C-76/00

OPINION OF MR JACOBS — CASE C-76/00 P

OPINION OF ADVOCATE GENERAL JACOBS delivered on 25 April 2002 1

1. This appeal arises out of the dismissal by Petrotub the Court of First Instance 2of the appli- cations of two Romanian producers of seamless pipes and tubes of iron or non- alloy steel for the partial annulment of a Council Regulation 3 imposing definitive anti-dumping duties on, inter alia, their products. The two actions were joined before the Court of First Instance and the The underlying issue two producers have brought a joint appeal in which they both allege that that Court erred in law in its interpretation and application of the requirement to state adequate reasons for a measure. None the less, it is preferable to deal with the 2. Although the appeal concerns the obli- arguments of each appellant separately, gation to state reasons and must be decided since they concern separate portions of on that basis, it may be helpful to set out the judgment under appeal. first the underlying issue, that of the calculation of dumping margins by the 'asymmetrical method', with the concomi- tant practice of 'zeroing' — a matter of some controversy in world trade.

anti-dumping duty may be applied to any such price is available, constructed on the dumped product whose release for free basis of specified rules). Article 2.4.2 pro­ circulation in the Community causes injury' vides that '... the existence of margins of and, under Article 1(2), 'a product is to be dumping during the investigation phase considered as being dumped if its export shall normally be established on the basis price to the Community is less than a of a comparison of a weighted average comparable price for the like product, in normal value with a weighted average of the ordinary course of trade, as established prices of all comparable export trans­ for the exporting country'. actions or by a comparison of normal value and export prices on a transaction-to-trans- action basis. A normal value established on a weighted average basis may be compared to prices of individual export transactions if the authorities find a pattern of export prices which differ significantly among different purchasers, regions or time periods, and if an explanation is provided as to why such differences cannot be taken into account appropriately by the use of a 4. The 1995 reenactment was prompted weighted average-to-wcighted average or both by inadequacies in the previous legis­ transaction-to-transaction comparison.' lation 5 and, perhaps more importantly, by the establishment of the '1994 Anti-dump­ ing Code' as part of the Uruguay round of multilateral trade negotiations. 6

basis. However, a normal value established models of his product to the Community, on a weighted average basis may be com- model A and model B: pared to prices of all individual export transactions to the Community, if there is a pattern of export prices which differs significantly among different purchasers, regions or time periods, and if the methods specified in the first sentence of this para- graph would not reflect the full degree of dumping being practised....' — For model A, the weighted average normal value is calculated to be 100, from a number of domestic trans- actions at prices ranging from 95 to 105. Two export transactions are made at a dumped price 8 of 80, and 10 at an undumped price of 110. 7. Those provisions are thus substantially the same, each providing for the same three possible methods of calculation, although the definition of the circumstances in which the third method may be used differs — 'if an explanation is provided as to why such differences cannot be taken into account appropriately by the use of [the first two methods]' as opposed to 'if the [first two — For model B, the weighted average methods] would not reflect the full degree normal value is calculated to be 90, of dumping being practised'. It may there- from a number of domestic trans- fore be worthwhile considering what the actions at prices ranging from 85 to three methods involve. 95. Ten export transactions are made at a dumped price of 70, and five at an undumped price of 100.

8. For that purpose, it is helpful to look at some actual (though totally fictional) fig- ures. A simple example 7might involve a third-country producer exporting two 9. A comparison of the weighted average normal value with the weighted average of

1 200 divided by 15 (total number of export transactions) = 80

(2 x 80) + (10 x110) = 160 + 1 100 = 1 260

Compared to the weighted average normal value of 90, this shows a ('positive') dumping margin of 150 for all 15 transactions, averaging at 10 per transaction.

1 260 divided by 12 (total number of export transactions) = 105

— If an overall margin is calculated for both models, the negative dumping margin of -60 is offset against the positive margin of 150 to give a total positive dumping margin of 90, or 3.33 Compared to the weighted average per transaction. normal value of 100, this shows no dumping; there is in fact a 'negative dumping margin' of -60 9 for all 12 transactions, averaging at -5 per trans- action. 10. Leaving to one side for the moment the rarely-used method of comparing prices on 9 — I shall leave dumping margins in absolute terms, although a transaction-to-transaction basis (the 'sec- they are normally expressed as a percentage of c.i.f. prices, which I am omitting from these examples. ond symmetrical method'), I turn to the

