US–Zeroing (EC)

The EU's challenge related first of all to the measures expressly identified by the United States as ‘measures taken to comply’ with the DSB rulings. The EU thus challenged the Section 129 determinations adopted by the United States to implement the recommendations and rulings of the DSB. The EU argued that one such determination was vitiated by a calculation error. Of a number of other determinations, the EU argued that they were inconsistent with the United States's obligations under the AD Agreement since no new injury analysis was performed, even though the recalculation of the dumping margins without zeroing led to the exclusion of a number of exporters previously found to be dumping. The EU also challenged the US determination of the all-others rate based on margins that were below de minimis or based on facts available. These sets of claims could be qualified as ‘ordinary’ implementation issues as they relate to the measures taken to comply with a DSB ruling and examine their consistency under the relevant Agreement. These claims are unrelated to the question of zeroing as it is undisputed that, in these redeterminations, the United States indeed no longer used ‘zeroing’.

In addition, the EU challenged a number of reviews that were undertaken by the United States as part of the normal ongoing life of the anti-dumping orders originally challenged by the EU. The EU thus claimed violations in respect of subsequent administrative reviews, changed-circumstances reviews, and sunset reviews adopted in relation to the 15 original investigations and the 16 administrative reviews at issue in the original proceedings, as well as liquidation and assessment instructions and final liquidation of duties resulting from those subsequent reviews. The EU argued that all of these subsequent reviews undertaken in the context of AD orders of which the original determination or the administrative review were found to be WTO-inconsistent, were sufficiently related to the challenged measures to be considered as ‘measures taken to comply’. Since the United States continued to use zeroing in these subsequent reviews, the EU argued that the United States continued to violate the AD Agreement. Similarly, the EU argued that any liquidation instructions based on a determination vitiated by zeroing that were made after the expiry of the RPT were WTO-inconsistent actions. Finally, the EU argued that the United States, by omission, had failed to implement the DSB's rulings. In particular, the EU noted that there was a gap between the end of the RPT for implementation and the adoption of some of the Section 129 determinations through which the United States sought to implement the WTO rulings.

The United States considered it had fully implemented the DSB rulings by amending, through several so-called ‘Section 129 determinations’, the challenged original determinationsFootnote ¹² and reviewsFootnote ¹³ and by making an effective end to the use of model zeroing in original investigations.Footnote ¹⁴ According to the United States, only these Section 129 determinations are ‘measures taken to comply’ in respect of which the Panel has a mandate under article 21.5 of the DSU. The United States acknowledged that the USDOC issued, in the ordinary course, administrative-review determinations with respect to anti-dumping duty orders relating to the original investigations at issue in the original proceedings. The USDOC continued to apply zeroing when calculating margins of dumping in those administrative reviews.Footnote ¹⁵ It also acknowledged that sunset reviews were conducted with respect to some of the measures at issue in the original proceedings. Twelve sunset-review determinations of these measures resulted in the continuation of the relevant AD duty order.Footnote ¹⁶ According to the United States, these subsequent reviews should not be considered as ‘measures taken to comply’ with the DSB rulings.

US–Zeroing (Japan)

Japan considered that the United States had failed to comply with the recommendations and rulings of the DSB. It argued that the United States had done nothing to implement the DSB rulings in respect of the ‘as such’ claims relating to the zeroing methodology in the context of transaction-to-transaction comparisons in original investigations, and under any comparison methodology in periodic and new-shipper reviews. It further alleged that the United States had not amended the importer-specific assessment rates in the context of five periodic reviews that were found to be WTO-inconsistent in the original proceedings. In addition, Japan considered that the United States committed the same zeroing violation in the case of four subsequent periodic reviews, one of which was only adopted in the course of the Panel's proceedings, which Japan considered to be ‘measures taken to comply’. Japan also argued that the United States failed to bring into compliance the sunset-review determination of 4 November 1999, which was found to be vitiated by the use of zeroing. Finally, Japan argued that by continuing with certain liquidation actions relating to reviews that were completed before the end of the RPT, but where such actions were taken after the expiry of the RPT, the United States acted in violation of Articles II:1(a) and II:1(b) of the GATT 1994.

