CFI, 5th chamber, October 12, 1999, No T-48/96

THE COURT OF FIRST INSTANCE OF THE EUROPEAN COMMUNITIES (Fifth Chamber, Extended Composition)

Background to the dispute

1 This action seeks the annulment of Council Regulation (EC) No 5-96 of 22 December 1995 imposing definitive anti-dumping duties on imports of microwave ovens originating in the People's Republic of China, the Republic of Korea, Malaysia and Thailand and collecting definitively the provisional duty imposed (OJ 1996 L 2, p. 1, hereinafter `the contested regulation'). That regulation follows the adoption of Commission Regulation (EC) No 1645-95 of 5 July 1995 imposing a provisional anti-dumping duty on imports of microwave ovens originating in the People's Republic of China, the Republic of Korea, Thailand and Malaysia (OJ 1985 L 156, p. 5, hereinafter `the provisional regulation'). Those regulations were adopted on the basis of Council Regulation (EEC) No 2423-88 of 11 July 1988 on protection against dumped or subsidised imports from countries not members of the European Economic Community (OJ 1988 L 209, p. 1, hereinafter `the basic regulation').

2 The applicant, Acme Industry Co. Ltd (hereinafter `Acme'), is a company incorporated under Thai law which manufactures and exports microwave ovens (hereinafter `MWOs') and a subsidiary of the Japanese holding company Nisshin Industry Co. Ltd (hereinafter `the Nisshin Group'). The Nisshin Group also controls Korea Nisshin Co. Ltd (hereinafter `Korea Nisshin'), a Korean manufacturer of MWOs, and Imarflex Mfg Co. (hereinafter `Imarflex'), a Japanese company which markets MWOs, including some of those manufactured by the applicant.

3 Following a complaint filed in June 1993 by the Groupement Interprofessionnel des Fabricants d'Appareils d'Équipement Ménager, the Commission published, on 18 December 1993, a notice of initiation of an anti-dumping proceeding concerning imports of MWOs originating in the People's Republic of China, the Republic of Korea, Thailand and Malaysia (OJ 1993 C 341, p. 12). The investigation covered the period from 1 October 1992 to 30 September 1993.

4 The Commission sent a questionnaire to the applicant to which the latter replied by letter of 4 February 1994. At the Commission's request, the applicant supplemented its reply by letter of 22 February 1994. No other Thai producer cooperated in the investigation procedure.

5 On 19 April 1994 the Commission carried out a first inspection at the premises of Imarflex in Osaka, Japan, in order to verify the applicant's replies to the questionnaire. On 22 April and 5 and 6 May 1994, the Commission also carried out inspections at the premises of Korea Nisshin and Acme respectively.

6 On 5 July 1995, the Commission adopted the provisional regulation which fixed the provisional rate of duty applicable to the MWOs manufactured by the applicant at 20.3%.

7 By letter of 14 July 1995 the Commission notified the applicant of the main facts and considerations on the basis of which it had imposed the provisional anti-dumping duty (hereinafter `the provisional disclosure letter').

8 By letter of 31 July 1995 the applicant commented on the provisional disclosure letter.

9 By letter of 24 October 1995 the Commission notified the applicant of the main facts and considerations on the basis of which it intended to propose to the Council that a definitive anti-dumping duty be imposed (hereinafter `the definitive disclosure letter').

10 By letter of 3 November 1995 the applicant commented on the definitive disclosure letter.

11 On 22 December 1995 the Council adopted the contested regulation imposing a definitive anti-dumping duty of 14.1% on imports of MWOs originating in Thailand and manufactured by the applicant. It appears from that regulation that the Commission and the Council were unable, as the applicant did not sell MWOs or products within the same economic sector of activity on its domestic market, to determine the normal value on the basis of the actual price charged on the Thai market. As a result, the institutions calculated the constructed normal value in accordance with Article 2(3)(b)(ii) of the basic regulation, taking the view that the amount corresponding to the sales, general and administrative expenses (hereinafter `SG & A expenses') and the profit margin should be established on `any other reasonable basis' and, in this case, considered appropriate to use the amount established for profitable sales on the domestic market in Korea, the only competitive market covered by the investigation where profitable sales of similar products were made in representative quantities. The constructed values of the models exported by the applicant were, therefore, calculated on the basis of `all costs, both fixed and variable, of materials and manufacture' (hereinafter `manufacturing costs'), increased by an amount corresponding to the SG & A expenses and a reasonable profit margin (recital 26 in the preamble to the contested regulation and recitals 46 and 36 in the preamble to the provisional regulation).

Procedure and forms of order sought

12 The applicant brought this action by an application filed with the Registry of the Court of First Instance on 29 March 1996.

13 The Commission and the French Republic were given leave to intervene in support of the form of order sought by the Council by orders of the President of the Fourth Chamber (Extended Composition) of 23 October and 9 December 1996 respectively. The second order also granted a request for confidential treatment of the applicant in relation to France.

14 On hearing the report of the Judge-Rapporteur, the Court of First Instance (Fifth Chamber, Extended Composition) decided to open the oral procedure without any preparatory enquiry. It did, however, put a number of written questions to the parties to be answered at the hearing.

15 The parties presented oral argument and replied to the Court's questions at the hearing on 11 March 1999.

16 The applicant claims that the Court should:

- annul the contested regulation in so far as it relates to the applicant;

- order the Council to pay the costs.

17 The Council contends that the Court should:

- dismiss the application;

- order the applicant to pay the costs.

18 The Commission and the French Republic, interveners, contend that the Court should:

- dismiss the application;

- order the applicant to pay the costs.

The substantive claim

19 In support of its action, the applicant relies on five pleas in law. In its first and second pleas, the applicant challenges the Council's calculation of the constructed normal value in so far as, first of all, it was determined in accordance with a method which did not comply with Article 2(3)(b)(ii) of the basic regulation and, secondly, it was calculated by using the Korean exporter's S G & A expenses and profits thus infringing the principle of equality. In its third plea, it submits that the Council failed to take account of the principle of equity by failing to apply certain provisions of Council Regulation (EC) No 3283-94 of 22 December 1994 on protection against dumped imports from countries not members of the European Community (OJ 1994 L 349, p. 1), as amended inter alia by Council Regulation (EC) No 355-95 of 20 February 1995 (OJ 1995 L 41, p. 2), and the 1994 Agreement on Implementation of Article VI of the General Agreement on Tariffs and Trade (OJ 1994 L 336, p. 103, hereinafter `the WTO anti-dumping code'), approved by Council Decision 94-800-EC of 22 December 1994 concerning the conclusion on behalf of the European Community, as regards matters within its competence, of the agreements reached in the Uruguay Round multilateral negotiations (1986-1994) (OJ 1994 L 336, p. 1). The fourth plea alleges infringement of Article 190 of the EC Treaty (now Article 253 EC). The fifth plea alleges infringement of Article 2(10) of the basic regulation in making a comparison between the constructed normal value and the export price.

20 The Court considers it appropriate to begin by examining the question of the applicability of Regulation No 3283-94 and the WTO anti-dumping code (third plea), then the complaints relating to the determination of the constructed normal value and the comparison thereof with the export price (first, second and fifth pleas) and, finally, the argument alleging that the reasons stated in the contested regulation are insufficient (fourth plea).

