WT/DS264/AB/RW
15 August 2006
(06-3893)
Original: English
AB-2006-3
Report of the Appellate Body
WT/DS264/AB/RW
Page i
Introduction 1
Arguments of the Participants and the Third Participants 4
Claims of Error by Canada – Appellant 4
Article 2.4.2 of the Anti-Dumping Agreement 4
Article 2.4 of the Anti-Dumping Agreement 8
Arguments of the United States – Appellee 10
Article 2.4.2 of the Anti-Dumping Agreement 10
Article 2.4 of the Anti-Dumping Agreement 15
Arguments of the Third Participants. 16
China 16
European Communities 16
India 19
Japan 19
New Zealand 22
Thailand 24
Issues Raised in This Appeal 26
Article 2.4.2 of the Anti-Dumping Agreement 26
Introduction 26
Findings of the Panel 28
Arguments of the Participants and the Third Participants 30
Is Zeroing Allowed under the Transaction-to-Transaction Comparison
Methodology Set Out in Article 2.4.2 of the Anti-Dumping Agreement? 31
The Use of "Zeroing" in the Section 129 Determination 31
Article 2.4.2 of the Anti-Dumping Agreement 33
Context 41
"Symmetry" between the Determination of Dumping and the Injury
and Causation Analyses 47
Object and Purpose of the Anti-Dumping Agreement 48
Historical Background 48
Conclusion on Canada's Claim under Article 2.4.2 50
Article 2.4 of the Anti-Dumping Agreement 51
Introduction 51
Findings of the Panel 51
Arguments of the Participants and the Third Participants 52
WT/DS264/AB/RW
Page ii
Whether Zeroing is Consistent with the "Fair Comparison" Requirement in
Article 2.4 of the Anti-Dumping Agreement 54
Article 2.4 of the Anti-Dumping Agreement 54
The Basis of the Panel's Finding under Article 2.4 of the Anti-Dumping Agreement 55
The Consistency of Zeroing under the Transaction-to-Transaction Comparison Methodology with Article 2.4 of the Anti-Dumping
Agreement 56
Findings and Conclusions 59
ANNEX I Notification of an Appeal by Canada
WT/DS264/AB/RW
Page iii
CASES CITED IN THIS REPORT
Short Title | Full Case Title and Citation |
EC – Audio Cassettes | GATT Panel Report, EC – Anti-Dumping Duties on Audio Tapes in Cassettes Originating in Japan, ADP/136, 28 April 1995, unadopted |
EC – Bed Linen | Appellate Body Report, European Communities – Anti-Dumping Duties on Imports of Cotton-Type Bed Linen from India, WT/DS141/AB/R, adopted 12 March 2001, DSR 2001:V, 2049 |
EEC – Cotton Yarn | GATT Panel Report, European Economic Community – Imposition of Anti- Dumping Duties on Imports of Cotton Yarn from Brazil, ADP/137, adopted 30 October 1995, BISD 42S/17 |
EC – Tube or Pipe Fittings | Panel Report, European Communities – Anti-Dumping Duties on Malleable Cast Iron Tube or Pipe Fittings from Brazil, WT/DS219/R, adopted 18 August 2003, modified by Appellate Body Report, WT/DS219/AB/R, DSR 2003:VII, 2701 |
Mexico – Anti-Dumping Measures on Rice | Appellate Body Report, Mexico – Definitive Anti-Dumping Measures on Beef and Rice, Complaint with Respect to Rice, WT/DS295/AB/R, adopted 20 December 2005 |
US – Corrosion-Resistant Steel Sunset Review | Appellate Body Report, United States – Sunset Review of Anti-Dumping Duties on Corrosion-Resistant Carbon Steel Flat Products from Japan, WT/DS244/AB/R, adopted 9 January 2004, DSR 2004:I, 3 |
US – Hot-Rolled Steel | Appellate Body Report, United States – Anti-Dumping Measures on Certain Hot-Rolled Steel Products from Japan, WT/DS184/AB/R, adopted 23 August 2001, DSR 2001:X, 4697 |
US – Softwood Lumber V | Appellate Body Report, United States – Final Dumping Determination on Softwood Lumber from Canada, WT/DS264/AB/R, adopted 31 August 2004, DSR 2004:V, 1875 |
US – Softwood Lumber V | Panel Report, United States – Final Dumping Determination on Softwood Lumber from Canada, WT/DS264/R, adopted 31 August 2004, modified by Appellate Body Report, WT/DS264/AB/R, DSR 2004:V, 1937 |
US – Softwood Lumber V (Article 21.5 – Canada) | Panel Report, United States – Final Dumping Determination on Softwood Lumber from Canada – Recourse to Article 21.5 of the DSU by Canada, WT/DS264/RW |
US – Zeroing (EC) | Appellate Body Report, United States – Laws, Regulations and Methodology for Calculating Dumping Margins ("Zeroing"), WT/DS294/AB/R, adopted 9 May 2006 |
US – Zeroing (EC) | Panel Report, United States – Laws, Regulations and Methodology for Calculating Dumping Margins ("Zeroing"), WT/DS294/R, adopted 9 May 2006, modified by Appellate Body Report, WT/DS294/AB/R |
WT/DS264/AB/RW
Page iv
ABBREVIATIONS USED IN THIS REPORT
Abbreviation | Description |
Anti-Dumping Agreement | Agreement on Implementation of Article VI of the General Agreement on Tariffs and Trade 1994 |
DSB | Dispute Settlement Body |
DSU | Understanding on Rules and Procedures Governing the Settlement of Disputes |
GATT 1994 | General Agreement on Tariffs and Trade 1994 |
Group of Experts Report | GATT Second Report of the Group of Experts, Anti-Dumping and Countervailing Duties, L/1141, adopted 27 May 1960, BISD 9S/194 |
Harmonized System | Harmonized Commodity Description and Coding System |
Original panel | Panel in the original US – Softwood Lumber V proceedings |
Panel | Panel in these US – Softwood Lumber V (Article 21.5 – Canada) proceedings |
Panel Report | Panel Report, US – Softwood Lumber V (Article 21.5 – Canada) |
Section 129 | Section 129 of the Uruguay Round Agreements Act, Public Law No. 103-465, 108 Stat. 4836, United States Code, Title 19, section 3538 (2000) |
Section 129 Determination | Notice of Determination Under Section 129 of the Uruguay Round Agreements Act: Antidumping Measures on Certain Softwood Lumber Products from Canada, United States Federal Register, Vol. 70, No. 83 (2 May 2005), pp. 22636-22646 (Exhibit CDA-1 submitted by Canada to the Panel) |
Statement of Administrative Action | Statement of Administrative Action to the Uruguay Round Agreements Act, H.R. Doc. No. 103-316 (1994) |
Tokyo Round Anti- Dumping Code | Tokyo Round Agreement on Implementation of Article VI of the General Agreement on Tariffs and Trade, BISD 26S/171 |
USDOC | United States Department of Commerce |
Vienna Convention | Vienna Convention on the Law of Treaties, done at Vienna, 23 May 1969, 1155 UNTS 331; 8 International Legal Materials 679 |
Working Procedures | Working Procedures for Appellate Review, WT/AB/WP/5, 4 January 2005 |
WTO | World Trade Organization |
WT/DS264/AB/RW
Page 1
WORLD TRADE ORGANIZATION APPELLATE BODY
Recourse to Article 21.5 of the DSU by Canada
Canada, Appellant
United States, Appellee
China, Third Participant
European Communities, Third Participant
India, Third Participant
Japan, Third Participant
New Zealand, Third Participant
Thailand, Third Participant
Introduction
AB-2006-3
Present:
Abi-Saab, Presiding Member Baptista, Member
Sacerdoti, Member
Canada appeals certain issues of law and legal interpretations developed in the Panel Report, United States – Final Dumping Determination on Softwood Lumber from Canada – Recourse to Article 21.5 of the DSU by Canada (the "Panel Report").1 The Panel was established to consider a complaint by Canada regarding the consistency with the Agreement on Implementation of Article VI of the General Agreement on Tariffs and Trade 1994 (the "Anti-Dumping Agreement") of a measure taken by the United States to comply with the recommendations and rulings of the Dispute Settlement Body (the "DSB") in the US – Softwood Lumber V proceedings.2
The original dispute concerned an anti-dumping investigation by the United States Department of Commerce (the "USDOC") that led to the imposition, in May 2002, of anti-dumping duties on imports of softwood lumber from Canada.3 Before the panel in US – Softwood Lumber V (the "original panel"), Canada claimed that, in imposing anti-dumping duties on imports of softwood lumber from Canada, the United States had acted inconsistently with several provisions of the Anti-Dumping Agreement, as well as with Articles VI:1 and VI:2 of the General Agreement on
1WT/DS264/RW, 3 April 2006.
2The recommendations and rulings of the DSB resulted from the adoption on 31 August 2004, by the DSB, of the Appellate Body Report, WT/DS264/AB/R, and the Original Panel Report, WT/DS264/R.
3Notice of Amended Final Determination of Sales at Less Than Fair Value and Antidumping Duty Order: Certain Softwood Lumber Products From Canada, United States Federal Register, Vol. 67, No. 99 (22 May 2002) 36068.
WT/DS264/AB/RW
Page 2
Tariffs and Trade 1994 (the "GATT 1994").4 The original panel found, inter alia, that the United States had acted inconsistently with Article 2.4.2 of the Anti-Dumping Agreement in determining the existence of margins of dumping on the basis of a methodology incorporating the practice of "zeroing".5 In the light of its finding on Article 2.4.2, the original panel applied judicial economy and declined to rule on Canada's claims under Article 2.4 of the Anti-Dumping Agreement ("fair comparison") in respect of zeroing.6 The original panel's finding of inconsistency with Article 2.4.2 was upheld by the Appellate Body.7
The original panel and Appellate Body reports were adopted by the DSB on 31 August 2004.8 On 6 December 2004, Canada and the United States jointly informed the DSB, pursuant to Article 21.3(b) of the Understanding on Rules and Procedures Governing the Settlement of Disputes (the "DSU"), that they had mutually agreed that the reasonable period of time to implement the recommendations and rulings of the DSB would be seven and one-half months, that is, from 31 August 2004 to 15 April 2005.9 The reasonable period of time was later extended to 2 May 2005 by agreement between the parties.10
On 2 May 2005, the USDOC issued a new final determination11 pursuant to Section 129 of the Uruguay Round Agreement Act12 (the "Section 129 Determination"). In the original determination, the USDOC had calculated the margins of dumping by comparing weighted-average normal value to the weighted average of export prices. By contrast, in the Section 129 Determination, the USDOC established the margins of dumping on the basis of a comparison of normal value and export prices on a transaction-to-transaction basis.13 On 19 May 2005, the United States notified the DSB that, with the Section 129 Determination, it had implemented the DSB's recommendations and rulings.14
4Original Panel Report, para. 3.1. See also Appellate Body Report, US – Softwood Lumber V, para. 3.
5Original Panel Report, para. 8.1(a)(i).
6Ibid., para. 8.1(c)(i).
7Appellate Body Report, US – Softwood Lumber V, para. 183(a).
8WT/DS264/9.
9WT/DS264/12. In the light of this agreement, the arbitration proceeding under Article 21.3(c) that had been initiated at the request of Canada was terminated. (WT/DS264/13)
10WT/DS264/15.
11Notice of Determination Under Section 129 of the Uruguay Round Agreements Act: Antidumping Measures on Certain Softwood Lumber Products from Canada, United States Federal Register, Vol. 70, No. 83 (2 May 2005), pp. 22636-22646 (Exhibit CDA-1 submitted by Canada to the Panel).
12Public Law No. 103-465, § 129, 108 Stat. 4836, United States Code, Title 19, section 3538 (2000).
13See Section 129 Determination, pp. 22637-22639.
14WT/DSB/M/189, para. 92.
WT/DS264/AB/RW
Page 3
Canada, however, considered that the United States had failed to bring its measure into conformity with its obligations under the Anti-Dumping Agreement. Canada therefore requested that the matter of compliance be referred to a panel pursuant to Article 21.5 of the DSU.15 On 1 June 2005, the DSB referred the matter to the original panel.16 In the Article 21.5 proceedings, Canada claimed that the use of zeroing by the USDOC in the Section 129 Determination is inconsistent with the United States' obligations under Articles 2.4 and 2.4.2 of the Anti-Dumping Agreement.17
The Panel Report was circulated to the Members of the World Trade Organization (the "WTO") on 3 April 2006. The Article 21.5 Panel (the "Panel") found that "the [US]DOC was entitled not to offset the non-dumped transactions against the dumped transactions when calculating the margin of dumping for each respondent foreign producer or exporter."18 Consequently, the Panel rejected Canada's claim that "the [US]DOC's use of zeroing in the [transaction-to-transaction] comparison methodology at issue is inconsistent with Article 2.4.2 of the [Anti-Dumping] Agreement."19 In addition, the Panel rejected Canada's claim that "the United States has violated the fair comparison obligation provided for in the first sentence of Article 2.4 of the [Anti-Dumping] Agreement."20
The Panel concluded that:
... the determination of the [US]DOC in the section 129 proceeding investigation is not inconsistent with the asserted provisions of Articles 2.4 and 2.4.2 of the [Anti-Dumping] Agreement.
We therefore consider that the United States has implemented the recommendations and rulings of the DSB in US – Softwood Lumber V, to bring its measure into conformity with its obligations under the [Anti-Dumping] Agreement.21
Having found that the United States did not act inconsistently with its obligations under the Anti- Dumping Agreement, the Panel did not make any recommendation under Article 19.1 of the DSU.22
15WT/DS264/16.
16WT/DS264/20/Rev.2.
17Panel Report, para. 3.1.
18Ibid., para. 5.66.
19Ibid.
20Ibid., para. 5.78.
21Ibid., paras. 6.1-6.2.
22Ibid., para. 6.3.
WT/DS264/AB/RW
Page 4
On 17 May 2006, Canada notified the DSB, pursuant to Article 16.4 of the DSU, of its intention to appeal certain issues of law and legal interpretations developed in the Panel Report, and filed a Notice of Appeal23 pursuant to Rule 20 of the Working Procedures for Appellate Review24 (the "Working Procedures"). On 24 May 2006, Canada filed an appellant's submission.25 On
12 June 2006, the United States filed an appellee's submission.26 On the same day, the European Communities, Japan, New Zealand, and Thailand each filed a third participant's submission27, and China and India each notified the Appellate Body Secretariat of its intention to appear at the oral hearing and make an oral statement.28
The oral hearing in this appeal was held on 24 June 2006.29 The participants and the third participants presented oral arguments and responded to questions posed by the Members of the Division hearing the appeal.
Arguments of the Participants and the Third Participants
Claims of Error by Canada – Appellant
Article 2.4.2 of the Anti-Dumping Agreement
Canada claims that the Panel erred in finding that the use of zeroing by the USDOC in the Section 129 Determination rests on a permissible interpretation of Article 2.4.2 of the Anti-Dumping Agreement. Canada requests the Appellate Body to find that the USDOC's use of zeroing in the Section 129 Determination is inconsistent with Article 2.4.2 and that, therefore, the United States has failed to implement the recommendations and rulings of the DSB in the original dispute.
Canada asserts that Article 2.4.2 of the Anti-Dumping Agreement prohibits the use of zeroing when investigating authorities calculate the margin of dumping by comparing normal value and export prices on a transaction-to-transaction basis in accordance with the methodology set out in the first sentence of that provision. Canada submits that the calculation of a "margin of dumping" within the meaning of Article 2.4.2 must be done for the "product as a whole". This means that an
23WT/DS264/25 (attached as Annex I to this Report).
24WT/AB/WP/5, 4 January 2005.
25Pursuant to Rule 21 of the Working Procedures. 26Pursuant to Rule 22 of the Working Procedures. 27Pursuant to Rule 24(1) of the Working Procedures. 28Pursuant to Rule 24(2) of the Working Procedures.
29The oral hearing was originally scheduled for 26 June 2006. However, the oral hearing had to be rescheduled owing to logistical difficulties due to meetings held at the WTO building in connection with Doha Development Agenda negotiations. Neither the participants nor the third participants objected to the change of date.
WT/DS264/AB/RW
Page 5
investigating authority must take into account all intermediate comparison results, regardless of whether they are positive or negative, in any aggregation leading to a final margin of dumping. Canada explains that this requirement is based on the correct interpretation of the term "margin of dumping" and the definition of "dumping" set out in Article 2.1 of the Anti-Dumping Agreement. According to Canada, in US – Softwood Lumber V and US – Zeroing (EC), the Appellate Body found that Article VI:1 of the GATT 1994 and Article 2.1 of the Anti-Dumping Agreement use the term "product" to refer to the product under investigation or the product "as a whole".30 Canada adds that the Appellate Body also found that Articles 6.10 and 9.2 of the Anti-Dumping Agreement provide contextual support confirming that the term "product" refers to the entire product under investigation.31
Canada points out that the Appellate Body has found that the use of "zeroing" is not permitted when making weighted average-to-weighted average comparisons of normal value and export prices in an original investigation32, or when making weighted average-to-transaction comparisons in an assessment proceeding.33 In Canada's view, it follows that, where an investigating authority makes multiple comparisons on a transaction-to-transaction basis and produces a number of intermediate values or results and then aggregates those results, Article 2.4.2 requires it to take into account the full amount of those results, be they positive or negative; an investigating authority cannot treat those
results that are negative as zero. Therefore, according to Canada, prior reasoning and interpretations of the Appellate Body apply to prohibit zeroing under the transaction-to-transaction comparison methodology in the same manner as they did under the weighted average-to-weighted average and weighted average-to-transaction methodologies.
Canada further submits that Article 2.4.2 contains no express language that permits the use of zeroing in the calculation of margins of dumping. Moreover, Canada argues that, as the Appellate Body has explained, calculating margins of dumping without aggregating the results of all "multiple comparisons" for the "product as a whole" introduces an "inherent bias" that "may distort not only the magnitude of a dumping margin, but also a finding of the very existence of dumping."34 In Canada's view, an interpretation of Article 2.4.2 that permits zeroing under the transaction-to-transaction
30See Appellate Body Report, US – Softwood Lumber V, paras. 92-93; and Appellate Body Report,
US – Zeroing (EC), paras. 125-126.
31See Appellate Body Report, US – Softwood Lumber V, para. 94.
32Canada's appellant's submission, para. 28 and footnote 39 thereto (referring to Appellate Body Report, US – Softwood Lumber V, paras. 97-98).
