What's New?
 - Sitemap - Calendar
Trade Agreements - FTAA Process - Trade Issues 

español - français - português
Search

BINATIONAL PANEL REVIEW
PURSUANT TO
THE NORTH AMERICAN FREE TRADE AGREEMENT
ARTICLE 1904

 

In the Matter of:

Brass Sheet and Strip
from Canada
  Secretariat File No.

USA-CDA-98-1904-03

(Continued)

During the course of determining the dumping margins, the ITA attempted to calculate a weighted average COP using the quarterly data which had been furnished. Error messages were allegedly encountered when Wolverine's data was inserted in the ITA's computer program. In a post-Final Determination memorandum which this Panel does not believe shoud be part of the administrative record, the ITA purported to explain that it could not reconcile the existence of a computed COP for certain CONNUMS which were not actually produced in certain quarters. The ITA believed that Wolverine had neglected to report the amount of production for those periods. Consequently, it deleted the product quantity (PRODQTY) 1 variable from the program. The result was that the program produced a "simple" average COP (giving equal weight to each quarter), rather than associating the costs to the amounts actually produced (a "fully" weighted average). 2

In the Preliminary Results, the ITA found Wolverine's margins to be de minimis.The consequences of deleting the PRODQTY line of the program do not seem to have been realized at this time by the parties.

(1) Did the ITA violate 19 USC §1677(m)(d)?

Wolverine states that it did not interpret the supplemental questionnaire as asking it to revise its reported data on a period of review (i.e., annual) COP basis. There is nothing in the record to gainsay this. The question does not refer to an "annual" weighted average cost. The question is too vague to infer that Wolverine should reasonably have known that an annual weighted average cost was being requested. While the ITA believes that it had asked for an annual COP figure, it admits that it did not request further information, nor did it identify Wolverine's quarterly COPs as being deficient in any manner. Further, the "simple" average calculated by the ITA comprised a period of six quarters, not four quarters as would be expected in an annual average COP. 

Wolverine has submitted that there was only one finding of an insufficient response by the ITA, and that came during the final stages of the revocation proceeding. The ITA claims that when it initially ran its margin calculation program, it noted certain error messages in the resulting output which were due to the content and format of the data submitted by Wolverine. The ITA concluded that the production quantities of zero represented flaws in Wolverine's response since a cost of production was reported for these same products. Accordingly, the initial ITA rationale was that it had to use a simple (or arithmetic) average to calculate the revised COP's for the new CONNUMs. During this Review, the ITA admitted that it had made an error in not properly weighing the collapsed COP's, and consequently requested an opportunity to correct this error.

The U.S. Industry's submissions on these facts are that Wolverine failed to follow the ITA's instructions to submit a single weighted average cost for each unique product as represented by a specific matching control number. The U.S. Industry states that Wolverine deliberately chose to ignore the ITA's instruction in two important respects when it first submitted its sales and costs data. First, Wolverine inserted a physical characteristic of its own choosing into the CONNUM system, claiming that the ITA should rely on a matching characteristic that would distinguish between products identified by Wolverine as non-reroll and reroll. Second, Wolverine submitted quarterly weighted average costs for each control number rather than a single weighted average cost for each control number. Despite what the U.S. Industry argues were clear and repeated requests, the U.S. Industry suggests that Wolverine chose to ignore the ITA's instructions, and refused to submit a single weighted average cost that would not distinguish between reroll and non-reroll.

It is acknowledged that the ITA originally intended to weight the cost figures by the production quantities provided by Wolverine. However, as noted, when the ITA attempted to weight these data, it encountered error messages in the resulting output which were due to the content and format of the data. The problem was that the cost data contained a significant number of zero production quantities throughout the cost response, which the ITA decided to ignore in its calculation. The U.S. Industry has taken the position that, given the problems with the production quantity data submitted by Wolverine, the ITA correctly disregarded Wolverine's reported zero production quantities and employed a simple average COP for the collapsed CONNUMs.

Wolverine says that the only notification it received that suggested its response was insufficient was in the June 24, 1998 Memorandum it received after the Final Determination was made and after Wolverine had brought the "clerical" error to the ITA's attention. However, Wolverine points out that "the June 24 Memorandum cites to no specific supplemental question that Wolverine allegedly did not answer." (Canadian Complainant's Brief at 25).

In determining when, if ever, the ITA realized that there was an insufficient response from Wolverine, the ITA's submissions on this point are important. At page 34 of its Response Brief, the ITA states:

The presence of zero production values with respect to some quarterly data did not compromise the results of the cost values calculated for Wolverine. These so-called missing values corresponded solely to CONNUM-specific quarterly data for quarters for which Commerce used no home market sales as matches. Furthermore, Commerce verified that the POR zero production values represent actual periods of non-production. Thus, although the presence of what the computer perceived as missing values was disconcerting, it did not, in fact, prevent the cost test function from running or compromise the results of the cost test....Because all the facts needed to calculate an accurate weighted average cost for the home market matches were, in fact, on the record, Commerce's resort to a simple average was a "solution" for which there was no real underlying problem. (Emphasis added).