1. The applicability of Regulation No 3283-94 and the WTO anti-dumping code (third plea in law)

Arguments of the parties

21 The applicant alleges in the first place that the Council failed to determine the SG & A expenses and the profit margin on the basis of Article 2(6)(iii) of Regulation No 3283-94. Although that regulation was not yet applicable, the institutions were none the less bound by a general principle of equity to apply it in this case in so far as its provisions were more favourable to the applicant than those of the basic regulation (Case 78-77 Lührs [1978] ECR 169, paragraph 13, and Case 236-78 FNROM [1979] ECR 1819).

22 The applicant claims that, according to the principles governing the application of legislation ratione temporis, a law amending a legislative provision applies as a rule to the future consequences of situations which arose under the old law (Case 68-69 Brock [1970] ECR 171, paragraph 6). The Council can only derogate from that principle for compelling reasons of Community interest (Case 92-77 An Bord Bainne [1978] ECR 497 and Case C-345-88 Butterabsatz Osnabrück-Emsland [1990] ECR I-159).

23 Secondly and in the alternative, the applicant claims that the Council was bound to interpret the provisions of the basic regulation in the light of Article 2, paragraph 2.2.2 of the WTO anti-dumping code, according to which the amounts corresponding to the profits established in the framework of the calculation of the constructed normal value may not exceed that `normally realised by other exporters or producers on sales of products of the same general category in the domestic market of the country of origin'. It considers that the amendments made to the GATT anti-dumping code during the course of the Uruguay Round of negotiations must be applied immediately to pending cases (Joined Cases 21-72 to 24-72 International Fruit Company and Others [1972] ECR 1219, paragraphs 14 to 18; Case 38-75 Nederlandse Spoorwegen [1975] ECR 1439; and Case 112-80 Dürbeck [1981] ECR 1095).

24 Thirdly, in its comments on the French Government's statement in intervention, the applicant questions the application to this case of a general principle of law enshrined, inter alia, in Article 15 of the International Covenant on Civil and Political Rights (Collected Treaties, Volume 999, page 171, hereinafter `the International Covenant'), according to which if, following the commission of an offence, provision is made by law for the imposition of the lighter penalty, the offender must benefit thereby.

25 The Council and the interveners object that Regulation No 3283-94 and the WTO anti-dumping code were inapplicable in this case.

26 The French Government denies that there is a general principle of retroactivity in Community law. It maintains that, in accordance with the principle of legal certainty, the retroactivity of Community acts is the exception and not the rule (Case 17-67 Neumann [1967] ECR 441 and Case 70-83 Kloppenburg [1984] ECR 1075, paragraph 12) and that there can be no retroactivity in any area of criminal law nor where a conflict exists with general principles of Community law such as the principle of legal certainty or the principle of the protection of legitimate expectations.

27 The Commission observes that the institutions complied not only with the provisions of Regulation No 3283-94 but also with the Community's international obligations under the WTO anti-dumping code. Regulation No 3283-94 expressly provides for the application of the basic regulation to procedures which, though initiated prior to 1 September 1994, were still pending as at 1 January 1995. Those provisions simply reflect the terms of the WTO anti-dumping code, which is, by its own express terms, inapplicable to investigations initiated as a result of applications made before it entered into force.

Findings of the Court

28 Whilst it is true that the contested regulation was adopted after the entry into force of the basic regulation on 1 January 1995, the fact remains that it was adopted following a procedure which was initiated in 1993 and continued beyond 1 January 1995. It is clear from the transitional provisions, and particularly Article 23 of Regulation No 3283-94 as amended by Regulation No 355-95 of 20 February 1995, cited above, that the basic regulation continues to govern procedures in relation to which an investigation which was under way as at 1 September 1994 had not been concluded by 1 January 1995 (see, on this point, Case T-232-95 CECOM v Council [1998] ECR II-2679, paragraph 35).

29 The applicant's arguments cannot call into question that literal interpretation of the clearly-worded transitional provisions of Regulation No 3283-94. First of all, the case-law on which it relies in support of its arguments alleging infringement of the principles of equity and application of laws ratione temporis concerns situations where the Community judicature was called upon to resolve temporal conflicts of law in the absence of express transitional provisions.

30 Secondly, the applicant's argument as to infringement of the general principle of law enshrined, inter alia, in the International Covenant is invalid. Without there being any need to consider whether that argument constitutes a new plea in law within the meaning of Article 48(2) of the Rules of Procedure of the Court, it should be noted that Article 15 of the Covenant relates only to persons charged with a criminal offence in the course of legal proceedings and is therefore unconnected with anti-dumping investigations which are not criminal in nature (see, by way of analogy, Case 374-87 Orkem v Commission [1989] ECR 3283, paragraph 31). Furthermore, Article 2(6)(iii) of Regulation No 3283-94, which the applicant claims should be applied, relates to the method of calculating the constructed normal value. It is therefore clearly unrelated to the imposition of penalties or sanctions to which the principle relied on is said to apply.

31 Thirdly, it follows from the transitional provisions laid down in Article 18(3) of the WTO anti-dumping code that its terms apply only to investigations undertaken following an application submitted on or after 1 January 1995, the date on which it came into force. The terms of the WTO anti-dumping code are therefore irrelevant to these proceedings.

32 It follows that the legality of the contested regulation cannot be assessed in the light of either Regulation No 3283-94 or the WTO anti-dumping code, but must be assessed primarily in the light of the provisions of the basic regulation which the institutions were correct to apply in this case and also, where appropriate, in the light of the general principles of Community law and all the relevant rules that were in force at the time.

33 It follows from the foregoing that the applicant's third plea in law, alleging breach of the principle of equity, must be rejected.

2. Determination of the constructed normal value (first and second pleas in law)

34 The arguments advanced by the applicant in its first and second pleas may be summarised as follows. First of all, it claims that the institutions infringed an agreement relating to the use of Imarflex's data in determining the constructed normal value. Secondly, it challenges the choice of method of determining the SG & A expenses and profit margin used by the Council. By determining the constructed normal value on `any other reasonable basis' within the meaning of the final part of Article 2(3)(b)(ii) of the basic regulation and by refusing to use Imarflex's data, the Council infringed that regulation. Thirdly, the applicant questions whether it was reasonable to use the Korean data to determine the constructed normal value which had a major effect on the calculation of the anti-dumping duty. The use of Imarflex's data would have resulted in a finding that the mark-up was 11.86% instead of 32.47% in relation to the Korean producers and, accordingly, in the imposition of an anti-dumping duty of 0.183% instead of the 14.1% rate that was ultimately imposed.

Preliminary observations

35 Before undertaking an examination of those complaints, it is appropriate to recall that Article 2(3)(b)(ii) of the basic regulation, which sets out three methods of calculating the constructed normal value, provides as follows:

`For the purposes of this regulation, the normal value shall be:

[...]

(b) when there are no sales of the like product in the ordinary course of trade on the domestic market of the exporting country or country of origin, or when such sales do not permit a proper comparison:

[...]

