Nucor Corp. v. United States

612 F. Supp. 2d 1264, 33 Ct. Int'l Trade 207, 33 C.I.T. 207, 31 I.T.R.D. (BNA) 1166, 2009 Ct. Intl. Trade LEXIS 32

Opinion

OPINION

RIDGWAY, Judge.

These consolidated actions are before the court on cross-motions for- judgment on the agency record. Domestic steel manufacturers Nucor Corporation, Gerdau AmeriSteel Corporation, and Commercial Metals Company (collectively, the “Domestic Producers”) and ICDAS Celik Enerji Tersane ve Ulasim Sanayi, A.S. (“ICDAS”) — a Turkish producer/exporter of the subject merchandise — separately challenge various aspects of the final results of the U.S. Department of Commerce’s seventh administrative review of the anti-dumping duty order on Certain Steel Concrete Reinforcing Bars From Turkey. See generally Certain Steel Concrete Reinforcing Bars From Turkey; Final Results, Rescission of Antidumping Duty Administrative Review in Part, and Determination To Revoke in Part, 70 Fed.Reg. 67,665 (Nov. 8, 2005) (“Final Results”). Also here in dispute are the results of a voluntary remand to Commerce on the issue of the date of sale for ICDAS’ U.S. sales, for use in Commerce’s antidumping margin calculations. See Final Results of Redetermination Pursuant to Court Remand (“Remand Results”).

In particular, the Domestic Producers contend that Commerce erred in the Final Results by classifying ICDAS’ sales through its U.S. affiliate as Export Price (“EP”) sales, rather than Constructed Export Price (“CEP”) sales. See Memorandum in Support of Plaintiffs’ Motion for Judgment on the Agency Record (“Domestic Producers Brief’); Plaintiffs’ Reply Brief (“Domestic Producers Reply Brief’) at 1-8. 1 The Domestic Producers further assert that ICDAS’ attacks on the Final Results are unfounded, and that the Final Results therefore should be sustained in all other respects — with one major exception. See Response Brief of the Domestic Producers (“Domestic Producers Response Brief’). Specifically, the Domestic Producers assert that the Final Results erred in using contract date as the date of sale for ICDAS’ U.S. sales, and that the Remand Results — where Commerce reversed itself — are correct. See Domestic Producers Brief at 2 n. 1; Domestic Producers Reply Brief at 1, 8-15.

For its part, ICDAS challenges four aspects of the Final Results: (1) Commerce’s disallowance of a start-up adjustment for ICDAS’ Biga melt shop; (2) Commerce’s decision to treat ICDAS’ foreign exchange gains within the category of “financial expenses,” and to cap ICDAS’ total financial expenses at zero; (3) Commerce’s use of the average cost of manufacturing for the entire period of review (“POR”), rather than ICDAS’ quarterly costs, in the agency’s “sales below cost” analysis; and (4) Commerce’s use of the date of entry, rather than the date of sale, to define ICDAS’ universe of sales. See Plaintiff ICDAS’ Memorandum in Support of its Motion for Judgment on the Agency Record Pursuant to Rule 56.2 (“ICDAS Brief’); Plaintiff ICDAS’ Reply Brief in Support of Its Motion for Judgment on the Agency Record Pursuant to Rule 56.2 (“ICDAS Reply Brief’). In addition, ICDAS contests Commerce’s decisioh in the Remand Results to use invoice date as the date of sale for ICDAS’ U.S. sales, rather than using contract date (as the agency did in the Final Results). See DefendanNIntervenor ICDAS’ Memorandum in Opposition to Plaintiffs’ Motion for Judgment on the Agency Record at 1-3, 5-30 (“ICDAS Response Brief’); DefendanNIntervenor ICDAS’ Supplemental Reply Brief Regarding the Date of Sale Issue (“ICDAS Supp. Reply Brief’). ICDAS maintains that Commerce properly treated all of ICDAS’ U.S. sales as Export Price (“EP”) sales, rather than Constructed Export Price (“CEP”) sales, and therefore opposes the Domestic Producers’ Motion for Judgment on the Agency Record. See ICDAS Response Brief at 1, 3-4, 30-40.

The Government maintains that the Final Results should be sustained in all respects, save three. See Defendant’s Response to Plaintiffs’ and Defendant-Intervenor’s Motions for Judgment Upon the Agency Record (“Def. Response Brief’). First, the Government requests that two issues be remanded to Commerce for further consideration — specifically, Commerce’s use of the POR average cost of manufacturing (rather than ICDAS’ quarterly costs) in the agency’s “sales below cost” analysis, and Commerce’s use of the date of entry (rather than the date of sale) to define ICDAS’ universe of sales. See Def. Response Brief at 1-3, 8-9, 11-12, 28-29, 36. In addition, the Government asserts that, as to the issue of the date of sale for ICDAS’ U.S. sales, the Remand Results (which used invoice date as the date of sale) — rather than the Final Results (which used contract date) — should be sustained. See Defendant’s Response to Defendanb-Intervenor’s Memorandum in Opposition to Plaintiffs Motion for Judgment on the Agency Record (“Def. Supp. Response Brief’).

Jurisdiction lies under 28 U.S.C. § 1581(c) (2000). 2 For the reasons set forth below, the Domestic Producers’ Motion for Judgment on the Agency Record challenging Commerce’s decision to treat sales made through ICDAS’ U.S. sales affiliate as EP sales must be denied. ICDAS’ Motion for Judgment on the Agency Record must similarly be denied as to ICDAS’ claims that Commerce improperly denied ICDAS’ request for a startup adjustment, and that Commerce erred in its treatment of ICDAS’ foreign exchange gains as well as in its decision to' cap ICDAS’ total financial expenses at zero. On the other hand, ICDAS’ Motion for Judgment on the Agency Record is granted as to ICDAS’ challenges to Commerce’s use of invoice date (rather than contract date) as the date of sale for ICDAS’ U.S. sales, Commerce’s use of the POR average cost of manufacturing (rather than ICDAS’ quarterly costs) in the agency’s “sales below cost” analysis, and Commerce’s use of the date of entry (rather than the date of sale) to define ICDAS’ universe of sales; and this matter is remanded to the Department of Commerce for further proceedings not inconsistent with this opinion.

I. Standard of Review

In reviewing a challenge to a final determination by the Commerce Department in an antidumping administrative review, the court must hold unlawful any agency determination, finding, or conclusion that is found to be “unsupported by substantial evidence on the record, or otherwise not in accordance with law.” 19 U.S.C. § 1516a(b)(l)(B)(i); see also Elkem Metals Co. v. United States, 468 F.3d 795, 800 (Fed.Cir.2006). Substantial evidence is “more than a mere scintilla”; rather, it is “such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.” Universal Camera Corp. v. Nat’l Labor Relations Bd., 340 U.S. 474, 477, 71 S.Ct. 456, 95 L.Ed. 456 (1951) (quoting Consol. Edison Co. v. Nat’l Labor Relations Bd., 305 U.S. 197, 229, 59 S.Ct. 206, 83 L.Ed. 126 (1938)); see also Nippon Steel Corp. v. United States, 337 F.3d 1373, 1379 (Fed.Cir.2003) (same).

Moreover, “the substantiality of evidence must take into account whatever in the record fairly detracts from its weight,” including “contradictory evidence or evidence from which conflicting inferences could be drawn.”' Suramerica de Aleaciones Laminadas, C.A. v. United States, 44 F.3d 978, 985 (Fed.Cir.1994) (quoting Universal Camera, 340 U.S. at 487-88, 71 S.Ct. 456). On the other hand, the mere fact that “it [may be] possible to draw two inconsistent conclusions from evidence in the record ... does not prevent Commerce’s determination from being supported by substantial evidence.” Am. Silicon Techs. v. United States, 261 F.3d 1371, 1376 (Fed.Cir.2001); see also Consolo v. Federal Maritime Commission, 383 U.S. 607, 620, 86 S.Ct. 1018, 16 L.Ed.2d 131 (1966) (same).

II. Background

In April 1997, the Department of Commerce issued an antidumping order covering rebar from Turkey. See Antidumping Duty Order: Certain Steel Concrete Reinforcing Bars From Turkey, 62 Fed.Reg. 18,748 (April 17, 1997). Subsequently, in every annual administrative review that Commerce has conducted for ICDAS since 1999 — including three consecutive administrative reviews, covering the periods April 1, 2001 through March 31, 2004 — Commerce consistently found that the dumping margin for ICDAS’ U.S. sales was zero or de minimis 3 (at least until the Remand Results here in dispute were issued). 4

The administrative review which is the subject of this action — the seventh such review — began in April 2004, when Commerce gave notice of the opportunity to request a review of the antidumping order on rebar from Turkey, for the period April 1, 2003 through March 31, 2004. See generally Antidumping or Countervailing Duty Order, Finding, or Suspended Investigation; Opportunity To Request Administrative Review, 69 Fed.Reg. 17,129 (April 1, 2004). At the request of both the Domestic Producers and ICDAS, inter alia, Commerce initiated an administrative review the following month. See generally Initiation of Antidumping and Countervailing Duty Administrative Reviews and Request for Revocation in Part, 69 Fed.Reg. 30,282 (May 27, 2004).

In the Preliminary Results of the administrative review, Commerce calculated a margin of 0.47% for ICDAS. In light of that de minimis margin, and the company’s record of zero or de minimis margins in the two prior administrative reviews, the agency also announced its intention to revoke the antidumping order as to ICDAS. See generally Certain Steel Concrete Reinforcing Bars from Turkey; Preliminary Results and Partial Rescission of Antidumping Duty Administrative Review and Notice of Intent to Revoke in Part, 70 Fed.Reg. 23,990, 23,991, 23,995 (May 6, 2005) (“Preliminary Results”); see also Gerdau Ameristeel Corp. v. United States, 519 F.3d 1336, 1337-38 (Fed.Cir.2008) (summarizing history of de minimis findings in administrative reviews of ICDAS, leading to Commerce’s determination to revoke antidumping order as to ICDAS).

Following briefing and oral argument by the parties before the agency, Commerce published the Final Results of the administrative review. See generally Certain Steel Concrete Reinforcing Bars From Turkey; Final Results, Rescission of Anti-dumping Duty Administrative Review in Part, and Determination To Revoke in Part, 70 Fed.Reg. 67,665 (Nov. 8, 2005) (“Final Results”); see also Issues and Decision Memorandum for the Antidumping Duty Administrative Review on Certain Steel Concrete Reinforcing Bars From Turkey — April 1, 2003, through March 31, 2004 (Nov. 2, 2005) (Pub.Doc. No. 256) (“Decision Memo”). 5

In reaching the Final Results, Commerce decided, inter alia, (1) to treat all of ICDAS’ U.S. sales as Export Price (“EP”), rather than Constructed Export Price (“CEP”), transactions; (2) to disallow a start-up adjustment for ICDAS’ Biga melt shop; (3) to treat ICDAS’ foreign exchange gains within the category of “financial expenses” for purposes of calculating ICDAS’ cost of production, and to cap ICDAS’ total financial expenses at zero; (4) to use contract date (rather than invoice date) as the date of sale for ICDAS’ U.S. sales, for purposes of calculating ICDAS’ antidumping duty margin; (5) to use in its “sales below cost” analysis the weighted average cost of manufacturing for the entire one-year period of review, rather than ICDAS’ quarterly average costs; and (6) to use the date of entry, rather than the date of sale, to define ICDAS’ universe of sales.

Based on Commerce’s analyses as reflected in its Final Results, the final dumping margin for ICDAS was calculated to be 0.16% — once again, a de minimis margin. See Final Results, 70 Fed.Reg. at 67,667. As a result of the company’s de minimis dumping margin in the review at issue, as well as its de minimis or zero margins in the two previous administrative reviews, Commerce revoked the antidumping order as to ICDAS, in accordance with the agency’s regulations. See Final Results, 70 Fed.Reg. at 67,666; 19 C.F.R. § 351.222(b)(2)® (providing for revocation of an order as to a particular exporter or producer of subject merchandise where, inter alia, the exporter or producer has “sold the merchandise at not less than normal value for a period of at least three consecutive years”). ■

The Domestic Producers and ICDAS brought the two actions consolidated here, challenging various aspects of the Final Results. One of the two issues raised in the Domestic Producers’ Complaint was Commerce’s use of the contract date as the date of sale for ICDAS’ U.S. sales. On behalf of Commerce, the Government requested and was granted a voluntary remand on that issue. On remand, Commerce reversed itself, changing its “date of sale” methodology for ICDAS’ U.S. sales, using the invoice date — rather than the contract date — as the date of sale. See Final Results of Redetermination Pursuant to Court Remand (“Remand Results”). As a result of that change, Commerce recalculated the dumping margin for ICDAS as above the de minimis level, and concluded that ICDAS does not qualify for revocation. See Remand Results at 2-3, 25.

The parties’ pending cross-motions for judgment on the agency record are directed to the Final Results of Commerce’s seventh administrative review, as well as the “date of sale” issue addressed in Commerce’s Remand Results. 6

III. Analysis

In their Motion for Judgment on the Agency Record, the Domestic Producers contend that, in the Final Results, Commerce wrongly treated sales made through ICDAS’ U.S. affiliate as Export Price (“EP”) — rather than Constructed Export Price (“CEP”) — sales. In its Motion for Judgment on the Agency Record, ICDAS argues, in turn, that Commerce erred in denying ICDAS’ request for a startup adjustment in the Final Results, that Commerce both improperly treated ICDAS’ foreign exchange gains within the category of “financial expenses” for purposes of calculating ICDAS’ cost of production in the Final Results and also improperly capped ICDAS’ total financial expenses at zero, that Commerce erred on remand in using invoice date (rather than contract date) as the date of sale for ICDAS’ U.S. sales, that Commerce erred in the Final Results by using in its “sales below cost” analysis the weighted average cost of manufacturing for the period of review (“POR”) (rather than ICDAS’ quarterly average costs), and that Commerce erred in the Final Results by using the date of entry (rather than the date of sale) to define ICDAS’ universe of sales.

Each of the parties’ claims is discussed in detail below. As set forth there, there is no merit to the Domestic Producers’ challenge to Commerce’s decision to treat sales made through ICDAS’ U.S. sales affiliate as EP sales. The Domestic Producers’ Motion for Judgment on the Agency Record therefore must be denied. Similarly lacking in merit are ICDAS’ challenge to Commerce’s denial of ICDAS’ request for a startup adjustment, and ICDAS’ challenge to Commerce’s treatment of ICDAS’ foreign exchange gains as “financial expenses” as well as the agency’s decision to cap ICDAS’ total financial expenses at zero. Accordingly, ICDAS’ Motion for Judgment on the Agency Record must be denied as to those claims. In contrast, Commerce’s use of invoice date (rather than contract date) as the date of sale for ICDAS’ U.S. sales, Commerce’s use of the POR average cost of manufacturing (rather than ICDAS’ quarterly costs) in the agency’s “sales below cost” analysis, and Commerce’s use of the date of entry (rather than the date of sale) to define ICDAS’ universe of sales cannot be sustained on the existing administrative record. ICDAS’ Motion for Judgment on the Agency Record therefore must be granted as to those claims, and this matter remanded to Commerce for further appropriate action.

A. Commerce’s Treatment of ICDAS’ U.S. Sales as Export Price (“EP”) Sales

Dumping takes place when merchandise is ' imported into the United States and sold at a price lower than its “normal value” — ie., the foreign market value of the subject merchandise. 19 U.S.C. §§ 1673, 1677(34). The difference between the normal value and the U.S. Price is the “dumping margin.” 19 U.S.C. § 1677(35). When normal value is compared to the U.S. Price and dumping is found, antidumping duties equal to the dumping margin may be imposed to offset the dumping. 19 U.S.C. § 1673(2)(B).

For purposes of an antidumping analysis, the U.S. Price is calculated using either the Export Price (“EP”) methodology or the Constructed Export Price (“CEP”) methodology. Commerce compares either the EP or the CEP with the “normal value” of the subject merchandise, to ascertain whether dumping is occurring, and, if so, to calculate the dumping margin. 19 U.S.C. §§ 1673, 1677a. If a transaction is classified as a CEP sale, the statute requires that certain additional deductions be taken from the sales price in order to arrive at the U.S. Price. 7 The bottom line is that use of CEP is more likely to result in a finding of dumping. See generally AK Steel Corp. v. United States, 226 F.3d 1361, 1364-65 & n. 4 (Fed.Cir.2000).

Export Price (“EP”) is defined in the statute:

The term “export price” means the price at which the subject merchandise is first sold (or agreed to be sold) before the date of importation by the producer or exporter of the subject merchandise outside of the United States to an unaffiliated purchaser in the United States or to an unaffiliated purchaser for exportation to the United States....

19 U.S.C. § 1677a(a). The statute defines Constructed Export Price (“CEP”) as well:

The term “constructed export price” means the price at which the subject merchandise is first sold (or agreed to be sold) in the United States before or after the date of importation by or for the account of the producer or exporter of such merchandise or by a seller affiliated with the producer or exporter, to a purchaser not affiliated with the producer or exporter....

19 U.S.C. § 1677a(b).

During the period of review, ICDAS exported merchandise to the United States both through a U.S. affiliate (which serves as importer of record, and is basically a “paper company”), as well as directly to unaffiliated customers. In the Final Results at issue here, Commerce classified all of ICDAS’ U.S. sales as EP sales, as it has since the 2001-2002 review, applying AK Steel (which includes a detailed analysis of the differences between EP and CEP sales) and emphasizing the locations of the transactions. 8 See generally Final Results, 70 Fed.Reg. 67,665; Decision Memo at 63-68. Commerce concluded:

While we note that it is undisputed that ICDAS’s U.S. importer is affiliated with ICDAS, this fact alone does not require a finding that the sales in question are CEP transactions. Under AK Steel, the salient issue is whether the sale at issue takes place inside or outside the United States, which the Court further discussed in Corus Staal, noting that “the focus of the inquiry is on the location of the sale not the role played by the affiliated importer.” See Corus Staal, 259 F.Supp.2d at 1259.

In this case, the record indicates that ICDAS’s sales through its affiliated importer were concluded in Turkey.... [T]he sales agreement was signed in Turkey by ICDAS personnel, the invoice was issued by an entity in Turkey (ie., the producer/exporter) to an entity in the United States (ie., the U.S. customer), and it was concluded outside the United States.

Regarding the petitioners’ arguments involving the transfer of title, we disagree that the evidence on the record shows that title passed to the customer inside the United States. We have examined the documents taken at verification and find that none of the contracts for ... entries [during the period of review] shows that title passed after entry.

Decision Memo at 66-67.

The Domestic Producers contend that Commerce erred in classifying sales made through ICDAS’ U.S. affiliate as EP sales. According to the Domestic Producers, those transactions instead should be treated as CEP sales. See generally Domestic Producers Brief, passim; Domestic Producers Reply Brief at 1-9.

In their briefs, the parties devote much ink to their competing interpretations of the Court of Appeals’ opinion in AK Steel, and, to a lesser degree, the opinion of this court in Corus Staal. See AK Steel, 226 F.3d 1361; Corus Staal BV v. U.S. Dep’t of Commerce, 27 CIT 388, 259 F.Supp.2d 1253 (2003) (concerning Final Determination in antidumping investigation). The Domestic Producers maintain that those two opinions require that sales made through ICDAS’ U.S. affiliate be classified as CEP sales. See Domestic Producers Brief at 8; see also id. at 9-17; Domestic Producers Reply Brief at 1-7.

The Domestic Producers initially asserted that AK Steel holds flatly that “any sale in which the contract is between a U.S. affiliate and an unaffiliated U.S. customer must be classified as CEP,” based solely on the domicile of the seller and without regard to the location of the sale or transaction. See Domestic Producers Brief at 11; see also id. at 8, 12. In their Reply Brief, the Domestic Producers moderated their stance slightly, arguing that “[t]he decision [in AK Steel ] appears to support two different, mutually exclusive tests for whether a transaction is CEP or EP” — one test based on the domicile of the party making the sale to the first unaffiliated customer, and one test based on the location of the sale or transaction. See Domestic Producers Reply Brief at 2.

Specifically, the Domestic Producers assert that “[portions of the opinion [in AK Steel ] clearly state that whenever the sale to the first unaffiliated U.S. customer is made by a U.S. selling affiliate, that transaction must be classified as CEP.” See Domestic Producers Reply Brief at 2. To illustrate this point, the Domestic Producers excerpt language from the Court of Appeals’ conclusion in AK Steel: “[I]f the contract for sale was between a U.S. affiliate of a foreign producer or exporter and an unaffiliated U.S. purchaser, then the sale must be classified as a CEP sale____ Similarly, a sale made by a U.S. affiliate or another party other than the producer or exporter cannot be an EP sale.” AK Steel, 226 F.3d at 1374 (quoted in Domestic Producers Reply Brief at 2).

At the same time, however, the Domestic Producers candidly concede — as they must — that “other portions of the opinion [in AK Steel ] appear to state that a sale by a U.S. affiliate can be classified as EP where title transfers and consideration is [given] outside of the United States.” See Domestic Producers Reply Brief at 2-3. To illustrate that point, the Domestic Producers point to another excerpt from AK Steel: “The term ‘outside the United States,’ read in the context of both the CEP and the EP definitions, ... applies to the locus of the transaction at issue, not the location of the company.” AK Steel, 226 F.3d at 1369 (quoted in Domestic Producers Reply Brief at 2-3).

Summing up their analysis of AK Steel, the Domestic Producers state: “Thus, at different points in the opinion, the Federal Circuit appears to endorse a bright-line rule whereby all sales through a U.S. selling affiliate are CEP; in others, it appears to make the distinction based solely on the location of title transfer.” Domestic Producers Reply Brief at 3. The Domestic Producers conclude — based on both their analysis of the language of the opinion, as well as their analysis of the facts and outcome of the case — that AK Steel’s references to “the location of the sale” were actually intended to refer to “the domicile of the seller”; and, moreover, that, under AK Steel, it is “the seller’s domicile, rather than the location of title-transfer” which is “the defining factor” in an EP/CEP analy sis. See Domestic Producers Brief at 13-17; Domestic Producers Reply Brief at 8.

To be sure, the detailed analysis in the Court of Appeals’ opinion in AK Steel is necessarily dense, and can therefore be challenging to follow at points. It is therefore difficult not to sympathize with the Domestic Producers, as they struggle to distill the implications of AK Steel for this case. And, as the Domestic Producers indicate, some statements in AK Steel appear (at least at first blush) to be somewhat in tension with other statements in the opinion. The Government and ICDAS correctly note, however, that the gravamen of AK Steel is the significance of the location of the sale or transaction — specifically, “whether the sale or transaction takes place inside or outside the United States.” See AK Steel, 226 F.3d at 1369-70 (characterizing location of sale or transaction as a “critical difference” between EP and CEP sales).

In AK Steel, the Court of Appeals focused repeatedly and definitively on the importance of the location of a sale or transaction in determining its classification as an EP or CEP sale. 9 Thus, the Court framed “[t]he question at the root of [the] appeal” in that case as whether a sale can be properly classified as an EP sale “if the sales contract ... is executed in the United States.” AK Steel, 226 F.3d at 1368 (emphasis added).

AK Steel’s analysis of the language of the statute similarly highlights the significance of the location of the sale or transaction. Reviewing the text of the statute, the Court of Appeals determined that “the plain meaning of the language enacted by Congress ... focuses on where the sale takes place.” AK Steel, 226 F.3d at 1369 (emphasis added) (discussing 19 U.S.C. § 1677a (defining “EP” and “CEP”)). 10 Underscoring the fact that the location of the sale or transaction is a “dispositive” factor in classifying sales as EP sales or CEP sales, the Court of Appeals continued:

The text of the [statutory] definition of CEP states that CEP is the “price at which the subject merchandise is first sold in the United States.” 19 U.S.C. § 1677a(a) (emphasis added). In contrast, EP is defined as the price at which the merchandise is first sold “outside the United States.” 19 U.S.C. § 1677a(b). Thus, the location of the sale appears to be critical to the distinction between the two categories.

