestos, Id. Para. 152.
16. Id. Para. 154.
17. See section IV.A below.
18. Even this claim must be taken with a grain of salt, since there is no unambiguous definition of "more alike in. their physical properties". If we double one dimension of a square one unit on a side, have we made it more like a square two units on a side (because it is now like it in one dimension) or less like it (because we have turned the smaller square into a rectangle)?
19. For similar remarks, see Robert Hudec, '"Like Product': The Differences in Meaning in GATT Articles I and III", in Thomas Cottier and Petros Mavroidis (eds), Regulatory Barriers and the Principle of Non-Discrimination in World Trade Law (Ann Arbor, MI: University of Michigan Press, 2000), p. 103. Incidentally, we shall see in section III.B below that we cannot make physical likeness provide the ultimate criterion even by interpreting it very narrowly, which was the stratagem adopted by the Panel and Appellate Body in Japan - Alcohol for dealing with Article III:2, First sentence.
20. Bananas III, para. 216.
21. Asbestos, para. 88, n. 57.
22. Id. paras 93, 98.
23. There could also be a weaker sort of collective action problem: it might be that each consumer will eschew plastic jugs if, but only if, he knows enough others are doing so, in which case we may need legal intervention to provide information about the numbers.
24. I ignore for the moment the possibility of considering regulatory purpose under the rubric of "less favourable treatment", which I mentioned in the Introduction and which I will address in the next paragraph.
25. For a very nice parallel discussion, see Frieder Roessler, The Legal Structure, Functions & Limits of the World Trade Order (London: Cameron May, 2000), pp. 122-130. This is as good a point as any to comment on an issue that may be troubling some readers: if we consider regulatory purpose under Article III rather than postponing it to Article XX, who has the burden of proof? The answer is that the complainant has the burden of proof of showing protectionist purpose, since protectionist purpose is the crux of the basic violation. But if that burden seems too heavy, we should remember the distinction between the burden of proof and the burden of going forward with the evidence. The "burden of proof" refers to the issue of who wins if, with all the evidence in, the probabilities seem equally balanced. The burden of proof of the basic Article III violation is and remains with the complainant. But the complainant may well at some point produce enough evidence of protectionist purpose so that the defendant regulating country will lose unless they produce some evidence in response. (This is one possible meaning of a "prima facie case", although not the only one.) At this point, the defendant has the burden of going forward with the evidence - not in the sense that they have any legal duty to go forward, but in the practical sense that if they do not, they will lose. Often, I suggest, the complaining country will be able to shift the burden of going forward onto the defendant just by putting before the tribunal the text of the statute and a few basic facts about who produces what. In truth, in cases where the putative regulatory justification is one that is recognized by Article XX, it will often not make much difference whether we consider regulatory purpose under Article III or Article XX. (The burden of proof differs, but the burden of proof is usually not what decides the case. There may be other differences as well, for example in the precise way in which evidence about possible alternative measures is relevant, but such differences are not worth pursuing here.) But the italicized qualification in the last sentence but two is important, since many perfectly sound regulatory policies are not recognized by Article XX. In any event, even when the differences in procedure and probable result are slight, we should apply the treaty as it is written.
26. It is an interesting question what native speakers think about the formulations that correspond to "so as to afford protection" in the equally authentic French and Spanish texts of die treaty. The French formulation sounds purposive to my very non-native ear. The Spanish perhaps somewhat less so?
27. There may also be a revenue goal, but if there are import-competing producers of the same good, then the protectionist goal is manifested in the choice to raise revenue through a tariff rather than a neutral tax. If there are no import-competing producers, then the tariff might indeed be a pure revenue measure, but of course it could then equally be in the form of a neutral tax. The tariff might also be an "optimum tariff" designed to exercise collective monopsony power, but this situation is rare enough, at least in a pure form, so that the generalization in the text stands.
28. Italian Discrimination Against Imported Agricultural Machinery, BISD 7S/60 (adopted 23 October 1958).
29. One very important element here is the presumption, in effect, that origin-specific regulations are illegal. For an explanation of why the language of Article III compels distinct treatment of origin-neutral and origin-specific measures, even though it does not mention the distinction explicitly, see section III.A below.