(ii) the constructed value, determined by adding cost of production and a reasonable margin of profit. The cost of production shall be computed on the basis of all costs, in the ordinary course of trade, both fixed and variable, in the country of origin, of materials and manufacture, plus a reasonable amount for selling, administrative and other general expenses. The amount for selling, general and administrative expenses and profit shall be calculated by reference to the expenses incurred and the profit realised by the producer or exporter on the profitable sales of like products on the domestic market. If such data is unavailable or unreliable or is not suitable for use they shall be calculated by reference to the expenses incurred and profit realised by other producers or exporters in the country of origin or export on profitable sales of the like product. If neither of these two methods can be applied the expenses incurred and the profit realised shall be calculated by reference to the sales made by the exporter or other producers or exporters in the same business sector in the country of origin or export or on any other reasonable basis.'

36 Having regard to the wording of that article, those three methods must be considered in the order in which they are set out (Case C-69-89 Nakajima v Council [1991] ECR I-2069, paragraph 61; Case C-105-90 Goldstar v Council [1992] ECR I-677, paragraph 35; and Case T-118-96 Thai Bicycle v Council [1998] ECR II-2991, paragraph 53). It is only where none of those methods can be applied that recourse must be had to the general provision at the end of Article 2(3)(b)(ii), to the effect that expenses and profit may be calculated `on any other reasonable basis' (Nakajima v Council, cited above, paragraph 61).

37 It is clear from the wording of Article 2(3)(b)(ii) of the basic regulation that each of the methods of calculating constructed normal value set out therein must be applied in such a way that the calculation remains reasonable, a concept, moreover, which is expressly referred to in the first two sentences and the final sentence of that provision (Nakajima v Council, cited above, paragraph 35). Accordingly, the institutions cannot take into consideration accounting data which are unreliable.

38 Article 2(3)(b)(ii) of the basic regulation also confers a wide discretion on the institutions in assessing accounting data submitted to them for the purposes of determining constructed normal value. The Court's review must therefore be restricted to verifying whether the procedural rules have been complied with, whether the facts on which the contested assessment is based are accurate or whether there has been a manifest error in the appraisal of those facts or a misuse of powers.

39 The Community judicature cannot intervene in assessments reserved to the Community authorities but must restrict its review to satisfying itself that the institutions took account of all the relevant circumstances and appraised the facts of the matter with all due care, so that normal value may be regarded as having been determined in a reasonable manner (see Case C-16-90 Nölle [1991] ECR I-5163, paragraphs 12 and 13, and Case T-164-94 Ferchimex v Council [1995] ECR II-2681, paragraph 67).

40 Under the basic regulation, it is for the Commission, as the investigating authority, to determine whether the product in question is being dumped and causing damage when it is released into free circulation in the Community. In that connection, the Commission must verify whether the export price to the Community of the product in question is lower than the normal value of a similar product and, in so doing, must use the data available at the time without imposing the burden of proof on one of the parties.

41 Thus, according to Article 7(2)(a) of the basic regulation, `the Commission shall seek all information it deems to be necessary and, where it considers it appropriate, examine and verify the records of importers, exporters, traders, agents, producers, trade associations and organisations.'

42 However, the basic regulation does not confer on the Commission investigating powers enabling it to compel producers or exporters in respect of whom a complaint has been filed to participate in an investigation or to produce information. Although the Commission may, if necessary, carry out investigations in third countries, that right may only be exercised subject to the agreement of the firms concerned and if no objection is raised by the government of the relevant country, which must be officially notified (Article 7(2)(b) of the basic regulation).

43 The reply to the questionnaire and the verification after which the Commission may act on the spot are therefore essential to the course of the procedure. Under Article 7(7)(b) of the basic regulation:

`In cases in which any interested party or third country refuses access to, or otherwise does not provide, necessary information within a reasonable period, or significantly impedes the investigation, preliminary or final findings, affirmative or negative, may be made on the basis of the facts available. Where the Commission finds that any interested party or third country has supplied it with false or misleading information, it may disregard any such information and disallow any claim to which this refers.'

44 Therefore, the risk that the institutions might take into account data other than that provided in reply to the questionnaire is inherent in the anti-dumping procedure and is intended to encourage genuine cooperation on the part of the undertakings targeted by the investigation.

45 In the light of those factors, it is appropriate to examine in turn the arguments concerning, first, breach of an agreement relating to the use of Imarflex's data, then choice of method for determining the SG & A expenses and profit margin and, finally, the unreasonableness of using the Korean data.

Breach by the Commission of an agreement relating to the use of Imarflex's data

Arguments of the applicant

46 The applicant claims that the Commission official handling the file agreed orally on 6 January 1994 during a meeting with the applicant's adviser to take into account the SG & A expenses and profit margin of Imarflex, the real exporter of the applicant's products, in order to determine the constructed normal value and that, on the basis of that agreement, it filled in the questionnaire using those data.

47 It argues therefore that the failure to comply with that alleged verbal agreement, which the institutions deny existed, amounts to an infringement of the principle of the protection of legitimate expectations.

Findings of the Court

48 If an applicant is to rely on a breach of the principle of the protection of legitimate expectations in the form of a failure to comply with an alleged verbal agreement between a Commission official and the applicant's adviser as to the data to be taken into account when calculating the constructed normal value, it must establish that the Community authorities gave it precise assurances such as to arouse legitimate expectations on its part (see Case T-571-93 Lefebvre and Others v Commission [1995] ECR II-2379, paragraph 72).

49 In fact, the purpose of the discussion of 6 January 1994 during which the alleged agreement relied on by the applicant was entered into was, according to the actual wording of a letter of 29 December 1993 from the applicant's adviser to the Commission, `to clarify some important part of the questionnaire'. While it is clear both from the file and from the explanations put forward at the hearing that, during that discussion, the parties referred to the use of Imarflex's data on SG & A expenses and its profit margin with a view to preparing the response to the questionnaire, the applicant none the less stated in the reply that it had understood that Commission staff would use the response to the questionnaire thus prepared `provided that the verification would allow to confirm the data on Imarflex's SG & A and profit'.

50 In those circumstances, the applicant cannot claim that the Commission gave it a precise assurance that it would determine the constructed normal value on the basis of Imarflex's data such as to cause it to entertain legitimate expectations. It follows that the applicant's complaint, alleging breach of the principle of the protection of legitimate expectations, must be rejected as unfounded.

The choice of method for determining SG & A expenses and profit margin

51 The applicant challenges not the use of the constructed normal value method but the institutions' decision to calculate that constructed normal value on `any other reasonable basis' and to use the Korean data for that purpose. It argues that the institutions should have calculated the constructed normal value by reference to Imarflex's SG & A expenses and the profit margin realised by Imarflex on its domestic market in accordance with the last of the three methods referred to in Article 2(3)(b)(ii) of the basic regulation, that it to say `by reference to the sales made by the exporter or other producers or exporters in the same business sector in the country of origin or export'. As part of its argument, the applicant lays emphasis first of all on Imarflex's role as exporter and secondly on the reliability of its data. Thirdly, it disputes the representative nature of the mark-up on the `CMO 552' model, on the basis of which the Commission concluded that Imarflex's data were unreliable and, fourthly and finally, it claims that Imarflex's data were, in any event, capable of being verified.