As well, the ITA states that resorting to a simple average is not called for by the "facts available" provision. 3 The ITA notes:

[E]ven if one assumes, arguendo, that Commerce's request that Wolverine submit a "single" weighted average was sufficiently clear that Wolverine's failure to calculate its own annual average for its COP data constituted refusal to provide requested information, the absence of such an annual average database on the record also did not create a "need" to use a simple average. It simply created a need for Commerce itself to perform the weight-averaging function on the data of record, which was adequate for that purpose.

For the following reasons, the Panel holds that the ITA did not violate 19 USC §1677(m)(d). The codified Tariff Act provision, 19 USC §1677(m), deals with conduct of investigations and administrative reviews. More specifically, §1677(m)(d) deals with deficient submissions and states:

If the administering authority...determines that a response to a request for information under this subtitle does not comply with the request, the administering authority...shall promptly inform the person submitting the response of the nature of the deficiency and shall, to the extent practicable, provide that person with an opportunity to remedy or explain the deficiency in light of the time limits established for the completion of investigations or reviews under this subtitle. If that person submits further information in response to such deficiency and either—

(1) the administering authority...finds that such response is not satisfactory, or

(2) such response is not submitted within the applicable time limits, 

then the administering authority...may, subject to subsection (e) of this section, disregard all or part of the original or subsequent responses.

Thus, §1677(m)(d) allows the ITA to disregard all or part of the original or subsequent responses of Wolverine only if it has already informed Wolverine of the nature of the deficiency and, to the extent practicable, provided Wolverine with an opportunity to remedy or explain the deficiency. If the further information is not satisfactory, or provided outside the specified time limits, the ITA may disregard the information provided by Wolverine (subject to the requirements of §1677(m)(e)). In sum, §1677(m)(d) clearly requires notification of a deficient response, and a reasonable opportunity to correct the response, before the ITA is allowed to disregard information submitted by Wolverine. Thus, the question to be answered is whether the ITA determined that there was an insufficient response from Wolverine and, if so, did it allow Wolverine to clarify its submissions.

While Wolverine is correct that the ITA was wrong in failing to notify it of its claimed justification for calculating a simple average COP, Wolverine cannot rely on the provisions of §1677(m)(d) to secure a remand. This provision is meant to address the actual filing of submissions by Wolverine, and not the ITA's incorrect use of submitted and verified data. Moreover, there is no clear evidence that at the relevant time the ITA failed to inform Wolverine of perceived deficiencies in its submitted data. More importantly, there does not appear to be any record in the course of proceedings which shows that the data submitted by Wolverine was considered by the ITA, prior to June 24, to be deficient. 

Furthermore, the June 24 memorandum, even if it could form part of the record of this proceeding (which the Panel does not believe it can), by Wolverine's own admission, does not explicitly state that any of Wolverine's submitted data was wrong. The ITA, in these proceedings, has taken the position that Wolverine's submitted data were sufficient to prepare a weighted average, and thus were not deficient in any manner. The result is that it is unclear, at best, that there has been a failure on the part of the ITA to fulfill the applicable statutory requirements.

In summary, the action of the ITA might have been brought within the criteria of 19 USC§1677(m)(d) had the assumed defect in Wolverine's submitted information ever been brought to Wolverine's attention. It was not. The ITA simply operated on the basis that there had been an inadvertent error in Wolverine's data. Nor did the ITA explain what it had done. At no stage, until the threshold of these proceedings, was Wolverine informed of what steps were taken by the ITA.

It was argued by the U.S. Industry that knowledge of what the ITA did ought to be imputed to Wolverine. This assertion was made on the sole basis that the computer program used by the ITA, if reviewed by a knowledgeable and alert programmer, would disclose what the ITA had done. It was suggested by the U.S. Industry that an independent review of the ITA's computer language by a programming expert is a normal step in conducting an antidumping case before the ITA. The concept of imputed knowledge, having particular regard to the general norm of procedural fairness, cannot be stretched so far as to include notice buried in computer programming language.

Chief Judge Carman's decision in Sugiyama Chain Co., Ltd. v. United States, 797 F. Supp. 989 (CIT 1992) is on point. In that case, the ITA had made a computer program error which skewed the dumping calculation. During the proceedings in that case, the ITA had incorrectly, as it did here as well, described its calculation methodology. After "rejecting out of hand" the ITA's exhaustion defense as "somewhat disingenuous in light of the circumstances", the Court held that the computer "methodology enunciated [by the ITA] cannot be sustained and a remand must issue." The Court agreed with he ITA that the proper course is for the agency to "examine [the calculation]; give the reason for its actions, and, if warranted, make any necessary corrections in the computer programming instructions." 797 F.Supp. at 997.

