Krupp Thyssen Nirosta GMBH v. United States

Opinion Barzilay, Judge: This case is before the court pursuant to Plaintiffs’ USCIT R. 56.2 Motion for Judgment Upon an Agency Record. Krupp Thyssen Nirosta GmbH (“KTN”) is a German producer of stainless steel and Krupp Hoesch Steel Products, Inc. (“KHSP”) is KTN’s United States sales affiliate. Plaintiffs challenge certain aspects of the final determination of the Department of Commerce’s International Trade Administration (“Commerce” or “Department”) in the antidumping investigation of stainless steel sheet and strip from Germany. See Final Determination of Sales at Less Than Fair Value; Stainless Steel Sheet and Strip in Coils from Germany, 64 Fed. Reg. 30,710 (1999) (“Final Determination”), as amended, 64 Fed. Reg. 40,557 (1999). The court has jurisdiction pursuant to 28 U.S.C. § 1581(c) (1994). I. Background On June 30, 1998, Commerce initiated antidumping duty investigations covering stainless steel sheet and strip in coils from several countries. See Initiation of Antidumping Duty Investigations: Stainless Steel Sheet and Strip in Coils From France, Germany, Italy, Japan, Mexico, South Korea, Taiwan, and the United Kingdom, 63 Fed. Reg. 37521 (1998). Commerce determined the period of investigation to be from April 1,1997 through March 31,1998. See Final Determination, 64 Fed. Reg. at 30,712. During the period of investigation, KTN was a subsidiary of Krupp Thyssen Stainless (“KTS”). KTS was a joint venture holding company owned by two German steel manufacturers, Krupp AG Hoesch-Krupp (“Krupp”) and Thyssen Stahl AG (“Thyssen”). Krupp owned sixty percent of KTS and Thyssen owned forty percent. Thyssen also owned re *667 sellers in Germany (“German Resellers”) and in the United States (“USR”). KTN sold the subject merchandise in Germany direct from its factory and inventory, through a KTN affiliated service center specializing in resale of second quality stainless products and through the German Resellers. KTN sold the subject merchandise in the United States through KHSf^ USR and two other Thyssen owned importers. Commerce determined that KTN was affiliated with Thyssen and, through Thyssen, with Thyssen’s affiliated resellers and service centers in Germany and the United States. See Final Determination, 64 Fed. Reg. at 30,713. Because KTN failed to supply the German Resellers’ downstream sales data, Commerce found KTN failed to act to the best of its ability, and applied partial adverse facts available in the margin calculation. See id. at 30,726. Following the preliminary dumping finding, Commerce conducted a number of verifications. Only the verification of USR’s data and the accompanying results are in contention in the case before the court. Commerce conducted its verification of USR’s sales data at the end of February, 1999, and its verification of USR’s cost data at the beginning of March, 1999. In the Final Determination, Commerce rejected USR’s data entirely, applying total adverse facts available. Commerce determined that KTN failed to cooperate to the best of its ability because it failed to ensure the accuracy of its data prior to verification. See id. at 30,739. II Standard of Review In reviewing a challenge to Commerce’s determination in an anti-dumping investigation, the court is to hold unlawful a determination, finding or conclusion by Commerce that is unsupported by substantial evidence or otherwise not in accordance with law. See 19 U.S.C. § 1516a(b)(l)(B)(i) (1994). Substantial evidence is “such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.” Consolidated Edison v. NLRB, 305 U.S. 197, 229 (1938); accord Matsushita Elec. Indus, v. United States, 750 F.2d 927, 933 (Fed. Cir. 1984). III. Discussion A. German Resellers Plaintiffs challenge Commerce’s Final Determination with respect to the German Resellers on two grounds. First, they argue that Commerce’s decision to use adverse facts available was not supported by substantial evidence. Second, Plaintiffs argue that even if the use of adverse facts available was supported by substantial evidence, the particular adverse facts chosen by Commerce were not. 1. Commerce’s decision to use adverse facts available is supported by substantial evidence. Commerce is required in antidumping investigations to compare “the U.S. prices of the subject merchandise to the prices (‘normal value’) for *668 the same or similar merchandise in the home market.” Mannesmannrohren-Werke A.G. v. United States, 77 F. Supp.2d 1302, 1304 (CIT 1999) (citing 19 U.S.C. §§ 1677a, 1677b (1994)). In order to calculate and verify the difference between these prices, Commerce requests various data from parties to the investigation. If parties fail to supply the data or when the data cannot be verified, Commerce is required by statute to “use the facts otherwise available in reaching the applicable determination. * * *” 19 U.S.C. § 1677e(a). Further, if Commerce “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 * * * [Commerce] may use an inference that is adverse to the interests of that party in selecting from among the facts otherwise available.” 