1. Introduction
Over the past two decades, one of the most controversial issues related to the GATT National Treatment (NT) rule has been whether regulatory purpose should be considered in determining the inconsistency of an origin-neutral measure with this rule. The significance of this issue arises from the fact that it has close bearing on the attempt to address the tension between trade liberalization and WTO Members' pursuit of non-trade interests through internal policy instruments. Past debates on the issue have predominantly drawn upon the primary means of treaty interpretation as laid down in the Vienna Convention on the Law of Treaties (Vienna Convention),Footnote 1 including the text, context, and objective of this rule. However, the application of this primary interpretative approach has not managed to resolve this issue in a clear-cut manner. The consultation of supplementary materials seems to be necessary for the interpretation of the issue.
This article proposes a comprehensive study of the preparatory work of the NT rule, which has received little attention to date from either WTO adjudicators or WTO observers.Footnote 2 By studying the preparatory work, this article aims to explore two issues: (i) whether regulatory purposes should be considered under this rule, and, if so, (ii) where (i.e. under which legal elements) should purpose inquiries be undertaken. The focus of this article is on the negotiating records of Articles III:1, 2, and 4 – the most frequently invoked provisions in NT disputes.
This article is organized as follows. Section 2 briefly outlines the major issues and controversies, and establishes the view that the issues remain unresolved by resorting to the primary means of treaty interpretation. Section 3 analyzes the negotiating records leading up to the conclusion of the NT rule during the four negotiation sessions from 1946 to 1948. It is argued that the drafting records indicate that while the intended function of Article III:2, 2nd sentence was to tackle de facto discrimination arising from the application of origin-neutral measures, Article III:2, 1st sentence was originally intended to combat de jure discrimination arising from the application of origin-specific measures only. Article III:4 received little attention from the negotiators, but can be interpreted as being intended to deal with both origin-specific and origin-neutral measures. Purpose inquiries were considered necessary only in cases involving origin-neutral measures. This article submits that the insertion of Article III:1 to the NT rule was intended to limit the function of this rule by treating origin-neutral measures as de facto discriminatory measures only when they had a protectionist purpose. Section 4 examines whether the WTO jurisprudence on the NT rule reflects the intentions of the original authors. It finds that the WTO rulings under Article III:2, 2nd sentence and Article III:4 are consistent with the draftsmen's intention. After arguing that the rulings under Article III:2, 1st sentence have not been following the interpretative guidance derived from the drafting history, this article offers recommendations for improvement. Section 5 concludes this article.
2. The debate on the role of regulatory purposes under the NT rule
2.1 The debate in a nutshell
The tension between multilateral disciplines and domestic autonomy growing out of the operation of the multilateral trading regime has been widely recognized.Footnote 3 WTO rules which compel Member governments to live up to their trade liberalization commitments inevitably impose certain constraints on their freedom to formulate and implement policy instruments for objectives unrelated to trade. The far-reaching effect of the NT rule in particular arises from the expansion of this rule to origin-neutral measures, with the potential coverage being broad enough to encompass almost all internal measures of all 153 current WTO Member states.Footnote 4 To address this tension, it has been frequently submitted that the adjudicators of WTO disputes involving non-trade concerns should endeavor to strike a balance between trade values and non-trade values, with the aim of preventing the pursuit of the former from unduly impeding the pursuit of the latter.Footnote 5
In the pre-WTO GATT era, the primary means of achieving such a balance under the NT rule was the introduction and application of the so-called ‘Aim and Effect’ test to deal with origin-neutral measures by two GATT panels. This test essentially requires the determination of ‘like product’ under Articles III:2 and 4 to consider both the aim and the effect of a contested measure pursuant to the underlying principle of ‘not applied … so as to afford protection to domestic production’ (hereinafter ‘SATAP’) as laid down in Article III:1. In US–Malt Beverages, the panel held that:
The purpose of Article III is … not to prevent contracting parties from using their fiscal and regulatory powers for purposes other than to afford protection to domestic production. Specifically, the purpose of Article III is not to prevent contracting parties from differentiating between different product categories for policy purposes unrelated to the protection of domestic production.Footnote 6
The panel observed that the application of this test could avoid unwanted encroachment on Members' regulatory autonomy.Footnote 7 The test was also applied in a subsequent GATT case – US–Taxes on Automobiles.Footnote 8
However, the WTO adjudicating bodies rejected the ‘Aim and Effect’ test at the very first NT dispute involving origin-neutral measures – the Japan–Alcoholic Beverages II case.Footnote 9 This rejection incited a spate of criticisms which, in general, contended that the illegality of an origin-neutral measure cannot be concluded without paying due consideration to the purposes of the measure. The main supporting argument in this critique is that while origin-specific measures bear an explicit and innate demonstration of protectionist intent, origin-neutral measures carry no clear sign of such intent.Footnote 10 Finding the inconsistency of the origin-neutral measures without purpose assessment may thus mistakenly invalidate those measures genuinely serving bona fide domestic policy objectives other than protectionism. Notwithstanding the fact that Article XX of the GATT provides certain justifications for such measures, it would be unlikely to resolve the tension by granting only ten exceptions, far less than what Member states may, in reality, need to pursue.Footnote 11 In the end, the entire debate on the validity of the ‘Aim and Effect’ test boils down to the issue of whether purpose inquiries are necessary under the NT rule. From that issue flows another: if purpose inquiries are necessary under the NT rule, then under which legal element of Articles III:2 and 4 should purpose inquiries be undertaken? The resolution of the issues predominantly hinges on how the NT rule is interpreted.
2.2 Interpreting the NT rule – mission unaccomplished
For the purpose of interpreting the WTO agreements, it has been endorsed that the materials contemplated in the Vienna Convention should be examined.Footnote 12 Article 31(1) of the Vienna Convention sets out primary rule of treaty interpretation:
A treaty shall be interpreted in good faith in accordance with the ordinary meaning to be given to the terms of the treaty in their context and in the light of its object and purpose.
In cases where the invocation of the primary rule renders the issues under interpretation ambiguous or obscure, Article 32 of the Vienna Convention provides that one can have recourse to other materials – such as the preparatory work – to accomplish the interpretation.Footnote 13
Arguably, obscurity or ambiguity arises in the interpretation of whether to undertake purpose inquiries under the NT rule by application of the primary materials of treaty interpretation. With respect to the text of the NT rule, the term ‘SATAP’ can be interpreted in two different ways: one interpretation is that this term merely requires the assessment of discriminatory effect of an origin-neutral measure,Footnote 14 while the other interpretation is that the term also mandates the consideration of policy aim of the measure.Footnote 15 Even if one were to accept that the ‘SATAP’ test should mandate investigations into both the effect and the purpose of a measure, one may still advance different ways of interpretation as to whether purpose should be assessed under Article III:2, 1st sentence and Article III:4. If a strict textualist interpretative approach is taken, then one may interpret the two provisions as not requiring purpose assessment on the ground that neither of them specifically invokes Article III:1, in particular the test of ‘SATAP’.Footnote 16 In contrast, if a contextual interpretative approach is taken, that is we take into account the underlying objective of the NT rule (set out in Article III:1) as not being intended to condemn internal measures in pursuit of non-protectionist policy purposes, then the two provisions can also be interpreted as to mandate consideration of regulatory purposes.Footnote 17 Considering the context beyond the NT rule itself, an interpretation that purpose assessment should not be conducted in the context of the NT rule can be developed on the basis of the argument that an examination of regulatory purposes under Article III would render Article XX redundant.Footnote 18 However, one may object to such interpretation by submitting that if purpose inquiries are to be implemented in de facto discrimination cases only, then doing so could at most imply a less frequent recourse to Article XX. Measures in violation of other GATT provisions (such as Articles II and XI:1) still need to invoke Article XX for justifications.Footnote 19 Based on this argument, one can still suggest that purpose inquiries be necessary under the NT rule.
Even if we were to accept that purpose inquiries are necessary under the NT rule, obscurity or ambiguity could also arise in the interpretation of where regulatory purposes should be examined under Article III:2, 1st sentence and Article III:4. Unlike Article III:2, 2nd sentence which imposes a separate test of ‘SATAP’, the two provisions do not contain an explicit reference to that term. Therefore, there can be an uncertainty as to under which legal elements of the two provisions the ‘SATAP’ test should be undertaken. Proponents of the ‘Aim and Effect’ approach have taken the position that ‘like product’ is the most appropriate legal element to consider policy purposes.Footnote 20 That is, the interpretation of ‘like product’ should take into account declared policy reasons underlying product differentiations. Besides the WTO tribunals' rejection of this approach, some commentators have argued that the determination of product likeness is merely a technical or threshold question which should not involve complex discussions on the policy objectives of a measure.Footnote 21 By opposing to the consideration of regulatory purposes in the step of determining ‘like product’, these commentators have implied an interpretation whereby other elements of Article III:2, 1st sentence and Article III:4 should be utilized for the assessment of purpose.