Finally, the actions of the ITA which form the underlying substance of Wolverine's complaint, namely, failing to provide Wolverine with a timely explanation of the basis on which it was dealing with the information received, serve to underpin the overarching allegation that Wolverine was not treated fairly. As noted, even if it were true that a computer expert might have discerned the change in the ITA's methodology by analyzing the programming language from the Preliminary Determination, this is not adequate to discharge the ITA's duty to advise Wolverine of how it was dealing with the information provided and, in particular, that in its assessment it was departing from its normal method of calculation. See discussion infra at 39.

(2) Did the ITA violate 19 USC §1677(e)(b) and unlawfully make an adverse inference against Wolverine?

Wolverine also argued that the ITA unlawfully made an inference adverse to the interests of Wolverine. Specifically, section 776(b) of the Tariff Act of 1930, 19 U.S.C. §1677(e)(b), provides that:

If the administering authority...finds that an interested party has failed to cooperate by not acting to the best of its ability to comply with a request for information from the administering authority..., the administering authority..., in reaching the applicable determination under this subtitle, may use an inference that is adverse to the interests of that party in selecting from among the facts otherwise available. Such adverse inference may include reliance on information derived from—

(1) the petition, 

(2) a final determination in the investigation under this subtitle,

(3) any previous review under section 1675 of this title or determination under section 1675b of this title, or

(4) any other information placed on the record."

Wolverine cites Olympic Adhesives, Inc. v. United States, a Federal Circuit decision, which holds that to avoid the threat of an adverse inference on the grounds of being non-responsive, a respondent "need only provide complete answers to the questions presented in an information request." 

The U.S. Industry disputes Wolverine's submission that the ITA even made an adverse inference against Wolverine. In this case, the U.S. Industry submits, the ITA did not rely on information from any other source and therefore did not employ an adverse inference against Wolverine as described in the statute. In fact, the ITA used the very cost information submitted by Wolverine, despite Wolverine's failure to submit its information in the form requested by the ITA. In summary, the U.S. Industry submits that there is no indication whatsoever that the ITA intended to rely on facts that would be adverse to Wolverine when collapsing the cost data.

For its part, the ITA states its position 1 as follows:

[T]he Department may only use an adverse assumption in selecting from the facts available if it has made a finding that a party has failed to cooperate in a proceeding. Commerce made no such finding in this review. Thus, because Commerce had all the data it needed to calculate a weighted average COP and because there was no basis for disregarding Wolverine's reported production data, Commerce erred in calculating, instead, a simple average.

It is clear from a review of the record that the ITA did not make a finding, as required by §1677(e)(b), that Wolverine failed to co-operate by not acting to the best of its ability to comply with a request for information from the administering authority. However, this is presumably because the ITA did not intend to make a finding adverse to the interests of Wolverine and, in the view of the Panel, it did not do so. What it did do was employ a methodology which had an adverse result to Wolverine in this case, but in doing so it drew no inference adverse to the interests of Wolverine in selecting from among the facts otherwise available. There was no evidence that ITA intended this adverse result. Indeed, when the ITA employed this methodology the first time in the Preliminary Results, before it collapsed the data, the result was not adverse to Wolverine.

Therefore, the Panel finds that the ITA, since it did not make an inference adverse to the interests of Wolverine, did not violate 19 USC §1677(e)(b).

(3) Did the ITA fail to use the COP data submitted by Wolverine in violation of 19 USC §1677(m)(e)? 

The Tariff Act (19 USC §1677(m)(e)) requires that the ITA shall not decline to consider information that is submitted by a party which is, inter alia, necessary to a determination, provided that the information is timely submitted, can be verified, and can be used without undue difficulty.

The statute states that: 

[i]n reaching a determination...the administering authority ...shall not decline to consider information that is submitted by an interested party and is necessary to the determination but does not meet all the applicable requirements established by the administering authority... if—

(1) the information is submitted by the deadline established for its submission,

(2) the information can be verified; 

(3) the information is not so incomplete that it cannot serve as a reliable basis for reaching the applicable determination,

(4) the interested party has demonstrated that it acted to the best of its ability in providing the information and meeting the requirements established by the administering authority...with respect to the information, and

(5) the information can be used without undue difficulties.

19 USC §1677(m)(e). 

The U.S. Industry argues that the ITA calculations, without weighing for production quantity, were reasonable under the circumstances and resulted in legally sustainable COPS. The U.S. Industry does not seriously dispute that using a weighted average COP would be proper, nor is there serious debate about whether the ITA set out to produce a weighted average. Nonetheless, it is argued that the data submitted by Wolverine were so flawed that ITA was justified in disregarding them.

