hidden costs of international dispute settlement: WTO review of domestic anti-dumping decisions, The

Tarullo, Daniel K

I. INTRODUCTION

The increasingly international scope of some economic activities is correlated with closer integration of national and international legal arrangements. The causal relationship between economic and legal change varies: Some actors seek legal change in order to facilitate international economic activity, even as other actors seek legal change to respond to that activity. Modes of integrating national and international legal systems also vary: Governments may formally harmonize laws, regulations, or standards. Certain government officials may informally coordinate activities within their areas of competence with their counterparts from other countries. And, of course, governments may enter legally binding arrangements that establish rules against which national action is measured. Trade agreements have traditionally assumed this last form.

Since the conclusion of the Uruguay Round of multilateral trade negotiations and the transformation of the makeshift General Agreement on Tariffs and Trade (GATT) into the World Trade Organization (WTO), international scrutiny of national compliance with international trade obligations has become markedly more legalized. WTO dispute settlement bodies (known as "panels") hear complaints from member states that actions of other member states violate WTO obligations, which are found in the original GATT text and the various specialized agreements of the WTO. The panels conduct quasi-judicial proceedings that produce results binding upon the disputants.1 An Appellate Body (AB) of seven members sitting in three-member panels hears appeals from panel findings. The AB is intended both to provide a check on wayward panels and to promote consistency in WTO dispute settlement practice. Among the many features of this more legalistic framework, panels conduct direct reviews of some national administrative actions.

This Article examines the short and, to this point, unhappy experience with the introduction into WTO jurisprudence of a standard of review for WTO panels to follow in considering one type of administrative action, the imposition of anti-dumping duties. The experiment with a standard of review arises in the context of WTO review of national anti-dumping measures, a long-time area of contention. "Dumping" was originally defined as selling products more cheaply abroad than at home. While not specifically proscribed by international agreements, dumping has been internationally identified as deserving of condemnation if it causes injury to an industry in the importing country.2 U.S. law, since emulated by other countries, added to the definition sales below fully allocated cost of production, even where the price charged for the merchandise was the same as that in the importing country. Anti-dumping law generally provides for imposition of an additional import duty to equalize the price of the imported goods with the "normal value," as calculated from foreign sales or from the cost of production. Most economists find the entire premise of anti-dumping law misguided-at least where there is no predatory intent or effect-because it discourages some forms of price competition in some circumstances. Certain domestic interests-particularly those in industries with high fixed costs-are equally insistent that anti-dumping laws are necessary to protect them from foreign producers suppressing prices by flooding domestic markets. The laws and regulations enacted by countries, which can be very complex, reinforce the suspicion of liberal traders that the laws are rigid, biased implementations of a misguided premise.

Not surprisingly, disputes over imposition of dumping duties have been frequent. Exporting countries have often complained that, quite apart from the principle that dumping is bad, importing countries misuse their anti-dumping laws. The first "code" negotiated in the GATT to supplement the rules of the original GATT agreement was one that limited the use of anti-dumping measures.3 New, more detailed agreements to limit national anti-dumping measures were included in both the Tokyo Round and Uruguay Round of trade negotiations. Meanwhile, use of anti-dumping measures had spread from the United States, European Union, and other industrialized countries to developing countries as well.4 Thus, while international disagreements over dumping continue to pit some industrialized countries against Japan and many developing countries, the lines are not as clearly drawn as they were twenty years ago.

With the prospect of a new WTO dispute settlement approach, which was to be binding and highly legalized, the United States sought some insulation of its administration of the dumping laws from what supporters of those laws feared would be aggressive review by WTO panels. Thus, the United States negotiated for a special standard of review,5 based upon U.S. administrative law concepts and described below. Of course, the review is used to determine whether the U.S. agencies have complied with the WTO Anti-Dumping Agreement ("the Agreement"), rather than with domestic statutes. Furthermore, neither a foreign country prevailing in the WTO nor its exporter has a right to seek enforcement of the WTO finding directly in the U.S. legal system. However, the legislation implementing the Uruguay Round empowers the United States Trade Representative to request and, if necessary, direct the relevant U.S. agencies to modify previous decisions in order to make them consistent with WTO panel decisions, so long as the panel decision does not require action inconsistent with U.S. statutes.6 Thus, a blend of international and national institutional features renders this arrangement a recognizable system of "judicial" review of administrative action. In fact, some lawyers representing foreign producers in U.S. dumping cases believe that the WTO review is generally a more effective option than is review in the U.S. Court of International Trade.

In this Article, I attempt answers to three questions: (1) How has the AB of the WTO handled reviews of administrative action under the Article 17.6(ii) standard? (2) Why has the AB taken this position? (3) Is the AB practice desirable? Part II presents a brief overview of WTO anti-dumping dispute settlement and examines in considerable detail the application of the special standard of review. It concludes that the special standard of review has had virtually no impact on the review of national anti-dumping measures by the WTO. Part III offers possible explanations for this outcome. Although it is not possible to answer the second question convincingly, Part III uses these potential explanations as a starting point for the normative analysis. It reviews the serious normative arguments on both sides of the third question and suggests that one's starting premises will probably determine one's conclusion. Part IV adds a geopolitical perspective to the analysis. It shows how the dynamic effects of the AB's decisions on the negotiating positions of the United States and other countries will, in light of the institutional features of the international trading system, produce "deadweight" losses to all WTO members. Thus, regardless of one's preferences among the normative principles discussed in Part III, the AB's disregard of negotiated provisions may be undesirable. The analysis also reveals the dangers in transposing features of a national legal system to the very different institutional context of the international legal system and, thus, suggests the potential for WTO governance problems extending well beyond review of anti-dumping laws.

II. OUTCOMES IN WTO ANTI-DUMPING DISPUTE SETTLEMENT CASES

A. Overview

Review of national anti-dumping actions in the WTO got off to a slow start, in part because the new Anti-Dumping Agreement applied only to investigations or reviews initiated after the January 1, 1995 effective date of the WTO agreements.7 Also, there was an unusually low number of anti-dumping investigations initiated worldwide in 1995.8 Until mid-1999, only three challenges to national anti-dumping actions had been initiated in the WTO.9 Four additional challenges were lodged in just the four-month span of July to October 1999.10 As of this writing, final WTO decisions have been issued in eleven cases,11 with three other cases pending before panels.12

To date, exporting countries challenging the imposition of anti-dumping duties by other countries have prevailed in every case. One must be cautious, however, in drawing conclusions from this statistic alone. After all, it is possible that only dumping determinations resting on the most egregious violations of the Agreement have been challenged, particularly in its early years. It is also possible that determinations by national anti-dumping authorities regularly violate the Agreement and are thus broadly vulnerable to challenge in the WTO. Also, because the exporter need prevail on only one issue in order to "win" a WTO case, the odds would seem to be with the exporters.13 Thus, notwithstanding the grumbling of some legislators, lawyers, and business people sympathetic to import-sensitive industries, the fact that the "exporter always wins" in anti-dumping cases does not alone tell us much. To evaluate the performance of the panels and the AB fairly, we must assess their reports more closely.

The eleven completed cases have involved scores of issues, which can be grouped into four categories. The first, and smallest, involves the dispute settlement procedures of the Anti-Dumping Agreement and the Dispute Settlement Understanding (DSU) themselves. The second category involves procedures followed by national authorities that were challenged as inconsistent with requirements imposed by the Agreement. The third involves challenges to the factual determinations of national authorities that dumping or injury due to dumping have in fact occurred. The fourth category, the subject of this Article, involves the legal interpretation of the provisions of the Agreement.

Needless to say, there is often no clear distinction between a review of whether a national authority has established the facts necessary for a dumping determination and a legal interpretation of the provisions of the Agreement that govern those factual determinations. Still, there are issues which dispute settlement panels and the AB either characterize explicitly as matters of legal interpretation or analyze in a manner indicating their belief that legal interpretation is required. A focus upon this set of issues is useful both for understanding how the WTO dispute settlement panels have in fact reviewed national administrative actions and for considering how they should review those actions. As previously mentioned, the Anti-Dumping Agreement contains special standards of review for WTO panels to apply in examining the imposition of anti-dumping measures by national authorities. Moreover, it is obviously easier to evaluate an overtly interpretive analysis than a factual review, because the latter involves a voluminous record that may not be completely available to non-participants in dispute settlement processes.

As for matters of legal interpretation, the standard of review is set forth in Article 17.6(ii):

[T]he panel shall interpret the relevant provisions of the Agreement in accordance with customary rules of interpretation of public international law. Where the panel finds that a relevant provision of the Agreement admits of more than one permissible interpretation, the panel shall find the authorities' measure to be in conformity with the Agreement if it rests upon one of those permissible interpretations.14

This standard, like its companion for matters of factual determination,15 is considerably more specific and, it would seem, more favorable to Member state "respondents" than the provisions applicable to most WTO disputes. Article 11 of the DSU instructs a panel to "make an objective assessment of the matter before it, including an objective assessment of the facts of the case and the applicability of and conformity with the relevant covered agreements."16 Another generally applicable standard cautions that rulings of the WTO Dispute Settlement Body, which "adopts" final reports of panels and the AB, "cannot add to or diminish the rights and obligations provided in the covered agreements."17 There is no generally applicable provision that sets forth any recognizable standard of review (e.g., de novo review), much less one as solicitous of national authorities as Article 17.6.18

The Article 17.6(ii) standard will doubtless remind many readers of the standard for judicial review of agency legal interpretations articulated by the United States Supreme Court in Chevron, U.S.A. v. National Resources Defense Council:

When a court reviews an agency's construction of the statute which it administers, it is confronted with two questions. First, always, is the question whether Congress has directly spoken to the precise question at issue. If the intent of Congress is clear, that is the end of the matter. . . . [I]f the statute is silent or ambiguous with respect to the specific issue, the question for the court is whether the agency's answer is based on a permissible construction of the statute.19

The similarities are not coincidental. U.S. trade negotiators sought constraints on the ability of the new, binding dispute settlement system of the WTO to overturn anti-dumping determinations by the Commerce Department and the International Trade Commission. While inclusion of a "permissive" standard of review in the Anti-Dumping Agreement was far from uncontroversial,20 it emerged in the form quoted above, somewhat different from the original U.S. proposal.21 Yet it was still a notable departure from the generally applicable rules for WTO dispute settlement.

In the "Statement of Administrative Action" accompanying President Clinton's transmittal to the Congress of the Uruguay Round Trade Agreements and implementing legislation, the Administration characterized the standard of review as "analogous to the deferential standard applied by U.S. courts in reviewing actions by Commerce and the Commission."22 Article 17.6(ii) was touted as "ensur[ing] that panels will not be able to rewrite, under the guise of legal interpretation, the provisions of the Agreement, many of which were deliberately drafted to accommodate a variety of methodologies."23 Of course, statements by the U.S. Government notwithstanding, the language of the Chevron standard and the language of Article 17.6(ii) are similar, rather than identical. Moreover, there is no authoritative statement of the intention underlying 17.6(ii), like or unlike that offered by the United States.

B. Application of the Article 17.6(ii) Standard

A universe of a dozen cases hardly provides the basis for definitive judgments on matters such as application of a specified standard of review. Still, an examination of those cases suggests that Article 17.6(ii) has had little effect upon determinations by WTO dispute settlement bodies. The standard itself is sometimes barely mentioned and is only infrequently applied in a serious way to specific interpretive problems. It is difficult to identify any issue in any of the cases in which this special standard has produced an outcome different from that which would have prevailed had there been no Article 17.6. In a number of instances, a panel or the AB has found a single, unambiguous meaning for provisions of the Agreement that seem readily susceptible to multiple readings. As actually applied in WTO anti-dumping cases, Article 17.6(ii) resembles not at all the Chevron-type standard contemplated by the United States at the time the Anti-Dumping Agreement was drafted.

The tendencies just noted are best illustrated through a careful look at the 2001 opinion of the AB in the Japan Steel case.24 This opinion, largely affirming the determinations of the dispute settlement panel, contains the AB's fullest explication to date of Article 17.6. It also involves half a dozen legal interpretations to which Article 17.6(ii) was obviously or arguably applicable. Despite discussing the Article 17.6 standard at some length, the AB did not adopt a uniform approach to these interpretive issues. However, the existence of 17.6(ii) has had little noticeable impact on any of the variations on the AB's interpretive method. In both these respects, Japan Steel parallels legal interpretation of the Anti-Dumping Agreement in other cases decided to date.

In its introductory discussion of the standard of review, the AB's principal concern is the relationship between Article 17.6 and Article 11 of the DSU. Specifically, the AB asks whether Article 17.6 is properly read as "complementing" Article 11 of the DSU, or whether it is in conflict with that more generally applicable provision.25 Both the tenor and the conclusion of this discussion minimize the differences between the standard of review applicable in WTO cases generally and that applicable in anti-dumping cases.

Although this Article does not consider at length the Article 17.6(i) standard of review for factual questions, it is worth noting that the AB apparently found no significant difference between this standard and that applicable to other WTO cases under Article 11 of the DSU. The AB noted that the latter calls for the panel to make an "objective assessment of the matter," while the former instructs a reviewing panel to determine if the factual determinations of national anti-dumping authorities were "unbiased and objective."26 "If these broad standards have not been met, a panel must hold the investigating authorities' establishment of the facts to be inconsistent with the Anti-Dumping Agreement."27 Though the AB never quite said that the two standards are functionally equivalent, it minimized the wording differences.28 Most significantly, the AB did not even mention the last part of Article 17.6(i), which forbids panels from overturning determinations if the establishment of facts was objective and unbiased, "even though the panel might have reached a different conclusion."29

In discussing Article 17.6(ii), the AB did assess the language that particularly distinguishes this standard from Article 11, noting that the second sentence "presupposes that application of the rules of treaty interpretation in Articles 31 and 32 of the Vienna Convention could give rise to, at least, two interpretations of some provision of the Anti-Dumping Agreement, which, under that Convention, would both be 'permissible interpretations'."30 Having acknowledged this instruction to reviewing panels that is unique in WTO agreements, the AB immediately downplayed its significance. Referring to the language of Article 11 of the DSU, the AB observed that "[n]othing in Article 17.6(ii) of the Anti-Dumping Agreement suggests that panels examining claims under the Agreement should not conduct an 'objective assessment' of the legal provisions of the Agreement, their applicability to the dispute, and the conformity of the measures at issue with the Agreement."31 Concluding its discussion, the AB stated that "Article 17.6(ii) simply adds" to the normal Article 11 approach the proviso that a panel shall uphold a national anti-determination measure "if it rests upon one permissible interpretation" of the Agreement.32

The AB's formulation is not, to invoke a different standard of review, clearly erroneous. Yet it appears to reduce the significance of Article 17.6(ii) to a mere ancillary analytical step. The AB's formulation certainly does not resemble the statements of the Chevron standard familiar to readers of U.S. administrative law cases, which appear to prescribe an overall approach to judicial review of agency legal interpretation. This apparently narrow reading of 17.6(ii) is matched by the AB's restrictive application of the standard in its analyses of specific questions of legal interpretation. Some of these analyses are hard to square with the terms of Article 17.6(ii) itself, as illustrated by the disposition of the legal issues in Japan Steel. The remainder of this section considers the different approaches taken by the AB in that case, along with examples of similar approaches in other WTO anti-dumping cases.

1. Overturn National Authorities, Disregard 17.6(ii)

In the first and last legal issues of the Japan Steel report, the AB found the interpretations of the importing authorities inconsistent with the Agreement without applying the Article 17.6(ii) standard, except in the most pro forma fashion.

Perhaps the best evidence for the proposition that the AB regards Article 17.6(ii) as relatively insignificant is found in its discussion of the very first substantive issue, which immediately follows the explication of 17.6 described above. There, the AB never refers to any standard of review in finding that the United States violated Article 6.8 of the Anti-Dumping Agreement by refusing to accept information from one of the Japanese respondent steel companies after the administrative deadline for submission of information had passed.33 The AB's reading of the applicable provisions does not seem compelled by the text of the Agreement. On the contrary, it seems rather strained. An exemplary issue in which to recognize the possibility of multiple "permissible" readings is instead treated as if the AB were free to impose whatever interpretation it prefers.

This issue was one of two involving the use by the U.S. Department of Commerce (DOC) of information other than that provided by the Japanese companies subject to the dumping investigation. Use of such information is usually detrimental to the respondent, particularly where the "facts available" have been offered by the domestic industry seeking the imposition of dumping duties. Article 6.8 of the Anti-Dumping Agreement permits the use of information from other sources where a respondent "refuses access to, or otherwise does not provide, necessary information within a reasonable period or significantly impedes the investigation."34 National authorities also incorporate information from other sources where information that was provided by the respondents cannot be "verified." Here, DOC rejected certain information provided by two respondents, because the period it had established at the outset for submission of information (87 days, in this case) had lapsed.35 Japan successfully argued before the dispute settlement panel that the United States could not reject information for the "sole" reason that it had been provided after the deadlines for submission of information.36

In upholding the panel, the AB looked not only to Article 6.8, but also to paragraph 3 of Annex II of the Agreement.37 Paragraph 3, which is referenced explicitly in Article 6.8, elaborates on the use of available information: "All information which is verifiable, which is appropriately submitted so that it can be used in the investigation without undue difficulties, which is supplied in a timely fashion . . . should be taken into account when determinations are made."38

The AB interpreted the italicized requirement of timeliness in light of the Article 6.8 standard that investigating authorities may use facts available where information is not submitted within a reasonable period of time.39 This, said the AB, "indicates that information which is submitted in a reasonable period of time should be used by the investigating authorities."40

This almost tautological beginning was followed by the AB's brief but critical interpretation of the terms "reasonable period" and "reasonable time." So brief was this analysis that it can easily be quoted in full:

The word "reasonable" implies a degree of flexibility that involves consideration of all of the circumstances of a particular case. What is "reasonable" in one set of circumstances may prove to be less than "reasonable" in different circumstances. This suggests that what constitutes a reasonable period or a reasonable time, under Article 6.8 and Annex II of the Anti-Dumping Agreement, should be defined on a case-by-case basis, in the light of the specific circumstances of each investigation.41

The AB then listed six factors to be considered by national authorities, on a case-by-case basis, in determining whether information had been submitted in a "reasonable period."42 Because DOC had not made such a determination, the AB found that "USDOC's action does not rest upon a permissible interpretation of Article 6.8 of the Anti-Dumping Agreement."43

This concluding sentence was presumably meant to foreclose any objection that the AB had ignored Article 17.6(ii). But it was about as pro forma an invocation of a standard of review as can be imagined. Insofar as the AB never considered other possible interpretations of Article 6.8, was it stating implicitly that its interpretation was the only permissible one? This seems a hard claim to sustain. In the first place, though the AB asserted that the "word 'reasonable' implies a degree of flexibility that involves consideration of all of the circumstances of a particular case," it did not rest that assertion on an examination of the "ordinary meaning" of the terms of the Agreement in their "context," as required by Article 31.1 of the Vienna Convention.44 As we will see, WTO dispute settlement bodies often do begin with a dictionary definition of a disputed term (though we will also see instances of a suspiciously selective use of dictionary definitions).

Looking at the definition of "reasonable" in the very dictionary consulted in other WTO cases, we find no meaning that includes "flexible" or a synonym for it. The entry for "reasonable" includes the following definitions: "having sound judgment; moderate, ready to listen to reason"; "in accordance with reason; not absurd"; "within the limits of reason; not greatly less or more than might be expected"; "inexpensive; not extortionate"; "tolerable, fair."45 While some common uses of the term do seem to imply an element of flexibility (e.g., "Be reasonable!"), not all do. Indeed, the dictionary definition "fair" seems wholly consistent with the U.S. practice in this case.

Article 31.2 of the Vienna Convention defines the "context" of a treaty term to include the text of the rest of the treaty.46 Paragraph 3 of Annex II is, as already shown, unhelpful in illuminating the meaning of "reasonable." The AB also cites Article 6.1.1 of the Anti-Dumping Agreement, which requires that national authorities give "due consideration" to a request by exporters for an extension of the period for submitting information to the authorities, and to grant such requests where "practicable" and "upon good cause shown."47 The AB cited this provision as "indicat[ing] that the time-limits imposed by investigating authorities for responses to questionnaires are not necessarily absolute and immutable"48 One might more readily have drawn the opposite conclusion. After all, Article 6.1.1 requires consideration of an extension where there has been a request for an extension, which presumably means a request before the period for response has lapsed. In Japan Steel, the exporter sought to add information only after the 87-day period for response set by DOC had expired.49 One might fairly read Article 6.1.1 as implying that acceptance of the information is not required under these circumstances.