33Ibid. (referring to Appellate Body Report, US – Zeroing (EC), para. 127).
34Ibid., para. 31 (quoting Appellate Body Report, US – Corrosion-Resistant Steel Sunset Review, para. 135).
WT/DS264/AB/RW
Page 6
methodology cannot be reconciled with the "inflation and distortion" of dumping margins that this practice causes.35
Canada asserts that the Panel's erroneous interpretation rests on four flawed conclusions. The first is the Panel's understanding that the term "margin of dumping", read in the context of Articles VI:1 and VI:2 of the GATT 1994, refers to a "price difference", and that this may be interpreted to allow investigating authorities not to reflect fully non-dumped transaction-specific comparison results.36 According to Canada, the Panel's understanding "ignores" the interpretation of the term "margins of dumping" by the Appellate Body as requiring an investigating authority to take into account the full value of all intermediate comparison results, be they positive or negative, so that the calculation of the final margin of dumping reflects fully the "product as a whole".37 Canada adds that, as a result of this incorrect understanding, the Panel rejected Canada's submission that the term "margins of dumping" could not have a different meaning for each of the two calculation methodologies set out in the first sentence of Article 2.4.2.
Secondly, Canada argues that the Panel improperly interpreted the term "product" as not referring to the "product as a whole". The Panel claimed that interpreting "product" in Articles VI:2, VI:3, and VI:6 of the GATT 1994 to refer to the "product as a whole" would mean that these provisions could not be read as requiring the "application of a duty on a single import transaction".38 Canada observes, however, that the Panel did not explain why these provisions needed to be interpreted as referring to the application of a duty on a single import transaction.
Thirdly, Canada submits that the Panel incorrectly concluded that the Appellate Body's reasoning in US – Softwood Lumber V does not apply to the transaction-to-transaction comparison methodology set out in the latter part of the first sentence of Article 2.4.2. Canada explains that, contrary to the Panel's view, the phrase "all comparable export transactions" was not central to the Appellate Body's reasoning in that case, which was, instead, based on the definitions of "dumping" and "margins of dumping". In addition, Canada notes that the Appellate Body recently found that the term "margin of dumping" prohibits zeroing under not just the weighted average-to-weighted average comparison methodology pursuant to Article 2.4.2, but also under a weighted average-to-transaction methodology in duty assessment proceedings pursuant to Article 9.3 of the Anti-Dumping
35Canada's appellant's submission, para. 31.
36Panel Report, paras. 5.27-5.28.
37Canada's appellant's submission, paras. 37-38.
38Ibid., para. 43 (referring to Panel Report, para. 5.23).
WT/DS264/AB/RW
Page 7
Agreement.39 Canada explains that the Appellate Body reached this conclusion by applying the same interpretation to the term "margin of dumping" in Article 9.3 as it did to the term "margins of dumping" in Article 2.4.2 with respect to original investigations, and that the phrase "all comparable export transactions" does not appear in Article 9.3. The Appellate Body's interpretation in US – Zeroing (EC) of "margin of dumping" in Article 9.3 therefore completely undermines the basis upon which the Panel restricted the reasoning of the Appellate Body in US – Softwood Lumber V to the weighted average-to-weighted average comparison methodology.
Fourthly, Canada asserts that the Panel's reliance on the so-called "broader contextual considerations" it identified is misplaced. According to the Panel, one such consideration was that, if "margins of dumping" were interpreted to prohibit zeroing under the weighted average-to-transaction comparison methodology provided in the second sentence of Article 2.4.2, such an interpretation would render this methodology redundant, because it would produce results that were "mathematically equivalent" to the weighted average-to-weighted average methodology.40
Canada argues that this "contextual consideration" rests on two faulty analytical foundations. First, the issue before the Panel was the use of zeroing in the Section 129 Determination under the transaction-to-transaction comparison methodology; the weighted average-to-weighted average and the weighted average-to-transaction methodologies were not before the Panel. Moreover, Canada and all but one of the third participants noted that their answers to the Panel's questions on the weighted average-to-transaction comparison methodology were hypothetical, because none of them had any experience with this calculation methodology. In Canada's view, it is inappropriate to reject an interpretation of a provision and its application in respect of a concrete measure on the basis of the interpretation and application of a provision not at issue, in the abstract, and without evidence.
The second fault in the Panel's analysis of the "mathematical equivalence" argument is that it rests on the false premise that the hypothetical proposed by the United States represented the only way in which the weighted average-to-transaction comparison methodology could be applied. However, Canada and certain third participants provided a variety of examples of how the weighted average-to-transaction comparison methodology would not result in "mathematical equivalence" without the use of zeroing. According to Canada, the Panel "ignored" these examples.41 Thus, even if the Panel's consideration of the weighted average-to-transaction comparison methodology were appropriate, the conclusions it drew, in Canada's view, were incorrect.
39Canada's appellant's submission, para. 49 (referring to Appellate Body Report, US – Zeroing (EC), paras. 135 and 263(a)(i)).
40Ibid., para. 52 (referring to Panel Report, para. 5.33).
41Ibid., para. 56.
WT/DS264/AB/RW
Page 8
Canada also draws attention to the Panel's conclusion that the term "margin of dumping" has a different meaning in Article 9.3 of the Anti-Dumping Agreement than in the rest of that Agreement. Canada argues that this is "manifestly at odds" with the findings of the Appellate Body in US – Zeroing (EC), where it determined that the term "margin of dumping" in Article 9.3 has the same meaning as "margins of dumping" in Article 2.4.2 and prohibits zeroing "in some forms of assessment proceedings".42
In addition, Canada notes that the Panel relied on an interpretation of Article 2.2 of the Anti- Dumping Agreement in the abstract as a "contextual consideration" that requires interpreting Article 2.4.2 to allow zeroing under the transaction-to-transaction comparison methodology. Canada asserts that Article 2.2 was not at issue in this case and that this "consideration", again, reflects the Panel's "misguided effort" to decide this case on the basis of provisions other than the one before it.43 Canada also challenges the Panel's reliance on a single line in a GATT report as negotiating history.44 According to Canada, this report should be accorded no weight as it is already clear that the ordinary meaning of Article 2.4.2 read in its context prohibits zeroing.
Finally, Canada argues that an interpretation of Article 2.4.2 that prohibits zeroing in one of the "normal" calculation methodologies set out in the first sentence of that provision (weighted average-to-weighted average), but permits it under the other (transaction-to-transaction), would lead to "absurd results" in anti-dumping investigations.45 Canada explains that such an interpretation would mean that the choice of a calculation methodology could be determinative as to whether dumping exists, and could create substantial differences in the size of the "margins of dumping".
For these reasons, Canada requests the Appellate Body to reverse the Panel's conclusion and to find, instead, that the use of zeroing in the Section 129 Determination is inconsistent with Article 2.4.2 of the Anti-Dumping Agreement.
2. Article 2.4 of the Anti-Dumping Agreement
Canada asserts that the Panel erred in finding that Article 2.4 of the Anti-Dumping Agreement permits the treatment of transaction-specific comparisons that result in a negative value as zero in the calculation of margins of dumping under the transaction-to-transaction comparison methodology.
42Canada's appellant's submission, para. 57 (referring to Appellate Body Report, US – Zeroing (EC), paras. 127-129).
43Ibid., para. 58.
44Canada's appellant's submission, para. 59 (referring to Panel Report, para. 5.63; and GATT Second Report of the Group of Experts, Anti-Dumping and Countervailing Duties, L/1141, adopted 27 May 1960, BISD 9S/194 ("Group of Experts Report")).
45Ibid., para. 62.
WT/DS264/AB/RW
Page 9
Canada requests the Appellate Body to reverse the Panel's finding and to find, instead, that the USDOC's use of zeroing in the Section 129 Determination is inconsistent with the "fair comparison" requirement in Article 2.4 of the Anti-Dumping Agreement.
Canada notes that the ordinary meaning of the term "fair" is "just, unbiased, equitable, impartial; legitimate, in accordance with the rules or standards".46 Therefore, based on this plain meaning, Article 2.4 must be interpreted as requiring an "equitable" comparison between the export price and the normal value. Canada refers to the Appellate Body's statement in US – Corrosion- Resistant Steel Sunset Review that zeroing inflates margins of dumping and creates an "inherent bias" in these comparisons.47 According to Canada, this bias is more pronounced under the transaction-to- transaction comparison methodology than under the weighted average-to-weighted average methodology because, under the former, zeroing is applied to every non-dumped transaction. Thus, Canada asserts that zeroing intermediate comparison values in a transaction-to-transaction comparison is not consistent with the requirement in Article 2.4 to provide a "fair" or "equitable" comparison between export price and normal value.
Canada submits, moreover, that the term "comparison" in Article 2.4 refers to "the full and true difference between an export price and a normal value".48 It argues that, because the calculation of margins of dumping under the transaction-to-transaction comparison methodology involves multiple comparisons between export prices and selected normal values, the calculation must include the "full results of all of these comparisons".49 In contrast, Canada argues, the USDOC "manipulated" the results of comparisons where the export price was higher than the home market price by disregarding the difference between these prices and replacing that difference with zero. Canada asserts that this "manipulation" is inconsistent with Article 2.4 because it does not lead to a "fair comparison".50
Canada argues that the Panel's conclusion with respect to Article 2.4 is premised on the Panel's incorrect interpretation of Article 2.4.2 as permitting zeroing under the transaction-to- transaction comparison methodology. Canada repeats in this claim the arguments that it put forward in its challenge against the Panel's conclusion under Article 2.4.2 and concludes that, given that
46Canada's appellant's submission, para. 66 (quoting The New Shorter Oxford English Dictionary,
Brown (ed.) (Clarendon Press, 1993), Vol. 1, p. 907 (Exhibit CDA-5 submitted by Canada to the Panel)).
47Ibid., para. 67 (quoting Appellate Body Report, US – Corrosion-Resistant Steel Sunset Review, para. 135).
48Ibid., para. 68 (referring to Appellate Body Report, EC – Bed Linen, para. 55; and Appellate Body Report, US – Corrosion-Resistant Steel Sunset Review, para. 135).
49Ibid., para. 68.
50Ibid., para. 69.
WT/DS264/AB/RW
Page 10
Article 2.4.2 does not allow the use of zeroing when making a transaction-to-transaction comparison, the USDOC's use of zeroing in the Section 129 Determination does not constitute a "fair comparison" within the meaning of Article 2.4.
Finally, because the USDOC's Section 129 Determination is inconsistent with the requirements of Articles 2.4 and 2.4.2 of the Anti-Dumping Agreement, Canada requests the Appellate Body to find that the United States has failed to implement the recommendations and rulings adopted by the DSB in the original dispute.
Arguments of the United States – Appellee
Article 2.4.2 of the Anti-Dumping Agreement
The United States asserts that Canada's appeal should be rejected because, as the Panel correctly found, Article 2.4.2 of the Anti-Dumping Agreement does not require the provision of offsets when the margins of dumping are established on the basis of transaction-to-transaction comparisons of normal value and export prices. According to the United States, Canada's interpretation contradicts the ordinary meaning of the relevant treaty texts and leads to implications that cannot be reconciled with other provisions of the Anti-Dumping Agreement.
The United States rejects Canada's reliance on Article VI of the GATT 1994 to support its assertion that margins of dumping must be established for the "product as a whole" under the transaction-to-transaction comparison methodology set out in Article 2.4.2 of the Anti-Dumping Agreement. The United States points out that the Panel, upon reviewing the use of the term "product" in Article VI of the GATT 1994, noted several instances where that term is used in a manner that refers to a single import transaction. The Panel noted, for example, the references in Article VI to "levy[ing]" a duty on a "product", which suggest transaction-specific events.51 For the United States, depending on the context in which it is used, the term "product" can have either a collective or an individual meaning.52 Where the term "product" is used in a collective sense, a particular model, type, or category constitutes a sub-group of the "product". This does not exclude that, in a different context, the object of a particular transaction can be considered to be the "product". In addition, the United States asserts that Canada's argument "ignores the sense in which the Appellate Body originally used the phrase 'product as a whole'".53 According to the United States, in EC – Bed Linen
51United States' appellee's submission, paras. 18-19 (referring to Panel Report, para. 5.23).
52The United States notes, for example, that Article 2.6 of the Anti-Dumping Agreement refers to the "product under consideration" in the collective sense. In contrast, Article VII:3 of the GATT 1994 uses the term "product" in the individual sense.
53United States' appellee's submission, para. 24.
WT/DS264/AB/RW
Page 11
and US – Softwood Lumber V, the Appellate Body used the phrase "product as a whole" for the purpose of distinguishing the margin of dumping for a product from the dumped amounts found to exist for sub-products based on multiple weighted-average comparisons. The United States alleges that Canada's use of the phrase "product as a whole" in this appeal "represents a dramatic departure from the original sense in which the Appellate Body used the phrase".54
The United States submits, furthermore, that the Panel correctly concluded that a margin of dumping may be established on a transaction-specific basis. The United States observes that the Panel "based its finding firmly in the text of Article VI of the GATT 1994"55, which provides that there is dumping when the export price is less than the normal value. Given this definition of dumping, the United States agrees with the Panel that it would be permissible for a Member to interpret the term "margin of dumping" as the amount by which the export price is less than the normal value. The United States disagrees, in this regard, with Canada's characterization of the results of transaction-specific comparisons as "intermediate values" and Canada's argument that, in establishing the margin of dumping, the investigating authority must take into account the full value of all intermediate comparison results, be they positive or negative.56 The United States also rejects Canada's argument that the grammatical construction of the first sentence of Article 2.4.2 means that the term "margins of dumping" must have the same meaning with respect to transaction-to-transaction comparisons as it has with respect to weighted average-to-weighted average comparisons. According to the United States, this argument wrongly assumes that the "product as a whole" is a generic concept, equally applicable to different comparison methodologies, even though the Appellate Body used that term in the context of weighted average-to-weighted average comparisons. Furthermore, the United States submits that Canada "effectively invents"57 an obligation found nowhere in the text of Article 2.4.2, which contains no express language requiring offsets in the calculation of "margins of dumping".
The United States notes that Canada seeks support in Article 6.10 of the Anti-Dumping Agreement for the proposition that a margin of dumping can be established only with respect to a "product as a whole". The United States responds that Article 6.10 merely provides that a Member must calculate a margin of dumping for each individual exporter or producer, but says nothing about whether the margin must be based on more than one transaction. For the United States, nothing
54United States' appellee's submission, para. 28.
55Ibid., para. 41.
56Ibid., para. 42 (referring to Canada's appellant's submission, para. 37).
57Ibid., para. 47.
WT/DS264/AB/RW
Page 12
prevents a margin of dumping from being calculated on both a transaction-specific basis and an exporter-specific basis at the same time.58
The United States considers that Canada's reliance on previous reports of the Appellate Body is misplaced. According to the United States, Canada essentially urged the Panel to separate the reasoning of the Appellate Body Report in US – Softwood Lumber V from the context in which it was articulated and to apply it to the different context of the present dispute. The United States emphasizes that, in US – Softwood Lumber V, the Appellate Body referred to the phrase "all comparable export transactions" in Article 2.4.2 as an element of context within which the term at issue, "margins of dumping", should be construed. This phrase, notes the United States, pertains only to the weighted average-to-weighted average comparison methodology. Furthermore, the United States maintains that Canada dismisses the significance of the Appellate Body's decision in US – Softwood Lumber V to refrain from ruling on the transaction-to-transaction comparison methodology that was raised as relevant context.
The United States also rejects Canada's reliance on the Appellate Body Report in US – Zeroing (EC), which the United States considers to be irrelevant to the question at issue in this dispute—that is, whether zeroing is allowed in transaction-to-transaction comparisons in an original investigation. Furthermore, the United States considers that the Appellate Body's finding in US – Zeroing (EC) "appears to be based on a misinterpretation of a finding of the original Appellate Body report" in US – Softwood Lumber V.59 The United States explains that, in US – Zeroing (EC), the concept of the "product as a whole" was "inexplicably broadened from the multiple averaging context to the context of 'multiple comparisons'"60, such that, if multiple comparisons are undertaken, zeroing is prohibited. According to the United States, such reasoning renders "illusory"61 a WTO Member's right to choose whether to use multiple comparisons in calculating the margin of dumping, because a Member must always aggregate transactions for calculating a margin of dumping. Finally, the United States considers that Canada's argument based on the Appellate Body Report in US –
Zeroing (EC) effectively makes redundant the phrase "all comparable export transactions". This is because Canada's reasoning leads to a general obligation to establish the margin of dumping for the "product as a whole", regardless of the comparison methodology. The phrase "all comparable export transactions", however, is used only with reference to the weighted average-to-weighted average comparison methodology.
58United States' appellee's submission, para. 52.
59Ibid., para. 64.
60Ibid., para. 65. (original emphasis)
61Ibid., para. 65.
WT/DS264/AB/RW
Page 13
The United States asserts that the Panel correctly found that simply extending the finding from the Appellate Body Report in US – Softwood Lumber V to the transaction-to-transaction comparison methodology, as Canada proposed, would result in a number of difficulties. First, the United States submits that Canada's argument that "the term 'product' in Article VI of the GATT 1994 necessarily refers to the entire universe of exported product subject to an antidumping investigation hinges importantly on the use of that term in paragraphs (1) and (2) of Article VI", but is undercut by paragraphs 6(a) and (b) that refer to "'levy[ing]' a duty on the 'importation of any product'".62 Moreover, the United States asserts that the term "margin of dumping", as used in Note Ad Article VI:1 of the GATT 1994, refers to a single export transaction through an associated importer and, therefore, contradicts Canada's suggestion that a margin of dumping necessarily means a margin of dumping for the "product as a whole". Secondly, the United States refers to Article 2.2 of the Anti-Dumping Agreement, which provides that constructed normal value may be used where there are no sales in the ordinary course of trade. According to the United States, the practice of many Members is to resort to constructed normal value on a model- or transaction-specific basis. However, under Canada's interpretation, the United States submits, an investigating authority would be required to use constructed normal value for the "product as a whole" even though its use might only be necessary for the model not sold in the ordinary course of trade.