AK Steel, 226 F.3d at 1369 (second emphasis added). 11 Echoing its characterization of the location of the sale or transaction as a “dispositive” factor, the Court expressly identified “whether the sale or transaction takes place inside or outside the United States ” as a “critical difference” between EP and CEP sales. AK Steel, 226 F.3d at 1369-70 (emphasis added).

The Court of Appeals’ review of the specific facts of AK Steel continues the drumbeat on the location of the sales or transactions. There, too, the Court of Appeals focused like a laser on the issue, ultimately concluding that the transactions in the ease were not EP sales, but CEP sales. See AK Steel, 226 F.3d at 1370-72, 1374. Observing that — as a practical matter — “whether a sale is ‘outside the United States’ depends, in part, on whether the parties are or are not located in the United States,” the Court ruled:

A transaction, such as those here, in which both parties are located in the United States and the contract is executed in the United States cannot be said to be “outside the United States.” Thus, such a transaction cannot be classified as an EP transaction. Rather, classification as an EP sale requires that one of the parties to the sale be located “outside the United States,” for if both parties to the transaction were in the territory of the United States and the transfer of ownership was executed in the United States, it is not possible for the transaction to be outside the United States.

AK Steel, 226 F.3d at 1370 (emphases added). Emphasizing that “Congress has made a clear distinction between [EP sales and CEP sales] based on the geographic location of the transaction,” the Court of Appeals stated that it would be “contrary to the plain meaning of the statute” to classify the transactions in AK Steel as EP sales “[w]hen ... there are contracts showing that the sales at issue took place in the United States between two entities with United States addresses, one of which was an affiliate of the producer/exporter.” AK Steel, 226 F.3d at 1370-71 (emphases added). 12

In AK Steel, the Korean producers argued that the statutory term “seller” was ambiguous, and that Commerce should be permitted to interpret it in terms of the U.S. affiliate’s activities. AK Steel, 226 F.3d at 1371. The Court of Appeals made short work of that argument, again highlighting the significance of the location of the sale or transaction. The Court of Appeals concluded: “If Congress had intended the EP versus CEP distinction to be made based on which party set the terms of the deal or on the relative importance of each party’s role, it would not have written the statute to distinguish between the two categories based on the location where the sale was made and the affiliation of the party that made the sale.” AK Steel, 226 F.3d at 1372 (emphasis added).

Finally, the significance of the location of the sale or transaction is highlighted in the ultimate statement of the holding of AK Steel:

Stated in terms of the EP. definition: if the sales contract is between two entities in the United States, and executed in the United States and title will pass in the United States, it cannot be said to have been a sale “outside the United States”; therefore, the sale cannot be an EP sale.

AK Steel, 226 F.3d at 1374 (initial emphasis added). As illustrated by the excerpts quoted in the discussion above (including notes 9 through 12), AK Steel’s emphasis on the location of a sale or transaction in classifying it as EP or CEP is much greater than the Domestic Producers acknowledge.

In short, there is no merit to the Domestic Producers’ claim that AK Steel mandates that any sale in which the contract is between two U.S. domiciled entities — ie., a U.S. affiliate of a foreign producer/exporter and an unaffiliated U.S. customer— must necessarily, by definition, be classified as a CEP sale, without regard to the location of that sale or transaction. Indeed, the Court of Appeals in AK Steel held that a critical inquiry in making an EP/CEP classification is the location of the sale or transaction — in particular, whether the sale or transaction takes place inside or outside the United States. See AK Steel, 226 F.3d at 1369; 19 U.S.C. § 1677a(a)-(b) (defining “export price” in terms of a sale made (or an agreement to sell reached) “outside of the United States,” and defining “constructed export price” in terms of a sale made (or an agreement to sell reached) “in the United States”).

The Court of Appeals held that, for purposes of 19 U.S.C. § 1677a, the term “outside of the United States” refers to “the locus of the transaction at issue, not the location of the company.” AK Steel, 226 F.3d at 1369. The Court noted that, in most situations, EP sales will involve one party domiciled outside the United States, because sales between two U.S. domiciled parties normally will take place inside the United States. AK Steel, 226 F.3d at 1370. Nevertheless, as all parties acknowledge, the Court of Appeals expressly reserved judgment as to whether “a sales contract between two U.S. domiciled entities that is entirely executed outside the United States” would be classified as an EP sale or a CEP sale. See AK Steel, 226 F.3d at 1370 n. 8 (emphasis added); Domestic Producers Brief at 17 n. 11; ICDAS Response Brief at 33-34; Decision Memo at 63. 13 Thus, under the AK Steel test, it may be possible for two U.S. entities to conduct an EP sale, provided that the sale is completed outside the United States. In any event, as discussed herein, Commerce determined in this case that all sales in question were between ICDAS (a Turkish producer/exporter) and unaffiliated U.S. purchasers, and that all sales in question were completed outside the United States.

The Domestic Producers’ argument was fully laid to rest by the Court of Appeals in Corus Staal. See Corus Staal BV v. United States, 502 F.3d 1370, 1377 (Fed.Cir.2007) (concerning second administrative review of antidumping order covering hot- rolled carbon steel flat products from the Netherlands). The Court of Appeals there underscored its holding in AK Steel, reiterating the significance of the location of a transaction in classifying it as an EP or CEP sale:

AK Steel does not stand for the proposition that all sales by foreign sellers to unaffiliated U.S. customers should be considered EP transactions. In fact, AK Steel states that transactions ... in which the sale made by a foreign producer or exporter occurs in the United States, should be treated as CEP transactions.

Corus Staal, 502 F.3d at 1377 (emphases added); see also id. (quoting AK Steel, 226 F.3d at 1369: “[T]he location of the sale appears to be critical to the distinction between the two categories.”). It is thus a major overstatement to assert (as the Domestic Producers do) that AK Steel mandates that “where the first sale to an unaffiliated party is made by a U.S. selling affiliate of the foreign producer/exporter, the sale must be classified as CEP,” without regard to the location of the transaction. See Domestic Producers Brief at 8. 14

As the Government correctly points out, the.statute requires that — -in determining whether a sale is an EP sale or a CEP sale — the first step in Commerce’s analysis is to identify when “the subject merchandise is first sold (or agreed to be sold).” See Def. Response Brief at 30-31 (discussing definitions of EP and CEP in 19 U.S.C. § 1677a(a)-(b), both of which include quoted phrase). The Government further notes that, in determining where merchandise is “first sold (or agreed to be sold),” Commerce must adhere to the plain language definitions of the terms “sold” and “agreed to be sold.” See Def. Response Brief at 31. AK Steel defined a “sale” in the context of 19 U.S.C. § 1677a to “require[ ] both a ‘transfer of ownership to an unrelated party and consideration.’ ” AK Steel, 226 F.3d at 1371 (citation omitted; emphases added in AK Steel). Similarly, in Corus Staal, the Court of Appeals defined an “agreement to sell” (for purposes of 19 U.S.C. § 1677a) as “a binding commitment that has not yet been consummated by the exchange of goods for consideration, i.e., the ‘sale’ itself.” Corus Staal, 502 F.3d at 1376-77. 15 Thus, “[a]s used in the statute, the terms ‘sale’ and ‘agreement to sell’ ... cover different types of transactions.” Corus Staal, 502 F.3d at 1377. Further, the Court of Appeals has held that “[njeither a sale nor an agreement to sell occurs until there is mutual assent to the material terms [of a deal] (price and quantity).” Corus Staal, 502 F.3d at 1376.

’ In the case at bar, all activities relevant to sales of ICDAS’ rebar to U.S. customers — including sales negotiations, issuance of invoices, and preparation of documentation to facilitate payment — were handled outside the United States, by ICDAS personnel in Turkey. See Section A Questionnaire Response of ICDAS Celik Enerji Tersane ve Ulasim Sanayi, A.S. (Pub. Doc. No. 67; Conf. Doc. No. 1) at A-8. 16 Specifi cally, all of ICDAS’ sales to the United States were based upon contracts which were negotiated and finalized in Turkey prior to ICDAS’ shipment of merchandise. See id. The Domestic Producers highlight the fact that, inter alia, ICDAS’ sales documentation shows that — for sales made through ICDAS’ U.S. affiliate — ICDAS first invoiced merchandise from itself to its U.S. affiliate, and then from the U.S. affiliate to ICDAS’ U.S. customer. See, e.g., Domestic Producers Brief at 11. But ICDAS’ U.S. affiliate is merely a “paper” company that has no employees or business premises in the United States, is not involved in the sales process, never takes possession of subject merchandise, and acts only as importer of record. See Section A Questionnaire Response of ICDAS Celik Enerji Tersane ve Ulasim Sanayi, A.S. (Pub. Doc. No. 67; Conf. Doc. No. 1) at A-8, A-14-15. 17 The Domestic Producers can point to no evidence refuting these record facts, and thus cannot establish that any entity other than ICDAS prepared all relevant invoice documentation. See id. at A-8 (stating that “[a]ll sales activities related to the sales to U.S. customers ... such as ... issuing of invoices ... occurred in Turkey”); ICDAS Sales Verification Report (Conf.Rec. No. 44) at 3 (confirming, in the course of verification process, that “the personnel in ICDAS’s export sales department act on behalf of [the importer] because [the importer] itself has no employees”). See generally Decision Memo at 66 (finding that “the sales agreement was signed in Turkey by ICDAS personnel, the invoice was issued by an entity in Turkey (i.e., the producer/exporter) to an entity in the United States (i.e., the U.S. customer), and [the sale] was concluded outside the United States”) (emphasis added).

Moreover, pursuant to the terms of ICDAS’ sales, all deliveries of merchandise (first from ICDAS to its U.S. affiliate, then immediately from the affiliate to the unaffiliated purchaser) occurred at the port of shipment in Turkey — outside the United States. See Section A Questionnaire Response of ICDAS Celik Enerji Tersane ve Ulasim Sanayi, A.S. (Pub. Doc. No. 67; Conf. Doc. No. 1), Exh. A-9 at 1, 2; ICDAS Sales Verification Report (Conf.Rec. No. 44) at Exh. 11. Under the circumstances, Commerce properly determined to treat ICDAS’ U.S. sales as EP transactions, in accordance with AK Steel, “because the[] sales were made pursuant to agreements made between [unaffiliated U.S. customers and] ICDAS personnel in Turkey.” See Decision Memo at 65 (emphasis added). 18

Apart from their arguments disputing the legal significance of the location of the transaction based on their interpretation of AK Steel (addressed above), the Domestic Producers further contend that “the record lacks the evidence necessary to make any reasonable determination regarding the location of the sales at issue.” Domestic Producers Brief at 17-18; see also id. at 2, 8-10, 17-21, 24; Domestic Producers Reply Brief at 1, 7-8. 19 Noting that AK Steel defined the term “sold” (for purposes of EP/CEP classification) by reference to the transfer of ownership or title, the Domestic Producers challenge the quantum of record evidence concerning the transfer of title in the sales at issue here. Specifically, the Domestic Producers point to the fact that Commerce collected sales trace information concerning two of ICDAS’ sales to unaffíliated U.S. customers during the period of review. The Domestic Producers further note that only one of those sales traces includes information concerning the transfer of title, and assert that it is not enough to support Commerce’s determination. See Domestic Producers Brief at 18; see also Domestic Producers Reply Brief at 7.

Contrary to the Domestic Producers’ claims, the record evidence is sufficient to establish the location of the limited number of transactions at issue. As a threshold matter, it is well-established that, in principle, “[i]t is up to Commerce, not the court, to weigh the ... evidence that was properly submitted during verification.” See Corus Staal, 27 CIT at 394, 259 F.Supp.2d at 1259. In this case, Commerce “examined the documents taken at verification and [concluded] that none of the contracts for POR entries shows that title passed after entry.” See Decision Memo at 67. In other words, Commerce determined that — as to each transaction at issue — title transferred outside the United States. See Def. Response Brief at 34-35. 20

One sale which was verified by Commerce specified that title passed when payment for the merchandise was received in full — which occurred well before the import entry date. See ICDAS Sales Verification Report (Conf.Doc. No. 44) at Exh. 14. Commerce thus confirmed that title for that sale passed before the goods entered this country — that is, outside the United States. See generally ICDAS Response Brief at 37; Def. Response Brief at 35. In addition, as to all other sales at issue, Commerce determined that all deliveries of ICDAS’ goods were made outside the United States, in accordance with the terms of each of the sales, which were governed by certain specific Incoterms provisions. See Decision Memo at 67.

The Domestic Producers challenge ICDAS’ reliance on Ineoterms, insisting that “Incoterms are not relevant to transfer of title.” See Domestic Producers Brief at 19-20. But the Domestic Producers’ argument glosses over certain pivotal points.

As ICDAS readily acknowledges, the relevant Incoterms (including FOB, CFR, and CIF) deal directly with the transfer of risk, rather than transfer of title. See ICDAS Response Brief at 39. However, under generally accepted principles of commercial law (reflected domestically in, inter alia, the Uniform Commercial Code, as well as in international lex mercatoria ), 21 in the absence of an express agree ment between the parties as to when title passes, title to goods transfers when the seller completes performance with respect to the physical delivery of the goods. See ICDAS Response Brief at 39. In each of the transactions here at issue, the goods were delivered outside the United States. Thus, as to each of those transactions, Commerce reasonably concluded that title transferred outside the United States as well.

In short, contrary to the Domestic Producers’ assertions, the administrative record in this matter adequately supports Commerce’s determination that each of the transactions at issue occurred outside the United States. 22 The record before Commerce plainly includes “such relevant evidence as a reasonable mind might accept as adequate” to support the agency’s determination on this point. See Consol. Edison Co., 305 U.S. at 229, 59 S.Ct. 206. The law requires no more.

B. Commerce’s Disallowance of Startup Adjustment for ICDAS’ Biga Melt Shop

During the administrative review proceedings, ICDAS requested that Commerce grant it a startup adjustment for ICDAS’ Biga melt shop, which began production of steel billets in December 2003. See Decision Memo at 76-81. 23 A startup adjustment is an adjustment to the costs incurred by a company for production that is affected by startup operations during the period covered by an administrative review. See generally 19 U.S.C. § 1677b(f)(l)(C)(i). In the Final Results, Commerce denied ICDAS’ request, stating that ICDAS had failed to meet the requirements of the statute. See Decision Memo at 76. ICDAS here challenges Commerce’s disallowance of the claimed startup adjustment. See ICDAS Brief at 2, 5, 22-32; ICDAS Reply Brief at 5-10.

The statute authorizes a startup adjustment only where a producer establishes both that it is “using new production facilities or producing a new product that requires substantial additional investment,” and that “production levels are limited by technical factors associated with the initial phase of commercial production.” See 19 U.S.C. § 1677b(f)(l)(C)(ii). In the case at bar, Commerce concluded that ICDAS did not meet its burden as to the second criterion. Specifically, Commerce determined that ICDAS did not demonstrate sufficiently limited production levels for its claimed startup period (December 2003); nor did ICDAS establish that any asserted limitations on production were attributable to “technical factors associated with the initial phase of production.” See Decision Memo at 79.

ICDAS basically raises two challenges to Commerce’s disallowance of the claimed startup adjustment. First, ICDAS argues that Commerce’s use of a full-month equivalent methodology to compare December 2003 production levels at the Biga facility to those of subsequent months is not in accordance with law, because — according to ICDAS' — it relies on “theoretical” data. See generally ICDAS Brief at 2, 5, 22-29; ICDAS Reply Brief at 5-8. And, second, ICDAS contends that Commerce erred in concluding that ICDAS did not provide sufficient information concerning technical factors limiting production. See generally ICDAS Brief at 5, 22-23, 29-31; ICDAS Reply Brief at 8-10.

Both arguments are unavailing.

1. Commerce’s Determination That Biga’s December 2003 Production Was Not Limited

The startup adjustment statute does not define what constitutes “limited production.” Nor does the statute dictate how Commerce is to measure levels of production. 24 However, the statute does delimit the duration of the startup period.

According to the statute, the startup period ends “at the point at which the level of commercial production that is characteristic of the merchandise, producer, or industry concerned is achieved.” 19 U.S.C. § 1677b(f)(l)(C)(iii). The statute thus does not extend the startup period up to the date on which the new facility reaches optimum functioning capacity. Indeed, the Statement of Administrative Action accompanying the Uruguay Round Agreements Act expressly provides that “[attainment of peak production levels will not be the standard for identifying the end of the startup period because the startup period may end well before a company achieves optimum capacity utilization.” See Statement of Administrative Action, H.R. Doc. No. 103-316, at 836 (1994), reprinted in 1994 U.S.C.C.A.N. 4040, 4173. 25

To be sure, as ICDAS repeatedly emphasizes, the absolute (unadjusted) production figures for the Biga melt shop for December 2003 were relatively low. Indeed, no production at all occurred in the first 10 days of the month, while ICDAS was conducting a series of test runs. Thus, the facility did not begin production until well into the start of the month; and the December 2003 figures reflect only a partial month of production.

Because the December 2003 production figures were based on a partial month, Commerce converted those figures to a full-month equivalent, so that Biga’s production data for December 2003 could be compared to the full-month production data for subsequent months. See Decision Memo at 80. Specifically, Commerce examined ICDAS’ actual production measured in units processed for the months of December 2003, and January through March 2004. Based on ICDAS’ actual production data, Commerce then calculated a full-month equivalent production figure for December 2003. In addition, Commerce calculated December 2003 production starts using a full-month equivalency, so that the agency would have data on production starts for that month to compare to production starts data for January, February, and March 2004. See Decision Memo at 80. 26

With the production data for December 2003 restated as full-month equivalents, Commerce could reasonably compare the Biga melt shop’s production in December 2003 (the claimed startup period) to the facility’s production levels in January, February, and March 2004. Based on its comparison of production data, Commerce concluded thaN-while Biga’s production in December 2003 clearly was not at the optimal level that ICDAS planned to achieve— the facility’s production in December 2003 in fact was not limited within the meaning of the statute, and that ICDAS therefore was not eligible for a startup adjustment. See Decision Memo at 80.

ICDAS charges that Commerce’s use of a full-month equivalent methodology is not in accordance with law, because it “improperly uses hypothetical production data.” ICDAS Brief at 27; see also id. at 2, 5, 26-28; ICDAS Reply Brief at 5-6, 8. To support its argument, ICDAS points to other instances in which Commerce has declined to grant a startup adjustment on the grounds that the production data proffered by producers were “theoretical” or “hypothetical.” See, e.g., ICDAS Brief at 24-25, 27 (citing Issues and Decision Memorandum for the Final Determination in the Antidumping Investigation of Chlorinated Isocyanurates from Spain, 2005 WL 2290648 (May 2, 2005) (“Chlorinated Isos from Spain”), at comment 9 (where Commerce denied requested startup adjustment, because respondent’s calculations relied “on a theoretical production capacity rather than the level of commercial production as required by the [Statement of Administrative Action]”); Issues and Decision Memorandum for the Final Results of the Administrative Review of Stainless Steel Bar from India, 2003 WL 24153851 (Aug. 4, 2003) (“Stainless Steel Bar from India”), at comment 2 (where Commerce rejected respondent’s production limitation analysis because respondent compared “its actual production levels to its theoretical production capacity”)).

But there was nothing “hypothetical” about the full-month equivalent methodology that Commerce used in the administrative review at issue here — at least not in the sense in which Commerce has previously used that term. True enough, Commerce in the past has rejected purely hypothetical production data, such as data based solely on production capacity, or speculative projections as to future production. See Decision Memo at 81. However, those situations were very different from what Commerce did in this case.

Here, Commerce did not use theoretical production data to project future production levels. Instead, the agency used actual production data to establish a full-month equivalent for a partial month of production in the past See Def. Response Brief at 19-20 (distinguishing full-month equivalent methodology in this case from situations presented in Chlorinated Isos from Spain and Stainless Steel Bar from India). Contrary to ICDAS’ implication, Chlorinated Isos from Spain and Stainless Steel Bar from India do not stand for the proposition that Commerce is precluded from using actual production data for a partial month of operations to extrapolate a full month equivalent for purposes of comparison with other actual production data for subsequent full months. Those two cases merely reflect Commerce’s policy of rejecting producers’ attempts to qualify for a startup adjustment by using projected or optimal production capacity to establish an artificially-inflated benchmark for commercial production.

In sum, in analyzing ICDAS’ request for a startup adjustment, Commerce used ICDAS’ actual production data, and adjusted those data to permit an apples-to-apples comparison of partial-month production for December 2003 with full-month production in the three months that followed. Commerce’s full-month equivalent methodology thus used ICDAS’ actual production data and reasonably adjusted them to effectuate Congress’ intent — that is, to determine whether Biga’s production in December 2003 was limited. See generally Domestic Producers Response Brief at 20-21. 27

Nothing about Commerce’s full-month equivalent approach is inconsistent with the language of the startup adjustment statute, which does not specifically define how Commerce is to measure whether, in any given case, production was “limited.” Under such circumstances, Commerce is entitled to substantial deference in interpreting the statute. And nothing about the agency’s interpretation here is inherently unreasonable. See generally Domestic Producers Response Brief at 20-21; Chevron U.S.A., Inc. v. Natural Resources Defense Council, 467 U.S. 837, 843, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984) (holding that, “if the statute is silent or ambiguous with respect to [a] specific issue, the question for the court is whether the agency’s answer is based on a permissible construction of the statute”); Suramerica de Aleaciones Laminadas, C.A. v. United States, 966 F.2d 660, 665 (Fed.Cir.1992) (observing that courts have duty to “respect legitimate policy choices made by the agency in interpreting and applying the statute”). Cf. INS v. Elias-Zacarias, 502 U.S. 478, 483-84, 112 S.Ct. 812, 117 L.Ed.2d 38 (1992) (stating that, where Congress has entrusted agency to administer statute in fact-intensive situations, agency’s conclusion should be reversed only if the record evidence is “so compelling that no reasonable factfinder” could reach the same conclusion). ICDAS’ attack on Commerce’s full-month equivalent methodology must therefore be rejected. 28

2. Commerce’s Determination That ICDAS Failed to Prove That Technical Factors Limited Biga’s Production

In addition to its challenge to Commerce’s full-month equivalent methodology, ICDAS also disputes Commerce’s conclusion that ICDAS failed to provide sufficient information to establish that any limited production at Biga was the result of “technical factors associated with the initial phase of commercial production.” See ICDAS Brief at 5, 22, 29-32; ICDAS Reply Brief at 8-10. As discussed below, however, ICDAS’ argument is lacking in merit. Accordingly, even assuming arguendo that ICDAS had established that Biga’s “production levels [were] limited” in December 2003 (the claimed startup period) (which, as discussed in section III.B.l immediately above, it did not), ICDAS nevertheless still would not be entitled to a startup adjustment, because ICDAS failed to meet its burden of proof to establish the cause of any assertedly limited production. See generally Def. Response Brief at 5-6, 9, 14-17;- Domestic Producers Response Brief at 3, 17-18, 21-26.

In denying the requested startup adjustment, Commerce found that ICDAS failed to respond to the agency’s inquiries concerning technical factors associated with the initial phase of commercial production. Indeed, the record on point consisted of only a single vague statement, with no documentary support. See Decision Memo at 80.

Section D of Commerce’s antidumping questionnaire explicitly requests that respondents provide support for any claimed startup adjustments. Thus, in its questionnaire to ICDAS, Commerce expressly asked ICDAS to provide detailed information and documentation to support ICDAS’ claim that Biga’s production for the month of December 2003 was limited by (in the words of the statute) “technical factors associated with the initial phase of commercial production.” Specifically, Commerce’s questionnaire requested that ICDAS:

8. [Ejxplain how the production levels were limited by technical factors associated with the initial phase of commercial production (as part of your analysis, describe the technical factors which limited production, demonstrate how these technical factors restricted the number of units processed by the company, and demonstrate how these technical factors are unique to the startup phase, not a result of chronic or normal production problems).

The sole information on point that ICDAS placed on the record was the following succinct statement, in the company’s Section D Questionnaire Response: ■

Production levels were limited by technical factors associated with the initial phase of commercial production because the company had to 1) develop the production parameters of the new operations; 2) install, adjust, calibrate and test the new equipment; and 3) train new employees to operate the new equipment. Operations typically incur such technical problems because of the newness of the facility.