The United States argued it had brought its measures into compliance. It asserted that the zeroing procedures challenged ‘as such’ by Japan in the original proceeding no longer exist, as on 27 December 2006 USDOC published a final notice announcing that it would no longer apply zeroing in weighted-average-to-weighted-average comparisons in original investigations.Footnote ¹⁷ It also held that it complied regarding certain of the challenged administrative reviews by withdrawing the WTO-inconsistent cash-deposit rates with prospective effect, replacing them with new cash-deposit rates determined in subsequent administrative reviews. The United States denied it was required to take any compliance action in respect of the importer-specific assessment rates determined in several administrative reviews. In addition, it asserted that a number of reviews challenged by Japan that were not challenged in the original proceedings should not be considered as ‘measures taken to comply’ within the meaning of Article 21.5 of the DSU, and therefore fell outside the scope of these proceedings. It also argued that a review which was not terminated at the time Japan requested the establishment of the implementation Panel was not within the Panel's terms of reference. Finally, the United States asserted that it was not required to take any action to comply with the DSB's recommendations and rulings regarding the 4 November 1999 sunset review, because the relevant likelihood of dumping determination continues to be based on a number of dumping rates not called into question by the findings of the Appellate Body.Footnote ¹⁸

US–Zeroing (EC)

In its Report on US–Zeroing (EC), the Panel upheld some, but not all of the EU's claims. As to what actions are to be considered as ‘measures taken to comply’, the Panel noted that ‘a nexus-based analysis, as articulated in Australia–Leather II (Article 21.5 – US), Australia–Salmon (Article 21.5 – Canada), and US–Softwood Lumber (Article 21.5 – Canada) is useful in examining which measures challenged by a complainant properly fall within the scope of an Article 21.5 proceeding’. It thus found that if the ‘contested measures are closely connected with the measures at issue in the original dispute or with the steps taken by the Member to implement the DSB's recommendations and ruling’, they are measures taken to comply.Footnote ¹⁹

The Panel examined the links, ‘in terms of their nature and of their effects’, that exist in general between the subsequent reviews challenged by the European Communities and the measures at issue in the original dispute and the recommendations and rulings of the DSB. It found that ‘successive determinations of different types are made in the context of a single trade remedy proceeding, involving the imposition and assessment of anti-dumping duties on imports of a particular subject product, from the same country. They all concern the imposition and collection of anti-dumping duties under a particular anti-dumping order. In this sense, these determinations form part of a continuum of events and measures that are all inextricably linked.’Footnote ²⁰ It then considered the links, ‘in terms of the timing of the determinations at issue’ and found that ‘only those subsequent reviews that were decided after such [DSB] adoption could be taken into consideration as part of a compliance panel's examination of the implementation of DSB recommendations and rulings’.Footnote ²¹ In other words, the Panel was of the view that only ‘measures adopted following the adoption of the DSB's recommendations may have a close link with the DSB's recommendations and rulings and with the steps, if any, taken by the implementing Member to achieve compliance with the recommendations and rulings, and therefore warrant inclusion in the scope of an Article 21.5 proceeding’.Footnote ²² This is so, according to the Panel, because ‘the application of a nexus-based test should primarily aim at bringing within the scope of the compliance dispute measures that potentially circumvent implementation or undermine measures officially taken to comply’. Therefore, ‘none of the subsequent reviews challenged by the European Communities that were decided before the adoption of the DSB's recommendations and rulings fall within our terms of reference’.Footnote ²³

Having established the measures that fell within its terms of reference, the Panel examined the relevant claims of the EU in respect of such measures. It agreed with the EU that the United States had violated its WTO obligations by making determinations of the final amount of duty assessment after the end of the RPT in certain administrative reviews, and by continuing to apply cash-deposit rates to certain imports established in an administrative review that was found to be WTO-inconsistent. Liquidation of anti-dumping duties calculated with zeroing pursuant to final duty-assessment determinations made before the end of the RPT was deemed not to be WTO-inconsistent. In essence, the Panel was of the view that any final determination in a related periodic review made after the end of the RPT must comply with the WTO prohibition on zeroing, even when it relates to imports that entered the country before the end of the RPT. However, according to the Panel, when the final determination using zeroing is made before the end of the RPT and thus at a point in time that the United States's violation was not subject to a sanction, it is irrelevant that the execution of that determination through liquidation instructions takes place after the end of the RPT. According to the Panel, ‘the relevant date for implementation of DSB recommendations and rulings concerning anti-dumping duties by a Member operating a retrospective duty assessment system is the date of the final determination of liability’.Footnote ²⁴

The Panel further found that the United States had violated Article 3 of the AD Agreement by maintaining the AD duty orders, in certain cases, without having made a new determination of injury and of the volume of dumped imports even though the redeterminations without zeroing led to the exclusion of a number of exporters that were no longer considered to have been dumped. However, the Panel found that the United States did not act inconsistently with Article 9.4 of the AD Agreement in the establishment of ‘all others’ rates in the Section 129 determinations in the relevant cases since it was of the view that Article 9.4 does not impose any obligations on authorities when all margins are either de minimis, zero, or based on facts available. The Panel did not make findings for technical reasons on a number of other claims.

