Mannesmann-Sumerbank Boru Endustrisi T.A.S. v. United States

OPINION GOLDBERG, Judge. In this action, the Court reviews two aspects of the Department of Commerce’s (“Commerce”) Certain Welded Carbon Steel Pipe and Tube and Welded Carbon Steel Line Pipe from, Turkey; Final Results and Partial Recission of. Countervailing Duty Administrative Reviews, 63 Fed.Reg. 18,885 (April 16, 1998)- (“Final Results”)- Specifically, Mannesmann-Sumerbank Boru Endustri-si T.A.S., Borusan Birlesik Boru Fabri-kalari A.S., and Borusan Ithalat Ve Da-gitim A.S. (“plaintiffs”) complain that Commerce (1) countervailed benefits under the Freight Rebate Program on the date of receipt instead of on the date of export; and (2) in calculating plaintiffs' ad valorem subsidy rate, erroneously excluded foreign exchange gains from the denominator of the equation. The Court exercises jurisdiction to review this motion for judgment on the agency record pursuant to 28 U.S.C. § 1581(c) (1994). The Court sustains in part and remands in part. I. BACKGROUND On March 7, 1986, Commerce published a countervailing duty order on certain welded carbon steel pipes and tubes and certain welded carbon steel line pipe from Turkey. Countervailing Duty Order, Certain Welded Carbon Steel Pipe and Tube Products from Turkey, 51 Fed.Reg. 7984 (Mar. 7, 1986). At the request of both domestic and Turkish producers and exporters, Commerce initiated a review of the order on April 24, 1997. 1 Initiation of Antidumping and Countervailing Administrative Reviews, 62 Fed.Reg. 19,988 (Apr. 24, 1997) (Certain Welded Carbon Steel Pipe and Tube). The review covered the period January 1, 1996 through December 31,1996. Commerce published' its preliminary results of the review on December 9, 1997. Certain Welded Carbon Steel Pipes and Tubes and Welded Carbon Steel Line Pipe from Turkey; Preliminary Results and Partial Recission of Countervailing Duty Administrative Reviews, 62 Fed.Reg. 64,- *1268 808 (Dec. 9, 1997) (“Preliminary Results” ). It published the Final Results on April 16, 1998. The facts relevant to the issues before the Court are summarized below. A. The Freight Rebate Program In October, 1993, the Government of the Republic of Turkey (“the Government”) instituted the Freight Rebate Program (“Program”) to promote exports of iron and steel products from Turkey. Under the Program, exporters were eligible to receive the Turkish lira (“TL”) equivalent of $30 or $50 per ton of merchandise exported; thirty percent (30%) was payable in cash, and the remainder in 1 and 2-year treasury bonds. An exporter was eligible to apply to the Government’s Central Bank for Program benefits whenever it completed an export transaction and payment was received from the customer. The Government did not specify what exchange rate would be used to calculate the TL equivalent of the U.S. dollar (“USD”). In December, 1994, the Government discontinued the Freight Rebate Program and announced that no benefits would be paid on shipments made after December 31, 1994. Several months later, in February, 1995, the Government declared that claims for benefits under the Program would be paid using the exchange rate prevailing on December 31, 1994. Subsequently, during the period of review, plaintiffs received cash payments 2 for export transactions completed in October, 1993 through December 31, 1994. See Preliminary Results, 62 Fed.Reg. at 64,811. In the Preliminary Results, Commerce determined that benefits paid under the Freight Rebate Program were countervail-able export subsidies. See id. Commerce also determined that exporters did not know the exact value of their benefits under the Program at the time they exported their merchandise. See id. Accordingly, citing its usual practice, Commerce countervailed benefits under the Program on the date the exporters received payment from the Government. See id. The practical effect of doing so is that Commerce countervailed cash payments associated with export transactions completed in October, 1993 through December, 1994 as part of an administrative review covering the period January 1 through December 31,1996. See id. In this appeal, plaintiffs contest Commerce’s decision to countervail Program benefits on the date plaintiffs received payment instead of on the date of export. B. “Kur Farki” Accounts To calculate plaintiffs’ ad valorem subsidy rate, Commerce divided the amount of benefit received (numerator) by the total value of exports to the United States (denominator). See Preliminary Results, 62 Fed.Reg. at 64,811. In calculating the denominator, Commerce excluded the amount listed in plaintiffs’ “kur farki,” or “exchange difference,” accounts. Pis.’ Br. in Supp. of Their Mot. for J. on the Agency R. Pursuant to Rule 56.2 (“Pis.’ Br.”), at 27. Commerce’s reasoning was that plaintiffs’ “kur farki” accounts do not represent actual sales revenue but rather foreign exchange differences, i.e. “income derived from fluctuations of the relative value of the dollar versus the TL,” and thus should not be included in the total value of exports to the United States. Final Results, 63 Fed.Reg. at 18,890. Plaintiffs contest Commerce’s treatment of the “kur farki” accounts. They explain that the kur farki account reflects the difference between the estimated TL amount recorded on the invoice date, when the sale is booked, and the TL amount actually received upon receipt of payment from the customer. 