Section D Questionnaire Response of ICDAS Celik Enerji Tersane ve Ulasim Sanayi, A.S. (Pub.Doc. No. 67) at D-40.

Although ICDAS bore the burden of proof on all elements necessary to establish its right to a startup adjustment, 29 Commerce followed up on ICDAS’ terse questionnaire response, on its own initiative. Specifically, Commerce took affirmative steps to seek to elicit the requisite detailed information and documentary support concerning the claimed limiting technical factors, issuing a supplemental Section D Questionnaire to ICDAS, 30 and inquiring again at ICDAS’ cost verification. ICDAS nevertheless failed to supply any further information. See Decision Memo at 79-80. Commerce ultimately concluded that the record lacked sufficient information to allow the agency to conclude that any. asserted limitation on production at the Biga facility in December 2003 was attributable to technical factors unique to startup, rather than “factors unrelated to startup, such as marketing difficulties or chronic production problems.” See Decision Memo at 79-80; Statement of Administrative Action, H.R. Doc. No. 103-316, at 838, reprinted in 1994 U.S.C.C.A.N. at 4174.

In this action, ICDAS raises two principal objections to Commerce’s conclusion. ICDAS first argues that the relationship between December 2003 production levels at the Biga facility and technical factors associated with the initial 'phase of commercial production should have been “self-evident” to Commerce, in light of the information that the agency had before it. See ICDAS Brief at 30; see also id. at 5, 22, 29-31; ICDAS Reply Brief at 8. In addition, ICDAS argues that Commerce’s failure to grant the requested startup adjustment amounts to the improper use of “facts otherwise available” or adverse inference. See ICDAS Brief at 31-32. Neither argument holds water.

a. ICDAS’ Claim That Limitation Due to Technical Factors Is ■ “Self-Evident”

Notwithstanding the fact that it bore the burden of proof, and despite Commerce’s requests for further detail and documentation (both through a supplemental questionnaire and at verification), ICDAS maintains that — other than the existing information on the record — it was not required to provide evidence that any asserted limitations on production at the Biga facility in December 2003 were due to factors unique to startup. ICDAS maintains that those factors and their limiting effects are “self-evident” from the record evidence, and that it thus “provided sufficient information for Commerce to address the startup issue.” See ICDAS Brief at 30, 31 n. 21. ICDAS further emphasizes that Commerce verified the fact that ICDAS conducted test runs at the Biga facility in early December, and that it did not begin actual production until later that month. See ICDAS Brief at 5, 22-23, 25, 30. Finally, ICDAS notes that the Biga facility produced only a limited number of types of billet in December 2003, but produced many more types in the months that followed. See ICDAS Brief at 22-23, 31 n. 20.

As the Domestic Producers observe, however, the information to which ICDAS points was not an adequate basis for a startup adjustment. See generally Domestic Producers Response Brief at 3, 18, 21-26; see also Def. Response Brief at 5-6, 9, 14-17, 22. By any measure, the information on which ICDAS relies was not sufficient to demonstrate that technical factors unique to startup — rather than “factors unrelated to startup, such as marketing difficulties or chronic production problems” — were the cause of assertedly limited production levels at Biga in December 2003, and to “document that fact to the Department’s satisfaction.” See Statement of Administrative Action, H.R. Doc. No. 103-316, at 838, reprinted in 1994 U.S.C.C.A.N. at 4174; Antidumping Duties; Countervailing Duties: Proposed Rule, 61 Fed.Reg. 7308, 7340 (Feb. 27, 1996) (Preamble).

The Domestic Producers sum up the state of the record thusly: “The verified evidence, as identified by ICDAS, is this: Biga Melt was a new facility.... The production equipment was newly installed.... Test runs were conducted prior to production .... [A limited number of] types of billet were produced there in December 2003; [many more] types were produced in succeeding months.... That is all.” See Domestic Producers Response Brief at 23. As discussed below, these basic facts— considered alone, or even in the aggregate — simply do not suffice to allow Commerce to grant the startup adjustment that ICDAS seeks.

For example, ICDAS’ first piece of evidence — that the Biga facility was entirely new — is logically relevant only to the first criterion of the startup adjustment standard (i.e., that “a producer [was] using new production facilities”), not to the sec ond criterion (ie., that “production levels [were] limited by technical factors associated with the initial phase of commercial production”), which is the criterion at issue here. Nothing about the newness of the Biga facility, in and of itself, demonstrates that production levels were limited due to technical factors unique to the startup phase. If newness were itself evidence that technical factors necessarily limit production in a facility’s startup phase, the second criterion of Congress’ startup adjustment standard would be entirely superfluous. See generally Domestic Producers Response Brief at 23-24; 19 U.S.C. § 1677b(f)(l)(C)(ii) (two-part standard for startup adjustment).

Similarly, ICDAS’ second piece of evidence — that the equipment at the Biga facility was newly-installed — does not, without more, demonstrate that any asserted limitations on initial production were attributable to technical factors associated with startup. It simply reinforces the fact that the facility itself was new. New equipment alone is not evidence of technical limitations affecting production. See generally Domestic Producers Response Brief at 24.

ICDAS’ third piece of evidence — that ICDAS devoted days to testing equipment at the Biga facility prior to beginning production — is also inapposite. This fact too merely demonstrates that the facility was new, and does not necessarily say anything about whether production levels were limited by technical factors unique to startup. See generally Domestic Producers Response Brief at 24.

The fourth piece of evidence cited by ICDAS is the disparity between the number of types of billet produced at the Biga facility in December 2003 and the number produced in later months. But this evidence is equally meaningless vis-a-vis the existence (or non-existence) of technical factors unique to the startup phase. There is nothing on the record to show that the relatively low number of types of billet produced at Biga in December 2003 was due to technical factors associated with startup. The record simply shows that additional types of billet were produced later. Indeed, there is no record evidence to indicate that the difference between the types of billet produced in December 2003 and the types produced in later months reflects anything more than a business decision on the part of ICDAS. See Domestic Producers Response Brief at 24; see also id. at 23 n. 9.

In sum, none of the evidence on which ICDAS relies speaks to whether any asserted limitation on production at Biga in December 2003 was attributable to technical factors unique to startup operations. 31 The evidence either simply reflects that the Biga facility was new (a fact which was known and undisputed, and which is relevant at most to the first criterion of the startup adjustment standard, which is not at issue here), or it is wholly irrelevant. Either way, ICDAS’ evidence does little or nothing to support its claim to a startup adjustment. 32

ICDAS failed to provide Commerce with the evidence required to allow the agency to grant ICDAS’ request for a startup adjustment. The statute requires more than mere evidence that a production facility is new. Rather, to justify a startup adjustment, a producer must provide specific, detailed information concerning how, and to what extent, technical factors associated with the initial phase of commercial operations limited initial production at its new facility. See 19 U.S.C. § 1677b(f)(l)(C)(ii). ICDAS’ argument that the relationship between December 2003 production levels at the Biga facility and technical factors unique to startup is “self-evident” therefore must fail.

b. ICDAS’ Claim That Commerce Resorted to “Facts Otherwise Available”

As its final challenge to Commerce’s denial of the requested startup adjustment, ICDAS argues that, even if ICDAS failed to fully respond to Commerce’s requests for information and documentation concerning “technical factors associated with the initial phase of commercial production” at the Biga facility, Commerce had no “basis for resorting to facts available or drawing any adverse inferences.” See generally ICDAS Brief at 31. 33 ICDAS faults Commerce’s determination because it does not include “any analysis under the antidumping law’s provisions on ‘facts available’ ” to justify “discarding]” information provided by ICDAS to support its adjustment request. Id. ICDAS further complains that Commerce “erred by failing to explain why the information that was allegedly withheld was necessary to reach a decision on the startup adjustment issue.” Id. But ICDAS’ arguments have no basis in either law or fact.

ICDAS’ position is apparently based on its belief that it provided sufficient information to Commerce to allow the agency to grant the requested startup adjustment. As discussed above, however, that is simply not the case. Despite Commerce’s repeated prodding, and notwithstanding the fact that ICDAS bore the burden of proof on the issue, ICDAS left the agency with only a very thin record on its request.

The record evidence — basically, little more than a single statement by ICDAS— was not sufficient to permit Commerce to properly consider ICDAS’ request for a startup adjustment. As Commerce stated, “without an explanation of how the claimed technical factors limited production levels, we are not able to determine whether ICDAS’s production levels were limited by technical factors associated with the initial phase of commercial production.” See Decision Memo at 80. But, contrary to ICDAS’ implication, that determination does not reflect Commerce’s invocation of “facts otherwise available” or adverse inference. Rather, it is simply an explanation that, because ICDAS failed to meet its burden of proof — in accordance with the Statement of Administrative Action and established agency practice — Commerce could not even evaluate ICDAS’ request, and was forced to deny the startup adjustment. See generally Domestic Producers Response Brief at 26-27; Def. Response Brief at 18.

Like its other challenges to Commerce’s denial of its requested startup adjustment, ICDAS’ claim that Commerce improperly resorted to “facts otherwise available” or adverse inference is similarly lacking in merit. In light of the record that the agency had before it, Commerce’s action denying the startup adjustment must be sustained.

C. Commerce’s Treatment of ICDAS’ Net Foreign Exchange Gain

In the course of the administrative review at issue here, Commerce conducted a cost investigation to determine whether ICDAS made sales of subject merchandise at prices below the cost of production. The statute defines cost of production as an amount equal to the sum of “the cost of materials and of fabrication or other processing ... employed in producing the foreign like product,” and includes “an amount for selling, general, and administrative expenses based on actual data pertaining to production and sales of the foreign like product.” 19 U.S.C. § 1677b(b)(3). However, the statute does not specify the method of determining those expenses for purposes of calculating cost of production. See generally Def. Response Brief at 22-23. Commerce has interpreted the statute to include financial expenses in the calculation of cost of production, and treats foreign exchange gains and losses as financial expenses. See, e.g., Silicomanganese from Brazil: Preliminary Results of Antidumping Duty Administrative Review, 68 Fed.Reg. 61,185, 61,187 (Oct. 27, 2003).

During the period of review in question, ICDAS realized a net foreign exchange gain, as a result of its foreign exchange income on sales, as well as its foreign exchange income on foreign currency bank checking accounts (which, ICDAS emphasizes, were “necessary for [the company’s] purchases and sales in foreign currencies”). See ICDAS Brief at 32; see also id. at 6; ICDAS Reply Brief at 10. In accordance with its standard practice, in calculating ICDAS’ cost of production here, Commerce treated the company’s foreign exchange gain within the category of “financial expenses,” and included it in the total financial expense ratio calculation in the Final Results. See generally Decision Memo at 86-88. Although ICDAS’ net foreign exchange gain exceeded its financial expenses, Commerce did not allow any of that gain to offset other expenses included in ICDAS’ cost of production, effectively “capping” ICDAS’ financial expenses at zero. See ICDAS Brief at 32-33; ICDAS Reply Brief at 10.

ICDAS asserts that its foreign exchange gains or losses do not result from separate cash management activities, but merely constitute adjustments necessary to ensure that other components of its cost of production are properly stated in a single currency. See generally ICDAS Brief at 6, 32-36; ICDAS Reply Brief at 10, 12-13. ICDAS therefore contests Commerce’s treatment of ICDAS’ net foreign exchange gain as part of “financial expenses” in the agency’s calculation of ICDAS’ cost of production. See generally ICDAS Brief at 2, 6, 32-35; ICDAS Reply Brief at 10-13. ICDAS further contends that Commerce erred in capping ICDAS’ net foreign exchange gain so as to set ICDAS’ “financial expenses” at zero. See generally ICDAS Brief at 2, 6, 35-37; ICDAS Reply Brief at 10-12. According to ICDAS, Commerce should have fully recognized the company’s net foreign exchange gain in the agency’s cost of production calculations. See generally ICDAS Brief at 2, 6, 32-33, 35-37; ICDAS Reply Brief at 10-12.

As discussed below, however, Commerce’s treatment of ICDAS’ net foreign exchange gain must be sustained.

1. Commerce’s Treatment of ICDAS’ Foreign Exchange Gain Within “Financial Expenses”

ICDAS’ threshold argument is that its foreign exchange gains or losses do not result from separate cash management activities, but — instead—constitute an adjustment necessary to ensure that other costs (such as the costs of manufacturing, sales, and general company operations) are properly stated in a single currency. See generally ICDAS Brief at 6, 32-36; ICDAS Reply Brief at 10-13. ICDAS maintains that its foreign exchange gains or losses therefore “should be fully accounted for” in the cost of production, rather than included in.“a discrete category of ‘financial expenses,’” which Commerce capped at zero. ICDAS Reply Brief at 11; see also id. at 10-12; ICDAS Brief at 2, 6, 32-33, 35-37.

ICDAS explains that its net foreign exchange gain during the period of review had “both a cost of manufacturing component and a sales-related component.” ICDAS Brief at 33. ICDAS’ manufacturing operations depend upon both raw material inputs and capital assets which are purchased in currencies other than ICDAS’ domestic currency, the Turkish Lira. According to ICDAS, consistent with generally accepted accounting principles, the company’s purchases of foreign raw materials and foreign capital assets are generally booked as accounts payable on the date they are received, using the appropriate foreign exchange rate on that date. However, actual payment generally is not made until some time later. As ICDAS notes, the change in the foreign exchange rate between the two dates results in a foreign exchange gain or loss for the company, related to its manufacturing operations. See generally ICDAS Brief at 33.

Like ICDAS’ manufacturing operations, ICDAS’ sales operations also produce foreign exchange gains or losses. When ICDAS makes a sale in foreign currency, the account receivable booked at the time of sale is converted to Turkish Lira on that date. However, the actual amount of Turkish Lira that ICDAS receives depends on the exchange rate when the buyer deposits its payment in foreign currency into ICDAS’ account. The difference between the exchange rate on the date of sale and the exchange rate on the date of payment results in a foreign exchange gain or loss for ICDAS, as a result of its sales operations. See generally ICDAS Brief at 33.

In addition to the foreign exchange gains and losses associated with ICDAS’ manufacturing and sales operations, foreign exchange gains and losses also result from ICDAS’ outstanding loans denominated in foreign currency. ICDAS asserts that such loans “relate to the general operation of the company,” and must be accounted for in Commerce’s cost of production calculations. See generally ICDAS Brief at 33-34. ICDAS explains:

[A]s a company conducting business in multiple currencies, ICDAS constantly faces the currency risk resulting from the mismatch between the currencies in which costs are incurred and revenues are earned. To mitigate this risk directly linked to its production and sale of merchandise, ICDAS ... incur[s] some of its debt in foreign currencies. Thus, when the Turkish Lira appreciates, and revenue in Turkish Lira terms declines relative to costs, the foreign currency loans provide a hedge and generate foreign currency gains to offset the foreign currency loss generated by the decline in sales income in Turkish Lira terms.

ICDAS Brief at 34.

ICDAS argues that foreign exchange gain or loss thus “is not a distinct element of cost.” ICDAS Brief at 34. ICDAS asserts that “[t]here is no check or account credit received from a financial institution for a foreign exchange gain, and no direct payment is made for a foreign exchange loss. Rather, the exchange rate gain or loss results from a series of accounting entries that are necessary and required by [generally accepted accounting principles] to ensure that all of the other elements of cost and income recorded by the company are properly stated in a single currency.” Id. ICDAS therefore characterizes net exchange rate gain or loss as “an overall adjustment necessary to ensure that all other costs for the calculation of ICDAS’ [cost of production] are properly stated in Turkish Lira.” Id.

As the Government correctly notes, however, ICDAS’ foreign exchange gains and losses are not inherent in its manufacturing and sales operations. Instead, they are the product of cash management decisions made by ICDAS in connection with its operations — cash management decisions which expose the company to those gains and losses. See Def. Response Brief at 9, 24; see generally Decision Memo at 86. In other words, ICDAS’ financing decisions — such as whether to pay for its purchases immediately, or to carry them as accounts payable; whether to make sales on a credit basis (i.e., as accounts receivable), or to require immediate payment; whether to borrow in a foreign currency, or in its own domestic currency; and whether to enter into foreign currency contracts — are related to, but separate and distinct from, the company’s manufacturing and sales activities. See generally Decision Memo at 86-87.

Thus, when ICDAS purchases raw materials or other inputs needed for production using a foreign currency, the company can — at the time of purchase — pay in cash immediately, based on the prevailing exchange rate between the Turkish Lira and the foreign currency. ICDAS could thereby avoid any exposure to exchange rate gains or losses in its manufacturing operations, if it wished to do so. See generally Def. Response Brief at 24; Decision Memo at 86-87. On the other hand, if ICDAS instead elects to pay for the purchase of the inputs at a later date (ie., to finance the purchase, or to set it up as an account payable, which is — in effect — buying on credit), the value of the inputs is booked in the equivalent domestic currency (i.e., in Turkish Lira) as of the date of sale, and not as of the date of actual payment. And the change in the foreign exchange rate between the date of purchase and the date of payment creates a foreign exchange gain or loss for the company. Accordingly, it is not the purchase transaction itself that results in a foreign exchange gain or loss, but — rather—ICDAS’ decision to defer payment and to finance the purchase instead. See generally Def. Response Brief at 24-25; Decision Memo at 86-87.

The same logic applies with equal force to ICDAS’ export sales transactions denominated in foreign currencies. As Commerce observed in the Final Results, ICDAS could demand immediate payment in such transactions, and would then avoid any exposure to foreign exchange rate gains or losses. On the other hand, if ICDAS instead elects to extend credit to customers in such transactions (i.e., by setting up accounts receivable from customers), the difference between the foreign exchange rate as of the date of sale and as of the date of payment results in a foreign exchange gain or loss for ICDAS. ICDAS’ decision to extend credit and thus to expose itself to foreign currency fluctuations in such transactions is a cash management decision. See generally Decision Memo at 87. Accordingly, it is not the sale transaction itself that results in a foreign exchange gain or loss, but — rather — ICDAS’ decision to extend credit to its customer (rather than requiring immediate payment).

In its briefs, ICDAS never directly confronts the fundamental logic of Commerce’s position, but instead repeatedly asserts (in essence) that the company’s foreign exchange gains or losses are “intertwined with and inseparable from ” its manufacturing, sales, and other operations. See ICDAS Reply Brief at 11; see also id. at 10, 12-13; ICDAS Brief at 6, 35. To the contrary, as Commerce explained in its Final Results, ICDAS could completely avoid exposure to foreign exchange risks, if it wished to do so, by making different cash management decisions — by, for example, making immediate payment for inputs that it purchases in foreign currencies, and by requiring that its customers make immediate payment in export sales, transactions denominated in foreign currencies. See Decision Memo at 86-87..

It is of no moment that, as ICDAS pointedly notes, the company is “not [in] the business of speculating with foreign currencies,” and that the company’s foreign exchange gains or losses associated with its operations “reflect the international currency market rate changes — which ICDAS can, neither predict nor control.” See ICDAS Reply Brief at 11. Although foreign exchange rates are not within IC-DAS’ control, Commerce’s point is that ICDAS can control whether or not to expose itself to the risk of gains or losses in such rates. ICDAS’ affirmative decisions to expose itself to such risks — for example, by delaying payment through the use of credit in purchasing inputs using foreign currencies, and by extending credit to its own customers in export sales transactions in foreign currencies- — are cash management decisions related to, but separate and distinct from, its underlying decisions to purchase inputs or to make sales. Commerce therefore treated ICDAS’ net foreign exchange gain within the category of financial expenses, for purposes of its cost of production calculations. It cannot be said that Commerce’s interpretation is an unreasonable interpretation of the statute. See Def. Response Brief at 22-23, 25-26 (citing Chevron, 467 U.S. at 842-43, 104 S.Ct. 2778).

In short, Commerce reasonably concluded in the Final Results that ICDAS’ net foreign exchange gain was “part of the company’s overall net financing expense.” See Decision Memo at 87. ICDAS’ arguments to the contrary are unavailing.

• 2. Commerce’s Decision Capping ICDAS’ Financial Expenses at Zero

Not only does ICDAS dispute Commerce’s treatment of the company’s foreign exchange gain within the category of “financial expenses” for purposes of calculating cost of production, but — in addition — ICDAS challenges Commerce’s decision to cap the company’s financial expenses at zero. See generally ICDAS Brief at 2, 6, 35-37; ICDAS Reply Brief at 10-13. According to ICDAS, Commerce should have fully recognized the company’s., net foreign exchange gain, by using all of that gain to offset expenses included in ICDAS’ cost of production. See generally ICDAS Brief at 2, 6, 35-37; ICDAS Reply Brief at 10-12.

As Commerce explained in the Final Results, there is typically a cost associated with financing a company’s operations, which is what the agency seeks to capture as part of “financial expenses.” See Decision Memo at 88. Commerce includes a cost of borrowing, as determined by various factors. If income is generated through those activities, the agency allows that income to be used to offset the cost of financing, up to the total financial expenses incurred. Id. But where — as here — the amount of relevant income exceeds the company’s financial expenses, Commerce recognizes that the company’s financial expenses were zero, and does not include a sum for financial costs in calculating the company’s cost of production. Id. Commerce does not allow financial expenses to be used to offset other expenses included in cost of production. As Commerce observed in the Final Results, “while certain types of income can legitimately be used to offset an expense, they can be used to do so only to the extent that there are costs to offset.” Id. It would therefore “be inappropriate ... to reduce other components of the [cost of production] by the net financing income,” as ICDAS urges. Id.; see generally Def. Response Brief at 25-27; Domestic Producers Response Brief at 29-33.

ICDAS argues that Commerce’s actions here run afoul of a new policy first articulated in Mushrooms from India, which concerns the agency’s treatment of foreign exchange gains or losses in calculating cost of production. See ICDAS Brief at 35-36 {citing Certain Preserved Mushrooms from India: Preliminary Results of Anti-dumping Duty Administrative Review, 68 Fed.Reg. 11,045, 11,048 (March 7, 2003) (“Mushrooms from India”)); ICDAS Reply Brief at 10-12 (same). Prior to Mushrooms from India, Commerce had required respondents to break down their foreign exchange gains and losses into separate components based on the source, and to include only those from certain sources in their reported costs. See generally Domestic Producers Response Brief at 29. But, in Mushrooms from India, Commerce explained that it was changing its practice: “Instead of splitting apart the foreign exchange gains and losses as reported in an entity’s financial statements, [Commerce] will normally include in the interest expense computation all foreign exchange gains and losses.” See Mushrooms from India, 68 Fed.Reg. at 11,048 (emphasis added).

In citing to Mushrooms from India, ICDAS conveys the impression that Commerce’s intent and practice is to fully recognize all foreign exchange gains and losses in calculating a respondent’s cost of production. See ICDAS Brief at 35-36; ICDAS Reply Brief at 10-12. However, as Commerce emphasized in the Final Results, Mushrooms from India did not address a net foreign exchange gain. See Decision Memo at 88. Instead, the case stands for the proposition that Commerce will include in its calculations all elements or components of foreign exchange gain and loss — not that the agency will necessarily offset a net gain against any and all other elements of cost of production. See generally Domestic Producers Response Brief at 30. Contrary to ICDAS’ claims, nothing in Mushrooms from India mandates that- Commerce recognize the entirety of ICDAS’ net foreign exchange gain by using all of it to offset expenses included in the company’s cost of production.

