1.6 Argentina; Brazil; China; Hong Kong, China; India; Japan; Korea, Rep. of; Mexico; Norway; Chinese Taipei; and Turkey reserved their third-party rights.

C. PANEL PROCEEDINGS

1.7 On 7 February 2005, the Panel received an amicus curiae brief from the Committee to Support US Trade Laws ("CSUSTL"). Recalling the statement of the Appellate Body that a panel has "the discretionary authority either to accept and consider or to reject information and advice submitted to it, whether requested by a panel or not",⁴ the Panel invited the parties and third parties to express their views on how it should handle CSUSTL's brief. After reviewing the comments of the parties and third parties, the Panel decided that it would not further consider the arguments in the CSUSTL brief except to the extent that the parties reflected those arguments in their written submissions and/or oral statements.⁵

1.8 The Panel met with the parties on 16-17 March 2005 and on 26-27 April 2005. The Panel met with the third parties on 17 March 2005.

1.9 The Panel submitted the Interim Report to the parties on 4 August 2005. The Panel submitted the Final Report to the parties on 28 September 2005.

II. FACTUAL ASPECTS

2.1 This dispute concerns the application by the United States of the so-called zeroing methodology when determining dumping margins in anti-dumping proceedings, including proceedings resulting in the imposition of anti-dumping measures and proceedings relating to the collection of anti-dumping duties.

2.2 The imposition and collection of anti-dumping duties in the United States can be broadly described in the following terms. In order to determine whether the imposition of anti-dumping measures on known exporters of a product under consideration may be justified, the United States examines whether dumping existed during a given period of investigation. This determination is made by the United States Department of Commerce ("USDOC") and is published in a Notice of Final Determination of Sales at Less Than Fair Value. The Notice of Final Determination of Sales at Less Than Fair Value sets out USDOC's assessment of the existence and level of dumping. The United States International Trade Commission ("USITC") will then determine whether the relevant United States industry was injured by reason of the dumped imports. When USDOC finds dumping and USITC finds that such dumping caused injury to the domestic industry, USDOC issues a Notice of Antidumping Duty Order imposing final measures, including a cash-deposit rate equivalent to the margin of dumping found to exist for each known exporter.

2.3 When calculating the magnitude of any margin of dumping in order to determine whether the imposition of anti-dumping measures on known exporters of a product under consideration may be justified, the United States uses a method that the European Communities refers to as model zeroing. The investigating authority, as well as determining the overall product scope of the proceeding (also referred to as subject product or subject merchandise)⁶, will, in applying the weighted average-to-weighted average comparison method, identify those sales of sub-products in the United States considered "comparable", and will include such sales in an "averaging group".⁷ An averaging group consists of merchandise that is identical or virtually identical in all physical characteristics.⁸ Each category of sub-product within the subject merchandise is assigned a control number, or CONNUM.⁹ The weighted-average-to-weighted-average comparison between normal value and export price is made within each averaging group.¹⁰ The amount by which normal value exceeds export price is considered by the United States to be a "dumping margin"¹¹ or dumped amount – referred to by the United States as the Potential Uncollectible Dumping Duties, or PUDD.¹² If export price exceeds normal value (the margin is negative), the "dumping margin" or dumped amount or PUDD for that averaging group is considered to be zero. The margin of dumping for the overall product is calculated by combining the averaging group results. The total of the dumped amounts or PUDDs (excluding the negative amounts or treating them as zero) is expressed as a percentage of the total export prices (including all averaging groups). An averaging group, as well as consisting of merchandise that is identical or virtually identical in all physical characteristics, also consists of merchandise that is sold at the same level of trade. In identifying sales to be included in an averaging group, the United States also takes into account, where appropriate, the region of the United States in which the merchandise is sold, and such other factors as are considered relevant.¹³

2.4 Subsequently, the United States will assess the liability for anti-dumping duties on specific entries of the subject product by individual importers. The United States system of duty assessment operates on a retrospective basis. Under this system, an anti-dumping duty liability attaches at the time of entry, but duties are not actually assessed at that time. Rather, the United States collects security in the form of a cash deposit at the time of entry, and determines the amount of duties due on the entry at a later date. Specifically, once a year (during the anniversary month of the orders) interested parties may request a review to determine the amount of duties owed on entries made during the previous year.¹⁴ The amount of anti-dumping duties owed by each individual importer (the assessment rate) is calculated on the basis of a comparison of each individual import to a contemporaneous average normal value. The total amount of dumping associated with each importer is then aggregated and expressed as a percentage of that importer's United States imports. This assessment rate is then applied to imports during the period reviewed. The amount of dumping found on all imports from a given exporter (regardless of the importer) is also used to derive a cash-deposit rate that will apply on future entries from that exporter. If no review is requested, the cash deposits made on the entries during the previous year are automatically assessed as the final duties. The final anti-dumping duty liability for past entries and the new cash-deposit rate for future entries is calculated by USDOC and published in a Notice of Final Results of Antidumping Duty Administrative Reviews.

2.5 When calculating the magnitude of any margin of dumping for the purpose of assessing an importer's final liability for paying anti-dumping duties and any future cash-deposit rate, the United States normally uses the average-to-transaction method¹⁵ and applies, what the European Communities refers to as, simple zeroing. When comparing a weighted-average normal value with an individual export transaction, the amount by which normal value exceeds export price is considered to be the "dumping margin"¹⁶ or "dumped" amount for that export transaction.¹⁷ If export price exceeds normal value (the margin is negative), the "dumping margin" or "dumped" amount for that export transaction is considered to be zero. The overall margin of dumping is calculated by combining the results of each comparison. The total dumping amount (excluding the negative amounts or treating them as zero) is expressed as a percentage of the total export price (including all export transactions). The importer is liable to pay the final anti-dumping duty. USDOC sends appraisement instructions to the US Bureau of Customs and Border Protection ("USCBP"), determining "an assessment rate" and thus the final anti-dumping duty to be paid.¹⁸ When USDOC gives percentage instructions, it calculates an assessment rate, to be applied by USCBP to all entries during the relevant period.¹⁹ For the purpose of calculating an assessment rate, USDOC divides the total margin or amount of dumping calculated during the period of review, by the entered values of the products sold.²⁰

2.6 The European Communities challenges certain United States legal instruments, procedures, methodologies and practice, "as such" and "as applied". In the 15 "as applied" cases, referred to by the European Communities as "original investigations", the challenged measures are: the 15 Notices of Final Determinations of Sales at Less Than Fair Value, including any amendments, and including all the Issues and Decision Memoranda to which they refer, and all the Final Margin Program Logs and Outputs to which they in turn refer, for all the firms investigated; each of the 15 Anti-dumping Duty Orders; each of the assessment instructions issued pursuant to any of the 15 Anti-dumping Duty Orders; and each of the USITC final injury determinations.²¹ In the 16 "as applied" cases, referred to by the European Communities as "periodic reviews", the challenged measures are: the 16 Notices of Final Results of Antidumping Duty Administrative Reviews, including any amendments, and including all the Issues and Decision Memoranda to which they refer, and all the Final Margin Program Logs and Outputs to which they in turn refer, for all the firms investigated; and each of the assessment instructions issued pursuant to any of the 16 Notices of Final Results.²² The legal instruments, procedures, methodologies and practice challenged by the European Communities "as such" are the following:

(a) Sections 731, 751(a)(2)(A)(i) and (ii), 771(35)(A) and (B), and 777A(d) of the Tariff Act of 1930, as amended;

(b) Section 351.414(c)(2) of the United States Department of Commerce Regulations;

(c) certain provisions of the "1997 edition of the Import Administration Anti-Dumping Manual";

(d) the "Standard AD Margin Program", which includes the "Standard Zeroing Procedures"; and

(e) the United States practice or methodology of zeroing.²³

2.7 The European Communities considers the above legal instruments, procedures, methodologies and practice, as well as each of the identified measures at issue in the 31 "as applied" cases, to be inconsistent with the United States' obligations under the AD Agreement, the GATT 1994 and the WTO Agreement.

