E.I. DuPont de Nemours & Co. v. United States

Opinion Restani, Judge: This case is before the court on plaintiff E.I. DuPont De Nemours & Company’s (“DuPont”) motion for judgment upon the agency record pursuant to USCIT R. 56.2. Plaintiff contests the anti-dumping duty margin assigned Aramid Products Vo.F. (“Aramid”) and Akzo Nobel Fibers Inc.1 by the United States Department of Commerce (“Commerce”) in the final results of the first administrative review of the antidumping duty order on aramid fiber formed of poly para-pheny-lene terephthalamide (“PPD-T aramid fiber”) from the Netherlands. Aramid Fiber Formed of Poly Para-Phenylene Terephthalamide from the Netherlands, 61 Fed. Reg. 51,406 (Dep’t Commerce 1996) (final results of admin, rev.) [hereinafter “Final Results”]. The court sustains Commerce’s determination. Background Commerce published its antidumping duty order for PPD-T aramid fiber from the Netherlands on June 24,1994. Aramid Fiber Formed of [20] Poly-Phenylene Terephthalamide from the Netherlands, 59 Fed. Reg. 32,678, 32,678 (Dep’t Commerce 1994). On June 6, 1995, Commerce published a notice of opportunity to request an administrative review of the antidumping duty order covering the period December 16, 1993 through May 31, 1995. Antidumping or Countervailing Duty Orders, Finding, or Suspended Investigation, 60 Fed. Reg. 29,821,29,821 (Dep’t Commerce 1995)(opportunity to request admin, rev.). Commerce published a notice of initiation of this first antidumping duty administrative review on July 14,1995. Initiation of Antidumping and Countervailing Duty Administrative Reviews and Request for Revocation in Part, 60 Fed. Reg. 36,260, 36,260 (Dep’t Commerce 1995). After completing the review, Commerce calculated the final antidumping margin of22.03%.2 Final Results, 61 Fed. Reg. at 51,410. DuPont’s motion before the court challenges the following aspects of the Final Results: (1) Commerce’s calculation of Aramid’s financing expenses based upon the consolidated financial statements of Akzo Nobel N.V, (2) Commerce’s exclusion of certain goodwill expenses from Ara-mid’s cost of production, (3) Commerce’s refusal to place Akzo’s standard costs from the original investigation on the record to test the reliability of Akzo’s standard cost data in this review, and (4) Commerce’s reliance on constructed value in some instances even though alternative home market sales were available. Standard of Review The standard of review for an agency’s determination requires the court hold any determination unlawful if unsupported by substantial evidence on the record or otherwise not in accordance with law. 19 U.S. C. § 1516a(b)(l)(B)(i)(1994); see also Samsung Elecs. Co. v. United States, 946 F. Supp. 5, 7 (Ct. Int’l Trade 1996), aff'd, 129 F.3d 135 (Fed. Cir. 1997). Discussion I. In early 1994, Akzo Nobel N.V increased its holdings of Aramid from 50% to 95% ownership by converting to equity outstanding loans given to Aramid to finance production of PPD-T fiber by Akzo. Section D Questionnaire Response (Oct. 11, 1995), at Ex. D-11T, 2, C.R. 5, Def.’s App., Ex. 3. Also in 1994, Aramid sustained a [ ]. Akzo’s Section A Questionnaire Response (Sept. 6, 1995), at Ex. A-13T, 12, C.R. 1, Pl.’s App., Tab 1. Aramid’s long term debt in 1994 amounted to more than [ ]% of the value of its sales and almost [ ]% of its costs of sales. Id. at 11-13. As part of the restructuring, changes occurred to the records of both companies, including Akzo consolidating Aramid’s operating results for financial reporting purposes, pursuant to Dutch GAAP See Section D Supplemental Questionnaire Response (Dec. 11,1995), at 2-10, [21] C.R. 12, Def.’s App., Ex. 8. It is Commerce’s reliance on the consolidated results that is at the center of this issue. In the Final Results, Commerce calculated the respondent’s net interest expense based on Akzo’s consolidated financial statements rather than on Aramid’s individual financial statements. 61 Fed. Reg. at 51, 407. Specifically, Commerce stated: In general, this practice recognizes the fungible nature of invested capital resources (i.e., debt and equity) within a consolidated group of companies. * * * The controlling entity within a consolidated group has the “power” to determine the capital structure of each member company within the group. In this case, Akzo Nobel maintains a controlling interest in Aramid and includes the company in its consolidated financial statements. Furthermore, the [Statement of Administrative Action] and new law do not address any specific change in the Department’s practice of calculating interest expense. Id. (citation omitted). DuPont contests Commerce’s decision, arguing that Commerce’s reliance on Akzo’s net interest expense is not in accordance with the 1994 amendments to the antidumping statute and distorts Aramid’s real financing costs.3 A. URAA Amendments DuPont argues that the Statement of Administrative Action (“SAA”)4 provides clear and authoritative guidance as to how Commerce should calculate Aramid’s financing costs in this case. Specifically, DuPont argues that the SAA requires that: (1) Commerce’s calculation must be based on the records of the producer (i.e. Aramid) and accurately capture all of the actual costs incurred in producing or selling the product under investigation or review, (2) Commerce must ensure that the cost information it uses in such calculations is a representative measure of materials, labor and other costs, including financing costs, incurred to produce the subject merchandise, and (3) Commerce must adjust respondent’s reported costs if it determines that such costs, including financing costs, have been shifted away from production of the subject merchandise or are artificially reduced.5 See Statement of Administrative Action, accompanying H.R. 5110,103rd Cong., at 834-35 (1994), re [22] printed in 1994 U.S.C.C.A.N. 3773, 4171-72 [hereinafter “SAA”]. Thus, DuPont asserts that Commerce must now emphasize real financing costs of the entity producingthe subject merchandise and may no longer calculate financing expenses on the basis of unrepresentative costs derived from consolidated financial statements. Contrary to DuPont’s assertion, neither the SAA nor the amended an-tidumping statute, 19 U.S.C. § 1677b(f)(l) (1994), mandate a change of practice related to the narrow issue before the court of whether the net interest expense calculated from consolidated financial statements is “representative” of the actual costs incurred to produce the subject merchandise. Rather, the cited portion of the SAA suggests only two distinct changes in the law that do not affect Commerce’s past practice at issue here: elimination of the statutory mínimums for the selling, general and administrative (“SG&A”) expense factor in the constructed value