ICDAS also quarrels with Commerce’s reliance on Cinsa to support the agency’s decision in the Final Results to limit the use of ICDAS’ net foreign exchange gain to offsetting financial expenses, rather than recognizing the entirety of that gain and allowing it to offset other expenses included in the company’s cost of production. See ICDAS Brief at 36-37 (citing Cinsa S.A. de C.V. v. United States, 21 CIT 341, 351, 966 F.Supp. 1230, 1239-40 (1997)); Decision Memo at 88 (same). The court in Cinsa rejected the plaintiffs claim that Commerce had erred in allowing an offset of interest income only to the extent of interest expenses. The Cinsa court explained:

[E]xpenses by their' nature cannot produce a negative effect on the [cost of production]. Expenses, as a component of costs, cannot become a profit by the nature of their designation. Cinsa is effectively requesting that Commerce and the Court recognize a negative cost. Based on sound accounting and economic principles, the Court declines to accept a finding of negative costs when calculating [cost of production]. Interest expense, as a component of [cost of production], is a discrete expense account and as such, cannot provide an offset to any other expense accounts. Once the interest expense account is reduced to zero through the offset of interest income, interest expense and interest income [have] no further effect on the calculation of [cost of production]____[0]nce interest expense is reduced to zero, no further inquiry is necessary as Commerce cannot enter a profit into the calculation of [cost of production],

Cinsa, 21 CIT at 351, 966 F.Supp. at 1239-40. So too, in the case at bar, Commerce reasoned — by analogy to Cinsa — that “financial expenses, as a component of [cost of production], are a discrete expense account and as such, cannot provide an offset to any other expense accounts.” Decision Memo at 88; see generally Def. Response Brief at 26-27.

ICDAS argues that Cinsa is irrelevant here because, according to ICDAS, foreign exchange gains or losses are not a subset of financial expenses. See ICDAS Brief at 36-37. As discussed in section III.C.l above, however, Commerce properly concluded that ICDAS’ foreign exchange gains here were the product of its cash management decisions, and thus properly treated ICDAS’ net foreign exchange gain as part of the company’s overall net financing expense for purposes of Commerce’s cost of production calculations. See Decision Memo at 87; see generally Def. Response Brief at 27. ICDAS’ attempt to distinguish Cinsa is therefore futile.

In sum, Commerce properly decided to include ICDAS’ net foreign exchange gain in the financial expense ratio calculation, and to limit the recognition of that gain to offset only ICDAS’ financial expenses (effectively capping those expenses at zero). The agency’s determinations to that effect were consistent with agency practice, and were both supported by substantial evidence and otherwise in accordance with law. ICDAS’ arguments to the contrary must be rejected.

D. Commerce’s Decision on Remand to Use Invoice Date As Date of Sale

ICDAS strenuously objects to Commerce’s determination on remand that, for purposes of the agency’s anti-dumping analysis, the appropriate date of sale for ICDAS’ U.S. sales is the date of invoice. See generally ICDAS Response Brief at 1-3, 5-30; ICDAS Supp. Reply Brief, passim. In particular, ICDAS protests Commerce’s use of the same date of sale — invoice date — for both ICDAS’ U.S. and home market sales, asserting that the ways that the two types of sales are negotiated, orders are finalized, and merchandise is produced “differ markedly.” See ICDAS Supp. Reply Brief at 1-2 (quoting Circular Welded Non-Alloy Steel Pipe from the Republic of Korea; Final Results of Antidumping Duty Administrative Review, 63 Fed.Reg. 32,833, 32,835-36 (June 16,1998) (“Pipe from Korea”)).

ICDAS contends that, with one exception, the contract date — rather than the invoice date — best reflects the date on which ICDAS and its U.S. buyers reached a meeting of the minds on the material terms of sale, and should be used as the date of sale for purposes of Commerce’s analysis. See ICDAS Response Brief at 1, 3, 30; ICDAS Supp. Reply Brief at 13; see also Remand Results at 2, 13-14. As to that one exception, involving a price increase in a single contract, ICDAS asserts that the proper date of sale is invoice date (in effect, the date of contract amendment). See ICDAS Response Brief at 3, 20-21, 30; ICDAS Supp. Reply Brief at 1, 5 n. 3, 13; see also Remand Results at 2, 13-14.

The antidumping statute on its face does not specify the manner in which Commerce is to determine the date of sale. However, by enacting the Uruguay Rounds Agreements Act, Congress “incorporated the trade agreements adopted by the World Trade Organization at the Uruguay Round negotiations into United States law.” Allied Tube and Conduit Corp. v. United States, 24 CIT 1357, 1367-68, 127 F.Supp.2d 207, 216 (2000) (Allied Tube I). One such WTO agreement expressly provides that “[njormally, the date of sale would be the date of contract, purchase order, order confirmation or invoice, whichever establishes the material terms of sale.” See Agreement on Implementation of Article VI of the General Agreement on Tariffs and Trade 1994, Art. 2.4.1 n. 8 (emphasis added). Further, the Statement of Administrative Action accompanying the Uruguay Round Agreements Act expressly defines date of sale as the “date when the material terms of sale are established.” See Statement of Administrative Action, H.R. Doc. No. 103-316, at 810, reprinted in 1994 U.S.C.C.A.N. at 4153. Through the Uruguay Round Agreements Act and the Statement of Administrative Action, Congress thus “expressed its intent that, for antidumping purposes, the date of sale be flexible so as to accurately reflect the true date on which the material elements of sale were established.” Allied Tube I, 24 CIT at 1370, 127 F.Supp.2d at 219 (emphasis added).

Consonant with Congress’ intent as manifested in the Uruguay Round Agreements Act, Commerce promulgated a regulation on date of sale, which provides that the date of sale is invoice date, except where another date better reflects the date on which the material terms of sale were established:

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® (emphasis added). In the Preamble to its date of sale regulation, Commerce contracting parties reached a “meeting of the minds” on the material terms of sale:

If the Department is presented with satisfactory evidence that the material terms of sale are finally established on a date other than the date of invoice, the Department will use that alternative date as the date of sale. For example, in situations involving large custom-made merchandise in which the parties engage in formal negotiation and contracting procedures, the Department usually will use a date other than the date of invoice. However, the Department emphasizes that in these situations, the terms of sale must be firmly established and not merely proposed. A preliminary agreement on terms, even if reduced to writing, in an industry where renegotiation is common does not provide any reliable indication that the terms are truly “established” in the minds of the buyer and seller. This holds even if, for a particular sale, the terms were not renegotiated.

Antidumping Duties; Countervailing Duties: Final Rule, 62 Fed.Reg. 27,296, 27,349 (May 19, 1997) (Preamble) (emphasis added).

ICDAS acknowledges that Commerce’s regulations afford the agency discretion in determining the date of sale to be used in its antidumping margin calculations. But, according to ICDAS, Commerce abused that discretion in the Remand Results here — both by using a date of sale methodology that is inconsistent with Congressional intent and the governing regulation, as well as established agency practice, and by making factual findings that are not supported by substantial evidence in the record. See ICDAS Response Brief at 2-3, 17 n. 12; ICDAS Supp. Reply Brief at 1-2.

ICDAS explains that, during the administrative review at issue, it used different bases to report the dates of sales for its home market sales and its U.S. sales, because those sales were made pursuant to two distinct sales processes. ICDAS made tens of thousands of home market sales during the relevant period, most of which were relatively small and were filled out of ICDAS’ existing inventory. ICDAS did not negotiate and sign formal written contracts with its customers in its home market, and, instead, did business by phone or through written confirmation. For home market sales, ICDAS reported the earlier of the date or invoice or the date of shipment as the date of sale for use in Commerce’s margin calculations. See generally ICDAS Response Brief at 5 (and authorities cited there).

In contrast, ICDAS had far, far fewer U.S. sales, but each of those sales was larger in volume by orders of magnitude. Merchandise that ICDAS sold to the U.S. market was generally manufactured to order, was sold in inches (rather than millimeters), and was subject to other special requirements not applicable to ICDAS’ home sales. In light of the long lead time required to produce, pack, and ship such large special orders, U.S. sales were made pursuant to a deliberate and formal negotiation process, and formal written contracts were executed by the parties. As ICDAS explains, the formal written contract afforded protection to both parties, memorializing their meeting of the minds on the quantity and specifications of the merchandise to be supplied, the unit price, the shipment date, and other material terms of their deal. The U.S. buyer thus was assured of the supply of merchandise needed to fill orders from its customers. And ICDAS was assured that it had a customer for the specified quantity of its U.S.-customized merchandise before it began production of that merchandise. For purposes of Commerce’s antidumping analysis, ICDAS reported the contract date as the date of sale for its U.S. sales, on the theory that the contract date better reflected the date on which the material terms of those sales were established. See generally ICDAS Response Brief at 5-7 (and authorities cited there).

In the course of the verification process, Commerce confirmed that ICDAS’ “export sales process differs from the domestic sales process in that: 1) orders are always in written form; 2) a contract is signed after confirmation of the order; 3) merchandise is sold on a theoretical-weight basis; and 4) merchandise is always produced to order.” See ICDAS Sales Verification Report (Conf.Doc. No. 44) at 4. In its Preliminary Results, Commerce nevertheless used invoice date as the date of sale for ICDAS’ U.S. sales. See Preliminary Results, 70 Fed.Reg. at 23,992. In the Final Results, however, Commerce reversed its position on the date of sale issue. Concluding that the material terms of ICDAS’ U.S. sales were established on the date of contract, Commerce used contract date as the date of sale in its Final Results, and calculated a de minimis dumping margin of 0.16% for ICDAS. See Decision Memo at 29; Final Results, 70 Fed.Reg. at 67,666-67.

Several days after the Final Results were published, the Domestic Producers filed a ministerial error letter disputing Commerce’s ruling on the date of sale issue,, pointing to a price change as to one of ICDAS’ U.S. contracts. Soon thereafter, the Domestic Producers' filed suit, challenging, inter alia, Commerce’s use of contract date as the date of sale for ICDAS’ U.S. sales.

Commerce’s request for a voluntary remand on the date of sale issue was granted. See generally ICDAS Response Brief at 10 (and authorities cited there); Remand Results at 2-3. On remand, Commerce reversed course once again, finding that invoice date — rather than contract date — should be used as the date of sale for ICDAS’ U.S. sales. In its Remand Results, Commerce stated:

[W]e find that the price change, while limited to a single contract, related to a significant percentage of ICDAS’s U.S. entries during the [period of review]. Under these circumstances, we determine that the contract date does not represent the date on which the parties had a real “meeting of the minds” because the material terms of sale not only could be, but were altered after the date in the ordinary course of business.

Remand Results at 20. Using invoice date as the date of sale for ICDAS’ U.S. sales, Commerce recalculated ICDAS’ dumping margin to be 1.63% — a figure which exceeds the de minimis threshold, rendering ICDAS ineligible for revocation of the antidumping order. See Remand Results at 2-3, 5, 24-25.

ICDAS charges that the Remand Results “contravene[ ] the antidumping statute and Commerce’s regulations by using a rigid date of sale methodology that relies entirely on a single price change and the volume of sales affected by that price change to reach the conclusion that invoice date is the date of sale for all of ICDAS’ U.S. sales.” ICDAS Response Brief at 2. ICDAS argues that, “[although the [Remand Results] purport[] to establish the date on which the parties had a ‘real meeting of the minds,’ [the Remand Results] fail to employ an appropriate test for ascertaining whether the contracting parties reached a binding agreement.” Id. According to ICDAS, the Remand Results “do[ ] not examine the parties’ expectations about what was to be purchased, how much would be purchased, and how long it would take to produce.” Id. Moreover, ICDAS asserts, the Remand Results fail to “consider whether the subsequent course of conduct between the parties reveals that the parties understood that they were bound by the terms of contract.” Id. at 2-3.

ICDAS requests that the date of sale issue be remanded to Commerce once again, “with specific instructions that Commerce ascertain the point at which ICDAS and its U.S. customers had a meeting of the minds by considering the nature of the U.S. sales process and the course of conduct between the parties.” See ICDAS Response Brief at 3. ICDAS predicts that “[b]ased upon such an analysis, Commerce should find that the date of sale generally is the contract date,” and that “consistent with Commerce’s past decisions, the agency should treat the sole price change that occurred as an amendment to the contract, and use amendment date as date of sale only for that particular transaction.” Id.

As set forth more fully below, Commerce’s determination on remand to use invoice date as the date of sale for all of ICDAS’ U.S. sales is not supported by substantial evidence. Nor is that determination otherwise in accordance with law. Accordingly, the issue must be remanded to Commerce once again, for its reconsideration.

1. Whether the Remand Results Are In Accordance With Law

As ICDAS notes, under Commerce’s approach in the Remand Results here, even a single change to a material term in a single transaction — without regard to the nature of the change or the circumstances surrounding it — may require an across-the-board use of invoice date as the date of sale for all sales to all customers during the period of review. As ICDAS observes, such an approach is fundamentally at odds with the antidumping statute and regulations, as well as Commerce’s past practice, because it involves nothing more than a superficial, black-and-white, all-or-nothing determination whether there has been any change in any material term in any contract at issue, rather than a reasoned, case-specific, fact-intensive analysis as to when the parties had a meeting of the minds on the material terms of sale, which is what the law requires. See generally ICDAS Response Brief at 16-17; see also Remand Results at 20 (noting that appropriate date of sale is “the date on which the parties had a real ‘meeting of the minds’ ”). 34

The Government seeks to dismiss ICDAS’ challenge to the Remand Results out of hand, boldly asserting that Commerce’s determination must be sustained because the agency has “absolute” discretion in determining date of sale. See Def. Supp. Response Brief at 8. 35 Apparently relying on the phrasing of Commerce’s date of sale regulation (which provides for use of a date other than invoice date “if the Secretary is satisfied ” that use of the alternafive date is more appropriate), and on a single sentence in Hevensa, the Government maintains that “although Commerce may exercise its discretion to use a different time than the invoice date as the date of sale, because this is a discretionary act, it is not required to do so.” See generally Def. Supp. Response Brief at 7-8 (emphases added) (citing 19 C.F.R. § 351.401(i) (emphasis added); Hornos Electricos de Venezuela, S.A. v. United States, 27 CIT 1522, 1536, 285 F.Supp.2d 1353, 1367 (2003) (Hevensa)); see also Domestic Producers Reply Brief at 9 (quoting Hevensa, 27 CIT at 1536, 285 F.Supp.2d at 1366-67, for proposition that, even if material terms of sale are not subject to change, “discretion ... means that [Commerce] may use a date of sale other than the invoice date, but is not required to do so”).

The Government’s position on Commerce’s authority is plainly far too expansive. As a threshold matter, there is no area in which any government agency has “absolute,” unfettered discretion. See, e.g., Beardmore v. Dep’t of Agriculture, 761 F.2d 677, 679 (Fed.Cir.1985) (holding that “an agency’s discretion is not unlimited”). Certainly no court in any international trade case (including Hevensa) has held that Commerce has “absolute,” unbridled discretion to apply invoice date as the date of sale across-the-board, with no regard for the record evidence in a case. Notwithstanding the Government’s implications, there is nothing in Hevensa to suggest that Commerce is free to arbitrarily choose to use as the date of sale some date other than the date when the material terms of sale were established. In other words, if a particular date is demonstrated to be the date when the material terms of sale were established, Commerce has no discretion to simply ignore that date and choose to use some other date as the date of sale.

Similarly, contrary to the Domestic Producers’ claims, neither Commerce’s date of sale regulation nor the Preamble to the agency’s antidumping regulations expresses a “strong preference” for use of invoice date as a respondent’s date of sale. See Domestic Producers Reply Brief at 9 (asserting that agency regulations “express a strong preference” for invoice date, and also discussing Preamble to regulations). In fact, neither Congress nor the agency in its regulations expresses any “preference” at all on the matter — “strong” or otherwise. 36

Rather than the “strong preference” claimed by the Domestic Producers, Commerce’s date of sale regulation and the Preamble to the agency’s antidumping regulations establish only a “rebuttable presumption” — and, indeed, one that has been successfully rebutted in numerous cases in the past, as illustrated in the discussion below. See, e.g., Colakoglu Metalurji A. S. v. United States, 29 CIT 1238, 1240, 394 F.Supp.2d 1379, 1380 (2005) (indicating that a “plain reading” of date of sale regulation indicates that it establishes only a “rebuttable presumption”); Remand Results at 3 (noting that date of sale regulation “provide[s] for a rebuttable presumption”). Thus, as the Preamble to Commerce’s date of sale regulation explains, where the agency “is presented with satisfactory evidence that the material terms of sale are ... established on a date other than the date of invoice, the Department will use that alternative date as the date of sale.” See Antidumping Duties; Countervailing Duties: Final Rule, 62 Fed.Reg. at 27,349 (Preamble) (emphasis added).

Equally unfounded is the Domestic Producers’ assertion (also reflected in the Remand Results) that judicial precedent restricts Commerce’s use of a date of sale other than invoice date to “unusual” situations. See Domestic Producers Reply Brief at 9 (citing Thai Pineapple Canning Indus. Corp., Ltd. v. United States, 24 CIT 107, 109, 2000 WL 174986 (2000), rev’d on other grounds, 273 F.3d 1077 (Fed.Cir.2001)); Remand Results at 19. 37 As even the Government acknowledges, the Preamble to the agency’s date of sale regulation expressly states that where — as here — a case “involv[es] large custom-made merchandise,” and where — as here — “the parties engage in formal negotiation and contracting procedures,” Commerce “usually will use a date other than the date of invoice” as the date of sale. See Def. Supp. Response Brief at 7 (quoting Anti-dumping Duties; Countervailing Duties: Final Rule, 62 Fed.Reg. at 27,349 (Preamble)). And, again, as noted above and discussed below, Commerce has used a date other than invoice date as the date of sale in numerous cases in the past. 38

The Domestic Producers further contend that — pursuant to Commerce’s assertedly “consistent practice” and judicial precedent — a foreign producer seeking to rebut the regulatory presumption that the invoice date is the date of sale must satisfy two criteria: “1) ‘produc[e] sufficient evidence,’ ie., establish a complete record that includes all relevant sales documents for all reported sales; and 2) satisfy the Secretary that the material terms were established at a date other than invoice date.” See Domestic Producers Reply Brief at 10 (quoting Allied Tube and Conduit Corp. v. United States, 25 CIT 23, 25, 132 F.Supp.2d 1087, 1090 (2001) (Allied Tube II); citing Hevensa, 27 CIT at 1537, 285 F.Supp.2d at 1366-67). But the Domestic Producers’ formulation of what is required to rebut the regulatory presumption misstates the existing state of the law.

The Domestic Producers cite no authority for their assertion that Commerce is required to use invoice date as the date of sale in a case unless the record in that ease “includes all relevant sales documents for all reported sales.” See Domestic Producers Reply Brief at 10 (emphasis added). Nor does it appear that the Domestic Producers can cite any such authority; independent research discloses no precedent to that effect. ICDAS, of course, bears the burden of proof on the date of sale issue; and where, as here, the applicable standard is whether “the Secretary is satisfied,” a determination as to the sufficiency of proof ordinarily “lies primarily within Commerce’s discretion.” See 19 C.F.R. § 351.401(i); Koyo Seiko Co., Ltd. v. United States, 551 F.3d 1286, 1292 (Fed.Cir.2008). Nevertheless, like all other agencies, Commerce is generally prohibited from “treating] similar situations in dissimilar ways.” See Burinskas v. Nat’l Labor Relations Bd., 357 F.2d 822, 827 (D.C.Cir.1966) (cited in Nakornthai Strip Mill Public Co., Ltd. v. United States, 32 CIT -, -, 587 F.Supp.2d 1303, 1307 (2008)). And a review of various cases in which Commerce has used a date of sale other than invoice date suggests that a number (if not all) of them involved administrative records that did not meet the Domestic Producers’ asserted criterion. Indeed, in the Final Results in this case, Commerce was satisfied with the existing record as a basis for its determination to use contract date as the date of sale. 39

Like their first criterion (discussed immediately above), the Domestic Producers’ second criterion similarly overstates the evidence required to rebut the regulatory presumption concerning date of sale. According to the Domestic Producers, the second criterion — which requires a demonstration that “the material terms were established at a date other than invoice date” — is “two-fold”: “a party must a) demonstrate that there were no actual changes to the material terms between the proposed date and the invoice date; and b) demonstrate that, in the absence of actual changes, the material terms were also not subject to change.” See Domestic Producers Reply Brief at 10 (citing Allied Tube II, 25 CIT at 25, 132 F.Supp.2d at 1090).

The Domestic Producers’ second criterion basically amounts to a claim that the regulatory presumption of invoice date can be overcome only if a foreign producer establishes that there were no changes whatsoever to any material term of any contract at issue (and, moreover, that there was no possibility of any such change). That position, however, is patently incorrect. As Commerce itself candidly conceded in the Remand Results here, “a single change in price does not automatically disqualify contract date from selection as the proper date of sale.” See Remand Results at 20. 40 And, in fact, Commerce has used a date other than invoice date as the date of sale in numerous cases in the past, notwithstanding changes in price or other material contract terms. See, e.g., Pipe from Korea, 63 Fed. Reg. at 32,835-36; Issues and Decision Memorandum for the Antidumping Duty Investigation of Sulfanilic Acid from Portugal; Final Determination, 2002 WL 31493754 (Sept. 18, 2002) (“Sulfanilic Acid from Portugal”), at comment 1; Issues and Decision Memorandum for the Final Re- suits of the Administrative Review of the Antidumping Duty Order on Low Enriched Uranium from France (2003-2004), 2005 WL 2305751 (Sept. 14, 2005) (“Uranium from France”), at comment 11; Issues and Decision Memorandum for the Final Results of the Fourth Administrative Review of Steel Concrete Reinforcing Bars from Latvia, 2006 WL 3702620 (Dec. 13, 2006) (“Rebar from Latvia”), at comment 2.

As ICDAS puts it, the “key element to consider” in determining date of salé is which date best reflects the point at which the parties had a meeting of the minds on the material terms of sale — not whether there is evidence of even a single change in a single material term of a single contract. See ICDAS Response Brief at 13 (quoting Sulfanilic Acid from Portugal, 2002 WL 31493754, at comment 1).

Although Commerce’s date of sale regulation reflects a presumption that the date of invoice will be the date of sale, the same regulation specifically provides for Commerce’s use of a different date where that other date “better reflects the date on which the exporter or producer establishes the material terms of sale.” See 19 C.F.R. § 351.401®. Accordingly, although Commerce “normally” presumes that invoice date is the date of sale, the invoice date in fact is merely the starting point of Commerce’s analysis. It is by no means intended to “foreclose® the possibility that another date could be chosen as the date of sale.” Allied Tube I, 24 CIT at 1371, 127 F.Supp.2d at 219; see also Colakoglu, 29 CIT at 1240, 394 F.Supp.2d at 1380 (explaining that date of sale regulation merely establishes “rebuttable presumption” favoring invoice date).

Commerce itself has labeled as “untenable” the “blanket use [of invoice date] as the date of sale in an antidumping analysis” where “the invoice date does not reasonably approximate the date on which the material terms of the sale were [established].” See Pipe from Korea, 63 Fed. Reg. at 32,835-36. And Commerce itself has recognized that its regulations do not tie the agency’s hands, but instead afford Commerce the “flexibility” needed to determine, and to use in its analysis as the date of sale, that date which best reflects the date on which the parties reached a meeting of the minds on the material terms of sale. See Issues and Decision Memorandum for the 1997-1998 Administrative Review of Circular Welded Non-Alloy Steel Pipe from Mexico: Final Results of Antidumping Duty Administrative Review, 2000 WL 777746 (June 15, 2000) (“Pipe from Mexico”), at Hylsa comment 1.

Flexibility in Commerce’s date of sale analyses is more than a mere regulatory preference; it rises to the level of a statutory mandate. Allied Tube I, 24 CIT at 1367-69, 127 F.Supp.2d at 216-17. As discussed above, in enacting the Uruguay Round Agreements Act, Congress made clear — both through the Statement of Administrative Action and through its incorporation into U.S. law of the date of sale provision in the Agreement on the Implementation of Article VI of the General Agreements on Tariffs and Trade 1994— that the date of sale is to be the date on which “the material terms of sale” are established; See Allied Tube I, 24 CIT at 1367-68, 127 F.Supp.2d at 216-17 (explaining, inter alia, that both the WTO trade agreement and Congress’ interpretation of the agreement in the Statement of Administrative Action unambiguously provide that “the date of sale is to be the date on which the material terms of sale are established”). In other words, rather than endorsing a mechanistic methodology conclusively establishing invoice date as the date of sale whenever there is even a single change in a material term of a single contract, Congress instead “expressed its intent that, for antidumping purposes, the date of sale be flexible so as to accurately reflect the true date on which the material elements of sale were established.” See Allied Tube I, 24 CIT at 1370, 127 F.Supp.2d at 219. 41

Indeed, flexibility in Commerce’s date of sale analysis is a natural corollary of Commerce’s overarching obligation to determine dumping margins as accurately as possible. See, e.g., NTN Bearing Corp. v. United States, 74 F.3d 1204, 1208 (Fed.Cir.1995); Koyo Seiko Co. v. United States, 36 F.3d 1565, 1573 (Fed.Cir.1994); Allied Tube I, 24 CIT at 1370-71, 127 F.Supp.2d at 218-19 (in context of review of agency determination on appropriate date of sale, emphasizing the “need to calculate anti-dumping duty margins on a fair and equitable basis”).