2.8 The European Communities sets out its claims and arguments with respect to the 31 "as applied" cases by first presenting certain general facts and evidence of relevance to each case. Secondly, the European Communities presents its claims and arguments with respect to the specific facts of two sample "as applied" cases. It then alleges that the same claims and arguments apply, mutatis mutandis, with respect to the measures at issue in the 29 other determinations it is challenging.

2.9 The first sample case presented by the European Communities is with respect to the Notice of Final Determination of Sales at Less Than Fair Value regarding stainless steel bar from Italy, issued by the USDOC on 23 January 2002²⁴, including the related Notice of Antidumping Duty Order issued by USDOC on 7 March 2002.²⁵ The rates of ad valorem anti‑dumping duty applied in this case were 2.50 per cent for Acciaierie Valbruna Srl/Acciaierie Bolzano D.p.A, 7.07 per cent for Acciaiera Foroni SpA, 3.83 per cent for Rodacciai S.p.A, 33 per cent for Cogne Acciai Speciali Srl and 3.81 per cent for all others.²⁶

2.10 In calculating the dumping margins applied in this case, the USDOC identified sub-groups of products within the product under investigation ("averaging groups") on a per model basis. Within each of the averaging groups, a weighted average export price was established and compared to the corresponding weighted average normal value. The results of these comparisons on an "averaging group" basis were added up to establish the dumping margin of the product under investigation as a whole for each individual exporter. In this process, any negative margins or negative amounts of "dumping" resulting from the comparison of weighted average normal values with weighted average export prices on an "averaging group" basis were, through application of the "Standard Zeroing Procedures", equated with zero. The European Communities describes this methodology as model zeroing.²⁷ Exhibits EC-2 to EC-15 include the final determinations and facts with respect to 14 other Notices of Final Determinations of Sales at Less Than Fair Value and Notices of Antidumping Duty Orders made between 24 July 1996 and 24 April 2003 where the same methodology described above, including the "Standard Zeroing Procedures", was applied.²⁸

2.11 The second sample case presented by the European Communities is with respect to the Notice of Final Results of Antidumping Duty Administrative Reviews regarding ball bearings from Italy, issued by the USDOC on 30 August 2002.²⁹ The dumping margin calculated by USDOC in these proceedings was 3.70 per cent for SKF Industrie SpA and 1.42 per cent for FAG Italia SpA.

2.12 In calculating the relevant dumping margins in this case, the USDOC identified sub-groups of products of normal vale sales based on physical characteristics and level of trade. Within each of the sub-groups, a weighted average normal value was compared to individual export prices. When adding up the results of the comparisons to determine the total amount or margin of dumping of the product under investigation, any negative amounts of "dumping" were, through application of the "Standard Zeroing Procedures", counted as zero. The European Communities describes this methodology as simple zeroing.³⁰ Exhibits EC-17 to EC-31 include the final determinations and facts with respect to 15 other Notices of Final Results of Antidumping Duty Administrative Reviews issued between 26 October 2001 and 24 March 2003 where the same methodology described above, including the "Standard Zeroing Procedures", was applied.³¹

III. REQUESTS FOR FINDINGS AND RECOMMENDATIONS

A. THE EUROPEAN COMMUNITIES

3.1 The European Communities requests the Panel to make the following findings:³²

(a) As a consequence of the use of model zeroing in the original investigations identified in Exhibits EC-1 to EC-15, the United States acted inconsistently with Articles 2.4 and 2.4.2 of the AD Agreement and Articles 3.1, 3.2, 3.5 and 5.8 and Articles 9.3, 1 and 18.4 of the AD Agreement; Articles VI:1 and VI:2 of GATT 1994; and Article XVI:4 of the WTO Agreement. The European Communities states that the Panel need make no findings with regard to Articles 3.1, 3.2 and 3.5 of the AD Agreement where the corrected margin of dumping for individual exporters is not less than 2 per cent; and Article 5.8 of the AD Agreement in cases where the corrected margin of dumping for the exporting country is not less than 2 per cent.

(b) The "Standard Zeroing Procedures" used in original investigations (or the United States practice or methodology of zeroing) and Sections 771(35)(A) and (B), 731 and 777A(d) of the Tariff Act are "as such" inconsistent with Articles 2.4 and 2.4.2 of the AD Agreement and Articles 5.8, 9.3, 1 and 18.4 of the AD Agreement; Articles VI:1 and VI:2 of GATT 1994; and Article XVI:4 of the WTO Agreement.

(c) As a consequence of comparing a weighted average normal value with individual export transactions, without explanation or justification, and the use of simple zeroing in the periodic reviews listed in Exhibits EC-16 to EC-31, the United States acted inconsistently with Articles 2.4, 2.4.2 and 9.3 of the AD Agreement and Articles 11.1, 11.2, 1 and 18.4 of the AD Agreement; Articles VI:1 and VI:2 of GATT 1994; and Article XVI:4 of the WTO Agreement.

(d) The "Standard Zeroing Procedures" used in periodic reviews (or the United States practice or methodology of zeroing) and Sections 771(35)(A) and (B), 731, 777A(d) and 751(a)(2)(A)(i) and (ii) of the Tariff Act and Section 351.414(c)(2) of the Regulations are "as such" inconsistent with Articles 2.4, 2.4.2 and 9.3 of the AD Agreement and Articles 11.1, 11.2, 1 and 18.4 of the AD Agreement; Articles VI:1 and VI:2 of GATT 1994; and Article XVI:4 of the WTO Agreement.

(e) The "Standard Zeroing Procedures" used or relied upon in new shipper, changed circumstances reviews and sunset reviews (or the United States practice or methodology of zeroing) and Sections 771(35)(A) and (B), 731, 777A(d) and 751(a) (2)(A)(i) and (ii) of the Tariff Act and Section 351.414(c)(2) of the Regulations are "as such" inconsistent with Articles 2.4, 2.4.2, 9.3, 9.5, 11.1, 11.2 and 11.3 of the AD Agreement and Articles 1 and 18.4 of the AD Agreement; Articles VI:1 and VI:2 of GATT 1994; and Article XVI:4 of the WTO Agreement.

B. UNITED STATES

3.2 The United States requests that the Panel reject the claims of the European Communities in their entirety.³³

IV. ARGUMENTS OF THE PARTIES

A. BURDEN OF PROOF AND STANDARD OF REVIEW

1. Burden of Proof

(a) The United States³⁴

4.1 The United States submits that the AD Agreement imposes obligations on investigating authorities that they must satisfy, but the burden of proving that those obligations have not been satisfied is on the complaining party. Accordingly, in this dispute, the burden is on the European Communities to prove that the United States acted in a WTO-inconsistent manner with respect to both its "as applied" and its "as such" claims. The burden is not on the United States to prove that it acted in a WTO-consistent manner.

(b) The European Communities³⁵

4.2 The European Communities argues that it has presented facts and evidence more than sufficient to make out a prima facie case, in relation to both its "as applied" and "as such" claims, and has fully explained what specific obligations in the AD Agreement it considers the United States has failed to comply with, and why. In the light of these circumstances, the European Communities argues that it is incumbent on the United States to substantiate its defence, failing which the claims of the European Communities must prevail. In particular, the European Communities submits that, where the United States has not contested the facts and adduced evidence, the Panel must base itself on the facts and evidence presented by the European Communities. The European Communities also notes that the United States' first written submission did not contest claims nor respond to arguments presented by the European Communities in relation to the "as applied" cases in original investigations. The European Communities further argues that the burden of establishing that the phrase "the existence of margins of dumping during the investigation phase" has the special meaning asserted by the United States, as provided for in Article 31(4) of the Vienna Convention, falls on or has been shifted to the United States, and has not been discharged by the United States.

2. Standard of Review

(a) The United States³⁶

4.3 The United States recalls that the standard of review with respect to a panel's task of assessing an investigating authority's establishment and evaluation of facts, and its interpretation of the provisions of the AD Agreement, is set forth in Articles 17.6(i) and 17.6(ii) of the AD Agreement. The United States highlights that the standard of review described in Article 17.6(i) has been interpreted by several panels as prohibiting a panel from engaging in de novo review. As regards the standard of review expressed in Article 17.6(ii), the United States asserts that the question it poses is whether an investigating authority's interpretation of the AD Agreement is a permissible interpretation. According to the United State, Article 17.6(ii) acknowledges that there may be provisions of the AD Agreement that "admit[] of more than one permissible interpretation." Where that is the case, and where the investigating authority has relied upon one such interpretation, a panel is to find that interpretation to be in conformity with the AD Agreement.