30. Plainly, failure to use less trade-restrictive measures that would achieve the asserted non-protectionist goal is strong evidence that the actual goal is protectionism. For fuller discussion, see Donald Regan, Judicial Review of Member-State Regulation of Trade Within a Federal or Quasi-Federal System: Protectionism and Balancing, Da Capo, 99 Mich L Rev 8 (August 2001), pp. 1853-1902, at 1870-1872, 1891-1892, 1899-1902.
31. See Regan, Mich L Rev, ibid., pp. 1853-1879.
32. Again, see id.
33. US - Malt Beverages, para. 5.25. I should mention somewhere the particular puzzle that is posed for purpose analysis by changed circumstances since the adoption of the law. Even if a regulation was adopted with an innocent purpose, it may be maintained for protectionist reasons. We could ask the question directly about the purpose of the non-repeal of the law - but often inaction has no such focused purpose as action. I have suggested elsewhere that we should deal with this problem by taking a hypothetical approach - asking whether it is reasonable to suppose that if the legislature enacted the regulation in the present, they might do so without protectionist purpose. For a fuller discussion, see Regan, note 28 above, pp. 1869-1870.
34. Japan - Taxas on Alcoholic Bevenges, WT/DS8 & DS10 & DS11/R, para. 6.17.
35. It has been suggested to me that perhaps an Article entitled "General Exceptions" should have some application to every prohibitory Article. This seems to read too much into a title. "General", in the context, means at most "broad, and formally universal", but not "universally efficacious in concreto". If a textual argument is required, it is worth comparing the title of Article XX, "General Exceptions", with the title of Article XXI, "Security Exceptions". If these titles are read as parallel in construction, then we see that "General" refers not to the range of other Articles that may be excepted from (that is taken care of by the language "nothing in this Agreement" in both Articles XX and XXI), but rather to the nature of the considerations by which the exceptions are justified.
36. Japan - Taxas on Alcoholic Bevenges, WT/DS8 & DS10 & DS11/R, para. 6.16.
37. Origin-specific measures may still be rescued by Article XX, of course.
38. 477 U.S. 131 (1986).
39. It might still be asked, "Why Article XX? Why not bring these considerations under Article III?" I argued above that all origin-specific measures violated Article III because they distinguished between products that were identical in all respects except their origin. We now see, from the baitfish case, that that is both true and not true. It is true in the sense that the Maine baitfish law will exclude some foreign non-infested baitfish that is in fact identical to local non-infested baitfish. It is not true in the sense that all the foreign baitfish, even the non-infested baitfish, is potentially infested - not known to be not infested and not knowably not infested - in a way that distinguishes it from the local baitfish. We could say that this epistemological difference is enough to make all foreign baitfish "unlike" all local baitfish, so that the law does not violate Article III. There is a choice to be made here about the relevance of epistemological considerations to the interpretation of "likeness". My inclination, when we are confronted with an origin-specific measure, is to stick with the idea that the law excludes some identical foreign baitfish, even though we cannot identify the particular baitfish which is thus "improperly" excluded. It seems more in keeping with the general structure of Articles III and XX to say that origin-specificity raises a strong presumption of illegality-both because it is rare that an origin-neutral purpose requires an originspecific measure and because an origin-specific measure is particularly likely to be resented by the burdened foreign interests. We embody this strong presumption of illegality in our procedures by requiring that a justification for an origin-specific measure be proved under Article XX, where only certain justifications are available and the regulator bears the burden of proof. No such presumption of illegality can plausibly attach to origin-neutral measures with disparate impact.
40. Japan - Taxes on Alcoholic Bevenges, WT/DS8 & DS10 & DS11/R, para. 6.17.
41. EC - Asbestos, para. 115.
42. For a fuller discussion of how purpose is interpreted in terms of political process, and the advantages of such an approach, see Regan, as note 28 above, pp. 1882-1889.