As the preceding couple of paragraphs demonstrate, there is considerable potential for sophistry in arguing about "permissible" interpretations of specific words or phrases in any legal text. A more straightforward critique of the AB's reasoning is based on the implications of its reasoning. The AB concludes that an administrative agency may never "reasonably" adhere strictly to the letter of limits it may establish for the submission of information by interested parties, even if those limits are themselves generous. This position is inconsistent with much administrative and judicial practice. Bright-line time limits may work regrettable hardships in some cases, but they can also conserve substantial resources of administrative agencies and interested parties. By maximizing incentives for parties to submit information expeditiously, time limits that are strictly enforced may be the most efficient and fair to anti-dumping parties as a whole. The case-by-case consideration of six factors required by the AB is not costless.

Strict time limits also restrict efforts by interested parties to game the system. In this case, in fact, it is possible that the exporters were engaged in just such a tactic. In response to the DOC request for "weight conversion factors," the exporters stated that it was "impracticable or impossible" to calculate such a factor.50 After preliminary determinations by the DOC, which used non-exporter-provided information to calculate the factor for these two exporters, the two became aware of the fact that a third exporter had submitted its "best estimate" of a conversion factor. One of the two exporters also indicated that it had "discovered" information it had formerly said did not exist.51 Then, in an effort to reduce the preliminary dumping margins, the two exporters sought to submit their own estimates of the required weight conversion factor.

Perhaps the exporters at first withheld information in an effort to avoid having DOC use any conversion factor and changed their approach only when it became apparent that DOC intended to proceed. Perhaps not. The true story in this case matters less than the larger point that deadlines which are themselves reasonable can make administrative action more efficient by requiring all parties to marshal their information and arguments sooner rather than later. There seems nothing in either the terms of the Anti-Dumping Agreement or the logic of the situation that compels a holding that deadlines can never be firm. Yet that is exactly what the AB held to be the only permissible interpretation of Article 6.8. In so doing, the AB not only ignored Article 17.6(ii). It also arguably ran afoul of the affirmation in Article 3.2 of the Dispute Settlement Understanding that WTO dispute settlement rulings "cannot add to or diminish the rights and obligations provided in the covered agreement."52

The AB also ignored Article 17.6(ii) in disposing of another issue in Japan Steel. Japan charged that the U.S. International Trade Commission (ITC) had violated the causation requirements of Article 3.5 in finding injury to a U.S. industry without adequately examining factors other than imports and without ensuring that injury caused by those other factors was not attributed to the imports.53 The panel had rejected Japan's argument, relying in part on a pre-WTO panel determination which said that the importing country's authorities were not required to "isolate" the injury caused by other factors, but simply had to "conduct an examination sufficient to ensure that . . . it did not find that material injury was caused by imports . . . when material injury . . . was in fact caused by factors other than these imports."54

The AB reversed, rejecting the reasoning in the pre-WTO dumping case:

In order that investigating authorities, applying Article 3.5, are able to ensure that the injurious effects of the other known factors are not "attributed" to dumped imports, they must appropriately assess the injurious effects of those other factors. Logically, such an assessment must involve separating and distinguishing the injurious effects of the other factors from the injurious effects of the dumped imports. If the injurious effects of the dumped imports are not appropriately separated and distinguished from the injurious effects of the other factors, the authorities will be unable to conclude that the injury they ascribe to dumped imports is actually caused by those imports.55

The AB acknowledged that "it may not be easy, as a practical matter, to separate and distinguish the injurious effects of different causal factors" but asserted that "this is precisely what is envisaged by the non-attribution language."56

The AB was "fortified" in its interpretation of Article 3.5 by its interpretation of the non-attribution language of the Safeguards Agreement in two recent reports.57 The AB conceded that this language is "by no means identical" to the non-attribution language of the Anti-Dumping Agreement, but that the similarities were most relevant.58 Never did it mention the Article 17.6(ii) standard of review and, specifically, the possibility that this standard might require a different outcome in an anti-dumping case than in a safeguards case.59

Although the AB's interpretation of Article 3.5 is certainly plausible, it does not seem compelled by the language. In the first place, the pre-WTO panel interpretation occurred during the negotiation of the Uruguay Round. The language on non-attribution in the earlier Tokyo Round Anti-Dumping Agreement was left essentially unchanged in the new Agreement concluded during the Uruguay Round.60 This fact suggests, though it obviously does not prove, that the negotiating countries were content to let stand the interpretation offered in that 1994 adopted panel report.

Furthermore, one can make a case for the proposition that, as an economic matter, there is a certain arbitrariness to any exercise that purports to separate the causal effects of specific market factors in the harm suffered by a domestic industry.61 Accepting for the sake of argument the possibility of a coherent causal analysis, there are plausible interpretations other than that insisted upon by the AB. For example, in cases where a domestic industry has endured a rapid and serious reversal of fortunes, a decision-maker might be able to avoid attributing injury from imports to other factors by drawing the connection between imports and injury in a particularly rigorous way, with direct evidence of the causal effects of the imports. This might well establish that the required standard of "injury" caused by imports was met without an effort to quantify how much injury was caused by other market phenomena. Though it is unclear whether the United States explicitly relied upon Article 17.6(ii), perhaps because it was arguing to uphold the panel determination, the possibility of "permissible" alternative interpretations was not addressed by the AB.

As we will see, the complete disregard of 17.6(ii) found in the two Japan Steel discussions is not the norm in panel and AB legal interpretations of the Anti-Dumping Agreement. However, in one other case a panel seemingly went out of its way not to show deference to an administrative interpretation of a nation's own regulations. It thereby found an inconsistency with WTO obligations, with nary a mention of Article 17.6(ii). In Korean DRAMs, the panel ignored the actual practice of DOC in revoking anti-dumping findings and instead concluded that the Department's regulation was on its face inconsistent with Article 11.2 of the Anti-Dumping Agreement.62

The panel construed 19 C.F.R. [sec] 353.25, which reads in relevant part: "The Secretary may revoke an order . . . if the Secretary concludes that . . . [i]t is not likely that those persons will in the future sell the merchandise at less than foreign market value."63 In a discussion seemingly modeled on the dialogue of Through the Looking Glass,64 the panel concluded that this regulation was a "mandatory requirement" inconsistent with the Article 11.1 rule that "an anti-dumping duty shall remain in force only as long as and to the extent necessary to counter-act dumping" and the Article 11.2 provision that "parties shall have the right to request the authorities to examine whether the continued imposition of the duty is necessary to offset dumping."65 To be fair to the panel, it is not easy to compare precisely the relative probabilities in these and similar tests. The point, though, is that the panel never examined whether the actual decision of DOC in the case was in fact consistent with Article 11 requirements. Nor did the panel inquire into how DOC had previously applied the guideline embodied in its regulation.

Coincidentally, the U.S. Court of International Trade decided the identical issue in the identical case.66 Though the court was clear that its decision depended on U.S. law and not WTO obligations, including a WTO panel decision,67 it cited the traditional canon of statutory construction that a statute should not be interpreted to conflict with an international obligation, absent express language to the contrary.68 Notwithstanding the panel's decision, to which it explicitly referred,69 the court found that the discretion retained by DOC saves the regulation from any possible conflict with the Anti-Dumping Agreement.70 Though the court's decision itself leaves unanswered the question of whether DOC action in this case violated WTO obligations, its deference to DOC's discretionary authority to interpret and implement its own regulations contrasts sharply with the panel decision.

2. Overturn National Authorities, Infer Application of 17.6(ii)

The AB's approach to the second legal issue discussed in Japan Steel is perhaps most characteristic of its posture towards Article 17.6(ii): the existence of the standard is acknowledged but has no palpable effect on the AB's analysis. Here, the AB upheld a panel determination and, in response to a complaint from the importing country that the standard of review was ignored, simply said that the panel had "in effect" concluded the country's interpretation to be impermissible. Yet in neither the original panel determination nor the appellate report was the standard actually applied.

This issue involved a variation on the use by DOC of "facts available" in the absence of information provided by a Japanese steel exporter. DOC ruled that KSC, the steel exporter, had not cooperated with the investigation because it had not "made every effort" to obtain certain cost and price information from its joint venture partner. In accordance with U.S. law,71 DOC consequently used "adverse" facts available concerning the prices and costs in determining the dumping margin. Japan contested the finding that KSC had not cooperated, but not the use of "facts available" as such.72 Thus the issue was whether non-cooperation justified using "adverse" facts available. The AB upheld the panel's conclusion that the DOC was not justified in finding a lack of cooperation. Rejecting the argument of the United States that this was a question of fact, the AB characterized "the issue as one of legal interpretation."73 Interestingly, the AB did not, as one might have expected, then cite to Article 17.6(ii). It did, consistently with Article 31 of the Vienna Convention, turn to the relevant text of the Agreement, paragraph 7 of Annex II: "It is clear, however, that if an interested party does not cooperate and thus relevant information is being withheld from the authorities, this situation could lead to a result which is less favourable to the party than if the party did cooperate."74

The AB appeared to regard the issue as one of construing the term "cooperate." Citing a dictionary meaning of cooperate, "to 'work together for the same purpose or in the same task,'" the AB found in this definition a suggestion that cooperation is a "process."75 Thus, "parties may very well 'cooperate' to a high degree, even though the requested information is, ultimately, not obtained."76 However, the Appellate Body declined to give an interpretation based only on the "ordinary meaning" of the relevant term, because "Paragraph 7 of Annex II does not indicate what degree of 'cooperation' investigating authorities are entitled to expect."77 Moreover, though the Body did not say so, acceptance of the concept of cooperation as process still requires a judgment as to whether the interested party was in fact "working together" with the administrative authorities in accumulating all relevant information.

Then, again as required by Article 31 of the Vienna Convention, the AB attempted to construe paragraph 7 in its context. The AB found particularly relevant other provisions of the Anti-Dumping Agreement that indicate some solicitude for interested parties being investigated.78 Paragraph 5 of Annex II prohibits authorities from disregarding information that "may not be ideal in all respects" if an interested party "has acted to the best of its ability."79 Paragraph 2 of Annex II allows authorities to require submission of information in a particular medium (such as computer language), but not "if the interested party does not maintain its computerized accounts . . . and if presenting the response as requested would result in an unreasonable extra burden on the interested party, e.g. it would entail unreasonable additional cost and trouble."80 Finally, Article 6.13 requires the authorities to take "due account of any difficulties experienced by interested parties, in particular small companies, in supplying information requested, and shall provide any assistance practicable."81

From the first two provisions the AB concluded that the Agreement reflects "a careful balance between the interests of investigating authorities and exporters."82 The AB read Paragraph 2 "as another detailed expression of the principle of good faith . . . that informs the provisions of the Anti-Dumping Agreement."83 Investigating authorities "are not entitled to insist upon absolute standards or impose unreasonable burdens" upon exporters.84 From Article 6.13, the AB understood the "cooperation" specified in the Agreement to be a "two-way process" requiring the authorities to "make certain allowances for" interested parties where there are "genuine 'difficulties'".85 In light of this "interpretation," the AB upheld the panel's conclusion that DOC had gone "far beyond any reasonable understanding of any obligation to cooperate implied by paragraph 7."86 Specifically, both the panel and the AB found unreasonable the insistence of DOC that KSC invoke its rights under a shareholders' agreement to force its joint venture partner to produce the relevant information. The panel had asserted that such a request would have "disrupted" the business relationship between KSC and its joint venture partner.87

Mindful, perhaps, that the panel had given short shrift to Article 17.6, the AB concluded with a single sentence addressing the standard of review: "In effect, the Panel held that USDOC's conclusion that KSC had failed to 'cooperate' in the investigation did not rest on a permissible interpretation of that word."88 Of course, the use of the phrase "in effect" is an acknowledgement that the panel had not, in fact, followed the Article 17.6(ii) standard of review. Nor did the AB. Its own interpretation of "cooperate" is plausible but hardly compelling. Insofar as paragraph 2 of Annex II imposes a constraint upon requests for information in one particular circumstance-where computerized data is involved-the absence of such a qualification in other provisions of Annex II and the body of the Agreement might fairly be read as indicating the absence of constraint. More to the point, the AB nowhere answers the question of why "cooperate" in paragraph 7 might not permissibly be read to mean, in the case under discussion, "cooperate through the use of all legal means available to the responding party." The discussion of the AB, like that of the panel it reviewed, looks suspiciously like a determination of the "best" interpretation of a term, not whether a term is unambiguous in its context or susceptible to more than one "permissible" interpretation.

A similar approach and outcome may be found in a case involving quite different issues, a challenge to the imposition by the European Communities of dumping duties on bed linen from India.89 This case involves the "zeroing" of "negative" margins for some product types, an issue that is arcane even by the standards of anti-dumping law. A bit more detail is thus in order. The issue of zeroing had been prominent during the Uruguay Round negotiations on the Anti-Dumping Agreement. Prior to the implementation of the Uruguay Round results, for example, DOC practice had been to compare a weighted-average foreign market value with individual export sales to the United States. If a specific export sale was at a price below the weighted-average foreign market price, the resulting dumping margin was calculated. If, on the other hand, a specific export price was above the average foreign market price, the margin was counted as "zero." Consequently, when all the margins of the specific sales (adjusted for the quantities sold at those prices) were totaled up and averaged, the instances of higher export prices were not permitted to offset instances of lower export prices. It was thus possible to have an affirmative dumping finding in a case where the weighted average of export prices was higher than the weighted average foreign market value. A number of countries sought change in this practice. As a result, Article 2.4.2 was included in the Agreement:

Subject to the provisions governing fair comparison in paragraph 4, the existence of margins of dumping during the investigation phase shall normally be established on the basis of a comparison of a weighted average normal value with a weighted average of prices of all comparable export transactions or by a comparison of normal value and export prices on a transaction-to-transaction basis.90

The interpretive problem arose because the EC calculated multiple margins, based on the different kinds of bed linen exported by India (referred to, somewhat misleadingly, as "models"). As explained by the AB:

[F]irst, the European Communities identified with respect to the product under investigation-cotton-type bed linen-a certain number of different "models" or "types" of that product. Next, the European Communities calculated, for each of these models, a weighted average normal value and a weighted average export price. Then, the European Communities compared the weighted average normal value with the weighted average export price for each model. For some models, normal value was higher than export price; by subtracting export price from normal value for these models, the European Communities established a "positive dumping margin" for each model. For other models, normal value was lower than export price; by subtracting export price from normal value for these other models, the European Communities established a "negative dumping margin" for each model. Thus, there is a "positive dumping margin" where there is dumping, and a "negative dumping margin" where there is not . . . [T]he European Communities then added up the amounts it had calculated as "dumping margins" for each model of the product in order to determine an overall dumping margin for the product as a whole. However, in doing so, the European Communities treated any "negative dumping margin" as zero-hence the use of the word "zeroing". Then, finally, having added up the "positive dumping margins" and the zeroes, the European Communities divided this sum by the cumulative total value of all the export transactions involving all types and models of that product [to obtain] an overall dumping margin for the product under investigation.91

Needless to say, India was unhappy with this procedure, since the "negative margins" for some product types were not allowed to offset the "positive margins" for other product types.

The EC argued that Article 2.4.2 does not address the question whether an importing country may "zero" out a "negative" margin for one model while combining the margins for each of the different models to determine if there is dumping of the product.92 The panel had rejected this argument. Though the panel read the requirement that a weighted average normal value be compared to all comparable export transactions to permit a model-by-model comparison, it stated that "Article 2.4.2 must be understood to apply to the entire process of determining the existence of margins of dumping for the product, and that there is no 'subsequent stage' which escapes entirely from the strictures of Article 2.4.2."93 This reasoning appears untenable. If there can be multiple margins for a single "product" based on differing models, there must be some second stage of analysis that yields a single margin for a product, in accordance with the overarching Article 2.1 requirement that dumping may be found "if the export price of the product exported from one country to another is less than the comparable price . . . for the like product when destined for consumption in the exporting country."94

Perhaps recognizing the infirmity of the panel's analysis, the AB used different reasoning to reach the same result. Like the panel, the AB emphasized that the Article 2.1 reference to a product means that the "margins" of dumping referred to in Article 2.4.2 can refer only to a single product.95 Unlike the panel, the AB remained agnostic on the issue of whether a comparison of different models of a single product is acceptable. The AB then focused on the Article 2.4.2 requirement that "all comparable export transactions" be included in the comparison,96 emphasizing the adjective "all": "By 'zeroing' the 'negative dumping margins', the European Communities, therefore, did not take fully into account the entirety of the prices of some export transactions, namely, those export transactions involving models of cotton-type bed linen where 'negative dumping margins' were found."97

Of course, as the EC pointed out, Article 2.4.2 requires that all comparable export transactions be considered. The AB dismissed this argument rather summarily:

In our view, the word "comparable" in Article 2.4.2 does not affect, or diminish in any way, the obligation of investigating authorities to establish the existence of margins of dumping on the basis of "a comparison of the weighted average normal value with the weighted average of prices of all comparable export transactions".98

In specifying the "ordinary meaning" of "comparable," the AB then cited to the dictionary definition "able to be compared."99 From this definition and a strained reference to the "context" of other parts of Article 2.4, the AB concluded that the EC's methodology impermissibly excluded some export transactions by zeroing negative margins for discrete models: "All types or models falling within the scope of a 'like' product must necessarily be 'comparable'."100

With this interpretation, the AB essentially read the word "comparable" out of Article 2.4.2 or, more precisely, rendered its presence meaningless. This in itself is contrary to the usual norm of international treaty interpretation that parties to a treaty are assumed to intend that its provisions have an effect and, thus, are not to be interpreted as meaningless.101 The significance of this norm in the present case is underscored by the drafting history. Insertion of the word "comparable" was the only change made to Article 2.4.2 between publication of the Draft Final Act of the Uruguay Round negotiations and the text as adopted.102 A panel in another anti-dumping case found this fact to suggest that inclusion of the word "was not merely incidental but reflected careful consideration by the drafters."103

Furthermore, the AB's reference to a dictionary definition of "comparable" was exceedingly selective. The very dictionary consulted by the AB lists as additional definitions "fit to be compared" and "worth comparing."104 The most recent edition of the same dictionary defines "comparable" as "able to be likened to another; similar" with the sub-sense "of equivalent quality."105 All these other definitions add a sense of selectivity to the concept of "comparable" that the AB overlooked in citing the almost tautological definition "able to be compared." A fairer reading of the "ordinary meaning" of the term in accordance with usual principles of interpretation would suggest that its presence in the sentence has some effect upon the meaning of the sentence. The meaning may indeed be susceptible to different interpretations, perhaps including the one favored by the AB. But the AB failed to admit of this possibility in an opinion that is hard to characterize as a straightforward application of the Vienna Convention, much less Article 17.6(ii).

The AB did mention 17.6(ii), but only in passing at the end of its analysis, in response to the EC's complaint that the panel had ignored the applicable standard of review: "It appears clear to us from the emphatic and unqualified nature of this finding of inconsistency that the panel did not view the interpretation given by the European Communities . . . as a 'permissible interpretation' within the meaning of Article 17.6(ii) of the Anti-Dumping Agreement."106

It is curious for the AB to dismiss the standard of review issue by reference to panel reasoning which it modified significantly. Moreover, in both declining to reprimand the panel for its failure to apply 17.6(ii) and in failing to apply the standard explicitly in its own report, the AB again minimized the significance of this provision in WTO jurisprudence.107

3. Overturn National Authorities After Considering Alternative Interpretations

Notwithstanding the short shrift given to Article 17.6(ii) in parts of Japan Steel and other cases, the AB and panels do at times consider interpretations offered by the national authorities of the importing party. Usually, though not always, the issue is decided adversely to the importing government. In these instances, however, the AB and panel analyses are more careful and ultimately convincing than those discussed previously.

Two other legal issues considered in Japan Steel manifest this approach, though not to the same degree. One involved the panel's holding that, when an importing country investigates some but not all exporters of the subject merchandise, it may not use margins established in any way on the basis of "facts available" to calculate a margin for the products of exporters not investigated. The parties agreed that the Agreement permits what is, in effect, a sampling technique by national authorities under circumstances defined in Article 6.10.108 The issue was whether the sample of exporters used to calculate an "all others" margin for exporters not investigated could include exporters for which any "facts available" had been used in the calculation of their own margins.

