RESTRICTED

GENERAL AGREEMENT ADP/137

4 July 1995

ON TARIFFS AND TRADE Special Distribution

(95-1814)

Committee on Anti-Dumping Practices

 

 

 

EC - IMPOSITION OF ANTI-DUMPING DUTIES ON

IMPORTS OF COTTON YARN FROM BRAZIL

REPORT OF THE PANEL

 

Table of Contents Paragraphs

I. INTRODUCTION 1-4

II. FACTUAL ASPECTS 5-10

III. FINDINGS REQUESTED 11-14

IV. PRELIMINARY OBJECTIONS 15-70

Introduction 15-70

(a) Arguments relating to the Panel's terms 20-50

of reference

(b) Arguments relating to the scope of the Panel's 51-63

factual review

(c) Arguments relating to whether Brazil's request covered 64-70

the Provisional Determination

V. MAIN ARGUMENTS 71-438

V.1 Alleged violation of Article 2:4 71-116

(a) Alleged violation of Article 2:4; failure to take into 71-72

consideration the particular market situation prevailing

in Brazil

Arguments by the parties 73-109

(b) Violation of Article 2:4: incorrect determination that 110-111

sales were not made in the ordinary course of trade

Introduction

Table of Contents (Cont'd) Paragraphs

Arguments by the parties 112-116

V.2 Violation of Article 2:4 and 2:6: Failure to make adjustments to 117-301

ensure a fair comparison between normal and export price

(a) Arguments relating to the fundamental principle of the 121-122

Agreement

Arguments by the parties 123-178

(b) Arguments relating to the factual aspects of this case 179-197

(c) Arguments relating to lack of conformity with practice in 198-209

major anti-dumping jurisdictions

(d) Arguments relating to the alternative estimates presented 210-214

by Brazil

(i) Estimates of dumping margins presented by Brazil 215-239

(ii) Arguments relating to "negative dumping" or 240-252

"zeroing"

(e) Brazil's allegation that use of average monthly exchange 253-271

rates was not sufficient

(f) Arguments relating to the inability to consider method of 272-282

establishment of official Brazilian exchange rate and to

the principle of "monetary neutrality"

(g) Arguments relating to the EC's contention that the 283-286

adjustments requested were beyond the scope of the

Agreement

(h) Arguments relating to the likely effect of the methodology 287-294

on future proceedings, i.e. the "floodgates" argument

(i) Arguments relating to treatment of Brazil being contrary to 295-296

the treatment of other countries in the case

(j) Arguments relating to the importance of the economic 297-299

situation and a change in the criteria by the EC i.e.

"moving the goalpost"

(k) Brazil's allegation that the EC failed to exercise its 300-301

discretion

 

Table of Contents (Cont'd) Paragraphs

V.3 Determination of Injury 302-383

(a) Violation of Articles 3:1 and 3:2: Injury findings not based 304-305

on positive evidence, and failure to make an objective

examination of the relevant facts

Arguments of the parties 306-331

(b) Violation of Articles 3:2, 3:3 and 3:4: Lack of an 332-344

acceptable explanation of how the facts supported

the injury determination

(c) Violation of Article 3.2 in combination with Article 8.2: 345-346

Discrimination against Brazilian exporters in the injury

findings

Arguments by the parties 347-374

(i) Arguments relating to the relevance of Article 8:2 348-355

(ii) Alleged discriminatory treatment of imports 356-371

from Brazil

(iii) Alleged discrimination with respect to the exchange 372-374

rate used

(d) Violation of Articles 3:2, 3:3 and 3:4: Quotas agreed under 375-383

Arguments by the parties 377-383

V.4 Violation of Article 13: Failure to give due consideration to the 384-438

status of Brazil as a developing country

Arguments by the parties 388-438

(a) Arguments relating to the first sentence of Article 13 389-410

(b) Arguments relating to the second sentence of Article 13 411-419

(c) Arguments relating to the coverage of Article 13 420-434

(d) Arguments relating to discrimination 435-436

(e) Arguments relating to the alleged refusal of the EC to 437-438

consider its obligation arising from the MFA

the bilateral textile agreement precluded a finding of injury

 

Table of Contents (Cont'd) Paragraphs

VI. FINDINGS 438-593

VI.1 Preliminary objections of the EC that (1) claims raised by Brazil 438-466

were outside the terms of reference of the Panel; (2) arguments

raised by Brazil that required the Panel to exceed the scope of review

VI.2 Breach of Article 2:4 - failure to properly consider the particular 467-483

market situation prevailing in Brazil

VI.3 Violation of Article 2:6 - Failure to take into account distortions 484-502

arising from high domestic inflation combined with fixed exchange

rates iin comparing normal value and export price

VI.4 Articles 3:1 and 3:2, not basing the injury findings on positive 503-506

evidence, and not making an objective examination of the relevant

facts

VI.5 Violation of Article 3:2, 3:3 and 3:4: "... the facts stated in an 517-531

injury analysis must reasonably support the determination that

is reached

VI.6 Articles 3:2, 3:3 and 3:4 - Quotas agreed under the MFA 532-544

precluded a finding of injury

VI.7 Article 3:2 in combination with Article 8:2 - Discrimination 545-564

against Brazilian imports

VI.8 Article 13 - the EC had breached Article 13 of the Agreement 565-590

by not giving "special regard" to the "special situation" of Brazil

and not exploring the possibility of constructive remedies proposed

by Brazilian exporters

VI.9 Conclusions 591-593

 

I. INTRODUCTION

1. On 11 November 1991, consultations under Article 15:2 of the Agreement on Implementation of Article VI of the General Agreement on Tariffs and Trade (hereinafter "the Agreement") were held between Brazil and the European Community (hereinafter "EC") regarding the provisional anti-dumping measure imposed by the EC on cotton yarn from Brazil. Consultations on definitive anti-dumping duties by the EC on this item were held on 27 October 1993. The consultations failed to result in a mutually satisfactory solution to the dispute.

2. Brazil requested the Committee on Anti-Dumping Practices (hereinafter "the Committee") for conciliation under Article 15:3 of the Agreement (ADP/106) and a special meeting was held for this purpose on 20 December 1993 (ADP/M/43). There was no mutually satisfactory solution to the dispute and on 5 April 1994, Brazil requested the establishment of a panel under Article 15:5 of the Agreement (ADP/121). This request was considered by the Committee at its regular meeting on 26 April 1994. At that meeting, the Committee established a Panel as requested by Brazil (ADP/M/44).

3. On 17 June 1994, the Committee was informed by the Chairman in document ADP/125 that the terms of reference and composition of the Panel were as follows:

Terms of reference:

"To examine, in the light of the relevant provisions of the Agreement on Implementation of Article VI of the General Agreement, the matter referred to the Committee by Brazil in document ADP/121 and to make such findings as will assist the Committee in making recommendations or in giving rulings."

Composition:

Chairman: Mr. Crawford Falconer

Members: Mr. Mohan Kumar

Mr. Paul O'Connor

 

4. The Panel heard the parties to the dispute on 12 and 14 September 1994 and 9 and 10 November 1994. A finding regarding one preliminary objection by the EC was submitted by the Panel to the parties on 9 November 1994. The Panel submitted its complete findings and conclusions to the parties to the dispute on 12 June 1995.

II. FACTUAL ASPECTS

5. On 22 March 1990, the EC published a notice in the Official Journal of the European Communities that it was initiating an anti-dumping proceeding against imports of cotton yarn originating in Brazil, Egypt, India, Thailand and Turkey. The dumping investigation covered the period from 1 January 1989 to 31 December 1989, and the injury investigation covered the years 1986 to 1989.

6. On 23 September 1991, the EC imposed provisional anti-dumping duties on imports of cotton yarn originating in Brazil, Egypt and Turkey and terminated the proceeding in respect of India and Thailand. Definitive anti-dumping duties were imposed by the EC on 23 March 1992 on cotton yarn originating in Brazil and Turkey. The duties were not imposed on Egypt because a de minimis dumping margin was determined for the imports from Egypt.

7. In the case involving Brazil, nine exporters returned completed questionnaire responses to the EC and those exporters were deemed to be co-operating. Three exporters, Nisshinbo, Renda and Kanebo, were selected as a sample and a separate dumping calculation was made for each of them. The dumping calculation for the remaining six co-operating exporters was based on the weighted average results for the three sampled exporters. Dumping calculation for the non-co-operating exporters was based on the facts available.

8. During 1989 (i.e. the investigation period for dumping), the official exchange rate between the cruzado and the United States dollar was temporarily frozen by the Brazilian authorities for January, February and March. In April, May and June 1989, the cruzado was allowed to depreciate gradually. For the rest of 1989, the cruzado depreciated more freely against the dollar.

9. In its calculations the EC used the official cruzado to United States dollar exchange rates published by Fundacao do Instituto Brasileiro de Geograpfia e Estatistica (FIBGE), a public organization linked to the Brazilian Ministry of Planning. The EC made its dumping calculation for the three sampled exporters by comparing each export transaction with an average monthly normal value expressed in cruzados; for one company (Kanebo), the EC used actual returns in cruzado, and for the other two it used end-of-month official exchange rates for the dumping calculation.

10. For October, November and December 1989, the EC calculated normal values for Nisshinbo and Kanebo on the basis of costs of production, on the grounds that domestic sales in those months had not been made in the ordinary course of trade.

III. FINDINGS REQUESTED

11. Brazil requested the Panel to find that the EC, in imposing and maintaining anti-dumping duties on imports of cotton yarn from Brazil, had violated the following provisions of the Agreement:

(i) Article 2:4, by failing to consider the particular market situation prevailing in Brazil;

(ii) Article 2:4 by incorrectly determining that certain domestic sales were not made in the ordinary course of trade;

(iii) Article 2:6, by failing to effect a fair comparison between normal value and export price;

(iv) Articles 3:1 and 3:2, by not basing the injury findings on "positive evidence", and not making an objective examination of the relevant facts;

(v) Articles 3:2, 3:3 and 3:4, by not giving a reasonable explanation of how the facts supported the injury determination;

(vi) Articles 3:2, 3:3 and 3:4, by failing to take into account that quotas agreed under the bilateral textile agreement precluded a finding of injury, especially in light of the provisions of Article 13;

(vii) Article 3:2 in combination with Article 8:2, by discriminating against Brazilian exporters;

(viii) Article 13, by not giving due consideration to the status of Brazil as a developing country.

12. Brazil requested the following remedy: (i) a recommendation that Council Regulation (EEC) No. 738/92 be withdrawn as far as imports of cotton yarn originating in Brazil were concerned; and (ii) a recommendation that duties paid under Council Regulation (EEC) No. 738/92 (i.e. both provisional and definitive duties) be repaid.

13. The EC requested the Panel to recommend that the EC's definitive duty Regulation (EEC) No. 738/92 of 23 March 1992 (hereinafter the "Definitive Determination" or "Definitive Regulation") had not violated any provision of the General Agreement and the 1979 Agreement claimed by Brazil.

14. The EC also raised preliminary objections relating to the coverage of the Panel's terms of reference and to the scope of factual review by the Panel (see below).

IV. PRELIMINARY OBJECTIONS

Introduction

15. The EC raised certain preliminary objections regarding Brazil's submission to the Panel. First, the EC argued that two "claims" put before the Panel by Brazil were not covered by the terms of reference of the Panel. The objections pertained to the "claims" concerning: (i) the EC's determination that domestic sales of Nisshinbo and Kanebo in October to December 1989 were not in the ordinary course of trade; and (ii) "negative dumping margins" or "zeroing". The EC argued that the Brazilian submission on both these points were outside the terms of reference of the Panel, and as there was by now well-established case law on this matter, the EC requested the Panel to issue preliminary rulings on these objections and determine that the Brazilian claims on these points were not admissible and/or were outside the terms of reference of the Panel.

16. Second, the EC raised certain preliminary objections pertaining to the scope of factual review by the Panel. For "negative dumping margins", in addition to the preliminary objection mentioned in the paragraph above, the EC raised a preliminary objection regarding the scope of factual review by the Panel. Moreover, with regard to the scope of factual review, the EC also objected to two sets of estimates of dumping margins that had been submitted by Brazil to the Panel. One set contained data on dumping margins provided by the EC in its disclosure letter of 4 January 1992, allegedly arranged in a chronological order by Brazil. The second set contained two different estimates of dumping margins presented to the Panel by Brazil on the basis of recalculations using two different exchange rates, namely, (i) exchange rates indexed to domestic inflation and (ii) lagged exchange rate, i.e the exchange rate prevailing two months after the shipment. The EC argued that consideration of both these sets of estimates of dumping margins would require the Panel to go beyond the proper scope of review to be respected by the Panel and would require the Panel to enter into a de novo review. The EC did not insist on a preliminary ruling from the Panel on the two issues, but argued that the Panel would exceed its competence if it were to rule on these points because the task of the Panel was not to re-do the factual investigation. The EC argued that for a proper review of the EC's determination in this case, the Panel should examine, as mentioned in Article 8:5 of the Agreement, whether the factual basis of the findings stated in the determination were discernable from the text of the public notice and reasonably supported those findings.

17. The EC further argued that Brazil's request to the Panel did not cover the EC's determination imposing provisional anti-dumping duties in this case. Thus, whatever the legal findings in this case, the validity of the EC's provisional duty Regulation (hereinafter "Provisional Determination" or "Provisional Regulation") could not be affected.

18. Brazil argued that it had not raised new claims in its submission to the Panel. In support of this contention, Brazil referred to its submissions to the Committee on Anti-Dumping Practices relating to consultations, conciliation and request for establishment of a panel in this case (ADP/106, 113 and 121). Regarding the estimates of dumping margins presented to the Panel, Brazil argued that one set of data was only a chronologically re-arranged data provided by the EC in this case, and the other set was merely an illustration of how attempts to correct for the situation which arose due to the freezing of the exchange rate would have resulted in de minimis dumping margins for Brazil. Brazil argued that its request covered both the Preliminary and Definitive Determinations in this case.

19. The arguments of the two parties relating to the EC's preliminary objections are provided below in three sections. The first section summarizes their arguments relating to the preliminary objection by the EC with regard to the Panel's terms of reference. The second section summarizes their arguments relating to the scope of a panel's factual review, and the third section summarizes the arguments relating to whether Brazil's request to the Panel covered the Preliminary Determination also.

(a) Arguments relating to the Panel's terms of reference

20. The EC recalled that the terms of reference of the Panel were:

"To examine, in the light of the relevant provisions of the Agreement on Implementation of Article VI of the General Agreement, the matter referred to the Committee by Brazil in document ADP/121 and to make such findings as will assist the Committee in making recommendations or in giving rulings."

The EC argued that previous panel Reports had clearly established that "a matter", including each claim composing that matter, cannot be examined by a panel under its terms of reference unless that same matter and claim were within the scope of, and had been identified in, the written statement or statements referred to in its terms of reference.

21. The EC argued that the Brazilian exporters raised no objections to "negative dumping margin" or "zeroing" when the EC imposed provisional or definitive anti-dumping duties, and raised concerns regarding cost of production used as normal value for Nisshinbo and Kanebo at the stage of the imposition of definitive duties. Neither of these claims were raised by Brazil in the consultations and conciliation that took place in the period up to the establishment of the terms of reference of this Panel, and were not mentioned in ADP/121. Therefore, the Brazilian submission to the Panel on these points clearly exceeded the terms of reference of the Panel, which were based on ADP/121. Taking into account the established procedural principles of previous panel reports, the Panel should thus refrain from giving a ruling on these items. Making a ruling on these questions would amount to a violation of paragraph A.1 of the 1989 Improvements to the GATT Dispute Settlement Rules and Procedures, and would diminish the rights of the EC under the General Agreement and the Agreement.

22. The EC said that in order to ensure a proper functioning of dispute settlement, it was essential to distinguish between "the matter", "a claim" and "an argument". This distinction contributed to an effective functioning of the consultation and conciliation stages, while protecting the rights of the defending party at the panel stage. This was also shown by the particular nature of dispute resolution under GATT and the Agreement, which required that a mutually satisfactory solution be sought during consultations and conciliation before resorting to a panel. The EC referred in this context to paragraph 6 of the Understanding Regarding Notification, Consultation, Dispute Settlement and Surveillance, adopted on 28 November 1979 (BISD 26S/210, hereinafter the "1979 Understanding"), which provided that contracting parties should attempt to obtain satisfactory adjustment of the matter in accordance with the provisions of Article XXIII:1 before resorting to Article XXIII:2. The EC argued that pursuant to Article 15:7 of the Agreement, this understanding applied mutatis mutandis to consultations, conciliation and dispute settlement under the Agreement. However, such a mutually satisfactory solution could not be reached if all the problems were not identified during the consultation/conciliation stage. The respondent was entitled to be clearly aware of the scope of the dispute by the time the panel was established. The respondent would be denied the rights of defense, and the purpose of the dispute settlement rules would be frustrated, if part of the problems were revealed only at the panel stage. This concern had been addressed by a number of panels which had clarified existing provisions and developed a number of principles intended to ensure a proper functioning of the dispute settlement mechanism.

23. The EC said that "the matter" was the part of the original problem on which consultations were sought and which remained outstanding at the end of the consultation/conciliation stage. Thus, "the matter" was the sum of the claims referred to a panel in the document attached to its terms of reference. The EC did not consider that there was much controversy on the definition of "the matter" in general or in the present proceedings.

24. The EC then argued that "a claim" was a specific legal claim (e.g., an infringement of the Agreement, an error of fact or of interpretation of the facts committed by the investigating authority) in relation to certain facts. It was a legal reason for, or a ground upon, which a party considered that an obligation under the Agreement had been violated. Thus, there could be a number of claims relating to the violation of a particular provision of the Agreement, each one of them being independent of the others. "An argument" was legal or factual reasoning advanced to support, clarify or explain a claim, or to rebut a claim or an argument from the other side. An argument could support one or more claims if the factual basis upon which the claims were based was the same. The link between an argument and the matter referred to a panel was not direct. Even though this link could be presumed, it was clearly established only through the claim.

25. The EC argued that in this dispute, Brazil's allegations relating to negative dumping margin or "zeroing", and to the determination of normal value for two firms on the basis of costs of production in the last three months of 1989, constituted "claims" and not "arguments". The first amounted to stating that through recourse to so-called "zeroing" (i.e. the use of a particular methodology which treated negative dumping margins as zero in the calculation of the average dumping margins), the EC had violated the requirements of Article 2:4 and 2:6 of the Agreement to make adjustments to ensure a fair and proper comparison.

26. The EC argued that the second allegation, which pertained to the use of constructed normal value for two of the investigated companies for the last three months of 1989, was "a claim" relating to the violation of Article 2:4 of the Agreement. It was not related to the question of taking into account a particular market situation and thus had no connection with the other claim of Brazil concerning Article 2:4. Rather, it was based on an alleged methodological mistake in the calculation of costs of production in comparison to domestic sales price. The EC argued that Brazil had in fact made two distinct claims under Article 2:4 before the Panel, i.e., the claim that since domestic sales prices did not permit a proper and fair comparison, the EC should not have used costs of production but sales to a third country in order to establish normal value, and the (new) claim that the establishment of normal value on the basis of cost of production the EC used a wrong methodology as regards "financing costs".

27. The EC argued that pursuant to Salmon AD and Non-Rubber Footwear panel reports, a claim must at the very least be identified in the document referred to in the terms of reference of a panel in order to be properly put before that panel. In this case, this meant that each claim had to be made clearly and separately by the complaining party in document ADP/121. However, nowhere in document ADP/121 nor in other documents issued during the consultation/conciliation stage was there any trace of the claims of Brazil regarding which the EC had raised preliminary objections.

28. Therefore, the EC argued that both the allegations to which it had objected were "claims", that these "claims" were not identified in document ADP/121, and that they were not properly before the Panel. Therefore, the EC asked the Panel to issue a preliminary ruling that these "claims" were not covered by the Panel's terms of reference.

29. Brazil agreed with the EC on the need for certainty and clarity in procedural matters before a panel, and argued that there was a need to distinguish between a claim and an argument. While new claims may not be raised in a first written submission to the Panel, new arguments to support the claims could be so raised. Brazil argued that the preliminary objections raised by the EC were unfounded because they failed to make a clear distinction between a claim and an argument. What the EC was arguing to be "claims" by Brazil were in reality "arguments" supporting claims. In this regard, Brazil argued that the EC had itself stated in its first submission that "Brazil is raising new arguments" and that "These new arguments relate to .." (emphasis added by Brazil). Brazil considered that this language was effectively an admission by the EC of the true nature of the two points raised. They were arguments supporting claims that had been made throughout consultation, conciliation and in the request for the establishment of the panel. As such, they were clearly within the Panel's terms of reference. Brazil argued that the EC had subsequently also followed the same approach when it had suggested that a number of references to "arguments" were "typographical" errors and should be "clarified" as "claims" in order to support its new position. Brazil argued that far from being mere clarification or typographical corrections, these changes were changes of substance which reflected the EC's attempts to substantiate its (unfounded) arguments on the question of whether or not the two arguments concerned were within the Panel's terms of reference.

30. Brazil argued that the importance of the terms of reference was that the parties to the dispute must have been given an opportunity to reach a mutually satisfactory resolution of the matter, and sufficient notice must have been given to the defending party and other contracting parties that could be affected by the panel decision and the outcome of the dispute. Brazil argued that the purpose of consultation and conciliation was to clarify the facts and arguments in dispute, and to frame the dispute concerning the matter in terms which a panel may resolve. Brazil considered that both of these objectives had been met in this case. Brazil considered that the EC had been given every opportunity to reach a mutually satisfactory resolution of the matter, but had consistently shown itself unwilling to do so.

31. Brazil argued that "the matter" was the sum of the claims referred to in the request for the establishment of the Panel, i.e., ADP/121 in this case. An argument and a claim corresponded to an argument and conclusion in a syllogism. A syllogism consisted of two or more premises (or facts), a number of connecting statements (or arguments) and one or more conclusions (claims) deduced from the premises. Therefore, a claim was a statement about certain facts or legal provisions. On the other hand, an argument was a logical construction supporting a claim. For example, Brazil was claiming that the EC breached Article 2:4 of the Agreement because it used costs of production as the basis for establishing normal value in October, November and December 1989 for the Brazilian exporters Nisshinbo and Kanebo. In support of that claim, Brazil had argued that the EC should have had regard to the existence of the frozen exchange rate in selecting the method for determining normal value, and that the EC incorrectly determined that sales made by those companies in the relevant months were not made in the ordinary course of trade. Both these arguments led to the same conclusion, i.e. the claim that, in using costs of production as the basis for establishing normal value, the EC had breached Article 2:4 of the Agreement. Similarly, Brazil argued that it had made a number of arguments in support of the claim that the EC should have taken account of distortions arising from the unstable financial environment prevailing at the time of the investigation, particularly because of the facts of high domestic inflation and exchange rate freeze. These arguments included that the EC's "zeroing" methodology led, in these circumstances, to particularly strong distortions in the calculation. Thus, Brazil argued that all of the factual and legal matters before the Panel in this case were clearly within the Panel's terms of reference.

32. Brazil disagreed with the EC's statement that the essence of a claim was that it was not "related" to another claim, in the sense that the panel may accept one claim and reject another. Brazil argued that two arguments may also be "unrelated" in this sense, and thus the method proposed by the EC offered no reliable way of distinguishing between a claim and an argument. Brazil considered that, in reality, the question of the distinction between a claim and an argument was as much a question of fact, as it could be of law. The key to determining what, in any particular case, may be considered a claim and what an argument may be found in a consideration of the objectives of the GATT rules on admissibility. Brazil argued that the objectives of the rules on admissibility had been more than adequately fulfilled in this case.

33. Brazil argued that its statements contained in documents ADP/106, 113 and 121 showed that the issues regarding which the EC was presenting preliminary objections were addressed by Brazil during consultations, conciliation and in its request for the establishment of the Panel. Regarding the EC's determination of normal value on the basis of cost of production for certain companies in the last quarter of 1989, Brazil argued that it had stated in ADP/121 that "Normal value for the remaining sales was determined on the basis of cost of production. Brazil considers that this determination equally infringed Article 2:4 of the Agreement" (paragraph 18). Paragraph 16 of ADP/113 showed that the same claim was made by Brazil during conciliation. In that paragraph, Brazil had stated that "Normal value for the remaining sales was determined on the basis of costs of production. Brazil considers that this determination also infringed Article 2:4 of the Code." The claim was also raised during consultations, as shown by paragraph 2 of ADP/106, which referred to Brazil's view that the EC had violated Article 2:4 of the Agreement.

34. Regarding "zeroing", Brazil argued that this claim did not concern the EC's methodology of discounting negative dumping in general terms, i.e. in non-inflationary and stable exchange rate environments. Rather, it was concerned with the EC's failure to take into account the particularly distortive effects of that methodology in the circumstances prevailing in the case under review, namely high domestic inflation and frozen exchange rates. Brazil argued that this claim was contained in ADP/121 (paragraphs 13, 19 and 26), which stated that "in order to take account of the distortions arising from a situation whereby very high inflation on the domestic market was coinciding with a freeze in exchange rates", that the EC should have ignored "alleged dumping merely caused by temporary and unexpected exchange rate fluctuations", and that particularly in respect of exporting countries experiencing high inflation "the methodology adopted should permit a proper comparison". The same claim was made by Brazil during conciliation in ADP/113 (paragraphs 11 and 23), which stated that "allowance should have been made in order to take into account distortions arising from the artificial fixing of exchange rates ... or by simply ignoring alleged "dumping" merely caused by temporary and unexpected exchange rate fluctuations". Also, paragraph 17 of ADP/113 stated that "the methodology adopted should permit a proper comparison". During consultations this claim was raised in ADP/106, paragraph 2, which clearly referred to the EC's failure to address "the distortions arising from the prevailing exchange rate system".

35. Brazil argued that in view of the availability of ADP/113 and ADP/121, sufficient notice had been given to the EC and to other contracting parties of the nature of the matter to be placed before the Panel.

36. Brazil further argued that the correct legal basis for determining the limits of a panel's ability to review the factual and legal matters placed before it was the relevant GATT dispute settlement provisions. In this respect, Brazil recalled that the Agreed Description of the Customary Practice of the GATT in the Field of Dispute Settlement (Article XXIII:2) provided: "The function of a panel has normally been to review the facts of a case and the applicability of GATT provisions and to arrive at an objective assessment of these matters. ... The panel can question both parties on any matter which it considers relevant to the dispute ... Panels often consult with and seek information from any relevant source they deem appropriate . ... Panel reports have normally set out findings of fact, the applicability of relevant provisions, and the basic rationale behind any findings and recommendations that it has made" (paragraphs 3, 6(iv) and  6(v); emphasis added by Brazil).

37. Brazil also noted that the Understanding Regarding Notification, Consultation, Dispute Settlement and Surveillance provided: ".... a panel should make an objective assessment of the matter before it, including an objective assessment of the facts of the case and the applicability of and conformity with the General Agreement" (paragraph 16; emphasis added by Brazil). In this regard, Brazil noted that paragraph 3 of the 1989 Improvements to the GATT Dispute Settlement Rules and Procedures the Contracting Parties agreed that the existing rules and procedures of the GATT in the field of dispute settlement (i.e. those indicated above) shall continue to apply.

38. Brazil further argued that the issues to which the EC was objecting had been before the EC since before the adoption of provisional measures, and that the attitude of the EC had always been simply, to dismiss requests by Brazil without consideration. Brazil had not found any resolve on the part of the EC to facilitate a solution to the problems which had arisen in this case. In this context, Brazil referred to the 1966 Decision on Procedures under Article XXIII of the GATT, the recitals of which provide as follows : "Recognizing that the prompt settlement of situations in which a contracting party considers that any benefits accruing to it Directly or indirectly from the General Agreement are being impaired by measures taken by another contracting party, is essential to the effective functioning of the General Agreement and the maintenance of a proper balance between the rights and obligations of all contracting parties; Recognizing further that the existence of such a situation can cause severe damage to the trade and economic development of the less-developed contracting parties; and Affirming their resolve to facilitate the solution of such situations while taking fully into account the need for safeguarding both the present and potential trade of less-developed contracting parties affected by such measures; in conducting its examination and having before it all the background information, the panel shall take due account of all the circumstances and considerations relating to the application of the measures complained of, and their impact on the trade and economic development of affected contracting parties" (emphasis added by Brazil).

39. Brazil considered that it would be more in keeping with the spirit of these provisions if the EC would address the substance of the two arguments raised by Brazil.

40. The EC noted that Brazil had not made its allegation to the use of "zeroing" in general, but only its application in the context of an inflationary environment or of a freezing of exchange rate. The EC considered that these two allegations were distinct claims. The latter claim related to the appraisal of a factual situation, whereas the former dealt with the use of a methodology ("zeroing") which was applied in all cases, irrespective of the existence of an inflationary situation. These two claims had no relation to one another, apart from the fact that they related to the alleged violation of the same provisions of the Agreement. On that basis, the Panel was free to reject one of them while, at the same time, accepting the other.

41. Regarding "zeroing", the EC said that the citations by Brazil to relevant parts of the documents as referring to this allegation revealed that they all referred exclusively to the claims relating to whether the freezing of exchange rate led to a particular market situation. In none of these citations did Brazil make any reference to the methodology (i.e. "zeroing") used by the EC in this case. Brazil could not now contend that it had a claim relating to "zeroing" in its requests for consultation, conciliation and in document ADP/121. Moreover, given that the claim on the use of "zeroing" was different from the those identified in document ADP/121, one could not deduce from the text of that document that the claim on "zeroing" was implicitly included in the matter. In light of the purpose of consultations and conciliation, as well as the terms of reference of a panel (i.e. to identify among the issues discussed those claims which were still outstanding), the EC considered that there could not be any "implicit" identification of a claim in the document referred to in the mandate of the Panel.

42. The EC recalled Brazil's statement that paragraph 18 of document ADP/121 identified the claim pertaining to the use of constructed normal value for two of the investigated companies for the last three months of 1989. That allegation seemed to be based on the use of the term "cost of production", which appeared in that paragraph. However, paragraph 18 related to a totally different claim, i.e. that normal value should have been based on sales to third countries instead of domestic sales or costs of production. It did not relate at all to the claim which dealt with the issue of using an allegedly incorrect methodology in the calculation of the costs of production.

43. The EC then addressed the issue of the extent of the obligations of parties to an international agreement and its consequences on scope of review at a general level. The EC argued that it was one of the basic principles of international law that the liberty of subjects of international law was presumed to be unlimited, except by obligations stemming from international norms binding on them. This principle dated back to the very origin of international legal relations and had been expressly confirmed as early as 1927 by the Permanent Court of International Justice in the "Lotus" case. Since international law did not create competence for parties but rather limited their original freedom through specific obligations, limits to the liberty of international actors could not be presumed (for example, "Lotus" case cited above). They must come from international obligations, as interpreted in conformity with generally recognized principles of interpretation of international law. Therefore, it was clear that the liberty for GATT contracting parties to take anti-dumping measures does not result from the GATT, nor from the Agreement. It pre-existed these agreements. This right was acknowledged in a number of provisions of the General Agreement. Article II:2(b) recognized the right for contracting parties to raise duties beyond their bound rate when applying anti-dumping measures, and Article VI:1 confirmed the legitimacy of anti-dumping measures. Therefore, anti-dumping was one of the field of activities of international subjects where there existed a total freedom to act, unless this freedom was limited by international obligations. In the present case, if no obligation was created under the Agreement, then the EC was free to proceed as it wished (provided it did not act contrary to other obligations under international law), i.e., the above-mentioned principle implied that Brazil's contention according to which a practice that was not expressly allowed by the Agreement was forbidden could not be accepted as a matter of international law. The Agreement was not drafted with the view to be comprehensive.

44. The EC argued that the interpretation of the Agreement should be made on the basis of the generally accepted principles of interpretation of international agreements mentioned in Article 31 of the Vienna Convention on the Law of Treaties. In other words, either the obligation stemmed from the actual text of the Agreement, or it could be deduced from that text on the basis of an interpretation made pursuant to the principles recalled in Article 31. The burden of proof was initially on the investigating authority to establish the existence of injurious dumping. This burden of proof may be discharged by a publication of the findings, including a sufficient statement of reasons. Once this had been done, the obligation to discharge the burden of proof shifted upon the party which found itself aggrieved and requested the establishment of a panel. If the complaining party claimed that the investigating authorities of another party had disregarded their obligations under the Agreement, it would have to establish the existence of the obligations, the extent to which the investigating authorities of the other party were subject to the alleged obligations, as well as the manner in which they had infringed such obligations.

45. The EC argued that there was no trace in the Agreement of the obligations which, according to Brazil, were allegedly borne by the EC in the present case. Moreover, even if it was an obligation under the Agreement that the investigating authorities take into account all aspects that Brazil was claiming, such as the particular treatment to be accorded to exchange rate fluctuations (quod non), the Agreement did not specify any instructions on how such obligations should be fulfilled. This implied that, even if Brazil were correct about its claims regarding the existence of certain obligations, the EC would remain free to implement them in the way it found the most appropriate to satisfy the requirements of the Agreement.

46. The EC argued that the above-mentioned principle had consequences for the work of panels and, in particular, for the scope of their review. Firstly, the task of panels should be limited to reviewing the conformity of the measures at issue with the obligations under the Agreement. Secondly, in carrying out their review of the conformity of the measures with the Agreement, panels were only asked to verify whether the determinations made by the investigating authority were based on an examination of all relevant facts and that the factual basis for a determination was discernible on the basis of the statement of reasons. Panels were consequently not asked to make their own independent evaluation of the facts before the anti-dumping authorities on whether the conditions for imposing anti-dumping duties contained in the Agreement were fulfilled or otherwise to substitute their own judgement as to the sufficiency of the particular evidence considered by the investigating authorities. Quite a number of panel reports had confirmed the above interpretation of the task of panels. Pursuant to this well established interpretation, the panel could review, on the basis of the reasoned determination, whether the facts which were taken into account by the investigating authority were all the facts relevant to the determination it had made. Thus, it was not enough for the investigating authority to mention, in its statement of reasons, "a sufficient number of accurate facts". It must be possible to identify, on the basis of the determinations, whether all relevant facts were considered.

47. The EC argued that panels were only asked to review the conformity of the determinations as made by the investigating authorities for consistency with the Agreement, and hence it was logically not within their task to consider whether another (or better) option was available than the one followed by the investigating authority, provided the latter was consistent with the Agreement. Moreover, the final consequence of the principle of freedom of subjects of international law in the absence of specific obligations limiting their liberty was that, where the Agreement was silent, investigating authorities were free to apply their anti-dumping legislation as they thought appropriate. However, given that the application of this legislation takes place within the context of an international framework on anti-dumping, the EC considered that there was a minimum obligation not to nullify the purpose of the Agreement when making determinations not otherwise subject to particular requirements under the Agreement. In such circumstances, the Panel was entitled, in the opinion of the EC, to determine whether the investigating authority had made a manifest error of fact or of interpretation of the facts or acted arbitrarily so that it nullified, by doing so, the purpose of the Agreement.

48. On the question of the scope of the Panel's review, Brazil argued that the objective of the EC's arguments was to restrict the Panel's ability to conduct a meaningful assessment of the matter placed before it by the Parties. Moreover, the Agreement was not silent, and even if it were, it did not follow that the parties were free to do anything. Contracting parties were required to ensure that they applied the Agreement in accordance with its general principles and in accordance with other relevant principles of international law. In Brazil's view the Panel was fully competent to examine all of the factual and legal issues that have been placed before it in this case.

49. Addressing the EC's argument that the Agreement was silent on the question of exchange rates, and that under international law "limits to the liberty of international actors cannot be presumed" (paragraphs 43 to 45 above), Brazil said that these arguments overlooked the fact that Brazil was not arguing that the EC should have used one exchange rate or another. Brazil's argument was that the Agreement imposed certain clear legal obligations, particularly the obligations that there be a fair and proper comparison between normal value and export price, with special regard to the special situation of developing countries. These were clear and objective obligations arising under the Agreement, which the EC had violated in this case. Brazil's arguments did not therefore require any "presumption" of obligations arising under the Agreement.

50. Brazil argued that it was not Brazil's view that a practice that was not expressly allowed by the Agreement was forbidden, nor did Brazil consider that pre-GATT anti-dumping actions were contrary to international law. These views were not even implied by any of the arguments made by Brazil. Brazil argued that it had clearly identified the provisions of the Agreement which it considered had been violated, as well as the factual circumstances of the relevant violations. In so doing, Brazil had more than discharged the burden of proof which may fall on it, either in the original investigation or during the course of the panel procedure.

(b) Arguments relating to the scope of the Panel's factual review

51. The EC argued the requirements under Article 8:5 were to set forth the findings and conclusions reached on all issues of fact and law considered material and the reasons and basis therefor. In view of this requirement, a proper review of the EC's Definitive Determination in this case meant that the Panel should examine whether the factual basis of the findings stated in the Determination was discernible from the text of the public notice and whether such factual basis reasonably supported those findings. In carrying out its review, however, the Panel should normally not be allowed to conduct a de novo review of the evidence relied upon by the EC or otherwise to substitute its own judgement as to the sufficiency of the particular evidence considered by the investigating authorities of the EC. To do so would ignore the principle that the task of the Panel was not to make its own independent evaluation of the facts before the EC on whether the conditions for imposing anti-dumping duties contained in the Agreement were fulfilled, but to review the definitive determination as made by the EC for consistency with the provisions of the Agreement. All the Panel needed to do was to satisfy itself that there was a sufficient reasoning in the EC's final determination as to the connection between the factual basis stated in that determination and the legal findings contained in it, and that the EC authorities had not relied upon incorrect factual information in making these findings.

52. The EC further argued that there were some limits as to what should be mentioned in the statement of reasons. While there should be sufficient information regarding the elements upon which the investigating authority reached its findings in order to allow the Panel to undertake an effective review, investigating authorities were not required by Article 8:5 to mention in the published statement of reasons each and every argument discussed in the investigation. If it were otherwise, this would make procedures administratively unmanageable. More important, such a degree of detail was unnecessary for the purpose of panel review and accordingly was not required by the Agreement.

53. The EC argued that the estimates of dumping margins and certain other documents submitted by Brazil as annexes to its main submissions to the Panel were of no relevance for the work of the Panel, which should base its determinations on the published findings of the investigating authorities. In fact, the documents submitted by Brazil (in particular "correspondence with the EC authorities") were part of the record of the case and were reviewed by the investigating authorities in due time in order to make their determinations. If the Panel were to review these documents with the intention of obtaining more than a confirmation of the findings contained in the published determinations, this would amount to a de novo examination, which the Panel was not entitled to carry out.

54. Brazil argued that while anti-dumping determinations must satisfy the requirements stated in Article 8:5, that Article did not represent an exhaustive statement of the role of a panel under the GATT dispute settlement procedures. It was illogical to start from a particular requirement, such as that contained in Article 8:5, and use it as the basis for general propositions about the limits of a panel's ability and duty to review the factual and legal matters placed before it. Brazil argued that according to the EC's interpretation, a correctly reasoned determination based on a sufficient number of accurate facts, would not be reviewable by a panel even in the presence of other facts, not referred to in the determination, which indicated that the determination violated the Agreement. Such a proposition (1) was contrary to the relevant GATT dispute settlement provisions; (2) was not supported by the case law; (3) would allow anti-dumping authorities undue opportunities to insulate their determinations from review by a panel; (4) would emasculate the GATT dispute settlement procedures in a manner not intended by the contracting parties; and (5) would almost certainly not be welcomed by the EC itself.

55. Brazil further argued the specific remarks contained in the Panel reports in question concerned only injury determinations and could not form the basis for the general propositions made by the EC regarding the limits of a panel's ability to review all legal and factual issues before it. Brazil argued that in the relevant Panel decisions referred to by the EC, the Panel had in fact concluded that claims at issue were within its terms of reference. Brazil noted that the authority cited by the EC suggested that a claim was covered by a request for the establishment of a panel, "however characterized" in that document. Furthermore, if the relevant reference in the request for the establishment of a panel could "reasonably be interpreted" as covering the claim in issue, the claim should be considered as within the panel's terms of reference. Also, the Improvements to the GATT Dispute Settlement Rules and Procedures provided that a request for a panel shall provide "a brief summary of the factual and legal basis of the complaint". Brazil considered that these requirements had been met in this case. Thus, the preliminary objections raised by the EC in relation to the Panel's terms of reference were unfounded and should be rejected.

56. Brazil argued that the paragraphs of the Resin panel report that had been referred to by the EC (i.e. paragraphs 226 to 228; please see reference in footnote 16 above) were concerned with an entirely different matter, namely the interpretation of Article 3:1 of the Agreement, which related to the determination of injury. Paragraph 226 was a summary of the arguments of the parties and contained no findings of the panel. It was therefore irrelevant to the present issue under consideration. Paragraph 227 was concerned with the United States' claim that the relevant Korean authorities (KTC) had not carried out an objective assessment of the factors they were required to consider under Article 3 of the Agreement. Under Article 3, a number of different factors must be weighed and taken into consideration before making a determination on injury. In paragraph 227, the report stated that: "The Panel considered that a review of whether the KTC's determination was based on positive evidence did not mean that the Panel should substitute its own judgement for that of the KTC as to the relative weight to be accorded to the facts before the KTC. To do so would ignore that the task of the Panel was not to make its own independent evaluation of the facts before the KTC to determine whether there was material injury to the industry in Korea but to review the determination as made by the KTC for consistency with the Agreement, bearing in mind that in a given case reasonable minds could differ as to the significance to be attached to certain facts." (emphasis added by Brazil) Thus, Brazil argued that in its reference to paragraph 227 of the Resin panel report, the EC had not mentioned the fact that this paragraph was concerned with the weighing of different factors in the context of an injury determination. Paragraph 227 was certainly not persuasive authority for the far reaching general propositions put forward by the EC in its preliminary objections.

57. Brazil argued that paragraph 228 of the Resin panel report was also concerned with an entirely different matter, namely the attempts by Korea to refer to extraneous materials in order to meet the argument raised by the United States that there were insufficient findings and reasoning in its determination. Korea could not in this way retroactively cure the breach of Article 8:5 of the Agreement. Paragraph 228 of the report therefore had no bearing on the arguments raised by the EC in its preliminary objections.

58. Brazil noted that the EC had also referred to paragraph 295 the report of the panel on Milkpowder, and argued that the cited paragraph dealt with precisely the point raised in paragraph 228 of the Resin panel report. Hence paragraph 295 of the report of the panel on Milkpowder also had no bearing on the arguments raised by the EC.

59. In the light of the above, Brazil considered that the authorities cited by the EC, namely Article 8:5 of the Agreement, and the Resin and Milkpowder panel reports, offered no support for the EC's argument that certain facts contained in Brazil's first written submission may not be considered by the Panel, and for the far reaching propositions made by the EC concerning the limits of the Panel's ability to review the factual matters placed before it.

60. Brazil further argued that the facts were not in dispute in this case. The effect of using the frozen exchange rate had been clear to all parties since well before the imposition of provisional measures, and had been repeatedly drawn to the attention of the EC. The EC did not dispute these effects, but was taking the position that, as a matter of principle, it was not required to take them into consideration. Therefore, Brazil was not asking the Panel to enter into a reassessment of the facts. Rather, Brazil was presenting to the Panel its legal arguments that, on the basis of the undisputed facts, known to both parties and accepted by them, the EC's determinations were not made in conformity with the Agreement.

61. Brazil argued that the data that it had presented to the Panel should not be called into question, since these were identical to or based on the figures provided by the EC in its disclosure letter. The integrity of those data could be easily demonstrated by comparing them with the data in the EC's disclosure in chronological order.

62. In the light of the above, Brazil considered that it was clear that the Panel was empowered to consider all of the factual arguments placed before it in this case.

63. With respect to the allegations of Brazil according to which the two panels reports cited by the EC in its first submission would not be relevant because they dealt with the weighing of facts in an injury determination, the EC argued that the findings of these two panels as regards the standard of review by a panel were of general applicability for all aspects of anti-dumping procedures (i.e. they were relevant also for dumping determinations, which involved choices such as, in the determination of normal value under Article 2:4, the selection between costs of production and sales to third countries).

(c) Arguments relating to whether Brazil's request covered the Provisional Determination

64. The EC argued that Brazil's request to the Panel did not cover the EC's determination imposing provisional anti-dumping duties in this case, because Brazil had not mentioned the Provisional Determination in the paragraphs summarizing its complaint in both its first submission to the Panel as well as its request for the establishment of the Panel. Therefore, the EC argued that whatever the legal findings in this case, the validity of the EC's Provisional Determination could not be affected.

65. Brazil argued that the scope of the Panel's review extended to both the Provisional Determination (Regulation 2818/91) and Definitive Determination (Regulation 738/92). In fact, Regulation 2818/91 was actually annexed (as Annex 1) to Brazil's first written submission, and that submission expressly referred to that Regulation. Furthermore, Brazil had also stated in that submission that "[i]t is Brazil's opinion that the anti-dumping duties imposed are not in conformity with the provisions of the Agreement ... ", and had referred to "the measures taken by the EC ...". These statements referred to both Regulation 2818/91 and Regulation 738/92. Furthermore, throughout its first written submission Brazil referred explicitly to, or used quotations from, Regulation 2818/91, and made arguments aimed directly at the EC's provisional determinations. In addition, the two Regulations were so inextricably linked, and the claims raised by Brazil so clearly directed at the foundation of the EC's methodology, that it was self evident that the EC's determinations in the two Regulations, on the points which formed the subject matter of Brazil's claim, must stand or fall together.

66. Brazil noted that the EC had referred to paragraph 13 of Brazil's first submission to argue that the Provisional Determination had not been mentioned in that paragraph, and therefore it was not part of the Brazilian complaint. Brazil argued that in that specific paragraph referred to by the EC, Brazil was not concerned with defining the scope of its claim because that had already been done. Rather, Brazil was concerned with identifying the legislative acts that it considered should not be applied to the Brazilian exporters, and the duties it considered should be repaid. Therefore, in that paragraph, Brazil had stated its view that Regulation 738/92 should not be applied to the Brazilian exporters, and that all duties paid under Regulation 738/92 should be reimbursed. Regulation 2818/91 was not mentioned in that paragraph because it was no longer in force and could not therefore be applied in any event to the Brazilian exporters. The request for reimbursement included all the amounts secured by way of provisional anti-dumping duty under Regulation 2818/91, but which were not paid until definitively collected. This was clear from the fact that, in the absence of any definitive measures, provisional measures would lapse and no duties would be payable. Thus, Regulation 2818/91 was irrelevant to the points being raised in the paragraph mentioned by the EC and the fact that no reference to it appeared in that particular paragraph had no bearing on the scope of the Panel's review.

67. Brazil argued that Regulation 2818/91 was expressly referred to in its request for the establishment of the Panel. Document ADP/121 at paragraphs 3, 9 and 48 related to both Regulation 2818/91 and Regulation 738/92. ADP/121 contained numerous arguments aimed directly against the determinations contained in Regulation 2818/91, which was inextricably linked with Regulation 738/92, and paragraph 11 of ADP/121 was analogous to the paragraph mentioned by the EC.

68. Brazil argued that paragraph 3 of ADP/113 (i.e. Brazil's request for conciliation under Article 15:3 of the Agreement) referred explicitly to the Provisional Regulation, and paragraphs 9 and 13 referred to the duties imposed and the anti-dumping "action" as a whole, i.e. to both the provisional and definitive measures. ADP/113 contained numerous references to the "measures" and to arguments aimed directly against the determinations contained in Regulation 2818/91, which was inextricably linked with Regulation 738/92. At paragraph 46, ADP/113 quoted directly from recital 46 of Regulation 2818/91. Moreover, as stated in paragraph 4 of ADP/113, consultations were held between Brazil and the EC under Article 15:2 of the Agreement on 14 November 1991. These consultations related only to the Provisional Regulation, which had been adopted by the EC on 23 September 1991. They did not relate to the Definitive Regulation, which was not adopted until 23 March 1992.

69. Brazil argued that, in the light of the above, the argument raised by the EC to the effect that the scope of the Panel's review was limited to the EC Regulation imposing definitive measures should be rejected. The Provisional Regulation had been subject to consultation and conciliation and was clearly identified in Brazil's request for the Establishment of the Panel and in Brazil's first written submission.

70. The EC argued that Brazil had ignored the fact that the provisional duty measures were neither contested as such in the consultations nor in the conciliation leading to the establishment of this panel. They did not figure in ADP/121 or in the first submission to the Panel by Brazil. Moreover, the references made to the Provisional Determination by Brazil were for the purpose of getting information only. Brazil had never clearly asked the Panel to find that these measures were taken in violation of the Agreement. Therefore, even if the Panel were to find in favour of Brazil on the definitive duty regulation (which the EC was sure the Panel could not), it could not recommend reimbursement of duties for the provisional duty measures.

V. MAIN ARGUMENTS

V. 1. Alleged violation of Article 2:4

(a) Alleged violation of Article 2:4: failure to take into consideration the particular market situation prevailing in Brazil

Introduction

71. Brazil argued that the EC had violated Article 2:4 of the Agreement because it had failed to take into consideration the particular market situation that prevailed in Brazil during the investigation period, and therefore calculated its dumping margin on the basis of a normal value that did not provide for a proper comparison. Brazil argued that the EC should have used sales to third markets as the basis for normal value in view of the particular market situation in this case.