Thirdly, the United States asserts that, if the "margins of dumping" must invariably be calculated for the "product as a whole", zeroing would also be prohibited under the second sentence of Article 2.4.2, which permits weighted average-to-transaction comparisons if certain conditions are met, including a "pattern of export prices which differ significantly among different purchasers, regions or time periods". Calculating, in a targeted dumping scenario, one margin of dumping for transactions falling within the specified pricing pattern and another for all other transactions would also be inconsistent with Canada's "product as a whole" theory.63 Moreover, according to the United States, this would nullify the second sentence of Article 2.4.2 because, without zeroing, the results of the weighted average-to-weighted average comparison methodology in the first sentence of Article 2.4.2 would be mathematically equivalent to those of the weighted average-to-transaction methodology in the second sentence of Article 2.4.2.
In addition, the United States refers to the problem that Canada's interpretation would raise in the context of prospective normal-value systems. The United States explains that, under a prospective normal-value system, a WTO Member determines the liability for anti-dumping duties at the time of importation by comparing the export price with a prospective normal value. If the export price is
62United States' appellee's submission, para. 69.
63Ibid., para. 84.
WT/DS264/AB/RW
Page 14
greater than the prospective normal value, the Member collects no anti-dumping duty, but does not give the importer a credit against other export transactions. Thus, the assessment is based on a single transaction.64 The United States argues that a general obligation to establish the margin of dumping for the "product as a whole" would require a Member applying a prospective normal-value system to take all export transactions into consideration. As a result, the anti-dumping duties owed by one importer would be offset by the amount of "non-dumping" determined with respect to another importer.65 In this respect, the United States concurs with the Panel that such a result "makes no sense".66
Finally, the United States refers to the past experience regarding the application of Article VI of the GATT 1947, which, in its view, contradicts Canada's interpretation of the term "margins of dumping". The United States refers to a 1960 report by a Group of Experts concerning Article VI of the GATT 1947, which states that the ideal method for applying anti-dumping duties was "to make a determination in respect of both dumping and material injury in respect of each single importation of the product concerned."67 In the United States' view, the Panel properly considered this report and found that "the Group of Experts clearly envisaged the calculation of transaction-specific margins of dumping."68 In addition, the United States refers to two pre-WTO panel reports69, in which the panels rejected arguments that failure to take into account negative margins was inconsistent with the Tokyo Round Agreement on Implementation of Article VI of the General Agreement on Tariffs and Trade70 (the "Tokyo Round Anti-Dumping Code"), the predecessor to the Anti-Dumping Agreement. The United States maintains that none of the parties in those disputes considered Article VI of the GATT 1947 as containing a "product as a whole" concept.
The United States also recalls that, at the time of the Uruguay Round negotiations, during which the Anti-Dumping Agreement was negotiated, it was common among major users of anti- dumping measures to use zeroing when aggregating individual comparisons. Thus, it is "completely inconsistent with these historical facts, to suggest that any Contracting Party or negotiator believed that, without any modification to Article VI [of the GATT 1947], the word 'product' in that provision
64United States' appellee's submission, para. 86.
65Ibid., para. 89.
66Ibid., para. 91 (quoting Panel Report, para. 5.57).
67Ibid., para. 34 (quoting Group of Experts Report, supra, footnote 44, para. 8).
68Ibid., para. 34 (quoting Panel Report, para. 5.64).
69GATT Panel Report, EC – Audio Cassettes, para. 360 (unadopted); GATT Panel Report,
EEC – Cotton Yarn, para. 502.
70BISD 26S/171.
WT/DS264/AB/RW
Page 15
as carried into the GATT 1994" effectively would prohibit zeroing.71 In the light of this background, the United States submits that "it hardly can be claimed that the understanding of the terms 'product' and 'margins of dumping' urged by the United States and supported by the Panel in the present dispute is impermissible" under the applicable standard of review set out in Article 17.6(ii) of the Anti- Dumping Agreement.72
For these reasons, the United States requests the Appellate Body to uphold the Panel's finding that the Section 129 Determination is not inconsistent with Article 2.4.2 of the Anti-Dumping Agreement.
2. Article 2.4 of the Anti-Dumping Agreement
The United States submits that the Panel correctly rejected Canada's argument that the Section 129 Determination is inconsistent with the "fair comparison" requirement in Article 2.4 of the Anti-Dumping Agreement. The United States asserts that Canada's argument assumes its own conclusion by simply insisting that a margin established without zeroing is fair. In addition, the United States submits that Canada's argument must fail in the light of the fact that zeroing in the context of transaction-to-transaction comparisons is permissible under Article 2.4.2. As the Panel correctly stated, in the Unites States' view, the very general "fair comparison" requirement of Article 2.4 cannot trump the more specific provisions of Article 2.4.2.73
According to the United States, Canada's argument is that, if a given comparison methodology yields a higher margin of dumping than another comparison methodology, then the first methodology is unfair. Such an argument, the United States observes, incorrectly assumes that the magnitude of the resulting margin of dumping is the basis for determining whether the selection of a comparison methodology is "fair" during the investigation phase. This would mean that investigating authorities would have to determine at least two putative margins of dumping for each exporter, using at least two comparison methodologies provided in Article 2.4.2, and ultimately rely on the methodology that generated the lowest margin. The United States sees nothing in the Anti-Dumping Agreement that imposes such a results-driven obligation.74 In addition, the United States refers to the contextual considerations referred to by the Panel in relation to Article 2.4.2. In particular, the United States reiterates that, should Canada's interpretation be accepted, the second sentence of Article 2.4.2 would be rendered inutile due to the mathematically equivalent results reached under the weighted
71United States' appellee's submission, para. 39. (original emphasis)
72Ibid., para. 40.
73Ibid., para. 98.
74Ibid., para. 100.
WT/DS264/AB/RW
Page 16
average to-transaction comparison methodology provided in this sentence and the weighted average- to-weighted average methodology provided in the first sentence of Article 2.4.2. Finally, the United States asserts that the Appellate Body reports cited by Canada75 do not support Canada's argument.76
Consequently, the United States requests the Appellate Body to uphold the Panel's finding that the Section 129 Determination is not inconsistent with the "fair comparison" requirement in Article 2.4 of the Anti-Dumping Agreement.
Arguments of the Third Participants
China
Pursuant to Rule 24(2) of the Working Procedures, China chose not to submit a third participant's submission. In its statement at the oral hearing, China agreed with Canada that the Panel erred in concluding that zeroing under the transaction-to-transaction comparison methodology was not inconsistent with Articles 2.4.2 and 2.4 of the Anti-Dumping Agreement.
2. European Communities
The European Communities argues that, "[i]n principle"77, an exporter's margin of dumping must be established for the product under investigation "as a whole". Thus, although an investigating authority may undertake multiple comparisons at an intermediate stage, it is only by aggregating the results of all such intermediate comparisons that an investigating authority can establish an exporter's margin of dumping for the "product as a whole". According to the European Communities, there is no justification for taking into account only some of the results of multiple comparisons while disregarding the other results. Moreover, the European Communities observes that other provisions of the Anti-Dumping Agreement are explicit regarding the permissibility of disregarding certain matters. As examples, the European Communities refers to Article 2.2.1 (that allows certain domestic sales to be disregarded), Article 9.4 (that directs investigating authorities to disregard zero and de minimis margins of dumping), Article 2.7 (that excludes application of the disciplines of Article 2 in cases involving non-market economies), and Article 6.8 and Annex II (that permit investigating authorities to disregard certain information).
75Canada referred to Appellate Body Report, EC – Bed Linen, para. 55; and Appellate Body Report,
US – Corrosion-Resistant Steel Sunset Review, para. 135. (See Canada's appellant's submission, paras. 67-68)
76United States' appellee's submission, para. 104.
77European Communities' third participant's submission, para. 6.
WT/DS264/AB/RW
Page 17
The European Communities disagrees with the United States that the term "margin of dumping" can be interpreted as applying on a transaction-specific rather than an exporter-specific basis. According to the European Communities, Article 6.10 confirms that an investigating authority is required, as a rule, to determine an individual margin of dumping for each known exporter of the product under investigation. Moreover, Article 2.4.2 provides that transaction-to-transaction juxtapositions may be the basis for the calculation of a margin of dumping. In other words, this indicates that there is one thing (one or more transaction-to-transaction comparisons) that is an underlying support for something else (a margin of dumping). Therefore, although a first step may involve more than one transaction-to-transaction comparison, in a subsequent step of calculation, all intermediate results of the first step must be aggregated. The European Communities asserts that such a proposition is supported by the first sentence of Article 2.4, which uses the singular when it provides that a "fair comparison" must be made between the export price and the normal value. Also, Article VI of the GATT 1994 consistently refers to "normal value" in the singular. The European Communities finds support for its interpretation in the term "normal value", in that it would be "highly problematic" for an investigating authority to assume that the price at which a single domestic transaction is concluded is equivalent to a "normal value" simply because, in the first step of the calculation, the transaction-to-transaction method of juxtaposition is used.78 The European Communities asserts that the phrase "ordinary course of trade" in Article 2.2 provides further contextual support because, if one transaction is isolated from many other transactions at varying prices, that single transaction cannot be taken as representative of the "ordinary course of trade" and cannot, on its own, be considered a "normal value".
Recalling the definitions of "dumping" and "margin of dumping" in Article VI of the GATT 1994, as implemented in Article 2 of the Anti-Dumping Agreement, the European Communities argues that the intermediate results of a series of transaction-to-transaction comparisons cannot be "margins of dumping" because they do not involve a comparison between normal value and export price. For the European Communities, it is not possible to measure fairly international price discrimination between two different markets if the fundamental methodology for defining and measuring behaviour in each of the markets is different. Nor does price variation or volatility justify zeroing. Moreover, the fact that market conditions in the United States and in the home market may be different (for example, demand may be different) has nothing to do with a comparison between normal value (in one market) and export price (in the other market), and is irrelevant for determining how dumping should be measured—assuming that all appropriate adjustments for any differences affecting price comparability have been made in accordance with Article 2.4.
78European Communities' third participant's submission, para. 17.
WT/DS264/AB/RW
Page 18
In the European Communities' view, the true question before the USDOC was how the United States market and the home market can be defined, independently, and in terms of product, geography, and time.79 What the USDOC has done, however, by using a transaction-to-transaction methodology with zeroing, is to posit that the home market should be temporally defined by reference to each transaction. If the United States wanted to address a situation where there is "a pattern of export prices which differ significantly among different ... time periods"80, then a targeted dumping analysis on the basis of a weighted average-to-transaction comparison would have been the only lawful approach.
The European Communities rejects the "mathematical equivalence" argument because the measure taken to comply at issue in this dispute is based on the transaction-to-transaction methodology, whereas the Panel's discussion of the "mathematical equivalence" was hypothetical and related to the two other methods of comparison. According to the European Communities, even assuming that zeroing is permitted under the targeted dumping provision, it does not follow that zeroing must also be permitted when using the transaction-to-transaction comparison: something that might be justified by exceptional circumstances is not, for that reason, justified in all circumstances and as a general and normal method of calculation. The European Communities further submits that because the number of transactions in the domestic and export markets, in practice, will rarely, if ever, be the same, the transaction-to-transaction comparison methodology without zeroing will not produce a mathematical result equivalent to that obtained when using either of the other two methodologies.
With respect to the United States' argument regarding prospective normal-value systems, the European Communities asserts that the Panel erred when it proceeded as if it were discussing final duty assessment based on contemporaneous normal value and export price (within the meaning of Articles 9.3.1 and 9.3.2), although what the Panel was in fact discussing was the initial payment based on prospective normal value. The European Communities also submits that Article 9.4 concerns the issue of sampling and, for that reason alone, is irrelevant to the present case, where that issue does not arise. Furthermore, the European Communities recalls that similar discussions regarding the situation of importers in a duty assessment proceeding were rejected by the Appellate Body in US – Zeroing (EC).81 In this regard, the European Communities notes that the Anti-Dumping Agreement addresses the behaviour of exporters (not the relative position of importers), and an investigating authority must calculate a margin of dumping for each exporter (not each importer). According to the European Communities, therefore, the total amount of duties collected from importers purchasing
79European Communities' third participant's submission, para. 45.
80Ibid., para. 46. (original underlining)
81Ibid., para. 70 (referring to Appellate Body Report, US – Zeroing (EC), paras. 130-131 and footnote 234).
WT/DS264/AB/RW
Page 19
from a particular exporter cannot exceed the ceiling of the margin of dumping for that exporter.82 The European Communities explains that, if necessary, the duty amounts calculated with respect to specific importers must be reduced pro rata until that ceiling is respected, but this does not mean that importers with a "negative" duty amount are "offsetting" dumping margins, as the United States argues.
The European Communities also takes issue with the Panel's reasoning regarding constructed normal value under Article 2.2. The European Communities states that it is clear that Article 2.2 of the Anti-Dumping Agreement is drafted in such a manner as to describe the various means by which a normal value may be established. This does not mean that the practice of making multiple comparisons at sub-group level is not possible, provided that an investigating authority eventually aggregates in full all of the results of such multiple comparisons in order to calculate a margin of dumping for the "product as a whole", and, in doing so, it is not entitled to "zero" negative intermediate results.83
The European Communities, therefore, considers that the Appellate Body should reverse the Panel's finding that the Section 129 Determination is not inconsistent with Articles 2.4.2 and 2.4 of the Anti-Dumping Agreement.
3. India
Pursuant to Rule 24(2) of the Working Procedures, India chose not to submit a third participant's submission. In its statement at the oral hearing, India presented its views as to why it considers that the Panel erred in finding that the use of zeroing in a transaction-to-transaction comparison methodology is not inconsistent with Articles 2.4.2 and 2.4 of the Anti-Dumping Agreement.
4. Japan
Japan disagrees with the Panel's finding that zeroing in a transaction-to-transaction comparison is permitted under Article 2.4.2 of the Anti-Dumping Agreement. Japan emphasizes that "dumping" is not determined on the basis of individual transaction-to-transaction comparisons, and the price differences that arise in individual comparisons do not, on their own, "show" that there is dumping.84 According to Japan, in order to establish "dumping" and "margins of dumping" under the Anti-Dumping Agreement and Article VI of the GATT 1994, an investigating authority must take into
82European Communities' third participant's submission, para. 70.
83Ibid., para. 74.
84Japan's third participant's submission, para. 11.
WT/DS264/AB/RW
Page 20
account the "product as a whole".85 Whenever the USDOC disregards negative comparison results, the dumping determination is based on only a partial aggregation of comparison results that takes into account some, but not all, export transactions. However, the meanings of "dumping" and "margins of dumping" should not change because only individual export transactions are involved in the comparison method. Japan further argues that the first sentence of Article 2.4.2 by no means indicates that the term "margins of dumping" encompasses "two quite different meanings" in the same sentence.86 In Japan's view, the prohibition of zeroing stems from the meaning of "dumping" in Article 2.1 of the Anti-Dumping Agreement and in Article VI:1 of the GATT 1994, and therefore should extend equally to the weighted average-to-weighted average and the transaction-to-transaction comparison methodologies.87 Japan adds that, by defining "dumping" in relation to a product, the Anti-Dumping Agreement ensures parallelism between the scope of a dumping determination and the scope of the regulatory consequences the determination entails.
Japan also takes issue with the Panel's analysis of broader contextual considerations. First, Japan considers that the Panel wrongly held that prohibiting zeroing in a transaction-to-transaction comparison "would deprive the second sentence of Article 2.4.2 of effect".88 Japan asserts that the Panel made this finding because it agreed with the United States' position that, without the use of zeroing, a weighted average-to-weighted average comparison and a weighted average-to-transaction comparison pursuant to the second sentence of Article 2.4.2 of the Anti-Dumping Agreement would lead to the same mathematical outcome. According to Japan, properly interpreted and absent zeroing, weighted average-to-transaction comparisons produce different outcomes from those resulting from weighted average-to-weighted average comparisons. Japan also notes that the "mathematical equivalence" argument involves the two comparison methodologies that are not at issue in this dispute, which concerns the transaction-to-transaction methodology under the first sentence of Article 2.4.2. Thus, even assuming arguendo that the second sentence of Article 2.4.2 permits zeroing on an exceptional basis, that exception is not relevant in this dispute. In any event, Japan argues that, by extending a right to zero from the second sentence to the first sentence of Article 2.4.2, the Panel improperly transforms an exceptional right into a general rule. Japan further submits that the structure of the exceptional weighted average-to-transaction comparison must be rooted in the circumstances that justify its use, namely, the existence of a particular pricing pattern "which differ[s] significantly among different purchasers, regions or time periods". This is also the interpretation
85Japan's third participant's submission, para. 35 (quoting Appellate Body Report, US – Zeroing (EC), para. 126). See also Appellate Body Report, US – Softwood Lumber V, paras. 93 and 99; and Appellate Body Report, US – Corrosion-Resistant Steel Sunset Review, paras. 126 and 127.
86Japan's third participant's submission, para. 37.
87Ibid., paras. 34 and 38 (referring to Appellate Body Report, US – Zeroing (EC), paras. 125-126).
88Ibid., para. 13 (referring to Panel Report, para. 5.52).
WT/DS264/AB/RW
Page 21
adopted, according to Japan, by the USDOC in its own regulations.89 Therefore, contrary to the Panel's assumptions, the universe of export transactions in an exceptional weighted average-to- transaction comparison is narrower than in a weighted average-to-weighted average comparison. Consequently, absent zeroing, the outcome of the two comparison methodologies differ. In addition, Japan notes that, even assuming arguendo that a weighted average-to-transaction comparison must be applied to all export transactions without zeroing, the second sentence of Article 2.4.2 is not rendered a nullity because it has not been shown that the outcome is always the same as under a weighted average-to-weighted average comparison.
Secondly, Japan states that the Panel's analysis of the implications for prospective normal- value systems confuses two concepts, the "amount of anti-dumping duty" and the "margin of dumping", which are distinct and separate.90 Japan submits that the rules in Article 9 on the imposition and collection of anti-dumping duties do not have a bearing on the rules in Article 2 on the determination of dumping margins. Japan argues that Members are entitled to impose variable duties on the basis of a prospective normal value, as envisaged by Article 9.4(ii), but the amount of variable duties initially imposed is not a transaction-specific "margin of dumping" determined under Article 2. Rather, relying upon the Appellate Body's ruling in US – Zeroing (EC), Japan argues that the margin of dumping operates as a "ceiling" on the amount of anti-dumping duties and is distinguishable from "the amount of duties initially imposed on entries of a product".91 Subject to this proviso, Members may also appropriately apportion liability among importers to avoid disturbing competitive conditions. According to Japan, there is no basis for the United States' argument that investigating authorities would be obliged to refund to importers an amount that exceeds the duties initially paid in the event that export prices were higher than normal value for the "product as a whole".92
Thirdly, Japan asserts that the Panel misread Article 2.2 of the Anti-Dumping Agreement, thereby incorrectly finding that, if the term "margin of dumping" always refers to the "product as a whole", constructed normal value can no longer be determined on a model-specific basis.93 Japan submits that, although the texts of Articles 2.2 and 2.4.2 are silent in terms of the comparison of normal value and export price on a model-specific basis, the Appellate Body in US – Softwood Lumber V held that such a comparison is permissible.94 Thus, Articles 2.2 and 2.4.2 must be
89Japan's third participant's submission, para. 14 (referring to United States Code of Federal Regulations, Title 19, Section 351.414(f)(2)).
