Union Steel Manufacturing Co. v. United States

Slip Op. 14- 27 UNITED STATES COURT OF INTERNATIONAL TRADE UNION STEEL MANUFACTURING COMPANY, LTD., Plaintiff, and WHIRLPOOL CORPORATION, Before: Timothy C. Stanceu, Judge Plaintiff-intervenor, Consol. Court No. 10-00106 v. UNITED STATES, Defendant, and UNITED STATES STEEL CORPORATION, NUCOR CORPORATION, and HYUNDAI HYSCO, Defendant-intervenors. OPINION AND ORDER [Affirming in part, and remanding in part, a remand redetermination issued by the U.S. Department of Commerce in an action contesting the final results of an administrative review of an antidumping duty order on certain corrosion-resistant carbon steel flat products from the Republic of Korea] Dated: March 4, 2014 Brady W. Mills and Donald B. Cameron, Morris, Manning & Martin, LLP, of Washington, DC, argued for plaintiff Union Steel Manufacturing Company, Ltd. With them on the brief were Julie C. Mendoza, R. Will Planert, and Mary S. Hodgins, of Washington, DC. Consol. Court No. 10-00106 Page 2 Donald B. Cameron, Morris, Manning & Martin, LLP, of Washington, DC, for consolidated plaintiff Dongbu Steel Company, Ltd. With him on the brief were Brady W. Mills, Julie C. Mendoza, R. Will Planert, and Mary S. Hodgins, of Washington, DC. Donald Harrison, Andrea Fraser-Reid Farr, and John Christopher Wood, Gibson, Dunn & Crutcher, LLP, of Washington, DC, for plaintiff-intervenor Whirlpool Corporation. L. Misha Preheim, Attorney, Commercial Litigation Branch, Civil Division, U.S. Department of Justice, of Washington, DC, argued for defendant. With him on the brief were Stuart F. Delery, Principal Deputy Assistant Attorney General, Jeanne E. Davidson, Director, Patricia M. McCarthy, Assistant Director, of Washington, DC. Of counsel on the brief was Daniel J. Calhoun, Office of the Chief Counsel for Import Administration, U.S. Department of Commerce, of Washington, DC. Ellen J. Schneider, Skadden, Arps, Slate, Meagher & Flom LLP, of Washington, DC, argued for defendant-intervenor and consolidated plaintiff United States Steel Corporation. With her on the brief were Jeffrey D. Gerrish and Robert E. Lighthizer of Washington, DC. Timothy C. Brightbill, Wiley Rein LLP, of Washington, DC, argued for defendant-intervenor Nucor Corporation. With him on the brief was Alan H. Price, of Washington, DC. Jarrod M. Goldfeder, Akin Gump Strauss Hauer & Feld LLP, of Washington, DC, argued for consolidated plaintiff and defendant-intervenor Hyundai HYSCO. With him on the brief was J. David Park, of Washington, DC. Stanceu, Judge: In this consolidated action, four plaintiffs challenge the determination (“Final Results”) the International Trade Administration, U.S. Department of Commerce (“Commerce” or the “Department”) issued to conclude the fifteenth administrative review of an antidumping order on certain corrosion-resistant carbon steel flat products (“CORE” or “subject merchandise”) from the Republic of Korea (“Korea”).1 Certain Corrosion-Resistant Carbon Steel Flat Products from the Republic of Korea: Notice of Final Results of the Fifteenth Admin. Review, 75 Fed. Reg. 13,490 (Mar. 22, 2010) (“Final Results”). The fifteenth review pertained 1 Due to the presence of common issues, the court, on May 13, 2010, consolidated three actions under Consol. Court No. 10-00106. Order, ECF No. 46. Consolidated with Union Steel Mfg. Co., Ltd. v. United States (Consol. Court No. 10-00106) are Dongbu Steel v. United States, (Court No. 10-00109), Hyundai HYSCO v. United States (Court No. 10-00127), and United States Steel Corp. v. United States, (Court No. 10-00139). Consol. Court No. 10-00106 Page 3 to entries of subject merchandise made during the period of August 1, 2007 through July 31, 2008 (“period of review” or “POR”). Id. Before the court is a determination (the “Remand Redetermination”) Commerce issued in response to the court’s remand order in Union Steel Mfg. Co., Ltd. v. United States, 36 CIT __, 837 F. Supp. 2d 1307 (2012) (“Union Steel Mfg. Co., Ltd.”). Results of Redetermination Pursuant to Remand (Sept. 24, 2012), ECF No. 161 (“Remand Redetermination”). For the reasons discussed in this Opinion and Order, the court sustains the Remand Redetermination as to: (1) the general and administrative (“G&A”) expense ratio Commerce used to determine the cost of production (“COP”) of the foreign like product for Union Steel Manufacturing Co., Ltd. (“Union”); (2) the Department’s decision to calculate Union’s interest expense ratio using financial statements from both the 2007 and 2008 fiscal years; (3) the Department’s application of its modified “sales-below-cost” and “recovery-of-costs” tests to calculate Union’s normal value; (4) the Department’s decision to create separate product categories for laminated CORE and non-laminated, painted CORE for use in a model-match methodology comparing Union’s home market and U.S. sales; and (5) the Department’s use of the “zeroing” methodology to calculate Union’s weighted average dumping margin. The court orders a second remand for further proceedings to address a number of the Department’s decisions in the remand determination: (1) the decision to make a major input adjustment when calculating Union’s interest expense ratio; (2) the application of the modified “quarterly cost” methodology wherever used in the normal value calculations for Hyundai HYSCO (“HYSCO”), including the difference-in-merchandise (“DIFMER”) adjustments and constructed value (“CV”) determinations; (3) the application of the modified “quarterly cost” methodology for all aspects of the normal value calculations for Union except the revised Consol. Court No. 10-00106 Page 4 sales-below-cost and recovery-of-costs tests; (4) the decision to depart from the normal method for selecting a comparison month when determining antidumping margins for Union and HYSCO; and (5) the decision to depart from the normal method by selecting the date of shipment, rather than the date of invoice, as the date of sale for certain sales that HYSCO made through a U.S. affiliate, Hyundai HYSCO USA, Inc. (“HHU” or “HYSCO USA”). Finally, pursuant to the court’s order in Union Steel Mfg. Co., Ltd., 36 CIT at __, 837 F. Supp. 2d at 1337-38, the court concludes that the second remand redetermination must recalculate the margin for Dongbu based on the redetermined margins for Union and HYSCO. I. BACKGROUND The background of this case is provided in the court’s previous opinion and is supplemented herein. Union Steel Mfg. Co., Ltd., 36 CIT at __, 837 F. Supp. 2d at 1311-12. Three of the four plaintiffs in this action, Union, HYSCO, and Dongbu Steel Co., Ltd. (“Dongbu”), are Korean producers and exporters of the subject merchandise. Issues & Decision Mem., A-580-816, ARP 07-08, at 2 (Mar. 15, 2010), available at http://ia.ita.doc.gov/frn/summary/KOREA-SOUTH/2010-6258-1.pdf (last visited Feb. 26, 2014) (“Decision Mem.”). Plaintiffs Union and HYSCO were the mandatory respondents in the fifteenth administrative review and plaintiff Dongbu was a non-examined respondent in the fifteenth review. Id. United States Steel Corporation (“U.S. Steel”), a petitioner in the fifteenth administrative review, is the fourth plaintiff and a defendant-intervenor in this action. Id. at 1; Compl. 1 (Apr. 21, 2010), ECF No. 6 (Court No. 10-00139). Nucor Corporation (“Nucor”), also a petitioner in the fifteenth administrative review, is another defendant-intervenor in this action. Id. at 1 n.1. Consol. Court No. 10-00106 Page 5 Union challenged the Department’s: (1) exclusive reliance on the 2008 financial statement of Union’s parent company to calculate Union’s interest expense ratio; (2) application of the quarterly costs and indexing methodologies to various calculations when determining normal value; (3) selection of the comparison month for normal value sales; (4) similar treatment of laminated CORE and non-laminated, painted CORE in the model-match process; and (5) use of zeroing in calculating the weighted average dumping margin for Union. Union Steel Mfg. Co., Ltd., 36 CIT at __, 837 F. Supp. 2d at 1310, 1312, 1314, 1324. HYSCO challenged the Department’s use of the quarterly costs and indexing methodologies and the Department’s decision to use a nonstandard method for determining the comparison month. Id., 36 CIT at __, 837 F. Supp. 2d at 1310, 1312, 1324. Dongbu raised the same challenges as Union and HYSCO and sought to be subject to a rate reflecting any modifications to the weighted average dumping margins for Union and HYSCO. Id., 36 CIT at __, 837 F. Supp. 2d at 1310, 1335-36.2 U.S. Steel challenged the Department’s determination of the date of sale for the sales of subject merchandise that HYSCO made through a U.S. affiliate. Id., 36 CIT at __, 837 F. Supp. 2d at 1310, 1312, 1334. After the parties submitted all briefings in this action, defendant United States requested a voluntary remand so that Commerce could reconsider the quarterly cost and indexing methodologies used in the Final Results in light of an intervening decision by this Court. In the court’s previous opinion in this case, Union Steel Mfg. Co., Ltd., it concluded that Dongbu was entitled to a recalculated margin reflecting any changes made on remand to the 2 Plaintiff Dongbu Steel Co., Ltd. (“Dongbu”) also claimed entitlement to an individually-determined dumping margin; the court denied relief on this claim due to Dongbu’s failure to exhaust administrative remedies. Union Steel Mfg. Co., Ltd. v. United States, 36 CIT __, __, 837 F. Supp. 2d 1307, 1334 (2012) (“Union Steel Mfg. Co., Ltd.”). Consol. Court No. 10-00106 Page 6 margins for mandatory respondents HYSCO and Union. Id., 36 CIT at __, 837 F. Supp. 2d at 1329-34. The court also concluded that all of the claims brought by Union, HYSCO, and U.S. Steel required a remand. Id., 36 CIT at __, 837 F. Supp. 2d at 1337-38. Specifically, the court determined that it could not uphold the Department’s decision to use financial data from Union’s parent company pertaining only to the 2008 fiscal year when determining Union’s interest expense ratio. Id., 36 CIT at __, 837 F. Supp. 2d at 1313-21. The court also granted defendant’s request for a voluntary remand on the quarterly cost and indexing issues but directed Commerce to reconsider the quarterly cost and indexing methodologies wherever Commerce applied the methodologies in the Final Results—including the sales-below-cost test, “recovery-of-costs” test, constructed value (“CV”) determinations, and difference-in-merchandise (“DIFMER”) adjustments. Id., 36 CIT at __, 837 F. Supp. 2d at 1321-25. As defendant requested, the court also remanded the Department’s decisions to: (1) depart from the normal method for selecting the comparison month when comparing U.S. and home market sales of Union and HYSCO; (2) treat Union’s laminated CORE and non-laminated, painted CORE as “identical” merchandise in the Department’s model-match methodology; (3) use the “zeroing methodology” in determining weighted average margins; and (4) use shipment dates as the date of sale for HYSCO’s sales through a U.S. affiliate. Id., 36 CIT at __, 837 F. Supp. 2d at 1325-29, 1334-35. In its Remand Redetermination, filed on September 24, 2012, Commerce redetermined Union’s interest expense ratio using both the 2007 and 2008 financial statements of Union’s parent company, Dongkuk Steel Mills Company Ltd. (“DSM”). Remand Redetermination 7-8. Commerce also modified its application of the quarterly cost and indexing methodologies when determining various aspects of the normal value calculation for Union and HYSCO. Id. at 16-18. Commerce added a separate product category for non-laminated, painted CORE in its revised Consol. Court No. 10-00106 Page 7 model-match methodology for Union. Id. at 32-34. Commerce again found it appropriate to depart from the normal method for selecting the comparison months of normal value sales for Union and HYSCO. Id. at 27-29. Commerce also continued to use zeroing when calculating Union’s weighted average dumping margin. Id. at 60. Finally, Commerce continued to use the date of shipment as the date of sale for the sales HYSCO made through a U.S. affiliate. Id. at 61-62. In the Remand Redetermination, Commerce revised Union’s weighted average dumping margin from 14.01% to 9.85% and HYSCO’s weighted average dumping margin from 3.29% to 1.46%. Id. at 67. Dongbu’s weighted average dumping margin, which was based on the margins for two mandatory respondents, declined from 8.65% to 5.66%. Id. Union, HYSCO, Dongbu, U.S. Steel, Nucor, and defendant each filed comments on the Remand Redetermination.3 Union Steel’s Comments on the U.S. Dept. of Commerce’s Sept. 24, 2012 Results of Redetermination Pursuant to Remand (Nov. 30, 2012), ECF No. 169 (“Union’s Comments”); Resp. of Hyundai HYSCO to Def.’s Redetermination on Remand (Nov. 30, 2012), ECF No. 170 (“HYSCO’s Comments”); Dongbu Steel’s Comments on the U.S. Dept. of Commerce’s Sept. 24, 2012 Results of Redetermination Pursuant to Remand (Nov. 30, 2012), ECF No. 168 (“Dongbu’s Comments”); Comments of U.S. Steel Corp. on the Results of Redetermination Pursuant to Remand Issued by the Dept. of Commerce (Dec. 3, 2012), ECF No. 171 (“U.S. Steel’s Comments”); Nucor Corp.’s Comments on Remand Results (Dec. 3, 2012), ECF No. 176 (“Nucor’s Comments”). Defendant responded to the 3 Plaintiff-intervenor Whirlpool Corporation, a U.S. importer of subject merchandise, did not make any submissions in support of the motions for judgment on the agency record in this action and did not submit comments on the Remand Redetermination. Consol. Court No. 10-00106 Page 8 various comments on February 15, 2013. Def.’s Resp. to Comments on the Dept. of Commerce’s Remand Results, ECF No. 204 (“Def.’s Resp.”). The court held oral argument on May 23, 2013 at which Union, HYSCO, U.S. Steel, Nucor, and defendant United States appeared to address the various remaining challenges to the Remand Redetermination. Order (Mar. 3, 2013), ECF No. 207; Mot. for Oral Argument (Feb. 27, 2013), ECF No. 206; Oral Tr. 3-4 (July 22, 2013), ECF No. 215. II. DISCUSSION The court exercises jurisdiction under section 201 of the Customs Courts Act of 1980, 28 U.S.C. § 1581(c) (2006), pursuant to which the court reviews actions commenced under section 516A of the Tariff Act of 1930 (“Tariff Act”), 19 U.S.C. § 1516a, including an action contesting the final results of an administrative review that Commerce issues under section 751 of the Tariff Act, 19 U.S.C. § 1675(a).4 When reviewing the Department’s redetermination, the court “shall hold unlawful any determination, finding, or conclusion found . . . to be unsupported by substantial evidence on the record, or otherwise not in accordance with law. . . .” 