US–Zeroing (Japan)

The Panel upheld almost all of the arguments of Japan. It found that the United States had failed to comply with the DSB's recommendations and rulings regarding the importer-specific assessment rates determined in a number of administrative reviews in respect of imports that were liquidated after the expiry of the RPT. It concluded that the United States was thus in continued violation of its obligations under Articles 2.4 and 9.3 of the AD Agreement and Article VI:2 of the GATT 1994. It also found that the United States acted inconsistently with Articles 2.4 and 9.3 of the AD Agreement and Article VI:2 of the GATT 1994 by applying zeroing in the context of those administrative reviews that were not challenged in the original dispute but which could be considered as measures taken to comply with the DSB rulings and recommendations because of their close nexus with the determinations challenged in the original proceedings.

In addition, the Panel held that the United States failed to comply with the recommendations and rulings of the DSB regarding the maintenance of zeroing procedures challenged ‘as such’ in the original proceedings. It found that the United States had not implemented the recommendations and rulings in the context of transaction-to-transaction comparisons in original investigations and under any comparison methodology in periodic and new-shipper reviews. It thus found that the United States remained in violation of Articles 2.4, 2.4.2, 9.3, and 9.5 of the AD Agreement and Article VI:2 of the GATT 1994. It also found that the United States was in violation of Articles II:1(a) and II:1(b) of the GATT 1994 with respect to certain liquidation actions taken after the expiry of the RPT, even when they were based on reviews terminated prior to the end of the RPT. The Panel thus adopted a different approach from that of the Panel on US–Zeroing (EC).Footnote ²⁵ Finally, the Panel found that the United States had not complied with the DSB recommendations and rulings with respect to the 1999 sunset review.Footnote ²⁶

US–Zeroing (EC)

A number of issues were raised on appeal.Footnote ²⁷ Most importantly, the EU argued the Panel erred in finding that subsequent reviews that predated the adoption of the recommendations and rulings of the DSB did not fall within its terms of reference, because they did not have a sufficiently ‘close nexus’ with the original measures at issue and the recommendations and rulings of the DSB. The EU also considered that the Panel erred by not extending the US compliance obligations to actions consequent to the assessment of duties, including the collection or liquidation of duties occurring after the end of the RPT when they related to administrative-review determinations completed before that date.

In addition, the EU challenged a number of other aspects of the Panel's findings, such as the Panel's refusal to make a finding in respect of measures taken to comply after the end of the RPT but before the request for establishment of an implementation Panel was submitted by the EU. The EU also challenged the Panel's refusal to examine an alleged calculation error in a measure taken to comply for reason of the fact that this error could have been challenged in the original dispute as it had also been present in the original measure. It also contested the Panel's findings in respect of the lack of disciplines in respect of the ‘all others’ rates when all margins are determined based on zeroing or are de minimis.

The United States challenged a number of the Panel's findings, most importantly the ‘close nexus’ test developed by the Panel that formed the basis for the inclusion of a significant number of reviews that had not been challenged in the original dispute since they postdated the original request for establishment.

US–Zeroing (Japan)

On 20 May 2009, the United States filed a Notice of Appeal,Footnote ⁵⁶ challenging several aspects of the Panel Report. First, the United States argued that a review which was not yet terminated at the time of Japan's request for establishment of a Panel was not within the Panel's terms of reference. The United States considered that the general terms ‘any subsequent closely connected measures’ was not sufficiently specific in the sense of Article 6.2 DSU. In addition, the United States argued that this review was a ‘future measure’ that had not yet come into existence at the time of Japan's Panel request, and therefore could not be considered to be part of the Panel's terms of reference.