72. The EC argued that Brazil was not correctly interpreting the term "particular market situation" in Article 2:4. This term pertained to only the domestic market situation of the exporting country, and the EC had met all the requirements under Article 2:4 for selection of the normal value. Therefore, this claim of Brazil should be rejected by the Panel.

Arguments by the parties

73. Brazil argued that the phrase "particular market situation" in Article 2:4 included the relevant situations external to the domestic market, such as exchange rates, which affect price comparability. Brazil argued that due to the exchange rate freeze in Brazil during the first quarter of 1989, the particular market situation in Brazil was such that under Article 2:4, only sales to third countries could be used as the normal value for determining the dumping margin. Brazil argued that the EC had violated Article 2:4 by not relying on normal value based on sales to third countries because the other methodologies for normal value provided under Article 2:4 would not have eliminated the effect of the distortions in the market.

74. Brazil argued that at the beginning of the investigation period, Brazil was experiencing a deep economic crisis, including very high inflation. Urgent economic policy measures of a general nature were therefore needed and were applied in a manner consistent with Brazil's obligations under the IMF, GATT and the Agreement. In particular, in the first quarter of 1989, the Brazilian Government froze exchange rates in order to decrease the money supply and thereby control inflation. However, domestic prices continued to rise, while export earnings converted into Cruzados remained stable. This evidently led to a gross distortion in the comparison between domestic and export prices.

75. Brazil argued that the exchange rate situation for Cruzado affected the dumping calculation for two reasons, i.e. temporary fluctuation and rapid depreciation of the Cruzado. The temporary exchange rate freeze generated artificially high dumping margins because the overvalued cruzado during the exchange rate freeze would tend to produce a higher dumping margin. Brazil said that if this argument were true, then higher dumping margins would have been found by the EC in the first six months of 1989 and low or zero dumping margins in the second half of 1989. A consideration of the chronologically arranged dumping margins showed that dumping margins had this characteristic.

76. Brazil argued that in a high inflation environment it was often logical for exporters to set export prices at a level which anticipated depreciation of the domestic currency, particularly if there was a temporary exchange rates freeze. The exporter had a reasonable expectation that by the time he would be paid the exchange rate would have returned to its normal level. Brazil argued that in such a situation, it would be reasonable to use the exchange rate prevailing at a later date, i.e. use a lagged exchange rate (Brazil's and the EC's arguments on this point are mentioned in greater detail in Section V.2).

77. Brazil emphasised that the overriding principle of Articles 2:4 and 2:6 of the Agreement, reiterated throughout the text of the Agreement, was that the methodology adopted should permit a proper comparison. Brazil believed that this fundamental principle had been violated in this case. Brazil argued that to ensure a proper comparison between normal value and export price, the EC should have taken further steps by acknowledging the particular market situation prevailing in Brazil and basing normal value on sales to third countries, or adjusting the normal value based on domestic data, or adjusting the exchange rate used.

78. Brazil argued that the EC's determination in this case was based on the legal rule that in determining the correct method for establishing normal value in the context of Article 2:4, the investigating authorities may have regard only to the circumstances prevailing on the domestic market. This rule, whose application resulted in the investigating authorities determining the existence of the exchange rate freeze was irrelevant to the operation of Article 2:4, was not in conformity with the Agreement because it was not a correct interpretation of Article 2:4. Article 2:4 was concerned with "the particular market situation" which was not in any way limited to the domestic market. Moreover, Article 2:4 was concerned with ensuring a proper comparison. Since this could only be understood as comparison with the export price, the contracting parties were required to consider in any selection of normal value whether or not the method of selecting the normal value would permit a proper comparison with the export price. It followed therefore that where the characteristics of the export market were such that a proper comparison with the normal value could not be achieved on the basis of one method provided for in Article 2:4, but could be reached by the use of an alternative method provided for in that Article, the alternative method was to be preferred.

79. Brazil argued that Article 2:4 provided that where the particular market situation prevailing in the exporting country did not permit a proper comparison, the margin of dumping had to be determined on the basis of a normal value other than the domestic price, i.e. on the basis of cost of production, or sales to third country markets. Brazil argued that during the first half of the investigation period, there was a particular market situation prevailing in Brazil within the meaning of Article 2:4 of the Agreement, as a result of which the establishment of normal value on the basis of domestic sales of the like product did not, in the exceptional conditions prevailing, permit a proper comparison with export prices. Despite this, the EC determined normal value during this period on the basis of domestic sales of the like product, without giving due consideration to the special circumstances prevailing, and thus infringed Article 2:4.

80. Brazil argued that the rule applied by the EC in this case was not in conformity with the EC's own legislation and established practice. The EC's anti-dumping Regulation provided that alternative methods for establishing normal value may be considered where "for any reason" domestic sales did not form a proper basis for determining the existence of dumping. Furthermore, in its anti- dumping investigations, the EC considered that if the volume of an exporter's domestic sales of the like product represented less than 5 per cent of the volume of its export sales, the appropriate method for determining normal value was not domestic sales, but costs of production. Hence, the EC itself interpreted Article 2:4 as referring to situations outside the domestic market, and as permitting investigating authorities to have regard to characteristics of the export market in selecting the method for establishing the normal value.

81. Brazil then addressed the EC's argument in the EC's conciliation statement that the "question of the exchange rate evolution in relation to domestic inflation is a completely separate matter that has nothing to do with the determination of whether or not domestic market conditions and price mechanisms are an appropriate basis for the determination of normal value." Although Brazil agreed that the determination of normal value and the comparison of normal value and export prices were distinct stages in the dumping calculation, Brazil argued that Article 2:4 of the Agreement was concerned with the selection of the method of establishing normal value and with ensuring a fair comparison between normal value and export price. It was not possible to make such a comparison without using an exchange rate, and the selection of the exchange rate, particularly in the light of domestic inflation, was critical for the comparison. Article 2:4 therefore recognized the link between the establishment of normal value and evolution of exchange rates. The "particular market situation" referred to in Article 2:4 included precisely the situation in which the Brazilian exporters found themselves: high inflation and frozen exchange rates. The Agreement established an obligation to take into account distortions arising from temporary exchange rate fluctuations, especially in the context of proceedings concerning exports from developing countries. If the interpretation of Article 2:4 by the EC were correct, Article 2:4 should refer to the "particular market situation in the domestic market" and to the need to ensure "a proper establishment of normal value". However, neither of these statements were contained in Article 2:4.

82. Brazil also recalled that from October to December 1989, the EC determined normal value for Nisshinbo and Kanebo on the basis of cost of production, on the grounds that domestic sales in those months had not been made in the ordinary course of trade. Brazil argued that this determination was inconsistent with Article 2:4 because in this case the nature of the particular market situation (high inflation combined with a frozen exchange rate) had the effect of making cost of production equally unreliable as a basis for normal value. The use of cost of production as the basis for normal values for Nisshinbo and Kanebo for October to December 1989 artificially inflated the dumping margin for those companies. Brazil argued that the use of normal value based on sales to third countries would have eliminated the effect of the distortions introduced by the particular market situation.

83. The EC argued that its determination was consistent with the requirements of Article 2:4. Exchange rates were not the subject of Article 2:4, and the phrase "particular market situation" included external factors only to the extent that they affected domestic sales. No argument had been presented in this case that external factors had had such an effect. Also, the phrase "particular market situation" in Article 2:4, interpreted in the light of its object, purpose and context, did not have the meaning attributed to it by Brazil. There was nothing in the text of Article 2:4 to suggest that the phrase "particular market situation" was meant to cover high inflation in the domestic market and "freezing" of currency exchange rate. The drafting history of Article 2:4.,to the extent it existed, did not support Brazil's claim either. The phrase "particular market situation" clearly referred to domestic sales and to the prices at which they were made, and it was on these sales and prices that a "particular market situation" must have an impact before a decision could be made that these prices were unusable. If there was such an impact, Article 2:4 permitted a comparison either with a comparable price of a like product when exported to any third country or with the cost of production in the country of origin plus a reasonable amount for sales, general and administrative expenses and for profits. Even if the domestic sales were not appropriate as normal value, the Agreement provided a choice, and not an hierarchy, between two alternative methodologies for determining normal value. The EC's usual practice in the situation when domestic sales cannot be used as basis for normal value was to resort to the method based on costs of production, not export prices to third countries.

84. The EC argued that in the case under consideration the EC based its determined normal value in accordance with Article 2:1 of the Agreement, i.e. on the comparable prices, in the ordinary course of trade, for the like products destined for consumption in Brazil. With the exception of Nisshinbo and Kanebo for which normal value for the last three months of 1989 was constructed using a method based on the cost of production, these prices were found to be at arm's length between independent parties under competitive conditions in a market economy. In this context, the evolution of the exchange rate of the cruzados against the currency in which exports were made had no direct bearing on the functioning of the domestic market, where transactions in the ordinary course of trade continued to be carried out in the domestic currency. It could not, therefore, be claimed that the external developments during the investigation period had affected the domestic transactions so that they could not be used by the EC to establish normal value. A high rate of domestic inflation need not necessarily imply that producers of cotton yarn were discouraged from selling to the domestic market. Since only the proceeds of exports in United States dollar were affected by the freezing of the exchange rate (when the exporter had to convert the proceeds into cruzados), the freeze could not have affected their incentives to sell in the local market. Therefore, Brazil could not claim that a "particular market situation" existed which made the calculation of normal value on the basis of sales in the domestic market unsuitable.

85. The EC further argued that while a particular market situation mostly resulted from the domestic market itself (e.g. monopoly situation), it could also arise due to an external factor, provided the external factor affected domestic sales and prices, which were the sales and prices considered by Article 2:4. In the present case, at no time did Brazil or the exporters supply evidence that the particular situation of the exchange rate of Cruzados affect the sales and price situation on the domestic market. Therefore, the EC was fully entitled to use either domestic sales or a constructed normal value.

86. The EC argued that Brazil did not appear to argue that the EC had calculated the normal value incorrectly. Rather, Brazil had stated that: "Clearly during the first half of the investigation period, there was a particular market situation prevailing in Brazil within the meaning of Article 2:4 of the Agreement, as a result of which the establishment of normal value on the basis of domestic sales of the like product did not, in the exceptional conditions prevailing, permit a proper comparison with export prices." This showed that Brazil was apparently not distinguishing between the determination of normal value and its comparison with the export price for the purpose of establishing the dumping margin. However, the term "particular market situation" referred to in Article 2:4 could not, and was not intended to, apply to the situation of high inflation and "freezing" of exchange rates for the purpose of establishing the normal value. The issue of comparison was a matter potentially relevant under Article 2:6 of the Agreement, but not in the context of Article 2:4 thereof. It should be concluded, therefore, that the alleged violation of Article 2:4 of the Agreement should be rejected as unfounded.

87. The EC disagreed with Brazil's view that limiting the coverage of the words "particular market situation" in Article 2:4 would require adding "in the domestic market" in that Article. The EC argued that it was not necessary to add "in the domestic market" in the text of Article 2:4 because this resulted from the context of the language "particular market situation" in the sentence (i.e. a few words after a reference to "domestic market") and from the context of this language in Article 2:4. This paragraph dealt with the determination of normal value and nothing else, in contrast to the determination of export price which was expressly addressed in Article 2:5. For the same reasons, the inclusion of the terms "establishment of normal value" was not necessary. The purpose of Article 2:4 was the determination of an appropriate normal value by selecting among three possibilities: sales at arms length on the domestic market (preferred, if at all possible); constructed normal value; or exports to third countries if domestic sales cannot be used. Therefore, Brazil's point regarding the drafting of Article 2:4 was not only redundant but also inappropriate from a systematic interpretation point of view.

88. Regarding consideration of small volume of domestic sales (i.e. 5 per cent) for ascertaining whether domestic sales could be used as normal value, the EC argued that this aspect (i.e. the consideration of the volume of domestic products in comparison to the volume of exports) was not as such a situation outside the domestic market. If the structure of the total sales of a company, i.e. the ratio between its sales on the domestic market and its export sales, was considered in order to ascertain the relative importance for a company's sales on its domestic market, a value of less than 5 per cent for this ratio showed that the quantities sold on the domestic market were so small that they could not be considered to be representative. Furthermore, even if this practice were to be considered as "a situation outside the domestic market", there would still not be any contradiction with the EC's interpretation because this would be a particular market situation that actually affected domestic sales.

89. The EC argued that the text of Article 2:4 established a hierarchy, for the purpose of establishing normal value, between use of sales in the ordinary course of trade in the domestic market and the other two methods mentioned in that provision. It did not, however, establish any hierarchy between the two latter methods (export prices to a third country and cost of production) which could be used to establish normal value when it was not possible or appropriate to rely on domestic sales prices. Thus, there could not be any question of the EC having to justify its choice between these two methods. The EC normally relied on cost of production, and in this case the EC did not use the price of exports to third countries for a number of reasons, including the fact that the prices to such third countries were also likely to be affected by the freezing of the exchange rates. Most, if not all, export sales of cotton yarn from Brazil were made in United States dollars. The prices of those exports were equally affected by the freezing. In such a case, even though the comparison would have been made in the same currency, sales to third countries would most probably not have permitted a proper comparison. These imports to third countries could have been also dumped and would hence not be representative of the normal value of Brazilian cotton yarn. On the contrary, the use of constructed normal value, as applied in this case, appeared to provide a more representative normal value.

90. Brazil argued that the EC compared broadly stable United States dollar export prices with the steadily rising Cruzado denominated normal values, through the medium of the artificially fixed cruzado to United States dollar exchange rates. This was not proper comparison under Article 2:4. Similarly, costs of production were based on adding together the costs of raw materials, manufacturing overheads and selling costs, and general and administrative expenses. Similar to the normal value based on domestic sales, the cost of production was also unreliable in this case because domestic inflation caused the prices of the components of cost to also increase. Therefore, neither normal values based on domestic sales nor normal values based on costs of production permitted a proper comparison with export price. In the absence of any allowance under Article 2:6, Brazil argued that the normal value should have been based on sales of cotton yarn to a third country.

91. Brazil further argued that the EC had itself acknowledged that the reason for which export prices to third countries were not usable in this case was that the prices to such third countries were also likely to be affected by the freezing of exchange rates. Hence, exchange rates were considered by the EC to be relevant when considering whether a normal value based on sales to any third country would permit a proper comparison, but were considered irrelevant when considering whether normal value based on domestic sales or costs of production would permit a proper comparison. Such a position was contradictory. As the EC's had itself acknowledged, exchange rates were a relevant factor in the application of Article 2:4. Brazil argued that far from giving rise to distortions, the use of sales to any third country would actually have eliminated the distortions in the calculation.

92. Brazil disagreed with the EC's argument that Article 2:4 could not apply to situations concerning exchange rates. There was nothing in Article 2:4 which prohibited a consideration of exchange rates in the application of that Article, and it was clearly possible that Article 2:4 included the exchange rate situation within the concept of particular market situation. Moreover, the EC had admitted that the phrase "particular market situation" may cover external factors (i.e. factors outside the domestic market), but that these must "affect domestic sales and prices" before they could be relevant. Brazil considered that exchange rates were clearly capable of affecting domestic sales and prices. For example, exchange rates affect the cost of imported raw materials. Brazil therefore concluded that, on the basis of the EC's own interpretation, exchange rates would be capable of being a relevant factor in the application of Article 2:4. However, the EC had admitted that the investigating authorities proceeded in this case on the basis that the exchange rate situation was irrelevant to determining the basis for establishing normal value, i.e. the EC did not include the exchange rate situation within the factors to be considered under Article 2:4 in the context of determining the correct basis for establishing normal value, but eventually decided that, in all the circumstances, the most appropriate basis was domestic prices (and, in certain cases, costs of production). Therefore, the EC treated the exchange rate situation as completely irrelevant, even though the exchange rate situation was clearly a particular market situation affecting price comparability within the meaning of Article 2:4. Accordingly, the EC had reached its determinations on the basis of an erroneous interpretation of Article 2:4, and its determinations were therefore not in conformity with the Agreement.

93. Regarding the hierarchy of methodologies established in Article 2:4, Brazil argued that Article 2:4 imposed an obligation to effect a proper comparison, and if one method of establishing normal value would lead to a proper comparison, but another would not, the first method was to be preferred. In this case, Brazil considered that, in the absence of any allowance under Article 2:6, sales to any third country would have permitted certainly a more proper comparison than costs of production. However, the EC did not even request the information it would have required in order to base normal value on sales to any third country, proceeding on the basis only of an entirely unsupported assumption that such sales "could have been also dumped". Brazil argued that the steps taken by the EC in this respect could not be considered as sufficiently reasonable to ensure that its determinations were in conformity with the obligations imposed by Article 2:4.

94. Regarding the EC's argument that there was nothing in Article 2:4, or in its drafting history, to suggest that it covered exchange rates, Brazil argued that Article 2:4 clearly stated that it was concerned with ensuring a "proper comparison". Since exchange rates were an essential factor in the comparison between normal value and export price, Brazil considered that these words were a very strong indication that exchange rates were relevant to the application of Article 2:4.

95. Brazil argued that under Article 2:4, if the normal value was based on domestic sales, these sales must be in the ordinary course of trade and permit a proper comparison. Brazil argued that the EC's interpretation of Article 2:4 ignored the words "proper comparison" altogether by effectively reading them out of Article 2:4, or reading them to mean something completely different from their plain and ordinary meaning, i.e. under the EC's interpretation, Article 2:4 just referred to the requirement that normal value be established without any reference to: (i) ensuring a proper comparison; or (ii) ensuring that any measures adopted did not constitute an unjustifiable impediment to international trade; or (iii) ensuring that they were in accordance with Article VI of the GATT. However, the text of Article 2:4 requiring proper comparison could not be overlooked, and thus Brazil did not consider that the EC's interpretation of Article 2:4 was correct. If the interpretation of the EC was correct, then why were the relevant words not repeated in Article 2:5 of the Agreement? Brazil argued that an interpretation of the Agreement which required words to be ignored or deleted should not be accepted.

96. Brazil argued that keeping in mind the object of the obligation to ensure a "proper comparison", the only possible reading of Article 2:4 was by considering what the normal value had to be compared with. The only possibility was that it had to be compared with the export price. Thus, Article 2:4 imposed an obligation on the EC to select a method of establishing normal value, which would not only be in accordance with the other requirements of that Article, but would also permit a proper comparison with the export price. In view of this, Brazil argued that the interpretation put forward by Brazil had to be preferred because it respected the plain and ordinary meaning of the words.

97. Brazil argued that the Anti-Dumping Committee's Decision of 5 May 1980 provided further confirmation that its interpretation of Article 2:4 was correct. That Decision stated that: "Due consideration should be given to all cases where, because special economic conditions affect prices in the home market, these prices do not provide a commercially realistic basis for dumping calculations" (emphasis added by Brazil).

98. Regarding the 5 per cent rule, Brazil argued that if a clear and watertight distinction must be made between Article 2:4 (domestic market) and Article 2:5 (export market), it could not be that the volume of export was relevant to Article 2:4 but the export market was not. The volume of export sales was clearly a characteristic outside the domestic market. According to the EC, since it was the ratio between domestic sales and export sales volume that was analyzed for its 5 per cent rule, somehow this brought the volume of export sales within the domestic market. However, a ratio was a comparison, and since the EC made a comparison between domestic and export volume in its 5 per cent rule, the export volume was relevant to the application of Article 2:4. In this respect, Brazil saw no difference between sales volume and sales value (i.e. export prices), which would have to be compared with normal value through the medium of an exchange rate. In view of this, Brazil argued that the EC's analysis of Article 2:4 was not logical because it always allowed export volume but never exchange rates to be considered in the application of that Article. Brazil argued that the link between exchange rates and price comparability was much stronger and more direct than the link between export volumes and price comparability. Brazil further argued that the 5 per cent rule was not concerned with the effects of export volume on domestic sales, but with the effects of export volume in conjunction with domestic volume on price comparability. This precisely confirmed the interpretation of Article 2:4 set out by Brazil.

99. Brazil argued that the EC took a rigid and isolated interpretation of Articles 2:4, 2:5 and 2:6, according to which Article 2:4 concerned normal value, Article 2:5 export price and Article 2:6 the comparison between the two. Brazil agreed with this general interpretation, but argued that there was no contradiction between, on the one hand, the determination of normal value and export price being distinct stages in the dumping calculation and, on the other hand, regard being given to characteristics of the export market (such as frozen exchange rates) in selecting the method for determining normal value. The exchange rate freeze had a profound effect on price comparability and Article 2:4 required that this factor should at least have been considered. Article 2:4 required that the issue of price comparability also be considered in the context of the selection of the method of establishing normal value. This followed from the plain and ordinary meaning of the words, and no special interpretative technique was required to reach this conclusion. Brazil argued that no other meaningful interpretation of Article 2:4 was possible. Hence, Article 2:4 should be interpreted teleologically, with "one eye" on the question of price comparability. There was no contradiction between the view that Article 2:4 was concerned principally with the establishment of normal value, and the view that Article 2:4 was also concerned with certain aspects of price comparability. The question of exchange rates may easily be distinguished from the question of establishing export price, standing as it did between the two elements of the calculation. Therefore, the interpretation put forward by Brazil in no way undermines the general structure of Articles 2:4, 2:5 and 2:6.

100. Brazil argued that the EC's own interpretation of Article 2:4 in other respects reflected the teleological interpretation, rather than the isolated interpretation. This was true, for example, of the EC's decision to establish normal value on a monthly basis. This decision was made in the light of problems of price comparability which only arise when the comparison takes place. In this respect, the EC purported to select a method for establishing normal value which was intended to ensure a "proper comparison". The same was true of the application by the EC of the 5 per cent criterion. Thus, although Articles 2:4, 2:5 and 2:6 focused on different issues, an interpretation of one must be informed by an interpretation of the others -- as was the established practice of the EC and other contracting parties. Brazil argued that this contextual approach was normal in the interpretation of international agreements, a view that the EC had endorsed in its statement that "it would not be coherent with the general practice of international law that the GATT or the Agreement be totally isolated from the general body of international law ..."

101. Brazil argued that the object and purpose of the Agreement also showed that the EC's method did not permit a proper comparison between normal value and export price. The object and purpose of the Agreement was not to penalize countries which had been obliged to apply temporary exchange control, or to exclude them from their export markets simply because such temporary exchange controls were applied. On the contrary, in light of the provisions concerning developing countries, special allowance should be made in such cases. Brazil noted that the first recital of the Agreement provided that "anti-dumping practices should not constitute an unjustifiable impediment to international trade". Brazil wondered how anti-dumping duties could be justified if they were imposed solely on the basis of margins artificially generated by an exchange rate freeze, especially when the United States dollar price remained broadly stable, and when zero or de minimis margins were calculated in the absence of the exchange rate freeze. Brazil was of the view that ninety days of exchange rate freeze could not justify five years of duties.

102. Brazil further argued that even if the Agreement was silent on a particular point, it did not follow that Contracting Parties enjoyed unlimited discretion in the relevant area. On the contrary, they were bound to ensure that they exercise their discretion in accordance with the other relevant provisions and/or principles of the Agreement, the GATT and relevant international law. For example, Article 5:1 of the Agreement required investigations to be initiated on a "written request ... on behalf of the industry affected". The Agreement did not elaborate explicitly on the meaning of this phrase. It did not follow, however, that the contracting parties had unlimited discretion in the application of this provision. On the contrary, the Panel on Swedish Steel had ruled that:

"5.9 .... The Panel concluded that "a written request ... on behalf of the industry affected" implies that such a request must have the authorization or approval of the industry affected before the initiation of an investigation.

5.10 .... Article 5:1 must be interpreted to require investigating authorities, before opening an investigation, to satisfy themselves that a written request is made on behalf of a domestic industry, defined in accordance with Article 4.

The Panel noted that the parties to the dispute did not disagree on the existence of this requirement. .... Rather, the parties seemed to disagree regarding the nature of the specific procedural steps to be taken by investigating authorities to meet this requirement. The Panel noted that the Agreement did not provide precise guidance in this respect and considered that the question of how this requirement was to be met depended upon the circumstances of each particular case. Rather than attempting to define any general guidelines, the Panel limited itself to examine whether in the case before it the relevant authorities of the United States had taken the such steps as could reasonably be considered sufficient to ensure that the initiation of this investigation was consistent with their obligation to satisfy themselves that the written request for the opening of an investigation had been made on behalf of the relevant domestic industry." (emphasis added by Brazil)

The Panel had then concluded that the relevant steps could not be considered reasonably sufficient, and recommended the withdrawal of the measures and the repayment of the duties.

103. Brazil argued that in precisely the same way, Article 2:4 required the EC to select a method of establishing normal value which would have ensured a "proper comparison". Brazil argued that the EC did not appear to disagree that this obligation existed, and that though the Agreement did not elaborate explicitly on the meaning of the phrase "proper comparison", it did not follow that the EC had unlimited discretion in the application of this provision. Brazil argued that for the reasons mentioned above, the steps taken by the EC could not be considered reasonably sufficient to ensure that Article 2:4 was applied in a manner consistent with this obligation.

104. The EC argued that by disregarding the effect of the exchange rate freeze on the determination of the normal value (both when normal value had been established on the basis of domestic sales and when costs of production had been used instead), the EC did not depart from its obligations under the Agreement. Given that the task of the Panel was to review whether the EC acted in conformity with the Agreement, not whether an alternative method existed and may have been used instead, the EC submitted that Brazil's claims in this respect should be rejected.

105. The EC argued that following the generally accepted principles of treaty interpretation, the text, object purpose and context of Article 2:4 could only mean that for the calculation of normal value, there was no legal obligation upon the parties to take into account factors like freezing of exchange rates. The phrase "because of the particular market situation" could only refer to the situation "in the domestic market of the exporting country". Therefore, if "sales" (in the domestic market) did not permit a proper comparison, then Article 2:4 provided two alternatives for the calculation of normal value: either exports to a third country or on the basis of the cost of production. While the Agreement actually established a hierarchy between sales in the ordinary course of trade (the preferred method to establish normal value) on the one hand and sales to third countries or costs of production, on the other hand, no such hierarchy existed between the latter two options (i.e. sales to third countries and costs of production). The Agreement did not even suggest circumstances where one would be more appropriate than the other or would not be applicable.

106. The EC argued that Article 2:4 was concerned only with the calculation of a normal value which was representative of the situation in the domestic market of the exporting country. This resulted from the wording of Article 2:4 itself; nowhere did that provision refer to the export market. This was also confirmed by the context, which related to the identification of normal value. Moreover, if one considered the mechanism set up by Article 2:4 to select a representative normal value, it was clear that the expression "particular market situation" related only to the domestic market of the exporting country. Article 2:4 provided that sales in the ordinary course of trade on the domestic market should, in principle, be used. However, two reasons would justify not using those sales: either they did not exist, or they would not allow a proper comparison because of the particular market situation. If, as Brazil contended, "particular market situation" also included situations whose effect was felt on export markets, the negotiators would not have included costs of production as an alternative option, because costs of production would not permit a proper comparison for the same reasons for which sales on the domestic market could not be used. Alternatively, they would have treated cost of production in the same way as domestic sales; otherwise, the language of Article 2:4 would be incoherent because it would not prevent a party from using costs of production to establish normal value, even though a "situation" which allegedly made domestic sales an unsuitable basis for comparison had the same effect on cost of production. Thus, the EC considered that Article 2:4 should be interpreted as referring exclusively to a "particular market situation" having an impact on domestic sales in the country of export. However, this did not mean that factors external to the domestic market (e.g. commodity prices on the world market) may not be taken into account. What this interpretation implied was that those external factors should have an impact on prices on the domestic market of the exporting country in order to be taken into account in establishing the appropriate normal value so as to effect a proper comparison. Therefore, neither language nor the context of Article 2:4, nor other principles of treaty interpretation, created an obligation for the EC to take into account factors affecting export markets in selecting the appropriate normal value.

107. The EC argued that Brazil had failed to show or to provide any evidence whatsoever that sales in domestic market for the period of the exchange-rate freezing were not usable. The domestic sales in Brazil were taking place as normal during that period, since the freezing of the exchange rate applied only to the conversion of the proceeds of exports to third countries. Domestic sales and their prices were, therefore, totally unaffected. Thus, Brazil had no right to claim that the EC should have disregarded these domestic sales.

108. The EC argued that Brazil was incorrectly interpreting the EC's argument regarding the effect of exchange rates on use of third country sales as a basis for normal value. Sales to third countries were different from sales in the domestic market, and the EC's argument was that exchange rates might affect the first but clearly not the second type of sales. The EC further argued that following the logic of Brazil's arguments regarding the exchange rate freeze, sales to third countries would have been also unusable as normal value. Therefore, the only alternative left was to base the calculation on cost of production, and that was what the EC did for two companies for the last three months of 1989, when it found domestic sales were made at a loss.

109. The EC argued that Brazil's interpretation of Article 2:4 did not take into account the object, purpose and, especially, the drafting history of that provision. In equating the freezing of exchange rates with the term "because of a particular market situation", Brazil was interpreting the term "proper comparison" by unduly stretching its meaning to "perfect" comparison. Brazil was then proceeding backwards by saying that since the EC did not take into account the freezing of the exchange rates it did not make a proper comparison and thus violated Article 2:4. Hence, Brazil was going as far as to suggest that the EC was prohibited from using domestic sales in this case and the only option available to it was to use sales to third countries. Such an interpretation by Brazil was incorrect because the phrase "particular market situation" could never cover freezing of exchange rates, the obligation to make a "proper" comparison (not a "perfect" one) may be fulfilled by comparing either sales to third countries or domestic cost of production, and Article 2:4 did not oblige the parties to employ one or the other of the above two alternative methods of calculation for the purpose of comparison. The EC argued that because Brazil had failed to show why sales to third countries should be preferred to domestic sales or domestic cost of production, the whole of its allegation should be rejected.

(b) Violation of Article 2:4: incorrect determination that sales were not made in the ordinary course of trade

Introduction

110. Brazil argued that the EC had violated Article 2:4 by calculating the normal values in October to December 1989 for two Brazilian companies on the basis of cost of production, because the EC had incorrectly determined that the domestic sales by these companies were made at a loss.

111. The EC argued that this claim was a new claim and therefore should not be admitted by the Panel. The EC further argued that in any case, its determination on the point raised by Brazil in this context was in conformity with Article 2:4 and that the Brazilian claim was based on an incorrect argument regarding the calculation of the cost of production.

Arguments by the parties

112. Brazil recalled that the EC had determined normal value for Nisshinbo and Kanebo in October, November and December 1989 on the basis of costs of production. This was done on the basis that sales in the domestic market by those companies in those months were not made in the ordinary course of trade, in so far as they were made at a loss. Brazil argued that the EC had incorrectly reached its conclusion in this regard and thus its findings on this point were in violation of Article 2:4 of the Agreement.

113. Brazil argued that the EC had reached this determination after comparing total costs of production per unit, including production finance costs, with domestic sales at ex-works level which did not include costs of credit incurred by the companies as a result of the payment terms granted in the domestic market. Thus, the EC had included in the total costs of production finance costs incurred partly as a result of the domestic payment terms granted, but had simultaneously discounted these costs from the domestic sales prices.

114. Brazil argued that working capital finance costs may have been incurred by a company not only because capital was tied-up in a company during the production process (including the storage of raw materials and finished products) but also because a company granted credit terms to its domestic customers. For both of these reasons a company may need to borrow money, thereby incurring finance costs. From the wording of the Definitive Regulation, it was Brazil's understanding that the costs of production used by the EC in determining whether or not domestic sales were profitable included both these finance cost elements. The method used by the EC was incorrect to the extent that the EC apparently included all finance costs in the costs of production figure, but excluded credit costs from domestic sales figures, in comparing the two to reach a determination on profitability. Thus, the EC was not comparing like with like. Since the rate of interest charged by a company to its customers generally reflected the rate of interest incurred by the company in borrowing capital until it received payment from its domestic customers, such a methodology would inevitably lead to an (incorrect) finding that sales were not profitable. Such a methodology produced a distortion in the calculation, particularly in an high inflationary environment. Brazil argued that had the EC taken into account the domestic payment terms, it would have found that the sales in question were in fact profitable and were made in the ordinary course of trade. Hence, those sales, and not cost of production, should have been used as the basis for normal value. Brazil argued that if those sales had been used, de minimis or no dumping margins would have been found for Nisshinbo and Kanebo in the final quarter of 1989. Therefore, the EC's findings on this point violated Article 2:4 of the Agreement.

115. The EC argued that Brazil's claim was founded on a wrong interpretation of the meaning of the term "production financing costs" as it appeared in Recital 14 of Regulation (EEC) No. 738/92 (i.e. cost of financing the productive investment plus any cost incurred to finance the actual production and/or maintain the stocks). The text of that recital showed that those costs, taken in relation to a constructed normal value at the ex-factory price, could not include costs related to credit to the customers. Consequently, Brazil's claim was based on an over-inclusive definition of "production financing costs" which was not supported by the usual meaning of these terms and found no justification in Recital 14 of the Definitive Determination. Without prejudice to the fact that this claim is not properly before the Panel, this showed that the claim was also unfounded on the merits and should in any case be rejected by the Panel.

116. Brazil disagreed with the explanation offered by the EC regarding this point. Brazil argued that in recital 14 of the Definitive Determination, the EC stated that "... the cost of production was calculated by taking into account all cost elements, including production financing costs." From "all cost elements" Brazil's understanding was that the EC included all financing costs in its calculation, including those incurred because credit terms are granted in the domestic market. Brazil therefore considered that the arguments put forward by the EC on this point should be rejected.

V. 2. Violation of Article 2:4 and 2:6: Failure to make adjustments to ensure a fair comparison between normal value and export price

Introduction

117. Brazil argued that the EC had reached its determination on the basis of a legal rule which could be stated as follows: in a dumping calculation the EC investigating authorities must apply the "official" exchange rate. This violated the EC's obligations under the Agreement, in particular the obligation to effect a proper and fair comparison between normal value and export price, as required by Articles 2:4 and 2:6 of the Agreement.

118. Brazil argued that the EC's refusal in the present case to adjust the exchange rates violated a fundamental principle of the Agreement, overlooked the well-known phenomenon of "exchange dumping", was not in conformity with the practice in major anti-dumping jurisdictions, had relied on certain principles such as "monetary neutrality" that were not valid in the context of anti-dumping proceedings, and the discretion available to the EC under the Agreement had not been exercised. Based on the data provided by the EC in its disclosure letter, Brazil presented its own calculations of adjusted dumping margins in support of its argument that if the EC had made appropriate adjustments, then the dumping margin would have been found to be de minimis.

119. The EC argued that it had met all the requirements of the Agreement, including those specified in Articles 2:4 and 2:6. The EC argued that the calculation of dumping margins had to be made on the basis of objective and verifiable information, and not on the basis of arbitrary and subjective aspects. Accepting Brazil's arguments in this regard would amount to introducing considerable amount of subjectivity and uncertainty into the system. It would go far beyond the scope of the Agreement, the possibilities and the competence of the investigating authorities, and the interests of the signatories to have security and predictability in international trade.

120. The parties presented arguments relating to: the fundamental principle of the Agreement; the factual aspects of this case; lack of conformity with practice in major anti-dumping jurisdictions; the alternative estimates presented by Brazil; Brazil's allegation that use of average monthly exchange rates was not sufficient; the inability to consider method of establishment of official Brazilian exchange rate and to the principle of "monetary neutrality"; the EC's contention that the adjustments requested were beyond the scope of the Agreement; the likely effect of the methodology on future proceedings, i.e. the "floodgates" argument; treatment of Brazil being contrary to the EC's treatment of other countries in the case; the importance of the economic situation and a change in the criteria by the EC i.e. "moving the goalpost"; and, Brazil's allegation that the EC failed to exercise its discretion.

(a) Arguments relating to the fundamental principle of the Agreement

Introduction

121. Brazil argued that the EC had failed to meet the fundamental principle of the Agreement, namely the obligations to conduct a fair comparison between normal value and export price. Given the special situation arising in Brazil due to the exchange rate freeze, the EC should have made appropriate adjustments to the exchange rates in order to make a fair comparison. Instead, the EC based its comparison on a rule that it should use only "official" exchange rates, and this rule prevented the EC from exercising the required flexibility to adjust the exchange rates in order to take account of the distortion created by the exchange rate freeze.

122. The EC argued that it met all the requirements of the Agreement, including those for conducting a fair comparison. Fair comparison required only that the comparison be made at the same level of trade and in respect of sales made at as nearly as possible the same time. In this case, the EC had used monthly exchange rates for comparing normal values with export price and had thus fulfilled the requirement of fair comparison. Also, the requirement in the Agreement was to conduct a fair comparison and not necessarily to use the "best" possible method of comparison. The EC had met the requirement that was specified in the Agreement.

Arguments by the parties

123. Brazil argued that even if it could be accepted that normal value could have been calculated on the basis of domestic sales or costs of production, the EC should have made adjustments in order to take account of temporary distortions and to ensure a fair comparison between normal value and export prices. Such adjustments could have been made either to the normal value or by the use of an alternative exchange rate. Since such adjustments were not made by the EC, Brazil considered that the EC had infringed Article 2:6 of the Agreement.

124. Brazil recalled that Article 2:6 provided that "[d]ue allowance shall be made in each case ... for ... other differences affecting price comparability." Brazil argued that in this case, domestic prices were steadily increasing in line with inflation while the export prices in United States dollars remained basically stable. The use of a frozen exchange rate in making the price comparison introduced a gross distortion into the comparison of the relevant data, for which the EC should have made due allowance. However, the EC investigating authorities had operated in this case on the basis of a legal rule that they must apply the "official" exchange rate in a dumping calculation. This rule was not to be found in the text of the Agreement, found no support in the object or the drafting history of the Agreement, in past GATT Panel decisions, in the Agreement on Implementation of Article VI of the General Agreement on Tariffs and Trade 1994 (hereinafter "WTO Anti-Dumping Agreement"), or in the practice of the contracting parties. This rule unduly fettered the investigating authorities' ability to make a determination in accordance with the provisions of the Agreement, even though the Agreement did not place any such obligation on the contracting party. This rule, which was incorrectly deemed to be mandatory, meant that no consideration could be given to the question of whether or not the frozen exchange rate should have been used. Thus, the EC had violated Articles 2:4 and 2:6 of the Agreement, in particular the obligation to effect a proper and fair comparison between normal value and export price.

125. Brazil argued that the contracting parties had consistently recognized the overriding principles that: (i) a proper and fair comparison had to be made between normal value and export price in the context of an anti-dumping proceeding; and, (ii) allowance had to be made for temporary exchange rate distortions where they otherwise led to an unfair comparison between normal value and export price. The principles of the Agreement established that allowance should be made for differences affecting price comparability, including temporary fluctuations that would otherwise produce artificially inflated dumping margins. Where appropriate, exchange rates applicable in a past or future period (i.e. lagged exchange rates) may be applied or inflation indexed or lagged official exchange rates may be used. Moreover, negative dumping should be taken into account, particularly where fluctuations arise as a result of an high inflation environment. In this case, the temporary nature of the distortion in question (i.e. the fixed exchange rates) represented a radical and unsustainable departure from the underlying economic environment, which was different from a fluctuation occurring in the context of a freely floating currency.

126. To further support its argument, Brazil referred to the provisions of Article 2.4.1 of the WTO Anti-Dumping Agreement. Brazil argued that though the WTO Anti-Dumping Agreement was not yet in force and its provisions therefore were not binding, it was acceptable to have regard to that Agreement to the extent that it revealed the intention of the contracting parties regarding the way in which anti-dumping investigations should be conducted. Brazil considered that the provisions of Article 2.4 of the WTO Anti-Dumping Agreement (the successor to the existing Article 2:6) did not substantively amend the Agreement but merely clarified the nature of the obligation under the existing Agreement to effect a fair comparison. Brazil argued that Article 2.4.1 of the WTO Anti-Dumping Agreement supported its arguments since it made clear that fluctuations in exchange rates shall be ignored, i.e. shall not be used as the basis for a dumping determination. Furthermore, the new Agreement made it clear that, faced with an overvalued domestic currency (as was the case in the proceeding under review), exporters shall be allowed at least sixty days to adjust their export prices. In this case the exchange rate was frozen for a longer period (i.e., 90 days), and was due to measures taken by the Brazilian government, publicly expressed to be temporary. The exporters could be absolutely certain that, within a short period of time, the Cruzado would once again begin to depreciate, i.e. they could be sure of the temporary nature of the distortion. However, no allowance was even considered by the EC in the present case.

127. Brazil argued that the normal level of the exchange rate was not prevailing during the period of the exchange rate freeze, and the exchange rate did not begin to stabilize until the second half of the year. It was only in August 1989 that the rate of depreciation of the cruzado (29.36 per cent) aligned with the domestic rate of inflation (29.34 per cent), an alignment that continued for the remainder of the year. Though the Brazilian exporters may not have been able to say at any given moment precisely what the normal level of the exchange rate would be, nor the precise level to which the exchange rate would move once the effects of the freeze were passed, they would, however, have been well aware that once the freeze was lifted, the exchange rate would again move in line with domestic inflation. It was common knowledge that the exchange rate freeze would not last because similar attempts by the Brazilian authorities to deal with financial instability in Brazil in the preceding years had also not been successful. Thus, "exchange dumping" occurred in the first part of the year. Brazil argued that there was a clear and very particular distortion arising in this case, for which the EC should have made allowance. How such allowance could actually have been made was a matter for the investigating authorities, acting within the limits of their discretion and taking into account the provisions of the Agreement. Brazil considered that use of indexed exchange rate would have been one reasonable method.

128. The EC argued that though Brazil was claiming that the contracting parties had "considered or recognized" the issue that the use of so-called "temporary" exchange rates should be avoided and that "lagged" exchange rates should be used, Brazil had cited no reference to any GATT documents to support its views.

129. The EC noted that Article 2 of the Agreement provided the rules for the determination of the dumping margin. However, neither paragraph 4 nor paragraph 6 of Article 2 explicitly mentioned the need to take into account inflation and currency exchange rates in the calculation of dumping. Therefore, it was necessary to interpret the provisions alleged by Brazil to have been violated in accordance with the general rules of treaty interpretation, i.e. in accordance with their ordinary meaning in the context of the Agreement, and in light of their object and purpose. The drafting history could also be taken into account as a supplementary means of interpretation.

130. The EC argued that there was nothing in the text of Article 2:6 to suggest that high inflation and "freezing" of exchanges rates should be taken into account when making the comparison. The allegation of Brazil unduly stretched the language of Article 2:6. The phrase "due allowance shall be made .." clearly referred to objective differences affecting price comparability due to differences in conditions and terms of sale, in taxation, physical characteristics, etc, of the products in question. That phrase did not, and could not be taken to, include subjective and highly volatile and unpredictable differences, such as those resulting from inflationary environments, exchange rate fluctuations and currency restrictions resulting from monetary policies of the Parties to the Agreement.

131. The EC noted that the first sentence of Article 2:6 required that a fair comparison be made, and the definition of the concept of fair comparison was not left to the discretion of the parties. Article 2:6, first sentence, clearly set out the conditions to be satisfied in order to achieve a fair comparison in conformity with the Agreement, i.e. fair comparison of the two prices (normal value and export price) required comparison at the same level of trade and in respect of sales made at as nearly as possible the same time. The Agreement required no more than that. This was confirmed by the introductory language of Article 2:6, first sentence: "in order to effect a fair comparison" (emphasis added by the EC). The language "to effect a fair comparison" could not be taken in isolation from the rest of the sentence. The EC therefore concluded that it had satisfied its obligation to make a fair comparison under the Agreement by making a comparison at the same level of trade and in respect of sales made at as nearly as possible the same time.

132. The EC noted that the second sentence of Article 2:6 required that "due allowance shall be made in each case, on its merits, for the differences in conditions and terms of sale, for the difference in taxation and for the other differences affecting price comparability". Consequently, the situations foreseen in that sentence clearly indicated that allowances were limited to objective differences and did not include subjective factors such as exchange rates fluctuations. As a result, Article 2:6, second sentence, did not require that the exchange rate be taken into account among the "other differences affecting price comparability". While the EC considered that the Agreement did not require the investigating authorities to take into account exchange rate fluctuations or freezing as such, it would be appropriate to take into account certain effects of exchange rates when fulfilling obligations under the first sentence of Article 2:6. This was more particularly the case with respect to the obligation to make a price comparison in respect of sales made at as nearly as possible the same time.

133. The EC argued that the choice of exchange rates was relevant to the fulfilment of the obligation laid down in the preceding sentence of Article 2:6, i.e., that "prices shall be compared... in respect of sales made at as nearly as possible the same time." In order to fulfil that obligation, the first and best choice was clearly the rate of exchange actually obtained by the exporter for the sale concerned. In this case, the EC established monthly average normal values in order to take into account the particular market situation existing in Brazil (high inflation) and thus permit a proper comparison between normal value and export prices (This was, however, strictly a matter relating to Article 2:4 only). The EC then compared such normal values with export prices in respect of sales made at as nearly as possible the same time in order to effect a fair comparison under Article 2:6. To convert the proceeds of such export sales from United States dollar to cruzado when making this comparison, the EC used actual returns in cruzado when possible (in this case, for Kanebo); when actual returns were not available, it used end-of-month official exchange rates.

134. The EC further argued that the drafting history of Article VI of the General Agreement appeared to confirm the view that the intention of the drafters was not to include so-called "exchange dumping". Where the drafters of the General Agreement wished to deal explicitly with the effects of currency devaluations and exchange rate fluctuations and their effects on the rights and obligations of the contracting parties, they had done so by inserting such provisions in the text of the GATT (for example, Articles II:6(a) and (b), Article VII:4(a) to (d), Article VIII:4(d) and the note ad Article VIII of GATT). The note ad Article VI on multiple currency practices (which referred to Article VI:2) was another example of a provision that dealt explicitly with the issue but did not cover the problem addressed by Brazil in the present case. Moreover, the text of Article VI:1, last sub-paragraph, of the General Agreement had not changed substantially in the 1962 and 1979 Anti-Dumping Codes. From the text and drafting history of the Codes commentators had concluded that: "The antidumping Code is silent as to how an appropriate rate is to be selected, even though it can have enormous impact on the dumping calculation. The Code imposes no requirements on signatories and offers them no guidance. Consequently, the law relating to exchange rates in antidumping investigations is entirely a matter for the municipal law of the Code signatories" (emphasis added by the EC). The EC argued that the commentator quoted here had also remarked that in the real world what he called a "proper exchange rate" could rarely be chosen.

135. The EC argued that if the language and drafting history did not support a particular interpretation of Article 2:6 of the Agreement, the examination of the object and purpose of this provision would be useful in the interpretation of the treaty. The EC noted that Article 2:6 of the Agreement imposed an obligation upon the signatories to make a fair comparison between the export price and the normal value. According to the terms of Article 2:6, such fairness required that the comparison be made "at the same level of trade" and in respect of sales made "at as nearly as possible the same time". It seemed, therefore, that the object and purpose of Article 2:6 to achieve a fair comparison would be fulfilled if the rate of exchange actually obtained by the exporter when converting the proceeds of a given export transaction into his national currency was used in the comparison. The rate of exchange actually obtained, however, should be a rate obtained for sales made as nearly as possible at the same time with the sales in the domestic market. Article 2:6 of the Agreement, on the contrary, did not require the use of deferred or lagged exchange rates. Dumping should be based on actual information established on the basis of objective and verifiable parameters over a representative past period, and thus account could not be taken of speculation on future rates or on the changes resulting from sovereign economic and monetary decisions of the authorities in the country of export. Accepting Brazil's arguments in this regard would amount to introducing a considerable amount of subjectivity and uncertainty into the system, going far beyond the scope of the Agreement, the possibilities and the competence of the investigating authorities, and the interests of the signatories to have security and predictability in international trade.

136. Therefore, the EC argued that the text, object, purpose and drafting history of Article 2:6 of the Agreement showed that there was no obligation directing or guiding the signatories as to how an appropriate exchange rate was to be selected for the purpose of making a fair comparison between the export price and the domestic price. Hence, the Agreement left this issue entirely to the domestic law of the contracting parties. This implied that whatever criteria was used under the domestic law of the importing country for the purpose of effecting the comparison, the only obligation that was imposed upon that country by the GATT was to use fair criteria in a transparent and non-discriminatory manner to all other signatories in comparable situations and not to commit manifest errors in establishing and appreciating the facts in the process.

137. The EC argued that in the present case, as in almost all cases, it had used the official exchange rate applicable at the time of sale on a transaction by transaction basis. For a normal value calculated on a yearly basis, normally the exchange rate applied to the export price was based on a yearly average. In this case, "in order to permit a proper comparison for export price" the EC had on its own initiative used an exchange rate based on a monthly basis (instead of yearly basis). Conscious of its obligation to effect a "fair" comparison, the EC had moved from an yearly to monthly basis, and had applied it in a transparent and non-discriminatory manner to the two countries (Brazil and Turkey) found to be in a comparable situation, i.e. both had high inflation in their domestic market.