19 U.S.C. § 1677e(b). Plaintiffs do not dispute that the use of facts available is necessary. Rather, Plaintiffs argue that the decision by Commerce to apply adverse facts available for KTN’s alleged refusal to cooperate to the best of its ability is nót supported by substantial evidence. The court does not agree. In the Final Determination, Commerce found that KTN, the wholly owned subsidiary of KTS, a joint venture between Krupp and Thyssen, was affiliated with the German Resellers, who were wholly owned by Thyssen. See Final Determination, 64 Fed. Reg. at 30,713. KTN asserts that while the two companies may be affiliated through their common parent company, Thyssen, KTN has no operational control over the German Resellers and therefore, could not be expected to compel them to produce their sales data. In other words, KTN claims that it was unable to comply with Commerce’s data requests regarding the German Resellers, and that its failure to provide the information does not evince a failure to act to the best of its ability. The affiliation issue and its attendant issues of potential or actual operational control were thoroughly addressed by petitioners, respondents and Commerce during the investigative process. See id. at 30,713, 30,721-24; CR 34, Memorandum from Analyst/IA to File (Nov. 13, 1998); Notice of Preliminary Determination of Sales at Less Than Fair Value and Postponement of Final Determination; Stainless Steel Sheet and Strip in Coils from Germany, 64 Fed. Reg. 92, 95 (1999). Although the issue of control and a party’s ability to cooperate are linked, the facts of this case, as revealed by the record, make an analysis of the issue of whether KTN had the ability to control the German Resellers unnecessary. Therefore, the court declines to address the issue of whether actual control must be shown, and instead bases its decision on whether the application of adverse facts available was appropriate on other grounds. In World Finer Foods v. United States, 2000 WL 897752 (CIT June 26, 2000), this court found Commerce’s decision to apply adverse facts available to an interested party improper, inter alia, because such party had written to Commerce explaining its inability to supply the data. In this case, however, KTN failed to notify Commerce Department officials in writing of its claimed inability to supply data from the German Resel *669 lers despite the questionnaire’s request that it do so. See Confidential Record Doc. (“CR”) 15, Letter from Hogan & Hartson to Dep’t Commerce (Sept. 8, 1998); see also Final Determination, 64 Fed. Reg. at 30,726-27. KTN not only failed to explain its claimed inability to retrieve data from the German Resellers; it also indicated by letter that it did “not intend to report sales by [the German Resellers]” [ ]. See CR 30, Letter from Hogan & Hartson to Dep’t Commerce (Oct. 29, 1998); see also Final Determination, 64 Fed. Reg. at 30,726-27. Thus, in this case the burden never shifted to Commerce to consider KTN’s ability to supply the German Reseller’s downstream sales data. See World Finer Foods, 2000 WL 897752, at * 3. Based on this record evidence, and absent any evidence in the record that KTN requested data from the German Resellers, the court finds that Commerce’s decision to apply adverse facts available with respect to the German Resellers pursuant to 19 U.S.C. § 1677e(b) was supported by substantial evidence. 2. Commerce must explain its selection of adverse facts available. Plaintiffs argue that even if Commerce may properly draw an adverse inference, it still must adhere to certain standards in selecting the adverse facts. 1 The court agrees. Even under the former best information available statute, which provided Commerce wider latitude, Commerce had to select facts from the record that met the basic goals of the statute: “to encourage compliance while determining current margins as accurately as possible.” National Steel Corp. v. United States, 20 CIT 100, 103, 913 F. Supp. 593, 596 (1996) (citing Rhone Poulenc, Inc. v. United States, 899 F.2d 1185, 1191 (Fed. Cir. 1990)); see also National Steel Corp. v. United States, 18 CIT 1126, 1132-33, 870 F. Supp 1130, 1136-37 (1994) (discussing limits on Commerce’s use of best information available). 2 The adverse facts selected, therefore, must be non-aberrant, rationally related to sales, and indicative of KTN’s customary selling practices. National Steel, 20 CIT at 103, 913 F. Supp. at 596. In the current case, Commerce selected the “highest NV [normal value] reported by control number” for “KTN’s and NSC’s sales to its affiliates for which KTN did not report home market downstream sales,” Final Determination, 64 Fed. Reg. at 30,728, but it failed to explain why such number was non-aberrant, rationally related to sales, and indicative of customary selling practices. As Plaintiffs have argued, the result of Commerce’s choice of adverse facts available is to increase the aver *670 age price of KTN’s sales to the German Resellers by a substantial amount. 