Second, in respect of the determination of the relevant moment in time for determining compliance, the United States considered that the Panel's basic error consisted of using the date of collection of duties as the determining factor for assessing whether the United States had brought itself into compliance. According to the United States, it is rather the date of importation that prevails. This implies that reviews relating to imports that entered the United States prior to the expiration of the RPT and that apply zeroing are not inconsistent with the US obligations, even when this leads to the collection of duties after the RPT. In addition, the United States submits that, even if the date of liquidation was relevant for assessing compliance, liquidation actions that take place after the RPT as a result of domestic litigation cannot provide a basis for a finding of noncompliance. Relying on the AB Report in US–Zeroing (EC) (EC – Article 21.5), the United States further maintains that the liquidation actions that have been delayed as a result of domestic litigation cannot be said to ‘derive mechanically’ from the challenged periodic reviews, and therefore cannot be deemed to be WTO-inconsistent.Footnote ⁵⁷ In addition, the United States challenges the finding concerning Reviews 4, 5, and 6, on the grounds that these reviews did not have effects after the expiration of the RPT because assessment of duties calculated in these reviews was enjoined prior to the conclusion of the RPT and continues to be enjoined.

Periodic reviews as ‘measures taken to comply’

The main issue in this case related to the manner in which the United States was supposed to have implemented adverse rulings relating to zeroing in anti-dumping investigations. The implementation issues that arose are specific to the United States, which is the only country that uses a retrospective anti-dumping system rather than a prospective anti-dumping system. To put it simply, in a retrospective system, the original investigation into dumping, injury, and the causal link will allow for the imposition of anti-dumping measures and will set the amount of cash deposits for imports entering the country following imposition of the measure. The actual amount of duties to be paid, however, will normally only be determined in the course of a periodic (yearly) review of the measure. At the time of the review, it will be examined whether, in the course of the past review period, imports were actually entering the country at dumped prices and what the level of the dumping margin was. Based on the level of the margin of dumping – which may well be higher than the margin of dumping determined at the time of the original investigation – a definitive determination of the amount of duties that are due will be made and on that basis the duties will be collected (‘liquidated’). At the same time, a cash-deposit requirement will be imposed for future imports entering the country, the level of which is set on the basis of the margin of dumping that was determined at the time of the review. And that story repeats itself until the end of the normal five-year life cycle of the anti-dumping order when a sunset review may be conducted to examine whether there is a likelihood of continuation or recurrence of dumping causing injury if the duty were removed.

The EU and Japan challenged a number of original determinations and a number of periodic reviews in which dumping was determined by the US authorities using the WTO-inconsistent zeroing methodology. Following the original dispute in which the EU and Japan prevailed, the United States revisited the dumping calculations of the original investigations to determine whether, even without zeroing, dumping existed. If so, the measure continued to exist. The United States considered that it was not required to do anything in case subsequent periodic reviews had superseded the reviews that had originally been challenged. And it thus continued to zero in the course of these subsequent reviews. Obviously, this type of implementation was considered to be completely meaningless to the EU and Japan since, at the end of the day, their exporters were still required to pay duties based on margins of dumping determined through the use of zeroing.

The EU and Japan argued that these subsequent reviews in which the United States continued to zero were ‘measures taken to comply’ with the DSB ruling and could thus be examined by the implementation Panel in the context of an Article 21.5 DSU implementation dispute. As explained above, the Panel and AB agreed, applying the ‘close nexus’ test: the subsequent reviews conducted in the context of the same AD orders that had been the subject of the original dispute were sufficiently closely linked in terms of the nature, effect, and timing with the previously challenged measures that they could be seen as ‘measures taken to comply’. The AB considered that this was true even for reviews that had been initiated well before the adoption of the Reports they were supposed to be ‘complying with’.

We do not disagree with the outcome of this dispute and consider that ultimately the AB got it right. However, we do find that there is something bizarre about the AB's reasoning – how can a measure be a ‘measure taken to comply’ with a Report when that Report has not even been adopted or even issued? It is simply counterintuitive to treat reviews initiated and completed even before adoption of the Panel and AB Reports as measures ‘taken to comply’ with these Reports.

The AB explained that any measure adopted in the course of the dispute but before adoption of a Report could well impact on the country's compliance with its WTO obligations and should therefore be treated as a ‘measure taken to comply’ if a sufficiently close link existed between the originally challenged measure and the alleged measure taken to comply. However, we wonder whether there was actually any need to develop the rather complicated reasoning that the AB adopted in this case in respect of these reviews arguing that they were ‘measures taken to comply’. In our view, the problem could have been dealt with in a simple and straightforward manner, avoiding problematic statements such as the one referred to before about measures complying with Reports even before the Report was released let alone adopted.