138. The EC argued that Brazil was not correct in stating that the EC had to apply official exchange rates in its dumping determinations. The EC had to carry out its investigation on the basis of objective and verifiable facts. If reliable data could support the fact that a particular exchange rate was actually used (as was the case with Kanebo), the EC would use such exchange rate in priority to a spot rate. This was not contradictory with the opinion expressed by the EC that using other than official exchange rates would amount to an interference in the monetary policy of countries under investigation (see section (f) below for the arguments of the parties relating to "monetary neutrality"). What the EC used for Kanebo was the actual return, verifiable in the books of the company. This had nothing to do with the suggestions made by Brazil, as appraisals based on "lagging" or "indexation" could be highly subjective and difficult to verify.

139. The EC argued that Brazil was contending that the EC should have used not the official exchange rate, but a rate "lagged" on the basis of a subjective reaction (the "anticipation" by the exporters) to a "foreseeable fact" (the "unfreezing" of the exchange rate later on in the year). However, the EC could not know and verify the existence and the extent of the "anticipation" effect. All it could know and verify were actual export prices in United States dollars, actual official exchange rates, and sometimes actual returns in cruzado (where available). Moreover, it was misleading to affirm or to imply that the development of dumping margins throughout the investigation period reflected the relationship between domestic inflation and "frozen/unfrozen" exchange rates, because dumping margins found were much higher than those which one could have expected had the United States dollar/cruzado exchange rate not been "frozen" for three months, but were instead left to fluctuate and change in parallel to domestic inflation. In addition, there were other decisions of the Brazilian monetary authorities, taken during the investigation period, which could affect Brazilian exports just as much as the "freezing" of exchange rate: the devaluations of the external value cruzado in respect of the United States dollar in January and in June 1989. In terms of impact on the dumping margins, these decisions had a favourable effect for the exporters but the EC did not take them into account because it could not do so on the basis of objective and verifiable facts. Indeed, the effect of these devaluations was also (like that of the "freezing" of the exchange rate) dependent on the subjective reactions of individual exporters, and this reaction could only be taken into account when finally reflected in changes in actual export prices and actual returns in domestic currency for those export sales, i.e. once it was translated into verifiable facts.

140. The EC disagreed with Brazil's argument that Article 2.4.1 of the WTO Anti-Dumping Agreement merely clarified the nature of the obligation under the existing agreement to effect a fair comparison. The EC argued that the new Agreement did not clarify but put into the text of Article 2.4.1 a completely new provision which, in any case, concerned fluctuations in exchange rates, not freezing of exchange rates.

141. The EC argued that the reference by Brazil to Article 2.4.1 of the WTO Agreement on Anti-Dumping was of no relevance to the present case. That provision was not yet in force nor was it applicable to the facts of the present case. Also, it did not support the theory Brazil was advancing, since it stated that "fluctuations in exchange rates shall be ignored". The "adjustment" required by Article 2.4.1 of the WTO Anti-Dumping Agreement was applicable only to "sustained movements" during the investigation period, not to "temporary exchange rate distortions" such as those claimed by Brazil in this case. Furthermore, the reference by Brazil to Article 2.4.1 of the WTO Anti-Dumping Agreement was irrelevant also because that provision allowed for a period of 60 days for exporters to adjust their prices, not the 180 days that were involved in this case (see section (b) below for more details on this point).

142. In response to a question by the Panel, the EC argued that it would not be coherent with the general practice of international law that the GATT or the Agreement be totally isolated from the general body of international law and principles and the parties left to absolute discretion. In the present case, the applicable principle of public international law was that of non-interference in the sovereign authority of States to manage their internal affairs in all areas of their competence, including finance and monetary issues. The EC further argued that the interpretation of GATT in the light of generally recognized principles of international law was confirmed in the WTO Agreements, for example, Article 3:2 of the WTO Understanding on Rules and Procedures Governing the Settlement of Disputes. The Agreement was an international agreement like any other and, accordingly, should be subject to the same principles, i.e. the interpretation of the Agreement should be based on the internationally recognized principles contained in Article 31 of the Vienna Convention on the Law of the Treaties (1969). Moreover, when the Agreement was absolutely silent (as it was in relation to the use of exchange rates), and in the absence of other principles of international law applicable, signatories should be subject to the control of manifest error of facts or of interpretation of the facts or of arbitrariness.

143. The EC further argued that the drafting history confirmed the conclusion that the Agreement was silent on the issue of exchange rates. Therefore, the parties were free to apply the system they wished, as long as their practice did not infringe other provisions of the Agreement. This discretion left to the contracting parties was not unknown or unusual in the system of the GATT or the Agreement or in international trade law. The EC said that while its normal practice was to use annual average exchange rates, in this case, it had determined the dumping margin on the basis of monthly exchange rates in order to address the existence of high inflation. Though the Brazilian economy suffered from high inflation, this situation was well known and was not confined to the investigation period. Hence, as far as possible, the EC took account of it by making comparisons of sales made at as nearly as possible the same time by establishing normal value on a monthly basis, not on the average of the whole investigation period, as was normally its practice.

144. The EC argued that on the basis of its arguments presented above, and because in this case Brazil had the burden of proof which it did not fulfil, it would be concluded that the EC had respected the single rule that may be relevant - though to a limited extent - for the selection of the exchange rate, namely the rules that the comparison should take place at as nearly as possibly the same time. The EC had made a fair comparison between the export price and normal value by using monthly official exchange rates. Thus the Panel should reject as unfounded in GATT law and practice Brazil's claims regarding the calculation of the dumping margin and the alleged violations of Articles 2:4 and 2:6 of the Agreement.

145. Brazil argued that there were three clear obligations under Article 2:6: due allowance must be considered; where necessary, due allowance must be made; and that allowance must result in a fair comparison. Brazil argued that the rule applied by the EC investigating authorities in this case, i.e. that the official exchange rate must always be used, was capable of conflicting with these three obligations. While such a rule may well provide a useful starting point in the great majority of cases and may be administratively convenient, it could not release the investigating authorities from the three obligations.

146. Brazil argued that the EC had proceeded on the basis that there was no option but to use the official exchange rate, and had sought to draw on an alleged principle of international law, of doubtful relevance based on unspecified legal grounds but certainly not expressed in the Agreement. In so doing, the EC had set aside the Agreement's most fundamental and express requirement, that of a fair comparison. This in itself was a violation of Article 2:6. Furthermore, the EC violated Article 2:6 because, in comparing normal value and export price, it made no allowance to take into consideration the exchange rate situation, even though such an allowance was expressly required by that Article. Moreover, the EC violated Article 2:6 because its failure to make due allowance for the exchange rate situation led to an unfair comparison between normal value and export price.

147. Brazil argued that at no time did the EC enter into any discussion of the precise mechanism for making allowance to take into consideration the exchange rate freeze. Brazil said that there were a number of options for indexation that the EC could have used. For example, indexation could have been done only for the period of the exchange rate freeze, or for the period during which the rate of inflation and the rate of depreciation remained clearly in disequilibrium as a result of the freeze, or for the year as a whole; Brazil noted that inflation and depreciation moved in parallel from August 1989 onwards. Brazil emphasized that its argument was not that the EC should have used a particular indexed exchange rate, or any particular solution. The Agreement imposed certain very clear legal obligations, particularly the obligations that there be a fair and proper comparison between normal value and export price, with special regard to the special situation of developing countries. These were clear and objective obligations arising under the Agreement, which had been violated by the EC in this case. The EC should have taken steps to ensure a fair and proper comparison and should have had special regard to the situation of Brazil as a developing country. The use of monthly average exchange rates and monthly average normal values did not address in any way the special situation that arose. The problem was not high inflation alone, but high inflation in conjunction with the frozen exchange rate. The use of monthly data did not address this issue. Furthermore, the use of monthly data could produce distortions where, for any reason, the data (especially domestic prices, costs of production and exchange rate) was out of step. In a high inflation environment, the EC's practice of treating a negative dumping margin as zero tended to artificially inflate the final dumping margin calculated. This could lead to a very wide range of results in the dumping margin, both negative and positive, and given the high inflation environment in this case, the EC's failure to take into account negative dumping inevitably led to a distorted and artificially high dumping margin which may be prejudicial for exporters. Thus, far from removing the key distortion from the calculation, the use of monthly data was actually capable of introducing additional distortions.

148. Brazil argued that the purpose of Article 2:6 was stated clearly in the opening sentence, i.e. it was to effect a fair comparison, and the EC had failed to fulfil this object and purpose in this case. Brazil argued that it did not follow from the alleged silence of the Agreement on a particular point that contracting parties enjoyed unlimited freedom in the relevant area. Hence, though the Agreement was silent on which exchange rate should be used, this did not mean that contracting parties were free to select any exchange rate they wished. Rather, they were bound to ensure that their selection of the exchange rate was in conformity with the Agreement. For example, they were bound to select exchange rates in a non-discriminatory manner and to have regard to the requirement that a proper and fair comparison be made. Since the rule applied by the EC made no reference to any of these obligations, it was an incomplete statement of the EC's obligations under the Agreement.

149. Brazil further argued that where the Agreement was silent on any particular point, the contracting parties were still bound to ensure that their practice did not conflict with other provisions of the Agreement, and particularly with the Agreement's fundamental principles. In this respect, Brazil referred to the comments of the panel in Swedish Steel, in particular to the discussion relating to the conclusions of that Panel. Brazil argued that the EC itself had recognized this point when it had admitted that, in selecting an appropriate exchange rate, the GATT imposed an obligation on the contracting parties to observe the principle of non-discrimination; similarly, the EC had acknowledged that the GATT imposed an obligation on the investigating authorities to use "fair criteria" in effecting the comparison, i.e. to ensure that a fair comparison is made between normal value and export price. In precisely the same way, Brazil considered that, in selecting an appropriate exchange rate, the Agreement imposed an obligation on the contracting parties to observe the principle that a proper and fair comparison must be made between normal value and export price.

150. Brazil argued that the EC was artificially restricting the application of the words "fair comparison", so that they applied only to the first sentence of Article 2:6. In Brazil's view, this could not be the correct interpretation of Article 2:6. Allowances granted under the second sentence of Article 2:6 must also ensure that a fair comparison was effected. If they did not, the relevant determination would not be made in accordance with that Article.

151. Brazil disagreed with the EC's contention that there was nothing in the text of Article 2:6 to suggest that it would cover the situation that arose in this case, and that Article 2:6 was restricted to "objective" differences affecting price comparability. The qualifying word "objective" did not appear in that Article. Further, Brazil recalled the EC's view that certain items (such as taxation) were "objective". Brazil argued that the EC had not offered any analysis that might distinguish between what it called objective and subjective differences, but had simply introduced a distinction, not apparent from Article 2:6 itself, which it applied to one type of difference but not to another.

152. Brazil argued that the obligation to make a comparison between normal value and export price at as nearly as possible at the same time was simply one particular aspect of the obligation to effect fair comparison. Thus, even if it could be argued that comparison was made at as nearly as possible at the same time, it did not follow that a fair comparison had been made. Brazil argued that the second sentence of Article 2:6 required allowance to be made, for example, for differences in taxation, and failure to make such an allowance would clearly infringe Article 2:6 because the resulting comparison between normal value and export price would not be fair. Yet taxation was not a question of level of trade or of the timing of the comparison. Also, an ex-factory comparison between different products, in respect of sales made at as nearly as possible the same time, would clearly not be a fair comparison within the meaning of Article 2:6. Similarly, an ex-factory comparison between a normal value based on domestic prices including a sales tax and export prices not including a sales tax, in respect of sales made at as nearly as possible the same time, would not be a fair comparison within the meaning of Article 2:6. This was why Article 2:6 expressly provided that: "Due allowance shall be made in each case, on its merits, for the differences in conditions and terms of sale, for the differences in taxation, and for the other differences affecting price comparability" (emphasis added by Brazil). Thus, it followed that the obligation to effect a fair comparison was not exhausted by the obligations to effect the comparison at the same level of trade and at as nearly as possible the same time.

153. Brazil argued that normal value expressed in domestic currency was not directly comparable with an export price expressed in an export currency because the two values were expressed in different currencies. This was a difference which clearly affected price comparability, and was "objective" similar to the difference due to taxation.

154. Brazil argued that the methodology which the EC claimed it adopted to deal with the fact of the exchange rate freeze (i.e. the use of monthly exchange rates) was in fact adopted for entirely different reasons. It seemed to be intended to deal with the distortions arising in an inflationary environment where the domestic currency was also depreciating against the export currency. This was not the situation in this case where the exchange rate was actually frozen. The EC's use of a monthly reference period did not therefore in any way address the special situation in which Brazil found itself. The EC had no regard to the fact of the exchange rate freeze and took no steps to deal with the distortions which resulted from the freeze.

155. Brazil argued that it was not contending that whenever domestic inflation was not fully reflected in the depreciation of the domestic currency against the export currency, an allowance should be made. Neither did it purport to establish in the abstract the circumstances in which such an allowance may be relevant. Rather Brazil's argument was, in the first place, that the EC should at least have considered such an allowance and, in the second place, that this specific case involved such clear and large distortions that some sort of allowance should have been made. The question of how the allowance might have been calculated was a matter for the EC investigating authorities, provided that it resulted in a fair comparison. This was clearly a broad term giving rise to a degree of discretion on the part of the EC. There may well have been several ways in which an allowance could have been made in order to take into account the distortions which arose. Brazil had offered two methods which it considered would have been reasonable. The point was, however, that the EC made no allowance whatsoever, indeed did not even consider the possibility of doing so. That was a violation of Article 2:6.

156. Brazil did not agree that if an investigating authority were to follow its suggestions, it would be required to guess future exchange rate movements (i.e. movements not known at the time of the investigation). Brazil argued that the investigation period evidently pre-dated the time at which the investigation was actually carried out, so only past exchange rate movements (which were known with certainty) were relevant. Furthermore, with regard to the action or adjustment that should have been made in this case, Brazil did not believe that it would have been necessary to "second guess" past exchange rate movements. Brazil considered that the very severe nature of the distortion that arose, and the catastrophic effect that it had on price comparability, gave rise to an obligation on the part of the EC to consider the various possible solutions and to use a methodology that would have taken the distortion into account. There were a number of different options available. Brazil did not consider that these would have involved any element of "second guessing". On the contrary, they would have involved the use of alternative data that was either known or could be known precisely, or which represented, in Brazil's view, a very fair approximation of the distortion arising. These options included, for example: using export prices to third countries as the basis for normal value; changing the investigation period so that it did not cover the period of the exchange rate freeze; using an indexed exchange rate; using a lagged exchange rate; or using figures deflated to the beginning of the investigation period.

157. Brazil considered that the EC's position was not supported by the authority which the EC had cited regarding the choice of an exchange in dumping calculations. The author cited by the EC had actually concluded that a system which did not have regard to variations of the order of 5 per cent and which did not have regard to the practical problems faced by exporters was defective. He had stated that: "A currency conversion system that ignores exchange rate movements of less than five per cent is simply not precise enough for a dumping regime that bases affirmative determinations on price differences of one-tenth that amount" (ibid., page 77), and that "exchange rate movements always cause uncertainty, and it is important that any regulations adopted provide room to accommodate the practical problems exporters face. A regulation should avoid penalizing mere "exchange rate dumping". The decline of export prices in terms of home market currency, or the rise of home market prices in terms of export market currency, is usually incidental to the pricing policies of exporters. These phenomena arise from the unintended failure of exporters to respond quickly to complex market forces beyond their control." Brazil argued that this case involved exchange rate distortions of the order of 90 per cent, i.e. greatly in excess of 5 per cent, which were ignored by the EC. It was also apparent that the EC made absolutely no attempt to accommodate the practical problems faced by the exporters.

158. Brazil argued that the drafting history of Article 2.4.1 of the WTO Anti-Dumping Agreement made clear that its interpretation of Article 2.4.1, as opposed to the interpretation put forward by the EC, was correct. An earlier version of Article 2.4.1 read: "... the authorities should ensure that margins of dumping do not result from movements of exchange rates during the investigation period to which the exporter or producer could not reasonably be expected to adjust the price of the exported product. The authorities should ignore the effects on the margins of dumping of temporary exchange rate fluctuations."

159. Regarding burden of proof , Brazil said that it had cited specific obligations under the Agreement that the EC had violated and had provided the factual basis to support its claims.

160. The EC argued that in accordance with its usual practice, it sought to use the actual return in domestic currencies for export sales. It did so in the case of Kanebo (i.e. it used the actual returns in cruzado obtained by the company when converting the proceeds of its exports in United States dollars), because this information was supplied by the company. For the other companies no such information was provided, (but that was the informed choice of the companies themselves). The EC did not have an obligation to extract their actual returns from their export sales, because Article 6:8 of the Agreement allowed use of information available. In this case, therefore, the EC resorted to end of the month exchange rates in order to compare it with the normal values for the other companies. The EC did so in order to fulfil its obligation to compare prices "in respect of sales made at as nearly as possible the same time".

161. The EC argued that it was only if information on actual returns were not provided by the respondent, or could not be identified, that the EC had to use other data. In such a case however, given that anti-dumping proceedings must be based on objective and verifiable information, the EC found it more appropriate to have recourse to official exchange rates. In this case, the EC, having established monthly average normal values in order to permit a proper comparison with export prices, resorted then to end-of-the-month exchange rates in order to compare those normal values with export prices and to fulfil its obligation under Article 2:6 "to effect a fair comparison" through comparing prices "in respect of sales made at as nearly as possible the same time".

162. The EC argued that when GATT negotiators considered that exchange rates had to be taken into account, they had made express references in the relevant provisions of the General Agreement. In the absence of any reference to exchange rates in the Agreement, the EC considered that it did not have to take it into account. Any attempt to have them addressed by investigating authorities would be contrary to the purpose of the Agreement.

163. The EC argued that its obligations under Article 2:6 were to effect a fair comparison of normal value and export price. Such fair comparison would be ensured if the two prices were compared at the same level of trade, and in respect of sales made at as nearly as possible the same time. The EC argued that the allowances provided for in the second sentence of Art. 2:6, and in particular the phrase "and for the other differences affecting price comparability", did not include freezing of exchange rates. Indeed, a textual and systematic interpretation of this sentence led to the conclusion that the allowances could be applied to the condition to compare the prices "at the same level of trade", but normally not to the condition to make it "at as nearly as possible the same time". Indeed, all the examples cited in the second sentence related to the level of trade.

164. The EC had arrived at that conclusion on the basis of an interpretation that took into account the text, object and purpose of the Agreement as well as its drafting history. All these criteria pointed clearly to the fact that the intention of the drafters was to exclude unknown, unpredictable, highly volatile and subjective criteria when making allowances. Freezing of exchange rates by definition may affect in an unverifiable, unpredictable and usually very individual manner the exports of the exporters of a country. In addition there were a number of ways available in the financial markets for the companies to protect themselves against such effects of monetary decisions of governments. All these made it almost impossible for the investigating authorities to venture into an attempt to find how a particular exchange rate freezing affected individual exporters.

165. The EC argued that it was quite reasonable to assume that in a country faced with a situation of hyperinflation, the exporters could not accept the risks of exchange rate volatility and converted the payments of their exports at whatever spot rate prevailed when they were received. There was therefore a great amount of uncertainty about the actual behaviour of the exporters when they concluded the export contracts, the amount of information on future movements of the exchange rates when they performed them and their attitude when they converted their payments into cruzado. The exports made during the freeze concerned export contracts concluded at a time when the freeze was not yet in force; for the exports made during the freeze, it was not certain whether the payments were made while the freeze was still in place. If payments were made during the freeze, they were usually converted into cruzado at whatever spot rate prevailed when they were received. The EC argued that it could be concluded that export contracts concluded during the freeze were actually fulfilled in later months when the freeze was not in force, so that the exports and receipts of such exports were in principle not affected by the freezing of the exchange rates.

166. The EC argued that all of Brazil's arguments on the determination of dumping amounted to saying that the EC should have determined normal value in a different way. However, the real question was whether the EC has infringed the Agreement by determining normal value and dumping the way it did. The answer to that question was certainly in the negative, as it had been demonstrated that the EC had respected the single rule that is relevant for the selection of the exchange rate, namely the rule that the comparison between export price and normal value must take place at as nearly as possible the same time. For the rest Brazil had not presented a trace of evidence that the EC had acted in a discriminatory or arbitrary fashion or had committed manifest errors of fact or in the appreciation of the facts. The EC argued that Brazil's interpretation was not correct because the text did not require a "perfect" but a "fair" comparison. In this case, the EC had made a "fair" comparison by acting in the way it did. On the basis of the above-mentioned elements, the EC argued that its dumping determination in the present case was made in conformity with its obligations under the Agreement and that, consequently, the allegations of Brazil regarding Article 2:4 and 2:6 should be rejected.

167. Brazil agreed with the EC that Article 2 of the Agreement required a proper and fair, not a perfect, comparison between normal value and export price, and argued that it was not claiming that a "perfect" comparison should have been made. Brazil argued that it was simply claiming that a proper and fair comparison should have been made. That had not happened in this case.

168. Brazil then argued that there was a distinction between "devaluation exchange dumping" and the opposite phenomenon "apparent exchange dumping". "Devaluation exchange dumping" occurred where there was a manipulation of exchange rates in order to achieve a competitive advantage for exports, and "apparent exchange dumping" occurred when the export price expressed in domestic currency fell below normal value as a result of distortions in the export currency/domestic currency exchange rate. The established view of the contracting parties was that the former issue did not fall under the Agreement concerning anti-dumping. The GATT and other organisations (notably the IMF) contained other mechanisms which were intended to regulate this aspect of international trade. Accordingly, investigating authorities had no legal basis or authority for making adjustments in the context of an anti-dumping proceeding to reflect such events. In any event, contrary to the views expressed by the EC, there were no competitive devaluations in this case and the EC did not "give" the benefit of any such events to the exporters.

169. Brazil argued that the Agreement contained a number of provisions which, in Brazil's view, clearly indicated that it covered situations of apparent exchange dumping. Though the Agreement did not cover devaluation exchange dumping, such dumping was not unregulated; it was subject to a different international regulatory regime.

170. Brazil argued that: the exchange rate freeze in the present case had the opposite effect to a devaluation; Brazil had raised no issue on devaluation exchange dumping in this case; the exchange rate was not manipulated by the Brazilian authorities, but the adjustments were made in line with IMF recommendations and fully in accordance with Brazil's obligations under the GATT; the two "devaluations" in January and June 1989 which represented a decrease in the prevailing trend of currency depreciation, were well below the level of inflation, and, in the context of the freeze, could not be characterized as advantageous for exporters (see the discussion in section (b) below); and the contracting parties took the view that devaluation exchange dumping was not covered by Article VI of the General Agreement.

171. Brazil argued that this view was supported by the drafting history of the GATT and the Agreement. This clearly showed that a proposed Article 17(6) would have introduced the concept of "Dumping by means of Depreciation of Currency". This concept was rejected by the contracting parties, as shown by the minutes of the Second Session of the Preparatory Committee of the United Nations Conference on Trade and Employment. The Committee had concluded only that multiple currency practices could constitute dumping by means of a partial depreciation, but this case did not involve any multiple currency practice issues. On the other hand "apparent" or "technical" exchange dumping was the opposite phenomena, which arose not because there was a devaluation, but because the domestic currency was temporarily overvalued. Unless investigating authorities exercised due care in the conduct of anti-dumping investigations, this could lead to findings of dumping which could not be justified by underlying macroeconomic considerations. In other words, as had been illustrated in this case, there could be apparent dumping margins when, in reality, none existed. The Agreement was clearly not intended to lead to the imposition of protective measures where margins were calculated solely as a result of such "apparent" dumping. The very first recital of the Agreement made clear the contracting parties' intention to ensure that "anti-dumping practices should not constitute an unjustifiable impediment to international trade".

172. Brazil recalled that in the panel report on Swedish Steel, one of the arguments put forward by Sweden was that the determination of "technical dumping" made in that case resulted from the steady depreciation of the export currency against both the domestic currency and the currency of the third country (Germany), exports to which were used in establishing the normal (or, in United States terms, the "fair") value. Sweden had argued that the United States investigating authorities should have made an allowance under Article 2:6 of the Agreement in order to account for the exchange rate movements, which the exporters had no chance to avoid, since these movements fell within the phrase "other difference affecting price comparability". Sweden had also argued that the exporters could not adjust their prices when there were such rapid exchange rate changes, during a short period of time, given the existence of commitments to customers that limited the ability of exporters to change prices at short notice. Notably, the United States seemed to have agreed that an adjustment could be available in principle, arguing only that no such claim was made during the proceeding. In the event the Panel did not find it necessary to rule on the issue.

173. Brazil argued that the present case did not really concern "apparent exchange dumping" in general terms and that the arguments being put forward by Brazil were not the same as the arguments put forward by, for example, Sweden in Swedish Steel. The Swedish Steel case involved one currency (the United States dollar) steadily depreciating against another (the Swedish Krona and German Mark). The present case involved a very special case of "apparent exchange dumping", where one exchange rate was frozen during the investigation period. Brazil considered that this case may be distinguished from the situation prevailing, for example, in Swedish Steel, in the light of the very large distortion in the calculation that resulted. In these circumstances, Brazil did not consider that any reasonable interpretation of Articles 2:4 and 2:6 (especially in conjunction with Article 13) could possibly lead to the conclusion that a proper and fair comparison, which had special regard for Brazil's status as a developing country, could result from the use of the frozen exchange rate.

174. In the light of the above, Brazil considered that in zeroing negatively dumped transactions in the circumstances of this case, the EC violated Articles 2:6 and 13 of the Agreement, in that it did not ensure a fair comparison, did not make due allowance for differences affecting price comparability, and did not have special regard for the special situation in which Brazil, as a developing country, found itself (more details on the arguments relating to "zeroing" are provided in section (d).

175. The EC argued that it did not see any plausible or credible difference between "devaluation exchange dumping" and "apparent exchange dumping", especially as regards their effects upon trade and the possibility of dumping. A devaluation gave competitive advantage to the exporting country. However, governments of all countries frequently did not openly and formally devalue their currency, but let it slide in the financial markets. This could also have effects on trade similar to a freezing of exchange rates or a system of applying periodical adjustments in the exchange rate. From the economic and legal point of view, therefore, it did not make any sense to distinguish between "devaluation exchange dumping" (because devaluation may be brought about also by informal sliding of the currency guided by the government) and "apparent exchange dumping". If the GATT and the Agreement did not cover the first case, as admitted by Brazil, it did not make any sense to cover the second one. In case of formal devaluation or informal sliding of a currency administered by the government of a party, the effects upon trade were much easier to identify and calculate. For example, in case of a 10 per cent devaluation, this may broadly increase by an equivalent percentage the export returns, in real terms, of that country. For the purposes of identifying dumping, therefore, "devaluation exchange dumping" could have been easier to control under the Agreement than what Brazil called "apparent exchange dumping", because in the latter case there was a lot of uncertainty and unpredictability about the effects of freezing, the period that needed to be allowed to lapse before one could say that the effects of the freezing had disappeared and also the technical possibilities which exist in the financial markets for exporters to hedge against such effects. Still, Brazil admitted that the contracting parties in 1947 intended to exclude the "devaluation" dumping from the GATT, but in its view the "apparent exchange dumping" could be covered by the Agreement. Brazil was making this allegation despite the fact that "apparent" dumping presented much more complex and unpredictable economic issues. Accepting Brazil's arguments on this artificial distinction, therefore, was likely to lead to arbitrary and unpredictable results.

176. Regarding the Swedish Steel panel report, the EC argued that Brazil had explicitly admitted that the report did not concern the same factual situation as the one Brazil was claiming to exist in the present case; the legal issues in the two situations were different; and, the panel did not find it necessary to rule on the issue. The EC argued that that report was the only citation by Brazil to support its arguments regarding the distinction between "devaluation exchange dumping" and "apparent exchange dumping".

177. The EC further argued that the theoretical and unrealistic nature of Brazil's claims was further compounded by its multiple references to complex economic studies, culminating in statements like "unless investigating authorities exercise due care in the conduct of antidumping investigations this may lead to findings of dumping which are not justified by underlying macro-economic considerations". However, Brazil had made no effort to define what this standard of "due care" or what the "underlying macroeconomic considerations" would be, before dumping may be found to exist.

178. Brazil argued that the Brazilian exports of cotton yarn to the EC radically declined following the imposition of measures by the EC. The EC's denial of the effect described by Brazil (and the distinction between the two types of exchange dumping) was difficult to understand, given that it was widely recognised to exist, as set out in the various documents submitted by Brazil to the Panel. The EC was addressing only one side of the equation, namely, depreciation. However, the critical point was one of timing, and the relationship between inflation in the domestic market and exchange depreciation. In this case, the exporters were caught just at the point (January 1989) where the actions of the Brazilian government introduced a distortion of the order of 40 per cent into the calculation. That distortion remained until the exchange rate was unfrozen and the exporters became free to adjust their United States dollar export prices. This was the problem that the EC should have addressed, but did not do so.

(b) Arguments relating to the factual aspects of this case

179. Brazil argued that the distortions in this case occurred as a result of measures taken by the Brazilian government, publicly expressed to be temporary. The exporters could be absolutely certain that the cruzado would again begin to depreciate within a short period of time. Thus, the fact that the exchange rate freeze was in place for 90 days did not detract from its evidently temporary nature. The temporary exchange rate freeze in the first three months of 1989 should have been taken into account by the EC and adjustments to the exchange rates should have been made to determine the dumping margin. This was not done by the EC, which operated on the principle that it had to use only "official" exchange rates. Therefore, the EC had violated Articles 2:4 and 2:6 of the Agreement.

180. Brazil said that its understanding was that the United States dollar price of Brazil's exports were generally fixed during the period of the exchange rate freeze as a consequence of contracts entered into prior to that period. The terms and provisions set out in these contracts included an agreed United States dollar amount. Brazil informed that Panel that its understanding was that the exporters did not use forward markets. Brazil argued that contractual obligations concluded before the exchange rate freeze explained why during the first quarter of 1989, the Brazilian exporters were obliged to continue to export, despite the extremely unfavourable terms of trade. In April 1989, the volume of exports by the three companies verified fell very considerably.

181. The EC noted that the cruzado depreciated by 14.1 per cent in United States dollar terms on 16 January 1989. The EC argued that this depreciation was excessive in respect of what could be justified by the inflation rate at that time, and this devaluation arguably obtained a considerable benefit in the Brazilian export trade. Despite that benefit, the Brazilian exporters were found to be dumping in January 1989. It appeared also that for the period of the freeze, the dumping margins found by the EC by far exceeded the inflation rates in Brazil during that same period. Following small adjustments in the exchange rate, the Brazilian currency was again devalued by 10.7 per cent in United States dollar terms on 30 June 1989. This devaluation also seemed to be excessive in relation to domestic inflation and, not surprisingly, the dumping margins found by the EC fell. But in the calculation of the dumping margin, the EC again gave the benefit of this devaluation to Brazilian exporters. The EC argued that circumstances in this case could thus be considered from a perspective different from that provided by Brazil, namely that dumping was occurring in the normal course of events and that dumping only fell when devaluations (excessive in relation to inflation) were made in January and June 1989. Hence, the EC argued that if Brazil's claim was that the freezing of the exchange rate was abnormal, it may reasonably be argued by the EC that the devaluations may also be considered to be abnormal. The facts provided at least as much support to the EC's claims as they did to those of Brazil.

182. The EC argued that the commercial banks in Brazil throughout 1988 and 1989 were permitted to provide forward exchange facilities to exporters, usually for a period of up to 180 days. The Brazilian exporters had the possibility of benefiting from these and a number of other facilities during the period of the exchange rate freeze. Since 1986, exporters were authorized to open foreign exchange accounts in United States dollars at the Central Bank of Brazil. In addition, foreign exchange contracts covering transactions could be closed either prior to the shipment of goods or within 10 working days of shipment. The foreign exchange proceeds from all exports were sold at freely negotiated rates within the limits of the official market established by the Central Bank. Export proceeds had to be surrendered, however, to the Central Bank within 180 days of shipment.

183. The EC argued that, as admitted by Brazil, it was quite possible that exports made during the first quarter of 1989 were made on the basis of contractual obligations entered into during the final part of 1988 (i.e. before the freezing was in place). The EC noted that Brazil had argued that payments for the great majority of transactions were likely in fact to have been made in months following the issue of invoices, and that exporters may have set their export prices with credit terms that anticipated depreciation of the cruzado. But subsequently, Brazil had said that its understanding was that the exporters did not use forward exchange markets. However, in a letter of 14 February 1992, the legal representative had indicated to the EC that such forward exchange contracts were used by his client. In view of these statements, the EC argued that if the Brazilian government and its exporters could not make up their minds on whether such hedging possibilities were actually used, it would not be possible for the EC to do so regarding highly subjective and unpredictable behaviour on the part of the exporters.

184. In response to a question by the Panel on whether the EC would have used the rate of exchange in a forex contract if it were available, the EC argued that if the data on the actual rate used for the transaction in question were available, then actual returns in domestic currency would also be available for the export sales considered, and the EC could use that for the purpose of the comparison with normal value (as it did in this case for Kanebo). However, if data on actual returns were not available or not reliable, it would not be possible to know the effect of the use of exchange rate facilities.

185. The EC addressed Brazil's argument that the freezing of the exchange rate lasted 90 days in 1989, and this should be considered to be temporary. The EC argued that 90 days could not be taken to be a temporary fluctuation. Moreover, as admitted by Brazil itself, the abolition of the freeze did not take place all at once in April 1989, but the exchange rate was apparently allowed gradually to depreciate over a period up to June 1989. Moreover, on 30 June 1989, Brazil again devalued the cruzado by 10.7 percent in United States dollar terms. Therefore, there was considerable uncertainty regarding the level and effects of the Brazilian authorities' intervention for a period ranging from January to June 1989, i.e. about 180 days, which could not be considered to be temporary by any standard.

186. Brazil argued that the Brazilian currency had been depreciating since January 1988 against the United States dollar at more than 15 per cent each month (more than 30 per cent in December 1989). On 16 January 1989, the cruzado was devalued by 14.1 per cent, and in January 1989 as a whole there was a depreciation of 30.67 per cent compared to the rate prevailing in December 1988. However, it was incorrect to view the events of January 1989 as a devaluation because domestic inflation during January 1989 was 70.28 per cent (i.e. 2.27 per cent per day). Therefore, although the domestic currency partially depreciated to reflect domestic inflation, it was clearly overvalued in January 1989 as a whole. Thus, it was wrong to portray the change in exchange rate in January 1989 as some kind of government sponsored devaluation designed to increase competitiveness on export markets. Indeed a devaluation produced no competitive advantage in export markets if the proceeds of export sales were not freely convertible into the domestic currency. In this case, the exporters were obliged to abide by the frozen exchange rate and far from favouring exporters, the measures taken by the Brazilian Government were very unfavourable to exporters. Brazil argued that a fair estimate of the extent of the overvaluation in January 1989 could be the difference between the estimates of inflation and devaluation given above, i.e. 39.61 per cent. Alternatively, an exchange rate for January 1989 could be derived by indexing the December 1988 exchange rate to domestic inflation. Moreover, as indicated earlier, the contracting parties had agreed that there was no place in the context of anti-dumping investigations for any kind of adjustment in response to IMF and GATT compatible devaluations. The EC did not therefore "give" any "benefit" to the exporters by not seeking to make such an adjustment. Also, the EC had in any event alleged that it was not empowered to make any such adjustment under the Agreement.

187. Brazil argued that it was not correct that the exporters were free to hold foreign currency abroad or to chose the time of conversion. In fact, the time during which they could hold foreign currency was limited. Under regulations applicable in Brazil at the time, the exporters were obliged to convert the proceeds of their export transactions into cruzado at the frozen rate of exchange within a specified period (usually 20 days).

188. Brazil argued that at the time of the freeze (immediately before which the currency was rapidly depreciating, as it had been doing for some time) the forward exchange markets were not well developed. This was because the nature of a forward exchange market involved the matching of the requirements of both buyers and sellers of foreign exchange. Where a currency was moving rapidly in a single direction the normal functioning of such markets could be seriously disrupted. In such circumstances the conclusion of a forward exchange contract may depend on the presence of speculators, of whom there were very few in such volatile conditions. Furthermore, access to the market was difficult and expensive and required a degree of resources and technical knowledge not readily available in small companies located in developing countries. Brazil therefore considered that the EC's view that the exporters should have used such markets was not reasonable, especially in the light of the requirement imposed by Article 13 that the EC should have had special regard to the situation of Brazil as a developing country. In any event, it appeared that the EC would not have accepted the use of such exchange rates even if they had been available. Brazil argued that on 14 February 1992 (well before the imposition of definitive measures by the EC) the legal representative of ABIT wrote to the EC explicitly requesting that forward exchange rates could be used. However, that request was not accepted by the EC. It appeared therefore, that whether or not such rates were available, the EC would not have accepted their use. Hence, the arguments of the EC on this point should be rejected.

189. Regarding the importance attached by the EC to the fact that dumping margins were found in January 1989, and to the fact that for the period of the freeze the dumping margins calculated exceeded the rate of domestic inflation, Brazil argued that it was not surprising that dumping margins were found in January 1989: the currency depreciated by only 30.67 per cent by comparison with December 1988, whilst inflation rose by 70.28 per cent. Brazil argued that following the apparent reasoning of the EC, that might lead to an expectation of a dumping margin of 39.61 per cent, but in fact the dumping margins found were all well below this figure (Nisshinbo 32.04 per cent, Renda 6.43 per cent, Kanebo 12.11 per cent). Furthermore, Brazil did not agree that the dumping margins found during the period of the freeze were generally well in excess of what might be expected given the development of inflation. Brazil referred to the data that it had provided to the Panel (see section (d) below) to argue that the facts showed clearly that the dumping margins found were consistently less than what might have been expected given the divergence between the cumulative rate of inflation and the frozen exchange rate.

190. Brazil argued that the data submitted by it to the Panel fully illustrated the way in which the FIBGE exchange rates developed during the relevant period, and their relation with the dumping margins calculated by the EC (arguments relating to this aspect are presented in more detail in section (d)). For all three Brazilian companies verified, it was clear that from January through May 1989, the dumping margin steadily increased as the gap between the inflated normal values and the frozen exchange rate widened. Only when the exchange rate finally stabilised (around June or August) was the situation normalised, and there was a sudden decrease in dumping margins. There was absence of dumping margins in June and they were very low at the end of the year. Brazil argued that the margins calculated for Nisshinbo and Kanebo in November and December 1989 resulted from the (incorrect) use of costs of production as the basis for normal value, a methodology notorious for producing high margins. Therefore, the EC was not correct that the facts provided at least as much support to the EC's claims as they did to those of Brazil.

191. The EC argued that a review of the development in 1989 of the average monthly exchange rates of the Brazilian cruzado against the United States dollar with the development of the Brazilian wholesale price index showed that, contrary to the allegation by Brazil about a special situation prevailing between January and April 1989, the differential between inflation and devaluation for this period never exceeded 3.32 per cent. Compared to the average deviation for the rest of this year (1989) of 3.34 per cent, this deviation seemed absolutely normal. It could be concluded from this that the monthly variation of the dumping margins could certainly not be caused by a distortion of the exchange rates and the rate of inflation.

192. Therefore, the EC argued that if an investigating authority had recourse to "lagged" or indexed exchange rates, it would be disregarding the behaviour of the exporters and applying a presumption that freezing of exchange rates must have created dumping margins which could not be attributed to the operators. In any event, before attempting to correct the difference between the development of the exchange rates and the domestic inflation, one had to identify whether any such difference existed. If not, the freedom of the exporters to react to the market situation, which was an important factor in identifying dumping, would be systematically ignored.

193. Regarding Brazil's argument that there was a sudden reduction in the dumping margin in June 1989, complete absence of a dumping margin in July 1989, and that margins were very low at the end of the year, the EC argued that one explanation of such an alleged drop in the dumping margin could also result from the fact that the export prices during the third and fourth trimester of 1989 increased by 14 per cent and 16 per cent on average (first trimester = 100). However, Brazil did not even consider this possible explanation.

194. Brazil argued that the overvalued domestic currency produced an upward pressure on export prices expressed in United States dollars. This did appear to have led to some increases in the United States dollar export price during the investigation period (although Brazil had not been able to confirm this point). The EC, however, offered no clear analysis on this question, nor did the Determinations imposing the measures contain any reasoning on this issue. The figures mentioned by the EC were not supported with any explanation. It was not clear what kind of average had been calculated, and in which currency or at what level of trade had the calculations been made. Nor was it clear for which company (or companies) and for what product categories the figures corresponded to. All of these factors resulted in material differences in the relevant prices. If they had not been taken into account by the EC (as seemed to be the case), the figures cited were meaningless and should be rejected.

195. Brazil argued that in any event, to the extent that there may have been increases in the United States dollar export prices, this was entirely consistent with the arguments presented by Brazil, and indeed supported them. The distortion introduced in January 1989 could only be removed in the long term by a combination of the releasing of the frozen exchange rate and some readjustment of export prices by the Brazilian exporters. Indeed, practice in the jurisdictions of other contracting parties was to look for positive efforts by exporters to adjust their prices in the medium to long term as one of the criteria for deciding whether or not to grant an allowance to take into account exchange rate movements. To the extent that such an adjustment could be identified in this case, it further supported the argument that an allowance should have been granted by the EC.

196. Brazil argued that the devaluation of June 1989 must be seen in the context of the freeze imposed during 1989. During the 6 months to December 1988, the currency depreciated by 216.59 per cent, whilst during the first 6 months of 1989, as a result of the controls imposed, it depreciated by only 51.9 per cent. It was self evident that once controls on the currency were lifted, inflationary pressure would once again cause it to depreciate. The important distinction was between releasing the brake on depreciation on the one hand, and accelerating depreciation on the other. In any event, inflation in June 1989 was 24.83 per cent. In this context, a devaluation of 10.7 per cent on 30 June 1989 was clearly not "excessive". Hence it was not correct to portray the June 1989 change as a devaluation designed to increase competitiveness on export markets. Furthermore, in any event, because of its rule that official exchange rates were to be used in the investigation, the EC was not at liberty to make any adjustment in relation to the circumstances described.

197. The EC noted Brazil's statement that its currency depreciated by 216.59 per cent during the six months to December 1988, whilst during the first six months of 1989 it depreciated by only 51.9 per cent. The EC said that from this data, one could argue that such an enormous amount of depreciation had given its exporters in any case such a competitive benefit to practice dumping that the freezing of the exchange rates, for the three months of 1989, was in any way unable to take it away. This also highlighted the fictitious distinction operated by Brazil between "devaluation exchange dumping" and "apparent exchange dumping" (arguments relating to these two concepts are given in section (a)).

(c) Arguments relating to lack of conformity with practice in major anti-dumping jurisdictions

198. Brazil argued that the principles of the Agreement mentioned above were fully supported by the existing practices of the main anti-dumping jurisdictions, e.g. the United States, Australia, Canada and the EC. Since the early 1970's, the United States had recognized the importance of assessing the impact of exchange rate fluctuations in anti-dumping investigations, i.e., when the assessment revealed that exchange dumping had occurred, duties were not imposed. Australia considered the impact of exchange rates not only during the determination of the dumping margin, but also during the material injury determination. In this respect, short term dumping resulting from temporary exchange rate fluctuations was considered too insubstantial to cause material injury. The effects of exchange dumping were considered no stronger than the effects of other variable factors with which industries were accustomed to dealing in the normal course of a constantly changing business environment.

199. Brazil argued that the Canadian provisions relating to exchange rate conversions were less developed. The general rule was that conversion should be carried out on the date of sale. However, the Canadian authorities recognized that special considerations applied in the case of "unstable economies", "rapidly devaluating currency" and "fluctuating currency exchange rates". The Canadian authorities did not share the view, taken in this case by the EC, that such factors should simply be ignored. Brazil further argued that even the EC's position in this case was contrary to its own well-established practice.

200. In view of the preceding considerations, Brazil argued that the determination of dumping by the EC in this case violated its obligations under the Agreement because the EC ignored the exchange rate fluctuations in this case.

201. The EC argued that since the Agreement left the issue of exchange rates used in determination of dumping entirely to the domestic law of the Parties to the Agreement, citing the practice on this issue of other major Agreement jurisdictions was of no relevance in deciding whether the method used by the EC in this case respected the requirements of Article 2:6 of the Agreement. Whatever criteria the domestic law of the importing country was using for the purpose of effecting the comparison, the only obligation imposed upon that country by the GATT was to use fair criteria in a transparent and non-discriminatory manner to all other signatory countries in comparable situations and not to commit manifest or arbitrary errors in establishing and appreciating the facts in the process. Both the GATT and the Agreement contained no specific provision on this issue and left it entirely to the domestic law of the signatories. Moreover, none of the examples cited by Brazil supported Brazil's claims. The cases cited only called for "flexibility" and for a case-by-case approach. But none of them required the use of unofficial "lagged" exchange rates for a period of 180 days. Also, Brazil's statement that the EC simply ignored exchange rate fluctuations was not correct because the recourse to monthly exchange rates in circumstances justifying such recourse was expressly designed to take those circumstances into account.

202. In response to a question by the Panel, the EC argued that Brazil did not claim that the practice of the parties referred to by Brazil constituted subsequent practice within the meaning of Article 31.3(b) of the Vienna Convention on the Law of Treaties. Rather, Brazil had simply stated that the principles stated by it in the submission were "fully supported by the existing practices of the main anti-dumping jurisdictions." Thus, it was clear that Brazil was giving examples of practices of the parties and did not claim that there had been any subsequent practice resulting in a particular interpretation or modification of the Agreement. In addition, the EC had used the same practice as in the present case several times in the past against Brazil without it ever complaining about the method of comparison used.

203. The EC argued that even if one were to admit that Brazil claimed that there was a subsequent practice, the conditions for having such a practice accepted here were not met. Article 31 of the Vienna Convention provided that in the interpretation of treaties, there shall be taken into account, together with the context "any subsequent practice in the application of the treaty which establishes the agreement of the parties regarding its interpretation". Article 31.3(b) hence required that the practice be one developed in the application of the treaty (practice of non-parties was not relevant) but also that the practice be an agreed practice. Furthermore, it must be agreed by all the parties to the treaty. This was confirmed by authors who considered that a practice must be concordant, common and consistent and could not in general "be established by one isolated fact or act or even by several individual applications". In the present case, Brazil had mentioned only four Parties to the Agreement, which were far from being similar nor did they reach a sufficient degree of consistency. In any event, the EC did not subscribe to the practices followed by the other Parties referred to by Brazil. Also, the principle of subsequent practice in the sense of Article 31 of the Vienna Convention had been raised several times in the past in dispute settlement proceedings under the General Agreement, including by the EC, but it had never been accepted by the Panels. For all these reasons, the practice of those countries was of no relevance to the interpretation of the Agreement.

204. Brazil disagreed with the EC's argument that the practice of the contracting parties was of no relevance to this case . Brazil argued that, for example, the EC's own Regulations on anti-dumping explicitly stated that: "Whereas, in applying these rules it is essential, in order to maintain the balance of rights and obligations which these Agreements sought to establish, that the Community take account of their interpretation by the Community's major trading partners, as reflected in legislation or established practice" (emphasis added by Brazil). This balance was absent from the EC's determinations in this case. Brazil then referred to the following text of the United States Federal Appellate Court in Melamine versus United States case: "The purpose of this regulation is clear. Anti-dumping investigations are meant to determine whether prices of merchandise sold in the United States are at less than "fair value". When exchange rates are fluctuating substantially, a given dollar price of a product in the United States could change technically from fair to "unfair" literally from day to day, even if the foreign price of the product denominated in the foreign currency, also remained constant. This result is not called for by the language or purpose of the Act. It would be unrealistic to expect business to change prices instantaneously to take account of fluctuating exchange rates. So too, weekly price changes could create substantial confusion and inconvenience for the customers of that business. The regulation, then, allows a reasonable period in which the business may take sustained exchange rate fluctuations into account. The regulation further instructs that temporary fluctuations should not be the sole basis for determination of less than fair value sales. Businesses are to be given time to assess whether one currency has truly appreciated against another before changing their pricing practices ... The purpose of the anti-dumping law, as its name implies, is to discourage the practice of selling in the United States at LTFV by the imposition of appropriately increased duties. That purpose would be ill-served by application of a mechanical formula to find LTFV, and thus a violation of the anti-dumping laws, where none existed. A finding of the LTFV sales based on a margin resulting solely from a factor beyond the control of the exporter would be unreal, unreasonable, and unfair." (emphasis added by Brazil)

205. Brazil further argued that while the discussion of indexed and lagged exchange rates and of the practice in other major anti-dumping jurisdictions were not essential for the above finding to be made, they were clearly useful in that they demonstrated that perfectly reasonable methods were available by which the EC could have taken the relevant distortion out of the calculation, and ensured that a proper and fair comparison was made. Brazil argued that though the interpretations given by other contracting parties did not represent binding legal precedent, neither did the reports of previous panels (especially if not yet adopted by the contracting parties), to which the EC had made several references in its submission. However, Brazil considered that it was acceptable, even desirable, to have regard to the practices of such authorities in assessing the reasonableness of a particular interpretation of the Agreement.