43. Except briefly in stating an argument of the European Communities, Asbestos, para. 86.
44. Canada - Certain Measures Concerning Periodicals, WT/DS31/AB/R, s.VI.B.3.
45. US - Malt Beverages, paras 5.23-5.26.
46. Chile - Alcohol, WT/DS87/AB/R.
47. Japan - Alcohol, WT/DS8 & DS10 & DS11/AB/R.
48. EC - Measures Concerning Meat unit Meat Products (Hormones), WT/DS26/AB/R (16 January 1998) [hereafter Hormones], para. 244.
49. Hormones, para. 245.
50. Hormones, Id.
51. Australia - Measures Affecting Importation of Salmon, WT/DS18/AB/R (20 October 1998), para. V.C.3.12.
52. Id., para. V.C.3.8.
53. See the discussion in section IV.B below.
54. See the discussion later in this section.
55. See text accompanying note 42 above.
56. Chile - Alcohol, WT/DS87 & DS110/AB/R (13 December 1999), para. 62 (emphases in original, including emphases added by the Chile Division to the Japan quote).
57. Chile - Alcohol, Id., para. 71 (emphasis in original).
58. Chile - Alcohol, Id., para. 72.
59. Chile - Alcohol, Id., para. 71.
60. It might seem that the Appellate Body in Korea-Taxes on Alcoholic Beverages, WT/DS75 & DS84/AB/R (18 January 1999), denied the possibility of justifying a very unequal tax scheme when it said, "[T]he reasons given by Korea as to why the tax is structured in a particular way do not call into question the conclusion that the measures are applied 'so as to afford protection to domestic production' " (para. 150). This could be read as saying that attempts at explanation are irrelevant in principle. But in context the better reading is just that Korea's attempts at explanation were so thin as to be unpersuasive. [See the Panel report, WT/DS75 & DS84/R (17 September 1998), paras 5.172-5.181.] And if the Korea Appellate Body in fact meant that attempts at explanation are irrelevant in principle, Then Chile - Alcohol, which is both later in time and more explicit, establishes the contrary.
61. Except, as I have noted previously, in summarizing an argument by the European Communities. EC - Asbestos, WT/DS135/AB/R (12 March 2001), para. 86.
62. Working Party Report, Border Tax Adjustments, adopted 2 December 1970, BISD 18S/97.
63. "Thus, a determination of 'likeness' under Article III:4 is, fundamentally, a determination about the nature and extent of a competitive relationship between and among products." Asbestos, para. 99.
64. Asbestos, Id., para. 99.
65. Asbestos, Id., para. 100.
66. Asbestos, Id., para. 101.
67. Asbestos, Id., para. 103.
68. Hudec, as note 17 above, pp. 112-113.
69. Asbestos, para. 133.
70. Asbestos, Id., para. 139, relying on earlier argument in paras 113-123. Technically, the Appellate Body argue only that Canada has not carried its burden of proof of showing a competitive relationship, but it is plain that their attitude to what Canada needs to do is coloured by their view about what is plausible.
71. The Appellate Body acknowledges in a footnote that informed consumers sometimes make choices that involve risks to their health, as with tobacco. Asbestos, Id., para. 122, n. 103.
72. Id., para. 122.
73. This position might be suggested by, e.g., Asbestos, id., paras 118, 121.
74. This position is suggested by, e.g., Asbestos, id., paras 114, 115.
75. Asbestos, Id., para. 100.
76. It is not surprising that regulatory purpose should come in by two routes. There is one underlying question, about "so as to afford protection". If we divide that into a two- aspect question, about "less favourable treatment" of "like" foreign products, it is natural that the underlying issue can manifest itself through either aspect. The one minor difference is this: If we come to the question of regulatory purpose through the meaning of "like product", we are directed primarily to look for plausible non-protectionist purposes. In contrast, if we come to the question of regulatory purpose through the meaning of "less favourable treatment", we are directed more immediately to the question whether the actual purpose was protectionism. But these are just opposite sides of the same coin.
77. I ignore the fact that France is part of the European Communities. This raises some genuine questions about which producers and which markets we should look at, but those questions are peripheral to the central point.