3 Depend *1269 ing on the date that the payment is received, the exchange difference can increase or decrease the invoice value. Therefore, the total amount in the kur farki account and the sales revenues account represents total actual income received from export sales transactions. Id. Plaintiffs complain that by excluding the amount in the “kur farki” accounts from the value of export sales, Commerce erroneously shrunk the denominator of the subsidy rate equation and in turn, artificially inflated the subsidy margin. See id. at 18,889. II. STANDARD OF REVIEW Commerce’s determination will be sustained if it is supported by substantial evidence on the record and is otherwise in accordance with law. See 19 U.S.C. § 1516a(b)(1)(B) (1994). III. DISCUSSION A. Commerce’s Determination to Countervail Benefits Under the Freight Rebate Program on the Date of Receipt Instead of on the Date of Export is Sustained. No statute or regulation governs the date on which Commerce must countervail benefits. Commerce describes its normal practice as a “cash flow” approach; that is, it countervails a benefit when that benefit affects the recipient’s cash flow. See Live Swine from Canada; Final Results of Countervailing Duty Administrative Reviews, 61 Fed.Reg. 52,408, 52,419 (Oct. 7, 1996) (“The Department’s practice is to countervail a benefit only when it affects the respondent’s cash flow.”); Fer-rochrome from South Africa; Final Results of Countervailing Duty Administrative Review, 56 Fed.Reg. 38,254, 38,255 (July 19, 1991) (“The Department’s standard practice is to recognize a benefit at the time it affects the cash flow of the recipient company.”); see also Countervailing Duties: Notice of Proposed Rule-making and Request for Public Comments (“Proposed Rule”), 54 Fed.Reg. 23,366, 23,375 (proposed May 31, 1989) (intended to be, but never codified at 19 C.F.R. pt. 355) (explaining “the Department’s general cash flow approach with regard to the timing of receipt of benefits”). According to Commerce, the cash flow effect usually occurs at the time of receipt. 4 See, e.g., Final Affirmative Countervailing Duty Determination: Certain Pasta (“Pasta”) from Turkey (“Pasta From Turkey”), 61 Fed.Reg. 30,366, 30,-367 (June 14, 1996) (deeming cash grants to have been bestowed at time of receipt); Final Affirmative Countervailing Duty Determination and Countervailing Duty Order; Lamb Meat from New Zealand, 50 Fed.Reg. 37,708, 37,714 (Sept. 17, 1985) (countervailing price support payments in year of receipt); see also Proposed Rule, 54 Fed.Reg. at 23,384 (proposing that “[t]he cash flow and economic effect of a benefit, normally occurs ... as a result of [a firm’s] receipt of the benefit”). There is an exception to this rule, however. As stated by Commerce, “where [a benefit] is provided as a percentage of the value of the exported merchandise on a shipment-by-shipment basis, and the exact amount of the countervailable subsidy is known at the time of export,” Commerce will “countervail an export subsidy on the date of *1270 export on an ‘earned basis’ rather than on the date the benefit is received.” Final Results, 63 Fed.Reg. at 18,888 (describing Commerce’s “long-standing practice”); see also Certain Iron-Metal Castings From, India: Final Results of Countervailing Duty Administrative Review (“Iron-Metal Castings From India”), 60 Fed.Reg. 44,-843, 44,844 (Aug. 29, 1995) (stating rule); Final Affirmative Countervailing Duty Determination and Countervailing Duty Order: Carbon Steel ButB-Weld Pipe Fittings From Thailand (“Pipe Fittings From Thailand”), 55 Fed.Reg. 1695, 1699 (Jan. 18, 1990) (discussing rule); “Proposed Rule” 54 Fed.Reg. at 23,384 (proposing that “in the case of an export benefit provided as a percentage of the value of the exported merchandise ... [the cash flow effect will be deemed to occur] on the date of export”). Plaintiffs dispute neither the existence nor the validity of Commerce’s practice as described above. Rather, plaintiffs assert that in countervailing Program benefits on the date of receipt instead of on the date of export, Commerce failed to conform to this established practice; and, plaintiffs argue, Commerce cannot do so without explanation. See Pls.’ Br., at 14-15 (citing Cinsa, S.A. de C.V. v. United States, 21 CIT-, -, 966 F.Supp. 1230, 1238 (1997) (“It is a well settled rule that an agency cannot arbitrarily change its methodology without explanation.”)). The issue before the Court, then, is whether Commerce’s Final Results are consistent with its established practice. The Court finds that Commerce’s determination is consistent, and thus that the Final Results are supported by substantial evidence and in accordance with law. Pursuant to its established practice, Commerce will countervail a benefit on the date of export only when an exporter satisfies three criteria: (1) the benefit is provided as a percentage of the value of the exported merchandise, or as in this case, based on export tonnage, see Preliminary Results, 62 Fed.Reg. at 64,811; (2) the benefit is provided on a shipment-by-shipment basis; and (3) the exact value of the benefit is known at the time of export. See Final Results, 63 Fed.Reg. at 18,888. In this case, Commerce found that plaintiffs failed to satisfy the third criteria; that