52 The Court considers it appropriate to begin by examining the question whether the data relating to Imarflex were reliable.

The reliability of Imarflex's data

- Arguments of the parties

53 The applicant submits, first, that the Council was wrong not to use Imarflex's data because of alleged contradictions between the information subsequently provided by it in response to the questionnaire, when inspecting Imarflex's premises, and in its letter of 31 July 1995. It considers that the discrepancies between the sum of the SG & A expenses and the profit margin, expressed as a percentage of the manufacturing costs (hereinafter `mark-ups') established at each of the three stages, were negligible. It notes that those mark-ups were 11.39% in its reply to the questionnaire and in its letter of 31 July 1995, and 14% when the on-the-spot inspection was carried out.

54 The institutions concluded that Imarflex's data were unreliable after the inspection had revealed that the mark-up in respect of the `CMO 522' MWO sold in Japan was 31.1%. The gap between that mark-up and those referred to above is effectively attributable to differences in the treatment for accounting purposes of sales rebates and after-sales repair and delivery costs.

55 In reply, the Council states that the data relating to Imarflex were unusable because they were imprecise and did not comply with the terms of the questionnaire.

56 The questionnaire required SG & A expenses specifically showing sales of MWOs to be itemised by category, and to be expressed as a percentage of net turnover. Although the applicant estimated SG & A expenses at 7.24% and the profit margin at 4.15% of the manufacturing costs, it did not distinguish between MWOs and other products sold, nor did it itemise SG & A expenses by category. At the request of the Commission, the applicant provided supplementary information by letter of 22 February 1994, but did not genuinely distinguish between MWOs and other products. Finally, during the inspection of Imarflex's premises, the investigators checked the data relating to the `CMO 552' model and found that there was a mark-up of over 30% on it. The applicant subsequently attempted to explain those differences by letter of 31 July 1995 and offered to use other methods of calculating the SG & A expenses and profit margin.

57 The Council contends that, following errors of classification, the various levels of SG & A expenses and profit margin reported by the applicant had to be considered unreliable. In particular, the applicant did not include after-sales repair and delivery costs when calculating the constructed normal value. The Council believes that that was an attempt on its part to conceal the existence of those costs.

58 The applicant denies that its intention was to conceal those costs and goes on to describe the method used by it to compile Imarflex's SG & A expenses and profit margin. In that connection, it notes that Imarflex does not have a computerised accounting system nor does it account for product lines separately and it could not therefore produce an itemised breakdown of the costs relating to MWOs only.

59 The applicant notes first of all that whilst the institutions take the view that repair costs are not `costs of materials and manufacture' within the meaning of the basic regulation, the fact remains that, from an accounting point of view, they are. Indeed, Imarflex and the applicant record after-sales repair costs relating to defective MWOs as `costs of sales' in their profit and loss accounts. Moreover, its accounts were audited by a certified accountant.

60 The applicant explains that Imarflex does not repair defective ovens manufactured by Korea Nisshin which it sells on the Japanese market but replaces them with new ovens. From an accounting point of view, that transaction is treated as a decrease in `inventory' and recorded under `costs of sales' in the profit and loss account. The applicant argues that repair costs are also `costs of materials and manufacture', within the meaning of the basic regulation, and the institutions should therefore not have required their inclusion in Imarflex's SG & A expenses. That method would result in those costs being recorded first in the applicant's manufacturing costs and then a second time in Imarflex's SG & A expenses when calculating the constructed normal value.

61 In reply the Council states that the after-sales repair costs should, in the response to the questionnaire, have been declared as SG & A expenses. Only the costs of maintaining the manufacturing equipment form part of manufacturing costs. Since the applicant added after-sales repair costs to Imarflex's manufacturing costs but not to its own manufacturing costs, it deducted those costs twice over in determining the constructed normal value.

62 The applicant claims, secondly, that it included transportation costs under the heading `costs of sales' in its reply to the questionnaire and denies attempting to conceal them. It argues first of all that the Nisshin Group companies do not all classify those costs as `sales costs-general and administrative expenses' in their profit and loss accounts. Furthermore, it emphasises that those accounts are audited by various certified accountants and that Japanese legislation does not require those companies to harmonise the presentation of their annual accounts.

63 The applicant acknowledges in the reply that it made a mistake in not including the delivery costs under SG & A expenses in its response to the questionnaire. It claims that it corrected that error in its submissions of 3 November 1995.

64 The applicant claims that in any event the error was negligible and could be rectified. The delivery costs merely have to be deducted from the `costs of sales' heading in Imarflex's profit and loss account and added to SG & A expenses. After deducting the financial costs, the SG & A expenses would amount to 11.76% and the profit margin to 4.32%, that is to say, a mark-up of 16.08%. Deduction of the delivery costs expressed as a percentage of the manufacturing costs (4.22%) would therefore result in SG & A expenses of 11.86%. That rate, which it notified to the Commission by letter of 3 November 1995, is very near to the 11.39% submitted in the reply to the questionnaire.

65 The Council points out that the applicant acknowledges that it did not comply with the terms of the questionnaire, which required that delivery costs be treated as SG & A expenses. The Commission established on the basis of the information gathered during the on-the-spot investigation that Imarflex's delivery costs in respect of MWOs sold to Japan were between 4.29% and 11.83%, that is, considerably higher than the average of 4.22% suggested by the applicant for Imarflex's overall sales. The Council concludes from all of those factors that the applicant attempted to conceal delivery costs.

66 The applicant claims, thirdly, that the sales rebates granted by Imarflex for cash payments fulfilled the conditions laid down in Article 2(3)(a) of the basic regulation and should therefore have been deducted from the normal value and not added to the SG & A expenses. It considers that, under the basic regulation and previous anti-dumping legislation, it did not have to include the amount of those rebates in Imarflex's SG & A expenses or to provide information on those rebates in its reply to the questionnaire.

67 The Council notes that sales rebates are business expenses, whether they are shown in the accounts as a deduction from turnover or as part of SG & A expenses. Under Article 2(3)(a) of the basic regulation, the normal value is net `of all discounts and rebates directly linked to the sales under consideration provided that the exporter claims and supplies sufficient evidence that any such reduction from the gross price has actually been granted'. The questionnaire stated how rebates must be declared in order to be able to take advantage of those provisions. The applicant never made any such request but confined itself to reporting Imarflex's net turnover.

68 Finally, the Council notes that the questionnaire also required, in relation to the declaration of net turnover, detailed information on reductions granted and deducted from the gross turnover, however they are treated under Article 2(3)(a) of the basic regulation. The applicant did not give any particulars on that point in its reply to the questionnaire. In its letter of 22 February 1994, the applicant indicated that it granted reductions on cash sales of 0.8% of Imarflex's total turnover. Then, in its letter of 31 July 1995, it stated that it granted a rebate of 1.5% of sales prices and also declared a cash rebate of 3%. The Council believes that those facts are sufficient to demonstrate the unreliability of the data relating to Imarflex.