What was the original problem? That anti-dumping duties were imposed based on a WTO-inconsistent method for determining dumping; that cash deposits were required and duties were collected on the basis of dumping margins calculated using this same methodology. We recall that GATT Article II allows for the imposition of duties and charges only up to the level of the bound tariff and that ‘other charges’ may only be imposed if they are included in the Schedule of Concessions. GATT Article II.2 allows, by way of exception, for the imposition of anti-dumping measures, as long as they are imposed in a GATT-consistent manner. Similarly, GATT Article VI and the WTO AD Agreement allow for the imposition of anti-dumping measures on imports of certain countries only, thus deviating from the general MFN rule of GATT Article I, and set forth certain disciplines in terms of the determination of the margin of dumping. The United States lost the original dispute because the ‘zeroing’ method for calculating the margin of dumping that was used in a number of anti-dumping cases was considered to be WTO-inconsistent. The United States was to correct this error and was given a reasonable period of time to do so. Upon expiry of this period of time, the question is whether the United States has complied with the ruling of the DSB that its method of determining margins of dumping and thus of imposing anti-dumping duties based on zeroing is WTO-inconsistent. It seems normal to examine the situation at the end of the RPT to determine whether duties are still in place, whether cash deposits are still required, and whether duties are still being collected in respect of the originally challenged AD orders and this based on the WTO-inconsistent method of zeroing. If that is the case – and the United States did not deny that this was indeed the case – then it seems pretty clear that the United States has failed to comply with the DSB ruling. What can be so complicated about this conclusion that would warrant a lengthy discussion about the meaning of the term ‘measures taken to comply’?

The basic question is whether there exists compliance. The imposition of charges (cash deposits) and the collection of additional import duties (anti-dumping duties) based on zeroed margins clearly demonstrate, not that the measures taken to comply were WTO-inconsistent, but that no measures were taken to comply. The existence of such GATT/WTO-inconsistent cash deposits and duties is inconsistent with the AD Agreement, GATT Article VI, and thus ultimately also with GATT Article II. Interestingly, this is also what the AB appears to ultimately find: that any action taken after the expiry of the RPT that concerns the same subject matter and is not consistent with the WTO obligations of a Member demonstrates a failure to comply with the DSB ruling dealing with that subject matter.

In our view, the AB inappropriately was led down a pseudo-technical path distinguishing between original determinations, periodic reviews, new periodic reviews, sunset reviews, prospective systems, retrospective systems, cash deposits, duty collection, duty liquidation, etc. If this had not been an anti-dumping dispute but a simple tariff violation, it is very doubtful that the AB would have complicated matters as much as it did in this case. In addition, its acceptance of these technical nonarguments required it to effectively disregard one of the three tests it had itself developed for determining whether there existed a sufficient nexus between the measure challenged and the ‘measure taken to comply’: the timing of the measure. Clearly, the sensible rationale for including the ‘timing’ of any measure as part of the analysis of whether a new measure was a measure taken to comply and thus part of the implementation Panel's terms of reference was whether it was taken following release and/or adoption of the Report. The AB in these zeroing cases rejected the idea that timing is a determinative factor but failed to explain what role, if any, this factor should play if it actually does not matter whether the measure was adopted even before the release of the Report to the parties. It seems that the three-pronged ‘nexus’ test that was developed by the AB in the context of Article 21.5 implementation disputes has effectively been reduced to a two-pronged test in these zeroing cases.

Timing of compliance – any action following the expiry of the RPT is captured

A second important clarification offered by the AB in these cases concerned the importance of the end of the RPT. We are of the view that the AB was right that the relevant moment in time for determining compliance is the end of the RPT and that any action taken after the end of the RPT that is inconsistent with a country's WTO obligations demonstrates a failure to comply. For many WTO Members, this solution is obvious and there is no reason why the same conclusion would not apply to the United States as well.