First, it is contended that a value for a CONNUM which was not produced for a given period cannot be calculated because a zero denominator will always produce a zero COP. Second, the use of the "window" period sales produces unusable COPS. Lastly, there are several contentions regarding the accuracy of Wolverine's data supplied in response to the questionnaire. It follows, the U.S. Industry argues, that the ITA was justified in employing some other methodology, e.g., not weighing the data. The U.S. Industry cites various examples where the ITA employed, and justified, a simple average COP, rather than a weighted average.

The ITA takes the position that its normal practice is to calculate a weighted average COP and that nothing in this case required a departure from that practice. Indeed, (as discussed infra), it was an error to use a simple average in the absence of some necessity requiring such a departure. Weighted averages smooth out the variations in product costs, which may vary from quarter to quarter, and thus give an accurate picture of the overall annual costs. 2 It is noted both that the ITA had requested a weighted average, and that during the proceeding, the U.S. Industry recognized that a weighted average should be calculated. Examples are given of the extraordinary circumstances which, in the past, have caused the ITA to use simple averages. It is argued, therefore, that the use of a simple average was not justified, as the data submitted in this case were adequate to compute a weighted average. Neither in its brief, nor at oral argument, did the ITA comment upon, or attempt to explain, the reasons why it computed a single "simple" average. 

Wolverine contends that since its data were verified, the ITA was without justification in ignoring its submitted COPs in computing a simple average. It asserts that during the verification process, the ITA had actually verified its claimed COP for at least one CONNUM which was not produced during the period for which verification was undertaken. Further, Wolverine disputes the contention that proper use of the computer program supplied results in error messages. In addition, under the case law3, and as required by section 1677m(e), if the ITA encountered difficulties in running the program, it was obliged, before ignoring Wolverine's data, to inform Wolverine and offer an opportunity to clarify why error messages were encountered. Even if Wolverine erred in submitting quarterly data, Commerce had never made a point of objecting to it, and, in any event, could easily have calculated proper annual weighted averages.

One thing is clear in this case. A mistake was made. All of the interested persons had asked for a weighted average and assumed that one would be computed. The problem with the U.S. Industry's argument is that the ITA did not intentionally reject Wolverine's data, nor did it intentionally employ a different methodology. True, Wolverine could have avoided this problem by furnishing a weighted average in response to the ITA's request.

It is noted that the ITA's request was that Wolverine report a weighted average cost of production. Specifically, in rejecting Wolverine's roll-reroll product categories, the ITA instructed Wolverine to collapse the two into a single CONNUM, and report a single weighted average for the category. In this context, the ITA surmises that Wolverine may have continued to report quarterly information because it focused upon the first part of the request (i.e., to collapse the two product codes), rather than the latter. Indeed (in contrast to its present position), the U.S. Industry after the preliminary determination, also asked ITA to require Wolverine to collapse the two product codes, and report an annual weighted average COP.

Had the ITA consciously intended to depart from its normal practice of using weighted averages, the issue here would be different. In that case, the ITA would perhaps have articulated its rationale for doing so.

Whether there were deficiencies in Wolverine's submissions which might have justified the ITA's rejection of Wolverine's data is not the point. The record supports the conclusion that the ITA intended to use the submitted data, which it had verified to its satisfaction. The panel will not revisit the ITA's findings.

The Panel concludes that the ITA's actions were not reasonable. To the extent that the deletion of the PRODQTY line of programming was based upon a mistaken notion that Wolverine neglected to report sales, it was contrary to the verified evidence. In any event, the alleged "error messages" should have triggered an inquiry of Wolverine by the ITA as to the source of the errors, as required by section 1677m(e). The result was, therefore, not supported by substantial evidence. Section 1677m(e) dictates that the ITA use the information furnished by the respondent if it was timely submitted, it was verified, and if it can be used without undue difficulty. How much difficulty constitutes "undue difficulty", is, of course, a somewhat subjective standard. However, given that Wolverine's data was verified to the satisfaction of the ITA, the bar should be set at a high level to justify the failure to contact Wolverine. 

If Wolverine's contention that it was able to run the program without any difficulty is accurate, seemingly it would have been a simple matter to rectify the problem.

The Panel concludes that the ITA did not comply with 19 USC §1677(m)(e) in that it declined to consider Wolverine's information which had been timely submitted, was verified and, could have been used without undue difficulty. Accordingly, the Final Results are not affirmed in this respect. 