19 U.S.C. § 1516a(b)(1)(B)(i). A. Union’s Interest Expense and G&A Expense Ratios When determining the cost of producing the foreign like product (“cost of production” or “COP”) in the Final Results, Commerce, according to 19 U.S.C. § 1677b(b)(3), calculated Union’s G&A expense ratio using Union’s unconsolidated financial statements for fiscal year 2008 and Union’s interest (“financing”) expense ratio using consolidated financial statements from Union’s parent for fiscal year 2008. Decision Mem. 42-43. Because both Union and its 4 All statutory citations herein are to the 2006 edition of the United States Code and all citations to regulations are to the 2010 edition of the Code of Federal Regulations. Consol. Court No. 10-00106 Page 9 parent company, DSM, use a fiscal year that matches the calendar year, these financial statements did not correspond precisely with the POR in this administrative review (August 1, 2007 to July 31, 2008). Id., 36 CIT at __, 837 F. Supp. 2d at 1314. In its Rule 56.2 brief, Union challenged the use of the 2008 financial statements on the grounds that these statements “‘were aberrational and do not reasonably reflect Union’s actual data pertaining to the production and sale of the subject merchandise during the POR.’” Id. (citation omitted). Union argued that Commerce should have calculated G&A and interest using only the 2007 statement or, in the alternative, a “blended rate,” for instance by combining the 2007 and 2008 financial data using a weighted average. Id., 36 CIT at __, 837 F. Supp. 2d at 1314-15 (citation omitted). In its previous opinion, the court ruled that using only the 2008 DSM financial statement to determine Union’s interest expense was unlawful for two reasons. “First, Commerce failed to consider an important aspect of the question before it, which was whether determining Union’s interest expense ratio solely on the basis of data in that financial statement produced the most accurate result.” Id., 36 CIT at __, 837 F. Supp. 2d at 1317. The court noted that “the interest cost ratio derived from the 2008 financial statement of DSM reflects a five-fold increase from the interest cost ratio derived from the 2007 statement and appears to have been affected significantly by currency-related losses that coincided with a massive post-POR decline in the value of the Korean won.” Id., 36 CIT at __, 837 F. Supp. 2d at 1319-20. Second, the court concluded that the Department’s use of the 2008 financial statement was based on the incorrect premise “that the Department has a consistent practice of using the single financial statement corresponding to the largest portion” of the POR. Id., 36 CIT at __, 837 F. Supp. 2d at 1319. The court directed Commerce to “reconsider its decision to base Union’s interest expense ratio entirely on data obtained from DSM’s 2008 financial statement” and to examine “the relative Consol. Court No. 10-00106 Page 10 merits of alternative methods.” Id., 36 CIT at __, 837 F. Supp. 2d at 1320. Although the court did not identify the same legal infirmities with respect to the G&A expense ratio, the court allowed Commerce to reconsider the G&A expense ratio on remand and deferred any ruling on whether or not that ratio complies with law. Id. 1. The Court Sustains the Redetermined G&A Expense Ratio for Union In the Remand Redetermination, Commerce also examined its G&A expense ratio for Union but continued relying on Union’s 2008 financial statement. Commerce stated that “[w]hile the change in Korean won to U.S. dollar exchange rates from the beginning of 2007 to the end of 2008 may be considered aberrational, which directly affected the interest expense rate calculation, there is nothing unusual with the G&A expenses for 2007 and 2008 that would lead us to the same conclusion.” Remand Redetermination at 8-9. The Department’s decision on its G&A expense complies with the courts’ remand order and is not contested by any party in this action. The court therefore sustains the G&A expense ratio that Commerce reached in the Remand Redetermination. 2. Commerce Permissibly Based Union’s Interest Expense Ratio on the 2007 and 2008 Financial Statements of DSM In the Remand Redetermination, Commerce reconsidered the interest expense ratio for Union and calculated a new interest expense ratio using both the 2007 and 2008 DSM financial statements, weighted according to the number of months of the POR occurring within each fiscal year.5 Remand Redetermination 7-8. Commerce explained that, notwithstanding the 5 Because the period of review (“POR”) was August 1, 2007 through July 31, 2008, Commerce based the weighted average rate on five months in 2007 and seven months in 2008. Results of Redetermination Pursuant to Remand 8 (Sept. 24, 2012), ECF No. 161 (“Remand Redetermination”). Commerce articulated its new methodology as follows: (continued . . .) Consol. Court No. 10-00106 Page 11 Department’s longstanding practice of using data from the fiscal year most closely corresponding to the POR, “inclusion of both fiscal years’ financial statements helps to ensure, based upon the unique facts present in this review as identified by the [Court of International Trade] in its Remand Order, that financial expenses for both fiscal years that encompass the POR are reflected in the financial expense ratio.”6 Id. at 8. The use of this blended calculation of 2007 and 2008 data in the Remand Redetermination significantly reduced Union’s interest expense ratio.7 Union Steel Accounting Cost Mem. 2 (Aug. 23, 2012) (Confidential Remand.R.Doc. No. 10). Union opposes the revised interest expense ratio, arguing that that the data from the 2007 DSM financial statement are superior to a blended calculation using both the 2007 and 2008 statements because the latter approach still includes aberrationally high interest rate expenses (continued . . .) The blended financial expense rate is the result of a two-step calculation. We first calculated separate financial expense rates for 2007 and 2008. For each of these calculations, we used annual net financial expenses in the numerator and annual [cost of goods sold] in the denominator. We relied on the 2007 rate calculation that was used in the preliminary results, and the 2008 rate calculation that was used for the final determination. Remand Redetermination 14-15. 6 Upon reconsidering the interest expense ratio on remand as directed by the court, Commerce chose the blended approach over maintaining its exclusive reliance on the 2008 statement of Dongkuk Steel Mills Company Ltd. (“DSM”). The Remand Redetermination discloses that Commerce made the change voluntarily instead of making the change “under protest,” i.e., solely in response to the court’s having ruled against its previous method. Remand Redetermination 14 (“While the Court’s Remand Order may have left open the option to use only DSM’s 2008 financial statements, we disagree that it is a feasible option considering the significant devaluation of the Korean won during the year and the fact that the 2008 financial expense rate was significantly higher than that in the previous year.”). 7 The actual interest expense and general and administrative expense ratios are not disclosed in this Opinion and Order due to a claim for proprietary treatment. Consol. Court No. 10-00106 Page 12 that occurred predominantly after the POR.8 Union’s Comments 1-4. Defendant counters that the Department’s new method accounts for the whole of the POR while mitigating the potential effect of DSM’s post-POR foreign exchange losses. Def.’s Resp. 6-7, 11-12. Defendant also points out that the blended rate methodology is consistent with one favored by Union in its Rule 56.2 brief. Id. at 7. In reviewing the revised financial ratio, the court accords Commerce considerable deference in the choice of methodology. See Corus Staal BV v. Dep’t of Commerce, 395 F.3d 1343, 1346 (Fed. Cir. 2005) (“Corus Staal”). The Department’s revised approach recognized the need to address the unusual circumstance of the aberrational interest expenses for DSM, which Commerce reasonably related to post-POR decline in the value of the Korean won. By drawing on expense data from both the 2007 and 2008 fiscal years, Commerce accounted for all twelve months of the POR while limiting any distorting effect of the post-POR foreign exchange losses. This approach necessarily includes in the calculation the finance cost data for twelve months outside of the POR, but no option would allow perfect coverage of the POR on the record facts. No DSM financial statement on the record corresponds temporally with the POR, and neither the 2007 nor the 2008 DSM statement reports monthly, or even quarterly, interest expenses.9 Both statements report data as of the end of DSM’s fiscal year. The court concludes that Commerce reached a decision of the choice of DSM financial statements that 8 At oral argument, Union’s attorney stated, consistent with the exhibit Union prepared for its rebuttal brief before Commerce, that the Korean won fell 7% in the seven months of 2008 that include the POR and 26% in the five months afterward. Oral Tr. 131; Union Steel Rebuttal Br. to U.S. Steel’s Post-Prelim. Case Br., Ex. 2 (Jan. 27, 2010) (Admin.R.Doc. No. 5219). 9 See Remand Redetermination 16 (“While [United States Steel Corporation] claims that DSM’s financial statements show that the losses were generated, there is no evidence that identifies either the timing or the magnitude of these losses throughout the fiscal year.”). Consol. Court No. 10-00106 Page 13 must be sustained. The Department’s approach to this issue addresses the problem of the aberrational interest expenses but also endeavors to cover the entire POR using a combination of the two DSM financial statements. According to Union, only by excluding consideration of the 2008 financial statement entirely may Commerce eliminate the distorting effect of aberrational finance costs occurring after the close of the POR. Union’s Comments 3-4. Union also argues, as it did when originally contesting the Final Results, that the 2007 financial statement is more appropriate than the 2008 statement. Id. at 4-5. Union notes that the majority of the home market reporting period, when calculated according to the so-called “window period” used for average-to-transaction sales comparisons, which includes the POR plus ninety days before and sixty days after the POR, falls in 2007. Id. The court is not persuaded by Union’s arguments. Use of the 2007 statement, although eliminating any distortion caused by the fall of the Korean won in the latter months of 2008, would create other distortions by covering only five months of the POR. Union’s argument based on the window period—any relevance of which is limited to instances in which there are no home market sales in the corresponding month—does not satisfy the basic objection that the 2007 financial statement covers only the first five months of the POR and thereby fails to cover the seven months of the POR that fell within the 2008 calendar year.10 Commerce acted within its discretion in using both statements to determine Union’s interest expense ratio. Balancing the competing considerations, Commerce thereby addressed the problem of aberrational expenses and also maintained reasonable contemporaneity with the POR. 10 The months outside the POR affect the margin calculation only in instances in which there are no sales of the foreign like product during the same month of the POR in which a U.S. sale of subject merchandise occurred. See 19 C.F.R. § 351.414(e)(2). Consol. Court No. 10-00106 Page 14 U.S. Steel and Nucor also oppose the use of both the 2007 and 2008 statements to determine the interest expense ratio. Both defendant-intervenors argue that the Department’s revised method departs, without providing a reasoned basis, from the Department’s longstanding practice of using the financial statement that corresponds most substantially with the POR. U.S. Steel’s Comments 17-26; Nucor Comments 8-12. However, as the court noted in Union Steel Mfg. Co., Ltd., Commerce does not appear to have a consistent practice of using financial statements from the fiscal year most closely corresponding to the POR when calculating interest expense ratios. Union Steel Mfg. Co., Ltd., 36 CIT at __, 837 F. Supp. 2d at 1319. Moreover, the Remand Redetermination provides a reasoned basis for the Department’s choice to rely on the blended approach rather than the 2008 statement, citing the unusual circumstances occurring in 2008, which included a steep decline in the value of the Korean won after the close of the POR. U.S. Steel contends that Commerce did not comply with the specific instructions