206. Brazil then argued that on the basis of the arguments advanced by the EC, the EC apparently considered the practice of the United States on exchange rates in the context of anti-dumping proceedings to be in contravention of this principle.

207. The EC argued that it had not stated or considered that the practice of the United States on exchange rates infringed the Agreement on the point of a consideration of exchange rate. The argument of the EC was a different one, namely that the Agreement was silent on the issue of exchange rates. Therefore, each Member was free to apply its own practice on this issue. The EC also disagreed with Brazil's allegation that the practice of the EC in this case was not in conformity with the EC's own legislation and established practice. The EC argued that there was nothing in the practice of the EC which was contrary to its legislation. Moreover, the practice was in conformity with its legislation as interpreted by its highest Court (the European Court of Justice; hereinafter "ECJ") as: "In that regard it should be pointed out that the purpose of defining a dumping margin is to correct the effect which imports of products from non-members countries at dumped prices actually had on the corresponding Community industry. That effect can only be assessed in the light of the official exchange rate on the basis of which international trade transactions are taking place." Following this practice, the EC applied in this case, as in other cases, the official exchange rate on the basis of which international trade transactions are taking place.

208. The EC argued that the investigating authority had to base its determinations on objective and verifiable facts. Official exchange rates were, within the framework of an anti-dumping procedure, objective and reliable facts. This was the reason for which the ECJ ruled that the effect on imports of products from non-member countries at dumped prices could "only be assessed in the light of the official exchange rates on the basis of which international transactions take place". Allowing investigating authorities to use any form of exchange rates, such as consumer purchasing-power parities as it was argued in the case before the ECJ, would open the door to subjective judgement and risks of discrimination. Consequently, the investigating authorities were under an obligation to use official exchange rates if recourse to exchange rates had to be made. This did not mean, however, that, if the respondent could provide reliable data on the actual returns on its sales to third countries, the investigating authorities would be precluded from using such returns. The obligation imposed by the ECJ applied where the investigating authority had to resort to exchange rates.

209. Brazil argued that the EC chose to use different methodologies in the two cases, which contradicted the EC's argument that it was constrained by the "principle of monetary neutrality" to use the exchange rate that it did. In reality, there appeared to be a degree of flexibility in the general practice of the EC - a flexibility which was clearly lacking in this case.

(d) Arguments relating to the alternative estimates presented by Brazil

Introduction

210. Brazil submitted to the Panel alternative estimates regarding dumping margins that had been calculated by the EC in this case. The data submitted by Brazil were (1) a chronological re-ordering of the estimates made by the EC in this case; (2) estimates of the dumping margins with exchange rates indexed to the inflation rate; and (3) estimates of dumping margins with lagged exchange rates, i.e. exchange rates with a lag of two months. Brazil argued that these estimates showed clearly that if the EC had made the required adjustments, Brazilian exports would have been determined to have de minimis dumping margins.

211. Brazil argued that in a high inflation environment, distortions could occur if an average annual normal value was compared with export prices converted using spot exchange rates. Transactions at the beginning of the year would tend to generate very high dumping margins and those at the end of the year would tend to generate very high negative dumping margins. Though the EC purported to reduce the risk of such distortions by calculating monthly average normal values and using monthly average exchange rates, they were not entirely eliminated. As a result, from one month to the next, it was common to find large variations in dumping margins in a high inflation environment. In these circumstances the EC's practice of treating a negative dumping margin as zero tended to artificially inflate the final dumping margin calculated, and thus had a particularly prejudicial effect. In view of the special situation of exchange rate freeze and high inflation in this case, had the EC not relied on its methodology of treating negative dumping as zero, then the dumping margins would have been even lower than those shown by the estimates provided by Brazil to the Panel.

212. The EC argued that if the Panel were to assess the estimates presented by Brazil then it would be conducting a de novo review. This would exceed the mandate of the Panel. Also, the EC said that the criteria for judging the EC's action was whether it had met the requirements of the Agreement, and not whether a "better" calculation method could have been used. The EC argued that it had met all the requirements of the Agreement in this case. Moreover, the EC presented the Panel with its own estimates of dumping margins which it had calculated using indexed exchange rates, and argued on the basis of those estimates that the results would not have changed if the EC had used the methodology suggested by Brazil, and in some instances the margin of dumping would have been even higher than those calculated by the EC in its investigation.

213. The EC argued that Brazil was introducing a new concept of "negative dumping" that did not exist under the Agreement. Under the Agreement, there was either dumping or no dumping. The EC also argued that the dumping margins calculated on the basis of its methodology provided a fair estimate of the incidence of dumping and showed the propensity of the exporter to dump.

214. The arguments below are presented in two sub-sections, the first of which relates to the estimates presented by Brazil and the second relates to the arguments regarding negative dumping.

(i) Estimates of dumping margins presented by Brazil

215. Brazil argued that when converting an export price into a domestic currency for the purposes of comparison with a normal value expressed in domestic currency, an important question was the relevant date that should be used for determining the exchange rate to be applied. Generally accepted practice was that the exchange rate prevailing on the date on which the material terms of the sale were agreed should be used. That was usually the date of invoice or of shipment, rather than payment. The accrued amount was posted in the exporter's sales account on the date on which the sale was agreed, using the then prevailing exchange rate. Gain or loss resulting from exchange rate movements between accrual and payment was posted in an exchange rate gain/loss account. By way of exception, where the exporter had concluded a forward exchange contract directly related to the export transaction, the exchange rate of that contract could be used. Brazil argued that this approach would relate primarily to low inflation environments. In a high inflation environment it was often logical for exporters to set export prices at a level which anticipated depreciation of the domestic currency. This effect was particularly acute where exchange rates had been temporarily frozen (as in this case): the exporter had a reasonable expectation that by the time he would be paid (for example, in two months time) the exchange rate would have returned to its normal level. In such a case it would be reasonable to use the exchange rate prevailing at a later date than the date of sale, i.e. to use a lagged exchange rate.

216. Brazil said that the data supplied by the EC in its disclosure letter was not arranged chronologically. Brazil provided the Panel with a chronological ordering of this data, in a manner that the monthly sub-totals were also shown. Brazil argued that the data confirmed that the temporary exchange rate freeze generated artificially high dumping margins, i.e. there was exchange dumping in this case. The dumping margins in the first part of the year were high and they were low or de minimis in the second part of the year. Brazil argued that, given the high inflation environment in this case, the EC's failure to take into account negative dumping inevitably led to a distorted and artificially high dumping margins in this case.

217. Brazil presented to the Panel recalculated margins that would have resulted from the use of an exchange rate indexed to domestic inflation and an exchange rate prevailing two months after the shipment to argue that if the EC had made such adjustments, the dumping margins would have been de minimis. Brazil submitted to the Panel the following estimates of average dumping margins (in per cent) for Nisshinbo, Renda, Kanebo, and other cooperating exporters: those calculated by the EC (12.08, 6.99, 15.84 and 12.89 respectively); those calculated by Brazil on the basis of indexed exchange rates (0.97, 0.85, 2.45 and 1.68 respectively); and those calculated by Brazil on the basis of lagged exchange rates (0.97, 0.76, 0.3 and 0.59 respectively). Brazil argued that the margins would be even lower if the (incorrect) use of cost of production as the basis of establishing normal value in October, November and December 1989 for Nisshinbo and Kanebo was replaced by domestic sales of these companies as the basis for normal values. Brazil recalled that certain countries had been eliminated from the proceedings by the EC on the grounds of de minimis dumping margins, namely, India (dumping margins ranging from 0.1 to 9.5 per cent and weighted average of 1.8 per cent), Thailand 0.1 to 7.9 per cent), and Egypt (0 to 0.4 per cent and weighted average of 0.1 per cent). Brazil argued that in view of the treatment of the other countries which were eliminated from the proceedings, the correctly calculated dumping margins for Brazil would have similarly led to a termination of the case for the Brazilian exporters without the adoption of the measures.

218. The EC first recalled its procedural objection to the calculations provided by Brazil to the Panel, and considered that the issue of alleged use of erroneous statistics was a typical example of a request by a party to have the investigation reviewed de novo by the Panel. The EC then argued that calculation methods used by Brazil were nowhere provided for in the Agreement and were not used by the EC in this case. A calculation method based on indexed or lagged exchange rates was bound to provide a biased and unacceptable comparison of the normal value with the export price and was contrary to the provisions of the Agreement. Therefore, the EC suggested to the Panel to dismiss as totally irrelevant the allegations of Brazil resulting from such an alternative calculation method.

219. Regarding the chronologically arranged data, the EC argued that it only showed that the dumping margins were higher when the exchange rate was held high by the freeze. This merely reflected the factual situation, and Brazil had not explained why this violated the Agreement. The EC also observed that during the period of the freeze, the three investigated companies did not appear to have been affected to the same degree by the freezing of the exchange rates. Nisshinbo's dumping margin in February to May 1989 was as high as that of Rendas but the latter's annual dumping margin was almost half of that of Nisshinbo. On the other hand, Kanebo's dumping margin during the same period of February to May 1989 was much lower (almost half) of that of the other two companies, but still its annual dumping margin was found to be almost as high as that of Nisshinbo. Moveover, during the first three months of 1989, the three companies were exporting increasingly higher quantities of cotton yarn, which were also higher in terms of CIF value. This indicated that there was not only "exchange rate" dumping, as Brazil argued, but real dumping practices were also involved. Moreover, the EC argued that the dumping margins of the three companies in January 1989 diverged greatly, despite the 14.1 per cent devaluation made on 16 January 1989, and continued to be high in May and June 1989 despite the abolition of the freezing of the exchange rates in April 1989. On the other hand, the dumping margins fell substantially or were completely eliminated in July 1989, most probably after the 10.7 per cent devaluation made on 30 June 1989. Thus, viewed individually, the dumping margins, quantities and value of the three companies during the entire investigation period exhibited certain features which were not only easily reconcilable with Brazil's claims, but which may be said to also support the EC's arguments. Furthermore, the chronologically arranged data provided by Brazil did not help answer the question before the Panel, which was whether the method used by the EC to calculate dumping margins in this case was in conformity with the Agreement or not.

220. The EC argued that the recalculations of the margins provided by Brazil were inaccurate: (1) Brazil had suggested that the exchange rate for December 1988 should, inter alia, have been used. However, December 1988 was outside the investigation period, and a methodology which required the use of data outside the investigation period undermined the certainty and objective verifiability of the investigation itself. Furthermore, the exchange rate used by Brazil for January 1989 did not represent the average rate for that month but probably the highest applicable for that month; (2) For one of the three companies investigated, the data did not reflect the data actually used by the EC. For Kanebo the EC used the actual returns in cruzados obtained by the company when converting the proceeds of its export sales in United States dollar. Thus, the whole of Brazil's argument on the more appropriate nature of "lagged" or "indexed" exchange rates fell altogether. For the other two companies, the EC could not use the same kind of data, because they were not made available to the EC, either through inability or unwillingness on the part of the companies. Thus, the EC, having established monthly average normal values, resorted to monthly exchange rates in order to compare those normal values with export prices. Furthermore, even though the EC was of the opinion that monthly average exchange rates would have been fully appropriate and in conformity with the Agreement, in this particular case it resorted to end-of-month official rates, which already contained an element of "lagging" when compared with the date of the actual exchange transactions. In these circumstances, it was hardly surprising that dumping margins for Kanebo were consistently higher than those calculated for the other companies.

221. With respect to lagged exchange rates, the EC wondered why two months should be better than three months or two weeks. The choice of making adjustments of exchange rates or of using lagged exchange rates created an additional layer of discretion. Given the wide varieties of factors to be taken into account in attempting to make monetary adjustments, exercise of such discretion was not required by the Agreement. In any event, an investigation could only be based on objective and verifiable facts, which alone would satisfy the need to assure fair and predictable treatment of all parties concerned.

222. Brazil disagreed with the EC's statement that the exporters were not "affected to the same degree" by the exchange rate freeze. Brazil argued that normally one would not expect the dumping calculation to produce the same absolute result for the three companies given their entirely independent cost and pricing structures, and the uneven distribution of their sales volumes over the year. However, a consideration of their experience showed a remarkable and striking correlation for the three, totally independent, sets of figures. Most notable were (1) the steady increase from January to May 1989; (2) the sudden reduction in June 1989; and (3) the complete absence of a dumping margin in July 1989.

223. Brazil argued that the data provided by it to the Panel showed a clear correlation between "a change in the factual situation" during the investigation period and its "consequences". From this, Brazil believed that the "consequences" illustrated (the trend in dumping margins) did not flow from factors peculiar to particular companies (such as changes in pricing policy). The "consequences" flowed from factors common to the three companies concerned, i.e. the business environment is which they were operating. The "change in the factual situation" which led to the trends illustrated was the exchange rate freeze and high domestic inflation. Brazil argued that it had demonstrated this not only by a careful theoretical explanation of the exact effects of the distortion in the context of the dumping calculation, but also in practice by showing how these effects are apparent from the disclosure data provided by the EC.

224. Brazil claimed that the EC was not entitled to simply ignore and refuse to consider the problems arising in this case. Brazil requested the Panel to consider whether or not a reasonable assessment of the facts before the EC authorities could have led to the conclusion that a proper and fair comparison had been made between normal value and export price, without any account having been taken of the exchange rate freeze. Brazil argued that the objective answer to that question must be that no reasonable assessment could reach the conclusion that a proper and fair comparison was made. Brazil considered that such a finding would be sufficient to dispose of the case.

225. Regarding the EC's argument that Brazil had used data from outside the investigation period (i.e. data based on January 1988 had been used), Brazil argued that the December 1988 rate of exchange was not essential to its arguments. Brazil was not arguing that indexing should have occurred, only that steps should have been taken to deal with the distortion that arose. There were a number of other ways, apart from indexing, in which Brazil considered this could reasonably have been done. Further, Brazil argued that there was no verifiability problem in relation to exchange rates that it had used. Data for December 1988 could be verified as easily as data for January 1989, since both came from the same source, i.e. the Brazilian official agency FIBGE.

226. Brazil argued that the disclosure by the EC that it had used actual returns from Kanebo and end-of-month exchange rates for Nisshinbo and Renda did not affect the chronologically ordered data provided by Brazil to the Panel, because that data was exactly the data on the basis of which the EC had reached its determination. The clear correlation (independently for each of the three verified companies) between the exchange rate freeze and the unrealistically high dumping margins found by the EC was not affected in any way. For Kanebo, Brazil had recalculated the margins and presented the estimates to the Panel by extracting the relevant data from Kanebo's questionnaire response and applying the appropriate exchange rate. Brazil argued that the figures still showed that the resulting margin would have been de minimis in the case of the lagged calculation, and would have been even smaller in the case of the indexed calculation.

227. Brazil further argued that the alleged use of the "actual returns" for Kanebo did not in any way undermine its arguments because the subject exports during the first quarter of 1989 were made on the basis of contractual obligations entered into during the final part of 1988 (i.e. before the freeze was in place). In addition, the exporters were not free to hold export earnings indefinitely, but were under an obligation to convert such earning into cruzado using the frozen exchange rate. Finally, there was no realistic prospect for the exporters to resort to a futures market in this case. Thus, the question of what the exporter's actual returns were had no bearing on Brazil's argument that in this case with a distortion, a proper and fair comparison between normal value and export price was not made.

228. The EC said that to verify the estimates presented by Brazil to the Panel, it had calculated two types of estimates of dumping margins. One was by using an exchange rate indexed to inflation for the first three months only (leaving the calculation for the other months of the reference period as in the EC original calculation), and second was by using an exchange rate indexed to inflation for the entire reference period. The actual exchange rates and inflation rates used by the EC were taken from the details used in the calculation of the revised estimates (using indexed and lagged exchange rates) of dumping margins provided by Brazil to the Panel. The EC argued that the results obtained by it were very different from those shown by Brazil.

229. The EC argued that if an indexed exchange rate were used for only the first three months of the investigation period (January to March 1989), there was actually a reduction of only about 0.4 per cent of the total dumping margin for all the companies concerned. The use of an indexed exchange rate for the whole investigation period on the other hand, led to increases of between 0.9 per cent and 1.4 per cent of the dumping margins for Nisshinbo, Kanebo and the other exporters. In both instances, however, the dumping margins of the companies remained high, and did not become de minimis. The EC did not understand how and by means of what type of data Brazil had arrived at the conclusions that if indexed exchange rates were used the dumping margins for the companies would have been 0.97 per cent for Nisshinbo, 0.85 per cent for Renda, 2.45 per cent for Kanebo and 1.68 per cent for the other cooperating companies. The calculations of the EC showed that the dumping margins were respectively: 12.99 per cent, 6.65 per cent, 17.24 per cent, and 13.78 per cent.

230. Thus, the EC argued that Brazil had failed to supply any real evidence or had supplied incorrect information to the Panel to support its allegation that if indexed exchange rates were used by the EC substantial dumping margins would not have been found. The EC recalled that Brazil had itself indicated to the Panel that it could accept a recalculation based on indexing of the exchange rate for the first 3 months of 1989 only, or indexing it for the whole reference period. This was precisely what the EC had done in the calculations above, but still the dumping margins remained almost the same. It followed that the whole factual basis of Brazil's contentions was baseless.

231. The EC then recalled that according to Brazil, a possible way of achieving a proper and fair comparison was to use the domestic rate of inflation to deflate all normal values, expressing them at 1 January 1989 levels and using an exchange rate prevailing in January 1989 as the basis for comparison with export prices. The EC provided to the Panel its calculations of dumping margins based on that suggestion, i.e. by eliminating the influence of inflation on export prices and normal values. Except for adjusting normal values and CIF values of each month in cruzado to the price levels of January 1989 by dividing them with an inflation index for each month on the basis of January 1989 = 1, all other information was the same as that used in the previous calculations and provided by Brazil to the Panel. The results of these calculations showed that the average dumping margin for the three companies concerned increased from 12.8 per cent to 19.4 per cent.

232. The EC argued that by submitting the above factual information to the Panel the EC had not changed its position on the principle that panels under the dispute settlement system of the Agreement were not required nor equipped to undertake a de novo review of the facts as made by the investigating authorities. The EC was still of the same view, and requested the Panel to confirm the above principle. In this case, the EC had exceptionally agreed to supply this information to the Panel in order to show the incorrect and baseless allegations of Brazil on the facts and legal issues. But this should not constitute a precedent.

233. The EC also pointed out to the Panel that in another anti-dumping case concerning ferro-silicon from Brazil, the exporters had provided the EC with different exchange rates from those given by the exporters and Brazil in the cotton yarn case for the same year (i.e. 1989).

234. Brazil argued that all of the data provided to the Panel by Brazil had been presented in a totally transparent and verifiable way. Though the EC was now stating that it did not understand how the calculations had been made, it had never asked for clarification of the figures or of any of the numerous and lengthy explanations already given by Brazil. In Brazil's view the relevant data was clear.

235. Regarding the EC's point that the data presented by Brazil were not verifiable, Brazil argued that the Directorate-General II of the EC (responsible for economic and monetary affairs) maintained records on global exchange rates, and it was the EC's practice to refer to such rates for the purposes of its anti-dumping investigations. Such data may either be obtained directly from the relevant central bank or other agency, or be taken from public documents, such as newspapers, where they are regularly published. Brazil did not therefore agree that there was any issue of verifiability in this case. Indeed, as stated by the EC at recital 9 of the Provisional Determination, the EC sought and verified all information it considered necessary for its determination.

236. Brazil said that it was difficult to comment on either the source or accuracy of data allegedly used by the EC in the course of other EC anti-dumping proceedings, especially because, as far as Brazil was aware, such data had never been published by the EC in its determinations, and the EC offered no documents or explanation clarifying the circumstances in which such data may or may not have been used by the EC. However, Brazil confirmed that the data submitted to the Panel by the EC on the exchange rates that the EC used in the calculation in the cotton yarn case corresponded to the exchange rates in the EC's disclosure letter. With regard to the exchange rates allegedly used by the EC in the ferro-silicon proceedings, Brazil argued that the EC had given no further indication of what these might be. There appeared to be a reasonable possibility that these may be average monthly exchange rates, as stated by the EC to be the case at recital 17 of Council Regulation (EEC) No 1115/91. However, Brazil noted that the basis on which the EC gave data for the first 8 months of 1989 in that case was not clear because recital 11 of that document stated that the investigation period covered the period 1 September 1989 to 30 April 1990.

237. Brazil argued that the calculations presented by the EC to the Panel should be rejected for at least two fundamental reasons. One, the EC had ignored domestic inflation in January 1989. January 1989 was the critical month in the calculations, for it was in that month that the exchange rate was frozen at a level showing depreciation of only 30.67 per cent, when monthly inflation was 70.28 per cent. The difference between these figures, i.e. 39.61 per cent, gave rise to the gross distortion in the dumping calculation. Two, the EC had made a simple month-by-month comparison, instead of a cumulative comparison made by Brazil. The EC's calculation assumed that from one month to the next, exporters should have made huge adjustments to their export prices in order to avoid a finding of dumping. However, the exporters were constrained by fixed contracts and, in any event, could not hope to adjust their prices by such a large amount within such a short space of time. In this respect, Brazil recalled that Article 2.4.1 of the Uruguay Round Agreement on the Interpretation of Article VI of the GATT 1994 referred to a period of at least two months for adjustment of prices by exporters.

238. Brazil further argued that the data presented by the EC were not transparent and no clear and precise explanation was offered of how the calculation had actually been made. The calculation appeared to have been made on a monthly basis for all companies together and for all types together. On these three points, the EC's recalculation departed radically from the method used in the EC's calculations that were contained in its disclosure letter. The latter were made separately for each company, for each type and on a transaction by transaction basis. For this reason alone, the new figures presented by the EC should be rejected.

239. Furthermore, Brazil argued that even the most simple verification showed that the figures presented by the EC to the Panel were not the same as those contained in the EC's disclosure letter. For example, the total quantity figures for exports in kilograms given in the EC's disclosure letter for January 1989 were as follows: Kanebo (19,440); Nisshinbo (75,025); Renda (54,979); Total (149,444). Yet the figure presented by the EC were completely different (350,674). Similar other checks produced the same result, and the data presented by the EC contained obvious internal inconsistencies. How was it, for example, that columns 9 and 12 of the data presented by the EC to the Panel showed identical monthly dumping margins and indexed monthly dumping margins? Did the "indexation" carried out by the EC have no effect whatsoever on the dumping margin? If so, why were the total figures (12. 8 per cent and 19.4 per cent) different? There was therefore some fundamental error or lack of clarity in the figures provided by the EC, and hence they should be rejected.

(ii) Arguments relating to "negative dumping" or "zeroing"

240. Brazil argued that the margins would have been further reduced if the EC would have taken account of the fact that the inflationary environment resulted in large differences in dumping margins from one transaction to another. If the EC would have made due allowance for negative dumping, then the overall margins calculated would have been zero. Brazil contended that the dumping margins calculated by the EC resulted solely from the temporary exchange rate distortion which the EC failed to take into consideration.

241. Brazil argued that the phenomena of "exchange" dumping was well documented, and had long been the subject of consideration by the contracting parties. Exchange dumping occurs when the export price expressed in domestic currency falls below normal value as a result of distortions in the export currency/domestic currency exchange rate. The issue of exchange dumping was at the boundary between the international regulation of trade (under the GATT) and international monetary regulation (under IMF). In this respect, there was a crucial distinction between a sustained exchange rate disequilibrium and a temporary exchange rate distortion. The contracting parties had historically taken the view (though not unanimously) that sustained disequilibria fell within the sphere of international monetary regulation. Such factors would not generally be taken into consideration in the context of an anti-dumping proceeding. On the other hand, the contracting parties had consistently recognized the overriding principle that a fair comparison had to be made between normal value and export price in the context of an anti-dumping proceeding, and that allowance had to be made for temporary exchange rate distortions where they would otherwise lead to an unfair comparison between normal value and export price.

242. Brazil argued that it was not challenging the EC's zeroing methodology in general terms, but only in terms of the application of that methodology resulting in distorted estimates of dumping margins in the special situation of this case. Thus, even if the EC's zeroing methodology could be defended in general terms, it could not be defended in cases where very large variations in positive and negative dumping margins occurred, the only plausible explanation for which was distortions in the financial environment. In this case the EC apparently failed to have any regard whatsoever to its own reasons for applying the zeroing methodology, but simply applied the methodology in a mechanical way without regard to the fact that the special situation prevailing made its use wholly inappropriate.

243. Brazil argued that in inflationary circumstances, it was the established practice of the EC to calculate monthly normal value and to use monthly average exchange rates (as it did in this case) in order to reduce the distortions which may arise. For the same reason, in other cases, the EC had made the comparison between normal value and export price in the export currency. Hence, Brazil argued, the EC recognized the link between domestic inflation, exchange rates and the method of establishing normal value, i.e. the need to avoid artificial dumping findings generated by exchange rate distortions, and the EC practice implicitly acknowledged the principle that exchange dumping should not lead to unnecessary protective measures.

244. In the event that the EC's preliminary objections relating to "negative dumping" or "zeroing" were not accepted by the Panel, the EC requested the Panel to consider the following points. The EC argued that in its system, whenever export transactions of an exporter found to dump were shown to have been made at prices above normal value, those transactions were regarded as instances where no dumping occurred. The absolute volume of these transactions were included in the determination of the average weighted dumping margin for the exporter, with the dumping margin corresponding to these transactions being treated as zero. This method was fair to the exporters since it took account of their overall behaviour, i.e. the proportion of their export sales that were dumped and those that were not dumped.

245. The EC argued that the Agreement imposed no obligation upon the Parties to take into account the so-called "exchange rate dumping". The concept of "negative dumping" or "exchange dumping" appeared nowhere in the GATT or the Agreement, and there was no reference to sales made at or above normal value. Therefore, at the level of principle, the arguments of Brazil on this point should be rejected. The EC argued that, in reality, there was no such thing as "exchange dumping", or at least such a notion was not relevant under the provisions of both the Agreement and EC law applying and implementing the Agreement. There was either dumping, calculated in conformity with the obligations imposed by the Agreement, or there was no dumping. The definition of dumping in the Agreement required that the export price be less than the comparable price of the like product destined for domestic consumption (Article 2:1). Hence, to establish dumping one had to first and foremost consider export transactions made at prices below normal value. As the Agreement was intended to regulate the application of legitimate trade legislations, the purpose of anti-dumping would be nullified if an exporter were allowed to offset sales at dumped prices by sales made above normal value. The EC's computation involved a weighing of the transactions, and therefore also of the exchange rates applied to such transactions, even where there was no dumping. This was above and beyond the obligations and requirements imposed by the Agreement, and therefore could not be a violation of the Agreement.

246. The EC argued that even if one accepted Brazil's contention regarding the so-called "exchange dumping" (quod non), and thus recalculate the dumping using one of the methodologies suggested by Brazil, a substantial part of the dumping margin found by the EC would remain. However, the method suggested by Brazil was inappropriate, because it would not account for important factors which impact on trade in the product under investigation. Trying to grant allowances for all such factors would mean embarking in a highly speculative exercise, with an unacceptable loss of certainty and predictability of anti-dumping proceedings.

247. Brazil did not agree with the EC's view that there was no such thing as exchange dumping, and argued that exchange rates had a profound effect on dumping calculations. Brazil argued that the effect of exchange rate changes on export prices occurs on a lagged basis, and that many perceived dumping cases may, in fact, be caused by the effects of currency fluctuations. Brazil further argued that an investigating authority did not enjoy absolute discretion in the application of exchange rates in the context of an anti-dumping proceeding. While the Agreement and EC law may not explicitly mention exchange dumping, they did require that exchange rates be applied in a manner which was consistent with the provisions of the Agreement, notably the requirement that there be a proper and fair comparison between normal value and export price.

248. Brazil argued that while a number of factors influenced currency fluctuations, domestic inflation was by far the most important factor. Brazil considered that, ceteris paribus, there was a close correlation between the rate of domestic inflation and the rate of depreciation of a freely floating currency against other currencies (taking into consideration the rates of inflation associated with those other currencies).

249. Brazil argued that the EC was not correct in stating that if margins were recalculated using one of the alternative methodologies proposed by Brazil, at least a part of the dumping margin would still remain. Brazil claimed that the EC had provided no data in support of such a contention. On the other hand, Brazil had provided estimates to support its contention that alternative methodologies would result in de minimis dumping margins.

250. Brazil argued that though it had not challenged the EC's zeroing methodology in general terms, this decision should not be taken as any indication on the part of Brazil that it agreed in any way with the EC's general zeroing methodology. Brazil was concerned with the distortion arising as a result of the application of the zeroing methodology in a volatile financial environment, such as that prevailing during the investigation period in this case. Brazil noted that the particular concern of the EC for its general use of the zeroing methodology was to counter the possibility of "manoeuvres" on the part of exporters. An example of this was found in the judgement of the Court of Justice of the European Communities in a case involving anti-dumping measures on mini-bearings from Japan: "The transaction-by-transaction method is the only method capable of dealing with certain manoeuvres in which dumping is disguised by charging different prices, some above the normal value, and some below it. The application of the weighted average method in such a situation would not meet the purpose of the anti-dumping proceeding, since that method would in essence mask sales at dumping prices by those at what are known as "negative" dumping prices, and would thus in no way eliminate the injury suffered by the Community industry concerned."

251. Brazil argued that a wide range of dumping margins were found for the Brazilian companies in this case: for Renda these varied from 59.06 per cent (10 February) to -29.96 per cent (20 December), a range of 89.02 per cent; for Kanebo the variation was between 59.93 per cent (22 May) to - 15.90 per cent (28 July) , a range of 75.83 per cent; and for Nisshinbo, the variation was between 70.30 per cent (8 May) and - 19.80 per cent (24 June), a range of 90.10 per cent. Brazil asked whether these large variations arose due to "manoeuvring" on the part of the Brazilian exporters? Such an allegation had never been formally made, and measures adopted by the EC contained no reasoning on this point. Also, was the EC suggesting that variations of this order of magnitude could be explained on the basis of alleged attempts on the part of the Brazilian exporters to successfully penetrate the EC market with dumped products? Brazil argued that such reasoning would not be credible. In reality, and quite obviously, the variations found by the EC in this case were not caused by "manoeuvring" on the part of the Brazilian exporters. They were caused by severe distortions in the financial environment.

252. Brazil argued that there were three main reasons for large variations to occur. The first was due to the result of the frozen exchange rate itself. It was notable that the highest margins were all found in the first half of the year, whilst the lowest margins were all found towards the second half of the year. The second reason was that the uneven distribution of domestic transactions within one month could have a very significant effect on the normal value. The third factor which tended to generate large variations was the use of very short reference periods in a financially unstable environment. Although the use of monthly average normal values was intended to eliminate certain distortions that may otherwise arise it also had another effect. Since one month was a relatively short period of time, it was very difficult to make a meaningful match between monthly domestic prices, costs of production and export prices. It was common for a production process (from purchase of raw materials to sale of finished goods) to span more than one month. Furthermore, domestic and export pricing decisions were often made in a non-synchronized manner, often for periods in excess of one month, and often in relation to sales that would not be shipped, invoiced or paid for until a later month. If these figures got out of step in a high inflationary environment the result was large variations in the positive and negative margins found and a distorted calculation. Since the EC gave no credit for negative dumping, large variations of this nature had a particularly prejudicial impact on the dumping calculation.

(e) Brazil's allegation that use of average monthly exchange rates was not sufficient

253. Brazil recalled that in recital 14 of the Definitive Determination, the EC stated that "Given the high inflationary situation prevailing in [Brazil] ... the Commission established normal value on a monthly basis in order to permit a proper comparison for export price." Brazil argued that in the special circumstances of this case, the practice adopted by the EC (i.e. the use of monthly exchange rates in arriving at normal value) was insufficient to deal with the distortions which arose. The EC should have therefore taken further measures to ensure a fair comparison between normal value and export price. However, the EC had refused, as a matter of principle, to do so. Thus, the EC's position in this case was contradictory: monthly average calculations were used because, in principle, the EC considered that exchange dumping should be avoided; however, the use of adjusted exchange rates was rejected by the EC because, in principle, the existence of exchange dumping was considered irrelevant. Thus, the position adopted by the EC was motivated by expediency and not by principle. Brazil argued that the data presented by it to the Panel showed that such methodology was insufficient to deal with the temporary exchange rate distortions that were present. In the situation in this case, the EC should have granted further allowances.

254. Brazil argued that the EC's argument on Article 2:6 appeared to consist of two steps. First, that the comparison be made between sales made "at as nearly as possible the same time" exhausted (together with the requirement that the comparison be made at the same level of trade) the requirement that a fair comparison be made. As argued above in section (a), Brazil did not agree with this interpretation of Article 2:6. Second, the EC had argued that by calculating normal value on a monthly average basis, and by using a monthly exchange rate, the requirement that the comparison be made between sales made "at as nearly as possible at the same time" was met. However, high inflation in conjunction with a freely depreciating exchange rate normally meant that domestic prices and export prices (expressed in domestic currency) both increase over time and run approximately parallel to each other. For example, domestic prices could increase steadily from 10 to 50 units of domestic currency over the year (as a result of domestic inflation), export prices in the export currency could remain fixed, and export prices expressed in domestic currency could increase from 12 to 60 units of domestic currency over the year (as a result of the depreciation of the domestic currency against the export currency). Prima facie, there was no dumping in this situation but a dumping margin would be found if there was an exchange rate freeze.

255. Brazil further argued that the EC's practice was to compare each individual export transaction with an average normal value. If an average normal value was compared with each individual export transaction using spot exchange rates in the example above, it was clear that there would be dumping at the beginning of the year, and no dumping at the end of the year. The result was a finding of dumping, where none in fact existed, since the EC's methodology gave no credit for negatively dumped transactions and treated these margins as equal to zero (this was the so-called methodology of "zeroing"). One way to try and reduce the distortions arising was to also use an annual average exchange rate, thus also levelling the export price curve, expressed in domestic currency. However, this approach assumed an even distribution of domestic (and to a lesser extent export) transactions during the year, and did not take into account other factors which may influence pricing changes during the investigation period. The use of monthly average normal values in conjunction with monthly exchange rates was intended to address some of these problems. However, because the reference period was short in the case of monthly averages, this method could also have the effect of producing more distortions where the two sides of the calculation were not correctly synchronised.

256. Brazil argued that it was clear from the above analysis that the methodology which the EC claimed it adopted to deal with the fact of the exchange rate freeze was in fact adopted for entirely different reasons. It was intended to deal with the distortions arising in a high inflationary environment where the domestic currency was also depreciating against the export currency. This was not the case in this proceeding. In this proceeding the exchange rate was actually frozen. The EC's use of a monthly reference period did not therefore in any way address the special situation in which Brazil found itself. Thus the EC had no regard to the fact of the exchange rate freeze and took no steps to deal with the distortions which resulted from the freeze.

257. The EC noted that Brazil was admitting that the use of monthly average normal values in conjunction with monthly exchange rates was intended to address some of the problems that arise in a determination of dumping margin. In this case, though the EC was not under a legal obligation (and contrary to its normal practice), it had used end of the month exchange rates or an exchange rate used on actual returns by the exporters. Such exchange rates were the closest rates in time to the export transactions the EC could have used in the circumstances of this case.

258. The EC argued that even Brazil had not contested that the use of monthly, instead of yearly, exchange rates permitted a fairer comparison. In fact, the EC had several times in the past applied the official monthly average exchange rate in its comparison of export price and normal value vis-à-vis Brazil, but Brazil had not complained about the fairness or conformity of this method with the provisions of the Agreement. In fact, Brazil's argument did not appear to be that the exchange rate used by the EC was incorrect, but only that it was "insufficient". Indeed, all Brazil could probably ask for from the EC would be for it to use not monthly exchange rates, but possibly daily or weekly rates. However, Brazil has instead claimed that the EC should have used "adjusted or lagged" unofficial exchange rates. The EC further argued that Brazil had not specified how far away in time such rates should be or who should choose them. In fact there could not be an agreement between parties on such issues, because they involved a high degree of arbitrariness and unpredictability, which also damaged the certainty and security required for the correct application of the Agreement. Under those circumstances, the EC was absolutely convinced that it had made a "fair" comparison of the normal value with the export price.

259. The EC argued that a claim for using "lagged" unofficial monthly exchange rates should be rejected for two reasons. First, it had no foundation in Article 2:6 or in the other provisions of the Agreement or of the General Agreement. Second, it was not the Panel's task to judge whether the EC could have used exchange rates which were "better" or "more convenient" in Brazil's view, but whether, in using monthly exchange rates, the EC did not allow a comparison at as nearly as possible the same time, thus frustrating a fair comparison. Article 2 of the Agreement required a fair, not a perfect, comparison between export price and normal value, and fair in Article 2:6 in this respect meant "as nearly as possible the same time as for the sale transactions". In view of the fact that the EC used monthly normal values to take account of the Brazilian inflation rate, end-of-month exchange rates were considered to be sufficiently close to the requirement to make a comparison at as nearly as possible the same time to make sure that any unfavourable effect on the exporter from the conversion of the export price to the domestic price would be excluded.

260. The EC recalled its arguments (presented above) that it could be reasonably argued that the devaluations were abnormal. Thus, in consideration of all the facts, the EC argued that the monthly official exchange rates gave a fair and proper representative picture of the period as a whole.

261. The EC argued that when exporters concluded the export contract, there was a great deal of uncertainty about their behaviour. Therefore, in accordance with its usual practice, the EC sought the actual returns in domestic currency for export sales. In this case, it used the actual returns in cruzados obtained by Kanebo when converting the proceeds of its exports in United States dollars because the information was supplied by that company. Such information was not provided by the other two companies that were verified (as a result of informed choice on the part of those companies), and the EC used official exchange rates for them. In fact, for these two, the EC resorted to the end-of-the-month official rates, which already contained an element of lagging when compared to the date of the actual exchange transactions. Thus, having established monthly average normal values in order to permit a proper comparison with export prices, the EC resorted to end-of-the-month exchange rates to compare those normal values with export prices and to fulfil its obligation under Article 2:6 "to effect a fair comparison" through comparing prices "in respect of sales made at as nearly as possible the same time".

262. In response to a question by the Panel, the EC stated that it had requested the companies subject to investigation to supply the date of the export transactions in question, and used the dates actually supplied by the companies, which appeared in Annex I to the disclosure letter. For Kanebo, the date appeared to be that of the exchange rate contract, for the other two, it was the invoice date. These dates were considered acceptable by the EC, for Kanebo because actual returns in cruzado were used, and for the other two because normal EC practice was to use the date of the invoice or the date of the bill of lading.

263. Brazil argued that all three Brazilian companies had cooperated in the investigation as shown by recital 19 of the Provisional Determination, which stated that: "The Commission sought and verified all the information it deemed necessary for the purposes of making a preliminary determination of dumping and consequent injury, for the parties which agreed to cooperate. To this end it carried out inspections at the premises of ... [Nisshinbo, Renda and Kanebo]" (emphasis added by Brazil). The exporters were therefore considered as fully cooperating, and all necessary data was deemed sought and verified. Neither the Provisional nor the Definitive Regulation contained any reasoning whatsoever which would support the contrary.

264. Brazil argued that the facts of the case showed that devaluations could not be considered abnormal. Rather, the fixed exchange rates were implemented at a time when the inflation rate was far higher than the rate of devaluation, and therefore, Brazil's arguments in relation to the effect of the fixed exchange rates remained unchanged. The EC's use of monthly exchange rates did not in any way address the special situation in which Brazil found itself. The EC had no regard to the fact of the exchange rate freeze and took no steps to deal with the distortions which resulted from the freeze.

265. Regarding the EC's argument that the use of end-of-month exchange rates already involved an "element of lagging", Brazil argued that the use of end-of-month exchange rates would have had no material impact on the arguments presented by Brazil. The end-of-month exchange rates for the critical period of the exchange rate freeze did not materially differ from the monthly average exchange rates, precisely because the exchange rate during this period was frozen. The comparison between the end-of-month and average exchange rates was as follows : January (1.0 and 0.9); February (1.0 and 1.0); March (1.0 and 1.0); and April (1.032 and 1.017). Thus, from the factual point of view, such an "element of lagging" during the critical period of the freeze would not have materially reduced the distortions arising. Furthermore, payments for the great majority of transactions were likely to have been made in months following the issue of invoices. The alleged "lagging" effect identified by the EC would not therefore have approached to any significant degree the adjustment that would need to have been made for the exchange rate distortions to be eliminated.

266. Brazil argued that from the legal point of view, the EC appeared to suggest that the use of these end-of-month rates by the EC was intentional. However, this argument seriously contradicted the whole basis of the EC's defence that it could not consider the alternatives put forward by Brazil because it was bound by the "international law principle of respect for monetary sovereignty" to use the "official" exchange rate, without any flexibility. If the EC's claim was that it did try to introduce an "element of lagging" into the calculation, the whole basis of the EC's reasoning in the Provisional and (especially) the Definitive Regulations was contradicted. Furthermore, if the purpose of the EC's alleged use of an "element of lagging" was to address the distortions which were present, it clearly failed to achieve this objective. If this option was available (as it apparently was), the EC should have properly considered how the option could be applied to remove the distortion and the Determinations should have contained some reasoning on this point. Also, if the EC considered the "element of lagging" allegedly used an acceptable exercise of discretion, it should also consider the lagging argued for by Brazil as acceptable.

267. Brazil argued that while the EC had stated before the Panel that for one company (Kanebo) it used "actual returns" and for two companies (Nisshinbo and Rend) it used end-of-month exchange rates (and not monthly average exchange rates), in the Provisional Determination (at recital 17), the EC had stated that : "As regards Brazil and Turkey, for the reasons explained in recital 14, the Commission considered it more appropriate to use an exchange rate based on a monthly average" (emphasis added by Brazil). This position was confirmed in the Definitive Determination (recitals 25 to 29) and in the EC's disclosure letter (at point 2 (b), page 2). This point had also been earlier re-iterated by the EC before the Panel. The consequences of this change of position by the EC was effectively an admission that the Provisional and Definitive Determinations contained a serious misstatement of fact on the key question of exchange rates, and that when the measures were imposed, the EC did not properly consider or have regard to the exchange rate problem which arose, even though this was a critical element in the determination. Brazil considered that this confirmed the arguments made by Brazil throughout this proceeding.

268. Brazil argued that since the Provisional and Definitive Determinations did not state accurately the factual basis on which the EC reached its determinations, all of the arguments and reasoning contained in those Determinations which were based on this incorrect statement of fact were vitiated. Given the absolutely central importance of the exchange rate in the determination, Brazil submitted that, on this basis alone, the EC's determinations could not be considered to have been made in accordance with the EC's obligations under the Agreement, particularly Article 8:5 of the Agreement, which provided that : "Public notice shall be given of any preliminary or final finding whether affirmative or negative and of the revocation of a finding. In the case of affirmative finding each such notice shall set forth the findings and conclusions reached on all issues of fact and law considered material by the investigating authorities, and the reasons and basis therefor" (emphasis added by Brazil).

269. Regarding Brazil's argument that the EC had changed position on the use of exchange rates from monthly average to end-of-month rates, the EC argued that Brazil had admitted that the dumping margin that would have resulted for Nisshinbo and Kanebo were not affected in any way by the two methodologies. Thus, the EC considered that this argument of Brazil was irrelevant for the purposes of this proceeding before the Panel.

270. The EC explained that it did not argue that the companies for which it did not use actual exchange rates (i.e. Nisshinbo and Renda), did not cooperate in the investigation. The EC argued that a distinction had to be made between non-cooperation and the non-submission of information which it would have been in the interest of the company to supply but for which non-submission was not regarded as constituting non-cooperation. There was non-cooperation if necessary information required by the investigating authority was not supplied. On the contrary, if a respondent did not supply any information which was not considered as strictly necessary by the investigating authority (for instance, because other sources of information were publicly available), even though it would be in the interest of the company to submit it, this company could still be considered as cooperating. The respondent companies were aware of the fact that, lacking that information, the EC would use information publicly available on the issue, in this case official exchange rates. The use of official exchange rates in this case was consequently not related to non-cooperation of the companies concerned, and did not contain any element of sanction for non-cooperation. It was based on the obligation to establish determinations on objective and verifiable facts. In the present case, when information on actual returns was not available, the only alternative information objectively available was official exchange rates.

271. Regarding the "element of lagging" which end-of-month rates contained, the EC argued that it was obvious that such rates were more likely to contain an "element of lagging" than beginning of the month or monthly average rates. Moreover, the EC remained convinced that the use of official exchange rates on a monthly basis was fully appropriate in circumstances such as those present in this case, and fully consistent with the Agreement. The argument about the end-of-the month rates having some lagged effect was also put forward to show that Brazil was incorrect in arguing that the EC used an excessively rigid method which led to absurd consequences.

(f) Arguments relating to the inability to consider method of establishment of official Brazilian exchange rate and to the principle of "monetary neutrality"

272. Brazil recalled that in recital 28 of the Definitive Determination, the EC had stated that "the establishment, by the competent authorities of the exchange rate of a third country's currency is a decision which cannot be subject to appreciation by the Community institutions in the framework of an anti-dumping proceeding." Brazil argued that by choosing an exchange rate different from the official rate, the EC would not have encroached or commented upon Brazil's economic policy. In any event, the EC had been expressly requested by the Brazilian authorities to consider an adjustment. Brazil also argued that the EC frequently used exchange rates in anti-dumping proceedings provided to it by its Department responsible for monetary and economic policy (DG II) even when those rates were not identical to the official rates published in the exporting country. Moreover, the EC was not being asked to consider an alternative source or method of calculation for the official exchange rate, but to apply the official FIBGE exchange rate in a manner which ensured a fair comparison between normal value and export prices.

273. Brazil also recalled that in recital 28 of the Definitive Determination, the EC had stated that "to adjust this exchange rate for the purposes of dumping calculations would be inappropriate and contrary to the principle of neutrality as regards the monetary aspects of an anti-dumping case". Brazil argued that no reference was made to such a principle in the Agreement and the legal basis of the EC's statement in this regard was unclear. If the principle was alleged to be the irrelevance of "exchange dumping" in an anti-dumping proceeding, the EC's statement was incorrect. The overriding requirement for a fair comparison necessarily required temporary exchange rate distortions to be taken into account where appropriate. Further, to the extent such a principle may exist, it could be complied with by basing normal value on sales to third countries, by making the necessary adjustments to normal value based on domestic sales or cost of production, by ignoring "exchange dumping" caused by temporary and unexpected exchange rate fluctuations, or by using an adjusted exchange rate. Thus, by not making the required adjustments or choosing the appropriate basis for normal value, the EC had violated Articles 2:4 and 2:6 of the Agreement.

274. The EC argued that neither GATT nor the Agreement addressed the monetary aspects of dumping. While the issue may not have been raised at the time of GATT negotiations in the context of the Bretton Woods system, the absence of inclusion of monetary aspects into the Agreement demonstrated the clear intent of the negotiators to leave monetary aspects of dumping out of the scope of the Agreement. Moreover, the Agreement required an objective examination of facts based on sufficient and/or positive evidence. On that basis, two options were available to the Parties to the Agreement: either taking into account monetary aspects in their practice on a unilateral basis with the substantial risk of departing from the obligation of objective examination, or considering that monetary aspects should be treated as externalities, i.e. as given for the investigating authorities (which would also be in conformity with the need to ascertain factual basis for the imposition of anti-dumping duties). The latter option was, in addition, in conformity with the international principle of sovereignty of States on monetary issues. The EC was not entitled to make assumptions on the merits of the monetary policy of third countries and should consequently refrain from doing anything of that kind within the framework of an anti-dumping procedure by "indexing" exchange rates. Therefore, the EC considered that a neutrality vis-a-vis monetary aspects was best respected by using official exchange rates, which had in addition the advantage of being established in total disregard of the existence of anti-dumping procedures. If the EC were to adjust exchange rates in one case, it would have to do it in every case. It was not certain that other countries would accept or suggest, like Brazil, the use of a different exchange rate than the one they officially set. Moreover, this would also imply adjustments in cases of undervalued currencies, if one were to be fair to the complainants. In any event, the choice of the adjusted rate would always be subject to contention by respondent or complainants.

275. The EC argued that its standard practice was to use official exchange rates set by the competent authorities of the signatories. This practice was consistent with the provisions of the Agreement, it had been applied in the past by the EC against several third countries, including Brazil, without them ever complaining about it, and it was also in line with the general principle of public international law of non-interference in the sovereign authority of States to manage their internal affairs in all areas of their competence. The absence of a specific provision in the Agreement on the use of exchange rates showed that the States did not wish to surrender their sovereign rights, including the right to decide the exchange rate of their currency. The EC argued that its practice fully respected all the above-mentioned principles, and therefore it was not clear in what respect it had violated the provisions of the Agreement by not using in its comparison "unofficial" or "lagged" exchange rates.

276. Brazil noted that the EC had cited no authority to support its arguments relating to the principle of monetary neutrality and to the non-interference in the sovereign authority of States to manage their affairs in all areas of their competence. Brazil argued that to the extent that the principles were relevant, they could not be used to justify administrative practices which led to no proper and fair comparison between normal value and export price.

