Revisiting Precedential Effect of WTO Decisions Introductory Remarks

[draft 22.05.2007]

Joanna Gomula

Revisiting Precedential Effect of WTO Decisions – Introductory Remarks

It would seem incorrect to discuss the precedential effect of WTO panel reports in the context of the recent anti-dumping disputes without referring to the standard set forth in the early days of the WTO, when, as it might have been hoped, this issue was definitively resolved. In the Japan-Alcoholic Beverages II ruling (adopted on 1 November 1996) the Appellate Body reversed the panel’s finding that adopted GATT panel reports constitute subsequent practice within the meaning of Article 31.3(b) of the Vienna Convention on the Law of Treaties. Instead, the Appellate Body concluded that adopted panel reports are an important part of the GATT acquis, create “legitimate expectations” among WTO Members and therefore should be taken into account where they are relevant to a dispute. The Appellate Body explained that adopted reports are not binding, except with respect to resolving the particular dispute between the parties. It compared their status to that of decisions of the International Court of Justice, noting that “[t]his has not inhibited the development by that Court (and its predecessor) of a body of case law in which considerable reliance on the value of previous decisions is readily discernible” (footnote 30). The Appellate Body also addressed the status of unadopted panel reports, agreeing with the panel that they have no legal status, although they could be a source of useful guidance for future panels.
Although the Appellate Body was referring to GATT reports only, there can be no doubt that its 1996 conclusions extended to WTO reports as well. The Appellate Body emphasized that the GATT reports’ “character and their legal status have not changed since the coming into force of the WTO Agreement”. Moreover, as was confirmed by the Appellate Body in later jurisprudence, the above equally applies to the reports of the Appellate Body itself. In US-Shrimp (Article 21.5) the Appellate Body endorsed the panel’s decision to take into account one of the Appellate Body’s earlier reports: “…the Panel did not err. The Panel was correct in using our findings as a tool for its own reasoning” (para. 111).
In the much later US-OCTG Sunset Reviews report (adopted on 17 December 2004) the Appellate Body defended the panel’s decision to treat the Sunset Policy Bulletin as a “measure”. The Appellate Body emphasized that at stake was the same instrument as in the earlier dispute in US-Corrosion-Resistant Steel Sunset Review and that “it was appropriate for the Panel, in determining whether the SPB is a measure, to rely on the Appellate Body’s conclusion in that case. Indeed, following the Appellate Body’s conclusions in earlier disputes is not only appropriate, but is what would be expected from panels, especially where the issues are the same” (para. 188).
As a result of the above, the legal basis for the resolution of a matter by WTO dispute settlement bodies is rather unusual: the bases for the “recommendations and rulings of the DSB” are to be found in the two adopted reports: one of the Appellate Body and the other of the panel, as modified, reversed or upheld by the Appellate Body. It is to these two-tier decisions that Members’ “legitimate expectations” would formally extend to, although in practice Appellate Body reports are the ultimate source of authority.
Not long after Japan-Alcoholic Beverages II there was a notable attempt to undermine the implications of this important report. In US-Shrimp the panel did indeed acknowledge the Appellate Body’s findings in US-Gasoline, but proceeded to ignore its analysis completely, despite the fact that at issue in both cases was the interpretation of the same provisions of Article XX. The panel, seeing its task as one of clarification of the scope of measures permissible under Article XX, noted the sequence of interpretation endorsed by the Appellate Body in US-Gasoline, only to reject it immediately thereafter. The panel was of the opinion that “as the conditions contained in the introductory provision apply to any of the paragraphs of Article XX, it seems equally appropriate to analyze first the introductory provision of Article XX” (para. 7.27, emphasis added).
In other words, the panel placed itself at the same footing as the Appellate Body, in choosing its preferred method of interpretation of a measure sought to be justified under Article XX, as if the systemic implications of the WTO dispute settlement system were not relevant at all. Or was it just trying to correct an interpretation that, in its opinion, was wrong? Further in its report the panel acted even more boldly. Although mindful of Japan-Alcoholic Beveregas II, it paid great deference to one of the unadopted reports in the US-Tuna/Dolphin dispute, “finding support” in it, and even quoting it at length (par. 7.46). As is clear from footnote 652, the panel did understand the distinction between the legal status of adopted and unadopted reports, but thought the Tuna-Dolphin report was “relevant in the present case and provided useful guidance”. By implication, the unadopted GATT’47 report was, in the panel’s view, more relevant and more useful than the more recently adopted Appellate Body adopted.