4.4 For the reasons discussed elsewhere, interpreting "during the investigation phase" under Article 2.4.2 as referring to an investigation under Article 5 is a permissible interpretation. Likewise, interpreting Article 2.4 as containing no independent and overarching requirements with respect to offsetting and symmetry is a permissible interpretation. Therefore, because the challenged measures rest on permissible interpretations, Article 17.6(ii) of the AD Agreement requires the Panel to find that those measures are in conformity with the AD Agreement. As found by the panel in Argentina – Poultry, "[I]n accordance with Article 17.6(ii) of the AD Agreement, if an interpretation is "permissible", then we are compelled to accept it."³⁷

(b) The European Communities³⁸

4.5 The European Communities argues that the submissions of the United States on the question of the standard of review for facts do not appear to serve any useful purpose. In particular, the European Communities argues that it has presented facts and evidence more than sufficient to make out a prima facie case, which facts and evidence have not for the most part been contested by the United States. Furthermore, the submissions of the United States are irrelevant to the "as such" claims, with respect to which the facts concern United States municipal anti-dumping law "as such", and have not been "established" by an investigating authority.

4.6 Second, on the questions of law, the European Communities argues that it is highly significant that the United States refers, in its first written submission, to "the possibility that customary rules of interpretation would not always yield definitive meanings of particular provisions of the AD Agreement."³⁹ The European Communities agrees with this statement of the United States, and ascribes to it particular relevance in the context of its claims with respect to Article 2.4.2. In particular, the European Communities argues that the statement effectively amounts to an admission from the United States that, after applying customary rules of interpretation of international law to the interpretation of the word "investigation" in Article 2.4.2, the conclusion must be that that word does not have the limited or defined meaning argued for by the United States in these panel proceedings. This being so, for a Member to give that word such a limited or defined meaning, and apply the AD Agreement accordingly, would not be a permissible interpretation of Article 2.4.2.

B. "AS APPLIED" CLAIMS

1. Claims with respect to Article 2.4 of the AD Agreement

(a) The European Communities⁴⁰

4.7 The European Communities argues that, because of the use of model zeroing, each of the measures at issue in the cases identified in Exhibits EC-1 to EC-15 (described by the European Communities as "original investigations"), violated United States' obligations under Article 2.4 of the AD Agreement; and that because of the use of simple zeroing, each of the measures at issue in the cases identified in Exhibits EC-16 to EC-31 (described by the European Communities as "periodic reviews"), similarly violated United States' obligations under Article 2.4 of the AD Agreement. The European Communities presents essentially the same claims and arguments, mutatis mutandis, with respect to both "original investigations" and "periodic reviews", namely, that, in each of the "as applied" cases, the United States violated (first claim) the first sentence of Article 2.4 because model zeroing and simple zeroing (absent targeted dumping) are unfair; and (second claim) the third to fifth sentences of Article 2.4 because model zeroing and simple zeroing (absent targeted dumping) effectively result in an impermissible adjustment to export price.

4.8 The European Communities contends that Article 2.4 of the AD Agreement establishes an overarching and independent obligation to make a fair comparison between export price and normal value. In contrast with the Tokyo Round Anti-Dumping Code, in the Uruguay Round AD Agreement, the words "fair comparison … between the export price and the normal value" were lifted up and placed on their own in a new first sentence of Article 2.4. According to the European Communities, the drafters would not have done this without a purpose. Thus, the change introduced in the Uruguay Round confirms that Article 2.4 contains an overarching and independent obligation to make a fair comparison that goes beyond the obligations to make due adjustment described in Article 2.4.

4.9 The European Communities does not consider that the meaning of the first sentence of Article 2.4 is limited by the second sentence of Article 2.4. In other words, the European Communities does not consider that the fair comparison referred to in the first sentence of Article 2.4 is only what is contemplated in the second sentence of Article 2.4. Rather, the European Communities considers that the first sentence of Article 2.4 contains an obligation that is overarching – in the sense that it is further elaborated in the other provisions of Article 2.4, including Articles 2.4.1 and 2.4.2; and independent – in the sense that the other provisions of Article 2.4, including Articles 2.4.1 and 2.4.2, do not exhaust the fair comparison requirement in the first sentence of Article 2.4.

4.10 The third sentence of Article 2.4 is a development of the first and second sentences of Article 2.4. Read in isolation, the third sentence would make little sense. The words "price comparability" that appear twice in the third sentence relate to the "fair comparison" between export price and normal value referred to in the first and second sentences. The third sentence, like the second sentence, refers expressly to "level[s] of trade". The word "allowances" in the fourth sentence refers back to "due allowance" in the second sentence. The fifth sentence also refers to "price comparability" and "level of trade" (twice), and to the making of a "due allowance as warranted under this paragraph". The words "this paragraph" refer to all of Article 2.4. The fifth sentence, as well as the third sentence, contain obligations regarding allowances or "adjustments" (to use the vocabulary of footnote 7), whether made to normal value or export price. The final sentence of Article 2.4 expressly refers again to the "fair comparison" obligation in the first sentence of Article 2.4; and the procedural rules it contains regulate all of the substantive obligations set out in the first to final sentences of Article 2.4.

4.11 If the meaning of the first sentence of Article 2.4 would be considered to be limited by the second sentence of Article 2.4, or for that matter the remainder of Article 2.4, or Article 2.4 including Articles 2.4.1 and 2.4.2, then the first sentence would serve no purpose. It could just as well be deleted without changing the obligations of Members. That would be to render the first sentence of Article 2.4 a nullity – which would not be a permissible interpretation.

4.12 The European Communities argues that model zeroing and simple zeroing (absent targeted dumping) fail to fully and duly account for actual prices of export transactions that take place during a period of investigation and thereby result in an inflated dumping margin. To the extent that they inflate the margin of dumping or convert a negative dumping margin into a positive dumping margin, such methodologies are inherently biased and therefore unfair within the meaning of Article 2.4 of the AD Agreement. The European Communities finds support for this conclusion in the findings of the Appellate Body in EC-Bed Linen⁴¹, US – Corrosion Resistant Steel Sunset Review⁴², and US –Softwood Lumber V.⁴³

4.13 Given the common and ordinary meaning of the word "fair", the European Communities argues that the obligation to make a fair comparison under Article 2.4 must necessarily normally involve a fairly balanced comparison, being one based on equivalent methodologies – that is, a symmetrical comparison. A symmetrical comparison for the purposes of calculating a margin of dumping and eventually imposing a duty, in relation to a given product or time, is necessarily one that precludes model zeroing and simple zeroing. Thus, the European Communities contends that the obligation imposed by Article 2.4 to conduct a fair comparison precluded the model zeroing and simple zeroing methodologies applied by the United States in the cases in dispute.

4.14 In this context, the European Communities refers to the language of Article VI of the GATT 1994, the preamble to the AD Agreement, and various provisions of the AD Agreement, such as Articles 2.1, 2.6 and 4, supporting the view that the various parameters by which markets are generally defined, namely, physical characteristics, geography and time, play a central role in the AD Agreement. This also reflects a general requirement that permeates the entire AD Agreement, according to which investigating authorities must apply basic economic concepts in a consistent manner. The method used by the United States is actually incapable of measuring international price discrimination, because it is internally inconsistent, and thus unfair.

4.15 In addition, the European Communities also argues that the practice of zeroing involves effectively making an adjustment to export price that is impermissible under Article 2.4, third to fifth sentences, and thus also Article 2.4, first sentence. Such an adjustment is not allowed because it is based on a factor which does not affect price comparability. Indeed, the fact that the "export price" is above the comparable "normal value" is not a difference which affects price comparability. It is actually part of the very price difference that the investigating authorities have to establish. The practice of zeroing involves an adjustment for "non-dumping", which Article 2.4 does not, in general, permit.