set forth in the court’s Opinion and Order, having failed to “address whether using the blended rate would produce a dumping margin that was more accurate than using a rate based only on the 2008 financial statements.” U.S. Steel’s Comments 20-21. This argument is not persuasive because the Remand Redetermination explains why Commerce considered a blended rate more accurate than one based solely upon the 2008 financial statement. Commerce explained that “including both fiscal years’ financial statements helps to ensure, based upon the unique facts present in this review as identified by the [Court of International Trade] in its Remand Order, that financial expenses for both fiscal years that encompass the POR are reflected in the financial expense ratio.” Remand Redetermination 15. Commerce added that “[t]he Department’s methodology of using a blended financial expense rate for these final remand results recognizes the unique Consol. Court No. 10-00106 Page 15 circumstances of the significant loss in the value of the Korean won that DSM experienced during 2008.” Id. at 15-16. U.S. Steel submits that the decline in the Korean won occurring in 2008 after the close of the POR was not the “key factor” responsible for DSM’s 2008 foreign exchange transaction and translation losses. U.S. Steel’s Comments 21. U.S. Steel argues that “[t]o the contrary, the foreign exchange transaction and translation losses were due primarily to DSM’s substantially expanded commercial activities in 2008.” Id. In support of this argument, U.S. Steel points to record evidence that “[d]uring fiscal year 2008, DSM’s sales increased by fully 50% compared to 2007—a period of extraordinary growth—and its costs to produce such sales,” i.e., the cost of goods sold (“COGS”), “increased by over 40%.” Id. In further support of its argument, U.S. Steel states that “the notes to the 2008 financial statements show that DSM’s increased translation losses were based on the company’s enormous expansion of foreign borrowing in 2008,” adding that “in 2008, DSM’s notes payable in dollars increased by over 100%, and its notes payable in Japanese yen increased by over 2,578%.” Id. at 22-23 (emphasis in original). Even though the 2008 financial statement at issue is not Union’s, but DSM’s, U.S. Steel also identifies aspects of Union’s business activity during the POR in support of an argument that “it is certainly proper to attribute the foreign exchange transaction losses in the 2008 financial statements to interest expense and business activities that took place during the POR.” Id. at 22. The court understands U.S. Steel’s argument to be, in essence, that Commerce should use only the 2008 DSM financial statement because DSM’s increased business activity in 2008, and not the decline in the Korean won following the close of the POR, was the primary cause of the increased foreign exchange translation and transaction losses reflected in that financial statement. This argument does not convince the court that Commerce erred in using a rate derived from Consol. Court No. 10-00106 Page 16 DSM’s 2007 and 2008 statements. In making its argument, U.S. Steel itself acknowledges “that the foreign-currency related losses account for the vast majority of DSM’s 2008 net financial expenses.” Id. at 22 n.13. U.S. Steel also acknowledges that the increased business activity was related to the foreign exchange-related transaction losses, stating that “[p]lainly, this tremendous burst in business activity would have involved an enormous increase in purchases and sales in other currencies” and that “[a]s a result, it is a key factor in explaining the increases in the foreign exchange transaction losses in the 2008 financial statements.” Id. at 21. That the increased business activity may have made the currency transaction and translation losses larger than these losses otherwise would have been does not compel a conclusion that the fall in the value of the Korean won, particularly the fall occurring after the close of the POR, was not a key factor in DSM’s increased currency transaction and translation losses in the 2008 fiscal year. On this record, Commerce permissibly could infer an effect of post-POR declines in the Korean won because of the unusual level of those declines and the huge increase in the interest expenses DSM incurred from fiscal year 2007 to fiscal year 2008. U.S. Steel next argues that “[t]here is also no evidence that the translation and transaction losses in question were ‘extraordinary’ and, as a result, rendered the 2008 financial statements unrepresentative of Union’s interest expense during the POR,” adding that “[i]t is highly significant that DSM itself treated these gains and losses as ordinary income and expenses in its financial statements.” U.S. Steel’s Comments 23. This argument fails because there is substantial record evidence to support a finding that the currency-related losses incurred by DSM in fiscal year 2008 were a one-time event justifying the use of both of the DSM financial statements. Under the Department’s approach, events transpiring during both fiscal years were Consol. Court No. 10-00106 Page 17 factored into the financial ratio. U.S. Steel does not make the case that Commerce was required on this record to use only the 2008 statement to achieve the most accurate margin possible. 3. On Remand, Commerce Must Reconsider its Decision to Make a Major Input Adjustment to Union’s Interest Expense Ratio and Address Nucor’s Objection to that Decision Nucor raises other objections to the Department’s method for redetermining Union’s interest expense ratio. Nucor argues that Commerce should not have made a “major input adjustment” when calculating this ratio, on the premise that the interest expense ratio should reflect actual, not hypothetical, costs. Nucor’s Comments 15. “While Nucor agrees with the Department’s adjustment to direct materials in order to account for what Union should have theoretically paid to its affiliated party suppliers, Nucor disagrees with the second adjustment made to Union’s COGS in calculating the financial ratio.” Id. According to Nucor, the adjustment resulted in a COP for Union that fails to comply with 19 U.S.C. § 1677b(f)(3) because it “does not fully reflect an arm’s length transaction.” Id. The Tariff Act allows Commerce to make a major input adjustment to the COP calculation when a major input (in this case, coil steel substrate used to produce CORE that Union obtained from a related party) is obtained from a related party and Commerce has reasonable grounds to believe or suspect that the amount represented as the value of the input is less than the cost of producing the input. 19 U.S.C. § 1677b(f)(3). In determining Union’s COP, Commerce made such an adjustment to Union’s cost of coil steel substrate by increasing the COGS that it used as the denominator for the ratio pertaining to the direct materials input. Remand Redetermination 14-15. Commerce also made the adjustment to the COGS denominator when calculating Union’s finance cost ratio. Id. When responding to the draft remand results Commerce circulated for comment before issuing the Remand Redetermination, Nucor objected to the Department’s making a major input Consol. Court No. 10-00106 Page 18 adjustment when calculating Union’s interest expense ratio, on essentially the same grounds asserted here. Nucor’s Comments 14-15 (citing Comments on Draft Remand Results 8-10 (Sept. 6, 2012) (Remand.R.Doc. No. 13). The Remand Redetermination does not respond to Nucor’s objection. Commerce has an obligation to address important factors raised by comments from petitioners and respondents. SKF USA Inc. v. United States, 630 F.3d 1365, 1374 (Fed. Cir. 2011) (citation omitted). Because Commerce did not meet this obligation, the court will order Commerce to consider and respond to Nucor’s comment in the second remand redetermination. In responding to Nucor’s comment that Commerce should not have made the major input adjustment in determining Union’s finance cost ratio, defendant states that “[t]o ensure greater accuracy, the adjustments made by Commerce in its remand calculation of Union’s interest expense ratio eliminated th[e] very distortion recognized by the statute.” Def.’s Resp. 10. The court must review the Department’s decision on the rationale Commerce put forth (absent in this case) rather than defendant’s post-hoc rationalization. SEC v. Chenery Corp., 332 U.S. 194, 196 (1947) (“[A] reviewing court, in dealing with a determination or judgment which an administrative agency alone is authorized to make, must judge the propriety of such action solely by the grounds invoked by the agency.”). Nucor argues in the alternative that even if it is permissible for Commerce to make a major input adjustment to the interest expense ratio calculation, Commerce still erred by making the adjustment using an impermissible method. According to Nucor, Commerce should not have made the adjustment to the COGS for each of the two fiscal years and instead should have made the adjustment only once, before weighting the average of the resulting ratios, arguing that the Department’s method double counted the major input adjustment. Nucor’s Comments 13-14; Consol. Court No. 10-00106 Page 19 Oral Tr. 164-65. Because the question of how any major input adjustment, if permissible, should be effectuated is dependent on the answer to the primary question Nucor raises, the court defers any consideration of the “double counting” issue pending consideration of the Department’s answer to that primary question. B. Use of Quarterly Cost Averaging Periods Commerce determines the normal value of subject merchandise when conducting an administrative review of an antidumping duty order and compares that normal value to the export price or constructed export price of the subject merchandise. 19 U.S.C. § 1675(a)(2). In certain circumstances, Commerce, when calculating normal value, may exclude home market sales of the foreign like product that are less than the cost of producing the foreign like product (“cost of production” or “COP”).11 Id. § 1677b(b)(1). Commerce collects data from each respondent to determine the COP on a product-specific “control number” (or “CONNUM”) basis. Decision Mem. 14-15. Commerce then uses this CONNUM-specific cost information in various aspects of its dumping margin calculations. Id. at 14. Normally Commerce will calculate cost as a weighted average over the entire POR, but in some circumstances Commerce determines it is preferable to calculate costs as a weighted average for a shorter period of time, such as on a quarterly basis, as it did in the fifteenth review. Id. at 15, 21. 11 The antidumping statute provides that Commerce may disregard home market sales when calculating normal value if certain conditions are met: If the administering authority determines that sales made at less than the cost of production— (A) have been made within an extended period of time in substantial quantities, and (B) were not at prices which permit recovery of all costs within a reasonable period of time, such sales may be disregarded in the determination of normal value. 19 U.S.C. § 1677b(b)(1). Consol. Court No. 10-00106 Page 20 In this action, the circumstance influencing the Department’s decision to use quarterly costs was the fluctuating cost for a major input, steel substrate. Id. at 22-23. Commerce directed respondents to submit cost data on a quarterly basis; Commerce then used the quarterly data, along with its “indexing” methodology,12 to calculate the quarterly weighted average cost to produce each foreign like product for each examined respondent. Id. at 21, 24. Union and HYSCO both challenged the Department’s application of indexed quarterly cost data to aspects of their respective normal value calculations. Union Steel Mfg. Co., Ltd., 36 CIT at __, 837 F. Supp. 2d at 1312. Defendant requested a voluntary remand to allow Commerce to reconsider the quarterly cost methodology as applied to the cost recovery test. Def.’s Mot. for Partial Voluntary Remand 1 (June 21, 2011), ECF No. 130. Commerce indicated that it would review its methodology in light of an intervening decision of this Court. Id. (citing SeAH Steel Corp. v. United States, 34 CIT __, __, 704 F. Supp. 2d 1353, 1364-70 (2010) (“SeAH”) (holding that the Department’s application of quarterly cost methodology for recovery of costs purposes did not comply with section 773(b)(2)(D) of the Tariff Act)).13 The court 12 Commerce described its indexing methodology as follows: [T]he Department indexed the quarterly material costs to a common period cost level, thereby neutralizing the effect of the significant cost changes for the input between quarters. Then . . . the Department calculated a period of review weighted-average per unit cost. Finally, the weighted average per unit cost for the period of review for the substrate input was indexed back to the appropriate quarter to keep the weighted-average per unit costs consistent with the main input’s significantly changing price levels occurring between quarters. Issues & Decision Mem., A-580-816, ARP 07-08, at 23 (Mar. 15, 2010), available at http://ia.ita.doc.gov/frn/summary/KOREA-SOUTH/2010-6258-1.pdf (last visited Feb. 26, 2014). 13 The Act provides that “[i]f prices which are below the per unit cost of production at the time of sale are above the weighted average per unit cost of production for the period of investigation or review, such prices shall be considered to provide for recovery of costs within a reasonable period of time.” 19 U.S.C. § 1677b(b)(2)(D). Consol. Court No. 10-00106 Page 21 granted the request for voluntary remand but also instructed Commerce to reconsider its quarterly cost methodology generally (including the use of indexing), wherever it was used in the Final Results. Union Steel Mfg. Co., Ltd., 36 CIT at __, 837 F. Supp. 2d at 1321-25. 