3 In arriving at the adverse inference average price, for 97 control numbers, Commerce used a price more than double the weighted average price, and of the 97, 21 were more than 250% of the weighted average price and 5 were more than 500% of the weighted average price. On remand, Commerce must explain why the numbers it selected were not aberrant and unduly punitive. In the explanation, Commerce must bear in mind the goal of the new facts available statute, which is to deter non-compliance while being “a reasonably accurate estimate of the respondent’s actual rate. * * *” F.LLI De Cecco di Filippo Fara S. Martino S.p.A. v. United States, _ F.3d _, 2000 WL 777170, at *5 (Fed. Cir. June 16, 2000). The court therefore remands to the Department of Commerce with instructions to explain its selection of adverse facts in light of this standard. B. USR 1. Commerce’s decision to reject USR’s entire database is not supported by substantial evidence. Plaintiffs argue that Commerce’s decision to reject USR’s entire database and apply total adverse facts available because of errors in the further manufacturing data field is not supported by substantial evidence on the record. Although Commerce disputes that the decision is so limited, the Final Determination could not be more clear. 4 Accordingly, the court declines to consider Commerce’s post hoc rationale as support for the Final Determination. See U.H.F.C. Co. v. United States, 916 F.2d 689, 700 (Fed. Cir. 1990). The main reason Commerce provided for rejecting USR’s entire database was that the errors in the further manufacturing response were pervasive and served to undermine the reliability of the entire response. See Final Determination, 64 Fed. Reg. at 30,739. Commerce, however, fails to support its assertion with any record evidence. Indeed, the calculations made by Plaintiffs based on record evidence suggest that a large portion of USR’s cost database, let alone the sales database, would not be affected by the errors in the further manufacturing field. See Plaintiffs Mot. for J. on an Agency Record at 8-9 (“Pis.’ Br.”). As respondents often must do, KTN worked with USR to develop a computer program to respond to Commerce’s questionnaire, particularly section E. The program was designed to match resold subject merchandise that underwent further processing with its source coü. See CR 97 at 2, Memorandum from Accountant to Director/IA (Mar. 18, 1999) (“USR Cost Verification Report”). 5 In both the Cost Verification Report and the Final Determination, Commerce identified the following errors: *671 1) in several cases the surcharge for processing small amounts of material was not reported or was incorrect; 2) in some cases the output weight exceeded the input weight; 3) some unprocessed materials were allocated costs; 4) bottom finished products were not allocated finishing costs; and 5) re-spinning processing was not captured. See USR Cost Verification Report at 2; see also Final Determination, 64 Fed. Reg. at 30,739. Based on the fact that a significant amount of KTN’s subject merchandise underwent further processing, Commerce decided that the entire cost database was unreliable. Plaintiffs argue that Commerce’s decision to reject USR’s entire database is unsupported by substantial evidence because the errors were limited to the further manufacturing adjustment. As Plaintiffs noted at the hearing held on this matter, the further manufacturing field represented one of over fifty fields in the cost information USR submitted. Additionally, Plaintiffs argue that because the further manufacturing adjustment has a minimal aggregate effect on USR’s reported sales value, use of partial facts available would be more appropriate. See Pis.’ Br. at 8. Based on the fact that seven of eighteen tested transactions contained errors, Commerce stated in the Final Determination that it disagreed with KTN that the errors affected only a small number of products. See Final Determination, 64 Fed. Reg. at 30,740. What Commerce, ignored, or failed to address, in the above statement was the actual assertion KTN raised, namely, whether or not the errors in the further manufacturing adjustment were pervasive or isolated. See, e.g., Motor Vehicles Mfrs. Ass’n v. State Farm Mutual Automobile Ins. Co., 463 U.S. 29, 43 (1983) (noting that a reviewing court should not make up for deficiencies in agency’s reasoning). Commerce’s reasoning would certainly be valid and supported by substantial evidence if it were used to support a finding that facts available should be used in lieu of the further manufacturing data USR submitted. But based on the record evidence, a conclusion that the remainder of USR’s cost data was tainted by the flawed computer programming is not warranted. See 19 U.S.C. § 1677m(e) (requiring Commerce to consider information not meeting all of the Department’s requirements under certain conditions); see also Borden, Inc. v. United States, 4 F. Supp.2d 1221, 1245-46, opinion after remand, 1998 WL 895890, aff'd sub nom. F.LLI De Cecco di Filippo Fara S. Martino S.p.A. v. United States, _ F.3d _, 2000 WL 777170 (Fed. Cir. June 16, 2000) (discussing five part test); Ferro Union, Inc. v. United States, 44 F. Supp. 2d 1310, 1330 (CIT 1999) (same); Mannesmannrohren-Werke