277. Brazil argued that the EC's position hinged on an alleged "international principle of sovereignty of States on monetary issues", but the EC had provided no sources to support it and had offered no explanation for the extent of this alleged principle. Brazil doubted whether it existed in a form that had any direct bearing on the issues arising in this case.

278. Brazil argued that the EC had not offered any explanation of how, in practice, use of an indexed or lagged rate would have interfered with Brazil's sovereignty over its internal monetary affairs. Brazil considered that no such interference would have been present. The EC was not being requested in any way to make "assumptions on the merits of the monetary policy" of Brazil. The EC was simply being requested to make a technical adjustment to the calculation in order to ensure a proper and fair comparison.

279. Brazil further argued that the EC had not offered any explanation of what the consequences of this principle were alleged to be. Did it mean that only the official exchange rate could be used? If so, it did not appear to rule out lagging of the official rate. On the other hand, if the EC considered that this principle effectively prevented lagging, how did the EC explain the fact that (contrary to what was stated in the Regulations) the EC itself apparently introduced an "element of lagging" in this case.

280. Brazil disagreed with the EC's view that it was the clear intent of the negotiators to leave monetary aspects of dumping out of the scope of the Agreement. Rather, the intent of the negotiators was to exclude depreciation exchange dumping from the scope of the Agreement. It was absolutely wrong to infer from this specific fact the general statement that "monetary aspects of dumping" were outside the scope of the Agreement. In any event, even if such a principle could be relevant to the interpretation of the Agreement, Brazil did not consider that it could be applied in a way to frustrate the fundamental objective of the Agreement, i.e. the requirement that there be a proper and fair comparison between normal value and export price.

281. Brazil argued that even if such a principle could be said to exist, to have been potentially infringed and to be of sufficient weight to prevail against other provisions of the Agreement, it would have ceased to apply once the Brazilian authorities expressly requested the EC to make the necessary adjustment. Such a request had been made by letter on 23 October 1991, five months before the imposition of definitive duties. In these circumstances, Brazil considered that the reasoning put forward by the EC could not be sustained and should be rejected. Brazil argued that the real reason behind the position adopted by the EC was the entirely unfounded view that if the EC were to adjust exchange rates in one case, it would have to do it in every case (this issue is discussed in more detail in section (h)).

282. The EC argued that in the specific case of the choice of exchange rates for the purpose of the comparison, it could not be argued that if the Brazilian Government requested the EC to disregard the official exchange rate set by the Brazilian authorities then the EC was free to do so. Obviously, this would have eliminated any question of fair treatment of the Brazilian exporters in this particular case, but would have posed immediately the question of fair treatment of the injured EC industry. Furthermore, it was in the long-term interest of exporters in anti-dumping cases to have a minimum of predictability in relation to a factual question, such as exchange rates, which experience showed was characterised by a very high degree of volatility. Consistent resort to official exchange rate was the only method which could assure exporters subject to an anti-dumping investigation of at least a minimum degree of certainty as to whether they were in fact dumping or not.

(g) Arguments relating to the EC's contention that the adjustments requested were beyond the scope of the Agreement

283. Brazil recalled that in its statement during the conciliation meeting of the Committee, the EC had stated that the adjustments requested would be "unduly stretching" the provisions of the Agreement, introducing new concepts for price determination" and would be "far beyond the scope" of the Agreement. Brazil argued that where such adjustments were necessary to ensure a fair comparison between normal value and export prices, they were clearly within the scope of (and required by) the Agreement. Brazil recalled in this context its statements regarding the practice of the contracting parties and Article 2.4.1 of the 1994 Agreement (mentioned above in section (a)), and that Article 2:6 of the Agreement expressly mentioned "differences in taxation" as one of the factors affecting price comparability for which adjustments should be made. Brazil argued that taxes were typical economic measures imposed by a Government and were therefore comparable, in this respect, to the fixing of exchange rates. It was not unreasonable to see frozen exchange rates in this case as a kind of government tax on exports, because exporters were obliged to exchange their foreign exchange at a reduced value (40 per cent lower) due to the excess of domestic inflation over the depreciation. Taxes and exchange rate freezes may both be considered as government measure entirely outside the control of exporters. Therefore, it seemed contradictory to make an allowance for taxes but to dismiss all possibility of any allowance in the case of a frozen exchange rate. By not making the adjustments to the exchange rates, the EC had violated Article 2:6 of the Agreement.

284. The EC argued that taken literally, an "exchange rate" was not a difference, but a factual situation. Therefore the fixed "exchange rate" applied by Brazil in this case was not a difference nor did it lead to any different treatment since cotton yarn producers could sell at any price they wished abroad. They were also free to convert the proceeds of their sales abroad into cruzados at any moment they wished to, or they were free to keep the proceeds in their account abroad or to hedge against risks of exchange rates variation. Hence, there was no difference in the sense of Article 2:6, for instance, like the obvious "difference" which resulted from a country applying different tax rates, which was an objective and verifiable difference where the individual exporter had no freedom of choice as to whether to comply with it or not.

285. Brazil disagreed with the EC's argument that an exchange rate was not a "difference" within the meaning of Article 2:6, but a "factual situation", and argued that a sales tax was also a factual situation but not a "difference" in that sense. Nevertheless, Article 2:6 required an allowance to be made where domestic sales were subject to a sales tax because of the difference between the way in which the normal value was expressed including sales tax and the way in which the export price was expressed not including sales tax. This difference affected price comparability, and therefore, in making the comparison, an allowance had to be granted. Brazil argued that in the case before the Panel also there was clearly a difference between the way in which normal value was expressed and the way in which export price was expressed. Normal value was expressed in cruzado and export price was expressed in United States dollars. This was an objective difference which affected price comparability. Therefore, in making the comparison (i.e. in applying an exchange rate), investigating authorities were required to make due allowance for this difference. Use of the frozen exchange rate by the EC in this case did not take into account the objective difference that were present. The EC should therefore at least have considered making due allowance to take into account the distortions which arose. Brazil had suggested a number of ways in which this could have been done.

286. The EC argued that its interpretation of the obligations under Article 2:6 was that exchange rates did not come under the second sentence of that Article. They were relevant in the context of the obligations imposed by the first sentence, and for the reasons already mentioned above, these obligations had been met by the EC in this case.

(h) Arguments relating to the likely effect of the methodology on future proceedings, i.e. the "floodgates" argument

287. Brazil argued that the position adopted by the EC appeared to be a concern that if Brazilian arguments were taken into consideration in this case there would be an opening of the floodgates, i.e. there would be endless wrangling about the correct exchange rates in future proceedings. However, it was not Brazil's position that any difference affecting price comparability should be taken into consideration, but that differences affecting price comparability arising from temporary exchange rate fluctuations should be taken into consideration. Also, Brazil's argument was not that the rate of inflation must always be fully reflected in the rate of depreciation for the official exchange rate to be used. Brazil's argument was that when a clear temporary exchange distortion occurred, this should be taken into consideration. For this, it was necessary to calculate an alternative or adjusted exchange rate for the period concerned, and there were a number of reasonable ways, including indexing to inflation, for doing so. The practice in the main dumping jurisdictions showed that there were a number of options available for authorities to deal with temporary exchange rate fluctuations, which could be tightly defined and circumscribed. Therefore, any administrative concerns of the EC in this case were ill founded. In any event, the EC should not rely on concerns about administrative workability as a basis for flouting a key principle of the Agreement, namely, the requirement for a fair comparison.

288. The EC disagreed with Brazil's argument that the EC practice on this issue was motivated by reasons of expediency and administrative convenience. The EC argued that there was nothing in the EC practice, in this or any other case, to suggest that the EC was acting on the basis claimed by Brazil. The mere fact that the investigation period in this case exceeded the normal period of one year (Recital 11 of Provisional Determination) may be an indication of the seriousness with which the EC had handled this case.

289. The EC argued that it was not sacrificing fairness in this case in order to increase certainty and predictability in other cases. The EC was concerned with certainty and predictability in all cases, including this one. A purpose of the Agreement was protection of the domestic industry from unfair trade, and under domestic law the investigating authorities must pay attention to and respect the rights and interests of all the parties involved in the administrative and judicial proceedings.

290. The EC then addressed Brazil's argument that the EC was free to use an exchange rate different from the Brazilian official exchange rate, and said that Brazil had failed to explain what this other exchange rate would consist of, by whom it would be fixed, and how such "freedom" to use a different exchange rate amounted to an infringement of the Agreement. Brazil's approach in this regard was short-sighted and disregarded the long-term interest of instilling into the system of the Agreement certainty, fairness, predictability, and conformity with the generally accepted principles of international law. The need to "second guess" the exchange rate to be used, and the arbitrary manner in which such a choice would be made, would not conform to the desired principles of international law.

291. Brazil shared the concerns referred to by the EC, but argued that these concerns were stated in a manner which reflected the EC's position as a contracting party that predominantly had recourse to anti-dumping measures, as opposed to a contracting party that predominantly found its exporters subject to anti-dumping measures. Brazil considered that the EC's statements on this point did not represent a well balanced view, in that the EC considerably overstated the extent to which Brazil's claim might lead to the results referred to. An exchange rate freeze which undervalued a currency could not be sustained for any significant period. Moreover, a situation with undervaluation, as a matter of law, would be irrelevant in the context of an anti-dumping proceeding. Thus, as a matter of fact and law, the only type of situation with which one could be concerned was the situation present in this case: a temporary exchange rate freeze which overvalued the domestic currency.

292. Brazil agreed that it was desirable to restrict the elements of subjectivity under the Agreement in order to prevent its abusive or creative application by contracting parties. However, the element of "subjectivity" involved in the method of assessment would not be more than that present in numerous other provisions of the Agreement, i.e. though the investigating authorities' determination on this point would involve an element of discretion, this was not different from the case for numerous other provisions of the Agreement. Moreover, if the investigating authorities exercise their discretion in accordance with the provisions of the Agreement, the discretion would have been properly exercised by them. That would represent the established manner in which the Agreement was currently applied, and would add nothing new. Furthermore, and most importantly, it could only ever be exercised in favour of the exporters concerned, who would almost certainly always be from developing countries experiencing special problems with inflation. As far as verification of the relevant parameters was concerned, since these would be determined by the investigating authorities themselves, no difficulty in verifying them would arise. At no time would the investigating authorities be required to second guess future exchange rate movements. The investigating authorities would not at any time be required to take into consideration future "highly volatile and unpredictable differences".

293. Brazil argued that it was taking a long term view of the need for a proper, fair and effective functioning of the Agreement. This view was apparently shared by the contracting parties when they clarified the correct interpretation of the 1979 Agreement in Article 2.4.1 of the Anti-dumping Agreement concluded under the Uruguay Round. Brazil was concerned in this case with a very real and well documented problem: apparent exchange dumping. This problem needed to be addressed and clarified in order to ensure a healthy and functional operation of the Agreement. This was a view that was apparently shared by the contracting parties (including the EC), who had offered further clarification of this issue in the Uruguay Round Agreement

294. The EC argued that Uruguay Round Agreement was not relevant to the present case. The EC argued that it had met all the requirements of the Agreement. The procedural requirements of the Agreement, including those of Art. 2:4 and 2:6 were not dictated by the interests of the investigating authorities but of all private parties involved in the proceedings.

(i) Arguments relating to treatment of Brazil being contrary to the treatment of other countries in the case

295. Brazil argued that the EC contradicted itself within the proceedings of this case which, inter alia, also included Egypt. In the Definitive Determination, the EC stated that: "With regard to the rate of exchange used for the calculation of the cost of imported cotton, the Commission found that the claim of the Egyptian producers/exporters was justified. Consequently the cost was re-calculated and this modification led to a decrease of normal value" (recital 20). Therefore, in the case of Egypt, the EC was prepared to accept a modified exchange rate. The lack of acceptance of exchange rate adjustment for Brazil therefore was discriminatory.

296. The EC argued that the "special" exchange rate in force for raw cotton transactions mentioned in Recitals 17 and 20 of the Definitive Determination was the "official" exchange rate set by the Egyptian authorities for import of raw cotton into Egypt. Brazil could not, therefore, claim that it had been discriminated against, because in both instances the EC used "official" exchange rates fixed by the responsible authorities of the two countries.

(j) Arguments relating to the importance of the economic situation and a change in the criteria by the EC i.e. "moving the goalpost"

297. Brazil recalled that in recital 17 of the Provisional Determination, the EC had stated that "the Brazilian exporters were not able, during the preliminary investigation, to produce any arguments that the official exchange rates did not correspond to a realistic economic situation. Therefore, the request for an adjustment of the export price has been rejected at this stage of the proceeding". Thus, Brazil argued that the EC strongly implied that if evidence had been available that the official exchange rate did not correspond to a realistic economic situation, the adjustment would have been granted. However, by the time of the Definitive Determination the EC had dropped this argument even though the exporters had irrefutably established that the use of unadjusted official exchange rate in this proceeding would have been economically unrealistic. Without giving the exporters the opportunity to present further arguments, the EC switched to the position that it could not consider the adjustment at all, this was effectively "moving the goalposts". Brazil argued that if economic reality was a relevant factor in determining whether or not to use an official exchange rate, the EC should have stated whether or not it considered the use of the official exchange rate in this proceeding to be economically realistic. Brazil said that if economic reality was not a relevant factor, then the EC should confirm that the Provisional Determination was incorrect on this point.

298. Brazil also said that if economic reality was a relevant factor in determining whether or not to use an official exchange rate, the EC should state whether or not it considered the use of the exchange rate in this proceeding to be economically realistic, especially in the light of the data presented by Brazil.

299. The EC argued that the exporters had ample opportunity to present their arguments at both stages leading to the imposition of provisional and definitive duties. Recital 17 of the Provisional Determination and recitals 26-28 of the Definitive Determination demonstrated that in both instances the Brazilian exporters were making the same type of allegations to the EC authorities. Moreover, the replies of the latter were the same in both instances. Therefore, it was not reasonable to claim that the EC had changed standards by "moving the goalposts". The fact was that Brazil was contesting the validity of the Definitive Determination with the Agreement, and it had failed to provide any convincing arguments or data, both at the time of the investigation and now before the Panel, to demonstrate that the EC's recourse to "official" exchange rates in its determination is unlawful. The EC argued that it had met all the requirements of the Agreement in this case.

(k) Brazil's allegation that the EC failed to exercise its discretion

300. Brazil argued that there was an important distinction between failing to effectively exercise a discretionary power and exercising it incorrectly. The Agreement established an obligation for anti-dumping authorises to ensure a fair comparison between normal value and export price, and this included the obligation to take into account distortions arising from temporary exchange rate fluctuations, especially in the context of proceedings concerning exports from developing countries. Anti-dumping authorities exercised discretion in how they fulfil these duties. Brazil argued that in this proceeding, the EC effectively failed to exercise its discretion by taking the view that the existence of "exchange dumping" was simply irrelevant to its determinations, even in relation to a developing country. According to the EC, the exchange rates at issue could not be the "subject of appreciation". In taking this position, the EC had infringed the Agreement. Such infringement in itself was sufficient for a recommendation of withdrawal of the measures, without it being necessary to reach any conclusions on the facts of the case. However, should it be considered that the EC did exercise its discretion, Brazil additionally believed that, given the facts, the course of action taken by the EC was so clearly erroneous, that it also infringed the Agreement. The position taken by the EC in this case effectively nullified the two most important principles applicable to this proceeding: the requirement that a fair comparison be made between normal value and export price and the requirement that special account be taken of the position of developing countries.

301. The EC argued that Brazil's claim did not add anything materially new to the previous arguments. They were based on the false assumption that the requirement "to ensure a fair comparison" included the obligation to take into account distortions arising from temporary exchange rate fluctuations. This argument had been dealt with above. The EC also argued that it had met the requirements of Article 13 of the Agreement, i.e. the provision which addressed the situation of developing countries.

V. 3. Determination of Injury

Introduction

302. Brazil claimed that the EC had violated Articles 3:1, 3:2, 3:3, 3:4 and 8:2 of the Agreement in so far as: (i) the injury findings were not based on positive evidence and the EC authorities did not make an objective examination of the relevant facts; (ii) the EC authorities did not give a reasonable explanation of how the facts supported the injury determination; (iii) the Brazilian exporters were discriminated against in the injury findings; and, (iv) the quotas agreed under the bilateral textile agreement precluded a finding of injury.

303. The EC argued that its determination had not violated Articles 3:1 to 3:4 and 8:2 of the Agreement. The EC argued Article 8:2 did not relate to a determination of injury, but only to the imposition of an anti-dumping duty. Further, in view of the text of Article 3 (which pertained to injury determination) and the application of the standards of review mentioned above (section IV), the investigating authority had a very broad discretion in determining the relative importance of each factor. Thus, the Panel could reach the conclusion that the provisions of Article 3 had been infringed only if the Panel were to consider that the EC did not take into account all relevant facts (which was not argued by Brazil in this respect) or, given the degree of discretion left by the Agreement, made a manifest error in interpreting the facts before it. The EC argued that Brazil had merely stated in its submission that the facts before the investigating authority did not support the determinations made. Brazil had not claimed that the EC had made a manifest error of fact or of interpretation of facts, and it could not be considered as having discharged its burden of proof before the Panel. The EC argued that Brazil did not seem to contend that all relevant facts were not taken into consideration. Rather, Brazil's arguments seemed to pertain to the relative weight given to each of the relevant facts.

(a) Violation of Articles 3:1 and 3:2: Injury findings not based on positive evidence, and failure to make an objective examination of the relevant facts

Introduction

304. Brazil argued that Article 3:1 imposed an obligation on the investigating authorities to make the injury determination on the basis of "positive evidence" and "an objective examination". Article 3:2 was a more detailed specification of that obligation. Brazil argued that the EC had violated Articles 3:1 and 3:2 in this case because it did not use the appropriate evidence. Brazil argued that the EC should have used Cacex statistics for its injury determination because they were the most accurate source of data, and were the data on which the EC itself appeared to rely on in the case of the EC/Brazil bilateral agreement on textiles under the MFA. Instead, the EC had used Eurostat statistics, which were less accurate and had been shown to be inaccurate in the past. The Eurostat statistics reported a higher import volume from Brazil compared to the Cacex data.

305. The EC argued that there was no evidence to show that the Eurostat statistics were not accurate. The Eurostat statistics were normally verified by the EC authorities, and showed data on imports that had entered into the EC. On the other hand, Cacex data were on exports from Brazil (and not imports into the EC). The EC argued that the determination of injury focused on import data and not on export data, and there could be various reasons for the data on exports and imports for any particular product within a particular time to be different. Also, Brazil's allegation regarding inaccuracy of Eurostat data was based on one other case which involved a different situation, i.e. in a case where product classification had resulted in an error which had been pointed out by the respondents themselves. In the present case, no such error had been pointed out, and the product categories were not such that errors of misclassification were likely to occur.

Arguments of the parties

306. Brazil argued that the data on the volume of EC imports of Brazilian cotton yarn used by the EC in its injury analysis were based on Eurostat statistics which reported a considerably higher figure than the official Brazilian export statistics (Cacex). Cacex data were gathered under a strict system imposed by the bilateral textile agreement which Brazil had concluded with the EC in the framework of the MFA. They were based on origin statements issued in Brazil after shipment, and the proof of the shipment was the bill of lading. Thus, Brazil was in a position to accurately monitor all exports of cotton yarn to the EC. Furthermore, these imports were also checked by the EC authorities when the goods were imported, and the EC had never challenged these statistics. On the other hand, Eurostat data were merely the result of adding up import data gathered by the customs authorities of each individual EC member State. Inaccuracies often occurred in this data and this had been acknowledged by the EC.

307. Brazil argued that evidence and arguments presented to the EC during the proceedings showed that Cacex figures were more accurate and reliable in this case. Indeed, if Eurostat figures were correct, Brazil would have exceeded its import quota under the bilateral textile agreement in both 1987 and 1988. The bilateral textiles agreement provided for immediate notification to the Brazilian authorities by the EC if quotas were exceeded, but no such notification was made to Brazil. Thus, the EC had itself considered the Cacex figures as being more reliable than Eurostat figures. Brazil argued that the use of the correct Brazilian Cacex figures would have shown that, similar to a number of developing countries involved in this case, the market share of Brazilian cotton yarn in the EC market was de minimis. Hence, termination of the proceedings against Brazil without the imposition of measures would have been the only justified outcome.

308. Brazil also argued that there was a marked contrast between the EC's adherence to the official FIBGE exchange rates (i.e. the official exchange rates mentioned in section V.2) and its outright rejection of the official Cacex statistics.

309. The EC recalled that the Brazilian exporters had also claimed before the investigating authorities of the EC that Cacex data were more reliable. However, the investigating authorities had considered that the import statistics of Eurostat were more reliable than Brazil's export data, because they represented the volume of imports cleared at the EC frontier and on which customs duties had been collected. It was the EC's practice to use Eurostat figures in verifying the quantities imported into the EC because these were official figures provided by the customs services of the member States and they were the only statistics which showed the quantity of goods which were "introduced into the commerce" of the EC, within the meaning of Article 2:1 of the Agreement. Therefore, the Eurostat statistics indicated the volume which had been proven to have entered for consumption into the EC market, and upon which the arguments relating to injury had been based. On the other hand, Brazil's export data consisted of records of goods being exported from Brazil, but with no proof that the goods had actually entered into the EC. The bill of lading was not, as claimed by Brazil, irrefutable evidence that the goods actually reached the destination mentioned on it, as it was not uncommon practice that goods at sea changed both ownership and country of destination.

310. The EC argued that it was not that the EC always relied on the Eurostat figures or that it always disregarded any other potential source of information on the volume of imports. For instance, when the relevant EC Combined Nomenclature (CN) code had a wider coverage than the product under investigation (and provided that there was complete cooperation from the exporters), the volume of exports could be assessed on the basis of information contained in the questionnaire responses. However, other sources may be resorted to only if it was proven that Eurostat data could not be used and that alternative reliable information was available. The EC argued that it had never "itself considered the Cacex figures as being more reliable than Eurostat", and had never acknowledged inaccuracies as regards imports of cotton yarn from Brazil in the years in question. The EC also said that it had not made any statement before the Panel or during the investigation that would support Brazil's argument that "it resulted from the evidence and arguments presented to the EC during the proceeding that in this case, Cacex figures were more reliable".

311. Regarding Brazil's argument that if Eurostat data were accurate then Brazil would have "exceeded" its import quotas under the bilateral textile agreement in 1987 and 1988, the EC argued that Brazil was ignoring that its bilateral textiles agreement with the EC contained so-called "flexibilities" provisions, which permitted the exporting country to make advance use of up to 5 per cent and carry over up to 7 per cent for each category of products under quota for each quota year. Since these provisions had been frequently resorted to by exporting countries, it was important to consider whether the total quantities actually imported during the period of application of the agreement did not exceed the quantities allowed for during the agreement's total period of application taken together.

312. Brazil argued that it had not alleged that the EC had made any statement that the Cacex data was more accurate than the Eurostat data. Brazil simply considered that all the evidence supported the view that the Cacex figures were more accurate than the Eurostat figures. Brazil argued that if the Eurostat figures were accurate (and were considered accurate by the relevant EC authorities), then the quotas agreed under the bilateral agreement would have been exceeded. If this were the case, certain provisions of the bilateral agreement would have come into operation. However, this did not occur and therefore Brazil inferred that, in the context of the bilateral agreement, the EC was itself proceeding on the basis that the Cacex figures, not the Eurostat figures, were accurate.

313. The EC recalled that the bilateral agreement between the EC and Brazil laid down quotas for the years 1987 to 1991, which overlapped with the period taken into account for determination of injury (1986 to 1989). The EC argued that the quantities allowed to be imported by the agreement were 98.145 tonnes in 1987 to 1989, and the quantities actually imported during that period were 97.907 tonnes on the basis of the Eurostat statistics. These figures were very close to one another and did not show any excess of imports for the period under consideration. Therefore, it was natural that the EC had not notified or complained to the Brazilian authorities for what Brazil was claiming as excessive imports in 1987 and 1988, since these imports were within the overall limits allowed by the bilateral textiles agreement. The EC therefore asked the Panel to reject as unfounded Brazil's argument that its market share of cotton yarn in the EC was de minimis.

314. Brazil argued that Article 6(1) of the bilateral agreement permitted the advance use of part of the following year's quota, equivalent to up to 5 per cent of the current year's quota. Article 6(3) permitted the transfer into Group I, category 1 (into which cotton yarn falls) from categories 2 and 3 of up to a further 2 per cent. The 7 per cent carryover provision of Article 6(2) could not be used, since the 1987 quota (32,165 tonnes) would have been fully used (Eurostat gave a figure of 35,430 tonnes). These Articles could not therefore form the basis for an increase in the quota sufficient to support the argument being put forward by the EC. In any event, these provisions could not have been used without prior notification by the Brazilian authorities (Article 6(5) of the bilateral agreement). No such notifications were given. It followed that the Eurostat figures clearly suggested that the quota was exceeded in 1988, even taking into consideration the various flexibility provisions referred to by the EC. However, no consultations were initiated by the EC under Article 7 of the bilateral agreement. Furthermore, no consultations were requested by the EC under Article 8(5) of the bilateral agreement (which dealt with discrepancies in statistics). Brazil inferred from this fact that, in reality, the EC operated the bilateral agreement by reference to the Cacex figures, not the Eurostat figures. In Brazil's view, therefore, there was a clear contradiction between, on the one hand, the EC's insistence that the Eurostat figures were correct in the context of this anti-dumping proceeding, and its apparent reliance on the Cacex figures in the context of the bilateral agreement. Brazil did not consider that recitals 37 to 38 of the Definitive Determination offered further clarification in this regard.

315. Brazil argued that while the EC was correct in noting that the total Eurostat figure of import volume of cotton yarn for the years 1987, 1988 and 1989 was less than the total quota for those years, this was not relevant to the argument being made by Brazil. The situation noted by the EC occurred only because exports fell dramatically in 1989, most probably as a result of the exchange rate freeze. What was relevant was the position in, for example, 1988. The quota in 1988 was 32,712 tonnes, the Cacex figure for that year was 26,736 tonnes, whilst Eurostat's figure was 35,430 tonnes. The position was similar in relation to 1987, where the Eurostat figures suggest that the quota was exceeded by 10.14 per cent.

316. Brazil argued that if the EC's argument that goods recorded as leaving Brazil for the EC may have subsequently been diverted to another destination were correct, then Cacex figures would be higher than Eurostat figures. However, in fact, the relevant Cacex figures were lower than the relevant Eurostat figures. Furthermore, even an argument that some goods recorded as leaving Brazil for other destinations in fact arrived in the EC would be incorrect because cotton yarn originating in Brazil could not be imported to the EC under the bilateral agreement unless accompanied by the correct documents issued by the Brazilian authorities. The Brazilian authorities applied stringent administrative procedures on its exports of cotton yarn to the EC and penalized the incorrect completion of bills of lading. The EC also penalized incorrect import declarations. These penalties varied across member States, and included fines, confiscation of goods and, in some cases, imprisonment.

317. Brazil argued that the most likely explanation for the discrepancies was the misclassification of goods by Eurostat, either according to the customs heading, or according to the country of origin. This had occurred in another case which involved Turkey, and had been acknowledged by the EC. Brazil argued that in that case, as a result of the correction to the data, the dumping duty applied to most Turkish exporters was reduced from 52.1 per cent to 10.1 per cent. The inaccuracy of the Eurostat figures was brought to the attention of the EC investigating authorities by ABIT on 28 October 1991 but the EC took no steps to request the member States to verify the accuracy of the Eurostat figures, as it had done in other anti-dumping proceedings when reasonable doubt existed concerning the accuracy of Eurostat data. This failure to verify was also a failure on the part of the EC to take the steps necessary to have special regard to the position of Brazil as a developing country.

318. Brazil further argued that in this case, the EC was reluctant to allow exceptions to the general rule that findings should be based on Eurostat data in the context of an anti-dumping proceeding. This was a rule easy to administer and could generally be applied in a non-discriminatory way. However, Brazil considered that such considerations could not prevail when it was demonstrated beyond reasonable doubt, as was done in this case, that the Eurostat figures were inaccurate. At the very least, the EC should have taken the usual steps to verify the accuracy of the data. This was all the more important in this case, since the figures had a critical impact on the injury analysis carried out by the EC.

319. The EC argued that Brazil was making an incorrect assumption in stating that the most likely explanation for the alleged discrepancies was the misclassification of goods by Eurostat either according to the customs heading, or according to the country of origin. In the case referred to and quoted by Brazil (polyester yarn from Turkey), the respondent had taken active steps to have specific customs declarations corrected for identified quantities. In the request for reclassification of certain exports filed with the United Kingdom authorities, the respondents in that case had provided positive information, and not mere assumptions or statements, that the Eurostat figures were not reliable. No such action was taken by the exporting companies or their counsel in the present case. Given the alleged importance of this aspect for the exporters, one could assume that if they had had serious evidence of misclassifications, they would have submitted them to the EC customs authorities. In this respect, the EC argued that the responsibility for correct classification of imported products was not with the national customs authorities or with Eurostat; rather, it was with the importer when filling in the customs declaration relating to a particular shipment.

320. The EC argued that in the polyester yarn case cited by Brazil as evidence of error in the Eurostat data, the procedure only covered yarns made of polyester staple fibres. The EC explained that the CN distinguished between man-made fibres and filament, and also between synthetic and artificial fibres. Moreover, within the synthetic fibres category (like polyester fibres) the CN identified four different sub-categories. As a result, there was a larger risk that misclassification occurred in the polyester yarn case than in the present case. In the present case, the risks of misclassification were very limited because (i) the product concerned was easy to identify and, hence, to classify; and (ii) the scope of investigation covered two complete CN codes (CN 5205 and 5206) with no sub-heading excluded in contrast to the product coverage in the polyester yarn case. Since the present case concerned all cotton yarns, misclassification would have no impact on the global import figures. Moreover, in the absence of positive evidence of misclassification, the EC had no reasons, on the basis of its practice (including polyester yarn from Turkey), to seek verification of import data.

321. The EC argued that Eurostat data were based on the declaration filed by the importers. Strict customs requirements, in the form of export and import licences were also applied by the EC authorities together with case-by-case customs verifications, which made Eurostat import data highly reliable. Moreover, customs authorities of the members States were required to regularly check import figures and correct them if necessary. The EC said that the product under investigation was subject to a bilateral export agreement concluded within the framework of the MFA. This agreement provided for very complete and strict means of control of the quantities imported (including for the product under investigation in the present case). Protocol A to the bilateral agreement provided that import licences were granted only on presentation of the original of the corresponding export licences (Article 11:1); exports of Brazilian origin which were not covered by Brazilian export licences issued in accordance with the provisions of Protocol A may be refused (Article 12:2). Moreover, each Brazilian export licence could only cover one of the categories of products made subject to quantitative limits (Article 6:2). As a result, each shipment had to be accompanied by the relevant export licence which could only cover a particular category of products. This system also implied that the nature and content of each shipment presented to customs was normally checked on a case-by-case basis to ensure the correspondence between the import licence, the export licence and the shipment, thus making mistakes on classification and quantities almost impossible.

322. The EC argued that Eurostat data in this case were based on these verified customs declarations. Moreover, in order to ensure that no reporting mistakes slipped in, Eurostat services in Luxembourg used monthly reports from customs authorities. These reports included volumes and values. Eurostat regularly controlled the data by calculating a monthly unit value for imports of a particular product from a particular origin and comparing this value with the unit value calculated on the same basis for the previous month. If, depending on the product, differences of a certain magnitude were identified, national customs authorities were requested to proceed to verifications. In the present case, the Eurostat services did not notice discrepancies. Under the Agreement, the EC had no obligation to accept a request from a complainant or a respondent unless it was satisfied that such a request was sufficiently justified on its merits.

323. The EC drew a distinction between internal verification of Eurostat data through, inter alia, the method described above and the verification of data apparently required by Brazil, i.e. verification of customs documents by national authorities, which was done when active steps were taken by the importer and a specific request to this effect was made on the basis of positive evidence. In the present case, in contrast with what had occurred in the polyester yarn (Turkey) case, no exporter had asked for such a verification.

324. The EC argued that all the evidence submitted by Brazil supporting its allegation of inaccuracy of the Eurostat figures consisted of saying that there was most probably a problem, because Brazilian export figures did not match Eurostat data. Brazil had however not provided evidence that the Brazilian data were more reliable. The EC further argued that the Cacex figures with respect to the product under investigation, while being different from the Eurostat figures, were also different from the export data notified by Brazil to the United Nations. For instance, for the year 1989 (the investigation period), the Eurostat figures for imports from Brazil into the EC were 26,863 tonnes. Brazil claimed that the Cacex export figures were 22,201 tonnes. However, the export figures for the same product notified by Brazil to the United Nations (SITC 65133 and 65134, which corresponded to CN 5205 and 5206) for the same year showed a quantity of 24,293 tonnes. Thus, the EC had serious doubts that the Cacex figures were reliable and should have been used instead of the Eurostat figures. The EC said that the difference between the Eurostat figures and the United Nations SITC figures could be explained by the fact that the product coverage of CN 5205 and 5206 was a bit broader than that of SITC 65133 and 65134.

325. The EC further argued that Brazil's allegations regarding the use of Cacex figures demonstrated only that the EC had not taken advantage of the enforcement provisions available under the bilateral agreement for a particular year (as the system was based on a global quota for the implementation of the agreement allocated on a yearly basis). This did not prove that the EC had relied on Cacex figures for the administration of the quotas. Brazil's allegation that the EC actually managed the quotas under the bilateral agreement on the basis of Cacex data was purely an assumption, based on incorrect factual statements. The EC argued that when the purpose of an investigation was to identify the quantities of a particular good entering the territory of the investigating authority, it was a priori more logical and reliable to take into account the statistics recording imports than those recording exports. The fact that Cacex figures were lower was not incompatible with the fact that imports for a given period were higher than exports for the same period, inter alia, for two reasons. The first was that goods could transit through a third country and be stocked there for some time before being exported to their final destination. The second was that Eurostat figures recorded goods put into free circulation in the EC. It may be in the interest of the importer to stock the goods in a customs warehouse for some time and clear them through customs only when they needed to be delivered. In such cases, import statistics could include quantities which were recorded on export statistics before the investigation period. This was all the more possible if one kept in mind the fact that under the bilateral agreement, export and import licences had a relatively long period of validity.

326. The EC argued that it was not correct that with the exports for a given year exceeding the annual quota, the enforcement mechanisms available to the EC under the bilateral agreement would have been triggered. The quota was set for a particular implementation period under the bilateral agreement and the fact that exports exceeded an annual allocation of the quota was not as such a justification for initiating consultations if requests for flexibility had been made; the EC noted that Brazil had itself agreed that the total quota was not exceeded for the years 1987 to 1989. Therefore, no reaction of the EC was necessary. Also, it was not correct that Brazil had never notified its intention to take advantage of the flexibility provisions. Brazil had actually asked for flexibility several times, and the EC had evidence of numerous such requests. The EC argued that as this was the main point on the basis of which Brazil alleged that the EC had used Cacex figures (instead of Eurostat figures) to administer the bilateral agreement, the argument of Brazil in this respect should be rejected as factually incorrect.

327. The EC argued that given that the Agreement did not contain any specific obligation as to what figures should be used by the investigating authority and given that the EC's obligation in this respect was limited to a global obligation to base its findings on objective and verifiable facts, the EC was of the opinion that the Panel should limit its review on this issue to whether the EC made a manifest error of facts or of appreciation of the facts. None of the arguments raised by Brazil on this issue established the existence of such an error. For those reasons, the EC considered that this claim of Brazil should be dismissed.

328. Brazil argued that the notifications provided by the EC appeared to relate largely to regional (i.e. within the EC) flexibility. The arguments raised by Brazil related to global flexibility. In this respect, it was notable that the EC had made no attempt to demonstrate how the documents provided by it might actually have affected the relevant global figures. The key point being made by Brazil was that even taking into account the relevant flexibility provisions, the EC had offered no explanation of how the Eurostat figures might be reached.

329. Brazil did not agree with the EC's argument that the difference between the Eurostat figures and the United Nations SITC figures could be explained by the fact that the product coverage of CN 5205 and 5206 was a bit broader than that of SITC 65133 and 65134. Brazil claimed that the product coverage of SITC 65133 and 65134 was similar, if not identical, to that of CN 5205 and 5206.

330. With regard to the figures allegedly provided by Brazil to the United Nations, Brazil stated that the basis on which products were classified under the United Nations system had been revised three times. This could be seen from the document provided by the EC to the Panel, which was headed "Rev 3", i.e. the third revision. The basis of the second revision could be seen in the final column of the same document headed "Rev 2". Obviously such reclassification could give rise to changes in the relevant figures. It was therefore quite possible that the United Nations figures referred to a different classification base than the CN figures, and therefore they were different in comparison to the Cacex figures as well as to those provided by the EC in this case.

331. Brazil also noted that figures provided by the EC to the GATT integrated data base, which served as reference for market access negotiations in the Uruguay Round, were different from Eurostat figures. According to Eurostat figures, EC imports of cotton yarn from Brazil in 1988 were of 35,430 tonnes, while the corresponding figure according to the data provided by the EC to the GATT integrated data base amounted to 28,544 tonnes.

(b) Violation of Articles 3:2, 3:3 and 3:4: Lack of an acceptable explanation of how the facts supported the injury determination

Introduction

332. Brazil argued that on the one hand the EC authorities determined that there was a decrease of imports of cotton yarn from Brazil both in absolute and relative terms, but on the other hand that price undercutting of these imports was the lowest among the countries that were investigated, and that export prices of cotton yarns from Brazil increased during 1989. However, the EC had failed to explain how it had taken these findings into account in its determination of injury. Brazil did not consider that the facts stated in the injury analysis reasonably supported the determination that exports from Brazil caused injury to the EC industry. Therefore, Brazil argued that the EC had violated Articles 3:2, 3:3 and 3:4 in its determination in this case.

333. The EC argued that its determination had not violated Articles 3:2, 3:3 and 3:4 in this case. The EC argued that under Article 3:2, as confirmed by a number of panel reports, the investigating authority was only required to consider whether there had been a significant increase in the volume of imports and whether there had been a significant price undercutting or price suppression. The EC had met these requirements and those specified under Articles 3:3 and 3:4. The EC also argued that both Articles 3:2 and 3:3 provided that "no one or several of these factors can necessarily give decisive guidance". Thus, affirmative injury determinations could be made even if one or more of the factors which had to be taken into consideration did not lead to such a conclusion. The EC also recalled its arguments regarding the standard of review by the Panel. In light of the above, the EC argued that the Panel should reject Brazil's claims regarding lack of an acceptable explanation of how the facts supported the injury determination.

Arguments of the parties

334. Brazil argued that the facts stated in an injury analysis must reasonably support the determination that was reached. Brazil argued that in this case, the EC's imports from Brazil had the lowest level of price undercutting, their volume was decreasing (both in absolute and relative terms), and the US$ prices of the Brazilian exports were comparatively stable and even increasing in 1989. Brazil argued that the EC had failed to explain how it had taken these findings into account in its determination of injury. The facts stated in the injury analysis did not reasonably support the determination that exports from Brazil, as opposed to those from the countries excluded from the proceeding, caused injury to the EC industry.

335. Brazil argued that Article 3:4 of the Agreement provided that "injury caused by other factors must not be attributed to the dumped imports". In this regard, Brazil noted that the EC had determined that "the investigation did not reveal any factors other than the dumped imports which caused material injury to the EC industry". However, the findings showed the existence of other causes of injury, particularly the existence of numerous non-dumped imports which undercut the EC producers' prices. In this respect, Brazil considered that the EC failed to reasonably explain how the facts supported its injury determinations.

336. The EC argued that if Brazil's complaint was that the EC did not give sufficient explanation of how the facts supported the injury determination (or that the EC had insufficiently explained its reasoning), this had nothing to do with Article 3, but related to Article 8:5. However, given that Brazil had not raised any claim on the basis of Article 8:5, this would be at this stage a new claim that should be disregarded by the Panel, and certainly not redefined by the Panel on its own initiative. If, on the contrary, Brazil was satisfied with the degree of information contained in the statement of reasons but disagreed with the conclusions reached on the basis of the facts of the case, the EC accepted that this was an issue to be reviewed on the basis of Article 3.

337. The EC argued that in both the Provisional Determination (recitals 26, 29-30 and 32) and the Definitive Determination (recitals 35 to 42), the EC had explained that the Brazilian market share was significant and Brazilian products were in competition with other imports and the like product produced in the EC. Regarding Article 3:2, the EC argued that the consideration of whether there had been an increase in the volume of imports was made on a cumulated basis. The cumulated market share of Brazil and Turkey was 9.6 per cent in 1986, 9.3 per cent in 1987, 9.8 per cent in 1988 and 8.7 per cent in 1989 (Definitive Duty Regulation, recital 41). Even though this market share slightly decreased over the reference period (1986-1989), it remained significantly high over the period. With regard to the effect of dumped imports on prices, the EC found significant price undercutting on the part of Brazilian exporters: between 2.6 per cent and 7.5 per cent (Provisional Determination, recital 32). The findings of the EC on price undercutting were confirmed by the Council (Definitive Determination, recital 42). Price depression and price undercutting on the EC market were considered as resulting from the low prices of dumped imports (Definitive Determination, recital 47). Thus, the EC had considered the two aspects of Article 3:2, i.e. volume of imports and price of dumped imports, and found that there was price undercutting

338. Regarding Article 3:3, the EC argued that it had evaluated the relevant economic factors for the EC industry, i.e. production and capacity utilization, sales and market share, prices (which were found to have significantly declined in 1988 and 1989), investments (which did show some increased productivity), profitability (which was found to have decreased between 1986 and 1989), and plant closures and job losses (Provisional Determination, recital 40). These findings were confirmed by the Definitive Determination, in particular the sharp decline of selling prices, financial losses (especially in 1988 and 1989), lack of return on investment, closure of a large number of plants and a substantial loss of jobs (recital 43). The EC argued that, as stated in the Definitive Determination (recital 45), the relevant economic factors of injury must not be evaluated in isolation since no one or several of them can necessarily give decisive guidance. Each of them had to be analysed in close conjunction with the others. On that basis, the EC considered that the negative factors (such as losses, plant closure and lost jobs) outweighed other factors such as EC producers' output and market share, which could have been evidence of a less negative situation.

339. The EC argued that Brazil had not offered any evidence to support its allegations regarding its claim that "the findings clearly showed the existence of numerous non-dumped imports which undercut the EC producers' prices". The EC argued that the Definitive Determination showed that the EC had examined a number of other relevant economic factors, such as a sharp decline of selling prices (recital 45), financial losses and lack of return in investment (recitals 44 and 45), and closure of plants and substantial loss of jobs (recital 45). With respect to other factors of Article 3:4, the EC had made a determination that they had not contributed to the injury suffered by the EC industry (for example in recital 44 of the Provisional Determination and 51 of the Definitive Determination). Furthermore, the exporters concerned had made a number of other allegations in the course of the investigation trying to show that reasons other than dumped imports were causing the injury to the EC industry. The EC had replied in detail to all these claims in recital 48 of its Definitive Determination, and those reasons reinforced the view that injury had indeed been caused by the dumped imports from Brazil (and Turkey). In addition, the impact of imports from other sources was assessed in recital 44 of the Provisional Determination. It was concluded that only one other country (Switzerland) exported cotton yarn to the EC in considerable volume during that period, but that these exports had not caused market disruption nor was there evidence of any dumping or injury caused by imports from this country. Furthermore, the market share of other countries exporting cotton yarn to the EC in the period in question was always substantially below 1 per cent.

340. The EC argued that Brazil had merely stated in its submission that the facts before the investigating authority did not support the determinations made. It did not claim however that the EC had made a manifest error of fact or of interpretation of facts. As a result, Brazil could not be considered as having discharged its burden of proof before the Panel with respect to this claim.

341. Brazil argued that the situation in this case represented a violation of both Article 3 and Article 8:5 of the Agreement. This was not a new claim by Brazil, but one which had been made by Brazil throughout the proceeding. Brazil argued that its claim was both that the facts of the case could not support the findings reached by the EC, and that the regulations imposing the anti-dumping measures did not state the findings and conclusions reached on all issues of fact and law in a manner that conformed with Article 8:5. Thus, the facts stated in the Provisional and Definitive Determinations did not sufficiently support the conclusions reached. In Brazil's view, although Article 8:5 might not be referred to explicitly in its first written submission to the Panel, the fact that this argument was being made by Brazil was clear from the following sentence of Brazil's first written submission, which stated that: "In this respect, Brazil considers the EC failed to reasonably explain how the facts supported its injury determinations." Brazil did not therefore agree with the suggestion by the EC that the clarification and development of Brazil's arguments on this point could be classified as a new "claim".

342. The EC reiterated its arguments regarding the sufficiency of explanation of how the facts supported the determination. The EC also argued that the Definitive Determination did not commit a manifest error in the appreciation of these facts by stating:

"... the Commission had concluded that the investigation had not revealed any factors other than the dumped imports from Brazil, Egypt and Turkey (which were causing injury). This conclusion was based mainly on the fact that the loss of profitability and the other negative economic circumstances had coincided with the continuing price depression and price undercutting of the Community market, owing to the low prices of dumped imports." (recital 47).

 

343. The EC argued that Brazil had not provided any proof to establish that this passage of the Definitive Determination was a manifestly erroneous interpretation of the facts before the investigating authorities or that there was no support for the conclusion that dumped imports from Brazil were causing material injury, particularly in view of the last sentence of Article 3:2 that "no one or several of these factors can necessarily give decisive guidance".

344. The EC argued that it had respected the provisions of Articles 3:2, 3:3 and 3:4 of the Agreement by clearly establishing in the Definitive Determination that the injury caused to its industry was caused by the dumped imports and not by other factors. By continuing to base itself on export and consumption statistics, Brazil was repeating the arguments already made by its exporters during the investigation. As argued above by the EC, those arguments were not correct. Brazil had offered no new evidence to support its claims, and seemed to ignore the fact that the last sentence of Articles 3:2 and 3:3 explicitly indicated that none of the factors mentioned there (for example, price undercutting) could give decisive guidance. Further, the EC argued that there was no valid reason to disregard Eurostat figures in the absence of positive evidence that they were erroneous. The market share and volume of exports of Brazil could therefore be based legitimately on Eurostat figures. The EC also reiterated its other arguments mentioned above to support its claim that it had not violated Articles 3:2, 3:3 and 3:4 in this case.

(c) Violation of Article 3.2 in combination with Article 8.2: Discrimination against Brazilian exporters in the injury findings

Introduction

345. Brazil argued that in conducting its injury analysis, the EC had discriminated against Brazil in comparison to other countries which were also subject to the same investigation, and therefore had violated Article 3:2 in combination with Article 8:2. Brazil argued that though 18.36 per cent of the Brazilian sampled exports were determined by the EC to be not dumped, the EC still treated all the imports from Brazil as dumped imports in its determination of injury. This was different from the treatment of imports from two other countries subject to the investigation (India and Thailand), for which the EC had considered only dumped imports. The imports from those countries were excluded from the investigation on the grounds that they were de minimis. Brazil argued that its exports to the EC should have been considered de minimis in the same manner as for those two other countries, a result particularly valid if only the dumped imports from Brazil were taken into account. Brazil also argued that the EC had used an official exchange rate for Brazil but a different, i.e. non-official, exchange rate for Egypt, whose dumping margin was found to be de minimis after resorting to the non-official exchange rates.

346. The EC argued that its determination was not inconsistent with the requirements of Articles 3:2 or 8:2 of the Agreement. The EC argued that Brazil had not been discriminated against under the Agreement. Countries which had very low import shares were excluded from the investigation under Article 5:3 of the Agreement because they had negligible import shares. The exchange rate used for the imports from Egypt was an official rate, and thus there was no discrimination against Brazil even on that account. Furthermore, the EC argued that Article 8:2, which was being cited by Brazil in support of its claim regarding discrimination, applied only in the process of collection of duties and not in the determination of injury. Therefore, the EC argued that Brazil's claims regarding Articles 3:2 in combination with 8:2 should be rejected by the Panel.

Arguments by the parties

347. The arguments of the parties are organized in three sections, which respectively cover the following issues: relevance of Article 8:2 in respect of Brazil's claim; alleged discriminatory treatment of imports from Brazil; and, alleged discrimination with respect to the exchange rate used by the EC.

(i) Arguments relating to the relevance of Article 8:2

348. Brazil argued that the fundamental GATT principle of non-discrimination contained in Article I of the General Agreement was embodied in Article 8:2 of the Agreement which stated that "when an anti-dumping duty is imposed in respect of any product, such anti-dumping duty shall be collected (...) on a non-discriminatory basis on imports of such product from all sources found to be dumped and causing injury" (emphasis added by Brazil). Brazil considered that this general principle of non-discrimination applied in the context of the application of all provisions of the Agreement. As a consequence, any decision to exclude a country from a proceeding on the grounds that the volume of its exports were de minimis, i.e to impose and collect duties, must be taken in a non-discriminatory manner. Brazil argued that the EC had not fulfilled this requirement because it discriminated against imports from Brazil in this case, and therefore violated Articles 3:2 in combination with Article 8:2.

