United States – anti-dumping and countervailing measures on large residential washers from korea

AB-2016-2

Report of the Appellate Body

Table of Contents

1   Introduction.. 12

2   Arguments of the Participants. 18

3   Arguments of the third participants. 18

4   Issues Raised in This Appeal. 18

5   Analysis of the Appellate Body. 20

5.1   Claims under the Anti-Dumping Agreement and related claims under the GATT 1994. 20

5.1.1   Background. 20

5.1.1.1   The Nails II methodology and the Washers anti‑dumping investigation. 20

5.1.1.2   The DPM and the first administrative review of the Washers anti‑dumping order 22

5.1.2   Article 2.4.2 of the Anti-Dumping Agreement 23

5.1.3   The relevant "pattern" for the purposes of the second sentence of Article 2.4.2 of the Anti‑Dumping Agreement 25

5.1.4   Whether the Panel erred in finding that "the DPM is inconsistent 'as such' with the second sentence of Article 2.4.2 [of the Anti-Dumping Agreement] because, by aggregating random and unrelated price variations, it does not properly establish 'a pattern of export prices which differ significantly among different purchasers, regions or time periods'". 30

5.1.5   Whether the Panel erred in finding that the W-T comparison methodology should only be applied to "pattern transactions". 31

5.1.6   The extent to which price differences are to be assessed quantitatively, qualitatively, and in light of the "reasons" for these price differences. 35

5.1.7   Whether an explanation needs to be provided with respect to both the W-W and the T-T comparison methodologies. 40

5.1.8   "Systemic disregarding". 43

5.1.8.1   Whether the Panel erred in finding that Korea failed to establish that the United States' use of "systemic disregarding" under the DPM is inconsistent "as such" with the second sentence of Article 2.4.2 of the Anti‑Dumping Agreement 43

5.1.8.2   Whether the Panel erred in finding that Korea failed to establish that the United States' use of "systemic disregarding" under the DPM is inconsistent "as such" with Article 2.4 of the Anti‑Dumping Agreement 55

5.1.9   Zeroing under the W-T comparison methodology. 58

5.1.9.1   Whether the Panel erred in finding that the United States' use of zeroing when applying the W‑T comparison methodology is inconsistent "as such" and "as applied" in the Washers anti‑dumping investigation with Article 2.4.2 of the Anti‑Dumping Agreement 58

5.1.9.2   Whether the Panel erred in finding that the United States' use of zeroing when applying the W‑T comparison methodology is inconsistent "as such" and "as applied" in the Washers anti‑dumping investigation with Article 2.4 of the Anti‑Dumping Agreement 65

5.1.9.3   Whether the Panel erred in finding that the United States' use of zeroing when applying the W‑T comparison methodology in administrative reviews is inconsistent "as such" with Article 9.3 of the Anti‑Dumping Agreement and Article VI:2 of the GATT 1994. 67

5.1.10   Separate opinion of one Appellate Body Member regarding zeroing under the W‑T comparison methodology. 69

5.2   Claims under the SCM Agreement and related claims under the GATT 1994. 71

5.2.1   Background. 71

5.2.2   Whether the Panel erred in its interpretation and application of Article 2.2 of the SCM Agreement by upholding the USDOC's determination that the RSTA Article 26 tax credit programme was regionally specific  73

5.2.2.1   Whether the term "certain enterprises" in Article 2.2 of the SCM Agreement is limited to entities with legal personality. 74

5.2.2.2   Whether the "designation" of a region for the purposes of Article 2.2 of the SCM Agreement must be affirmative and explicit, or may also be carried out by implication. 78

5.2.2.3   Whether the concept of "geographical region" for the purposes of Article 2.2 of the SCM Agreement depends on the territorial size of the area covered by a subsidy. 79

5.2.2.4   Conclusions. 81

5.2.3   Whether the Panel failed to conduct an objective assessment of the matter before it in articulating its findings on regional specificity. 82

5.2.4   Whether the Panel erred in its interpretation and application of Article 19.4 of the SCM Agreement and Article VI:3 of the GATT 1994 by upholding the USDOC's determination that the tax credits received by Samsung under Articles 10(1)(3) and 26 of the RSTA were not tied to particular products. 83

5.2.4.1   The Panel's affirmation of the USDOC's test for ascertaining whether the RSTA Article 10(1)(3) and Article 26 tax credits were tied to particular products. 86

5.2.4.2   The Panel's affirmation of the USDOC's dismissal of certain evidence submitted by Samsung  92

5.2.4.3   Conclusions. 95

5.2.5   Whether the Panel erred in its interpretation and application of Article 19.4 of the SCM Agreement and Article VI:3 of the GATT 1994 by upholding the USDOC's attribution of the tax credits received by Samsung under Article 10(1)(3) of the RSTA to Samsung's domestic production only. 95

6   Findings And Conclusions. 101

6.1   The relevant "pattern" for the purposes of the second sentence of Article 2.4.2 of the Anti-Dumping Agreement 101

6.2   The scope of application of the W-T comparison methodology in the second sentence of Article 2.4.2 of the Anti-Dumping Agreement 102

6.3   Prices which differ "significantly" under the second sentence of Article 2.4.2 of the Anti-Dumping Agreement 102

6.4   The explanation to be provided under the second sentence of Article 2.4.2 of the Anti-Dumping Agreement 103

6.5   "Systemic disregarding". 103

6.6   Zeroing under the W-T comparison methodology. 104

6.7   Article 2.2 of the SCM Agreement 105

6.8   Article 19.4 of the SCM Agreement and Article VI:3 of the GATT 1994. 106

6.9   Recommendation. 106

ABBREVIATIONS USED IN THIS REPORT

PANEL EXHIBITS CITED IN THIS REPORT

CASES CITED IN THIS REPORT

World Trade Organization

Appellate Body

1  Introduction

1.1.  The United States and Korea each appeals certain issues of law and legal interpretations developed in the Panel Report, United States – Anti-Dumping and Countervailing Measures on Large Residential Washers from Korea^[1] (Panel Report). The Panel was established on 22 January 2014 to consider a complaint by Korea^[2] with respect to the consistency of the United States' measures imposing definitive anti-dumping and countervailing duties on imports of large residential washers (LRWs) from Korea with the Agreement on Implementation of Article VI of the General Agreement on Tariffs and Trade 1994 (Anti‑Dumping Agreement), the Agreement on Subsidies and Countervailing Measures (SCM Agreement), and the General Agreement on Tariffs and Trade 1994 (GATT 1994).

1.2.  On 3 September 2014, after consultations with the parties, the Panel adopted additional working procedures for the protection of business confidential information (BCI Procedures).^[3] On 12 December 2014, the Panel rejected China's request for enhanced third party rights.^[4] On 13 February 2015, the Panel also rejected the European Union's request to amend certain aspects of its working procedures and found that there was no need to modify its BCI Procedures in light of the "reservations" raised by the European Union.^[5]

1.3.  With regard to the United States' measures imposing definitive anti-dumping duties on imports of LRWs from Korea^[6], Korea challenged before the Panel certain aspects of the methodologies used by the United States Department of Commerce (USDOC) to determine whether to apply the weighted average‑to-transaction (W-T) comparison methodology. Specifically, Korea challenged: (i) the so-called "Nails II methodology"^[7] used in the anti-dumping investigation conducted by the USDOC concerning imports of LRWs from Korea^[8] (Washers anti‑dumping investigation); (ii) the so-called "Differential Pricing Methodology" that replaced the Nails II methodology as of March 2013 (DPM) "as such"; (iii) the DPM "as applied" in the first administrative review of the anti‑dumping order imposing anti-dumping duties on LRWs from Korea issued by the USDOC on 15 February 2013^[9] (Washers anti-dumping order); and (iv) the ongoing and future application of the DPM in connection with the Washers anti‑dumping investigation. Korea also challenged the USDOC's use of "zeroing" in the context of the W-T comparison methodology.^[10] Specifically, Korea challenged: (i) "as such" the rule or norm pursuant to which the USDOC engages in zeroing; and (ii) zeroing "as applied" in the Washers anti-dumping investigation.^[11]

1.4.  With regard to the United States' measures imposing definitive countervailing duties on imports of LRWs from Korea in connection with the Washers countervailing duty investigation^[12], Korea challenged under the SCM Agreement the USDOC's determinations that two tax credit programmes^[13] were specific. Moreover, Korea raised claims under the SCM Agreement and the GATT 1994 challenging the manner in which the USDOC calculated the ad valorem subsidy rate for Samsung Electronics Co., Ltd (Samsung)^[14] under those programmes.^[15]

1.5.  In the Panel Report, circulated to Members of the World Trade Organization (WTO) on 11 March 2016, the Panel found as follows concerning the anti-dumping measures at issue:

a.    with regard to the Washers anti-dumping investigation:

i.      the United States acted inconsistently with the second sentence of Article 2.4.2 of the Anti-Dumping Agreement by applying the W-T comparison methodology to transactions other than those constituting the patterns of transactions that the USDOC had determined to exist^[16];

ii.     Korea failed to establish that the United States acted inconsistently with the second sentence of Article 2.4.2 of the Anti-Dumping Agreement by determining the existence of a "pattern of export prices which differ significantly" among purchasers, regions or time periods on the basis of purely quantitative criteria, without any qualitative assessment of the reasons for the relevant price differences^[17];

iii.    the United States acted inconsistently with the second sentence of Article 2.4.2 of the Anti-Dumping Agreement by merely focusing on the difference between the margin of dumping calculated using the weighted average-to-weighted average (W‑W) comparison methodology and the margin of dumping calculated using the W‑T comparison methodology, and by failing to consider whether the factual circumstances surrounding the relevant price differences were suggestive of something other than "targeted dumping"^[18]; and

iv.   Korea failed to establish that the United States acted inconsistently with the second sentence of Article 2.4.2 of the Anti-Dumping Agreement by failing to explain why the relevant price differences could not be taken into account appropriately by the transaction-to-transaction (T-T) comparison methodology^[19];

b.    with regard to the DPM:

i.      the DPM is inconsistent "as such" with the second sentence of Article 2.4.2 of the Anti-Dumping Agreement because it applies the W-T comparison methodology to "non-pattern transactions" when the aggregated value of sales to purchasers, regions, and time periods that pass the "Cohen's d test"^[20] accounts for 66% or more of the value of total sales^[21];

ii.     Korea failed to establish that the DPM is inconsistent "as such" with the second sentence of Article 2.4.2 of the Anti-Dumping Agreement by determining the existence of "a pattern of export prices which differ significantly" among purchasers, regions or time periods on the basis of purely quantitative criteria, without any qualitative assessment of the reasons for the relevant price differences^[22];

iii.    the DPM is inconsistent "as such" with the second sentence of Article 2.4.2 of the Anti-Dumping Agreement because, in applying the "meaningful difference test"^[23], the DPM focuses on the difference between the margin of dumping calculated using the W-W comparison methodology and the margin of dumping calculated using the W-T comparison methodology or the "mixed" comparison methodology.^[24] The Panel also found that the DPM fails to provide for any consideration of whether the factual circumstances surrounding the relevant price differences are suggestive of something other than "targeted dumping"^[25];

iv.   Korea failed to establish that the DPM is inconsistent with the second sentence of Article 2.4.2 of the Anti-Dumping Agreement when, having concluded that the W-W comparison methodology cannot appropriately take into account the observed pattern of significantly different prices, it does not also consider whether the relevant price differences could be taken into account appropriately by the T-T comparison methodology^[26];

v.    the DPM is inconsistent "as such" with the second sentence of Article 2.4.2 of the Anti-Dumping Agreement because, by aggregating random and unrelated price variations, the DPM does not properly establish "a pattern of export prices which differ significantly among different purchasers, regions or time periods"^[27];

vi.   Korea failed to establish that the United States' use of "systemic disregarding"^[28] under the DPM is inconsistent "as such" with the second sentence of Article 2.4.2 of the Anti-Dumping Agreement^[29]; and

vii.  Korea failed to establish that the United States' use of "systemic disregarding" under the DPM is inconsistent "as such" with Article 2.4 of the Anti-Dumping Agreement^[30]; and

c.    with regard to zeroing:

i.      the United States' use of zeroing when applying the W-T comparison methodology is inconsistent "as such" with Article 2.4.2 of the Anti-Dumping Agreement^[31];

ii.     the United States' use of zeroing when applying the W-T comparison methodology is inconsistent "as such" with Article 2.4 of the Anti-Dumping Agreement^[32];

iii.    the USDOC acted inconsistently with Article 2.4.2 of the Anti-Dumping Agreement by using zeroing when applying the W-T comparison methodology in the Washers anti‑dumping investigation^[33];

iv.   the USDOC acted inconsistently with Article 2.4 of the Anti-Dumping Agreement by using zeroing when applying the W-T comparison methodology in the Washers anti‑dumping investigation^[34]; and

v.    the United States' use of zeroing when applying the W-T comparison methodology in administrative reviews is inconsistent "as such" with Article 9.3 of the Anti-Dumping Agreement and Article VI:2 of the GATT 1994.^[35]

1.6.  In addition, the Panel found as follows concerning the countervailing duties at issue:

a.    the USDOC's original and remand determinations that the "RSTA Article 10(1)(3) tax credit programme"^[36] is de facto specific because Samsung received subsidies under that programme in disproportionately large amounts are inconsistent with Article 2.1(c) of the SCM Agreement^[37];

b.    the USDOC acted inconsistently with Article 2.1(c) of the SCM Agreement by failing to take account of the two mandatory factors referred to in the final sentence of that provision in its determination of de facto specificity^[38];

c.    Korea failed to establish that the USDOC's determination of regional specificity in respect of the "RSTA Article 26 tax credit programme"^[39] is inconsistent with Article 2.2 of the SCM Agreement^[40];

d.    Korea failed to establish that the USDOC's failure to tie the subsidies claimed by Samsung under the RSTA Article 10(1)(3) and Article 26 tax credit programmes to Samsung's digital appliance products (including LRWs) is inconsistent with Article 19.4 of the SCM Agreement and Article VI:3 of the GATT 1994^[41]; and

e.    Korea failed to establish that the USDOC acted inconsistently with Article 19.4 of the SCM Agreement and Article VI:3 of the GATT 1994 by limiting the denominator to the sales value of products produced by Samsung in Korea when allocating the benefit conferred to Samsung under the RSTA Article 10(1)(3) tax credit programme.^[42]

1.7.  In accordance with Article 19.1 of the Understanding on Rules and Procedures Governing the Settlement of Disputes (DSU), and having found that the United States had acted inconsistently with certain provisions of the Anti-Dumping Agreement, the SCM Agreement, and the GATT 1994, the Panel recommended that the United States bring its measures into conformity with its obligations under those Agreements.^[43]

1.8.  On 19 April 2016, the United States notified the Dispute Settlement Body (DSB), pursuant to Articles 16.4 and 17 of the DSU, of its intention to appeal certain issues of law covered in the Panel Report and certain legal interpretations developed by the Panel and filed a Notice of Appeal^[44] and appellant's submission pursuant to Rule 20 and Rule 21, respectively, of the Working Procedures for Appellate Review^[45] (Working Procedures). On 25 April 2016, Korea notified the DSB, pursuant to Articles 16.4 and 17 of the DSU, of its intention to appeal certain issues of law covered in the Panel Report and certain legal interpretations developed by the Panel and filed a Notice of Other Appeal^[46] and other appellant's submission pursuant to Rule 23 of the Working Procedures. On 9 May 2016, Korea and the United States each filed an appellee's submission.^[47] On 9 May 2016, China filed a third participant's submission.^[48] On 10 May 2016, Brazil, Canada, the European Union, Japan, Norway, and Viet Nam each filed a third participant's submission.^[49] On 17 June 2016, India, Saudi Arabia, Thailand, and Turkey each notified its intention to appear at the oral hearing as a third participant.^[50]

1.9.  On 22 April 2016, the Appellate Body Secretariat transmitted the Working Schedule for Appeal drawn up by the Appellate Body Division hearing this appeal, setting out the deadlines for filing written submissions.

1.10.  On 5 April 2016, Korea and the United States had jointly addressed a letter to the Chair of the Appellate Body, attaching a request that the Division that would eventually hear an appeal in this dispute adopt, pursuant to Rule 16(1) of the Working Procedures, additional procedures for the protection of BCI on the record of this dispute (joint request). Korea and the United States requested the Division to adopt additional procedures for the protection of BCI on the basis of the BCI procedures adopted by the Panel, and attached draft procedures to the joint request. They explained that BCI procedures in this appeal would serve "the interest of fairness and orderly procedure in the conduct of an appeal", according to Rule 16(1) of the Working Procedures. On 8 April 2016, the European Union addressed a letter to the Chair of the Appellate Body commenting on the joint request. The European Union expressed the view that BCI procedures at the appellate stage should not be based on the Panel's BCI procedures.

1.11.  By letter dated 19 April 2016, the United States sought guidance from the Appellate Body on how to proceed with filing its appellant's submission, which was to be filed that day and contained information that was designated as BCI in the Panel proceedings. In a letter issued on the same day, the Chair of the Appellate Body, on behalf of the Division hearing this appeal, informed the United States, Korea, and the third parties that, pending a final decision on the joint request, the Division had decided to provide provisional additional protection to information marked as BCI in the United States' appellant's submission and in an eventual other appellant's submission by Korea.^[51]

1.12.  On 21 April 2016, the Chair of the Appellate Body addressed a letter on behalf of the Division hearing this appeal to the participants, asking them to substantiate further why certain information contained in their submissions and in the Panel record warranted special protection at the appellate stage beyond that already provided under the confidentiality standards set out in Articles 17.10 and 18.2 of the DSU, and the Rules of Conduct.^[52] Korea and the United States responded to this request with separate communications on 26 April 2016. On the same date, the Division invited the third participants to provide further comments on the joint request and on Korea's and the United States' responses of 26 April 2016. On 27 April 2016, the European Union provided comments and, on 28 April 2016, China indicated that it had no objection to the requested additional protection of BCI.

1.13.  On 9 May 2016, the Division hearing this appeal issued a Procedural Ruling on the protection of BCI on the record in this dispute, which is contained in Annex D of the Addendum to this Report, WT/DS464/AB/R/Add.1.

1.14.  By letter dated 19 May 2016 to the Chair of the Appellate Body, Korea requested authorization, pursuant to Rule 18(5) of the Working Procedures, to correct certain clerical errors in its Notice of Other Appeal, other appellant's submission, and appellee's submission. In accordance with Rule 18(5), the Division, by letter dated 20 May 2016, provided the United States, the third participants, and third parties with an opportunity to comment in writing on the request by 23 May 2016. No objections to Korea's request were received and, on 25 May 2016, the Division authorized Korea to correct the clerical errors in its Notice of Other Appeal, other appellant's submission, and appellee's submission, as identified in its letter of 19 May 2016.

1.15.  In a letter from the Appellate Body dated 15 June 2016, the United States and Korea were asked to keep their opening statements at the oral hearing to 30 minutes each and the third participants to keep their opening statements to a maximum of five minutes each. In a communication dated 16 June 2016, China requested the Division to allocate further five minutes for its opening statement. China also stated that it looked forward to a full opportunity to engage on the relevant issues during the hearing, in light of its direct and immediate interest in the issues raised in this appeal.^[53] By letter dated 17 June 2016, the Presiding Member explained that, in light of the many issues raised in this appeal and in order to be able to complete the hearing within a reasonable time-frame, the Division did not consider that it would be appropriate to extend the time allocated for opening statements.