The AB affirmed the panel's conclusion in favor of Japan. The key provision for interpretation was Article 9.4, which provides that, in establishing a weighted-average margin on the basis of a limited number of exporters, "the authorities shall disregard . . . margins established under the circumstances referred to in paragraph 8 of Article 6."109 Article 6.8, as noted in the first Japan Steel issue discussed above, permits the use of information from sources other than the exporter under investigation where that exporter "refuses access to, or otherwise does not provide, necessary information."110 The AB noted that Article 6.8 applies where "recourse to facts available is needed to cure the lack of even a very small amount of information."111 The AB reasoned that, inasmuch as Article 9.4 refers simply to "the circumstances referred to in Article 6.8," it must therefore prohibit the inclusion in an "all others" average of margins based even on a small amount of information from facts available.112

In itself, this seems a reasonable, perhaps even necessary, reading. More significant for present purposes is that the AB explicitly considered and carefully rejected the alternative reading offered by the United States-that the Article 9.4 prohibition applies only where an exporter's margin has been based entirely on facts available. The AB pointed out that the United States was, in effect, adding an adverb such as "entirely" or "wholly" in front of the word "established" in Article 9.4.113 The AB also responded to the U.S. argument that its reading would make calculation of an "all others" reading impractical in those investigations where all sampled exporters' margins are derived using some "facts available." The AB pointed out that it is also possible that in some cases all the margins of exporters actually surveyed would have been based entirely on "facts available."114 The AB frankly acknowledged that there is a "lacuna" in Article 9.4 and that it might someday be required to decide how that lacuna might be bridged, but concluded that such an exercise was unnecessary in the present case.115

The AB also responded directly to the U.S. claim that the panel's determination was inconsistent with Article 17.6(ii). The AB, referring to the analysis described in the preceding paragraph, concluded that Article 9.4 is not "susceptible, under the customary law rules of treaty interpretation, of the interpretation on which the United States' measure rests."116 The panel, in turn, had identified its own task as determining whether Article 9.4 "admits of the U.S. interpretation,117 a phrase taken from Article 17.6(ii). Though the AB mentioned 17.6(ii) only at the end of its analysis rather than at the outset-where one might expect a reviewing body to place the applicable standard of review-the discussion of the issue itself is consonant with the 17.6(ii) requirements in a way that other Japan Steel discussions were not.

The second issue in Japan Steel that was treated more in accordance with the terms of Article 17.6(ii) was a challenge to the DOC practice of excluding from its calculation of "normal value" in the home market any sales by the exporter to an affiliated entity unless the prices of those sales are, on average, at least 99.5 percent of the prices charged to unaffiliated customers. The United States justified this practice by reference to Article 2.1, which states that dumping exists where the export price is "less than the comparable price, in the ordinary course of trade, for the like product when destined for consumption in the exporting country."118 The panel disagreed, on the ground that the "arm's length" test applied by the United States "does not, in fact, test for differences in prices of sales to affiliated customers . . . [but] only tests whether prices to affiliated customers are lower, on average, than prices to unaffiliated customers."119

The AB affirmed on somewhat different grounds. The AB stipulated that transactions between affiliated parties "might not be in the ordinary course of trade,"120 and that the Anti-Dumping Agreement did not provide a methodology for determining when such a circumstance occurred.121 The AB then reached its key conclusion:

Although we believe that the Anti-Dumping Agreement affords WTO members discretion to determine how to ensure that normal value is not distorted through the inclusion of sales that are not "in the ordinary course of trade," that discretion is not without limits. In particular, the discretion must be exercised in an even-handed way that is fair to all parties affected by an anti-dumping investigation. If a Member elects to adopt general rules to prevent distortion of normal value through sales between affiliates, those rules must reflect, even-handedly, the fact that both high and low-priced sales between affiliates might not be "in the ordinary course of trade."122

Because the United States applied its "bright line" 99.5% test only to lower-priced sales, and applied a much vaguer test that did not automatically exclude higher-priced sales to affiliated customers, the AB concluded that there was a lack of even-handedness that systematically disadvantaged exporters.123

Though neither cited Article 17.6(ii) as such, both the panel and the AB explicitly stated that the U.S. practice did not rest on a "permissible interpretation" of "sales in the ordinary course of trade."124 Each considered and rejected the U.S. interpretation that nothing in the Anti-Dumping Agreement compelled the use of the same test for determining whether artificially high and artificially low prices to affiliated parties were outside the ordinary course of trade. Thus, as with the "all others" interpretation, the AB provided an analysis that implicitly rejected the U.S. interpretation as not "permissible." However, here the AB did not make quite as careful a case for its reading from the text and context of the Anti-Dumping Agreement. Instead, its decision rested on concepts of fairness and even-handedness which, although appealing on one level, may be problematic in the context of an international dispute settlement body's interpretation of a text agreed upon in a multilateral trade negotiation. This possibility is discussed later in Part III.

Other anti-dumping cases contain a similar approach to questions of legal interpretation. Both significant legal issues in the unappealed Korea Steel panel report were decided against the importing country, but only after consideration and rejection as "impermissible" of alternative interpretations offered by the respondent United States. In the first of the two issues, DOC had adjusted the comparison of home market "normal value" and export price to account for the bankruptcy of a U.S. customer which had, accordingly, not paid for steel it had purchased from the Korean company.125 This step increased the dumping margin by lowering the calculated "price" received by the company for its remaining U.S. sales.126 The United States argued that this adjustment was permissible under Article 2.4, which provides that "[d]ue allowance shall be made in each case, on its merits, for differences which affect price comparability, including differences in conditions and terms of sale."127 Specifically, the United States reasoned, a sale on credit was a term or condition of sale, and thus the credit expense of an unpaid debt was included within the Article 2.4 terms.128

The panel rejected this argument. Its decision was carefully reasoned, economically sensible, and at least formally consistent with the standard of Article 17.6(ii), though one could imagine a more generous stance toward the U.S. interpretation. Based in part on dictionary definitions of "term" and "condition," the panel inferred that "'conditions and terms of sale'" in Article 2.4 "refers to the bundle of rights and obligations created by the sales agreement."129 While "differences in payment terms in the two markets" constitutes a different term or condition of sale, non-payment is not a condition or term, but a breach of the conditions and terms of the contract.130 The panel believed that its conclusion, based on the meaning of the contested phrase, was buttressed by the context of that phrase in Article 2.4, where it appears as one of several "differences affecting price comparability." Thus, argued the panel, "the requirement to make due allowance for differences that affect price comparability is intended to neutralise differences in a transaction that an exporter could be expected to [make]."131

The panel explicitly considered the U.S. contention-based on a different dictionary definition of "condition" as a "mode or state of being"-that an adjustment for bad debt was a permissible interpretation of Article 2.4. It is unclear whether, ultimately, the panel felt there was no ambiguity in the Article 2.4 phrase or whether it believed the U.S. reading to be impermissible even if some ambiguity existed. The panel, unlike the AB in parts of Japan Steel, frankly acknowledged that the language under consideration could have many meanings, and even stipulated the definition offered by the United States for purposes of argument.132 While the panel thought that it might be permissible to adjust "where the contractual provisions governing sales in the two markets were identical but the seller was aware from circumstances existing at the time of the sale that those provisions would likely entail different costs,"133 there was no such evidence in this case. The panel noted further that actual bad debt experience might be taken as evidence relevant to establishing the existence of different credit conditions in the different markets. But here, the United States did not use the bad debts as evidence of different circumstances; it simply treated the bad debt as a direct selling expense.

The second legal issue in Korea Steel involved the use of two separate "averaging periods" by DOC in its calculation of margins. The two periods were divided by the onset of the Korean financial crisis in November 1997, as a result of which the value of the Korean won declined forty percent in a two-month period. The United States justified this method on the ground that Article 2.4 states that price comparisons should be "in respect of sales made at as nearly as possible the same time" and that the depreciation of the won during the period of investigation meant that the transactions from these different periods were not "comparable" within the meaning of Article 2.4.2.134

On this second legal issue, the panel's own reasoning is much less careful than in its first analysis. Indeed, the economic merits of the U.S. position-so often not on the side of anti-dumping authorities-seem fairly strong. Though the panel faithfully considered the U.S. argument that its interpretation of "comparable" is "permissible" and admirably declined to hide behind a selective use of dictionary definitions, the legal basis for its decision against the United States is not altogether clear. The panel acknowledged that the Article 2.4 language made the timing of sales relevant in deciding the comparability of export and home market transactions.135 But, said the panel, the very contemplation of averaging in Article 2.4.2 meant that sales made at different times may still be comparable. The panel concluded from this observation that "the requirement that a comparison be made between sales made at as nearly as possible the same time requires as a general matter that the periods on the basis of which the weighted average normal value and the weighted average export price are calculated must be the same."136 But why, exactly? The panel, having implicitly acknowledged the ambiguity in Article 2.4.2, simply rejected the U.S. position. To further muddy the waters, the panel then went on to say that if there were both significant changes in "normal value" and significant differences in the ratio of home market to export sales, then "comparability" might be affected.137 Yet the panel pointed to no text or context to explain why this situation would be different.

The panel, in a case involving an Indian challenge to U.S. anti-dumping and countervailing duty measures on steel plate, also rejected a U.S. interpretation, but also only after a careful consideration of the U.S. argument.138 The contested issues in this case again involved the use of "facts available." DOC had disregarded all information submitted by the Indian exporter and based its determination exclusively on facts available. The United States argued that it was entitled to do so under Article 6.8 of the Agreement so long as any "essential" element of the requested information was not provided in a timely fashion. The panel concluded that this interpretation is not consistent with the requirement in paragraph 3 of Annex II to the Agreement (which, as earlier noted, elaborates upon the Article 6.8 obligation) that "all information" that is appropriately and timely submitted be taken into account by the investigating authorities.139 The panel dutifully cited to the Article 17.6(ii) standard at the beginning of its discussion,140 though like other panels, it did not refer to the standard in its actual analysis of the contested provision. Without using the 17.6(ii) terms, the panel's reasoning does show that the U.S. reading was impermissible.141

Finally, in Egypt-Steel Rebar from Turkey,142 a panel rejected Egypt's contention that an importing country is never required by Article 2.4 of the Agreement to make a price adjustment for credit costs where constructed value is the basis for comparison with export prices. Egypt claimed that, because a constructed value is a "notional price," its level "cannot be influenced by any conditions and terms of the relevant sales."143 Egypt's submissions to the panel explicitly argued that its reading was a "permissible interpretation" of Article 2.4.144 The panel fairly summarized Egypt's arguments but found that such a reading was "not possible."145 The panel stated clearly that Article 2.4 requires "a fact-based, case-by-case analysis of differences that affect price comparability."146 The panel supported this conclusion with the Article 2.4 language that "[d]ue allowance shall be made in each case, on its merits" for differences, as well as other clauses of Article 2.4 indicating that case-by-case consideration of possible adjustments is required.147

4. Uphold National Authorities

The United States, the responding country in Japan Steel, did prevail on two legal issues. In neither case, however, did the specific standard of review in Article 17.6(ii) appear to play much of a role. Moreover, in practical terms neither result was a clear victory for the United States.

One issue arose because of application of the 99.5% test discussed above. Once DOC had excluded sales based on this test, it used the first "downstream" sale by the affiliate to an unrelated purchaser as its basis for comparison. The panel found the use of these downstream sales a violation of Article 2.1 on the grounds that they were not sales of the "exporter or producer" for which the margin was to be calculated.148 The AB found that Article 2.1 set four conditions for selecting sales in calculating normal value, corresponding to each of four clauses in the definition of normal value.149 The AB went on to note that Article 2.1 is "silent as to who the parties to relevant sales transactions should be."150 The AB saw "no reason to read into Article 2.1 an additional condition that is not expressed"151 and concluded: "In these circumstances, we find that the reliance by USDOC on downstream sales to calculate normal value rested upon an interpretation of Article 2.1 of the Anti-Dumping Agreement that is, in principle, "permissible" following application of the rules of treaty interpretation in the Vienna Convention."152

At least in the present case, the practical value of this ruling is uncertain, for two reasons. First, because the AB previously rejected DOC's use of the 99.5% guideline for excluding sales to affiliates, it is not clear that the United States will still be able to use downstream sales in calculating normal value. Second, the qualification "in principle" may turn out to be a significant one. The AB noted Japan's complaint that DOC had not properly adjusted the prices of the downstream sales in accordance with Article 2.4. The AB opined that when investigating authorities use the prices of downstream sales as a basis for establishing normal value, they have "a particular duty to ensure the fairness of the comparison because it is more than likely that downstream sales will contain additional price components which could distort the comparison."153 Because this issue had not been addressed by the panel, the AB declined to decide it,154 and it remains open for later consideration.

The practical benefits of the second legal interpretative "victory" of the United States are even less certain. The AB found that the so-called "captive production" provision of U.S. law does not on its face violate the Anti-Dumping Agreement. The captive production provision requires that, under certain conditions defined by statute, the ITC shall-in calculating market share and other indicators of the injury to a domestic industry required before dumping duties may be imposed-"focus primarily on the merchant market for the domestic like product."155 That is to say, the "captive market" in which companies internally consume the merchandise they produce shall not be the primary focus of the ITC inquiry. Insofar as intra-corporate consumption of a good is less susceptible to import competition than the "merchant market" of unrelated customers, this provision presumably increases the chances for an affirmative determination of injury.

The AB expressed considerable misgivings over the captive production provision.156 However, because interpretation of the provision within U.S. domestic law was unsettled, and because the provision did not require an "exclusive focus" upon the merchant market, the AB declined to find it a facial violation of the Article 3.1 requirement of an "objective examination" of the "domestic industry," in turn defined in Article 4.1 as the "domestic producers as a whole of the like products."157 The value of this apparent triumph for the United States was immediately diminished, however, by the AB's assessment that the ITC had violated Articles 3.1 and 3.4 in its application of the captive production provision in this case.158 Finally, it is worth noting that the AB neither invoked Article 17.6(ii) nor used terms such as "permissible interpretation." This suggests that the standard of review had little impact on the AB's analysis, even though its reference to the panel analysis159 and its construction of key terms in Articles 3 and 4 made clear that it was engaged in legal interpretation.160

There are few issues from other WTO anti-dumping cases in which the importing authorities' legal interpretation on a seriously contested issue has been accepted. In some instances a panel or the AB has more or less summarily dismissed an argument of the exporting country.161 There is one issue on which the panel accepted a contested interpretation of the authorities and, more significantly, indicated that this interpretation was "permitted" but not necessarily required.162 Poland argued that Thailand had drawn on an excessively narrow category of products to determine the imputed profit to be included in a constructed value that was used by Thai authorities in the absence of sufficient actual sales for establishing "normal value." Looking at the language of Article 2.2.2, the panel found no guidance as to the requisite breadth of the product category on which the authorities should draw.163 Looking to the context of other provisions in the Agreement, the panel seemed to find a modest preference for the narrower category favored by Thailand.164 Thus, although the panel was explicit that the Thai interpretation was "permitted," yet not required, the panel itself apparently thought this the better reading. So, although this issue was decided favorably to the importing authorities using Article 17.6(ii) language (but without reference to the provision itself), even this case does not present an instance in which the Article 17.6(ii) standard demonstrably affected the outcome.

III. ASSESSING THE OUTCOMES

With the possible exception of the Thailand-Steel from Poland panel decision previously discussed, the existence of the Article 17.6(ii) standard has had no perceptible impact on WTO review of national anti-dumping actions. The point is not that the importing country loses every issue (though they have lost at least one significant issue in every case). Rather, the preceding section demonstrates that the special standard of review has not, as U.S. negotiators surely hoped, changed any importing country losses to wins. Where the AB reads a provision of the Anti-Dumping Agreement in the same way as the national authorities, or finds the reading of a complaining country implausible, Article 17.6(ii) is not necessary for the importing country to prevail. The AB has barely mentioned the standard of review and has not followed it in numerous instances where the importing country appears to have offered a plausible interpretation of the Agreement to support its action. The most explicit application of the standard is found where the arguments of the importing country seem particularly weak, and thus quite clearly do not offer a "permissible" reading of the text.

The obvious questions are, first, why the U.S. effort to export this feature of administrative law to the international legal arena has to this point failed and, second, whether the AB's practice should be applauded or deplored. I will first dispose of two possible explanations, one that the AB simply does not know any better and the other that international law compels the AB to ignore Article 17.6(ii). Having concluded that the AB has chosen this path, I then consider the normative arguments for and against it. Although this is not the place to rehearse the extensive normative debate over WTO dispute settlement, it is important to illustrate briefly how some of the important arguments concerning a "strong" or "activist" AB apply in the discrete context of Article 17.6(ii) reviews.

A. Explanations

1. AB Misunderstanding

One possibility is that members of the AB simply may not understand that standards of review can differ. Perhaps U.S. negotiators erred in attempting to insert into an international agreement a standard of review of a sort found in domestic administrative law. It is clear that the expectations of U.S. negotiators in arguing for 17.6(ii) have not been realized and, were the popular image of the AB members as "trade bureaucrats" accurate, their backgrounds might be a plausible explanation. However, the AB is dominated by people with extensive legal backgrounds. Of the thirteen people who have sat on the AB since its inception, seven had been law professors.165 Two were judges, including an associate justice of the Supreme Court of the Philippines and a judge from Australia, a common-law country.165 Two others had been trained as lawyers, including the sole American to have been appointed to the AB.167 Thus only two of the thirteen appointees to the AB did not have legal training and experience. The panel that decided Japan Steel was composed of the former Philippine Supreme Court justice and two law professors.168

Of course, given the differences among legal systems, legal training is no guarantee of the sensitivity to differing standards of review expected of American judges. Yet an American lawyer has been on the AB from the outset, and some other members-such as Claus-Dieter Ehlermann and Mitsuo Matsushita-are very sophisticated legal scholars with extensive exposure to multiple legal systems, including that of the United States. Moreover, standard of review of administrative actions is a topic much discussed in many legal systems, not just that of the United States.169 Most significantly, discussion of Article 17.6(ii) in Japan Steel itself suggests that the AB understood the importance of differences in standards of review. Recall that the AB began its analysis of the issues with a discussion of how Article 17.6 fit with the generally applicable instructions to WTO panels in dispute settlement cases.170 So, while we cannot discount entirely the possibility that cross-cultural differences on judicial standards of review explain the AB's disregard of 17.6, this is a less than satisfying explanation.171

Another possible "misunderstanding" by the AB is the mistaken belief that it needed to disregard the 17.6(ii) standard in order to overrule the national anti-dumping authorities on the issues discussed in Part II. As experience in the United States under Chevron demonstrates, there is ample room for decision-makers to overrule administrative interpretations even as they apply the generous standard of review. A review of recent decisions rendered by the Court of International Trade and the Court of Appeals for the Federal Circuit reveals that both courts do frequently defer to Commerce and USITC. Yet there are recent decisions in which the courts apply the Chevron standard and still find against the administrative agency, either under the first172 or second173 parts of the standard. Because the AB could have put forth an analysis with a restrictive interpretation of "permissible," its rather obvious disregard of Article 17.6(ii) in many instances is somewhat puzzling. Perhaps a lack of familiarity with judicial practice in common law countries is a partial explanation, though I rather doubt it. In any case, the failure of the AB to mask its de-emphasis of 17.6 provides a firmer basis for concluding that it did, in fact, intend just such a de-emphasis.

2. Legal Compulsion

A second possibility is that the AB believes Article 17.6(ii), in its context, to have little effect as a matter of law, despite its apparent similarities with the Chevron standard. There have been suggestions that the first sentence of 17.6(ii)-requiring application of the rules of interpretation of customary international law-may resolve virtually all interpretive questions.174 Were that the case, the second sentence of 17.6(ii) would almost never be operative. These suggestions, which pre-date the AB's report in Japan Steel, do not accord with most commentary on the rules of interpretation laid out in the Vienna Convention. More importantly, statements of the AB itself are inconsistent with this view of Article 17.6(ii).