- Findings of the Court

69 In its provisional disclosure letter, the Commission, having indicated that it did not rely on Imarflex's data, stated as follows:

`[...] it was established on [the] spot that the mark-up taken by Imarflex when selling MWOs on the Japanese market was around 30%. Despite that relatively high mark-up, the SG & A rate reported by Imarflex was 7.24% and the profit rate was only 4.15%. It is clear that these rates, compared to the mark-up, are contradictory and therefore not reliable.'

70 In its definitive disclosure letter, the Commission took the following view:

`[...] first of all [the] general statement concerning the mark-up of 30% on sales of MWOs in Japan supported by detailed example which was in contradiction with the 11.39% mark-up reported by Acme in the response to the questionnaire was given during the on [the] spot verification in Japan;

in the second place, the mere fact that costs on sales such as delivery costs and repair costs are not booked in the SG & A [expenses] but elsewhere does not mean that they don't have to be reported as SG & A expenses;

in the third place, the documents provided by your correspondence of 30 July clearly show that the mark-up for each of the MWO types sold is substantially above the mark-up that was reported in response to the questionnaire, even taking in consideration the unverified rebate which was also not reported in the reply to the questionnaire as correct;

in the fourth place, the fact that Imarflex is "offering" four different SG & A [expenses] and profit rates varying from 11.29% as reported in the reply to 9.96% in Annex 10 of Acme's letter of 31 July 1995 to 13% in Annex 11 of the same letter and to 15% as used by the US authorities shows that there is clearly something wrong. In particular, all these "proposals" are still far below the margins between purchase prices and sale prices of each of the models sold.

As it is impossible for the Commission services to verify the different proposals of Acme on the spot, it can only be concluded that the SG & A [expenses] and profit rate verified initially were unreliable and that they have as such to be rejected.'

71 Finally, at recital 26 in the preamble to the contested regulation, the Council stated that Imarflex's data had turned out to be unreliable and that it had been `considered reasonable to maintain, as in the case of Malaysia, the general methodology for determining normal value for Thailand as outlined in recitals 46 and 47 of the provisional duty regulation'.

72 The applicant did not provide any evidence enabling those statements to be verified.

73 It is not denied that the applicant, in its reply to the questionnaire dated 4 February 1994, did not submit Imarflex's SG & A expenses and profit margin in accordance with the terms of the questionnaire and, in particular, did not break down those expenses by category or produce information relating exclusively to MWOs sold by Imarflex in Japan.

74 Although the Commission invited it to supplement its initial reply on those points, the applicant did not, in its letter of 22 February 1994, comply with the requirements of the questionnaire.

75 It is also common ground that information relating to the SG & A expenses and profit margin submitted by the applicant in its reply to the questionnaire was invalidated during the on-the-spot inspection. Having regard to the discrepancies found on completion of the verification of the costs relating to the `CMO 552' model, the Commission was entitled to cast doubt on the reliability of the applicant's replies.

76 It was only following the provisional disclosure letter that the applicant, by letters of 31 July and 3 November 1995, shed any light on Imarflex's SG & A expenses and profit margin.

77 It is therefore not contested that the applicant did not declare the after-sales repair costs, delivery costs and sales rebates as part of Imarflex's SG & A expenses, notwithstanding the terms of the questionnaire. It is also common ground that those costs affect the calculation of constructed normal value to a significant degree.

78 The applicant has expressly recognised that it committed an error in failing to declare delivery costs but attempts to justify its omission of after-sales repair costs by arguing that its accounting practices and those of Imarflex are sound. The costs relating to product repairs, incurred after the sale of the product, are by definition selling expenses within the meaning of Article 2(3)(b)(ii) of the basic regulation. It follows that those costs must be included in SG & A expenses when calculating the constructed normal value. The allegation that Imarflex and the applicant treat those costs as `cost of sales' for accounting purposes is irrelevant in that respect. The way in which undertakings treat certain types of costs when drawing up their annual accounts cannot call into question the classification of those costs for the purposes of an anti-dumping investigation. Furthermore, the applicant did not, either during the procedure or at the hearing, adduce any evidence to support that allegation.

79 With regard to sales rebates, Article 2(3)(a) of the basic regulation provides that discounts and rebates directly linked to sales may be deducted from the normal value provided that the exporter claims and supplies sufficient evidence that any such reduction from the gross price was actually granted. In this case, the applicant does not dispute that, in reply to the questionnaire, it deducted of its own accord certain rebates on sales, without providing any explanations or evidence in support of its action. Furthermore, it accepted in its pleadings and during the hearing that the rebates which were deducted unilaterally had been granted for cash. By their nature, cash reductions do not affect the price payable by Imarflex's customers but correspond to the value which Imarflex attributes to early payment of the invoiced price. They are not therefore deductible from the constructed normal value. It follows that the institutions were right to consider that those sales rebates should have been included as part of Imarflex's SG & A expenses.

80 The Court also observes that, in its statements following the provisional disclosure letter, the applicant `proposed' various mark-ups. Thus, having opted for a mark-up of 11.39% in its reply to the questionnaire, the applicant went on to propose further mark-ups of 9.96, 13 and 15% in its letter of 31 July 1995, and then of 11.86 and 21.75% in its letter of 3 November 1995.

81 Finally, in its reply, the applicant, under cover of explanations relating to its letter of 3 November 1995, put forward for the first time a new basis of calculation giving a mark-up of 22.26%.

82 Without there being any need to rule on other facts or on the good faith of the applicant, it is sufficiently clear from the foregoing that the institutions did not commit a manifest error of assessment in finding that the data relating to Imarflex and provided by the applicant in reply to the questionnaire was unreliable.

83 It follows that the applicant's arguments pertaining to the reliability of Imarflex's data must be rejected as unfounded.

The representative nature of the mark-up on the `CMO 552' model

- Arguments of the parties

84 The applicant questions the representativeness of the mark-up on the `CMO 552' model on the basis of which the Commission decided that Imarflex's data were unreliable. That model, which was selected by the Commission during the on-the-spot investigation, is the one with the highest mark-up and represents only 3.2% of all models sold by Imarflex.

85 The mark-up on that model is significantly higher than both the average mark-up of 22.26% on all MWOs sold by Imarflex in Japan and that of 16.08% on all products sold by Imarflex. That discrepancy is the result of the fact that delivery costs are higher for MWOs than for other products.

86 From that, the applicant concludes that the institutions made an obvious error of assessment in using the mark-up relating to the `CMO 552' model in order to reject Imarflex's data.

87 In reply, the Council states that the figures for the `CMO 552' model were provided by Imarflex managers during the on-the-spot investigation `by way of an example'. The Commission checked those figures on the spot and decided that they were accurate. The finding that the mark-up was higher than 30%, clearly contradicting the applicant's response to the questionnaire, was sufficient for the conclusion to be drawn that the figures previously submitted by the applicant were unreliable. Their unreliability was confirmed, inter alia, by the fact that the applicant, contrary to its earlier statements, was subsequently able to provide data relating only to MWOs.