1.16.  On 17 June 2016, the Chair of the Appellate Body notified the Chair of the DSB that the Appellate Body would not be able to circulate its Report in this appeal within the 60-day period stipulated in Article 17.5 of the DSU, or within the 90-day period pursuant to the same provision, and informed the Chair of the DSB that the circulation date of the Appellate Body Report in this appeal would be communicated to the participants and third participants after the oral hearing.^[54] The Chair of the Appellate Body explained that this was due to a number of factors, including the substantial workload of the Appellate Body in 2016, scheduling difficulties arising from overlap in the composition of the Divisions hearing the different appeals, the number and complexity of the issues raised in this and concurrent appellate proceedings, together with the demands that these concurrent appeals place on the WTO Secretariat's translation services, and the shortage of staff in the Appellate Body Secretariat. On 7 July 2016, the Chair of the Appellate Body notified the Chair of the DSB that the Report in this appeal would be circulated to WTO Members no later than Wednesday, 7 September 2016.^[55] 

1.17.  The oral hearing in this appeal was held on 20-21 June 2016.^[56] The participants and six third participants (Brazil, Canada, China, Japan, Norway, and Viet Nam) made opening oral statements. The participants and third participants responded to questions posed by the Appellate Body Division hearing this appeal.

2  Arguments of the Participants

2.1.  The claims and arguments of the participants are reflected in the executive summaries of their written submissions provided to the Appellate Body.^[57] The Notices of Appeal and Other Appeal, and the executive summaries of the participants' claims and arguments, are contained in Annexes A and B of the Addendum to this Report, WT/DS464/AB/R/Add.1.

3  Arguments of the third participants

3.1.  The arguments of the third participants that filed a written submission are reflected in the executive summaries of those submissions provided to the Appellate Body^[58] and are contained in Annex C of the Addendum to this Report, WT/DS464/AB/R/Add.1.

4  Issues Raised in This Appeal

4.1.  With regard to the anti-dumping measures, the following issues are raised in this appeal:

a.    whether the Panel erred in its interpretation and application of the second sentence of Article 2.4.2 of the Anti-Dumping Agreement in relation to the Nails II methodology and the DPM. In particular:

i.      whether the Panel erred in its interpretation of the relevant "pattern" for the purposes of the second sentence of Article 2.4.2 in finding that "the relevant 'pattern' … comprises only low-priced export transactions to each particular 'target' (be that a purchaser, or a region, or a time period), while other higher-priced export transactions to other purchasers, regions, or time periods are 'non-pattern' transactions"^[59] (raised by the United States);

ii.     whether the Panel erred in finding that the DPM is inconsistent "as such" with the second sentence of Article 2.4.2 because, by aggregating random and unrelated price variations, it does not properly establish "a pattern of export prices which differ significantly among purchasers, regions or time periods" (raised by the United States);

iii.    whether the Panel erred in finding that the United States acted inconsistently with the second sentence of Article 2.4.2 because the USDOC applied the W-T comparison methodology to all export transactions, including transactions other than those constituting the patterns of transactions that it had determined to exist in the Washers anti-dumping investigation (raised by the United States);

iv.   whether the Panel erred in finding that the DPM is inconsistent "as such" with the second sentence of Article 2.4.2 because it applies the W-T comparison methodology to "non-pattern transactions" when the aggregated value of sales to purchasers, regions, and time periods that pass the Cohen's d test accounts for 66% or more of the value of total sales (raised by the United States);

v.    whether the Panel erred in finding that Korea failed to establish that the United States acted inconsistently with the second sentence of Article 2.4.2 in the Washers anti-dumping investigation because the USDOC determined the existence of "a pattern of export prices which differ significantly" based on purely quantitative criteria, without any qualitative assessment of the "reasons" for the relevant price differences (raised by Korea);

vi.   whether the Panel erred in finding that Korea failed to establish that the DPM is inconsistent "as such" with the second sentence of Article 2.4.2 because it determines the existence of "a pattern of export prices which differ significantly" based on purely quantitative criteria, without any qualitative assessment of the "reasons" for the relevant price differences (raised by Korea);

vii.  whether the Panel erred in finding that Korea failed to establish that the United States acted inconsistently with the second sentence of Article 2.4.2 in the Washers anti-dumping investigation because the USDOC did not explain why the relevant price differences could not be taken into account appropriately by the T-T comparison methodology (raised by Korea); and

viii. whether the Panel erred in finding that Korea failed to establish that the DPM is inconsistent with the second sentence of Article 2.4.2 because when, having concluded that the W-W comparison methodology cannot appropriately take into account the observed price differences, it does not also consider whether the relevant price differences could be taken into account appropriately by the T-T comparison methodology (raised by Korea);

b.    whether the Panel erred in finding that the use of "systemic disregarding" in the context of the DPM, whereby, when the W-T comparison methodology applied to "pattern transactions" is combined with the W-W comparison methodology applied to "non‑pattern transactions", an overall negative comparison result arising from the application of the W-W comparison methodology is disregarded or set to zero, is not inconsistent "as such" with Article 2.4 and the second sentence of Article 2.4.2 of the Anti-Dumping Agreement (raised by Korea); and

c.    whether the Panel erred in its interpretation and application of Articles 2.4, 2.4.2, and 9.3 of the Anti-Dumping Agreement and Article VI:2 of the GATT 1994 when it found that the use of zeroing when applying the W-T comparison methodology is inconsistent "as such" with these provisions and that the United States acted inconsistently with Article 2.4 and the second sentence of Article 2.4.2 by using zeroing when applying the W-T comparison methodology in the Washers anti-dumping investigation (raised by the United States).

4.2.  With regard to the countervailing duties, the following issues are raised in this appeal:

a.    whether the Panel erred in its interpretation and application of Article 2.2 of the SCM Agreement by upholding the USDOC's determination that the RSTA Article 26 tax credit programme was regionally specific (raised by Korea);

b.    whether the Panel failed to conduct an objective assessment of the matter before it in articulating its findings on regional specificity, thereby acting inconsistently with its duties under Article 11 of the DSU (raised by Korea);

c.    whether the Panel erred in its interpretation and application of Article 19.4 of the SCM Agreement and Article VI:3 of the GATT 1994 by upholding the USDOC's determination that the tax credits received by Samsung under Articles 10(1)(3) and 26 of the RSTA were not tied to particular products (raised by Korea);

d.    whether the Panel failed to conduct an objective assessment of the matter before it in finding that tax credits bestowed under Article 10(1)(3) of the RSTA are not research and development (R&D) subsidies, thereby acting inconsistently with its duties under Article 11 of the DSU (raised by Korea); and

e.    whether the Panel erred in its interpretation and application of Article 19.4 of the SCM Agreement and Article VI:3 of the GATT 1994 by upholding the USDOC's attribution of the tax credits received by Samsung under Article 10(1)(3) of the RSTA to the sales value of Samsung's products manufactured in Korea only (raised by Korea).

5  Analysis of the Appellate Body

5.1.  We first address the claims raised on appeal under the Anti‑Dumping Agreement and the related claims under the GATT 1994, before turning to the claims on appeal under the SCM Agreement and the related claims under the GATT 1994.

5.1  Claims under the Anti-Dumping Agreement and related claims under the GATT 1994

5.1.1  Background

5.1.1.1  The Nails II methodology and the Washers anti‑dumping investigation

5.2.  The Washers anti‑dumping investigation was initiated on 19 January 2012 and concerned LRWs produced by three Korean companies, including LG Electronics, Inc. (LGE) and Samsung.^[60] During the investigation, the USDOC applied the Nails II methodology to determine whether "targeted dumping" was occurring and whether to use the W-T comparison methodology to determine dumping margins for LGE and Samsung.

5.3.  The Nails II methodology applied a two‑stage test run on a model basis, with each model being assigned a control number (CONNUM).^[61] First, following an allegation of "targeted dumping" by a member of the domestic industry, the "standard deviation test" aimed to establish whether there were price differences by considering whether the weighted average price to one "targeted" group (i.e. customer, time period, or region) in a particular CONNUM was below a benchmark price equal to one standard deviation below the weighted average mean price in that CONNUM. Second, the "gap test" aimed to establish whether the observed price differences were significant. Where the gap between the weighted average price of the sales to the "targeted" group and the next highest weighted average price of sales to a "non-targeted" group exceeded the average gap among the weighted average prices between the "non-targeted" groups, those sales were considered to have passed the gap test.

5.4.  If both tests were met^[62], the USDOC evaluated the difference between the weighted average dumping margin calculated using the W-W comparison methodology (without zeroing) and that calculated using the W-T comparison methodology (with zeroing). The W-T comparison methodology was applied to all export sales where there was a meaningful difference between the two results.

5.5.  During the Washers anti‑dumping investigation, a domestic producer of LRWs had alleged that "targeted dumping" was occurring with respect to LRWs produced in, and exported from, Korea by LGE and Samsung.^[63] LGE and Samsung asserted that the alleged "targeted" sales corresponded to, inter alia, normal promotional practices.^[64] The USDOC, however, took the view that it was not required "to consider why [the price] differences exist".^[65] Applying the Nails II methodology, the USDOC found "a pattern of U.S. prices … that differs significantly among certain time periods, customers, and regions" for LGE as well as for Samsung.^[66] The identified "pattern" was "defined by all of the respondent's U.S. sales".^[67] Moreover, the USDOC explained that using the W‑W comparison methodology "conceal[ed] differences" between the export prices and that there was a "meaningful difference" between the margin of dumping calculated using the W-W comparison methodology and the margin of dumping calculated using the W-T comparison methodology.^[68] Consequently, the USDOC applied the W-T comparison methodology with zeroing to all of LGE's and Samsung's export transactions to establish margins of dumping for these two exporters.^[69]

5.6.  On 26 December 2012, the USDOC issued a Notice of Final Determination finding that LRWs were being sold or were likely to be sold in the United States at less than fair value.^[70] On this basis, and on the basis of the material injury determination of the United States International Trade Commission, the USDOC issued, on 15 February 2013, the Washers anti‑dumping order.^[71]

5.1.1.2  The DPM and the first administrative review of the Washers anti‑dumping order

5.7.  The USDOC initiated an administrative review of the Washers anti‑dumping order on 1 April 2014.^[72] In this administrative review, the USDOC applied the DPM, which replaced the Nails II methodology as of March 2013^[73], to determine whether to apply the W‑T comparison methodology to establish dumping margins for LGE.^[74]

5.8.  Unlike the Nails II methodology, where an allegation of "targeted dumping" from the domestic industry was required, the USDOC applies the DPM on its own motion.^[75] The DPM consists of three main components.^[76]

5.9.  First, the "Cohen's d test" evaluates the extent of the difference between the mean price of a test group that comprises the sales to a particular purchaser, region, or time period and the mean price of a comparison group that comprises all other sales of comparable merchandise.^[77] The Cohen's d test is applied by the USDOC when the test and comparison groups each have at least two observations (i.e. two transactions) and when the sales quantity for the comparison group accounts for at least 5% of the total sales quantity of the comparable merchandise. The Cohen's d coefficient is then calculated to evaluate the extent to which the net prices to a particular purchaser, region, or time period differ significantly from the net prices of all other sales of comparable merchandise. The extent of the difference is quantified by one of three fixed thresholds defined by the Cohen's d test: small (0.2); medium (0.5); or large (0.8). The difference is considered significant if the Cohen's d coefficient is equal to or exceeds the large threshold.

5.10.  Unlike the Nails II methodology, the DPM is not concerned with prices that are "targeted" at a particular customer, region, or time period. Rather, the focus is on any differences in prices, irrespective of whether such prices are above or below the average and without the prior identification of a specific purchaser, region, or time period. The DPM thus analyses any given export transaction in three different ways (by purchaser, region, and time period) in order to identify six possible types of price variation, namely: (i) higher prices to a particular purchaser; (ii) lower prices to a particular purchaser; (iii) higher prices in a particular region; (iv) lower prices in a particular region; (v) higher prices during a particular time period; and (vi) lower prices during a particular time period.^[78]

5.11.  Second, the "ratio test" assesses the extent of the significant price differences for all sales as measured by the Cohen's d test. When the result of the ratio test is above 66%, in the sense that the value of the transactions to purchasers, regions, and time periods that pass the Cohen's d test accounts for 66% or more of the value of the total sales, the application of the W-T comparison methodology (with zeroing) is being considered for all export transactions. When it is between 33% and 66%, a combined application of the W-T comparison methodology and the W-W comparison methodology is being considered (this is referred to as the "mixed" comparison methodology by the Panel). When it is below 33%, the application of the W-T comparison methodology is not being considered.

5.12.  Third, the "meaningful difference test" examines whether the use of the W-W comparison methodology can appropriately account for the differences found to exist in the first two stages of the analysis. By applying this test, the USDOC compares the weighted average dumping margin obtained using the W-T comparison methodology with that resulting from the use of the W-W comparison methodology. If there is a meaningful difference between the two results, this demonstrates that the W-W comparison methodology cannot account for the identified price differences. A difference is considered meaningful if there is a 25% relative change in the weighted average dumping margin or if the weighted average dumping margin moves across the de minimis threshold. If the meaningful difference test is met and the result of the ratio test is above 66%, the W-T comparison methodology is applied to all export transactions. By contrast, if the meaningful difference test is met and the result of the ratio test is between 33% and 66%, the W‑T comparison methodology is applied to those sales that pass the Cohen's d test and the W‑W comparison methodology is applied to the remaining sales.^[79]

5.13.  Applying the DPM in the first administrative review of the Washers anti‑dumping order, the USDOC determined that 47.12% of LGE's sales passed the Cohen's d test.^[80] The result of the ratio test was thus between 33% and 66%. The USDOC further determined that the meaningful difference test was met.^[81] Consequently, the USDOC applied the W-T comparison methodology (with zeroing) to those sales that passed the Cohen's d test and the W-W comparison methodology (without zeroing) to the remaining sales of LGE. When combining the overall comparison results of the W-T and the W-W comparison methodologies, the USDOC disregarded the overall negative comparison result arising from the W‑W comparison methodology (this is referred to as "systemic disregarding"). The preliminary results of the first administrative review of the Washers anti‑dumping order were presented on 9 March 2015 and the final results were published on 8 September 2015.^[82]

5.1.2  Article 2.4.2 of the Anti-Dumping Agreement

5.14.  Article 2.4.2 of the Anti‑Dumping Agreement reads: 

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.

5.15.  The first sentence of Article 2.4.2 provides for two symmetrical comparison methodologies that "shall normally" be used by investigating authorities to establish "margins of dumping": (i) the W-W comparison methodology, whereby dumping margins are established on the basis of a comparison of a weighted average normal value with a weighted average of prices of all comparable export transactions; and (ii) the T-T comparison methodology, whereby normal value and export prices are compared on a transaction-specific basis. As the Appellate Body has explained, the W-W and T-T comparison methodologies "fulfil the same function" and there is no "hierarchy between the two".^[83] Accordingly, the Appellate Body found:

An investigating authority may choose between the two depending on 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.^[84]

5.16.  The second sentence of Article 2.4.2 provides for a comparison methodology that is asymmetrical: the W-T comparison methodology, whereby a weighted average normal value is compared to prices of individual export transactions. This comparison methodology may be used if the following two conditions are met: first, "the authorities find a pattern of export prices which differ significantly among different purchasers, regions or time periods"; and, second, "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."^[85]

5.17.  The second sentence of Article 2.4.2 allows investigating authorities to address pricing behaviour that is focused on, or "targeted" to, purchasers, regions, or time periods by having recourse to the W-T comparison methodology. The function of the second sentence of Article 2.4.2 is, therefore, to enable investigating authorities to identify so-called "targeted dumping" and to address it appropriately.^[86] The Appellate Body has stated that "[t]his provision allows Members, in structuring their anti‑dumping investigations, to address three kinds of 'targeted' dumping, namely dumping that is targeted to certain purchasers, targeted to certain regions, or targeted to certain time periods."^[87] The Appellate Body has further suggested that the second sentence of Article 2.4.2 allows investigating authorities to "unmask targeted dumping".^[88]

5.18.  Whereas the first sentence of Article 2.4.2 provides that investigating authorities "shall normally" use the W-W or the T-T comparison methodology, the second sentence of Article 2.4.2 stipulates that a weighted average normal value "may be compared" to prices of individual export transactions, provided the two above-mentioned conditions are met. In particular, the requirement in the second sentence of Article 2.4.2 that an explanation be provided contemplates that there may be circumstances in which an investigating authority identifies a "pattern of export prices which differ significantly among different purchasers, regions or time periods", but where "such differences" could be taken into account appropriately by the W-W or T-T comparison methodology.^[89] It follows that the W-T comparison methodology is an "exception" to the comparison methodologies in the first sentence that are normally to be used. In US ‒ Zeroing (Japan), the Appellate Body stated that "[t]he asymmetrical methodology in the second sentence is clearly an exception to the comparison methodologies which normally are to be used."^[90] In the same vein, in US – Softwood Lumber V (Article 21.5 – Canada), the Appellate Body stated that the W-T comparison methodology "may be used only in exceptional circumstances" and that it is "an exception".^[91]

5.19.  We start our analysis with the United States' claims that the Panel erred in its interpretation of the relevant "pattern" and in finding that the DPM does not properly establish a "pattern", before addressing the United States' claims pertaining to the scope of application of the W-T comparison methodology. We then turn to Korea's claims concerning the identification of a pattern of prices which differ "significantly" and its claims regarding the explanation to be provided under the second sentence of Article 2.4.2. Finally, we examine Korea's claims regarding the use of "systemic disregarding" in the context of the DPM and then the United States' claims regarding the use of zeroing in the application of the W-T comparison methodology.

5.1.3  The relevant "pattern" for the purposes of the second sentence of Article 2.4.2 of the Anti‑Dumping Agreement

5.20.  The first condition set forth in the second sentence of Article 2.4.2 of the Anti‑Dumping Agreement is that the investigating authority identify "a pattern of export prices which differ significantly among different purchasers, regions or time periods". The Panel's reasoning and conclusions regarding the relevant "pattern" are interspersed throughout its Report. First, the Panel agreed with the parties that the term "pattern" refers to a "regular and intelligible form or sequence discernible in certain actions or situations" and that random price variation does not constitute a pattern.^[92] On this basis, the Panel found that, "[i]f particular prices are observed to differ in respect of a particular purchaser, region or time period, those prices may be treated as a regular and intelligible form or sequence relating to that purchaser, region or time period."^[93] The Panel further considered that the relevant "pattern" is composed of a subset of export transactions set aside for specific consideration in the second sentence of Article 2.4.2.^[94] The Panel clarified that, if particular prices are observed to differ by purchaser, region, or time period, those prices may be treated as a "pattern". The Panel stated that, although those prices are identified by reference to other prices pertaining to other purchasers, regions, or time periods, those other prices are not part of the relevant "pattern".^[95] The Panel did not specify, however, whether the subset of export transactions set aside for specific consideration necessarily comprises export prices which differ significantly because they are significantly lower than other export prices or whether a pattern could comprise prices which differ significantly because they are significantly higher than other export prices.