is, plaintiffs did not know the exact amount of their benefit at the time of export. See id. The parameters of Commerce’s third criteria — that an exporter know the exact amount of the benefit at the time of export (hereinafter “knowledge requirement”)— are not well-defined. Specifically, Commerce has never explicitly indicated whether, to satisfy the requirement, an exporter must know the value of a benefit in its home currency, U.S. dollars, or either one. Indeed, many determinations in which Commerce has countervailed an export subsidy on the date of export (based on a finding that an exporter knew the exact value of the benefit on the date of export) do not discuss this issue. See, e.g., Iron-Metal Castings From India, 60 Fed. Reg. at 44,844 (exact amount of benefit known at the time of export because calculated as a percentage of FOB price and “paid upon export”); Pipe Fittings from Thailand, 55 Fed.Reg. at 1699 (exact amount of benefit known at time of export because tax certificates based on fixed percentage of FOB value of each export shipment and not dependent on company’s ultimate tax liability). In prior determinations, Commerce has deemed the knowledge requirement met when exporters knew the value of their benefit in their home currency at the time of export. For example, in Final Affirmative Countervailing Duty Determination; Cabron [sic] Steel Plate From Brazil (“Brazil /”), 48 Fed.Reg. 2568 (Jan. 20, 1983), one of the determinations on which plaintiffs rely, Commerce countervailed export credit premiums on the date of export. The amount of credits for which an exporter was eligible was based on a percentage of the value of exports (in dollars), as converted into eruzerios using the exchange rate prevailing on the date of shipment. Based on this, Commerce deter *1271 mined that exporters knew the exact value of their benefit, in cruzerios, at the time of shipment. See id. at 2578; accord Final Affirmative Countervailing Duty Determination; Prestressed Concrete Steel Wire Strand From Brazil, 48 Fed.Reg. 4516, 4521 (Feb. 1, 1983) (“Brazil II”) (same); see also Final Negative Countervailing Duty Determination; Paint Filters and Strainers From Brazil, 52 Fed.Reg. 19,-184, 19,186 (May 21, 1987) (noting that Commerce has “consistently calculated the benefit under this [IPI export credit premium] program based on the date the premium was earned, not on the date it was received”). In comparison, Commerce deemed the knowledge requirement met for a different program examined during this review, the Export Performance Credit Program (“EPCP”), because exporters knew the value of those benefits in U.S. dollars at the time of export. Under the EPCP, “[ejxporters ... receive[d] a percentage of the U.S. dollar value of their exports in TLs based on the foreign exchange rate prevailing at the time of payment.” Final Results, 63 Fed.Reg. at 18,888. Commerce explained that “although at the time of receipt the exporters received more TLs than they would have been paid upon exportation, because the benefit was tied to the U.S. dollar, the value of the TL amount remained the same in U.S. dollar terms.” Id. Thus, because exporters knew the value of the benefit under the EPCP at the time of export in U.S. dollars, Commerce countervailed EPCP benefits on the date of export. See id. The Court need not decide at this time whether, in order to satisfy Commerce’s knowledge requirement in a manner consistent with past practice, exporters must know the value of a benefit in their home currency, U.S. dollars, or either one. In this case, plaintiffs did not know the value of their benefit at the time of export in either Turkish lira or U.S. dollars. Plaintiffs did not know, at the time of export, the amount of TL they would receive because Program benefits were stated in USD, not TL. Likewise, plaintiffs did not know, at the time of export, the value of their benefits in USD. Although Program benefits were stated in USD, the Turkish Government did not specify the exchange rate that would be used to convert USD to TL, and thus plaintiffs could not be sure they would receive the full TL equivalent of $30 or $50. Plaintiffs’ arguments to the contrary are unavailing. Significantly, the language plaintiffs use to argue their case belies their claim that they knew the exact value of their benefit at the time of export. First, plaintiffs argue they satisfy Commerce’s knowledge criteria because they “had every reason to believe that the full value of the benefit [under the Program] would be either $30 or $50 per ton.” Pis.’ Br., at 18. They further explain they “did not know ... that the Government of Turkey would later renege on its commitment to pay the full dollar value of the benefit promised under the program.” Id. What plaintiffs believed or assumed the value of their benefits would be, however, is not the same as the proposition that plaintiffs knew the exact value of the benefit at the time of export, especially considering Turkey’s high rate of inflation at the time. As previously emphasized, “the GRT made no initial commitment to use the exchange rate prevailing on the date payment was made.” Def.’s Mem. In Opp’n to Pis.’ Mot. for J. Upon the Agency R. Pursuant to Rule 56.2 (“Def.’s Br.”), at 19 (emphasis added). Without this critical information, exporters assumed, but did not know, that the government would choose an exchange rate that guaranteed them the full dollar value of their benefit. Plaintiffs further argue they knew the exact value of the benefit at the time of export because it “was fixed for a period of time in U.S. dollars before being converted into local currency” and was therefore “protected against inflation until December 31, 1994.” Pis.’ Br., at 19-20. This argument, too, misses the mark. Benefits under the Freight Rebate Program were *1272 merely stated in U.S. dollars; they were not tied or fixed to the U.S. dollar per se. An example of a Government program tied to the U.S. dollar is the aforementioned Export Performance Credit Program. Under the EPCP, exporters knew they would receive the full TL equivalent of a given dollar amount because benefits were to be paid using the exchange rate prevailing at the time of payment, 5 In contrast, under the Freight Rebate Program, exporters did not know for certain the exchange rate the government would use to convert the rebate payments, as stated in U.S. dollars, to TLs. Consequently, it was impossible for exporters to know the exact value of their benefit at the time of export. Thus, it was consistent for Commerce to countervail benefits under the EPCP on the date of export, but to countervail benefits under the Program on the date of receipt. See, e.g., Certain Welded Carbon Steel Pipes and Tubes and Welded Carbon Steel Line Pipe From Turkey; Final Results of Countervailing Duty Administrative Reviews, 62 Fed. Reg. 43,984, 43,990 (Aug. 18, 1997) (finding exporters did not know the amount of benefit at the time of export because exchange rate not specified and high rate of inflation; countervailing benefit at the time of receipt); Pasta from Turkey, 61 Fed.Reg. at 30,367 (countervailing benefits under the Pasta Export Grants program on the date cash was received because, as explained in Preliminary Affirmative Countervailing Duty Determination: Certain Pasta (‘Pasta”) from Turkey, 60 Fed.Reg. 53,747, 53,749 (Oct. 17, 1995), “the amount the exporter will actually receive in TL is not certain until the time of receipt because it is subject to fluctuations in the exchange rate”). Finally, plaintiffs’ reliance on countervailing duty investigations and orders involving Brazil (between 1983 and 1987) and Mexico (during the 1980s) is misplaced. Plaintiffs claim those determinations show that Commerce has countervailed benefits on the date of export even when the benefit was denominated in local currency, the economy was characterized by high inflation, 6 and there was a long delay between the time of export and the receipt of the benefit. See Pis.’ Br., at 15-16. Those cases are not relevant, however, to the question presented by this case; that is, whether, consistent with Commerce’s prior practice, an exporter can “know” the exact value of a benefit when no exchange rate has been specified and the rate of inflation is high. For example, a majority of the Brazilian determinations cited by plaintiffs address a much different issue, that of whether a delay between the date of export and the date of receipt during which inflation may *1273 have eroded the real value of the benefit qualifies an exporter for a deduction for “any loss in the value of the countervaila-ble subsidy resulting from its deferred receipt, if the deferral is mandated by Government order.” 19 U.S.C. § 1677(6)(B) (1994). In response to that question, Commerce found that any delay in an exporter’s receipt of the benefit was “frequently the result of a company’s application for it,” not a government mandate. Brazil I, 48 Fed.Reg. at 2578. Notably, the amount of the benefit in those cases was calculated using the exchange rate prevailing on the date of shipment. Thus, Commerce determined that the amount of the benefit was solidified on the date of shipment, “even if the application is made months later and the exchange rate has changed substantially.” Id.; accord Final Affirmative Countervailing Duty Determination; Cash-Iron Pipe Fittings from Brazil, 50 Fed. Reg. 8755, 8759 (Mar. 5, 1985); Final Affirmative Countervailing Duty Determination; Oil Country Tubular Goods From Brazil, 49 Fed.Reg. 46,570, 46,575 (Nov. 27, 1984); Certain Carbon Steel Products From Brazil; Final Affirmative Countervailing Duty Determinations, 49 Fed.Reg. 17,988, 17,996 (Apr. 26, 1984); Final Affirmative Countervailing Duty Determinations; Certain Stainless Steel Products From Brazil, 48 Fed.Reg. 21,610, 21,612 (May 13, 1983); Brazil II, 48 Fed.Reg. at 4521. In conclusion, Commerce’s determination to countervail benefits under the Freight Rebate Program on the date of receipt instead of on the date of export was consistent with its long-standing practice. The Court sustains that portion of the Final Results as supported by substantial evidence and in accordance with law. . B. Commerce’s Determination to Exclude Plaintiffs’ “Kur Farki” Accounts From the Denominator of the Subsidy Equation is Remanded. 19 U.S.C. § 1671 mandates that the countervailing duty rate imposed on merchandise imported into the United States be “equal to the amount of the net countervailable subsidy.” 