Commentators on Articles 31 and 32 of the Vienna Convention are skeptical that these rules-and other rules of treaty interpretation that complement these two articles-always produce a single interpretation of a treaty provision. After a thorough examination of the rules and their application, the author of the leading treatise on the Vienna Convention concluded that the Convention made a significant advance in treaty interpretation by establishing a clear set of "guidelines."175 However, his "review of recent international law case law on treaty interpretation reveals only too clearly that widely differing results can still be achieved even if a conscious effort is being made to apply the Convention rules."176 He concluded that the generality of the rules assured that "serious divisions of opinion" would arise over questions of treaty interpretation.177 Other commentators have reached similar conclusions,178 which accord with the almost axiomatic view among American legal academics that texts-even when read in their context and with the assistance of legislative or preparatory history-frequently offer multiple interpretive possibilities.

The AB itself seems not to subscribe to the view that treaty ambiguities will always, or nearly always, be resolved by reference to the Vienna Convention rules. In its discussion of 17.6(ii) m Japan Steel, the AB emphasized that the "second sentence of Article 17.6(ii) presupposes that application of the rules of treaty interpretation in . . . the Vienna Convention could give rise to, at least, two interpretations of some provisions" of the Anti-Dumping Agreement.179 This acknowledgement that the second sentence of Article 17.6(ii) must have some effect is also consistent with the AB's embrace elsewhere of the oft-repeated view that, in interpreting an international agreement, "[a]n interpreter is not free to adopt a reading that would result in reducing whole clauses or paragraphs to redundancy or inutility."180

Of course, the fact that Article 17.6(ii) has some legal effect does not by any means imply that the effect is necessarily similar to that of the Chevron standard in U.S. administrative law. The context is different, juxtaposed as the standard is to the requirement for interpretation of the Agreement in accordance with international law norms. The wording, while very similar to one of the formulations used by the Supreme Court in Chevron, is not identical. The United States originally proposed to shield administrative interpretations that were "reasonable," as opposed to "permissible."181 Thus the AB might reasonably have concluded that the 17.6(ii) standard, including the scope of "permissible" definitions, is quite different from the applicable standard of review in domestic U.S. law. Yet the AB has never pursued these questions. Indeed, but for the introductory discussion in Japan Steel, the AB has not explicated Article 17.6(ii). In and of itself this is noteworthy, because the normal practice of both panels and the AB in the first seven years of the WTO dispute settlement system is to parse the language of contested provisions of WTO agreements in excruciating detail.182

A different legal rationale for disregarding Article 17.6(ii) is almost the converse of the argument just considered. Rather than rejecting 17.6(ii) because there can never be ambiguity where Vienna Convention principles are followed, the AB may be ignoring 17.6(ii) because it is superfluous. The customary international law interpretive principle of in dubio mitius holds that, where a treaty provision is ambiguous, "that meaning is to be preferred which is less onerous to the party assuming an obligation."183 Because a member's chosen interpretation of an ambiguous provision of the Anti-Dumping Agreement is presumably that least burdensome to that member, the AB might regard 17.6(ii) as simply a specific expression of an interpretive principle which it applies more generally. Of course, insofar as the AB does not, in fact, closely follow the in dubio mitius principle, this possible legal rationale for disregarding 17.6(ii) is not applicable.

B. Normative Arguments for Disregarding 17.6(ii)

It appears, then, that the AB neither misapprehends the concept of a standard of review nor believes itself compelled by the first sentence of 17.6(ii) to relegate the second sentence to a position of near irrelevance. Without some other explanation, we must conclude that the AB has chosen to read 17.6(ii) out of the Agreement. This hypothesis is supported by some of the AB's actions: its suggestion in Japan Steel that 17.6(ii) was essentially congruent with Article 11 of the DSU;184 its invocation of 17.6(ii) only where a more convincing case for rejecting the interpretation of administering authorities was available; its uncharacteristic failure to interpret key terms of the standard itself, such as what constitutes a "permissible" interpretation. An imputation of a choice to disregard a provision of the Agreement is obviously not readily proven (absent an unexpected confession to that effect). Ultimately, though, the question of the AB's actual motivation(s) is of secondary concern to the question whether the AB's practice is desirable. Accordingly, my focus in this section is upon identifying normative positions that are consistent with AB practice, rather than upon trying to prove that these positions inform that practice.

1. Preference for Free Trade

One popular view of the WTO in general and its dispute settlement system in particular is that the organization is biased towards free trade. Given the organization's origins and purpose, it would not be surprising if, when confronted with an action by a member state that impeded trade, the dispute settlement bodies were inclined to find a violation of some WTO obligation. Concerns of this sort have multiplied as more forms of domestic regulation are brought within the WTO's ambit. One need not attribute bad faith to the WTO Secretariat and to WTO panelists to believe that, in any given dispute, they will be more likely to see the harm to liberal trade norms than to environmental or health or redistributive norms.185 A preference for free trade that animates the AB in its decision-making could explain the inattention to Article 17.6. Because good faith application of 17.6 would, in at least some cases, uphold national actions that restrain trade, the AB may have chosen simply to disregard the special standard of review where it might have affected the outcome.

It is, of course, very difficult to prove the negative-that is, to prove that an overarching preference for free trade has not influenced the AB's treatment of 17.6. Indeed, I rather suspect it has. But there is a difference between a general influence and one that produces specific outcomes. If pro-trade bias were the dominant factor in WTO dispute settlement, exporting countries would win every decision. Moreover, recent AB decisions involving challenges to national health and environmental regulation suggest that the AB is becoming more willing to let stand at least some kinds of restrictions on trade.

In the Asbestos case, the AB reversed a panel finding that asbestos fibers and non-asbestos fibers used for similar purposes were "like" products for purposes of determining whether a French ban on the asbestos fibers was a violation of the national treatment requirements of GATT Article III:4.186 Though the panel had permitted the ban as falling within the GATT Article XX(b) health and safety exception, the AB went further by finding that the dangerousness of the asbestos to humans was itself a relevant physical characteristic in deciding whether there was a denial of national treatment at all.187 A few years before, in the celebrated Shrimp-Turtle case, the AB interpreted another Article XX exception to permit a U.S. prohibition upon importation of shrimp caught in a manner that placed an endangered turtle species at risk.188 The AB reached this determination, which effectively "overruled" an earlier GATT case, by reasoning that the terms in GATT articles do not have a "static" meaning, but that the meaning can evolve based on subsequent international conventions and practice.189 This potentially far-reaching analysis was conducted to uphold a policy that restricted trade.

The pro-trade orientation of a trade organization is surely relevant to decisions on what matters should be subjected to WTO rules.190 This orientation may well have affected the AB in the anti-dumping cases discussed earlier, but it seems too blunt to be a wholly satisfactory explanation. Something more than antipathy to all trade restrictions is influencing the AB in disregarding 17.6.

2. Hostility to Administration of Trade Remedy Laws

Another possible explanation of the AB's practice in marginalizing Article 17.6(ii) is a more specific aversion to national administration of trade remedies such as anti-dumping laws. This reading of AB practice is supported by its decisions in disputes involving other trade remedy laws, notably countervailing duty and "safeguards" laws.191 As with reviews of dumping determinations, every WTO case involving these other two trade laws has resulted in at least partial victory for the complaining, exporting country. While U.S. import actions have been the subject of more than half these decisions,192 the AB has also upheld findings of GATT and Safeguards Agreement violations by other countries.193 In these cases the AB and the panels have scrutinized national administrative actions quite closely. The AB has, in at least some of these cases, appeared quite aggressive in finding violations.194

Reasons why the AB may be particularly hostile to administration of national trade laws are not hard to imagine. The most obvious is that the national authorities administering these laws are widely perceived to favor domestic industries in their decisions. The familiar political economy story of how a small, well-organized group with concentrated interests can prevail politically over the diffuse interests of the public at large combines with the equally familiar story of" agency capture to produce a picture of institutional bias. Because the trade laws do allow so much discretion to those administrators, the AB may see considerable opportunity for mischief. To take one example from the Japan Steel case itself, the determination of when a responding party has provided adequate information, so as to preclude the use of best information available, is easily manipulated by an agency looking to disadvantage foreign competitors of domestic industries.

Close review of trade administrative actions may thus be central to a core purpose of the WTO-to constrain protectionism by the political systems of member states, in which foreign producers are either unrepresented or underrepresented. An international review of domestic administrative action may also be an institutional antidote to the power of concentrated domestic interests and, thereby, promote the lower prices and increased consumer choice that result from import competition. That is, the international review mechanism may work not just to stop one country from unfairly harming the interests of another. It may also retard special interests in one country from harming diffuse interests in that country itself.

From this perspective, a Chevron-type standard is wholly misplaced. As one pair of authors has commented, "some of the most common and most powerful justifications of the Chevron doctrine carry very little weight once transplanted to the context of GATT/WTO dispute settlement."195 The authors are referring to such traditional bases for judicial deference as specialized agency expertise and the political accountability of administrative agencies (at least administrative agencies that are part of the Executive Branch and thus ultimately answerable to the President). The "expertise" of an institutionally biased agency may lie as much in finding ways to disadvantage foreign producers as in knowing the most appropriate way to allocate costs across a wide range of products in a dumping case. And the political accountability of the agency is, at least under the political economy assumption mentioned above, a reason to limit rather than respect the exercise of agency discretion.196 Thus the standard of review in anti-dumping cases ought to be less, not more generous to the national administering authorities.197

Whether this reasoning actually lies behind the AB's disregard of Article 17.6, it presents a serious normative position. Moreover, it connects to some broader conceptual positions that could support a narrow reading of 17.6. For example, a different pair of authors has set forth a theory of the WTO, based on public choice analysis, as a structure that promotes international trade and democratic government by reducing the power of protectionist interest groups.198 Norms drawn from both microeconomics and institutional economics could also lend support: Where health or safety regulation is at stake, there is no obvious identity of interest between foreign producers and domestic consumers. Where, on the other hand, the issue is one of straightforward restraints on imports, there is such an identity of interest. Accordingly, governance institutions should be most activist in containing import-restricting actions that obviously impede the efficient allocation of resources. If domestic institutions cannot be configured with the proper set of incentives to achieve this end, then perhaps international institutions can do the job.

Finally, some variations on cosmopolitan concepts of justice might even be invoked in support of strict international scrutiny of domestic trade law enforcement. A domestic system that regularly subordinates the interests of some interested actors is not, from a cosmopolitan point of view, a just system. Again, international institutions can (and should) strive for a fair result taking all interests into account. It is true that current cosmopolitan arguments are generally oriented towards promoting justice for obviously disadvantaged groups around the world.199 This version of cosmopolitanism does not, at first glance, seem to provide much argument for a stringent international review process to favor Japanese steel producers and Italian floor tile producers over American steel producers and Argentine floor tile producers respectively. However, if one believes that the import-unfriendly administration of trade remedy laws has disproportionate effects on developing country exports, then such cosmopolitan views may to some extent bolster the AB's approach.200

Also relevant here is a kind of natural-rights version of cosmopolitanism, which has been associated for many years with Ernst-Ulrich Petersmann, former legal director of the GATT. Petersmann has urged that every economic actor should be protected from the arbitrary actions of politicians in all countries. To this end, he has advanced the sweeping proposal that WTO rules be incorporated and privately enforceable in the domestic legal systems of all WTO members.201 This position is often and understandably criticized as grounded in ideological opposition to government economic intervention of all sorts, including that arising from a robust democratic process. Yet, in its own way, it also reflects a cosmopolitan vision, in that it seeks to offset the influence of organized interests in powerful countries by giving direct legal rights to every economic actor of every nationality. While Petersmann's proposals are obviously beyond the power of the AB to achieve, his perspective could support the AB practice of strict scrutiny of trade restrictions taken by national administrative authorities.

3. Constitutional Aspirations

The AB's place in what has increasingly been referred to as the "trade constitution"202 suggests another possible explanation and justification for its disregard of Article 17.6(ii). De-emphasis of the standard of review may advance the broader agenda of strengthening the capacity of the WTO to restrict national practices inconsistent with liberal trade values, of establishing a more significant role for the AB in shaping the WTO, or both.

The "trade constitution" is a phrase with many possible meanings. In one sense it is unremarkable-a simple description of the powers and prerogatives of the WTO and its constituent entities. Usually, though, the phrase is intended to convey the much more expansive sense of "constitutionalism" akin to that in some domestic law systems-a set of founding and fundamental principles that trump other law. In the WTO context, the emphasis is upon the superiority of WTO rules over inconsistent national law or practice. This concept differs from conventional understandings of international law as binding on nations, in that the WTO dispute settlement system provides final and authoritative judgments as to the consistency of national practice with WTO obligations. The constitutional processes of the WTO, including implementation of WTO rulings in domestic law systems, are a significant change from traditional international law. In the words of one commentator: "[T]he WTO may be explained as a trade constitution having the capacity to provide a legal framework for the universalisation of binding norms of substantive law, which have a significant impact on private individuals at the domestic level."203

This vision of a trade constitution is controversial, even among academics,204 and there is certainly no direct evidence that the AB has embraced this vision. But the activism of the European Court of Justice in its formative years is often held up as a model for the AB, even by scholars whose better judgment counsels against those comparisons.205 To realize this vision, even in a softer form, the AB would have to play the key role through its interpretation of WTO obligations in the context of specific disputes. Many constitutional visions rest heavily on an adjudicatory body invalidating actions of legislatures or subordinate polities that are inconsistent with the constitution. In the WTO there simply are no other candidates for such a policing role. The dispute settlement bodies are the only part of the WTO that operates continuously and effectively. The quasi-legislative, rule-making function is the province of traditional negotiations among all 132 Member states of the WTO-traditional in that no state may be bound by an agreement to which it has not consented. New rules, or modifications to old rules, are created only in the massive "rounds" of trade negotiations that occur no more than once a decade, when increasingly complicated packages of agreements are finally (and painfully) pulled together.206 Occasional proposals to establish some form of interim rule-making, even to "elaborate" or "implement" existing rules, have thus far come to naught. The "executive" of the WTO-the Managing Director and a relatively small secretariat-are not among the more influential staffs of international organizations and play only a facilitating role in trying to get member states to act. Since the conclusion of the Uruguay Round, however, the dispute settlement bodies are not only continuously active; they also produce binding outcomes. Most of the time, then, the real action at the WTO is to be found mainly in the panels and AB.

In these circumstances, the attraction of a more assertive role for the AB must be strong. There is some limited evidence for the proposition that the AB has indeed assumed such a role. In the Shrimp-Turtle case discussed earlier, the AB provided a rationale for a "dynamic" interpretation of GATT obligations that allowed it to redefine a clause drafted in the 1940s in light of the changed circumstances of the intervening half century-specifically, the emergence of ecological issues as a subject of international arrangements. This potentially seminal decision by the AB could provide the WTO with an adaptive capacity between rounds of trade negotiations. By applying this reasoning in support of an environmentally-motivated trade restriction, the AB avoided serious controversy, but may now have set a precedent for applying similar reasoning in ways that invalidate national actions. It is also true that, because of the difficulty of achieving unanimity among member states, "overruling" the dispute settlement bodies is very difficult.

The practice of national agencies administering trade statutes in a way perceived as protectionist is a particularly inviting target for proponents of a strong WTO constitution and a more activist AB possessed of constitutional aspirations. In fact, the AB has applied an interpretive approach in anti-dumping cases that is consonant with a robust trade constitution. In a series of decisions in different subject areas, the AB has taken pains to establish that an overall duty of "good faith" attaches to all actions by WTO members in their trade-related dealings with one another.207 The good faith principle is well-established in traditional international law, including the injunction in Article 26 of the Vienna Convention that "[e]very treaty in force is binding upon the parties to it and must be performed by them in good faith."208 Thus the AB is not breaking new doctrinal ground. Needless to say, however, the concept is one with potentially far-reaching implications for the prerogatives of decision-makers.209

In Japan Steel, the AB concluded that the "organic principle of good faith" operated to restrain national administrative authorities "from imposing burdens on exporters which, in the circumstances, are not reasonable."210 Thus, even where there was no provision of the Anti-Dumping Agreement specifically limiting the demands that national authorities may place on responding companies, the AB inferred such a limit from the good faith principle. This principle, which may presumably be invoked at the convenience of the AB, might be justified as necessary to ensure the integrity of the WTO. It also, obviously, gives the AB great scope for shaping WTO obligations.

The good faith principle might provide the basis for a more far-reaching interpretive approach to dumping cases: Administrative interpretation of national laws or the Anti-Dumping Agreement intended to increase protection for domestic industries may be characterized for this reason alone as not made in "good faith." Accordingly, such interpretations would not be entitled to deference from WTO dispute settlement bodies. For this reason-rather than because the Vienna Convention provides a roadmap to determinate interpretations of most treaty provisions-the second sentence of Article 17.6(ii) might rarely, if ever, be applicable: "permissible" interpretations may be limited to those made in good faith, which is not present when administering authorities make interpretations with the aim of finding dumping or higher dumping margins.

C. Normative Arguments Against Disregarding 17.6(ii)

It may be that the AB underappreciates its ability to overturn national administrative determinations while applying 17.6(ii). Even if true, this possibility begs the question of why the AB wants to minimize the impact of that provision. The more plausible answers to this question, not mutually inconsistent, are the last two mentioned in the preceding discussion-hostility towards national administration of trade remedies and aspiration to a more powerful constitutional role in the world trading system. We may not be able to say with assurance which explanations are correct as a positive matter. As noted already, however, these explanations are congruent with important normative positions that have previously been staked out in the ongoing debate over the role of the WTO in general and the dispute settlement system in particular. Misgivings about the AB's approach may also be derived from this ongoing debate.

1. Sovereignty

An argument based on sovereignty holds that member states of the WTO have relinquished their sovereignty only to the narrow extent of agreeing to the specific rules included in the WTO agreements. Thus, whenever dispute settlement bodies interpret a WTO agreement to impose obligations on member states that are not clearly set forth in the agreement, they encroach upon national sovereignty. This argument, which has been made during discussions at some professional conferences in connection with WTO review of anti-dumping cases, is a kind of originalism transposed to an international context. One glaring problem for this position is that the member states of the WTO have agreed not only to the specific rules in the agreements, but to the dispute settlement process that decides disagreements between member states. The degree to which they have relinquished their sovereignty thus depends on the scope of the dispute settlement system. This issue returns us to the question of what Article 11 of the DSU and Article 17.6 of the Anti-Dumping Agreement mean. Formalistic concepts of surrendered sovereignty cannot provide the answer.

2. Legitimacy

A second argument is that WTO dispute settlement procedures lack "legitimacy" and should therefore be characterized by what, in an American context, is known as judicial self-restraint. The WTO AB and panel members are not elected or otherwise accountable to the citizens of the member states. When a panel finds a national action unjustified under WTO rules, it is displacing the result reached through the "legitimate" processes of a national system (at least where that system is workably democratic). Moreover, the dispute settlement processes of the WTO take place largely behind closed doors, reinforcing the image of an unaccountable, closeted group of (mostly or totally) foreign lawyers deciding key issues of public policy.

Those who make arguments based on legitimacy provide a counterpoint to the "constitutional" positions described earlier.211 The increasing reach of WTO agreements into regulatory and other internal affairs of member states has led to a situation in which "there is no real consensus among WTO members on many of the complex regulatory issues" that will be raised in the dispute settlement system.212 Thus the efforts of a "judiciary" to resolve these issues will, particularly in the absence of an effective "legislating" body, be substantively and politically unacceptable. A somewhat different, but complementary, position holds that the conflicts between trade and other values emphasizes the dangers of reifying decisions resolving such conflicts, a tendency that is certain to restrict "democratic contestability."213 To constitutionalize the WTO is to change (unwisely) its character as "a complex, messy negotiated bargain of diverse rules, principles, and norms into a single structure."214 From this perspective, the practices of the European Union that are most appropriate for adoption by the WTO are not those ECJ decisions which moved the European treaty system closer to a federal system, but those which defer to national governments based on the principle of subsidiarity.215

The anti-constitutionalists have not spoken specifically to the Article 17.6(ii) decisions discussed in this article.216 And, insofar as anti-dumping does not implicate the delicate balancing of domestic regulatory programs with trade principles that they identify as particularly problematic, most anti-constitutionalists might not be overly troubled by the Article 17.6(ii) decisions. However, one need not extrapolate too dramatically from their position to fashion the argument that Article 17.6(ii) is itself an explicit invocation of subsidiarity principles. Moreover, the trade laws governed by the Anti-Dumping Agreement themselves reflect a careful balance of political forces, mostly in democracies, which will be superseded by an activist AB.