5.21.  Moreover, in the specific context of the DPM, which seeks to identify prices that differ significantly because they are higher or lower than other export prices, the Panel considered that "prices that are too high and prices that are too low do not belong to the same pattern".^[96] In light of the use of the words "or" and "among" in the phrase "among different purchasers, regions or time periods" in the second sentence of Article 2.4.2, the Panel also found that a pattern cannot be found to exist across purchasers, regions, and time periods, "cumulatively".^[97] The Panel considered that a pattern of prices which differ significantly among different purchasers must be found in "the price variation within a group of purchasers, as between one or more particular purchasers in relation to all other purchasers of the same group" (with the same being true for a pattern of prices which differ significantly among different regions or time periods).^[98]

5.22.  On appeal, the United States claims that the Panel erred in its interpretation of the relevant "pattern" under the second sentence of Article 2.4.2 because it concluded that "the relevant 'pattern' … comprises only low-priced export transactions to each particular 'target' (be that a purchaser, or a region, or a time period), while other higher-priced export transactions to other purchasers, regions, or time periods are 'non-pattern' transactions."^[99] In particular, the United States submits that a pattern includes both lower and higher export prices that "'differ significantly' from each other".^[100] In the context of the DPM, the United States also argues that the focus does not need to be on export sales that are priced lower than other export sales; rather, a pattern may also be identified by considering prices that are higher than other export prices.^[101] At the oral hearing, the United States further clarified that identifying a pattern does not require an assessment of how export prices relate to normal value.^[102] Moreover, the United States submits that the relevant "pattern" is one that would transcend multiple purchasers, regions, or time periods.^[103] According to the United States, the phrase "among different purchasers, regions or time periods" in the second sentence of Article 2.4.2 is to be construed as allowing an investigating authority to find a pattern of export prices which differ significantly among different purchasers, different regions, or different time periods, or any combination of these categories.^[104]

5.23.  For its part, Korea argues that the pattern is the group of prices that stand out in some discernible way from the other prices, but does not exclude the possibility that a pattern could comprise export prices that differ significantly because they are higher than other export prices.^[105] At the oral hearing, Korea also agreed with the United States that export prices are not benchmarked against normal value to identify a pattern in accordance with the second sentence of Article 2.4.2.^[106] Regarding the issue of whether a pattern can transcend multiple purchasers, regions, or time periods, Korea submits that the pattern focuses on purchasers, regions, and time periods, which Korea sees as three "fundamentally independent and distinct" categories.^[107]

5.24.  Turning to our analysis, we observe that, according to the second sentence of Article 2.4.2, the investigating authority is required to "find a pattern of export prices which differ significantly among different purchasers, regions or time periods". Assuming the requisite explanation under the second sentence of Article 2.4.2 is provided by the investigating authority, the identification of the pattern is the trigger for the application of the W-T comparison methodology.

5.25.  The word "pattern" is not explicitly defined in the text of the Anti‑Dumping Agreement. We agree with the Panel that, in the context of the second sentence of Article 2.4.2, a "pattern" can be defined as "[a] regular and intelligible form or sequence discernible in certain actions or situations".^[108] As the United States notes^[109], this definition of the word "pattern" is frequently used in conjunction with the word "of", such as is the case in the second sentence of Article 2.4.2 where the reference is to a "pattern of export prices".^[110] Moreover, this definition accords with the French and Spanish versions of the Anti‑Dumping Agreement. The French version refers to the term "configuration", which can be defined as a "[f]orme extérieure d'un ensemble; relief" (the exterior shape of a system; relief); and the Spanish version refers to the term "pauta", which can be defined as an "[i]nstrumento o norma que sirve para gobernarse en la ejecución de algo" (an instrument or norm that serves to govern the execution of something).^[111] Understanding the word "pattern" as a regular and intelligible form means that there must be regularity to the sequence of "export prices which differ significantly" and this sequence must be capable of being understood.^[112] In particular, the word "intelligible" excludes the possibility of a pattern merely reflecting random price variation, something that is not challenged on appeal.^[113]

5.26.  The relevant "pattern" in the second sentence of Article 2.4.2 is one of export prices which differ significantly. The verb "differ", which can be defined as "[t]o have contrary or diverse bearings, tendencies, or qualities; to be not the same; to be unlike, distinct, or various, in nature, form, or qualities, or in some specified respect"^[114], expresses a relative concept. As such, prices that are found to differ necessarily differ from other prices. However, by its terms, the second sentence of Article 2.4.2 refers to a pattern of "prices which differ". This wording indicates that the focus in the pattern is on the prices that are found to differ, not on all prices. Therefore, whereas an investigating authority would analyse the prices of all export sales made by the relevant exporter or producer to identify a pattern^[115], the distinguishing factor that allows that authority to discern which export prices form part of the pattern would be that the prices in the pattern differ significantly from the prices not in the pattern.

5.27.  Our interpretation accords with the ordinary meaning of the word "pattern" as used in the context of the second sentence of Article 2.4.2. As explained, the pattern is a regular and intelligible form or sequence of "export prices which differ significantly", which means that there must be regularity to the sequence of prices and that this sequence must be capable of being understood. We consider that a pattern that would comprise both the prices found to differ significantly from other prices and those other prices (namely, a pattern comprising all the transactions to all purchasers, in all regions, and in all time periods) would effectively be composed of prices that do not form a regular and intelligible sequence.

5.28.  In addition, an interpretation of the term "pattern" as comprising only those prices which differ significantly from other prices gives meaning and effect to the second sentence of Article 2.4.2, whose function is to allow investigating authorities to identify and address "targeted dumping". Indeed, by comprising only the transactions found to differ from other transactions, the pattern focuses on the "targeted" transactions. This is also consistent with the Appellate Body's statements in US – Zeroing (Japan) that "[t]he prices of transactions that fall within [the] pattern must be found to differ significantly from other export prices" and that "[t]his universe of export transactions would necessarily be more limited than the universe of export transactions to which the symmetrical comparison methodologies in the first sentence of Article 2.4.2 would apply."^[116]

5.29.  We thus agree with the Panel that, under the second sentence of Article 2.4.2, "a sub-set of export transactions is set aside for specific consideration."^[117] We further agree with the Panel that, once prices are identified as being different from other prices, "they constitute the relevant 'pattern'" and that, "[a]lthough those prices are identified by reference to other prices pertaining to other purchasers, regions or time periods, those other prices are not part of the relevant 'pattern'."^[118] The text of the second sentence of Article 2.4.2 does not expressly specify whether the prices need to differ significantly because they are lower than other prices, or whether they may differ because they are higher than other prices. Nor does the text of the second sentence of Article 2.4.2 specify whether those prices found to differ need to be below normal value. However, the Anti‑Dumping Agreement as a whole is concerned with injurious "dumping"^[119], and Article 2.4.2 sets out the methodologies that investigating authorities may use to establish margins of dumping. Article VI:1 of the GATT 1994 and Article 2.1 of the Anti‑Dumping Agreement refer to export prices that are lower than normal value as "dumped" prices. Significantly, the function of the second sentence of Article 2.4.2 is to allow investigating authorities to identify and address "targeted dumping". Therefore, although we recognize that a pattern may be identified in a variety of factual circumstances, we consider that the relevant "pattern" for the purposes of the second sentence of Article 2.4.2 comprises prices that are significantly lower than other export prices among different purchasers, regions or time periods. We fail to see how an investigating authority could identify and address "targeted dumping" by considering significantly higher export prices. If the prices found to differ significantly are higher than other export prices, the other (lower) export prices would not "mask" the (higher) dumped prices found to form the pattern.

5.30.  Turning to the issue of whether a pattern can be found to exist across purchasers, regions, or time periods, we recall that, pursuant to the second sentence of Article 2.4.2, a pattern involves export prices which differ significantly in relation to specified sub-groups, namely, "among different purchasers, regions or time periods". As the Panel noted, these terms determine how the relevant "pattern" must be identified.^[120]

5.31.  Starting with the text of the second sentence of Article 2.4.2, we observe that, depending on the context in which it is used, the conjunction "or" can be exclusive or inclusive.^[121] We further observe that, as the Panel considered, the term "among" refers to something "in relation to the rest of the group [it belongs] to".^[122] The use of this word in the second sentence of Article 2.4.2 emphasizes membership of a group, and something belongs to a group when it shares certain common characteristics with the other members of that group or has some form of relationship with them. As such, the word "among" serves to specify the dimensions in relation to which export prices which differ significantly may be discerned (i.e. purchasers, regions, or time periods). This understanding of the word "among" suggests that each category should be considered on its own, in the sense that a pattern of prices which differ significantly among different purchasers must be found in the price variation within purchasers, as between one or more particular purchasers and the other purchasers (with the same applying to regions and time periods, respectively^[123]).

5.32.  Importantly, the terms "or" and "among" in the second sentence of Article 2.4.2 draw meaning from the immediate context in which they appear. In particular, the need to identify a pattern provides contextual support for an interpretation of the terms "or" and "among" as requiring the investigating authority to consider each category (purchasers, regions, and time periods) on its own. As we have explained, the sequence of "export prices which differ significantly" must be both regular and intelligible. As such, a pattern cannot merely reflect random price variation. This means that an investigating authority is required to identify a regular series of price variations relating to one or more particular purchasers, or one or more particular regions, or one or more particular time periods to find a pattern. A single "pattern" comprising prices that are found to be significantly different from other prices across different categories would effectively be composed of prices that do not form a regular and intelligible sequence.

5.33.  Therefore, we consider that the words "or" and "among" as used in the phrase "among different purchasers, regions or time periods" cannot be interpreted to mean that the three categories can be considered cumulatively to find one single pattern. This means that some transactions that differ among purchasers, taken together with some transactions that differ among regions, and some transactions that differ among time periods, cannot form a single pattern. Accordingly, "a pattern" has to be identified among different purchasers, or among different regions, or among different time periods, and cannot transcend these categories. In EC ‒ Bed Linen, the Appellate Body also understood the three categories to work independently from one another. In that case, the Appellate Body noted that there are "three kinds of 'targeted' dumping, namely dumping that is targeted to certain purchasers, targeted to certain regions, or targeted to certain time periods".^[124] As the Panel correctly observed, the Appellate Body did not identify any other types of "targeted dumping".^[125]

5.34.  Finally, we note the United States' argument that the word "among" is used once in the second sentence of Article 2.4.2 and is not repeated before each category, which would suggest that those categories may be considered collectively in identifying a pattern. According to the United States, for the Panel's reading of the word "among" to be correct, one would expect that word to appear before the mention of each category, i.e. "among different purchases [sic], among different regions or among different time periods".^[126] We consider, however, that this repetition would have conveyed an identical meaning to that of the existing text. We thus agree with the Panel that, "[i]f particular prices are observed to differ in respect of a particular purchaser, region or time period, those prices may be treated as a regular and intelligible form or sequence relating to that purchaser, region or time period" and that "[t]he price differences are 'regular' and 'intelligible' because they pertain only to that particular purchaser, region or time period."^[127] We further agree with the Panel that "a 'pattern' can only be found in prices that differ significantly either among purchasers, or among regions, or among time periods."^[128]

5.35.  Consequently, in order to find a pattern, the export prices to one or more particular purchasers must differ significantly from the prices to the other purchasers, or the export prices in one or more particular regions must differ significantly from the prices in the other regions, or the export prices during one or more particular time periods must differ significantly from the prices during the other time periods. Our interpretation does not exclude the possibility that the same exporter or producer could be practicing more than one of the three types of "targeted dumping". We also do not exclude the possibility that a pattern of significantly differing prices to a certain category may overlap with a pattern of significantly differing prices to another category. For instance, the same transactions could "target" certain purchasers in certain regions, in which case the investigating authority might find that a pattern of significantly differing prices among different purchasers and a pattern of significantly differing prices among different regions exist.

5.36.  For the reasons set out above, we consider that a "pattern" for the purposes of the second sentence of Article 2.4.2 comprises all the export prices to one or more particular purchasers which differ significantly from the export prices to the other purchasers because they are significantly lower than those other prices, or all the export prices in one or more particular regions which differ significantly from the export prices in the other regions because they are significantly lower than those other prices, or all the export prices during one or more particular time periods which differ significantly from the export prices during the other time periods because they are significantly lower than those other prices. For the purposes of this Report, we refer to these transactions forming the relevant "pattern" as "pattern transactions".

5.37.  Consequently, we uphold the Panel's conclusions regarding the relevant "pattern" set out in, inter alia, paragraphs 7.24, 7.27-7.28, 7.45-7.46, 7.141-7.142, and 7.144 of its Report.

5.1.4  Whether the Panel erred in finding that "the DPM is inconsistent 'as such' with the second sentence of Article 2.4.2 [of the Anti-Dumping Agreement] because, by aggregating random and unrelated price variations, it does not properly establish 'a pattern of export prices which differ significantly among different purchasers, regions or time periods'"

5.38.  Having set out the interpretation of the relevant "pattern" for the purposes of the second sentence of Article 2.4.2 of the Anti-Dumping Agreement, we now turn to the claim raised by the United States that pertains to the identification of a pattern under the DPM. We recall that the United States requests us to reverse the Panel's finding that the DPM "is inconsistent 'as such' with the second sentence of Article 2.4.2 because, by aggregating random and unrelated price variations, it does not properly establish 'a pattern of export prices which differ significantly among different purchasers, regions or time periods'."^[129]

5.39.  As the Panel observed, the DPM examines any given export transaction in three different ways (by purchaser, region, and time period) in order to identify six possible types of price variation that the USDOC considers to pass the Cohen's d test: (i) prices that are "too high" to a particular purchaser; (ii) prices that are "too low" to a particular purchaser; (iii) prices that are "too high" in a particular region; (iv) prices that are "too low" in a particular region; (v) prices that are "too high" during a particular time period; and (vi) prices that are "too low" during a particular time period.^[130] The Panel noted that the USDOC then aggregates the value of these six different types of price variation to determine whether a pattern exists. As we have set out above, the Panel found that a pattern can only be found in prices that differ significantly either among purchasers, or among regions, or among time periods – not across these categories "cumulatively" – and that "prices that are too high and prices that are too low do not belong to the same pattern".^[131] With these considerations in mind, the Panel concluded that the DPM is inconsistent "as such" with the second sentence of Article 2.4.2 because it aggregates random and unrelated price variations and thus does not properly establish "a pattern of export prices which differ significantly among purchasers, regions or time periods".^[132]

5.40.  The United States acknowledges that the USDOC aggregates the results of the Cohen's d test. The USDOC does so without double-counting those export sales that pass the Cohen's d test for more than one category.^[133] According to the United States, however, there is no aggregation of random and unrelated price variations. Rather, the results of the Cohen's d test represent different aspects of the exporter's overall pricing behaviour.^[134] The United States also argues that the DPM does, in fact, consider export prices relative to other prices in the same category (i.e. by purchasers, regions, or time periods).^[135]

5.41.  The fact that the DPM starts by considering export prices relative to other prices in the same category before aggregating the results of the Cohen's d test is not dispositive of whether the DPM properly identifies a pattern for the purposes of the second sentence of Article 2.4.2. We have found above that a pattern can only be found in prices which differ significantly either among purchasers, or among regions, or among time periods, not across these categories. We have also found that the relevant "pattern" for the purposes of the second sentence of Article 2.4.2 is comprised of the export prices to one or more particular purchasers which differ significantly from the prices to the other purchasers because they are lower than those other prices (with the same applying to regions and time periods, respectively).

5.42.  It is undisputed between the participants that the DPM aggregates prices found to differ among different purchasers, among different regions, and among different time periods for the purposes of identifying a single pattern. As the Panel correctly considered, the DPM "effectively identifies a 'pattern' of export prices across different categories (purchasers, regions or time periods), rather than 'among' the constituents of each category".^[136] It is equally undisputed that the DPM aggregates prices that are higher and lower than other export prices within a given category. As we have just set out, finding a pattern across the three categories is inconsistent with the second sentence of Article 2.4.2. Moreover, a proper interpretation of the relevant "pattern" for the purposes of the second sentence of Article 2.4.2 means that it cannot be identified by considering prices that are higher than other prices.

5.43.  Consequently, we uphold the Panel's finding, in paragraph 8.1.a.ix of its Report^[137], that "the DPM is inconsistent 'as such' with the second sentence of Article 2.4.2 [of the Anti-Dumping Agreement] because, by aggregating random and unrelated price variations, it does not properly establish 'a pattern of export prices which differ significantly among different purchasers, regions or time periods'".

5.1.5  Whether the Panel erred in finding that the W-T comparison methodology should only be applied to "pattern transactions"

5.44.  We turn now to consider the United States' appeal of the Panel's findings regarding the scope of application of the W-T comparison methodology. The issue raised by the United States on appeal is whether the W-T comparison methodology can be applied to all transactions or whether it should only be applied to those transactions that form the relevant "pattern". In the Washers anti‑dumping investigation, only certain transactions passed the standard deviation and gap tests under the Nails II methodology. The USDOC nonetheless applied the W-T comparison methodology to all of Samsung's and LGE's export transactions.^[138] For its part, the DPM applies the W-T comparison methodology to all export transactions in certain situations, namely, when the value of the sales that pass the Cohen's d test accounts for 66% or more of the value of the total sales.^[139]

5.45.  The Panel found that the word "individual" in the phrase "individual export transactions" in the first part of the second sentence of Article 2.4.2 of the Anti‑Dumping Agreement indicates that the W-T comparison will not involve all export transactions, but only certain transactions identified individually.^[140] The Panel further found that the only textual basis for individual identification of export transactions is that they form the pattern under the second sentence of Article 2.4.2. The Panel found support for its interpretation in the requirement that lies on investigating authorities to explain why "'such differences' cannot be taken into account appropriately by one of the symmetrical comparison methodologies."^[141] The Panel considered that such explanation refers back to the "pattern transactions" and does so precisely because only those "pattern transactions" should be subject to a W-T comparison. The Panel also relied on the exceptional nature of the W-T comparison methodology, which suggests that "something different from the first sentence should be undertaken, i.e. assessment of only the pattern transactions set aside for specific consideration."^[142] Finally, the Panel relied on the object and purpose of the second sentence of Article 2.4.2 to enable investigating authorities to "'unmask' so-called 'targeted dumping'"^[143], as well as on the Appellate Body report in US – Zeroing (Japan), where the Appellate Body read "'individual export transactions' … as referring to the transactions that fall within the relevant pricing pattern".^[144]

5.46.  Accordingly, the Panel found that the W-T comparison methodology should only be applied to those transactions that constitute the "pattern of export prices which differ significantly among different purchasers, regions or time periods".^[145] Consequently, the Panel found that the United States acted inconsistently with the second sentence of Article 2.4.2 in the Washers anti‑dumping investigation because the USDOC applied the W-T comparison methodology to all export transactions, including transactions other than those constituting the patterns of transactions that the USDOC had determined to exist.^[146] The Panel also found that the DPM is inconsistent "as such" with this provision because it applies the W-T comparison methodology to all export transactions, where the value of the sales that pass the Cohen's d test accounts for 66% or more of the value of the total sales.^[147]

5.47.  The United States claims that there is no textual and contextual support for the Panel's finding that the W-T comparison methodology should only be applied to the transactions constituting the pattern. In particular, the United States contends that the relevant definition of the word "individual" is "single; separate" and that this word merely suggests that prices of single, separate export transactions may be compared to a normal value established on a weighted average basis, not that the scope of application of the W-T comparison methodology is limited to certain transactions.^[148] In this context, the United States also takes issue with the fact that the Panel assumed that the application of the W-T comparison methodology is limited to "pattern transactions" because the second sentence of Article 2.4.2 provides no other basis for identifying "individual" export transactions.^[149] The United States further submits that the Panel misapplied Article 31 of the Vienna Convention on the Law of Treaties^[150] (Vienna Convention) by relying on the object and purpose of the second sentence of Article 2.4.2, rather than that of the Anti‑Dumping Agreement itself.^[151] Finally, the United States argues that the Panel's reliance on the Appellate Body's statements in US – Zeroing (Japan) was "misplaced" and that these statements do not support the Panel's interpretation.^[152]

5.48.  Korea submits that the United States reads the term "individual" more narrowly than the Appellate Body did in US – Zeroing (Japan)^[153] and unduly focuses on specific words without considering the second sentence of Article 2.4.2 as a whole.^[154] In particular, Korea is of the view that the textual reference to "such differences" in the context of the explanation to be provided under the second sentence suggests that this provision creates a group of sales that meet the conditions for the exception, and to which the W-T comparison methodology may be applied, and another group of sales that do not.^[155] Korea further argues that the United States' interpretation of the second sentence of Article 2.4.2 would lead to absurd results, in that an authority would be allowed to apply an exceptional comparison methodology to all export transactions once it has found a pattern to exist, irrespective of how few transactions constitute this pattern.^[156]

5.49.  Starting with our analysis of the text of the second sentence of Article 2.4.2, we recall that, pursuant to this provision, "[a] normal value established on a weighted average basis may be compared to prices of individual export transactions". By its express terms, the second sentence of Article 2.4.2 refers to a comparison to the prices of individual export transactions. The word "individual" can be defined as "[e]xisting as a separate indivisible entity; numerically one; single, as distinct from others of the same kind; particular"^[157]; or, as argued by the United States, as "single; separate".^[158] This definition alone does not clarify whether the W-T comparison methodology can be applied to all transactions or whether it should only be applied to certain transactions, namely, the "pattern transactions". The term "individual export transactions" in the second sentence of Article 2.4.2 draws meaning from the immediate context in which it appears.^[159]

5.50.  The second sentence of Article 2.4.2 requires, inter alia, that the investigating authority provide "an explanation … 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."^[160] The term "such differences" refers back to the "export prices which differ significantly" and, therefore, form part of the pattern. As set out above, the requirement in the second sentence of Article 2.4.2 that an explanation be provided contemplates that there may be circumstances in which an investigating authority identifies a "pattern of export prices which differ significantly among different purchasers, regions or time periods" but where "such differences" could be taken into account appropriately by the W-W or T-T comparison methodology.^[161] It follows that recourse to the W-T comparison methodology is permissible only to the extent that it is necessary to remedy the inability of the normally applicable comparison methodologies to take into account appropriately the identified "pattern".