calculation (“CV”)6 and codification of Commerce’s practice of applying GAAP unless such an application would distort costs.7 Moreover, the language relied upon by DuPont does not contradict Commerce’s past practice. In following its past practice of applying the net interest expense based on consolidated financial statements, Commerce aimed to allocate costs in a manner that “reasonably reflects” the “actual costs.” In addition, as discussed below, in allocating financing expenses, “all available evidence” properly includes the degree of financial integration of a corporate group and the degree of control exercised by the parent of the group — factors relied upon heavily by Commerce in determining how to calculate financing expenses in a parent-subsidiary situation. B. Commerce’s Interpretation is Reasonable Having determined that the statute does not directly address this specific issue, the court will uphold Commerce’s interpretation of the stat [23] ute if it is reasonable. Here, Commerce applied its long standing policy of calculating interest expense for COP/CV purposes from the borrowing costs incurred by the consolidated group of companies rather than the individual producer. Commerce has repeatedly explained its rationale for this policy: We recognize the fungible nature of a corporation’s invested capital resources, including debt and equity, and we do not allocate corporate financing expenses to individual divisions of a corporation on the basis of sales per division. Instead, we allocate the interest expense related to the debt portion of the capitalization of the corporation, as we appropriate, to the total operations of the consolidated corporation. * * * # * * * [Respondent] maintains that all interest expenses incurred by [the parent] pertain solely to the parent’s operations. Under this principle, [respondent] would have us accept that its parent funds its own operations from borrowing while, at the same time, funding its investment in [respondent] solely through equity capital. Such a principle ignores the fact that [the parent]’s capital structure is comprised of both debt and equity. Therefore, it is neither possible, nor appropriate, in our analysis to allow the company to pick and chose which portions of its parent’s operation should incur the additional interest costs associated with borrowed funds. Silicon Metal from Brazil, 61 Fed. Reg. 46,763, 46,773-74 (Dep’t Commerce 1996) (final results of admin, rev.); see also Small Diameter Circular. Seamless Carbon and Alloy Steel, Standard, Line and Pressure Pipe fromltaly, 60 Fed. Reg. 31,981, 31,990 (Dep’tCommerce 1995) (final de-term.)(“[Commerce’s] longstanding practice is to calculate interest expense for COP/CV purposes from the borrowing costs incurred by the consolidated group.”). This methodology, upheld by the Court of International Trade, is based on the fact that the consolidated group’s controlling entity has the power to determine the capital structure of each member of the group. Gulf States Tube Div. of Quanex Corp. v. United States, Slip Op. 97-124, 38-39,1997 WL 633221, at *18 (Ct. Int’l Trade Aug. 29, 1997). Majority equity ownership is prima facie evidence of such corporate control. Id. at 40, 1997 WL 633221, at *17; New Minivans from Japan, 57 Fed. Reg. 21,937, 21,946 (Dep’t Commerce 1992) (final determ.). DuPont does not directly challenge this general principle. Instead, DuPont relies on both an administrative determination, Polyethylene Terephthalate Film, Sheet, and Strip from the Republic of Korea, 56 Fed. Reg. 16,305 (Dep’t Commerce 1991) (final determ.) [hereinafter “PET Film”], and a case, Gulf States Tube Division of Quanex Corp. v. United States, Slip Op. 97-124, 1997 WL 633221 (Ct. Int’l Trade Aug. 29,1997), as support for the proposition that notwithstanding a respondent’s membership in a broader corporate structure that affects its capital resources and management control, Commerce should depart from its general rule when company-specific data provides the most accurate [24] picture of a respondent’s true financial requirements and borrowing costs.8 Neither supports plaintiffs position. In PET Film, Commerce noted that in this case, the company’s own financial statements and not those of the group were more accurate because the * * * [g]roup prepares combined rather than consolidated financial statements, based largely on limited brother-sister corporate relationships and the existence of some degree of common management control between the group’s companies. Respondent did not present any evidence of inter-company production financing arrangements, either through debt or equity, within the * * * group that would lead us to conclude that [the company’s] PET film financing costs were most accurately depicted at the combined group level. PET Film, 56 Fed. Reg. at 16,313. The specific facts relied upon in PET Film, the lack of control by any one company within the group, the lack of consolidated financial statements, and the lack of inter-company financing arrangements, create an exception to the general rule because they are the opposite of the facts needed for the general rule’s application — mainly, control through majority equity ownership. Here, however, there is a consolidated company, with Akzo owning 95% of Aramid, consolidated financial statements and intercompany financing. Clearly, the facts of this case distinguish it from the fact specific exception found in PET Film. DuPont also cites Gulf States9 for the same proposition that Commerce shall depart from its general practice of calculating expenses using a parent company’s consolidated financial statements where substantial evidence on the record demonstrates that this practice would distort actual financing expenses. Moreover, DuPont argues that the evidence on the record here, in light of Gulf States demonstrates a distortion of Aramid’s financing expenses. This reasoning is problematic. The court in Gulf States, after noting that such a departure may be possible in some situations, did not depart from the general practice. Slip Op. 97-124, at 41-42, 1997 WL 633221, at *20. Instead the court noted: Although it might also be reasonable for Commerce to have based its calculations on [respondent’s own] consolidated statements, the [25] Court may only determine whether Commerce’s method was a reasonable one. In view of the presumption that majority equity ownership of a subsidiary as prima facie evidence of corporate control, Commerce’s use of [the parent’s] consolidated expense is reasonable. Id. Implicit in this holding is that the respondent must rebut the presumption of control to indicate that the use of consolidated expenses distort the actual costs. See id.; see also New Minivans from Japan, 57 Fed. Reg. at 21,946 (respondent did not