In summary, the analysis above demonstrates that the interpretation of the NT rule on the basis of the primary interpretative materials leaves the issues under interpretation ambiguous or obscure. This makes the examination of the preparatory work of the NT rule desirable.
As a matter of fact, the preparatory work of the GATT for interpretation purposes has received a good deal of attention. As Professor Jackson has observed, interpreting the GATT by reference to the preparatory work ‘can bring added certainty and predictability to the institution’.Footnote 22 In practice, GATT/WTO tribunals have made frequent reference to the preparatory work in interpretations.Footnote 23 Thus, it is clear that the negotiating records of the NT rule could be valuable sources for clarifying the issues of whether and where regulatory purposes of an origin-neutral measure should be considered in deciding the NT-consistency of the measure.
3. The preparatory work of the NT rule on the role of regulatory purpose
The NT rule was originally drafted during the negotiations of a Charter for an International Trade Organization (ITO) and tariff concessions on a multilateral basis (i.e. the GATT).Footnote 24 The preparatory work of the rule covers four sessions of negotiations from October 1946 to March 1948, including the first (London) session of the Preparatory Committee (15 October–26 November 1946), the New York session of the drafting sub-committee (20 January–25 February 1947), the second (Geneva) session of the Preparatory Committee (10 April–31 October 1947), and the Havana conference (21 November 1947–24 March 1948). The negotiating records leading up to the conclusion of the NT rule are quite complex, reflecting the intractable negotiations during the four conferences. The final version of the NT rule, which is the one existing today, grows out of the compromise of various conflicting views. Thus, it is not only important to pay attention to the changes in the major drafts of the NT clause resulting from each discussion, but it is also essential to understand how the final compromise was reached. In order to show the dynamics of the negotiations, we discuss the preparatory work in a chronological order below. All the relevant parts of the major drafts of the NT clause will be highlighted in boxes.
3.1 London conference
The negotiations in London started on the basis of the ‘Suggested Charter’ prepared by the US in which the NT clause was set out in Article 9.
Article 9 National Treatment on Internal Taxation and Regulation (Suggested Charter)
1. The products of any Member country imported into any other Member country shall be exempt from internal taxes and other internal charges higher than those imposed on like product of national origin, and shall be accorded treatment no less favorable than that accorded like product of national origin in respect of all internal laws, regulations or requirements affecting their sale, transportation or distribution or affecting their mixing, processing, exhibition or other use…
2. The Members recognize that the imposition of internal taxes on the products of other Member countries, for the purpose of affording protection to the domestic production of competitive products, would be contrary to the spirit of this Article…Footnote 25 (emphasis added)
This version of the NT clause demonstrated the position of the US on the issues of whether and where regulatory purposes should be considered. By using ‘for the purpose of affording protection’, the US evidently suggested the relevance of regulatory purposes to at least some findings of violations of the NT rule. It seems that purpose inquiries were required only in cases where internal taxes and ‘competitive products’ were at issue. On the contrary, the assessment of regulatory purposes seems to be regarded as unnecessary in cases involving ‘like product’. Therefore, at the inception of the negotiations, there was a clear distinction made in relation to the role of regulatory purposes, between, on the one hand, internal fiscal and non-fiscal measures on ‘like product’ (covered by Article 9(1)) and, on the other hand, internal fiscal measures on ‘competitive products’ (covered by Article 9(2)). The reason for such distinctions was however unclear at that stage.
3.1.1 The discussions on the NT rule at the London session
The suggested draft of the NT rule was discussed in the first meeting of the Technical Sub-Committee of Committee II.Footnote 26 The relationship between Articles 9(1) and (2) became one of the major issues in this meeting. While delegates shared the view that internal taxation should not be used for protectionist purposes,Footnote 27 they were divided on the different functions of the two paragraphs. One view, led by the US delegate, was that while Article 9(1) should mainly deal with internal measures formulated by central governments, Article 9(2) should mainly target on measures imposed by local or state or provincial governments ‘over which the central government lacked authority’.Footnote 28
The delegate for Netherlands took a different position on the division of labor between the two paragraphs. Having pointed out that both of them should be applicable to ‘obligations of both central and local governments’, the Netherlands delegate observed that:
The purpose of paragraph 1…was to prohibit higher taxes on imported products than on like domestic products, while the purpose of paragraph 2 was to guard against the more concealed types of discriminatory taxation.Footnote 29 (emphasis added)
The Netherlands delegate explained his observation with the following example:
If a country produced product A, but not product B, the two products being similar but not the same, it might wish to place an internal tax on B so as to protect A.Footnote 30 (emphasis added)
Thus, the Netherlands delegate contemplated a distinction between internal taxes discriminating against imports identical to domestic products, and internal taxes discriminating against imports not identical but similar to domestic products. Using the contemporary WTO jargon, it can be argued that this example illustrated precisely the distinction commonly acknowledged between origin-specific measures and origin-neutral measures. In other words, where the contested products are the same, then the discriminatory treatment between imported and domestic goods must be made on the basis of nothing but their nationality, constituting a case of overt discrimination (i.e. de jure discrimination). However, if the products have some differences, such as being physically different from each other, then the measures treating them differently may be based on criteria (i.e. their physical differences) other than their origin. The delegate correctly suggested that these measures, if proved to be used for the purpose of protecting domestic goods, became a covert form of discrimination (i.e. de facto discrimination). If the aforesaid was the true position of the Netherlands delegate, then this delegate indicated that the division of labor between the two paragraphs was that the 1st paragraph was meant to deal with origin-specific measures (or de jure discrimination), while the 2nd paragraph was intended to tackle origin-neutral measures (or de facto discrimination). This observation therefore implied that inquiries as to purpose were unnecessary in assessing whether origin-specific measures violated the rule in the Article 9(1), whereas such inquiries were necessary in judging whether origin-neutral measures were inconsistent with the rule in Article 9(2). This observation is significant as it is in line with the current views on the role of regulatory purposes within the NT regime. However, in suggesting that the 1st paragraph deal with discrimination between the same products, the delegate seems to have interpreted the term ‘like product’ as identical products. This interpretation is at odds with current views on the scope of ‘like product’ which, even the narrowest understanding of it in the context of Article III:2, 1st sentence, is not limited to identical products.
On the issue of ‘like product’, the US delegate provided some interpretative views in this meeting. According to the US delegate, the determination of ‘like product’ would have to be left to the development of ‘case law’.Footnote 31 This delegate however suggested that significant price differences between domestic and imported products in concern would disqualify them from ‘likeness’, provided that the differences were not attributed to discriminatory internal taxes.Footnote 32 Setting aside the reasonableness of this interpretation, the view of the US delegate indicated the relevance of market-related factors (i.e. the price difference between products in the market) to the determination of product likeness. There was however no sign of intention that non-market related factors (i.e. regulatory purposes of a measure) should play a role in finding product likeness.
3.1.2 The revisions of the NT clause at the London session
The suggested text of the NT rule was subsequently revised to reflect the different views addressed in the previous discussions.
Article 9 National Treatment on Internal Taxation and Regulation (London draft)
1. The products of any member country imported into any other member country shall be exempt from internal taxes and other internal charges of any character whatsoever higher than those imposed directly or indirectly on identical or similar products of national origin.
2. The products of any member country imported into any other member country shall be accorded treatment no less favorable than that accorded identical or similar products of national origin in respect of all internal laws, regulations or requirements affecting their sale, offering for sale, transportation distribution or use of any kind whatsoever…
3. The members agree that neither internal taxes or other charges nor internal laws, regulations or requirements should be used to afford protection directly or indirectly for any national product.
4. Each member agrees that it will take all measures open to it to ensure that the objectives of this Article are not impaired in any way by taxes, charges, laws, regulations or requirements of subsidiary governments within the territory of the member government.Footnote 33 (emphasis added)
First, to accommodate the diverging views on the different functions of Articles 9(1) and (2) of the ‘Suggested Charter’, Article 9(2) was divided into two paragraphs – Articles 9(3) and (4) of this revised NT clause. Article 9(4) was drafted to explicitly deal with internal fiscal and non-fiscal measures imposed by local or subsidiary governments, hence reflecting the view held by the US delegate. Article 9(3) was drafted to set out a general principle of not applying internal fiscal and non-fiscal measures for protectionist purposes, which reflected the view of the Netherland delegate; that is, that this provision was to combat more disguised discrimination (i.e. de facto discrimination). Noticeably, while the NT draft of the ‘Suggested Charter’ only subjected internal taxes to the discipline of not being used for protectionist purposes, Article 9(3) of this revised draft brought non-fiscal measures into that discipline. Negotiating records do not show any explanations as to why this particular change was made. However, considering the fact that non-fiscal measures can be perfect substitute for fiscal measures in pursuit of protectionism, this change appears to reflect sensible policy considerations. Despite the reasonableness of this change, there was nonetheless a clear emphasis on the discipline of internal fiscal measures. At the end of the London session, the Sub-Committee reported that the NT clause ‘was generally agreed to in principle … in that internal taxes and charges should not be used to afford protection to domestic products’Footnote 34 (emphasis added). The record of the negotiations does not contain any similarly clear statement that internal non-fiscal measures should not be used to afford protection to domestic products. As will be further demonstrated below, internal non-fiscal measures received considerably less attention than internal fiscal measures during the negotiations.