Consistent with Congressional intent, Commerce in the past has repeatedly “recognize^] the need for flexibility in those circumstances in which an alternative date better reflects the date of sale.” See Issues and Decision Memorandum for the Antidumping Duty Administrative Review of Certain Welded Carbon Steel Pipes and Tubes from Thailand, 65 ITADOC 60910 (Oct. 4, 2000) (“Pipes and Tubes from Thailand”), at comment 1; 42 see also Pipe from Korea, 63 Fed.Reg. at 32,835 (explaining that, “[i]n granting this flexibility, the [date of sale] regulations anticipate the possibility of inappropriate comparisons via the strict use of invoice date as the date of sale”). Commerce therefore has a “well-established and long-standing practice” of looking beyond the invoice date to the parties’ actual course of conduct, as well as the parties’ expectations concerning the transaction, to determine whether an earlier date — such as the contract date — represents the point at which the parties reached a meeting of the minds on the material terms of sale. See Sulfanilic Acid from Portugal, 2002 WL 31493754, at comment l. 43

Commerce’s date of sale analysis in the Remand Results here cannot be squared with the relatively long line of cases in which Commerce has determined that proof that material contract terms could (or even did) change nevertheless does not automatically warrant use of invoice date as the date of sale. 44 That line of cases illustrates that a change in a material contract term, while relevant, does not end Commerce’s date of sale analysis. 45 Commerce is still required to undertake a factual analysis of the expectations and conduct of the contracting parties, to ascertain when they reached a true meeting of the minds on the material terms of sale.

In Pipe from Korea, for example, the foreign respondents argued that Commerce should use invoice date as the date of sale for their U.S. sales, because “the possibility for change [of the material terms of sale] exists and sometimes does occur.” See Pipe from Korea, 63 Fed.Reg. at 32,835. Notwithstanding the fact that material terms of sale not only could be, but in fact were, changed, Commerce nonetheless did not use invoice date as the date of sale in that case. Id., 63 Fed.Reg. at 32,836. Instead, Commerce analyzed the facts surrounding the change in contract terms in the context of the parties’ expectations, focusing particularly on critical differences in the respondent’s sales processes for U.S. and home market sales:

In this case, the sales processes for U.S. and home market sales differ markedly. Sales in the home market are typically out of inventory with the purchase order/contract, invoice and shipment dates all occurring within a relatively short period of time. In contrast, U.S. sales are usually conducted on a made-to-order basis (CEP sales out of inventory being an exception). The material terms of sale in the U.S. are set on the contract date and any subsequent changes are usually immaterial in nature or, if material, rarely occur .... As can be seen from the foregoing, “invoice” dates in both markets, while the same in name, are materially quite different.... Notwithstanding the respondents’ comment that the terms of sale are subject to change and that, therefore, the final terms are not known until the date of invoice, we find that, in this case, there is no information on the record indicating that the material terms of sale change frequently enough on U.S. sales so as to give both buyers and sellers any expectation that the final terms will differ from those agreed to in the contract.

Pipe from Korea, 63 Fed.Reg. at 32,836 (emphases added). Based on its analysis, Commerce determined in Pipe from Korea that the contract date was the proper date of sale,; reasoning that — due to the sales process used for the- “made-to-order” merchandise, as well as the low frequency of changes in the case — the parties to the sales at issue could not have expected that the material terms of their deals were not fixed as of the date of contract.

The record here indicates that — like the respondent in Pipe from Korea — ICDAS too has markedly different sales processes for its U.S. and home market sales. 46 Similarly, given the extraordinarily low frequency of change in the instant ■ ease-^-a single price increase, in a single contract, over the course of a full year — it is not at all clear that either ICDAS or its buyers could have had any reasonable expectation that the material terms of sale for the made-to-order rebar that ICDAS sold to the U.S. would vary in any respect from those specified in the parties’ contracts 47 (just as Commerce found that the parties in Pipe from Korea could not have expected that they would not be bound by the terms of their contracts). 48

Like the underlying facts of Pipe from Korea, the facts of Sulfanilic Acid from Portugal also closely parallel the facts of the case at bar in certain key respects. Commerce’s decision in that case further illustrates the importance of “look[ing] to the course of conduct between the parties in evaluating whether a written document represents a binding agreement.” See Sulfanilic Acid from Portugal, 2002 WL 31493754, at comment 1 (citing Polyvinyl Alcohol from Taiwan (final determination)).

In Sulfanilic Acid from Portugal, the original contract was amended twice to increase the price. However, rather than mechanistically applying the invoice date as the date of sale, Commerce there reviewed the parties’ course of conduct to ascertain the parties’ expectations concerning the contract. Commerce found that, despite the material “modifications to the original sales contract,” “the parties acted in a manner consistent with a ‘meeting of the minds’ to be bound by the terms of the original contract.” See Sulfanilic Acid from Portugal, 2002 WL 31493754, at comment 1. To account for the modification, Commerce treated the price changes as amendments to the original contract, and treated the dates of the contract amendments as the dates of sale in its antidumping analysis. Id. Nowhere has Commerce explained why such an approach would be inappropriate in this case. 49

In short, Commerce here has failed to justify its use of the same date of sale for both ICDAS’ home market sales and its U.S. sales, given that the sales and production processes used for the two markets “differ markedly.” See Pipe from Korea, 63 Fed.Reg. at 32,835-36. Nor has Commerce adequately explained why a single price increase as to a single contract warrants the blanket application of invoice date as the date of sale for all of ICDAS’ U.S. sales to all of its customers. Cases such as Pipe from Korea suggest that, notwithstanding the price increase, Com merce should have used contract date as the date of sale for all of ICDAS’ U.S.. sales, including those under the contract affected by the price change. Alternatively, Commerce conceivably might have followed its approach in Sulfanilic Acid from Portugal (and other similar cases), and treated ICDAS’ price change as an amendment to the original contract. 50

Here, Commerce has failed even to consider, much less explain, how ICDAS’ eleventh-hour efforts to negotiate a modest price increase as to a single contract— after ICDAS had substantially performed its contractual obligations — could have upset the expectations of ICDAS and its U.S. buyers that the material terms of ICDAS’ U.S. sales were established with the execution of their legally-binding contracts.

Although the Remand Results pay lip service to determining “the date on which the material terms of sale (ie., price and quantity) [were] established” (that is, to ascertaining “the date on which the parties had a real ‘meeting of the minds’ ”), that was clearly not the object of the analysis that Commerce conducted on remand. See Remand Results at 18, 20. Instead, as the Remand Results themselves state, Commerce confined its analysis on remand to the much narrower question “whether ICDAS’s material terms of sale changed after the contract date.” See Remand Results at 21; see also id. (stating that “the issue at hand is ... whether the material terms of sale changed”). To be sure, evidence that “material terms of sale changed after the contract date” is relevant to determining “the date on which the parties had a real ‘meeting of the minds’ ” (that is, to ascertaining “the date on which the material terms of sale (ie., price and quantity) [were] established”). But Commerce’s date of sale analysis cannot begin and end with a simple, cut-and-dried, “yes/ no” determination as to whether there was any change to any material term of any contract at issue.

In sum, a second remand is necessary here, to permit Commerce to revise its date of sale analysis for ICDAS’ U.S. sales. Under both settled law and agency practice, that analysis does not, and should not, hinge on a single change in price or quantity, or the volume of sales affected by that change, with no regard for any other relevant facts. See Pipe from Mexico, 2000 WL 777746, at Hylsa comment 1 (noting, in principle, that the fact “that the material terms of sale can change up to the time of invoice is not dispositive”). Even when there is evidence of change in a material term, Commerce still must consider whether — as evidenced by their understanding of the sales process, as well as their course of conduct — the parties had the expectation that the material terms of sale were fixed on the date of contract.

On remand, Commerce shall focus its date of sale analysis on “the date on which the parties had a real ‘meeting of the minds,’ ” to ascertain “the date on which the material terms of sale (ie., price and quantity) [were] established.” See Remand Results at 18, 20. In its analysis, Commerce shall consider, inter alia, the differences between the sales and production processes for ICDAS’ U.S. sales compared to its home market sales, the parties’ understanding of the sales process, the parties’ course of conduct, and the practical effect (if any) of the single price increase on parties’ expectations as to the legally binding nature of their contracts, as well as all other relevant facts.

2. Whether the Remand Results Are Supported by Substantial Evidence

For the reasons outlined immediately above, Commerce’s Remand Results are not in accordance with law. But ICDAS contends that the Remand Results are further flawed because they are not supported by substantial evidence in the record. See generally ICDAS Response Brief at 2-3, 21-30; ICDAS Supp. Reply Brief at 1-2, 6-13.

On remand, ICDAS filed detailed comments, arguing, inter alia, that the nature of ICDAS’ U.S. sales process and the course of conduct between the parties compel the conclusion that the contract date is the date on which the parties reached a meeting of the minds on the material terms of sale. See generally ICDAS Remand Comments (Remand Pub. Doe. No. 4). However, in the Remand Results, Commerce essentially ignored the substance of ICDAS’ comments, on the theory that the existence of a single increase in price obviated the need to consider the comments as well as the underlying record evidence that ICDAS cited. See, e.g., Remand Results at 21 (stating that “[t]he relevant issue at hand is whether ICDAS’s material terms of sale changed after the contract date”); id. (asserting that “the issue at hand is ... whether the material terms of sale changed”). Apart from emphasizing that single price change, the Remand Results are largely limited to Commerce’s (generally unsuccessful) efforts to distinguish the cases that ICDAS cites. See Remand Results at 22-24. 51 The Remand Results reflect virtually no effort on the part of Commerce to grapple with the significant body of evidence cited by ICDAS in support of its position. 52

As noted above, evidence of a price change is unquestionably relevant to determining the date on which contracting parties had a “meeting of the minds,” establishing the material terms of sale. But, as a matter of law, discerning whether or not there was such a change does not conclude Commerce’s date of sale inquiry. Moreover, any evaluation of the record evidence supporting Commerce’s conclusion on date of sale must necessarily “take into account whatever in the record fairly detracts from [the] weight of that evidence,” including “contradictory evidence or evidence from which conflicting inferences could be drawn.” Suramerica, 44 F.3d at 985 (quoting Universal Camera, 340 U.S. at 487-88, 71 S.Ct. 456).

Because Commerce failed to properly consider ICDAS’ remand comments and the voluminous evidence on which those comments relied, Commerce’s remand de termination to use invoice date as the date of sale for both ICDAS’ U.S. sales and its home market sales cannot be sustained on the strength of the existing record. As discussed in greater detail below, Commerce failed to give appropriate consideration to certain critical ways in which the two sales processes “differ markedly.” See Pipe from Korea, 63 Fed.Reg. at 32,-836.

a. The Use of Formal Negotiation and Contracting Procedures

As ICDAS correctly notes, the use of formal negotiation and contracting procedures is conduct that bears directly on the expectations of the parties to a sale. ICDAS underscores the critical differences between the process used to negotiate and finalize ICDAS’ U.S. sales and its process for handling its home market sales. See generally ICDAS Response Brief at 5-9, 23-25; ICDAS Supp. Reply Brief at 7-9. 53

The evidence indicates that ICDAS and its home market buyers did not expect the terms of a sale to become final until the invoice was issued, because home market sales were made through an informal negotiation process. Home market sales were negotiated by phone or fax over a very short period of time; the merchandise was generally shipped out of inventory within one day of order confirmation; and, because there were no formal written contracts between ICDAS and its domestic buyers, the invoice “represent[ed] the first document generated chronologically in the sales process that reflects the actual price and quantity of a sale.” See ICDAS Sales Verification Report (Conf.Doc. No. 44) at 4. Under such circumstances, Commerce’s use of invoice date as the date of sale for ICDAS’ home market sales was entirely proper.

But the record evidence shows that ICDAS’ U.S. sales involved a markedly different process. See ICDAS Sales Verification Report (Conf.Doc. No. 44) at 4. That evidence casts a long shadow over Commerce’s use of invoice date as the date of sale for ICDAS’ U.S. sales, and suggests strongly that the proper date , of sale for ICDAS’ U.S. sales is contract date (or, in the case of the price increase, the date of invoice, or contract amendment).

Specifically, the record evidence on ICDAS’ U.S. sales process establishes that— following an initial, informal exchange of information — the buyer would submit a formal order, “always in written form.” See ICDAS Sales Verification Report (Conf.Doc. No. 44) at 4. Thereafter, ICDAS would make a formal written offer. Id. The parties would then come to a preliminary agreement on price, quantity, and shipping terms. Id.; see also Issues and Decision Memorandum for the Final Results of the Antidumping Duty Administrative Review: Certain Welded Carbon Steel Pipe and Tube from Turkey, 2005 WL 3417291 (Dec. 12, 2005), at comment 1 (finding date of e-mail order confirmations to be appropriate date of sale, because negotiations on material terms of sale occurred prior to written confirmation).

Significantly, the evidence demonstrates that ICDAS and its U.S. buyer always went beyond their preliminary agreement, to enter into a formal written contract, signed by both parties. See ICDAS Sales Verification Report (Conf.Doc. No. 44) at 4. That contract memorialized the parties’ meeting of the minds on all essential terms of the contract, and more — including the quantity and dimensions of the rebar to be supplied, the price, the shipment date, the packing requirements, the method of payment, the mode of shipment, and the risk of damage or loss. In short, the record evidence cited by ICDAS indicates strongly that the material terms of its U.S. sales were indeed “firmly established and not merely proposed” in the contracts between ICDAS. and its buyers. See Antidumping Duties; Countervailing Duties: Final Rule, 62 Fed.Reg. at 27,349 (Preamble). Neither the Remand Results nor the briefs filed with the court by the Government and the Domestic Producers cites any evidence or authority even hinting that the language of the formal written contracts between ICDAS and its U.S. buyers was anything other than clear, unambiguous, and legally binding. 54

Because Commerce failed to properly consider' the record evidence indicating that the formal written contracts between ICDAS and its U.S. customers documented the parties’ genuine meeting of the minds on all material terms of sale as of the date of contract, Commerce’s determination on remand to use invoice date as the date of sale is not supported by substantial evidence. See Antidumping Duties; Countervailing Duties: Final Rule, 62 Fed.Reg. at 27,349 (Preamble) (explaining that “in situations, involving large custom-made merchandise in which the parties engage in formal negotiation and contracting procedures, the Department usually will use a date other than the date of invoice” as the date of sale) (emphasis added).

b. The Lead Time Required to Produce Mad&-to-Order Merchandise

Compounding its failure to properly consider the record evidence on the differences between the sales process used in ICDAS’ home market sales and its process in U.S. sales (and, in particular, the evidence on ICDAS’ consistent use of formal written contracts for its U.S. sales), Commerce similarly failed to properly consider the record evidence on the nature of the merchandise that ICDAS sold in the U.S. market and the process required to produce that merchandise, as well as the impact of those factors on the expectations of the contracting parties. See generally ICDAS Response Brief at 5-9, 25-27; ICDAS Supp. Reply Brief at 9-12. For example, as ICDAS correctly observes, when merchandise must be made-to-order, a seller is generally unwilling to incur costs and begin production unless and until the material terms of sale have been established. The situation is very different for sales made out of inventory.

The undisputed evidence indicates that the average quantity per invoice for ICDAS’ U.S. sales was exponentially greater than that for the average home market sale. Moreover, the average lead time needed to complete those massive U.S. orders was orders of magnitude longer than for home market sales. See ICDAS Response Brief at 6-8, 25-26; ICDAS Supp. Reply Brief at 9. 55 The record thus supports ICDAS’ assertion that — in sharp contrast to ICDAS’ home market sales— the parties in ICDAS’ U.S. sales transactions required that the material terms of sale be fixed on the date of contract, because they understood that they were contracting for massive shipments of made-to-order rebar that would require significant time and expense to produce. See Pipe from Korea, 63 Fed.Reg. at 32,836 (using invoice date as the date of sale for home market sales typically filled from inventory, while using contract date as the date of sale for U.S. sales “usually conducted on a made-to-order basis”); Rebar from Latvia, 2006 WL 3702620, at comment 2 (find that lead time of one week to one month was significant aspect of respondent’s sales process). 56

Pointing to the record evidence, ICDAS explains that both ICDAS and its U.S. customers required the certainty early in the process that only a formal, written, legally-binding contract could provide, because they knew that the rebar had to be, inter alia, manufactured-to-order and rolled in inches (rather than millimeters); tested; marked differently than rebar sold in the home market; and packed for export. See generally ICDAS Response Brief at 27-28.

The record evidence supports ICDAS’ assertion that certainty was vital to ICDAS’ U.S. buyers, who depended on ICDAS to supply the contracted-for rebar that they required to meet the needs of their own customers. To fill their orders on time, ICDAS’ U.S. buyers expected ICDAS to begin production with sufficient lead time to produce, pack, and ship the contracted-for rebar in conformance with the specific terms of the contract.

Certainty was equally critical to ICDAS, because — as the record evidence indicates — ICDAS lacked the ability to stock in advance the thousands of metric tons of specific sizes of rebar sold to the U.S. market during the period of review. To fill U.S. sales, ICDAS therefore had to incur the cost of production for contracts with values in the millions of dollars. ICDAS relied on the enforceability of its contracts, because it was keenly aware that — if a U.S. buyer sought to renege after the contracted-for rebar had been produced — it would be commercially infeasible for ICDAS either to resell in the home market rebar that had been specifically produced to U.S. dimensions or to warehouse such massive quantities of merchandise for future U.S. sales. See Issues and Decision Memorandum for the Final Results of the Antidumping Duty Administrative Review of Certain Hot-Rolled Carbon Steel Flat Products from Thailand, 2004 WL 3524375 (April 13, 2004), at comment 1 (finding contract date the appropriate date for date of salé, where merchandise for U.S. customers was manufactured-to-order). As the record evidence establishes — for these sound business reasons— ICDAS’ standard practice has been to begin production for a U.S. sale only after ICDAS has in hand a formal, written, signed contract. See ICDAS Sales Verification Report (Conf.Doc. No. 44) at 4. 57

The record evidence cited by ICDAS and summarized above strongly supports ICDAS’ assertion that — as to its U.S. transactions — the material terms of sale were firmly established as of the date of contract. Factors such as the made-to-order nature of the merchandise for U.S. sales (particularly combined with ICDAS’ inability to stock significant quantities of that merchandise in inventory), and the lead time required to produce that merchandise, lend credence to ICDAS’ case. See Antidumping Duties; Countervailing Duties: Final Rule, 62 Fed.Reg. at 27,349 (Preamble) (explaining. that “in situations involving large custom-made merchandise in which the parties engage in formal negotiation and contracting procedures, the Department usually will use a date other than the date of invoice” as the date of sale) (emphases added); Rebar from Latvia, 2006 WL 3702620, at comment 2 (using contract date or contract amendment date as date of sale, at least in part due to lead time of one week to one month). The Remand Results say relatively little about ICDAS’ evidence on this point; and what the Remand Results do say does little to diminish the force of ICDAS’ argument, and even less to support Commerce’s conclusion.

Because Commerce failed to properly consider the record evidence concerning how — as a practical matter — the lead time required to produce made-to-order rebar affected the contracting parties’ expectations as to when the terms of sale became final, Commerce’s determination on remand to use invoice date as the date of sale for ICDAS’ U.S. sales is not supported by substantial evidence.

c. The Contracting Parties’ Course of Conduct and the Price Increase

Evidence of the course of conduct between ICDAS and its U.S. customers further buttresses ICDAS’ assertion that the parties understood that they were bound by the terms of their contracts, which memorialized their agreement on material terms of sale. But, once again, the Remand Results gave the evidence short shrift. See generally ICDAS Response Brief at 28-30; ICDAS Supp. Reply Brief at 11-12.

ICDAS points to the sales traces of ICDAS’ U.S. sales that were selected by Commerce for verification as evidence that the terms of sale were firmly established on the date of contract. See ICDAS Response Brief at 28. As ICDAS notes, the sales traces demonstrate that ICDAS took all necessary steps to honor the terms of its contracts, incurring significant costs. Specifically, ICDAS produced the contracted-for rebar, tested the rebar, trucked the tens of thousands of tons of the merchandise to the port, paid brokerage and handling fees, made arrangements for marine insurance and a ship to transport the merchandise to the United States, and took all actions necessary to comply with applicable Turkish customs and loading requirements. As ICDAS succinctly observes: “This is not the behavior of a producer that believes it is free to change or breach its contracts.” See ICDAS Response Brief at 28 (emphasis added).

ICDAS emphasizes that its course of conduct during the sale in which there was a price increase was consistent with its conduct during its other sales, and demonstrates that both ICDAS and the U.S. customer understood that the terms of their original contract were firm and binding, notwithstanding ICDAS’ subsequent successful efforts to negotiate a modest price increase. See ICDAS Response Brief at 28-29 (citing Issues and Decision Memorandum for the Final Determination in the Antidumping Duty Investigation of Certain Large Diameter Carbon and Alloy Seamless Standard, Line and Pressure Pipe from Mexico — April 1, 1998, through March 31, 1999, 2000 WL 959479 (June 26, 2000), at comment 2 (using date of sales acknowledgment as date of sale, even though producer admitted possibility- of change)). , ■■

As ICDAS notes, the facts of the sale in question largely parallel those of every other U.S: sale that ICDAS made during the period of review. See ICDAS Response Brief at 28-29. The record evidence establishes that ICDAS entered into a formal, written, legally-binding contract with its customer, establishing the material terms of the sale. Pursuant to the contract, ICDAS was required to produce the rebar to the agreed-upon specifications; inspect and test the physical and chemical composition of the rebar; color code each lot; pack the rebar for quick discharge at the specified U.S. port, in accordance with U.S. customs and other regulatory requirements; transport the rebar to the Turkish port; and load the rebar aboard a vessel for transport, in accordance with Turkish customs requirements. See ICDAS Sales Verification Report (Conf.Doc. No. 44), Exh. 14 at contract.

The record further indicates that — after incurring the full (and tremendous) cost of production — ICDAS began to transport the thousands of metric tons of rebar specified in the contract, from the rolling mill to the Turkish port. The inland freight process lasted a week, and involved dozens of trips. ’ Finally, inland transport was completed, and marine bills of lading were issued, indicating that the ocean carrier had taken control of the merchandise for loading and transport. Only then — after ICDAS had released the merchandise to the carrier — -was an invoice issued reflect ing an increase in price. See ICDAS Sales Verification Report (Conf.Doc. No. 44), Exh. 14 at domestic inland freight documentation. 58

The fairly compelling evidence documenting ICDAS’ course of conduct (described above) leaves little room to suggest that ICDAS “perceived any flexibility in the terms of the contract.” See Issues and Decision Memorandum for the Final Determination of the Antidumping Duty Investigation of Polyethylene Retail Carrier Bags from Thailand, 2004 WL 3524397 (June 18, 2004) (“Polyethylene Retail Carrier Bags from Thailand”), at comment 2. Indeed, ICDAS argues persuasively that— quite to the contrary — haying produced the rebar to order, hauled it to the Turkish port, and tendered it for loading and transport, it is abundantly clear that ICDAS believed that the contract terms were firm and binding, without regard to its attempt to negotiate a modest increase in price after it had rendered substantial performance.