4.16 The European Communities does not consider that it is possible to read the second to fifth sentences of Article 2.4 as only imposing obligations on Members to make allowance or adjustment for differences that affect price comparability. These provisions also impose an obligation not to make an allowance or adjustment when there is no such difference. Otherwise, faced with differences affecting price comparability, a Member would first be required to make a series of due allowances or adjustments, in order to render normal value and export price comparable; but having done that, the Member would then be free to make further allowances or adjustments, even if not for differences affecting price comparability, and even if reversing the effects of the first set of allowances or adjustments – without acting inconsistently with Article 2.4. That would introduce a distortion or inherent bias in the comparison and would be to render the provisions of Article 2.4 a nullity.

4.17 This is exactly what the United States practice of zeroing at issue in these proceedings does. After making adjustments for differences affecting price comparability, the investigating authority makes certain intermediate comparisons between export transactions and a "normal value"; and then makes further adjustments if the value of the export transaction exceeds that of the normal value. Thus, to the extent that the United States applied model zeroing or simple zeroing in the cases in dispute, the United States effectively made an improper adjustment to export price and violated its obligations under Article 2.4.

4.18 In response to the United States invocation of the second sentence of Article 2.4.2 as context – a provision that, in the view of the European Communities, the United States has never claimed that it was applying in the cases before the Panel – the European Communities considers that an adjustment or allowance in the context of targeted dumping might be "due", within the meaning of the third sentence of Article 2.4, if it is made under the conditions laid out in the second sentence of Article 2.4.2. According to the European Communities, if two distinct patterns of export prices are identified, in accordance with the second sentence of Article 2.4.2, then there is a difference. Article 2.4.2 refers expressly to "a pattern of export prices which differ"; and to "differences" that cannot otherwise be taken into account. Such differences may affect price comparability. It is self-evident that prices in what has been identified as distinct market A cannot – without some further consideration and explanation - necessarily simply be directly compared with prices in what has been identified as distinct market B; and that an investigating authority would be justified if it decided that the data from markets A and B could not be simply lumped together and compared with market C. An adjustment in such circumstances would be made for a difference (between markets A and B) that affected price comparability (between markets A and B; and between A, B and C). This is also clear from Article 2.4.2, which uses the words "compared" and "comparison". Thus, according to the European Communities, to the extent that the United States applied simple zeroing in the cases at issue, without justifying resort to an asymmetrical comparison methodology, the United States effectively made an improper adjustment to export price and violated Article 2.4, third to fifth sentences, and thus also Article 2.4, first sentence.

4.19 The European Communities does not agree that its interpretation of the AD Agreement renders the words: "Subject to the provisions governing fair comparison in paragraph 4" in Article 2.4.2 redundant. The "provisions governing fair comparison" are sentences 2 to 6 of paragraph 4 of Article 2. The first sentence of Article 2.4 sets forth the fair comparison obligation itself – it does not contain provisions governing that comparison. The words "subject to" simply indicate that the rules in Article 2.4.2 are subject to sentences 2 to 6 of Article 2.4. In other words, in case of conflict, the provisions of sentences 2 to 6 of Article 2.4 prevail over those of Article 2.4.2. To put matters another way, the rules in Article 2.4.2 cannot be applied in such a way as to frustrate or compromise the fair comparison required by Article 2.4.

4.20 In addition, the European Communities argues that Articles 2.4 and 9.3 require that the total amount of duty collected with respect to exports made by one specific exporter during the assessment period must not exceed the relevant margin of dumping of the same exporter. The only relevant margin of dumping permitted to be calculated under the AD Agreement, whether in an original investigation or a periodic review, is that pertaining to the exporter, not that pertaining to individual importers.

4.21 The European Communities rejects the contention of the United States that importer-specific final assessment on a transaction by transaction basis involving zeroing is contemplated under the AD Agreement by virtue of Article 9.4(ii). First, the European Communities notes that the prospective normal value or variable duty methodology recognised in Article 9.4(ii) is not in fact what was applied in any of the periodic reviews at issue. Secondly, Article 9.4 is not concerned directly with the concept of "a prospective normal value and variable duty". Rather, it sets out specific obligations concerning imposition and collection where the authorities have limited their examination in accordance with the second sentence of paragraph 10 of Article 6 – that is, when sampling has been used. Article 9.4(ii) mentions in passing the situation where the liability for payment of anti‑dumping duties has been calculated on the basis of a prospective normal value – but that is not what the provision is directly concerned with. As such, all that Article 9.4(ii) does is to confirm that a prospective normal value and variable duty is possible.

4.22 In any event, the collection of anti-dumping duties on the basis of prospective normal values is only one possible intermediate stage of the prospective system since it is subject to "a prompt refund, upon request" under Article 9.3.2 of the AD Agreement, which is itself subject to the full disciplines of Article 2. Thus, it is legally erroneous to rely on one possible intermediate stage of the prospective system, to justify the final result reached by the United States in the application of its retrospective system. There is nothing in Article 9.4 that releases authorities from the obligations in Article 9.3, including Articles 9.3.1 or 9.3.2. Thus, Members must ensure that the obligations in Article 9.3.2 are complied with whenever the "amount of the anti-dumping duty is assessed on a prospective basis". Provision must be made for a prompt refund; and in any re-calculation of the exporter specific margin of dumping the authorities must ensure that the amount of the anti-dumping duty does not exceed the margin of dumping as established under Article 2, including Article 2.4.2.

4.23 The reason why Article 9.3.2 provides for the possibility of refund in a prospective system is simply that the initial anti-dumping duty is based on the margin of dumping or normal value established during the original investigation. Therefore, there will be excess collection whenever the amount of duty collected exceeds the actual margin of dumping. The possibility of excess collection arises for all forms of duty (ad valorem, variable or other) and for a number of reasons. This is why Article 9.3.2 obliges Members applying the prospective system to put in place a mechanism of prompt refund. The mechanism is available irrespective of the form of the duty (ad valorem, variable or other). If the amount of duty collected is higher than the new dumping margin, the difference will be repaid.

4.24 Viewed in this light, the prospective system is no different from the retrospective system. The retrospective system also requires the calculation of a new margin of dumping in order to establish the final duty liability on the basis of current data.

4.25 There are, however, some important differences between the prospective system and the retrospective system operated by the United States. The prospective refund system can only result in a reduction of the total duty paid. On the other hand, the retrospective system, at least as applied by the United States, may lead to either a decrease or an increase in the margin of dumping, resulting in a decrease or increase of the total duty payable. Another difference is that a request for refund in the prospective system will, as a matter of logic, only ever be made by the exporter/importers, whereas, in the United States retrospective system, a revised determination of the dumping margin may also be requested by the domestic industry.

4.26 The type of duty has no bearing on the requirement to provide for a prompt refund mechanism. As a previous Panel stated: "(…), a properly designed variable duty system would include a refund mechanism consistent with Article 9.3.2, (…)".⁴⁴ The European Communities agrees. The European Communities does not see how the choice of a specific form of duty could in any way alter a Member's obligation under Article 9.3 of the AD Agreement not to collect in excess of the actual exporter specific margin of dumping. Thus, the obligation applies with equal force to variable duties. To decide otherwise would imply that the choice of a particular form of duty entitles Members to collect anti-dumping duties irrespective of the actual margin of dumping and even in cases where the exporter has eliminated its margin of dumping.

4.27 Finally, the European Communities argues that if simple zeroing is allowed in periodic reviews conducted under Article 9.3.1, it would effectively open up a gross distortion between prospective and retrospective systems of duty assessment. The position being advocated by the European Communities, on the other hand, would maintain the situation in which the two systems lead to an equal amount of duty collected in identical situations. The European Communities does not agree with the United States that it would be possible for the European Communities to achieve the same result as the United States in an assessment proceeding by conducting a changed circumstances review where, according to the United States, the prohibition of zeroing would not apply. The European Communities disagrees, first, with the assertion that the prohibition of zeroing does not apply in a changed circumstances review – the entire discipline of Article 2 is applicable in all margin of dumping calculations. Second, the European Communities notes that, in any event, a changed circumstances review would only cover future shipments whereas an assessment proceeding covers past sales.