1. The Department’s Revised Sales-Below-Cost and Cost Recovery Tests On remand, Commerce continued to use quarterly weighted average COPs to identify Union’s and HYSCO’s home market sales that were made below cost. Remand Redetermination 17. However, Commerce removed indexing from its quarterly cost methodology in most instances, instead relying on “Union’s and HYSCO’s reported historic quarterly-cost data.” Id. For the sales-below-cost test, Commerce “compared the un-indexed historical quarterly COPs to sales prices in each respective quarter to determine whether they were below cost.” Id. Seeking conformity with the decision in SeAH, Commerce also amended its cost recovery test, stating in the Remand Redetermination that it “calculated an un-indexed weighted average per-unit COP for the entire period of review using the historical quarterly costs and production quantities as reported by respondents.” Id. at 17. In doing so, Commerce “first isolated Union’s and HYSCO’s sales” that were “disregarded because they failed the below-cost test.” Id. Commerce then “calculated CONNUM-specific annual weighted average prices” for each of these disregarded sales “and compared them, on a CONNUM-specific basis, to the annual weighted-average un-indexed costs.” Id. at 17-18. If, on comparison, the annual weighted average price for a CONNUM exceeded the annual weighted average unindexed cost for that CONNUM, Commerce “restored all sales of that CONNUM to the normal value pool of sales available for comparison with U.S. sales.” Id. As a result, Commerce included a greater number Consol. Court No. 10-00106 Page 22 of sales in both Union’s and HYSCO’s dumping margin calculations in the Remand Redetermination than it had in the Final Results. Id. at 18. i. The Court Sustains the Application of the Department’s Revised Sales-Below-Cost and Recovery-of-Costs Tests as Applied to Union Union does not object to the Department’s revised sales-below-cost and cost recovery tests. Union’s Comments 5-6; Oral Tr. 6-10. Accordingly, the court sustains this aspect of the Remand Redetermination as it pertains to Union. ii. Commerce Must Reconsider its Revised Recovery-of-Costs Test as Applied to HYSCO To calculate the annual weighted average cost under its revised methodology, Commerce combined cost averages for the different quarters of the POR. Where a home market sale occurred in a quarter for which the record contained no cost data for the CONNUM that was sold, Commerce, in its own words, “fill[ed] gaps with a surrogate or indexed cost.”14 Remand Redetermination 20. This occurred where a specific CORE product (identified by CONNUM) was sold but not produced in a particular quarter. In such an instance, Commerce substituted cost data pertaining to the actual CONNUM with cost data for a similar CONNUM. Remand Redetermination at 17 n.3. In its justification for this practice, Commerce explained that “[b]ecause it is necessary to calculate a POR-average cost for the cost recovery test, it is necessary to include a cost for every quarter in which a sale of a CONNUM occurred.” Id. at 20. HYSCO contests the method Commerce used in the Remand Redetermination to calculate annual weighted average costs under the revised recovery-of-costs test. HYSCO’s Comments 2-5. Specifically, HYSCO objects to the Department’s using as a surrogate the cost 14 Also, where HYSCO reported a sale made before the POR for which the merchandise was entered during the POR, Commerce “calculated the [cost of production] by indexing the first quarter costs back to derive the pre-POR quarterly costs,” as there was no quarterly [cost of production] data on the record for these pre-POR sales. Remand Redetermination at 17 n.2. Consol. Court No. 10-00106 Page 23 data for a similar CONNUM where cost data for the actual CONNUM sold were present on the record, albeit not cost data for the particular quarter in which the sale occurred. Id. at 4-5. HYSCO argues that “[t]he Department’s method makes no sense when actual costs for those CONNUMs are on the record.”15 Id. Defendant counters that the court should uphold the Department’s method, maintaining that “[h]aving found significant cost changes between quarters, Commerce determined it to be necessary to include a cost—even if a surrogate cost from the most similar CONNUM—for every quarter in which a sale of a CONNUM occurred to obtain the most representative POR-average cost in applying the recovery-of-cost[s] test.” Def.’s Resp. 14. The relevant statutory provision imposes a straightforward rule: a sale of the foreign like product in the home market may not be excluded from the normal value calculation as a sale below cost if the price paid is “above the weighted average per unit cost of production for the period of . . . review.” 19 U.S.C. § 1677b(b)(2)(D). Under this rule, Commerce must calculate the weighted average per unit cost of producing the good sold in the home market, and it must do so on a POR-wide (in this case, yearly) basis. The statute makes no exception for a situation in which the COP, or an element of that COP (in this case, the cost of the surrogate), fluctuates significantly during the POR. Nor does the statute provide an exception for a situation in which the good is sold but not produced in the home market during a period within the POR that is less than one year. Commerce nevertheless determined that where both of these situations existed, it was appropriate to determine “the weighted average per unit cost of production for the period 15 HYSCO also argues that Commerce acted inconsistently in addressing the problem of missing CONNUM-specific quarterly data and the problem caused where there was an absence of pre-POR cost data. Resp. of Hyundai HYSCO to Def.’s Redetermination on Remand 3-4 (Nov. 30, 2012), ECF No. 170 (“HYSCO’s Comments”). The court does not address this argument because it orders a remand for other reasons, as discussed herein. Consol. Court No. 10-00106 Page 24 of . . . review,” Id. § 1677b(b)(2)(D), using data other than data on how much it cost to produce the actual good sold in the home market. Remand Redetermination 17 n.3. Commerce is empowered to use “facts otherwise available” when data necessary for a required determination are missing from the record. 19 U.S.C. § 1677e(a). But in this case, the record contained the data required to perform the calculation required by § 1677b(b)(2)(D): cost of production data that pertained to both the CONNUM sold in the home market and the period specified in the statute, i.e., the one-year period of review. Remand Redetermination at 17 & n.3. The Department’s reasoning that “[b]ecause it is necessary to calculate a POR-average cost for the cost recovery test, it is necessary to include a cost for every quarter in which a sale of a CONNUM occurred,” id. at 20, is not consistent with the language or purpose of the statute. Because the Department’s method does not comply with the rule established by § 1677b(b)(2)(D), the court must order Commerce to perform its recovery-of-costs test for HYSCO in a way that complies with the statutory directive. 2. On Remand, Commerce Must Reconsider its Application of Quarterly Cost Methodology to Union’s and HYSCO’s DIFMER Adjustments and Constructed Value Determinations With the exception of the Department’s revised recovery-of-costs test, the Remand Redetermination retained the use of the quarterly cost methodology in all instances where it was used in the Final Results, including difference-in-merchandise (“DIFMER”) adjustments when comparing similar merchandise and constructed value (“CV”) determinations. In their comments on the Remand Redetermination, neither HYSCO nor Union objected specifically to the Department’s use of unindexed quarterly cost data for DIFMER adjustments or CV. However, both maintained general objections to the Department’s deviation from the normal method of using POR-wide cost averages. See HYSCO’s Comments 2-8; Union’s Comments 5-6; Oral Tr. 9, 17-18, 25-30; see also Union Steel Mfg. Co., Ltd., 36 CIT at __, Consol. Court No. 10-00106 Page 25 837 F. Supp. 2d at 1324. As it did in the Final Results, the Department concluded that a 25% or greater fluctuation in the cost of steel substrate together with a “reasonable correlation” with price of the good justified the Department’s use of the quarterly cost methodology in the Remand Redetermination, with the exception of the recovery-of-costs test, as discussed supra. Remand Redetermination 24-26. Both plaintiffs argue that, for various reasons, the circumstances of this administrative review do not warrant the application of shorter cost averaging periods. Union’s Comments 5-14; HYSCO’s Comments 5-8. For example, both plaintiffs argue that there was not a sufficient nexus between steel substrate costs and CORE prices to justify the use of shorter averaging periods and that the Department’s previous practice required more than a reasonable correlation. Union’s Comments 10-14; HYSCO’s Comments 6-8. Union argues that the 25% threshold for substrate cost fluctuation was met in its case only because Commerce performed a major input adjustment and instead should have looked at actual, not adjusted, costs. Union’s Comments 7-9. The court does not address all of the objections raised by Union and HYSCO because it finds insufficient the Department’s explanation for the continued use of quarterly cost averaging periods for DIFMER and CV. That explanation is that it is “the Department’s normal practice” to calculate COP, CV, and the DIFMER adjustment “in the same manner.” Remand Redetermination 20-21. Standing behind its “normal practice” and providing no further explanation, Commerce does not address the question of why using unindexed quarterly cost data in the specific context of DIFMER and CV ensures a more accurate dumping margin. Also related to the question of accuracy is the fact that the Remand Redetermination does not inform the court whether using quarterly costs influenced the Department’s decision to use surrogate costs for CV and DIFMER in the manner similar to that the court found objectionable above, i.e., Consol. Court No. 10-00106 Page 26 in the application of the recovery-of-costs test to HYSCO when quarterly CONNUM-specific data were not available. The court directs Commerce to address these matters on remand. 3. Commerce Must Reconsider Its Departure from the Normal Method for the Identification of the Contemporaneous Month The Department’s decision to proceed with a quarterly cost methodology influenced another decision in the Remand Redetermination, which involved the manner of making price-to-price comparisons between home market and U.S. sales. When comparing export prices to home market sales, Commerce is limited in its averaging of home market prices “to a period not exceeding the calendar month that corresponds most closely to the calendar month of the individual export sale.”16 19 U.S.C. § 1677f-1(d)(2). In the Final Results, Commerce followed its standard practice in administrative reviews and calculated dumping margins using an average- to-transaction comparison method according to 19 C.F.R. § 351.414(c)(2). Union Steel Mfg. Co., Ltd., 36 CIT at __, 837 F. Supp. 2d at 1325. Under this method, Commerce compares the price of a weighted average of sales of the foreign like product in the home market to an individual U.S. sale of subject merchandise. 19 C.F.R. § 351.414(b)(3). Generally, Commerce will average, and use for comparison, only those home market sales that incurred during the “contemporaneous month” of the U.S. sale. Id. § 351.414(e)(1). Where there are no home market sales of the foreign like product during the same month as the U.S. sale, the Department’s regulation prescribes a “normal” method to determine the contemporaneous month: the so-called 16 Under section 773(a)(1)(A) of the Tariff Act, normal value is to be based on home market prices, made in the ordinary course of trade, existing “at a time reasonably corresponding to the time of the sale used to determine the export price or constructed export price . . . .” 19 U.S.C. § 1677b(a)(1)(A) (emphasis added). Consol. Court No. 10-00106 Page 27 “90/60-day window” method.17 Id. § 351.414(e)(2). However, Commerce departed from this “normal” method in the Final Results, concluding that it was inappropriate to compare U.S. sales with home-market sales occurring outside the quarter in which the U.S. sale occurred. Union Steel Mfg. Co., Ltd., 36 CIT at __, 837 F. Supp. 2d at 1325-26. Instead, Commerce determined that “it is appropriate in this case to match sales only within the same quarter.” Id., 36 CIT at __, 837 F. Supp. 2d at 1326. In contesting the Final Results, Union and HYSCO objected to the Department’s deviation from the 90/60-day window period rule. Id., 36 CIT at __, 837 F. Supp. 2d at 1325. Reviewing these claims, the court noted that “the Department’s decision to use a non-standard method of determining the contemporaneous month was solely a consequence of the decision to apply a non-standard quarterly cost methodology.” Id., 36 CIT at __, 837 F. Supp. 2d at 1326. Commerce itself, in its request for voluntary remand, questioned aspects of its quarterly cost methodology in the fifteenth review, and the court ordered Commerce to reconsider the quarterly cost methodology as used in the Final Results. Id. at 1325-26. The court also ordered Commerce, on remand, to reconsider the decision to depart from the normal method of determining the contemporaneous month as prescribed in the Department’s regulation. Id. 17 Section 351.414(e)(2) of the Department’s regulations describes the “90/60-day window” rule as follows: (2) Contemporaneous month. Normally, the Secretary will select as the contemporaneous month the first of the following which applies: (i) The month during which the particular U.S. sale under consideration was made; (ii) If there are no sales of the foreign like product during this month, the most recent of the three months prior to the month of the U.S. sale in which there was a sale of the foreign like product. (iii) If there are no sales of the foreign like product during any of these months, the earlier of the two months following the month of the U.S. sale in which there was a sale of the foreign like product. 