AG, 77 F. Supp. 2d at 1313 (same). Because the conduct of an-tidumping investigations are largely within Commerce’s discretion, the court will provide Commerce an opportunity on remand to explain this issue further. As for the five errors Commerce specifically identified as reasons for rejecting USR’s entire database, see Final Determination, 64 Fed. Reg. at 30,739, the court finds that they would constitute substantial evi *672 dence on which to base a finding that facts available should be used instead of USR’s further manufacturing data. From the record it appears that Commerce was able to verify, successfully, a great many of the remaining fields in USR’s data. 6 On remand, however, it will be left to Commerce to explain whether the remaining fields, in fact, were verified or verifiable. See Ferro Union, Inc., 74 F. Supp. 2d at 1296 (noting Commerce’s statutory responsibility under 19 U.S.C. § 1677m(e) to use a party’s information even if incomplete). If Commerce can explain why the other items in USR’s cost database were unreliable or not verifiable, the court will reexamine whatever determination Commerce makes with respect to the cost database and any use of facts available. If, however, Commerce is unable to point to any record evidence demonstrating the unreliability or unverifiability of other data contained in the cost database, it shall use the remainder of USR’s cost data. See 19 U.S.C. § 1677m(e). Commerce conducted a separate verification of USR’s sales data. See CR 95, Memorandum from Analyst/IA to Program Manager/IA (Mar. 15, 1999) (“USR Sales Verification Report”). In rejecting USR’s entire response, however, Commerce also rejected USR’s sales data. The USR Sales Verification Report notes several errors in USR’s sales responses. Among these were failures to report early payment discounts, certain misreported commission payment dates, the inability to substantiate that some of the material shipped was non-prime merchandise, and the inability to provide documentation for an invoice sourced from KTN. See USR Sales Verification Report at 2. None of these errors appear to be attributable to the computer program. It appears that the computer program KTN designed to respond to Commerce’s questionnaire executed correctly for the sales data, although problems were encountered in the cost verification. 7 Commerce conducted a number of sales traces and noted some discrepancies, but nothing approaching the magnitude encountered in the cost verification. See USR Sales Verification Report at 11-14. Further, there is no indication in the record that the invoices and other materials Commerce used to trace the sales were generated by the computer program KTN designed to respond to the questionnaire. This leads to an important gap in Commerce’s reasoning. If the computer program was designed to respond to Commerce’s questionnaire for the investigation, it could not have been used to generate the actual sales documents Commerce was able to verify. Thus, on remand Commerce must cite record evidence that supports its conclusion that errors in the computer program would have affected KTN’s sales data. Commerce may find on remand that it is appropriate to apply partial facts available to fill any gaps in the sales *673 data it could not successfully verify, but it may not disregard the sales data absent evidence in the record that the sales data was fatally tainted by the errors in the computer program. See 19 U.S.C. § 1677m(e); see also Ferro Union, Inc., 74 F. Supp. 2d at 1297 (approving Commerce’s use of partial adverse facts where party failed to provide information in one respect). 2. Commerce’s decision to apply an adverse inference is not supported by substantial evidence. On remand, Commerce may resort to facts available provided it complies with the concerns annunciated above by the court. Since some facts available may be used, it is appropriate at this time to examine whether Commerce may draw an adverse inference in its selection of facts available. Based upon the reasoning provided by Commerce thus far, the court finds Commerce has not supported its decision to draw an adverse inference with substantial evidence. In the Final Determination, Commerce stated that the decision to draw an adverse inference was based on its finding that KTN failed to act to the best of its ability. That finding, in turn, was based on KTN’s “failure to conduct even rudimentary checks for the accuracy of the reported further-processing data.” See Final Determination, 64 Fed. Reg. at 30,739. It is instructive to compare Commerce’s decision to draw an adverse inference with respect to the German Resellers, discussed above, to Commerce’s decision to draw an adverse inference with respect to USR. In deciding that KTN failed to act to the best of its ability in providing the German Reseller’s data, Commerce looked at correspondence from KTN’s counsel, which repeatedly stated that KTN would provide legal and factual support for its position that the data was not owed. More importantly, KTN did not notify the case analyst at Commerce, per the questionnaires’ instruction, that it was having trouble securing the German Reseller’s data. Based upon this evidence combined with the other Thyssen affiliates’ compliance, it was permissible for Commerce to conclude that KTN failed to act to the best of its ability in obtaining the German Reseller’s data. In contrast, the sole evidence on which Commerce based its decision to draw an adverse inference for USR’s database was the possibility of KTN discovering and correcting errors in the computer program prior to verification. While the antidumping laws leave a great deal of discretion to the agencies, that discretion has certain limits. In particular, as this court has noted on two prior occasions, “[u]nder the URAA, Commerce is now required to make more subtle judgments than under the previous best information available (‘BIA’) standard.” Ferro Union, Inc., 44 F. Supp. 2d at 1329; Mannesmannrohren-Werke AG, 77 F. Supp. 2d at 1314. Indeed, the “antidumping statute may now be less administratively convenient, but Commerce must conform * * * to the new provisions.” Ferro Union, Inc., 44 F. Supp. 2d at 1329. *674 In this case, it is unclear whether Commerce’s decision to draw an adverse inference is even based on evidence in the record. Certainly the number of errors appearing in the further manufacturing field of the cost database are part of the record. Commerce’s conclusion, however, that rudimentary checks would have been able to identify the problem prior to verification is not based on evidence, but rather an unwarranted inference drawn from the evidence. Commerce’s position is that had KTN used the time in between its receipt of the cost verification agenda and the commencement of verification, it would have discovered the errors just as Commerce did. There is no indication in the record of whether this statement accurately describes KTN’s resources and its ability to run the checks prior to verification that would have discovered the errors. In making the finding required by 19 U.S.C. § 1677e(b) that a party has failed to cooperate by not acting to the best of its ability, Commerce must decide what the party was able to do. While the former decision is necessarily subjective, the latter is not. Commerce has not pointed to any evidence in the record to support its determination that KTN could have run the checks prior to verification. Commerce failed to address the critical issue, which was whether KTN, in fact, had the resources available to it prior to verification to discover the errors. Commerce argues that KTN approached the verifiers on the third and final day of verification claiming to have isolated and corrected the error in the computer program. Based on its ability to correct the problem in such a short time frame, Commerce concluded that a check before the verification commenced would also have enabled KTN to correct the problem. Even if it is true that such a correction would have been possible, errors in a party’s response, alone, are not evidence that the party failed to act to the best of its ability. See, e.g., Borden, Inc., 4 F. Supp. 2d 1221. Commerce has not pointed to a scintilla of evidence beyond the errors in the computer program indicating KTN did not cooperate to the best of its ability in compiling and submitting USR’s response. On remand, Commerce must point to record evidence demonstrating that KTN had the necessary resources to perform the checks prior to verification. If Commerce can point to such evidence, the court will reexamine the use of an adverse inference with respect to USR’s further manufacturing data. Commerce, however, will need to point to additional evidence beyond any failure to perform checks if it wants to draw an adverse inference with respect to any additional facts available that may be needed. If Commerce is unable to provide any additional evidence demonstrating KTN failed to act to the best of its ability, it may not draw an adverse inference in applying whatever additional facts available are needed. *675 3. Commerce’s decision to allocate all sales of unknown origin among three suppliers is not in accordance with law. Prior to verification, KTN informed Commerce that a large number of sales could not be traced to the original supplier. See Final Determination, 64 Fed. Reg. at 30,742. The reason for KTN’s inability to link the sales was due to USR’s reporting system, which removed the original supplier’s identity once the merchandise was transferred between different USR warehouses. See id. Although the original information could have been traced manually, the large number of sales involved would have placed a significant burden on KTN. See id. Commerce drew an adverse inference in applying facts available by allocating all sales of unknown origin among the three respondent firms in the investigations from Germany, Mexico and Italy. See id. at 30,743. In doing so, Commerce inflated the margin by including within the investigation sales that should not have been counted, as the product involved in at least some of these sales was sourced from countries not within the anti-dumping duty investigation. Plaintiffs argue