I - 87

11. It thus appears that the two methods achieve the same result if, as I have assumed, the same export transactions, both dumped and undumped, are all taken fully into account. However, the trans- actions used need perhaps not be the same (2 x 20) + (10 x -10) = 40 + -100 = 40 in both cases, since the first symmetrical -100 = -60 (average -5). method is stated to apply to 'all com- parable transactions' whereas the asym- metrical method compares average normal prices with 'individual' export transactions, which does not appear to preclude a selection by the investigating body. In the event of any such selection, it might seem plausible that the aim and effect would be — For model B (normal value = 90), the to exclude certain undumped exports and 10 transactions at 70 show a positive thus increase the resulting positive dumping margin of 20, the five at 100 show a margin, rather than the contrary. negative margin of 10. The total mar- gin is:

12. Moreover, calculations may be affected by a further technique known as 'zeroing'. (10 x 20) + (5 x -10) = 200 + -50 = 200 This involves reducing all negative dump- -50 = 150 (average 10). ing margins to zero, instead of offsetting

I-88

PETROTUB AND REPUBLICA v COUNCIL

them against the positive dumping margins. 0 + 150 = 150, instead of 90 (an All undumped exports are thus treated as if average of 5.56 per transaction, instead they were made at normal value. 10 When of 3.33) 'zeroing' is applied, the results produced by the two methods may differ. I shall again assume that the calculations apply to all the transactions for each model.

— If an overall margin is established for both models on the basis of the separ- ate margins established for each, it is — For both models together, now

10—If an anti-dumping duty is applied, however, it applies equally to undumped exports, although there is a (cum- bersome} procedure, under Article 11(8) of the basic regulation, whereby importers may request a refund of anti-dumping duties if they can show that the dumping (12 x 20) + (15 x -10) = 240 + - 1 5 0 = margin has been reduced or eliminated. 240 + 0 = 240 (average 8.89).

I -89

OPINION OF MR JACOBS — CASE C-76/00 P

15. Thus, it appears, where 'zeroing' is not B above) if there is 'targeted' dumping — applied, the two methods produce the same that is to say if an exporter sells at a loss in result if the same transactions are all taken a particular region or during a particular fully into account but, where it is applied, period because he believes he can increase the asymmetrical method will always pro- his sales locally or periodically without duce a higher result, since the first sym- endangering sales in other regions or during metrical method, by averaging export other periods. 11 prices, is bound to take negative dumping margins into account to a certain extent while the asymmetrical method will never take any into account. (Obviously, the issue will arise in any event only when there is a mixture of dumped and undumped exports; if all exports are dumped, there will be no negative margins and no 'zeroing'.)

17. The controversy over 'zeroing' has been considered by the Dispute Settlement Body of the World Trade Organisation (WTO) in cases involving the European Community. In 1995, in the Cotton Yarn case, that body's predecessor appeared to accept the validity of the technique used in conjunc- tion with the asymmetrical method. 12 In 1997, however, in Bed Linen, the reports of both the panel and the appellate board roundly condemned the practice in a case involving the first symmetrical method, 13

and it appears that the Community no comparison of individual normal values longer uses it in conjunction with that and individual export prices to the Com- method. 14 munity on a transaction-to-transaction basis. This method is less easy to illus- trate — and, as the Council and Commis- sion have stressed, to apply — because it involves identifying individual domestic transactions which are comparable to indi- vidual export transactions. Thus, it is 18. It must be stressed at this point that the necessary to take into account factors such practice of 'zeroing' is not itself an issue in as quantity, date and so forth. The dump- the present proceedings, although it may ing margin revealed for each export trans- still be an integral part of the asymmetrical action will differ depending on the method as applied by the Community and domestic transaction with which it is com- neither the Council nor the Commission pared, so close comparability is extremely has denied that it was in fact used in this important. If that condition can be met, case. What is in issue is whether adequate however, it might seem likely that the reasons were given by the Council in the second symmetrical method would present contested regulation for its choice of the a truer picture of the dumping practised asymmetrical method and, more particu- than either of the other methods. larly, whether the Court of First Instance was right to find that adequate reasons had been given. Nevertheless, those matters can be more reliably assessed if the con- sequences entailed by the choice of method are borne in mind.