present evidence sufficient to show that Commerce’s practice would result in distortion because it did not rebut prima facie evidence of corporate control found in consolidation); Small Diameter Circular Seamless Carbon and Alloy Steel, Standard, Line and Pressure Pipe from Italy, 60 Fed. Reg. at 31,990 (claim that parent did not exercise control over respondent’s operations rejected because majority equity ownership is prima facie evidence of corporate control and respondent did not provide evidence of other distortion). Here, DuPont has not cited evidence which would overcome the presumption of corporate control, as it is insufficient to demonstrate that had Commerce relied only on Aramid’s financial statements, that determination would have been supported by substantial evidence as well. Instead, the record supports Commerce’s findingthat Akzo was the controlling entity with 95% ownership of Aramid. Aramid is the consolidated subsidiary of Akzo Nobel. Section A Questionnaire Response (Sept. 6,1995), at 2, C.R. 1, Def.-Intervenors’ App., Tab C.R. 1. With the exception of loans from decades past, all of Aramid’s borrowing was handled by Akzo. Id. at 2, Ex. A-13T, 9. Akzo finances the entire group of Akzo companies and determines which part of an operation is financed through equity or debt. Rebuttal Brief (Aug. 5,1996) at 22, C.R. 30, PL’s App., Tab 30. This is uncontested prima facie evidence of control. In sum, DuPont is seeking to segregate out costs by each division of a consolidated company in order to prove dumping. Antidumping law, structured for product to product comparisons, however, does not mandate such an investigation. The law “does not counteract all conduct of harm to plaintiff Furthermore, it does not contain every safeguard possible to prevent avoidance of dumping margins by foreign producers in situations where their conduct might seem to plaintiff to warrant duties. The antidumping laws counteract dumping in a very specific way.” Monsanto Co. v. United States, 12 CIT 937, 940, 698 F. Supp. 275, 278 (1988). Thus, Commerce’s determination is sustained. II. For purposes of the administrative review, Akzo submitted amortized goodwill calculated under U.S. GAAP Section D Supplemental Questionnaire Response (Dec. 11, 1995), at 8, C.R. 12, Def.-Intervenors’ App., Tab 12. Thus, Akzo Nobel amortized the goodwill over its asserted useful life. Id. The goodwill included the following elements: (1) write-down of fixed assets, (2) write-off of obsolete inventory and waste dis [26] posal, (3) write-down of beginning inventory, and (4) residual goodwill. See Def.-Intervenors’ Brief, Ex. 1,1-3; Adjustments to COP and CV Information for Final Results (Aug. 29,1996), at 1-2, C.R. 32, Def.-Inter-venors’ App., Tab 32; Cost of Production Analysis for Final Results of Review (Sept. 25,1996), at 1-4, C.R. 33, Def.-Intervenors’ App., Tab 33. Akzo included the amortized portion of the goodwill charge in its reported general and administrative expenses. Section D Supplemental Questionnaire Response (Dec. 11,1995), at Ex. D-33,1,C.R. 12, Def.-In-tervenors’ App., Tab 12. Commerce agreed with Akzo’s approach for the preliminary determination, but became concerned that the amortization of the goodwill over a substantial number of years10 might fail to capture the full costs of the assets. Cost of Production Analysis for Final Results — Issues Regarding Write Down of Fixed Assets (July 11,1996), at 2, C.R. 26, Def.-Intervenors’ App., Tab 26. In a July 11th memo, Commerce questioned whether it was appropriate to accept Akzo’s goodwill accounting treatment or adjust Aramid’s costs to recognize the depreciation expense for the devalued assets. Id. This concern arose because Akzo reported that the assets were used during the period of review to produce the subject merchandise and had a very short remaining useful life.11 Section D Supplemental Questionnaire Response, at 5, C.R. 24, Def.-Intervenors’ App., Tab 24. Commerce concluded that: [i]n determining the cost of producing the subject merchandise for antidumping investigations, the cost of these fixed assets must be recognized in full, whether through annual depreciation amounts or certain one time charges. We do not consider it appropriate for Akzo to writedown the value of Aramid’s fixed assets and then not recognize fully the resultant expense for purposes of computing COP and CV Accordingly we consider it appropriate to account for these fixed asset writedowns in the cost of producing the subject merchandise during the POR. Cost of Production Analysis (Sept. 25,1996), at 3, C.R. 33, Def’s App., Ex. 9. Therefore, Commerce increased Akzo’s fixed overhead costs by depreciating Aramid’s assets over the remaining useful life, based on Aramid’s historical accounting treatment of the assets. Id.; see also Adjustments to COP and CV Value Information (Aug. 29,1996), at 1, C.R. 32, Def.-Intervenors’ App., Tab 32. This in effect shifted the focus of the calculation from the consolidated level (Akzo) to the individual company level (Aramid). Commerce also increased product costs by including, as a one time charge, the amounts related to the write-off of obsolete material and waste disposal. Cost of Production Analysis, at 4, C.R. 33, Def.-Interve-nors’ App., Tab 33. The two remaining components, a write-down of be [27] ginning inventory to reflect reduced depreciation expenses and residual goodwill related to Akzo’s conversion of loans to equity, were excluded by Commerce and are discussed in detail below. A. Treatment of Aramid’s Beginning Inventory Write-down Following Akzo’s acquisition of the additional 45% equity interest in Aramid, Aramid wrote-down the book value of its fixed assets to conform with Akzo’s accounting methods and policies. Supplemental Questionnaire D Response, at 2, C.R. 12, Def’s App., Ex. 8. As a result, Aramid’s annual depreciation expense decreased, which lowered the standard cost of producing the subject merchandise. Id. at Ex. 27-T, 2. Because Aramid records its finished goods and work-in-progress inventory at standard cost, Aramid then wrote-down the value of its 1994 beginning inventory to reflect the lower depreciation expense contained within the standard cost of production. See id.; see also Cost of Produc-tionAnalysis, at 4, C.R. 33, Def.’s App., Ex. 9. The 1994beginninginven-tory was revalued based upon the 1994 standard costs, which included the lower depreciation expense resulting from the fixed asset write-down. Supplemental Section D Questionnaire Response, at 2, Def. App., Ex. 8. Commerce in the Final Results and as discussed above, disallowed the goodwill charge to equity arising from the write-down in fixed assets and instead accounted for the write-down in the cost of producing the subject merchandise during the period of