19 C.F.R. § 351.414(e)(2). Consol. Court No. 10-00106 Page 28 On remand, Commerce retained the contemporaneous month methodology it applied in the Final Results. Remand Redetermination 27. Commerce explained that it continued to find it “appropriate to depart from the normal method of determining the contemporaneous month” because it continued to “determine that the changes in Union’s and HYSCO’s [cost of manufacturing] during the POR due to fluctuating raw material input costs [were] significant.” Id. at 27-28. According to Commerce, “[w]hen significant cost changes have occurred during the POR, these same conditions are typically accompanied by changes in prices as the market reacts to changing economic conditions.” Id. at 28. In these situations, Commerce argues, “price-to-price comparisons should be made within the shorter cost-averaging period to lessen the margin distortions caused by changes in sales prices which result from significantly changing costs.” Id. Thus, in the Department’s view, “comparing home market sales from one quarter to U.S. sales during another quarter of the POR when the unadjusted comparison market price does not reflect the contemporaneous price changes that have occurred through the date of the U.S. sale distorts the dumping analysis” and it is therefore “appropriate in this case to match sales only within the same quarter.” Id. at 29. Both Union and HYSCO continue to oppose the departure from the 90/60-day window period.18 Union maintains that the Department’s methodology is unlawful because the shortened comparison window led to fewer identical sales matches, contravening the statutory preference for matching identical merchandise. Union’s Comments 14-19. HYSCO submits that the Department’s methodology is unreasonable because the artificial three-month comparison 18 Neither United States Steel Corporation nor Nucor Corporation addressed the comparison window issue in their written comments on the Remand Redetermination. Comments of U.S. Steel Corp. on the Results of Redetermination Pursuant to Remand Issued by the Dept. of Commerce 1 (Dec. 3, 2012), ECF No. 171; Nucor Corp.’s Comments on Remand Results 1 (Dec. 3, 2012), ECF No. 176. Consol. Court No. 10-00106 Page 29 window creates distortions in the final margin that undermine the statutory interest in accuracy. HYSCO’s Comments 8-11. The court concludes that it cannot affirm the Department’s departure from the 90/60-day comparison window in favor of a quarterly comparison window on the rationale Commerce provided in the Remand Redetermination. As the Department’s regulation, 19 C.F.R. § 351.414(e)(2), establishes a normal method for selecting the comparison month, Commerce must provide an explanation justifying a deviation from this method. See NMB Singapore Ltd. v. United States, 557 F.3d 1316, 1328 (Fed. Cir. 2009) (“Once Commerce establishes a course of action, however, Commerce is obliged to follow it until Commerce provides a sufficient, reasoned analysis explaining why a change is necessary.”); see also Save Domestic Oil, Inc. v. United States, 357 F.3d 1278, 1283-84 (Fed. Cir. 2004) (“[I]f Commerce has a routine practice for addressing like situations, it must either apply that practice or provide a reasonable explanation as to why it departs therefrom.”). The rationale Commerce offers does not provide such an explanation. The Department’s methodology raises two issues that the Remand Redetermination fails to resolve. First, by shortening the standard window period by half (i.e., from six months to the three months), the Department’s method sacrifices identical matches for the sake of some form of contemporaneity. It thus caused Commerce to resort more often to matches of similar merchandise or to a CV calculation even though actual price comparisons inherently yield a more accurate margin than do comparisons of similar merchandise or the use of CV. See 19 U.S.C. § 1677b(a), (e). The Remand Redetermination explains that cost fluctuations in the price of steel substrate, which Commerce presumed to affect price, justify shortening the comparison window. Commerce, however, provided no reasoning beyond this presumption, nor Consol. Court No. 10-00106 Page 30 does Commerce explain how its method produced the most accurate margin possible. The court therefore considers the Department’s rationale conclusory. Commerce must address this deficiency on remand. The second issue raised by the Department’s choice of contemporaneous month is whether Commerce permissibly shortened the comparison window period in the particular way that it did: by limiting comparisons of a U.S. sale to home market sales occurring only in the quarter in which the U.S. sale occurred. By doing so, Commerce largely dispensed with the hierarchy, reflected in 19 C.F.R. § 351.414(e)(2), of matching a U.S. sale with earlier months of home market sales before resorting to subsequent months in a situation where no match could be made in the month in which the U.S. sale occurred. The Remand Redetermination does not explain why Commerce chose a method that deviated significantly from this hierarchy by shortening the comparison window. For example, if a U.S. sale made in the first month of a quarter had no match in the month in which it occurred, it could not be matched with a home market sale (or sales) occurring in the immediately preceding month, despite the preference embodied in the regulation for the use of earlier months before the use of later months.19 Under 19 At oral argument, the court obtained clarification on the Department’s methodology and the three potential scenarios that could arise thereunder. First, if a U.S. sale occurred during the first month of a quarter, Commerce looked to foreign market sales in the current month, followed by the two subsequent months, in chronological order, to identify the contemporaneous month (a “0/60 window”). Oral Tr. 53. If a U.S. sale occurred during the second month of a quarter, Commerce looked to foreign market sales in the current month, followed by the month prior and month subsequent, in that order, to identify the contemporaneous month (a “30/30 window”). Id. at 53-54. And if a U.S. sale occurred during the third month of a quarter, Commerce looked to foreign market sales in the current month, followed by the two prior months, in reverse chronological order, to identify the contemporaneous month (a “60/0 window”). Id. at 54. If Commerce was unable to find a match of identical products (identified by “control number,” or “CONNUM”) using its quarterly window period methodology, it applied the same methodology as to matches of similar CONNUMs before resorting to the use of constructed value. Id. Consol. Court No. 10-00106 Page 31 the regulation, a match occurring as early as the third month before the month of the U.S. sale is preferred to a match occurring in the month following the month of the U.S. sale. See 19 C.F.R. § 351.414(e)(2). Accordingly, in a second remand redetermination, Commerce must reconsider its methodology for identifying reasonably corresponding contemporaneous months for Union’s and HYSCO’s sales of subject merchandise and, in so doing, address the two specific issues the court has raised. The cases Commerce cited for the point that courts have sustained its method are not binding on the court. This includes the decision in Garofalo, which the United States Court of Appeals for the Federal Circuit (“Court of Appeals”) issued pursuant to Federal Circuit Rule 36 and therefore is not precedential. See Fed. Cir. R. 36 (allowing entry of a judgment of affirmance without opinion when certain conditions exist and an opinion would have no precedential value). Moreover, the court must consider the methodology on the facts of this case and according to the rationale Commerce put forth. Finally, the Department’s insistence that neither Union nor HYSCO identified unreasonable matches resulting from the Department’s methodology does not convince the court that this methodology must be sustained. See Remand Redetermination 31. The methodology has the inherent characteristic of reducing the number of identical matches, as defendant concedes.20 Def.’s Resp. 27 (“[T]he shortened window may not have provided for every identical match that could have been made using the standard window period . . . .”). 20 At oral argument, Union’s counsel was unable to identify the extent of the Department’s use of constructed value resulting from the quarterly comparison window but did indicate that there were fewer identical matches than there would have been under the normal 90/60-day comparison window. Oral Tr. 43-44. Consol. Court No. 10-00106 Page 32 Before the court, defendant, citing 19 C.F.R. § 351.414, argues that “Commerce reasonably exercised the discretion permitted by the regulation’s use of the term ‘normally’” after determining that the factual scenario presented warranted a shortening of the comparison window period. Def.’s Resp. 26. Defendant submits that this Court has held that the use of the term “normally” affords Commerce discretion in determining when the normal situation does not apply, based on the circumstances presented. Id. (citing KYD v. United States, 33 CIT 299, 311, 631 F. Supp. 2d 1371, 1382; (2009), SeAH, 34 CIT at __, 704 F. Supp. 2d at 1376). The non-precedential cases defendant cites, like those cited by Commerce in the Remand Redetermination, do not offer the court a basis upon which to sustain the Department’s decision as to contemporaneous month. The regulation affords Commerce discretion in its sales comparison methodology, but in this case, the particular method Commerce adopted as an alternative to the normal method raises unanswered questions that relate to the accuracy of the result. C. The Remand Redetermination Lawfully Determined that the Non-Laminated, Painted CORE May Not Be Compared to the Laminated Core as Products “Identical in Physical Characteristics” Prior to the court’s first opinion in this action, defendant requested, and the court granted, a voluntary remand to allow Commerce to reconsider aspects of the “model-match” methodology applied to Union under section 771(16)(A) of the Tariff Act, 19 U.S.C. § 1677(16)(A).21 Union Steel Mfg. Co., Ltd., 36 CIT at __, 837 F. Supp. 2d at 1327. The model-match methodology is used by Commerce to compare prices of subject merchandise with home market prices of the foreign like product. Specifically, Commerce wanted to reconsider its decision to treat Union’s 21 Section 1677(A) of the Tariff Act directs that Commerce, in determining the foreign like product, first seek to compare a U.S. sale of subject merchandise with a comparison-market sale of merchandise “which is identical in physical characteristics with, and was produced in, the same country by the same person as, that merchandise.” 