I - 91

OPINION OF MR JACOBS — CASE C-76/00 P

The case at first instance tions failed to explain why a compari- son of the weighted average normal value with the prices of individual exports reflected the real extent of the dumping better than the symmetrical 21. On 31 August 1996 the Commission methods, a failure which, was inter initiated an anti-dumping proceeding con- alia, in breach of Article 2.4.2 of the cerning imports of certain seamless pipes 1994 Anti-dumping Code. and tubes of the types in issue originating in Russia, the Czech Republic, Romania and the Slovak Republic. In Regulation No 981/97 ('the provisional regulation'), 15 it imposed a provisional anti-dumping duty on imports from, inter alia, Petrotub and Republica, on which the rate of duty was calculated by the asymmetrical method. The findings were confirmed, and a defini- tive anti-dumping duty was imposed, by the 105 Although, according to settled case- Council in the contested regulation. law, the provisions of the basic regu- lation must be interpreted in the light of the 1994 Anti-dumping Code (Case C-69/89 Nakajima v Council [1991] ECR1-2069, paragraphs 30 to 32), the rules governing anti-dumping meas- 22. In Case T-33/98, Petrotub sought the ures are contained in that regulation annulment of Article 1 of the contested alone. The obligation referred to in regulation in so far as it was concerned. In Article 2.4.2 of the 1994 Anti-dump- one of its pleas it alleged that the Council ing Code to explain why the sym- had infringed Article 2(11) of the basic metrical methods of comparison can- regulation by, inter alia, failing to provide not show the real extent of the dump- justification for its choice of the asym- ing does not therefore, as such, con- metrical method in order to determine the stitute a rule which is to be applied, anti-dumping margin. The Court of First and Article 2(11) of the basic regu- Instance dealt with that aspect in para- lation clearly does not mention any graphs 104 to 115 of the judgment under specific obligation to give such an appeal, as follows: explanation.

lation is inadequate, it should be with the adjusted individual export borne in mind that the statement of prices in accordance with Article 2(11) reasons required by Article 190 of the of the basic regulation. This was EC Treaty (now Article 253 EC) must necessary in order to reflect the full show clearly and unequivocally the degree of clumping being practised reasoning of the Community authority and because there was a pattern of which adopted the contested measure, export prices which differed signifi- so as to inform the persons concerned cantly between different customers of the justification for the measure and regions." adopted and thus enable them to defend their rights and the Commu- nity judicature to exercise its powers of review. The extent of the obligation to state reasons must be assessed in the light of the context and the procedure in which the contested regulation was adopted and the body of legal rules governing the field concerned (see, most recently, Case T-48/96 Acme Industry v Council [1999] ECR It maintained that view in the pro- II-3089, paragraph 141). visional disclosure of 2 June 1997.

— for one company, there was no difference in dumping margin between both methods as all export transactions were made at 111 In its final observations on dumping of dumped prices; 8 September 1997, the applicant con- tended that the dumping margin should be determined by applying the weighted average to weighted average method of comparison. — for three companies, a pattern of export prices which differed sig- nificantly by destination or time period was found.

112 In recital 22 of the contested regu- lation, the Council stated:

In view of the above, and in accord- ance with Article 2(11) of the basic regulation, the method comparing the weighted average normal value by "One company claimed that the cal- time period to individual adjusted culation of the dumping margin export prices on a transaction-by- should not be made on the basis of a transaction basis was retained for the comparison of weighted average nor- purposes of the definitive deter- mal values with the adjusted export mination."

I - 94

PETROTUB AND REPUBLICA v COUNCIL

113 The contested regulation thus sets out with individual export prices) might the reasons for which the Community not reflect the real extent of the institutions decided to apply the cri- dumping engaged in, the Court finds terion of comparison of weighted that this is a separate plea in law average normal value with the prices which was not raised until the stage of of individual exports. the reply. This plea must therefore be rejected as inadmissible pursuant to Article 48(2) of the Rules of Pro- cedure.'

I - 95

OPINION OF MR JACOBS — CASE C-76/00 P

insufficient merely to refer to a provision or 27. Article 2(11) allows the asymmetrical repeat its wording when the provision or method 'if the methods' — in the plu- wording involves a legal or factual assess- ral — 'specified in the first sentence of this ment which is essential for the decision to paragraph would not reflect the full degree be taken; the institution must at least of dumping'. The Council should thus have indicate the logical link between the provi- given reasons for discarding both sym- sion and the assessment, so that the Court metrical methods before applying the asym- can verify that the assessment was made on metrical method. Although that issue came a correct basis and that the institution has to the fore at the stage of the reply at first not merely referred to the text as a self- instance, it was implicit in the use of the justifying argument. plural 'methods' in the application and should in any event have been dealt with by the Court of First Instance pursuant to its obligation to apply Article 2(11) correctly. The Court of First Instance was therefore wrong to dismiss the argument as inadmiss- ible at paragraph 115 of the judgment under appeal and to find that the statement of reasons was adequate.