review. Commerce allowed, however, the goodwill charge arising from the beginning inventory write-down. Commerce articulated the basis for this contested decision as: As a result of Aramid writing down its fixed assets, the company also wrote down the value of its beginning inventory as of January 1,1994. Aramid values its inventory at standard cost (variances are not included in inventory), part of which is depreciation. Since annual depreciation was reduced due to the fixed asset write-down, Aramid adjusted its inventory value to reflect this reduced standard cost for depreciation. # Although the beginning inventory was written down and some of this inventory was in an intermediate stage (e.g. polymer) and, therefore, used to produce subject merchandise during the POR, the writedown did not affect production costs during the POR. The write-down to beginning inventory resulted in beginninginventory as of January 1,1994, which was stated at 1993 standard costs, being valued at 1994 standard costs. Because the reported POR production was valued using the same 1994 standard costs, the beginning inventory adjustment had no effect on the current years costs. Accordingly, we made no adjustment for the inventory write-down. Cost of Production Analysis, at 3-4, C.R. 33, Def.’s App., Ex. 9. [28] DuPont objects to Commerce’s determination that the 1994 inventory write-down had no impact on the production costs incurred during the POR.12 DuPont argues that under the basic rules of cost accounting, cost of goods sold during a given period is the sum of beginning inventory plus period manufacturing costs less ending inventory. See Charles T. Horngren & George Foster, Cost Accounting AManagerial Emphasis 32 (6th Ed. 1987). DuPont argues that Akzo’s write-down of 1994 beginning inventory resulted in a corresponding reduction in the cost of sales for the period of review, by reducing the beginning inventory component of the “cost of sales” equation. Thus, because Aramid values inventories at the lower of cost or net realizable value, where cost is defined as the full manufacturing cost related to the stage of processing, any reduction in the value of inventory as of December 31,1993 undertaken by Akzo is a reduction from Aramid’s full manufacturing cost, which Commerce should not have accepted. Moreover, DuPont claims that Commerce is wrong in asserting that it “in effect” accounted for the portion of the inventory write-down arising from the fixed asset reduction. DuPont claims that in no way was the write-down of 1994 beginning inventory dictated by the revaluation of fixed assets. The 1994 beginning inventory write-down was a reduction of 1993 costs and was unaffected by the 1994,1995, 1996 and 1997 depreciation components of the fixed asset write-down. It is these 1993 costs, including 1993 depreciation, that are at issue during the first part of the period of review regarding the actual costs of sales of PPD-T ara-mid fiber from inventory. Thus, DuPont argues that to include the 1994 beginning inventory write-down in the cost of production would not double-count the fixed asset reduction. DuPont, however, mischaracterizes Commerce’s treatment of the inventory write-down. If Commerce were to follow DuPont’s proposal to increase Aramid’s COP to include the beginning inventory write-down, Commerce’s calculations would be distorted because it would double count the amount of depreciation included in Aramid’s COP and CV As a result of Commerce’s disallowance of Aramid’s adjustment, Commerce increased the reported cost of production on a per kilogram basis for all of 1994, including whatever was in progress at the time, specifically the 1993 year end inventory, by adding back the depreciation associated with the assets that were written-down. This increase affected all of 1994 because the one increase in cost of production con [29] cerned what was in inventory to begin with as the beginning inventory is used to produce merchandise during the year. Thus, contrary to DuPont’s assertion, because Commerce based the cost of manufacturing upon Aramid’s 1994 standard costs, increased by the amount of the disallowed adjustment, there was no need to further adjust Aramid’s reported inventory figure. To do otherwise would increase the COP and CV amount by double counting the depreciation expense. B. Residual Goodwill Expense The last remaining element of the goodwill expense is the residual goodwill or the difference between the value of the forgiven debt and the net book value of Aramid in its financial records. Cost of Production Analysis, at 1-3, C.R. 33, Def.’s App., Ex. 9. Commerce did not account for Akzo’s goodwill amortization expense because: [t]he remaining goodwill represents a cost to Akzo not Aramid, the producer. This goodwill amount reflects the amount Akzo paid for its additional equity interest in Aramid in excess of Aramid’s book value. Thus, we do not consider it to be a production cost incurred by Aramid for the subject merchandise. Cost of Production Analysis, at 4, C.R. 33, Def.’s App., Ex. 9. Plaintiff argues that this determination is not supported by substantial evidence because the residual goodwill is directly related to Ara-mid’s production of the subject merchandise.13 Plaintiff suggests that the residual goodwill is a direct consequence of the restructuring, was created by eliminating from Aramid’s hooks loans that Akzo made to Aramid, and was simultaneously recognized as an offsetting cost on Akzo’s books. Moreover, these loans were made to assist in financingthe production of PPD-T aramid fiber. Thus, DuPont argues that the loans and therefore the related goodwill that eliminated the loans, are directly tied to the production of the subject merchandise. This argument is unpersuasive in light of Commerce’s treatment of the fixed asset write-down. The write-down of fixed assets, as the largest component of Commerce’s adjustment, appeared only on Aramid’s balance sheet as a direct charge to equity, while the goodwill amount was only on Akzo’s balance sheet. Accordingly, Commerce’s decision to revert to Aramid’s historic accounting methodology by depreciating assets over the remaining useful life was appropriate only if it eliminated the entire goodwill amount from the calculation. There was no basis on the record that would condone the result advocated by DuPont of segregating the write-down of Aramid’s assets from the goodwill and recalculating the goodwill amount carried by Akzo. Neither element of the [30] calculation would trace to either company’s audited records. Thus, because Commerce shifted the focus to the individual company level, it then had to carry through this approach and eliminate the goodwill amount from the reported G&A expense. This correlative elimination was the only way