5.51.  Furthermore, the first sentence of Article 2.4.2 provides for two comparison methodologies that "shall normally" be used by investigating authorities to establish margins of dumping. Under the first sentence of Article 2.4.2, the consideration is in respect of all export transactions. By contrast, the emphasis in the second sentence of Article 2.4.2 is on a "pattern". Both conditions in the second sentence of Article 2.4.2, namely, that a pattern of significant price differences be established and that an explanation be provided, pertain to "pattern transactions". Moreover, the W-T comparison methodology is considered to be an exception to the symmetrical comparison methodologies in the first sentence. The structure of Article 2.4.2, which distinguishes the normally applicable methodologies from the exceptional W-T comparison methodology, thus serves as further indication that the W-T comparison methodology should only be applied to those transactions justifying its use, namely, the "pattern transactions". We, therefore, agree with the Panel that "[t]he exceptional nature of this comparison methodology suggests that something different from the first sentence should be undertaken, i.e. assessment of only the pattern transactions set aside for specific consideration."^[162]

5.52.  For the reasons set out above, we agree with the Panel that: (i) the use of the word "individual" in the second sentence of Article 2.4.2 indicates that the W-T comparison methodology does not involve all export transactions, but only certain export transactions identified individually; and (ii) the "individual export transactions" to which the W-T comparison methodology may be applied are those transactions falling within the relevant "pattern".^[163] Accordingly, we read the phrase "individual export transactions" as referring to the universe of export transactions that justify the use of the W-T comparison methodology, namely, the "pattern transactions". Our interpretation gives meaning and effect to the second sentence of Article 2.4.2, whose function is to allow investigating authorities to identify and address "targeted dumping". It also accords with the object and purpose of the Anti‑Dumping Agreement. Although the Anti-Dumping Agreement does not contain a preamble expressly setting out its object and purpose, it is apparent from the text of this Agreement that it deals with injurious dumping by allowing Members to take anti‑dumping measures to counteract injurious dumping and imposing disciplines on the use of such anti-dumping measures.^[164]

5.53.  Applying the W-T comparison methodology to "pattern transactions" only is also in line with the Appellate Body's observations in US ‒ Zeroing (Japan), where the Appellate Body "[read] the phrase 'individual export transactions' … as referring to the transactions that fall within the relevant pricing pattern", and considered that "[t]his universe of export transactions would necessarily be more limited than the universe of export transactions to which the symmetrical comparison methodologies in the first sentence of Article 2.4.2 would apply."^[165] The Appellate Body added that, "[i]n order to unmask targeted dumping, an investigating authority may limit the application of the W-T comparison methodology to the prices of export transactions falling within the relevant pattern."^[166]

5.54.  We note that the United States makes a number of arguments that take issue with the Panel's reliance, in paragraphs 7.25 and 7.27 of its Report, on the Appellate Body's statements in US ‒ Zeroing (Japan).^[167] First, the United States argues that this case did not involve the application of the second sentence of Article 2.4.2. While this is correct, the Appellate Body was nonetheless addressing legal issues raised on appeal, as the panel had drawn contextual support from the second sentence of Article 2.4.2 for its finding that zeroing is permitted under the T‑T comparison methodology. Second, the United States asserts that the Appellate Body merely suggested that an investigating authority may limit the application of the W-T comparison methodology. While the Appellate Body's use of the word "may" could, if read in isolation, be perceived as being ambiguous, the remaining parts of the relevant paragraph in its report make clear that the Appellate Body excluded the possibility that the W-T comparison methodology might apply to "non‑pattern transactions".^[168] Finally, we disagree with the United States that the Panel's understanding of the scope of application of the W-T comparison methodology was derived exclusively from that Appellate Body report. The Panel conducted its own analysis of the second sentence of Article 2.4.2, and relied on the Appellate Body report in US – Zeroing (Japan) as "[f]urther support" for its interpretation.^[169]

5.55.  Based on the foregoing considerations, in particular in light of the function of the second sentence of Article 2.4.2 to allow investigating authorities to identify and address "targeted dumping", we consider that the W-T comparison methodology should only be applied to those transactions that justify its use, namely, those transactions forming the relevant "pattern".

5.56.  We, therefore, uphold the Panel's finding, in paragraph 7.29 of its Report, that "the W-T comparison methodology should only be applied to transactions that constitute the 'pattern of export prices which differ significantly among different purchasers, regions or time periods'." We further uphold the Panel's consequential findings: (i) in paragraph 8.1.a.i of its Report^[170], that "the United States acted inconsistently with the second sentence of Article 2.4.2 of the Anti‑Dumping Agreement, by applying the W-T comparison methodology to transactions other than those constituting the patterns of transactions that the USDOC had determined to exist in the Washers anti-dumping investigation"; and (ii) in paragraph 8.1.a.vi of its Report^[171], that "the DPM is inconsistent 'as such' with Article 2.4.2 of the Anti-Dumping Agreement, because it applies the W‑T comparison methodology to non-pattern transactions when the aggregated value of sales to purchasers, regions, and time periods that pass the Cohen's d test account[s] for 66% or more of the value of total sales".

5.1.6  The extent to which price differences are to be assessed quantitatively, qualitatively, and in light of the "reasons" for these price differences

5.57.  Korea claims that the Panel erred in finding that a "pattern of export prices which differ significantly" can be established "solely based on quantitative differences … regardless of the factual context of the prices and the differences".^[172] This claim on appeal raises the issue of whether a "pattern of export prices which differ significantly" can be established based on purely quantitative criteria, without a qualitative assessment and without considering the "reasons" for the price differences.^[173]

5.58.  According to the Panel, the text of Article 2.4.2 of the Anti-Dumping Agreement contains no requirement to consider the "reasons" for the price differences in order to identify a pattern, and an examination of the relevant numerical price values suffices.^[174] The Panel further found that "an authority might properly find that certain prices differ 'significantly' if those prices are notably greater – in purely numerical terms – than other prices, irrespective of the reasons for those differences."^[175] As the Panel noted, those "reasons" could, however, be relevant in the context of the explanation to be provided under the second sentence of Article 2.4.2.^[176] The Panel acknowledged that, in certain factual circumstances, the size or scale of a price difference may need to be assessed in light of the prevailing factual circumstances.^[177] For example, a relatively minor numerical difference between two large prices may not be significant, whereas the same numerical difference between two much smaller prices may be significant. The Panel, however, found that this relates to how, not why, the relevant prices differ.^[178] Accordingly, the Panel rejected Korea's claim that the United States acted inconsistently with the second sentence of Article 2.4.2 in the Washers anti‑dumping investigation because the USDOC determined the existence of a pattern of export prices which differ significantly on the basis of purely quantitative criteria, without any qualitative assessment of the "reasons" for the relevant price differences.^[179] The Panel also found that the DPM is not inconsistent "as such" with this provision because it determines the existence of a pattern of export prices which differ significantly on the basis of purely quantitative criteria, without any qualitative assessment of the "reasons" for the relevant price differences.^[180]

5.59.  Korea claims that the Panel mischaracterized its claim as if it were solely that the authority must state the reasons why prices differ.^[181] Korea argues that, while it discussed the factual context of the Washers anti-dumping investigation in terms of the "reasons" why prices may be differing before the Panel, this was not its sole or even principal argument.^[182] Moreover, Korea requests us to reverse the Panel's findings that the authorities need not consider qualitative factors as part of making a proper finding of export prices that "differ significantly" and constitute a "pattern".^[183] Specifically, Korea argues that the Panel erred in finding that the authority is not required to examine the "reasons" for the price differences to establish the existence of a pattern.^[184] Korea recalls that the term "significantly" entails both qualitative and quantitative aspects. Yet, while acknowledging this, in Korea's view, the Panel effectively held that the term "significantly" can be analysed in purely quantitative terms.^[185] According to Korea, qualitative considerations are relevant both as part of determining whether price differences are "significant" and form a "pattern" and as part of determining whether a W-W comparison can "take into account appropriately" the significant price differences.^[186]

5.60.  The United States, in turn, is of the view that Korea attempts to expand its claims beyond what is set forth in its panel request.^[187] Moreover, while the United States acknowledges that the term "significantly" has both quantitative and qualitative aspects^[188], according to the United States, this does not mean that the authority must examine the "reasons" for the price differences, and a pattern may be discerned through a simple examination of the relevant numerical price values.^[189] At the oral hearing, the United States clarified that, in its view, there may be instances where a purely quantitative analysis suffices to find that price differences are significant.^[190]

5.61.  Addressing first whether the Panel mischaracterized Korea's claim, we note that Korea's request for the establishment of a panel refers to "commercial reasons", "market explanations", and "economic or market factors" that the USDOC allegedly fails to consider.^[191] In our view, these references show that Korea claimed that qualitative aspects should be considered. Even if these references should be understood to go to the more narrow issue of why differences might exist between export prices, Korea's panel request is otherwise broadly drafted. The "commercial reasons", "market explanations", and "economic or market factors" that the USDOC allegedly fails to consider appear to be mere examples of how, according to Korea, the USDOC fails to identify a pattern in accordance with the second sentence of Article 2.4.2^[192], and as such are in the nature of arguments rather than claims.^[193] The Panel rejected Korea's claims that the USDOC acted inconsistently with the second sentence of Article 2.4.2 in the Washers anti‑dumping investigation and that the DPM is inconsistent "as such" with this provision "by determining the existence of a 'pattern of export prices which differ significantly' among purchasers, regions or time periods on the basis of purely quantitative criteria, without any qualitative assessment of the reasons for the relevant price differences".^[194] We thus understand the Panel to have concluded that both the Nails II methodology and the DPM involve a purely quantitative analysis, which the Panel did not find to be inconsistent with the second sentence of Article 2.4.2. As such, the Panel sufficiently addressed Korea's claims and arguments, and not only whether the USDOC would have needed to consider the "reasons" for the price differences. Therefore, we disagree with Korea that the Panel mischaracterized its claim.

5.62.   Turning to the issue of whether the Panel erred in finding that a pattern of export prices which differ significantly can be established on the basis of purely quantitative criteria, without any qualitative assessment of the reasons for the relevant price differences, we recall that, pursuant to the second sentence of Article 2.4.2, the relevant "pattern" is one of export prices which differ significantly. The term "differ" is thus qualified by the term "significantly". As the Panel correctly recognized^[195], "significant" can be defined as "important, notable or consequential".^[196] We thus understand the word "significantly" to speak to the extent of the price differences and to suggest something that is more than just a nominal or marginal difference in prices. Under the second sentence of Article 2.4.2, the something that must be important, notable, or consequential is the difference in export prices.

5.63.  Furthermore, the term "significantly" has both quantitative and qualitative dimensions.^[197] Accordingly, assessing the extent of the differences in export prices to establish whether those export prices differ significantly for the purposes of the second sentence of Article 2.4.2 entails both quantitative and qualitative dimensions. As part of the qualitative assessment, circumstances pertaining to the nature of the product or the markets may be relevant for the assessment of whether differences are "significant" in the circumstances of a particular case. The significance of differences may indeed be affected by objective market factors, such as the nature of the product under consideration, the industry at issue, the market structure, or the intensity of competition in the markets at issue, depending on the case at hand. Hence, what may be deemed "significant" price differences in one instance may fail to meet the same threshold when different variables are considered. For example, the Panel observed that, in a more price-competitive market, smaller differences may be significant.^[198] Unless the investigating authority considers such qualitative aspects, it will not know if and how these aspects are relevant to its assessment of whether prices differ significantly. Therefore, we disagree with the Panel to the extent it considered that an investigating authority may properly find that certain prices differ significantly within the meaning of the second sentence of Article 2.4.2 if they are notably greater in purely numerical terms.^[199]

5.64.  Our understanding of the term "significant" accords with previous Appellate Body reports. For example, in the context of Article 6.3(c) of the SCM Agreement^[200], in US – Large Civil Aircraft (2^nd complaint), the Appellate Body found that "an assessment of whether a lost sale is significant can have both quantitative and qualitative dimensions."^[201] Regarding the qualitative dimension, the Appellate Body referred to the "highly price-competitive" nature of the market.^[202] We also note that the panel in US – Upland Cotton, in considering a case of significant price suppression under the same provision, found that "it may be relevant to look at the degree of the price suppression … in the context of the prices that have been affected" to assess whether the price suppression is significant.^[203] As the panel reasoned:

The "significance" of any degree of price suppression may vary from case to case, depending upon the factual circumstances, and may not solely depend upon a given level of numeric significance. Other considerations, including the nature of the "same market" and the product under consideration may also enter into such an assessment, as appropriate in a given case.^[204] 

5.65.  The words "significantly" and "pattern" in the second sentence of Article 2.4.2, however, do not imply an examination into the cause of (or reasons for) the differences in prices. The second sentence of Article 2.4.2 requires an investigating authority to find "a pattern of export prices which differ significantly among different purchasers, regions or time periods". The text does not impose an additional requirement to ascertain whether the significant differences found to exist are unconnected with "targeted dumping". As the Panel correctly observed, in US – Large Civil Aircraft (2^nd complaint), there was no suggestion by the Appellate Body that the qualitative dimension of the significance of lost sales extends to consideration of the cause of (or reasons for) those lost sales.^[205] Similarly, the Panel correctly observed that the US – Upland Cotton panel did not refer to the underlying cause of (or reasons for) price suppression as being relevant to the potential significance of the degree of price suppression.^[206] The text of the second sentence of Article 2.4.2 also does not imply an examination of the motivation for, or intent behind, the differences in prices. We thus see merit in the United States' argument that, under the second sentence of Article 2.4.2, the investigating authority is charged with finding whether a pattern of export prices exists, not whether an exporter or producer has intentionally patterned its export prices to "target" and "mask" dumping.^[207]

5.66.  Based on the foregoing, we find that the requirement to identify prices which differ significantly means that the investigating authority is required to assess quantitatively and qualitatively the price differences at issue. This assessment may require the investigating authority to consider certain objective market factors, such as circumstances regarding the nature of the product under consideration, the industry at issue, the market structure, or the intensity of competition in the markets at issue, depending on the case at hand. The investigating authority is, however, not required to consider the cause of (or reasons for) the price differences. Therefore, we agree with the Panel that an investigating authority is not required to consider the cause of (or reasons for) the price differences to establish the existence of a pattern under the second sentence of Article 2.4.2 of the Anti-Dumping Agreement. However, we reverse the Panel's finding in respect of the Washers anti-dumping investigation, in paragraph 8.1.a.ii of its Report^[208], to the extent that the Panel found that "a pattern of export prices which differ significantly among purchasers, regions or time periods" can be established "on the basis of purely quantitative criteria". We also reverse the Panel's finding in respect of the DPM, in paragraph 8.1.a.v of its Report^[209], to the extent that the Panel found that "a pattern of export prices which differ significantly among purchasers, regions or time periods" can be established "on the basis of purely quantitative criteria".

5.1.7  Whether an explanation needs to be provided with respect to both the W-W and the T-T comparison methodologies

5.67.  We turn now to consider Korea's claims regarding the second condition set forth in the second sentence of Article 2.4.2 of the Anti‑Dumping Agreement, pursuant to which the investigating authority is to provide "an explanation … 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 issue raised by Korea on appeal is whether this provision requires that an explanation be provided with respect to either the W-W or the T-T comparison methodology, or with respect to both of these two methodologies when the application of the W‑T comparison methodology is considered under the second sentence of Article 2.4.2. We recall that the USDOC's explanation in the Washers anti‑dumping investigation focused on the W-W comparison methodology and did not make reference to the T-T comparison methodology.^[210] Nor does the DPM require the USDOC to provide an explanation with respect to the T-T comparison methodology.^[211]

5.68.  The Panel observed that, under the second sentence of Article 2.4.2, an explanation is required of why "a" W-W "or" T-T "comparison" cannot take into account appropriately the relevant price differences. The Panel considered that the use of the indefinite article "a", combined with the disjunctive "or" and the term "comparison" in the singular, suggests that the requisite explanation need only be provided in respect of one type of comparison (W-W or T-T), not both.^[212] Furthermore, the Panel relied on the Appellate Body's finding in US ‒ Softwood Lumber V (Article 21.5 – Canada) that "[a]n investigating authority may choose between the two [comparison methodologies in the first sentence of Article 2.4.2] depending on which is most suitable for the particular investigation."^[213] The Panel considered that this choice under the first sentence of Article 2.4.2 would likely be made before the application of the second sentence is considered. The Panel further considered that, if an authority were to opt for the W-W comparison methodology to avoid an overly burdensome comparison process, it would seem anomalous for that authority to have to incur the burden of reverting to the T-T comparison methodology in the context of the second sentence of Article 2.4.2 before applying the W-T comparison methodology. The Panel added that requiring an explanation with respect to both the W-W and the T-T comparison methodologies would undermine the investigating authority's "initial discretion" to choose between the W-W and T-T comparison methodologies.^[214]

5.69.  Accordingly, the Panel found that Korea failed to establish that the United States acted inconsistently with the second sentence of Article 2.4.2 in the Washers anti‑dumping investigation because the USDOC failed to explain why the relevant price differences could not be taken into account appropriately by the T-T comparison methodology.^[215] The Panel also found that Korea failed to establish that the DPM is inconsistent with this provision because it does not consider whether the relevant price differences can be taken into account appropriately by the T-T comparison methodology.^[216]

5.70.  Korea advances a series of arguments that take issue with these findings. Korea submits that the word "a" in the phrase "the use of a weighted average-to-weighted average or transaction-to-transaction comparison" in the second sentence of Article 2.4.2 reflects the fact that the investigating authority will be using "a" single comparison methodology (rather than both at the same time) and that the word "or" reflects this choice between the W-W and T-T comparison methodologies.^[217] In addition, Korea disagrees with the Panel that providing an explanation with respect to both methodologies would be burdensome. According to Korea, given that the USDOC uses the T-T comparison methodology in "unusual situations", it could easily explain why this methodology would not work in a particular situation.^[218] Korea adds that the Panel created an "artificial distinction that the authority 'would likely' have chosen a preferred method under the first sentence before turning to the second sentence", whereas the authority may consider all three options at once.^[219]

5.71.  For its part, the United States is of the view that the Panel's interpretation follows a proper application of the customary rules of interpretation of public international law, is logical, and accords with prior Appellate Body guidance concerning Article 2.4.2.^[220] In particular, the United States argues that if, as the Appellate Body has found, an authority is free to choose between the W-W and T-T comparison methodologies in the first sentence of Article 2.4.2 and those methodologies yield systematically similar results, there would be no purpose in requiring that an explanation be provided with respect to both methodologies.^[221] Furthermore, according to the United States, Article 2.4.2 describes a logical progression, in which the investigating authority first selects whether to use the W-W or T-T comparison methodology, and thereafter examines whether the W-T comparison methodology can be applied.^[222] Finally, the United States submits that, under Korea's proposed interpretation, the investigating authority would effectively be required to explain its initial choice between the W-W and T-T comparison methodologies, something that is not required under the first sentence of Article 2.4.2.^[223]

5.72.  We recall that the second sentence of Article 2.4.2 requires that "an explanation [be] 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." Depending on the context in which it is used, the conjunction "or" can be exclusive or inclusive.^[224] We note the United States' argument that the first sentence of Article 2.4.2 provides an option between the W‑W and T-T comparison methodologies using the conjunction "or" and that the word "or" thus has the same meaning in the context of the explanation to be provided under the second sentence of Article 2.4.2.^[225] However, the mere fact that the conjunction "or" is used in the first and second sentences of Article 2.4.2 does not imply that it has the same meaning in both sentences.^[226] We also observe that, if the second sentence of Article 2.4.2 were to include the conjunction "and" instead of the conjunction "or", this would suggest that the authority is required to use the W-W and the T-T comparison methodologies in combination.^[227] Therefore, in the context of the second sentence of Article 2.4.2, using the conjunction "and" instead of the conjunction "or" was not viable to indicate that both methodologies should be addressed in the investigating authority's explanation.