However, were it not for the above explicit and rather inelegant “straying away” from the Appellate Body’s analysis, the panel might have been exonerated for its approach. When one reviews the arguments of the parties, one is left with the impression that even the parties themselves did not believe that US-Reformulated Gasoline would have much bearing on the resolution of the dispute. The arguments focused on the extraterritorial nature of the measure, in the context of general principles of international law relating to sovereignty, and on whether such type of measure is permissible under the WTO system. That is, the discussion itself revolved around the conclusions of the unadopted Tuna/Dolphin reports.
But, not surprisingly, when the above conclusion was appealed, the Appellate Body had little understanding for the panel and was merciless in pointing out the flaws of the panel’s analysis and findings. It noted with obvious satisfaction that these flaws “flow almost naturally from the fact that the Panel disregarded the sequence of steps essential for carrying out such an analysis” (para. 117). The Appellate Body was adamant to emphasize that the sequence of interpretation of Article XX does make a difference (par. 119). It pointed out that “the consequences of the interpretative approach adopted by the Panel are apparent in its findings” (par. 121), which, according to the Appellate Body, amounted to the creation of another “test” for measures taken under Article XX. Naturally, the panel’s findings were reversed and the Appellate Body completed the legal analysis with respect to the interpretation of the measure in light of Article XX.
In general, however, it seems that WTO panels have at least made some attempts to follow the guidelines set by the Appellate Body in earlier decisions, and to pay appropriate respect to GATT/WTO acquis (in particular, in the versions endorsed by the Appellate Body). This may be out of natural deference for a superior judicial organ or out of common sense, because in case of an appeal, it is the Appellate Body that has the last say and the power to modify or reverse a panel’s findings.
It has happened, however, that the usefulness of reliance on previous reports has been questioned by the parties. This was the case in US-Softwood Lumber V where the United States requested the Appellate Body not to “import wholesale the findings and reasoning” from EC-Bed Linen. Despite this, the Appellate Body decided to take into account the reasoning and findings contained in that report, “as appropriate” (para. 112). As the Appellate Body explained, neither of the parties had argued that “the reasoning and findings of the Appellate Body should not be taken into account in this case” (footnote 175).
The above should be borne in mind when discussing the precedential effect in WTO decisions in light of the Zeroing cases. The obvious question that comes to mind at this stage is: has there been any special development that would justify a re-assessment of the principle dating back to Japan-Alcoholic Beverages II? In my opinion, there is nothing to indicate that there has been a change in the legal status of WTO decisions. Instead, what the Zeroing cases have revealed is a number of other problems related to the WTO dispute settlement mechanism. I would like to emphasize two of these problems at the outset.
First, there is the problem of the complexity of trade remedy cases, and anti-dumping cases in particular. The Appellate Body has not always managed to ensure total uniformity and consistency of its jurisprudence. This is not necessarily a criticism. Like any other judicial tribunal, the Appellate Body cannot avoid controversial or even erroneous interpretations. Like any other judicial tribunal, with great responsibility for the resolution of often highly politically charged disputes, it has to be particularly careful. This has resulted in some, perhaps excessive, caution, which in turns leads to confusion and inconsistencies.
Second, there is the problem of implementation of WTO decisions. The system is not designed in a way that would force Members to bring their measures into prompt compliance with WTO law if they lose a dispute. This in itself provides even less motivation for the adjustment of measures that have not been complained of, with respect to Members who are not parties to a dispute. Partly, the reason is the precedential effect of WTO decisions as such: a WTO report is binding only with respect to a particular dispute and as between the parties to that dispute. However, this is not unusual and the problem exists also in other jurisdictions. For example, one may recall the experience of the International Court of Justice in three cases: Breard (1998, discontinued), the LaGrand (1999) and Avena (2003), where the respondent continued to committed similar breaches of Article 36 of the Vienna Convention on Consular Relations, albeit with respect to nationals of different states.