Commerce could justify its determination to isolate the write-down amount and continue depreciating the assets as if the acquisition had not occurred. Accordingly, the court sustains Commerce’s treatment of goodwill. III. A. Background In the beginning of the administrative review process, DuPont made numerous suggestions to Commerce aimed at ensuring Akzo’s cost data within the new corporate structure, as well as the manner in which it was reported, was adequate and reliable. Memorandum from Case Analyst to File of 7/27/95, at 1, PR. 10, Def.’s App., Ex. 10 (exparte meeting with DuPont); Letter from DuPont to Commerce of 8/1/95, at 2, PR. 5, Def.’s App., Ex. 2; Letter from DuPont to Commerce of 9/14/95, at 4-5, C.R. 2, Def.’s App., Ex. 11; Letter from DuPont to Commerce of9/28/95, at 9-10, C.R. 4, Def.’s App., Ex. 12; Letter from DuPont to Commerce of 11/9/95, at 6-7, C.R. 10, Def’s App., Ex. 13. In response, Commerce added questions to the original Section D Questionnaire, as well as to the Section D Supplemental Questionnaire. Section D Questionnaire (Aug. 7, 1995), PR. 7, Def.’s App., Ex. 1; Section D Supplemental Questionnaire (Nov. 20,1995), C.R. 11, Def.’s App., Ex. 4. Akzo, in turn, provided detailed responses to Commerce’s Questionnaires. Section D Questionnaire Response (Oct. 11,1995), C.R. 5, Def.’s App., Ex. 3; Section D Supplemental Questionnaire Response (Dec. 11,1995), C.R. 12, Def.’s App., Ex. 8. On January 31, 1996, Commerce received from DuPont the first of two letters, suggesting that Akzo had manipulated its cost allocation system in order to shift costs away from certain high volume products that the company had sold in the U.S. market during the POR. Letter from DuPont to Commerce of 1/31/96, at 2, C.R. 13, Def.’s App., Ex. 14. DuPont asserted that, in order to prove its allegation, Commerce must place in the administrative record of the first review the product-specific cost data that Akzo originally submitted in the LTFV investigation. Id. at 11-12. In the second letter, which constituted a follow-up submission, DuPont recommended including a specific set of questions in the Second Supplemental Section D Questionnaire. Letter from DuPont to Commerce of 2/7/96, at 2-9, C.R. 15, Def.’s App., Ex. 15. Akzo responded to DuPont’s allegations in its submission of February 28, 1996. See Letter from Akzo to Commerce of 2/28/96, C.R. 18, Def.’s App., Ex. 16 [hereinafter “February Memorandum”]. First, DuPont claimed one test of the reasonableness of product-specific costs would compare the gross profits and net profits of the U.S. and home market products against the overall profit figures for world-wide sales. Id. at 5. [31] Akzo responded by noting that it is unrepresentative to compare the results- for a small percentage of total sales with the worldwide results, particular when the product mix is considered. Id. at 6-7. Second, DuPont claimed that Commerce should compare the cost of high volume exports to the U. S. with the costs of products that are not exported or are exported in low volume. Id. at 6. Akzo responded that the fact that high volume items have a lower cost than more sophisticated products is the norm, and any other result would be unusual. Id. at 8-10. The largest volume of products sold in the U.S. have the lowest cost because they are the easiest to produce and involve the fewest production steps, resulting in the highest yield rates. Id. at 9. Thus, the cost differences between high volume and low volume goods were fully justified, as it directly related to a number of legitimate factors. Id. at 8-10. Third, DuPont suggests testing the reasonableness by comparisons of Akzo’s submitted costs, taking into account different physical characteristics. Id. at 6. Akzo’s February Memorandum demonstrates that DuPont’s claim works on a flawed assumption of how the goods are produced and does not account for additional factors that affect the calculation of the weighted average cost. Id. at 12-14. Commerce both reviewed the submissions and held a series of ex parte meetings with the parties. See Memorandum from Case Analyst to File of 3/7/96, at 1, PR. 51-52, Def’s App., Ex. 17. In its preliminary determination on April 9, 1996, Commerce stated that DuPont’s cost shifting allegations were untimely for this review. Aramid Fiber Formed of Poly Para-Phenylene Terephthalamide from The Netherlands, 61 Fed. Reg. 15,766,15,767 (Dep’t Commerce 1996) (prel. determ.). On May 16, 1996, however, Commerce held a meeting with DuPont to further discuss the cost issues, Memorandum from Case Analyst to File of 5/20/96, at 1, PR. 68, Def.’s App., Ex. 18, and issued another Section D Supplemental Questionnaire to which Akzo responded, Supplemental Section D Questionnaire (May 30, 1996), C.R. 23, Def.’s App., Ex. 5. In the Final Results, Commerce concluded that there were no reasonable grounds to challenge Akzo’s methodology of allocating production costs and cost variances based on product-specific standard costs. 61 Fed. Reg. at 51,407. Commerce reasoned that this approach was consistent with Akzo’s normal accounting practices and noted that it had previously accepted this method for a different Akzo-affiliated company in another determination. Id. Furthermore, Commerce did not accede to plaintiffs request to include the data of standard costs from the LTFV investigation in this review, noting that it would only “serve to confuse and complicate the issue rather than provide sufficient conclusive proof of cost shifting.” Id. at 51,408. Commerce offered two reasons why the requested comparison between the cost data for the investigation and the review would be meaningless: First, in the initial investigation, the Department relied on third country sales and cost data as the basis for fair value. In this review, [32] however, we are relying on Akzo’s home market sales as the basis for normal value * * *. Because of this change in comparison markets, it is highly doubtful that, between the two proceedings, there would be a sufficient number of common non-U.S. products from which to draw any conclusions regarding revisions to standard costs. Secondly, Akzo grouped together products defined as “identical” in accordance with our hierarchy of physical characteristics. Akzo then computed a single weighted-average standard cost for these products based on production quantities. From the investigation to this review, changes in relative production quantities of the various Akzo products within any single product group could significantly change the weighted-average standard costs Akzo submitted. Thus, we determined that conducting a meaningful comparison of cost figures between the two segments would be difficult. Id. Asa result, Commerce concluded that it was inappropriate to include the standard costs