Second, Article 9(1) of the ‘Suggested Charter’ was divided into two paragraphs (i.e. Article 9(1) and Article 9(2)) to set out obligations on internal fiscal measures and non-fiscal measures separately. Although no explanations were documented for this change, it appears to have been made to make the obligations clearer than in the original Article 9(1).Footnote 35
Third, the term ‘like product’ was replaced by the term ‘identical or similar products’, while the term ‘competitive products’ was replaced by ‘any national product’. The former revision demonstrated that ‘like product’ covered similar products and hence was not limited to identical products. This revision indicated a rejection of the observation of the Netherland delegate discussed above which suggested ‘like product’ being identical products. However, this revision did not affect the view that Articles 9(1) and (2) were intended to deal with overt forms of discrimination (i.e. de jure discrimination), while Article 9(3) was meant to tackle covert forms of discrimination (i.e. de facto discrimination). Thus, in dealing with explicit origin-specific discriminatory measures between ‘like product’ under Articles 9(1) and (2), no purpose inquiries were needed. The latter revision, which replaced ‘competitive products’ with ‘any national product’, was unexpected. Without a specified product scope, the exact scope of application of Article 9(3) became unclear.
This revised draft of the NT clause, which was not substantially discussed subsequently, was adopted without any meaningful changes.Footnote 36 The final report of the London session moved the NT clause from Article 9 to Article 15.Footnote 37
3.2 New York conference
3.2.1 The discussions and revisions of the NT rule at the New York session
At an early stage of the New York session, the Technical Sub-Committee tabled another revised version of the NT rule.
Article 15 National Treatment on Internal Taxation and Regulation (New York draft)
1. The Members agree that neither internal taxes nor other internal charges nor internal laws, regulations or requirements should be used to afford protection directly or indirectly for any national product.
2. The products of any Member country imported into any other Member country shall be exempt from internal taxes and other internal charges of any kind higher than those imposed, directly or indirectly, on like product of national origin.
3. The products of any Member country imported into any other Member country shall be accorded treatment no less favorable than that accorded to like product of national origin in respect of all laws, regulations or requirements affecting their internal sale, offering for sale, transportation, distribution or use of any kind whatsoever….Footnote 38 (emphasis added)
This revised version moved paragraph 3 of the London draft to the very beginning as paragraph 1 due to the general nature of that paragraph.Footnote 39 No changes were made to the text of that paragraph. Furthermore, the Committee made a preliminary decision to change the term ‘identical or similar product’ back to the original expression – ‘like product’.Footnote 40 Finally, and most significantly, Article 9(4) of the London draft, which targeted on measures applied by subsidiary governments, was deleted, with the understanding that a more general provision was to be added to the ITO Charter to lay down the obligations of central governments on ensuring the activities of subsidiary governments to be in conformity with the provisions of the Charter.Footnote 41 Recall the discussions on the relationship between the then Articles 9(1) and (2) at the London session. By discarding the view that the separate functions of Articles 9(1) and (2) of the ‘Suggested Charter’ were to deal with measures of national and sub-national governments respectively, this third revision appears to confirm the view expressed by the Netherlands delegate at the London session that the separate function of Article 9(1) (here broken up into Articles 15(2) and (3)) was to deal with overt discrimination arising from origin-specific measures, while the separate function of Article 9(2) (here Article 15(1)) was to deal with more concealed forms of discrimination arising from origin-neutral measures.
This revised NT clause was not further discussed and was incorporated into the final report of the New York conference.Footnote 42
3.3 Geneva conference
The negotiations on the NT clause in Geneva started with a US submission which, amongst other amendments, proposed to delete Article 15(1) while adding a new sentence to Article 15(2). Article 15(2) (now Article 15(1)) reads:
1. The Products of any Member country imported into any other Member country shall be exempt from internal taxes and other internal charges of any kind higher than those imposed, directly or indirectly, on like product of national origin. Moreover, in cases in which there is no substantial domestic production of like product of national origin, no Member shall impose new or higher internal taxes on the products of other Member countries for the purpose of affording protection to the production of competitive products.Footnote 43 (the new sentence underlined and emphasis added)
3.3.1 The discussions on the NT rule at the Geneva session
The US proposal became the central issue in the first formal discussion of the NT rule in Geneva. The US delegate explained that the new sentence was added to replace the 1st paragraph which was too general to be treated as imposing binding obligations.Footnote 44 In his view, the NT rule should not be limited to ‘like product’. Rather, it should be extended to combat the use of internal taxes for the purpose of protecting domestic goods by disfavoring highly competitive imports.Footnote 45 The UK delegate lent full countenance to both the proposal and the explanation tabled by the US delegate.Footnote 46 To illustrate the issue in concern, the UK delegate provided the famous ‘Apple–Orange’ example as follows:
Let us suppose that some country in its negotiations has secured the binding of the duty on oranges. Country A gets a duty on the binding of oranges from Country B. Now, Country B after that can proceed to put on an internal duty of any height at all on oranges, seeing that it grows no oranges itself. But by putting on that very high duty on oranges, it protects the apples which it grows itself. The consequence is that the binding duty which Country A has secured from Country B on its oranges is made of no affect, because in fact the price of orange is pushed up so high by this internal duty that no one can buy them. The Consequence is that the object of that binding is defeated. It seems to me that that is a point that has to be taken care of, and I think that the United States amendment is well-conceived to take care of it.Footnote 47
This illustration can be regarded as an elaboration of the example tabled by the Netherlands delegate at the London session. Both examples concerned the situation where, in the absence of domestic production of certain products which were ‘like’ imports, an importing country may employ discriminatory internal fiscal measures for the purpose of protecting competitive products produced substantially at home. According to the UK delegate, the newly added sentence set out a more concrete and clearer obligation than Article 15(1), and thus found no reason to retain it.Footnote 48 This was true in the sense that this new sentence, by using the term ‘competitive products’ instead of the term ‘any national product’, clarified the product scope of Article 15(1). Therefore, if, as discussed above, the previous negotiations revealed the draftsmen's intention to have Article 15(1) to tackle more disguised discrimination (i.e. de facto discrimination), then it can be argued that the new sentence was intended to perform that function within the boundary imposed by the term ‘competitive products’.Footnote 49 The US delegate later testified to a US Senate Committee that the new sentence was meant to deal with internal measures having the indirect effect of disfavoring imports vis-à-vis competitive domestic goods.Footnote 50 However, while proposing the new sentence to clarify the disciplines on internal fiscal measures, the US delegate did not make an equivalent proposal to better reflect the disciplines on internal non-fiscal measures as also contemplated in Article 15(1). It was unclear why only internal fiscal measures were dealt with. As stated above, the reason might be the greater emphasis placed on disciplining internal fiscal measures than on regulating non-fiscal measures throughout the entire negotiations.
Many delegates disagreed significantly with the US proposal.Footnote 51 The delegate for Norway maintained that the deletion of the 1st paragraph should not be dependent upon the insertion of the new sentence in the 2nd paragraph.Footnote 52 In the view of this delegate, ‘paragraph 2 and the first sentence of paragraph 3 are the complete application of the principle laid down in paragraph 1’ (emphasis added).Footnote 53 Therefore, it seems that what was in the mind of the Norwegian delegate was an application of the NT rule to ‘like product’ only. Taking into account the view on the division of labor between Article 15(1) and Articles 15(2)and(3) as discussed above, the statement of this delegate implied an intention to limit the application of the NT rule to de jure discrimination only.
The delegate for Cuba agreed with the proposal submitted by the Norwegian delegate, but for a more specific reason. What concerned the Cuban delegate was that the obligations imposed by the 1st paragraph could be too broad to fulfill and, in particular, would impair the ability of underdeveloped countries (like Cuba) to pursue industrialization.Footnote 54 Therefore, like the Norwegian delegate, it seems that the Cuban delegate also supported a narrower application of the NT rule with no intention to expand this rule to cover more-concealed forms of discrimination (i.e. de facto discrimination).
Like the delegates for Norway and Cuba, the Chilean delegate also opposed vigorously the insertion of the new sentence on the ground that this new sentence ‘goes beyond the scope of the problem which is laid down in paragraph 1 of Article 15’.Footnote 55 Although the Chilean delegate did not expound his understanding on the scope of Article 15(1), the reason provided by this delegate indicated his concurrence with the view that the NT rule should be confined to the obligations set out in Articles 15(2) and (3), which were intended to condemn de jure discrimination only.