19 U.S.C. § 1677(A). Consol. Court No. 10-00106 Page 33 plastic-laminated CORE (specifically, polyethylene terephthalate (PET) or polyvinyl chloride (PVC) plastic film) as “identical in physical characteristics” with certain painted, non-laminated CORE. Id. During the administrative review, Commerce had rejected a proposal made by Union to treat laminated CORE as a separate model-match category. Id. Before the court, Union argued that on the basis of obvious physical differences between laminated and painted CORE, the Department’s finding that the two groups of products are “identical” is unsupported by substantial record evidence. Id. In remanding the issue, the court held that Commerce “may not compare as identical merchandise Union’s sales of the painted and non-laminated CORE with Union’s sales of laminated CORE absent a finding, supported by substantial evidence on the record, that the physical differences distinguishing the two product groups are minor and not commercially significant.” Id. On remand, Commerce altered its model-match methodology so that Union’s laminated CORE products and non-laminated, painted CORE products are no longer considered identical in physical characteristics under § 1677(16)(A). Remand Redetermination 32-33. Commerce found, based on an examination of Union’s antidumping questionnaire responses, that “laminated CORE products by their very nature are not painted products” as the laminated products are “coated by attaching a plastic film to a CORE substrate, and lamination is done in lieu of painting.” Id. at 35. Commerce also found that “‘[l]aminating the steel substantially increases both the production costs and the sales prices of laminated products vis-à-vis other painted products.’” Id. at 36 (quoting Union Steel Questionnaire Resp. 6 (Feb. 5, 2009) (Admin.R.Doc. No. 4845) (“Union’s Questionnaire Resp.”). Based on an examination of Union’s price and cost data, Commerce noted that “PET film and PVC film are more expensive than the various paints used to produce non-laminated, painted CORE products.” Id. at 35 Consol. Court No. 10-00106 Page 34 (citing Union Steel Supplemental Questionnaire Response 15 (Apr. 9, 2009) (Admin.R.Doc. No. 4915) (“Union’s Supplemental Questionnaire Resp.”)). Commerce identified record evidence showing that “Union’s customers sometimes require specific characteristics related to PET or PVC film, while other painted products are not suitable” and that “Union specifically distinguishes laminated products from other painted products in product codes.” Id. at 36 (citing Union’s Supplemental Questionnaire Resp. at 13-14, B-21). Commerce concluded that because “[r]ecord evidence establishes significant differences in the physical characteristics between laminated CORE products and non-laminated, painted CORE products . . . compelling reasons exist for the Department to alter its classification of the physical characteristics in this remand proceeding.” Id. at 37. The Department’s separation of laminated CORE products from non-laminated, painted CORE products for purposes of its model-match methodology complies with the court’s remand order and is supported by substantial record evidence of the physical and commercial differences between the respective products. No party opposes this redetermined model-match methodology in comments filed before the court. See Union’s Comments 19-20. The court, therefore, affirms this aspect of the Remand Redetermination. D. The Court Sustains the Department’s Use of Zeroing in the Final Results to Calculate Union’s Weighted Average Dumping Margin In the Final Results, Commerce applied its “zeroing” methodology to calculate Union’s weighted average dumping margin, under which it determines a dumping margin for each sale of subject merchandise and then converts negative margins to zero margins before calculating a weighted average percentage margin. See Union Steel Mfg. Co., Ltd., 36 CIT at __, 837 F. Supp. 2d at 1327-28. Noting that Commerce abandoned the use of zeroing in antidumping investigations while continuing to apply zeroing in periodic administrative reviews, Union Consol. Court No. 10-00106 Page 35 challenged the Department’s use of zeroing, arguing that Commerce inconsistently and unlawfully interpreted the statutory provision defining “dumping margin,” 19 U.S.C. § 1677(35)(A), contrary to basic rules of statutory construction.22 Id., 36 CIT at __, 837 F. Supp. 2d at 1328 (citation omitted). Defendant requested, and the court granted, a voluntary remand on the zeroing issue in light of the decision of the in JTEKT Corp. v. United States, 642 F.3d 1378, 1383-85 (Fed. Cir. 2011) (“JTEKT Corp.”), in which the Court of Appeals questioned the legality of the Department’s construction of section 1677(35)(A) and required the Department to provide an adequate explanation of the Department’s inconsistent interpretation of the statute vis-à-vis administrative reviews and investigations. Id. The court directed the Department, on remand, to “either modify its decision to apply the zeroing methodology in the fifteenth review or, alternatively, provide an explanation that satisfies the requirements the Court of Appeals imposed in . . . JTEKT Corp.” Id., 36 CIT at __, 837 F. Supp. 2d at 1329. In the Remand Redetermination, Commerce continued to use zeroing but, as directed by the court’s remand order, provided an explanation as to why it considered the use of that methodology to be in compliance with the Tariff Act despite the Department’s discontinuation of zeroing in antidumping duty investigations. Remand Redetermination 44, 48-60 (stating that the Department’s interpretation reasonably resolves the ambiguity in section 771(35) of the Act “in a way that accounts for the inherent differences between the result of an average-to-average comparison,” as used in investigations, and “the result of an average-to-transaction comparison,” as used in administrative reviews). 22 The relevant statutory provision defines a dumping margin as “the amount by which the normal value exceeds the export price or constructed export price of the subject merchandise.” 19 U.S.C § 1677(35)(A). Consol. Court No. 10-00106 Page 36 The Court of Appeals decided Union Steel, LG Hausys, Ltd. v. United States, 713 F.3d 1101, 1101 (Fed. Cir. 2013) (“Union Steel”), on April 16, 2013, after Commerce had already issued its Remand Redetermination. In Union Steel, the Court of Appeals affirmed the Department’s use of the zeroing methodology in circumstances analogous to those presented by this case involving an administrative review of an antidumping duty order. The court considers Union Steel dispositive of the zeroing issue presented by this case and sustains the use of zeroing in the Final Results. E. The Court Must Order a Second Remand of the Department’s Decision to Use the Date of Shipment as the Date of Sale for HYSCO’s Constructed-Export-Price Sales In the Final Results, Commerce used the date the subject merchandise was shipped from Korea to HYSCO’s U.S. affiliate, Hyundai HYSCO USA (“HHU” or “HYSCO USA”), as the date of sale for HYSCO’s constructed-export-price (“CEP”) sales of subject merchandise. Union Steel Mfg. Co., Ltd., 36 CIT at __, 837 F. Supp. 2d at 1334. In contesting the Final Results, U.S. Steel claimed the decision to use the date of shipment as the date of sale, rather than the date that the subject merchandise was invoiced, violated the Department’s regulation, 19 C.F.R. § 351.401(i),23 and was unsupported by substantial evidence. Id. Defendant requested, and the court granted, a voluntary remand to allow Commerce to reconsider the date-of-sale decision Commerce made in the Final Results for HYSCO’s CEP sales. Id., 36 CIT at __, 837 F. 23 Section 351.401(i) provides as follows: In identifying the date of sale of the subject merchandise or foreign like product, the Secretary normally will use the date of invoice, as recorded in the exporter or producer’s records kept in the ordinary course of business. However, the Secretary may use a date other than the date of invoice if the Secretary is satisfied that a different date better reflects the date on which the exporter or producer establishes the material terms of sale. 19 C.F.R. § 351.401(i). Consol. Court No. 10-00106 Page 37 Supp. 2d at 1335. On remand, Commerce again decided to use the date of shipment as the date of sale. Remand Redetermination 62. In opposing the Remand Redetermination, U.S. Steel argues, inter alia, that 19 C.F.R. § 351.401(i) implements the congressional intent that the date of sale is the date on which the material terms of sale are established, as demonstrated by the Statement of Administrative Action Accompanying the Uruguay Round Agreements Act (“SAA”). U.S. Steel’s Comments 5-6 (citing Statement of Admin. Action Accompanying the Uruguay Round Agreements Act at 810), reprinted in 1994 U.S.C.C.A.N. 4040, 4153 (“Article 2.4.1 specifies that currency conversions should be made using the rate of exchange on the date of sale, which is defined as a date when the material terms of sale are established.”) According to U.S. Steel, “[t]here is not a single shred of evidence showing that the price to be charged to HYSCO’s unaffiliated customer was fixed at the time of shipment or at any time prior to the invoice date: no emails, offers, confirmations, contracts or other documents of any kind.” Id. at 7. The Remand Redetermination maintains that using the date of shipment as the date of sale is the Department’s consistent and “normal” practice “[i]n cases where the shipment date precedes the invoice date.” Remand Redetermination 61-62. Commerce explained that “[i]n this administrative review, the Department finds no reason to depart from its normal practice not to consider dates subsequent to the date of shipment from the factory as appropriate for date of sale because once merchandise is shipped the material terms of sale are presumed to be established.” Id. at 62 (emphasis in original). The court considers, first, whether the decision challenged by U.S. Steel is governed by 19 C.F.R. § 351.401(i). It might be questioned whether this regulation applies to CEP sales, which are sales made to the first unrelated buyer in the United States, 19 U.S.C. § 1677a(b), or Consol. Court No. 10-00106 Page 38 whether the regulation applies only to export price sales. Like the transactions at issue in this case, CEP sales typically involve a U.S. reseller affiliated with the exporter or producer, and the invoice may be issued by the reseller after the date the merchandise is shipped to the United States. The first sentence of Section 351.401(i) might be read to contemplate an export price sale in identifying as the presumptive date of sale “the date of invoice, as recorded in the exporter or producer’s records kept in the ordinary course of business.” 19 C.F.R. § 351.401(i) (emphasis added). The second sentence also vaguely suggests an export price sale in referring to a date that “better reflects the date on which the exporter or producer establishes the material terms of sale.”24 Id. (emphasis added). Nevertheless, the text of 19 C.F.R. § 351.401(i) does not expressly limit the provision to export price sales, and no other regulatory provision addresses the date-of-sale determination. Notably, the Remand Redetermination does not conclude that the scope or intent of 19 C.F.R. § 351.401(i) is limited to export price sales. Rather, the Remand Redetermination considers the regulation to apply to the CEP issue before the court, Remand Redetermination 61, instructing that “Section 351.401(i) of the Department’s regulations states that the Department ‘normally will use the date of invoice’ as the date of sale, unless ‘the Secretary is satisfied that a different date better reflects the date on which the exporter or producer establishes the material terms of sale.’” Id. (quoting 19 C.F.R. § 351.401(i)). The Remand Redetermination adds that “[i]f the Department determines that another date better reflects the date on which the exporter or producer establishes the material terms of sale, the Department may use this date.” Id. The 24 The reference to only the “exporter or producer” raises a question in that “establishing” the material terms of sale would appear to require the participation of a buyer. However, the general context may recognize the routine situation in which the producer/exporter accepts a customer’s order by issuing a commercial document such as an invoice. Consol. Court No. 10-00106 Page 39 Remand Redetermination proceeds to discuss the Department’s date-of-shipment practice in the context of 19 C.F.R. § 351.401(i). Id. at 61-66. Because the Remand Redetermination applies 19 C.F.R. § 351.401(i), as well as the date-of-shipment practice, in deciding the date of sale for HYSCO’s CEP sales, the court reviews the Department’s decision on the basis Commerce put forth, i.e., as grounded in both the regulation and the claimed date-of-shipment practice. And because the factual situation presented here involves CEP sales, applying the regulation to that factual situation requires the court to construe the second sentence of the regulation more broadly than in the literal sense— i.e., the court must construe the sentence to refer to a date on which the material terms of sale were established, despite the limiting reference to the exporter or producer. See 19 C.F.R. § 351.401(i) (“However, the Secretary may use a date other than the date of invoice if the Secretary is satisfied that a different date better reflects the date on which the exporter or producer establishes the material terms of sale.”) (emphasis added). This broader construction also conforms to the underlying purpose of the regulation. As U.S. Steel points out, Congress intended the date of sale for antidumping law purposes to be the date on which the material terms of sale are established. See also Corus Staal, 502 F.3d at 1376 (“Neither a sale nor an agreement to sell occurs until there is mutual assent to the material terms (price and quantity).”). The second sentence of 19 C.F.R. § 351.401(i) requires a finding of fact. This case, therefore, raises a question of whether the regulation permitted Commerce to use a date of sale other than the date of invoice without making a lawful finding according to that sentence, i.e., a finding supported by substantial record evidence, that some date other than the invoice date (in this case, the date of shipment) better reflected the date on which the material terms of sale were established. It could be argued that the word “normally” in the first sentence of the regulation Consol. Court No. 10-00106 Page 40 plausibly could be interpreted to allow deviation from the normal “date-of-invoice” method even in situations other than the specific situation identified in the second sentence. Under such an interpretation, Commerce arguably could rely on its date-of-shipment practice without making the specific finding required by the second sentence. Alternatively, the regulation could be construed such that the only exception to the normal method identified in the first sentence of § 351.401(i), which presumes date of invoice will control, is a finding made pursuant to the second sentence. Supporting the latter construction is the use of the word “however,” a word of qualification, to introduce the second sentence and relate the second sentence to the first. This construction, consistent with the SAA and the Antidumping Agreement, links the first sentence (and thereby the regulation as a whole) to the concept of a date on which the material terms of sale are established. The plain language of 19 C.F.R. § 351.401(i) is susceptible to both interpretations. Therefore, the court considers