90Ibid., para. 71. (emphasis omitted)
91Ibid., para. 73 (referring to Appellate Body Report, US – Zeroing (EC), para. 130).
92Ibid., para. 76 (referring to Appellate Body Report, US – Zeroing (EC), footnote 234 to para. 131).
93Panel Report, para. 5.62.
94Japan's third participant's submission, para. 81 (referring to Appellate Body Report, US – Softwood Lumber V, para. 80).
WT/DS264/AB/RW
Page 22
interpreted harmoniously to accommodate the right to calculate constructed normal value for some models, but not for others, provided that the margin of dumping is determined on the basis of the product under investigation as a whole.
Finally, Japan argues that, contrary to the Panel's finding, the use of zeroing in a transaction- to-transaction comparison violates the "fair comparison" requirement in Article 2.4 of the Anti- Dumping Agreement. Japan asserts that zeroing involves an "inherent bias"95 that is the very "antithesis" of fairness.96 Japan explains that this is because the positive comparison results included in the determination relate to export transactions with prices that are lower than normal value; in contrast, the excluded negative results relate to export transactions with prices higher than normal value. The export transactions selected for inclusion in the determination, therefore, relate to the "sub-part" of the product that is the most likely to generate an affirmative dumping determination and inflate the margin of dumping, because, for excluded transactions, the export prices are treated as if
they were less than what they actually are.
Consequently, Japan considers that the Appellate Body should reverse the Panel's findings and find, instead, that the Section 129 Determination is inconsistent with Articles 2.4.2 and 2.4 of the Anti-Dumping Agreement.
5. New Zealand
At the outset, New Zealand emphasizes that it does not intend to "defend 'zeroing'".97 Rather, it is New Zealand's concern that interpretations of the Anti-Dumping Agreement developed in the context of the weighted average-to-weighted average comparison methodology are not applied in such a way as to prohibit permissible applications of the transaction-to-transaction methodology.
New Zealand argues that the transaction-to-transaction comparison methodology "does not involve making 'multiple comparisons' producing a number of 'intermediate values or results'".98 Thus, the transaction-to-transaction comparison methodology is not an averaging methodology requiring "multiple comparisons", and should not be construed as such. New Zealand submits that the transaction-to-transaction comparison methodology involves the establishment of dumping margins on a transaction-specific basis. If the export price is less than the comparable home market price, dumping has occurred. In New Zealand's view, Canada's interpretation blurs the distinction between
95Appellate Body Report, US – Corrosion-Resistant Steel Sunset Review, para. 135.
96Japan's third participant's submission, para. 21.
97New Zealand's third participant's submission, para. 2.03.
98Ibid., para. 1.05.
WT/DS264/AB/RW
Page 23
the two methodologies set out in the first sentence of Article 2.4.2, namely, the weighted average-to- weighted average and transaction-to-transaction comparison methodologies.
According to New Zealand, the focus of an anti-dumping investigation is to determine whether there has been injurious dumping. Focusing on the export transactions below normal value under the transaction-to-transaction comparison methodology does not necessarily inflate or distort the margin of dumping. New Zealand explains that, in a transaction-to-transaction analysis, only the volume of dumped products, rather than the entire volume of imported goods, should be relied upon to ascertain the appropriate dumping margin. The volume and prices of the export transactions that are above normal value "are fully taken into account" in the non-attribution analysis required under Article 3.5 of the Anti-Dumping Agreement.99 New Zealand asserts, in this regard, that there is "no distortion" of the margins of dumping where "there is symmetry or parallelism between the dumping investigation and the injury/causation investigation".100 In New Zealand's view, the transaction-to- transaction comparison methodology can, and should, be interpreted and applied in a way that ensures "consistent treatment" of the product under investigation.101 According to New Zealand, "Canada's concern that a transaction-specific methodology inflates dumping margins is not warranted", because the dumping margin is not artificially assigned to the entire volume of imports.102
New Zealand submits that Canada incorrectly draws from the notion of "product as a whole" a requirement to examine average pricing behaviour of an exporter throughout the period of investigation. While such an interpretation may be possible in the context of a weighted average-to- weighted average comparison methodology, there is no basis for making this a requirement under a transaction-to-transaction methodology. In New Zealand's view, this would reduce the transaction-to- transaction comparison methodology to an averaging methodology. New Zealand disagrees with Canada's proposition that the term "product as a whole" relates to the overall volume of imports and asserts that, instead, it refers to the scope of the product subject to investigation. In addition, New Zealand argues that Canada is wrong in dismissing the broader contextual considerations relied on by the Panel. New Zealand notes, in this regard, that the provisions of the Anti-Dumping Agreement "work in an integrated fashion to provide relief for instances of injurious dumping".103
99New Zealand's third participant's submission, para. 1.09.
100Ibid., para. 1.10.
101Ibid., para. 3.33.
102Ibid., para. 1.10.
103Ibid., para. 3.35.
WT/DS264/AB/RW
Page 24
6. Thailand
Thailand disagrees with the Panel's analysis and urges the Appellate Body to reverse the Panel's finding that the use of zeroing under the transaction-to-transaction comparison methodology is consistent with Articles 2.4.2 and 2.4 of the Anti-Dumping Agreement.
Thailand submits that the Panel misunderstood the implications of the Appellate Body's previous rulings that Article 2.1 of the Anti-Dumping Agreement requires an investigating authority to calculate a single margin of dumping for the product under investigation as a whole.104 Thailand explains that, if a single margin of dumping must be calculated for the product under investigation as a whole, then any interim steps to that end do not constitute determinations of dumping within the meaning of Article 2.1. In calculating a single margin of dumping, therefore, an investigating authority cannot prejudge the issue of whether dumping exists by including in its calculation only transactions that are likely to lead to a finding of dumping and by excluding those that reduce the likelihood of such a finding. Thailand emphasizes that the Appellate Body clarified in US – Zeroing (EC) that its reasoning in US – Softwood Lumber V was not limited to issues involving the use of zeroing in weighted average-to-weighted average comparisons, but applies to all instances in which investigating authorities make multiple comparisons at an intermediate stage and aggregate the results of those comparisons in calculating margins of dumping. Thailand finds contextual support in Articles 5.8, 6.10, and 9.2 of the Anti-Dumping Agreement for the notion that margins of dumping must be calculated for the product under investigation as a whole.
In addition, Thailand argues that the Panel's reliance on broader contextual considerations was misplaced. Thailand asserts that there is mathematical equivalence between the weighted average-to- weighted average and the weighted average-to-transaction comparison methodologies only if the same weighted-average normal value is used in both methodologies. According to Thailand, this would not occur in practice, even under the United States' system, because it does not use the same normal value for each methodology. Moreover, the exchange rate used to convert the domestic and export currencies could also change. Therefore, Thailand does not see why allowing zeroing is necessary in order to give effect to the weighted average-to-transaction comparison methodology referred to in the second sentence of Article 2.4.2. Moreover, that sentence refers to the use of weighted average-to- transaction comparisons in situations where export prices differ significantly among different purchasers, regions, or time periods, and an explanation is given as to why such differences would not be taken into account by either of the comparison methodologies referred to in the first sentence of Article 2.4.2. The second sentence does not, however, refer to or permit the use of zeroing, anymore
104Thailand's third participant's submission, paras. 21-22 (referring to Appellate Body Report, US – Zeroing (EC), para. 126).
WT/DS264/AB/RW
Page 25
than the first sentence of that provision. Thailand adds that, in any event, the argument that zeroing may be necessary in order to give effect to the weighted average-to-transaction comparison methodology referred to in the second sentence of Article 2.4.2, would suggest that zeroing is therefore not permitted under either of the comparison methodologies referred to in the first sentence of Article 2.4.2, that is, weighted average-to-weighted average and transaction-to-transaction.
Thailand points out, furthermore, that there is nothing inherent in a prospective normal-value system of collecting duties on importation that is inconsistent with a prohibition on zeroing. Under any system of collecting anti-dumping duties—prospective or retrospective—the amount collected on each importation represents merely an estimate of the final duty liability and does not necessarily reflect either the difference between export price and normal value for that transaction or the dumping margin for the "product as a whole". For this reason, the Anti-Dumping Agreement provides for the possibility of a review or refund procedure under Articles 9.3.1 (retrospective) and 9.3.2 (prospective) to determine final duty liability so as to ensure that, in accordance with the chapeau of Article 9.3, the amount of the anti-dumping duty does not exceed the margin of dumping established in accordance with Article 2.
Moreover, Thailand states that the reference to the term "margin of dumping" in Article 2.2 does not mean that, if that term were interpreted to refer to the margin of dumping for the product as a whole, the investigating authority would, as the Panel suggests, be prohibited from calculating constructed normal value on a model-specific basis. Thailand submits that an investigating authority has ample discretion to interpret the phrase "sales of the like product" in the first sentence of Article 2.2 to refer to sales of particular models within the like product and to use constructed normal value on a model-specific basis provided that the margin of dumping is determined on the basis of all intermediate model-specific comparisons.
Finally, Thailand argues that the Panel erred in concluding that, because it had found zeroing to be consistent with Article 2.4.2, it could not conclude that it was inconsistent with Article 2.4. Instead, Thailand asserts, the Panel should have interpreted Article 2.4 as an independent obligation that informs all of Article 2. It should have concluded that it required an impartial, objective, and unbiased analysis in determining the existence of dumping. Because zeroing involves the systematic exclusion, from the determination of whether dumping exists, of sales that are likely to lead to a contrary outcome, the use of zeroing is neither impartial, objective, or unbiased, and is therefore inconsistent with Article 2.4.
WT/DS264/AB/RW
Page 26
Issues Raised in This Appeal
The following issues are raised in this appeal:
whether the Panel erred in finding that the use of zeroing when margins of dumping are established by comparing normal value and export prices on a transaction-to- transaction basis is not inconsistent with Article 2.4.2 of the Anti-Dumping Agreement; and
whether the Panel erred in finding that the use of zeroing when margins of dumping are established by comparing normal value and export prices on a transaction-to- transaction basis is not inconsistent with the requirement of "fair comparison" in Article 2.4 of the Anti-Dumping Agreement.
Article 2.4.2 of the Anti-Dumping Agreement
Introduction
Canada claims that the use of zeroing by the United States Department of Commerce (the "USDOC") in the Section 129 Determination105 is inconsistent with the United States' obligations under Articles 2.4 and 2.4.2 of the Anti-Dumping Agreement. We begin with the claim under Article 2.4.2 because it is the main focus of Canada's argumentation. Canada's claim under Article 2.4 is addressed in Section V of this Report.
This is not the first time that the issue of zeroing has been raised on appeal. The Appellate Body has previously examined the use of zeroing by investigating authorities in the context of original investigations106, sunset reviews107, and duty assessment proceedings.108 The use of zeroing was
105Section 129 Determination, supra, footnote 11.
106Appellate Body Report, EC – Bed Linen, para. 66; Appellate Body Report, US – Softwood Lumber V, para. 117.
107Appellate Body Report, US – Corrosion-Resistant Steel Sunset Review, paras. 127-128 and 133-138.
108Appellate Body Report, US – Zeroing (EC), paras. 132-135.
WT/DS264/AB/RW
Page 27
found to be inconsistent with Article 2.4.2 in both disputes involving original investigations, and it was found to be inconsistent with Article 9.3 in the dispute involving duty assessment proceedings.109
In the two previous appeals involving zeroing in an original investigation, the investigating authorities had established the margins of dumping by comparing weighted-average normal value with a weighted average of export prices. One of these disputes concerned the USDOC's original dumping determination against softwood lumber imports from Canada. In that appeal, the Appellate Body upheld the panel's finding that the use of zeroing under the weighted average-to-weighted average comparison methodology was inconsistent with Article 2.4.2.110 The Section 129 Determination, which is the subject of this appeal, was taken to comply with the findings of the Appellate Body in the original US – Softwood Lumber V dispute.111 In the Section 129 Determination, the USDOC switched to a different methodology. The USDOC explained that, "[i]n light of the Appellate Body's findings and recommendations" concerning the weighted average-to- weighted average comparison methodology, "[i]t ha[d] determined to apply the transaction-to- transaction methodology in th[e] Section 129 Determination" to establish the margins of dumping.112 Certain Canadian parties objected to the USDOC switching methodologies in the Section 129 Determination.113 This argument was rejected by the USDOC, which stated that the Section 129 Determination is a "new", "second", or "different" determination and, consequently, it "may modify its calculations or methodologies to effectuate its compliance with a WTO decision".114 The USDOC added that "the calculations or methodologies that are necessary to implement [the Appellate Body's] decision rest largely with the discretion of the [USDOC], the United States Trade Representative and [the United States] Congress".115 In addition, the USDOC stated that, under United States law, it could calculate the margins of dumping using either the weighted average-to-weighted average or the
109In an appeal involving the use of zeroing in a sunset review, the Appellate Body stated that it saw "no obligation under Article 11.3 [of the Anti-Dumping Agreement] for investigating authorities to calculate or rely on dumping margins in determining the likelihood of continuation or recurrence of dumping". It added, nevertheless, that "should investigating authorities choose to rely upon dumping margins in making their likelihood determination, the calculation of these margins must conform to the disciplines of Article 2.4." (Appellate Body Report, US – Corrosion-Resistant Steel Sunset Review, para. 127) The Appellate Body was unable to complete the analysis and rule on the consistency of zeroing in that particular case because of insufficient factual findings. (paras. 133-138)
110Appellate Body Report, US – Softwood Lumber V, para. 117.
111United States' appellee's submission, para. 11.
112Section 129 Determination, p. 22637.
113Ibid., p. 22640.
114Ibid., p. 22640 (referring to Statement of Administrative Action to the Uruguay Round Agreements Act, H.R. Doc. No. 103-316 (1994) ("Statement of Administrative Action"), pp. 1025 and 1027).
115Ibid., pp. 22640-22641. Canada did not claim before the Panel that the USDOC was precluded from changing methodology. Consequently, this issue is not before us.
WT/DS264/AB/RW
Page 28
transaction-to-transaction comparison methodology.116 At the oral hearing, the United States acknowledged that the Section 129 Determination was the first time that the transaction-to-transaction comparison methodology had been used by the USDOC.
In what follows, we summarize briefly the findings of the Panel and the arguments of the participants and the third participants before turning to the issue before us.
Findings of the Panel
Canada argued before the Panel that the use of zeroing in the context of the transaction-to- transaction methodology is inconsistent with the first sentence of Article 2.4.2 of the Anti-Dumping Agreement. The Panel rejected Canada's claim. According to the Panel, "neither the ordinary meaning of the first sentence of Article 2.4.2 as a whole, nor the ordinary meaning of the phrase 'margins of dumping' in particular, require[s] that all transaction-specific comparisons under the [transaction-to-transaction] comparison methodology must be treated as 'intermediate values' and aggregated, without zeroing, in order to arrive at a single margin of dumping for the product as a whole."117 In addition, the Panel did not consider that the Appellate Body's ruling in the original dispute was applicable. First, the Panel noted that the Appellate Body's ruling in the original dispute was limited to the use of zeroing under the weighted average-to-weighted average comparison methodology.118 Secondly, the Panel observed that the Appellate Body's finding that zeroing is prohibited under the weighted average-to-weighted average comparison methodology was based on the phrase "all comparable export transactions" in the first part of the first sentence of Article 2.4.2 of the Anti-Dumping Agreement. That phrase, the Panel explained, appears in connection with the weighted average-to-weighted average comparison methodology and does not apply to the description of the transaction-to-transaction methodology in Article 2.4.2.119
116The USDOC, however, explained that the Statement of Administrative Action expresses a preference for the weighted average-to-weighted average comparison methodology, which was reflected in the United States anti-dumping regulations adopted after the Uruguay Round. These regulations provide that the USDOC "will only use the transaction-to-transaction means of comparison 'in unusual situations.'" (Section 129 Determination, p. 22638 (quoting United States Code of Federal Regulations, Title 19, section 351.414(c)) According to the USDOC, the regulations' preamble explains that the preference for the weighted average-to- weighted average comparison methodology "was directly tied to difficulties the agency had in the past with regard to the transaction-to-transaction methodology and concerns about the difficulty of guaranteeing that 'merchandise in both markets' would be 'identical or very similar' in order for such a comparison to work appropriately." (Section 129 Determination, pp. 22638-22639)
117Panel Report, para. 5.65.
118Ibid., para. 5.20.
119Ibid., para. 5.21.
WT/DS264/AB/RW
Page 29
Moreover, the Panel stated that "broader contextual considerations demonstrate that the application of the Appellate Body's interpretation beyond the confines of the [weighted average-to- weighted average] comparison methodology would lead to absurd results that could never have been intended by the Appellate Body, let alone the drafters of the [Anti-Dumping] Agreement."120 In particular, the Panel referred to the results that would obtain if, as Canada proposed, the interpretation of the term "margins of dumping" developed by the Appellate Body in US – Softwood Lumber V were applied to the other methodologies in Article 2.4.2. The Panel focused, in particular, on the weighted average-to-transaction comparison methodology set out in the second sentence of Article 2.4.2, explaining that it would be "deprive[d] ... of effect" if zeroing were to be prohibited under this methodology.121 This is because the results under the weighted average-to-transaction comparison methodology would be mathematically equivalent to those under the weighted average- to-weighted average methodology set out in the first sentence of that provision.
The Panel also found support for its interpretation in Articles 2.2 and 9.3 of the Anti- Dumping Agreement. According to the Panel, "[i]f the reference in Article 2.2 to 'margin of dumping' were understood to mean a single margin of dumping for the product as a whole", it would mean that an investigating authority would have to use constructed normal value for all models even if the conditions for using constructed normal value were applicable only to one model.122 Similarly, the Panel stated that, if this interpretation of "margins of dumping" were to apply to Article 9.3.2, "which expressly applies in respect of prospective duty assessment systems, [it] would mean that a refund becomes payable if an anti-dumping duty is paid in excess of the single margin of dumping for the product as a whole, calculated by aggregating the results of all intermediate comparisons, without zeroing."123 According to the Panel, "this makes no sense in the context of a prospective normal value duty assessment system, because ... the 'margin of dumping' at issue is a transaction-specific price difference calculated for a specific import transaction."124
120Panel Report, para. 5.65.