349. The EC argued that Brazil's interpretation of Article 8:2 appeared to be over-inclusive when compared with the interpretation warranted by the internationally accepted principles of interpretation found in Article 31.1 of the Vienna Convention on the Law of Treaties. Firstly, Article 8:1 was concerned with the conditions for the imposition of duties. This was confirmed by the wording of Article 8:1 which imposed a prerequisite that "all requirements for the imposition [of any anti-dumping duty] have been fulfilled". Therefore, interpreting Article 8:2 as relating to the imposition of duties would make Article 8:1 redundant. Secondly, as far as Article 8.2 was concerned, this provision clearly referred to collection of duties. Interpreted according to its ordinary meaning and in its context (i.e. including paragraph 1), this provision clearly dealt with non-discrimination among countries found to be dumping and causing injury at the stage of the collection of duties, not during the process of making a determination on whether to impose these duties. The wording on which Brazil was relying in order to claim that no discrimination should take place at the time of the imposition of anti-dumping duties stated that no discrimination should apply among those sources of products found to be dumped and causing injury by the investigating authority. The words "found to" had to be interpreted as meaning that this provision applied only to the collection of the duties. The issue of the determination of dumped imports was addressed in other provisions of the Agreement.

350. The EC argued that its position was not that an investigating authority should be free to discriminate between exporting countries. The EC considered that the Agreement contained an obligation not to discriminate between two parties in the same situation. This principle was implicit in Article 8:1, where it was required that all the requirements for the imposition of anti-dumping duties be fulfilled before measures could be imposed. Article 8:2 recalled this principle for a situation where the respondents were in the same situation, i.e. they had all been found to be dumping and causing injury. However, parties which were not in the same factual situation did not have to be treated identically. In the present case, Brazil had not shown that discrimination took place at the level of the imposition of duties. Only Brazil, not India or Thailand, was found to be causing injury by the investigating authority (more detail on this issue is given in section (b) below).

351. The EC argued that "exclusion" of exporters from an investigation on de minimis grounds was done in conformity with Article 5:3 of the Agreement, which read, in the relevant part: "There should be immediate termination in cases where the margin of dumping or the volume of dumped imports, actual or potential, or the injury is negligible." This was a specification of the more general obligation imposed by Article 5:3 that "An application shall be rejected and an investigation shall be terminated promptly as soon as the authorities concerned are satisfied that there is not sufficient evidence of either dumping or of injury to justify proceeding with the case." In the present case, the EC had argued that "the effects of dumped imports of cotton yarn, originating in the countries concerned and taken in isolation, have to be considered as causing material injury to EC industry" and that "given the negligible market shares held by the exports at dumped prices from India and Thailand, ... these exports have not significantly contributed to the injury suffered by the Community industry."

352. Brazil argued that to the extent that the legal basis for the decision to exclude certain countries from the proceeding in this case was based on Article 5:3 of the Agreement, Brazil considered that Article 5:3 must be also applied in a non-discriminatory way. To the extent that the EC had failed to do this, Brazil argued that the EC had violated the obligation imposed by Article 5:3 of the Agreement.

353. Brazil argued that even if Article 8 were interpreted as referring to the imposition of duties in Article 8:1, and Article 8:2 to only the collection of duties, it did not follow that discrimination was permissible in the context of a Decision not to impose duties on the grounds that imports from a certain country were de minimis. Discrimination was prohibited by Article I of the GATT in both the imposition and the collection of duties.

354. Brazil recalled that the text of Article 8:2 stated "[w]hen an anti-dumping duty is imposed in respect of any product, such anti-dumping duty shall be collected in the appropriate amounts in each case, on a non-discriminatory basis on imports of such products from all sources found to be dumped and causing injury ..." (emphasis added by Brazil). Brazil argued that this text showed that the intention of the parties was that duties should be both imposed and collected in a non-discriminatory manner. If Article 8:2 were concerned solely with collection as opposed to imposition, the words underlined would not make sense. They would be expected to read instead "from all sources on which duties have been imposed". The wording of Article 8:1 was simply intended to clarify that the authorities of the importing country may decide whether or not to impose duties, i.e. they were not automatically bound to do so. However, it was clear that they must reach that decision in a non-discriminatory manner. The particular provisions of Article 8 cannot be interpreted in such a way so as to override the general principle of non-discrimination contained in Article I of the GATT. Brazil argued that Article 8:2 applied to all of the determinations on the basis of which a decision had been reached on whether or not to impose measures. It was a particular expression of non-discrimination established by Article I of the General Agreement.

355. The EC argued that Brazil's arguments in relation to Articles 5:3 and 8 were exclusively based on the premise that the Cacex figures should have been used instead of the Eurostat figures. As the EC had demonstrated that there was no reason for doing so, the question of a discrimination was not valid. In the present case, the Indian and Thai market shares during the investigation period (1989) were respectively 0.7 and 0.1 per cent, whereas the market share of Brazilian imports in 1989 were 2.25 per cent. Thus, Brazil was not in the same situation as these two other countries. There was consequently no question of discrimination, neither under Article 5:3, as the volume of Brazilian dumped exports were much higher that the volume of Indian and Thai dumped exports during the investigation period, nor under Article 8, as both India and Thailand were found not to be causing injury (Recital 46 of the Provisional Duty Regulation) and consequently did not have to be subject to the imposition of the anti-dumping measures.

(ii) Alleged discriminatory treatment of imports from Brazil

356. Brazil recalled the statement in the Definitive Determination that the injury analysis of imports from two countries, India and Thailand, was based on the volume of "dumped imports". In contrast, the injury analysis of imports from Brazil was based on all imports, whether or not dumped. The same applied to the market share analysis. Brazil argued that the Definitive Determination was unclear on whether the EC was taking all imports or only dumped imports into consideration and it had failed to explain how "dumped imports" from other countries had been calculated. Brazil further argued that the practice of the EC authorities to either "cumulate" or consider de minimis a given amount of dumped imports had to be conducted in a non-discriminatory manner. By not doing so, the EC had violated Article 3:2 in combination with Article 8:2.

357. Recalling that the EC authorities had concluded that imports from other countries did not significantly contribute to injury and that price undercutting was deemed to be the main factor causing injury, Brazil argued that imports from Brazil had the same effect on the EC industry as imports from the countries with a lower market share, because although imports from those countries accounted for a slightly higher market share, price undercutting by those imports was much higher. Brazil argued that the EC authorities had failed to explain reasonably why the other imports were excluded, while imports from Brazil were included in the injury analysis.

358. Brazil argued that the EC's own disclosure letter had shown that 18.36 per cent of the verified imports from Brazil were not dumped. Despite this, the EC had treated all of its imports from Brazil as dumped imports. This was shown in recital 29 of the Provisional Determination which referred to the volume of "dumped imports" for Brazil, Egypt and Turkey, and by recital 41 of the Definitive Determination which maintained this approach in relation to Brazil and Turkey. Even a consideration of the data made it clear that, in relation to Brazil, the EC did not deduct from the Eurostat figures an amount to reflect the quantities (or percentage) of cotton yarn that the EC's investigation had shown not to be dumped. Had this been done, the 1989 Eurostat figure would have been reduced by at least 845.644 tonnes for Brazil, i.e. by 18.36 per cent or the percentage of verified exports found not be dumped. This was apart from any amounts from Turkey, and (in relation to the provisional findings) Egypt, that were found not to be dumped.

359. Brazil noted that the EC did not "cumulate" India and Thailand because of their small market share of "dumped" imports (0.1 per cent and 0.7 per cent, respectively). Brazil argued that the Brazilian cotton yarn imports into the EC were as marginal as those from at least one of the other countries, and therefore should also have been considered de minimis. This was even more true if it was considered that the import and market share figures relating to Brazil were based on total imports and not on "dumped" imports. If only "dumped" imports had been considered, Brazil's market share would have been lower.

360. Brazil argued that the analysis of the volume of imports from India and Thailand was conducted by the EC in a different manner in comparison to the volume of imports from Brazil, Turkey and (in relation to the provisional determination), Egypt. The volume of "dumped" imports reported by the EC for India and Thailand were substantially below the volume of imports reported by Eurostat, and the only explanation for this difference was that, in the case of India and Thailand, the EC did deduct from the Eurostat figures the quantities (or, most probably, the percentage) of cotton yarn that the EC's investigation had shown not to be dumped. Therefore, the EC's determination on the relevance of the volume of imports to the injury analysis was clearly discriminatory, and hence in violation of Article I of the GATT.

361. Brazil argued that the practical consequences of the approach adopted by the EC were serious. Had the EC used the accurate Cacex figures and reduced this amount by the percentage of imports found not to be dumped (i.e. by 18.36 per cent), as was apparently done for the excluded countries, the market share calculated for Brazil in 1989 would have been 1.53 per cent. It was the EC's well established practice that market shares of this order should be considered de minimis. Brazil argued that this was also confirmed by the Uruguay Round Anti-Dumping Agreement which referred to a market share of less than 3 per cent.

362. The EC argued that there was no evidence that any imports from Brazil were not dumped. In view of the large number of exporters involved, the EC had used sampling in this case for all countries except Thailand. In the case of Thailand, only two companies had cooperated and therefore there was no necessity for selection. The Brazilian trade association was informed of the methodology and had raised no objection. The EC had found dumping in the case of all Brazilian exporters who were visited in order to verify their data. The EC applied a weighted-average of their dumping margin to those cooperating producers which were not subject to verification, and applied a residual duty to non-cooperating exporters. In these circumstances Brazil could not claim that there was a difference between "dumped imports" and "all imports". Their volume was exactly the same.

363. The EC argued that the Agreement required that a causal link be established between the products exported by "all sources found to be dumped" and injury. It was necessary therefore to establish an average dumping margin for each exporter. The EC did this by calculating a weighted margin where the individual margin of dumping found was weighted by the volume of the corresponding transaction and all other transactions had a weight of zero because no dumping was found for those transactions. This made it possible to establish the impact of the dumping practised by the exporter on his price behaviour, as it was this price behaviour that may have a negative impact (i.e. cause injury) on the domestic industry. This method, which was consistent with the object and purpose of the Agreement, resulted in imposition of duties (which apply to future imports, and therefore affect the overall price behaviour of the exporter in the future) on the basis of the overall price behaviour of the exporter in the past. The other key feature of this method was that it allowed the setting of duties at a level which took account of the exporter's propensity to dump, because the average dumping margins for the exporters concerned were lower as there were fewer transactions in respect of which a margin of dumping was found: thus, duties were correspondingly lower in proportion to how less often the exporter practised dumping. This reflected accurately the link between the injury caused by the "dumped imports" and the duty which aimed at removing that injury. Without this averaging method, there would be absurd consequences, which would not be in conformity with the object and purpose of the Agreement, i.e. dumping margins (and therefore duties) would be higher and assuming that the dumping margin remained the same, the fewer the transactions where a dumping margin was found the higher the duty would be, because the margins would no longer be "diluted" by the other export sales by that exporter.

364. The EC argued that since all imports from Brazil were determined to be dumped on the basis of sampling, and given that all the three companies subject to verification were found to be dumping, the issue was not one of discrimination. The issue was to determine whether the EC applied its sampling methodology inconsistently with the Agreement, and whether the EC was entitled to agree with the exporters concerned on a particular sampling technique and to apply the results of the investigation conducted on that basis to all imports from Brazil. The EC recalled that the Agreement was silent on the use of sampling techniques. Previous panel reports had confirmed that there was no basis in the Agreement upon which to review the use of sampling methodologies applied by Parties. They concluded that a review of the sampling methodology had to examine whether the methodology could be considered to be sufficient to serve its stated purpose. Consequently, it seemed that previous panels had considered that their review of sampling methodologies should be limited. In view of the standards of review described above, the EC suggested that the Panel should examine whether in applying its methodology, the EC made a manifest error of facts, of interpretation of facts, or acted arbitrarily vis-à-vis Brazilian exporters. The EC recalled that the role of a panel was not to identify whether the EC could have applied a better methodology, but whether, in the present case, the EC acted consistently with its obligations under the Agreement. Brazil had not claimed that the sampling methodology was contrary to the Agreement, nor had it provided elements supporting the fact that, in applying the sampling methodology, the EC had acted arbitrarily or made a manifest error of fact or of appreciation of facts. The absence of arbitrary action or manifest error (or appreciation) of fact by the EC could be almost irrebutably presumed from the fact that respondents had been consulted and had agreed to the use of the sampling methodology applied by the investigating authority in the present case. Therefore, the arguments of Brazil should be rejected.

365. For the reasons mentioned above, the EC argued that it took into account all imports, whether dumped or non-dumped, in its analysis of the volume to be considered for injury purposes. As shown by Recital 27 of the Provisional Determination, Indian and Thai exports were excluded from cumulation on the basis of a de minimis total market share. If one considered the total imports from India and Thailand during 1989, their market shares were respectively 1 per cent and 0.5 per cent. Hence, both satisfied the de minimis threshold applied by the EC in cases involving several respondent countries. In Recital 30 the investigating authority, as regards India and Thailand, took into account the market share of dumped imports only, but that Recital addressed a totally different issue than that of cumulation. Nonetheless, the total market share was in any case low enough to justify the exclusion of India and Thailand on the basis of a de minimis market share. As a result, if Brazil claimed that the figures mentioned in Recital 30 referred only to dumped imports, this was correct and the EC did not deny it. However, if Brazil claimed that the EC relied on market shares of dumped imports from India and Thailand to exclude them from the determinations, this clearly was not correct. The total market shares of these two countries were clearly de minimis. Therefore, their exclusion on the basis of de minimis total market share was fully justified.

366. The EC argued that Brazil was not correct in arguing that its market share was de minimis. On the basis of the EC's verified data on import volume and total EC consumption, Brazil's market share was found to be 2.25 per cent in 1989 and, therefore, way above the market share of 0.7 per cent for India and 0.1 per cent for Thailand (recital 39 of the Definitive Determination). For that reason, even admitting that Brazil's price undercutting was slightly lower than that found for India and Thailand, cumulation of Brazil's imports with those of Turkey was amply justified on the basis of an overall appreciation of volume and price levels, as required by Article 3:2 of the Agreement. For the same reasons, the exclusion of India and Thailand was also justified because of their de minimis market share. Therefore, there was no discrimination as to the identification of a de minimis market share.

367. The EC further argued that given the market share held by Brazil, the question whether total imports or only dumped imports should have been taken into account was totally irrelevant for the determination by the Panel. However, even if one were to agree to the figures submitted by Brazil, its market share would still be 1.53 per cent. This market share was far above the percentage usually considered by the EC as de minimis in investigations involving more than one country. For that matter, it was also well above the total market shares held by India and Thailand respectively. Therefore, the outcome of the case would not have been affected whether total imports or dumped imports only were taken into account.

368. Addressing Brazil's argument that the EC should have considered Brazil's market share as de minimis pursuant to its allegedly "well established practice", the EC argued that there was no provision in the Agreement defining what should be considered as a de minimis market share. Consequently, any practice of the EC with regard to the definition of a de minimis market share was only a unilateral practice and the Agreement did not contain any obligation to invariably follow a particular practice which it did not otherwise mandate. In this regard, the inclusion of a de minimis percentage in the WTO Agreement was of no relevance for the interpretation of the Agreement. It was clear from Article 18:3 of the WTO Agreement that the provisions of that agreement should not apply to existing measures. That provision confirmed a fortiori that, given that the situation resulting from the new rules will be totally different, it was not possible to draw any conclusion for the application of the Agreement from the options taken in the WTO Agreement. Moreover, Article 5.8 of the WTO Agreement referred to 3 per cent of the imports of the like products, which was in any event lower, if not much lower, than a 3 per cent share of the market. Finally, the cases referred to by Brazil related to different circumstances than those prevailing in the present case. Most of them were relatively old cases dating back to 1985 or 1987. In the Styrene Monomer case, the EC did not expressly state that the market share of exports was de minimis. In addition, there was no evidence of price undercutting or price depression, which was most probably the main reason why the investigation was terminated. In Standard Wood Particle Board case, the EC did not expressly describe the market share of the exporting countries concerned as de minimis. Indeed, other factors contributed to the conclusion that there was no material injury, including factors relating to the EC industry. With respect to the Audio-Cassettes case, it should be noted that even in the presence of a market share of 1.5 to 1.61 per cent, provisional measures were imposed on imports from Hong Kong. Definitive duties were not imposed not only because of the relatively small market share, but also because of the absence of brand recognition of Hong Kong products on the EC market. The EC argued that it had a much more developed "well established practice" of considering a market share as de minimis only if it is below 1 per cent. Moreover, the relevance of market share was part of a global analysis and, pursuant to the Agreement, it was not necessarily a decisive guidance.

369. Brazil argued that the disclosure letter from the EC had itself stated that 18.36 per cent of the verified exports from Brazil were found to be non-dumped. Therefore, Brazil argued that the EC was not correct in stating that the volume of dumped import and all imports from Brazil were exactly the same.

370. Brazil argued that it was not in a different position by comparison with other countries that were investigated. Recital 9 of the Provisional Determination showed that many cooperating companies in all countries subject to investigation were not verified and, accordingly, sampling was used in all of the countries subject to the investigation. In the case of India and Thailand, the volume of imports into the EC used in the injury analysis was substantially below the figure provided in Eurostat. The only possible explanations for this were that non-dumped imports of verified companies were deducted from the Eurostat total, or imports of non-verified companies were considered as not dumped, or a percentage of imports from non-verified companies were considered not dumped, such percentage being the percentage of imports from verified companies found not to be dumped. Thus, the volume of imports used in the injury analysis for these countries was based on dumped imports. In the case of Brazil, however, the volume of imports used in the injury analysis was based on all imports, whether or not dumped. This was clearly discriminatory.

371. The EC argued that if imports were excluded from an investigation because their volume was negligible this was because they could not be considered as contributing to the injury in any significant manner. The EC reiterated its arguments that the factual situation was that Indian and Thai imports had a negligible market share and could therefore be excluded from the determination of injury. In contrast, Brazil's market share was much higher. In view of the arguments given above, the EC argued that its consideration of imports from Brazil in the determination of injury did not violate Article 3, 5 or 8 in this case.

(iii) Alleged discrimination with respect to the exchange rate used

372. As mentioned in section V.2 above, Brazil argued that the EC had discriminated against Brazil in comparison to Egypt by not considering the possibility of using non-official exchange rates for Brazil but actually using non-official exchange rates for Egypt.

373. The EC argued that in the case of Egypt, the cost of cotton imported from third countries (invoiced in US dollars) was recalculated on the basis of the special exchange rate in force, i.e. the official rate set by the Egyptian authorities for raw cotton transactions (recitals 17 and 20 of the Definitive Determination). The reason for doing so was the unreliability of raw cotton prices on the Egyptian market (recital 13 of the Provisional Determination), and had nothing to do with the developing status of Egypt. The treatment on this issue of Brazil and Egypt was identical, since both cases the official exchange rates were applied. Therefore, there was no question of any discrimination applied against Brazil.

374. The EC argued that the weighted average of the dumping margins for Egyptian companies investigated was 0.1 per cent ad valorem, and was thus considered to be de minimis (recital 31 (ii) of the Definitive Regulation). If the factual situation was that the conditions for the application of Article 5:3 were met, as they were in this case, then ipso facto there could not be any question of a violation of the obligation of non-discrimination imposed by Article 8:2. The latter, in fact, applied in respect of "products from all sources found to be dumped and causing injury", whereas in the present case imports from India and Thailand were found not to have caused injury (because of their negligible market shares), and imports from Egypt were found not to be dumped (because of the de minimis margin found)

(d) Violation of Articles 3:2,  3:3 and 3:4: Quotas agreed under the bilateral textile agreement precluded a finding of injury

Introduction

375. Brazil argued that the Multifibre Arrangement (MFA) effectively modified the Anti-Dumping Agreement with respect to products which were subject to the MFA. Thus, Brazil argued that in this case the EC had violated Articles 3:2 to 3:4 of the Agreement because it determined injury from, and consequently imposed anti-dumping duties on, imports from Brazil that were subject to import quotas under the Multifibre Arrangement (MFA). Furthermore, any issue of trade disruption on account of these products should have been settled under the MFA.

376. The EC argued that the Anti-Dumping Agreement was separate from the MFA, and the EC had not violated either of these two legal agreements in this case. There was no constraint imposed on anti-dumping actions by the MFA against products that were shown to be dumped and causing injury under the Anti-Dumping Agreement. The anti-dumping measures taken by the EC in this case were in conformity with the Anti-Dumping Agreement, and therefore, this claim of Brazil should be rejected.

Arguments by the parties

377. Brazil said that trade in textiles (including cotton yarn) between the EC and Brazil was regulated by bilateral agreements negotiated under the MFA (Multi-Fibre Arrangement). The quotas and monitoring provisions laid down under the MFA and bilateral agreements had been established to take the fullest possible account of the serious economic and social problems affecting the textile industry in both importing and exporting countries, and in particular, to eliminate real risks of market disruption on both the textile market of the EC and the textile trade in Brazil. Any trade disruption should be settled within the framework of the special procedure set up by the relevant agreement. Article 9.1 of the MFA expressly provided that: "In view of the safeguards provided for in this Arrangement, the participating countries shall, as far as possible, refrain from taking additional trade measures which may have the effect of nullifying the objectives of this Arrangement." Nowhere in the Definitive Determination had the EC offered any reason explaining why it was not possible to refrain from taking additional measures in this case. The EC simply stated that quantitative restrictions could not prevent injury resulting from unfair trading practices such as dumped imports at very low prices. Accordingly, Brazil considered that the EC's position on this point was contrary to the provisions of the Anti-Dumping Agreement.

378. Brazil argued that the bilateral agreement between the EC and Brazil effectively modified Article 3:2 of the Agreement, in so far as it applied to anti-dumping investigations conducted by the EC in respect of products exported from Brazil that fell under the bilateral agreement. Brazil argued that the volume of non-dumped imports could not be taken as a factor indicating injury in an anti-dumping proceeding, as long as the volume of those imports were within a quota agreed in the framework of the MFA. Such imports had to have the additional characteristic that they must be dumped, before they could be taken into consideration in the context of the injury analysis. However, in this case, the EC reached its determination on the basis that the entire volume of imports from Brazil was dumped. This violated Article 3:2 because the EC had itself determined that at least 845,644 Kg , i.e. 18.36 per cent, of the sampled imports were not dumped. In this situation, the only reasonable assumption would have been that at least the same percentage of total imports were not dumped.

379. The EC argued that this argument was raised inadmissibly in the present proceedings, because it related to an alleged violation of an agreement other than the one this Panel was established to examine. The MFA had its own dispute settlement procedures. Further, in recital 49 of the Definitive Determination the EC had explained that while quantitative restrictions, such as MFA quotas, protected the EC industry from excessive volumes of imports, they did not prevent injury resulting from unfair trade practices, such as dumped imports at very low prices even within the quotas allowed to be imported. Moreover, Article 9.1 of the MFA did not prevent anti-dumping action being taken when justified, because it was a best-endeavour clause that stated "... shall, as far as possible, refrain ...". Therefore, there was no question of a breach of the Agreement or of the MFA.

380. The EC argued that the principle according to which a Panel should only review the conformity of measures with provisions of the agreement under which it had been set up was well established. For instance, the temptation to invoke provisions of GATT before a panel established by the Committee on Anti-Dumping Practices was prevented by footnote 14 to Article 15 of the Agreement, which provided that: "[i]f disputes arise between Parties relating to rights and obligations under this Agreement, Parties should complete the dispute settlement procedures under the Agreement before availing themselves of any rights which they have under the GATT." The EC argued that this text implied that a party was not entitled to invoke a provision of GATT before a panel established under the Agreement. If this was the case regarding GATT, whose Article VI was further elaborated by the Agreement, a fortiori it should not be possible for a party to both the MFA and the Agreement to invoke a provision of the MFA before a panel established under the Agreement. Moreover, if it was a provision of the MFA which was allegedly violated, the issue had to be subject to dispute settlement under the MFA. Further, the EC argued that by making such allegations, Brazil was anticipating the integrated dispute settlement mechanism introduced by the WTO Agreement. If such a system had already existed under GATT 1947 and the Tokyo Round Agreements, the Uruguay Round negotiators would not have found it necessary to expressly organize it in the Understanding on Dispute Settlement.

381. The EC argued that the rules for the application of successive treaties relating to the same subject matter were spelt out in Article 30 of the Vienna Convention on the Law of Treaties. It was possible for two or more parties to a multilateral agreement to enter into a treaty later in time while the earlier treaty remained in force. In such a case, the provisions of the earlier treaty would apply between the parties to the later treaty only to the extent that those provisions were compatible with those of the later treaty. This implied not only that the two treaties should be related to the same subject matter but also that the provisions of the earlier treaty should be affected in some way by the provisions of the later treaty. The EC argued that the MFA or the bilateral textile agreement between the EC and Brazil on the one hand, and the Anti-Dumping Agreement on the other hand did not relate to the same subject matter. The former dealt with the establishment of quantitative restrictions in the field of textiles, the latter was concerned with the administration of a legitimate unfair trade practices instrument.

382. The EC further argued that even if, through an extensive and unwarranted interpretation of the terms, the Panel were to consider that the agreements related to the same "subject matter", the EC considered that neither the MFA nor the bilateral agreement between the EC and Brazil actually affected the rights and obligations of the Parties to the Anti-Dumping Agreement. There was consequently no question of applying the latter in a manner compatible with these agreements. Article 9.1 of the MFA only created a "best efforts" commitment. The commitment under Article 9.1 of the MFA was not elaborated further, either by other provisions of the MFA or by the bilateral agreement between the EC and Brazil. The bilateral agreement between Brazil and the EC was silent on the use of anti-dumping measures. Moreover, Article 1:1 of the bilateral agreement between the EC and Brazil provided that the provisions of the Geneva Arrangement (the MFA) applied to trade in textile products between the parties "subject to the provisions of this arrangement"; Articles 1:2 and 3 of that agreement prohibited quantitative safeguard measures or measures having equivalent effect to quantitative restrictions. However, apart from this specific prohibition, the bilateral agreement did not prevent any recourse to other trade measures, including anti-dumping measures. Also, the principles of interpretation did not permit to identify any precise circumstance where the parties would be mandated to refrain from having recourse to anti-dumping measures.

383. Brazil did not agree with the EC's comments regarding the link between the Agreement and the MFA. Brazil argued that there was a relationship between the Agreement and the MFA, which was relevant to the determinations made in this case. This had been disregarded by the EC.

V.4. Violation of Article 13: Failure to give due consideration to the status of Brazil as a developing country

Introduction

384. Brazil recalled that Article 13 stated that: "It is recognized that special regard must be given by developed countries to the special situation of developing countries when considering the application of anti-dumping measures under the Code. Possibilities of constructive remedies provided for by this Code shall be explored before applying anti-dumping duties where they would affect the essential interests of developing countries."

385. Brazil claimed that the EC had violated Article 13 because it did not give special consideration to the position of Brazil as a developing country and also failed to properly consider the constructive remedies put forward by the Brazilian exporters. In particular, the EC had violated Article 13 because it did not have proper regard to the relevance of the exchange rate freeze to the application of Article 2:4, any possibility of an allowance to take the exchange rate freeze into account in the context of Article 2:6 of the Agreement, the existence of the MFA, the importance of cotton yarn to Brazil's economy, the commercial realities of the difficult situation in which the Brazilian exporters found themselves, the long-term implications for Brazil's export sector (notably its cotton yarn and other textiles exports), and the possibility of finding a constructive solution which would not involve the imposition of measures.

386. Brazil argued that the EC's dismissal of the ABIT's submission concerning the existence of the exchange rate freeze in the application of Article 2:6 was a clear breach of that Article, as well as a breach of Article 13. This was because: the frozen exchange rate was clearly a "special situation" affecting Brazil in its capacity as a developing country; the EC failed to have "regard" to this special situation, in the sense that there was no meaningful assessment of the merits of the submission made by ABIT, i.e. they were simply dismissed as irrelevant; the EC failed to have "regard" to the special situation, in the sense that it took no steps to take it into account and provided no adequate explanation for its failure to do so; the EC failed to have "special" regard to the special situation, in that it took no steps to take into account the frozen exchange rate, and did not alter its normal practice; and the measures taken by the EC affected the essential interests of Brazil, yet the EC did not pro-actively explore the possibilities for finding a constructive remedy in the context of the application of Article 2:6.

387. The EC argued that it had met all the requirements of Article 13. The EC gave due consideration to the status of Brazil as a developing country at the time of considering the application of anti-dumping measures. Article 13 did not cover the issue of the calculation of the dumping margin, which had to be based on the facts of the case. Article 13 only applied at the stage when the investigating authority was considering the application of the anti-dumping measure. The EC had considered the possibility of the undertaking offered by Brazil but had not accepted it for the reasons mentioned in section M of the definitive determination and referred to earlier in section V.3 of this report.

Arguments by the parties

388. The arguments by the parties are presented in the following sub-sections: arguments relating to the first sentence of Article 13; arguments relating to the second sentence of Article 13; arguments relating to the coverage of Article 13; arguments relating to discrimination; and, arguments relating to the alleged refusal of the EC to consider its obligation arising from the MFA.

(a) Arguments relating to the first sentence of Article 13

389. Brazil recalled that the measures taken by the Brazilian authorities were publicly stated to be temporary, and it was clear to the Brazilian exporters that within a short period the cruzado would be allowed to continue depreciating. The temporary freeze was not a sustained change in the exchange rate. Commercial logic dictated that the exporters maintain their position in the export market and wait for the exchange rate to "catch up". Indeed, many of the exporters were bound to honour long terms contracts denominated in United States dollars. As expected, after three months, the Brazilian authorities unfroze the exchange rate. It was notable that during this period nothing changed from the point of view of the EC industry, i.e. the Brazilian export price remained basically the same in terms of the United States dollar. Nonetheless, contractual obligations concluded before the exchange rate freeze explained why during the first quarter of 1989 the Brazilian exporters were obliged to continue to export, despite the extremely unfavourable terms of trade. In April 1989, the volume of exports by the three companies verified fell very considerably. The EC could have taken this situation into account in the context of its obligations under Article 13 to have special regard to the situation of developing countries, but the EC refused to do so. According to the logic adopted by the EC, in order to avoid a finding of dumping, the exporters would have had to breach long terms contacts and temporarily increase massively their United States dollar export price to a point at which they would have been eliminated from the market.

390. Brazil argued that it was not commercially viable to constantly move in and out of an export market : the disruption to customers' sources of supply means that the export market will never really get off the ground. For countries with high inflation which periodically needed to tighten exchange controls (i.e. most developing countries), the EC's methodology was pre-programmed so that a finding of dumping was almost inevitable. The price at which the exporters would have had to sell to escape a finding of dumping based on the EC's methodology would have eliminated them from the market. Thus, the EC's position amounted to saying that cotton yarn producers could sell at any price they wished abroad, but in the circumstances prevailing it would have been impossible for them to avoid a finding of dumping. In the light of the EC's investigations, the price they would eventually have to pay for this "freedom" would be effective exclusion from the EC market for at least 5 years as a consequence of imposition of anti-dumping duties. Brazil argued that the EC's approach on this point directly contradicted the requirement clearly stated in Article 13 of the Agreement that special consideration be given to the position of developing countries.

391. Brazil argued that the phrase "situation of developing countries" envisaged by Article 13 encompassed situations which were connected to the status of a particular country as a developing country: a situation where the exports were raw material or semi-processed products, which were of strategic importance to the exporting country's economic development, and where there was financial instability (such as high inflation and a weak currency). Temporary measures taken to try and reduce that instability clearly fell within the meaning of Article 13. In this respect, Brazil noted that the Agreement referred to the: "particular trade, development and financial needs of developing countries" (emphasis added by Brazil)

392. Brazil argued that the presence of a "special situation" within the meaning of Article 13 in this case was also shown by the Decision of 5 May 1980 by the Committee on Anti-dumping Practices (hereinafter "the 1980 Decision"), which stated that: "In developing countries, governments play a large role in promoting economic growth and development in accordance with their national priorities, and their economic regimes for the export sector can be different from those relating to their domestic sectors resulting inter alia in different cost structures. This Agreement is not intended to prevent developing countries from adopting measures in this context, including measures in the export sector, as long as they are used in a manner which is consistent with the provisions of the General Agreement on Tariffs and Trade, as applicable to those countries" (emphasis added by Brazil). Brazil considered that a special situation within the meaning of Article 13 was present in this case.

393. Brazil argued that the special situation prevailing in this case was brought to the attention of the EC in writing on several occasions. However, the rule adopted by the EC in this case expressly prohibited the investigating authorities from having regard to the special situation prevailing. The Definitive Determination stated that the frozen exchange rate could not be "the subject of appreciation". Thus, Brazil considered that there was no proper "regard" of its special situation, in the sense that there was no meaningful review by the EC authorities of the arguments presented by ABIT and the Brazilian authorities. The EC authorities simply proceeded on the basis that all these arguments were irrelevant. They failed entirely to weigh the arguments presented. Therefore, Brazil considered that the obligation imposed by Article 13 was not complied with.

394. Brazil argued that the word "regard" in Article 13 required that at least the relevant circumstances must be looked at by the investigating authority, and where it was established that there was a "special situation" for the developing country, the investigating authorities must review this situation. This placed an obligation on the investigating authority to have regard to the facts reasonably available to it. These facts included at least those apparent from the complaint, from the responses to the questionnaires, from the submission by the exporters or EC producers, and from information submitted by other interested parties. The obligation required at least that: (1) the representatives of the investigating authority take due consideration of the submissions made orally and in writing; (2) the investigating authorities acknowledge receipt of and respond to relevant written submissions; (3) the public determinations of the investigating authorities refer to relevant written submissions, and (4) the public determinations of the investigating authorities be properly reasoned with regard to the facts submitted in writing, i.e. that the relevant facts reasonably support the determinations of the investigating authority. Brazil argued that a breach of Article 13 would be established where it could be demonstrated that an investigating authority failed to have "regard" to the special situation of a developing country.

395. Brazil argued that the word "regard" had a second, additional, meaning, i.e. it not only required the investigating authorities to look at the special situation prevailing, it required them to go further. It required them to pay attention to, take into account and, where appropriate, adapt their determinations to the special situation prevailing. Thus, faced with such a special situation that materially affected the determination, it would not be enough. for the investigating authority to state simply that it had read and considered the arguments submitted, but rejected them. There must be some coherent and valid reasoning to support a rejection of such submissions by interested parties. Therefore, even if it could be said that, in this case, the EC investigating authorities had regard to the arguments raised by ABIT in the first sense outlined above, i.e. that they looked at these arguments, it could not be said that they had regard to them in the second sense. The fact of the exchange rate freeze was not taken into account by the investigating authorities in any way, and no coherent and valid reasoning was given to support this determination. The frozen exchange rate was simply held to be not susceptible to appreciation. Therefore, in this additional sense, Brazil considered that the EC did not make its determination in accordance with the provisions of the Agreement.

396. Regarding the meaning of "special", Brazil argued that Article 13 did not refer simply to the situation of developing countries, but to the special situation of developing countries. Furthermore, it required investigating authorities not just to have regard to the situation of developing countries, but to have special regard. Thus, the obligation imposed on the investigating authorities by Article 13 was more onerous than that imposed by certain other provisions of the Agreement. For example, Article 3 of the Agreement required that the investigating authorities must have regard to the volume of dumped imports in an injury determination. Brazil considered that the obligation contained in Article 13 of the Agreement should be read as a stronger one than that contained in Article 3, since it required that the investigating authorities should have special regard to the situation of developing countries. Therefore, where a special situation existed, an investigating authority which did not depart from its standard practice would, prima facie, have failed to have special regard to that situation.

397. In view of this, Brazil argued that the EC's use of a monthly average normal value and monthly average exchange rates could not be characterized as having "special regard" to the "special situation", because monthly average exchange rates and/or monthly average normal values were often used by the EC in anti-dumping cases, for reasons other than the existence of high inflation, and also in investigations against developed countries. There was therefore nothing "special" about this methodology. Furthermore, the use of monthly average normal values and exchange rates evidently in no way addressed the special situation present in this case, namely the existence of the exchange rate freeze.

398. The EC noted that the first sentence of Article 13 started with the phrase "it is recognized" (in the French version "il est reconnu"), which seemed already to qualify any obligation which may be contained in the sentence. It would appear that this sentence was only the introduction of the actual obligation found in the second sentence of Article 13. Otherwise the negotiators could have used the word "shall". Moreover, the verb "to recognize" was used in a number of instances in the General Agreement and the Codes with respect to statements of principle. For instance, in Article XVI:3 of the GATT the word "recognize" was used with respect to a general consideration on the potentially harmful effects of export subsidies, while the actual obligation was stated in the following paragraph. Article XVIII:1, paragraphs 1, 2, 3, 5, 8 showed a similar situation and a similar feature was also found in Article XXVIIIbis. The same word was used in Article 11:1 of the 1979 Subsidies Code to acknowledge the positive role of subsidies. On that basis, the EC was of the view that the words "it is recognized" was evidence of the intention of the negotiators to consider the first sentence of Article 13 as a statement of principle, while the actual obligation in that Article was laid down in the second sentence.

399. The EC recognized that in the phrase "special regard", the term "special" was attached to the word "regard". However, the word "regard" could not be interpreted as meaning more than "consider", or, as Brazil had stated, "look at". Put in its context (i.e. at the time the investigating authority considered imposing measures), it meant that the investigating authorities should be ready to review the application of anti-dumping duties or to accept undertakings.

400. The EC argued that the text of Article 13 was clear that special regard had to be given to "the special situation of developing countries" (emphasis added by the EC). It did not state "to special situations in developing countries". What was referred to was the situation of developing countries in general, not a particular event or characteristic of the economy of the country under investigation. The "special situation" was therefore the global one of being "developing", as opposed to "developed". In the light of other provisions relating to developing countries such as the provisions of Part IV of the General Agreement, this particular situation could be considered as covering such situations as dependency on exportation of a limited range of primary products, lack of diversification of the economy or difficulty to gain access to export markets.

401. The EC further argued that the situation of a developing country could be illustrated by referring to elements which had been recognized, i.e. in Part IV of GATT, to be features peculiar to developing countries. In this respect, dependency on exports of a limited range of primary products, lack of diversification of the economy, difficulty to gain access to export markets are factors which can be found in developing countries, but normally not in developed countries. This was very different from the situation which Brazil defined to be a special one. High inflation was not a characteristic of developing countries. The situation described by Brazil was one which could be found in both developed and developing countries. Brazil's interpretation of Article 13 was that one special factual situation in a developing country should trigger any special regard. Accepting this interpretation would result in discrimination between developing countries on the basis of their comparative situation, a result obviously in contradiction with the purpose of Article 13.

402. The EC argued that interpreting the "freezing" of exchange rates to be a "special situation" relevant under Article 13 was contrary to the text of Article 13, as well as with the object and purpose of that provision. Article 13 was designed to give effect, within the Agreement, to the more general principle of special and differential treatment in favour of developing countries in the multilateral system. According to this principle, such treatment justified derogations from the general rules because developing countries were by definition in a special situation by virtue of the very fact of being "not yet developed". Thus, while the text of Article 13 contained express limitation as to the scope of the obligations it lays down (considering whether and which measures are appropriate), it did not contain any limitation as to the factual situation to which it applied.

403. The EC argued that no matter how "special" the situation of an exporting country, a determination of dumping must be based on objective and verifiable (and verified) facts. It was only once the investigation had been carried out that special regard must be given to the special situation of developing countries. In addition, it was clear that Article 13 did not create any further obligation on the EC than a "best endeavour". Furthermore, there was nothing special to developing countries in what Brazil did in respect of exchange rates. Such measures could have been taken by any country, whether developed or not and the fact that a developing country adopted such measures does not per se make them "special".

404. The EC further argued that if the developing country status of the exporting country had an impact on the factual situation of the exporters concerned, this would be reflected in the investigation, but only because of the factual situation (for example, due to the particular cost structure), and not because of that country's legal status under the GATT and the Agreement. The EC argued that this was also confirmed by paragraph 1(ii) of the 1980 Decision, which provided for the possibility of taking into account factual elements encountered exclusively or predominantly in developing countries.

405. Brazil agreed with the EC's view that the status of a developing country may be relevant in the context of Article 2:6 to the extent that it impacted on the factual situation of the exporters, but argued that the exchange rate freeze had a direct impact on the factual situation of the exporters: it led to a gross distortion in the comparison between normal value and export price. It should therefore have been taken into account in the determination.

406. Brazil disagreed with the EC's suggestion that reference to the status of Brazil as a developing country would mean that the application of Article 2:6 could not be based on objective and verifiable facts. Brazil argued that the exchange rate (and inflation) were both easily verifiable, and the adjustment sought by Brazil was based on facts far more objective than those which the EC apparently understood Article 13 to refer to. The EC apparently saw Article 13 as coming into play only once the dumping and injury calculations had been completed. It was a kind of "amnesty" effectively given at the sole discretion of the investigating contracting party. Brazil asked how the discretion was to be exercised, and stated that presumably all developing countries had to be treated in a non-discriminatory manner. Brazil asked whether this meant that undertakings applied to all or none within the same proceeding, and also asked what steps would be taken if one country's essential interests were affected but another's were not affected. Brazil also asked what would be done in parallel proceedings (or those proximate in time) for the same (or similar) products. In view of the situations highlighted by its questions, Brazil did not see the EC's approach as likely to lead to an "objective and verifiable" application of Article 13. Brazil argued that the distortion that arose in this case was clear. It could and should have been dealt with in an objective and verifiable way by applying Articles 2:4 or 2:6 in the manner advocated by Brazil.

407. Further addressing the EC's reference to "objective and verifiable" facts, Brazil argued that the principle concern of the contracting parties in this respect was to avoid the abusive application of the Agreement. Brazil argued that it was not advocating an interpretation of the Agreement that could possibly lead to a subjective or abusive application. The kind of adjustment at issue could only ever operate to the advantage of exporters, and this was exactly the sort of tangible adjustment or "special regard" envisaged by Article 13.

408. The EC argued that exchange rate fluctuations or "freezing" were not special to the developing status of Brazil, and that the method used by it in this case permitted a fair comparison under Article 2:6, namely that the comparison be made "in respect of sales made at as nearly as possible the same time". In fulfilling its obligations under Article 2, the EC was under no obligation to have regard to a "freezing" of exchange rates in the manner advocated by Brazil, irrespective of the "specialty" or not of such a "freezing". Also, it was under no obligation to have regard to Article 13 in this specific context, as its obligations under Article 13 were fulfilled elsewhere in the proceedings.

409. The EC argued that if one were to accept Brazil's interpretation of "special situation of developing countries" this would correspond to a particular factual (economic) situation in which a developing country found itself at a given point in time and in the context of specific circumstances, such as those of an anti-dumping investigation. The logical consequence of this was that an investigating authority could legitimately examine such a factual situation, and conclude that it was not "special" in terms of Article 13. This was obviously contrary to the object and purpose of Article 13 and of the special and differential treatment of developing countries. Consideration of the speciality of the situation of a developing country was due in all cases. What was not due in all cases was non-application of anti-dumping measures, or of measures other than anti-dumping duties. In this respect, Article 13 laid down an obligation to "consider" and "explore".

410. The EC further argued that the "special regard ... to the special situation of developing countries" referred to in Article 13 of the Agreement could not be taken to mean that no measures at all, or ineffective measures, should be taken once the existence of dumping and injury caused by developing countries had been established.

(b) Arguments relating to the second sentence of Article 13

411. Brazil recalled the second sentence of Article 13 and argued that the EC had violated the requirement in that sentence because it had failed to properly consider the constructive remedies put forward by the Brazilian exporters. Brazil argued that ABIT and the Brazilian authorities had made a number of constructive suggestions to the EC concerning the allowance that could be made to take into consideration the exchange rate freeze, and the EC was also approached in connection with the possibility of an undertaking being offered by the Brazilian exporters. However, the EC did not consider the suggestions and proposals made in this case.

412. Brazil argued that the correct interpretation of the second sentence of Article 13 was that the obligation for "constructive remedies provided for by this Code" should be explored before the imposition of duties, and that it required investigating authorities to have regard to the problems arising in investigations involving developing countries throughout a proceeding. Where these problems arose, there was an additional obligation imposed on the investigating authorities to pro-actively and constructively seek a solution. However, the EC did not do so in this case and therefore violated Article 13.

413. The EC argued that the condition for the obligation to explore constructive remedies in the second sentence of Article 13 was that the (not "an" or "any") essential interests of developing countries be affected by anti-dumping duties. Hence, the situation foreseen by the second sentence of Article 13 was a serious one: it addressed a global threat to the essential interests of developing countries. The EC argued that such a situation could exist, for instance, if anti-dumping duties were imposed by the country which was almost the only purchaser of a product on which the economy of a developing country was totally dependent. While cotton yarn was probably an important industry in certain regions of Brazil, it could not be claimed that Brazil relied almost exclusively on cotton yarn for its balance of payments, nor that the EC market was its almost exclusive export outlet.

414. The EC further argued that in any event, a second important aspect of the second sentence of Article 13 was that "constructive remedies" had to be found in the Agreement. In other words, they could only consist of price undertakings under Article 7. Moreover, the obligation imposed by the second sentence of Article 13 was only one of exploring possibilities of constructive remedies. There was consequently an obligation to consider the question (obligation de moyens); there was however no obligation to reach an agreement on a constructive remedy (obligation de résultat). The EC argued that it had satisfied these obligations, as could be seen from section M (Undertakings) of the Definitive Regulation. Not only the status of developing country of Brazil had been considered, but also the possibility of constructive remedies provided under the Agreement had been studied, even though no agreement could be found on the form of the undertaking proposed by the Brazilian exporters.

415. The EC argued that in all cases involving developing countries, its normal practice was to explore the possibility of undertakings, rather than to resort to outright imposition of anti-dumping duties. Such undertakings, however, must meet the conditions of Article 10 of Regulation 2423/88, which implemented into EC law the provisions of Article 7 of the Agreement. In this particular case, the Brazilian exporters had offered voluntary quantitative export restrictions, but they were not accepted because the EC was not satisfied that such an undertaking would eliminate the injurious effects of the dumping (section M of the Definitive Regulation, third and fourth subparagraphs). More specifically, the decision not to accept this offer of an undertaking was based on considerations of general policy and on the circumstances of this particular case. In general, the EC was wary of quantitative undertakings, because of the greater risk of trade distortion and trade diversion that they carried. Thus, any such undertakings would be the exception rather than the rule, and any offer of this kind would be scrutinized with particular attention. In this case, however, the key reason for not accepting the offer was that it appeared insufficient to fulfil the purpose of anti-dumping measures, i.e. to remove the injurious effects of dumping, in the light of the fact that the main element of the injury finding vis-a-vis Brazil was the pressure exerted by those imports on prices in the EC.

416. Regarding the timing of the application of Article 13, the EC argued that the basis for the argument in the second sentence of Article 13 was clear by reading the word "they" in "where they would affect" as referring to "anti-dumping duties". This was the only possible interpretation of this provision in view of the terms of this provision (see also section (c) below).

417. Brazil argued that Article 7 of the Agreement, which dealt with price undertakings, may be taken as an important example of an Article of the Agreement where the investigating authorities must have regard to the possibility of constructive remedies. However, Brazil did not consider that the second sentence of Article 13 may be read as referring only to Article 7. There was nothing in the second sentence of Article 13 which may be taken as limiting it in that way. Had that been the intention of the Parties, it would have been a simple matter to refer explicitly to Article 7 in Article 13. The Parties had included numerous cross references in the Agreement, but did not include such a cross reference in the second sentence of Article 13 (see also section (c) below).

418. The EC argued that there was no need for Article 13 to spell out the provisions of the Agreement to which it applied. The plain meaning of the text of Article 13 was obvious. It only applied at the time anti-dumping measures were to be applied. Brazil was unduly stretching the meaning of the term "consideration." No such "consideration" of the application of anti-dumping measures, whether it concerned developing countries or not, could possibly take place before a certain factual basis for such measures had been established (see also section (c) below).

419. In response to a question by the Panel, the EC said that there was no explicit reference, either in the Provisional Determination or in the Definitive Determination to the effect that the status of Brazil as a developing country had been considered. This was for the following reasons. First, all the countries under investigation were developing countries. Given the nature of the obligations under Article 13, this consideration was implicit. Second, this argument was not expressly raised by the respondents which explains also why it was not reported in the determination. Indeed, the arguments raised by the respondents with respect to exchange rates had no relation with the status of Brazil as a developing country. Third, the most concrete obligation of the investigating authority under Article 13 was that of exploring constructive remedies as provided for by the Agreement. As a matter of fact, all that an investigating authority could do to implement its obligation of giving special regard was to consider either not applying duties or consider acceptance of undertakings. In the present case, the EC had considered the proposals for undertakings. Hence, the EC effectively explored the possibilities of constructive remedies, in application of Article 13, second sentence.