(4) By Calculating a Simple Average Cost of Production, the Investigating Authority Also Did Not Follow Its Established Administrative Practice

Wolverine also challenges the ITA's decision to calculate a simple average cost of production as violating the agency's established administrative practice. 4 The ITA agrees that it is the "normal practice" of the ITA to use annual weighted average costs to determine COP, and argues that it had originally intended to follow that practice in this review as well. 5

The U.S. Industry argues that the ITA "could have resorted to any number of methodologies" other than a weighted average. 6 The U.S. Industry points to numerous other antidumping determinations where the ITA has, indeed, used a simple, rather than weighted average. 7

The ITA and Wolverine argue that in the prior cases where the ITA had resorted to a simple average, it did so either because special circumstances made a simple average more accurate, or because the necessary data were unavailable. 8 The U.S. Industry claims that the data submitted by Wolverine in this case, including the production quantities for the window periods, rendered the data unusable for calculating an accurate weighted average. This, it is claimed, necessitated using another methodology, such as a simple average. 

As already noted, the ITA and the U.S. Industry agree that the average ultimately used by the Department was neither a purely simple nor weighted average, but rather a "simple" average of weighted quarterly data. 9 The shipment quantity had already been used by Wolverine to weight the submitted costs. Also, the reported raw material costs had already been weighted by Wolverine. 

It is hornbook law that an administrative agency must not depart from its established policy. "The dominant law clearly is that an agency must either follow its own precedents or explain why it departs from them." Davis, K. & Pierce, R., II Administrative Law Treatise, § 11.5 at 206-07 (3d ed. 1994)(and cases cited therein). See, e.g., Atchison, Topeka & Sante Fe R. Co. v. Wichita Bd. of Trade, 412 U.S. 800, 808 (1973)(agency has a "duty to explain its departure from prior norms") and Greater Boston Television Corp. v. FCC, 44 F.2d 841, 852 (D.C. Cir.) cert. denied, 403 U.S. 923 (1971). In the antidumping context, the CIT has held that "Commerce has the flexibility to change its position providing that it explains the basis for its change and providing that the explanation is in accordance with law and supported by substantial evidence." Cultivos Miramonte S.A. v. United States, 980 F. Supp. 1268, 1274 & n. 6 (CIT, 1997) (footnotes omitted.) 10

The ITA's established practice has been to calculate COP on a weighted average basis, rather than a simple average of costs. E.g., Certain Welded Carbon Steel Pipe and Tube from Turkey, 61 Fed. Reg. 69067, 69075 (December 31, 1996)("the Department's normal practice is to calculate weighted-average costs of production").

In addition, the weighted average is normally calculated over the relevant sales period which is normally the year under the administrative review. The rationale for using an annual weighted average is that use of a simple average tends to produce a distorted COP if there are fluctuations in cost. See Fujitsu General Ltd. v. United States, 883 F. Supp. 728, 735 (CIT, 1995)("[R]andom fluctuations in COP...are precisely the type of routine, random fluctuations that justify Commerce's preference of using annual weighted average COP.")

In the view of this Panel, the established practice of the ITA has been to employ a weighted average where possible in calculating COP. Where that is not possible, the ITA has on occasion used simple averages in lieu of a weighted average. However, where it has resorted to simple averages, as in the prior determinations cited by the U.S. Industry, it has done so with an appropriate explanation. 

We agree with the ITA and Wolverine that the prior cases cited by the U.S. Industry constitute exceptions to the rule. In several of the cases, e.g., Fresh and Chilled Atlantic Salmon from Norway, 56 Fed. Reg. 7661, 7666, 7672 (February 25, 1991), a simple average was more accurate give the number of respondents or the conditions of the market. In a few of the other cases, it is true, as the U.S. Industry states, that sufficient data had not been submitted to enable the ITA to use a weighted average. E.g., Silicon Metal from Brazil, 62 Fed. Reg. 42759, 42761 (August 8, 1997). 

However, a review of the determinations cited by the U.S. Industry does indicate that, even where the necessary data had not been submitted and the ITA resorted to a simple average, the ITA consistently explained the rationale for use of a simple average. Therefore, regardless of whether the ITA had reasonably resorted to a simple average in the face of allegedly unusable data in the instant case, the threshold question here is whether it explained its departure from its established prior practice. Even if the Wolverine data were truly unusable and, as in some of the prior cases, the necessary data had not been submitted, the ITA has failed to provide any explanation in its final results for having deviated from its established practice. 

Were this the only aspect of this case in issue, we would remand to the ITA only for an explanation of why it calculated COP on a "simple" rather than a weighted average in this case and, in doing so, failed to follow established agency practice. This threshold question of the vice of an unexplained departure from agency practice is, however, subsumed in the broader remand.

Consequently, anticipating this, it is more appropriate here to conclude that the proper course of action is to order a remand of the calculation issue to the ITA so that, as it proposed in Sugiyama, it can "examine it, give reasons for its actions, and, if warranted, make any necessary corrections in the computer programming instructions." Compare Sugiyama Chain Co., Ltd. v. U.S., 797 F. Supp at 997 (CIT 1992).