121Ibid., para. 5.52.
122Ibid., para. 5.61.
123Ibid., para. 5.57.
124Ibid.
WT/DS264/AB/RW
Page 30
For these reasons, the Panel concluded:
We therefore find that the [US]DOC was entitled not to offset the non-dumped transactions against the dumped transactions when calculating the margin of dumping for each respondent foreign producer or exporter. Accordingly, we reject Canada's claim that the [US]DOC's use of zeroing in the [transaction-to-transaction] comparison methodology at issue is inconsistent with Article 2.4.2 of the [Anti-Dumping] Agreement.125
Arguments of the Participants and the Third Participants
Canada requests the Appellate Body to reverse the Panel's finding and find, instead, that Article 2.4.2 of the Anti-Dumping Agreement prohibits the use of zeroing when calculating the margin of dumping on the basis of the transaction-to-transaction comparison methodology. According to Canada, the Appellate Body's reasoning in the original dispute concerning the weighted average-to-weighted average comparison methodology is equally applicable to the use of zeroing when calculating the margin of dumping on the basis of the transaction-to-transaction methodology.126 In addition, Canada argues that the Panel erred in relying on the so-called "broader contextual considerations", including the United States' mathematical equivalence argument.127
The United States asserts that the Appellate Body should dismiss Canada's appeal and maintains that the Appellate Body's reasoning in the original dispute in relation to the weighted average-to-weighted average comparison methodology does not apply to the transaction-to- transaction methodology set out in the second part of the first sentence of Article 2.4.2.128 The United States requests the Appellate Body to uphold the Panel's finding that zeroing is allowed under the transaction-to-transaction comparison methodology set out in Article 2.4.2 of the Anti-Dumping Agreement because it is a permissible interpretation of that provision in accordance with Article 17.6(ii) of the Anti-Dumping Agreement.129
China, the European Communities, India, Japan, and Thailand support Canada's appeal and consider that the Appellate Body should find that the use of zeroing is not permissible under the transaction-to-transaction comparison methodology set out in Article 2.4.2 of the Anti-Dumping
125Panel Report, para. 5.66.
126Canada's appellant's submission, para. 29.
127Ibid., paras. 51-58.
128United States' appellee's submission, paras. 3 and 67.
129Ibid., paras. 30 and 40.
WT/DS264/AB/RW
Page 31
Agreement.130 New Zealand disagrees with Canada's appeal and argues that the transaction-to- transaction comparison methodology in Article 2.4.2 does not require that all transaction-specific comparisons be treated as intermediate values and aggregated, without zeroing.131 The basis for New Zealand's position is somewhat different than that of the United States. New Zealand emphasizes that the transaction-to-transaction comparison methodology does not necessarily inflate the margins of dumping if "there is symmetry between the goods determined to be dumped and the volume of dumped goods" considered for purposes of the injury and causation analyses.132
Is Zeroing Allowed under the Transaction-to-Transaction Comparison Methodology Set Out in Article 2.4.2 of the Anti-Dumping Agreement?
The Use of "Zeroing" in the Section 129 Determination
Before turning to Article 2.4.2, we consider it useful to identify what is understood by "zeroing" in terms of the measure that is the subject of this dispute. As we explained earlier, the USDOC calculated the margins of dumping in the original investigation by using the weighted average-to-weighted average comparison methodology set out in the first sentence of Article 2.4.2 of the Anti-Dumping Agreement. In the Section 129 Determination, the USDOC used the other comparison methodology provided in the first sentence of Article 2.4.2, that is, it compared normal value and export prices on a transaction-to-transaction basis. The Panel described the USDOC's approach in the Section 129 Determination in the following terms:
130China's statement at the oral hearing; European Communities' third participant's submission, para. 77; India's statement at the oral hearing; Japan's third participant's submission, para. 86; Thailand's third participant's submission, para. 52.
131New Zealand's third participant's submission, para. 4.01.
132Ibid., paras. 4.02-4.03.
WT/DS264/AB/RW
Page 32
In its Section 129 Determination, the [US]DOC calculated a single margin of dumping for softwood lumber for each respondent foreign producer or exporter. It calculated that margin of dumping by determining the total amount of dumping on the basis of individual comparisons of export price and normal value for each export transaction, and then expressing that total amount as a proportion of the total value of all export sales, including those sales for which export price exceeded normal value. In order to establish the amount of dumping, the [US]DOC summed up the amounts by which, on individual transactions, export price was less than the normal value. The [US]DOC did not include in that summing up the amounts by which, in individual transactions, export price exceeded the normal value. In other words, the [US]DOC did not offset the amounts attributable to non-dumped transactions against the amounts attributable to dumped transactions.133 (emphasis added)
The Panel went on to describe the issue before it as follows:
The issue presented by Canada's Article 2.4.2 claim is whether it was permissible for the [US]DOC to not make such offsets when calculating the margin of dumping for each producer/exporter. In other words, we must decide whether the [US]DOC was permitted to sum only the amounts derived from [transaction-to-transaction] comparisons showing dumping, or whether it was required to also include the amounts derived from comparisons involving non- dumped transactions in that aggregation.134 (footnote omitted; emphasis added)
At the oral hearing, the United States did not contest that it aggregated the results of the transaction-to-transaction comparisons. However, it explained that the aggregation was not performed by the USDOC for purposes of establishing the margin of dumping under Article 2.4.2, but, rather, to determine whether the margin of dumping exceeded the de minimis level for purposes of Article 5.8. We find it difficult to reconcile this explanation with the USDOC's approach in the Section 129 Determination. The Section 129 Determination plainly shows that the USDOC aggregated the results of the individual transaction-to-transaction comparisons. Indeed, according to the Section 129 Determination, the USDOC determined an overall margin of dumping for each exporter and an overall "all others rate", both of which it labelled "weighted-average margin".135 The use of aggregation is confirmed in the description provided by the United States in its appellee's submission of how the USDOC calculated the margins of dumping in the Section 129 Determination:
133Panel Report, para. 5.16.
134Ibid.
135See Section 129 Determination, p. 22645.
WT/DS264/AB/RW
Page 33
[The USDOC] compared each export transaction to the most appropriate normal value transaction, based on the [USDOC]'s matching criteria, to determine whether it was sold at less than normal value. For comparisons for which the U.S. sale was made at less than normal value, the results were aggregated and divided by the total of the respondent company's U.S. sales to determine whether the dumping margin for that respondent was above the de minimis level.136 (emphasis added)
Therefore, the Section 129 Determination involved a two-step process: (i) a comparison of normal value and export prices on a transaction-specific basis; and (ii) the aggregation of the results of those transaction-specific comparisons in which export price was less than normal value "when calculating the margin of dumping for each respondent foreign producer or exporter."137 The results of the transaction-specific comparisons in which export prices exceeded the normal-value price were disregarded, as they were not included in the aggregation.138 Zeroing thus occurred during the second step in the process when the transaction-specific comparisons were aggregated. Accordingly, the issue before us in this appeal is whether it is consistent with Article 2.4.2 of the Anti-Dumping Agreement for an investigating authority to disregard the results of those transaction-specific comparisons in which export prices exceeded normal value, when aggregating the results of those comparisons for purposes of establishing the margins of dumping for each respondent foreign producer or exporter, under the transaction-to-transaction comparison methodology.139
2. Article 2.4.2 of the Anti-Dumping Agreement
Having identified the issue before us, we turn to Article 2.4.2 of the Anti-Dumping Agreement. Our examination will begin with the first sentence of that provision, which is directly at issue in this appeal. We will then address the second sentence of Article 2.4.2. Although the comparison methodology provided in the second sentence was not applied in the Section 129 Determination, both the United States and the Panel drew guidance from it for their interpretation of the first sentence. Finally, we will address the context provided by other provisions of the Anti-
136United States' appellee's submission, para. 11. The manner in which the USDOC proceeded in the Section 129 Determination is similar to the description of the transaction-to-transaction comparison methodology set out in A Handbook on Anti-Dumping Investigations, written by the WTO Secretariat. According to the Handbook, when there is more than one export transaction subject to the investigation, multiple comparison results under the transaction-to-transaction comparison methodology are aggregated in order to arrive at "one margin of dumping for the subject product". (J. Czako, J. Human and J. Miranda, A Handbook on Anti-Dumping Investigations (World Trade Organization, 2003), pp. 127-130)
137Panel Report, para. 5.16.
138"Aggregation" in this Report refers to the calculation of the absolute amount of dumping. This amount is placed in the numerator when calculating the margins of dumping as a percentage of the total value of export transactions. It is undisputed that the USDOC included the value of all export transactions in the denominator. The issue before us relates to the inclusion of all comparison results in the numerator.
139Panel Report, para. 5.66.
WT/DS264/AB/RW
Page 34
Dumping Agreement and the GATT 1994, the object and purpose, and certain historical materials referred to by the United States and the Panel.
Article 2.4.2 provides :
Subject to the provisions governing fair comparison in paragraph 4, the existence of margins of dumping during the investigation phase shall normally be established on the basis of a comparison of a weighted average normal value with a weighted average of prices of all comparable export transactions or by a comparison of normal value and export prices on a transaction-to-transaction basis. A normal value established on a weighted average basis may be compared to prices of individual export transactions if the authorities find a pattern of export prices which differ significantly among different purchasers, regions or time periods, and if an explanation is provided as to why such differences cannot be taken into account appropriately by the use of a weighted average-to-weighted average or transaction-to-transaction comparison. (emphasis added)
This provision establishes three methodologies that investigating authorities may use to calculate the "margins of dumping during the investigation phase". The first sentence sets out two comparison methodologies (weighted average-to-weighted average and transaction-to-transaction) involving symmetrical comparisons of normal value and export prices. These two methodologies "shall normally" be used by investigating authorities to establish margins of dumping. The second sentence of Article 2.4.2 sets out a third methodology (weighted average-to-transaction), which involves an asymmetrical comparison and may be used only in exceptional circumstances. It may be used only if the following two conditions are met: (i) "the authorities find a pattern of export prices which differ significantly among different purchasers, regions or time periods"; and (ii) "an explanation is provided as to why such differences cannot be taken into account appropriately by the use of a weighted average-to-weighted average or transaction-to-transaction comparison."
The First Sentence of Article 2.4.2
Turning to the transaction-to-transaction methodology, Article 2.4.2 provides that "margins of dumping" may be established "by a comparison of normal value and export prices on a transaction-to- transaction basis". The reference to "export prices" in the plural suggests that the comparison will generally involve multiple transactions, as was the case in the anti-dumping investigation before us. At the same time, the reference to "a comparison" in the singular suggests an overall calculation exercise involving aggregation of these multiple transactions. The transaction-specific results are mere steps in the comparison process. This tallies with the term "basis" at the end of the sentence, which suggests that these individual transaction comparisons are not the final results of the
WT/DS264/AB/RW
Page 35
calculation, but, rather, are inputs for the overall calculation exercise.140 Thus, the text of Article 2.4.2 implies that the calculation of a margin of dumping using the transaction-to-transaction methodology is a multi-step exercise in which the results of transaction-specific comparisons are inputs that are aggregated in order to establish the margin of dumping of the product under investigation for each exporter or producer. Contrary to the United States' submission141, the results of the transaction- specific comparisons are not, in themselves, "margins of dumping".
Furthermore, the reference to "export prices" in the plural, without further qualification, suggests that all of the results of the transaction-specific comparisons should be included in the aggregation for purposes of calculating the margins of dumping.142 In addition, the "export prices" and "normal value" to which Article 2.4.2 refers are real values, unless conditions allowing an investigating authority to use other values are met.143 Thus, in our view, zeroing in the transaction-to- transaction methodology does not conform to the requirement of Article 2.4.2 in that it results in the real values of certain export transactions being altered or disregarded.
This interpretation is consistent with previous Appellate Body statements concerning the other methodology set out in the first sentence of Article 2.4.2, namely, the weighted average-to- weighted average comparison methodology. We recall that the Appellate Body has held, in connection with the weighted average-to-weighted average comparison methodology, that, "[i]f an investigating authority has chosen to undertake multiple comparisons, the investigating authority necessarily has to take into account the results of all those comparisons in order to establish margins of dumping for the product as a whole under Article 2.4.2."144 Both methodologies set out in the first sentence of Article 2.4.2 fulfil the same function, namely, establishing "the existence of margins of dumping". As Canada pointed out before the Panel, the grammatical construction of this sentence strongly confirms "that the term 'margins of dumping' could not have different meanings for each of the two calculation methodologies to which it applies because that term appears unmodified in the
140The European Communities emphasizes this point in its third participant's submission, arguing that "[t]he word 'basis' indicates that there is one thing (one or more transaction-to-transaction juxtapositions) that is an underlying support for something else (a margin of dumping)." It adds that, although "in a first step of the calculation, there may be one or more transaction-to-transaction juxtapositions. ... [U]ltimately, in a subsequent step of the calculation, ... there is 'a comparison' (that is, a single comparison) of 'normal value' (also in the singular) and export prices, which ... requires an aggregation of all intermediate results established in the first step of the calculation." (European Communities' third participant's submission, para. 12) (footnote omitted))
141United States' appellee's submission, para. 52.
142The United States has not asserted that an investigating authority could limit its examination to some export transactions under the transaction-to-transaction methodology, rather than examining all of them.
143Article 2.2 allows investigating authorities, under certain conditions, to use constructed normal value. Article 2.3 permits the use of constructed export prices under certain conditions.
144Appellate Body Report, US – Softwood Lumber V, para. 98. (original emphasis)
WT/DS264/AB/RW
Page 36
sentence in question."145 The disjunctive "or", which is placed between the weighted average-to- weighted average and transaction-to-transaction comparison methodologies, while denoting the existence of alternative means, does not sever the logical relationship between the term "margins of dumping" and the transaction-to-transaction methodology, which is provided precisely to establish those margins. This indicates that the term "margins of dumping" has the same meaning regardless of which of the two methodologies in the first sentence of Article 2.4.2 is used to establish them. In other words, it is a unitary concept and the two methodologies provided in the first sentence of Article 2.4.2 are alternative means to capture it.
The United States and the Panel consider that the term "margins of dumping" must have a different meaning when applied to the transaction-to-transaction comparison methodology because of the absence of the phrase "all comparable export transactions", which appears in only the first part of the first sentence of Article 2.4.2, dealing with the weighted average-to-weighted average comparison methodology.146 In the Panel's view, "[t]his difference in language reflects a fundamental distinction between the nature of the [weighted average-to-weighted average] and [transaction-to-transaction] comparison methodologies."147
We do not agree with the conclusions that the United States and the Panel draw from the phrase "all comparable export transactions". The Appellate Body has recognized that Article 2.4.2 allows investigating authorities to use "multiple averaging" under the weighted average-to-weighted average comparison methodology.148 In the case of that methodology, transactions may be divided into groups, for instance, according to model or product type. Because of this possibility, the phrase "all comparable export transactions" implies that two requirements must be met when investigating authorities make the comparison by grouping transactions and averaging them. First, they must include in each group only those export transactions that are "comparable". Secondly, they must include "all" comparable export transactions corresponding to that group, and none of these export transactions may be left out arbitrarily. Such a scenario does not arise in the same way when comparisons are made under the transaction-to-transaction comparison methodology. As transactions are not divided into groups under the transaction-to-transaction comparison methodology149, the
145Canada's appellant's submission, para. 39 (referring to Panel Report, footnote 28 to para. 5.20).
146United States' appellee's submission, para. 55 (referring to Panel Report, para. 5.21).
147Panel Report, para. 5.30.
148Appellate Body Report, US – Softwood Lumber V, para. 81. The Appellate Body added that "the results of the multiple comparisons at the sub-group level are, however, not 'margins of dumping' within the meaning of Article 2.4.2." (para. 97)
149However, product types could be used as a criterion in selecting which domestic transaction to match with a particular export transaction when carrying out a transaction-to-transaction comparison, but the initial comparison is not made on a sub-group basis.
WT/DS264/AB/RW
Page 37
phrase "all comparable export transactions" is not pertinent to that methodology and, consequently, no inference may be drawn from the fact that this phrase does not appear in relation to the transaction-to- transaction methodology. Accordingly, we disagree with the United States' and the Panel's view that the phrase "all comparable export transactions" would be deprived of effect and meaning if zeroing were prohibited under the transaction-to-transaction comparison methodology.
Furthermore, we note that the United States made a similar argument in US – Zeroing (EC) to support its contention that the term "margins of dumping" had a different meaning in Article 9.3 of the Anti-Dumping Agreement.150 The United States emphasized that the phrase "all comparable export transactions", which appears in Article 2.4.2 in connection with the weighted average-to- weighted average comparison methodology, does not appear in Article 9.3. This argument was rejected by the Appellate Body, which found that the USDOC's use of zeroing in the underlying assessment reviews was inconsistent with Article 9.3.151 The Appellate Body underscored that its previous finding concerning the inconsistency of zeroing under the weighted average-to-weighted average comparison methodology "was based not only on Article 2.4.2, first sentence, but also on the context found in Article 2.1 of the Anti-Dumping Agreement."152
The use of zeroing under the transaction-to-transaction comparison methodology also raises an issue concerning the relationship between the two methodologies set out in the first sentence of Article 2.4.2 of the Anti-Dumping Agreement. If zeroing were allowed under the transaction-to- transaction comparison methodology, while being impermissible under the weighted average-to- weighted average methodology, having recourse to one or the other methodology provided in the first sentence of Article 2.4.2 could produce results that are systematically different. Canada points out that "[t]he Panel's conclusion that zeroing is prohibited under the [weighted average-to-weighted average] methodology, but permitted under the [transaction-to-transaction] methodology, would mean that the choice of a calculation methodology could be determinative as to whether dumping exists and could create substantial differences in the size of 'margins of dumping'".153 We share Canada's concerns. The first sentence of Article 2.4.2 sets out the two methodologies that "shall normally" be used by investigating authorities to establish "margins of dumping". Although the transaction-to- transaction and weighted average-to-weighted average comparison methodologies are distinct, they fulfil the same function. They are also equivalent in the sense that Article 2.4.2 does not establish a hierarchy between the two. An investigating authority may choose between the two depending on
150Appellate Body Report, US – Zeroing (EC), para. 34.