(c) Arguments relating to the coverage of Article 13

420. Regarding the phrase "when considering the application of anti-dumping measures under this Code" in the first sentence, Brazil argued that the provisional (and definitive) measures taken by the EC in this case had to be understood as "the application of anti-dumping measures" within the meaning of Article 13 of the Agreement. Brazil noted that the EC member State Customs authorities applied anti-dumping measures from the date on which anti-dumping measures were stated to be in force, and argued that the phrase "when considering the application of anti-dumping measures under this Code" meant that the obligation to have special regard to the special situation of developing countries arose during the period when the investigating authority was considering the application of measures, i.e. before the application of measures actually occurred. There was nothing in Article 13 of the Agreement that would support the view that the obligation only came into operation once a determination on dumping and injury had been made. Therefore, in the absence of any other limitation, it applied throughout the proceeding and at least up to the imposition of measures. Thus, Article 13 was relevant to the application of Article 2, and particularly Articles 2:4 and 2:6, of the Agreement.

421. Brazil argued that the recital showing the general objectives of the Agreement also supported this contention; the recital stated that Parties were "[t]aking into account the particular trade, development and financial needs of developing countries". This matter of general principle was of general relevance to the application of the Agreement, particularly so in the case of the Agreement, since the recitals were remarkably short. Hence, these matters mentioned in the recitals were of special importance to the Parties, i.e. they were the key principles of general application on the basis of which the Agreement was constructed.

422. Brazil argued that the position of Article 13 in the Agreement also supported the view that it was of general application. If Article 13 was intended to deal with an additional substantive requirement that needed to be met before the imposition of measures, of the same nature as the requirements that there be determinations of dumping and material injury, then one would expect to find it at Article 4 of the Agreement, i.e. immediately following Articles 2 and 3, which dealt with the substantive determinations on dumping and injury that must be made before measures are imposed. Article 13 was the final Article in Part I of the Agreement (Parts II and III containing respectively institutional and final provisions). This supported the view that it should be read as a provision of general application to the other provisions of the Agreement. Also, the general nature of the obligations imposed by Article 13 supported the view that it should be considered as of general application in the context of the interpretation of other provisions of the Agreement. The EC's own Anti-dumping Regulation did not interpret Article 13 as imposing a discrete obligation, to be applied in isolation from other provisions of the Agreement.

423. Brazil argued that the provisions of Part IV of the GATT also supported the interpretation of the principle contained in Article 13 as of general application to other provisions of the Agreement. For example, Article XXXVI:l(e) of the GATT provided that: "international trade .... should be governed by .... rules and procedures - and measures in conformity with such rules and procedures .... consistent with the objectives set forth in this Article" Those objectives contained numerous references to the need to have special regard to the situation of developing countries. Furthermore, Article XXXVI required "positive" and "conscious and purposeful efforts" to give effect to these principles and objectives "to the fullest extent possible" (Article XXXVII).

424. Brazil argued that in its Decision of 5 May 1980, the Committee on Anti-Dumping Practices had recognized that the correct interpretation of Article 13 was that it was of general application to other provisions of the Agreement. In that Decision the Committee considered Article 13 relevant in the context of determining the correct method for establishing normal value, i.e. to the application of Article 2:4 of the Agreement. That Decision could not have been reached on the basis that Article 13 imposed a discrete and isolated obligation that only applied after the determination of dumping and-injury. On the contrary, the decision explicitly recognized the relevance of Article 13 at an early stage of the investigation, namely the determination of the method for establishing normal value. The Committee could only have reached that decision on the basis that Article 13 was of general application to other provisions of the Agreement, including Article 2:6.

425. Furthermore, Brazil argued that Part IV of the GATT applied to Article VI of the GATT. In that respect, Brazil noted that Article 1 of the Agreement expressly stated that: "The imposition of an anti-dumping duty is a measure to be taken only under the circumstances provided for in Article VI of the General Agreement ... "

426. From the above analysis, Brazil argued that the obligations imposed by Article 13 of the Agreement were of general application, in the sense that they were relevant in the context of the interpretation and application of all provisions of the Agreement.

427. The EC argued that Brazil was expanding the obligation of Article 13 beyond the stage at which the Agreement set it, i.e. the time where the application of anti-dumping duties was considered. This was supported by the terms of Article 13, taken in their ordinary meaning and in their context.

428. The EC argued that the central word in the language of the first sentence was "measures". The use of the term "measures" clearly limited the obligation of the EC to the stage following the conclusion of the investigation, either preliminary or final. This was confirmed, for example, by the title of Article 10, which referred to provisional measures. The first sentence of Article 10:1 confirmed the fact that this obligation would only arise at the end of the investigation. The first sentence of Article 10:2 defined what provisional measures may consist of (a provisional duty or a security). If the negotiators had wanted to include also earlier stages of the procedure, they would have used a more general wording such as "action", as they did, for instance, in the title of Article 12. Therefore, the obligation of Article 13, first sentence, was applicable to the EC only once the investigation was completed and the EC was about to apply ("considering the application of") measures on the basis of the results of the investigation and not during the investigation of dumping or injury.

429. The EC argued that there was no relationship between Article 13 and other provisions in the Agreement, such as Article 2:6. Article 13 expressly and exclusively referred to the application of anti-dumping measures (first sentence) or, more specifically, of anti-dumping duties (second sentence). Article 2:6 was concerned instead with an important aspect of the calculation of the dumping margin. Such a calculation, in the system of GATT and the Agreement could only be based on an analysis of objective and verifiable facts, and any consideration of the status of a country under GATT and the Agreement could not have any impact on such a factual determination. At the stage of considering the application of measures, and even more so when considering the application of measures or the application of a particular kind of measure (i.e. duties), it may be appropriate to have regard to the status of the country of export.

430. The EC argued that the Committee on Anti-Dumping Practices' Decision of 5 May 1980 did not concern the application of Article 13 to the entire Agreement, but the interpretation and application of the Agreement in its entirety, in the light of considerations similar to those on which Article 13 is based. Moreover, the situations contemplated by the Decision were totally different from those arising in this case. This meant that the Decision had to be read and interpreted on its merits, and not in the light of Article 13, or vice versa.

431. The EC argued that the 1980 Decision was not concerned with an interpretation of Article 13. The initial text of paragraph 1 made it clear that the Decision interpreted the Agreement as a whole "in relation to developing countries." It merely referred to Article 13 to say that the Committee was "cognizant" of the commitment in Article 13, i.e., that the Committee "knows" of that commitment. This appeared even clearer if one looked at the content of the Decision, which dealt with all sorts of matter, including administrative problems in adapting domestic legislation and administrative structures to the Agreement, as well as technical assistance to developing countries.

432. The EC argued that at most one could argue that the 1980 Decision had been taken for the same reasons for which Article 13 had been inserted in the Agreement, i.e., because developing countries were in a "special situation". In that context, paragraph 1 (ii) of the 1980 Decision was another special and differential treatment provision, which did concern the comparison in Article 2 of the Agreement, but did not establish a link between Articles 2 and 13. Article 2 had to be interpreted, as far as developing countries were concerned, inter alia, in the light of paragraph 1 (ii), but not of Article 13.

433. Brazil argued that if Article 13 were to be interpreted in the manner proposed by the EC, it represented no benefit to developing countries and was completely redundant. This was because Article 8:1 of the Agreement already provided that the decision to impose duties and the level of duties were matters to be decided at the discretion of the authorities of the importing country. What did Article 13 add to this provision if the interpretation put forward by the EC was correct ? Brazil considered that provisions of the Agreement should not be interpreted in a way which rendered them effectively empty of meaning.

434. Brazil then argued that while the EC was contending that there was no relationship between Article 13 and Article 2:6, the EC did appear to accept that there was a link between Article 13 and some other provisions of the Agreement (such as mentioned in the context of the arguments relating to the second sentence); presumably the provisions were those which, according to the EC, concerned the "application" of measures. However, the EC had offered no explanation of which provisions were linked to Article 13. Brazil argued that if Article 13 was supposed to be limited in the manner claimed by the EC, it should have contained a cross-reference to other Articles. Therefore, it would be wrong to infer a specific cross reference were none existed, and the second sentence of Article 13 may be understood as a particular, but by no means an exhaustive statement of the obligation arising under the first sentence of Article 13.

(d) Arguments relating to discrimination

435. Brazil noted that it was the only developing country involved in this proceeding which was subject to anti-dumping duties, and that due consideration had been given to the developing status of other countries involved in this proceeding to the extent that they had been excluded from the proceeding on the grounds of de minimis imports or, after recalculation of the dumping margin by taking into account special exchange rates, on the ground of de minimis dumping margin. However, arguments and evidence put forward by Brazil had consistently been ignored or neglected by the EC authorities. Thus, Brazil considered that the EC had not only failed to fulfil its obligations under Article 13 of the Agreement but had also clearly discriminated against Brazil as compared to other developing countries involved in this proceeding.

436. The EC argued that it had infringed neither Article 13 (either the first or the second sentence) nor the 1980 Decision in this case. Brazil could not continue to claim that a freeze of the exchange rate for 90 days (which in fact lasted 180 days if the gradual elimination of the freeze by the Brazilian authorities was also counted) amounted to a temporary fluctuation. In addition, as mentioned earlier (section V.2), dumping was occurring even if one were to follow Brazil's arguments and discount the effect of the freezing of exchange rates. Moreover, there was no evidence to suggest that Brazil had been discriminated compared to the other developing countries. The market share of imports from Brazil was not de minimis, while this was found to be so for the other developing countries. These other countries were excluded from the proceedings on that ground and not because of their developing country status. To do otherwise would have infringed the provisions of the Agreement.

(e) Arguments relating to the alleged refusal of the EC to consider its obligation arising from the MFA

437. Brazil argued that Article 13 of the Agreement and Article 6:1 of the MFA were both expressions of the general principles set out in Part IV of the General Agreement. Therefore, the EC should have had regard to the MFA in determining whether or not the volume of dumped imports could be a factor in a determination of injury, particularly in view of Brazil's status as a developing country. Thus, the EC refused to have regard to Brazil's status as a developing country in so far as it took additional trade measures despite the existence of MFA quantitative restrictions and failed to consider the de minimis nature of Brazil's market share.

438. The EC recalled its arguments made with regard to the relevance of the MFA in the context of injury determination.

VI. FINDINGS

 

VI.1 Preliminary objections of the EC that (1) claims raised by Brazil were outside the terms of reference of the Panel; (2) arguments raised by Brazil that required the Panel to exceed the scope of review

438. The EC raised preliminary objections to certain of the claims and arguments made by Brazil. The preliminary objections were distinguished viz that, (1) claims raised by Brazil were outside the terms of reference of the Panel, and (2) that certain arguments raised by Brazil required the Panel to exceed its scope of review.

(1) Claims raised by Brazil were outside the terms of reference of the Panel

439. The EC argued that the terms of reference were defined by the document prepared by Brazil requesting that the Committee establish a panel to examine the "matter" (document ADP/121). The scope of the "matter" was defined as the sum of claims contained in that document. Therefore, if Brazil sought to raise any claims which were not identified in the document requesting the establishment of a panel, such claims were outside the terms of reference of the Panel.

440. The EC argued that Brazil had sought to raise two "claims" which had not been identified in the document requesting the establishment of a panel. The claims that the EC argued were outside the terms of reference of the Panel were (i) the allegation that the EC violated Article 2:4 of the Agreement on Implementation of Article VI of the General Agreement on Tariffs and Trade by making an "incorrect determination that sales were not made in the ordinary course of trade", and (ii) Brazil's allegations in relation to the EC's consideration of "negative dumping". The EC, however, conceded that not all arguments made in support of claims identified in the document requesting the establishment of a panel had to be identified in the document requesting the establishment of a panel.

(i) The EC violated Article 2:4 of the Agreement on Implementation of Article VI of the General Agreement on Tariffs and Trade by making an "incorrect determination that sales were not made in the ordinary course of trade"

441. The EC raised a preliminary objection with regard to the "claim" by Brazil that the EC violated Article 2:4 of the Agreement by making an "incorrect determination that sales were not made in the ordinary course of trade". The EC argued that the "claim" had not been raised by Brazil in the consultations and conciliation stages of the dispute, nor mentioned in ADP/121 and therefore that the "claim" was not within the terms of reference of the Panel. The EC referred to several panel reports in support of its preliminary objection.

442. In response, Brazil argued that the contention described by the EC as a "claim" was an argument. Brazil argued that its point that the EC had violated Article 2:4 "by incorrectly determining that sales were not made in the ordinary course of trade" was within the terms of reference of the Panel as it was to be found in paragraph 18 of ADP/121. Brazil also noted that the point was to be found in paragraph 16 of ADP/113 and paragraph 2 of ADP/106 (respectively, the documents requesting conciliation and consultation under Article 15 of the Agreement). Brazil argued therefore that the alleged "claim" was clearly within the terms of reference of the Panel.

443. Brazil argued that in this case the Panel should find that ADP/121 expressly included the point, and alternatively, that the relevant test for the Panel was whether ADP/121 could reasonably be interpreted as covering the point. Brazil further argued that the point was raised in consultation when in paragraph 2 of document ADP/106 Brazil stated that the EC had breached Article 2:4 by "failing to consider the particular market situation prevailing in Brazil". Brazil also argued that in any event, ADP/121 and document ADP/113 had provided sufficient notice to the EC or could be reasonably interpreted as covering the issue.

444. The Panel examined first whether the point that the EC violated Article 2:4 of the Agreement by making an "incorrect determination that sales were not made in the ordinary course of trade" could be properly characterized as a claim. The Panel considered that a claim was the specification of the particular legal and factual basis upon which it was alleged that a provision of the Agreement had been breached.

445. The Panel considered, further, that depending on the circumstances, there could be more than one legal basis for alleging a breach of the same provision of the Agreement and that, accordingly, a claim in respect of one of these would not also constitute a claim in respect of the other. A separate and distinct claim would be required.

446. The Panel considered that the absence of sales of the like product in the ordinary course of trade in the domestic market of the exporting country was a particular legally defined situation under Article 2:4 of the Agreement which obliges investigating authorities to determine normal value on the basis of either a constructed value or a comparable price of the like product when exported to any third country.

447. The Panel considered that the point that the EC violated Article 2:4 of the Agreement by making an "incorrect determination that certain domestic sales were not made in the ordinary course of trade" referred to the legally defined situation in paragraph 446 above which was a particular legal and factual basis upon which a breach of the Agreement had been alleged. Therefore, the Panel concluded that Brazil's point that the EC had violated Article 2:4 of the Agreement by making an "incorrect determination that certain domestic sales were not made in the ordinary course of trade" was a separate and distinct claim.

448. The Panel noted that, with regard to Article 2:4, in the course of presentation of its case to the Panel Brazil had made two alternative claims, as follows:

"... the EC ... has violated the following provisions of the Agreement:

- Article 2:4, by failing to consider the particular market situation prevailing in Brazil.

...

- Article 2:4, by incorrectly determining that certain domestic sales were not made in the ordinary course of trade ...".

449. The Panel then examined whether the latter claim, which was the subject of a preliminary objection by the EC, was within its terms of reference.

450. In considering both parties' arguments, the Panel decided that for a claim to be before the Panel, the claim must be within the terms of reference. In this case, the terms of reference are:

"To examine, in the light of the relevant provisions of the Agreement on Implementation of Article VI of the General Agreement, the matter referred to the Committee by Brazil in document ADP/121 and to make such findings as will assist the Committee in making recommendations or in giving rulings."

The Panel decided therefore that for a claim to be before the Panel it would have to be specified in the document requesting establishment of a panel, that document defining the "matter" referred to the Committee by Brazil.

451. The Panel then examined the document requesting the establishment of a panel to determine whether the claim that the EC violated Article 2:4 of the Agreement by making an "incorrect determination that certain domestic sales were not made in the ordinary course of trade" was specified therein.

452. In this regard, the Panel noted that Brazil relied on the first two sentences in paragraph 18 of the document requesting the establishment of a panel to argue that the document stated the claim in issue. The Panel therefore examined paragraph 18 of ADP/121 where Brazil claimed that:

"Normal value for the remaining sales was determined on the basis of cost of production. Brazil considers that this determination equally infringed Article 2:4 of the Agreement. Even though, in the event of a particular market situation, Article 2:4 in principle offers the possibility to establish normal value on the basis of cost of production, in this proceeding the nature of the particular market situation had the effect of making equally unreliable, as a basis for normal value, both domestic sales and cost of production. As a matter of fact, where the particular market situation, as was the case of this proceeding, consists of a combination of high domestic inflation and a freeze of exchange rates, any determination based on domestic data will result in a gross distortion making it impossible to carry out a proper comparison."

 

453. The Panel noted, first, that the express wording in neither paragraph 18 nor any other paragraph of the document requesting the establishment of a panel made any reference to the EC having made an "incorrect determination that certain domestic sales were not in the ordinary course of trade". The Panel noted, further, that in paragraph 18 of the document requesting the establishment of a panel Brazil claimed that "the nature of the particular market situation had the effect of making equally unreliable, as a basis for normal value, both domestic sales and cost of production" (emphasis added by the Panel). This was confirmed by the express wording of the claims summarized in paragraph 10 of the document requesting the establishment of a panel, where Brazil specified that its claim under Article 2:4 was that the EC had breached that Article "by failing to consider the particular market situation prevailing in Brazil" (emphasis added by the Panel).

454. The Panel found therefore that Brazil's claim that the EC had made an "incorrect determination that certain domestic sales were not in the ordinary course of trade" was not expressly referred to anywhere in ADP/121, and accordingly dismissed Brazil's argument on this point.

455. The Panel noted Brazil's alternative argument that even if not expressly mentioned in the text of ADP/121, the claim "can reasonably be interpreted" as being covered by that document.

456. The Panel considered that it was not sufficient that a contention simply "can reasonably be interpreted" as amounting to a claim, as that implied that there could be indeterminacy or ambiguity regarding the ambit of a claim. This would, in the view of the Panel, run counter to the fundamental purpose of the terms of reference, which was to give advance notice to the defendant and to third parties of the claim at issue. This purpose could only be effectively served if there was no ambiguity regarding the ambit of the claim at issue. The Panel considered that, in order to ensure this, a claim had to be expressly referred to in ADP/121 in order to be within its terms of reference. The Panel accordingly also dismissed Brazil's argument on this point.

457. The Panel therefore found that the claim that the EC had made an "incorrect determination that certain domestic sales were not in the ordinary course of trade" was not within the Panel's terms of reference.

(ii) "negative dumping" margins

458. The EC objected to Brazil's contention that in a high inflation environment distortions could occur if an average normal value was compared with individual export prices and if no due allowance, pursuant to the second sentence of Article 2:6, was made for "negative" dumping margins (i.e. those cases in which the export price was higher than the average normal value). The EC argued that this was a "claim" that had not been identified in ADP/121. Brazil had argued that this contention was an argument rather than a claim, and alternatively that it was a claim that had been raised in the document requesting the establishment of a panel.

459. The Panel recalled its view expressed in paragraph 444 above that a claim was the specification of the particular legal and factual basis upon which it was alleged that a provision of the Agreement had been breached.

460. The Panel noted that in ADP/121, Brazil had claimed that the EC had failed to make adjustments that would take into account distortions arising from high domestic inflation combined with fixed exchange rates in comparing normal value and export price. The EC had not made any preliminary objection in relation to that claim.

461. The Panel also recalled that the EC had noted that the terms of reference were defined by the document requesting the establishment of a panel to examine the "matter". The scope of the "matter" was defined as the sum of the claims in that document. Therefore, any claims which were not identified in ADP/121 were outside the terms of reference of the Panel. The EC had conceded that all arguments made in support of claims identified in ADP/121 were not required to be identified in ADP/121. The EC had noted that an "argument" was legal or factual reasoning advanced to support, clarify or explain a claim, or to rebut a claim or an argument.

462. The Panel noted that Brazil had asserted that due to the high inflation environment, the methodology used by the EC (which practice Brazil argued artificially inflated dumping margins), led to a distorted and artificially high dumping margin. Brazil asserted that the dumping margins would have been eliminated if the EC had accepted that the high inflation environment resulted in large differences in the dumping margin from one transaction to another and had made due allowance for "negative dumping". The Panel also recalled that in ADP/121, Brazil had observed that in "a situation where it was impossible for exporters to avoid a dumping finding because of the methodology used by the EC", the EC should have used either third country sales as the basis of normal value or should have made due allowances to take account of distortions in the comparison arising from Brazil's economic policy.

463. The Panel was of the view that the reference in ADP/121 to "the methodology used by the EC that made it impossible for exporters to avoid a dumping finding" was a reference to the EC's practice of so called "zeroing". That reference was developed into an argument that the dumping margins would have been eliminated if the EC had accepted that the high inflation environment resulted in large differences in the dumping margin from one transaction to another, and if the EC had also made due allowance for "negative dumping". The Panel was of the view that Brazil's assertion was legal or factual reasoning advanced to support its claim that the EC had acted inconsistently with Article 2:6 second sentence when it failed to make due allowances to eliminate distortions arising from high domestic inflation combined with fixed exchange rates when comparing normal value and export price. Accordingly, the Panel dismissed this preliminary objection by the EC.

(2) Arguments raised by Brazil that required the Panel to exceed its scope of review

464. The EC argued that the Panel's review of the EC's Determination should be limited to examination of the Determination in order to determine whether the requirements of Article 8:5 had been complied with. The Panel should not conduct a de novo review of the evidence relied on by the EC nor should the Panel substitute its own judgement as to the sufficiency of the evidence relied on by the investigating authorities.

465. The Panel noted that the evidence specifically objected to by the EC was (i) a spreadsheet which allegedly contained chronologically arranged data copied from the EC's disclosure letter to the Brazilian exporters, which spreadsheet had been submitted in support of Brazil's claim of breach of Article 2:6; and (ii) a spreadsheet which set out "dumping margins" recalculated by Brazil.

466. In light of the Panel's resolution of Brazil's claim concerning Article 2:6 of the Agreement, (see paragraphs 484-502 below) it was not necessary for the Panel to determine these preliminary objections by the EC.

VI.2 Breach of Article 2:4 - failure to properly consider the particular market situation prevailing in Brazil

467. The Panel then proceeded to examine Brazil's claim in relation to the EC's determinations of normal value. Brazil claimed that the EC had failed to properly consider the particular market situation prevailing in Brazil and had thereby acted inconsistently with Article 2:4. Brazil advanced several arguments in support of its claim.

468. Brazil noted that in early 1989, due to very high inflation, the Brazilian Government froze the exchange rate at one Cr$ to one US$ in an attempt to decrease the money supply and to control inflation. The exchange rate freeze continued for a period of three months. During this period domestic inflation continued. Receipts from export sales (which were paid in US$), when converted into Cr$, remained stable. Following the unfreezing of the exchange rate, the Cr$ depreciated. Brazil argued that this combination of a fixed exchange rate and domestic inflation led to a gross distortion in the comparison between domestic prices (when used as the basis of normal value) and export prices, and this resulted in an inflated dumping margin.

469. Brazil contended that when the EC determined normal value based on domestic sales prices, the EC had acted inconsistently with the Agreement. The EC had based normal value on domestic sales prices because the EC had misinterpreted the phrase "particular market situation" in Article 2:4 to mean circumstances having an impact only on the sales in the domestic market of the exporting country. On the basis of that interpretation of the phrase "particular market situation" the EC had determined that the exchange rate freeze was irrelevant to the determination of normal value. Brazil argued the phrase "particular market situation" in Article 2:4 was not limited to the situation in the domestic market, but included external factors. Brazil further argued that Article 2:4 was also concerned with ensuring a "proper comparison" with the export price. Brazil was of the view that in selecting the method for establishing normal value, Parties to the Agreement were required to have regard to the question of whether the method selected would permit a proper comparison with the export price. Brazil considered that it was not the object and purpose of the Agreement to apply anti-dumping duties in situations where dumping margins were found solely because of the effect of exchange rate freezes. Brazil said that although the ascertainment of normal value and export price were distinct stages in a dumping investigation, there was still an obligation to consider the circumstances of the export market when determining a method for ascertainment of normal value.

470. Brazil noted that the EC's customary practice was to take circumstances on the export market into account. For instance, the EC examined whether the domestic sales represented less than five per cent of the volume of export sales for the purpose of determining whether domestic sales would permit a proper comparison with export price. Brazil said, therefore, that the EC was inconsistent with its own practice by not considering the circumstances of the export market when determining normal value in this case.

471. In relation to the instances in which the EC had used constructed normal values for Brazilian exporters, Brazil argued that the requirement of "proper comparison" in Article 2:4 also applied to the choice between constructed normal value or third country sales as alternative methods for establishing normal value. Where the external factors were such that a proper comparison with a normal value could not be achieved on the basis of one of the alternative methods set down in Article 2:4, but a proper comparison could be obtained by the use of the other method provided for in that Article, that other method must be used. Brazil said that where the EC had used constructed normal values in this investigation and had then compared those normal values with the export prices it had breached Article 2:4 as this comparison did not "permit a proper comparison".

472. In response, the EC noted that Article 2 provided that the first method for establishing normal value was on the basis of actual domestic sales. The EC argued that Article 2:4 provided that if the "particular market situation" had such an impact on the domestic sales that the domestic sales were not suitable for use as the basis of normal value, one of two equal alternative methods (i.e. constructed normal value or third country sales) could then be used. The EC noted that Brazil had not disputed that the domestic sales used as the basis of normal value were in the ordinary course of trade. It had instead said that the "particular market situation" in Article 2:4 prevented those sales being used as the basis of normal value. The EC argued that the phrase "particular market situation" in Article 2:4, interpreted in light of its object, purpose and context, did not have the meaning attributed to it by Brazil. Article 2:4 was concerned with the establishing of normal value in certain cases. The phrase "particular market situation" was not intended to cover both high inflation in the domestic market and "freezing" of the exchange rate. Moreover, a situation occurring outside Brazil could not be considered to be a "particular market situation" under Article 2:4 unless it had an impact on sales made within Brazil. Therefore, the freezing of Brazil's exchange rate would be relevant to the determination of normal value only if the freezing was shown to have had an impact on domestic sales. In this case the evolution of the exchange rate had no direct bearing on the functioning of the domestic market, where transactions were carried out in the domestic currency.

473. The EC further argued that the Agreement required that the normal value and the export price must first be determined and then be compared. That requirement made clear that the determination of normal value and export price were distinct and separate steps. The EC said that the fact that determination of normal value and export price were distinct and separate steps supported its assertion that a situation having an impact only on the export market (and the export price), need not be taken into account when determining normal value.

474. Regarding Brazil's argument that the EC took into account the volume of export sales as a factor affecting normal values, the EC argued that when it examined whether the domestic sales represented less than five per cent of export sales, it did so to determine whether the volume of sales within the domestic market was so low that the prices were not representative of normal value. If the volume of sales was at a very low level, that low volume would affect domestic sales, and would therefore make those sales in the domestic market unsuitable as the basis of normal value.

475. The EC argued if the negotiators of the Agreement had intended that circumstances having an effect on the export market alone could create a "particular market situation", the negotiators would have made clear that in such a situation third country sales should be used as the basis of normal values. This was because the use of a constructed normal value would not eliminate the effect of such external factors in the export market on the domestic market.

476. The Panel noted that Article 2:1 set out the primary method for determination of the margin of dumping. Article 2:1 provided, in relevant part, that:

" ... a product is to be considered as being dumped ... if the export price of the product ... is less than the comparable price, in the ordinary course of trade, for the like product when destined for consumption in the exporting country."

Article 2:4 further provided that:

"When there are no sales of the like product in the ordinary course of trade in the domestic market of the exporting country or when, because of the particular market situation, such sales do not permit a proper comparison, the margin of dumping shall be determined by comparison with a comparable price of the like product when exported to any third country ... or with the cost of production in the country of origin plus a reasonable amount for administrative, selling and any other costs and for profit ...".

The Panel noted that Article 2:4 provided alternative methods for determining the margin of dumping. Examination of Articles 2:1 and 2:4 together revealed that to determine normal value, the investigating authorities should first examine the sales in the domestic market, in order to establish whether there were sales of the like product in the ordinary course of trade. Article 2:4 required the investigating authorities to use either constructed normal value or third country sales as the basis of normal value if there were no sales in the ordinary course of trade, or if the particular market situation was such that the sales concerned would not permit a proper comparison.

477. The first type of situation requiring that either constructed normal value or third country sales be used as the basis of normal value arose when there were no sales of the like product in the domestic market in the ordinary course of trade in the domestic market. In that first situation, the investigating authorities were only required to examine the domestic market in order to establish whether sales in the ordinary course of trade had taken place in the domestic market. The second type of situation contemplated by Article 2:4 was where sales had taken place in the domestic market in the ordinary course of trade, but, due to the prevailing particular market situation, the use of those sales would not permit a proper comparison.

478. The Panel noted that it was the second situation that was relevant for this claim. The Panel further noted that recourse to use of constructed value or third country sales in this latter situation was governed by whether or not the sales concerned would permit a proper comparison, due to the particular market situation. In the Panel's view, the wording of Article 2:4 made it clear that the test for having any such recourse was not whether or not a "particular market situation" existed per se. A "particular market situation" was only relevant insofar as it had the effect of rendering the sales themselves unfit to permit a proper comparison. In the Panel's view, therefore, Article 2:4 specified that there must be something intrinsic to the nature of the sales themselves that dictates they cannot permit a proper comparison.

479. Even assuming arguendo that an exchange rate was relevant under Article 2:4, it would be necessary, in the Panel's view, to establish that it affects the domestic sales themselves in such a way that they would not permit a proper comparison. Brazil had asserted that exchange rates were capable of affecting domestic sales and prices, because for example, the cost of raw materials could be affected by fluctuations in the exchange rate. In particular, domestic sales and prices could be affected if imported raw materials were used in domestic production. However, Brazil had not argued that the cost of raw materials used in manufacture of cotton yarn were in fact so affected. For the Panel to engage in such an exercise, it would have to exceed its scope of review. The Panel noted also that Brazil had not specified its claim in the form of any argument showing that the prices used as a basis of normal value were themselves so affected by the combination of high domestic inflation and a fixed exchange rate such that those sales did not permit a proper comparison. Accordingly, the Panel concluded on the basis of Brazil's submission, that this did not demonstrate that the EC had acted inconsistently with the requirements of Article 2:4.

480. The Panel recalled that Brazil had also argued that if domestic sales were used as the basis of normal value without either an allowance being made pursuant to Article 2:6 for the distorting effect of the exchange rate, or third country sales being used as the basis of normal value, those domestic sales would "not permit a proper comparison" with export prices (under Article 2:4). The Panel understood Brazil's other argument to be that the term "proper comparison" was modified by the requirement that due allowances be made under Article 2:6 and that the EC should have used third country sales as the basis of normal value.

481. In the Panel's view, the obligation in Article 2:6 became operative only after selection of either constructed normal value or third country sales as the basis from which it was then possible to proceed to the determination of the comparable price. The wording of Article 2:6 second sentence dictates this view: "due allowance" can only be made in respect of a basis that has already been established. The words "due allowance" in Article 2:6 made clear that the requirements of the second sentence could not apply to the choice of the basis itself of a comparable price, which is strictly governed by the requirements of Article 2:4.

482. The Panel noted that Brazil had also argued that the EC's decision to construct certain normal values was also in breach of Article 2:4, because the EC had failed to comply with the phrase "... permit a proper comparison" when deciding whether to use a constructed normal value or third country sales as the basis for normal value. The Panel understood Brazil's argument in this respect to mean that for the EC to comply with the phrase "permit a proper comparison", it had to use third country sales as the basis of normal value. In the Panel's view, Article 2:4 provides for two methods that parties can have recourse to in case where there were no domestic sales in the ordinary course of trade or because of the particular market situation such sales would not permit a proper comparison, i.e. third country sales or constructed normal values. Article 2:4 establishes no legal hierarchy between the two alternative methods. Consequently, taking into account that Brazil contests the EC's recourse to one of the two methods, the Panel concluded that there was no basis for Brazil's argument. As a further consequence the Panel did not need to determine whether the term "proper comparison" was modified by the requirement in Article 2:6 to make due allowances. The Panel considered that it had disposed of Brazil's argument.

483. Accordingly, in light of its conclusions in paragraphs 477-482 the Panel dismissed Brazil's claim that the EC had failed to properly consider the particular market situation prevailing in Brazil and had thereby acted inconsistently with Article 2:4.

VI.3 Violation of Article 2:6 - Failure to take into account distortions arising from high domestic inflation combined with fixed exchange rates in comparing normal value and export price

484. Brazil claimed that the EC had violated Article 2:6 of the Agreement because it had failed to take into account distortions arising from high domestic inflation combined with frozen exchange rates in comparing normal value and export price. Brazil advanced several arguments in support of its claim.

485. Brazil argued that the EC had acted in a manner inconsistent with Article 2:6 of the Agreement because it had failed to make due allowance for differences affecting price comparability arising out of a combination of high inflation and an exchange rate freeze in Brazil. Brazil said that during the first three months of 1989, the Brazilian Government had frozen the exchange rate at 1 Cr$ to 1 US$. At the end of March 1989 the freeze had been lifted, and the Cr$ had depreciated. Brazil argued that in the circumstances of this case, the use of the official exchange rate by the EC in comparing normal value and export price, introduced a gross distortion into the comparison of normal value with export price. The EC had failed to make adjustments to take account of those gross distortions. Brazil said that appropriate adjustments could have been made to the normal value by the use of an alternative (indexed or lagged) exchange rate to the official exchange rate selected by the EC, or by many other methods. In failing to make due allowance pursuant to Article 2:6 to eliminate these distortions, the EC had failed to give effect to the object and purpose of Article 2:6, to effect a fair comparison. In response to an argument by the EC, Brazil argued that the object and purpose of Article 2:6, to effect a fair comparison, was not discharged by meeting the requirements of the first sentence of that Article. Brazil argued that allowances for the other differences affecting price comparability were also required to give effect to the object and purpose of Article 2:6 to effect a fair comparison. If a difference affecting price comparability was not the subject of an allowance, Article 2:6 would be breached.

486. Brazil also argued that the EC's practice of so-called "zeroing" in this case was inconsistent with Article 2:6 in that the EC had failed to make due allowance for a difference affecting price comparability. Brazil was of the view that the application of the EC's so-called "zeroing" methodology was always unfair to an exporter. Brazil was not arguing against zeroing per se. Rather, Brazil argued that in this case the large variations in dumping margins found by the EC were due to severe distortions in the Brazilian financial environment. In the circumstances of a volatile financial environment, so-called "zeroing" produced a distortion which should have been the subject of a due allowance.

487. The EC argued that the first sentence of Article 2:6 defined how a fair comparison was to be effected. If normal value and export price were compared at the same level of trade and in respect of sales made at as nearly as possible at the same time, then a fair comparison would be achieved.

488. The EC considered that the Agreement did not require the investigating authorities to take into account exchange rate fluctuations or freezing as such. In this case, it was considered appropriate to take into account certain effects of exchange rates when fulfilling the obligations created under the first sentence of Article 2:6. This was more particularly the case with respect to the obligation to make a price comparison in respect of sales made at as nearly as possible the same time. The EC first had sought to use actual returns received by exporters. Actual returns were used in the case of one exporter from Brazil. When information on actual returns was not provided by the exporter, or could not be identified, official exchange rates were used. The EC applied end-of-month official exchange rates to the monthly average normal values, mindful of the requirement to compare prices at as nearly as possible the same time.

489. The EC further argued that allowances required to be made under the second sentence of Article 2:6 were limited to objective differences, and would not include differences in exchange rates. These were subjective differences. Consequently, Article 2:6 did not require that exchange rate fluctuations be the subject of allowances as "other differences affecting price comparability".

490. The EC argued that the so-called "zeroing" related to a process of averaging the determined dumping margins for each exporter. There was no obligation under the Agreement to take account of so-called "negative dumping margins". Investigating authorities were only required to consider sales which were at less than normal value.

491. The Panel recalled that Article 2:6 relevantly provides that,

"In order to effect a fair comparison between the export price and the domestic price in the exporting country ... the two prices shall be compared at the same level of trade, normally at the ex-factory level, and in respect of sales made at as nearly as possible at the same time. Due allowance shall be made in each case, on its merits, for the differences in conditions and terms of sale, for the differences in taxation, and for the other differences affecting price comparability."

 

492. In the view of the Panel the argument of Brazil was that the requirement to make due allowance for differences affecting price comparability had to be interpreted in light of the object and purpose of Article 2:6, which was to effect a fair comparison. However, Brazil had not made any independent arguments designed to establish that apart from the requirements of the first sentence, and the allowances required by the second sentence of Article 2:6, there was a further requirement that any comparison of normal value and export price must be "fair". The Panel was of the view that although the object and purpose of Article 2:6 is to effect a fair comparison, the wording of Article 2:6 "[i]n order to effect a fair comparison" made clear that if the requirements of that Article were met, any comparison thus undertaken was deemed to be "fair". The Panel, therefore, considered that its task was to determine whether the action of the EC, in the circumstances of this case, had resulted in a breach of the requirement of the second sentence of Article 2:6, to make due allowances for differences affecting price comparability, in relation to (1) the EC's alleged failure to make due allowances for the effect of the exchange rate freeze combined with high inflation, and (2) the EC's alleged failure to make due allowance for the effects of the application of its so-called "zeroing" methodology.

(1) The EC's alleged failure to make due allowances for the effect of the exchange rate freeze combined with high inflation

493. Brazil claimed that the EC had violated the second sentence of Article 2:6 of the Agreement because it had failed to take into account distortions arising from high domestic inflation combined with fixed exchange rates in comparing normal value and export price. Brazil argued that in the circumstances of this case, the use of the official exchange rate introduced a gross distortion into the comparison of normal value with export price. According to Brazil, the EC, by not making allowances to eliminate the effect of those gross distortions, had failed to comply with the requirements of the second sentence of Article 2:6.

494. Brazil claimed in essence that the EC had violated the second sentence of Article 2:6, by failing to make due allowance for the effect of the exchange rate freeze. In the view of the Panel, the requirement to make due allowances applies only to "differences affecting price comparability". This required that only relevant differences in the factors that affected price determination in the respective markets should be taken into account. The objective of Article 2:6 second sentence is, thus, to establish two accurate prices, the amount of which is expressed in the respective currency of each market and is independent of and prior to the application of the exchange rate. The exchange rate in itself is not a difference affecting price comparability. It is a mere instrument for translating into a common currency prices that have previously been rendered comparable in accordance with the second sentence of Article 2:6. In the view of the Panel, an exchange rate's function is to make it possible to subsequently effect an actual comparison on a common basis as provided under the other relevant provisions of the Agreement. Furthermore, the use of an accurate exchange rate was not, in the Panel's view, a matter of making an "allowance" for a "difference". In the Panel's view the application of an exchange rate did not amount to the making of a compensatory adjustment in accordance with the ordinary meaning of the word "allowance". Application of an exchange rate to those prices is not a matter of making allowance for a difference but would indeed obviate the need to do so precisely because its purpose and effect would eliminate any such difference. Consequently, the Panel dismissed Brazil's argument that the exchange rate freeze was required to be the subject of a due allowance under the second sentence of Article 2:6.

495. The Panel recalled that Brazil had also argued that notwithstanding the EC's use of actual returns or official exchange rates, Article 2:6 second sentence required that the EC make due allowances to eliminate the distortions present in the comparison of normal value with export price which arose because the exchange rate had evolved differently to the exporters' expectations. For the reasons stated in paragraph 494 above, the Panel also dismissed this argument by Brazil.

496. The Panel recalled that Brazil had argued that the provisions of Article 2.4.1 of the Agreement on Implementation of Article VI of the GATT 1994 were relevant to interpretation of Article 2:6 of the Agreement on Implementation of Article VI of the GATT 1947. However the Panel did not agree that analysis of Article 2.4.1 was of assistance in the interpretation of Article 2:6. The Panel recalled that its terms of reference were to:

"examine, in light of the relevant provisions of the Agreement on Implementation of Article VI of the General Agreement, the matter referred to the Committee by Brazil in document ADP/121 and to make such findings as will assist the Committee in making recommendations on giving rulings."

The Agreement mentioned in the terms of reference was the Agreement on Implementation of Article VI of the General Agreement 1947. In the view of the Panel the Agreement on Implementation of Article VI of the GATT 1994 does not fall under any of the categories specified in Articles 30-32 of the Vienna Convention on the Law of Treaties and is, consequently, irrelevant for the interpretation of the Agreement on Implementation of Article VI of the GATT 1947 for the purposes of resolving the present dispute. Accordingly, the Panel dismissed this argument by Brazil.

497. In addition, Brazil had argued that the practices of Australia, Canada and the United States in relation to exchange rates were relevant to interpretation of the obligations contained in Article 2:6. The Panel did not agree. The practices of three of the total signatories to an Agreement did not constitute subsequent practice in the application of the treaty in accordance with Article 31(3)(b) of the Vienna Convention on the Law of Treaties. Accordingly, the Panel dismissed this argument by Brazil.

(2) The EC's alleged failure to make due allowance for the effects of the application of its so-called "zeroing" methodology

498. Brazil argued that in an environment of high inflation the EC's so-called "zeroing" methodology had an especially prejudicial effect on the calculation of the dumping margin. Brazil argued that in a high inflation environment, high positive and negative dumping margins would occur. As the EC's so-called "zeroing" methodology gave no "credit" for negative dumping, the result was a particularly prejudicial effect on the dumping margin. Brazil argued that even if so-called "zeroing" could be defended in most circumstances, it could not be defended in cases where due to high inflation very high fluctuations in positive and negative dumping margins occurred. Brazil argued that the EC's failure to make a due allowance to take account of the effect of the EC's so-called "zeroing" methodology when applied to Brazil's inflationary circumstances resulted in the EC acting inconsistently with Article 2:6.

499. The Panel understood Brazil's argument in relation to so-called "zeroing" to be that the EC was under an obligation to make a due allowance for a difference affecting price comparability, because application of the EC's system of averaging would tend to produce an exaggerated dumping margin. The Panel understood that the averaging practice operated as follows. The EC established an average normal value for each exporter. It then compared the export price for individual export transactions during that period to that average normal value. When the comparison revealed that the export price was less than the normal value, the EC considered that the exported goods were dumped. When the comparison of the average normal value and the export price revealed that the export price was equal to or greater than the normal value the EC considered that dumping did not occur. After comparison of all export transactions with the average normal value the EC then calculated a weighted average margin of dumping for the exporter by totalling the amount of dumping for the individual export transactions and dividing this total by the c.i.f. value of all the export sales of that exporter.

500. The Panel noted that the first sentence of Article 2:6 concerned the actual comparison of prices at the same level of trade and in respect of sales made at as nearly as possible the same time. The Panel considered that the second sentence of Article 2:6 concerned allowances to be made for the relevant differences in the factors that affected price determination in the respective markets sufficient to ensure the required comparability of prices. The Panel took the view that the second sentence of Article 2:6 required that allowances necessary to ensure price comparability be made prior to the actual comparison of the prices, in order to eliminate differences which could affect the subsequent comparison.

501. The Panel considered that "zeroing" did not arise at the point at which the actual determination of the relevant prices was undertaken pursuant to the second sentence of Article 2:6. In the Panel's view, "zeroing" was undertaken subsequently to the making of allowances necessary to ensure price comparability in accordance with the obligation contained in second sentence of Article 2:6. It related to the subsequent stage of actual comparison of prices; a stage which was not governed by the second sentence of Article 2:6. Therefore, the Panel dismissed Brazil's argument that the EC failed to make due allowances for the effects of its so-called "zeroing" methodology.

502. Based on its conclusions in paragraphs 499, 500 and 501 the Panel dismissed Brazil's claim that the EC had violated Article 2:6 of the Agreement because it had failed to take into account distortions arising from high domestic inflation combined with frozen exchange rates in comparing normal value and export price.

VI.4 Articles 3:1 and 3:2, not basing the injury findings on positive evidence, and not making an objective examination of the relevant facts

503. The Panel then turned to examine Brazil's claim that the EC had breached its obligations under Articles 3:1 and 3:2 by not basing its findings of injury on positive evidence and not making an objective examination of the relevant facts. The Panel noted that Brazil had made several arguments in support of its claim.

504. Brazil argued that the data used by the EC (hereinafter referred to as "Eurostat") as evidence of import volume was unreliable, and thereby did not represent positive evidence of import volume. Brazil made two arguments in support of its claim. The first argument was that in the course of the anti-dumping investigation "... on imports of certain polyester yarns (man made staple fibres) originating in Taiwan, Indonesia, India, the People's Republic of China and Turkey, (published in EEC Regulation no. 830/1992), ("the polyester yarns case") Eurostat data had been shown to be unreliable. Brazil's second argument was that Brazilian export data (hereinafter referred to as "Cacex"), compiled under strict surveillance pursuant to the Arrangement Regarding International Trade in Textiles, showed a much lower volume of exports to the EC than Eurostat data during the years under investigation. Brazil argued therefore that if Eurostat data had been reliable, Brazil's imports into the EC would have breached Brazil's quota under the MFA in 1987 and 1988. Brazil observed that the MFA treaty's "flexibilities" provisions (which permitted a country to exceed its quota) could not be used unless notifications were made. Brazil argued that it had not made such notifications, nor could it have used the "flexibilities" provisions, as its imports would have exceeded the amount permitted under the "flexibilities" provisions. The EC had not complained that Brazil had exceeded its MFA quota. Brazil argued, therefore, that the EC must have implicitly accepted that the Cacex data was more reliable. Therefore, Cacex data must have been more reliable than Eurostat.

505. Brazil made a consequent argument that if Cacex data had been used as the basis of import volume, Brazil would have been found to have had de minimis import volume. Accordingly, the only appropriate outcome would have been termination of the investigation in respect of Brazil.

506. The EC argued that Brazil's arguments in relation to the use of the Eurostat data constituted a request for a de novo review of part of the investigation, because Brazil's argument required an assessment of the reliability of two sets of data. It further submitted that the polyester yarns case was distinguished from the present case. The polyester yarns case concerned a product which was capable of being classified under four different tariff sub-headings, but the investigation did not cover all the goods in the four different tariff sub-headings. In that case some imports (not covered by the investigation) had been misclassified into tariff sub-headings covered by the investigation. In this case all cotton yarn imports were covered by the investigation, and therefore a similar type of error could not occur. Moreover, in the polyester yarns case the exporters had presented proof which established that Eurostat data was unreliable, whereas in this case the Brazilian exporters had merely made unsubstantiated assertions concerning Eurostat data.

507. The EC also argued that Eurostat data was inherently more reliable than Cacex data, as Eurostat data was import data, based on goods entered for consumption into the EC. Cacex data was export data based on bills of lading, which was at best proof of exportation, and not proof of the goods having arrived at the declared final destination. Another possible explanation for differences in the two figures was that goods could have been stocked in a third country or warehouse prior to being entered into the EC for consumption. The EC argued that in any event, the accuracy of Eurostat data was regularly verified. Monthly reports of the volumes and values of imports were checked by comparison of a monthly unit value of a particular product from a particular origin with the unit value calculated on the same basis for the previous month. If substantial differences between the two sets of monthly data emerged, Customs authorities were automatically alerted and verifications of the entries were conducted. In the case of cotton yarn from Brazil no such discrepancies emerged. The EC also noted that no submissions had been made during the investigation to show that a similar type of error had occurred.

508. The EC further argued that if Brazil's exports had exceeded the MFA, no inference could be drawn from the EC not asserting its rights under the MFA. The EC argued that in any event, Brazil had used the MFA's "flexibilities" provisions permitting it to "carry over" quota excesses and anticipate the following years' entitlements, and tendered to the Panel documents it described as notifications of the use of the "flexibilities" provisions. The EC argued that Brazil's imports did not exceed the allowable amounts under the MFA. The Panel noted that Brazil had responded to the Panel that the documents submitted to the Panel by the EC as proof of the use of the "flexibilities" provisions related largely to regional, i.e. within the EC, flexibility. Brazil said that the documents tendered by the EC did not contradict its arguments.

509. The Panel noted that the claim by Brazil related to Eurostat data used by the EC as the basis of import volume. Brazil claimed that the EC had not based its findings of import volume on positive evidence and had not conducted an objective examination of the relevant facts.

510. The Panel understood that Brazil's argument was that Eurostat data should not have been used because it was unreliable. Brazil argued that the EC should have based its findings of import volume from Brazil on Cacex export data. If the EC had used Cacex data a lower volume of imports would have been found. A consequence of using Cacex data would have been that the lower volume of imports would have been found to be de minimis and this would have required Brazilian imports to be excluded from the investigation. The Panel recalled that Brazil had stated that "... all the evidence supports the view that the Cacex figures are more accurate than the use of Eurostat figures...". Brazil claimed that the failure by the EC to base its findings on positive evidence and an objective examination entailed also a breach of Article 3:2. The Panel understood this to mean that as the requirement of positive evidence had not been complied with, the EC could not have properly made findings concerning "effects" of the dumped imports pursuant to Article 3:2. The Panel examined Brazil's claim on that basis.

511. The Panel's starting point was to examine Article 3:1, which relevantly provides;

"A determination of injury for purposes of Article VI of the General Agreement shall be based on positive evidence and involve an objective examination of both (a) the volume of the dumped imports ...".