5.73.  Turning to the indefinite article "a" and the singular form of the word "comparison" in this provision, we disagree with the Panel that these suggest that an explanation with regard to one of the two normally applicable methodologies comports with the second sentence of Article 2.4.2. We are particularly mindful of the fact that the equally authentic French version of the second sentence of Article 2.4.2 refers to "les méthodes de comparaison", using a definite article ("les") and "comparison methods" in the plural form.^[228]

5.74.  Furthermore, the W-T comparison methodology in the second sentence of Article 2.4.2 is an exception to the comparison methodologies that are set out in the first sentence and are normally to be used.^[229] Interpreting the second sentence of Article 2.4.2 as requiring that an explanation be provided with respect to both the W-W and the T-T comparison methodologies gives a proper recognition to the text of that provision and to the distinction between the normally applicable methodologies in the first sentence of Article 2.4.2 and the exceptional W-T comparison methodology in the second sentence. If the W-T comparison methodology were to apply in an instance where an explanation is provided with respect to one of the two normally applicable comparison methodologies, but the other could appropriately take the relevant price differences into account, the W-T comparison methodology would no longer be used as an exception. Although the W-W and T-T comparison methodologies are likely to yield substantially equivalent results, the possibility that, in a particular case, they might yield different results and might impact differently the possible use of the W-T comparison methodology, should not be entirely excluded.

5.75.  Finally, we disagree with the Panel's reasoning that the investigating authority's "initial discretion" between the W-W and T-T comparison methodologies under the first sentence of Article 2.4.2 would be undermined by requiring that an explanation be provided with respect to both these methodologies. We recall that the W-W and the T-T comparison methodologies "fulfil the same function" and that there is no "hierarchy between the two".^[230] As such, an investigating authority may choose between these two methodologies "depending on which is most suitable for the particular investigation".^[231] However, we consider that the investigating authority's option between the W-W and T-T comparison methodologies under the first sentence of Article 2.4.2 is unrelated to the question of whether these two methodologies are not appropriate to unmask "targeted dumping" such that the investigating authority contemplates the application of the W‑T comparison methodology. Requiring that an explanation be provided in respect of both the W‑W and T-T comparison methodologies, when the application of the W-T comparison methodology is considered under the second sentence of Article 2.4.2, does not mean that the investigating authority is deprived of its discretion should it decide to apply the first sentence of Article 2.4.2 instead of turning to the W-T comparison methodology in the second sentence.^[232]

5.76.  For these reasons, we consider that an investigating authority has to explain why both the W-W and the T-T comparison methodologies cannot take into account appropriately the differences in export prices that form the pattern. In circumstances where the W-W and T-T comparison methodologies would yield substantially equivalent results and where an explanation has been provided with respect to one of these two methodologies, the explanation to be included with respect to the other may not need to be as elaborate.

5.77.  Based on the foregoing, we reverse the Panel's finding, in paragraph 8.1.a.iv of its Report^[233], that "Korea failed to establish that the United States acted inconsistently with the second sentence of Article 2.4.2 [of the Anti-Dumping Agreement] in the Washers anti-dumping investigation by failing to explain why the relevant price differences could not be taken into account appropriately by the T-T comparison methodology". For the same reasons, we reverse the Panel's finding, in paragraph 8.1.a.viii of its Report^[234], that "Korea failed to establish that the DPM is inconsistent with the second sentence of Article 2.4.2 [of the Anti-Dumping Agreement] when, having concluded that the W-W comparison methodology cannot appropriately take into account the observed pattern of significantly different prices, it does not also consider whether the relevant price differences could be taken into account appropriately by the T-T comparison methodology".

5.1.8  "Systemic disregarding"

5.1.8.1  Whether the Panel erred in finding that Korea failed to establish that the United States' use of "systemic disregarding" under the DPM is inconsistent "as such" with the second sentence of Article 2.4.2 of the Anti‑Dumping Agreement

5.78.  We turn now to consider Korea's appeal of the Panel's findings under the second sentence of Article 2.4.2 of the Anti‑Dumping Agreement in respect of "systemic disregarding". We begin by recalling that the USDOC applies a "mixed" comparison methodology as part of the DPM. In cases where the value of the transactions that pass the Cohen's d test is between 33% and 66% of the value of the total export transactions, the W‑T comparison methodology with zeroing is applied to these transactions. For the remaining transactions, i.e. the transactions that do not pass the Cohen's d test, the W‑W comparison methodology is used.

5.79.  In the first administrative review of the Washers anti‑dumping order, the USDOC applied this "mixed" comparison methodology with respect to LGE.^[235] For the transactions to which the W‑W comparison methodology was applied, the USDOC took the sum total of the "evidence of dumping" (i.e. the positive comparison results) and "offset" this amount with export sales that were greater than normal value (i.e. the negative comparison results) up to the amount of the sum total of the "evidence of dumping". For the transactions to which the W‑T comparison methodology was applied, the USDOC took the sum total of the "evidence of dumping" and made no "offsets" for export sales above normal value (i.e. the USDOC used zeroing).^[236] The USDOC then combined the sum total of the comparison results of the W‑W and W‑T comparison methodologies to determine the margin of dumping for the exporter (LGE) and the product under investigation (LRWs) "as a whole". In aggregating the comparison results, the USDOC did not permit the overall negative comparison result arising from the W‑W comparison methodology to "offset" the "evidence of dumping" from the application of the W‑T comparison methodology.^[237] As we have observed above, it is in this context that the issue of "systemic disregarding" arises under the DPM.

5.80.  The Panel noted that the W‑T comparison methodology provided in the second sentence of Article 2.4.2 is an exceptional and alternative comparison methodology.^[238] The Panel considered that, when an investigating authority determines the margin of dumping for an individual exporter or foreign producer under the second sentence of Article 2.4.2, the investigating authority is entitled to have particular regard and, therefore, limit its analysis to the pricing behaviour of that exporter or foreign producer in respect of the transactions that form the "pattern". Furthermore, the Panel noted that, irrespective of the methodology applied, Articles 2.1 and 6.10 of the Anti‑Dumping Agreement require that margins of dumping be established for the product under investigation "as a whole" for the individual exporter or foreign producer concerned. Thus, the Panel reasoned that, while the numerator may be established from the "evidence of dumping" in "pattern transactions", the denominator of the equation has to reflect the value of total exports of that individual exporter or foreign producer.^[239]

5.81.  The Panel further considered that, consistent with the focus of the second sentence of Article 2.4.2 being on the pricing behaviour in respect of "pattern transactions", one could take the view that the combined application of the W‑T and W‑W (or T‑T) comparison methodologies is not envisaged by that provision. However, since Korea had not raised a claim to that effect, the Panel did not see the need to rule on this matter.^[240]

5.82.  Recalling that the second sentence of Article 2.4.2 is designed to enable an investigating authority to focus on "pattern transactions" in order to "unmask targeted dumping", the Panel found that, where an investigating authority applies the W‑T comparison methodology to "pattern transactions" and the W‑W (or T‑T) comparison methodology to "non‑pattern transactions", the effect of this approach would be to "zoom in" on the "targeted dumping" identified in respect of "pattern transactions", but subsequently to "zoom out" and "re-mask" the identified "targeted dumping", if the investigating authority had to take into account the result of the W-W (or T-T) comparison methodology applied to "non-pattern transactions" when making its overall determination of dumping.^[241] The Panel further found that such an approach would lead to "mathematical equivalence" ─ the same result that would arise from a straightforward application of the W‑W comparison methodology to all transactions.^[242] "Systemic disregarding", according to the Panel, "enables an investigating authority to reveal any dumping in respect of pattern transactions that would otherwise be masked by the negative dumping in respect of non‑pattern transactions".^[243]

5.83.  The Panel rejected Korea's argument that the use of different weighted average normal values could avoid mathematical equivalence. The Panel found that Korea had not identified any textual basis in Article 2.4.2 for concluding that the "normal value established on a weighted average basis" referred to in the second sentence should differ, within the same anti‑dumping proceeding, from the "weighted average normal value" referred to in the first sentence.^[244] Neither was the Panel persuaded by Korea's argument that mathematical equivalence could be avoided if the investigating authority undertook a "granular analysis" of the transactions involved in the W‑T comparison methodology and a detailed approach to price adjustments, i.e. by rethinking the adjustments that might be necessary to ensure price comparability.^[245] The Panel took the view that the second sentence does not envisage that any price adjustments be made in addition to those made pursuant to an investigating authority's general obligation to make a "fair comparison" under Article 2.4 of the Anti‑Dumping Agreement. According to the Panel, "there is nothing in the text of the second sentence to suggest that an authority could or should make the type of adjustments proposed by Korea in order to allow the authority to unmask targeted dumping."^[246]

5.84.  In light of the above, the Panel rejected Korea's claim that the USDOC's use of "systemic disregarding" when combining the overall comparison results arising from the W‑W and W‑T comparison methodologies under the DPM is inconsistent "as such" with the second sentence of Article 2.4.2.^[247]

5.85.  On appeal, Korea asserts that the Panel's approach essentially creates two margins of dumping, one for the subset to which the W‑T comparison methodology is applied, and another for the subset using the W‑W comparison methodology. Korea argues that, although the Panel correctly found that zeroing could not apply to either of these two subsets, it allowed the "functional equivalent of zeroing" by combining the two subsets and by refusing to allow any "offsets" from the subset based on the normal comparison methodologies when determining the overall dumping and margin of dumping.^[248] Korea argues that, apart from allowing an exception to the normal comparison methodologies, neither the second sentence of Article 2.4.2 nor the whole of Article 2.4.2 creates any exception to the basic concepts of "dumping" and "margin of dumping".^[249] Further, Korea asserts that the Panel's reasoning has no basis in the text or context of Article 2.4.2^[250] and that "'[d]umping' only exists as a final conclusion based on all export transactions overall for the exporter and for the product as a whole."^[251]

5.86.  Moreover, Korea argues that, unlike what the Panel found, "[t]he purpose of the second sentence of Article 2.4.2 … is simply to allow the authority to undertake the more careful examination of individual export prices that the W‑T comparison method[ology] makes possible."^[252] According to Korea, the purpose of the second sentence of Article 2.4.2 is not to "'unmask' so‑called 'targeted dumping'", but rather "to 'unmask' individual export prices".^[253] Korea further states that the second sentence of Article 2.4.2 is not about end results, but rather about the comparison methodology necessary to ensure that individual export prices are more carefully examined by the investigating authority in specified circumstances.^[254]

5.87.  For its part, the United States contends that, where an investigating authority applies the W‑T comparison methodology to fewer than all export prices, the Anti‑Dumping Agreement does not obligate the investigating authority to "offset" or "re‑mask" the "evidence of dumping" that has been "unmasked" through the use of the W‑T comparison.^[255] Therefore, the United States submits that Korea's argument in essence leads to an interpretation of the second sentence of Article 2.4.2 as requiring the mandatory "re‑masking" of below‑normal value export sales, which is not supported by the text of the Anti‑Dumping Agreement, and which would render the second sentence of Article 2.4.2 inutile.^[256] The United States contends that the Panel was right to interpret the second sentence as being an exception to the first sentence of Article 2.4.2, and as setting forth a special methodology for establishing margins of dumping.^[257]

5.88.  As regards the function of the second sentence of Article 2.4.2, the United States submits that Korea's position that the purpose of the second sentence is for the investigating authority to undertake a "more careful examination" means nothing if the provision requires "re‑masking" of "targeted dumping".^[258] Moreover, the United States submits that, under Korea's approach, there would never be any reason for an investigating authority to resort to the alternative, W‑T comparison methodology when the T‑T comparison methodology already provides the investigating authority with the possibility of undertaking a "granular examination of individual export prices".^[259] Thus, the United States submits that the only logical conclusion is that the second sentence of Article 2.4.2 is intended "to enable investigating authorities to 'unmask' so‑called 'targeted dumping'".^[260]

5.89.  For the purposes of our analysis and in order to address the discrete claims raised by Korea on appeal, we first recall below the relevant jurisprudence that dumping and margins of dumping are established for the product under investigation "as a whole" and for each exporter or foreign producer. Next, in light of this jurisprudence, we turn to consider the establishment of dumping and margins of dumping under the W‑T comparison methodology set forth in the second sentence of Article 2.4.2 of the Anti-Dumping Agreement.

5.1.8.1.1  Dumping and margins of dumping under the first sentence of Article 2.4.2 of the Anti-Dumping Agreement

5.90.  The Appellate Body has consistently held that the concepts of "dumping" and "margins of dumping" are the same throughout the Anti‑Dumping Agreement. Recalling that Article 2.1 of the Anti‑Dumping Agreement defines "dumping" "[f]or the purpose of this Agreement"^[261], the Appellate Body found that the definitional content of "dumping" must be capable of application throughout the Anti‑Dumping Agreement in a coherent fashion.^[262] In US – Zeroing (Japan), the Appellate Body clarified that the terms "dumping" and "dumped imports" have the same meaning in all provisions of the Anti‑Dumping Agreement and for all types of anti‑dumping proceedings, including original investigations, new shipper reviews, and periodic reviews and that, in each case, "they relate to a product because it is the product that is introduced into the commerce of another country at less than its normal value in that country."^[263] The Appellate Body further indicated that "[t]he definitions in Article 2.1 [of the Anti-Dumping Agreement] and Article VI:1 [of the GATT 1994] are no doubt central to the interpretation of other provisions of the Anti‑Dumping Agreement, such as the obligations relating to, inter alia, the calculation of margins of dumping, volume of dumped imports, and levy of anti‑dumping duties to counteract injurious dumping".^[264]

5.91.  Moreover, the Appellate Body has considered the Anti‑Dumping Agreement to prescribe that "dumping determinations be made in respect of each exporter or foreign producer examined … because dumping is the result of the pricing behaviour of individual exporters or foreign producers."^[265] According to the Appellate Body, in order to assess properly the pricing behaviour of an individual exporter or foreign producer, and to determine whether the exporter or foreign producer is in fact dumping the product under investigation and, if so, by which margin, it is obviously necessary to take into account the prices of all the export transactions of that exporter or foreign producer.^[266]

5.92.  Article 2.4.2 of the Anti‑Dumping Agreement explains how investigating authorities must proceed to ascertain margins of dumping. More particularly, Article 2.4.2 contains two sentences, which set out the methodologies that investigating authorities may use to establish margins of dumping.

5.93.  The first sentence of Article 2.4.2 states that the existence of margins of dumping "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." The Appellate Body has considered that, in establishing dumping and margins of dumping for each exporter or producer and for the product under investigation "as a whole" under the first sentence of Article 2.4.2, an investigating authority is under an obligation to take into account the entire "universe of export transactions", that is, all the export transactions of the "like" product by a given exporter or foreign producer.

5.94.  In the context of the W‑W comparison methodology, the Appellate Body has found that there is nothing in Article 2.4.2 or in any other provision of the Anti‑Dumping Agreement that provides for the establishment of the existence of margins of dumping for types or models of the product under investigation, and all references to the establishment of the existence of margins of dumping are references to the product under investigation.^[267] In particular, the Appellate Body has found that "[w]hatever the method used to calculate the margins of dumping, … these margins must be, and can only be, established for the product under investigation as a whole."^[268]

5.95.  In establishing margins of dumping for the product under investigation under the W‑W comparison methodology, the Appellate Body has also recognized that, while the investigating authority may undertake multiple averaging or multiple comparisons at the sub‑group level, it is "required to compare the weighted average normal value with the weighted average of prices of all comparable export transactions".^[269] Thus, it was in the context of "model zeroing" that the Appellate Body in EC – Bed Linen found that, "[b]y 'zeroing' the 'negative dumping margins', the European Communities … did not take fully into account the entirety of the prices of some export transactions, namely, those export transactions involving models of cotton‑type bed linen where 'negative dumping margins' were found."^[270] Accordingly, the Appellate Body found that the existence of margins of dumping was not established on the basis of a comparison of the weighted average normal value with the weighted average of prices of all comparable export transactions.^[271] The Appellate Body has explained that, for the purposes of the W‑W comparison methodology, "the word 'all' in 'all comparable export transactions' makes it clear that Members … 'may only compare those export transactions which are comparable, but [they] must compare all such transactions'."^[272]

5.96.  Moreover, the Appellate Body has emphasized that, although an investigating authority may undertake multiple averaging to establish margins of dumping for a product under investigation, "the results of the multiple comparisons at the sub‑group level are … not 'margins of dumping' within the meaning of Article 2.4.2."^[273] In the absence of a textual basis in Article 2.4.2 that would justify taking into account the results of only some multiple comparisons in the process of calculating margins of dumping, while disregarding other results^[274], the Appellate Body has stated that "it is only on the basis of aggregating all … 'intermediate values' that an investigating authority can establish margins of dumping for the product under investigation as a whole."^[275]

5.97.  In respect of the T‑T comparison methodology, the Appellate Body has not considered the absence of the phrase "all comparable export transactions" as suggesting that an investigating authority is free to disregard certain export transactions from the applicable "universe of export transactions" in order to establish margins of dumping. According to the Appellate Body, the phrase "all comparable export transactions" is not pertinent for this methodology because "under the T‑T comparison methodology, all export transactions are taken into account on an individual basis and matched with the most appropriate transactions in the domestic market."^[276] The Appellate Body has, thus, found that the text of Article 2.4.2 implies that the calculation of margins of dumping using the T‑T comparison methodology is a multi‑step exercise, whereby the results of transaction‑specific comparisons are inputs that are aggregated in order to establish the margin of dumping for the product under investigation and for each exporter or foreign producer.^[277]

5.98.  In light of the above, an investigating authority is not allowed to disregard, in establishing dumping and margins of dumping under the W‑W or T‑T comparison methodology provided in the first sentence of Article 2.4.2 of the Anti-Dumping Agreement, any transactions that make up the "universe of export transactions" to be examined thereunder, that is, all the export transactions of the "like" product by a given exporter or foreign producer.