from the investigation in the record for this review. Id. Before the court, DuPont reiterates its claim that Akzo may have understated its production costs by inappropriately decreasing its normal standard costs for certain products sold in the U.S. market and increasing the standard costs for the products not sold in the United States. DuPont, relying on a proposed regulation, Proposed Rules, Antidumping and Countervailing Duty Proceedings, 61 Fed. Reg. 4826 (Dep’t Commerce 1996) (new regulation 19 C.F.R. § 351.306 to allow information from earlier proceeding to be placed in the record of a subsequent proceeding), and the general principle that an agency decision must have a rational basis, asks the court to determine that Commerce erroneously rejected its request to include Akzo’s standard costs from the LTFV investigation on the record in this review. B. Discussion Commerce enjoys broad discretion in conducting investigations and reviews under the antidumping statute, particularly in assessing the scope of its investigations and decisions regarding relevant evidence. See Ceramica Regiomontana, S.A. v. United States, 10 CIT 399, 404-05, 636 F. Supp. 961, 966 (1986), aff’d, 810 F.2d 1137 (Fed. Cir. 1987). Commerce’s longstanding practice, upheld by this court, is to treat each segment of an antidumping proceeding, including the antidumping investigation and the administrative reviews that may follow, as independent proceedings with separate records and which lead to independent determinations. See Outokumpu Copper Rolled Products AB v. United States, 17 CIT 848, 854, 829 F. Supp. 1371, 1377 (1993) (anti-dumpinginvestigations and administrative reviews are wholly independent proceedings); see also Sugiyama Chain Co. v. United States, 18 CIT 423, 431, 852 F. Supp. 1103, 1110 (1994) (Commerce cannot base decision to terminate administrative review on factors related to or discovered in other review periods); Tapered Roller Bearings and Parts [33] Thereof, Finished and Unfinished, from the People’s Republic of China, 61 Fed. Reg. 65,527, 65,538 (Dep’t Commerce 1996)(final results of admin, rev.) (Commerce refused to apply verification results from one review period to two other review periods). Here, DuPont requests that the court order Commerce to consider the cost data provided in the prior LTFV investigation in reaching a decision about cost shifting in the present administrative review. DuPont’s request clearly requires conduct contrary to Commerce’s normal practice. Notwithstanding the fact that Commerce was under no general obligation to incorporate the datafrom the prior investigation into the current record, Commerce granted DuPont numerous opportunities to present its allegations. DuPont, however, at no time presented persuasive evidence of cost shifting. Specifically, DuPont did not direct Commerce’s attention to any flaws in Akzo’s rebuttal comments in the February Memorandum or any additional proof in the current administrative record of cost manipulation, thus providing Commerce with no reason to expand the record. Moreover, DuPont fails to demonstrate that Commerce’s determination lacked a rational basis. Commerce provided two reasons why in this particular case, it would not grant such a request: (1) the initial investigation relied on third country sales, while the administrative review relied on home market sales as the basis for normal value, and (2) any change in the production quantities of the products within a single product group could change the weighted-average standard costs submitted. Final Result, 61 Fed. Reg. at 51,408. As to the first reason, DuPont suggests that the destination of the product, whether a third country or the home market, would not change the cost of producing the merchandise. This may be a valid point, but it is not one that negates Commerce’s reasoning. Commerce specifically noted that the difference in markets was relevant because “it is highly doubtful that, between the two proceedings, there would be a sufficient number of common non-U. S. products from which to draw any conclusions regarding revisions to standard costs,” not because the production costs would change by destination. Id. This reasoning, in addition to the likely change in production quantities within product groups based on changes in supply and demand, support Commerce’s decision to restrict the breadth of the investigation. IV A. Background On October 10,1996, DuPont submitted a timely request to correct a ministerial error contained in Commerce’s model-match and margin calculation computer program used in the Final Results. DuPont’s Ministerial Error Allegation (Oct. 10,1996), C.R. 35, Def.’s App., Ex. 19. DuPont’s request alleged that as a result of the computer error, Commerce prematurely resorted to CV when the most similar home market matches were sold below cost and should have instead resorted to the other home market sales that were above cost. Id. at 2-3. [34] In response, Commerce explained: [i]t is the Department’s longstanding practice to match US sales to comparison market sales based “on the similarity of the merchandise only and not on whether the most similar model is sold above cost, section 771(16), [19 U.S.C. § 1677(16) (1994)], appears to direct us to the use of constructed value when the most similar model is sold below cost.” Decision Memorandum Regarding the Ministerial Error Allegation (Nov. 5, 1996), at 2, C.R. 37, Def’s App., Ex. 20. Thus, because the request was not to correct a ministerial error but rather to change an agency policy, Commerce did not make any changes to the final margin calculations. Id. B. Discussion Section 1675(h) of Title 19 provides that “ministerial errors” include “errors in addition, subtraction, or other arithmetic function, clerical errors resulting from inaccurate copying, duplication, or the like, and any other type of unintentional error which the administering authority considers ministerial.” Id. (emphasis added). Commerce may correct ministerial errors in final determinations. 19 U.S.C. § 1675(h) (1994). Moreover, the court has recognized that the statute grants Commerce broad discretion to determine what constitutes a ministerial error. See Cemex, S.A. v. United States, Ct. No. 93-10-00659, Slip Op. 95-72, 1995 WL 251561, at *5 (Ct. Int’l Trade Apr. 24, 1995) (“Commerce is given fairly broad discretion to determine which types of unintentional error to regard as ministerial.”). Plaintiff contests Commerce’s rejection of its request to correct the ministerial error. In its brief before the court, defendant bifurcates DuPont’s allegation of one ministerial error into two separate potential errors. Both alleged ministerial errors will be addressed below. 