C. THE U.S. INDUSTRY'S ALLEGATIONS OF ERROR BY THE ITA IN COMPUTING COP

As the ITA had disregarded home market sales made by Wolverine as below cost of production ("COP") in the prior administrative review, the agency had reasonable grounds to "believe or suspect that sales of the foreign like product...have been made at prices ...less than the cost of production...." 19 U.S.C. §1677b((b)(1). The ITA therefore conducted a COP investigation in the administrative review before us.

The U.S. Industry argues that the ITA incorrectly calculated Wolverine's cost of production in two respects: (1) the Agency failed to include an amount for the Canadian portion of Wolverine's indirect selling expenses ("ISE"); and (2) the Agency failed to adjust Wolverine's COP for an improper reduction to material costs associated with non-subject merchandise. 11

First, the U.S. Industry argues that ITA erred when it failed to include Canadian indirect selling expenses in Wolverine's COP despite the fact that such expenses were included in calculated net home market prices for the sales-below-cost test. In the U.S. Industry's view, this failure led to an understatement of the dumping margin when a COP not including Canadian ISE, was compared to net prices which did include Canadian ISE. See U.S. Industry Case Brief at 40-43.

Wolverine and Commerce both respond that the calculated COP did in fact include Canadian ISE, accomplished through an allocation of the verified selling, general and administrative expenses reported in the consolidated financial statements of Wolverine's U.S. parent company. Adding the indirect selling expenses again to COP would thus result in an impermissible double-counting. See Rule 57(2) Response Brief of the Investigating Authority at 48-56; Canadian Complainant's Reply Brief at 20-24. Wolverine also argued that the U.S. Industry had failed to exhaust its administrative remedies on this issue.

Second, the U.S. Industry argues that Commerce erred in not adjusting Wolverine's metal costs to disallow a credit for certain raw material inputs which result from the production of non-subject merchandise, and that this error caused Commerce to understate Wolverine's COP and thus the final dumping margin. U.S. Industry Case Brief at 43-46. This non-subject merchandise, known as toll sales, consists of metal inputs processed by Wolverine for outside parties into a finished product. These parties retain title to the material during processing. Canadian Complainant's Reply Brief at 24 n. 65. The U.S. Industry argues that "Wolverine reduced its cost of materials for the subject merchandise because of an offset due to nonsubject merchandise". U.S. Industry Case Brief at 45. 

(1) The U.S. Industry Has Exhausted Its Administrative Remedies

The ITA and Wolverine argue that the U.S. Industry failed to raise its toll sales argument during the administrative proceeding before the Agency, thereby precluding it from raising this argument before the Panel since it failed to exhaust its administrative remedies. Wolverine also argues that the U.S. Industry failed to raise its argument concerning Canadian ISE during the administrative proceeding.

For the reasons discussed below, the Panel finds that the U.S. Industry did raise both of these arguments before Commerce during the administrative proceedings in this case, and that reliance on the exhaustion of administrative remedies doctrine by both Wolverine and Commerce is misplaced. 

As already discussed at length supra, this Panel may not consider an argument that the Court of International Trade would be constrained from hearing by virtue of the fact that a party had failed to exhaust its administrative remedies. 

It is also worth noting, however, that the Panel has found no authority, nor has any authority been cited to it, in which the CIT applied the exhaustion doctrine to prevent a party from merely buttressing its legal arguments concerning factual issues which it had previously raised before the ITA.

Thus, the question which this Panel must decide is whether the U.S. Industry raised indirect selling expense and toll sales arguments during the 1996 administrative review that were specific enough to permit it to escape the application of the exhaustion doctrine. After examining the administrative case brief filed by the U.S. Industry with the ITA on April 13, 1998 during the annual review, ("Administrative Case Brief"), the Panel concludes that it clearly has. 12

With respect to the toll sales issue, both Wolverine and the ITA claim that the U.S. Industry raised an argument regarding a mill loss adjustment in its administrative case brief, while it now raises an argument regarding a different adjustment for scrap relating to toll sales. 13

An examination of the U.S. Industry's Administrative Case Brief, however, reveals that it clearly describes the issue complained of as a toll sales adjustment, just as in the brief submitted by the U.S. Industry to this Panel. 14 See Administrative Case Brief at 4-7.

In general, the Administrative Case Brief makes the same factual arguments on these two issues which the U.S. Industry had advanced in its briefs to this Panel, both with respect to improper adjustments to COP caused by the inclusion of toll sale material in the cost calculation, and with respect to the ITA's alleged error in failing to include the Canadian portion of Wolverine's ISE in its calculation of Wolverine's COP. See Administrative Case Brief at 4-7, 10-11. The only real difference between the U.S. Industry's Administrative Case Brief and its brief to this Panel is the addition of citations to the legal authorities which Commerce supposedly violated. Moreover, the Panel notes that Comments 3 and 4 of the Final Determination discuss the toll sales and ISE arguments in terms that are practically identical to those used by the U.S. Industry in its submissions to the Panel. See Final Results of Antidumping Duty Administrative Review and Notice Not to Revoke Order in Part, 63 Fed. Reg., 33,037, 33,039-40 (1998). Thus, it is clear that the agency itself in fact considered the issues. 15

Under these circumstances, there is no question but that the U.S. Industry did raise the specific issues under review during the administrative procedure before Commerce. This Panel must thus proceed to consider the issues raised by the U.S. Industry on the merits.