5.1.8.1.2  The W‑T comparison methodology in the second sentence of Article 2.4.2 of the Anti‑Dumping Agreement

5.99.  We now turn to address how the W‑T comparison methodology works under the second sentence of Article 2.4.2 of the Anti‑Dumping Agreement. In so doing, we focus on the "universe of export transactions" that needs to be compared with normal value in order to establish dumping and margins of dumping, also keeping in mind the function of this provision. We then consider whether the second sentence of Article 2.4.2 allows combining the W‑T comparison methodology with one of the two symmetrical comparison methodologies provided in the first sentence. Finally, we address the participants' arguments in respect of "mathematical equivalence".

5.100.  The second sentence of Article 2.4.2 of the Anti‑Dumping Agreement 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.

5.101.  The comparison envisaged under the second sentence of Article 2.4.2 is, thus, an asymmetrical comparison between a normal value "established on a weighted average basis" and prices of "individual export transactions". We have concluded above^[278] that these "individual export transactions" refer to the "universe of export transactions" that justify the use of the W-T comparison methodology, namely, the "pattern transactions". We have reached this conclusion based on the text of the second sentence of Article 2.4.2, read in light of its function of allowing an investigating authority to identify and address "targeted dumping", and in keeping with the overall structure of Article 2.4.2 of the Anti-Dumping Agreement.^[279]

5.1.8.1.2.1  Dumping and margins of dumping under the second sentence of Article 2.4.2 of the Anti-Dumping Agreement

5.102.  The Panel considered that, "at first glance, there appears to be some tension between (a) the Appellate Body's understanding of the limited scope of application of the W‑T comparison methodology and (b) its reference to the 'fundamental disciplines' that 'dumping' and 'margins of dumping' pertain to an exporter or foreign producer, and to the product [under investigation] (as a whole), taking into account all export transactions of the exporter or foreign producer concerned."^[280] In this regard, both the Panel^[281] and Korea^[282] referred to the findings of the Appellate Body in US – Zeroing (Japan), where the Appellate Body concluded:

Thus, it is evident from the design and architecture of the Anti‑Dumping Agreement that: (a) the concepts of "dumping" and "margins of dumping" pertain to a "product" and to an exporter or foreign producer; (b) "dumping" and "dumping margins" must be determined in respect of each known exporter or foreign producer examined; (c) anti‑dumping duties can be levied only if dumped imports cause or threaten to cause material injury to the domestic industry producing like products; and (d) anti‑dumping duties can be levied only in an amount not exceeding the margin of dumping established for each exporter or foreign producer. These concepts are interlinked. They do not vary with the methodologies followed for a determination made under the various provisions of the Anti‑Dumping Agreement.^[283]

5.103.  We observe that these statements in respect of the first sentence of Article 2.4.2, that dumping and margins of dumping have to be established for the product under investigation "as a whole", should be read in the context of the Appellate Body's finding against "model zeroing". "Model zeroing" occurs in situations where the investigating authority divides the product under investigation into product types or models for the purposes of calculating a weighted average normal value and a weighted average export price and sets to zero the negative comparison results arising in respect of certain product types or models. In EC – Bed Linen, in addressing "model zeroing", the Appellate Body stated that, "with respect to Article 2.4.2, the European Communities had to establish 'the existence of margins of dumping' for the product ‒ cotton‑type bed linen – and not for the various types or models of that product" since, "[h]aving defined the product as it did, the European Communities was bound to treat that product consistently thereafter in accordance with that definition."^[284] The Appellate Body found that, "[b]y 'zeroing' the 'negative dumping margins', the European Communities, therefore, did not take fully into account the entirety of the prices of some export transactions, namely, those export transactions involving models of cotton‑type bed linen where 'negative dumping margins' were found."^[285] The Appellate Body concluded that, in so doing, "the European Communities did not establish 'the existence of margins of dumping' for cotton‑type bed linen on the basis of a comparison of the weighted average normal value with the weighted average of prices of all comparable export transactions – that is, for all transactions involving all models or types of the product under investigation."^[286]

5.104.  The establishment of dumping and margins of dumping, for the product under investigation "as a whole" and by taking into account all export transactions of a given exporter or foreign producer, is to be carried out in respect of the applicable "universe of export transactions" for each of the comparison methodologies set forth in Article 2.4.2. It is in the context of establishing margins of dumping for the product under investigation "as a whole" under the normally applicable W-W and T-T comparison methodologies that the Appellate Body has consistently found that, under the first sentence of Article 2.4.2, an investigating authority is under an obligation to consider the entire "universe of export transactions" for a given exporter or foreign producer. However, and as the Appellate Body has previously stated, under the W‑T comparison methodology set forth in the second sentence of Article 2.4.2, the applicable "universe of export transactions" is more limited than those under the W‑W and T‑T comparison methodologies.^[287]

5.105.  Once the applicable "universe of export transactions" has been determined under the second sentence of Article 2.4.2 for the purposes of the application of the W‑T comparison methodology, dumping and margins of dumping pertaining to an exporter or foreign producer and to the product under investigation are limited to this identified "universe of export transactions", i.e. the "pattern transactions".

5.106.  In light of the above, not only is the application of the W‑T comparison methodology limited to "pattern transactions", but also the second sentence of Article 2.4.2 provides for establishing margins of dumping on the basis of the limited "universe of export transactions" to which the W‑T comparison methodology applies – i.e. the "pattern transactions".^[288] While the first sentence of Article 2.4.2 provides for symmetrical comparison methodologies that "shall normally" be used to establish the existence of margins of dumping by considering the universe of all export transactions, the second sentence provides for an asymmetrical comparison methodology, which, as we have explained above and as the Appellate Body has stated, "is clearly an exception to the comparison methodologies which normally are to be used".^[289]

5.1.8.1.2.2  The function of the second sentence of Article 2.4.2 of the Anti‑Dumping Agreement

5.107.  The function of the second sentence of Article 2.4.2 lends further support to a reading of this provision as providing for the establishment of margins of dumping on the basis of the limited "universe of export transactions" that form the pattern. We have considered above that the function of the second sentence of Article 2.4.2 is to allow an investigating authority to identify and address "targeted dumping". This is supported by the Appellate Body's statement in EC – Bed Linen that "[t]his provision allows Members, in structuring their anti‑dumping investigations, to address three kinds of 'targeted' dumping, namely dumping that is targeted to certain purchasers, targeted to certain regions, or targeted to certain time periods."^[290]

5.108.  The second sentence of Article 2.4.2 does not expressly refer to "targeted dumping", but allows the use of the W‑T comparison methodology "if the authorities find a pattern of export prices which differ significantly among different purchasers, regions or time periods". The notion of "targeted dumping" is implied in the reference in the second sentence of Article 2.4.2 to "a pattern of export prices which differ significantly among different purchasers, regions or time periods". Therefore, it is only certain export transactions, those that constitute the pattern by differing significantly from the remaining export transactions of an exporter or foreign producer, that are aimed at, or "targeted" to, a purchaser, region, or time period within the meaning of the second sentence of Article 2.4.2.

5.109.  In keeping with its function of allowing an investigating authority to effectively address "targeted dumping", the second sentence of Article 2.4.2 permits an investigating authority to establish margins of dumping by means of the application of the W‑T comparison methodology exclusively to "pattern transactions". The second sentence of Article 2.4.2 says nothing about including transactions that are not part of the pattern in the comparison process that is required to establish margins of dumping. If an investigating authority were required to conduct comparisons with export transactions outside of the pattern – i.e. for "non‑pattern transactions" – by applying one of the two normally applicable comparison methodologies, and then aggregate the result of this comparison with the result of the W‑T comparison methodology applied to "pattern transactions", the "targeted dumping" identified from the consideration of "pattern transactions" would be "re‑masked" by the comparison results arising from "non‑pattern transactions", in situations where the latter produces an overall negative comparison result.

5.110.  Korea disagrees that the function of the second sentence of Article 2.4.2 is to "unmask targeted dumping" and asserts that its purpose "is simply to allow the authority to undertake the more careful examination of individual export prices that the W‑T comparison method[ology] makes possible".^[291] We see no textual basis in the second sentence of Article 2.4.2 to conclude, as Korea asserts, that the function of the second sentence is to allow an investigating authority to undertake a more careful and "granular" examination of individual export prices. In any event, we do not see how Korea's proposed interpretation is different from what is already contemplated under the T‑T comparison methodology. In this respect, we are not convinced that, under Korea's approach, the W‑T comparison methodology would fulfil a function that is not already fulfilled by the T‑T comparison methodology.

5.111.  Therefore, we do not think that the Panel erred in identifying the function of the second sentence of Article 2.4.2 as addressing or "unmasking targeted dumping" and in considering that this function informs an interpretation of the second sentence of Article 2.4.2 as providing for the establishment of margins of dumping under the W‑T comparison methodology by comparing normal value only with "pattern transactions", while excluding from consideration "non‑pattern transactions".^[292] Even if one were to accept Korea's proposed interpretation that the function of the second sentence is to allow an investigating authority to conduct a more careful or "granular" examination, we do not consider that such an interpretation would be inconsistent with the reading of the second sentence of Article 2.4.2 as establishing margins of dumping based on "pattern transactions" only. Indeed, under the second sentence of Article 2.4.2 of the Anti-Dumping Agreement, an investigating authority will carefully consider all export transactions and separate those individual export transactions that form the relevant "pattern" within the meaning of that provision from those that are not part of the identified "pattern". Margins of dumping are then established by comparing the weighted average normal value only with transactions that are included in the identified "pattern".

5.1.8.1.2.3  The composition of numerator and denominator in establishing margins of dumping

5.112.  Korea further contends that, in determining the margin of dumping, the numerator should comprise "the net amount of the comparisons of all export prices to normal value"^[293] and the denominator should comprise the "total sales by the exporter".^[294] According to Korea, "a proper 'margin of dumping' consists of: (i) a numerator that considers all export sales, without pretending that some of the export sales were at lower prices (by denying offsets); and (ii) a denominator that also consists of all export sales."^[295]

5.113.  For its part, the United States does not disagree with Korea that all of an exporter's export transactions should be "taken into account" in the determination of dumping.^[296] The United States adds that the USDOC's approach does, in fact, take account of all export transactions. However, the United States submits that Korea's approach means that the "evidence of 'targeted dumping' must, as a matter of an obligation under the AD Agreement, be re‑masked by aggregating all results for all transactions in the numerator of the calculation of the margin of dumping."^[297]

5.114.  We have explained above that the second sentence of Article 2.4.2 permits the establishment of the existence of margins of dumping for the product under investigation and for each exporter or foreign producer by applying the W‑T comparison methodology to "pattern transactions" only. Thus, in calculating the margin of dumping as a percentage of the exports of a given exporter or foreign producer, the numerator comprises only "pattern transactions", while "non‑pattern transactions" are excluded. The denominator, however, must reflect the universe of all export transactions of a given exporter or foreign producer and comprises the value of all the sales of a given exporter or foreign producer of the "like" product.

5.115.  We observe that Article 6.10 of the Anti-Dumping Agreement states that "[t]he authorities shall, as a rule, determine an individual margin of dumping for each known exporter or producer concerned of the product under investigation." The margin of dumping determined under the second sentence of Article 2.4.2 in order to address "targeted dumping" is for each exporter or producer, and not just for the "targeted sales" by that exporter or producer. On the one hand, the existence of a "pattern" within the meaning of the second sentence of Article 2.4.2, and thus "targeting" by exporters or producers, justifies taking the dumping amount from the W‑T comparison that is applied to the "pattern transactions", while excluding from consideration "non‑pattern transactions". On the other hand, this dumping amount that is based on the "targeted" sales must be divided by all the export sales of a given exporter or producer in order to determine the margin of dumping and the corresponding anti‑dumping duty for that exporter or producer. Therefore, the Panel did not err in finding that "while the net amount of dumping may be established from considering the evidence of dumping in pattern transactions … the calculation of the margin as a percentage of the exports of that exporter or foreign producer must reflect the price of its total exports."^[298]

5.116.  Under the W‑T comparison methodology provided in the second sentence of Article 2.4.2, the margin of dumping, which is expressed as a percentage of the total value of export transactions of an exporter or foreign producer, would be established by considering "pattern transactions", while excluding "non-pattern transactions" in the numerator of the equation. The denominator, however, will reflect all export transactions of an exporter or foreign producer. In so doing, while "targeted dumping" is identified and addressed by including in the numerator the "pattern transactions", the denominator, in reflecting the value of all export transactions of the "like" product by a given exporter or foreign producer, ensures that, for the universe of "pattern transactions" to which the W‑T comparison methodology is applied, the margin of dumping is calculated for that exporter or foreign producer and for the product under investigation "as a whole". This exercise is, therefore, consistent with the concepts of "dumping" and "margins of dumping" as pertaining to an exporter or foreign producer and to the product under investigation "as a whole" and with the function of the second sentence of Article 2.4.2 of the Anti-Dumping Agreement as identifying and addressing "targeted dumping".

5.117.  We, therefore, conclude that, in calculating the margin of dumping as a percentage of the exports of a given exporter or foreign producer, the numerator comprises only the "pattern transactions", while excluding "non‑pattern transactions". The denominator, however, is composed of all export transactions of a given exporter or foreign producer and comprises the value of all the sales of a given exporter or producer of the "like" product.

5.1.8.1.2.4  Combining comparison methodologies

5.118.  Korea contends that the Panel erred in permitting the USDOC to establish margins of dumping under the DPM by disregarding an overall negative comparison result arising from the application of the W‑W comparison methodology to "non‑pattern transactions" when combining it with the overall comparison result of the W‑T comparison methodology.^[299] We note that Korea's argument on "systemic disregarding" of "non‑pattern transactions" is premised on the assumption that the second sentence of Article 2.4.2 of the Anti-Dumping Agreement permits the combining of comparison methodologies in establishing margins of dumping.

5.119.  The Panel acknowledged that "[o]ne might take the view … that the combined application of the W‑T and W‑W (or T‑T) comparison methodologies is not envisaged by that provision".^[300] However, since Korea had not raised a claim to this effect, the Panel did not see the need to rule on that matter and proceeded with the assumption that a combined application of comparison methodologies was not excluded by the second sentence of Article 2.4.2.

5.120.  We have found above that the second sentence of Article 2.4.2, read in the context of the structure of Article 2.4.2 and consistently with its function of allowing an investigating authority to identify and address "targeted dumping", permits the establishment of margins of dumping by the application of the W‑T comparison methodology to "pattern transactions", while excluding from consideration "non‑pattern transactions", when the conditions stated in that provision have been satisfied. Thus, having concluded that the applicable "universe of export transactions" for the purposes of establishing dumping and margins of dumping under the second sentence of Article 2.4.2 is limited to "pattern transactions", we do not consider that this provision allows the combining of comparison methodologies, that is, the combined application of the W‑T comparison methodology applied to "pattern transactions" with either the W‑W or the T‑T comparison methodology applied to "non‑pattern transactions".

5.121.  Although the second sentence mentions the W‑W and T‑T comparison methodologies, this reference appears in the context of the explanation requirement of why neither of these symmetrical comparison methodologies is capable of taking into account the "export prices which differ significantly". The second sentence does not provide for the application of the W‑W and T‑T comparison methodologies anew. Instead, these two symmetrical comparison methodologies are referenced as they are provided for in, and subject to the requirements of, the first sentence of Article 2.4.2. Consistent with the jurisprudence of the Appellate Body, we have explained above that, under the first sentence of Article 2.4.2, the W‑W and T‑T comparison methodologies apply to the universe of all export transactions, and not to "pattern" or "non‑pattern" transactions within the meaning of the second sentence. Thus, we do not see anything in the text of the second sentence, read in the context of the entire Article 2.4.2, that supports a reading of the W‑W and T‑T comparison methodologies as applying to a reduced "universe of export transactions" (i.e. the "non‑pattern transactions") pursuant to the reference to these two comparison methodologies in the second sentence of Article 2.4.2. The text of the second sentence, thus, does not provide for the application of either of the two symmetrical comparison methodologies to "non‑pattern transactions".

5.122.  Moreover, conducting a separate comparison under one of the two symmetrical comparison methodologies for "non‑pattern transactions" not only lacks support in the text of the second sentence of Article 2.4.2, but also undermines the function of the second sentence. In order to enable an investigating authority to identify and address "targeted dumping", the second sentence of Article 2.4.2 defines a more limited "universe of export transactions", which, as we have considered above, coincides with the "pattern transactions" that are targeted to purchasers, regions, or time periods. It is on the basis of these "pattern transactions" that margins of dumping are then established under the second sentence. To aggregate the results of a symmetrical comparison methodology applied to "non‑pattern transactions" may either "mask" the "targeted dumping" (if the overall comparison result is negative) or increase the margin of dumping (if the overall comparison result is positive) in a manner that is not provided for in the second sentence of Article 2.4.2 and that would compromise its function of effectively addressing "targeted dumping".

5.123.  The second sentence of Article 2.4.2 does not envisage "systemic disregarding" as described by the Panel.^[301] This provision does not provide for a mechanism whereby an investigating authority would conduct separate comparisons for "pattern transactions" under the W‑T comparison methodology and for "non‑pattern transactions" under the W‑W or T‑T comparison methodology, and exclude from its consideration the result of the latter if it yields an overall negative comparison result, or aggregate it with the W‑T comparison results for the "pattern transactions" if it yields an overall positive comparison result. This is further supported by our conclusion above that, if the conditions set forth in the second sentence of Article 2.4.2 are met, an investigating authority is allowed to establish margins of dumping by applying the W‑T comparison methodology only to "pattern transactions", while excluding from consideration "non‑pattern transactions".

5.124.  Hence, we conclude that Article 2.4.2 does not permit the combining of comparison methodologies for the purposes of establishing dumping and margins of dumping in accordance with the second sentence. Instead, the second sentence allows an investigating authority to establish the existence of margins of dumping by comparing a "normal value established on a weighted average basis" with "pattern transactions" only. We have explained above that this is consistent with the function of the second sentence of Article 2.4.2 and that, for the purposes of this provision, dumping and margins of dumping pertaining to an exporter or a foreign producer, and to the product under investigation "as a whole", refer to a more limited "universe of export transactions", that is, the "pattern transactions". In light of the above, we consider the question of whether the second sentence of Article 2.4.2 of the Anti-Dumping Agreement allows "systemic disregarding" as defined by the Panel as moot.

5.1.8.1.2.5  Mathematical equivalence

5.125.  In light of the above considerations, comparing normal value with "pattern transactions" only will not normally yield results that are mathematically or substantially equivalent to the results obtained from the application of the W‑W comparison methodology to all export transactions.^[302] Mathematical equivalence or substantial equivalence arises only if one were to take the view that the second sentence of Article 2.4.2 of the Anti-Dumping Agreement envisages the combining of comparison methodologies, thereby requiring an investigating authority to aggregate the results of the W‑T comparison methodology applied to "pattern transactions" with the results of the W‑W comparison methodology applied to "non‑pattern transactions".^[303] In contrast, we have rejected an interpretation of the second sentence of Article 2.4.2 as permitting the combination of comparison methodologies.

5.126.  Korea submits that the Panel found mathematical equivalence despite the fact that the argument of mathematical equivalence has been considered and rejected by the Appellate Body previously on numerous occasions when considering the second sentence of Article 2.4.2 as context for the permissibility of zeroing under the first sentence.^[304] Korea further asserts that there is no necessary mathematical equivalence if the investigating authority changes its assumptions about normal value or makes adjustments for export prices for the subset of sales subjected to the W‑T comparison methodology.^[305]

5.127.  The United States contends that, if zeroing is prohibited under the W‑T comparison methodology, and if the result of the W‑W comparison is allowed to "offset" the result of the W‑T comparison, then the overall result arising from this combination and that derived from the application of the W‑W comparison methodology to all export prices will be mathematically equivalent^[306], which would render the second sentence of Article 2.4.2 inutile, contrary to the principle of effectiveness.^[307]

5.128.  The function of the second sentence of Article 2.4.2 should not be addressed by focusing on mathematical equivalence.^[308] As the Appellate Body indicated in US – Softwood Lumber V (Article 21.5 – Canada), that the comparison methodologies in the first sentence of Article 2.4.2 and the W‑T comparison methodology will not normally produce equivalent results is a consequence of the fact that the third comparison methodology addresses "targeted dumping" by focusing on "pattern transactions".^[309] Korea's argument that mathematical equivalence can be avoided by changing normal value or considering adjustments to export prices, thereby leading to different results in some cases, still does not explain how this exercise accords with the function of the second sentence of Article 2.4.2 of the Anti-Dumping Agreement of allowing an investigating authority to identify and address "targeted dumping". We fail to see how "different results" will in themselves address "targeted dumping", unless such results are calculated based on "pattern transactions".