that Commerce’s use of an adverse inference was contrary to law, because tracing all of the sales of unknown origin manually would have imposed an unfair burden on KTN. If the basis for Commerce’s decision had been as Plaintiffs state, the court would be deeply concerned about the propriety of drawing an adverse inference in this instance based on a failure to cooperate to the best of one’s ability. The Final Determination, however, discloses that Commerce did not base its decision on KTN’s ability to trace all of the unidentified sales manually. Rather, Commerce’s decision was based on the fact that at verification, “of seven ‘unidentified supplier’ transactions sampled at verification, [Commerce was] able to trace immediately the outside supplier for three * * * using nothing more than a personal computer in USR’s offices.” Id. (citing USR Sales Verification Report at 10). Based on this, Commerce concluded that nearly half of the unidentified suppliers could have been traced without any additional burden, and that the failure to do so demonstrated that KTN had not acted to the best of its ability. See id. Thus, Commerce’s decision to draw an adverse inference is based on substantial evidence. Nevertheless, while the court finds substantial evidence to support Commerce’s decision to draw an adverse inference, by law, Commerce is obligated to ensure that the facts chosen are not unduly harsh or punitive, bearing in mind that even in applying an adverse inference the goal is to determine margins as accurately as possible. See F.LLI De Cecco di Filippo Fara S. Martino S.p.A., _ F.3d _, 2000 WL 777170, at *5 (“[T]he purpose of section 1677e(b) is to provide respondents with an incentive to cooperate, not to impose punitive, aberrational, or uncorroborated margins.”); see also Ferro Union, Inc., 74 F. Supp. 2d at 1297 (“The facts available statute is designed to reach a reliable margin, even in the face of an uncooperative respondent.”). The court is unsatisfied at *676 this point that Commerce’s decision is not unduly harsh or punitive. The record indicates that Commerce verified the number of USR’s suppliers as well as the relative percentages of sales attributable to each. See Final Determination, 64 Fed. Reg. at 30,743. Despite this, Commerce used an allocation methodology that attributed to KTN more than double the amount of the verified percentage of KTN’s sales by volume to USR during the period of investigation. The court remands this issue to Commerce to explain how the result meets the purpose of the facts available statute. In that regard, it would be useful for Commerce to explain the resulting impact on the margin that its allocation methodology had compared to using an allocation methodology more in line with the verified relative percentages. 4. Commerce must explain its decision not to deduct the verified movement and selling expenses from gross unit prices. In calculating KTN’s margin, Commerce failed to deduct KTN’s verified movement and selling expenses. Commerce assigned the adverse facts available margin to the weighted average unit value for USR’s sales. See id. at 30,714. Commerce defends its decision by arguing that the calculated margin already had deductions for movement and selling expenses. See Def’s Br. at 29. KTN, in response, argues that Commerce’s defense supports KTN. KTN argues that by applying the adverse facts available margin to USR’s sales that were net of movement and selling expenses without using USR prices that were net of movement and selling expenses, Commerce arbitrarily inflated the dumping calculation. See Pis.’ Reply at 13. Depending upon the remand determination, it may not be necessary for Commerce to visit this issue. If, however, it is necessary, Commerce must explain how its methodology does not result in an arbitrary result. If Commerce determines that this was an error, Commerce may correct the error in the remand results. 5. USR’s Sales of Non-Subject Merchandise. In the Final Determination, Commerce rejected KTN’s argument that a quantity adjustment was appropriate because certain sales involving non-subject merchandise should be excluded. See Final Determination, 64 Fed. Reg. at 30743. Upon further review, Commerce concedes that its failure to exclude the non-subject merchandise was in error and requests a remand to recalculate the dumping margin to exclude sales of the non-subject merchandise. See Def.’s Br. at 29. The court agrees with the reasoning provided by Plaintiffs and Defendant for remanding this aspect of the Final Determination and will so order. IV Conclusion For the foregoing reasons, the Final Determination is remanded in part, and sustained in part. Accordingly, it is hereby Ordered that this matter is remanded to Commerce; and it is further Ordered that Commerce’s decision to apply an adverse inference with respect to the German Resellers is sustained; and it is further *677 Ordered that Commerce must explain why the adverse facts selected for the German Resellers were rationally related to KTN’s sales and indicative of its customary selling practices; and it is further Ordered that