19 U.S.C. § 1671(a) (1994). The statute does not provide, however, a methodology for calculating “net countervailable subsidy,” or more particularly, the denominator of the subsidy equation. Therefore, in the absence of statutory direction, “Commerce’s methods will be upheld if they reflect a reasonable means of assessing the benefits and if there is support in the record as a whole for its determination.” SSAB Svenskt Staal AB v. United States, 15 CIT 202, 206, 764 F.Supp. 650, 655 (1991) (internal quotation marks omitted). In this case, Commerce calculated plaintiffs’ “ad valorem subsidy rate by dividing the benefit [numerator] by the firm’s total exports to the United States [denominator], adjusted for foreign exchange differences.” Preliminary Results, 62 Fed.Reg. at 64,811. “[W]here the foreign exchange difference was a positive amount,” Commerce deducted it from the denominator. Final Results, 63 Fed.Reg. at 18,890. Plaintiffs assert that Commerce erred in deducting the foreign exchange gain in plaintiffs’ “kur farki” accounts from the denominator of the subsidy equation for two reasons: (1) because the foreign exchange gains in plaintiffs’ “kur farki” accounts are actual sales revenue; and (2) because Commerce acted contrary to prior practice. The Court will address each argument in turn. First, plaintiffs argue that the amount in the “kur farki” accounts is actual sales revenue and therefore by definition, must be included in the total value of exports to the United States. According to plaintiffs, the amount Commerce chose to use in the denominator — invoice amount — is purely arbitrary, and does not reflect the actual value of total sales. Plaintiffs explain that the difference between what is booked on the date of invoice and what is booked on the date of receipt of payment is merely a reflection of the fact that, under an accrual-based accounting system, sales are first booked in the company’s accounts on or about the date of invoice, prior to the time of payment. *1274 Since the sales must be booked in Turkish lira, they must be converted into Turkish lira. However, the exchange rate that is used in Turkey is entirely artificial; it is picked according to an accounting convention. The actual revenues from the sale in local currency cannot be established until payment is made and the funds are actually exchanged into local currency. Pis.’ Reply Br. in Supp. of Their Mot. for J. on the Agency R. Pursuant to Rule 56.2 (“Pis.’ Reply Br.”), at 10. In support of their claim that the “kur farki” accounts qualify as actual sales revenue, plaintiffs point to Turkish accounting principles. They claim that their three-step method of booking sales is required under Turkey’s generally accepted accounting principles (GAAP) and that under Turkey’s standardized accounting plan, gross sales include “exchange rate differences related to export sales within the relevant period.” Pis.’ Br., at 27 n. 55. Commerce, on the other hand, argues that plaintiffs’ “kur farki” accounts are not actual sales revenue. It emphasizes that plaintiffs negotiated their export sales price in USD, not TL. See, Public Doc. App. to Accompany the Non-Confidential Version of Pis.’ Br. in Supp. of Their Mot. for J. on the Agency R. Pursuant to Rule 56.2 (“Pis.’ App.”), Ex. 8 (stating that prices were negotiated in U.S. dollars); Def.’s Br., at 30. Further, Commerce notes, plaintiffs were paid in USD, not TL. See Final Results, 63 Fed.Reg. at 18,890; Def.’s Br., at 32. Thus, Commerce reasons, because there was no change in the negotiated price paid by plaintiffs’ customers, the difference between the amount of TL invoiced and the amount of TL received is due to currency fluctuation, not to an increase in sales revenue. See Final Results, 63 Fed.Reg. at 18,890; Def.’s Br., at 32. Moreover, Commerce argues, “[s]uch foreign exchange difference becomes particularly significant in Turkey’s highly inflationary economy.” Final Results, 63 Fed.Reg. at 18,890. In sum, in Commerce’s view, that plaintiffs received more TL in payment that the amount originally booked is evidence of the devaluation of the TL relative to the USD in connection with high inflation, and is not the result of an increase in sales revenue. See Final Results, 63 Fed.Reg. at 18,890; Def.’s Br., at 31. Commerce’s analysis is insufficient, however. As an initial matter, the Court recognizes the intuitive force of plaintiffs’ arguments. From the U.S. customer’s point of view, the USD amount paid is the value of any sale. From the exporter’s perspective, however, the ultimate value of any sale is the total amount of TL received. More importantly, Commerce fails to explain why plaintiffs’ record-keeping and accounting methods do “not detract from” its finding that plaintiffs’ “kur farki” accounts are not actual sales revenue. Final Results, 63 Fed.Reg. at 18,890. In this case, Commerce asked plaintiffs to report the “actual value” of sales as recorded in plaintiffs’ accounting records. See Pis.’ App., Ex. 5. As Commerce has previously noted, an exporter’s own records may be the most accurate representation of its finances. See, e.g., Final Affirmative Countervailing Duty Determination: Certain Steel Products From Austria, 58 Fed. Reg. 37,217, 37,237 (July 9, 1993) (finding “that it is more appropriate to use respondents’ sales value as recorded in their financial statement and accounts in the denominator when calculating the ad valorem subsidy rate” when doing so “would provide a more accurate representation of respondents’ selling practices.”). And in other types of investigations, Commerce requires respondents to report data consistent with their home country GAAP. See Asociacion Colombiana de Exportadores de Flores v. United States, 22 CIT-, -, 40 F.Supp.2d 466, 473 n. 11 (1999) (noting that in antidumping cases Commerce normally “requires respondents to report preproduction expenses consistent with their home country generally accepted accounting principles (‘GAAP’), which typically are reflected in respondents [sic] ordinary books and records”); Cultivos *1275 Miramonte S.A. v. United States, 21 CIT -, -, 980 F.Supp. 1268, 1276 (1997) (same); see also Thai Pineapple Pub. Co. v. United States, 20 CIT 1312, 1318, 946 F.Supp. 11, 18 (1996) (explaining that Commerce’s rationale for adhering to “an individual firm’s recording of costs in accordance with the GAAP of its home country” in the antidumping context is that “normal accounting practices should, in most circumstances, provide an objective standard by which to measure costs”). Yet, in this case, Commerce rejected Turkish accounting methods and, at the same time, plaintiffs’ financial records kept in accordance with those conventions. See Final Results, 63 Fed.Reg. at 18,890. Commerce’s rationale for doing so is that a “lack of clearly defined commercial accounting principles and the predominance of tax law mean that Turkish law should be treated with extreme caution, and international accounting standards are preferred.” Id. (citing Price Waterhouse, Doing Business in Turkey, Chapter 11 (1992)). It is true that Commerce is not required to accept Turkish GAAP or Turkey’s standardized accounting plan. See Ad Hoc Comm. of Florida Producers of Gray Portland Cement v. U.S., 22 CIT -, -, 25 F.Supp.2d 352, 363 (1998) (noting that in the antidumping context, “Commerce is not required to follow GAAP in all cases.”); Aimcor, Alabama Silicon, Inc. v. United States, 18 CIT 1117, 1122, 871 F.Supp. 447, 452 (1994) (“Plaintiffs can provide no authority requiring Commerce to consistently apply GAAP in all aspects of countervailing duty investigations.”). If Commerce rejects those conventions, however, it must provide an adequate explanation for why relying on them would be dis-tortive. See Floral Trade Council v. United States, 41 F.Supp.2d 319, 336-37 (CIT1999) (stating that in the antidumping context Commerce must use home country GAAP unless it explains how respondent’s records do not adequately reflect the costs of production and sale). In this case, Commerce failed to adequately explain its rejection of Turkish GAAP and plaintiffs’ accounting methods. Significantly, Commerce relies on a single source: Doing Business in Turkey. At the time Commerce initiated its review, this source was already five years old. Moreover, the passage on which Commerce relies speaks to audit reports. See Pis.’ Br., at Ex. E. Commerce points to nothing in its source that calls into question the reliability of Turkish GAAP’s three-step accounting method or the standardized accounting plan’s inclusion of exchange rate differences in gross sales. Finally, although Commerce states that “international accounting standards are preferred” to Turkish accounting standards, it provides no evidence that international standards dictate that foreign exchange gains be excluded from total sales figures or counsel against the three-step accounting method. Moreover, Doing Business in Turkey, on which Commerce relies, notes that “[vjoluntary audits are requested usually by multinational companies.” Pis.’ Br., at Ex. E. And plaintiffs claim that “the accounts at both Borusan and Mannesmann are audited in accordance with IAS [International Accounting Standards].” Pis.’ Reply Br., at 12. The Court now turns to plaintiffs’ second argument. Plaintiffs claim that “Commerce normally adds exchange gains and losses on receivables to the total book value of sales in determining the total sales denominator used to compute the ad valo-rem value of the benefit.” Pis.’ Br., at 29. In support of their claim, plaintiffs rely on Pasta From Turkey, 61 Fed.Reg. 30,366. Plaintiffs also cite a number of other cases wherein “Commerce included foreign exchange gains and losses on export sales in the sales denominator” even though that fact is “not specifically reflected in the Federal Register.” 7 Pis.’ Br., at 29 n. 62. *1276 On this issue Commerce’s determination is again insufficient. Commerce fails to explain in the Final Results or the government’s brief why its methodology for calculating the denominator of the subsidy equation in this case differs from prior determinations. The Court has found at least two determinations in which Commerce has included foreign exchange gains and losses in the denominator. For example, to calculate the countervailable subsidy for the Pasta Export Grants in Pasta From Turkey, Commerce included respondent’s “Other Income” account in the total sales denominator. See Pasta From Turkey, 61 Fed. Reg. at 30,372. Petitioners in that case complained that the “Other Income” account contained subsidies which were not properly part of the denominator. See id. In response to that complaint, however, Commerce stated that “only the amount of foreign exchange gains from Maktas’ [sic] ‘Other Sales’ is included in the Maktas denominators used to calculate the benefit of the used