This being said, we agree with the United States that this puts it at a disadvantage in comparison with Members that use the prospective system of duty collection. Under a prospective system, imports entering the country prior to the end of the RPT will be subject to the payment of duties in a final manner and such duties will effectively escape the WTO disciplines since they were collected prior to the end of the RPT. In a retrospective system, such duties are collected only in a provisional manner. If a subsequent review that determines the final duty liability is conducted after the end of the RPT, it will be subject to the WTO disciplines and the duties may not be collected in full. In sum, it is clear that the AB's approach implies that the United States will have to accept the consequences of its retrospective system, which disadvantages the United States in this particular case. However, in many other ways, this retrospective system has important ‘advantages’ from the point of view of protecting the domestic industry. There are good reasons why domestic industry associations argue so strongly in favour of maintaining this administratively burdensome system, the usefulness of which is currently under discussion in the United States.

In this respect, it is not clear that the AB's approach does not disadvantage the retrospective system. The AB argues that in a prospective system (used everywhere else in the world), interested parties may also request refunds after the end of the RPT. However, such refund requests are rather uncommon. Moreover, we wonder whether an importer could request the refund of excess duties paid in the course of the RPT. After all, throughout the RPT a Member would be allowed to continue to act in a manner inconsistent with its WTO obligations without there being any remedy. On what basis would such a refund be requested? The Panel's position on this question seems more appropriate: that every system has its advantages and disadvantages and that each Member must accept the consequences, whether good or bad, of the system it adopts. This was essentially also the view adopted in the context of the US–FSC (Foreign Sales Corporations) dispute where, to put it somewhat simplistically, the United States suffered the negative consequences of the international tax regime that it had adopted.

A related matter of some importance concerns the impact that delays caused by domestic-court proceedings have on compliance. We consider that the AB was right to say that it does not matter why liquidation was delayed since any action taken after the expiry of the RPT has to be WTO-consistent. But why then did the AB not immediately say so when this first came before it in the Article 21.5 US–Zeroing (EC) case? After all, it is merely the logical consequence of its general approach to focus on the date of the expiry of the RPT and on the actions adopted since that time, rather than on the legal basis for the action in isolation (the determination of the margin of dumping).

In the Article 21.5 US–Zeroing (EC) case, the AB stated that it did ‘not express any opinion on the question of whether actions to liquidate duties that are based on administrative review determinations issued before the end of the reasonable period of time, and that have been delayed as a result of judicial proceedings, fall within the scope of the implementation obligations of the United States, as we do not need to do so in the context of our analysis of this issue in this case’.Footnote ⁷³ This was surprising since its entire reasoning, which focused on the fact that any action taken after the expiry of the RPT has to be WTO-consistent, clearly suggested that it had in fact expressed itself on this issue.

This is a typical example of the AB throwing in a statement to say that it is not expressing a view on a certain matter when it actually has already done so, thereby creating the impression that it may adopt a different approach with respect to this question. Why else would you add a paragraph saying that your views should not be read to imply that this or that is also inconsistent? The signal that is being given is clearly that a different conclusion may need to be drawn in such circumstances. But then, six months later when the issue is again before the AB in the Article 21.5 US–Zeroing (Japan) case, the AB does actually follow its own logic and does consider that such actions are also WTO-inconsistent irrespective of whether the delay was caused by court proceedings. Why did the AB not say that before and why did it suggest that a different outcome may actually be warranted if ‘security and predictability’ are the main guiding principles of the dispute-settlement system?Footnote ⁷⁴

This is not the first time the AB has done so. Its entire zeroing jurisprudence is the consequence of such failure to express itself clearly in respect of this issue from the first time the issue was raised before it. The AB stated that it did not express a view on whether zeroing was inconsistent with the fair-comparison requirement of Article 2.4 in, for example, US–Zeroing (EC) and stated that it was not expressing a view on whether zeroing in transaction-to-transaction comparisons is also prohibited in US–Softwood Lumber V, to later find that zeroing is unfair and that zeroing violates the fair-comparison requirement. This false restraint by the AB is not consistent with its important role in securing the predictability of the system.

The ‘all others’ rate – how to deal with a loophole?