Commerce must explain why the adverse facts selected for the German Resellers were not unduly harsh or punitive; and it is further Ordered that Commerce must explain whether the fields, beside the further manufacturing fields, in USR’s cost database were verified or verifiable; and it is further Ordered that if Commerce finds the remaining fields in USR’s cost database were verified or verifiable it must use the actual numbers; and it is further Ordered that Commerce must explain why USR’s sales database was affected by the errors in USR’s cost database; and it is further Ordered that if Commerce is unable to point to substantial evidence demonstrating that the sales database was affected by errors in the cost database, it must use USR’s actual data, subject to filling any gaps, as noted in the court’s opinion, with facts available; and it is further Ordered that Commerce must point to substantial evidence in the record demonstrating that KTN had the ability to run checks to discover and correct the errors in the computer programming prior to verification; and it is further Ordered that if Commerce is unable to point to substantial evidence in the record of KTN’s ability to discover and correct the errors in the computer programming prior to verification, or other facts evincing a failure to cooperate, Commerce must not draw an adverse inference in choosing from the facts available for the further manufacturing data; and it is further Ordered that Commerce must point to additional evidence in the record, beside the errors in the computer program, before drawing adverse inferences for any remaining use of facts available; and it is further Ordered that Commerce’s decision to draw an adverse inference with respect to the allocation of sales of unidentified origin is sustained; and it is further Ordered that Commerce must explain why the allocation methodology for the sales of unidentified origin is not unduly harsh or punitive; and it is further Ordered that Commerce must explain its decision not to deduct verified movement and selling expenses from gross unit prices; and it is further Ordered that Commerce exclude the non-subject merchandise from its calculation of the dumping margin; and it is further Ordered that Commerce’s remand results are due to be filed within sixty days from the date of this decision. Any comments by the parties to the remand results are due on or before 20 days from the date of Com *678 merce’s filing and are limited to 20 pages, any rebuttal comments are due 20 days thereafter and are limited to 10 pages. Footnotes [1] The Defendant argues that KTN should be precluded from raising this argument because it was not raised at the administrative level. Plaintiffs argue that they did not have an opportunity to raise this argument administratively because the decision was made after the administrative briefing period had closed. The court finds that Plaintiffs are not precluded from raising this argument. Because Commerce chose to use adverse facts, Commerce must have considered the issue and Plaintiffs may challenge the decision. See Ferro Union, Inc. v. United States, 44 F. Supp. 2d 1310, 1333 n. 48 (CIT 1999) (citing Natural Resources Defense Council, Inc. v. EPA, 824 F.2d 1146, 1151 (D.C. Cir. 1987)) (exhaustion not required when agency has in fact considered an issue)). Additionally, in its case brief, KTN challenged the methodology Commerce used in the Preliminary Determination, arguing that the methodology produced aberrant results. See CR 98 at 24-27, Brief from Hogan & Hartson to Dep’t Commerce (Mar. 23,1999). [2] The Court of Appeals for the Federal Circuit recently held that Commerce’s discretion under the new statutory regime is similarly bound. In F.LII De Ceceo Di Filippo Fara S. Martino S.pA. v. United States, the court noted that the corroboration requirement evinced Congressional intent to prohibit Commerce from selecting unreasonably high rates._F.3d_, 2000 WL 777170, at *5 (Fed. Cir. June 16, 2000). [3] u%. [4] Commerce expressly noted that the issues surrounding the sales verification were moot and were not addressed in light of the decision to apply total adverse facts available. See Final Determination, 64 Fed. Reg. at 30,744. Although not expressed as such, the errors cited by Commerce in the Final Determination as grounds for rejecting USR’s entire database all stem from the further manufacturing field. [5] This information was not kept in the ordinary course of USR’s business operations. [6] Plaintiffs argue that Commerce verified Q% of the remaining data. See Pls.’ Reply Br. at 2. [7] It is not clear from the Final Determination whether the computer program was even used to generate any of the sales data Commerce verified. Although the USR Sales Verification Report notes that the program was used to link item files with invoice history files, there is no indication that the program did not execute successfully in this regard. See USR Sales Verification Report at 10-11. Further, there is no indication that the actual documents Commerce used to verify USR’s sales data would be unreliable due to programming errors.