1. First Error DuPont alleges that Commerce’s computer program erred because when the most similar foreign model was sold below cost, Commerce should have used another lower-ranked match when that same foreign model was used as the best match for a different U.S. sale elsewhere in the review.14 Commerce, however, correctly decided as to this alleged error that DuPont seeks a change of a policy, not a correction of a ministerial error. It is Commerce’s long-standing policy that: Once the model matches are established and the COP test is completed, Commerce is not required to reexamine all of the undifferentiated model data in order to make new matches and price comparisons on the basis of whatever subset of lower-ranked such [35] or similar merchandise survives the COP test. Section 1677b(b) [of Title 19] does not direct such a result. Zenith Elecs. Corp. v. United States, 18 CIT 1145, 1154, 872 F. Supp. 992, 1000 (1994). Moreover, Commerce limits the universe of potential matches to those sales within 90 days before and 60 days after the month of the U.S. sale. This period is known as the “90/60 day contemporaneity window.” See 19 U.S.C. § 1677b(a)(l)(A); Antifriction Bearings (Other Than Tapered Roller Bearings) and Parts Thereof from France, Germany, Italy, Japan, Singapore and the United Kingdom, 62 Fed. Reg. 2081, 2111 (Dep’t Commerce 1997) (final results of admin, rev.) (applying post-URAAlaw, Commerce “resorted directly to CV where we have disregarded all contemporaneous identical HM sales as below cost” in the 90/60 day contemporaneity window); Certain Circular Welded Carbon Steel Pipes and Tubes from Thailand, 61 Fed. Reg. 1328, 1332 (Dep’t Commerce 1996) (final results of admin, rev.)(“consistent with established Department practice, we will also continue to limit the universe of sales from which we select the comparison model to those sales made during the contemporaneous 90/60 window.”). Thus, the dispute is whether the matches DuPont considers as suitable alternatives to the originally matched butbelow cost sales are “lower ranked” or are the within the same level of similar merchandise as the original match. According to DuPont, a suitable “second best” match would be a home market sale that was considered to be the best match for the same U.S. model in another period involving a different window of contemporaneity. Such a match would not be “lower ranked” DuPont argues because it was considered a similar or identical match at other times during the period of review. The foreign models considered within each 90/60 contemporaneity window, however, are analyzed separately; thus, the matching process within one 90/60 day window should not be confused with the matching process within another window. The fact that a certain match was the most similar merchandise in one window, does not mean that the same match would be the most similar in another window. Essentially, DuPont is asking Commerce to continue its model-match exercise after disregarding the below cost sales at the first level to find additional matches. Thus, Commerce correctly denied DuPont’s request as it clearly sought to change an established policy, not to correct a ministerial error. 2. Second Error The court finds that Commerce prematurely resorted to CV when, in a slight variation of the error alleged above, the computer program rejected the most similar matching home market models, the sales of which occurred in a different month, but within the same 90/60 day window. See Pl.’s Conf. Br. at 58 (examples of mistake). While defendant consents to a remand to correct this error, defendant-intervenors would prefer completing the appeal process quickly rather than obtaining the [36] limited benefit they would incur from the remand through the decrease in the dumping margin.15 Although DuPont claimed for the first time at oral argument that the limited remand would benefit it, DuPont has not presented any evidence of this to the court. Thus, as it appears only defendant-intervenors would benefit and they do not seek remand, Commerce’s determination is sustained it its entirety. Footnotes [1] Akzo Nobel Fibers Inc. is a subsidiary of Akzo Nobel N.V During the period of review, Akzo Nobel N.Y increasedits holding in Aramid to 95%. [2] Commerce conducted this review under the Tariff Act of 1930, as amended by the Uruguay Round Agreements Act (“URAA”). [3] DuPont argues that the consolidated net interest expense grossly understates Aramid’s real financing costs. Under the consolidated expense, Akzo’s reported net interest expense as a percentage of cost of goods sold is []%. Section D Questionnaire Response, at Ex. D-19,1, C.R. 5, Def.’s App., Ex. 3. For purposes of including an amount for financing expenses in Aramid’s constructed value, Akzo used consolidated company data to report a percentage of net interest expenses on assets other than accounts receivable and finished goods inventory amounting to [ ]%. Id. at 2. In contrast, DuPont argues that if Commerce used Aramid’s own financial statements, its net interest expense would, as apercent-ageofits cost of sales, be [ 1%. Section A Questionnaire Response (Sept. 6,1995), at Ex. A13T, 12-13, C.R. 1, Pl.’s App., Tab 1; Section D Supplemental Questionnaire Response (June 10,1996), at 12, C.R. 24, PL’s App., Tab 24. [4] The Statement of Administrative Action is Man authoritative expression by the United States concerning the interpretation and application of the Uruguay Round Agreements and this Act in any judicial proceeding in which a question arises concerning such interpretation or application.” Delverde, SrL v. United States, Slip Op. 97-163, at 33 n.18, 1997 WL 784640, at *15 n.18 (Ct. Int’l Trade Dec. 2,1997) (quoting 19 U.S.C. § 3512(d)). [5] The SAA states: a. Calculation of Costs in General Commerce will consider all available evidence submittedby the exporter or producer on a timely basis regarding the proper allocation of costs. * * * [22] In determining whether a company’s records reasonably reflect costs, Commerce will consider U.S. generally acceptedaccountingprinciples employed by the industry in question. * * * Costs shall be allocatedusinga method that reasonably reflects and accurately captures all of the actual costs incurred in producing and selling the product under investigation or review. In determining whether to accept the cost allocation methods proposed by a specific producer, Commerce will consider the production cost information available to the producer and whether such information could reasonably be used to compute a representative measure of the materials, labor and other costs, including financing costs, incurred to produce the subject merchandise, or the foreign like product. * * * Also, if Commerce determines that costs, including financing costs, have been shifted away from production of the subject merchandise, or the foreign like product, it will adjust costs appropriately, toensure they are not artificially reduced. SAA, at 834-35, 1994 U.S.C.C.A.N. at 4171-72 (emphasis added). [6] Defendafit-intervenors argue that the URAA’s required use of “actual data” was intended to eliminate the 10% and 8% statutory mínimums in the calculation of general expenses in the CV determination and was not intended to change Commerce’s cost of production (“COP”) calculation. 