The delegate for Belgium was not in favor of the deletion of the 1st sentence but did support the addition of the new sentence.Footnote 56 In response to the aforesaid statement of the Cuban delegate, the Belgian delegate pointed out that limiting the function of the NT rule by deleting the 1st paragraph did not necessarily allow discriminatory measures for protectionist purposes. The Belgian delegate stated that internal measures which were ‘directly and specifically meant for the protection’ would still be caught by the 2nd and the 3rd paragraphs.Footnote 57 This was a remarkable statement, which articulated the limited function of the 2nd and the 3rd paragraphs as being to tackle overt forms of discrimination whose purpose was conspicuously to protect domestic production. The other reason why the Belgian delegate was in opposition to the deletion of the 1st sentence was because it would constrain the NT discipline in dealing with measures other than taxes.Footnote 58 Therefore, in the view of this delegate, to keep the 1st sentence status quo would impose broader NT obligations on the use of internal non-fiscal measures. If the 1st sentence was meant to tackle de facto discrimination as discussed above, then the Belgian delegate's view was a suggestion that internal fiscal and non-fiscal measures should be equally restrained from being applied for such discrimination. This was the only instance that the author found in the official negotiating records of any clear support for equal disciplines to be imposed on the protectionist use of implicit discriminatory internal fiscal and non-fiscal measures. In contrast, a majority of the negotiators paid no attention to non-fiscal measures. The Chinese delegate even expressly objected to regulating non-fiscal measures as strictly as fiscal measures by proposing that the NT rule be confined to taxation only.Footnote 59
3.3.2 The revisions of the NT rule at the Geneva session
Subsequent to the discussions in this meeting, the Sub-Committee decided to consider redrafting the 2nd paragraph, and based on a revised draft to consider whether to delete the 1st paragraph.Footnote 60 A revised draft was then prepared for further negotiations. This draft moved the NT clause from Article 15 to Article 18.
Article 18 National Treatment on Internal Taxation and Regulation (Geneva draft)
1. The products of any Member country imported into any other Member country shall be exempt from internal taxes and other internal charges of any kind in excess of those applied directly or indirectly to like product of national origin. Moreover, in cases in which there is no substantial domestic production of like product of national origin, no Member shall apply new or increased internal taxes on the products of other Member countries for the purpose of affording protection to the production of directly competitive or substitutable products which are not similarly taxed…
2. The products of any Member country imported into any other Member country shall be accorded treatment no less favorable than that accorded to like product of national origin in respect of all laws, regulations and requirements affecting their internal sale, offering for sale, purchase, transportation, distribution, or use…Footnote 61 (emphasis added)
Obviously, this revised draft fully reflected the proposal lodged by the US. Other parties reserved the above-discussed disagreements on this draft.Footnote 62 Subsequently, negotiations of the NT rule at the Geneva session were devoted to the conclusion of the GATT by incorporating the drafted articles of the ITO Charter concerning general commercial policies into it. This revised draft of the NT rule was incorporated in totality into the GATT after an intense discussion at the tenth meeting of the Tariff Agreement Committee.Footnote 63 However, no further discussions occurred in that meeting on the issues relating to the deletion of paragraph 1 and the insertion of the additional sentence into paragraph 2.Footnote 64 This revised draft was finally incorporated into the Geneva Report, which also recorded the various dissenting views retained by the other parties.Footnote 65
3.3.3 Tentative conclusions on the role of regulatory purpose
As to the conclusion of the Geneva conference, at least three insights can be drawn in terms of the intentions of the draftsmen on the issues of whether and where to consider regulatory purposes.
First, there was a common view that the 1st paragraph which set out the underlying principle of the NT rule should be deleted. Some delegates (i.e. the delegates for the US and the UK) believed that this paragraph was too generally drafted and should be replaced by the new sentence to paragraph 2, which specified the obligations envisioned in the 1st paragraph. For these delegates, the function of paragraph 15(1) of the New York draft and of its replacement in the 2nd sentence of the new draft of Article 18(1) was to tackle de facto discrimination arising from origin-neutral measures. In dealing with such discrimination, there was a specified legal requirement of examining regulatory purposes of the measures. Other delegates (i.e. the delegates for Norway and Chile) took the position that the scope of paragraph 15(1) should be as narrow as the specific obligations set out in paragraph 15(2) on internal fiscal measures (without the new sentence) and in paragraph 15(3) on internal non-fiscal measures. One delegate (i.e. the Cuban delegate) submitted that if the 1st paragraph went beyond the obligations laid down in paragraphs (2) and (3), then the 1st paragraph should be deleted. These delegates were, thus, of the view that the NT rule should be limited to attacking de jure discrimination arising from origin-specific measures. Only the Belgian delegate opposed the deletion of the 1st paragraph for the reason that its deletion would render the NT rule less effective in regulating implicitly discriminatory non-fiscal measures than in combating implicitly discriminatory fiscal measures.
Second, apart from the Belgian delegate's concern with non-fiscal measures, there were almost no specific discussions of the NT disciplines on the protectionist use of non-fiscal measures. This lack of consideration might arise from the fact that fiscal measures were more frequently used domestic policy instruments than non-fiscal measures at the time of the negotiations. It may also arise from the unwillingness of negotiators to constrain non-fiscal measures as much as fiscal measures.
Third, despite the recognition of the potential interpretative problems associated with the terms ‘like product’ and ‘competitive products’, there was no effort documented to provide further clarifications on those terms. The US delegate's suggestion of considering price differences between products remained to be the only interpretative view tabled in the meetings. From this proposal, it seemed that the interpretation of ‘like product’ was not intended to take into account factors other than market-based ones.
3.4 Havana conference
3.4.1 The discussions and revisions of the NT rule at the Havana session
At the Havana session, the draftsmen maintained the conflicting views on the addition of the new sentence to paragraph 2 (now paragraph 1 due to the deletion of the 1st paragraph of the New York draft). As noted by the Chairman of the Third Committee, ‘there was general agreement regarding the principle of Article 18, but differences as to degree and scope of its application.’Footnote 66 The US delegate stressed that the proposed new sentence did not depart from the NT principle but strengthened it.Footnote 67 However, other delegates continued to treat this new sentence as being unacceptable due to the broad obligations it might impose.Footnote 68 Therefore, the negotiators were still divided on the issue of whether the NT rule should be applied to de facto discrimination. To elaborate their views, the parties proceeded to propose substantive amendments to the Geneva draft. The US position was unchanged, proposing to keep the two-sentence structure of paragraph 2 without a general principle clause (i.e. paragraph 1 of the New York draft).Footnote 69 Contrary to the US proposal, the Norwegian delegate submitted a proposal where all the main provisions of the NT rule made explicit references to the underlying principle of this rule; that is, to prohibit internal measures with protectionist purpose only.Footnote 70 This delegate explained that the main purpose of the proposed amendments was to prevent non-protectionist measures from being wrongly condemned by this rule.Footnote 71 After considering all these views, the Working Party of Committee III produced a revised draft of the NT rule.
Article 18 National Treatment on Internal Taxation and Regulation (Havana draft)
1. The Members recognize that internal taxes and charges, and laws, regulations or requirements affecting the internal sale, offering for sale, purchase, transportation, distribution, mixture, processing or use of products, should not be applied to imported or domestic products so as to afford protection to domestic production.
2. The products of any Member country, imported into any other Member country shall not be subject, directly or indirectly, to internal taxes or other internal charges of any kind, in excess of those applied, directly or indirectly, to like domestic products. Moreover, no Member shall otherwise apply internal taxes or charges, to imported or domestic products in a manner contrary to the principles set forth in Paragraph 1.
Interpretative note to paragraph 2: A tax conforming to the requirements of the first sentence of paragraph 2 would be considered to be inconsistent with the second sentence only in cases where competition was involved between, on the one hand, the taxed product, and on the other hand, a directly competitive or substitutable product which was not similarly taxed.”
4. The products of any Member country imported into any other Member country shall be accorded treatment no less favorable than that accorded to like product of national origin in respect of all laws, regulations, and requirements affecting their internal sale, offering for sale, purchase, transportation, distribution, or use …Footnote 72 (emphasis added)
This revised draft adopted the US proposed structure of the NT rule while reinserting the general principle clause to reflect the Norwegian proposal. It was this draft, standing without substantial revisions, which became the final version of Article III of the GATT.Footnote 73
3.4.2 Conclusions and Implications on the role of regulatory purpose
A number of conclusions and implications can be drawn from this NT revision which eventually produced a compromise acceptable to all negotiating parties.