25. Second, a mere reference to the exist- ence of a pattern of export prices which differs significantly by region or time period is not itself an adequate statement of reasons. Before using the asymmetrical method, the institutions must also satisfy themselves that the symmetrical methods 28. Petrotub's second, third and fourth would not reflect the full degree of dump- arguments concern the finding in para- ing, and adequate reasons must be given to graph 114 of the judgment under appeal establish the logical link between that the decision to use the asymmetrical Article 2(11) of the basic regulation and method was adequately reasoned, in so far the institution's decision. as that conclusion was based on the circumstances set out in paragraphs 108 to 113.

30. Paragraph 110 refers to the statement manœuvres in which dumping is disguised that, as regards Petrotub, the pattern of by charging different prices, some above export prices differed considerably as the normal value and some below it' 1 7— between the two periods considered and implying conduct chosen specifically to that, for all the Romanian companies taken disguise dumping or at least having no together, the difference between the results other reasonable explanation. Yet price of the first symmetrical method and the differences are often caused by changes in asymmetrical method was such that it- market conditions or differences in bar- could be concluded that the former did gaining power vis-à-vis different customers, not reflect the full extent of the dumping. In and the asymmetrical method is not justi- recital 22 of the contested regulation, cited fied. Targeted dumping, however, would at paragraph 112 of the judgment under justify use of that method, with 'zeroing' so appeal, the Council implicitly based its that negative dumping margins on other reasoning on the fact that the asymmetrical sales do not disguise the positive margins method produced an arithmetically greater on the dumped sales. result. That was not, however, an adequate statement of reasons.

Court of First Instance thus erred in law in substance of the contested regulation in the paragraph 105 of the judgment under form of a challenge to its reasoning. appeal by failing to take account of Second, in several regards it fails to specify Article 2.4.2 of the 1994 Anti-dumping the rule of law which the Court of First Code. Instance is alleged to have infringed. Third, it merely reproduces arguments already raised at first instance without challenging the legal assessment in the judgment under appeal. Fourth, it raises again an argu- ment — as to the need to provide reasons 34. In the same 1996 communication to the for not using the second symmetrical WTO, the European Communities stated method — which was already inadmissible that the term 'full degree of dumping', used at first instance because it was raised too in the basic regulation but not in the 1994 late in the proceedings. Fifth, it bases a new Anti-dumping Code, simply refers to tar- argument on evidence — the 1996 state- geted dumping. The reasons given in the ment to the WTO — which was not parts of the contested regulation on which adduced at first instance, without providing the Court of First Instance relied do not any justification for that omission. provide any explanation of how the alleged dumping was 'targeted'.

37. I disagree with those submissions. Admissibility of the appeal

sidiary pleas, that the Court of First mitted to the Court of First Instance is Instance erred in law by finding (i) that inadmissible. 19 However, that does not Article 2.4.2 of the 1994 Anti-dumping seem to be the case here. Moreover, that Code was not a rule which had to be ruling has been applied above all where the applied and (ii) that Petrotub's plea con- arguments put forward do not identify the cerning the second symmetrical method contested aspects of the judgment under could not be entertained because it was appeal but merely make an imprecise claim out of time. of illegality coupled with a blanket refer- ence to, or reproduction of, the submissions made at first instance. In the present case, there is a clear allegation that the Court of First Instance erred in specific aspects of its assessment. The justifiable concern to ensure that an appeal is not merely a 'second bite at the cherry' should not blind us to the fact that there must of necessity be considerable overlap between the argu- 39. I have no difficulty in seeing those ments advanced at first instance and those arguments as identifying specific aspects of raised on appeal. 20 the judgment under appeal which are con- tested, together with the legal grounds on which they are contested, as errors in law. Admittedly the appellant's arguments are not set out as clearly as they might have been for that purpose, but the Court is not thereby unduly hampered in its task. In so far as the Council claims that the appellant is disputing the reasons themselves rather than the adequacy of their statement — thus in fact challenging the substance of the measure — I consider it preferable to 41. With regard to the need to state reasons examine the reasoning in detail before for not applying the second symmetrical forming a view on the nature of the method, and in particular the admissibility challenge. of the plea at first instance, Petrotub advances a two-pronged argument: first, the plea was already present in the appli- cation and thus not raised too late and, second, the Court of First Instance should have examined it in any event because it relates to an essential procedural require- ment and is thus a question of public policy. 40. As regards the allegation of 'mere repetition' of arguments already raised at 19 — See, most recently, the Order of the Court in Case first instance, it is true that the Court has C-351/99 P Eridama and Others v C o u n c i l [ 2 0 0 1 ] ECR I-5007, paragraph 36. consistently held that an appeal which 20 — See also paragraphs 29 and 30 of my Opinion in Case simply repeats or reproduces verbatim the C-154/99 P Politi v ETF [2000] ECR I-5019 with which the Court apparently agreed at paragraph 12 of its pleas in law and arguments already sub- judgment.