78. US - Malt Beverages, BISD 39S/206, para. 5.17.
79. Howse and Regan, note 7 above, at n. 22.
80. In rejecting this argument from US - Malt Beverages, I cast no doubt on the result in United States - Section 337 of the Tariff Act of 1930, BISD 36S/345 (adopted 7 November 1989). The reason that product-by-product or case-by-case differences in treatment mattered in that case, even if, as the United States claimed, the overall statistics showed no disadvantage to the ensemble of foreign products subjected to patent challenge, was that the randornness itself, at the product or case level, imposed uncertainty costs on investors. There is no comparable uncertainty in my asbestos hypothetical.
81. This "fail-safe" idea was brought clearly into focus for me by conversation with Rob Howse. Notice that the existence of the "fail-safe" mechanism is not an argument for a more relaxed view about what are "like products" (and a fortio not an argument for a more relaxed view about Article III, because Article XX waits in the wings). The point of a "fail-safe" mechanism is to have multiple safeguards against error (in this case, against inappropriate interference with innocent regulation). There is a third possibility about Asbestos para. 100, that it is just reminding us of the standard but important point that different treatment is not necessarily less favourable treatment. At least in principle, some treatment could be different but equally favourable. This is not very plausible as an interpretation of para. 100, since this point has not even a possible application to Asbestos itself. No matter how we slice up the universe of products in Asbestos, there is never a pair of products or product groups which we could plausibly think of as being treated differently but equally favourably. (Nor does this point seem more likely to be overlooked because "like products" is given a broad scope.)
82. Japan - Alcohol, WT/DS8 &DS11 amp; DS11/AB/R (4 October 1996), H.2.
83. Japan - Alcohol, Id., H.1(a).
84. Japan - Alcohol, Ibid. H.1(a).
85. Japan - Alcohol, Ibid. H.1(a).
86. Japan - Alcohol, Ibid. H.1(a).
87. Japan - Alcohol, Id., H.2.
88. Bananas III, WT/DS27/AB/R (9 September 1997), para. 216. For an explanation of the "apparently", see note 91 below.
89. We could of course now deny the Bananas III premise; we could say that because Article III and Article III:4 have fundamentally different structures, there is no inference from how Article III:1 figures in the one to how it figures in the other. In fact, I would not argue that the Asbestos treatment of Article III:1 and Article III:4 itself entails any conclusion about Article III:1 and Article III:2, first sentence. But the Asbestos treatment of Article III:1 and Article III:4 certainly suggests and allows the view that we should look to Article III:1 in interpreting Article III:2, first sentence, if there is independent reason to take that position, as we have seen there is.
90. Argentina - Measures Affecting the Export of Bovine Hides and the Import of Finished Leather, WT/DS155/R (adopted 16 February 2001).
91. Argentina - Measures Affecting the Export of Bovine Hides and the Import of Finished Leather, Id., para. 11.133.
92. Argentina - Measures Affecting the Export of Bovine Hides and the Import of Finished Leather, Id., para. 11.137.
93. If we look carefully, we will see that all the Appellate Body ever said about Article III:4 in Bananas III was that the analysis did not require a separate step devoted to "so as to afford protection". I think saying that and only that, without explication of the role of "so as to afford protection" in the interpretation of "like products", was potentially very misleading and probably misled many readers. Still what they actually precisely said is completely consistent, both with what I am arguing here and with what they say about the relevance of Article III:1 to Article III:4 in Asbestos.
94. Japan - Alcohol, H.2.
95. Japan - Alcohol, Ibid., n. 52, quoting from the drafting history at the Havana Conference, E/CONF.2/C.3/59, p. 8.
96. To be sure, treating the interpretative note as a self-sufficient spelling-out of Article III:2, second sentence, might not do justice to the precise logical import of the phrase "only in cases" in the Ad Note. "Only in" makes the conditions it names necessary, not sufficient. On the other hand, the use of "only in" does not deny that the named conditions are intended to be sufficient, if the general structure suggests that. I doubt the treaty drafters were always perfectly punctilious about necessary and sufficient conditions.
97. Japan - Alcohol, H.2(c).
98. 'Robert Hudec, GATT/WTO Constraints on National Regulation: Requiem for an "Aims and Effects" Test, 32 Int'l Lawyer 3 (Fall 1998), pp. 619-649, at 631.
99. Japan - Alcohol, H.2(c), quoting the Panel report, para. 6.35.