3. Respecting the Terms of Negotiated Agreements

This last point suggests a third argument-an application to the AB, as an international instrumentality, of the maxim pacta sunt servanda,217 which is conventionally identified as a core principle of international law applicable to states. The argument runs as follows: Whatever the AB may think of the anti-dumping laws or the protectionist instincts of national bureaucracies, the member states of the WTO agreed to Article 17.6(ii). In so doing, they opted to allow those national administrators considerable leeway to interpret member state obligations under the Anti-Dumping Agreement. Accordingly, the AB should respect the outcome of negotiations by the Member States and apply 17.6(ii), even if the AB believes the provision to be ill-advised.218

This argument recalls the Supreme Court's own treatment of the Chevron doctrine, which in recent years appears to have been grounded more in Congressional intent than in presumptions of agency expertise or direct democratic accountability:

We accord deference to agencies under Chevron . . . because of a presumption that Congress, when it left ambiguity in a statute meant for implementation by an agency, understood that the ambiguity would be resolved, first and foremost, by the agency, and desired the agency ... to possess whatever degree of discretion the ambiguity allows.219

Thus the reason for deference under Article 17.6(ii) is not necessarily that national authorities are more expert in trade law than panels and the AB, and certainly not that national authorities are more responsive to all relevant actors. The reason for deference is simply that the member states have instructed the panels to give deference.220 Moreover, this feature of the Anti-Dumping Agreement is but one piece of a carefully balanced package of concessions among all the members of the WTO. To ignore or willfully de-emphasize this provision would be to upset the reasonable expectations of one or more member states and, thereby, to undermine the principle of reciprocal benefits that has been fundamental to the world trading system since 1947.

IV. TRADE LAW THROUGH A GEOPOLITICAL LENS

The foregoing analysis of rationales for, and arguments against, the AB's treatment of Article 17.6(ii) is based upon more or less neutral principles. In this kind of analysis, protectionism, sovereignty, and rent-seeking are presumptively of normative relevance regardless of who is practicing protection, losing sovereignty, or seeking rents. The reality of international relations renders this kind of analysis at best incomplete, and perhaps misguided.

Examination of AB practice through a geopolitical lens suggests an additional explanation of, and justification for, that practice-resistance to the hegemonic imposition of the terms of an international agreement. In the discussion to follow, I conclude that neither a positive nor a normative case can be made for this account, at least on the basis of the simple fact of U.S. power in the WTO. This does not mean, however, that geopolitical considerations are irrelevant. Even more directly than in most domestic polities, the relative power of actors determines their influence in the development of governance arrangements and the resolution of differences. The differences between national and international law-making assure a more direct transmission of an actor's response from one law-making venue to another. In this context, the dynamic effects of AB activism upon U.S. behavior may result in costs to the world trading system that exceed the putative benefits for liberal trade when the WTO dispute settlement process overturns national trade law actions.

A. Resistance to Hegemony

As is apparent from the negotiating history and from the cases themselves, the United States is clearly the featured player in the story of Article 17.6(ii). This provision is the outcome of a demand by the United States for a generous standard of review for administrative interpretations of legal obligations. The U.S. proposal was based on a hornbook statement of the Chevron doctrine as developed by the Supreme Court. A disproportionate number of challenges to legal interpretations by national administrators involve U.S. anti-dumping actions.

It is unremarkable that a member state should be involved in disputes over the interpretation of a provision it had demanded in a previous negotiation. When the protagonist is the United States, however, the disputes may assume a distinctive character because of the country's unique international position. U.S. power to shape the world trading system has, since the inception of the GATT in 1947, been simultaneously relied upon for progress and resented by much of the rest of the world. In more recent years, the relative degree of resentment has increased, perhaps reinforced by a more generalized anxiety in the face of the apparently untrammeled U.S. power of the post-Cold War era. Charges of U.S. unilateralism, familiar to trade policy analysts, have spread to other areas. Given the depth and pervasiveness of concerns, one must ask whether a complete understanding of a discrete issue, such as WTO review of national anti-dumping actions, requires examination of the issue in geopolitical terms.

Legal scholarship does not often invoke blunt geopolitical explanations for international legal developments, and for good reason. "Power" can be invoked tautologically to explain virtually any outcome and to overlook other factors. However, dismissing such explanations in favor of purely jurisprudential or ideological accounts of international dispute settlement risks missing a fundamental contemporary fact of life in all international organizations. There are two salient inquiries: Are the AB's anti-dumping decisions informed by a sense of confrontation with the hegemonic power of the United States? And, even if this first question cannot be answered, is the normative assessment of the AB's practice changed by the geopolitical context in which it arises?

As to the first inquiry, there is no direct evidence that the AB's disregard of Article 17.6(ii) derives from an attempt specifically to contain U.S. assertiveness in its administration of trade laws or U.S. efforts to make its jurisprudential standards applicable internationally. The AB has not made any statements-even obliquely-directed against the United States as such (e.g., "even the largest members are required to observe the obligations set forth in the Agreement"). Moreover, the fact that other countries have also fared badly in WTO disputes involving administration of their trade laws shows that American trade law has not been singled out for unsympathetic review.

It is possible that the AB is inclined to a dismissive treatment of 17.6(ii) precisely because it was inserted at the behest of the United States to preserve the discretion to use the very law so roundly criticized by many other countries. In denying discretion to interpret WTO obligations to all anti-dumping administrators, the AB may believe itself to be resisting the relationship between national and international law favored by the United States. Again, though, there is no specific evidence for this hypothesis. Additionally, the United States has fared well in seeking enforcement of other provisions included in WTO agreements at its behest.221 Thus, while resistance to hegemonic preferences is a factor of increasing relevance in analyzing actions of international institutions, it does not seem a satisfactory explanation here and is certainly not as convincing as the hypothesis of AB antipathy to trade law remedies.222

Even if the motives for past AB practice remain obscured, one may ask how future decisions might reveal whether America's hegemonic status affects AB practice and thereby help pin down an explanation. One possibility is that examples of resistance to U.S.-inspired rules or procedures will extend to other areas.223 If so, the nature of the constitutional struggle in the WTO may be clarified. But the failure of this extension to materialize would not disprove the hypothesis, since the AB may be wary of picking too many fights. Another possibility is that AB decisions finding U.S. anti-dumping actions inconsistent with WTO obligations will result in increasingly acrimonious discussion in the Dispute Settlement Body, during which the dynamic of a U.S. confrontation with the rest of the world may become obvious. But this dynamic could arise whatever the AB's actual motivation for its decision, so long as the members of the DSB perceived the conflict as an occasion for trying to contain the United States.

Thus it may be impossible to discern the influence of geopolitical considerations in AB decisions, either now or in the future. Regardless of the limits on a positive inquiry, however, the implications for normative analysis are substantial. Indeed, each of the major arguments for and against the AB's marginalization of Article 17.6(ii) is altered if (and this is clearly a big "if") U.S. dominance of the trading system is stipulated.

Because the United States arguably uses its administrative trade law remedies more extensively than any other WTO member, heightened AB scrutiny of these administrative processes may be seen as a particularly effective means of controlling the hegemon's protectionist impulses. The rationale based on constitutionalism is largely transformed by introduction of the geopolitical perspective. The argument becomes less one for universalizing binding norms of substantive law and more one for containing the exercise of hegemonic power. Seen through this lens, Article 17.6(ii) is one element of a WTO rules-system that disproportionately reflects U.S. negotiating aims, an outcome attributable to a combination of U.S. economic power and of influence exercised through non-economic channels. The constitutional task is then to use institutional prerogatives, notably the final authority of the dispute settlement system, to offset at least part of the hegemon's power to shape the outcome of negotiations. By this reasoning, the United States may be able to exploit its power in the fluidity of a negotiating end-game to foist a version of its preferred standard of review on the rest of the world, but the AB can rebalance the negotiated outcome by essentially ignoring it.224

The geopolitical perspective also changes normative arguments against the AB disregard of 17.6(ii). For example, the pacta sunt servanda argument loses some force if the agreement in question has been concluded at least in part through coercion. Suppose that the negotiation of many provisions, including 17.6(ii), was the result of the same distortion of the domestic political process that is thought to produce protectionist actions. Why should the AB, or anyone else for that matter (other than the groups which benefit), respect this outcome? Or suppose that the entire Uruguay Round was in some sense a contract of adhesion imposed by the United States (and possibly the European Union), leaving many developing countries with the Hobson's choice of acceding to an unsatisfactory package of agreements or being left out of the trading system altogether. Might there not be a countervailing argument for "judicial activism" to redress these shortcomings of the negotiating process?

Similarly, arguments based upon the sovereignty of the United States-already a bit weak, as explained earlier-look weaker still from the vantage point of countries that believe their own range of sovereign prerogative has long since been diminished by U.S. power to shape global institutions. The argument for subsidiarity over constitutionalism is also affected, since the case for deferring to a nation's own judgments about competing values becomes much less compelling when one believes that this nation has used its power to impose its judgments on the rest of the world.

Though the normative arguments against AB disregard of 17.6(ii) are weakened if the fact of U.S. hegemony is stipulated, they are hardly refuted. To disregard a rule for dispute settlement that was plainly negotiated, and at least formally agreed, is to call into question the entire basis for positive international law. Is every international rule negotiated at the request of a powerful country a less privileged form of international law for that reason alone? If not, what principles distinguish among such rules and who applies those principles? The implications of these questions for international law are obviously far-reaching. Not surprisingly, one's end position on this normative divide depends substantially on one's starting premises. Free traders and those suspicious of American influence will incline towards support for the AB practice, while representatives of import-sensitive industries, those suspicious of international intrusions on national politics, and maybe even those worried about the integrity of international legal principles will incline towards opposition. There is, of course, nothing unusual or misdirected about a debate whose resolution depends on first premises, so long as the premises and their implications have been fully explored. The next section adds something a bit less predictable-an argument why those favoring liberal trade, and perhaps even those wary of U.S. power, should oppose AB activism in some circumstances. This argument is all the more interesting because it is based on the very asymmetry of power that seemed just a moment ago to support this activism.

B. The Pragmatics of Asymmetrical Power

Even those who dismiss the normative arguments against AB activism might conclude on pragmatic grounds that Article 17.6(ii) should be applied rather than disregarded, and perhaps applied in a manner not too different from that expected by the Clinton Administration in its statements to Congress.225 The reason is that, whatever the policy or moral merits of applying 17.6(ii), failure to implement it may have negative dynamic effects that exceed the benefits to free trade and hegemonic containment that AB activism can achieve. These potentially negative dynamic effects derive from, and underscore, two important features of the legal regime for international trade.

First, states are the only legal actors in the WTO system. Hence the parties who "litigate" in the WTO dispute settlement system are all members of the "legislative" body that makes WTO rules. Thus outcomes in one arena are more likely to be directly reflected in the other. In a domestic setting, particularly one with divided constitutional powers, litigants are almost always private actors or the executive branch of government, and rarely legislators. A reversal in court may, to be sure, motivate the losing party to seek or oppose legislation to offset the effects of the judicial decision. But the losing party must work through sympathetic legislators, rather than directly. Although legislators beholden to powerful groups will often work closely with them, these legislators will rarely share an identity of interests with the groups. Politicians usually have other interests-constituency, party, and policy-that at least on the margin will lead them to strategies and tactics that differ from those of the losing party. By contrast, a losing party in the WTO is itself a "legislator" in WTO negotiations. Although a government may not have quite as consistent a perception of its interests as an individual, or even a corporation,226 the same configuration of political forces that produced domestic anti-dumping laws will be at work.

Second, the WTO "legislators" must agree unanimously to new rules. In theory, amendments to existing WTO agreements may be approved through supermajority votes in the Ministerial Conference if a consensus cannot be achieved.227 In practice, amendments to the WTO (and the GATT before it) are just not made, whether through consensus or voting. Thus if a losing party is sufficiently exercised with an AB decision that it is willing to invest the energy needed to change the underlying rules, it must pursue this aim during the next round of multilateral negotiations. As a formal matter, new agreements are not binding on any member that does not agree to them.

At first glance, these characteristics of the international trading system suggest net benefits to the WTO members that have prevailed in the cases discussed in Part II. By disregarding Article 17.6(ii), the AB has effectively revised the Uruguay Round Anti-Dumping Agreement. If the United States still wants the protections for its anti-dumping law administration that it thought it was getting in 1995, it will have to negotiate again for these protections in the new round of negotiations. This will require either foregoing some other negotiating aim or offering additional concessions to other countries. It looks, then, as if the United States must "pay" twice to obtain discretion to choose among "permissible" legal interpretations of its international obligations. The rest of the world looks likely to benefit.228 On the other hand, if the United States decides it is unwilling to take less or give more to obtain the Article 17.6(ii) discretion, and seeks the concessions it would have pursued absent the AB practice on 17.6(ii), the AB practice will stand. Members that challenge U.S. anti-dumping administration will have obtained a concession without paying for it.

The account in the preceding paragraph tells only half the story, however. The analysis of static effects must be complemented with an assessment of dynamic effects. Suppose first that the AB's disregard of 17.6(ii) is read as hostility specific to national administration of trade laws. The United States may conclude that there is no point in negotiating more specific language on the relevant standard of review, because the AB will again disregard the provision. As a result, the United States will not seek such a change, even if it attaches greater value to having a 17.6 standard of review than to some of its other negotiating aims. The United States might have been willing to "pay" more for such a standard than these other aims. Other countries, in turn, may assess the costs of these other aims (i.e., U.S. requests for concessions from other countries) as equal to or higher than a 17.6 standard. These countries would, in exchange for a particular concession to the United States, rather have given an effective 17.6 standard of review. Hence a deal with maximum reciprocal benefits-as valued by the negotiating parties-may be foreclosed.

The impact of the AB practice may not be limited to potential renegotiation of Article 17.6. Having seen the AB nullify a provision that was supposed to protect national prerogatives, the United States may conclude that the "costs" of negotiating further restrictions on the use of its trade laws will increase as the WTO dispute settlement system consistently decides cases against the importing nation. That is, if the United States expects the AB to construe broadly the constraints imposed by WTO agreements on national administration of import laws, it will assess the costs of these constraints as higher than if 17.6 were "honestly" applied. To the degree that other governments desire change in national administration of trade remedy laws, they will then have to "pay" more during trade negotiations to obtain agreement on those changes. Conceivably, the other governments will value the expected extension of the constraints by the AB less than the additional costs which the United States anticipates bearing. Again, the United States and its trading partners will be unable to reach an optimal trade deal, in which each country bargains for the maximum concessions from others for the minimum "price" it has to pay.

Note the relationship between AB practice and the potential for sub-optimal trade negotiations. By significantly "changing" the rules from what the United States expected it had negotiated, the AB has denied WTO members the ability to specify obligations in the way that maximizes the benefits each can obtain. Other WTO members cannot "give" the United States discretion to choose its own permissible interpretation of Anti-Dumping Agreement provisions, because the United States believes the AB will interpret that discretion away. To borrow terminology from contract law analysis, the AB has converted the standard of review in anti-dumping disputes from a "default rule" (one which the AB would apply in the absence of a negotiated rule) to an "immutable rule" (one which the parties cannot change through negotiations).229 The AB has, in effect, foreclosed a deal that the negotiating parties might have thought worth concluding.230

The United States and its negotiating partners do not know how activist the AB will be in interpreting new import law obligations. Their uncertainty introduces more distortions into the negotiating process. All sides may resolve the uncertainty by assuming the worst from their perspectives, because no government wants to end up gaining less than it expected from the trade agreement. Certainly the Commerce Department, members of Congress, and import-sensitive industries will expect the AB to be aggressive across the board in interpreting constraints on national trade remedy laws. Thus the United States, and perhaps other countries that are frequent users of trade laws, will assess the costs of new commitments on the use of those laws as higher than the literal terms of the new obligations would suggest. On the other hand, countries whose exports are more frequently subject to trade laws may assume that the AB will not extend its activism to all new import law obligations. Thus the negotiating parties may differ as to what obligations they are actually negotiating and as to the value of those obligations. Agreement will be even more difficult to achieve.

Suppose now that the AB's disregard of 17.6(ii) is read as part of a broader effort by the AB to establish the WTO as a potent constitutional regime. In that circumstance, the inhibitions upon negotiating new agreements may spread beyond the area of trade remedies. The "costs" of trade agreements generally will have become much harder for governments to gauge, since creative interpretation by the AB may effectively enlarge or diminish the scope of all WTO obligations. The AB's emphasis upon the obligation of Member States to act in "good faith" gives some idea of the potential scope of such interpretation. Consequently, the anticipated benefits of trade agreements will have to be significantly higher, in order to provide a premium for the uncertainty attaching to the obligations being assumed. Again, optimal trade deals that might otherwise be negotiated will not be concluded. If the uncertainty spreads broadly enough across issue areas, it is even possible that no deal at all will be reached.

Recalling contract law analysis again, bargains between contracting parties are not always Pareto efficient. Immutable contract rules are generally justified precisely for this reason, on grounds of either parentalism or third-party effects.231 Whatever the merits of these rationales in a domestic contract law context, they seem inapplicable within the institutional configuration at issue here. It is hard to see the argument for the AB protecting the WTO members from themselves by effectively prohibiting negotiation of a generous standard of review for national trade law actions. The governments of WTO members are presumptively much more competent than the AB to decide whether acceptance of such a standard is worthwhile in a particular negotiating context.

One might argue that, if the AB raises the perceived costs of future concessions enough, and thereby precludes negotiation of a new agreement, then less influential countries will be spared the choice of signing an agreement they dislike or being left out of the world trading system. This argument is also unavailing in the current context. The AB decisions that would prompt a hard-nosed U.S. position would have accumulated before decisions on the outlines of a new agreement are ripe. The AB will not have the information necessary to make an informed judgment on whether a new round will produce the international trade equivalent of a contract of adhesion. Moreover, the very idea of the AB acting so as knowingly to scuttle a new round of negotiations seems far-fetched, to say the least.

There is certainly a good argument that negotiations among trade ministers of WTO members produce negative externalities for non-participants. The tenets of political economy discussed earlier suggest that certain groups within each WTO member are disproportionately influential in setting the trade agenda. One should expect that the results of trade rounds will be disproportionately weighted towards the aims of these groups, and that the interests of some citizens will have been discounted or ignored. One might even apply a rudimentary public choice analysis and suggest that the outcome of a trade negotiation maximizes the political welfare of governments in power within WTO members, rather than global or national economic welfare. Interesting as pursuit of these points might be, they are essentially irrelevant to the question of whether the WTO AB should effectively remove certain subjects from the multilateral negotiating table. The picture of seven unelected former national officials, often framed as a caricature, seems appropriate here. One is hard-pressed to articulate the governance principles by which a panel of three members of this group should decide when the unrepresented interests of unorganized consumers should be invoked to overturn or disregard a negotiated provision of a trade agreement.

Although U.S. experience with Article 17.6(ii) has been the starting point for analysis of the dynamic costs of AB practice, the discussion to this point is essentially applicable to all countries: To the degree AB decisions either remove the ability of WTO members to grant specific concessions or introduce significant uncertainty into the breadth of concessions that may be granted, new trade agreements will be harder to conclude. The size of the hurdle, though, may depend substantially on the size of the WTO members whose negotiating aims are most directly affected. Here is where asymmetrical power and influence matter. Practically speaking, a smaller state may not be able to negotiate changes, or refuse changes it dislikes. It may then be faced with the choice of either signing on to the new agreements anyway or risk being left behind by the world trading system.

This was quite literally true in the Uruguay Round, which substituted the "GATT 1994" for the original GATT and thus ended the obligations of GATT members under the original agreement. Had a dissident state chosen not to accept the whole package of agreements concluded in the Uruguay Round, it would have been left with no multilateral trade rights. Of course, the more influential the state, the less likely it is to be left behind, because a "world" trade organization without a major trading country will be problematic. In the case of the United States (or the European Union), which accounts for more than a fifth of the world's gross domestic product, concurrence in the results of trade negotiations is essential.232 Hence, significant U.S. aims or objections can dramatically affect the shape and prospects of negotiations. If the uncertainty in U.S. calculations because of AB interpretive practice is great enough, for example, the entire scope of the negotiations may be affected, as the United States demands more for changes in the Anti-Dumping or Safeguards agreements.233

It now seems clear that there will be potentially significant costs arising from the dynamic effects of AB activism. To the degree that AB practice obstructs the exchange of maximally valued concessions, there will be the international negotiating equivalent of "deadweight" losses.234 Quantifying these losses is obviously impossible. Even describing them with precision, particularly in advance, is difficult. However, measurement constraints do not justify disregarding the dynamic. We can at least identify some parameters relevant to an estimate of these costs.