it appropriate to examine the Department’s stated intent at the time it promulgated the regulation. In the preamble to the 1997 promulgation notice (“Preamble”), Commerce explained that in “[i]n paragraph (i) [of § 351.401], we merely have provided that, absent satisfactory evidence that the terms of sale were finally established on a different date, the Department will presume that the date of sale is the date of invoice.” Antidumping Duties; Countervailing Duties, 62 Fed. Reg. 27,296, 27,349 (May 19, 1997) (“Preamble”). The Preamble thus explains the intended relationship of the two sentences in 19 C.F.R. § 351.401(i). Commerce intended the presumption in favor of invoice date to apply except in the circumstance identified in the second sentence, i.e., when Commerce can make a finding of fact based on “satisfactory evidence” that “the terms of sale were finally established on a different date.” Id. The Preamble underscores this point by stating that a respondent “will Consol. Court No. 10-00106 Page 41 be free to argue that the Department should use some date other than the date of invoice, but the respondent must submit information that supports the use of a different date” and that a respondent’s description of its selling processes, like all submitted information, “will be subject to verification.” Id. The Preamble clarifies that the second sentence of 19 C.F.R. § 351.401 requires a case-specific factual determination, not a mere presumption based solely on the sequence of the date of shipment and the date of invoice. In the second sentence, the regulation refers to “a different date” than the date of invoice but makes no specific reference to the date of shipment. In fact, the Preamble discusses the Department’s reasons for rejecting the suggestion of some commenters that “if the Department uses a uniform date of sale, it should use date of shipment rather than date of invoice.” Id. The Preamble responded to the suggestion as follows: For several reasons, the Department has not adopted this suggestion. First, date of shipment is not among the possible dates of sale specified in note 8 of the AD Agreement.25 Second, based on the Department’s experience, date of shipment rarely represents the date on which the material terms of sale are established. Third, unlike invoices, which can usually be tied to a company’s books and records, firms rarely use shipment documents as the basis for preparation of financial reports. Thus, reliance on date of shipment would make verification more difficult. Preamble, 62 Fed. Reg. at 27,349. In this way, the Preamble casts doubt on the notion that the date-of-shipment practice conforms to the regulation. It is questionable that Commerce, at the time of promulgation, intended 19 C.F.R. § 351.401(i) to authorize a practice under which the date of sale is presumed to be the date of shipment, which according to Commerce “rarely 25 The Antidumping Agreement, to which the passage refers, states that “[n]ormally, the date of sale would be the date of contract, purchase order, order confirmation, or invoice, whichever establishes the material terms of sale.” Agreement on Implementation of Article VI of the General Agreement on Tariffs & Trade 1994 (Antidumping Agreement), 1868 U.N.T.S. 201, Annex 1A, Art. 2.4.1, n.8. Consol. Court No. 10-00106 Page 42 represents the date on which the material terms of sale are established.” Id. At the time of promulgation, Commerce explicitly recognized that the shipment date may precede the invoice date but still rejected the notion that the shipment date should be presumed to control in that circumstance. Addressing a comment that shipment date, not invoice date, should be presumed to control because use of the latter as the date of sale could lead to manipulation of dumping calculations by respondents, Commerce opined that “most firms have a standard invoicing practice (e.g., three days after shipment, every two weeks).” Id. Despite these indications to the contrary, the Remand Redetermination applies the general principle that “once merchandise is shipped the material terms of sale are presumed to be established.” Remand Redetermination 62. By so doing, the Remand Redetermination adopted a presumption in favor of the shipment date even though Commerce intended when promulgating § 351.401 that the presumption would be in favor of the invoice date, regardless of the sequence of the two dates. Although it appears that the Department’s date-of-shipment practice is inconsistent with the interpretation of 19 C.F.R. § 351.401(i) in the Preamble, the court need not decide the question of whether the regulation renders that practice impermissible per se. Rather, it is sufficient for the court to conclude that the Department may not substitute the presumption on which this practice is based, i.e., that material terms of sale are established as of the date of shipment, in lieu of a finding under the second sentence of § 351.401(i) that is supported by substantial record evidence. The regulation presumes that date of invoice will control in the absence of a finding in the particular circumstances of a case “that a different date better reflects the date on which the exporter or producer establishes the material terms of sale.” 19 C.F.R. § 351.401(i). Consol. Court No. 10-00106 Page 43 In applying the date-of-shipment practice to determine a uniform date of sale for HYSCO, the Remand Redetermination concluded that there was an absence of evidence that the sales prices “actually changed” after the date of shipment and also found that the “price remained stable after the date of shipment.” Remand Redetermination 62. These findings do not satisfy the condition imposed by 19 C.F.R. § 351.401(i), which does not permit Commerce to base a decision to use the date of shipment as the date of sale on an absence of record evidence that the sales prices changed after the date of shipment. The finding that the “price remained stable after the date of shipment,” even if presumed valid, would not establish conclusively the intent of the three parties to the HYSCO sales as to when the material terms of sale were established. Moreover, this finding is unsupported by substantial record evidence. At oral argument, counsel for defendant did not dispute that Commerce reached the finding based on certain sales documentation pertaining to a single sale. Oral Tr. 100. This evidence is minimally probative on the question of whether prices were stable after shipment and even less probative on the larger question of whether the parties intended the prices to be established as of the time of shipment. The Remand Redetermination also finds that “[t]he evidence with regard to price shows that the transfer price from HYSCO to HYSCO USA did not vary from the offer to the commercial invoice, which was issued on the shipment date.” Remand Redetermination 65 (citing HYSCO Section A Questionnaire Resp., Ex. A-10 (Feb. 12, 2009) (Admin.R.Doc. No. 4850) (“HYSCO’s Section A Resp.”). This evidence, which does not pertain directly to the prices at issue, has little probative weight on the question of whether the parties intended the price to an unaffiliated customer to be established as of the date of shipment. Consol. Court No. 10-00106 Page 44 More probative on the question of when the three parties to the HYSCO sales intended to establish the material terms of sale is HYSCO’s Section A Questionnaire Response. The questionnaire response on page A-22 describes the HYSCO’s U.S. sales process as follows: Upon receipt of the customer’s inquiry, HHU will begin to negotiate the terms of sale with the customer, while making a corresponding inquiry with HYSCO. Although the negotiations may continue until actual shipment, once an initial agreement is reached between the customer and HHU as well as between HHU and HYSCO, HYSCO issues an order confirmation and schedules production of the merchandise. Once production is complete, HYSCO ships the merchandise to the United States, while issuing the relevant sales documentation, including the commercial invoice, packing list, and mill certificate, to HHU . . . . Upon arrival of the merchandise in the United States, HHU issues an invoice to the customer and is responsible for collecting payment from the customer. Id. at A-22 (emphasis added). The questionnaire response elaborated further on page A-23 that “[n]egotiations with customers can continue through the entire sales process” and that “[f]or U.S. sales, quantity can also change up until the merchandise is shipped from HYSCO’s factory, and price can change up until HHU issues its invoice to the unaffiliated U.S. customer.” HYSCO’s Section A Resp. at A-23, A-24. According to defendant, the passage on page A-22 indicates that negotiations do not continue after actual shipment. Oral Tr. 96-97. The court, however, looks at both of the passages quoted and concludes that there is not substantial evidence to support a finding that the prices to the unaffiliated buyers in HYSCO’s sales were established as of the date of shipment. The passage on page A-22 suggests, but does not state definitively, that negotiations did not continue after shipment. The passage on page A-23 speaks directly to that question, clarifying that price, as opposed to quantity, could change until HHU issues the invoice. Notably, the sentence in the passage on page A-22, on which defendant relies for its contention that negotiations did not continue after shipment, states that production is scheduled according to what is described as an “initial” agreement, suggesting that the “agreement” reached as of that point is subject to change after production. See HYSCO’s Section A Resp. A-22 Consol. Court No. 10-00106 Page 45 (“Although the negotiations may continue until actual shipment, once an initial agreement is reached between the customer and HHU as well as between HHU and HYSCO, HYSCO issues an order confirmation and schedules production of the merchandise.” (emphasis added)). The two passages can be read to be consistent with one another, and together they connote that quantity is established as of the date of shipment but price is not. As Commerce itself acknowledged in the Remand Redetermination, “nothing on the record refutes HYSCO’s statement that negotiations for sales to U.S. customers continued after shipment . . . .” Remand Redetermination 62 (emphasis added). The Remand Redetermination discusses record evidence but does not explicitly find that the parties to the HYSCO sales intended all material terms of sale to be established on the shipment date.26 It is in this respect that the date-of-shipment practice was unlawful as applied to HYSCO’s CEP sales in the fifteenth review: Commerce impermissibly interpreted § 351.401(i) to be satisfied by the presumption underlying that practice, applying the principle that “once merchandise is shipped the material terms of sale are presumed to be established.” Remand Redetermination 62. This principle is contrary to § 351.401(i) as explained in the Preamble. The regulation presumes invoice date, not shipment date, is the date of sale absent a valid finding that some other date better reflects the date on which the material terms of sale were established. 26 This case is distinguishable from U.S. Steel Corp. v. United States, 37 CIT __, Slip Op. 13-156 (Dec. 27, 2013), which affirmed the Department’s decision to use shipment date as the date of sale for a respondent in the seventeenth administrative review of the same antidumping duty order at issue here. Id. In contrast to the facts in this case, which concerns the fifteenth administrative review, the respondent in U.S. Steel Corp. confirmed that it “and its unaffiliated customers generally set price and quantity terms through e-mail correspondence” prior to shipment and that changes to material terms could only be made before shipment. Id., 37 CIT at __, Slip Op. 13-156, at 8-9 (citations omitted). Consol. Court No. 10-00106 Page 46 In contrast, the Department’s decision to use date of shipment for HYSCO’s sales relies on a regulatory interpretation of § 351.401(i) that requires no such finding and instead is satisfied by a presumption that the date of shipment is the date on which the material terms of sale are established. The court does not grant this interpretation the measure of deference afforded by Auer v. Robbins, 519 U.S. 452, 461 (1997), because “there is reason to suspect that the agency’s interpretation ‘does not reflect the agency’s fair and considered judgment on the matter in question.’” Christopher v. SmithKline Beecham Corp., 567 U.S. __, __, 132 S. Ct. 2156, 2169 (2012) (quoting Auer, 519 U.S. at 462).27 Instead, the court examines the Department’s interpretation for “those factors which give it the power to persuade” in accordance with the level of deference afforded by Skidmore v. Swift & Co., 323 U.S. 134, 140 (1944). Because the Remand Redetermination devotes only scant attention to 19 C.F.R. § 351.401(i), Skidmore deference to a regulatory interpretation in the Remand Redetermination does not suffice to sustain the department’s decision on date of sale. Defendant argues that “Commerce’s normal practice is not to consider dates subsequent to the date of shipment,” a practice which, according to defendant, “has been ‘implicitly approved by the courts.’” Def.’s Resp. 39 (quoting Mittal Steel Point Lisas Ltd. v. United States, 31 CIT 638, 647, 491 F. Supp. 2d 1222, 1231 (2007) (citing AIMCOR v. United States, 27 In Auer v. Robbins, 519 U.S. 452 (1997), the United States Supreme Court held that an agency’s interpretation of its own regulation is entitled to deference unless it is “plainly erroneous or inconsistent