43. Finally, I consider that the 1996 com- munication to the WTO is a document to which the Court's attention may be drawn 'Regulations, directives and decisions... without thereby rendering any argument of shall state the reasons on which they are the appellant inadmissible. The argument based...' The obligation therefore applies raised on appeal is that the Court of First equally to a decision — which is, accord- Instance was wrong to consider that ing to Article 189 of the EC Treaty (now Article 2.4.2 of the 1994 Anti-dumping Article 249 EC), 'binding in its entirety Code was not a rule to be applied; the upon those to whom it is addressed' — Court of Justice may decide that issue, if and to a regulation — which has 'general necessary, by reference to that article alone. application' and is 'binding in its entirety

21 — See Case C-367/95 P Commission v Sytraval and Brink's 22 — For an excellent and thoughtful review of the subject in a France [1998] ECR I-1719, paragraph 67 of the judgment, comparative context, see Martin Shapiro, 'The giving and Case C-265/97 P VBA v Florimex [2000] ECR I-2061, reasons requirement', University of Chicago Legal Forum paragraph 114. 1992, p. 179.

I - 100

PETROTUB AND REPUBLICA v COUNCIL

23 — See, most recently, the judgment of 7 February 2002 in Case C-328/00 Weber ECR I-1461, paragraph 42; cf. also paragraph 106 of the judgment under appeal and, for a 25 — Case 5/67 liens [1968] ECR 83, at p. 95. recent instance with regard to anti-dumping measures, 26 — See, for example. Case C-75/92 Cao Yao v Council [1994] Case T-87/98 International Potash Company [2000] ECR ECR I-3141, paragraph 26 of the judgment. II-3179, paragraph 65. 27 — See, in a slightly different context, Case C-478/93 Nether- 24 — See, for example, Case 24/62 Germany v Commission lands v Commission [1995] ECU I-3081, paragraphs 49 [1963] ECR 6 3 , at p. 69. and 50 of the judgment.

I - 101

OPINION OF MR JACOBS — CASE C-76/00 P

51. The duty to state reasons in the present case must thus also be viewed in the light of Article 2.4.2 of the 1994 Anti-dumping Code, which allows the use of the asym- metrical method only 'if an explanation is provided' why differences appearing in a 49. Furthermore, the procedure for adopt- pattern of export prices cannot be taken ing anti-dumping measures is set out in into account appropriately using the first or considerable detail in the basic regulation, second symmetrical methods. In the con- which lays down a series of steps to be text of such an instrument of public inter- followed in the normal course of events, national law, I find it difficult to conceive together with alternative steps which may that the explanation in question can have be taken in specified circumstances where a been intended to allow the investigating departure from that normal course is authority merely to state its reasons pri- warranted. vately — still less implicitly — to those interested parties who were involved in the investigation.

50. That procedure derives moreover in a large measure from the 1994 Anti-dumping 52. Such factors increase rather than limit Code, which is both binding on the Com- the extent to which reasons must be stated munity and specifically referred to in the for any contested steps in the procedure or any departure from the normal procedure.

28 — Case 24/62 Germany v Commission, cited above in note 24, and Case 294/81 Control Data v Commission [1983] 29 — See the judgment in Nakajima, paragraphs 30 and 31, ECR 911, paragraphs 14 and 15. cited in paragraph 105 of the judgment under appeal.

54. Moreover, not all the detailed rules set out in the basic regulation will be relevant- in every case. Certain steps are always 56. Where a standard procedure is thus laid essential, the need for others will depend on down with alternative procedures to be the circumstances. It is for example essen- followed only in certain specified circum- tial that the Community institutions deter- stances, it seems reasonable to require no mine the existence both of dumping and of specific statement of reasons when follow- injury to the Community industry, since ing the standard procedure 30 but a rather those, together with the causal relationship fuller one when departing from it. The between them, are the fundamental con- ditions for the imposition of a duty, and it 30 — See, for example, Case C-315/90 Cimelec v Commission must be clear from the regulation that their 11991] ECR I-5589, paragraph 27 of the judgment.