151Ibid., para. 135.
152Ibid., para. 126.
153Canada's appellant's submission, para. 62.
WT/DS264/AB/RW
Page 38
which is most suitable for the particular investigation. Given that the two methodologies are alternative means for establishing "margins of dumping" and that there is no hierarchy between them, it would be illogical to interpret the transaction-to-transaction comparison methodology in a manner that would lead to results that are systematically different from those obtained under the weighted average-to-weighted average methodology.
In sum, the results of the transaction-specific comparisons cannot be considered "margins of dumping" within the meaning of Article 2.4.2. The "margins of dumping" established under the transaction-to-transaction comparison methodology provided in Article 2.4.2 result from the aggregation of the transaction-specific comparisons. Article 2.4.2 does not permit an investigating authority, when aggregating the results of transaction-specific comparisons, to disregard transactions in which export price exceeds normal value.
(b) The Second Sentence of Article 2.4.2
We turn to the second sentence of Article 2.4.2, on which the Panel relied as context.154 We recall that the comparison methodology provided in the second sentence of Article 2.4.2 was not the one applied in the Section 129 Determination.155 The second sentence of Article 2.4.2 provides:
A normal value established on a weighted average basis may be compared to prices of individual export transactions if the authorities find a pattern of export prices which differ significantly among different purchasers, regions or time periods, and if an explanation is provided as to why such differences cannot be taken into account appropriately by the use of a weighted average-to-weighted average or transaction-to-transaction comparison.
The Panel placed much emphasis on the results that would obtain if the interpretation of the term "margins of dumping", developed by the Appellate Body in previous disputes dealing with the weighted average-to-weighted average comparison methodology, were extended to the other methodologies provided in Article 2.4.2. This reasoning assumes that a finding that zeroing is prohibited under the transaction-to-transaction comparison methodology would imply a prohibition of zeroing under the weighted average-to-transaction methodology, and that this would undermine the effective interpretation of the entire provision. The Panel was persuaded by the United States' submission that, if zeroing were prohibited also under the weighted average-to-transaction comparison methodology, that methodology would yield results that would be mathematically equivalent to those obtained by applying the weighted average-to-weighted average methodology,
154Panel Report, para. 5.31.
155See supra, para. 82.
WT/DS264/AB/RW
Page 39
rendering the second sentence of Article 2.4.2 inutile.156 Such a result, according to the Panel, undermines Canada's argument that the Appellate Body's interpretation of the term "margins of dumping" applies not only to the weighted average-to-weighted average comparison methodology, but also to the other methodologies provided in Article 2.4.2. The United States provided a hypothetical example to support its argument, which it has also submitted on appeal.157 The Panel "accept[ed]" the United States' mathematical equivalence argument, noting that "Canada and certain third parties ha[d] provided convoluted explanations of how the [weighted average-to-transaction] comparison methodology might be applied, without zeroing, so as to give results that are mathematically different from the results of the [weighted average-to-weighted average] comparison methodology."158
We disagree with the Panel's analysis of the "mathematical equivalence" argument for several reasons. First, the United States acknowledges that it has never applied the methodology provided in the second sentence of Article 2.4.2159, nor has it provided examples of how other WTO Members have applied this methodology.160 Thus, the United States' argument on "mathematical equivalence" rests on a non-tested hypothesis.161 Secondly, we note that the methodology in the second sentence of Article 2.4.2 is an exception. Article 2.4.2 clearly provides that investigating authorities "shall normally" use one of the two methodologies set out in the first sentence of that provision. Neither the participants, nor the third participants, disagree with this description of the relationship between the two sentences of Article 2.4.2. Being an exception, the comparison methodology in the second sentence of Article 2.4.2 (weighted average-to-transaction) alone cannot determine the interpretation of the two methodologies provided in the first sentence, that is, transaction-to-transaction and weighted average-to-weighted average.
Thirdly, the United States' "mathematical equivalence" argument assumes that zeroing is prohibited under the methodology set out in the second sentence of Article 2.4.2. The permissibility of zeroing under the weighted average-to-transaction comparison methodology provided in the second sentence of Article 2.4.2 is not before us in this appeal, nor have we examined it in previous cases. We also note that there is considerable uncertainty regarding how precisely the third methodology should be applied. For example, Canada and Japan suggested that the weighted average-to- transaction methodology could be applied only to the pattern of exports transactions that have prices
156Panel Report, paras. 5.33 and 5.52.
157United States' appellee's submission, para. 78 (Table 1).
158Panel Report, para. 5.52.
159United States' response to questioning at the oral hearing.
160Neither has Canada, nor the third participants.
161Canada's appellant's submission, para. 54.
WT/DS264/AB/RW
Page 40
that differ significantly among different purchasers, regions, or time periods.162 Before the Panel, the United States indicated that the USDOC would apply the weighted average-to-transaction comparison methodology to export transactions falling within the "pricing pattern" and would examine the other export transactions using the weighted average-to-weighted average methodology. The results obtained under each methodology would be combined to establish the margins of dumping for the respondent exporters or producers.163 At the oral hearing, however, the United States failed to explain how precisely the results of the two comparison methodologies would be combined, acknowledging that it had never applied the weighted average-to-transaction methodology. These uncertainties, which we are not called upon to resolve, undermine the Panel's reasoning based on the "mathematical equivalence" argument.
Fourthly, Canada and several third participants argued before the Panel that, even assuming that zeroing were prohibited also under the weighted average-to-transaction comparison methodology, mathematical equivalence would be limited to a specific set of circumstances. Canada and these third participants offered their own hypothetical scenarios showing that the weighted average-to-transaction comparison methodology would not yield necessarily the same results as the weighted average-to- weighted average methodology, even if the prohibition to use zeroing were to extend to the former. Thailand also explains that the mathematical equivalence argument works only under very specific assumptions, one of them being that the weighted-average normal value used in both the weighted average-to-weighted average and weighted average-to-transaction comparison methodologies be the same.164 The Panel rejected these explanations because, in its view, they "failed to address the
[United States'] mathematical equivalence argument on a ceteris paribus basis."165 In our view, the
Panel's approach is misguided. One part of a provision setting forth a methodology is not rendered inutile simply because, in a specific set of circumstances, its application would produce results that are equivalent to those obtained from the application of a comparison methodology set out in another part of that provision. In other words, the fact that, under the specific assumptions of the hypothetical scenario provided by the United States, the weighted average-to-transaction comparison methodology could produce results that are equivalent to those obtained from the application of the weighted
162Canada's closing statement before the Panel, Exhibit CDA-8 submitted by Canada to the Panel; Japan's response to Question 44 posed by the Panel, Panel Report, pp. E-70-E-73.
163United States' response to Question 17 posed by the Panel, Panel Report, pp. E-24-E-25, paras. 25-30. Japan points out that the United States anti-dumping regulations provide that the USDOC "normally will limit the application of the average-to-transaction method to those sales that constitute targeted dumping". (United States Code of Federal Regulations, Title 19, section 351.414(f)(2) (quoted in Japan's third participant's submission, para. 58 (emphasis omitted))
164Thailand's third participant's submission, para. 33. Canada and the third participants do not contest that mathematical equivalence does result under the specific assumptions used in the hypothetical provided by the United States.
165Panel Report, para. 5.52.
WT/DS264/AB/RW
Page 41
average-to-weighted average methodology is insufficient to conclude that the second sentence of Article 2.4.2 is thereby rendered ineffective.166 It has not been proven that in all cases, or at least in most of them, the two methodologies would produce the same results. Even if that were the case, it would not be sufficient to compel a finding that zeroing is permissible under the transaction-to- transaction comparison methodology, because this methodology is not involved in the "mathematical equivalence" argument.
In sum, we find the concerns of the Panel and the United States over the third comparison methodology (weighted average-to-transaction) being rendered inutile by a prohibition of zeroing under the transaction-to-transaction methodology to be overstated. It could be argued, on the contrary, that the use of zeroing under the two comparison methodologies set out in the first sentence of Article 2.4.2 would enable investigating authorities to capture pricing patterns constituting "targeted dumping", thus rendering the third methodology inutile.
3. Context
We now turn to the other provisions of the Anti-Dumping Agreement and the GATT 1994 that have been raised as relevant context in order to complete our interpretation of Article 2.4.2. Before proceeding, we observe that Article 2.4 of the Anti-Dumping Agreement, which is the subject of a separate claim by Canada, is also part of Article 2.4.2, in that Article 2.4.2 is expressly made "subject to the provisions governing fair comparison in paragraph 4 [of Article 2]", as well as possibly relevant as context. We address Article 2.4 when we examine Canada's claim under that provision in Section V of this Report.
Article 2.2 of the Anti-Dumping Agreement
The United States and the Panel find support for their interpretation in Article 2.2 of the Anti- Dumping Agreement, which provides:
166It could be argued that in targeted dumping the universe of export transactions constituting the "pricing pattern" would be smaller than the total number of export transactions. Thus, the universe of export transactions to which the weighted average-to-transaction comparison methodology applies would be different from the universe of transactions examined under the weighted average-to-weighted average methodology. In these circumstances, the two methodologies would not yield equivalent results, except by coincidence.
WT/DS264/AB/RW
Page 42
When there are no sales of the like product in the ordinary course of trade in the domestic market of the exporting country or when, because of the particular market situation or the low volume of the sales in the domestic market of the exporting country, such sales do not permit a proper comparison, the margin of dumping shall be determined by comparison with a comparable price of the like product when exported to an appropriate third country, provided that this price is representative, or with the cost of production in the country of origin plus a reasonable amount for administrative, selling and general costs and for profits. (footnote omitted)
According to the Panel, "[i]f the reference in Article 2.2 to 'margin of dumping' were understood to mean a single margin of dumping for the product as a whole", it would mean that "once the conditions for use of a constructed normal value were triggered, a constructed normal value would necessarily be used in all aspects of the determination of the margin of dumping for the product as a whole."167
The United States asserts that such a result would be contrary to the practice of WTO Members, which it explains as follows: "[I]f the home market sales of a particular model were not in the ordinary course of trade, the importing Member might resort to using a constructed normal value for that particular model; however, normal value for other models might still be based on home market sales."168 The United States adds that, if the term "margins of dumping" were to refer to the product as a whole throughout the Anti-Dumping Agreement, as Canada argues, then investigating authorities would be required to use constructed normal value "even if the condition precedent for using Article 2.2 relates only to 25 of the 100 comparisons." This, according to the United States, would be "inconsistent with the principle that constructed normal value is to be used only in limited circumstances."169
As Canada170 and several third participants171 point out, there is nothing in the text of Article 2.2 that prohibits an investigating authority from dividing the product under investigation into product types or models.172 Under the weighted average-to-weighted average comparison methodology, the prohibition of zeroing in Article 2.4.2 is triggered, not at the stage of determining whether to use a constructed normal value for a specific model or type, but, rather, when the results of
167Panel Report, para. 5.61.
168United States' appellee's submission, para. 74. 169Ibid., para. 75 (referring to Panel Report, para. 5.62). 170Canada's statement at the oral hearing.
171See European Communities' third participant's submission, para. 74; Japan's third participant's
submission, paras. 81-85; and Thailand's third participant's submission, para. 51.
172In the context of Article 2.4.2, the Appellate Body has said that an investigating authority may divide the product under investigation into product types or models for purposes of calculating the weighted-average normal value and weighted-average export price. (Appellate Body Report, US – Softwood Lumber V, para. 80)
WT/DS264/AB/RW
Page 43
the comparisons for each model or type are aggregated for purposes of establishing the margin of dumping for the product under investigation. Therefore, a prohibition of zeroing under the transaction-to-transaction comparison methodology provided in Article 2.4.2 does not limit the ability of an investigating authority, when the conditions are met, to use constructed normal value for one particular model or type, but not for others. We fail to see, therefore, the relevance of Article 2.2 for the interpretation of Article 2.4.2 as regards the permissibility of zeroing under the transaction-to- transaction comparison methodology.
Articles 5.8, 6.10, and 9.3 of the Anti-Dumping Agreement
Our interpretation that zeroing is not permissible when calculating margins of dumping by comparing normal value and export prices on a transaction-to-transaction basis is consistent with other provisions of the Anti-Dumping Agreement as well. For instance, Article 5.8 requires that an anti-dumping investigation be terminated if the investigating authority determines that the margin of dumping is de minimis, which is then defined as less than two per cent expressed as a percentage of the export price. As the United States recognizes173, a determination under Article 5.8 requires aggregation.174
Article 6.10 of the Anti-Dumping Agreement states that investigating authorities "shall as a rule, determine an individual margin of dumping for each known exporter or producer concerned of the product under investigation." The Panel acknowledged that this provision "arguably entails a need to aggregate the results of the comparisons made in respect of different transactions in order to establish 'an individual margin of dumping' for a particular exporter or producer."175 Nevertheless, the Panel reasoned that "the mere fact that Article 6.10 uses the term 'product under investigation' is insufficient to conclude that this provision dictates the use of a particular methodology for calculating an overall margin of dumping whereby the numerator of that margin must include the sum total of all (positive and negative) differences between export prices and the normal value."176 The United States adds that "Article 6.10 simply provides that a Member must calculate a margin of dumping for each individual exporter or producer – as opposed to one margin for all exporters or producers [but] says nothing about whether the margin must be based on more than one transaction, and it does not prohibit the calculation of a margin of dumping on a transaction-specific basis."177
173United States' response to questioning at the oral hearing.
174Thailand's third participant's submission, para. 29.
175Panel Report, para. 5.25.
176Ibid.
177United States' appellee's submission, para. 51 (referring to Appellate Body Report,
Mexico – Anti-Dumping Measures on Rice, paras 208 and 217; and Panel Report, para. 5.25).
WT/DS264/AB/RW
Page 44
We agree with the United States and the Panel that Article 6.10 does not expressly require taking into account both negative and positive comparison results from transactions relating to a given exporter or producer. However, it does not express the reverse proposition either. Rather, the import of Article 6.10 is to reinforce the notion that the "margins of dumping" are the result of an aggregation, in this case, of transaction-specific comparisons.
We find further contextual guidance in Article 9.3. This provision states that "[t]he amount of the anti-dumping duty shall not exceed the margin of dumping as established under Article 2." Thus, Article 9.3 sets the margins of dumping determined under Article 2 as a ceiling for the total amount of anti-dumping duty that may be imposed on individual exporters or foreign producers.178 Like Articles 5.8 and 6.10, this provision suggests that the margin of dumping is the result of an overall aggregation and does not refer to the results of the transaction-specific comparisons. Indeed, we do not see how the margin of dumping could operate as a ceiling under Article 9.3 if the result of each transaction-by-transaction comparison is deemed to be a margin of dumping as the United States suggests.
In our view, interpreting the term "margins of dumping" in Article 2.4.2 as referring to the aggregation of the results of individual transaction-to-transaction comparisons fits well into the context of the other provisions of the Anti-Dumping Agreement, including Articles 5.8, 6.10, and 9.3. In this regard, we agree with New Zealand that "[t]he provisions of the Anti-Dumping Agreement work in an integrated fashion", and that an interpretation of Article 2.4.2 "cannot be considered in isolation from the rest of the Anti-Dumping Agreement."179 Articles 5.8, 6.10, and 9.3 do not support the United States' interpretation of Article 2.4.2 that the results of transaction-specific comparisons are "margins of dumping".
Prospective Normal-Value Systems – Article 9.3.2 of the Anti- Dumping Agreement
Turning to the Panel's concerns relating to Article 9.3, we note that the Anti-Dumping Agreement allows WTO Members to assess duties either prospectively or retrospectively. According to the Panel, interpreting the term "margin of dumping" in the manner suggested by Canada would lead to absurd results under the prospective duty assessment system because it would mean that "one importer could request a refund on the basis of a margin of dumping calculated by reference to non- dumped transactions made by other importers."180
178Appellate Body Report, US – Zeroing (EC), para. 130. 179New Zealand's third participant's submission, para. 3.35. 180Panel Report, para. 5.57.
WT/DS264/AB/RW
Page 45
Article 9.3.2 of the Anti-Dumping Agreement applies when investigating authorities assess duties prospectively. It provides:
When the amount of the anti-dumping duty is assessed on a prospective basis, provision shall be made for a prompt refund, upon request, of any duty paid in excess of the margin of dumping.
As several third participants have indicated181, the Panel confuses duty collection at the time of importation with the determination of the final margin of dumping and assessment of final duties in administrative reviews. The margin of dumping is established during the investigation phase. Once an anti-dumping order has been issued, Article 9.2 allows an investigating authority to collect an anti- dumping duty in "the appropriate amounts in each case" on all subsequent entries of that product. Under a prospective normal-value system, the anti-dumping duty collected at the time of importation is subject to review and importers have the right to request a refund when the duties paid exceed the actual margin of dumping, pursuant to Article 9.3.2 of the Anti-Dumping Agreement.182 Accordingly, the operation of prospective normal-value systems has no bearing on the permissibility of zeroing under the transaction-to-transaction comparison methodology in Article 2.4.2.
Article VI of the GATT 1994
The United States takes issue with Canada's reliance on Article VI of the GATT 1994 to support its argument that "margins of dumping" must be established for the "product as a whole". According to the United States, the Panel identified several instances in Article VI where the term "product" "is used in a manner that refers to a single import transaction, as opposed to 'the entire universe of exported product subject to an anti-dumping investigation.'"183 As Canada pointed out at
181See European Communities' third participant's submission, para. 69; and Thailand's third participant's submission, para. 44.
182In US – Zeroing (EC), the Appellate Body stated that:
... under the methodology currently applied by the USDOC to assess anti- dumping duties, the aggregation of the results of the multiple comparisons performed at an intermediate stage might result in a negative value, for a given importer, if zeroing is not allowed. Of course, this would not mean that the authorities would be required under the Anti-Dumping Agreement or Article VI of the GATT 1994 to compensate an importer for the amount of that negative value (that is, when export prices exceed normal value).
(Appellate Body Report, US – Zeroing (EC), footnote 234 to para. 131)
Although that case involved a retrospective system, the reasoning is equally applicable to prospective normal- value systems.
183United States' appellee's submission, para. 18 (quoting Panel Report, para. 5.23). The Panel referred to Article VI:2 ("may levy on any dumped product an anti-dumping duty"); Article VI:3 ("[n]o countervailing duty shall be levied on any product"; Article VI:6(a) ("[n]o Member shall levy any anti-dumping or countervailing duty on the importation of any product"); and Article VI:6(b) ("a Member to levy an anti- dumping or countervailing duty on the importation of any product").