- Findings of the Court

88 The file shows that the Commission examined the data relating to the `CMO 552' model by way of a sample, in order to verify the applicant's statements contained in its reply to the questionnaire. That verification enabled the investigators to find contradictions between that data and its reply to the questionnaire. It also appears that those contradictions are the result of the applicant's methodology in replying to the questionnaire and, in particular, declaring certain SG & A expenses. Those deficiencies are not therefore specific to the `CMO 552' model. They are general in nature and affect the reply to the questionnaire as a whole. Accordingly, the question whether the model selected by way of example was representative of the products exported to the Community is irrelevant to the validity of the overall assessment which the institutions were right to make as to the unreliability of the accounting data relating to Imarflex and submitted to the Commission for its appraisal.

89 It follows that the applicant's argument that the mark-up on the `CMO 552' model was not representative must also be rejected.

Imarflex's position as an exporter

- Arguments of the parties

90 The applicant emphasises that, to the extent that Imarflex was responsible for exporting its MWOs, it should have been considered to be the exporter of the products in question within the meaning of Article 2(3)(b)(ii) of the basic regulation. In that connection, it points out that the Commission designated Imarflex as the exporter in recitals 3 and 4 in the preamble to the provisional regulation and carried out an inspection at Imarflex's premises, so that it must have been certain that Imarflex was the exporter.

91 According to the applicant, it was not open to the institutions to disregard Imarflex's data on the ground that the MWOs were physically exported from Thailand. To interpret the `country of export' in Article 2(3)(b)(ii) of the basic regulation as referring only to the place to which the products covered by the investigation were physically exported would be contrary to the aims of the anti-dumping legislation.

92 Where, within a single economic entity, the functions of manufacture and export are each entrusted to two different companies, the institutions are bound to refer to the prices charged by the exporter. In this case, the institutions failed to take into account the manner in which manufacture and export of MWOs are distributed within the Nisshin Group. The applicant itself merely manufactured MWOs, whereas Imarflex carried out all the commercial functions of an exporter entailing SG & A expenses, such as negotiating prices and terms of sale, research and development, invoicing customers and receiving payments.

93 In the alternative, the applicant claims that the institutions should have used Imarflex's data by analogy.

94 The Council contends that under Article 2(6) of the basic regulation, a third country can be considered to be the country of export only if the goods are trans-shipped through its territory. The MWOs manufactured by the applicant were dispatched directly from Thailand to the Community and were never shipped through Japan. Furthermore, the applicant overstates the role of Imarflex as exporter. The applicant itself issues certain invoices relating to exports and bears distribution costs as well as substantial export expenses. The institutions were therefore entitled to conclude that Japan was not the country of export.

95 Furthermore, the institutions looked at Imarflex's data only in order to assess whether they could be used as `any other reasonable basis' within the meaning of the last part of Article 2(3)(b)(ii) of the basic regulation. The Council observes that, in any event, the institutions should not have taken Imarflex's data into account because such data were unreliable.

96 The French Government maintains that the last sentence of Article 2(3)(b)(ii) gives the institutions an alternative. They can use either `sales made by the exporter or other producers or exporters in the same business sector in the country of origin or export' or `any other reasonable basis', and there is no order of priority between the two methods. Accordingly, even if it were shown that the first method was reliable and its use more reasonable than the use of `any other reasonable basis', that would not mean that the contested regulation was vitiated by a manifest error of assessment such as to entail its annulment.

97 For the Commission, both the literal meaning of the word `exporter' and the basic regulation preclude the interpretation suggested by the applicant.

- Findings of the Court

98 The Court has already held that the institutions were entitled to conclude that the data relating to Imarflex were unreliable. Therefore, those data could not be used for the purposes of determining the constructed normal value, and whether or not Imarflex was the exporter, as the applicant maintains, is irrelevant in that connection.

99 It follows that the argument that Imarflex was the exporter must be rejected.

Reliability of Imarflex's data

- Arguments of the parties

100 In the applicant's view, the Commission could easily have verified the reliability of the information which it considered dubious. The institutions cannot treat the reply to the questionnaire as a formal legal document which can be rejected in its entirety once certain errors or gaps have been found in it. The questionnaire is no more than a tool allowing the parties to provide relevant data to the best of their ability within the short period of time allocated to them to fill it out.

101 At every stage of the procedure, the Commission should have allowed the applicant to provide such further particulars as it deemed appropriate. The applicant maintains that the institutions used the supplementary observations and alternative suggestions made by it as an argument to highlight the existence of contradictions and reject Imarflex's data. In the applicant's submission, that practice is unfair. Parties to anti-dumping proceedings should be able to explore alternative solutions at the administrative stage, without having to fear that their proposals will subsequently be held against them.

102 The Council contends that the purpose of the questionnaire is not to allow the manufacturer to state whatever he sees fit, and to place with the Commission the responsibility of enabling him to make good any deficiencies found by it. The questionnaire clearly indicated the consequences to which the parties were exposing themselves if they gave incomplete or incorrect answers. The Council maintains that, although the Commission can, as in the present case, point to deficiencies which affect the reply to the questionnaire and request further information, it is not the Commission's task to remedy them.

- Findings of the Court

103 The Court has already held that the institutions were entitled to reject the data relating to Imarflex on account of their unreliability.

104 Furthermore, the questionnaire contained a reference to the responsibility of the applicant and the risks incurred by it in the event of providing an incomplete answer, worded as follows:

`The purpose of this questionnaire is to permit the Commission to obtain the information it deems necessary for its investigation. It is in your own interest to reply as accurately and as completely as possible and to attach supporting documents. If all the information requested is not received by the Commission, within the time-limit specified, the latter may make preliminary or final decisions on the basis of the factual data available. If the Commission finds that false or misleading information has been supplied, it may disregard such information and, subsequently, any claim to which this information refers. In view of this, if you have any difficulties or for any reason you feel unable to complete any part of this questionnaire you should inform the Commission within 15 days of the date of the covering letter. The Commission's services will endeavour to assist you. Please state the difficulties encountered or clarification requested.'

105 Furthermore, it is clear from the applicant's letters of 31 July and 3 November 1995, and from its statements at the hearing, that it treated the investigation procedure as if it involved negotiations, issuing successive `proposals' on the level of Imarflex's SG & A expenses and profit margin. In that respect, it is appropriate to note that it concluded its letter of 3 November 1995 as follows:

`Our final proposal

As shown in the above, we requested the SG & A / profit of 11.39% (corrected as 11.86%) in our original reply. Later we proposed 13% and 15% to you. The difference among them come from the difference in approach. 11.86% was calculated based on the SG & A / profit of Imarflex as a whole. 13% was calculated based on the data of individual sales of microwave ovens. 15% came from the past example.

In view of a limited time and the data which could satisfy you as reliable data, we propose finally that you will calculate our constructed values by using SG & A/ profit of 21.75% as appropriate figures. This 21.75% comes from CMO-552 that you verified in Japan, and was disclosed in ANNEX 11 of our letter dated 31 July 1995. As far as this model is concerned, you know the actual data on its selling and purchasing prices. And the rebate and the delivery cost are near the data calculated as a whole in ANNEX 2. The repair cost is a reasonable amount to your knowledge about business.