(b) The United States⁴⁵

4.28 The United States argues that the "fair comparison" language of Article 2.4 refers to price adjustments and does not create an "overarching and independent" obligation that applies beyond the content of Article 2.4 itself. The text of Article 2.4 establishes the obligation that a fair comparison be made between normal value and export price and provides detailed guidance as to how that fair comparison is to be made. The focus of Article 2.4 is on how the authorities are to select transactions for comparison and make the appropriate adjustments for differences that affect price comparability. As the panel in Egypt – Steel Rebar explained: "[A]rticle 2.4 in its entirety, including its burden of proof requirement, has to do with ensuring a fair comparison, through various adjustments as appropriate, of export price and normal value."⁴⁶ Thus, the United States contests the suggestion of the European Communities that the first sentence of Article 2.4 creates an obligation – the extent of which is unstated in the AD Agreement– that is independent of the remainder of Article 2.4, that applies after the comparisons between normal value and export price are made, and that is results-oriented.

4.29 The United States argues that the "fair comparison" referred to in Article 2.4, first sentence, is informed by the second sentence of Article 2.4, which provides that "[t]his comparison shall be made ...". In this way, the second sentence of Article 2.4 describes the basic framework by which fair comparisons between export price and normal value must be made. However, the remaining sentences of Article 2.4 also inform the meaning of the first sentence of Article 2.4. Those remaining sentences of Article 2.4 (dealing, inter alia, with terms of sale, taxation, and physical characteristics) provide additional, relevant discussion of allowances that may be due in order that the comparison to be made between export price and normal value will be fair. Thus, the second sentence, along with the remaining sentences, "limit" the first sentence.

4.30 The United States argues that its interpretation of Article 2.4 is supported by the rationale underlying the introduction of the provision following the Uruguay Round. Article 2.6 of the Tokyo Round Antidumping Code ("AD Code") stated, "In order to effect a fair comparison ... ." While the United States agrees that Article 2.6 of the AD Code addressed how to make a fair comparison, according to the United States, the language was ambiguous as to whether a fair comparison was required. Thus, all of Article 2.6 of the AD Code could have been read as non-mandatory.

4.31 The ambiguity contained in Article 2.6 of the AD Code was eliminated in the AD Agreement by separating and revising the first sentence of Article 2.4 of the AD Agreement so as to make explicit the requirement to make a fair comparison. However, the remainder of Article 2.4 of the AD Agreement, like its predecessor, defines the elements of a fair comparison. Thus, Article 2.4 of the AD Agreement is clearly mandatory – it requires Members to make a fair comparison and instructs them how to do so.

4.32 This interpretation of Article 2.4 is consistent with its drafting history. In what is known as the "Dunkel Draft", Article 2.4 read as follows:

"A fair comparison shall be made between the export price and the normal value. The two prices shall be compared at the same level of trade ... ."

Arguably, that formulation was ambiguous as to the elements that make up a fair comparison. That ambiguity was eliminated in the final draft, however, by revising the text to read as follows:

"A fair comparison shall be made between the export price and the normal value. This comparison shall be made at the same level of trade ... ." (Emphasis added).

Substitution of the phrase "this comparison" establishes a reference back to the subject of the prior sentence – i.e., a fair comparison – which is what is being defined.

4.33 Further support for this reading of Article 2.4 is found in the first sentence of Article 2.4.2 of the AD Agreement which refers to "the provisions governing fair comparison in paragraph 4." The plural term "provisions," as well as the reference to "paragraph 4," rather than a particular portion of paragraph 4, clarify that the entirety of Article 2.4 constitutes the provisions "governing fair comparison."

4.34 To the extent that the European Communities argues that a requirement to make "symmetrical" comparisons between normal values and export prices can be found in the fair comparison language of Article 2.4, the United States contends that such an argument cannot be reconciled with the text of Article 2.4.2. The first sentence of Article 2.4.2 provides that those "symmetrical" comparisons are "subject to" the provisions governing "fair comparison." In the view of the United States, the drafters never intended "fair comparison" to cover symmetrical comparisons, because such coverage would have rendered this language superfluous. Furthermore, according to the United States, the AD Agreement explicitly provides for the use of asymmetrical comparisons in at least two places, neither of which is identified as an exception to the "fair comparison" requirement of Article 2.4. First, asymmetrical comparisons are expressly provided for in the targeted dumping provision – the second sentence of Article 2.4.2. While this provision is an express exception to the symmetry requirements of the first sentence of Article 2.4.2, there is no basis for interpreting it to be an exception to the fair comparison requirements of Article 2.4. Second, in the application of antidumping duties to imports from producers for which an individual dumping margin has not been separately calculated, Article 9.4(ii) of the AD Agreement expressly provides for the use of asymmetric comparisons by Members with prospective normal value systems. Nothing in Article 9.4 suggests that this methodology was provided as an exception to the fair comparison requirement of Article 2.4 or the criteria for using the targeted dumping methodology of Article 2.4.2.

4.35 The United States argues that the European Communities' claim that zeroing is permissible, when the conditions under Article 2.4.2 for conducting a targeted dumping analysis are fulfilled, cannot be squared with the European Communities' contention that zeroing is prohibited under Article 2.4 because it amounts to "an arbitrary and artificial reduction" of the export price and that "an adjustment to export price, normal value or otherwise for differences that do not affect price comparability is inconsistent with Article 2.4." The United States contends that the targeted dumping methodology is not an exception to the "fair comparison" requirement of Article 2.4; it is an exception to the symmetrical comparison requirements for investigations set forth in the first sentence of Article 2.4.2. According to the United States, having asserted that zeroing is an "impermissible adjustment to export price" not related to a difference that affects price comparability as provided in Article 2.4, the European Communities has failed to explain how this "adjustment" becomes "permissible" when the targeted dumping methodology is used. In any case, the United States is of the view that the so-called zeroing does not constitute an adjustment to price within the meaning of Article 2.4.

2. Claims with respect to Article 2.4.2 of the AD Agreement

(a) The European Communities⁴⁷

4.36 The European Communities argues that, because of the use of model zeroing, each of the measures at issue in the cases identified in Exhibits EC-1 to EC-15 (described by the European Communities as "original investigations"), violated United States' obligations under Article 2.4.2 of the AD Agreement; and that because of the use of simple zeroing, each of the measures at issue in the cases identified in Exhibits EC-16 to EC-31 (described by the European Communities as "periodic reviews"), similarly violated United States' obligations under Article 2.4.2 of the AD Agreement.

(i) Article 2.4.2 prohibits model zeroing

4.37 The European Communities argues that the wording of Article 2.1 of the AD Agreement implies that an anti-dumping proceeding concerns a product (the subject product), and that, therefore, the margin of dumping to which Article 2.4.2 refers must be the margin of dumping for the subject product. According to the European Communities, this means that "dumping" for the purposes of the AD Agreement can be found to exist only for the product under investigation as a whole, and cannot be found to exist only for a type, model or category of that product.

4.38 In the original investigations at issue, the United States established "averaging groups" and made a separate calculation of "dumping" for each one. While the use of averaging groups is permissible under Article 2.4.2 of the AD Agreement, it is not a requirement. Article 2.4 of the AD Agreement requires only that due allowance and a fair comparison be made. Due allowance could have been made by adjusting the relevant prices to take account of differences in physical characteristics affecting price comparability, before making the (single) comparison referred to in Article 2.4 of the AD Agreement, and thus before calculating a (single) margin of dumping.

4.39 Having decided to establish averaging groups in the original investigations at issue, the United States considered itself under an obligation to compare weighted-average "normal value" with a weighted-average of prices of all comparable export transactions, for each averaging group. For this purpose, the United States considered that the weighted-average "normal value" was that for the relevant averaging group. The "comparable export transactions" were those relating to the same averaging group, country and exporter. The "weighted-average of prices of all comparable export transactions" was calculated in respect of all export transactions, including export transactions that exceeded the "normal value" used for that averaging group. There was no zeroing within an averaging group. Export prices in excess of normal value were fully incorporated, before the comparison between "normal value" and export price, mathematically, in the calculation of the weighted-average export price.