First, the negotiations at Havana did not affect the conclusions reached at the end of the Geneva conference. As noted by the Third Committee, the only substantial change to the Geneva draft was the outright elimination of the condition imposed on the application of Article 18(2), 2nd sentence (Article 18(1), 2nd sentence of the Geneva draft); that is, that there must be no substantial domestic production of ‘like product’.Footnote 74 Therefore, there was still the same intention with respect to the division of labor between the new provisions as there was with the earlier draft from which they were derived. On the one hand, Article 18(2), 1st sentence and Article 18(4) were aimed at origin-specific measures causing de jure discrimination and, on the other hand, Article 18(1) and the 2nd sentence of Article 18(2) were aimed at origin-neutral measures which might cause de facto discrimination. Evidently, in the determination of whether origin-neutral measures constituted a breach of the NT rule as de facto discrimination, it was contemplated that an assessment of the purpose of the measure would be necessary and this was intended to be mandated by the inclusion of the term ‘so as to afford protection’. Without a doubt, this term was used interchangeably with the term ‘for the purpose of affording protection’ as had been used in the ‘Suggested Charter’.
Second, the reinsertion of the general principle clause (i.e. Article 18(1)) must have some meaning. In other words, it could not be the case that this clause was added for no purpose or that despite having been added, it was intended to perform no substantive functions since the substance of that clause had been embedded into Articles 18(2) and (4). If that were the case, it would seem unlikely that the compromise would have been reached. If that were the case, the Geneva draft of the NT rule, which abandoned the general principle clause, should have been adopted and the reintroduction of that clause should have not occurred at the Havana session. What could be the compromise? In the face of the vigorous push for the broad application of the NT rule (especially from the US and the UK), those delegates in favor of a narrow application of this rule to de jure discrimination between ‘like’ products gradually realized that they were fighting for a losing case. Eventually, they decided to withdraw their disagreement on the basis of the revised draft which reintroduced the general principle clause.Footnote 75 With the reintroduction of this clause, these delegates pushing for the narrow scope of the NT rule found another way to realize their pursuit; that is, internal measures, be it fiscal or non-fiscal ones, serving policy purposes having nothing to do with protectionism should not be treated as contrary to the requirement of NT. Accordingly, the reinsertion of the general principle clause can be regarded as being intended to limit the scope of the NT rule to dealing with discriminatory internal measures with protectionist purpose only.Footnote 76 As there could be little doubt that regulatory purposes are a non-issue in de jure discrimination cases, the anti-protectionist principle embodied in this clause should be fundamental to the operation and application of the NT rule in de facto discrimination cases arising from origin-neutral measures.
Third, no further discussions were made on the terms ‘like product’ and ‘directly competitive or substitutable products’ (hereinafter DCSPs). Thus, the insight drawn at the conclusion of the Geneva conference stood; that is, that the interpretation of the terms should be based on market-related factors rather than policy-related factors. It may be argued that this conclusion tends to be weak as it is based on the single statement made by the US delegate. However, one should also take into account the fact that there was no sign of intention in the negotiating records indicating that policy-related factors are relevant to the determination of product likeness.
Fourth, the negotiating parties continued to pay substantially less consideration to non-fiscal measures than to fiscal measures. As a matter of fact, the text of the current Article 18(4), 1st sentence was neither seriously discussed nor substantively revised during the entire negotiations. Thus, it remained to be the case that, like Article 18(2), 1st sentence, the disciplines on internal non-fiscal measures under Article 18(4) were intended to deal with origin-specific measures causing de jure discrimination only. However, it is possible to interpret Article 18(4) in a way to also deal with origin-neutral measures by virtue of the operation of Article 18(1) which was intended to cover de facto discrimination. Without paying sufficient attention to internal non-fiscal measures, the negotiators have failed to construct a similar two-sentence structure in the context of Article 18(4) as they have done under Article 18(2). Thus, an interpretation of Article 18(4) as being applicable to origin-neutral measures must be based on reading the general obligation contemplated in Article 18(1) into Article 18(4). By doing so, the application of Article 18(4) to tackle de facto discrimination arising from origin-neutral measures must involve considerations of regulatory purposes.
4. Interpretative guidance and NT jurisprudence
This section examines whether the WTO NT jurisprudence in relation to purpose inquiries (i.e. whether and where to consider purposes) is consistent with the interpretative guidance extracted from the preparatory work.
4.1 Rejecting the ‘Aim and Effect’ approach to determining ‘like product’
The negotiating history of the NT rule suggests that the determination of ‘like product’ should rely on market-based factors rather than policy-based factors. The WTO tribunals' dismissal of the traditional ‘Aim and Effect’ test and adoption of a market-based approach to determining ‘like product’ and ‘DCSPs’ are thus in harmony with the draftsmen's intention.
In Japan−Alcoholic Beverages II, the panel upheld the EC's contention that the ‘Aim and Effect’ test was not an appropriate test under the NT rule.Footnote 77 On appeal, the Appellate Body indicated its agreement with the panel's rejection of the test. The Appellate Body ruled that it was a consistent practice of GATT panels to rely on the factors envisaged in the ‘1970 Working Party Report on Border Tax Adjustments’Footnote 78 (hereinafter the BTA report).Footnote 79 In its view, the practice should be continued and factors including products' physical properties, nature and quality, their end-uses, and consumers' tastes and habits should be examined to determine product likeness.Footnote 80 This approach to judging ‘like product’ has gradually evolved into a market-based approach, which has been applied consistently by the WTO tribunals. The discussion of the EC–Asbestos Footnote 81 case in Section 4.3.1(a) below demonstrates that the tribunals have established a market-based approach to determining product likeness and attempted to distance themselves from the ‘Aim and Effect’ test.
The rejection of the ‘Aim and Effect’ test is not only laudable because it conforms to the intention of the original authors of the NT rule; in practice, the application of the test could give rise to the following problems. First, applying this test under Article III:2, 1st sentence would render the 2nd sentence of that article superfluous. The protective purpose and effect of an origin-neutral measure are in fact the two determinants of whether the measure constitutes de facto discrimination. Assessing and making decisions on them under the product likeness test would thus potentially bring all internal fiscal measures under the surveillance of the 1st sentence. Even in cases where products are found to be ‘unlike’ under the 1st sentence due to the finding(s) of non-discriminatory effect and/or legitimate purposes, this observation remains valid as any further examinations under the 2nd sentence would most likely constitute a simple repeat of what has been done under the 1st sentence. Accordingly, the examination of the aim and effect of a measure under the product likeness test would deprive the other legal elements that substantively implement the principle ‘SATAP’ of effet utile.Footnote 82 Second, the ‘Aim and Effect’ approach to determining product likeness under the 1st sentence would expand the scope of ‘like product’ to encompass ‘DCSPs’ in cases where discriminatory effect and protectionist purpose are found. This is in stark conflict with the settled case law that ‘like product’ in the 1st sentence represents a much narrower category than ‘DCSPs’.Footnote 83 For example, domestic apples and imported oranges would be found to be ‘like’ insofar as the latter are found to be subject to a discriminatory tax for the purpose of protecting the former. However, they should, at most, be treated as ‘DCSPs’ rather than ‘like’, if the ‘likeness’ test does not involve examinations of the purpose and effect of the tax measure. Third, since the traditional ‘Aim and Effect’ test relies exclusively on the purpose and effect of a measure, the application of the test would render all other factors (including the ones contemplated in the BTA report) relevant to the determination of ‘like product’ superfluous.
Indeed, the merit of the ‘Aim and Effect’ test lies in mandating purpose inquiries. However, it is submitted that the test incorporates the assessment of purpose into the wrong element of the text of Article III. The product comparison should be regarded as a threshold question to address the preliminary issue of whether a contested measure falls under the ambit of the NT rule.Footnote 84 In other words, the legal elements of “like product” in Article III:2, 1st sentence and Article III:4, and “DCSPs” in Article III:2, 2nd sentence should be used to assess whether there is a competitive relationship between contested domestic and imported products as a pre-requisite step to then applying the substantive NT rules in the other legal elements. It is one thing to ask whether imported products are ‘like’ domestic goods in a given market so as to determine whether they should be entitled to equal competitive condition, it is quite another to ask whether imports that are entitled to competitive conditions as favorable as those applicable to domestic products should nonetheless be treated differently from domestic goods so as to pursue certain bona fide regulatory purposes unrelated to protectionism. In short, the ‘like product’ test merely serves to activate the application of the NT rule. The assessment of the effect and aim of an origin-neutral measure should be remitted to the other legal elements of Articles III:2 and 4.
4.2 Purpose inquiries under Article III:2, 2nd sentence through the test of ‘SATAP’
As far as Article III:2, 2nd sentence is concerned, the evaluation of the preparatory work revealed that this sentence was intended to deal with implicit discriminatory fiscal measures. The fact that the term ‘SATAP’ was used interchangeably with ‘for the purpose of affording protection’ leaves little doubt that purpose inquiries should constitute an indispensible component of the ‘SATAP’ test in determining the NT-legality of contested measures. The current jurisprudence in this regard, primarily cultivated by the Japan–Alcoholic Beverages II decision and the Chile–Alcoholic Beverages Footnote 85 decision, is consistent with the approach envisioned by the draftsmen of the NT rule.