I - 103

OPINION OF MR JACOBS — CASE C-76/00 P

I - 104

PETROTUB AND REPUBLICA v COUNCIL

ested party to ascertain the circumstances which the Commission founded its choice, in which the rules have been applied. 32 and thus cannot elucidate the scarcely more informative statements in the contested regulation. The only additional indication which may be inferred from the latter — and which seems clear from the Commis- sion's final disclosure referred to in para- graph 110 of the judgment under appeal — 60. The reason for choosing the asym- is that the first symmetrical method was metrical method given in the provisional judged inadequate to reveal the full extent regulation is that it 'was necessary in order of dumping because it produced a lower to reflect the full degree of dumping being margin. (It may be noted here that the only practised and because there was a pattern new reason given explicitly in the contested of export prices which differed significantly regulation — that the choice of method between different customers and regions'. made no difference to the result in the case In the contested regulation, the reasons of one company — should in fact be a given for not acceding to the request that reason for not departing from the normal the first symmetrical method be used are basis of comparison.) that 'for one company, there was no difference in dumping margin between both methods as all export transactions were made at dumped prices' and that 'for three companies, a pattern of export prices which 63. At first sight, therefore, the contested differed significantly by destination or time regulation does not contain any statement period was found'. from which the Court or any interested party could ascertain the reasons for which the asymmetrical method was used. It is however necessary to examine the three specific points on which adequate reason- ing was or may have been required. To what extent must reasons be stated on 61. Is such reasoning adequate with regard those points and, in the light of the whole to the context and to all the legal rules context, were such reasons given? governing the matter?

68. Nor is it relevant that Petrotub did not ask for any explanation in the course of the 66. The Council and the Commission have preliminary procedure. Determination of argued that, because there is no indication the dumping margin is an essential step in as to which of the symmetrical methods is the procedure, and it is necessary to give to be used in which circumstances, they the reasons for any departure from the may choose freely between them. To that normal basis of comparison for that pur- extent I agree. I do not however accept pose. In the circumstances, the relevant their conclusion that it is sufficient to explanation could have been succinct, but explain why the symmetrical method it could not be omitted. chosen was unsatisfactory in order to proceed to the use of the asymmetrical method. That is simply not consistent with the scheme or wording of the provisions, which clearly allow use of the asymmetrical method only if neither symmetrical method would 'reflect the full degree of dumping' 69. I therefore take the view that, by not or allow differences in the pattern of export explaining why the second symmetrical prices to be 'taken into account appropri- method was not appropriate, the Council ately'. failed to state adequate reasons for depart-

I - 106

PETROTUB AND REPUBLICA v COUNCIL

ing from the normal basis of comparison or time period was found' is not the same laid down in Article 2(11) of the basic as the statement in the provisional regu- regulation. Since the statement of such lation that 'there was a pattern of export reasons is an essential procedural require- prices which differed significantly between ment, the Court of First Instance should different customers and regions'. The have examined the question regardless of apparent discovery of significant differ- the stage in the proceedings at which it was ences by time period between the pro- raised. visional and the final regulation calls, prima facie, for an explanation.

I - 107

OPINION OF MR JACOBS — CASE C-76/00 P

74. From what the Council has stated in its 77. On a purely formal level, that view is pleadings and at the hearing, the fact that perhaps défendable. However, the border- the first symmetrical method produced a line between the adequacy of a statement of significantly lower dumping margin than reasons and the validity of the assessment the asymmetrical method seems to have which it embodies cannot be entirely clear- been the decisive reason for concluding that cut in a formal manner. Where no reasons the former did not reflect the full degree of at all are given for departing from the dumping, and no other reason appears to normal basis of calculation, there is clearly have been given at any stage. a failure to fulfil the obligation imposed by Article 253 EC. But simply to state, for example, that the Commission's computer program determined that the asymmetrical method was necessary would fall in the same category, since it would again give no indication from which to ascertain whether the departure from the normal basis was 75. That reason is not stated explicitly in justified. In those circumstances, even the contested regulation, but it may be though a formal reason has been given, inferred from the statement that 'for one the proper challenge would be to its company, there was no difference in dump- adequacy as a statement of reasons and ing margin between both methods' as not to its validity as a correct assessment. justification for rejecting the claim that the first symmetrical method should have been used and, more clearly, from the Commission's statement in its final disclos- ure that the difference in dumping margin obtained by the two methods was such that it could be concluded that the first sym- metrical method did not enable the real extent of the dumping to be reflected.