4.40 The United States then combined the margins calculated for each averaging group, in order to calculate the margin of dumping for the subject product. In this second stage of the calculation the weighted-average normal value was that for the subject product. The comparable export transactions must, by definition, necessarily have been those that related to the same subject product, country and exporter. The "prices" of those export transactions were reflected in the total value of exports, which total value was incorporated, mathematically, in the margins calculated for each averaging group, whether positive or negative. Just as in the first stage, in the second stage "negative margins" for averaging groups should have been incorporated, mathematically, in the calculation of the weighted-average of prices of all comparable export transactions for the subject product before the comparison between normal value and export price.

4.41 Further support for this view may be found in a consideration of the ordinary meanings of the words "comparison" and "margin" in Article 2.4.2 of the AD Agreement – at least to the extent that the United States asserts that the results of intermediate comparisons are "margins" within the meaning of that provision. Article 2.4.2 requires a simple comparison between normal value and export price. If there is a difference, according to the text, that difference is a margin. A margin is the amount by which one thing differs from another. Normal value may exceed export price. Or export price may exceed normal value. In both cases there is a margin. It is not possible to conclude in either case that there is no margin, or that the margin is zero. That would only be the case if normal value and export price were equal. Article 2.4.2 does not prejudge how the two elements to be compared are juxtaposed, nor, thus, whether each one of a series of intermediate "margins" is expressed as positive or negative.

4.42 Nor is it possible to prematurely conclude, before the final calculation is complete, that, in relation to some discrete model or type or category of exports, when "export price" exceeds "normal value", there is no or a zero margin of dumping, because Article 2.4.2 of the AD Agreement is precisely concerned with determining whether or not there is a margin of dumping for the subject product.

4.43 In summary, the European Communities submits that, in the second stage of the calculation, the United States was bound by the obligations contained in Article 2.4.2 of the AD Agreement, and particularly the obligation to make a fair comparison between a weighted-average normal value and a weighted-average of prices of all comparable export transactions. The European Communities argues that in the EC – Bed Linen⁴⁸ and US – Softwood Lumber V⁴⁹ cases, the Appellate Body rejected the notion that, in circumstances where there is more than one averaging group based on differences in physical characteristics, and two stages to the calculation, the second stage falls outside Article 2 of the AD Agreement and indeed outside the AD Agreement altogether. The precise and detailed rules set out in the AD Agreement would be pointless if, in the final step of the calculation, the importing Member would be free to make an unfair comparison.

4.44 The European Communities considers that the use of the plural "margins of dumping" in Article 2.4.2 reflects the possibility of different margins for different countries and exporters. The European Communities further considers that an investigating authority is not entitled to consider certain transactions to have become "non-comparable" simply because the results of certain intermediate comparisons are negative. There are other types of non-comparable transactions, such as : transactions in relation to merchandise outside the scope of the proceedings; transactions made at a time outside the investigation period; transactions in relation to another geographical region; transactions that relate to another exporting country; etc.

4.45 The European Communities considers that the view that "dumping" and "margins of dumping" can only be established for the product under investigation as a whole is in consonance with the need for consistent treatment of a product in an anti-dumping investigation. Thus, having defined the product under investigation, the investigating authority must treat that product as a whole for, inter alia, the following purposes: determination of the volume of dumped imports, injury determination, causal link between dumped imports and injury to domestic industry, and calculation of the margin of dumping. Furthermore, according to Article VI:2 of the GATT 1994 and Articles 9.2 and 9.3 of the AD Agreement, an anti-dumping duty can be levied only on a dumped product. For all these purposes, the product under investigation is treated as a whole, and so-called "non-dumped" export transactions should not be excluded. There is no basis in the AD Agreement for treating a transaction as "non-dumped" for one purpose and "dumped" for other purposes, as the United States did in the cases subject to this dispute.

4.46 The European Communities considers that, unlike other provisions of the AD Agreement that are explicit regarding the permissibility of disregarding certain matters (such as Article 2.2.1; Article 9.4; Article 2.7 and Annex II, paragraph 5) Article 2.4.2 contains no express language permitting an investigating authority to disregard certain results of multiple comparisons at the aggregation stage.

4.47 For all of these reasons, the European Communities concludes that, having defined the subject product in the original investigations at issue, the United States was not entitled to set at zero the negative margins calculated for certain averaging groups based on differences in physical characteristics. The United States had become bound by its own logic. The use of such a method by the United States in this case was not in conformity with obligations imposed on the United States by Article 2.4.2 of the AD Agreement.

(ii) Article 2.4.2 is not limited to original investigations

4.48 The European Communities argues that Article 2.4.2 applies to all types of investigations undertaken pursuant to Article VI of the GATT 1994 and the AD Agreement in which margins of dumping are calculated or relied upon, including, original investigations, periodic reviews, new shipper reviews, changed circumstances reviews and sunset reviews. The European Communities argues that the United States has failed to establish that the word "investigation" or the phrase "existence of margins of dumping during the investigation phase" in Article 2.4.2 is limited to original investigations.

4.49 The European Communities observes that the term "margin of dumping" is defined in Article VI:2 of the GATT 1994, which definition is implemented and further elaborated in Article 2. The same defined term, "margin of dumping", is also used in Article 9.3, which cross-refers to all of Article 2. It also appears in Articles 9.1 and 9.5.

4.50 Article 2.4.2 states what it applies to, not what it does not apply to. Unlike the United States Statement of Administrative Action ("SAA"), it does not contain the words "(not reviews)". The United States is therefore using a contrario reasoning in an attempt to defeat the definition of "margin of dumping" and the cross-reference in Article 9.3. If the Members had really wanted to achieve what the United States asserts, it would have been a simple matter for them to have used express exclusionary language, such as that in Article 2.2.1; Article 2.7; Article 9.4(ii); or Annex II, paragraph 5. The Members did not do that.

4.51 The GATT 1994 and the AD Agreement define eight terms, but do not define the word "investigation". Its meaning in Article 2.4.2 must be determined in accordance with the rules of treaty interpretation found in customary international law. According to the European Communities, the ordinary meaning of the word "investigation", referring to The New Shorter Oxford English Dictionary, indicates a systematic examination or inquiry or a careful study of or research into a particular subject. The European Communities believes that what makes an "investigation" an "investigation" is the nature of the activity carried out by the investigating authority, not the material scope of what is examined. For example, a cursory glance is not an investigation. The random or capricious fabrication of a margin of dumping would not be based on an investigation. On the other hand, an "objective examination" of "positive evidence", involving the "considerations", "assessments", "evaluations", "demonstrations", "determinations" and "special care", as for example, expressly referred to and required by Article 3 of the AD Agreement, does involve an "investigation". The European Communities is of the view that this type of activity takes place not only during an original investigation, but also in a periodic review, a new shipper review, a changed circumstances review and a sunset review.

4.52 The European Communities contests the United States assertion that the term "investigation" has a particular meaning – namely original investigation - throughout the AD Agreement.⁵⁰ Rather, the European Communities considers that the word "investigation" is used in different senses (such as: review investigation; injury investigation; country specific investigation; company specific investigation; on-the-spot investigation; preliminary investigation; etc.) in the AD Agreement and in past panel, Appellate Body and GATT panel reports, as well as by USDOC, the ITC and in United States municipal anti-dumping law.

4.53 The European Communities contests the United States assertion that the meaning of the word "investigation" or the phrase "existence of margins of dumping during the investigation phase" in Article 2.4.2 is limited by the context of Article 5.1. To this end, the European Communities argues that the obligations set out in Article 5.1 of the AD Agreement apply only to an investigation to determine the existence, degree and effect of any alleged dumping – that is, an original investigation. The European Communities is of the view that it is precisely the presence of the words "to determine the existence, degree and effect of any alleged dumping" that limit the meaning of the word "investigation" in Article 5.1 of the AD Agreement and in the other paragraphs of that article. The absence of these words in Article 2.4.2 confirms that the word "investigation" or the phrase "existence of margins of dumping during the investigation phase" in Article 2.4.2 does not have the limited or defined meaning that the United States asserts. The European Communities contends that if the United States would be correct in its assertions, the words "to determine the existence, degree and effect of any alleged dumping" in Article 5.1 would be redundant and without meaning.