4.2.1 Japan–Alcoholic Beverages II
The Japan–Alcoholic Beverages II case concerned the Japanese Liquor Tax Law which applied different tax rates to various alcoholic beverages categorized on the basis of their alcohol strength.Footnote 86 The Japanese measure was challenged as being in violation of both sentences of Article III:2. In its interpretation of ‘SATAP’ under the 2nd sentence, the Appellate Body held that:
This is not an issue of intent. It is not necessary for a panel to sort through the many reasons legislators and regulators often have for what they do…Although it is true that the aim of a measure may not be easily ascertained, nevertheless its protective application can most often be discerned from the design, the architecture, and the revealing structure of a measure.Footnote 87 (emphasis added)
Thus, the Appellate Body dismissed the relevance of the subjective intent behind the formulation of measure. However, in fashioning a test of the measure's protective application, the Appellate Body was most likely thinking of its protectionist purpose.Footnote 88 In other words, the Appellate Body did not rule out the examination of relevant objective evidence, such as the design and structure of the measure, to assist in discerning the genuine purpose of the measure.Footnote 89
4.2.2 Chile–Alcoholic Beverages
If the Appellate Body's decision in Japan–Alcoholic Beverages II was not sufficiently clear in relation to the issue of whether to assess regulatory purposes under Article III:2, 2nd sentence, it subsequently made its position crystal clear in Chile–Alcoholic Beverages.
This case concerned Chile's traditional and new tax systems on alcoholic beverages. As the traditional system was essentially an origin-specific measure, our interest here is in the new system only.Footnote 90 Under its new tax system on alcoholic beverages, Chile differentiated alcoholic beverages according to alcohol strength and imposed different ad valorem excise tax rates on different categories.Footnote 91 This Chilean measure was challenged as being contrary to Article III:2, 2nd sentence. In deciding whether the measure was applied ‘SATAP’, the Appellate Body rejected Chile's claim that the measure's purpose was not relevant to the assessment of its protective application.Footnote 92 In stark contrast, the Appellate Body elucidated that at the time that it was rejecting the relevance of subjective intent and creating the test of protective application in Japan–Alcoholic Beverages II, it meant to say that the objective expression of the regulatory purpose as a whole would be pertinent to the ‘SATAP’ testFootnote 93 (emphasis added). The Appellate Body stressed that:
a measure's purposes, objectively manifested in the design, architecture, and structure of the measure, are intensely pertinent to the task of evaluating whether or not that measure is applied so as to afford protection to domestic production.Footnote 94
The Appellate Body's statement leaves us in no doubt that an examination of the underlying purpose of an origin-neutral measure in determining ‘SATAP’ is inextricable from the analysis of whether the measure is inconsistent with Article III:2, 2nd sentence. This approach is in harmony with the intention of the original authors of the NT rule.
4.3 Purpose inquiries under Article III:4 through the test of ‘less favorable treatment’
Article III:4, which deals with internal non-fiscal measures, was not seriously discussed during the negotiations. The analysis of the preparatory work indicated that this provision can be applied to both origin-specific measures and origin-neutral measures. To decide whether an origin-neutral measure is in conflict with Article III:4, an assessment of regulatory purposes is necessary. The negotiating records also suggest that a determination of ‘like product’ should consider market-based factors rather than policy-based ones. This would mean that the element of ‘less favorable treatment’ should be employed to undertake purpose inquiries. Three WTO cases demonstrate that the tribunals have been heading in the direction that the draftsmen wanted them to take: the EC–Asbestos case, the Dominican Republic–Cigarettes Footnote 95 case, and the EC–Biotech Products Footnote 96 case.
4.3.1 EC–Asbestos
In EC–Asbestos, France imposed a ban on the importation, exportation, and use of all varieties of asbestos fibres and any product containing asbestos fibres, irrespective of their origin.Footnote 97 Amongst other claims, Canada challenged this ban under Article III:4.Footnote 98 In assessing the NT-legality of the ban, the WTO tribunals have applied Article III:4 to origin-neutral measures.
(a) Market-based approach to determining ‘like product’ – no purpose inquiries
In deciding product likeness, the Appellate Body endorsed a market-based approach by suggesting the use of the same tests as those applied to test “DCSPs” under Article III:2, 2nd sentence. A brief comparison between the Appellate Body's approach to determining ‘like product’ in EC–Asbestos and its approach to ‘DCSPs’ in previous NT cases helps bear this point out. In short, both determinations hinge on the competitive relationship between products at issue in the marketplace.Footnote 99 Both the test of product likeness and the test of products' competitiveness should examine the factors envisioned in the BTA report.Footnote 100 Of these factors, products’ end-uses and consumers’ tastes and habits constitute the key criteria for testing product likeness/competitiveness.Footnote 101 However, it is likely that potentially competitive products may not be treated as competitive by consumers due to government interventions in the market in concern. For this reason, the latent demand of consumers and products’ competitiveness or substitutability in a third/surrogate market are also relevant to determining product likeness/competitiveness under Article III:4 and Article III:2, 2nd sentence respectively.Footnote 102 Thus, the author submits that, by establishing a quasi ‘DCSPs’ test, the Appellate Body has adopted a market-based approach to “like product” under Article III:4, implying the irrelevance of policy-related factors.
However, the Appellate Body also ruled that, in addition to the BTA report factors, all other pertinent evidence should be examined in the product likeness decision.Footnote 103 Following this ruling, the Appellate Body found that the panel erred in declining to consider evidence related to health risks associated with chrysotile fibres not associated with the substitute fibres.Footnote 104 Some commentators have argued that in approving consideration of the lethal effect of products, the Appellate Body had implicated purpose inquiries (i.e. protecting human health) in deciding ‘like product’.Footnote 105 In this author's view, this argument is not acceptable for three reasons.
First, the Appellate Body merely considered the evidence relating to the toxicity of chrysotile fibres as part of the criteria of products' physical characteristics and consumers' tastes and habits, and hence did not treat such evidence as a separate criterion in determining ‘like product’.Footnote 106 In other words, it is argued that only when such evidence may affect the competitive relationship between products in the market would this evidence be regarded as relevant to the finding of product likeness. By its ruling, the Appellate Body relied on market-based factors rather than policy-based ones.
Second, even taking into account policy-related evidence, the Appellate Body demonstrated that such evidence was not conclusive in deciding product likeness, unlike market-based evidence reveals products competitiveness, which was conclusive. Indeed, the Appellate Body's reversal of the panel's finding of product likeness in this case appears to rely heavily upon the evidence disclosing the deadly effects of chrysotile fibres.Footnote 107 However, this reliance by no means constituted an attempt to make determinative such evidence. The Appellate Body was hesitant in naming the contested fibres ‘like product’ for two reasons: (1) their striking physical difference resulted from the toxic properties of chrysotile fibres rather than PCG fibres; and, more importantly, (2) the strong presumption of product ‘unlikeness’ stemming from this physical difference was not successfully removed as a result of the panel's inadequate examination of the products' end-uses and the lack of evidence relating to consumers' tastes and habits.Footnote 108 In other words, the ‘like product’ claim was unsuccessful because Canada failed to discharge its burden of producing adequate evidence relating to the two most important criteria in judging product likeness under Article III:4, namely products' end-uses and consumers' tastes and habits. This confirmed that the Appellate Body adhered to a market-based approach to determining ‘like product’. Furthermore, the very reason why the Appellate Body had approved the consideration of latent competition and third-market competition was because the evidence relating to consumers' tastes and habits in the market of a regulating Member may not reflect genuine consumer perceptions on the competitiveness of contested products due to government interventions.Footnote 109 Therefore, an examination of evidence relating to latent competition and third-market competition would reinforce the market-oriented approach by ruling out any considerations of regulatory distinctions structured by governments. At the end of the day, whatever objectives government interventions might have, they should not stop the application of the NT rule to products capable of becoming competitive in the market.
Third, in a concurring statement, one Member of the Appellate Body was inclined to find ‘unlikeness’ between chrysotile fibres and PCG fibres on the sole basis of the deadly nature of the former.Footnote 110 In not sharing this opinion, the majority of the Appellate Body sent the message that in deciding product likeness under Article III:4, policy justifications for a measure, if they do play a role, should not override evidence substantiating products' competitiveness in the marketplace.
In the light of these arguments, one may conclude that the Appellate Body in EC–Asbestos deliberately avoided the traditional ‘Aim and Effect’ test as it did in previous Article III:2 cases.Footnote 111 This approach conforms with the intention of the draftsmen of the NT rule by not relying on policy-related factors in determining product likeness.
(b) ‘Less favorable treatment’ – purpose inquiries?