79. In that regard, it may be relevant that 82. It may consequently be unnecessary to if — as is apparently systematically the consider specifically whether, in addition to case — 'zeroing' is used with the asym- being inadequate in the light of metrical method, it seems to lead auto- Article 2(11) of the basic regulation, the matically to a higher result whenever some statement of reasons was also inadequate in transactions with a negative dumping mar- the light of Article 2.4.2 of the 1994 gin are included in the calculation. 35 Anti-dumping Code in that no explanation However, the mere fact that some trans- was given as to why the differences found actions show a negative margin does not in the pattern of export prices could not be necessarily mean that there has been tar- taken into account appropriately by the use geted dumping and it is therefore necessary of either of the symmetrical methods. That to indicate, for this purpose also, the requirement is in my view in practice very grounds on which it was concluded that much the same as the obligation, when such targeting existed. departing from the normal basis of comparison, to state why that normal basis would not reveal the full — true — extent of dumping. None the less, the fact that the reasons which must be stated flow not only from the basic Community regulation but also from an international instrument bind- ing on the Community can only strengthen the obligation. 80. Be that as it may, a mere assertion that the method producing the higher result was used does not in any event enable the Court or anyone else to decide whether there might have been a manifest error of assess- ment. Thus, the proper challenge is not, as the Council argues, that the wrong test was — Conclusion used, but that insufficient reasons were stated in order to determine whether the right test was used.

Background, proceedings at first instance, judgment under appeal '68 The applicant maintains that its domestic sales made on the basis of compensatory arrangements accounted for some 24% of domestic sales of the product concerned during the investi- 84. Republica's appeal concerns the calcu- gation period. The compensation sys- lation of normal value on the domestic tem is imposed on the applicant by market, prior to comparison with export major clients, such as Romanian util- prices for the purpose of determining ities companies, and the prices charged dumping margins. Under Article 2(1) of under that system, which are non-ne- the basic regulation, normal value is gotiable, are considerably lower than normally based on the prices paid or normal market prices. Those sales payable, in the ordinary course of trade, ought therefore to have been excluded by independent customers in the exporting for the purposes of determination of country. However, under the third sub- normal value in accordance with the paragraph: third subparagraph of Article 2(1) of the basic regulation.

69 The applicant made no reference to 'Prices between parties which appear to be those sales in its reply to the ques- associated or to have a compensatory tionnaire because it was never asked to arrangement with each other may not be do so. It admits raising the question of considered to be in the ordinary course of sales under compensatory arrange- trade and may not be used to establish ments for the first time during the

I - 110

PETROTUB AND REPUBLICA v COUNCIL

86. It might be wondered at this point what motivated Republica to argue that sales made on the basis of compensatory arrangements should be excluded from the 'One company claimed only at a very late calculation since, as stated in the extract stage in the proceeding (at the hearing for just cited, the sales in question, some 24% comments on provisional disclosure) that... of domestic sales, were at considerably all sales made using compensation as means lower than normal market prices. If they of payment should... be excluded as not had been excluded, that should have led to being in the ordinary course of trade. [That a higher normal value and thus a higher claim was] not made in an opportune dumping margin, contrary to Republica's manner, having been made neither in the apparent interest. The issue underlying response to the questionnaire, nor on-the- Republica's claim is thus far from being spot, nor at any subsequent stage of the as clear as that underlying Petrotub's claim. proceeding when the company was invited to make comments.... during the course of the investigation, it was found that sales made using compensation were indeed made in the ordinary course of trade. Consequently, [the claim was] rejected.'

87. However, a possible explanation was suggested by the Council at the hearing: exclusion of those sales might have meant 89. In Case T-34/98, Republica sought that there were insufficient sales in the annulment of the contested regulation in

I - 111

OPINION OF MR JACOBS — CASE C-76/00 P

I -112

PETROTUB AND REPUBLICA v COUNCIL

on the basis of compensatory arrange- to exclude sales made using compensation ments made during the investigation from the determination of normal value. period, affected the prices charged in such transactions, as required by the third subparagraph of Article 2(1) of the basic regulation.

93. Once it had accepted that it was dealing with 'sales made using compensa- 75 Moreover, in the absence of any tion' and thus with 'prices between parties contrary indication from the applicant, which appear to be associated or to have a the Council gave, in the contested compensatory arrangement with each regulation, an adequate statement of other', the Council was not free to use the reasons for its refusal to exclude such prices to establish normal value with- compensatory sales from the deter- out the proof required by the basic regu- mination of normal value, by stating lation that such prices were 'unaffected by that "it was found that sales made the relationship'. The Court of First- using compensation were indeed made Instance therefore erred in finding that, in the ordinary course of trade".' because '... the applicant has produced no evidence to show or any reason to conclude that the compensatory arrangements... affected the prices charged in such trans- actions...', the Council's statement of rea- sons was adequate.