(2) Indirect Selling Expenses in the Home Market

The Tariff Act, 19 U.S.C. §1677b((b)(3), directs the ITA to include in its calculation of COP an amount for selling, general, and administrative expenses ("SG&A"). 16

As noted, the U.S. Industry argues that the ITA intended to include in COP an SG&A amount for a portion of U.S. corporate expenses associated with the Canadian operations, in addition to the amounts already reported for general and selling expenses by Wolverine for its Canadian operations. 

The ITA disagrees and states that to include Canadian ISE, in addition to the amounts allocated for consolidated SG&A, would result in a double-counting of these expenses, since the consolidated SG&A already includes the Canadian ISE.

In its Preliminary Results, the ITA described its allocation of SG&A expenses from the company's U.S. corporate headquarters: 

[W]e allocated a portion of the [SG&A] expenses for the corporate headquarters in Huntsville/Decatur, Alabama to Wolverine's [COP]. This additional allocation was based on SG&A and cost of sales information taken from Wolverine's financial statements.

ITA's Preliminary Analysis Memorandum at 3 (January 20, 1998). In its Final Results, the ITA clarified its reasoning:

Respondent's financial statements demonstrate that indirect selling expenses were included in general and administrative expenses. Adding an additional amount for indirect selling expenses to the COP would result in double-counting. 

Final Results, 63 Fed. Reg. at 33,040.

As documented in its cost verification report17, the ITA verified Wolverine's Fergus facility general expenses and ISE, verified that these ISE were included in the SG&A expenses that were traced to the Canadian financial statements, and verified that such expenses were included in the total SG&A expenses reported in the consolidated financial statements for Wolverine's U.S. parent company. 

The ITA used a three-step process to allocate a portion of the total consolidated SG&A expenses to the reported COP. 18 First, the cost of sales ("COS") for the Fergus operations were converted into a U.S. dollar amount. Second, this COS amount was divided by the total consolidated company COS (a U.S. dollar amount) to calculate a ratio representing the portion of total costs attributable to the Fergus operations. Finally, this ratio was applied against the total SG&A expenses for the consolidated company to calculate the SG&A portion attributable to Fergus. The application of this ratio can be seen at line 2460 of the log of the ITA's final analysis computer program.

While the Panel can see that the ITA's treatment of the consolidated SG&A expenses is not as clear as it could be in the Preliminary Results, the fact that the Canadian companies' SG&A expenses (including indirect selling expenses) are included in the consolidated SG&A of the U.S. parent company headquartered in Huntsville/Decatur is supported by the record evidence. To add in an amount for reported Canadian indirect selling expenses would result in an unlawful double-counting of these expenses. 

Therefore, this Panel finds the ITA's conclusion that adding additional amounts for indirect selling expenses would have double-counted these expenses is supported by substantial evidence on the record and otherwise in accordance with law. The ITA's determination is hereby affirmed by the Panel in this regard.

(3) The Toll Sales Handling Charge 

Wolverine's factory produces merchandise for sale in the home market. It also processes other companies' raw materials into similar merchandise for a fee, commonly referred to as "tolling". Wolverine argues here that such toll sales and their respective costs are not included in the ITA's normal value calculations. Wolverine's Reply Brief at 24. 

The U.S. Industry argues that Wolverine artificially lowered its costs of materials for subject merchandise by an offset to costs realized from toll sales, i.e., non-subject merchandise. The U.S. industry argues that, because the adjustment is derived from toll sales, such an adjustment is unlawful. U.S. Industry Case Brief at 43-46. 

The Investigating Authority determined that the mill loss adjustment did not affect the cost of materials. 

The Department verified that the reported per-unit materials cost was accurate. Although a mill loss adjustment was made to the metal pools account which reflected decreased quantities, this adjustment does not affect the cost of materials account. 

Final Results at 33,039.

The U.S. Industry cites no evidence of record to refute the ITA's determination that the mill loss adjustment has no effect on Wolverine's reported material costs. Nor does it demonstrate how the adjustment is anything more than an inventory accounting convenience. Moreover, the reasonableness of Wolverine's reporting of material costs is supported by the ITA's Commerce's verification report 19, its analysis memorandum 20, and the collected cost verification exhibits 21. Therefore, this Panel concludes that the ITA's determination regarding this issue is supported by substantial evidence on the record and otherwise in accordance with law. The ITA's decision is hereby affirmed with regard to this issue.