19 U.S.C. § 1677b(e)(2)(A) (1994) (calculating CV requires the use of “actual amounts incurred and realized by the specific exporter or producer * * * for selling, general, and administrative expenses.”); see also S. Rep. No. 103-412, at 74 (1994) (“requirements of Article 2.2.2 with respect to the use of actual data for administrative, selling and general costs and for profits conflict with current law”). This argument is tenable on its face. The problem lies in interpreting the section of the SAA cited by plaintiff as the legislative history for this change. Plaintiff cites the section of the SAA titled “Calculation of Costs In General.” There is, however, a separate section on “Profit and Selling, General and Administrative Expenses for Constructed Value” that addresses this specific change. See SAA, at 839,1994U.S.C.C.A.N.at4175. Thus,thequotedsectionoftheSAAand its reliance on “actual costs” likely applies beyond the new method of calculated SG&A expenses for constructed value. [7] See H.R. Rep. No. 103-826(1), at 90-91 (1994), reprinted in 1994 U.S.C. C.A.N. 3773,3862-63 (“Under new section 773(f), Commerce will continue its current practice of calculating costs on the basis of records kept by the exporter or producer of the merchandise, provided such records are kept in accordance with generally accepted accounting principles of the exporting (or producing) country and reasonably reflect the costs associated with production and sale of the merchandise.”). [8] DuPont also relies on Micron Technology, Inc. v. United States, 893 F. Supp. 21, 30 (Ct. Int’l Trade 1995), opinion after remand, 1995 WL 631998, aff'd, 117 F.3d 1386 (Fed. Cir. 1997), as justification for a departure from Commerce’s normal practice because the court may approve the agency’s departure when more appropriate methods are available. Contrary to DuPont’s assertion, the court in Micron Technology stated that “I t)he choice between alternative methodologies, each of which is supported by substantial evidence and in accordance with law, is clearly left to Commerce’s discretion.” Id. The court is not to substitute its own judgment as to which choice to make. [9] DuPont claims that Gulf States was decided after the URAA amendments to the antidumping statute and resoundingly confirms DuPont’s interpretation of the amended statute. The fact that Gulf States was decided after the URAA does not indicate that the court decided the case by interpreting any new changes to the statute. Rather, the relevant fact is that the investigation was initiated on July 13,1994. Gulf States, Slip Op. 97-124, at 4,1997 WL 633221, at *2. Thus, itis quiteclearthatbecause the investigation beganbefore January 1,1995, Gulf States involves only the oldlaw. See 19 U.S.C. § 1671 (1994) (note); see also Floral Trade Council v. United States, Ct. No. 96-09-02281, Slip Op. 97-176, at 3 & n.1 (Ct. Int’l Trade Dec. 18,1997) (discussing when agency determinations and reviews are governedby post-Uruguay Round Agreements law). [10] Akzo projected a [ I year useful life. Section D Supplemental Questionnaire Response, at 8, C.R. 12, Def.-Intervenors’ App., Tab 12. [11] The assets had a remaining useful life of [ ] years. Supplemental Section D Questionnaire Response (June 10, 1996), at 6, C.R. 24, Def-Intervenors’ App., Tab 24. [12] DuPont does not challenge directly the portion of the inventory write-down arising out of Aramid’s change in treatmentofl ] grants. Prior to the 1994corporate reorganization, Aramid recorded the fixed assets purchased with the grants at the asset’s actual cost prior to the [ 1, and recorded the value of the grants in its account called the “investment equalization account.” Supplemental Section D Questionnaire Response (Dec. 11,1995), at 3-6, C.R. 12, Def’s App., Ex. 8. Consequently, in its cost accounting treatment, Aramid recorded the depreciation expense without considering the benefit of the grants received. In December 1993, as a result of the reorganization, Aramid changed the cost accounting treatment of the grants so that their impact would be included in the product specific standard costs of production. Id. Because of the [ ], Aramid also [].Id. Thus, because Commerce determined that the amortization of these grants [ 1 production costs, Commerce accepted Aramid’s revised treatment of the grants in the first administrative review and did not add the portion of the [ ] arisingfrom the grants tothereportedCOP andCV Costof Production Analysis, at4, C.R. 33, Def.-Intervenors’ App., Tab 33. The portion related to the grants reflected only a difference in the way the grants were accounted for, with no change in results, and thus this aspect of Commerce’s determination is sustained. [13] DuPont also asserts that the SAA requires Commerce to use data that accurately captures all costs incurred in the production of the subject merchandise under review. DuPont notes the lack of a “corporate structure qualification” to this requirement and thus argues that Commerce’s determination should not have been based on where the costs were recorded, but rather solely on the relation to the cost of production. DuPont also relies on two cases, NTN Bearing Corp. of America v. United States, 905 F. Supp. 1083, 1096 (Ct. Int’l Trade 1995) and Cemex, S.A. v. United States, Ct. No. 93-10-00659, 1995 WL 251561, at *6 (Ct. Int'l Trade Apr. 24,1995), for the proposition that whether a cost is recorded is “irrelevant to whether these costs are actually incurred and should be included in [the] cost of production.” Here, however, Commerce did not base its determination on where the cost was recorded. Instead, it determined that the costs were not related to the cost of production of the subject merchandise incurred by Aramid. [14] DuPont provides the following as an example of the error. Commerce compared U.S. sales of model [) to home marketsalesofmodel[ ] in [ ] differentmonths, but resortedto constructed value rather thanmatch U.S. model [ 1 to the same home market model [ ] in [ ] other months involving multiple U.S. sales observations. See Pl.’s Brief, Attachment C. [15] Akzo notes that there are only a few products and less than [ ] transactions where the originally selected model was sold in a different month, but in the same window, and the margin actually decreases by less than [)% when this correction is made.