512. The Panel noted that all the arguments made by Brazil concerned the accuracy of the Eurostat data. The Panel therefore examined whether Brazil had established that the Eurostat data was unreliable or incorrect. The Panel considered that if a finding of injury was based on incorrect facts it would not be based on "positive evidence". The phrase "positive evidence" required at least that the evidence upon which a finding of injury was based must not be incorrect.

513. The Panel noted that Brazil's arguments concerning the incorrectness of Eurostat data related to differences between Cacex and Eurostat data. Cacex data was export data. Eurostat data was import data. Articles 3:1 and 3:2 required the investigating authorities to consider the effects of the dumped imports. Nowhere in Article 3 were investigating authorities required to consider exports. The Panel did not rule out that export data could be a relevant consideration for doubting the accuracy of import data; this however, could not happen in the absence of other evidence. The Panel noted that Brazil failed to provide any basis for considering that the import data was inaccurate other than the mere existence of its own export data which differed from the import data of the EC. Accordingly, the Panel dismissed Brazil's arguments based merely on the differences between Eurostat import data and Cacex export data. The Panel noted that Brazil had made submissions to the EC regarding the accuracy of its import data. In recital 38 of the Definitive Regulation the EC appeared to have examined those submissions. In the absence of further arguments by Brazil the Panel concluded that it had not been established that the EC had failed to conduct an objective examination in this respect.

514. The Panel recalled that Brazil had argued that the EC had implicitly accepted the accuracy of Cacex data because the EC had not chosen to enforce its rights pursuant to the MFA. Even if Brazil was correct, the EC's decision not to complain about Brazil exceeding the quota did not prove that Eurostat data was incorrect. The Panel noted that the EC in fact argued that Brazil had not exceeded the quotas because it had utilized the "flexibilities" provisions of the MFA. In any case, the Panel considered that it could not be simply presumed that the reason for the EC deciding not to exercise its rights under the MFA was that it did so because it accepted the export data of Brazil. In the absence of any specific arguments necessary to establish the latter argument the Panel dismissed Brazil's argument.

515. The Panel noted that Brazil had also argued that in the polyester yarns investigation Eurostat data had been determined to be inaccurate. The Panel first noted that the EC had explained that the polyester yarns case concerned only certain types of yarn, and that types of yarn not covered by the investigation had mistakenly been classified to the same tariff items as the product under investigation. In the cotton yarn case, all types of cotton yarn were covered by the investigation, which removed the risk of the same type of error occurring. Brazil had not argued against this explanation by the EC. It, therefore, appeared to the Panel that the argument raised by Brazil based on the errors encountered in the polyester yarns case was not relevant to this case.

516. Following its conclusions in paragraphs 512-515, the Panel considered that it had disposed of Brazil's claim. Accordingly, the Panel concluded that Brazil had failed to establish that the EC had breached its obligations in Articles 3:1 and 3:2 to base its findings of injury on positive evidence and to make an objective examination of all relevant facts and dismissed Brazil's claim.

VI.5 Violation of Article 3:2, 3:3 and 3:4 - "... the facts stated in an injury analysis must reasonably support the determination that is reached"

517. Brazil claimed that the EC had violated Articles 3:2, 3:3 and 3:4, because the facts stated in an injury analysis must reasonably support the determination that is reached, and the EC's decision was not in conformity with that requirement.

518. The Panel noted that Brazil particularised its claim as that the EC had found Brazil's imports of cotton yarn to be causing injury despite the Brazilian imports of cotton yarn having the lowest level of price undercutting, a finding of a decrease in the volume of Brazilian imports of cotton yarn (in absolute and relative terms), and the fact that the export price in US$ of Brazilian exports was comparatively stable, and actually increased in 1989. Brazil argued that the EC had made a manifestly erroneous interpretation of the facts, in that these facts precluded a finding of material injury caused by Brazilian imports.

519. Brazil also argued that when the EC had conducted its analysis of a causal link between Brazilian dumped imports and the injury suffered by the EC industry, the EC had violated Article 3:4 because it had failed to reasonably explain how other causes of injury, particularly numerous non-dumped imports from countries not under investigation had been taken into account. Brazil noted that the EC had determined that the investigation did not reveal any factors other than the dumped imports which caused material injury to the Community industry. Brazil noted that the EC's findings clearly showed the existence of numerous non-dumped imports from countries not under investigation which undercut the Community producer's prices.

520. The EC argued that Articles 3:2 and 3:3 permitted affirmative injury findings to be made even if not all the factors required to be considered were found to be present. The EC argued that it had considered all the factors that it was required to by Articles 3:2 and 3:3. The EC referred the Panel to its findings concerning the volume and price effects of Brazilian imports, the injury suffered by the EC industry, and its finding of a causal link between the Brazilian imports and the injury suffered by the EC industry. The EC further noted that Brazil had characterised this claim as a failure to make an objective examination of the relevant facts. Irrespective of how Brazil's claim could be characterized the EC argued that its findings properly explained its injury determination.

521. The Panel firstly noted that Brazil's claim was not based on Article 3:1 of the Agreement. Brazil particularized its claim as follows: Brazil had the lowest level of price undercutting, that dumped imports had decreased in volume (in absolute and relative terms), that Brazil's export prices were stable during the investigation period, and that the EC's findings clearly showed the existence of numerous non-dumped imports from other countries not under investigation which undercut the Community producers prices.

522. In the Panel's view Brazil's arguments did not suggest that the EC had not considered all those matters it was legally required to. Rather, Brazil argued that the facts stated in the injury finding did not reasonably support the injury finding.

523. Therefore, the Panel next proceeded to analyse Brazil's arguments.

524. Brazil argued that Brazilian imports had the lowest level of price undercutting, and because of that the facts stated in the injury analysis did not reasonably support the finding made. The Panel noted that in the Provisional Determination (recital 32) the EC found price undercutting for Brazilian imports of between 2.6 and 7.5 per cent. This was confirmed in the Definitive Determination (recital 42). The Panel noted that in recital 47 of the Definitive Determination the EC had found price depression and price undercutting on the EC market caused by the dumped imports. The Panel recalled that Brazil had not argued that the EC's findings of price depression and price undercutting were not properly reached. In the Panel's view Brazil had argued that, as its imports had the lowest level of undercutting, the EC was effectively precluded from finding that Brazil's imports caused injury. The Panel observed that Article 3:4 made clear that it should be read in conjunction with Articles 3:2 and 3:3. Articles 3:2, 3:3 and 3:4 revealed that in determining whether the dumped imports had caused material injury the investigating authorities were required to examine whether "effects" as defined in Articles 3:2 and 3:3 were present. The Panel also noted that Articles 3:2 and 3:3 provided that no "one or several of these factors can necessarily give decisive guidance." The Panel considered that the provisions made clear that not all of the types of effects described in Articles 3:2 and 3:3 need to be present in order for a finding to be made that the dumped imports caused material injury pursuant to Article 3:4. The Panel was of the view that provided that "effects" (whether volume or price effects, or both) as provided for in Article 3:2 and relevant economic factors showing that industry was suffering material injury as provided in Article 3:3, were present, and those "effects" were found to have caused the material injury suffered by the local industry per Article 3:4, the investigating authorities were properly entitled to make a finding that the dumped imports had caused material injury. From examination of Articles 3:2 and 3:4, it appeared clear to the Panel that a finding of an increase in volume of dumped imports by the investigating authorities was not necessary for them to make a finding of material injury caused by dumped imports. In addition, the Panel could not derive from its interpretation of Article 3:4 a requirement that if imports from a particular source had the lowest level of price effects a finding of injury in relation to those imports was precluded. The Panel concluded that whether or not Brazilian imports had increased in volume and whether or not Brazilian imports had the lowest level of price "effects" pursuant to Article 3:2 did not establish that the EC's finding was inconsistent with the Agreement. Therefore, the Panel decided to dismiss Brazil's argument.

525. The Panel then turned to Brazil's argument that its dumped imports had decreased in volume, and due to that the injury finding was not reasonably supported by the facts. The Panel recalled that in paragraph 524 it had noted that Articles 3:2 and 3:3 provided that no "one or several of these factors can necessarily give decisive guidance". In recital 29 of the Provisional Determination the EC determined that there had been an overall increase in the volume of cumulated dumped imports (from Brazil, Turkey and Egypt) of 8 per cent over the period 1986-1989. The EC found (in recital 30) that the market share of the cumulated dumped imports increased from 11.4 per cent in 1986 to 12.1 per cent in 1989. In the Definitive Determination, the EC had eliminated Egyptian cotton yarn from the cumulated imports. Following the elimination of Egyptian imports, in recital 41 of the Definitive Regulation the EC had found that the market share of the remaining cumulated dumped imports was 9.6 per cent in 1986 decreasing to 8.7 per cent in 1989. The Panel was of the view that provided that one "factor" or type of price effect was established it was not necessary that an increase in volume also be found before the investigating authority could establish whether the dumped imports had caused material injury to the EC industry. Equally therefore, it also followed from the Panel's analysis of Articles 3:2, 3:3 and 3:4 that the presence of a decrease in volume did not of itself preclude a finding that the dumped imports had caused material injury to the EC industry. Accordingly, the Panel decided to dismiss this argument by Brazil. However, the Panel noted that in responding to a question by the Panel the EC had stated "Regarding the volume to be considered for injury purposes, the Community took into account all imports, whether dumped or non-dumped, for the reasons mentioned above". Articles 3:2 and 3:4 of the Agreement required that the investigating authorities examine the volume and effects of the "dumped" imports. The Panel noted that the EC stated in its response that it had, for the purposes of its injury analysis, taken into account the effects of all imports from Brazil, whether dumped or non-dumped. As Brazil had not made a claim that the EC had thereby acted inconsistently with the Agreement, the Panel could not pronounce itself on any such claim.

526. The Panel next turned to Brazil's argument that Brazil's export prices were stable during the investigation period, and because of that the facts stated in the injury analysis did not reasonably support the finding made. The Panel recalled that in recital 51 of the Definitive Determination the EC had based its finding of injury caused by the dumped imports on price undercutting. The Panel considered that stable prices could be consistent with the establishment of price effects in accordance with Article 3:2 based on price undercutting. Recalling its reasoning in paragraph 524 above, the Panel considered that Brazil's argument was not inconsistent with a finding of material injury caused by the price undercutting. Accordingly, the Panel also dismissed this argument by Brazil.

527. The Panel next turned to Brazil's argument that the EC's findings clearly showed the existence of numerous non-dumped imports from countries not under investigation which undercut the Community producers' prices, and due to the existence of those non-dumped imports a finding of injury was not reasonably supported by the facts.

528. The Panel recalled that Article 3:4 of the Agreement set out the legal requirements for consideration of a causal link between material injury and the dumped imports, including whether "other factors" were responsible for the material injury suffered by the domestic industry. The Panel noted that the second sentence of Article 3:4 provided that

"...[t]here may be other factors5 which at the same time are injuring the industry, and the injuries caused by other factors must not be attributed to the dumped imports."

Footnote 5 provided "[s]uch factors include, inter alia, the volume and prices of imports not sold at dumping prices...". Therefore, the second sentence required that if there were other imports being sold at non-dumped prices, which were injuring the industry, the injury caused by the dumped imports should not be attributed to those imports.

529. In recital 44 of the Provisional Determination the EC had found that there was no evidence of any significant market disruption relating to imports from the only other significant importer not subject to the investigation (Switzerland). The Panel recalled that in this context, Brazil had also not argued that there were sources of non-dumped imports other than those identified by the EC, nor had it argued that Brazil was not responsible for the injury attributed to it by the EC. The Panel noted that in recital 46 of the Provisional Determination the EC had concluded that the material injury suffered by the EC industry was not caused by any "other factors", and was only caused by the dumped imports.

530. From examination of recitals 44 and 46 of the Provisional Determination it appeared clear to the Panel that the EC had examined whether injury suffered by the EC industry had been caused by "other factors", as required by Article 3:4. The Panel, therefore, dismissed Brazil's argument related to numerous non-dumped imports.

531. In light of its conclusions in paragraphs 524, 525, 526, 529 and 530, the Panel dismissed Brazil's claim that the EC had violated Articles 3:2, 3:3 and 3:4.

VI.6 Articles 3:2, 3:3 and 3:4 - Quotas agreed under the MFA precluded a finding of injury

532. Brazil claimed that the Arrangement Regarding International Trade in Textiles precluded a finding of injury under Articles 3:2, 3:3 and 3:4.

533. Brazil argued that the obligations under the MFA were required to be taken into account by the EC in conducting its injury investigation pursuant to Article 3 of the Anti-Dumping Agreement. The Definitive Determination did not explain why it was not possible to refrain from taking anti-dumping measures in this case, as required by Article 9:1 of the MFA. The EC had not had proper regard to the textile quotas under the MFA when conducting its injury investigation. The EC had merely stated that quantitative restrictions could not prevent injury arising from dumped imports at low prices. Brazil, therefore, argued that the EC's injury investigation failed to have regard to the obligations under the MFA, and consequently the EC had violated the obligations imposed by Article 3 of the Anti-Dumping Agreement.

534. Brazil further argued that the MFA modified Articles 3:2, 3:3 and 3:4 of the Anti-Dumping Agreement, insofar as it applied to anti-dumping investigations conducted by the EC in respect of products exported from Brazil falling under the Bilateral Agreement. Brazil argued that trade in textiles between Brazil and the EC was regulated by bilateral agreements negotiated under the MFA. Brazil referred the Panel to Article 9:1 of the MFA which provided:

"In view of the safeguards provided for in these Arrangements, the participating countries shall, as far as possible, refrain from taking additional trade measures which may have the effect of nullifying the objectives of this Arrangement."

Brazil therefore argued that any trade disruption arising from trade in textiles should be settled within the framework of the special procedures provided in the MFA.

535. The EC argued that Brazil's first argument was inadmissible to the present proceedings. This was because the argument related to an alleged violation of an agreement other than the Anti-Dumping Agreement, and because the MFA had its own dispute settlement procedures, which was the proper forum for an allegation of a breach of the MFA. The EC also argued that Article 9:1 did not preclude a finding of injury, it merely imposed a "best efforts" obligation, when it provided that investigating authorities should "as far as possible, refrain" from taking anti-dumping action. The EC also argued that in any case it had fulfilled the relevant obligations of the MFA. This was because in recital 49 of the Definitive Determination, the EC had determined that although the MFA quotas protected the EC industry from excessive volumes of imports, they did not protect the EC industry from dumped imports at very low prices.

536. The EC also noted that footnote 14 to Article 15 of the Anti-Dumping Agreement provides;

"[i]f disputes arise between Parties relating to rights and obligations under this Agreement, Parties should complete the dispute settlement procedures under this Agreement before availing themselves of any rights which they have under the GATT."

The EC argued that a fortiori it should also not be possible for a party to both the MFA and the Anti-Dumping Agreement to invoke a provision of the MFA before a panel under the Anti-Dumping Agreement. A claim of breach of the MFA should be determined under the dispute settlement provisions of the MFA.

537. Concerning the second argument made by Brazil, the EC argued that the MFA did not modify the Anti-Dumping Agreement. The EC argued that Article 30 of the Vienna Convention on the Law of Treaties dealt with the application of successive treaties relating to the same subject matter. Article 30 provided that in the case of treaties relating to the same subject matter, the provisions of an earlier treaty apply between the parties to the earlier and later treaty, only to the extent that the provisions of the earlier treaty were compatible with the later treaty. The EC argued that an essential precondition for application of Article 30 was that the two treaties should relate to the same subject matter, not merely that the earlier treaty should be affected in some way by the provisions of the later treaty. In this case the two treaties did not relate to the same subject matter. Therefore, the MFA did not modify the Anti-Dumping Agreement.

538. The Panel noted that in support of its claim, Brazil had argued, firstly, that the MFA's obligations should have been taken into account by the EC when analysing injury to the EC industry, and, secondly, that the MFA modified the requirements of Articles 3:2, 3:3 and 3:4 of the Anti-Dumping Agreement. The Panel proceeded to deal with the first argument.

539. The Panel recalled that it had been established by the Committee in order to:

"examine, in the light of the relevant provisions of the Agreement on Implementation of Article VI of the General Agreement, the matter referred to the Committee by Brazil in document ADP/121 and to make such findings as will assist the Committee in making recommendations or in giving rulings".

The Panel noted that Brazil alleged that the EC failed to have regard to obligations under the MFA in conducting its injury investigation. Brazil's argument, therefore, depended upon an interpretation of the MFA and its obligations. The Panel recalled that it was established to examine the matter referred to the Committee by Brazil in light of the relevant provisions of the Anti-Dumping Agreement. As this argument in support of Brazil's claim depended on examining the matter referred to the Committee by Brazil in light of obligations under the MFA, it appeared that it was, therefore, not within the Panel's terms of reference to determine whether obligations under the MFA were required to be taken into account by the EC. The Panel, therefore, decided that the argument by Brazil that the EC failed to have regard to obligations under the MFA was outside its terms of reference.

540. Brazil had also argued that the MFA modified the requirements of Articles 3:2, 3:3 and 3:4 of the Anti-Dumping Agreement. The Panel noted that Article 30 of the Vienna Convention on the Law of Treaties dealt with the application of successive treaties relating to the same subject matter. A precondition for the application of Article 30 of the Vienna Convention on the Law of Treaties was that the two treaties relate to the same subject matter. The Panel was of the view that when Article 9:1 of the MFA spoke of "additional trade measures" it included anti-dumping measures. For the purposes of analysing this argument by Brazil, the Panel assumed (but did not find it necessary to decide) that the Anti-Dumping Agreement and the MFA were treaties "relating to the same subject-matter".

541. The Panel was of the view that Article 30 of the Vienna Convention on the Law of Treaties assumes that obligations created by successive treaties relating to the same subject matter would be incompatible obligations. Article 9:1 of the MFA provides:

"In view of the safeguards provided for in this Arrangement the participating countries shall, as far as possible, refrain from taking additional trade measures which may have the effect of nullifying the objectives of this Arrangement."

From examination of this Article, the Panel noted that the obligation contained therein was not mandatory in nature. It did not establish a prohibition on the use of other "trade measures". It simply provided that because of the safeguards established by the MFA (i.e. quantitative restrictions), participating countries shall as far as possible refrain from taking additional trade measures which may have the effect of nullifying the objectives of the MFA. Accordingly, Article 9:1 of the MFA did not as a matter of law prohibit the taking of additional trade measures.

542. The Panel recalled that Brazil's claim was that the MFA precluded a finding of injury under Articles 3:2, 3:3 and 3:4. The Panel noted that Articles 3:2, 3:3 and 3:4 dealt with the determination of injury. Article 3:2 dealt with the establishment of the volume or price effects of the dumped imports, Article 3:3 dealt with the state of the industry in the importing country, and Article 3:4 dealt with the establishment of a causal link between the effect of the dumped imports and the injury suffered by the domestic industry. The Panel recalled that in paragraph 524 of this report it had determined that it was not necessary that both price and volume effects be established in relation to the dumped imports. The establishment of either price or volume effects was sufficient for a finding that the dumped imports had effects pursuant to Article 3:2.

543. The Panel noted that in recital 32 of the Provisional Determination the EC had found that the dumped imports from Brazil had undercut the prices of the Community industry. That finding was confirmed in recital 42 of the Definitive Determination. The EC had, therefore, determined that the imports of cotton yarn from Brazil had had a price effect pursuant to Article 3:2 of the Agreement on Implementation of Article VI of the GATT 1947. The EC had not determined that the imports of cotton yarn from Brazil had had volume effects pursuant to Article 3:2.

544. The Panel, therefore, dismissed Brazil's argument that the EC was precluded from making a determination of injury with respect to products falling within volume quotas under the MFA. In view of its conclusions reached above, the Panel dismissed Brazil's claim that the EC had breached Articles 3:2, 3:3 and 3:4 of the Agreement, because the MFA precluded a finding of injury.

VI.7 Article 3:2 in combination with Article 8:2 - Discrimination against Brazilian imports

545. The Panel then turned to examine Brazil's claim that the EC had violated Article 3:2 in combination with Article 8:2 because it had discriminated against products imported into the EC from Brazil. In support of its claim Brazil made several arguments.

546. Brazil argued that the EC had treated Brazil differently from "other exporters", and had thereby discriminated against Brazil. Brazil said that the EC had sampled Brazil's imports but had considered all Brazil's imports to be dumped, despite the EC having established that 18.36 per cent of Brazil's imports were not dumped. Although other exporters' (particularly India and Thailand) import volumes were sampled, the EC had deducted the non-dumped imports from total import volumes for the other exporters, but had not done so for Brazilian imports into the EC. On the basis of those adjusted total import volumes, India and Thailand were considered to be of de minimis import volume, because their imports could not have significantly contributed to injury. Accordingly, the EC had not added the volume of India and Thailand's imports to those of the other importers subject to the investigation (including Brazil) for the purpose of conducting a cumulative analysis of the effect of all the dumped imports on the EC industry.

547. Brazil argued that if it had been treated in the same manner as India and Thailand, its imports would not have been cumulated, because Brazil's import volume would have been within the EC's usual threshold for de minimis volume. Therefore, if it had been treated in the same manner as India and Thailand, Brazil's imports would have been excluded from the investigation as being of de minimis volume. Brazil also argued that the price effects of the imports excluded from the investigation were more significant than the price effects of the Brazilian imports.

548. Brazil argued that in so discriminating against Brazil, the EC had violated an obligation arising from Article 8:2 that duties should be both imposed and collected in a non-discriminatory manner. In support of its argument that Article 8:2 was the source of such an obligation, Brazil argued further that the obligation contained in Article 8:2 to not discriminate was not limited to the collection of duties, and applied to the decision to impose duties.

549. Brazil also contended that it also followed from the arguments set out in paragraphs 546 and 547 that the Panel should conclude that the EC's determination was in violation of Article I of the GATT, as it was discriminatory.

550. Brazil also argued that decisions whether imports from a particular country were of negligible volume (pursuant to Article 5:3), were required to be taken in a non-discriminatory manner. Likewise the EC was required to apply its cumulation criteria in a non-discriminatory manner.

551. The EC argued that Article 8:2 did not support the contentions made by Brazil. A correct interpretation of Article 8:2 revealed that Article 8:2 applied to the collection of duties. A decision to collect duties occurred subsequent to a decision to impose duties. Article 8:1 was the provision in Article 8 concerned with the imposition of duties. Article 8:1 would be redundant if Article 8:2 was interpreted as referring to the entire process of investigation. The EC noted that implicit in the Agreement was an obligation to not discriminate between two parties in the same situation, but argued that such an obligation did not arise from Article 8:2. Therefore, Brazil could not claim (even if discrimination had occurred, which the EC denied), that Article 8:2 had been breached.

552. The EC also noted that because of the large number of exporters from Brazil, Egypt, India and Turkey, sampling was used to determine whether dumping was occurring. The decision to use sampling had been accepted by all exporters from Brazil that were sampled. All exporters from Brazil that were visited for purposes of verification were found to have been dumping. The weighted average of those co-operating and verified Brazilian exporter's dumping margins was applied to co-operating Brazilian exporters who were not visited for verification. A residual duty was applied to non-co-operating Brazilian exporters. On the basis of that methodology, all the imports from Brazil were considered to have been dumped. The EC noted that the United States - Imposition of Anti-Dumping Duties on Imports of Fresh and Chilled Atlantic Salmon from Norway Panel (hereinafter referred to as "the Salmon Panel") stated that sampling was permitted under the Agreement provided that the methodology could be considered to be sufficient to serve its stated purpose. The EC argued that as Brazilian exporters accepted the sampling methodology, Brazil was not in the same position as Thailand, whose imports had not been sampled during the investigation. Consequently Brazil could not claim that the EC's decision to sample exports from Brazil was discriminatory.

553. The EC also stated that the total exports from India and Thailand held respectively a 1 per cent and 0.5 per cent market share in the EC. Indian and Thai exporters were excluded from cumulation because the dumped exports from those countries were determined to be of such a low volume that they had not significantly contributed to injury. The injury analysis was conducted on the basis of all imports from those countries. India and Thailand were excluded from the investigation due to the low total volume of their (dumped and non-dumped) imports. Even if the non-dumped imports had been deducted from the total volume of Brazil's imports, Brazil's imports could not have been considered to be de minimis. This was because market share of Brazil's exports to the EC (based on Eurostat data) was 2.25 per cent. Even if Cacex data had been used, and Brazil's imports had been adjusted to eliminate the non-dumped imports, Brazil's EC market share would have still been 1.53 per cent, which was above the EC's usual threshold for cumulation. Therefore, Brazil's imports would still have been cumulated. Consequently, on this basis also Brazil could not claim that it had been discriminated against.

554. The Panel noted that the legal basis of Brazil's claim regarding discrimination depended on whether Article 8:2 was the source of an obligation of non-discrimination in relation to both the investigation leading to imposition and the collection of anti-dumping duties. Accordingly, the Panel then turned to interpret Article 8:2 to determine whether Article 8:2 was the source of such an obligation.

555. The Panel noted that Article 8:2 of the Agreement relevantly provides:

"When an anti-dumping duty is imposed in respect of any product, such anti-dumping duty shall be collected in the appropriate amounts in each case, on a non-discriminatory basis on imports of such product from all sources found to be dumped and causing injury, except as to imports from those sources, from which price undertakings under the terms of this Code have been accepted."

The Panel then turned to examine the ordinary meaning of Article 8:2. The Panel considered that the ordinary meaning of Article 8:2 made clear that the provision was concerned with the collection of duty. The Article stated "[w]hen an anti-dumping duty is imposed in respect of any product, such anti-dumping duty shall be collected in the appropriate amounts in each case" (emphasis added). The highlighted words indicated that the obligation contained in Article 8:2, to collect duty in a non-discriminatory manner, only arose once a decision to impose duty existed. This was clear because without a decision to impose duty, no duty could be collected. The words "on a non-discriminatory basis on imports ... found to be dumped and causing injury ..." (emphasis added), also made clear that the investigating authorities must have concluded that the goods were dumped and causing injury, prior to its decision to impose duty. The ordinary meaning of Article 8:2, therefore, made clear that the obligation in Article 8:2 applied only after a decision to impose duties had been taken.

556. This was confirmed by the context of Article 8:2. The Panel noted that Article 8:1 provides

"The decision whether or not to impose an anti-dumping duty in cases where all the requirements for the imposition have been fulfilled and the decision whether the amount of the anti-dumping duty to be imposed shall be the full margin of dumping or less, are decisions to be made by the authorities of the importing country or customs territory. It is desirable that the imposition be permissive in all countries or customs territories Parties to this Agreement, and that the duty be less than the margin, if such lesser duty would be adequate to remove the injury to the domestic industry" (emphasis added).

557. The Panel considered that Article 8:1 was concerned with two kinds of decisions. The first was the taking of a decision whether to impose an anti-dumping duty after all conditions for the imposition had been fulfilled ("[t]he decision whether or not to impose..."). The second type of decision was at what level the anti-dumping duty should be set (i.e. "... whether the amount of the anti dumping duty shall be the full margin or less ..."). The ordinary meaning of Article 8:2 in the context of Article 8:1 revealed that after decisions had been taken to impose a duty (per Article 8:1), and to set the level of the duty (per Article 8:1), Article 8:2 came into effect to require that such duties should be collected in a non-discriminatory manner. Therefore, Articles 8:1 and 8:2 were concerned with different types of decisions taken at different points in time. This confirmed that the obligation contained in Article 8:2 not to discriminate only arose at the time of the collection of the anti-dumping duty, the decision whether to impose duty and the correct amount of the duty having been taken in accordance with Article 8:1.

558. Based on the ordinary meaning of Article 8:2 in its context, as set out in paragraphs 555 and 557 the Panel concluded that the obligation of non-discrimination contained in Article 8:2 arose only at the stage of collection of duties. The Panel, therefore, concluded that Article 8:2 was not the source of an obligation to the effect that the decision whether to impose duties should be taken in a non-discriminatory manner, as argued by Brazil. The obligation not to discriminate contained in Article 8:2 was limited to the collection of duties. The Panel concluded, therefore, that the ordinary meaning of Article 8:2 in its context, did not support the interpretation argued for by Brazil.

559. The Panel noted that Brazil's claim was described in the document constituting the terms of reference of the Panel as being "Violation of Article 3:2 in combination with Article 8:2 : Discrimination against Brazilian exporters in the injury findings" (emphasis added). It appeared to the Panel that Article 3:2 was mentioned in the claim because Brazil considered that the EC had practised discrimination at the time it had examined the "effects" of the dumped imports. However, no specific arguments had been made by Brazil that Article 3:2 in itself contained a prohibition on discrimination. Because Brazil's arguments in relation to the allegedly discriminatory behaviour of the EC were contingent on the Panel interpreting 3:2 in combination with Article 8:2 as expressing an obligation of non-discrimination in relation to both the imposition and the collection of anti-dumping duties, there was no basis for the Panel to examine whether the EC had discriminated against Brazilian imports during the conduct of its investigation.

560. The Panel next proceeded to examine Brazil's contention that the EC's injury determination was clearly discriminatory, and, therefore, in violation of Article I of the GATT. The Panel recalled that in its preliminary ruling it had dismissed a claim made by Brazil because that claim was not within the Panel's terms of reference. The Panel noted that Brazil's contention that the EC's injury determination was clearly discriminatory, and, therefore, in violation of Article I of the GATT did not concern a provision of the Anti-Dumping Agreement. Rather, Brazil's contention concerned a provision of the GATT. The Panel considered that its earlier definition of a claim was equally applicable to a claim relating to a provision of the GATT.

561. The Panel proceeded to examine whether Brazil's contention that the EC's injury determination was clearly discriminatory, and therefore in violation of Article I of the GATT was a claim. The Panel considered that Brazil's contention, although relying on the same arguments as its claim concerning breach of Articles 3:2 and 8:2, was a distinct claim, as it concerned a legal and factual allegation of a breach of a provision of the GATT. The Panel then proceeded to examine whether this claim was within its terms of reference.

562. Noting that in its preliminary ruling the Panel had determined that for a claim to be within the Panel's terms of reference it had to be included in the document constituting the terms of reference of the Panel, the Panel examined that document to see if the claim appeared therein. The Panel concluded that although in that document Brazil had claimed that Article 8:2 embodied a fundamental GATT principle of non-discrimination, Brazil had particularised its complaint as being of a breach of Article 8:2 (in combination with Article 3:2) alone. Brazil had not anywhere in the document constituting the terms of reference of the Panel made a claim of a breach of Article I of the GATT. As the claim was not mentioned within the document constituting the terms of reference of the Panel, the Panel concluded that this claim was not within its terms of reference.

563. The Panel then recalled Brazil's argument concerning Article 5:3. The Panel noted that it did not appear that Brazil had made any claim before the Panel that Article 5:3 had been breached. It instead appeared that Brazil had been advancing the argument to support its contention of a broad ranging obligation not to discriminate. If the argument was properly characterised as constituting a claim, the Panel noted that such a claim did not appear in the document constituting the terms of reference of the Panel. Again recalling its ruling at the second meeting of the parties on one of the preliminary objections of the EC, and its reasoning and conclusion concerning Brazil's claim of a breach of Article I of the GATT, the Panel observed that if the Panel had interpreted Brazil's various statements concerning Article 5:3 as constituting a claim, the Panel would have determined that such a claim was not within its terms of reference.

564. Recalling its conclusions in paragraphs 558, 562 and 563 the Panel dismissed Brazil's claim that the EC had acted inconsistently with Article 3:2 in combination with Article 8:2 of the Agreement when it had conducted its injury analysis.

VI.8 Article 13 - the EC had breached Article 13 of the Agreement by not giving "special regard" to the "special situation" of Brazil and not exploring the possibility of constructive remedies proposed by Brazilian exporters

565. The Panel turned to the claim of Brazil that the EC had breached Article 13 of the Agreement by not giving "special regard" to the "special situation" of Brazil and not exploring the possibility of constructive remedies proposed by Brazilian exporters. Brazil made several arguments in support of its claim.

566. Brazil argued that Article 13 imposed two obligations. The first obligation was to have special regard to the special situation of developing countries when considering the application of measures under the Agreement. The second obligation was to explore constructive remedies.

567. Brazil argued that the first obligation required interpretation of three concepts contained in Article 13. The first was Article 13's use of the phrase "special regard". In this context the phrase "special regard" required investigating authorities to examine information submitted to them, to acknowledge the receipt of and to respond to submissions, to consider those submissions, to note those submissions in the public statement of reasons of the investigating authorities, and to provide a properly reasoned public statement of reasons. Unless all of those requirements were complied with the obligation to have "special regard" would be breached. In this case the EC's adherence to an administrative "rule" to apply only official exchange rates had prevented the EC from giving "special regard" consistent with Brazil's interpretation of the phrase.

568. Brazil argued that the phrase "special situation of developing countries" meant situations connected to the status of Brazil as a developing country. As such a country, it was notable that Brazil was in such a special situation when it was recalled that cotton yarn was a product of particular strategic importance to Brazil, and that Brazil was faced with an economic crisis in 1989.

569. Brazil argued that the word "special" in the phrase "special regard" in Article 13 made clear that the result of the treatment of developing countries should be "special". Also, the EC's use of monthly average normal values did not sufficiently address the "special situation" in which Brazil found itself.

570. Brazil asserted that the second obligation created by Article 13 was to explore possibilities of constructive remedies before applying anti-dumping duties. Brazil said that the words "[p]ossibilities of constructive remedies ... shall be explored ..." created an obligation to seek out a solution other than the imposition of anti-dumping duties. Brazil also argued that the words "... constructive remedies provided for by this Code ..." were not intended to be limited to a reference to undertakings alone. Undertakings were only one of the kinds of constructive remedies provided for by the Agreement. If the drafters of the Agreement had intended to limit the words "... provided for by this Code ..." to undertakings alone, a cross reference to Article 7 (which dealt with undertakings) would have been inserted. Suitable constructive remedies in this case would have been the adjustments or allowances in relation to the normal values and export prices requested by the exporters in this case.

571. Brazil also argued that the words "... when considering the application of anti-dumping measures ..." did not mean that the obligation under Article 13 only arose immediately prior to the imposition of anti-dumping duties. In the circumstances of this case Brazilian imports had been subject to securities since the time of the Provisional Determination. Therefore, the words "... when considering the application of anti-dumping measures ..." should be interpreted to mean at any stage during the investigation process.

572. Therefore, Article 13 interrelated with all the other provisions of the Agreement. Because it was clear that Article 13 was not an obligation only arising subsequent to a final determination of dumping causing material injury, it applied throughout the conduct of an anti-dumping investigation. The third recital of the preamble to the Agreement specified that "[t]aking into account "the particular trade, development and financial needs of developing countries..." were to be taken into account during an investigation. Brazil also argued that the positioning of Article 13, at the end of Part I of the Agreement, was consistent with Article 13 being a statement of general principle.

573. Brazil also argued that the "Decisions by the Committee on Anti-Dumping Practices taken on 5 May 1980" (ADP/2) made clear that Article 13 applied to other provisions of the Agreement. Brazil argued that the decision expressly recognised that Article 13 applied to the determination of the method for establishing normal value, the making of allowances and adjustments to normal value and export price, and to the determination of injury.

574. If it was accepted that Article 13 created a broad obligation that applied throughout the investigation process, Brazil argued that in applying the terms of Article 2:6, without having "special regard" to the "special situation" of Brazil, the EC had again breached Article 13. This was because the EC failed to give meaningful regard to the special situation of the exchange rate freeze in deciding what due allowances to make. The EC had also failed to adequately explain its failure to take the exchange rate freeze into account. The failure to make adjustments, and to explain why it had not done so had affected the essential trade interests of Brazil.

575. Brazil argued that, pursuant to Article 13 of the Agreement, the EC was obliged to have regard to the MFA, in combination with the principles of Part IV of the GATT. In this context Brazil noted that Part IV of the GATT 1947 required that developed country contracting parties give effect to the principles and objectives stated in Article XXXVI in their dealings with less developed country contracting parties.

576. The EC argued that if Article 13 was interpreted consistently with Article 31:1 of the Vienna Convention on the Law of Treaties, in good faith in accordance with its ordinary meaning of the provision in its context and in light of its object and purpose, the first sentence of Article 13 made clear that it was merely introductory of the second sentence. The introduction to the first sentence "[i]t is recognized..." (in the French text "il est reconnu...") qualified any obligation that could be inferred from the remainder of the sentence.

577. The EC argued that the requirement that "special regard" could not be interpreted as meaning more than "consider". In its context the phrase meant nothing more than that the investigating authorities should be prepared to review whether anti-dumping duties should be imposed, or to consider the acceptance of undertakings, at the time of imposition of anti-dumping measures.

578. The EC argued that the phrase "...special situation of developing countries..." referred to the situation of developing countries in general, not to a particular event in a developing country under investigation. The phrase only included situations which arose because the economy of the country was developing, not any situation which arose in a developing country.

579. The EC argued that any obligation contained in the first sentence of Article 13 only arose at the time of consideration of imposition of measures. The word "measures" clearly limited the obligation to the stage following conclusion of an investigation. Article 10 of the Agreement referred to "Provisional measures". The first sentence of Article 10:1 confirmed that the obligation under Article 13 only arose after an investigation. Article 10:2 defined what provisional measures could consist of. Therefore, the obligation of Article 13 first sentence only applied once an investigation was completed, prior to imposition of anti-dumping duty, and not at the stage of investigation of dumping or injury. The second obligation contained in the second sentence only arose when the essential interests of a developing country could be affected by the imposition of anti-dumping duties. Such a situation could arise if dumping duties were imposed by the only purchaser of the only exported product of a developing country. Whilst cotton yarn was an important product for regions of Brazil, cotton yarn was not Brazil's exclusive source of foreign exchange, nor was the EC its only purchaser.

580. The EC also argued that Article 13 provided that "constructive remedies" must be remedies provided by the Agreement. Therefore "constructive remedies" could only include price undertakings. In addition, the obligation was to consider only the possibility of entering into undertakings. There was no obligation to enter into an undertaking, and the EC had indeed considered the possibility of entering into undertakings. Consequently, it had discharged its obligations under Article 13.

581. The EC argued that the 5 May 1980 decision of the Committee (ADP/2) was not concerned with the interpretation of Article 13. The Committee had merely noted that Article 13 created obligations, when it stated that it was "cognizant" or aware of the obligations in Article 13. The decision interpreted the Agreement as a whole. The precise terms of the decision also made this clear, as the decision dealt with such matters as administrative difficulties in adapting domestic legislation and domestic structures to the Agreement, and the provision of technical assistance. Sub-paragraph 1(ii) was an example of a special and differential treatment provision. It dealt with the comparison or normal value and export price under Article 2 of the Agreement, but did not establish a link between Articles 2 and 13. This meant that Article 2 should be interpreted in light of paragraph 1(ii) of the decision, but not in light of Article 13.

582. The Panel noted that Brazil's claim rested on the interpretation of Article 13 of the Agreement. The parties to the dispute had disagreed on the nature and scope of the obligations contained in Article 13. The Panel noted that Article 13 provides as follows:

"It is recognized that special regard must be given by developed countries to the special situation of developing countries when considering the application of anti-dumping measures under this Code. Possibilities of constructive remedies provided for by this Code shall be explored before applying anti-dumping duties where they would affect the essential interests of developing countries."

The Panel was of the view that Article 13 should be interpreted as a whole. In the view of the Panel, assuming arguendo that an obligation was imposed by the first sentence of Article 13, its wording contained no operative language delineating the extent of the obligation. Such language was only to be found in the second sentence of Article 13 whereby it is stipulated that "possibilities of constructive remedies provided for by this Code shall be explored before applying anti-dumping duties where they would affect the essential interests of developing countries".

583. The Panel noted that a precondition for the obligation in the second sentence was that the application of anti-dumping measures "would affect the essential interests of developing countries". Brazil had argued that cotton yarn was a strategic industry for Brazil, and that, therefore, application of anti-dumping duties would be of immense importance to its economy. The Panel was of the view that "essential interests" was defined by its context, like the phrase "special situation of developing countries" in the first sentence of Article 13. The Panel considered that the essential interests of a developing country which could be affected by the imposition of anti-dumping duties could include a strategic industry dependent on export trade. If cotton yarn was such an industry in Brazil, the essential precondition to the activation of the particular obligation contained in the second sentence of Article 13 was satisfied.

584. The Panel noted that if the application of anti-dumping measures "would affect the essential interests of developing countries", the obligation that then arose was to explore the "possibilities" of "constructive remedies". It was clear from the words "[p]ossibilities" and "explored" that the investigating authorities were not required to adopt constructive remedies merely because they were proposed.

585. The Panel was of the view that Article 1 of the Agreement provided the context for Article 13. Article 1 provided that anti-dumping duties could not be applied prior to determination of dumping, material injury and a causal link between the two. In addition, the words "where they would affect the essential interests of developing countries" establish the condition under which this obligation becomes operative. Clearly this condition could only be ascertained subsequent to the determination of the amount of the anti-dumping duty to be applied in order to know whether the imposition of the anti-dumping duty would affect the essential interests of developing countries. This made clear that the second sentence of Article 13 gave rise to an obligation to consider the possibility of adopting constructive remedies after determination of dumping, material injury and a causal link between the two. The Panel recalled that Brazil had argued that suitable constructive remedies in this case would have been the adjustments or allowances to the normal values and export prices requested by the exporters in this case. The adjustments or allowances mentioned by Brazil had been adjustments or allowances to the normal value and export price. Such adjustments or allowances were not consistent with the interpretation that "constructive remedies" should be remedies provided for by the Agreement which could resolve an anti-dumping investigation in which it had already been established that imports were dumped, that material injury had been suffered by the domestic industry and that there was a causal link between the two. Those adjustments or allowances mentioned by Brazil were only relevant to the stage of investigation of dumping or injury, whereas the "constructive remedies" in the context of Article 13 only applied once an investigation was completed. Accordingly, such adjustments or allowances would not be "constructive remedies provided for by this Code". Equally, a determination of negligible margins of dumping or low volume of market share, was required, pursuant to Article 5:3, to be made at a stage of the investigation process prior to the time at which parties were obliged to consider the possibility of constructive remedies; consequently, they should not be considered as "constructive remedies provided for by this Code" either.

586. The Panel noted that Brazil had argued that paragraph 1 of the "Decisions by the Committee on Anti-Dumping Practices taken on 5 May 1980" (ADP/2) operated to apply the obligation contained in Article 13 to other Articles of the Agreement. The Panel, therefore, proceeded to explore whether paragraph 1 of the 5 May 1980 decision of the Committee had the effect of altering any of the obligations contained in Article 13 so that Article 13 applied to other Articles of the Agreement. For the purposes of the following analysis, irrespective of the legal status of the decision, the Panel proceeded to examine whether paragraph 1 of the decision was in conflict with Article 13.

587. The Panel noted that paragraph 1 of the decision relevantly provides,

"1. The Committee, cognizant of the commitment in Article 13 of the Agreement on Implementation of Article VI of the General Agreement on Tariffs and Trade that special regard must be given by developed countries to the special situation of developing countries when considering the application of anti-dumping measures under the Agreement,takes the following decision concerning the application and interpretation of the Agreement in relation to developing countries:

...

(ii) In the case of imports from a developing country, the fact that the export price may be lower that the comparable price for the like product when destined for domestic consumption in the exporting country does not per se justify an investigation or the determination of dumping unless the other factors mentioned in Article 5:1 are also present. Due consideration should be given to all cases where, because special economic conditions effect prices in the home market, these prices do not provide a commercially realistic basis for dumping calculations. In such cases the normal value for the purposes of ascertaining whether the goods are being dumped shall be determined by methods such as a comparison of the export price with the comparable price of the like product when exported to any third country or with the cost of production of the exported goods in the country of origin plus a reasonable amount for administrative, selling and other costs and for profits."

The Panel noted that the determination of whether the decision operated as an agreement regarding the interpretation of Article 13, which thereby implied a link between Article 13 and other provisions of the Agreement rested on the interpretation of the word "cognizant". The Panel considered that the word "cognizant" in this context meant that the obligation noted in the first sentence of Article 13 was being recollected, which was thereafter set out in the first paragraph of the decision. Immediately following the restatement of the first sentence of Article 13, first paragraph of the Committee's decision interpreting the Agreement was set out in sub-paragraphs (i)-(iv). In this context, the word "cognizant" did not suggest that the decision operated as an amendment to Article 13. The word "cognizant" meant that recalling the first sentence of Article 13, and mindful of that sentence, the Committee had made a statement interpreting the Agreement. The Panel was of the view that paragraph 1 of the Committee's decision did not alter or amend Article 13 of the Agreement. Rather, it acknowledged the existence of Article 13, and then set out four sub-paragraphs concerning the application of the Agreement in relation to developing countries. The decision only operated to refer to the obligation noted in the first sentence of Article 13, it did not operate to modify Article 13. This was also clear from the structure of the decision, in that it restated the first sentence of Article 13, then interpreted the Agreement. The interpretations set out thereafter did not relate to the obligation created in Article 13, nor to the obligation noted in the first sentence of Article 13. Consequently, the fact that sub-paragraph 1(ii) of the decision dealt with calculation of normal value, did not modify the obligation in Article 13.

588. In light of its conclusions concerning the obligations contained in Article 13, the Panel turned to consider whether the EC had complied with the requirements of Article 13 as interpreted by the Panel.

589. The Panel note that in Section M entitled "UNDERTAKINGS" of the Definitive Determination the EC had stated,

"Both the Turkish and Brazilian authorities, having been informed on the essential facts and considerations on the basis of which it was intended to recommend the imposition of definitive duties, offered, on behalf of the exporters concerned, a form of undertaking.

...

As regards the Brazilian offer which provides for voluntary quantitative export restrictions, the Commission was not satisfied that its acceptance would eliminate the injurious effects of the dumping.

The Council noted that for these reasons both these offers of undertaking have, after consultation, been rejected."

In the view of the Panel at the stage of considering the application of anti-dumping duties, the EC had considered whether it could enter a quantitative undertaking and had considered that such an undertaking would not eliminate the injury caused by the dumped imports. The Panel recalled that it had concluded in paragraph 584 that there was no obligation to enter into the constructive remedies, merely to consider the possibility of entering into constructive remedies.

590. Based on its conclusions in paragraphs 585, 587 and 589, the Panel dismissed the claim by Brazil that the EC had breached Article 13 of the Agreement by not giving "special regard" to the "special situation" of Brazil and not exploring the possibility of constructive remedies proposed by Brazilian exporters.

VI.9 Conclusions

591. The Panel recalled its conclusions with respect to the preliminary objections of the EC that:

(a) the claim that the EC had made an "incorrect determination that certain domestic sales were not in the ordinary course of trade" was not within the Panel's terms of reference, and was therefore not properly before the Panel,

(b) the assertion that in a high inflation environment distortions could occur if an average normal value was compared with individual export prices and if no due allowance was made pursuant to Article 2:6 for "negative" dumping margins, was an argument advanced to support the claim that the EC had failed to taken into account distortions arising from high domestic inflation combined with fixed exchange rates in comparing normal value and export price, and accordingly could be considered by the Panel, and

(c) in light of the Panel's resolution of Brazil's claim concerning Article 2:6 of the Agreement, it was not necessary for the Panel to determine the EC's remaining "preliminary objections" concerning certain evidence submitted by Brazil which the EC argued extended the scope of review that panels were properly permitted to undertake.

592. The Panel further recalled its conclusion with respect to Brazil's claim that the EC breached Article 2:4 that:

(a) it dismissed Brazil's claim that the EC had failed to properly consider the particular market situation prevailing in Brazil and had thereby acted inconsistently with Article 2:4.

593. The Panel also further recalled its conclusion with respect to Brazil's claim that the EC breached Article 2:6 that:

(a) it dismissed Brazil's claim that the EC had violated Article 2:6 of the Agreement because it had failed to take into account distortions arising from high domestic inflation combined with frozen exchange rates in comparing normal value and export price.

594. The Panel also further recalled its conclusions with respect to the injury determination by the EC that:

(a) it dismissed Brazil's claim that the EC had breached its obligations in Article 3:1 and 3:2 to base its findings of injury on positive evidence and to make an objective examination of the relevant facts,

(b) it dismissed Brazil's claim that the EC had violated Articles 3:2, 3:3 and 3:4, because the facts stated in an injury analysis must reasonably support the determination that is reached, and the EC's decision was not in conformity with that requirement,

(c) it dismissed Brazil's claim that the EC had acted inconsistently with Article 3:2 in combination with Article 8:2 of the Agreement when it had conducted its injury analysis, and

(d) it dismissed Brazil's claim that the EC had breached Articles 3:2, 3:3 and 3:4, because the MFA precluded a finding of injury.

595. The Panel also further recalled its conclusions with respect to Brazil's claim that the EC had not complied with Article 13 of the Agreement that:

(a) it dismissed the claim by Brazil that the EC had breached Article 13 of the Agreement by not giving "special regard" to the "special situation" of Brazil and not exploring the possibility of constructive remedies proposed by Brazilian exporters.

596. For the foregoing reasons, the Panel concluded that Brazil had not, through its claims and arguments in support of those claims, established that the imposition of anti-dumping duties on imports of cotton yarn from Brazil by the EC was inconsistent with the EC's obligations under the Agreement.