WT/DS264/AB/RW
Page 46
the oral hearing, in its Report in US – Softwood Lumber V, the Appellate Body referred to Article VI:1 and 2 of the GATT 1994, together with Article 2.1 of the Anti-Dumping Agreement, to interpret the term "margins of dumping" in Article 2.4.2. The Appellate Body did not address the meaning of "product" in the other paragraphs of Article VI or in other provisions of the GATT 1994.
In addition, the United States argues that extending the phrase "product as a whole" to include all import transactions represents "a dramatic departure from the original sense in which the Appellate Body used the phrase."184 According to the United States, "the Appellate Body originally adopted the phrase to distinguish the margin of dumping for a 'product' from dumped amounts found to exist for individual sub-products based on multiple [weighted average-to-weighted average] comparisons."185 New Zealand makes a similar argument, stating that the phrase "product as a whole" was used by the Appellate Body in previous cases to refer to "the scope of the product subject to the investigation", while, in this case "there is no suggestion of 'model zeroing'".186 We note that in US – Softwood Lumber V, the Appellate Body referred generally to the use of zeroing in relation to the use of "multiple comparisons" when it stated that, "[i]f an investigating authority has chosen to undertake multiple comparisons, the investigating authority necessarily has to take into account the results of all those comparisons in order to establish margins of dumping for the product as a whole under Article 2.4.2."187 We fail to see, therefore, how concluding that zeroing is prohibited in respect of aggregation under the transaction-to-transaction comparison methodology would represent a "dramatic departure" from earlier decisions by the Appellate Body.
Article II of the GATT 1994
The United States also points to Article II of the GATT 1994 arguing that, if the references to the "product" in that provision were understood to mean "the product as a whole", "[i]t would be permissible to impose a tariff in excess of the bound rate of duty on particular entries so long as it was 'offset' by the tariff on other entries such that the tariff for the 'product as a whole' does not exceed the bound rate."188 This result, the United States adds, "would be extremely surprising to WTO Members".189 Article II of the GATT 1994 was not at issue in previous appeals relating to zeroing and there is no basis in the Appellate Body's analysis of certain provisions of the Anti-Dumping Agreement in those appeals for the United States to draw the implications for Article II and WTO
184United States' appellee's submission, para. 28.
185Ibid.
186New Zealand's third participant's submission, para. 3.29. (original underlining)
187Appellate Body Report, US – Softwood Lumber V, para. 98. (original emphasis; underlining added)
188United States' appellee's submission, para. 23.
189Ibid.
WT/DS264/AB/RW
Page 47
Members' tariff Schedules that it seeks to draw. We also note that WTO Members' Schedules base their product categorization on the Harmonized Commodity Description and Coding System (the "Harmonized System") and specify their tariff commitments by product subheadings. There is no indication in previous Appellate Body reports addressing the Anti-Dumping Agreement or in the Harmonized System to indicate that levying tariffs in excess of a bound rate on the importation of a product could be "offset" or justified by levying tariffs below the bound rate on another importation of that product. Nor is there any basis for an analogy between establishing the margin of dumping for the "product as a whole" under Article VI:2 of the GATT 1994 and the Anti-Dumping Agreement, on the one hand, and the application of a tariff on a product above the bound rate within the meaning of Article II of the GATT 1994, on the other hand.
Thus, our examination of the relevant context in the Anti-Dumping Agreement and the GATT 1994 does not support the United States' interpretation that the use of zeroing is permissible under the transaction-to-transaction comparison methodology in Article 2.4.2 of the Anti-Dumping Agreement.
4. "Symmetry" between the Determination of Dumping and the Injury and Causation Analyses
New Zealand emphasizes that the concern with zeroing in previous disputes focused on the lack of symmetry between the transactions included in the dumping determination and those considered in the injury determination.190 According to New Zealand, the use of zeroing in the context of the transaction-to-transaction comparison methodology does not raise concerns provided that the treatment of transactions in which the export price exceeds normal value is symmetrical in the dumping determination and in the injury, causation, and non-attribution analyses.191 The approach advocated by New Zealand is not before us in this appeal.192 The Section 129 Determination, which is the measure before us, concerns the determination of dumping exclusively and does not address injury or causation.193 Nor did the United States suggest that it followed the approach advocated by New
190New Zealand's third participant's submission, para. 3.33.
191It is unclear whether, under New Zealand's approach, account is taken of export prices above normal value in the causation or non-attribution analysis, or both. In addition, at the oral hearing, New Zealand seemed to suggest that these transactions would be relevant when applying the "lesser duty" rule pursuant to Article 9.1.
192We therefore express no views on whether such an approach is consistent with the obligations under the Anti-Dumping Agreement.
193Canada challenged the injury and causation determinations in separate disputes. (See Request for the Establishment of a Panel by Canada, United States – Investigation of the International Trade Commission in Softwood Lumber from Canada, WT/DS277/2, 4 April 2003; Request for the Establishment of a Panel by Canada, United States – Investigation of the International Trade Commission in Softwood Lumber from Canada, Recourse to Article 21.5 of the DSU by Canada, WT/DS277/8, 15 February 2005)
WT/DS264/AB/RW
Page 48
Zealand.194 Thus, New Zealand's argument about symmetry is irrelevant for purposes of determining whether the use of zeroing by the USDOC is consistent with Article 2.4.2.
5. Object and Purpose of the Anti-Dumping Agreement
We turn to examine what guidance is provided by the object and purpose of the Anti- Dumping Agreement for the interpretation of Article 2.4.2. The Anti-Dumping Agreement does not contain a preamble or an explicit indication of its object and purpose. Neither participant referred to the object and purpose in its written submission. At the oral hearing, Canada and certain third participants indicated that the object and purpose of the Anti-Dumping Agreement could be discerned from Article 1 of the Anti-Dumping Agreement. The United States and New Zealand, in contrast, said guidance could be found in Article VI of the GATT 1994. We do not consider it necessary for purposes of resolving the issue before us on appeal to engage in an in-depth analysis of the object and purpose of the Anti-Dumping Agreement.
6. Historical Background
Finally, we examine the "historical" arguments put forward by the United States, which were also examined by the Panel. The United States referred to a 1960 Second Report by the Group of Experts on Anti-Dumping and Countervailing Duties195 (the "Group of Experts Report") that, according to the Panel, "clearly envisaged the calculation of transaction-specific margins of dumping."196 This, in the Panel's view, "would suggest that the Group of Experts did not consider that there was anything in the definition of dumping set forth in Article VI of the GATT [1947] that would preclude the calculation of such transaction-specific margins", which "in turn would suggest that the GATT Contracting Parties would have disagreed with Canada's reliance on the same provision of the
194Before the Panel, the United States explained that "[i]n calculating an overall margin of dumping, the United States fairly reflects the total quantity of all imports of the product under consideration ... in the denominator of its calculations". It then explained that, if the transactions with an export price above normal value were to be taken into account in the injury determination, it would result in "'double-counting' of the[] non-dumped transactions". (United States' response to Question 19 posed by the Panel, Panel Report, pp. E-27- E-28, paras. 40-41.) At the oral hearing in this appeal, the United States qualified its response to the Panel, by stating that the Section 129 Determination was the first time that the USDOC had applied the transaction-to- transaction comparison methodology and that it was unclear how the United States International Trade Commission would treat the transactions with export prices above normal value in the injury and causation analyses.
195Group of Experts Report, supra, footnote 44.
196Panel Report, para. 5.64.
WT/DS264/AB/RW
Page 49
GATT 1994 to support its argument that 'margins of dumping' must always be calculated 'for the product as a whole' by aggregating all transaction-specific comparisons."197
On appeal, Canada asserts that "the Panel's dubious reliance on a single line in a GATT report as negotiating history should be afforded no weight as it is already clear that the ordinary meaning of Article 2.4.2 read in its context prohibits zeroing."198 The United States points to various historical materials including, in addition to the 1960 Group of Experts Report, two pre-WTO panel reports that dealt with the issue of zeroing199 and several proposals submitted during the Uruguay Round, which allegedly demonstrate that zeroing is not prohibited by the Anti-Dumping Agreement.200 The United States adds that, throughout the history of the GATT, it was recognized that zeroing was allowed under Article VI of the GATT 1947. This provision, the United States emphasizes, was not modified during the Uruguay Round.201
The same historical materials submitted in these Article 21.5 proceedings were also raised by the United States before the Appellate Body in the original dispute.202 The Appellate Body stated in response that "[t]he material to which the United States refer[red] does not ... resolve the issue of whether the negotiators of the Anti-Dumping Agreement intended to prohibit zeroing."203 The Appellate Body noted that, "[i]n any event", it had "concluded, based on the ordinary meaning of Article 2.4.2 read in its context, that zeroing is prohibited when establishing the existence of margins of dumping under the weighted-average-to-weighted-average methodology."204 In our view, the historical materials referred to by the Panel and the United States are of limited relevance. The Group of Experts Report dates back to 1960. Both pre-WTO panel reports examined the issue under the
197Panel Report, para. 5.64. We note that the Group of Experts Report states that, although the ideal method to determine dumping and injury was in respect of each single importation of the product concerned, this "was clearly impracticable, particularly as regards injury". (Group of Exports Report, para. 8 (reproduced in Panel Report, para. 5.63)) The fact that making a determination of dumping and injury for each importation was not considered practical could also be seen as implying that the Group of Experts did not, in fact, endorse transaction-specific determinations of dumping.
198Canada's appellant's submission, para. 59.
199GATT Panel Report, EC – Audio Cassettes, para. 360 (unadopted), circulated 28 April 1995; GATT Panel Report, EEC – Cotton Yarn, para. 502, adopted 30 October 1995.
200United States' appellee's submission, paras. 30-40.
201Ibid., paras. 37-39.
202Appellate Body Report, US – Softwood Lumber V, para. 107 and footnote 168 thereto. The United States argued in the original Appellate Body proceedings that recourse to the circumstances of the conclusion of the Anti-Dumping Agreement was appropriate as a supplementary means of interpretation under Article 32 of the Vienna Convention on the Law of Treaties (the "Vienna Convention"). (Done at Vienna, 23 May 1969, 1155 UNTS 331; 8 International Legal Materials 679) At the oral hearing in those proceedings, the United States acknowledged that the historical materials did not constitute "travaux préparatoires".
203Appellate Body Report, US – Softwood Lumber V, para. 108.
204Ibid.
WT/DS264/AB/RW
Page 50
provisions of the Tokyo Round Anti-Dumping Code, which did not contain a provision equivalent to Article 2.4.2 of the Anti-Dumping Agreement. The latter Agreement entered into force in 1995, as part of the Uruguay Round results, long after the 1960 Group of Experts Report and after the panels referred to by the United States and the Panel had been established.205 Furthermore, one of the two panel reports was not adopted. Finally, the negotiating proposals referred to by the United States are inconclusive and, in any event, reflected the positions of some, but not all, of the negotiating parties.206 In sum, the historical materials do not provide any additional guidance for the question whether zeroing under the transaction-to-transaction comparison methodology is consistent with Article 2.4.2 of the Anti-Dumping Agreement.
7. Conclusion on Canada's Claim under Article 2.4.2
On the basis of the above analysis, we conclude that zeroing is not permitted under the transaction-to-transaction methodology set out in the first sentence of that provision. The "margins of dumping" established under this methodology are the results of the aggregation of the transaction- specific comparisons of export prices and normal value. In aggregating these results, an investigating authority must consider the results of all of the comparisons and may not disregard the results of comparisons in which export prices are above normal value.
We have found that Article 2.4.2 does not admit an interpretation that would allow the use of zeroing under the transaction-to-transaction comparison methodology. Therefore, the contrary view is not a permissible interpretation of Article 2.4.2 within the meaning of Article 17.6(ii) of the Anti- Dumping Agreement.207
For these reasons, we reverse the Panel's finding, in paragraph 5.66 of the Panel Report, that "the [US]DOC was entitled not to offset the non-dumped transactions against the dumped transactions when calculating the margin of dumping for each respondent foreign producer or exporter." We reverse also the Panel's conclusion, in paragraph 6.1 of the Panel Report, that "the determination of the [US]DOC in the section 129 proceeding investigation is not inconsistent with ... Article[] 2.4.2 of
205The relevance of these panel reports is further diminished by the fact that the Tokyo Round Anti- Dumping Code was not part of the GATT 1947 and was terminated in 1996 without being incorporated into the WTO covered agreements.
206The United States refers to: Communication from Japan, Background Notes to Japan's Proposals on the Anti-Dumping Code, MTN.GNG/NG8/W/30, 20 June 1988; Communication from the Delegation of Singapore, Proposed Elements for a Framework for Negotiations Principles and Objectives for Anti-Dumping Rules, MTN.GNG/NG8/W/55, 13 October 1989; and Communication from the Delegation of Hong Kong, Principles and Purposes of Anti-dumping Provisions, MTN.GNG/NG8/W46, 3 July 1989, p. 7. (United States' appellee's submission, footnote 36 to para. 35 and footnote 39 to para. 38)
207The Appellate Body has explained that "a permissible interpretation is one which is found to be appropriate after application" of the customary rules of interpretation reflected in Articles 31 and 32 of the Vienna Convention. (Appellate Body Report, US – Hot-Rolled Steel, para. 60 (original emphasis))
WT/DS264/AB/RW
Page 51
the [Anti-Dumping] Agreement." We find, instead, that the use of zeroing by the USDOC in the Section 129 Determination is inconsistent with the United States' obligations under Article 2.4.2 of the Anti-Dumping Agreement.
Article 2.4 of the Anti-Dumping Agreement
Introduction
We turn next to Canada's claim under Article 2.4 of the Anti-Dumping Agreement. Before examining Article 2.4, we provide a brief summary of the Panel's findings and the arguments of the participants and the third participants.
Findings of the Panel
Canada argued before the Panel that the use of zeroing in the transaction-to-transaction comparison methodology is "by definition" inconsistent with the "fair comparison" requirement in Article 2.4.208 Canada submitted that zeroing does not take into account all comparisons, and that this introduces an "inherent bias" that inflates the margin of dumping. The Panel considered that "[t]he principle of effective treaty interpretation implies that the 'fair comparison' obligation in Article 2.4 must not be interpreted in a manner so as to trump the more specific provisions of Article 2.4.2."209 Thus, because it had concluded that the use of zeroing under the transaction-to-transaction comparison methodology was consistent with Article 2.4.2, the Panel found that the use of zeroing under such a comparison methodology was not "unfair" under Article 2.4.210 The Panel concluded that, "[s]ince zeroing therefore cannot be prohibited as 'by definition' unfair in the context of Article 2.4.2, Article 2.4 cannot provide for the unqualified, 'by definition' prohibition suggested by Canada."211 On this basis, the Panel rejected Canada's claim that the United States acted inconsistently with the "fair comparison" requirement in the first sentence of Article 2.4 of the Anti- Dumping Agreement.212
208Panel Report, para. 5.72. Canada had already asserted before the original panel that the USDOC's use of zeroing in aggregating the results of weighted average-to-weighted average comparisons in the original investigation was inconsistent with the "fair comparison" requirement under Article 2.4 of the Anti-Dumping Agreement. Having found that the use of zeroing was inconsistent with Article 2.4.2 of the Anti-Dumping Agreement, the original panel found it neither appropriate, nor necessary, to rule on Canada's Article 2.4 claim. (Original Panel Report, US – Softwood Lumber V, para. 7.226)
209Panel Report, para. 5.75. (footnote omitted)
210Ibid., para. 5.75.
211Ibid., para. 5.76.
212Ibid., para. 5.78.
WT/DS264/AB/RW
Page 52
Arguments of the Participants and the Third Participants
Canada submits that the Panel improperly interpreted Article 2.4 to permit zeroing under the transaction-to-transaction comparison methodology. Canada emphasizes the Appellate Body's repeated findings that zeroing inflates margins of dumping and creates an "inherent bias" in these comparisons.213 According to Canada, "[t]he USDOC manipulated the results of comparisons where the export price was higher than the home market price by disregarding the difference between these prices and replacing that difference with 'zero'."214 Canada argues that this manipulation is a violation of Article 2.4 because it is not a "fair comparison".215 Canada contends that the Panel wrongly relied on its conclusion that Article 2.4.2 permits zeroing under the transaction-to-transaction comparison methodology to find that zeroing was consistent with Article 2.4 of the Anti-Dumping Agreement. Because the Panel's interpretation of Article 2.4.2 is, according to Canada, incorrect, its reliance on the principle of effective interpretation also fails.
The United States responds that the Panel properly found the use of zeroing by the USDOC in the Section 129 Determination to be not inconsistent with the "fair comparison" requirement of Article 2.4 of the Anti-Dumping Agreement. The United States asserts that Canada's argument assumes its own conclusion by insisting that a margin of dumping established without "zeroing" is "fair".216 The United States agrees with the Panel's reliance on the principle of effective treaty interpretation, arguing that, because zeroing is permissible under the transaction-to-transaction comparison methodology, it cannot be deemed "unfair" simply because it results in a margin of dumping higher than that which would obtain without the use of zeroing.217 The United States further argues that, if, as Canada suggests, the size of any resulting margin of dumping were the basis for determining "fairness", investigating authorities would have to determine "at least two putative margins of dumping for each exporter" using the comparison methodologies provided under Article 2.4.2 and "ultimately rely on the methodology that generated the lowest margin."218 According to the United States, the text of the Anti-Dumping Agreement imposes no such results-
213Canada's appellant's submission, para. 67 (quoting Appellate Body Report, US – Corrosion-Resistant Steel Sunset Review, para. 135).
214Ibid., para. 69.
215Ibid.
216United States' appellee's submission, para. 97.
217Ibid., para. 99.
218Ibid., para. 100.