ICDAS thus maintains that the record evidence reflects that, at all times, it “acted in a manner consistent with a ‘meeting of the minds’ to be bound by the terms of the original contract” (see Sulfanilic Acid from Portugal, 2002 WL 31493754, at comment 1), because it sought a price increase only after it had honored its contractual obligations. Moreover, as ICDAS is quick to note, there appears to be a paucity of evidence to support the notion that ICDAS would — or could — have breached the contract in the event that its U.S. customer had rebuffed its attempts to negotiate a slightly higher price. 59

Section III.D.l above concludes that Commerce’s remand determination on date of sale is not in accordance with law, and that the issue therefore must therefore be remanded yet again to the agency, for further consideration. As discussed immediately above, the remand determination similarly is not supported by substantial record evidence — a conclusion which would independently warrant a second remand.

On remand, in determining the date on which the parties had a real meeting of the minds (to ascertain the date on which the material terms of sale (i.e., price and quantity) were established), Commerce shall consider in detail — and in the context of other similar cases (such as Pipe from Korea) — all record evidence concerning, inter alia, ICDAS’ use of formal negotiation and contracting procedures for its U.S. sales; the made-to-order nature of the rebar ICDAS produced for the U.S. market, the lead time required to produce that rebar, and the implications of those facts for the expectations of the contracting parties; the lag time between contract date and invoice date for ICDAS’ U.S. sales; the contracting parties’ general course of conduct in ICDAS’ U.S. sales, and precisely how — if at all — that conduct differed in the case of the contract as to which there was a price increase; the fact that there was a single price increase, as to a single contract; the use of an invoice (rather than a formal contract amendment) to reflect the price increase; the timing of the price increase, relative to the timing of actions that ICDAS took to fulfill its obligations under the contract; the ability of ICDAS (as a practical matter) to resell either in the home market or the U.S. market the volume of rebar subject to the price increase or to warehouse that rebar, for whatever period necessary, had ICDAS not fulfilled its contractual obligations by completing the subject sales; and the specific effect — if any — of the price increase on the expectations of contracting parties.

Further, if Commerce determines that the date of sale is some date other than contract date (or, in the case of the contract affected by the price increase, invoice date), Commerce shall expressly identify all record evidence indicating that ICDAS’ U.S. contracts were not legally-binding instruments, as well as all legal authority on which the agency relies to support that conclusion.

Finally, in weighing the record evidence on date of sale in the course of this second remand, Commerce shall be mindful that the “substantial evidence” standard requires consideration of the entirety of the administrative record, “tak[ing] into account whatever in the record fairly detracts from [the] weight [of the evidence on which it relies to support its determination]” — which includes “contradictory evidence or evidence from which conflicting inferences could be' drawn.” See Suramerica, 44 F.3d at 985 (quotation omitted).

E. Commerce’s Use of POR Average Costs in Its “Sales Below Cost” Analysis

In order to make fair comparisons between U-S. sales and normal value, and between home market- sales and costs, Commerce must determine the appropriate time period(s) for its weighted-average cost calculations. Arguing that — over the course of the period of review (“POR”) at issue here — it experienced a “57% increase in the cost of scrap (which is the single primary input for rebar),” 60 ICDAS challenges Commerce’s determination to use the weighted-average cost of manufacturing for ICDAS for the entire POR — rather than the company’s weighted-average cost of manufacturing for each quarter of the POR — in the agency’s “sales below cost” analysis. See generally ICDAS Brief at 1, 4-5, 7-22; ICDAS Reply Brief at 1-5. According to ICDAS, Commerce’s use of the POR average cost improperly inflates “normal value”; and “[w]hen ICDAS’ U.S. sales prices are compared with the inflated ‘normal values’ in the same or adjacent months,” Commerce’s analysis yields “significant distortions that preclude the ‘fair comparisons’ required by antidumping law.” See ICDAS Brief at 4-5, 9-13; ICDAS Reply Brief at 3; 19 U.S.C. § 1677b(a) (requiring Commerce to make “a fair comparison” between export price or constructed export price and normal value in determining whether dumping is occurring).

Specifically, ICDAS asserts that its scrap costs “increased consistently throughout the POR and skyrocketed during the last quarter for an overall increase of 57%.” ICDAS Brief at 4; see also id. at 1, 4, 7-9, 11; ICDAS Reply Brief at 2-3. As a result, ICDAS notes, the POR average cost that Commerce used in its calculations — which is largely driven by “skyrocketing scrap prices in the fourth quarter of the POR” — is both significantly higher than ICDAS’ quarterly average costs during the first and second quarters of the POR, and also lower than ICDAS’ quarterly average costs in the second half of the POR. ICDAS Brief at 4, 8-9, 11-12; ICDAS Reply Brief at 2-4. Because Commerce used the POR average cost in its home market “sales below cost” analysis in the Final Results, the agency excluded many of ICDAS’ first and second quarter home market sales from its analysis as below cost — even though those sales actually were above ICDAS’ quarterly average costs. ICDAS Brief at 4, 8-9, 11-12; ICDAS Reply Brief at 3. ICDAS asserts that Commerce thus artificially “inflated” the benchmark “normal value” to which ICDAS’ U.S. sales prices were compared, producing “significant distortions” in the agency’s analysis. ICDAS Brief at 4-5, 11-12; ICDAS Reply Brief at 3.

ICDAS further explains that, because Commerce used the date of entry to determine the universe of sales in this administrative review (see section III.F, infra), all of ICDAS’ U.S. sales used in Commerce’s antidumping analysis fall within the first and second quarters of the POR— which is when the POR average cost is significantly higher than ICDAS’ quarterly average costs. ICDAS Brief at 4, 9, 11-12; ICDAS Reply Brief at 4. As noted above, the high POR average cost is attributable to “skyrocketing scrap prices in the fourth quarter of the POR” — when, as ICDAS notes, “there were no U.S. sales used in Commerce’s dumping analysis.” ICDAS Brief at 1, 4, 8-9, 11-12; ICDAS Reply Brief at 4.

In the Final Results, Commerce stated that it found no reason to “deviate from its normal practice of using POR annual average costs,” which — the agency asserts— “events] out swings in the production cost experienced by the respondent over short periods of time.” Decision Memo at 10; see also Domestic Producers Response Brief at 4-5. According to Commerce, “relying on monthly or quarterly cost averaging periods creates uncertainty as to how accurately the average costs during the shorter period relate to the sales that occurred during the same period.” Decision Memo at 10; see also Domestic Producers Response Brief at 4-5. Commerce further asserted that, “[o]ver an extended period of time, ... factors [such as the raw material inventory turnover period, the inventory valuation method used by the company, the extent to which raw materials are purchased pursuant to long-term contracts, whether finished merchandise is sold to order or from inventory, and the finished goods inventory holding period] tend to smooth out, resulting in an average cost that reasonably reflects the cost of production for sales made throughout the year.” See Decision Memo at 10-11.

ICDAS disputes any suggestion that the effect of Commerce’s use of the POR average cost here was to “even out swings in the production cost experienced by [ICDAS] over short periods of time,” and to “smooth out the effect of fluctuating raw material costs.” See ICDAS Brief at 13 (discussing Decision Memo at 11). Nor does ICDAS accept that, in this case, the use of the POR average cost “[o]ver an extended period of time reasonably reflects the cost of production for sales made throughout the year.” See ICDAS Brief at 13 (discussing Decision Memo at 11). Asserting that “[t]he reality ... starkly contradicts Commerce’s assertions,” ICDAS argues that “[t]he consistent and significant increase in ICDAS’ weighted-average quarterly costs throughout the entire year of the POR, and especially during the last quarter of the POR, cannot be fairly characterized as mere ‘swings’ or ‘fluctuations.’ ” ICDAS Brief at 13. ICDAS concludes that Commerce’s use of the POR average cost in this case not only “does not ‘reasonably reflect the cost of production for sales’ made throughout the POR,” it actually “skews the cost.” ICDAS Brief at 13. In particular, ICDAS charges that Commerce’s use of the POR average cost (rather than ICDAS’ quarterly average costs) is both “inconsistent with case precedent and Commerce’s past decisions,” and “not supported by substantial evidence.” ICDAS Brief at 9-10, 21.

According to ICDAS, “[t]he touchstone for application of multiple averaging periods is whether the use of full POR cost results in ‘fair comparisons.’ ” ICDAS Brief at 10 (citing 19 U.S.C. § 1677b(a); Agreement on Implementation of Article VI of the General Agreement on Tariffs and Trade 1994, Art. 2.4 (requiring that “[a] fair comparison shall be made between the export price and the normal value”)). Citing Certain Pasta from Italy, ICDAS emphasizes that “Commerce has acknowledged the fundamental importance of ‘fair comparisons’ in determining the appropriate time period for weighted-average cost calculation.” ICDAS Brief at 10 (citing, inter alia, Issues and Decision Memorandum for the Third Antidumping Duty Administrative: Review; Final Results of Review re: Certain Pasta from Italy (Period of Review: July 1, 1998 through June 30, 1999), 2000 WL 1880666 (Dec. 13, 2000) (“Issues and Decision Memorandum for Certain Pasta from Italy”), at comment 18); see generally ICDAS Brief at 9-13 (discussing requirement of “fair comparison” in context of this case); ICDAS Reply Brief at 2 (same).

ICDAS concedes that, “[w]here there are only inconsistent fluctuations in both directions, Commerce uses a single weighted-average cost for the entire POR.” ICDAS Brief at 10 (citing Fujitsu Gen. Ltd. v. United States, 88 F.3d 1034, 1038-39 (Fed.Cir.1996)). But, according to ICDAS, Commerce’s use of the POR average cost in this case “contravenes [the] essential requirement of ‘fair comparisons.’ ” ICDAS Brief at 11. ICDAS maintains that “because of the significant and consistent increase in both the cost of scrap ... and the weighted-average [cost of manufacturing], particularly the dramatic rise in the fourth quarter of the POR, there are significant differences” between the POR average cost and ICDAS’ quarterly average costs, which “create significant distortions in Commerce’s dumping analysis.” ICDAS Brief at 10-11. Underscoring the significance of those asserted “distortions,” ICDAS asserts: “[I]t could make the difference between (i) ICDAS having a de minimis dumping margin and therefore qualifying for revocation ..., and (ii) ... ICDAS having to endure another three years of reviews before again becoming eligible for revocation.” ICDAS Brief at 12-13.

Citing SRAMs from Taiwan and DRAMs from Korea, ICDAS points out that — consistent with the agency’s obligation to make “fair comparisons” — Commerce in prior cases has used weighted-average costs for periods shorter than the POR “when normal values, export prices, or constructed export prices differ significantly over the course of the period of investigation.” See ICDAS Brief at 13-14 (quoting 19 C.F.R. § 351.414(d)(3)); Notice of Final Determination of Sales at Less Than Fair Value: Static Random Access Memory Semiconductors from Taiwan, 63 Fed.Reg. 8909, 8926 (Feb. 23, 1998) (“SRAMs from Taiwan”); Final Determination of Sales at Less Than Fair Value: Dynamic Random Access Memory Semiconductors of One Megabit and Above from the Republic of Korea, 58 Fed.Reg. 15,467, 15,476 (March 23, 1993) (“DRAMs from Korea”). 61 ICDAS notes one case in which Commerce used two averaging periods in its final determination, because the value of the exporters’ currency declined more than 40% over the course of the POR. See ICDAS Brief at 14 (citing Notice of Final Determination of Sales at Less Than Fair Value: Stainless Steel Sheet and Strip in Coils From the Republic of Korea, 64 Fed.Reg. 30,664, 30,674-76 (June 8, 1999) (“Stainless Steel Coils from Korea”); Notice of Amendment of Final Determination of Sales at Less Than Fair Value: Stainless Steel Plate in Coils From the Republic of Korea; and Stainless Steel Sheet and Strip in Coils From the Republic of Korea, 66 Fed.Reg. 45,279, 45,280 (Aug. 28, 2001) (“Stainless Steel Coils from Korea Amended”)). 62

More specifically, invoking Commerce’s determination in Certain Pasta from Italy as the agency’s standard for the use of multiple cost-averaging periods, ICDAS states that “Commerce has used ‘monthly or quarterly costs in instances of non-inflation ... when there is a single primary-input product and that input experiences a significant and consistent decline or rise in its cost throughout the reporting period.’ ” See ICDAS Brief at 14-15 (quoting Issues and Decision Memorandum for Certain Pasta from Italy, 2000 WL 1880666, at comment 18 (emphasis added by ICDAS)). ICDAS maintains that it has satisfied Commerce’s “test” applied in Certain Pasta from Italy, and that Commerce therefore should have used ICDAS’ quarterly average costs (rather than POR average cost) in the agency’s analysis in this case. 63

Commerce’s Final Results acknowledged that — although the agency’s normal practice is to use a single weighted-average cost for the entire POR — it has occasionally used shorter averaging periods in “unusual cases” in the past. See Decision Memo at 11 (citing several prior administrative determinations). Commerce sought to distinguish those prior cases from this case on their facts:

Since this case does not involve a high technology product which experienced drastic cost and price changes over a short period of time due to rapid technological advancements in the production process, we do not consider the facts in the SRAMs from Taiwan, DRAMs from [Korea], or EPROMs from Japan to be relevant.

Decision Memo at 11 (discussing, inter alia, Erasable Programmable Read Only Memories (EPROMs) from Japan; Final Determination of Sales at Less Than Fair Value, 51 Fed.Reg. 39,680, 39,682, 39,686 (Oct. 30, 1986), at Fujitsu Comment 1 (finding that significant changes in cost of production during short period of time due to technological advancements and changes in production process justified use of weighted-average costs for multiple periods of less than one year)). Commerce similarly sought to distinguish Stainless Steel Coils from Korea, asserting that— unlike that case — the instant case does not “involve changes in currency values.” See Decision Memo at 11 (discussing Stainless Steel Coils from Korea, 64 Fed.Reg. at 30,675-76).

As ICDAS notes, however, the legal relevance of the factual distinctions drawn by Commerce is entirely unclear. See ICDAS Brief at 15 (dismissing Commerce’s factual distinctions as “meaningless”); see also Burinskas, 357 F.2d at 827 n. 5 (noting that, to prevail, “[m]ore than enumeration of factual differences” is necessary; relevance of those differences must also be explained). ICDAS argues persuasively that “[t]he reason for significant cost changes is irrelevant. What matters is that there were significant changes in the cost of the input — because such changes affect whether ‘fair comparisons’ are made.” ICDAS Brief at 15. 64

Rather than applying the agency’s “test” from Certain Pasta from Italy, Commerce stated in the Final Results that it was analyzing ICDAS’ request for the use of quarterly average costs pursuant to the criteria set forth in the agency’s determination in the Brass Sheet and Strip from the Netherlands Final case — a case which Commerce characterized as “more on point” here. See Decision Memo at 11 (discussing Notice of Final Results of Anti-dumping Duty Administrative Review and Determination Not to Revoke the Anti-dumping Duty Order: Brass Sheet and Strip from the Netherlands, 65 Fed.Reg. 742, 747-48 (Jan. 6, 2000) (“Brass Sheet and Strip from the Netherlands Final”)); see also Domestic Producers Response Brief at 3, 5 (stating that Commerce “has formulated a strict test to govern departure” from its practice of using a single cost-averaging period, and “has uniformly required that three conditions be met”). Contrary to Commerce’s assertions, however, the test that the agency actually applied in the Final Results in fact is not the test applied in Brass Sheet and Strip from the Netherlands Final. It is instead “a new approach to multiple cost averaging periods.” See ICDAS Brief at 16.

Specifically, in Brass Sheet and Strip from the Netherlands Final, Commerce decided to use monthly weighted-average metal costs 65 based on the agency’s findings that “(1) the cost of copper and zinc [used in the production of brass sheet and strip] are treated as pass-through items when brass is sold to customers; (2) these metal costs represent a significant percentage of the total cost of producing brass sheet and strip; and (3) the cost of the metal dropped consistently and significantly throughout the POR.” Brass Sheet and Strip from the Netherlands Final, 65 Fed. Reg. at 748. In contrast, Commerce here stated that it was analyzing “[1] the significance of the change in the COM [ie., the total cost of manufacturing], [2] whether the change in [the cost of scrap] occurred consistently and significantly throughout the POR, and [3] whether the direct material inputs causing the cost fluctuation can be directly tied to the related sales transactions.” See Decision Memo at 11. Based on its analysis of those three criteria, Commerce declined to use quarterly average costs, and instead adhered to its “normal practice using POR weighted-av erage costs for the foreign like product in the overall weighted-average dumping margin” for ICDAS. See Decision Memo at 12.

Commerce’s second criterion in this case is the same as the third criterion in Brass Sheet and Strip from the Netherlands Final. And Commerce’s third criterion in this case generally seems to parallel the first criterion in Brass Sheet and Strip from the Netherlands. But the relationship between the first criterion in this case and the second criterion in Brass Sheet and Strip from the Netherlands is — at best — unclear.

As a threshold matter, ICDAS disputes Commerce’s third criterion in this case, asserting that Commerce historically has not treated a “pass-through item” as “a separate prerequisite to the use of shorter cost periods.” ICDAS Brief at 16; see also id. at 17, 21. “Nonetheless,” ICDAS argues, “even under Commerce’s new approach [i.e., applying the three criteria as set forth in the Decision Memo at 11], quarterly average costs should be used in this case since Commerce’s test is met.” See ICDAS Brief at 17.

Commerce here first analyzed “the significance of the change in the [cost of manufacturing].” See Decision Memo at 11. . Commerce determined that ICDAS’ cost of manufacturing “both decreased and increased during the first three quarters of the POR,” and that “[i]t was not until the third and fourth quarters of the POR that the [cost of manufacturing] increased steadily.” See Decision Memo at 11. Commerce agreed with ICDAS that “the annual average [cost of manufacturing] is higher than the quarterly average [cost of manufacturing] for the first two quarters of the POR,” but “disagree[d] that the difference is significant.” See id. at 11-12. 66

ICDAS analyzes “the significance of the change in the [cost of manufacturing]” rather differently, and reaches a very different conclusion. According to ICDAS, “from the first to the last quarter of the POR, ... ICDAS’ weighted-average total [cost of manufacturing] (denominated in U.S. dollars) increased by 53%.” See ICDAS Brief at 17. ICDAS asserts that, “[b]y any measure, an increase of more than 50% is a ‘significant’ increase.” Id. 67

ICDAS specifically takes issue with Commerce’s statement that ICDAS’ cost of manufacturing “both decreased and increased during the first three quarters of the POR,” and that “[i]t was not until the third and fourth quarters of the POR that the [cost of manufacturing] increased steadily.” See ICDAS Brief at 18-19 (discussing Decision Memo at 11). ICDAS asserts that Commerce’s statement “does not accurately describe the change in ICDAS’ [cost of manufacturing] during the POR,” and cannot be reconciled with the record evidence. See ICDAS Brief at 18-19. ICDAS acknowledges that, when denominated in Turkish Lira, there was a “minor dip” in the weighted-average cost of manufacturing from the first" to the second quarter of the POR. See ICDAS Brief at 18. But ICDAS emphasizes that — treating the first quarter weighted-average cost of manufacturing as the. baseline — the third quarter weighted-average cost of manufacturing in Turkish Lira shows an increase, and the fourth quarter weighted-average cost of manufacturing reflects a “dramatic” increase. See ICDAS Brief at 18. ICDAS concludes that “[t]he unmistakeable trend is a significant upward movement for ICDAS’ quarterly weighted-average [cost of manufacturing].” ICDAS Brief at 19.

Moreover, ICDAS highlights the fact that ICDAS’ costs were incurred primarily in dollars, including all of ICDAS’ purchases of imported scrap. See ICDAS Brief at 19. ICDAS contends that it is therefore more appropriate to analyze ICDAS’ costs in U.S. dollars. See id. And, according to ICDAS, “a dollar-based analysis confirms that ... [total cost of manufacturing] ... increased in every quarter.” Id.

The second criterion that Commerce purports to have analyzed here was “whether the change in [the cost of scrap] occurred consistently and significantly throughout the POR.” See Decision Memo at 11. According to the Decision Memo, Commerce “computed the difference in the cost of the input raw materials for the first two quarters of the POR using quarterly average cost data versus annual average cost data, and noted that ... the difference ranged from approximately five to 10 percent of the [cost of manufacturing].” Id. at 12. The Decision Memo states that, in the past, Commerce “has not considered one to ten percent increases” to be “significant.” Id. Commerce concluded that, in the instant case, there was “no significant change in the cost of scrap during the POR.” Id. 68

ICDAS flatly rejects Commerce’s analysis of the change in the cost of scrap during the POR. See generally ICDAS Brief at 18-20. ICDAS points to its evidence indicating that, from the first to the last quarter of the POR, its scrap costs increased by 57%. See ICDAS Brief at 17. 69 And, here too, ICDAS argues that, “[b]y any measure, an increase of more than 50% is a ‘significant’ increase,” relying on prior agency determinations where — under similar facts — Commerce has used shorter cost-averaging periods. See ICDAS Brief at 17. In Canned Pineapple from Thailand, for example, where there was an “almost fifty percent” increase in the cost of fresh pineapple (over an 18-month period), Commerce (on remand) used separate weighted-average costs for two periods of time. See Thai Pineapple Canning Indus. Corp. v. United States, 273 F.3d 1077, 1081-82 & n. 1 (Fed.Cir.2001). In another case, Commerce used multiple cost-averaging periods where the currency value decreased by more than 40%. See Stainless Steel Coils from Korea, 64 Fed.Reg. at 30,675-76; Stainless Steel Coils from Korea Amended, 66 Fed.Reg. at 45,280. 70

ICDAS zeroes in on the methodology that Commerce used to evaluate the significance of changes in costs in this ease, asserting that it differs from the approach that the agency has employed in the past. 71 According to ICDAS, when determining the significance of a change in costs during the POR in previous- cases, “Commerce has compared the costs at two points: the beginning and the end of the POR.” ICDAS Brief at 20; id. at 18; see also, e.g., Fujitsu, 88 F.3d at 1038-39 & n. 4 (analyzing change in cost from the beginning to the end of the POR); Issues and Decision Memorandum for Certain Pasta from Italy, 2000 WL 1880666, at comment 18 (discussing analysis which found that “prices [of semolina, the single primary input for pasta] from the beginning to the end of the POR ” decreased by 10% to 12%) (emphasis added).

In cases such as Fujitsu and Certain Pasta from Italy, ICDAS notes, Commerce calculated the difference between the weighted-average costs at the beginning and the end of the POR, and then divided that figure by the beginning cost. See ICDAS Brief at 20. In contrast, Commerce here took a “quite different” approach. See ICDAS Brief at 20. In'this review, Commerce calculated the difference between the weighted-average quarterly costs (for each of the first and second quarters of the POR) and the annual average cost, and then divided by the annual average cost. See ICDAS Brief at 20. ICDAS charges that Commerce’s “statistical manipulation” has the effect of “depressing] the magnitude of the resulting percentage.” See ICDAS Brief at 20. According to ICDAS, “Commerce’s sleight of hand enabled it to calculate a difference in this case that ‘ranged from approximately five to 10 percent’ — which Commerce then inappropriately claims are percentages found not to be significant in Certain Pasta from Italy (which used Commerce’s prior approach).” See ICDAS Brief at 20.

The Domestic Producers seek to defend Commerce’s determination on this second criterion, asserting that the increases in the price of scrap that ICDAS experienced were neither consistent nor significant over the course of the POR. See Domestic Producers Response Brief at 6-7, 9-12, 15. According to the Domestic Producers, “scrap prices both fell and rose at different times over the period of review.” See Domestic Producers Response Brief at 7. As 'ICDAS notes, however, the Domestic Producers’ argument focuses largely on changes in monthly scrap costs; and “ICDAS has not requested the use of monthly costs.” ICDAS Reply Brief at 3 (emphasis added); compare Domestic Producers Response Brief at 10 (analyzing scrap costs “on a monthly, rather than quarterly basis”), 11 (emphasizing results of examination of “monthly data”). And, while the Domestic Producers endorse Commerce’s cost comparison calculations in an effort to demonstrate that increases in the cost of scrap were not significant, the Domestic Producers do not dispute ICDAS’ claim that Commerce’s comparison in this case differed from the agency’s approach in pri- or cases. Nor do the Domestic Producers make any attempt to rebut ICDAS’ claim that Commerce, in effect, manipulated the analysis. See ICDAS Reply Brief at 4.