So, what is there in the Zeroing disputes that requires the revisit of the precedential effect of WTO decisions? Let us beginby US-Zeroing (Ecuador), which stands out because of certain unusual circumstances surrounding the dispute.Ecuador complained of the use by US authorities of “zeroing” in the calculation of dumping margins for three anti-dumping measures on frozen shrimp. Ecuador alleged a violation of Article 2.4.2 of the Anti-Dumping Agreement. This provision requires, as a rule, the establishment of the existence of dumping margins by weighted average comparison, or transaction-to-transaction comparison, of normal value and export price, allowing for certain exceptions. Zeroing, in turn, means in essence the calculating as “zero” the negative margins resulting from such a comparison, which in turn leads to a higher overall dumping margin than if the negative margins are reflected in the calculation.
It is important to note that neither Article 2.4.2, nor any other provision of the Anti-Dumping Agreement, refers to “zeroing”. This problem was raised and discussed during the Uruguay Round (in the context of the later Article 2.4.2), but not definitely resolved (see e.g. The GATT Uruguay Round: A Negotiating History (1986-1992). Antidumping, ed. T.P. Stewart, Kluwer Law and Taxation Publishers 1993, pp. 156-161). To complicate matters further, there are different types of zeroing, and this practice has been differently assessed depending on whether it relates to original investigations or other types of proceedings, such as administrative review proceedings.
Ecuador’s complaint concerned the use of zeroing in calculating dumping margins in an original investigation, using the weighted to weighted average comparison. The measure was almost identical to that reviewed in US-Softwood Lumber V. In an unusual procedural development, the parties reached an agreement (reproduced in Attachment 2 to the panel report), pursuant to which the United States undertook not to contest Ecuador’s claim that the measures were inconsistent with Article 2.4.2, on the grounds stated in US-Softwood Lumber V. The United States also undertook to recalculate dumping margins and issue a new determination in order to render the anti-dumping measures not inconsistent with the findings of the panel. The panel decided to accept the parties’ proposal and adopted an expedited timetable, as requested.
This dispute was initiated in mid-2006, at a time of great turbulence in WTO jurisprudence in this area. This year saw three other zeroing disputes, often with overlapping proceedings concerning various aspects of zeroing, with inconsistent panel and Appellate Body reports issued.
The “beginnings” of WTO jurisprudence on zeroing did not predict such a dramatic outcome. In EC-Bed Linen (adopted on 12 March 2001) the panel concluded, and the Appellate Body upheld this conclusion, that the EC had acted inconsistently with Article 2.4.2 by zeroing negative dumping margins for individual product types. The second “zeroing” dispute, US-Softwood Lumber V (adopted on 31 August 2004), caused a minor stir when one panel dissented from the conclusion that zeroing in weighted average-to-weighted average comparison violated Articles 2.4.2 and 2.4. The Appellate Body agreed with the majority, adopting the view that dumping margins must be calculated for the product under investigation “as a whole”, and not for product types, models, or categories. According to the Appellate Body, because zeroing may result in export prices being treated as less than what they actually are, which leads to an inflation of the margin of dumping for the product as a whole, treating of comparisons for which the weighted average normal value is less than the weighted average export price as “zero” is contrary to Article 2.4.2.
The problems seem to have begun with US-Zeroing (EC) (adopted on 9 May 2006). There, the panel concluded that “model” zeroing (as opposed to “simple” zeroing) was inconsistent with Article 2.4.2, as was the zeroing methodology in original investigations. The panel majority (one member dissented) found that the inconsistency of the zeroing method did not extend to administrative reviews. The Appellate Body reversed the latter finding and determined that the use of “simple” zeroing (comparing a weighted-average normal value with an individual export transaction) in administrative reviews is contrary to Article 9.3 of the Anti-Dumping Agreement and Article VI:2 of GATT 94. The Appellate Body did not take a view on whether Article 2.4.2 was applicable to administrative reviews under Article 9.3 (paras. 160-164). On the other hand, it accepted as challengeable “as such” in WTO proceedings the zeroing methodology in original investigations, in which the weighted-average-to-weighted-average comparison method is used to calculate margins of dumping (para. 205).