First, there will be some "redistribution" effected by the AB practice. While the facts of gains and losses for variously situated parties are clear, the magnitudes will have to be educated guesses. Everyone loses in dynamic terms. We can say with fair assurance that the centrality of a hegemonic nation raises the chances of significant global dynamic losses. Without the United States, no round will be completed. To the degree that AB activism and attendant uncertainty have significantly raised the perceived costs of U.S. concessions beyond what they are worth to other countries, the United States will need proportionately greater concessions from others or will substitute proportionately less valuable concessions of its own. The United States obviously loses in static terms. States that only moderately value constraints on U.S. trade law administration gain little in static terms. States that greatly value constraints upon U.S. trade law administration gain considerably in static terms but lose in dynamic terms. We cannot say in the abstract whether, even for these countries, their static gains will exceed their dynamic losses. Moreover, given that Japan would be one of the principal static gainers, the "redistributive" effects of the AB practice are not just from richer to poorer countries.

Second, unlike many private contractual situations, it will be difficult as a practical matter for WTO members to minimize the dynamic losses by providing contingencies for unfavorable legal interpretations. Two companies that prefer a particular arbitration clause can include the clause, which reflects part of their optimal deal, but also include a clause reallocating rights and obligations in the event the arbitration clause is held invalid. If the arbitration clause is never invoked or held valid, then the parties' first-best deal will have been realized. This promising possibility is not readily transferable to the WTO context. With over 130 members, providing a contingent WTO agreement would require a significant additional negotiating effort in what is already a difficult, prolonged exercise. Indeed, to the degree that uncertainty about AB intentions spreads from one provision to whole agreements, it may be impossible as a practical matter to provide for every contingency. Furthermore, unlike judicial decisions on the validity of an arbitration clause, application of a standard of review may be ambiguous. Indeed, I am certain that some readers will insist that the AB has, in fact, "applied" Article 17.6(ii), regardless of whether that standard of review has determined any outcomes.

The redistributive elements of AB activism may not outweigh the dynamic losses even from the perspective of those sympathetic to countries whose exports are most regularly subjected to U.S. trade remedies. Possibilities for contracting around AB decisions, so as to minimize the dynamic losses, are limited. The hidden costs of AB activism in reviewing national administration of trade laws-in the form of impediments to subsequent negotiating rounds-may well be significant, if hard to specify.235

V. CONCLUSION

This Article began with three questions: How has the WTO Appellate Body handled reviews of administrative action under the Article 17.6(ii) standard? Why has the AB taken this position? Is the AB practice desirable? My analysis offers a confident answer to the first question and a speculative answer to the second. The AB has essentially disregarded the 17.6(ii) standard. Its actions seem most consistent with an aversion to national import-restraining statutes such as the anti-dumping law, shaped perhaps by constitutional aspirations for the world trading system.

My answer to the third question is actually less a direct response than a basic template to use in developing an answer. Use of this template requires a reader to select her own normative preferences. Part III showed that there are serious normative principles that can be invoked in support of both sides of a debate on disregarding Article 17.6(ii). My purpose here has not been to urge my own normative preferences, but to suggest theories, which yield solid arguments on the issue at hand, and principles, which even if stipulated for purposes of argument are less relevant for deciding how to apply 17.6(ii).

The analysis in Part IV established that the answer to the third question does not depend solely on one's core normative preferences. The characteristics of the international trading system, including the geopolitical reality of the United States as a dominant actor, mean that AB practice will impose deadweight costs on the system, particularly in subsequent negotiations. These dynamic effects will likely be undesirable from the standpoint of the free trade advocates most likely to hold the normative principles that would otherwise support the AB disregard for 17.6(ii). Part IV filled out the template for answering my third question by identifying the empirical data necessary for determining the likely magnitude of the deadweight loss: the interpretation of the AB action in the United States, the relative strength of the domestic political support for 17.6(ii), the expectations of the United States and other states as to future AB activism, and the reconfiguration of negotiating aims in light of the other factors.

The degree of speculation involved in each of these inquiries makes any notion of "calculating" deadweight loss at best quixotic. The analysis generates orders of magnitude, not precision. What is most important is to keep the dynamic effects in mind. A free trade advocate mindful of these dynamic effects would surely have counseled the AB to be less blatant in its disregard of Article 17.6. A series of decisions that duly analyzed the importing country's interpretation and explained why it was unreasonable would have helped. Even one decision in which the AB explicitly deferred to an interpretation it thought less desirable but "permissible" would have significantly reduced expectations concerning future activism.

Legal institutions exist in specific political circumstances. A court in a domestic polity has a different scope for action than the WTO "court" in a trading system defined by nation-states and dominated by the United States. Courts are, in this sense, political institutions. The more successful see opportunities to advance the principles to which they subscribe but also understand their political limits. This suggests the need for caution in transposing specific institutional features from domestic to international settings because these limits differ so dramatically. Those-ironically including successive U.S. administrations-who argued for a binding WTO dispute resolution procedure may have underappreciated the differences. At times, it appears, so has the Appellate Body.

1. In formal terms, a panel report must be "adopted" by the Dispute Settlement Body of the WTO. See Understanding on Rules and Procedures Governing the Settlement of Disputes, Apr. 15, 1994, art. 2.1, Marrakesh Agreement Establishing the World Trade Organization [hereinafter WTO Agreement], Annex 2, in RESULTS OF THE URUGUAY ROUND OF MULTILATERAL TRADE NEGOTIATIONS 404 (1994) [hereinafter RESULTS OF THE URUGUAY ROUND], 33 I.L.M. 1226, 1226 [hereinafter DSU]. However, in a change from GATT practice prior to the Uruguay Round, the Dispute Settlement Body must adopt the report unless there is a consensus against adoption (or the report is appealed). DSU, supra, art. 16.4, at 417, 33 I.L.M. at 1235. In other words, the prevailing party must agree to a decision not to adopt a report giving it a victory. Thus, adoption is close to automatic.

2. See Agreement on Implementation of Article VI of the General Agreement on Tariffs and Trade 1994, Apr. 15, 1994, WTO Agreement, Annex 1A, in RESULTS OF THE URUGUAY ROUND, supra note 1, at 168 [hereinafter Anti-Dumping Agreement].

3. The original Antidumping Code was concluded as part of the Kennedy Round of multilateral trade negotiations, which ended in 1967. See Agreement on Implementation of Article VI of the General Agreement on Tariffs and Trade (June 30, 1967), 651 U.N.T.S. 320 (1968).

4. During the most recent period for which WTO statistics are available, India was the most frequent user of anti-dumping laws, followed by the United States and Argentina. See Press Release by World Trade Organization, Press/287, WTO Members Report on Anti-Dumping Activity (Apr. 22, 2002).

5. See Anti-Dumping Agreement, supra note 2, art. 17.6, at 193.

6. 19 U.S.C. [sec] 3538 (2001). Canada challenged [sec] 3538 in the WTO because it requires the U.S. administering authorities to continue to apply the trade remedy even after a WTO ruling, until the USTR has formally requested a reversal. A WTO panel ruled against Canada. See WTO Panel Report, United States-Section 129(c)(1) of the Uruguay Round Agreements Act, WT/DS221/R, [para] 7.1 (July 15, 2002).

7. Anti-Dumping Agreement, supra note 2, art. 18.3, at 193. The negotiating parties built some lag into the Agreement because investigations and reviews typically take about a year.

8. According to WTO figures, 157 investigations were initiated in 1995, by far the lowest number since the Uruguay Round was completed. The second lowest number was 224, in 1996. See World Trade Organization, Anti-Dumping Initiations by Importing Country from 01/01/95 to 30/06/02, at www.wto.org/english/tratop_e/adp_e/adp_stattab2_e.htm (last visited Jan. 22, 2003).

9. Each of these complaints resulted in determinations in the dispute settlement process. See WTO Appellate Body Report, Guatemala-Anti-Dumping Investigation Regarding Portland Cement from Mexico, WT/DS60/AB/R (Nov. 2, 1998) [hereinafter Guatemala Cement I]; WTO Panel Report, Mexico-Anti-Dumping Investigation of High Fructose Corn Syrup (HFCS) from the United States, WT/DS132/R (Jan. 28, 2000); WTO Panel Report, United States-Anti-Dumping Duty on Dynamic Random Access Memory Semiconductors (DRAMS) of One Megabit or Above from Korea, WT/DS99/R (Jan. 29, 1999) [hereinafter United States-Korean DRAMS Panel Report].

10. Again, each of these complaints has now been decided on the merits. See WTO Appellate Body Report, Thailand-Anti-Dumping Duties on Angles, Shapes and Sections of Iron or Non-Alloy Steel and H-Beams from Poland, WT/DS122/AB/R (Mar. 12, 2001) [hereinafter Thailand-Steel from Poland Appellate Body Report]; WTO Appellate Body Report, European Communities-Anti-Dumping Duties on Imports of Cotton-Type Bed Linen from India, WT/DS141/AB/R (Mar. 1, 2001); WTO Panel Report, United States-Anti-Dumping Measures on Stainless Steel Plate in Coils and Stainless Steel Sheet and Strip from Korea, WT/DS179/R (Dec. 22, 2000); WTO Panel Report, Guatemala-Definitive Anti-Dumping Measures on Grey Portland Cement from Mexico, WT/DS156/R (Oct. 24, 2000).

11. In addition to the cases cited in the preceding two footnotes, final decisions have also been issued in four other cases. See WTO Appellate Body Report, United States-Antidumping Measures on Certain Hot-Rolled Steel Products from Japan, WT/DS184/AB/R (July 24, 2001) [hereinafter United States-Japan Steel Appellate Body Report]; WTO Panel Report, Egypt-Definitive Anti-Dumping Measures on Steel Rebar from Turkey, WT/DS211/R (Aug. 8, 2002) [hereinafter Egypt-Steel Rebar from Turkey Panel Report]; WTO Panel Report, United States-Anti-Dumping and Countervailing Measures on Steel Plate from India, WT/DS206/R (June 28, 2002); WTO Panel Report, Argentina-Definitive Anti-Dumping Measures on Imports of Ceramic Floor Tiles from Italy, WT/DS189/R (Sept.28, 2001).

12. WTO Note by the Secretariat, United States-Sunset Review of Anti-Dumping Duties on Corrosion-Resistant Carbon Steel Flat Products from Japan, WT/DS244/5 (July 22, 2002); WTO Note by the Secretariat, Argentina-Definitive Anti-Dumping Duties on Poultry from Brazil, WT/DS241/4 (July 5, 2002); WTO Request for Consultations by Brazil, European Communities-Anti-dumping duties on Malleable Cast Iron Tube or Pipe Fittings from Brazil, WT/DS219/1 (Jan. 9, 2001).

13. As a matter of fact, most of these cases have resulted in revocation or significant modification of the dumping finding. See generally WTO, Update of WTO Dispute Settlement Cases, WT/DS/OV/6 (May 3, 2002). Not surprisingly, perhaps, this outcome has been achieved more quickly in some cases than others, and disputes are still pending in a couple of instances. Still, the WTO dispute settlement process has clearly mattered. There is not a pattern of national authorities simply substituting one decision for the WTO-inconsistent decision in order to maintain the dumping finding. This was the initial outcome in the Thailand-Steel from Poland case, but the parties subsequently settled the case in an apparently mutually agreeable fashion. Id. at 125.

14. Anti-Dumping Agreement, supra note 2, art. 17.6(ii), at 193.

15. As to factual matters, Article 17.6(i) requires that "the panel shall determine whether the authorities' establishment of the facts was proper and whether their evaluation of those facts was unbiased and objective. If the establishment of the facts was proper and the evaluation was unbiased and objective, even though the panel might have reached a different conclusion, the evaluation shall not be overturned." Id. art. 17.6(i), at 193.

16. DSU, supra note 1, art. 11, at 413, 33 I.L.M. at 1233.

17. Id. art. 3.2, at 405, 33 I.L.M. at 1227.

18. In a 1998 report regarding a case not involving the Anti-Dumping Agreement, the AB did construe the Article 11 standard of "objective assessment" to be "neither de novo review as such, nor 'total deference.'" WTO Appellate Body Report, European Communities-Measures Concerning Meat and Meat Products (Hormones), WT/DS26/AB/R, [para] 117 (Jan. 16, 1998). Unfortunately, having said what the "objective assessment" is not, the AB did not explicate what it is.

19. Chevron U.S.A., Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 842-43 (1984). Later in its opinion, the Court cited with approval earlier cases that referred to "reasonable" (rather than "permissible") interpretations made by administrative agencies and to "reasonable" policy choices by agencies. Id. at 844-45. The Court made no indication, either in Chevron or in later opinions, that the use of "permissible" or "reasonable" is supposed to make a difference in application of the relevant standard of review.

20. See Steven P. Croley & John H. Jackson, WTO Dispute Procedures, Standard of Review, and Deference to National Governments, 90 AM. J. INT'L L. 193, 198-99 (1996).

21. For a succinct history of the provision, see Gary N. Horlick & Peggy A. Clarke, Standards for Panels Reviewing Anti-Dumping Determinations under the GATT and WTO, in INTERNATIONAL TRADE LAW AND THE GATT/WTO DISPUTE SETTLEMENT SYSTEM 315-24 (Ernst-Ulrich Petersmann ed. 1997). One sometimes hears the argument that the change in wording during the negotiations renders the 17.6(ii) standard inconsequential. This argument is considered infra at 152.

22. President's Message to Congress Transmitting the Uruguay Round Trade Agreements, Texts of Agreements Implementing Bill, Statement of Administrative Action and Required Supporting Statements, H.R. Doc. No. 103-316, at 818 (2d Sess. 1994).

23. Id. In arguments presented to the dispute settlement panel in the United States-Japan Steel case, the United States has amplified its view of the significance of Article 17.6(ii):

It reflects the negotiators' understanding that they had left a sufficient number of issues ambiguous such that they needed to make special provision for cases in which customary rules would not provide an unequivocal result. The drafters of the Agreement wisely recognized that they could not possibly foresee, and write rules for, the precise methodology to be applied to every issue that would arise in the conduct of highly technical and complex anti-dumping proceedings. Equally wisely, they understood that, with regard to many of these complex issues, national authorities already did things differently, and that the Agreement should allow sufficient flexibility for authorities to continue such variations on the anti-dumping theme.

WTO Panel Report, United States-Anti-Dumping Measure on Certain Hot-Rolled Steel Products from Japan, WT/DS184/R, Annex A-2, [para][para] 83-84 (Feb. 28, 2001) [hereinafter United States-Japan Steel Panel Report].

24. United States-Japan Steel Appellate Body Report, supra note 11.

25. Id. [para] 52.

26. Id. [para][para] 54-56.

27. Id. [para] 56.

28. The AB's apparent conflating of the 17.6(i) test with the standard in DSU Article 11 is echoed in a subsequent decision in which a panel cited AB discussion of the Article 11 "objective assessment" clause in deciding the scope of its review under 17.6(i). See Egypt-Steel Rebar from Turkey Panel Report, supra note 11, [para][para] 7.11-.12. Moreover, the panel stated that, notwithstanding the putative rule against de novo review of national administrative determinations, it might conduct "a detailed review" of the evidence in order to determine if an objective and unbiased authority could reach the conclusion arrived at by Egyptian authorities. Id. [para] 7.14. It appears that, far from applying a more deferential review standard as seemingly required by 17.6(i), the panel conducted something closer to a de novo review.

29. Anti-Dumping Agreement, supra note 2, art. 17.6(i), at 193.

30. United States-Japan Steel Appellate Body Report, supra note 11, [para] 59. The invocation of the Vienna Convention reflects the widespread assumption that the "customary rules of interpretation of public international law" incorporated by Article 17.6(ii) refer to the Convention. See id. [para] 57.

31. Id. [para] 62.

32. Id.

33. See United States-Japan Steel Appellate Body Report, supra note 11, [para][para] 63-90. As noted later, at times the analysis of panels and the Appellate Body have appeared to apply a standard similar to 17.6(ii) even when they fail to invoke that section explicitly. In the Appellate Body decision of the Article 6.8 issue, the Body neither cited to 17.6(ii) nor used language from which a reader could infer that such a standard was in fact being applied. Nor, as the discussion in the text will explain, was the Body's conclusion consistent with a 17.6(ii) standard.

34. Anti-Dumping Agreement, supra note 2, art. 6.8, at 180.

35. United States-Japan Steel Appellate Body Report, supra note 11, [para] 66.

36. See id. [para] 87.

37. See Id. [para][para] 80-83.

38. Anti-Dumping Agreement, supra note 2, Annex II, [para] 3, at 196 (emphasis added).

39. Id. art. 6.8, at 180.

40. United States-Japan Steel Appellate Body Report, supra note 11, [para] 82.

41. Id. [para] 84.

42. Id. [para] 85.

43. Id. [para] 89.

44. See Vienna Convention on the Law of Treaties, art. 31.1, May 23, 1969, 1155 U.N.T.S. 332, 340 (1980) [hereinafter Vienna Convention].

45. THE CONCISE OXFORD DICTIONARY OF CURRENT ENGLISH 1144 (9th ed. 1995).

46. Article 31.2 defines the context as also including certain contemporaneous agreements or instruments which are not relevant in the present instance. See Vienna Convention, supra note 44, art. 31.2, 1155 U.N.T.S. at 340.

47. The full text of Article 6.1.1 reads: "Exporters or foreign producers receiving questionnaires used in an anti-dumping investigation shall be given at least 30 days for reply. Due consideration should be given to any request for an extension of the 30-day period and, upon cause shown, such an extension should be granted whenever practicable." Anti-Dumping Agreement, supra note 2, art. 6.1.1, at 178.

48. See United States-Japan Steel Appellate Body Report, supra note 11, [para] 74.

49. See United States-Japan Steel Panel Report, supra note 23, [para] 7.40.

50. United States-Japan Steel Appellate Body Report, supra note 11, [para] 65.

51. Id. [para] 68.

52. DSU, supra note 1, art. 3.2, at 405, 33 I.L.M. at 1227.

53. Article 3.5 reads, in relevant part: "The authorities shall also examine any known factors other than the dumped imports which at the same time are injuring the domestic industry, and the injuries caused by these other factors must not be attributed to the dumped imports." Anti-Dumping Agreement, supra note 2, art. 3.5, at 173.

54. GATT Dispute Panel Report, United States Complaint Concerning Atlantic Salmon Anti-Dumping Duties, Apr. 27, 1994, GATT B.I.S.D. 41S/Vol. 1/229, at [para] 555, quoted in United States-Japan Steel Appellate Body Report, supra note 11, [para] 218. This panel report had been formerly "adopted" by the Contracting Parties to the GATT.

55. United States-Japan Steel Appellate Body Report, supra note 11, [para] 223.

56. Id. [para] 228.

57. Id. [para] 229.

58. Id. [para] 230.

59. As part of the Uruguay Round package of agreements and declarations, trade ministers recognized "the need for the consistent resolution of disputes arising from anti-dumping and countervailing duty measures." Declaration on Dispute Settlement Pursuant to the Agreement on Implementation of Article VI of the General Agreement on Tariffs and Trade 1994 or Part V of the Agreement on Subsidies and Countervailing Measures, Apr. 15, 1994, WTO Agreement, in RESULTS OF THE URUGUAY ROUND, supra note 1, at 453. Some have speculated that this statement means that Article 17.6 should also be applied to the Agreement on Subsidies and Countervailing Measures, a position that is perhaps stronger in the context of the review of injury determinations, standards for which are very similar in the two agreements. In any case, there is not even any indirect textual support for the proposition that the Safeguards Agreement and the Anti-Dumping Agreement need to be interpreted consistently with one another, particularly since the language at issue differs. Moreover, even if there were a reason to attempt consistent interpretation of the two agreements, it should presumably not be at the cost of the standard of review set forth in Article 17.6(ii).