V. DISPOSITION

For the reasons stated above, the Panel hereby remands as follows: 

That part of the ITA's determination where it calculated Wolverine's cost of production on a simple, rather than weighted average, basis. The ITA shall examine the COP calculation contained in the Final Results and determine whether a weighted average rather than a simple average should have been used in the calculation. If a weighted average should have been used, then the ITA shall make the necessary changes to the computer program. If a simple average should have been used, then the ITA shall provide reasons for why it has departed from the established practice of employing a weighted average in calculating COP.

The Panel affirms the Commerce Department's determination in all other respects. 

The Panel hereby directs the Commerce Department to provide the results of this remand within sixty (60) days of the date of this Order. Any objections to the Investigating Authority's remand results shall be filed by an interested person within twenty (20) days after the filing of the remand. Any responses to the objections shall be filed by an interested person ten (10) days thereafter.

  

Signed in the original by:

  

Robert E. Ruggeri, Esq., Chairman 

Frank Foran, Q.C.

Daniel Pinkus, Esq.

Maureen Rosch

T. Bradbrooke Smith, Q.C.

 

Issued on the 16th day of July, 1999

1 Memorandum from Paul Stolz to Holly Kuga, 3 (dated June 24, 1998)(Admin. Rec. Non-Pub.  35/36/34). 

2 What the ITA calls a "simple" average, the U.S. Industry calls a "hybrid weighted average cost".  This is because, as they both admit, the "simple" average taken by the ITA was, more  precisely, a straight average taken of Wolverine's weighted average quarterly cost data.  Some element of weighting was therefore included, but not to the degree the ITA has  traditionally done in calculating a weighted average.

3 Id. at 39.

1 Rule 57(2) Response Brief at 40.

2 Fujitsu General Ltd. v. United States, 883 F. Supp. 728 (CIT 1995).

3 Borden, Inc. v. United States, 4 F. Supp. 2d at 1246 (CIT 1998).

4 Canadian Complainant's Brief at 34-35.

5 Rule 57(2) Response Brief of the Investigating Authority at 24-25.

6 Case Brief of the U.S. Industry at 37.

7 Id. at 37-39.

8 Canadian Complainant's Brief, nn 87 and 88 at 34; and Rule 57(2) Response Brief of the  Investigating Authority, 22-34.

9 Response Brief of the Investigating Authority, n. 13 and Case Brief of the U.S. Industry, 9-10.

See, also, Citrosuco Paulista, S.A. v. United States, 704 F. Supp. 1075, 1088 (CIT, 1998) and  Hussey Copper Ltd. v. United States, 17 CIT 993, 997, 834 F. Supp. 413, 418 (CIT,  1993)(Purpose of this rule is to "insure consistency in an agency's administration of the  statute.").

10 Reference Re Canada Temperance Act, [1939] 4 D.L.R. 14 (Ont. C.A.) affd. (P.C.)

11 As noted, supra, a third issue was raised and subsequently dropped by the U.S. Industry. 

12 The U.S. Industry argued below that the ITA should add certain ISE's to COP. Administrative  Case Brief at 11 (Admin. Rec. Non-Pub 33/01/25). Wolverine claims now that the U.S.  Industry's current argument before this panel is different from that raised below.  Wolverine Reply Brief at 21-23. We disagree: the issue had been raised by the U.S.  Industry sufficiently before the agency to satisfy the purposes of the exhaustion doctrine.

13 Rule 57(2) Response Brief of ITA at 59; Wolverine Reply Brief at 22-23. Also, Transcript of Oral  Argument at 107-109 (ITA) and 148-150 (Wolverine).

14 The Panel notes that Commerce did not advance an exhaustion argument with respect to the ISE  issue, and that Wolverine's argument in this respect was limited to the assertion in its brief  described above that the U.S. Industry had failed to raise the issue in the underlying  proceeding below.

15 See Holmes Products Corp. v. United States, 16 C.I.T. 1101, 1992 Ct. Intl. Trade LEXIS 258, *7-8 (exception to exhaustion doctrine where agency had, in fact, considered the issue below.)

16 The Statement of Administrative Action, H.R. Doc. No. 103-316, vol. 1. at 809 reprinted in 1994 U.S.C.C.A.N. 4153 (Art. 2.4) specifies "a general requirement that comparisons be  fair...[and] admonishes national authorities not to double-count adjustments." (Emphasis  added).

17 ITA Cost Verification Report (undated) 5-6 (Admin. Rec. 32/54/23); PD-71 at 51, Fiche 18, Fr. 62.

18 Final Results Analysis Memorandum at 3. 

19 ITA Cost Verification Report at 10-11 (Admin. Rec. 32/54/23); PD-71 at 51, Fiche 18, Fr. 59-60.

20 Memo to File from Paul M. Stolz at 1-2 (June 6, 1998)(Fiche 18, fr. 64-65).

21 (Confidential Record, fiches 37-51).