The third, and final, criterion that Commerce analyzed was “whether the direct material inputs causing the cost fluctuation can be directly tied to the related sales transactions.” See Decision Memo at 11. In its Decision Memo, Commerce emphasized that, in Brass Sheet and Strip from the Netherlands Final, the price of the raw material inputs was a direct pass-through item. See Decision Memo at 12. In other words, as a service to its customers, the respondent in that case purchased the input metals on the customers’ behalf and then billed the customers for the cost of the metals, the terms of which were set forth on the finished brass sales invoice. See Decision Memo at 12. According to the Decision Memo, the price of the raw material inputs there thus “could be directly tied to each related sales transaction.” See Decision Memo at 12. In contrast, in this case, Commerce determined that ICDAS’ sales transactions could not be “directly tied to a particular shorter period’s cost.” See Decision Memo at 12. Commerce stated:

Without a direct link between the input raw material ■ costs and the directly related sales transactions, as was the case in Brass Sheet and Strip from the Netherlands Final, there is no certainty that in adopting [ICDAS’] quarterly cost approach, sales occurring in a given quarter are directly the result of the recorded raw material costs for the same quarter.

Decision Memo at 12. Commerce therefore concluded that “deviating from [the agency’s] normal practice in an attempt to make a more accurate comparison of sales prices and costs [might] well result in a comparison that is less accurate due to the many factors that influence a fair comparison of production and sales.” See Decision Memo at 12.

ICDAS acknowledges that “[t]he fact that there is a pass-through item may serve as one of the reasons supporting the use of shorter cost periods to achieve ‘fair comparisons.’” ICDAS Brief at 16. But, as noted above, ICDAS vigorously disputes that a “pass-through item” is “a separate prerequisite to the use of shorter cost periods.” Id. As ICDAS observes, Commerce has used multiple, shorter cost-averaging periods in other cases where there was no evidence that the identified primary input was a pass-through item. See ICDAS Brief at 16; Issues and Deci sion Memorandum for Certain Pasta from Italy, 2000 WL 1880666, at comment 18 (analyzing request for use of monthly average costs, without making finding as to whether semolina — the single primary input for pasta — was a pass-through item, or even mentioning any pass-through requirement); SRAMs from Taiwan, 63 Fed.Reg. at 8926 (determining that use of quarterly average costs was appropriate, without making finding that single primary input was a pass-through item, or even mentioning any pass-through requirement). 72 More generally, ICDAS argues that Commerce’s focus on “whether the direct material inputs causing the cost fluctuations can be directly tied to the related sales transaction” is an entirely new approach, and that “there is no basis for imposing in this case a new requirement of a ‘direct tie’ between the direct material inputs (scrap in this case) that causes the cost changes and the related home market sales of re-bar.” See ICDAS Brief at 16-17, 21.

ICDAS in any event asserts that — even if a “direct tie” criterion is applicable in the case at bar — “the record provides ample evidence to satisfy Commerce’s new requirement.” ICDAS Brief at 21. ICDAS acknowledges that “it is not possible in every case to have an item-by-item match between sales and costs,” but maintains that this fact “should not automatically lead to the conclusion that only POR costs can be used.” See ICDAS Brief at 21. ICDAS argues that “[i]t would be impossible to match each pound of pasta produced with each pound of semolina input in Certain Pasta from Italy or to match each piece of rebar with each piece of scrap in this case.” See ICDAS Brief at 21. According to ICDAS, its evidence demonstrating that the changes in home market sales prices closely track changes in the cost of scrap is “a more than sufficient showing ... to satisfy Commerce’s [assertedly new] test.” See ICDAS Brief at 21.

The Domestic Producers argue that “ICDAS’ submitted evidence is not sufficient to demonstrate that its scrap purchases were incorporated into merchandise sold within the same quarter.” Domestic Producers Response Brief at 7; see also id. at 6-7, 13-15. Specifically, the Domestic Producers contend that ICDAS’ inventory management practices suggest that ICDAS did not incorporate purchased scrap into rebar which was sold in the same quarter. See Domestic Producers Response Brief at 14-15. But ICDAS brushes off such talk as 'mere “speculation,” and maintains that the Domestic Producers say nothing that “undercuts] the record evidence that ICDAS’ quarterly average home market sales prices closely tracked scrap costs.” See ICDAS Reply Brief at 4. Indeed, according to ICDAS, the Domestic Producers’ argument “highlights the big problem here: Commerce has used high scrap costs incurred at a later date in its cost comparison with rebar sales that occurred months earlier — when ICDAS could not possibly have anticipated that it would have such high scrap costs.” Id. (emphases added). 73

Finally, the Domestic Producers argue that the use of multiple cost-averaging periods “is especially inappropriate to the steel industry.” Domestic Producers Response Brief at 15; see also id. at 7,13-17. The Domestic Producers assert that “the fluctuations of which ICDAS complains are normal in the steel market.” See Domestic Producers Response Brief at 7. The Domestic Producers pointedly note that other respondents in the review at issue “experienced the same fluctuations in scrap costs, but nevertheless felt no need to ask [Commerce] for use of multiple cost-averaging periods.” See Domestic Producers Response Brief at 7. 74 The Domestic Producers state that they know of no case where Commerce has deviated from the practice of single POR-wide cost-averaging due to changes in scrap costs, and they caution that deviating from that practice for ICDAS here would require Commerce to calculate the cost of production on a quarterly basis for all other respondents in this review as well, and — indeed—would have potentially wide-ranging implications for all current and future steel cases “where the fluctuation of scrap costs exceeds some indeterminate ... threshold.” See Domestic Producers Response Brief at 7, 15-16; see generally Decision Memo at 8-10 (summarizing Domestic Producers’ arguments as to consequences for other steel cases, as well as other respondents in this case).

ICDAS summarily rejects the Domestic Producers’ “Chicken Little” predictions. ICDAS maintains that “[w]here, as here, Commerce’s test for using multiple cost periods has been met, then Commerce must use multiple cost periods.” ICDAS Reply Brief at 4. ICDAS emphasizes that a quarterly cost approach is needed to properly analyze ICDAS’ sales, “not only because of the mismatch between sales and costs, but also because all of ICDAS’ U.S. sales used in Commerce’s dumping analysis fall within the first and second quarters of the POR when the POR average cost was significantly higher than the respective quarterly average costs.” ICDAS Reply Brief at 4. ICDAS makes the fair point that “[t]he fact that some other respondents ... did not provide (and Commerce did not request) quarterly cost data and did not request use of quarterly costs, is not the fault of ICDAS,” and should not properly affect the outcome on this issue. ICDAS Reply Brief at 4-5.

As to the Domestic Producers’ claims about the broader implications of ICDAS’ request, ICDAS states that it is “telling” that the Domestic Producers cited no authority for their assertion that using quarterly costs for ICDAS “would ... effectively establish ] a new methodology for the steel industry.” See ICDAS Reply Brief at 5 (quoting Domestic Producers Response Brief at 16). ICDAS further asserts that the Domestic Producers have failed to “provide[] a single example of a steel case that shares similar characteristics with this rebar case — ie., a single primary input (such as scrap) that experiences a significant and consistent change in price — that would warrant the use of multiple cost averaging periods.” See ICDAS Reply Brief at 5.

Even more fundamentally, ICDAS correctly observes that Commerce did not rely on the Domestic Producers’ argument as a basis for the agency’s decision not to use quarterly average costs in this case. See ICDAS Reply Brief at 5. Accordingly, just as Commerce’s determination cannot be sustained based on the Domestic Producers’ claim that scrap does not qualify as a “single primary input,” so too Commerce’s determination cannot be sustained on the basis of the Domestic Producers’ claims that wide-ranging implications would assertedly flow from the use of ICDAS’ quarterly costs here. See Burlington Truck Lines, Inc. v. United States, 371 U.S. 156, 168-69, 83 S.Ct. 239, 9 L.Ed.2d 207 (1962) (ruling that an agency’s decision may only “be upheld, if at all, on the same basis articulated ... by the agency itself’).

In its brief, the Government completely side-steps the merits of the issue. Instead, the Government seeks a voluntary remand, invoking SKF, which recognizes the right of an agency to “request a remand, without confessing error, to reconsider its previous position.” Def. Response Brief at 12 (quoting SKF USA, Inc. v. United States, 254 F.3d 1022, 1027-28 (Fed.Cir.2001)); see also id. at 1-2, 8, 11-12. The Government concedes that “Commerce’s Final Results do not adequately address ICDAS’s arguments,” and that a “more in-depth analysis is required due to the technical nature of ICDAS’s arguments and the cost methodology which Commerce employed in this determination.” Def. Response Brief at 11. The Government requests a remand to “allow Commerce to provide a thorough explanation concerning the complexities of the issue and its effect upon ICDAS’s cost of production.” Def. Response Brief at 11-12. Moreover, notwithstanding their brief (arguing at length that ICDAS’ request for use of quarterly costs should be rejected on the merits, and that the agency’s use of POR average costs should be sustained), the Domestic Producers advised in the course of oral argument that they do not oppose the agency’s request for a voluntary remand on this issue. See Domestic Producers Response Brief at 2-17, 40; Transcript of Oral Argument (“Tr.”) at 54.

ICDAS takes a radically different stance. While ICDAS agrees that Commerce’s use of ICDAS’ quarterly average costs versus the POR average cost must be remanded to the agency, ICDAS argues that the Court should first “reach[ ] a decision on the merits and enter[] an order with instructions that, on remand, Commerce must use quarterly costs in its anti-dumping analysis for ICDAS.” ICDAS Reply Brief at 5; see also id. at 1-2; ICDAS Brief at 5. 75

Under SKF, an agency is generally entitled to a voluntary remand to reconsider its position, “if the agency’s concern is substantial and legitimate.” See SKF, 254 F.3d at 1028-29. But ICDAS questions the legitimacy of Commerce’s concern here. ICDAS argues, in essence, that the parties have been down this road before, that Commerce is simply seeking a “do over,” and that the agency is not entitled to repeated bites at the same apple. 76 ICDAS raises the spectre of an agency with a result in search of a rationale. 77

ICDAS is, quite understandably, frustrated with the position in which it finds itself. But the Government must be presumed to have acted in good faith. Certainly ICDAS has pointed to no evidence to substantiate any suggestion of prejudgment on the part of Commerce. See generally Habas Sinai Ye Tibbi Gazlar Istihsal Endustrisi AS. v. United States, 31 CIT-,-, 2007 WL 3378201 at **4-5 (2007) (and cases cited there). Moreover, although ICDAS is — perhaps with good reason — skeptical of the outcome of the voluntary remand that Commerce seeks, the Government has given the Court its express assurances that the agency plans to use the remand proceeding to take a fresh look at the issue. See Tr. at 48^9 (denying that results of remand are “predetermined,” emphasizing that Commerce “could change its mind” on remand, advising that Commerce has assured counsel for the Government “that this is not a result oriented remand request,” and stating that the agency has not “already made up its mind”). And this is not a case in which it can be said that a remand to the agency would be futile. See generally Nippon Steel Corp. v. United States, 458 F.3d 1345 (Fed.Cir.2006).

A court must tread lightly in administrative cases, taking care not to infringe on an agency’s mandate and discretion by reweighing facts and substituting its judgment for that of the agency. As the Government notes, that is particularly true where, as here, “the issue is based on a ‘factual determination’ which requires agency expertise.” See Def. Response Brief at 12. The Court of Appeals has underscored that Commerce is the “master” of the antidumping law, and that “[f]actual determinations supporting anti-dumping margins are best left to the agency’s expertise.” See F.lli De Ceceo Di Filippo Fara S. Martino, S.p.A., 216 F.3d 1027, 1032 (Fed.Cir.2000); Micron Tech., Inc. v. United States, 117 F.3d 1386, 1394 (Fed.Cir.1997).

The Government’s request for a voluntary remand is therefore granted. On remand, the Commerce Department shall consider anew the use of ICDAS’ quarterly average costs versus the POR average cost in calculating ICDAS’ cost of production. 78 In addition, Commerce shall clarify the test that it is applying for the use of multiple cost-averaging periods, fully articulate the rationale for its redetermination on the issue, and recalculate ICDAS’ dumping margin, if appropriate.

F. Commerce’s Use of Date of Entry to Defíne the Universe of Sales

ICDAS’ final challenge to the Final Results contests Commerce’s decision to use the date of entry — rather than the date of sale — to define the “universe of sales” subject to the administrative review at issue here. See generally Decision Memo at 20-25 (analyzing “Universe of Sales” issue); ICDAS Brief at 6, 37-40; ICDAS Reply Brief at 13-15.

In all six prior reviews up to the administrative review here, Commerce consistently had defined the universe of sales by using the date of sale. See Decision Memo at 22; ICDAS Brief at 2, 6, 37, 39-40; ICDAS Reply Brief at 13, 15; Def. Response Brief at 28; Domestic Producers Response Brief at 34. Indeed, even in this administrative review, Commerce used the date of sale to define the universe of sales in its Preliminary Results. See ICDAS Brief at 6; ICDAS Reply Brief at 13. As ICDAS puts it, only in the Final Results did Commerce “abruptly decide[] to reverse its decision in the Preliminary Results, and to apply a new universe of sales methodology based on date of entry.” ICDAS Brief at 37. 79 ICDAS pointedly notes that “Commerce decided to make this change even though Commerce recognized that it ... could make the difference between revocation of the [antidumping] order as to ICDAS and an above-de minimis margin.” ICDAS Brief at 37 (citing Decision Memo at 20); see also id. at 40; ICDAS Reply Brief at 14-15.

ICDAS candidly acknowledges that “Commerce has discretion to reconsider the methodologies it uses in its antidumping analysis.” ICDAS Brief at 37. But, as ICDAS emphasizes, that discretion “is not unbounded.” See ICDAS Brief at 37 (citation omitted). In particular, ICDAS asserts that an agency’s discretion is limited “where a respondent has detrimentally relied on an old methodology used in previous reviews.” ICDAS Brief at 37 (quoting Anshan Iron & Steel Co. v. United States, 27 CIT 1234, 1241-42 (2003)). In addition, as ICDAS notes, an agency must “explain the basis for its change.” ICDAS Brief at 37-38 (quoting Anshan Iron, 27 CIT at 1242); see also id. at 38 n. 24 (same). Invoking Shikoku, ICDAS asserts that Commerce violated both of these limitations on its discretion in this case. See ICDAS Brief at 38-40 (citing Shikoku Chems. Corp. v. United States, 16 CIT 382, 795 F.Supp. 417 (1992)). 80 ICDAS concludes:

Basic principles of fairness demand that Commerce be prevented “from changing its methodology at this late stage.” Shikoku, 16 CIT at 388, [795] F.Supp. at 421. Commerce cannot be permitted to abruptly change its methodology unless it can articulate specific reasons that differentiate this review from all the pri- or reviews. No such reasons have been provided. And certainly, there is no equitable reason why ICDAS must endure another three years of administrative reviews due to an eleventh hour discretionary change of methodology.

ICDAS Brief at 40.

Although the Domestic Producers seek to minimize ICDAS’ claim of reliance, 81 ICDAS argues that, “[a]fter many reviews, ICDAS ha[d] developed a strong expectation that Commerce would employ the same universe of sales methodology.” See Domestic Producers Response Brief at 34-37 & n. 14 (characterizing ICDAS’ example of reliance as “trivial”); ICDAS Brief at 39. ICDAS asserts that its expectation “affected the manner in which [it] has dealt with this case.” ICDAS Brief at 39. As an example, ICDAS states that, in requesting revocation of the antidumping order, it submitted a certification that it believed that its dumping margin for the review would be zero or de minimis. ICDAS Brief at 39. ICDAS explains that it submitted that certification “on the assumption that Commerce would continue to use the same methodologies that it had used for many years, including the universe of sales methodology.” ICDAS Brief at 39-40.

Apart from its assertions of reliance, ICDAS further argues that Commerce here “failed to offer a reasoned explanation for its change in methodology.” ICDAS Brief at 39; see also id. at 2, 6, 39-40; ICDAS Reply Brief at 13-14. ICDAS contends that Commerce identified no legal authority requiring it to make the change. See ICDAS Brief at 39. Moreover, according to ICDAS, Commerce failed to identify any new facts which would warrant the change. Id. Although the Domestic Producers go to some lengths to seek to defend Commerce’s treatment of the universe of sales issue in the Final Results, the Government itself readily concedes that the Final Results failed to adequately explain the reasons for Commerce’s change in methodology. Compare Domestic Producers Response Brief at 4, 34-40 (asserting that Commerce “clearly and completely explained its reason for changing its methodology”) with Def. Response Brief at 28-29 (flatly admit ting that “Commerce’s Final Results did not explain the change in methodology”). The Government therefore requests a. voluntary remand. Def. Response Brief at 1-3, 9, 28-29.

The Domestic Producers see no need for a voluntary remand, and urge that Commerce’s use of the date of entry to define the universe of sales in the Final Results be sustained in all respects. See Tr. at 77-79. ICDAS too opposes the Government’s request for a voluntary remand — albeit for a very different reason. ICDAS maintains that the requested remand is not required under ’ SKF, “because the Government does not state that it wishes to ‘reconsider ’ its position, only that it wishes to farther explain and justify it.” See ICDAS Reply Brief at 14 (quoting SKF, 254 F.3d at 1029). And ICDAS contends that “[t]here is no legal or factual basis that could justify Commerce’s last-minute reversal in methodology at the very point when ICDAS became eligible for revocation.” ICDAS Reply Brief at 14-15. Quoting Shikoku, ICDAS maintains that “ ‘[i]t is simply too late to mandate another three years of administrative reviews because of a last minute “improvement” in Commerce’s methodology,’ ” and that — accordingly — the proper course is to “remand this issue to Commerce with instructions to use date of sale to define the universe of sales.” See ICDAS Reply Brief at 15 (quoting Shikoku, 16 CIT at 387-88, 795 F.Supp. at 421-22); see also id. at 14.

As discussed in section III.E immediately above, however, under SKF, an agency is generally entitled to a voluntary remand to reconsider its position, “if the agency’s concern is substantial and legitimate.” See SKF, 254 F.3d at 1028-29. ICDAS’ skepticism notwithstanding, 82 the Government must be presumed to be acting in good faith. To be sure, ICDAS has pointed to no specific evidence to indicate that Commerce has prejudged the outcome of the remand that it requests. See generally Habas, 31 CIT at -, 2007 WL 3378201 at * 4-5 (and cases cited there). Further, as with the issue of ICDAS’ request for the use of its quarterly average costs (rather than the POR average cost), the Government has given the Court its express assurances that the agency plans to use the remand proceeding to take a fresh look at the issue. See Tr. at 48-49 (disavowing any notion that remand results have been “predetermined,’.’ and stating that Commerce has assured Government counsel that request for remand “is not ... ' result oriented”). Thus, once again, it cannot be said with any assurance that a remand to the agency would be futile. See generally Nippon Steel Corp., 458 F.3d 1345.

The Government’s request for a voluntary remand is accordingly granted. On remand, the Commerce Department shall consider anew Commerce’s use of the date of sale versus the date of entry to define ICDAS’ universe of sales for the administrative review here at issue, weighing all appropriate factors (including past agency practice). In addition, Commerce shall fully articulate the rationale for its redetermination on the issue, and recalculate ICDAS’ dumping margin, if appropriate.

IV. Conclusion

For all the reasons set forth above, the Domestic Producers’ Motion for Judgment on the Agency Record challenging Commerce’s decision to treat sales made through ICDAS’ U.S. affiliate as EP sales must be denied. ICDAS’ Motion for Judgment on the Agency Record similarly must be denied as to ICDAS’ claims that Commerce improperly denied ICDAS’ request for a startup adjustment, and that Commerce erred in its treatment of ICDAS’ foreign exchange gains as well as in its decision to cap ICDAS’ total financial expenses at zero. On the other hand, ICDAS’ Motion for Judgment on the Agency Record is granted as to ICDAS’ challenges to Commerce’s use of invoice date (rather than contract date) as the date of sale for ICDAS’ U.S. sales, Commerce’s use of the POR average cost of manufacturing (rather than ICDAS’ quarterly costs) in the agency’s “sales below cost” analysis, and Commerce’s use of the date of entry (rather than the date of sale) to define ICDAS’ universe of sales; and this matter is remanded to the Department of Commerce for further action not inconsistent with this opinion.

A separate order will enter accordingly.

1 . Export Price ("EP”) and Constructed Export Price ("CEP”) refer to Commerce’s two methods for calculating prices for merchandise imported into the United States. Commerce compares those prices to normal values to determine whether merchandise has been dumped in the United States.

Both EP and CEP are calculated using the price at which the subject merchandise is first sold to a U.S. buyer not affiliated with the foreign producer or exporter. Generally, a U.S. sale is treated as an EP sale when the first sale to an unaffiliated U.S. buyer occurs before the goods are imported into the United States. In turn, a U.S. sale generally is calculated as a CEP sale when the first sale to an unaffiliated U.S. buyer occurs after importation. See generally AK Steel Corp. v. United States, 226 F.3d 1361, 1367-74 (Fed.Cir.2000) (analyzing 19 U.S.C. § 1677a(a)-(b) (2000), and distinction between EP and CEP sales).

Thus, in general, "EP treatment is limited to transactions that occur between a seller outside the United States and an [unaffiliated] buyer inside the United States, before the date of importation.” Corus Staal BV v. United States, 502 F.3d 1370, 1377 (Fed.Cir.2007) (citing 19 U.S.C. § 1677a(a)); cf. AK Steel, 226 F.3d at 1370 n. 8 (hypothesizing “a sales contract between two U.S. domiciled entities that is entirely executed outside the United States,” but expressly declining to decide "whether such a sale would be classified as an EP or CEP sale”). And, for reasons summarized in section III.A below, the classification of U.S. sales as CEP sales (rather than EP sales) is more likely to result in a determination that merchandise has been dumped in the United States. See Corus Staal, 502 F.3d at 1376.

2 . All citations to federal statutes are to the 2000 edition of the United States Code.

3 . See 19 C.F.R. § 351.106 (2003) (providing that a de minimis dumping margin is one below 0.5%).

All citations to federal regulations are to the 2003 edition of the Code of Federal Regulations.

4 . See Certain Steel Concrete Reinforcing Bars from Turkey; Amended Final Results of Anti-dumping Duty Administrative Review, 66 Fed.Reg. 63,364 (Dec. 6, 2001) (amended final results for administrative review covering April 1, 1999 through March 31, 2000, finding dumping margin of zero); Certain Steel Concrete Reinforcing Bars from Turkey; Final Results, Rescission of Antidumping Duty Administrative Review in Part, and Determination Not To Revoke in Part, 68 Fed.Reg. 53,127 (Sept. 9, 2003) (final results for administrative review covering April 1, 2001 through March 31, 2002, finding de minimis dumping margin of 0.10%); Certain Steel Concrete Reinforcing Bars from Turkey; Final Results, Rescission of Antidumping Duty Administrative Review in Part, and Determination Not To Revoke in Part, 69 Fed.Reg. 64,731 (Nov. 8, 2004) (final results for administrative review covering April 1, 2002 through March 31, 2003, finding dumping margin of zero); Certain Steel Concrete Reinforcing Bars from Turkey; Final Results, Rescission of Antidumping Duty Administrative Review in Part, and Determination To Revoke in Part, 70 Fed.Reg. 67,665 (Nov. 8, 2005) (final results for administrative review covering April 1, 2003 through March 31, 2004, finding de minimis dumping margin of 0.16%).

A pending companion case challenges the results of the seventh administrative review (covering 2003-2004) — the administrative review at issue here. See Habas Sinai ve Tibbi Gazlar Istihsal Endustrisi A.S. v. United States, No. 05-00613 (Ct. Int'l Trade filed Nov. 10, 2005). An action challenging the results of the sixth administrative review (covering 2002-2003) also remains pending. See Gerdau AmeriSteel Corp. v. United States, No. 04-00608 (Ct. Int'l Trade filed Dec. 6, 2004); see also Gerdau Ameristeel Corp. v. United States, 519 F.3d 1336 (Fed.Cir.2008).