4.54 The European Communities considers that Article 2.2, which is also part of Article 2, and thus also part of the implementation of the definition of "margin of dumping", is more immediate and relevant context. It points to the use – eight times – of the word "investigation" in Article 2.2. It notes that these are provisions in respect of which implementation in the United States also for retrospective assessments was, according to the SAA, "required or appropriate". It particularly notes the use of the phrases "in the course of the investigation" in Article 2.2.1.1 and "during the investigation" in footnote 6. The European Communities also cites as relevant context the repeated use of the word "investigation" in Article 6.

4.55 Similarly, the European Communities contests the United States assertion that Article 1 of the AD Agreement means that the word "investigation" in Article 2.4.2 refers to an "original investigation" only. First, according to the European Communities Article 1 does not exhibit the characteristics of a definition. Secondly, the word "original" does not appear in Article 1. Likewise, the European Communities also contests that footnote 1 of Article 1 defines the term "investigation" for the purpose of the entire AD Agreement, including Article 2.4.2. Rather, it defines the word "initiated" – other types of investigation, such as sunset reviews, changed circumstances reviews, new shipper reviews and periodic reviews also being "initiated". Footnote 1 refers to Article 5.1 only for the purposes of determining what is "the procedural action by which a Member formally commences an investigation".

4.56 With respect to the word "phase" in Article 2.4.2, and considering its ordinary meaning, especially in conjunction with the word "during", referring to the New Shorter Oxford English Dictionary, as indicating a stage in the passage of time, the European Communities argues that it has no incidence on the meaning of the word "investigation" in Article 2.4.2. The European Communities points to several other words that are used only once in the AD Agreement without that having any legal significance. It also points to several other concepts that are referred to in the AD Agreement in different ways (as in "period" and "phase") without that having any legal significance. Thus, according to the European Communities, the word "phase" cannot, as a matter of law, lead to the conclusion that the term "investigation" is defined in Article 5.1 or elsewhere for the purposes of the entire AD Agreement, including Article 2.4.2. Finally, the European Communities points out that the word "phase" does not appear in Article 2.4.1 – which is treated the same way by the United States, simply on the basis of an erroneous assumption that "investigation" always means "not reviews".

4.57 The European Communities makes similar arguments with respect to the word "existence". The European Communities contends that it is simply not possible to entirely dissociate the "existence" of dumping and the "degree" of dumping. The ordinary meaning of "exist" being "having objective reality", it does not, when used in conjunction with the phrase "dumping margin", encompass the concept of an amount or margin somewhere between zero and infinity. In this respect, the European Communities points to the references in Article 7 to the "amount" of dumping; and in Article 9.1 to the "amount" and "full margin of dumping"; and in Article 3.4 to the "magnitude of the margin of dumping". The AD Agreement uses the word "existence" in other provisions, without that being of any legal consequence, as does the SCM Agreement, including in the provisions relating to imposition and collection. The proposition that this one word in Article 2.4.2 could have as a consequence that the basic principles for calculating a margin of dumping should be rendered entirely worthless in a system applying retrospective assessment is tenuous and implausible. In any event, the European Communities argues that the United State is wrong to assert that the only time that the "existence" of a margin of dumping is determined is in an original investigation. Issues relating to the "existence" of dumping may also arise, for example, in the context of a changed circumstances review, or in a sunset review. The European Communities requests the Panel not to base itself on the esoteric mathematical and philosophical combination dictionary meanings referred to by the United States. The European Communities also invites the Panel to pay due regard to the French and Spanish texts of the AD Agreement, the titles of which make it clear that all of the provisions of Article 2 – including those that the United States admits apply in a retrospective assessment – are concerned with the "existence" of dumping. Finally, the European Communities points out that the word "existence" does not appear in Article 2.4.1 – which is treated the same way by the United States, simply on the basis of an erroneous assumption that "investigation" always means "not reviews".

4.58 The European Communities points to the object and purpose of retrospective assessment, as it appears from the AD Agreement, and as set out in the Manual and confirmed by the United States in these proceedings: it is just a temporal up-date. The European Communities also points to the object and purpose of the AD Agreement, which at the very least requires basic consistency in the application of economic concepts. The European Communities considers that the United States has failed to articulate any alleged object or purpose to explain why, at the moment of final payment, the basic method for calculating a margin of dumping should suddenly change, and be substantially unregulated by the AD Agreement.

4.59 The European Communities considers that since the United States has failed to establish that the word "investigation" or the phrase "existence of margins of dumping during the investigation phase" has the limited meaning argued for by the United States – and that since a consideration of the ordinary meaning, context, and object and purpose confirms that there is no such limited meaning – there is no need to have recourse to the preparatory work. In these circumstances, the preparatory work could only ever be used to confirm the meaning resulting from Article 31 of the Vienna Convention. That is precisely what it does. The negotiators were acutely aware of and sensitive to the issue of definitions: it really matters whether or not something is defined; and the fact that some terms are defined and others not must be given meaning. There was general consensus on the need for a consistent, balanced and fair application of anti-dumping measures. There was broad consensus on both sides of the debate that international markets and business had evolved, and that the AD Agreement should be up-dated accordingly. At no point in the debate was it ever suggested that there should be different treatment for original investigations and retrospective assessments on the fundamental question of how to calculate a margin of dumping. This is so notwithstanding the fact that assessment and refund issues were repeatedly discussed in detail and at length, with regard to the "duty as a cost" and "lesser duty" issues. There is a clear and strong indication of consensus that the interests of both parties in the asymmetry and zeroing debate could be accommodated in the targeted dumping provisions that eventually became the second sentence of Article 2.4.2 of the AD Agreement. There is an overwhelming indication of consensus that the presence of the word "investigation", used repeatedly in what was to become Article 2.2 of the AD Agreement, did not mean that those provisions were to be irrelevant when a margin of dumping was calculated in retrospective assessments. Right from the start, there was a "strong convergence of views" on the need to strengthen the rules applicable to the pre-investigation phase, without, however, placing an unreasonable burden on complainants or authorities.

4.60 The European Communities considers that the negotiating history of Article 2.4.2 does not record which Member or Members – if any – proposed the particular form of words "during the investigation phase" or why. However, had the United States actually put to the other Members, during the negotiations, an express provision excluding the disciplines of Article 2.4.2 from retrospective assessments – the European Communities contends that there is no doubt at all, based on a fair consideration of the negotiations, that the other Members would never have agreed to it. This was confirmed by each of the 11 third parties in response to a question posed by the European Communities during the meeting with the third parties: none of them thought that the AD Agreement means that investigating authorities are free to calculate margins of dumping in all reviews and all retrospective assessments as they wish, and free from the disciplines of Article 2, including Article 2.4.2. Furthermore, the European Communities argues that subsequent practice of the Members with respect to the implementation of Article 2.4.2 is supportive of its case. In this regard, the European Communities notes that its review of the notifications of the anti-dumping legislation of 105 Members indicates that no Members take the view being advocated by the United States in the present dispute. The European Communities cites this in support of its case as "subsequent practice" within the meaning of Article 31(3)(b) of the Vienna Convention on the Law of Treaties; and also as evidence of the intention of the Members, within the meaning of Article 31(4) of the Vienna Convention.

4.61 The European Communities does not consider that its interpretation of the word "investigation" in Article 2.4.2 renders the words "during the investigation phase" in Article 2.4.2 meaningless. In this regard, the European Communities suggests four different alternative, but not necessarily mutually exclusive, possible meanings of these terms: during the investigation period; during the period in which the particular type of investigation must be concluded; not during the pre-investigation phase; or that the words are merely descriptive, in the sense that the United States considers the words "to determine the existence, degree and effect of any alleged dumping" in Article 5.1 to be merely descriptive.

4.62 With regard to the first meaning, if, referring to a dictionary, "phase" indicates a distinct period in a process of change or development, and "during" indicates a temporal connotation, then an investigation phase is something that spans a period of time, such as the "investigation period". The word "phase" and the word "period" have very similar meanings, especially when the word "phase" is associated with the word "during", giving it a temporal connotation. The two words may be considered synonymous in the context of Article 2 of the AD Agreement, referring to a "stage" in the passage of time.