Another missed opportunity in clarifying the rules has been the AB's reasoning on the ‘all others’ problem in this case. Article 9.4 of the AD Agreement provides that the ‘all others’ rate that applies to nonsampled exporters should not exceed the average margin of dumping, while excluding margins that are zero or de minimis or based on facts available. But what to do if all margins are either zero or de minimis or are based on facts available? That was the problem the Panel was confronted with. It reached the conclusion that in such a case there is no obligation to respect under Article 9.4. The Panel found that Article 9.4 merely ‘establish[es] a methodology for the calculation of a “ceiling” which the “all others” rate may not exceed’, but does not ‘specify a method for or imposes disciplines on the calculation of the “all others” rate itself.’Footnote ⁷⁵ According to the Panel, in cases where all the margins of dumping for the investigated exporters are zero, de minimis, or based on facts available, ‘there are simply no margins of dumping from which the investigating authority … may calculate the maximum allowable ‘all others’ rate and therefore, in such circumstances, ‘Article 9.4 simply imposes no prohibition, as no ceiling can be calculated.’Footnote ⁷⁶

The AB disagrees:

Thus, the AB considers that Article 9.4 imposes an obligation, but it refuses to clarify what this obligation is and what an authority is to do in such circumstances. It hides behind the fact that no alternative methodologies have been suggested by the parties. In our view, it would have been very useful and even necessary given the objective of providing ‘security and predictability to the multilateral trading system’ had the AB clarified its view of what, if any, obligation Article 9.4 imposes in such circumstances. Unfortunately, the AB said A without saying B.

Needed: more effective surveillance and analysis

The role of the compliance Panels that are the subject of this paper was to determine whether the United States took action (implemented new measures) to become compliant with the AD Agreement. They concluded that the United States had not done so. As of the time of writing, both the EU and Japan had launched Art. 22.6 DSU proceedings, seeking authorization to retaliate against the United States for the losses incurred as a result of noncompliance by the United States. Paralleling the limited economic analysis of the effects of zeroing discussed above, the parties have wildly different estimates of the economic impacts of US zeroing, with the EU claiming upwards of $300 million per year and the United States arguing it is closer to one-hundredth of that figure. Much will depend on the answer to the question posed above – i.e. the extent to which the use of zeroing in each case inflated the calculated dumping margin. The availability of information on this matter (more accurately, the lack thereof) will affect the feasibility of generating an accurate estimate of the amount of compensation-cum-retaliation that is appropriate.

This information problem also applies at the compliance Panel stage. A fundamental problem affecting the efficacy of the DSU is that the process of determining if compliance has occurred is left to Panels that are required to limit their remit to the issues raised in the original Panel. What may be missing – and is in the case of zeroing – is information on what is being implemented by the country found to have violated the WTO. For example, it is not clear whether the United States has in fact stopped using zeroing in original investigations. In at least one recent case, it has reportedly reverted to the practice.Footnote ⁸¹ It is generally held that the United States continues to use zeroing in reviews, but it is not necessarily true that it does so in all reviews, and we do not know for all these reviews how much the dumping margin is affected by the practice.

Zeroing affects not only EU and Japanese exporters. The first zeroing cases were brought by India and Brazil (against the EU), and, as noted above, more recently Thailand, Mexico, Korea, and Vietnam have now brought cases against the United States. It is rather inefficient if continued noncompliance by the United States results in recurring cases that are all dealing with an issue that has already been litigated. Rather than force individual exporting countries to initiate their own disputes, imposing unnecessary costs on the system and the countries concerned, it would be much better that if the United States has no intention of complying, the focus be put on compensating affected trading partners. Given that the US methodology is (can be) applied to all exporters, it would be more efficient to address this matter in a cooperative manner, rather than through a recurring series of cases that impose an excess burden on the DSB and on affected exporting countries and their exporters.Footnote ⁸²

Of course, the best way to resolve the matter would be for the United States to bring its measures into compliance. At the time of writing, the United States is reportedly still working towards this – which can be inferred from the decision in early September 2010 by the EU to request that the Art. 22.6 arbitration be suspended for a one-year period to provide the United States more time to reform its AD regime.Footnote ⁸³ In the absence of full compliance, one way to facilitate either an agreement with exporters on compensation or to reduce the costs to the WTO system of continued litigation is to make available information on a country-by-country basis how much dumping margins have been affected by zeroing. That information can then be used to provide the affected countries with compensation, or be the basis for retaliatory action by the affected countries. At some point, if enough countries retaliate the United States may be induced to change its policy. However, as noted by Bown and Prusa (Reference Bown and Prusa2010) the amounts involved – even though subject to great uncertainty at this point because of the lack of data – may not be large enough to have much of an impact on US exporters and thus to change the political-economy equilibrium in the United States on zeroing. But a system of transparent compensation-cum-retaliation based on the actual effects of zeroing would be better than another possible reaction to continued noncompliance by the United States: emulation by trading partners.