With the circulation of the Appellate Body report in US-Zeroing (EC) on 18 April 2006, the situation becamevery complex. On 8 March 2006 the panel in US-Zeroing (Japan)had issued its interim report. It invited the parties to submit their written comments on the relevant issues addressed by the Appellate Body in the former report. Japan argued that, in line with the Appellate Body’s reasoning, the panel should adopt a broader application of the concept of the “product as a whole”, which the panel refused to do. However, unlike the US-Shrimp/Turtle panel, which was abrupt in its approach to the question of precedential effect of Appellate Body reports, the panel in US-Zeroing (Japan) appears to have been facing a true dilemma. In particular, the panel recognized the “important systemic considerations in favour of following adopted panel and Appellate Body reports” (para. 7.99) and acknowledged that this supports the objective of providing security and predictability to the multilateral trading system (footnote 733). However, the panel balanced this against its obligation under Article 11 of the DSU to make an objective assessment of the matter before it and to clarify the existing provisions of covered agreements in accordance with Article 3.2 of the DSU (footnote 733). It seems that the panel tried to explain that it was because of its strong conviction that it could not accept the Appellate Body’s reasoning in the given case, and not because of an inherent right to ignore Appellate Body’s decisions, that it was unable tofollow the Appellate Body’s approach (paras. 7.100-7.102).
However, at regardless of its intentions, the end result was that the panel engaged in a process of undermining the systemic considerations that it had acknowledged were so important. In fact, the panel was acting as if it were itself an “appeal body”with respect to the Appellate Body. It concluded that there is no general requirement to determine dumping and margins of dumping for the product as a whole, which would entail a general prohibition of zeroing. Therefore, in its opinion, by maintaining simple zeroing procedures the United States had not acted inconsistently with Articles 2.1 and 2.4.2, and Articles VI:1 and VI:2 of GATT 94.
In the meantime, on 3 April 2006, the panel report was circulated in US-Softwood Lumber V (Article 21.5) (adopted on 1 September 2006). Like its US-Zeroing (Japan) counterpart, the panel here had refused to rely on the Appellate Body’s findings in US-Softwood Lumber V. In its opinion, these findings applied to the weighted-to-weighted average comparison methodology only. The panel pointed out that the Appellate Body had “explicitly declined to apply its findings regarding zeroing under the W-W comparison methodology to zeroing under the T-T comparison methodology” (para. 5.20). However, the panel did not allege that it was applying a better interpretation; rather, it tried to distinguish the matter before it from that in US-Softwood Lumber V.On appeal, the Appellate Body reversed and found that zeroing in transaction-to-transaction comparisons in original investigations was inconsistent with Articles 2.4.2 and 2.4.
It was against this background that the panel in US-Zeroing (Ecuador) was seized with its dispute. Although slightly disconcerted by “the lack of substantive disagreement between the parties” and worried that it may be faced with a mutually agreed solution, the panel accepted that it could nevertheless discharge its responsibilities under Article 11 of the DSU. The panel’s caution might have been excessive: the circumstances did not seem to indicate a mutually agreed solution; if at all, then to have the dispute resolved by the panel. In essence, the United States simply decided not to defend its measure, endorsed the facts and acknowledged the similarity of the measure with that reviewed in US-Softwood Lumber V, provided that the measure would be assessed in line with the findings of the latter report. If anything, the dispute might have qualified for expeditious arbitration under Article 25 of the DSU, as the issues were “clearly defined by the parties”.
Because the parties had requested that the panel focus on the examination of the measure in light of US-Softwood Lumber V, the panel felt obliged to present its view on the binding force of this report (par. 7.37):

While we are not, strictly speaking, bound by the Appellate Body’s reasoning in US-Softwood Lumber V, we are reminded that adopted Appellate Body Reports create legitimate expectations among WTO Members and that ‘following the Appellate Body’s conclusions in earlier disputes is not only appropriate, but is what would be expected from panels, especially where the issues are the same’.