subsidy programs.” Id. Thus, as the Court reads Pasta From Turkey, Commerce included foreign exchange gains associated with certain sales in the denominator of the subsidy equation. Commerce included foreign exchange gains in the denominator in at least one other determination as well. In Final Affirmative Countervailing Duty Determination; Brass Sheet and Strip From Brazil {“Brass Sheet”), petitioners complained that respondents inflated the denominator of the subsidy equation by including inflation as revenue. See 51 Fed.Reg. 40,837, 40,841 (Nov. 10, 1986). Commerce disagreed, however, and clarified that respondents’ sales figures included, not an adjustment for inflation, but an adjustment for foreign exchange gains. It explained that [respondents’ export sales are adjusted for an exchange gain resulting from the lag in fixing the dollar-cruzeiro exchange rate between the date of export and the date of receipt of funds. Continued devaluation of the Brazilian cruzeiro against the dollar increases the number of cruzeiros per dollar between the date of export and the date the exchange contract is concluded. This is standard accounting practice for foreign exchange transactions. Id. This passage explicitly demonstrates that Commerce accepted the inclusion of foreign exchange gains in the denominator of the subsidy equation. And, notably, Commerce described the practice of adjusting sales figures to include foreign exchange gains as “standard” accounting procedure. See id. Moreover, in Brass Sheet, Commerce accepted an adjustment to the export sales denominator for foreign exchange gains due to the “continued devaluation of the Brazilian cruzeiro against the dollar.” Brass Sheet, 51 Fed.Reg. at 40,841. As the government itself notes in its brief in this case, “extreme inflation ... is characterized by a devaluation of the currency.” Def.’s Br., at 31 (citing North Star Steel Ohio, A Division of North Star Steel Co. v. United States, 17 CIT 459, 467, 824 F.Supp. 1074, 1081 n. 7 (1993)). Thus, in Brass Sheet, Commerce accepted the inclusion of exchange gains in the denominator precisely because of inflation. Yet, a “significant” factor in Commerce’s decision to exclude the “kur farki” accounts from the denominator of the subsidy equation in the determination before the Court was “Turkey’s highly inflationary economy.” Final Results, 63 Fed.Reg. at 18,890. This is inconsistent. Together, Pasta From Turkey and Brass Sheet suggest that Commerce has, on at least two occasions in the past, included foreign exchange gains in the total sales denominator of the subsidy equation. “Commerce can reach different determinations in separate administrative reviews but it must employ the same methodology or give reasons for changing its practice.” Cinsa, 21 CIT at -, 966 F.Supp. at 1238. Commerce is not excused from this “well settled rule” of administrative law, id., simply because it also excluded foreign exchange gains from the denominator in *1277 one previous administrative review of this order. 8 Thus, on remand, Commerce must include plaintiffs’ foreign exchange gains in the denominator of the subsidy margin or provide an adequate explanation of how this case differs from prior determinations. If Commerce takes the latter course of action, it must also explain why Turkish GAAP and plaintiffs’ accounting methods are unreliable or distortive. Depending on Commerce’s remand results, it may also need to consider whether the denominator and numerator of its subsidy equation “match.” In order for Commerce to calculate an accurate subsidy margin, the numerator and denominator must both take into account factors affecting value such as, in this case, inflation and foreign exchange movements. Cf. Certain Iron-Metal Castings From India; Final Results of Countervailing Duty Administrative Review, 62 Fed.Reg. 32,297, 32,302 (June 13, 1997) (explaining that “it is imperative that both the numerator (the benefit) and denominator (the universe of sales to which the benefit applies) used in our calculation of a subsidy reflect the same universe of goods. Otherwise the rate calculated will either over- or understate the subsidy attributable to the subject merchandise.”). Commerce claims in the present action that if it were to include the “kur farki” accounts in the denominator, it “would be equivalent to indexing export sales for inflation and thus, would inflate the denominator while the program benefits (the numerator) would remain un-indexed.” 9 Def.’s Br., at 31; Preliminary Results, 62 Fed.Reg. at 64,809. At the same time, plaintiffs argue that if Commerce “insists on reducing the total export value by the foreign exchange difference, then it must compute and deduct from the numerator (the countervailable benefit) the foreign exchange difference included in the benefit.” Final Results, 63 Fed.Reg. at 18,890. IY. CONCLUSION For the foregoing reasons, the Court sustains in part and remands in part. A separate Order will be entered accordingly. ORDER Upon consideration of plaintiffs’ motion for judgment upon the agency record, plaintiffs’ briefs in support of said motion, and defendant’s brief in opposition thereto; upon all other papers and proceedings had herein, and upon due deliberation; it is hereby ORDERED that the United States Department of Commerce’s (“Commerce”) determination, in Certain