There are some parallels between the zeroing saga and the use of quantitative restrictions in the 1970s and 1980s that resulted in the Mutifibre Arrangement (MFA). Quantitative restrictions on textiles and clothing – which clearly violated the GATT but were accommodated by the system – emerged gradually but became widespread over time. A similar phenomenon could arise with zeroing, except in reverse, with developing countries adopting the practice. This would be very unfortunate.Footnote ⁸⁴ The textiles episode generated a GATT Textile Surveillance Body to monitor the implementation of the MFA and a Textiles Monitoring Body that was responsible for surveillance of all measures taken under the Uruguay Agreement on Textiles and Clothing (ATC) to liberalize trade in textiles and clothing. Agreement to establish a multilateral surveillance mechanism that documents the continued use by the United States of zeroing could increase the transparency of US policy in this area and help in quantifying the likely effects of zeroing on dumping margins. Such information would facilitate the work of both compliance and arbitration Panels, and help to inform the policymaking process in the United States.

The need for greater transparency goes beyond zeroing. AD has become by far the most frequently used contingent trade-policy instrument by WTO members and there is only limited multilateral surveillance of what countries are doing and what the effects are. The conventional wisdom is that the aggregate effects of AD are small. Egger and Nelson (Reference Egger and Nelson2010) estimate that between 1960 and 2000 anti-dumping actions reduced imports in the countries applying these measures by 1.6 percentage points. It has long been pointed out by economists that such aggregate estimates ignore both the incidence of actions – the targeted exporters may be severely affected even if on average most are not – and the chilling effects of the uncertainty of market-access conditions created by the threat of AD.Footnote ⁸⁵ A recent empirical analysis that explicitly considers the counterfactual – how much trade would be expected to be observed without the threat of AD – by incorporating information on the date when a country adopts AD legislation concludes that use of AD measures reduced imports by some 6–7% in active users, such as Brazil and India (Vandenbussche and Zanardi, Reference Vandenbussche and Zanardi2010). However, the results also show that the reduction in imports resulting from AD actions in major users (which constitute a relatively small number of countries, albeit ones that are large) only partially offsets the increase in trade that resulted from overall liberalization of trade policy during the sample period in the countries considered. The estimated offsets ranged between one-sixth and one-half.

These results provide support for the notion that AD can be a useful (political) safety valve that allows countries to implement and sustain more general trade liberalization. As noted by Egger and Nelson (Reference Egger and Nelson2010) this is an argument that has a long pedigree in the economic literature, dating back to Viner (Reference Viner1923). Finger and Nogués (Reference Finger and Nogués2006) provide case-study evidence how AD and similar measures were used in Latin America as a proliberalization, safety-valve device. However, the trend in the prevalence of AD is clearly upward, with the 2009–2010 period seeing an additional boost in response to calls by firms for assistance to cope with the global economic crisis (Bown, Reference Bown, Chad2010). The spread of AD may generate incentives for ‘retaliation’ – tit-for-tat responses by targeted countries – and this may serve to constrain the growth of AD. There is some evidence that retaliation threats are one reason for the use of AD by major exporters, and that it may help restrain the use of AD by importing countries (Blonigen and Bown, Reference Blonigen, Bruce and Bown2003; Feinberg and Reynolds, Reference Feinberg and Reynolds2006).

Whatever the dynamics may turn out to be, one conclusion we draw from the continuing zeroing saga is that greater transparency would be beneficial. What the WTO does in terms of surveillance is very narrow – essentially it is limited to a biannual compilation of AD investigations initiated and measures imposed without detail on trade volumes affected and no information on the methodologies used to determine dumping margins. The Temporary Trade Barriers database created by Chad Bown offers greater detail in terms of information on products affected and the amount of trade involved,Footnote ⁸⁶ but there is no effort currently to collect and report information on how margins were calculated. This is a major task that in practice will require the cooperation of the investigating authorities. As far as zeroing is concerned, if the United States were to decide to continue to zero in reviews and to go down the renegotiation/compensation route, such an exercise would need to be done just once, assuming that continued zeroing is limited to the existing stock of AD orders – i.e. to the reviews. Whatever the eventual outcome of this extended litigation and the decision of the United States regarding the use of zeroing, greater surveillance and information will be beneficial to the system.