4.63 The most natural reading of Article 2.4.2, and the reading that is grammatically correct, is that the words "during the investigation phase" are associated with the word "existence" rather than with the word "established". In other words, the provision does not refer to something that has to happen (the "establishment" of the margins of dumping) during the period of time in which an investigating authority must normally make its determination (generally 12 to 18 months). Rather, it refers to something that has to be established (whether or not there is dumping, and if so, what the margin of dumping is) by reference to a certain period of time: the investigation period. If the drafters had intended to associate the phrase "during the investigation phase" with the word "established", the provision could have been drafted differently. For example, it could have read "… the existence of margins of dumping shall normally be established during the investigation phase …".

4.64 Contextual support for such an approach may be found in various other provisions of Article 2 of the AD Agreement. Article 2 provides valuable and persuasive context, because Article 2.4.2 is part of Article 2, and because Article 2 implements the definitions of dumping and margin of dumping set out in Article VI of the GATT 1994. Article 2.2.1 imposes obligations on investigating authorities concerning the circumstances in which sales of the like product in the domestic market "may be disregarded" as below cost. As such, it is of double interest as context, given that the practice of zeroing with which these proceedings are concerned also involves the investigating authority disregarding or adjusting, at least in part, certain export sales or other data. The object and purpose of this rule is clear. By requiring an investigating authority to consistently use data arising during the investigation period, the AD Agreement is imposing basic obligations of consistency, objectivity and fairness. Precisely the same is true of Article 2.4.2 of the AD Agreement. If the data used to establish normal value arose during the investigation period, but, in the final step of comparison, the existence of margins of dumping were permitted to be established by reference to another period, there would clearly be a rupture in the basic structure and continuity of the method for establishing whether or not there is dumping, and if so what the margin of dumping is, in conformity with the definitions set out in Article VI of the GATT 1994, as implemented in Article 2 of the AD Agreement. It is thus entirely consistent with the object and purpose of Article 2.4.2 of the AD Agreement that it should continue to require at the comparison stage, just as at the normal value stage, that the margins of dumping must be established by reference to data arising during the investigation period (or phase). Precisely the same observations may be made with respect to the references to the investigation period in Article 2.2.1.1, final sentence; footnote 6, second phrase; and Article 2.4.1, final phrase of the AD Agreement.

4.65 Other references in Article 2 to the investigation period are references to the methodology for establishing normal value. The provision that primarily relates to export price – Article 2.3 of the AD Agreement – contains no reference to data arising during the investigation period. Thus, aside from Article 2.4.2 of the AD Agreement, there is no other express textual provision that would oblige an investigating authority to use export price data arising during the investigation period (or phase). The phrase "at as nearly as possible the same time" in Article 2.4 is qualified, and in any event does not relate to the exclusion of transactions outside the chosen period. Thus, absent the words "during the investigation phase" in Article 2.4.2, an investigating authority might, for one reason or another, seek to draw on export price data arising outside the investigation period. Viewed in this light, the phrase during the investigation phase (or period) in Article 2.4.2 of the AD Agreement has an important – even vital – role to play.

4.66 Further contextual support for this view may be found in the reference to "the period of investigation" in Article 9.5 of the AD Agreement. What happens in a new shipper review is that an unrelated exporter that did not export during the original investigation period can, on request, obtain "a determination of dumping". In the opinion of the European Communities, that "determination of dumping" must be consistent with all of the disciplines set out in Article 2 of the AD Agreement – including Article 2.4.2. What makes an unrelated "new shipper" new is precisely the fact that, during the original investigation period (or phase) it did not export. That is why, in the original investigation, no specific margin of dumping could have been calculated in relation to such new shipper. Viewed in this light, it is clear that the phrase "during the period of investigation" in Article 9.5 refers back to the phrase "during the investigation phase" in Article 2.4.2, and confirms that the word "phase" in Article 2.4.2 is in fact synonymous with the word "period".

4.67 The European Communities does not consider it remarkable that, in some instances, different language is used to refer to concepts that are the same, such as a stage in the passage of time (that is, in this case, "period" and "phase"). Where the drafters sought a particularly high level of consistency, they defined the relevant terms (there are eight such definitions in the GATT 1994 and the AD Agreement). There are many other examples in the AD Agreement of different words or phrases being used to refer to the same thing. All of these terms fall to be interpreted according to the Vienna Convention, being given their ordinary meaning and having regard to context and object and purpose. There is nothing in the Vienna Convention that excludes the possibility that such different terms in fact have the same meaning. In such circumstances, the mere fact that the terms are different is of no further legal consequence.

4.68 The European Communities observes that, if the drafters had intended to take a step of such great importance as jettisoning, in retrospective assessments, the basic principles governing the determination of dumping margins, as defined in Article VI of the GATT 1994 and elaborated in Article 2 and particularly Article 2.4.2, thus rendering the results of original investigations worthless, they would not have chosen to: a) rely on an over subtle distinction between the words "during … period" and the words "during … phase", neither the word "period" nor "phase" being defined, and those phrases bearing the same ordinary meaning, if reference is made to a dictionary; b) in circumstances where the use of one word or the other logically had no incidence whatsoever on the meaning of the word "investigation" in Article 2.4.2 – investigation itself not being a defined term: c) by associating the phrase with the word "existence" as opposed to "established", in a manner that is grammatically incorrect given their alleged objective; and d) whilst at the same time including the directly contradictory cross-reference in Article 9.3 to the whole of Article 2. Rather, given the importance of the step the drafters were allegedly taking, they would rather have simply expressly stated that the rules in Article 2.4.2 could be disregarded in retrospective assessments. Just as they did in Article 2.2.1 with regard to sales not in the ordinary course of trade. Just as they did in Article 9.4 with regard to zero, de minimis and facts available margins. Just as they did in Article 2.7 with regard to non-market economies. That, at least, is what the third parties in this case seem to think.

4.69 A second possible meaning of the words "during the investigation phase" in Article 2.4.2 is that the investigating authority must make the determination within the 12 to 18 month period during which investigations must generally be concluded.

4.70 In the context of an original investigation, Article 5.10 of the AD Agreement contains the procedural rule that: "Investigations shall, except in special circumstances, be concluded within one year, and in no case more than 18 months, after their initiation." Article 2.4.2 takes this procedural rule one step further, and makes it a substantive rule, clearly of application to investigations of all types. It provides that the "margins of dumping" must be established during the 12 to 18 month original investigation phase. Thus, if a margin of dumping resulting from an original investigation were determined 2 years after the initiation of the original investigation, the Member responsible would act inconsistently with the AD Agreement. The same would be true if a Member purported to alter the margin of dumping determined during the original investigation, after the end of the 18 month period, without conducting a new investigation. It should not be forgotten that, with a new investigation, there comes, in principle, a new investigation period (or review period) – that is, fresh data (see further below) – and a new obligation to make any necessary margin of dumping determinations required by Article 2 within the new investigation phase.

4.71 The reasons for this rule are self-evident. First, it would be Kafkaesque if a producer or exporter were to find itself forever "under investigation" without ever being subject to a measure (or having the investigation terminated). That would be inconsistent with the basic requirements of certainty and predictability in international trade, which the Appellate Body has so clearly stated lie at the heart of the WTO Agreements. Second, although the AD Agreement contains no rule concerning the establishment of periods of investigation, it is well established that an investigating authority is not entirely unfettered in this matter. In principle, the period of investigation must end as close as reasonably possible to the date on which the investigation is initiated. There are very good reasons for this. Anti-dumping measures can only reasonably be imposed on the basis of sufficiently fresh data. Absent some special justification, it would not be reasonable or acceptable under the AD Agreement for an investigating authority to impose anti-dumping measures today, based on data from 10 years ago. Requiring the dumping margin determination to be made within the investigation phase is an essential part of this "fresh data" rule: there would be no point in requiring the investigation to be initiated shortly after the end of the investigation period, if the dumping margin determination could be delayed until several years later. Third, given the importance of this rule, it is entirely appropriate that it should be expressed not only in the essentially procedural terms of Article 5.10