WT/DS264/AB/RW
Page 53
driven obligation.219 Finally, the United States submits that the Appellate Body reports on which Canada relies do not support its "fair comparison" argument.220
In their third participant's submissions, the European Communities, Japan, and Thailand contend that the Panel erred in its interpretation of Article 2.4, arguing that zeroing in transaction-to- transaction comparisons is inconsistent with the "fair comparison" requirement in Article 2.4.221 The European Communities submits that Article 2.4 establishes an overarching and independent obligation to make a "fair comparison" between export price and normal value. The European Communities considers that the negotiating history of Article 2.4 confirms such an interpretation, noting that under the Tokyo Round Anti-Dumping Code, "the equivalent or similar provisions to the first and second sentences of Article 2.4 of the Uruguay Round Anti-Dumping Agreement were contained in the same sentence" but the "fair comparison" requirement was placed on its own in a new first sentence of Article 2.4 in the Uruguay Round.222 Furthermore, the European Communities argues that, because of the "arbitrary and artificial reduction of the value of certain export transactions", zeroing does not accord the same treatment to both domestic and export sales.223
Japan submits that the "discernable standard" the Panel used to assess fairness was wrong because, contrary to the Panel's conclusion, zeroing is not permitted in a transaction-to-transaction comparison under the first sentence of Article 2.4.2. Regarding the meaning of "fair", Japan refers to the findings by the panel in EC – Tube or Pipe Fittings that an "investigating authority must act in an unbiased, even-handed manner and must not exercise its discretion in an arbitrary manner", arguing that such findings suggest a meaning rooted in the basic requirements of good faith and fundamental fairness.224 Japan also emphasizes the Appellate Body's findings in EC – Bed Linen and US – Corrosion-Resistant Steel Sunset Review that zeroing inflates margins of dumping and creates an "inherent bias".225 Thailand makes similar arguments in its submission, arguing that a "fair
219United States' appellee's submission, para. 100.
220Ibid., paras. 104-106.
221New Zealand does not present arguments under Article 2.4 in its third participant's submission. New Zealand nevertheless rejects, in its arguments responding to Canada's claims under Article 2.4.2, the notion that zeroing is unfair.
222European Communities' third participant's submission, para. 28.
223Ibid., para. 30.
224Japan's third participant's submission, para. 90 (quoting Panel Report, EC – Tube or Pipe Fittings, para. 7.178). (emphasis added by Japan)
225Ibid., para. 96 (referring to Appellate Body Report, US – Corrosion-Resistant Steel Sunset Review, para. 135; and Appellate Body Report, EC – Bed Linen, para. 55).
WT/DS264/AB/RW
Page 54
comparison", rather than being a "subjective" test as suggested by the Panel, requires objectivity, lack of bias, and even-handedness.226
Whether Zeroing is Consistent with the "Fair Comparison" Requirement in Article 2.4 of the Anti-Dumping Agreement
We begin our analysis with the text of Article 2.4 of the Anti-Dumping Agreement and with a brief review of how it has been interpreted in previous appeals. Next, we examine the basis of the Panel's finding on this issue. Finally, we conduct our own examination of Canada's claim, taking into account the conclusion that we reached in the previous section of this Report regarding Article 2.4.2.
1. Article 2.4 of the Anti-Dumping Agreement
Article 2.4 of the Anti-Dumping Agreement provides:
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, normally at the ex-factory level, and in respect of sales made at as nearly as possible the same time. Due allowance shall be made in each case, on its merits, for differences which affect price comparability, including differences in conditions and terms of sale, taxation, levels of trade, quantities, physical characteristics, and any other differences which are also demonstrated to affect price comparability. In the cases referred to in paragraph 3, allowances for costs, including duties and taxes, incurred between importation and resale, and for profits accruing, should also be made. If in these cases price comparability has been affected, the authorities shall establish the normal value at a level of trade equivalent to the level of trade of the constructed export price, or shall make due allowance as warranted under this paragraph. The authorities shall indicate to the parties in question what information is necessary to ensure a fair comparison and shall not impose an unreasonable burden of proof on those parties. (footnote omitted)
We recall that Article 2.4.2 begins with the phrase "[s]ubject to the provisions governing fair comparison in paragraph 4". Thus, the application of the comparison methodologies set out in Article 2.4.2 of the Anti-Dumping Agreement, including the transaction-to-transaction methodology applied in the investigation underlying this dispute, is expressly made subject to the "fair comparison" requirement set out in Article 2.4.
In EC – Bed Linen, the Appellate Body explained that "Article 2.4 sets forth a general obligation to make a 'fair comparison' between export price and normal value", adding that "[t]his is a general obligation that ... informs all of Article 2, but applies, in particular, to Article 2.4.2 which is
226Thailand's third participant's submission, paras. 54-56.
WT/DS264/AB/RW
Page 55
specifically made 'subject to the provisions governing fair comparison in [Article 2.4]'."227 Moreover, in US – Zeroing (EC), the Appellate Body agreed with that panel's conclusion that "the 'fair comparison' language in the first sentence of Article 2.4 creates an independent obligation, and, secondly, that the scope of this obligation is not exhausted by the general subject matter expressly addressed by paragraph 4 (that is to say, the price comparability)."228
2. The Basis of the Panel's Finding under Article 2.4 of the Anti-Dumping Agreement
The Panel's finding on Canada's claim under Article 2.4 is premised on its finding that the use of zeroing in transaction-to-transaction comparisons is not inconsistent with Article 2.4.2:
Since we have already concluded that [a transaction-to-transaction comparison] with zeroing (resulting in higher margins) is not inconsistent with Article 2.4.2, one cannot conclude that failure to use a comparison methodology that would have resulted in lower margins (i.e., [transaction-to-transaction] without zeroing) is "unfair".229
The Panel did not offer any other reasons that could independently support its finding under Article 2.4.
Because the Panel's conclusion under Article 2.4 was premised precisely on its finding that zeroing under the transaction-to-transaction comparison methodology is consistent with Article 2.4.2, which we have reversed, this conclusion can no longer stand.
Moreover, the Panel's reasoning essentially makes the "fair comparison" requirement in Article 2.4 dependent on Article 2.4.2. According to the Panel, "[t]he principle of effective treaty interpretation implies that the 'fair comparison' obligation in Article 2.4 must not be interpreted in a manner so as to trump the more specific provisions of Article 2.4.2".230 Apparently, the Panel considered Article 2.4.2 as lex specialis. This, however, is not a correct representation of the
227Appellate Body Report, EC – Bed Linen, para. 59.
228Appellate Body Report, US – Zeroing (EC), para. 146 (referring to Panel Report, US – Zeroing (EC), paras. 7.253-7.258). With regard to the remainder of Article 2.4, the Appellate Body found in US – Zeroing (EC) that "disregarding a result when the export price exceeds the normal value (zeroing) cannot be characterized as an allowance or an adjustment covered by the third sentence of Article 2.4, including its a contrario application." (para. 158) The Appellate Body therefore agreed with the panel that "conceptually, zeroing is not an adjustment or an allowance falling within the scope of Article 2.4, third to fifth sentences." (Ibid. )
229Panel Report, para. 5.75.
230Ibid. (footnote omitted)
WT/DS264/AB/RW
Page 56
relationship between the two provisions. Rather, the introductory clause to Article 2.4.2 expressly makes it "[s]ubject to the provisions governing fair comparison" in Article 2.4.
3. The Consistency of Zeroing under the Transaction-to-Transaction Comparison Methodology with Article 2.4 of the Anti-Dumping Agreement
We turn to examine whether the use of zeroing in the Section 129 Determination, in addition to being contrary to Article 2.4.2, is inconsistent with the "fair comparison" requirement in Article 2.4.
The term "fair" is generally understood to connote impartiality, even-handedness, or lack of bias.231 For the reasons stated below, we consider that the use of zeroing under the transaction-to- transaction comparison methodology is difficult to reconcile with the notions of impartiality, even- handedness, and lack of bias reflected in the "fair comparison" requirement in Article 2.4.
First, the use of zeroing under the transaction-to-transaction comparison methodology when aggregating the transaction-specific comparisons for purposes of calculating the "margins of dumping", distorts the prices of certain export transactions because export transactions made at prices above normal value are not considered at their real value. The prices of these export transactions are artificially reduced when zeroing is applied under the transaction-to-transaction comparison methodology.232 As the Appellate Body explained in the original dispute, "[z]eroing means, in effect, that at least in the case of some export transactions, the export prices are treated as if they were less than what they actually are."233
Secondly, the use of zeroing in the transaction-to-transaction comparison methodology, as in the weighted average-to-weighted average methodology, tends to result in higher margins of dumping. As the Appellate Body underscored in US – Corrosion-Resistant Steel Sunset Review, the use of zeroing:
231The relevant dictionary meaning of "fair" is "just, unbiased, equitable, impartial; legitimate, in accordance with the rules or standards", and "offering an equal chance of success". (Shorter Oxford English Dictionary, 5th edn, W.R. Trumble, A. Stevenson (eds) (Oxford University Press, 2002), Vol.1, p. 915)
232European Communities' third participant's submission, para. 30.
233Appellate Body Report, US – Softwood Lumber V, para. 101. (original emphasis)
WT/DS264/AB/RW
Page 57
... will tend to inflate the margins calculated. Apart from inflating the margins, such a methodology could, in some instances, turn a negative margin of dumping into a positive margin of dumping. ... Thus, the inherent bias in a zeroing methodology of this kind may distort not only the magnitude of a dumping margin, but also a finding of the very existence of dumping.234
Moreover, it has been argued in these Article 21.5 proceedings that the effect of zeroing is even more pronounced under the transaction-to-transaction comparison methodology than under the weighted average-to-weighted average methodology.235 Japan explains that, under the weighted average-to-weighted average comparison methodology, the effect of zeroing is "moderated" because certain transactions in which export prices exceed normal value are included in the aggregation for purposes of calculating the margins of dumping. This is because zeroing is not performed at the level of individual transactions, but after the transactions have been grouped (for instance, by model or type) and then averaged. 236 In contrast, the application of zeroing under the transaction-to-transaction comparison methodology excludes ab initio the results of all the comparisons in which the export prices are above normal value.237 In fact, at the oral hearing, the United States acknowledged that the use of zeroing under the transaction-to-transaction comparison methodology in the Section 129 Determination resulted in margins of dumping for respondent exporters or producers that were higher than those in the original determination, which had been calculated using zeroing under the weighted average-to-weighted average methodology.238
In sum, the use of zeroing under the transaction-to-transaction comparison methodology artificially inflates the magnitude of dumping, resulting in higher margins of dumping and making a positive determination of dumping more likely. This way of calculating cannot be described as impartial, even-handed, or unbiased. For this reason, we do not consider that the calculation of "margins of dumping", on the basis of a transaction-to-transaction comparison that uses zeroing,
234Appellate Body Report, US – Corrosion-Resistant Steel Sunset Review, para. 135.
235Canada's appellant's submission, para. 67. 236Japan's third participant's submission, para. 100. 237Ibid.
238Indeed, the Section 129 Determination resulted in an increase in the margins of dumping over those calculated in the original determination. The "all others" rate increased from 8.43 per cent to 11.54 per cent. Individual exporter rates increased as follows: Abitibi from 12.44 per cent to 13.22 per cent; Canfor from 5.96 per cent to 9.27 per cent; Slocan from 7.71 per cent to 12.91 per cent; Tembec from 10.21 per cent to 12.96 per cent; West Fraser from 2.18 per cent to 3.92 per cent; and Weyerhaeuser from 12.39 per cent to 16.35 per cent. (See Section 129 Determination, p. 22645; Notice of Amended Final Determination of Sales at Less Than Fair Value and Antidumping Duty Order: Certain Softwood Lumber Products From Canada, United States Federal Register, Vol. 67, No. 99 (22 May 2002) 36068, 36069; and Canada's appellant's submission, footnote 7 to para. 12) In the Section 129 proceeding, the USDOC used the information gathered in the original investigation and did not reopen the record. (Section 129 Determination, p. 22645)
WT/DS264/AB/RW
Page 58
satisfies the "fair comparison" requirement within the meaning of Article 2.4 of the Anti-Dumping Agreement.
According to the United States, the basis for determining whether a comparison methodology is "fair" within the meaning of Article 2.4 cannot be whether it yields a higher margin of dumping than another comparison methodology.239 Otherwise, the United States adds, investigating authorities "would have to determine at least two putative margins" and rely on the methodology that generated the lowest margin.240 We understand the United States to argue that, even if a comparison methodology that uses zeroing results in higher margins of dumping, it does not become "unfair" by this mere fact alone, provided that it is WTO-consistent. This proviso, however, has not been met because, as we have found, the use of zeroing under the transaction-to-transaction comparison methodology is inconsistent with Article 2.4.2 of the Anti-Dumping Agreement.
The Panel followed a logic similar to that of the United States when it reasoned that:
... the fact that comparison methodology A produces a higher margin of dumping than comparison methodology B would only make comparison methodology A unfair if comparison methodology B were the applicable standard. If, however, the [Anti-Dumping] Agreement were to permit either comparison methodology A or B, this would not be the case.241 (footnote omitted)
Given our finding under Article 2.4.2, the Panel's own logic would lead to a finding of unfairness.
The United States reiterates here some of the contextual considerations examined by the Panel in its analysis of Canada's Article 2.4.2 claim, including the "mathematical equivalence" argument. We have examined these contextual considerations in the previous section and concluded that they did not support the United States' interpretation that the use of zeroing under the transaction-to- transaction comparison methodology is consistent with Article 2.4.2. For the same reasons, we do not find that they are relevant to our examination of Canada's claim under Article 2.4.
Therefore, we reverse the Panel's finding, in paragraphs 5.78 and 6.1 of the Panel Report, that "the determination of the [US]DOC in the section 129 proceeding investigation is not inconsistent with ... Article[] 2.4 of the [Anti-Dumping] Agreement". We conclude that the USDOC's use of zeroing under the transaction-to-transaction comparison methodology in the Section 129
239United States' appellee's submission, paras. 96-97.
240Ibid., para. 100.
241Panel Report, para. 5.74.
WT/DS264/AB/RW
Page 59
Determination is inconsistent with the "fair comparison" requirement in Article 2.4 of the Anti- Dumping Agreement.
Findings and Conclusions
For the reasons set out in this Report, the Appellate Body:
reverses the Panel's finding, in paragraphs 5.66 and 6.1 of the Panel Report, that the USDOC's Section 129 Determination is not inconsistent with Article 2.4.2 of the Anti-Dumping Agreement and finds, instead, that the use of zeroing by the USDOC in the Section 129 Determination is inconsistent with the United States' obligations under Article 2.4.2 of the Anti-Dumping Agreement;
reverses the Panel's finding, in paragraphs 5.78 and 6.1 of the Panel Report, that the USDOC's Section 129 Determination is not inconsistent with Article 2.4 of the Anti- Dumping Agreement and finds, instead, that the use of zeroing in the Section 129 Determination is inconsistent with the "fair comparison" requirement in Article 2.4; and
consequently, reverses the Panel's conclusion, in paragraph 6.2 of the Panel Report, that "the United States has implemented the recommendations and rulings of the DSB in US – Softwood Lumber V, to bring its measure into conformity with its obligations under the [Anti-Dumping] Agreement".
The Appellate Body recommends that the Dispute Settlement Body request the United States to bring its measure into conformity with its obligations under the Anti-Dumping Agreement.
WT/DS264/AB/RW
Page 60
Signed in the original in Geneva this 28th day of July 2006 by:
Georges Abi-Saab Presiding Member
Luiz Olavo Baptista Giorgio Sacerdoti Member Member
WT/DS264/AB/RW
Page 61
ANNEX I
WT/DS264/25
17 May 2006
(06-2427)
Original: English
UNITED STATES – FINAL DUMPING DETERMINATION ON SOFTWOOD LUMBER FROM CANADA
Recourse to Article 21.5 of the DSU by Canada
Notification of an Appeal by Canada
under Article 16.4 and Article 17 of the Understanding on Rules and Procedures Governing the Settlement of Disputes (DSU),
and under Rule 20(1) of the Working Procedures for Appellate Review
The following notification, dated 17 May 2006, from the Delegation of Canada, is being circulated to Members.
Pursuant to paragraph 4 of Article 16 of the Understanding on Rules and Procedures Governing the Settlement of Disputes (the "DSU") and Rule 20 of the Working Procedures for Appellate Review, Canada appeals certain issues of law and certain legal interpretations covered in the Panel Report in United States – Final Dumping Determination on Softwood Lumber From Canada
Recourse to Article 21.5 of the DSU by Canada.1
The Panel found that the U.S. Department of Commerce's ("USDOC") section 129 dumping determination on softwood lumber was not inconsistent with Articles 2.4.2 and 2.4 of the Agreement on Implementation of Article VI of the General Agreement on Tariffs and Trade 1994 ("ADA").2 Consequently, the Panel determined that the United States had implemented the rulings and recommendations of the Dispute Settlement Body ("DSB").3 These findings are in error as they are based on erroneous findings on issues of law and related legal interpretations.
1 Report of the Panel, WT/DS264/RW, circulated April 3, 2006.
2 Ibid., at paras. 5.27-5.28, 5.30, 5.65-5.66, 5.78, 6.1.
3 Ibid., at para. 6.2. See DSB, Minutes of Meeting (31 August 2004), WT/DSB/M/175, September 24,
2004, 4(a), at para. 42.
WT/DS264/AB/RW
Page 62
Canada requests that the Appellate Body review, and reverse, these findings of the Panel.
Specifically, Canada asks the Appellate Body to find:
first, the Panel erred in interpreting Article 2.4.2 to permit the USDOC to treat transaction-specific comparisons that result in a negative value (i.e., non-dumped transactions) as zero in calculating "margins of dumping" under the transaction-to- transaction methodology. These erroneous findings appear throughout the panel report including paragraphs 5.17-5.30, 5.65-5.66, and 6.14; and
second, the Panel erred in finding that Article 2.4 permits the treatment of transaction-specific comparisons that result in a negative value as zero in the calculation of "margins of dumping" under the transaction-to-transaction methodology. These findings are set out in paragraphs 5.72-5.78 and 6.1 of the panel report.
In the light of these errors, Canada respectfully requests that the Appellate Body:
reverse the findings of the Panel set out above;
find that the USDOC's use of zeroing under the transaction-to-transaction methodology in the section 129 determination, as described above, is inconsistent with Articles 2.4.2 and 2.4 of the ADA; and
determine that, as a consequence, the United States has failed to implement the recommendations and rulings of the DSB.
4 The USDOC confirmed that it treated negative transaction-specific comparison results as zero in its section 129 determination, at one point describing the process as follows: "… because the Appellate Body report requires the offset for non-dumped sales only for a weighted-average-to-weighted-average comparison, we have not applied the offset for non-dumped sales in our transaction-to-transaction comparison." See Notice of Determination Under Section 129 of the Uruguay Round Agreements Act; Antidumping Measures Concerning Certain Softwood Lumber Products From Canada, 70 Fed. Reg. 22,636 at 22,639 (Dep't Commerce May 2, 2005).