Welded Carbon Steel Pipe and Tube and Welded Carbon Steel Line Pipe from Turkey; Final Results and Partial Redssion of Countervailing Duty Administrative Reviews, 63 Fed.Reg. 18,885 (April 16, 1998) {‘Final Results”), to countervail benefits under the Freight Rebate Program on the date of receipt is sustained; ORDERED that Commerce’s determination, in the Final Results, to exclude plaintiffs’ “kur farki” accounts from the denominator of the subsidy equation is remanded; and that upon remand, Commerce shall include plaintiffs’ foreign exchange gains in the denominator of the subsidy margin or (1) provide an adequate explanation of how this case differs from prior determinations, and (2) explain why Turkish GAAP and plaintiffs’ accounting methods are unreliable or distortive; *1278 ORDERED that Commerce shall, within sixty (60) days of the date of this Order, issue a remand determination in accordance with the instructions provided herein; ORDERED that the parties may, within twenty (20) days of the date on which Commerce issues its remand determination, submit memoranda addressing Commerce’s remand determination, not to exceed twenty (20) pages in length; and it is further ORDERED that the parties may, within twenty (20) days of the date on which memoranda addressing Commerce’s remand determinations are filed, submit response briefs, not to exceed twenty (20) pages in length. Footnotes [1] Commerce initiated this review after December 31, 1994, and thus the applicable statutory provisions are those that existed on January 1, 1995, the effective date of the amendmenls made by the Uruguay Round Agreements Act, Pub.L. No. 103-465, 108 Stat. 4809 (1994). See Torrington Co. v. United States, 68 F.3d 1347, 1352 (Fed.Cir.1995). [2] Plaintiffs also received bonds and interest payments on bonds during the period of review. Commerce determined, however, that "the benefits from the bonds are bestowed on the date of maturity.” Preliminary Results, 62 Fed.Reg. at 64,811. [3] As the Court understands plaintiffs’ explanation, every transaction involves three steps ("three-step accounting method”). First, an exporter submits a Customs Export Declaration (GCB) to the Turkish Customs Service. Second, about two weeks later, the exporter *1269 issues an invoice. At the same time, the exporter books the sale in TL using the "book rate," the exchange rate in effect on the date of the GCB. Third, the exporter receives payment on the invoice in U.S. dollars and converts the USD to TL; the exporter records the difference between the TL amount booked on the invoice date and the amount of local currency actually received as payment in the "kur farki” account. See Pis.' Br., at 11-12. [4] In 1998, Commerce promulgated 19 C.F.R. § 351.524 (1999). Although it post-dates this review, it is noteworthy in that it states that "[t]he Secretary will allocate (expense) a recurring benefit to the year in which the benefit is received.” Id. [5] Notably, plaintiffs do not contest Commerce's assertion in the Final Results that the Government of Turkey specified that benefits under the Export Performance Credit Program would be paid using the exchange rate prevailing at the time of payment. Rather, plaintiffs assert that, despite the fact that the Government did not malte the same commitment with respect to the Freight Rebate Program, there is no record evidence that they believed the Government would pay differently under the programs. See Pis.’ Br., at 22 ("At the time of export, the exporters had every reason to believe that the exchange rate used to convert benefits into local currency would, under both programs, ensure that the full dollar value of the benefit was preserved. There is no evidence on the administrative record that exporters believed or had any forewarning that such would not be the case [with the Freight Rebate Program].”). [6] In discussing inflation, plaintiffs reference the International Financial Statistics Yearbook. Pis.’ Br., at 16 n. 38 & Attach. B. Defendant, by way of motion dated November 6, 1998, asks the Court to strike this data pursuant to CIT Rule 81(m), as well as Memorandum of 3/6/96 from Kristin H. Mowry to Susan H. Kuhbach (Pis.’ Attach. C) and Financial Accounting Standards Board, Original Pronouncements, (John Wiley & Sons, Inc. 1996) (Pis.' Attach. F), because the documents are not part of the administrative record. It is unnecessary for the Court to rule on defendant’s motion to strike; even if plaintiffs' submissions were part of the administrative record, they would not alter the Court's decision in this case. Cf. AK Steel Corp. v. United States, 21 CIT-,-, 988 F.Supp. 594, 608 (1997) (refusing to rule on motion to strike because the Court would still rule in favor of Commerce on all issues). [7] The Court has reviewed these determinations, and found no discussion of the issue at hand. [8] Certain Welded Carbon Steel Pipes and Tubes and Welded Carbon Steel Line Pipe From Turkey; Final Results of Countervailing Duty Administrative Reviews, 63 Fed.Reg. 43,984 (Aug. 18, 1997). [9] The Court does not necessarily accept the veracity of this statement, especially considering Brass Sheet, in which Commerce made clear that including foreign exchange gains in the export sales denominator is different from indexing that number for inflation. See Brass Sheet, 51 Fed.Reg. at 40,841.