Since the Appellate Body found the contested products ‘unlike’, it would be unnecessary for it to make decisions under the element of ‘less favorable treatment’. However, the Appellate Body provided a remark that has become one of the most oft-cited rulings on this element. The Appellate Body observed that:
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 favorable treatment” than it accords to the group of “like” domestic products. The term “less favorable 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 favorable treatment” of the group of “like” imported products, there is, conversely, “protection” of the group of “like” domestic products.Footnote 112 (original italics and underlines added)
This remark can be interpreted in two ways. The first interpretation is that by reading the test of ‘SATAP’ into ‘less favorable treatment’, the Appellate Body implicitly approved the assessment of regulatory purposes under the latter.Footnote 113 This interpretation is supported by the preparatory work of the NT rule. The second interpretation is, however, that by requiring an examination of whether a group of ‘like’ imports are less favorably treated than a group of domestic products, the Appellate Body was merely concerned with the issue of whether a measure has a discriminatory effect on imports (i.e. inflicting disproportionate impact on imports while favoring a substantial number of domestic goods); that is, without a view to approving purpose inquiries.Footnote 114 It is unfortunate that the Appellate Body did not have the chance to further clarify whether purpose should be examined under the rubric of ‘less favorable treatment’ in this case. Clarification of this issue was not provided until the Dominican Republic–Cigarettes Case.
4.3.2 Dominican Republic–Cigarettes
The Dominican Republic–Cigarettes Case involved a range of measures dealing with the import and sale of cigarettes. One of these measures, which was challenged under Article III:4, was the bond requirement imposed on both domestic producers and importers of cigarettes in order to ensure the payment of the consumption tax of cigarettes.Footnote 115 One aspect of the bond requirement that was subject to the challenge was the imposition of a fixed amount of RD$5million on both domestic producers and importers.Footnote 116 Honduras claimed that by setting a fixed amount, the bond requirement effectively subjected imported cigarettes to a higher per-unit cost due to the higher market share occupied by domestic cigarettes as compared to imported ones.Footnote 117 In rejecting the Honduras' assertion, the Appellate Body offered the following reasoning:
the existence of a detrimental effect on a given imported product resulting from a measure does not necessarily imply that this measure accords less favorable treatment to imports if the detrimental effect is explained by factors or circumstances unrelated to the foreign origin of the product…Footnote 118 (emphasis added)
The Appellate Body's ruling explicitly abandoned the second interpretation of the ‘less favorable treatment’ test (contemplated at the end of the discussion of the EC–Asbestos case above). This ruling suggests that a finding of ‘less favorable treatment’ would require not only an examination of the disparate impact of a measure, but also an examination of the factors responsible for the impact.Footnote 119 All such factors may need to be taken into account before concluding whether imported products are less favorably treated than domestic ones. Insofar as the impact cannot be attributed to the foreignness of imported goods, an affirmative conclusion on ‘less favorable treatment’ cannot be made. Following this ruling, it seems that an origin-neutral internal regulation that incidentally inflicts a disparate impact upon imports would not be found as a case of treating imports less favorably than ‘like’ domestic ones so long as such an impact can be explained by certain non-protectionist policy objectives. Accordingly, the Appellate Body successfully created the flexibility to undertake formal purpose inquiries under Article III:4 so that non-protectionist origin-neutral measures would not be incorrectly invalidated.
Unfortunately, the Appellate Body did not have a chance to specifically consider any regulatory purpose of the measure in this case, which may still have invited doubt as to whether regulatory purposes of measures would fall into the Appellate Body's list of factors unrelated to origin. This doubt was subsequently dispelled by the panel in EC–Biotech Products.
4.3.3 EC–Biotech Products
The EC–Biotech Products case concerned the EC's regime for the approval of biotech products and measures restricting the marketing of these products.Footnote 120 In the context of Article III:4, certain product-specific measures affecting the approval of specific biotech products were challenged as treating biotech products less favorably than non-biotech products.Footnote 121 The panel disapproved this claim on the ground that:
Argentina is not alleging that the treatment of products has differed depending on their origin. In these circumstances, it is not self-evident that the alleged less favorable treatment of imported biotech products is explained by the foreign origin of these products rather than, for instance, a perceived difference between biotech products and non-biotech products in terms of their safety, etc.Footnote 122 (emphasis added)
Thus, in applying the approach developed by the Appellate Body in Dominican Republic–Cigarettes, the panel endeavored to clarify that approach by specifying one of the factors unrelated to origin, namely the difference between biotech products and non-biotech products in terms of their potential risks to humans. This clarification confirmed that evidence relating to regulatory purposes of measures (i.e. protecting human health) can serve as an explanation for alleged disparate impacts upon imports. Therefore, as of the panel's decision in EC–Biotech Products, the WTO jurisprudence on Article III:4 has achieved conformity with the intention of the draftsmen of the NT rule by opening the door for purpose inquiries under “less favorable treatment” in dealing with origin-neutral non-fiscal measures.
4.4 Application of Article III:2, 1st sentence to origin-neutral measures
The intended application of Article III:2, 1st sentence was confined to de jure discrimination in which purpose inquiries were considered unnecessary. However, the WTO tribunals have also applied that sentence to origin-neutral measures. As the negotiating records suggest, such a broad interpretation of the 1st sentence would require the implementation of the explicit requirement of purpose inquiries set out in Article III:1. In this connection, it is worth pointing out that the extension of the 1st sentence to deal with origin-neutral measures would not find the same policy justification as the expansion of Article III:4 for this purpose. The two-sentence structure of Article III:2 suggests a clear division of labor between the 1st sentence and the 2nd sentence. As long as de facto discrimination can be adequately dealt with by the 2nd sentence, an interpretation of the 1st sentence would not be necessary to tackle the same type of measures. In other words, it would suffice if the combined application of the two sentences of Article III:2 could capture discriminatory internal fiscal measures for protectionist purposes. The fact that the 2nd sentence is mainly responsible for combating de facto discrimination provides a good reason to limit the application of the 1st sentence to de jure discrimination.
The case law has demonstrated the efforts of the WTO tribunals to avoid the application of Article III:2, 1st sentence to origin-neutral measures. The approaches that the tribunals have undertaken are to narrowly squeeze the scope of ‘like product’ (hereinafter the ‘narrow likeness approach’) and to create flexibility in deciding ‘like product’ on a case-by-case basis (hereinafter the ‘case-by-case approach’). However, as will be argued below, the WTO tribunals' approaches are unable to exclude origin-neutral measures from the reach of the 1st sentence. Therefore, purpose inquiries remain necessary under that sentence.
4.4.1 Japan–Alcoholic Beverages II
In Japan–Alcoholic Beverages II, the Appellate Body upheld the panel's rejection of a separate test of ‘SATAP’ under Article III:2, 1st sentence due to the lack of reference to Article III:1.Footnote 123 However, the Appellate Body recognized that in the absence of a separate consideration of the general principle, the 1st sentence tended to be too strict. Therefore, the Appellate Body emphasized that:
the first sentence of Article III:2 must be construed narrowly so as not to condemn measures that its strict terms are not meant to condemn. Consequently … the definition of “like product” in Article III:2, first sentence, should be construed narrowly. How narrowly is a matter that should be determined separately for each tax measure in each case. We agree with the practice under the GATT 1947 of determining whether imported and domestic products are “like” on a case-by-case basis.Footnote 124 (emphasis added)
This raises the question of which measures found inconsistent with the 1st sentence (by virtue of the tests of ‘like product’ and ‘in excess of’) are not meant to be condemned by the NT rule. The answer lies in what would not be examined to determine the NT-illegality of a measure under that sentence. Obviously, in the absence of a separate test of ‘SATAP’, the factor that one would fail to consider is the regulatory purpose of the measure. Without a formal investigation of purpose, one could mistakenly invalidate origin-neutral measures serving non-protectionist policy objectives. To avoid the occurrence of such an error, the Appellate Body developed the requirement that ‘like product’ under the 1st sentence must be narrowly interpreted so as to limit the reach of that sentence. In practice, the establishment of product likeness under that sentence would require virtually identical physical characteristics and common end-uses of the products in concern.Footnote 125
Could the ‘narrow likeness approach’ satisfactorily avoid this potentially incorrect invalidation of origin-neutral measures? In general, the answer tends to be negative. Indeed, the requirement of close to identical physical characteristics could be expected to significantly reduce the origin-unrelated bases for product distinctions. In some cases, however, even the most minor difference between products may constitute the very basis upon which a legitimate policy is formulated. Examples include tax distinctions based on alcohol content (i.e. for the purpose of discouraging alcohol consumption) or on the fuel efficiency of certain car components (i.e. for the purpose of protecting the environment). Thus, no matter how tightly the Appellate Body squeezes the scope of ‘like product’, it may not prevent origin-neutral measures from falling into the ambit of the 1st sentence. As long as the 1st sentence has the potential to capture origin-neutral measures, the assessment of regulatory purposes of the measures would be essential. Not doing so would not only fail to respect the intention of the draftsmen, but would likely result in the over-inclusiveness of this rule; that is, to wrongly condemn origin-neutral measures serving bona fide policy objectives unrelated to protectionism.