5.1.8.1.2.6   Conclusions

5.129.  We have concluded above that the second sentence of Article 2.4.2 of the Anti‑Dumping Agreement allows an investigating authority to establish margins of dumping by applying the W‑T comparison methodology only to "pattern transactions" to the exclusion of "non‑pattern transactions". We have also concluded that the second sentence of Article 2.4.2 does not permit the combining of comparison methodologies. Accordingly, we have found that this provision does not envisage "systemic disregarding", as described by the Panel. We do not consider that the second sentence of Article 2.4.2 envisages a mechanism whereby an investigating authority would conduct separate comparisons for "pattern transactions" under the W‑T comparison methodology and for "non‑pattern transactions" under the W‑W or T‑T comparison methodology, and exclude from its consideration the result of the latter if it yields an overall negative comparison result or aggregate it with the W-T comparison result for the "pattern transactions" if it yields an overall positive comparison result. Thus, in circumstances where the requirements of the second sentence of Article 2.4.2 have been fulfilled, an investigating authority is allowed to establish margins of dumping by comparing a weighted average normal value with export prices of "pattern transactions" and dividing the resulting amount by all the export sales of a given exporter or foreign producer.

5.130.  In light of the above, we, therefore, moot the Panel's finding, in paragraph 8.1.a.x of its Report^[310], that "Korea failed to establish that the United States' use of 'systemic disregarding' under the DPM is 'as such' inconsistent with the second sentence of Article 2.4.2". Instead, as explained above, when the requirements of the second sentence of Article 2.4.2 of the Anti‑Dumping Agreement are fulfilled, an investigating authority may establish margins of dumping by comparing a weighted average normal value with export prices of "pattern transactions", while excluding "non-pattern transactions" from the numerator, and dividing the resulting amount by all the export sales of a given exporter or foreign producer.

5.1.8.2  Whether the Panel erred in finding that Korea failed to establish that the United States' use of "systemic disregarding" under the DPM is inconsistent "as such" with Article 2.4 of the Anti‑Dumping Agreement

5.131.  We turn now to consider Korea's appeal of the Panel's findings under Article 2.4 of the Anti‑Dumping Agreement in respect of "systemic disregarding".

5.132.  The Panel recalled its findings that Article 2.4.2 of the Anti‑Dumping Agreement enables investigating authorities to establish the existence of margins of dumping by focusing on "pattern transactions" and that, if an investigating authority chooses to combine the application of the W‑W comparison methodology to "non‑pattern transactions" with the application of the W‑T comparison methodology to "pattern transactions", "systemic disregarding" enables it to avoid concealing any dumping identified in respect of "pattern transactions" with the negative dumping in respect of "non‑pattern transactions". Accordingly, the Panel rejected Korea's argument that "'systemic disregarding' is unfair and contrary to Article 2.4 because it inflates the margin of dumping and ignores the negative amount of dumping in respect of non‑pattern transactions".^[311]

5.133.  On appeal, Korea argues that the Panel's finding that there is nothing "unfair" about "systemic disregarding" and that, therefore, it is not inconsistent with Article 2.4, repeats the same legal errors that Korea has identified with regard to the Panel's interpretation of the second sentence of Article 2.4.2 that dumping can exist within a subset of the export sales.^[312] Korea, therefore, requests the reversal of the Panel's findings under Article 2.4.^[313] In addition, Korea requests us to complete the legal analysis based on undisputed facts and on the Panel's findings in respect of the second sentence of Article 2.4.2, and to find that "systemic disregarding" is inconsistent with the "fair comparison" requirement in Article 2.4.^[314] Korea asserts that, for a measure to meet the "fair comparison" requirement in Article 2.4, it must be impartial, even‑handed, and unbiased, something that "systemic disregarding" is not.^[315]

5.134.  The United States submits that the Panel was correct in finding that the USDOC's approach to the application of a "mixed" comparison methodology is not inconsistent "as such" with Article 2.4 and that Korea's arguments lack merit.^[316] The United States contends that, since it is "fair" to take steps to "unmask targeted dumping" by faithfully applying a comparison methodology consistent with the second sentence of Article 2.4.2, when the conditions for its use are met, doing so is entirely consistent with the obligation of an investigating authority under Article 2.4 to be impartial, even‑handed, and unbiased.^[317] Thus, the United States submits that, since the Panel's findings need not be reversed, there is no need for us to complete the legal analysis.^[318]

5.135.  We begin by recalling that the first sentence of Article 2.4 provides that "[a] fair comparison shall be made between the export price and the normal value." We note that the introductory clause of Article 2.4.2 expressly makes this provision "[s]ubject to the provisions governing fair comparison" in Article 2.4. In this regard, the Appellate Body has 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 specifically made 'subject to the provisions governing fair comparison in [Article 2.4]'."^[319] Therefore, the application of all three comparison methodologies (i.e. W‑W, T‑T, and W‑T) set out in Article 2.4.2 is expressly made subject to the "fair comparison" requirement in Article 2.4.

5.136.  In explaining the "fair comparison" requirement in Article 2.4, the Appellate Body in US ‒ Softwood Lumber V (Article 21.5 – Canada) stated that "[t]he term 'fair' is generally understood to connote impartiality, even‑handedness, or lack of bias."^[320] The Appellate Body found in that dispute 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"^[321], since "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."^[322]

5.137.  We have concluded above that the second sentence of Article 2.4.2 allows an investigating authority to establish dumping and margins of dumping for an exporter or foreign producer and for the product under investigation "as a whole" by applying the W‑T comparison methodology to "pattern transactions" only. We have also explained that, in doing so, while the denominator of the equation will comprise all export transactions of an exporter or foreign producer, the numerator is composed of "pattern transactions" only, while excluding "non-pattern transactions". Moreover, we have concluded that the second sentence of Article 2.4.2 does not permit the combining of comparison methodologies.

5.138.  The obligation to undertake a "fair comparison" between normal value and export prices arises in respect of the applicable "universe of export transactions" to which each of the three comparison methodologies set out in Article 2.4.2 applies. The exceptional nature of the W‑T comparison methodology, consistent with the function of the second sentence of Article 2.4.2 as allowing an investigating authority to identify and address "targeted dumping" by considering a "universe of export transactions … which … would be different from the universe of transactions examined"^[323] under the normally applicable comparison methodologies, also confirms that the "fair comparison" requirement in Article 2.4 applies only in respect of "pattern transactions". Accordingly, we consider that Articles 2.4 and 2.4.2 not only inform each other, but must be read together harmoniously. The exclusion of "non‑pattern transactions" from the establishment of dumping and margins of dumping under the second sentence of Article 2.4.2, thus, comports with the notions of impartiality, even‑handedness, and lack of bias reflected in the "fair comparison" requirement in Article 2.4. Once an investigating authority has identified the "pattern transactions" within the meaning of the second sentence of Article 2.4.2 for the purposes of the application of the W‑T comparison methodology in order to address "targeted dumping" to the exclusion of "non‑pattern transactions", the investigating authority can be said to have adopted an approach that is impartial, even‑handed, and unbiased. Such an approach neither distorts the prices of certain export transactions (i.e. the "non‑pattern transactions"), nor inflates the magnitude of dumping, as Korea asserts^[324], since there is nothing "to disregard", as the second sentence of Article 2.4.2 does not contemplate, in the first place, a comparison outside of the identified "pattern".

5.139.  We recall that the Panel rejected Korea's argument that "systemic disregarding" under the DPM is "unfair and contrary to Article 2.4".^[325] Although the Panel considered that an investigating authority could establish the existence of margins of dumping by focusing on "pattern transactions" to the exclusion of "non‑pattern transactions"^[326], the underlying assumption of the Panel was that the combined application of comparison methodologies is not excluded under the second sentence of Article 2.4.2.^[327] We have, however, explained above that, while we agree that the second sentence of Article 2.4.2 allows an investigating authority to establish margins of dumping based on a comparison between normal value and "pattern transactions" only, we do not consider that the second sentence allows for the combined application of the W‑T comparison methodology to "pattern transactions" and the W‑W (or T‑T) comparison methodology to "non‑pattern transactions".

5.140.  In light of these considerations, we conclude that the establishment of margins of dumping by comparing a weighted average normal value with export prices of "pattern transactions", while excluding "non-pattern transactions" from the numerator, and dividing the resulting amount by all the export sales of a given exporter or foreign producer, is consistent with the "fair comparison" requirement in Article 2.4. Having concluded that the second sentence of Article 2.4.2 does not permit an investigating authority to combine the W‑T comparison methodology with the W‑W or T‑T comparison methodology and, thus, does not provide for "systemic disregarding" as described by the Panel, we moot the Panel's finding, in paragraph 8.1.a.xi of its Report^[328], that "Korea failed to establish that the United States' use of 'systemic disregarding' under the DPM is 'as such' inconsistent with Article 2.4" of the Anti‑Dumping Agreement.

5.1.9  Zeroing under the W-T comparison methodology^[329]

5.1.9.1  Whether the Panel erred in finding that the United States' use of zeroing when applying the W‑T comparison methodology is inconsistent "as such" and "as applied" in the Washers anti‑dumping investigation with Article 2.4.2 of the Anti‑Dumping Agreement

5.141.  The Panel found that, since the second sentence of Article 2.4.2 of the Anti-Dumping Agreement puts particular emphasis on the exporter's pricing behaviour in respect of "pattern transactions", the entirety of the evidence of dumping in respect of that pattern must be taken into account. According to the Panel, the focus of the W‑T comparison methodology is on the prices of the "individual" export transactions within the pattern, which suggests that each "pattern transaction" should be considered in its own right and with equal weight, irrespective of whether the export price is above or below normal value. Thus, the Panel found that there is no basis in the text of the second sentence of Article 2.4.2 to conclude that the export prices of certain individual transactions (e.g. those below normal value) should be accorded greater significance than the export prices of other individual export transactions (e.g. those above normal value). The Panel took the view that the phrase "individual export transactions" in the first part of the second sentence of Article 2.4.2 suggests that each and every "pattern transaction" should be fully taken into account in the assessment of the exporter's pricing behaviour in respect of that pattern.^[330] The Panel also found that there is "no consideration of whether transactions within the pattern are priced at significantly different levels relative to one another"^[331] and, thus, no basis to conclude that one (pattern) transaction priced significantly lower than "non‑pattern transactions" might mask evidence of dumping in respect of another (pattern) transaction priced significantly lower than "non‑pattern transactions".^[332]

5.142.  Accordingly, the Panel concluded that the USDOC's use of zeroing when applying the W‑T comparison methodology is inconsistent "as such" with the second sentence of Article 2.4.2.^[333] For the same reasons, the Panel concluded that the USDOC acted inconsistently with the second sentence of Article 2.4.2 of the Anti-Dumping Agreement by using zeroing when applying the W‑T comparison methodology in the Washers anti‑dumping investigation.^[334]

5.1.9.1.1  Zeroing under the second sentence of Article 2.4.2 of the Anti‑Dumping Agreement

5.143.  The United States claims that the Panel erred in finding that the use of zeroing in connection with the application of the W‑T comparison methodology is inconsistent with the second sentence of Article 2.4.2 of the Anti‑Dumping Agreement, when the conditions for the use of the methodology have been fulfilled. According to the United States, an examination of the text and context of Article 2.4.2 leads to the conclusion that zeroing is permissible, and indeed necessary, when applying the W‑T comparison methodology, if that "exceptional" comparison methodology is to be given any meaning.^[335]

5.144.  Korea responds that the United States' argument does not address the Appellate Body's consistent logic that any finding of dumping must reflect all export transactions for each exporter and for the product under investigation "as a whole"^[336] and that denying "offsets" improperly disregards the actual prices of some of the export transactions.^[337] Moreover, Korea contends that these principles apply consistently throughout the Anti‑Dumping Agreement.^[338] According to Korea, there is nothing in the text or context of the second sentence of Article 2.4.2 that suggests that dumping or margin of dumping should have any different meaning for the second sentence of Article 2.4.2 than for the rest of the Anti‑Dumping Agreement.^[339]

5.145.  We have recalled above the main Appellate Body findings regarding the use of zeroing under the W‑W and T‑T comparison methodologies provided in the first sentence of Article 2.4.2. In finding that zeroing is not permitted under either of the two symmetrical comparison methodologies set forth in the first sentence, the Appellate Body considered that the concepts of "dumping" and "margins of dumping" are the same throughout the Anti‑Dumping Agreement and that margins of dumping are established for the product under investigation "as a whole", and for each exporter or foreign producer, by including in the comparison all the export transactions of the "like" product by that exporter or foreign producer.^[340]

5.146.  Thus, according to this jurisprudence, in establishing dumping and margins of dumping under the W‑W and T‑T comparison methodologies provided in the first sentence of Article 2.4.2, an investigating authority needs to take into account all transactions that make up the applicable "universe of export transactions" to be examined thereunder.

5.147.  We have also explained above that, under the second sentence of Article 2.4.2, dumping and margins of dumping pertaining to all export transactions of an exporter or foreign producer and to the product under investigation "as a whole" are limited to the applicable "universe of export transactions" for that provision, namely, the more limited universe of "pattern transactions".^[341] Thus, dumping and margins of dumping under the W‑T comparison methodology applied pursuant to the second sentence of Article 2.4.2 are to be determined by conducting a comparison between normal value and "pattern transactions", without having to take into account "non‑pattern transactions".^[342]

5.148.  We now turn to address the United States' claim on appeal that the Panel erred in finding that the use of zeroing in connection with the application of the W‑T comparison methodology is inconsistent with the second sentence of Article 2.4.2. In doing so, the question that we need to address is whether, in the application of the W‑T comparison methodology, an investigating authority is required to aggregate the results of all the transaction‑specific comparisons that arise from the consideration of "pattern transactions", or whether it can exclude those transactions within the pattern that yield negative intermediate comparison results, i.e. whether zeroing is permitted under the second sentence of Article 2.4.2.

5.149.  We recall that the Appellate Body has consistently held that the concepts of "dumping" and "margins of dumping" are the same throughout the Anti‑Dumping Agreement. The Appellate Body has considered that "Article 2.1 of the Anti‑Dumping Agreement and Article VI:1 of the GATT 1994 are definitional provisions" that "set out a definition of 'dumping' for the purposes of the Anti‑Dumping Agreement and the GATT 1994".^[343] The Appellate Body has also found that "the terms 'dumping' and 'margins of dumping' in Article VI of the GATT 1994 and the Anti‑Dumping Agreement apply to the product under investigation as a whole and do not apply to sub‑group levels."^[344]

5.150.  In US – Zeroing (EC), the Appellate Body found that the use of zeroing by the USDOC under the W‑T comparison methodology in administrative reviews was inconsistent with Article 9.3 of the Anti‑Dumping Agreement and Article VI:2 of the GATT 1994.^[345] These findings by the Appellate Body do not, however, directly address the question of whether the second sentence of Article 2.4.2 permits the use of zeroing in the application of the W‑T comparison methodology. Under the second sentence, the application of the W‑T comparison methodology serves to allow an investigating authority to identify and address "targeted dumping", which would otherwise be "masked" by the application of the two symmetrical comparison methodologies under the first sentence of Article 2.4.2. In contrast, in the administrative reviews at issue in US – Zeroing (EC), the United States applied the W‑T comparison methodology to assess the liability for anti‑dumping duties on specific entries of the subject product for each individual importer – i.e. on the basis of the transactions of each individual importer from the exporter, for whom a margin of dumping under Article 2 already existed. Thus, the two provisions fulfil different functions under the Anti‑Dumping Agreement.

5.151.  In interpreting the second sentence of Article 2.4.2, we have come to the conclusion that the term "individual export transactions" refers to the pattern of export prices identified by the investigating authority which differ significantly from other export prices. These export prices differ significantly because they are significantly lower than other export prices. This conclusion is supported by the text and context of Article 2.4.2, taking into account the function of the second sentence of allowing an investigating authority to identify and address "targeted dumping".^[346] We find no such textual and contextual support to conclude that the term "individual export transactions" in the second sentence of Article 2.4.2 refers only to those transactions that form part of the identified "pattern" but are priced below normal value. Rather, we agree with the Panel that the term "individual" suggests that "each pattern transaction should be considered in its own right, and with equal weight, irrespective of whether the export price is above or below normal value."^[347]

5.152.  Under the second sentence of Article 2.4.2, the relevant "pattern" is composed of a set of significantly lower prices to purchasers, regions, or time periods, and margins of dumping are established by conducting a comparison between normal value and those export transactions that are included in the pattern. The second sentence of Article 2.4.2 is "exceptional" because it allows investigating authorities to establish margins of dumping, while excluding from the dumping comparison those transactions that do not form part of the pattern. This exception is spelled out in the text of the second sentence of Article 2.4.2 that uses the terms "individual export transactions" and "a pattern of export prices which differ significantly". We have concluded above that these terms refer to the same set of export prices. Moreover, the reference in the second sentence to the term "individual export transactions" is in the context of highlighting the asymmetrical nature of the W-T comparison methodology, whereby a normal value established on a weighted average basis is compared to prices of individual export transactions as opposed to a comparison between normal value and export price on a W-W or T-T basis. Thus, we do not see any basis to read the term "individual export transactions" as permitting the exclusion of those individual "pattern transactions" that are priced above normal value from the establishment of margins of dumping in the application of the asymmetrical W-T comparison methodology.

5.153.  Zeroing within the pattern necessarily amounts to a definition of "pattern" that is limited to those export transactions to one or more particular purchasers, regions, or time periods that are below normal value, as it is only those sales that would be taken into account to establish margins of dumping when using zeroing. However, as we have found above, the second sentence of Article 2.4.2 does not define the pattern in reference to normal value. Rather, the reference to purchasers, regions, or time periods indicates that, while export prices within a pattern must differ significantly from other export prices, the pattern is composed of all the export prices to one or more particular purchasers, regions, or time periods, not just those that are below normal value. To allow zeroing within an identified "pattern" would disconnect the notion of pattern that is identified under the second sentence (as all export sales to one or more particular purchasers, regions, or time periods) from the pattern to which the W‑T comparison methodology is applied for establishing margins of dumping in order to address "targeted dumping" (by considering only those sales to one or more purchasers, regions, or time periods that are below normal value). However, the text of the second sentence of Article 2.4.2 does not support an interpretation according to which the pattern to which the W‑T comparison methodology applies for establishing margins of dumping is different from the pattern that triggers the application of the second sentence and that reveals the existence of "targeted dumping".

5.154.  We also recall that, in examining the permissibility of zeroing in the context of the T‑T comparison methodology, the Appellate Body stated in US – Softwood Lumber V (Article 21.5 ‒ Canada) that "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."^[348] The Appellate Body, thus, concluded that "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."^[349] While this finding concerned the permissibility of zeroing under the T‑T comparison methodology, similar considerations apply to the reference to "prices of individual export transactions" in the plural in the second sentence of Article 2.4.2, which as noted above is not qualified by any reference to normal value.