We strongly hope that you will accept this proposal and settle this case amicably and reasonably.'

106 Those facts show that, by behaving in that way, the applicant failed to appreciate the nature of the investigation. In those circumstances, it cannot complain that the institutions did not carry out a second verification or failed to make adjustments of such a kind as to remove the defects affecting the reliability of the information which it initially submitted.

107 The applicant's arguments relating to the reliability of Imarflex's data must therefore be rejected as unfounded.

The unreasonableness of using the Korean data

Arguments of the parties

108 The applicant claims first of all that the contested regulation is contradictory and is not sufficiently reasoned. It points out that the institutions applied the Korean mark-up to a producer based in Malaysia, principally because it was connected to one of the Korean producers, even though it was exporting MWOs from Malaysia, without shipping them through Korea. The same logic should have led the institutions to refer to Imarflex's data, as Imarflex is an exporter connected with the applicant.

109 The applicant further maintains that the use of the Korean manufacturers' SG & A expenses and profit margin is not a reasonable basis for calculating the constructed normal value and infringes the principles of equal treatment and non-discrimination (Opinion of Advocate General Lenz in Nakajima v Council, cited above, [1991] ECR I-2112, point 85). The institutions exceeded the limits of their discretion when choosing the reasonable basis for calculating the constructed normal value and made a clear error of assessment (Nölle, cited above, and Ferchimex v Council, cited above). The institutions could not, in the applicant's submission, reasonably use the Korean producers' mark-up because their situation was so different from that of the applicant. In that connection, the applicant refers to two substantial differences.

110 First of all, the Korean producers differ from the Nisshin Group in size. For that reason, they bear administrative and general costs which are considerably higher than those of the Nisshin Group. That difference is apparent from a comparison of the respective share capital, turnover, and staff of the Korean companies Samsung, Daewoo and LG on the one hand, and of the Nisshin Group on the other. In that regard, the applicant produced an opinion of Professor Sekkat from the Free University of Brussels, according to which the turnover per employee is not a relevant indicator of the efficiency of Korean manufacturers.

111 Secondly, the applicant notes that the Korean manufacturers' distribution methods differ from its own. Certain Korean manufacturers sell MWOs on their domestic retail market, thereby incurring sales costs which are significantly higher than those of the applicant, which specialises in original equipment manufacturer (`OEM') exports, and of Imarflex, which sells 90% of its MWOs to a single customer in Japan.

112 The Council notes that the institutions used the data relating to the Korean market because this was the only market covered by the investigation in which profitable sales of similar products had been made in representative quantities. In addition, the Korean market was large and competitive.

113 It also considers that the data provided by the applicant on the Korean manufacturers is unspecific and incapable of verification. The turnover per employee achieved by those companies is higher than that of the Nisshin Group which suggests that economies of scale have positive effects and runs counter to the applicant's argument.

114 Furthermore, the Council rejects the argument based on differences in sales methods and notes in that connection that the purpose of determining the constructed normal value is to establish the SG & A expenses which the applicant would have borne had it sold MWOs on the Thai market. It observes that the institutions applied a specific adjustment to the applicant so as to take into account the fact that it exported significant quantities of MWOs, essentially to a single client within the Community.

Findings of the Court

115 The Court has already held that the institutions were entitled to reject the data relating to Imarflex. It follows that the applicant's argument that the institutions should have relied on Imarflex's data is irrelevant.

116 The principles of equality and non-discrimination, on which the applicant also relies, require that comparable situations should not be treated differently and different situations should not be treated in the same way, unless such treatment is objectively justified (Joined Cases C-133-93, C-300-93 and C-362-93 Crispoltoni and Others [1994] ECR I-4863, paragraph 51, and Thai Bicycle v Council, cited above, paragraph 96).

117 The last part of Article 2(3)(b)(ii) of the basic regulation permits the institutions to rely on accounting data from third parties subject only to the condition that the method employed was reasonable. It follows that the possibility of relying on accounting data from third parties whose situation is necessarily different from that of the companies targeted by the anti-dumping investigation is an inherent feature of the basic regulation and such reliance cannot therefore be considered to constitute in itself a breach of the principles of equality and non-discrimination.

118 It is apparent, moreover, from a reading of recitals 36 and 46 in the preamble to the provisional regulation that `the Commission considered it appropriate to take the domestic SG & A costs and profit established on profitable sales in Korea. This approach was considered reasonable as the Korean market was the only one covered by the present anti-dumping proceeding in which profitable sales of the like product were made in representative quantities. Furthermore, as outlined under (...) recitals 12 and 13, the Korean market is large and the economic operators concerned operate in a competitive environment.'

119 The applicant did not submit any evidence showing that there was a manifest error such as to undermine the validity of that assessment.

120 Furthermore, in this case the institutions made adjustments to the constructed normal value so as to take into account certain differences between the respective positions of the Korean producers and the applicant and, in particular, the physical characteristics of the MWOs in question, the import charges and the commercial stages at which sales are made (recital 27 in the preamble to the contested regulation and recital 50 in the preamble to the provisional regulation).

121 It must therefore be accepted that, by using the data relating to the Korean manufacturers in order to calculate the constructed normal value, the institutions did not make a manifest error of assessment or infringe the principles of equality and non-discrimination.

122 The arguments relied on by the applicant relating to the unreasonableness of using the Korean data must therefore be rejected.

123 It follows that the complaints directed at the choice of method for calculating the SG & A expenses and the profit margin are unfounded. The first and second pleas in law must therefore be rejected in their entirety.

3. Account taken of import duties and indirect charges when calculating the constructed normal value and comparison with the export price (fifth plea in law)

Arguments of the parties

124 The applicant challenges the Council's method of taking account of import duties and indirect charges when calculating the constructed normal value. It notes that the Council added the 35% import tax which was applied in Thailand to the cost of the raw materials. In order to calculate the SG & A expenses and profit margin, the Council increased the manufacturing costs by the mark-up derived from the Korean data. In so doing, the SG & A expenses and profit margin were increased by an amount equivalent to that yielded by the import duty multiplied by the mark-up.

125 Since import duties are reimbursed upon export, the Council deducted them from the constructed normal value. However, the applicant emphasises that no correction was made in order to neutralise the corresponding increase in the SG & A expenses and profit margin. That method, it claims, leads to an overestimation of the normal value and increases the dumping margin by more than 3%.

126 The applicant notes that Article 2(10)(b) of the basic regulation does not preclude the normal value from being reduced not only by the amount of import duties and indirect charges but also by the amount by which those duties and charges have increased the SG & A expenses and profit margin. The purpose of that article is to prevent the difference between the sale price on the domestic market (including indirect charges and import duties) and the lower export price, net of those duties and indirect charges, from being treated as amounting to dumping.

127 The applicant considers that the import duties should not have been incorporated in the calculation of the constructed normal value. Being established in a free-trade zone, the applicant imports materials free of import duties. However, the Council proceeded on the basis that the applicant had actually paid those duties which were then reimbursed to it upon export.