If the ‘narrow likeness approach’ cannot by itself completely avoid wrong cases (i.e. the potential over-inclusiveness of the NT rule), would the use of this approach together with the ‘case-by-case approach’ be able to do so? To answer this question, we need to discuss how the ‘case-by-case approach’ can contribute to eliminating wrong cases. In general, the WTO panels have left open the question of what physical differences would disqualify products from ‘likeness’.Footnote 126 The case-based approach can thus be employed to ensure that tax distinctions made on the basis of physical characteristics in connection with certain policy pursuits are not captured by Article III:2, 1st sentence. Since the ‘narrow likeness approach’ would allow remaining physical differences between ‘like’ products, the ‘case-by-case approach’ tends to be more essential to the Appellate Body's attempt to avoid wrong cases. In essence, this approach would allow a particular physical difference to be treated as not significant enough to disqualify products from ‘likeness’ in one case, but to be treated as significant enough to render products ‘unlike’ in another case. Accordingly, this approach grants a certain degree of flexibility in product likeness decisions so that WTO judges can exercise their judicial discretion to decide which physical differences matter depending on the facts of each case. In cases where a physical difference can be reasonably associated with a declared policy purpose, this difference would suffice to render products ‘unlike’, despite their commonality in all other aspects of physical characteristics. In contrast, if a remaining physical difference is not policy-related, then it can reasonably be ignored, leaving the nationality of products the sole conceivable basis for differential tax treatments. In this connection, it must be noted that the assessment of whether a physical difference is policy-related or not does not necessarily involve any substantive investigation into the genuine purpose of a measure. What is required is merely a simple inquiry into whether the physical difference has any bearing on alleged policy objectives. As long as the difference is apparently policy-related, the products in concern would be treated as ‘unlike’ (i.e. in the name of the physical difference), triggering the application of the 2nd sentence of Article III:2. Thus, this case-based approach to avoiding wrong cases, if applied, tends to be more moderate than the traditional ‘Aim and Effect’ test in the sense that the former does not involve formal assessments of the substance of regulatory purposes. The aforesaid analysis may also be applied to the end-uses of products. Thus, the ‘case-by-case approach’ may be used to determine whether certain remaining differences in end-uses are significant enough to disqualify products from ‘likeness’ depending upon whether or not the end-uses are policy-related.
4.4.2 The weaknesses of the Appellate Body's ‘case-by-case approach’
Despite this more moderate feature of the tribunals' ‘case-by-case approach’ compared with the application of the traditional ‘Aim and Effect’ test, this approach suffers from at least three weaknesses.
First, although this approach to finding product likeness does not rely on a comprehensive examination of the genuine purpose of a measure, it nonetheless implicitly allows regulatory purpose to play a decisive role in finding product likeness. It does so by associating the significance of products' physical differences or/and end-uses with declared policy objectives. Thus, this approach is essentially a policy-based approach rather than a market-based approach to determining ‘like product’, which thereby contradicts the WTO jurisprudence as well as the preparatory work of the NT rule.
Second, this approach entails a backdoor assessment of whether a physical difference or a difference in end-use is policy-related, which harms the predictability and transparency of the application and interpretation of the NT rule.Footnote 127 Uncertainty about which physical characteristics or end-uses matter in the finding of ‘like product’ further impairs predictability and transparency.
Third, and more fundamentally, this approach would be unlikely to eliminate all wrong cases because not all legitimate product distinctions are made on the basis of physical characteristics or end-uses. Examples include product distinctions made by reference to price (i.e. for the purpose of regulating the consumption of luxury goods and raising fiscal revenues) or to manufacturing process (i.e. for the purpose of protecting the environment). These examples illustrate the possibility that criteria unrelated to products' physical characteristics or end-uses may become the very basis for legitimate product distinctions. In such cases, the tribunals could not achieve what they attempted to do by playing the accordion of ‘like product’ in the context of Article III:2, 1st sentence.
4.2.3 Improving the WTO tribunals' approach to assessing regulatory purposes under Article III:2, 1st sentence
Considering these deficiencies in the WTO tribunals' approaches, the author submits two alternative ways for improvement.
(a) Clear division of labor – origin-specific measures only
The first approach is to limit the function of Article III:2, 1st sentence to dealing with origin-specific measures only. By doing so, all origin-neutral measures would fall under the scrutiny of Article III:2, 2nd sentence where purpose inquiries can be formally undertaken under the test of ‘SATAP’.
While this approach is in line with the interpretative guidance derived from the preparatory work of the NT rule, one obstacle to its acceptance lies in its disharmony with modern practice. In no disputes have the WTO tribunals indicated such a narrow reading of Article III:2, 1st sentence. On the contrary, the common understanding has been that the 1st sentence can be applied to origin-neutral measures. Decisions on the NT-illegality of the discriminatory tax against vodka in favor of shochu in Japan–Alcoholic Beverages II and the discriminatory tax against beet sugar in favor of cane sugar in Mexico–Soft Drinks Footnote 128 signify the applicability of the 1st sentence to origin-neutral measures. Thus, a more feasible approach would be to allow that sentence to be invoked to tackle origin-neutral measures and, through the operation of Article III:1, incorporate formal purpose inquiries into that sentence.
(b) Purpose inquiries under ‘in excess of’
An alternative approach is to examine regulatory purpose under the test of ‘in excess of’ under Article III:2, 1st sentence. As a matter of fact, since it would be unlikely to apply the traditional ‘Aim and Effect’ test under ‘like product’, the test of ‘in excess of’ provides the last option for the assessment of regulatory purposes under the 1st sentence.
Compared with the first approach, this approach has a better chance of being accepted by the WTO adjudicators. It can be argued that although ‘SATAP’ does not constitute a separate legal element due to the textual constraint of the 1st sentence, the general principle set out in Article III:1 must inform the interpretation of the specific obligations contemplated in that sentence. While the interpretation of product likeness is merely to activate the application of the NT rule, the substantive issues mandated by the term ‘SATAP’ must be examined under the test of ‘in excess of’. This interpretation is in line with the Appellate Body's interpretation of ‘less favorable treatment’ under Article III:4 in EC–Asbestos. That is, as ‘SATAP’ informs the interpretation of ‘less favorable treatment’,Footnote 129 it must also be read into ‘in excess of’. Furthermore, in deciding “in excess of”, one WTO panel has taken into account the discriminatory effect of the contested measures. In Mexico–Soft Drinks, the panel, in deciding whether the Mexican measure applied taxes to imported sweeteners (i.e. beet sugar) in excess of those applied to ‘like’ domestic sweeteners (i.e. cane sugar), not only relied on the significant tax differentials, but also considered the fact that while beet sugar tended to be found exclusively in imported products, cane sugar was essentially produced domestically in Mexico.Footnote 130 If the assessment of the disparate impact of the measure was undertaken to decide whether the measure was applied ‘SATAP’, then it seems that the panel has implicitly incorporated the element of ‘SATAP’ into ‘in excess of’. In integrating the test of ‘SATAP’ into ‘in excess of’, it can be argued that the panel also opened the door to purpose inquiries – the central mandate of ‘SATAP’ – in deciding ‘in excess of’. In this case, in assessing the discriminatory effect of the Mexican measure, the panel may arguably have implicitly considered that this measure, by subjecting all imported beet sugar to considerably higher taxes than those applied to domestic cane sugar, had the purpose of protecting the domestic production of cane sugar. Finally, this approach also finds support from the preparatory work of the NT rule which, as discussed above, indicates that it would be necessary to consider regulatory purposes insofar as Article III:2, 1st sentence is invoked to tackle de facto discrimination through the operation of Article III:1.
5. Conclusion
It has been widely discussed as to how to resolve the tension between multilateral trade liberalization and domestic pursuit of policy objectives unrelated to trade. As far as the NT rule is concerned, one solution to this tension has centred on the issue of whether regulatory purposes of origin-neutral measures should be considered in determining the NT-legality of the measures. Another issue closely related to the previous one is under which legal element, if purpose inquiries are necessary, they should be undertaken. In the context of Articles III:1, 2, and 4, while past debates have failed to resolve the two issues, the analysis of the preparatory work of the NT rule in this article has provided significant interpretative guidance. The WTO adjudicators have been successfully following the guidance by approving purpose inquiries under Article III:2, 2nd sentence through the test of ‘SATAP’, and under Article III:4 through the test of ‘less favorable treatment’. However, due to their rejection of the relevance of regulatory purposes to the finding of violations of Article III:2, 1st sentence, the WTO tribunals have failed to follow the instructions offered by the original authors of the NT rule. In the absence of formal purpose assessment, the tribunals have developed the ‘narrow likeness approach’ and the ‘case-by-case approach’ in order to avoid wrong cases (i.e. over-inclusiveness of the NT rule). Having considered the deficiencies of the approaches, this article has proposed two alternative approaches under which the tribunals should either limit the application of the 1st sentence to origin-specific measures, or undertake formal purpose inquiries under the element of ‘in excess of’ of that sentence in dealing with origin-neutral measures. By doing so, they cannot only achieve a higher degree of predictability and credibility, but also consistency between the jurisprudence and the preparatory work of the NT rule.