5.155.  Turning to the function of the second sentence of Article 2.4.2, we observe that the second sentence provides for an exception to the normally applicable symmetrical comparison methodologies of the first sentence in order to allow investigating authorities to identify and address "targeted dumping", whereby the "targeted dumping" coincides with the identified "pattern" of significantly different export prices. By conducting a comparison between normal value and all transactions included in the identified "pattern", an investigating authority is able to address the "targeted dumping" that is identified and that corresponds to that particular "pattern". Indeed, the second sentence of Article 2.4.2 allows an investigating authority to identify and address "targeted dumping" that corresponds to a properly defined pattern, which includes sales that are both above and below normal value, and not to a pattern composed exclusively of sales that are below normal value. We are, therefore, of the view that there is nothing more that needs to be "unmasked" once the dumping comparison has been conducted between normal value and "pattern transactions" to the exclusion of "non-pattern transactions". In this respect, while zeroing within a pattern that includes sales above normal value increases the margin of dumping, it does not "unmask" the "targeted dumping" that corresponds to the properly identified "pattern" of significantly lower sales, whereby such pattern includes sales below and above normal value.

5.156.  The United States' argument that, if zeroing is not allowed, the second sentence of Article 2.4.2 "would no longer be 'exceptional' and would no longer provide a means to 'unmask targeted dumping'"^[350], does not take into account that under the W‑T comparison methodology provided in the second sentence of Article 2.4.2 an investigating authority is allowed to establish margins of dumping by taking into account only "pattern transactions" to the exclusion of all other transactions that fall outside of the pattern. The use of the W‑T comparison methodology under the second sentence of Article 2.4.2 allows an investigating authority to "unmask" and address "targeted dumping" in keeping with its function and, accordingly, it does not deprive the second sentence of its effet utile. The second sentence of Article 2.4.2 has meaning and effect because it allows for the identification of the relevant "pattern" within the meaning of that provision and it allows an investigating authority to address "targeted dumping" by applying the W-T comparison methodology to the limited universe of "individual export transactions" that form the identified "pattern". As noted above, however, the "targeted dumping" to be "unmasked" corresponds to the properly identified "pattern", and not to a set of sales below normal value within that pattern for which there exists neither a textual nor a contextual basis in the second sentence. Therefore, when the W‑T comparison methodology is applied to "pattern transactions", zeroing is neither necessary to "unmask targeted dumping", nor permitted under the second sentence of Article 2.4.2.

5.157.  The United States recalls the Appellate Body's observation in US – Softwood Lumber V (Article 21.5 – Canada) that "[i]t could be argued … 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."^[351] According to the United States, an implication of this observation by the Appellate Body is that "it is possible to use zeroing 'to capture pricing patterns constituting "targeted dumping"'."^[352] Moreover, the United States observes that "the Appellate Body has never found that it is permissible to use zeroing in connection with the alternative, average‑to‑transaction comparison methodology set forth in the second sentence of Article 2.4.2, when the conditions for use of that methodology have been established, just as it has never found that it is impermissible to do so, because it has never had occasion to examine that issue."^[353]

5.158.  We have found above that under the second sentence of Article 2.4.2 an investigating authority may focus exclusively on "pattern transactions", while excluding from its consideration "non‑pattern transactions". This enables an investigating authority to "capture pricing patterns constituting 'targeted dumping'"^[354] without any need to resort to the use of zeroing in the application of the W‑T comparison methodology.

5.159.  We note that the application of the W‑T comparison methodology provided for in the second sentence of Article 2.4.2 was not at issue before the Appellate Body in US – Softwood Lumber V (Article 21.5 – Canada). In that dispute, the Appellate Body was addressing the second sentence of Article 2.4.2 in view of the Panel's contextual reliance on that provision. The Appellate Body focused on the effet utile of the second sentence of Article 2.4.2. The Appellate Body considered that to use zeroing under the comparison methodologies provided in the first sentence to "unmask targeted dumping" would have deprived the second sentence, whose function is to address "targeted dumping", of its effet utile. The Appellate Body, however, did not pronounce on how "targeted dumping" should be addressed under the second sentence of Article 2.4.2. In particular, the Appellate Body did not suggest that giving effect to the second sentence of Article 2.4.2 would require an investigating authority to establish margins of dumping by applying the W‑T comparison methodology with zeroing to the applicable "universe of export transactions", i.e. the "pattern transactions".

5.160.  We have concluded above that, under the second sentence of Article 2.4.2, dumping and margins of dumping pertaining to all export transactions of an exporter or foreign producer and to the product under investigation are limited to "pattern transactions". The exceptional W‑T comparison methodology in the second sentence of Article 2.4.2 requires a comparison between a weighted average normal value and the entire universe of export transactions that fall within the pattern as properly identified^[355] under that provision, irrespective of whether the export price of individual "pattern transactions" is above or below normal value. While the results of the transaction‑specific comparisons of weighted average normal value and each individual export price falling within the pattern will be intermediate results, the aggregation of all these results is required and will determine dumping and margins of dumping for the product under investigation as it relates to the identified "pattern". Zeroing the negative intermediate comparison results within the pattern is neither necessary to address "targeted dumping", nor is it consistent with the establishment of dumping and margins of dumping as pertaining to the "universe of export transactions" identified under the second sentence of Article 2.4.2.

5.1.9.1.2  Mathematical equivalence

5.161.  The United States further submits that, assuming that under both the W‑W and W‑T comparison methodologies the calculation of the margins of dumping is based on the same normal value and export sales data, if zeroing is prohibited under both comparison methodologies, then the results of the W‑W comparison methodology and the W‑T comparison methodology will be mathematically equivalent with respect to the total amount of all comparison results, the total amount of dumping, and the weighted average dumping margin for each exporter for the product under investigation.^[356] Korea, for its part, contends that the United States' argument on mathematical equivalence starts with the false assumption that the method of determining normal value must remain the same.^[357] According to Korea, the possibility of changing the normal value or the adjustments to the export prices breaks mathematical equivalence.^[358] Korea asserts that "[a]ny equivalence reflects assumptions about how the authority is making the comparison."^[359]

5.162.  We note that the United States' argument on mathematical equivalence is premised on its understanding of what constitutes the relevant "pattern" for the purposes of the second sentence of Article 2.4.2 of the Anti-Dumping Agreement. The United States submits that, "[o]n its face", the second sentence of Article 2.4.2 "contemplates a pattern of export prices that would transcend multiple purchasers, regions, or time periods"^[360] and that "[s]uch a 'pattern' necessarily includes both lower and higher export prices that 'differ significantly' from each other."^[361]

5.163.  We have concluded above that the "pattern of export prices which differ significantly" within the meaning of the second sentence of Article 2.4.2 comprises only a subset of all the export transactions and that these significantly different export prices are significantly lower export prices, which would be used by an exporter or producer to "target" purchasers, regions, or time periods. In this respect, the W‑W and W‑T comparison methodologies are designed to operate based on different "universes of export transactions": the first, based on all export transactions; and the second, based on individual export transactions that form the relevant "pattern". Accordingly, the second sentence of Article 2.4.2 allows an investigating authority to establish margins of dumping by comparing "pattern transactions" with normal value, while excluding from its consideration "non‑pattern transactions". Comparing normal value with "pattern transactions" only will not normally yield results that are mathematically or substantially equivalent to the results obtained from the application of the W‑W comparison methodology to all export transactions.^[362]

5.164.  The United States further submits that the Panel's approach of applying the W‑T comparison methodology to a subset of transactions without zeroing is in essence equivalent to the application of the W‑W comparison methodology to the same subset of transactions without zeroing. According to the United States, the Panel, in doing so, "effectively rewrote the second sentence of Article 2.4.2, changing it from allowing the application of the average-to-transaction comparison methodology under certain circumstances to allowing the application of the average‑to-average comparison methodology to a subset of transactions under certain circumstances."^[363] Thus, the United States contends that the Panel "invented a new methodology" and "read the average-to-transaction comparison methodology out of the second sentence of Article 2.4.2 of the AD Agreement altogether, contrary to the principle of effectiveness."^[364]

5.165.  The mere fact that the result of the application of the W-T comparison methodology to "pattern transactions" may be equivalent to the result of comparing the weighted average normal value with the weighted average export price of all "pattern transactions", is neither relevant under the second sentence that provides for the application of the W-T comparison methodology to "pattern transactions" only, nor does it read the W‑T comparison methodology out of the second sentence of Article 2.4.2. As we have explained above, the function of the second sentence of Article 2.4.2 is to allow an investigating authority to address "targeted dumping" by identifying a "pattern of export prices which differ significantly" and to which the W‑T comparison methodology is applied. Once the pattern of export prices within the meaning of the second sentence has been identified by the investigating authority, the fact that the application of the W‑T comparison methodology to that pattern of export prices leads to equivalent results as the application of the W‑W comparison methodology to the same pattern, neither undermines the effet utile of the second sentence of Article 2.4.2 of the Anti-Dumping Agreement, nor does it lead to equivalent results between the application of the symmetrical comparison methodologies normally used under the first sentence to the universe of all export transactions and the application of the W‑T comparison methodology used under the second sentence of Article 2.4.2 to the limited universe of "pattern transactions".

5.1.9.1.3  Negotiating history

5.166.  The United States also argues that the negotiating history of the Anti‑Dumping Agreement confirms that zeroing is permissible when applying the asymmetrical and exceptional W‑T comparison methodology set forth in the second sentence of Article 2.4.2 of the Anti‑Dumping Agreement.^[365] The United States refers to four documents in support of its arguments and submits that certain proposals from GATT Contracting Parties seeking changes to the Tokyo Round Anti‑Dumping Code addressed concerns about the use of an asymmetrical comparison methodology, in which negative dumping margins would be treated as zero instead of being added to the other transactions to "offset" the dumping margin.^[366] The United States considers that such proposals reveal that those GATT Contracting Parties viewed asymmetry and zeroing as one and the same problem.

5.167.  We have found above that under the second sentence of Article 2.4.2, an investigating authority can use the W‑T comparison methodology to identify and address "targeted dumping" by establishing margins of dumping based on the pattern of export prices which differ significantly and which are "targeted" at purchasers, regions, or time periods. We have also concluded that this exercise would allow an investigating authority not only to identify but also address "targeted dumping" without the need to have recourse to zeroing. We have, thus, reached the conclusion that zeroing under the W‑T comparison methodology is not allowed based on the text and context of Article 2.4.2 read in light of the object and purpose of the Anti-Dumping Agreement. We, therefore, consider that it is not necessary to have recourse to the negotiating history of the Anti‑Dumping Agreement in order to confirm the meaning of the second sentence of Article 2.4.2.

5.168.  Nonetheless, we address here the United States' arguments regarding the negotiating history of the Anti‑Dumping Agreement. We observe, first, that the negotiating proposals referred to by the United States reflect the positions of only some of the GATT Contracting Parties and not all.^[367] Second, the Appellate Body has already considered some of the exhibited materials to conclude that these materials did not resolve the issue of whether the negotiators of the Anti‑Dumping Agreement intended to prohibit zeroing.^[368] Finally, we observe that, even if these GATT Contracting Parties associated the asymmetrical comparison methodology with zeroing, this does not necessarily support a reading of the asymmetrical comparison methodology in the finally agreed version of the second sentence of Article 2.4.2 as permitting zeroing. Zeroing may have been considered by some GATT Contracting Parties as a possible way to "unmask" and address "targeted dumping". We note, however, from a reading of these documents that these GATT Contracting Parties were opposed to both zeroing and the inclusion of an asymmetrical methodology (with zeroing) in the Anti‑Dumping Agreement.

5.169.  We, thus, consider that these documents do not support a reading of the second sentence of Article 2.4.2 as permitting zeroing. On the one hand, they could be read, as the United States suggests, as supporting the view that the asymmetrical comparison methodology was associated with zeroing. On the other hand, they also could be read as explaining why the final version of the Anti‑Dumping Agreement included the second sentence of Article 2.4.2 as a compromise provision addressing "targeted dumping" by means of an asymmetrical comparison methodology, but without zeroing.

5.170.  While the text of the second sentence of Article 2.4.2 allows an investigating authority to focus on "pattern transactions" and exclude from its consideration "non‑pattern transactions" in establishing dumping and margins of dumping under the W‑T comparison methodology, it does not allow an investigating authority to exclude certain transaction‑specific comparison results within the pattern, when the export price is above normal value. As we have considered above, the second sentence of Article 2.4.2 allows an investigating authority to exclude from its consideration "non‑pattern transactions" and to establish dumping and margins of dumping based exclusively on a comparison of a weighted average normal value with all the identified "pattern transactions", in order to identify and address "targeted dumping". In so doing, however, the second sentence of Article 2.4.2 does not allow an investigating authority to exclude from the applicable "universe of export transactions" individual transactions that form part of the pattern, but that are priced above normal value.

5.171.  In light of these considerations, we uphold the Panel's findings, in paragraphs 8.1.a.xii and 8.1.a.xiv of its Report^[369], that "the United States' use of zeroing when applying the W‑T comparison methodology is inconsistent 'as such' with Article 2.4.2 of the Anti-Dumping Agreement" and that "the USDOC acted inconsistently with Article 2.4.2 of the Anti-Dumping Agreement by using zeroing when applying the W‑T comparison methodology in the Washers anti‑dumping investigation".

5.1.9.2  Whether the Panel erred in finding that the United States' use of zeroing when applying the W‑T comparison methodology is inconsistent "as such" and "as applied" in the Washers anti‑dumping investigation with Article 2.4 of the Anti‑Dumping Agreement

5.172.  We turn now to consider the United States' appeal of the Panel's findings under Article 2.4 of the Anti‑Dumping Agreement in respect of zeroing.

5.173.   The Panel recalled that the Appellate Body had previously upheld claims under Article 2.4 against the use of zeroing after finding that zeroing is inconsistent with the first sentence of Article 2.4.2 of the Anti-Dumping Agreement. In particular, the Panel cited the Appellate Body's findings in US – Zeroing (Japan) that, "[i]f anti‑dumping duties are assessed on the basis of a methodology involving comparisons between the export price and the normal value in a manner which results in anti‑dumping duties being collected from importers in excess of the amount of the margin of dumping of the exporter or foreign producer, then this methodology cannot be viewed as involving a 'fair comparison' within the meaning of the first sentence of Article 2.4."^[370]

5.174.  The Panel considered that the use of zeroing in the context of the W‑T comparison methodology would not lead to a "fair comparison" since individual "pattern transactions" priced above normal value would not be properly taken into account when an investigating authority has particular regard to the exporter's pricing behaviour within that pattern. Thus, the Panel found that the use of zeroing in the context of the W‑T comparison methodology is inconsistent "as such" with Article 2.4 and, for the same reasons, the USDOC acted inconsistently with Article 2.4 by using zeroing in the Washers anti‑dumping investigation.^[371]

5.175.  Having found the use of zeroing to be inconsistent with both Article 2.4 and the second sentence of Article 2.4.2, the Panel exercised judicial economy with regard to Korea's dependent claims under Articles 1 and 2.1 of the Anti‑Dumping Agreement and Article VI:1 of the GATT 1994.^[372]

5.176.  On appeal, the United States asserts that the Panel's findings of inconsistency regarding the use of zeroing in the application of the W‑T comparison methodology under Article 2.4 rest on its earlier findings of inconsistency under the second sentence of Article 2.4.2.^[373] Thus, the United States submits that, since the Panel's findings under Article 2.4.2 "are erroneous and should be reversed, the Panel's findings under Article 2.4, which are based on those … flawed findings, likewise are erroneous and should be reversed."^[374] Moreover, the United States submits that the second sentence of Article 2.4.2 provides means to "unmask targeted dumping" in exceptional situations and it is, thus, "fair" to take steps to "unmask targeted dumping" by faithfully applying the W‑T comparison methodology, when the conditions for its use are met. According to the United States, doing so is entirely consistent with the obligation that an investigating authority be impartial, even‑handed, and unbiased.^[375]

5.177.  In our analysis of Korea's appeal in respect of "systemic disregarding" under Article 2.4, we have noted that the introductory clause of Article 2.4.2 expressly makes this provision "[s]ubject to the provisions governing fair comparison" in Article 2.4. In that context, we have explained that, in respect of the second sentence of Article 2.4.2, the "fair comparison" requirement in Article 2.4 applies only in relation to "pattern transactions", which is also consistent with the function of the second sentence of Article 2.4.2 of allowing an investigating authority to identify and address "targeted dumping". We have considered that Articles 2.4 and 2.4.2 of the Anti‑Dumping Agreement not only inform each other, but must be read together harmoniously. Accordingly, we have found that the exclusion of "non‑pattern transactions" from the establishment of dumping and margins of dumping under the second sentence of Article 2.4.2 is consistent with the notions of impartiality, even‑handedness, and lack of bias reflected in the "fair comparison" requirement in Article 2.4.

5.178.  Bearing in mind the above considerations, we now turn to analyse the consistency of zeroing in the application of the W‑T comparison methodology with the "fair comparison" requirement in Article 2.4. We have concluded above that the second sentence of Article 2.4.2 allows an investigating authority to establish dumping and margins of dumping by applying the W‑T comparison methodology to "pattern transactions", while excluding from its consideration "non‑pattern transactions". We have also concluded that, in doing so, an investigating authority is not allowed to use zeroing within the identified "pattern" and, accordingly, we have upheld the Panel's findings in this regard.

5.179.  In EC – Bed Linen, the Appellate Body explained that "a comparison … that does not take fully into account the prices of all comparable export transactions – such as the practice of 'zeroing' … – is not a 'fair comparison' between export price and normal value, as required by Article 2.4 and by Article 2.4.2."^[376] Additionally, in US – Softwood Lumber V (Article 21.5 ‒ Canada), the Appellate Body considered that, since "the use of zeroing under the transaction-to-transaction comparison methodology artificially inflates the magnitude of dumping", it "cannot be described as impartial, even‑handed, or unbiased" and, accordingly, it does not "satisf[y] the 'fair comparison' requirement within the meaning of Article 2.4".^[377]

5.180.  Setting to zero the intermediate negative comparison results has the effect of not only inflating the magnitude of dumping, thus resulting in higher margins of dumping, but it also makes a positive determination of dumping more likely in circumstances where the export prices above normal value exceed those that are below normal value. Moreover, by setting to zero "individual export transactions" that yield a negative comparison result, an investigating authority fails to compare all comparable export transactions that form the applicable "universe of export transactions" as required under the second sentence of Article 2.4.2, thus failing to make a "fair comparison" within the meaning of Article 2.4.

5.181.  We disagree with the United States' contention that, given that the function of the second sentence of Article 2.4.2 is to "unmask targeted dumping" and that zeroing is necessary in the application of the exceptional W‑T comparison methodology to give effect to the second sentence, such an approach is entirely consistent with the obligation that an investigating authority be impartial, even‑handed, and unbiased. We have considered above that the effet utile of the second sentence in addressing "targeted dumping" is fulfilled once an investigating authority has identified the relevant "pattern" within the meaning of the second sentence of Article 2.4.2 and has established dumping and margins of dumping by applying the W‑T comparison methodology exclusively to "pattern transactions". In this respect, we have explained above that zeroing under the W‑T comparison methodology is not required in order for the second sentence of Article 2.4.2 to fulfil its function of allowing an investigating authority to identify and address "targeted dumping".

5.182.  In light of the above and given that we have upheld the Panel's findings on zeroing under the second sentence of Article 2.4.2, we also uphold the Panel's findings, in paragraphs 8.1.a.xiii and 8.1.a.xv of its Report^[378], that "the United States' use of zeroing when applying the W‑T comparison methodology is inconsistent 'as such' with Article 2.4 of the Anti‑Dumping Agreement" and that "the USDOC acted inconsistently with Article 2.4 of the Anti-Dumping Agreement by using zeroing when applying the W‑T comparison methodology in the Washers anti‑dumping investigation".