60. The only change in the relevant language of Article 3, [para] 3 of the Tokyo Round Anti-Dumping Agreement was the insertion of the adjective "other," so that the clause in Article 3.5 of the current agreement reads as follows: "the injuries caused by these other factors must not be attributed to the dumped imports." Anti-Dumping Agreement, supra note 2, art. 3.5, at 173 (emphasis added).

61. See Daniel K. Tarullo, Beyond Normalcy in the Regulation of International Trade, 100 HARV. L. REV. 546, 589-93 (1987).

62. See United States-Korean DRAMS Panel Report, supra note 9.

63. Id. [para] 4.474 (quoting 19 C.F.R. [sec] 353.25 (1996)).

64. Consider the following passage: "For example, in common parlance, a statement that it is 'likely' to rain implies a greater likelihood of rain than a statement that rain is not unlikely, or not 'not likely.'" United States-Korean DRAMS Panel Report, supra note 9, [para] 6.46.

65. Id. [para] 6.51; Anti-Dumping Agreement, supra note 2, arts. 11.1-.2, at 187-88. In this paragraph the panel actually cited only to Article 11.2, but in its earlier discussion the panel had referred also to Article 11.1, which it regarded as "implemented" by Article 11.2. United States-Korean DRAMS Panel Report, supra note 9, [para] 6.40.

66. See Hyundai Electronics Co. v. United States, 53 F. Supp. 2d 1334 (Ct. Int'l Trade 1999). It is unusual for the U.S. Court of International Trade to decide the same issues as a WTO panel.

67. Id. at 1343.

68. Id. at 1344. The canon in question is usually referred to as the Charming Betsy doctrine, after its first articulation in Murray v. Schooner Charming Betsy, 6 U.S. (2 Cranch) 64 (1804).

69. Hyundai, 53 F. Supp. 2d at 1342-44.

70. Id. at 1344-45.

71. See 19 U.S.C. [sec] 1677e(b) (2000).

72. United States-Japan Steel Appellate Body Report, supra note 11, [para] 19.

73. Id. [para] 97.

74. Id. [para] 98 (quoting Anti-Dumping Agreement, supra note 2, Annex II, [para] 7, at 196) (emphasis added by Appellate Body).

75. Id. [para] 99.

76. Id.

77. Id. [para] 100.

78. United States-Japan Steel Appellale Body Report, supra note 11, [para][para] 100-04.

79. Anti-Dumping Agreement, supra note 2, Annex II, [para] 5, at 196.

80. Id., Annex II, [para] 2, at 196.

81. Id. art. 6.13, at 181.

82. United States-Japan Steel Appellate Body Report, supra note 11, [para] 102.

83. Id. [para] 101.

84. Id. [para] 102.

85. Id. [para] 104.

86. Id. [para][para] 107-10.

87. United States-Japan Steel Panel Report, supra note 23, [para] 7.73.

88. United States-Japan Steel Appellate Body Report, supra note 11, [para] 109 (emphasis added).

89. WTO Appellate Body Report, European Communities-Anti-Dumping Duties on Imports of Cotton-Type Bed Linen from India, WT/DS141/AB/R (Mar. 1, 2001) [hereinafter EC-Bed Linen Appellate Body Report].

90. Anti-Dumping Agreement, supra note 2, art. 2.4.2, at 171. Article 2.4.2 goes on to provide for an exception to this rule that was not applicable in the present case.

91. EC-Bed Linen Appellate Body Report, supra note 89, [para] 47.

92. WTO Panel Report, European Communities-Anti-Dumping Duties on Imports of Cotton-Type Bed Linen From India, WT/DS141/R, [para][para] 6.112-.113 (Oct. 30, 2000).

93. Id. [para] 6.116.

94. Anti-Dumping Agreement, supra note 2, art. 2.1, at 168 (emphasis added).

95. EC-Bed Linen Appellate Body Report, supra note 89, [para] 51.

96. Anti-Dumping Agreement, supra note 2, art. 2.4.2, at 171.

97. EC-Bed Linen Appellate Body Report, supra note 89, [para] 55.

98. Id. [para] 56.

99. Id. [para] 57.

100. Id. [para] 58.

101. The Latin maxim sometimes associated with this norm is ut res magis valeal quam pereat. This maxim was specifically considered by the International Law Commission to be embodied in the general rule of interpretation that eventually became Article 31.1 of the Vienna Convention. Report of the International Law Commission to the General Assembly, reprinted in [1966] 2 Y.B. Int'l L. Comm'n 219, U.N. Doc. A/CN.4/SER.A/1966/Add.1.

102. Compare The Uruguay Round Trade Negotiations Committee, Draft Final Act Embodying the Results of the Uruguay Round of Multilateral Trade Negotiations 20 December 1991, art. 2.4.2, in "THE DUNKEL DRAFT" From the GATT Secretariat, MTN.TNC/W/FA, F.4 (1992), with Anti-Dumping Agreement, supra note 2, art. 2.4.2, at 171.

103. WTO Panel Report, United States-Anti-Dumping Measures on Stainless Steel Plate in Coils and Stainless Steel Sheet and Strip from Korea, WT/DS179/R, [para] 6.111 n.114 (Dec. 22, 2000) [hereinafter United States-Korean Steel Panel Report]. Later in its report, the panel also commented that the definition "able to be compared (with)" as found in the dictionary cited by the Appellate Body "does not cast great light on the meaning of the term as used in Article 2 of the AD Agreement." Id. [para] 6.120.

104. THE CONCISE OXFORD DICTIONARY OF CURRENT ENGLISH 269 (9th ed. 1995).

105. THE CONCISE OXFORD DICTIONARY OF CURRENT ENGLISH 290 (10th ed. 1999).

106. EC-Bed Linen Appellate, Body Report, supra note 89, [para] 65.

107. The Appellate Body took a similar approach in its review of a panel determination in the Thailand-Steel from Poland Appellate Body Report, supra note 10. In that case, the analysis of both the panel and the Appellate Body were more convincing than in the United States-Japan Steel and EC-Bed Linens cases in establishing a lack of ambiguity in the disputed provision of the Anti-Dumping Agreement. However, the panel essentially ignored the standard of review in its analysis and referred to it earlier only by saying "we are mindful of the standard of review in Article 17.6(ii)." The Appellate Body upheld the panel's analysis despite the latter's failure to apply 17.6(ii) explicitly by asserting that the "Panel's interpretation that Article 3.4 requires a mandatory evaluation of all the individual factors listed in that Article clearly left no room for a 'permissible' interpretation that all individual factors need not be considered." Id. [para] 127. In fairness to the panel and the Appellate Body, neither Thailand nor the EC (which, as a third party participant, supported Thailand on the contested point) made the Article 17.6(ii) argument as explicitly or as well as they might have. The closest Thailand itself came to tying Article 17.6(ii) to the Article 3.4 issue was in its second oral statement, where it asserted less than convincingly that Thailand believed there to be another permissible interpretation of the term "evaluate." WTO Panel Report, Thailand-Anti-Dumping Duties on Angles, Shapes and Sections of Iron or Non-Alloy Steel and H-Beams from Poland, WT/DS122/R, Annex 2.9, [para][para] 53, 54 (Sept. 28, 2000) [hereinafter Thailand-Steel from Poland Panel Report].

108. Article 6.10 reads, in relevant part:

In cases where the number of exporters, producers, importers or types of products involved is so large as to make such a determination impracticable, the authorities may limit their examination either to a reasonable number of interested parties or products by using samples which are statistically valid on the basis of information available to the authorities at the time of the selection.

Anti-Dumping Agreement, supra note 2, art. 6.10, at 180.

109. United States-Japan Steel Appellate Body Report, supra note 11, [para] 114 (quoting Anti-Dumping Agreement, supra note 2, art. 9.4, at 185-86).

110. Anti-Dumping Agreement, supra note 2, art. 6.8, at 180.

111. United States-Japan Steel Appellate Body Report, supra note 11, [para] 119.

112. Id. [para] 122.

113. Id. [para] 121.

114. Id. [para] 125. The situation contemplated by the Appellate Body is not likely. It is theoretically possible, however, perhaps where a dispute between the United States and the government of the exporters resulted in the home government prohibiting all its exporters from cooperating in the anti-dumping investigation.

115. Id. [para][para] 125-26.

116. Id. [para] 130.

117. United States-Japan Steel Panel Report, supra note 23, [para] 7.86.

118. Anti-Dumping Agreement, supra note 2, art. 2.1, at 168 (emphasis added).

119. United States-Japan Steel Panel Report, supra note 23, [para] 7.110.

120. United States-Japan Steel Appellate Body Report, supra note 11, [para] 141.

121. Id. [para] 147.

122. Id. [para] 148.

123. Id. [para] 154.

124. Id. [para] 158; United States-Japan Steel Panel Report, supra note 23, [para] 7.112.

125. See United States-Korean Steel Panel Report, supra note 103, [para][para] 6.56-.57. The panel also considered a similar dispute with respect to sales that had been made to the same U.S. customer through an affiliated company. Here, though, the panel found the issue to be one of fact, rather than law, a conclusion that might be contested, but that will not be contested in this Article.

126. DOC technically added the allocated amount of bad debt to the home market "normal value" of the steel, in order to make that value comparable with the export value. See id. [para][para] 6.64-.66. Whether one adds the cost to the home market price or subtracts it from the export market price, the result is the same-a higher margin of dumping.

127. Anti-Dumping Agreement, supra note 2, art. 2.4, at 170.

128. United States-Korean Steel Panel Report, supra note 103, [para] 6.60.

129. Id. [para] 6.75.

130. Id.

131. Id. [para] 6.77.

132. Id. [para][para] 6.75-.76.

133. Id. [para] 6.76.

134. United States-Korean Steel Panel Report, supra note 103, [para][para] 6.107-.108.

135. Id. [para] 6.120.

136. Id. [para] 6.121.

137. Id. [para] 6.123.

138. See WTO Panel Report, United States-Anti-Dumping and Countervailing Measures on Steel Plate from India, WT/DS206/R, [para][para] 7.56-.61 (June 28, 2002) [hereinafter United States-Steel Plate from India Panel Report].

139. Id. [para] 7.57.

140. Id. [para] 7.7.

141. It should be noted that the panel did not accept India's reading of Article 6.8 either. Id. [para][para] 7.59-.62. Moreover, as noted later, the panel rejected India's challenge that the applicable U.S. statutory provisions were, on their face, inconsistent with Article 6.8. Id. [para][para] 7.81-100.

142. See Egypt-Steel Rebar front Turkey Panel Report, supra note 11.

143. Id. Annex 1-2, [para] 46.

144. Id.

145. Id. [para] 7.351.

146. Id. [para] 7.352.

147. Id. (quoting Anti-Dumping Agreement, supra note 2, art. 2.4, at 170). In fact, although Egypt lost on this point, it won on the issue, because the panel found that Turkey had not adequately raised the issue before Egyptian national import authorities. Id. [para] 7.387.

148. See United States-Japan Steel Appellate Body Report, supra note 11, [para][para] 159-61.

149. Id. [para] 165. "First, the sale must be 'in the ordinary course of trade'; second, it must be of the 'like product'; third, the product must be 'destined for consumption in the exporting country'; and, fourth, the price must be 'comparable'." Id.

150. Id.[para] 166.

151. Id.

152. Id. [para] 172.

153. Id. [para] 168.

154. United States-Japan Steel Appellate Body Report, supra note 11, [para] 180.

155. 19 U.S.C. [sec] 1677(7)(C)(iv) (2000).

156. United States-Japan Steel Appellate Body Report, supra note 11, [para][para] 204-10. The AB's concern was that a focus upon the poorly performing part of an industry may give a misleading impression of the industry as a whole, even when there was a nominal examination of the entire industry.

157. Id. [para][para] 208-09; Anti-Dumping Agreement, supra note 2, art. 4.1, at 174.

158. United States-Japan Steel Appellate Body Report, supra note 11, [para][para] 210-15. Similarly, the panel's conclusion in United States-Steel Plate from India Panel Report, supra note 138, that U.S. statutory provisions were not inconsistent on their face with Article 6.8 was offset to some degree by its conclusion that DOC had applied the statute in violation of that section of the Agreement. See text supra at 143.

159. United States-Japan Steel Appellate Body Report, supra note 11, [para] 184.

160. See id. [para][para] 192-94.

161. See, e.g., United States-Korean Steel Panel Report, supra note 103, [para][para] 6.134-.136 (Korean reliance upon Article 2.4 to attack injury determination is misplaced, but because no claim was made under Article 3 injury requirements, the panel did not have the issue within its terms of reference).

162. Thailand-Steel from Poland Panel Report, supra note 107, [para][para] 7.113-.114. Although the panel decision was appealed, this issue was not included in the appeal.

163. Id. [para] 7.111.

164. Id. [para] 7.113.

165. Mssrs. Abi-Saab (Egypt), Baptista (Brazil), Ehlermann (Germany), Lacarte-Muro (Uruguay), Matsushita (Japan), Sacerdoti (Italy), and Taniguchi (Japan). Biographical information on members of the Appellate Body is included in WTO press releases announcing their appointments, available on the WTO web site. Press Release by WTO, Press/246, WTO Appoints New Appellate Body Members (Sept. 25, 2001); Press Release by WTO, Press/179, WTO Completes Appointment of Appellate Body Members (May 25, 2000); Press Release by WTO, Press/32, WTO Announces Appointments to Appellate Body (Nov. 29, 1995), available at http://www.wto.org.

166. Mssrs. Feliciano and Lockhart, respectively.

167. Mssrs. Bacchus (U.S.) and Beeby (New Zealand).

168. The panel was composed of Mssrs. Feliciano, Taniguchi, and Lacarte-Muro. See United States-Japan Steel Appellate Body Report, supra note 11.

169. See, e.g., MARK ARONSON & BRUCE DYER, JUDICIAL REVIEW OF ADMINISTRATIVE ACTION 217-21 (1996) (discussing changes in standards of review in Australia); HILARY DELANY, JUDICIAL REVIEW OF ADMINISTRATIVE ACTION-A COMPARATIVE ANALYSIS (2001) (comparing standards of review in common law countries); JURGEN SCHWARZE, EUROPEAN ADMINISTRATIVE LAW 264-68 (1992) (explaining differing levels of review for different kinds of administrative cases).

170. United States-Japan Steel Appellate Body Report, supra note 11, [para][para] 50-62, discussed supra at 119-20. The controversy surrounding U.S. proposals for a special standard of review, as recounted in Croley & Jackson, supra note 20, reinforces my inference that the potential importance of different standards of review is understood in the WTO.

171. For similar reasons, the inconsistent treatment of legal issues in United States-Japan Steel and other AB cases seems hard to attribute solely to shortcomings in the experience and training of the members.

172. See, e.g., GTS Indus. v. United States, 182 F. Supp. 2d 1369 (Ct. Int'l Trade 2002).

173. See, e.g., Thai Pineapple Canning Indus. Corp. v. United States, 273 F.3d 1077 (Fed. Cir. 2001).

174. In one of the questions it posed to the Government of Thailand, the panel in Thailand-Steel from Poland stated that "[s]ome commentators believe that rarely if ever can there be more than one permissible interpretation of any given treaty provision." Thailand-Steel from Poland Panel Report, supra note 107, Annex 2-6, [para] 53; cf. Croley & Jackson, supra note 20, at 200 (suggesting "it is not clear in light of [the Vienna] Convention whether or how a panel could ever reach the conclusion that provision of an agreement admit of more than one interpretation" because the Convention provides rules for resolving ambiguities in the text of an international agreement). The authors, based on private conversations, state that "some of the negotiators seem to feel that this is precisely the case: there never can be resort to the second sentence." Id. at 201.

175. IAN SINCLAIR, THE VIENNA CONVENTION ON THE LAW OF TREATIES 153 (2d ed. 1984).

176. Id.

177. Id. at 154.

178. See, e.g., MAARTEN BOS, A METHODOLOGY OF INTERNATIONAL LAW 149 (1984) (concluding, after considering Articles 31 and 32, that "articles on interpretation, besides solving problems of interpretation may also provoke them!"); IAN BROWNLIE, PRINCIPLES OF PUBLIC INTERNATIONAL LAW 632 (1990) (noting that the textual approach of the Vienna Convention "in practice often leaves the decision-maker with a choice of possible meanings and in exercising that choice it is impossible to keep considerations of policy out of account"); EDWARD S. YAMBRUSIC, TREATY INTERPRETATION: THEORY AND REALITY 169 (1987) (stating that the Vienna Convention articles have failed to provide a general rule of interpretation). The closest commentators come to supporting the single permissible reading position are two statements found in later editions of 1 OPPENHEIM'S INTERNATIONAL LAW (Sir Robert Jennings & Sir Arthur Watts eds., 9th ed. 1992): (1) "The application of the basic rule of interpretation laid down in Article 31 . . . will usually establish a clear and reasonable meaning." Id. at 1275; (2) "The finding whether a treaty is clear or not is not the starting point but the result of the process of interpretation. It is not clarity in the abstract which is to be ascertained, but clarity in relation to particular circumstances and there are few treaty provisions for which circumstances cannot be envisaged in which their clarity could be put in question." Id. at 1267. The first statement, of course, leaves open, as does the Vienna Convention itself, that recourse to Article 32 is sometimes necessary. The second statement, with its implication that clarity emerges after recourse to the Vienna Convention rules, might provide some support for the single permissible reading view. However, the subsequent reference to circumstances that could put clarity into question seems to undermine this view. Moreover, the treatise, in later recounting eleven non-exclusive "supplementary" means of interpretation, id. at 1276-82, actually reinforces the Brownlie view, quoted supra.

179. United States-Japan Steel Appellate Body Report, supra note 11, [para] 59.

180. WTO Appellate Body Report, Japan-Taxes on Alcoholic Beverages, WT/DS8/AB/R, at 11 (Oct. 4, 1996); see also WTO Appellate Body Report, United States-Standards for Reformulated and Conventional Gasoline, WT/DS2/AB/R, at 15 (Apr. 29, 1996) [hereinafter United States-Reformulated Gasoline Appellate Body Report].

181. See Croley & Jackson, supra note 20, at 199.

182. Croley and Jackson briefly but convincingly argue against the use of undocumented negotiating history, such as the views of some negotiators that the eventual Article 17.6(ii) language was meant to preclude multiple permissible readings of Anti-Dumping Agreement provisions, in interpreting treaty provisions. See Croley & Jackson, supra note 20, at 201 n.37.

183. 1 OPPENHEIM'S INTERNATIONAL LAW, supra note 178, at 1278.

184. United States-Japan Steel Appellate Body Report, supra note 11, [para] 62.

185. A reader who doubts the suggestion in the text should perform the very simple mental exercise of considering whether an identical set of environmental regulations would be interpreted identically by the Commerce Department and the Environmental Protection Agency.

186. WTO Appellate Body Report, European Communities-Measures Affecting Asbestos and Asbestos-Containing Products, WT/DS135/AB/R, [para] 132 (Mar. 12, 2001).

187. Id. [para] 113-14.

188. WTO Appellate Body Report, United States-Import Prohibition of Certain Shrimp and Shrimp Products, WT/DS58/AB/R (Oct. 12, 1998) [hereinafter United States-Shrimp-Turtle Appellate Body Report].

189. Id. [para] 130.

190. See generally Daniel K. Tarullo, Norms and Institutions in Global Competition Policy, 94 AM. J. INT'L L. 478 (2000).

191. Countervailing duty (or countervailing "measure") laws are used by importing countries to offset the purported value of a subsidy on imported merchandise. Safeguards measures may be imposed by importing countries where imports are causing serious injury to a domestic industry, even where there is no finding of an "unfair" practice such as dumping or subsidization. Use of these measures is governed by the WTO Agreement on Subsidies and Countervailing Measures and the Agreement on Safeguards, respectively. See Agreement on Subsidies and Countervailing Measures, Apr. 15, 1994, WTO Agreement, in RESULTS OF THE URUGUAY ROUND, supra note 1, at 264; Agreement on Safeguards, Apr. 15, 1994, WTO Agreement, in RESULTS OF THE URUGUAY ROUND, supra note 1, at 315 [hereinafter Safeguards Agreement].