The appeal

97. The Council, supported by the Com- mission, considers that the statement of 96. For essentially the same reasons as reasons in the contested regulation was those I have set out with regard to Petro- adequate; the aspect in question was not tub, 39 I disagree with those objections. At challenged during the administrative pro- first instance, Republica argued that the cedure, and Republica is wrong in its contested regulation failed to state contention that the logical link between adequate reasons for finding that the trans- the text and the assessment must be actions in issue were made in the ordinary explained. There was no need to add the course of trade; the Court of First Instance reasons for concluding that the sales in accepted that the reasoning was adequate question took place in the ordinary course in that it stated that 'it was found that sales of trade. In addition, recital 19 indicates made using compensation were indeed that Republica's claim was rejected both made in the ordinary course of trade', because it was not justified and because it having also found as a fact that the was made too late in the procedure; the applicant had produced no evidence to Court of First Instance could have held simply that sufficient reasons had been stated on the first aspect. Even if its actual 38 — Case 203/85 Nicolet [1986] ECR 2049, paragraph 10 of finding were held to be an error in law, the the judgment; Case T-164/94 Ferchimex v Council [1995] ECR II-2681, paragraph 118 of the judgment, cited by the appeal should still be dismissed on the Court of First instance in paragraph 114 of the judgment ground of lateness of the application. In its under appeal. 39 — See paragraphs 38 to 40. rejoinder, the Council argues that the plea

98. The main issue between the parties is what they have described as the burden of proof, although the obligation to state 100. It is admittedly to be expected that reasons operates at a different level from exporters' interests will be the opposite of proof, which is a matter of substantive Republica's in the present case. They will assessment. However, the question to be usually wish to have transactions with answered is whether the Council was associated parties or compensatory trans- obliged to state its reasons for finding that actions — which are likely to be at lower the disputed sales were made in the ordi- prices than other transactions — included nary course of trade if Republica had not in the calculation of normal value which in substantiated its claim that they were not. turn will be lowered, thus reducing the extent of any dumping margin calculated on that basis. The onus of establishing that there are grounds for departing from the standard procedure will thus fall on the exporter and not on the Community insti- tution in those circumstances. Here, how- ever, for whatever reason, the position is reversed and that is accepted by the Coun- cil. 99. I take the view, in line with my reason- ing with regard to Petrotub, that there was such an obligation. The first subparagraph of Article 2( 1 ) lays down the basic rule for determining normal value (on the basis of 'prices paid or payable, in the ordinary course of trade, by independent customers in the exporting country'). Under the third 101. It is thus the Community institution subparagraph, prices of the kind in issue which is seeking to depart from the stan- 'may not be considered to be in the dard method of calculation and its reasons ordinary course of trade and may not be for doing so must be stated in a manner used to establish normal value unless it is sufficient to enable the Court to exercise its determined that they are unaffected by the power of review and interested parties to relationship'. It is thus an exceptional ascertain the circumstances in which the departure from the normal procedure and Community anti-dumping rules were requires an explanation going beyond a applied. If the Council had alleged and

103. Consequently, the question whether Republica's request was submitted at too late a stage in the preliminary procedure is of no relevance. 106. Since I have reached the view that the appeal should succeed independently in respect of both appellants, there is no need to consider this argument. However, if the Court were to uphold the appeal as regards one appellant and dismiss it as regards the 104. I am therefore of the opinion that the other, that argument must fail in accord- contested regulation did not contain an ance with the Court's consistent case-law, adequate statement, such as to enable the most recently reaffirmed in Nachi Court to exercise its power of review and Europe. 40 interested parties to ascertain the circum- stances in which the Community anti- dumping rules were applied, of the reasons 40 — Judgment in Case C-239/99 Nachi Europe [2001] ECR for departing from the normal basis of I-1197, in particular at paragraphs 22 to 27; see also my Opinion in the same case, especially at paragraphs 36 to determining normal value to be used in 53.

I - 116

PETROTUB AND REPUBLICA v COUNCIL

Conclusion

107. Since the contested regulation contains inadequate reasoning with regard to both appellants, which the Court of First Instance wrongly accepted as adequate, it is necessary to set aside the judgment under appeal. The state of the proceedings is such that the Court may itself give final judgment and annul the contested regulation with regard to Petrotub and Republica. Both the appellants have asked for costs at first instance and on appeal.