with the regulation.” Id. at 461. The Supreme Court declined to grant Auer deference in Christopher v. SmithKline Beecham Corp., 567 U.S. __, 132 S. Ct. 2156 (2012), to an interpretation by the Department of Labor of its own regulation and instead reviewed the interpretation for persuasiveness, deciding that the interpretation lacked “the hallmarks of thorough consideration.” Id. 567 U.S. at __, 132 S. Ct. at 2169. The Supreme Court concluded that the interpretation “should instead be given a measure of deference proportional to its power to persuade.” Id., 567 U.S. __, __, 132 S. Ct. 2156, 2160 (2012) (citing United States v. Mead Corp., 533 U.S. 218, 228 (2001)). Consol. Court No. 10-00106 Page 47 141 F.3d 1098, 1104-05) (Fed. Cir. 1998))). However, none of the cases defendant cites holds that the date-of-shipment practice, as described in the Remand Redetermination, is permissible under 19 C.F.R. § 351.401. Defendant also argues that 19 C.F.R § 351.401(i) “establishes only a presumption that Commerce use the date of invoice as the date of sale.” Id. at 40. Implicit in this argument is that the presumption established by the regulation may be rebutted by something short of a finding that a date other than the date of invoice better reflects the date on which the material terms of sale were established. For the reasons the court has discussed, the regulation is properly interpreted according to the intent Commerce had at the time of promulgation, under which the second sentence is exhaustive of the situations in which the normal practice described in the first sentence will not apply. Defendant argues, further, that “available record evidence establishes that price remained stable for at least a period of time,” noting the Department’s finding that “[t]he transfer price from HYSCO to HYSCO USA did not vary from the offer to the commercial invoice, which was issued on the shipment date.” Def.’s Resp. 41. According to defendant, “[w]hile Commerce normally does not use transfer price as a basis for United States price, the way in which Commerce relied upon that information in the remand results is reasonable because it is the only evidence on the record indicative of HYSCO’s actual price-setting behavior.” Id. This argument ignores the presence on the record of HYSCO’s Section A Questionnaire Response, which indicates that prices could change up until the invoice is issued to the unaffiliated purchaser. Regarding the questionnaire response, defendant at oral argument cited the above-quoted passage on page A-22 as creating an ambiguity on whether negotiations on price could continue after the date of shipment. Oral Tr. 96-97. Defendant suggested, in addition, that pages A-22 Consol. Court No. 10-00106 Page 48 and A-23 could be read to be consistent with one another and to mean “that the price per unit doesn’t change once the product is shipped, but because quantity can vary slightly within a certain tolerance level . . . the end, the total price may change at the end of the day.” Id. at 105. These arguments rest on a post hoc rationale not articulated in the Remand Redetermination, which itself conceded that no record information rebuts HYSCO’s statement that negotiations for sales to U.S. customers continued after shipment. Remand Redetermination 62. The court must review the Department’s decision on the rationale the agency puts forth. See SEC v. Chenery Corp., 332 U.S. 194, 196 (1947) (“[A] reviewing court, in dealing with a determination or judgment which an administrative agency alone is authorized to make, must judge the propriety of such action solely by the grounds invoked by the agency.”). Moreover, as the court discussed supra, the evidence of record does not support a finding that the parties to the sales intended the price term to be established as of the date of shipment. Defendant’s argument that the statement on page A-23 of the questionnaire response that the price term could change up until the invoice was issued refers only to the total price on the shipment, and not the per-unit price, is not convincing because it is based on a strained reading of the questionnaire response. There is no indication in the questionnaire that HYSCO intended the statement it made on page A-23 to be limited in the way defendant suggests. Defendant also suggested at oral argument that reliance on the invoice date in the fifteenth review might have caused some sales to be unexamined because Commerce used date of shipment as the date of sale in the previous review. Oral Tr. 100-01. If Commerce shared this concern, it chose not to mention it in the Remand Redetermination. In any event, the problem that defendant identifies does not allow the court to disregard 19 C.F.R. § 351.401(i), a Consol. Court No. 10-00106 Page 49 regulation that is binding on both the court and Commerce, and which controls the disposition of the date of sale claim U.S. Steel raises in this case. In support of the Department’s use of the date of shipment in the First Remand Redetermination, HYSCO argues that substantial record evidence supported “the conclusion that the date of shipment from HYSCO’s factory in Korea, which preceded the date of invoice to the unaffiliated U.S. customer, reflects the date on which HYSCO established the material terms of sale for U.S. sales.” HYSCO’s Comments 11. The record lacks substantial evidence to support such a conclusion. Most significantly, HYSCO’s statement on page A-23 of the Section A Questionnaire Response is to the contrary, and, as the Remand Redetermination acknowledged, nothing on the record refutes that statement. Remand Redetermination 62. Contending that it “met its burden of establishing that the date of sale occurred on a date after the invoice date,” HYSCO points to the sample sales documentation it provided and the fact that Commerce did not seek additional documentation. HYSCO’s Comments 12 (quotations omitted). According to HYSCO, “[t]he Department typically will not change its date of sale methodology unless it finds that a date of sale was determined erroneously or the respondent changed its business practice, neither of which applied here.” HYSCO’s Comments 12. See also Oral Tr. 124. This argument misses the essential point that Commerce, regardless of its “typical” methodology, was required to determine the date of sale in this review in compliance with 19 C.F.R. § 351.401(i). In summary, the court concludes that Commerce erred in using the date of shipment as the date of sale for HYSCO’s sales without a valid finding satisfying the second sentence of 19 C.F.R. § 351.401(i). On remand, Commerce must redetermine HYSCO’s margin according to a method of determining the date of sale that complies with the Department’s regulation. Consol. Court No. 10-00106 Page 50 F. On Second Remand, Commerce Must Recalculate the Rate Assigned to Dongbu to Incorporate Any Adjustments Resulting from Changes in the Margins for Union and HYSCO In the Final Results, Commerce assigned to Dongbu a margin of 8.65%, a simple average of the margins assigned to Union and HYSCO. Union Steel Mfg. Co., Ltd., 36 CIT at __, 837 F. Supp. 2d at 1335. In moving for judgment on the agency record, Dongbu incorporated the arguments raised by Union and HYSCO, seeking the benefit of any modifications to the margins assigned thereto on remand.28 Id., 36 CIT at __, 837 F. Supp. 2d at 1335-36. The court ordered Commerce, on remand, to reflect any changes to the margins for Union and HYSCO in the margin applied to Dongbu. Id., 36 CIT at __, 837 F. Supp. 2d at 1336. The court noted, additionally, that the Department’s recalculation of the rate applicable to Dongbu also must incorporate any modifications to the rate applied to HYSCO as a result of U.S. Steel’s claim regarding the date of sale. Id. In the Remand Redetermination, Commerce recalculated the rate assigned to Dongbu to reflect changes to the weighted average dumping margins calculated for Union and HYSCO arising from the final remand results. Remand Redetermination 66. Commerce indicated that Dongbu’s rate remains “a simple average of the revised weighted average dumping margins calculated for the mandatory respondents.” Id. In its comments on the Remand Redetermination, Dongbu incorporates by reference those comments filed by Union Steel and HYSCO while asking the court to order Commerce, on second remand, to recalculate Dongbu’s margin to incorporate any additional changes made to 28 Because Dongbu’s motion for judgment on the agency record limited the claim to those arguments raised by Union and HYSCO, the court concluded that other claims set forth in Dongbu’s complaint that were not raised in Union’s or HYSCO’s motions for judgment on the agency record had been abandoned. Union Steel Mfg. Co., Ltd., 36 CIT at __, 837 F. Supp. 2d at 1336. Consol. Court No. 10-00106 Page 51 Union’s and HYSCO’s margins. Dongbu’s Comments 2. Because Dongbu’s request is appropriate in light of Dongbu’s status as an unexamined respondent, the court will require Commerce, on second remand, to recalculate Dongbu’s margin based on any changes to Union’s and HYSCO’s margins. As Dongbu is to receive the benefit of any downward revisions to the weighted average dumping margins for Union or HYSCO, it also must be made subject to any increases in those margins. Accordingly, any changes in Union’s and HYSCO’s margins should be reflected in the derivative margin applied to Dongbu, including any changes resulting from the reconsideration of U.S. Steel’s claim challenging HYSCO’s date of sale. III. CONCLUSION AND ORDER For the reasons discussed in the foregoing, the court affirms in part, and remands in part, the Remand Redetermination, Results of Redetermination Pursuant to Remand (Sept. 24, 2012), ECF No. 161, issued by the International Trade Administration, U.S. Department of Commerce (“Commerce” or the “Department”) in the fifteenth administrative review of an antidumping order on certain corrosion-resistant carbon steel flat products (“CORE”) from the Republic of Korea. Accordingly, upon consideration of the Remand Redetermination, the comments of the parties thereon, and all papers and proceedings herein, and upon due deliberation, it is hereby ORDERED that the Remand Redetermination submitted by Commerce on September 24, 2012, be, and hereby is, affirmed in part and remanded in part for reconsideration and redetermination in accordance with this Opinion and Order; it is further ORDERED that the Remand Redetermination be, and hereby is, sustained as to: (1) the general and administrative (“G&A”) expense ratio Commerce used to determine the cost of production of the foreign like product for Union Steel Manufacturing Co., Ltd. (“Union”); (2) the Department’s decision to calculate Union’s interest expense ratio using financial statements from both the 2007 and 2008 fiscal years; (3) the Department’s application of its modified “sales-below-cost” and “recovery-of-costs” tests to calculate the normal value of Union’s subject merchandise; (4) the Department’s decision to create separate product categories for laminated CORE and non-laminated, painted CORE for use in a model-match methodology comparing Union’s home market and U.S. sales; and (5) the Department’s use of the “zeroing” methodology to calculate Union’s weighted average dumping margin; it is further Consol. Court No. 10-00106 Page 52 ORDERED that on remand, Commerce, in accordance with this Opinion and Order, shall reconsider its decision to make a major input adjustment when calculating the interest expense ratio for Union and address Nucor Corporation’s related objections; it is further ORDERED that on remand, Commerce, in accordance with this Opinion and Order, shall reconsider the application of its modified “quarterly cost” methodology to (1) the revised “sales-below-cost” and “recovery-of-costs” tests used in the normal value calculations applied to Hyundai HYSCO (“HYSCO”) and (2) the constructed value determinations and difference-in- merchandise (“DIFMER”) adjustments applied to Union and HYSCO; it is further ORDERED that, on remand, Commerce, in accordance with this Opinion and Order, shall reconsider the decision to depart from the “normal” method prescribed by regulation, 19 C.F.R. § 351.414(e)(2), for identifying the reasonably corresponding contemporaneous month used in comparing Union’s and HYSCO’s U.S. and home market sales; it is further ORDERED that, on remand, Commerce, in accordance with this Opinion and Order, shall reconsider the decision to depart from the normal date prescribed by regulation, 19 C.F.R. § 351.401(i), when determining date of sale for HYSCO’s U.S. sales sold through its U.S. affiliate, Hyundai HYSCO USA, Inc.; it is further ORDERED that, on remand, Commerce shall redetermine the weighted average dumping margins for Union and HYSCO as necessary and shall redetermine the margin for Dongbu Steel Company, Ltd., as a non-examined respondent, based on any adjustments made to Union’s or HYSCO’s weighted average dumping margins in the results of the second remand; it is further ORDERED that Commerce shall file the results of the second remand within ninety (90) days of the date of this Opinion and Order; it is further ORDERED that plaintiffs, plaintiff-intervenor, and defendant-intervenor each may file comments on the results of the second remand within thirty (30) days from the date on which these results are filed; and it is further ORDERED that defendant may file any response to those comments on the results of the second remand within fifteen (15) days from the date on which the last comment is filed. /s/Timothy C. Stanceu Timothy C. Stanceu Judge Dated: March 4, 2014 New York, New York