192. WTO Appellate Body Report, United States-Countervailing Measures Concerning Certain Products from the European Communities, WT/DS212/AB/R (Dec. 9, 2002); WTO Appellate Body Report, United States-Countervailing Duties on Certain Corrosion-Resistant Carbon Steel Flat Products from Germany, WT/DS213/AB/R (Nov. 28, 2002); WTO Appellate Body Report, United States-Definitive Safeguard Measures on Imports of Circular Welded Carbon Quality Line Pipe from Korea, WT/DS202/AB/R (Feb. 15, 2002); WTO Appellate Body Report, United States-Safeguard Measures on Imports of Fresh, Chilled or Frozen Lamb Meat from New Zealand and Australia, WT/DS177/AB/R (May 1, 2001) [hereinafter United States-Lamb Appellate Body Report]; WTO Appellate Body Report, United States-Definitive Safeguard Measures on Imports of Wheat Gluten from the European Communities, WT/DS166/AB/R (Dec. 22, 2000); WTO Appellate Body Report, United States-Imposition of Countervailing Duties on Certain Hot-Rolled Lead and Bismuth Carbon Steel Products Originating in the United Kingdom, WT/DS138/AB/R (May 10, 2000); WTO Panel Report, United States-Preliminary Determination with Respect to Certain Softwood Lumber from Canada, WT/DS236/R (Sept. 27, 2002).

193. WTO Appellate Body Report, Chile-Price Band Systems and Safeguard Measures Relating to Certain Agricultural Products, WT/DS207/AB/R (Sept. 9, 2002); WTO Appellate Body Report, Argentina-Safeguard Measures on Imports of Footwear, WT/DS121/AB/R (Dec. 14, 1999) [hereinafter Argentina-Footwear Appellate Body Report]; WTO Appellate Body Report, Korea-Definitive Safeguard Measures on Imports of Certain Dairy Products, WT/DS98/AB/R (Dec. 14, 1999) [hereinafter Korea-Dairy Appellate Body Report].

194. Of particular note is the trio of safeguards cases (United States-Lamb; Argentina-Footwear; Korea-Dairy Products) in which the Appellate Body concluded that the introductory language of Article XIX of GATT ("If, as a result of unforeseen developments" imports cause serious injury to a domestic industry, then safeguards may be imposed) requires a specific finding as a condition of making the requisite finding. See MICHAEL J. TREBILCOCK & ROBERT HOWSE, THE REGULATION OF INTERNATIONAL TRADE 228 (2d ed. 1999). The Uruguay Round Safeguards Agreement contains no such condition, consistent with the widespread belief prior to the Round that, as a result of GATT practice, the language did not establish such an independent requirement. See Safeguards Agreement, supra note 191.

195. Croley & Jackson, supra note 20, at 211.

196. Note that the political representativeness argument could be different in the case of an environmental or health regulation, for example. In that context, pro-environmental or pro-consumer interests might be sufficiently well represented to assure that the resulting regulation is in fact designed to promote those interests. On the other hand, it is also possible that domestic producer and consumer interests might reach a compromise that most disadvantages the party not represented in the process-foreign producers. Just this sort of process may lie behind the United States-Reformulated Gasoline case, where the EPA promulgated a regulation that was a careful balance between environmental and domestic producer interests, while disadvantaging foreign producers relative to domestic producers. See United States-Reformulated Gasoline Appellate Body Report, supra note 180.

197. Another argument sometimes offered against the use of a Chevron standard in the WTO is that deference to national authorities would permit multiple interpretations of the Anti-Dumping Agreement, a situation thought to be somewhere between undesirable and unbearable. It is easy to understand why U.S. courts do not apply Chevron where multiple agencies interpret a single statute-citizens could be subject to different obligations depending on which agency was regulating them. But in the international context this concern is much less important. The Anti-Dumping Agreement places restraints on the use of national anti-dumping laws; it does not require uniformity in those laws. National anti-dumping laws already vary in some particulars. By deferring to "permissible" national interpretations of the Anti-Dumping Agreement, the WTO would simply be stipulating that one of its obligations does not cover a specific national practice.

198. John O. McGinnis & Mark L. Movsesian, The World Trade Constitution, 114 HARV. L. REV. 511, 514 (2000). McGinnis and Movsesian adopt an explicitly anti-regulatory position that does not seem necessary for their core point. This analysis, in turn, is related to Jonathan Macey's suggestion in a domestic context that traditional canons of statutory construction allow judges to interpret statutes so as to promote, at least sometimes, the public interest, at the expense of private interests that may have succeeded in obtaining special interest legislation. Jonathan R. Macey, Promoting Public-Regarding Legislation Through Statutory Interpretation: An Interest Group Model, 86 COLUM. L. REV. 223, 226 (1986).

199. See, e.g., Charles Jones, Patriotism, Morality, and Global Justice, in GLOBAL JUSTICE 125 (Ian Shapiro & Lea Brilmayer eds., 1999); Robin West, Is the Rule of Law Cosmopolitan?, 19 QUINNIPIAC L. REV. 259 (2000).

200. Whether, as an empirical matter, administration of trade remedy laws disproportionately disadvantages exports from developing countries is uncertain. Traditionally the United States and the European Union (EU) have been the WTO members that invoke these laws most frequently and with the greatest effect upon trade. However, over the years Japan has been the most important target of U.S. and EU use of these laws. Moreover, as explained supra, note 4, developing countries are increasingly using anti-dumping laws themselves. Still, it is noteworthy that one of the highest priorities of most developing countries for a new round of trade negotiations is achieving further disciplines on the use of anti-dumping laws, and that the U.S. Congress has called upon the President to resist any agreement that would require changes in U.S. laws. If, as seems likely, the AB is aware of these negotiating maneuvers, it may believe that the proposition stated in the text is generally valid.

201. See, e.g., Ernst-Ulrich Petersmann, Human Rights and International Economic Law in the 21^sup st^ Century, 4 J. INT'L. ECON. L. 3 (2001); Ernst-Ulrich Petersmann, How to Constitutionalize International Law and Foreign Policy for the Benefit of Civil Society?, 20 MICH. J. INT'L L. 1 (1998).

202. See,.e.g., GAIL, E. EVANS, LAWMAKING UNDER THE TRADE CONSTITUTION (2000); JOHN H. JACKSON, THE WPRLD TRADE ORGANIZATION: CONSTITUTION AND JURISPRUDENCE (1998); McGinnis & Movsesian, supra note 198, at 511.

203. EVANS, supra note 202, at 4.

204. See, e.g., Robert Howse & Kalypso Nicolaidis, Legitimacy through "Higher Law?": Why Constitulionalizing the WTO is a Step Too Far, in THE ROLE OF THE JUDGE: LESSONS FOR THE WTO (Thomas Cottier & Petros C. Mavroidis eds.) (forthcoming Mar. 2003) (manuscript ch. 19, on file with Law and Policy in International Business). Most members of the United States Congress, even those who generally support liberalized trade, would be apoplectic at the thought of the WTO moving in the direction of trade constitutionalism in the strong sense described in the text.

205. See, e.g, Thomas Cottier & Matthias Oesch, The Paradox of Judicial Review in International Trade Regulation: Towards a Comprehensive Framework, in THE ROLE OF THE JUDGE, supra note 204 (manuscript ch. 18). In private conversations-though not, so far as I am aware, in published work-some supporters of the constitutionalist position cast the Appellate Body in a kind of heroic light, making history in the same way that the European Court of Justice (and even the United States Supreme Court in the U.S. system) pushed its role beyond what had been contemplated in the Treaty of Rome and, in doing so, contributed enormously to the European integration project.

206. The Kennedy Round ended in 1967, the Tokyo Round in 1979, and the Uruguay Round in 1994. In December 2001, trade ministers launched the Doha Round, which is supposed to end in 2005. If past experience is any indication, the actual completion date is likely to be years later.

207. See, e.g., United States-Shrimp-Turtle Appellate Body Report, supra note 188, [para] 158; WTO Appellate Body Report, United States-Tax Treatment for "Foreign Sales Corporations," WT/DS108/AB/R, [para] 166 (Feb. 24, 2000). For a review of the good faith obligation in WTO jurisprudence, see Thomas Cottier & Krista N. Schefer, Good faith and the Protection of Legitimate Expectations in the WTO, in NEW DIRECTIONS IN INTERNATIONAL ECONOMIC LAW 47, 47-68 (Marco Bronckers & Reinhard Quick eds., 2000).

208. Vienna Convention, supra note 44, art. 26, at 339. The Vienna Convention contains two other obligations of good faith: the Preamble (universality of principle) and Article 31 (treaties to be interpreted in good faith). Good faith is also used as a benchmark for determining rights under Articles 46 and 69. Id. pmbl., arts. 31, 46, 69, at 332, 340, 343, 348-49.

209. An account of the principle of good faith in international law may be found in SHABTAI ROSENNE, DEVELOPMENTS IN THE LAW OF TREATIES 1945-1986 135-79 (1989).

210. United States-Japan Steel Appellate Body Report, supra note 11, [para] 101.

211. The criticisms and recommendations of those who argue based on legitimacy are reminiscent of debates over judicial review in the United States. One observer, for example, argues that the panels and AB should take a conservative approach to their work and not step across the boundary "between policymaking and mere judicial interpretation." CLAUDE E. BARFIELD, FREE TRADE, SOVEREIGNTY, DEMOCRACY: THE FUTURE OF THE WORLD TRADE ORGANIZATION 56 (2001). The formulation recalls the complaints of some American conservatives about the decisions of the Warren Court (though not contemporary conservatives, who are quite pleased with judicial activism, even if they cannot bring themselves to admit it). Barfield has proposed changes in the WTO to restrain the dispute settlement bodies, such as allowing the Dispute Settlement Body to block panel or AB decisions with a vote of one-third or more WTO members accounting for at least one-quarter of total trade among WTO members. Id. at 127. The implicit premise is that legitimacy can be restored by subordinating the unelected "judiciary" to the representatives of member states. Other observers prescribe a different route to legitimacy-more transparency and broader participation by affected parties in the dispute settlement process. See, e.g., Howse & Nicolaidis, supra note 204. This kind of proposal recalls the time-honored method of dealing with substantive differences, or perceived legitimacy problems, by concentrating upon creating an open process.

212. BARFIELD, supra note 211, at 7.

213. Howse & Nicolaidis, supra note 204.

214. Id.

215. Id. Howse and Nicolaidis point out, among other things, that the ECJ's decisions are embedded in a much richer institutional setting than prevails in the WTO (with an active Commission and Council) and that EU treaties do provide for direct application of law made at the European level in the member states. Id.

216. Barfield does express passing sympathy for the view that the WTO decisions on anti-dumping and safeguards have imposed obligations beyond those agreed to by member states in the Uruguay Round. See BARFIELD, supra note 211, at 190-91 n.17.

217. The phrase may be loosely translated as "agreements should be obeyed."

218. This argument could also draw on the statement in Article 3.2 of the DSU to the effect that "[r]ecommendations and rulings of the DSB cannot add to or diminish the rights and obligations provided in the covered agreements," a provision that is often cited but of uncertain impact on the practice of panels and the AB. DSU, supra note 1, art. 3.2, at 405, 33 I.L.M. at 1227.

219. Smiley v. Citibank, 517 U.S. 735, 740-41 (1996). Cf. Cass R. Sunstein, Law and Administration After Chevron, 90 COLUM. L. REV. 2071, 2076 (1990) ("Chevron is best defended as a sensible reconstruction of congressional instruction in light of the relevant institutional capacities").

220. Again, the formulation in the text does not answer the question of what, precisely, the nature of that deference is.

221. See, e.g., WTO Appellate Body Report, India-Patent Protection for Pharmaceutical and Agricultural Chemical Products, WT/DS50/AB/R (Dec. 19, 1997) (intellectual property obligations). Opposition to the intellectual property obligations imposed on developing countries by the Uruguay Round agreements has grown, as a result of which the Ministerial Conference launching the next round of trade negotiations has taken a decidedly conciliatory view of developing country obligations. See, e.g., World Trade Organization Ministerial Conference, Declaration on the TRIPS Agreement and Public Health, WT/MIN(01)/DEC/2 (Nov. 20, 2001). To date, however, the dispute settlement process has not interpreted the provisions of the Agreement on Trade Related Aspects of Intellectual Property Rights (TRIPS) in a way that could fairly be interpreted as resisting U.S. policy preferences as achieved in the TRIPS agreement.

222. Some readers may protest that, with an American as one member of the Appellate Body, and with Americans having held important WTO positions during its first seven years, any notion of "containing America" in that organization is absurd. It is important to note that, as individuals, some Americans share concerns over U.S. assertiveness or unilateralism in a number of areas, including trade. Indeed, committed free traders in the United States have long criticized what they find to be inconsistencies or mercantilist tendencies in U.S. trade policy.

223. In fact, the AB has actually supported a U.S. effort to export another of its domestic judicial practices to the WTO. The United States has urged dispute settlement bodies to accept amicus briefs filed by non-governmental entities. The AB has done so in a number of cases. See, e.g., WTO Appellate Body Report, European Communities-Trade Description of Sardines, WT/DS231/AB/R, [para][para] 153-57 (Sept. 26, 2002). However, many other WTO members, particularly developing countries, have objected strenuously to this practice. See WTO Appellate Body Under Fire Again for Amicus Rulings in Dispute Proceedings, 19 Int'l Trade Rep. (BNA) 1858 (Oct. 31, 2002). Thus, while this may be a good example of resistance to U.S. efforts to remake the WTO "judicial" system in its domestic image, the resistance did not extend to the AB itself.

224. This argument suggests a possible third stage in the process described in an interesting recent article by Richard Steinberg, who argues that trade negotiations are launched in accordance with "law-based" principles but concluded in accordance with power principles. As a result, the results of trade negotiations generate asymmetrical benefits and may sometimes even be non-Pareto improving. See generally Richard H. Steinberg, In the Shadow of Law or Power? Consensus-Based Bargaining and Outcomes in the GATT/WTO, 56 INT'L ORG. 339 (2002). The third stage in this process may add institutional authority to the principles determining the outcomes of disputes under the negotiated rules.

225. Supra at 117.

226. Differences would most obviously arise if elections had led to a change in government. But even within a government, the weight of specific constituency and policy factors varies depending on the context in which they arise.

227. WTO Agreement, supra note 1, art. X, in RESULTS OF THE URUGUAY ROUND, supra note 1, at 12, 33 I.L.M. at 1149.

228. In this discussion, references to "costs," "pay," "price," and other economic terms are meant to be understood from the perspective of the negotiating states and are, accordingly, more political than economic. So, for example, a restriction on national discretion to administer the trade laws that would look to most economists like a gain may be seen by trade negotiators as a loss because configuration of domestic political forces makes an agreement with such provisions harder to ratify.

229. See Ian Ayres & Robert Gertner, Filling Gaps in Incomplete Contracts: An Economic Theory of Default, 99 YALE L.J. 87, 87 (1989).

230. The effects hypothesized in the text contrast to the well-known phenomenon whereby negotiators intentionally leave provisions ambiguous because they cannot agree on the underlying issues. This tactic works where the relevant political constituencies (interest groups, legislatures, etc.) can be convinced that the ambiguous provision in fact is close to what they had wanted. As part of an inter-agency group reviewing the Uruguay Round agreements, I heard several private explanations from the negotiators to the effect that an ambiguous phrase really achieved the U.S. negotiating end but that the language needed to be subtle enough to give political "cover" to negotiators from other countries who had effectively conceded to the United States. I recall at the time wondering if the other countries' negotiators were giving a similar explanation to their inter-ministerial review groups. The negotiators' tactic was to use the relative ignorance of other parties to argue that the agreement yielded the desired benefits. In the case of established AB practice, however, even precision in the negotiated language is not believed because experience teaches it will be applied differently than it reads.

231. See Ayres & Gertner, supra note 229, at 88-89.

232. For the same reason, it is virtually inconceivable that the provision, contained in Article X(5) of the Agreement Establishing the World Trade Organization, for forcing a member to choose between leaving the WTO and accepting an amendment approved by three-fourths of the membership would ever be applied to the United States or the European Union. WTO Agreement, supra note 1, art. X(5), in RESULTS OF THE URUGUAY ROUND, supra note 1, at 12, 33 I.L.M. at 1149.

233. Just such an effect may already be in play. Due at least in part to perceptions of AB activism in trade law administration cases, the U.S. Congress has established as a "principal negotiating objective" in new trade rounds that U.S. negotiators seek to "preserve the ability of the United States to enforce rigorously its trade laws." Trade Act of 2002, Pub. L. No. 107-210, [sec] 2102(b)(14)(A), 116 Stat. 933, 1001 (2002). Although not binding, this provision reflects substantial political weight and likely prefigures close congressional monitoring of any proposals to change the Anti-Dumping, Subsidies and Countervailing Measures, or Safeguards Agreements. Interestingly enough, the initial set of U.S. proposals concerning the new trade round's negotiations on trade remedy laws does not include any proposal to clarify the standard of review. Instead it asserts as a "principle" that WTO dispute settlement bodies should "follow the appropriate standard of review." See Daniel Pruzin & Gary G. Yerkey, Proponents Welcome, Criticize U.S. Paper for WTO Rules Negotiations on Antidumping, 19 Int'l Trade Rep. (BNA) 1818 (Oct. 24, 2002). The release of these proposals followed by just a few weeks a speech given by the then-Chairman of the Senate Finance Committee, in which he called for changes in the DSU to reinforce the standard of review. See Gary G. Yerkey, Sen. Baucus Calls WTO "Kangaroo Court" with Strong "Bias" Against the United States, 19 Int'l Trade Rep. (BNA) 1679 (Oct. 3, 2002). In a report to Congress required by Section 2105(b)(3) of the Trade Act of 2002, 116 Stat. at 1016, the Administration set forth a negotiating strategy to respond to certain "problematic" findings of the WTO dispute settlement process, including those involving Article 17.6(ii). See Executive Branch Strategy Regarding WTO Dispute Settlement Panels and the Appellate Body: Report to the Congress Transmitted by the Secretary of Commerce (Dec. 30, 2002), at 10, at http://www.ita.doc.gov/FinalDec31ReportCorrccted.pdf. Thus, it is clear that the U.S. negotiating approach has been affected by the AB decisions. What remains to be seen is the degree to which the negative dynamic possibilities suggested in the text will in fact be realized.

234. Again, the economic terms used in this discussion are analogies only. The deadweight "loss" referred to in the text is calculated based on the hypothetical valuations of various possible negotiating concessions. This is not deadweight economic loss. However, if one assumes that most concessions offered during trade negotiations are efficiency-enhancing (a contestable proposition, but one many economists would embrace), then the foreclosure of specific trade deals as described in the text is roughly congruent with deadweight loss in an economic sense.

235. There is also a scenario in which a WTO member's response to the AB's disregard of 17.6(ii) could come prior to a negotiating round. If the AB practice becomes sufficiently controversial, the domestic politics of that member could prompt defiance of the AB. In the dispute settlement context of the WTO, that would mean declining to implement the decision domestically and refusing to pay compensation in the form of trade concessions in other areas. It might even include threats to the prevailing party not to exercise its right to retaliate in the event of non-compliance with an AB decision. The justification would be that the AB had consistently ignored the rules negotiated by WTO members and that it had thereby exceeded its authority under the DSU by imposing new obligations on disputants. In the absence of an institutional mechanism within the WTO to rein in this "lawlessness," the aggrieved member could argue that it was compelled to take self-help measures. An ensuing downward spiral of threats and retaliation could ultimately threaten the WTO itself. This scenario is remote, to be sure. For liberal trade advocates, though, it is the classic case of a low probability event which, if it were to occur, would have devastating consequences. One might think that, as a hegemon, the United States may be more willing and able to defy AB decisions than would any other WTO member. If so, even those who wish to use the dispute settlement processes might prefer having the AB avoid decisions that too obviously neutralize provisions for which the United States negotiated. This outcome is not so clear as the increased costs of agreement, however, for if the United States has already used its power to shape the trading system to serve its interests, it may not want to risk crippling that system.

DANIEL K. TARULLO*

* Professor of Law, Georgetown University Law Center. Thanks to Chinedu Amankulor, Devaleena Das, and Kevin Robinson for their fine research assistance. I am also grateful to John Greenwald, Gary Horlick, Rob Howse, Steve Salop, and participants in a Georgetown University Law Center Faculty Workshop for comments and suggestions. As always, commenters are absolved of any responsibility for the final product.

Copyright Georgetown University Law Center Fall 2002
Provided by ProQuest Information and Learning Company. All rights Reserved