128 Finally, the applicant entertains doubts as to whether the Council's approach is well founded in so far as it adds the SG & A expenses and profit margin (including charges and duties paid in Korea) to the applicant's manufacturing costs, increased by Thai import duties which it never paid. It claims that when calculating the SG & A expenses and profit margin, the Council should have neutralised the impact of Korean import duties and indirect charges.

129 The Council notes that, even though the applicant was established in a free-trade zone, it would have had to pay import charges at the rate of 35% if it had sold MWOs in Thailand. Article 2(3)(b)(ii) of the basic regulation provides that costs of production are to be calculated on the basis of all costs incurred in the ordinary course of trade in the country of origin. In accordance with Article 2(10)(b) of that regulation, the institutions deducted the amount of import duties which had been included in the manufacturing costs from the constructed normal value. It was not necessary to make a further adjustment in order to neutralise the impact of import duties on the calculation of the SG & A expenses and profit margin. The SG & A expenses and profit margin calculated for the Korean companies were expressed as a percentage of the manufacturing costs - including indirect charges and import duties - of the MWOs bound for the Korean market. It was therefore necessary, for a proper comparison, to add those Korean data to the applicant's manufacturing costs, including indirect charges and import duties.

Findings of the Court

130 The construction of the normal value is aimed at determining the sale price of a product as it would be if that product were sold in its country of origin or export. Consequently, it is the costs relating to sales on the domestic market which must be taken into consideration (Nakajima v Council, cited above, paragraph 64). It is undisputed that, if the applicant had sold the MWOs in question on the Thai market, an import duty of 35% on the value of the items imported would have been payable. The institutions were therefore entitled to include the amount corresponding to payment of those import duties in the applicant's manufacturing costs.

131 Since the applicant's manufacturing costs include the import duties payable in Thailand, it was necessary to apply the Korean mark-up calculated on the basis of manufacturing costs which themselves include import duties or indirect charges payable in Korea. Any other method would destroy the symmetry between the Korean data (SG & A expenses and profit margin) and the Thai data (manufacturing costs of the applicant) of the constructed normal value.

132 In order to ensure that the comparison made between normal value and export price is valid, Article 2(9)(a) of the basic regulation provides for the possibility of making certain adjustments, inter alia, on the basis of differences relating to import charges and indirect taxes.

133 Those adjustments may not be made automatically; it is for the party which seeks to benefit from them to prove that they are justified (Case 255-84 Nachi Fujikoshi v Council [1987] ECR 1861, paragraph 33; Case T-171-94 Descom Scales v Council [1995] ECR II-2413, paragraph 66).

134 In addition, Article 2(10) of the basic regulation provides, inter alia, as follows:

`Any adjustments to take account of the differences affecting price comparability listed in paragraph 9 (a) shall, where warranted, be made pursuant to the rules specified below.

[...]

(b) Import charges and indirect taxes:

Normal value shall be reduced by an amount corresponding to any import charges or indirect taxes, as defined in the notes to the Annex, borne by the like product and by materials physically incorporated therein, when destined for consumption in the country of origin or export and not collected or refunded in respect of the product exported to the Community.'

135 It follows from those provisions that the institutions are not bound to deduct from the normal value an amount higher than that of the import charges or indirect taxes.

136 Therefore, by deducting from the normal value the amount of import duties which would have been payable if the applicant had sold the MWOs on its own domestic market, the institutions did not infringe Article 2(10) of the basic regulation.

137 It follows that the plea in law alleging breach of Article 2(10)(b) of the basic regulation must be rejected.

4. Inadequate statement of reasons (fourth plea in law)

Arguments of the parties

138 The applicant claims that the contested regulation is insufficiently reasoned because the Council did not examine one of its arguments. It denies making the argument referred to in recital 26 in the preamble to the contested regulation in which it `claimed that [the use of Imarflex's data] was in line with Article 2(6) of the basic anti-dumping regulation, as exports to the Community of MWOs produced in Thailand were actually shipped from Japan'. In the course of the investigation, it maintained that the Council should use Imarflex's SG & A expenses and profit margin on the Japanese market in order to determine the constructed normal value. Since the Council never replied to that argument, the contested regulation should be annulled.

139 The Council submits that it did not misunderstand the applicant's argument. In that connection, it refers to the wording of the final disclosure letter in which the Commission replied to all the arguments raised by the applicant.

140 It contends that the institutions were not bound, in the contested regulation, to reply to each of the arguments raised in the course of the procedure (Case 185-83 University of Groningen v Inspecteur der Invoerrechten en Accijnzen, Groningen [1984] ECR 3623, paragraph 38; Case 303-87 Universität Stuttgart v Hauptzollamt Stuttgart-Ost [1989] ECR 705, paragraph 13; and Case 246-86 Belasco and Others v Commission [1989] ECR 2117, paragraph 55).

Findings of the Court

141 It is settled case-law that the statement of reasons required by Article 190 of the Treaty must show clearly and unequivocally the reasoning of the Community authority which adopted the contested measure, so as to inform the persons concerned of the justification for the measure adopted and thus to enable them to defend their rights and the Community judicature to exercise its powers of review. However, regulations cannot be required to specify the often very numerous, complex and pertinent matters of fact and law dealt with, because the question whether they comply with the requirements of Article 190 of the Treaty must be assessed in the light not only of their wording but also of the context and procedure in which they were adopted and the body of legal rules governing the field concerned (Case 203-85 Nicolet Instrument v Hauptzollamt Frankfurt am Main-Flughafen [1986] ECR 2049, paragraph 10; Case 240-84 NTN Toyo Bearing and Others v Council [1987] ECR 1809, paragraph 31; Nachi Fujikoshi v Council, cited above, paragraph 39; Case T-2-95 IPS v Council [1998] ECR II-3939, paragraph 357).

142 That requirement is satisfied in this case. The statement of reasons in the contested regulation and the references to the provisional regulation set out to a sufficient standard the reasons for which the data relating to Imarflex were rejected. Moreover, the institutions replied to the arguments put forward by the applicant during the investigation, particularly in the provisional and definitive disclosure letters. It follows that the statement of reasons enabled the applicant to ascertain the reasons why Imarflex's data were not used as a basis for calculating the constructed normal value and enabled the Court to exercise its power of review.

143 The plea in law alleging breach of Article 190 of the Treaty must therefore be rejected.

144 It follows from the foregoing that the action must be dismissed in its entirety.

Costs

145 Under Article 87(2) of the Rules of Procedure of the Court of First Instance, the unsuccessful party is to be ordered to pay the costs if they have been applied for in the successful party's pleadings. Since the Council has asked for the applicant to be ordered to pay the costs, the applicant should be ordered to bear its own costs and to pay the costs incurred by the Council.

146 The first subparagraph of Article 87(4) of those Rules of Procedure provides that Member States and institutions which intervened in the proceedings are to bear their own costs. The Commission and the French Republic must therefore be ordered to bear their own costs.

On those grounds,

THE COURT OF FIRST INSTANCE

(Fifth Chamber, Extended Composition)

hereby:

1. Dismisses the action;

2. Orders the applicant to bear its own costs and to pay those of the Council;

3. Orders the Commission and the French Republic to bear their own costs.