5 . Because this action was remanded to Commerce for further consideration of the issue of the date of sale to be used for ICDAS’ U.S. sales, two administrative records have been filed with the court — the initial administrative record (which comprises the information on which the agency’s Final Results were based), and the supplemental administrative record (on which the Remand Results were based). Moreover, because confidential information is included in both administrative records, there are two versions of each — a public version and a confidential version.

Citations to public documents in the initial administrative record and the supplemental administrative record are noted as "Pub. Doc. No._” and "Remand Pub. Doc. No._,” respectively. Citations to the confidential versions are, in turn, noted as “Conf. Doc. No. _" and "Remand Conf. Doc. No.

6 . At the request of the Domestic Producers, and with the consent of all parties, the Court preliminarily enjoined liquidation of the entries subject to the seventh administrative review, which is the review at issue here. However, the Court declined to extend the injunction to entries made after the period of review. See Nucor Corp. v. United States, 29 CIT 1452, 1453, 1470, 412 F.Supp.2d 1341, 1343, 1357 (2005).

7 . Specifically, any selling commissions, any expenses associated with the sale (such as credit expenses), any costs of further manufacture, and the profit allocated to those costs and expenses must be deducted from CEP sales. See 19 U.S.C. § 1677a(d). No such deductions are taken from EP sales. See AK Steel, 226 F.3d at 1364 n. 4.

8 . See Certain Steel Concrete Reinforcing Bars from Turkey: Preliminary Results of Anti-dumping Duty Administrative Review and Notice of Intent Not To Revoke, 68 Fed.Reg. 23,972, 23,974 (May 6, 2003) (preliminary results for 2001-2002 period of review); Certain Steel Concrete Reinforcing Bars from Turkey; Preliminary Results and Partial Rescission of Antidumping Duty Administrative Review and Notice of Intent Not To Revoke In Part, 69 Fed.Reg. 25,063, 25,065 (May 5, 2004) (preliminary results for 2002-2003 period of review).

9 . The Court of Appeals noted, for example, that "the statute appears to allow for a sale made by the foreign exporter or producer to be classified as a CEP sale, if such a sale is made 'in the United States.’ ” AK Steel, 226 F.3d at 1367 n. 6 (emphasis added). Similarly, in an aside, the Court of Appeals observed that “[sjales in the United States between unaffiliated purchasers and unaffiliated sellers are never at issue.” AK Steel, 226 F.3d at 1367-68 (emphasis added).

10 . See also AK Steel, 226 F.3d at 1373 (observing that "the distinction [between CEP sales and EP sales] based on the location of the sale was already present” prior to the 1994 amendments to the statute) (emphasis added).

11 .Further, in rejecting an alleged ambiguity in the statute, the Court of Appeals again highlighted the significance of the location of the sale or transaction, ruling that "[t]he language of the CEP definition leaves no doubt that the modifier 'in the United States' relates to 'first sold.’ The term 'outside the United States,' read in the context of both the CEP and the EP definitions [at 19 U.S.C. § 1677a(a)-(b)], ... applies to the locus of the transaction at issue, not the location of the company.” AK Steel, 226 F.3d at 1369 (emphasis added).

The Court of Appeals also explained that "[a] sales contract executed in the United States between two entities domiciled in the United States cannot generate a sale ‘outside the United States.’ ” AK Steel, 226 F.3d at 1370 (final emphasis added). Emphasizing yet again the significance of the location of the sale or transaction, the Court added: "In general, a producer/exporter in a dumping investigation will always be located outside the United States. Thus, it must be the locus of the transaction that is modified by 'outside the United States' in the EP definition” which appears at 19 U.S.C. § 1677a(a). AK Steel, 226 F.3d at 1370.

12 . AK Steel’s analysis of the participation of an affiliate as a seller simply reinforces the significance of the location of the sale or transaction. Noting the relationship between the two considerations, for example, the Court of Appeals stated that "[limiting affiliate sales to CEP flows logically from the geographical restriction of the EP definition [to sales or transactions that take place ‘outside the United States'].” AK Steel, 226 F.3d at 1370-71 (emphasis added).

As noted above, the Court then reiterated that "[t]he location of the sale " is a "critical” factor in distinguishing between EP and CEP sales, and concluded that it would be "contrary to the plain meaning of the statute” to classify the transactions in AK Steel as EP sales “[w]hen ... there are contracts showing that the sales at issue took place in the United States between two entities with United States addresses, one of which was an affiliate of the producer/exporter.” AK Steel, 226 F.3d at 1371 (emphasis added).

13 . The Court of Appeals’ reservation of judgment on this point, alone, suffices to refute the Domestic Producers’ claim that AK Steel established a bright line rule that any sede in which the contract is between a U.S. affiliate and an unaffiliated U.S. customer is — by definition — a CEP sale.

14 . The Domestic Producers contend that the facts and outcomes of AK Steel and Corns Staal buttress the Domestic Producers' interpretation of those cases. See generally Domestic Producers Brief at 8-10, 13-17, 24 (discussing AK Steel and Court of International Trade's opinion in Corus Staal)- Domestic Producers Reply Brief at 1-6 (same). The Domestic Producers concede that the language of the opinions "emphasize[s] the 'location of the sale.' ” See Domestic Producers Brief at 14. But, according to the Domestic Producers, "a review of the courts' actions demonstrates that the courts intended this phrasing to reflect the place of the seller's domicile.” Id. In particular, the Domestic Producers state that "neither Court analyzed ... the location of title transfer (which would appear to be required by the Federal Circuit’s definition of 'sale' as the transfer of ownership).” Id. The Domestic Producers' arguments again are wide of the mark.

As the Government pointedly observes, the Court of Appeals in AK Steel could have chosen whatever verbiage it wished: "If the Federal Circuit had intended the phrasing to reflect the seller's domicile, it would have stated location of affiliate, or domicile of affiliate, not location of the sale.” See Def. Response Brief at 33-34. Further, as the Government notes, AK Steel not only used the term "location of the sale,” it specifically defined it — as the place of “the transfer of ownership or title.” See Def. Response Brief at 34 (quoting AK Steel, 226 F.3d at 1371). The fact of that express definition makes it crystal clear that the Court of Appeals meant what it said, and said what it meant, in referring to the "location of the sale”; and, contrary to the Domestic Producers' claims, the Court of Appeals did not mean "the place of the seller’s domicile.”

Moreover, contrary to the Domestic Producers’ claims, both AK Steel and Corns Staal specifically took note of the location of the transactions at issue in those cases. In AK Steel, the Court of Appeals noted that there were "contracts showing that the sales at issue took place in the United States.” AK Steel, 226 F.3d at 1371 (emphasis added); see also id., 226 F.3d at 1368 (noting that issue on appeal is "whether a sale to a U.S. purchaser can be properly classified as a sale by the producer/exporter, and thus an EP sale, even if the sales contract is between the U.S. purchaser and a U.S. affiliate of the producer/exporter and is executed in the United States ") (emphasis added); id. at 1370 (explaining that "[a] transaction, such as those here, in which both parties are located in the United States and the contract is executed in the United States cannot be said to be 'outside the United States' "). See generally Def. Response Brief at 34. Similarly, in Corns Staal, this court expressly took note of evidence indicating that "the terms of ... sale were agreed upon prior to the shipment of the merchandise,” and that "the sale was made by the producer 'outside of the United States,' ” supporting Commerce's determination "that there was a transfer of ownership in the Netherlands.” Corus Staal, 27 CIT at 393, 259 F.Supp.2d at 1258.

In sum, there is simply no truth to the Domestic Producers’ claim that neither AK Steel nor Corus Staal analyzed the location of the sales there at issue. See generally Def. Response Brief at 34.

15 . As such, it appears that, because — like a "sale” — an "agreement to sell" between a producer/exporter and an unaffiliated purchaser in the United States (or an unaffiliated purchaser for export to the United States) which is reached before importation is within the scope of 19 U.S.C. § 1677a(a), a transaction may be classified’as EP even if ownership does not transfer and consideration is not paid until a later date. .

16 . The Domestic Producers seek to analogize the facts of this case to the facts of AK Steel, arguing that the same result should obtain. According to the Domestic Producers, the sales here at issue — like those at issue in AK Steel — were so-called " 'back-to-back' sales." See AK Steel, 226 F.3d at 1365 (noting "back-to-back” nature of subject sales). The Domestic Producers further assert that the Court of International Trade's opinion in Corns Staal established a "bright-line” rule classifying all "back-to-back” sales as CEP sales. See generally Domestic Producers Brief at 11-17; Domestic Producers Reply Brief at 3-7. The Domestic Producers’ arguments are lacking in merit. See generally Def. Response Brief at 32-33; ICDAS Response Brief at 35-36.

Contrary to the Domestic Producers’ assertions, there are critical differences between the facts of AK Steel and those of this case. Thus, for example, the "back-to-back” sales in AK Steel involved a Korean producer which sold steel to an affiliated Korean exporter, which in turn sold it to a U.S. affiliate, who then made sales to unaffiliated U.S.' purchasers. See AK Steel, 226 F.3d at 1365. Those "back-to-back” sales — which involved a chain of actions taken both outside and inside the United States — are clearly distinguishable from the transactions at issue here, which were completed entirely outside the United States. See generally Def. Response Brief at 32.

Further, notwithstanding the Domestic Producers’ claims to the contrary, the Court of International Trade’s opinion in Corns Staal did not establish a bright line rule governing "back-to-back” sales, without regard to where they occur. The Domestic Producers read the court’s language out of context, and omit reference to the court’s discussion of AK Steel's emphasis on the significance of “the location of the sale.” See Def. Response Brief at 32-33 (discussing Corus Staal, 27 CIT at 392-93, 259 F.Supp.2d at 1258-59). Moreover, as the Domestic Producers themselves concede, the transactions at issue in Corns Staal were not "back-to-back” sales. See Domestic Produc ers Brief at 16 n. 10; Domestic Producers Reply Brief at 4. Accordingly, even if the Court of International Trade had articulated the bright line rule that the Domestic Producers assert, the court's language presumably would have been dicta. Even more to the point, there is certainly nothing in the Court of Appeals’ opinions in either AK Steel or Corus Staal establishing any such per se rule.

17 . As ICDAS points out, Commerce’s classification of the sales at issue here as EP sales is consistent with agency practice in other cases, recognizing that — under AK Steel — the critical difference between EP and CEP sales is "whether the sale or transaction takes place inside or outside the United States.” See generally ICDAS Response Brief at 38.

Thus, for example, in Canned Pineapple Fruit from Thailand, Commerce treated transactions as EP sales, where the foreign company’s U.S. affiliate handled customs clearances, issued invoices, and received payment from U.S. customers, but was "not involved in the sales process, never [took] possession or inventory of subject merchandise, [had] no physical presence in the United States, and act[ed] as an importer of record only,” because Commerce found that the sales in question took place outside the United States. See Issues and Decision Memorandum for the Final Results of the Antidumping Duty Administrative Review: Canned Pineapple Fruit from Thailand, 2001 WL 1241130 (Oct. 9, 2001) (“Canned Pineapple Fruit from Thailand”), at comment 16.

In the instant case, Commerce properly classified ICDAS' sales through its U.S. affiliate as EP sales — as the agency has done since the 2001-2002 review — because, like the sales in Canned Pineapple Fruit from Thailand, all sales here in question were made outside the United States. Similarly, ICDAS’ U.S. affiliate, like the U.S. affiliate in Canned Pineapple Fruit from Thailand, is a mere "paper company” with no employees and no physical premises in the United States, and no role in the sales process, and serves only as an importer of record. The Domestic Producers point to no evidence concerning ICDAS’ "paper company” U.S. affiliate which undercuts Commerce’s determination that the sales here at issue all occurred outside the United States.

Further, Commerce’s classification of transactions as EP sales was sustained in Corns Staal, even though the U.S. affiliate there accepted payment for transmissions to the foreign producer, and had an active role in certain administrative and sales functions. See Corus Staal, 27 CIT at 393-94, 259 F.Supp.2d at 1259. Thus, the acceptance of payment by a U.S. affiliate (like ICDAS' U.S. affiliate here) does not suffice to preclude classification of transactions as EP sales, where — as here — the transactions in question occur outside the United States. See generally ICDAS Response Brief at 38 n. 28; see also Issues and Decision Memorandum for the Final Results of the Ninth Administrative Review of Certain Corrosion-Resistant Carbon Steel Flat Products from Canada for Dofasco, Inc. and Sorevco, Inc. (Collectively,, Dofasco), 2004 WL 3524484 (Jan. 16, 2004), at comment 1 (where U.S. affiliate invoiced customer and accepted payment, transactions were classified as CEP sales, not because of issuance of invoices or receipt of payment in themselves, but because "transfer of ownership was executed in the United States”).

18 . Commerce’s classification of ICDAS’ U.S. sales as EP sales achieves the policy aims of the antidumping statute. As AK Steel explains, Commerce distinguishes between EP and CEP sales in order to "prevent foreign producers from competing unfairly in the United States market by inflating the U.S. Price [sic] with amounts spent by [a] U.S. affiliate on marketing and selling the products in the United States.” AK Steel, 226 F.3d at 1367.

In this case, the terms of the sales, including price, were set outside the United States. As a "paper company” with no employees, inventory, or premises, ICDAS’ U.S. affiliate had no selling or marketing functions, and incurred no costs beyond those normally associated with serving as importer of record. ICDAS' sales through its U.S. affiliate therefore are not at an inflated price, and there is thus no policy reason to treat them as CEP sales. See ICDAS Response Brief at 36 n. 25.

19 . The Domestic Producers’ briefs advanced two alternative arguments predicated on their principal claim on this issue — in other words, one argument assuming that they did not prevail on their claim that AK Steel’s reference to "the location of the sale” is actually a reference to the domicile of the seller, and a second argument assuming that they did prevail on that claim. In light of the analysis above (rejecting the Domestic Producers' claim that "the location of the sale” refers to the seller's domicile), there is no need to here consider the second of the Domestic Producers' alternative arguments — specifically, their argument that, "to the extent that location of the sale’ refers only to the domicile of the seller, [Commerce] improperly failed to apply the test, instead relying on just the sort of ‘activities’ analysis invalidated by AK Steel." See Domestic Producers Brief at 18; see also id. at 9, 21-23; Domestic Producers Reply Brief at 8 n. 3.

20 . As explained above, AK Steel defined "sale” (for purposes of EP/CEP analysis) by reference to the transfer of ownership or title to the goods at issue. See AK Steel, 226 F.3d at 1371 (discussing "transfer of ownership or title,” as well as "transfer of ownership to an unrelated party and consideration”) (citation and emphases omitted). However, in Corus Staal, the Court of Appeals distinguished an "agreement to sell” from a "sale,” defining an "agreement to sell” (for purposes of EP/CEP analysis) as "a binding commitment that has not yet been consummated by the exchange of goods for consideration, i.e., the sale’ itself.” See Corus Staal, 502 F.3d at 1376-77. Thus, it appears that — to establish the existence of an "agreement to sell” — a producer/exporter need only adduce evidence of "mutual assent to the material terms [of a deal] (price and quantity).” Corus Staal, 502 F.3d at 1376. It would seem that, by definition, establishing the existence of an "agreement to sell” does not require either proof of transfer of ownership or title, or proof of payment of consideration.

21 . As Commerce noted, several of ICDAS' U.S. sales were expressly governed by U.S. law (specifically, Texas state law, which includes the relevant sections of the UCC), pursuant to choice-of-law clauses. See Decision Memo at 67. Those contracts included no explicit agreement as to when and where transfer of title occurred. Accordingly, as discussed above, because the goods were delivered to the unaffiliated buyer outside the United States, title also transferred outside the United States, pursuant to the UCC.

22 . The mere fact that "it [may be] possible to draw two inconsistent conclusions from evidence in the record ... does not prevent Commerce's determination from being supported by substantial evidence.” Am. Silicon Techs., 261 F.3d at 1376; see also Consolo v. Federal Maritime Commission, 383 U.S. at 620, 86 S.Ct. 1018. Cf. INS v. Elias-Zacarias, 502 U.S. 478, 483-84, 112 S.Ct. 812, 117 L.Ed.2d 38 (1992) (stating that, where Congress has entrusted agency to administer statute in fact-intensive situations, agency's conclusion should be reversed only if the record evidence is “so compelling that no reasonable factfinder" could reach the same conclusion).

23 . Billets are short, thick bars of steel in the shapes of cylinders or rectangular prisms, which are produced from ingots. See McGraw-Hill Dictionary of Scientific and Technical Terms 235 (6th ed.2003) (defining “billet”). Billet is the single primary input for rebar production. ICDAS Brief at 7.

24 . Although the startup adjustment statute itself does not specify how Commerce is to measure levels of production, the Statement of Administrative Action accompanying the Uruguay Round Agreements Act states generally that, for purposes of startup adjustment analyses, "[production levels will be measured based on units processed.” See Statement of Administrative Action, H.R. Doc. No. 103-316, at 836 (1994), reprinted in 1994 U.S.C.C.A.N. 4040, 4173.

Congress has directed that the Statement of Administrative Action is to be "regarded as an authoritative expression by the United States concerning the interpretation and application of the Uruguay Round Agreements and [the Uruguay Round Agreements] Act in any judicial proceeding in which a question arises concerning such interpretation or application.” 19 U.S.C. § 3512(d).

25 . In drafting the regulations governing start-up adjustments, Commerce expressly took note of Congress' recognition that "any determination of the appropriate startup period involves a fact-intensive inquiry.... For this reason, the Administration intends that Commerce determine the duration of the startup period on a case-by-case basis.” Antidumping Duties; Countervailing Duties: Proposed Rule, 61 Fed.Reg. 7308, 7340 (Feb. 27, 1996) (Preamble) (quoting Statement of Administrative Action).

26 . Thus, for example, Commerce began with actual monthly data on production starts at the Biga facility, for December 2003, as well as January, February, and March 2004, which ICDAS provided to the agency in its Questionnaire Responses. See Decision Memo at 80. According to ICDAS, the Biga facility operated for only 21 days in December 2003. Commerce therefore took the December 2003 production starts data supplied by ICDAS and divided that figure by 21, to derive an actual daily average. Commerce then multiplied that actual daily average by 31 days, to arrive at a full month figure for production starts in December 2003, for comparison to production starts data for the first three months of 2004. See Decision Memo at 80.

Clearly, comparing partial month production data to full month production data in the startup cost adjustment analysis would distort that analysis. For example, comparing production start data for a 21-day period (December 2003) to production start data for a 31-day period (January 2004) — without adjusting for the 10-day difference — would not be an apples-to-apples comparison, and would yield skewed results. All other things being equal, production starts during the 21-day period obviously would be lower than production starts during the 31-day period. If Commerce were limited to comparing partial month data to full-month data, any respondent company with a new facility could greatly enhance its chances of being granted a startup cost adjustment simply by delaying the start of production till late in the first month of operation, since that would help ensure that the first (partial) month’s production levels would be lower than those of subsequent full months of production.

When Commerce evaluates whether a respondent is entitled to a startup adjustment, it is critical that Commerce have production data that are reasonably comparable. The full-month equivalency methodology employed by Commerce here is a reasonable means to help ensure fair comparisons.

27 . Although its point is not entirely clear, ICDAS also objects to Commerce's full-month equivalent methodology on the grounds that it "creates a mismatch between the cost of production data and production starts data." See ICDAS Brief at 28. According to ICDAS, Commerce's methodology "created phantom production starts for the first ten days of December with no associated production costs at all." Id. As the Government explained, however, production costs play no role in determining whether a startup adjustment is granted. The focus of Commerce's startup adjustment analysis is solely on levels of production. See Def. Response Brief at 21. Accordingly, any asserted "mismatch” did not prejudice ICDAS' request.

Similarly, ICDAS points to "the extremely high costs that ICDAS incurred [in December 2003] relative to subsequent months," and asserts that the purpose of the startup adjustment statute is "to take into account that a firm may experience unusually high costs when it is 'starting up' ... new production facilities.” See ICDAS Brief at 26 (quoting Statement of Administrative Action, H.R. Doc. No. 103-316, at 835, repiinted in 1994 U.S.C.C.A.N. at 4172); see also ICDAS Reply Brief at 10 (same). Whatever may be the underlying pwpose of the statute, the fact nevertheless remains that the language of the statute on its face authorizes Commerce to grant a startup adjustment only where a producer establishes that "production levels [were] limited by technical factors associated with the initial phase of commercial production.” See 19 U.S.C. § 1677b(f)(l)(C)(ii). This ICDAS failed to do.

28 . ICDAS further asserts that, even under Commerce’s assertedly "flawed” full-month equivalent methodology, the December 2003 production levels for the Biga facility were "significantly limited.” In support of its position, ICDAS compares full-month equivalent data on production starts for December 2003 to data on production starts for the first three months of 2004. With those data in mind, ICDAS emphasizes that "Commerce has found in a variety of antidumping contexts that differences greater than 20% to 25% are 'significant.' " See generally ICDAS Brief at 29 & n. 18.

However, as the Government observes, it is not enough for ICDAS to prove a difference (even a significant difference) between production starts in December 2003 and those in subsequent months. The issue is whether, in December 2003, Biga achieved “a level of commercial production that is characteristic of the merchandise.” See Antidumping Duties; Countervailing Duties: Final Rule, 62 Fed.Reg. 27,296, 27,364 (May 19, 1997) (Preamble). ' Comparing production starts data for December 2003 to production starts data, for subsequent months is essentially meaningless if, for example, production levels in those subsequent months were optimal. See Statement of Administrative Action, H.R. Doc. No. 103-316, at 836, reprinted in 1994 U.S.C.C.A.N. at 4173 (stating that “[attainment of peak production levels” is not the standard for identifying the end of a startup period). As the Government bluntly, puts it, "ICDAS does not provide any basis to conclude that just because the [December 2003] production levels were not optimal, they should still be considered startup levels.” See generally Def. Response Brief at 21-22; see also Domestic Producers.Response Brief at 21 n. 8 (arguing that "ICDAS' attempt to compare December 2003 production to March 2004 production should be rejected, as it appears to simply be an attempt to compare the commercial production levels achieved in December 2003 with 'optimum' production levels of March 2004”).

29 . ICDAS argues that "Commerce failed to identify any actual evidence ... that reasonably leads to its conclusion that the Biga melt shop had achieved a commercial level of production" in December 2003. See ICDAS Brief at 24. As the Government observes, however, ICDAS — in effect — seeks to turn the burden of proof on this issue on its head. See Def. Response Brief at 14-16.

The Statement of Administrative Action unequivocally places the burden of proving the -right to a startup adjustment squarely on the shoulders of the party seeking the adjustment:

The Administration intends that the burden will be on companies to demonstrate their entitlement to a-startup adjustment. Specifically, companies must demonstrate that, for the period under investigation or review, production levels were limited by technical factors associated with the initial phase of commercial production and not by factors unrelated to startup, such as marketing difficulties or chronic production problems. In addition, to receive a startup adjustment, companies will be required to explain their production situation and identify those technical difficulties associated with startup that resulted in the underutilization of facilities. This is consistent with the general rule in antidumping practice that a party seeking an adjustment has the burden of establishing entitlement to that adjustment as both a legal and factual matter.

Statement of Administrative Action, H.R. Doc. No. 103-316, at 838, reprinted in 1994 U.S.C.C.A.N. at 4174 (emphases added). See also PAM, S.p.A. v. U.S. Dep’t of Commerce, 27 CIT 671, 677, 265 F.Supp.2d 1362, 1367-68 (2003) (sustaining Commerce’s decision denying startup adjustment where plaintiff failed to prove it was entitled to adjustment); Agro Dutch Foods Ltd. v. United States, 24 CIT 510, 518 n. 10, 110 F.Supp.2d 950, 958 n. 10 (2000) (noting that, as to claim for startup adjustment, "[t]he burden of creating an adequate record lies with Agro Dutch, not with Commerce”).

Indeed, in drafting the regulations governing startup adjustments, Commerce expressly rejected one commenter’s suggestion that "once a respondent [had] made a prima facie case of entitlement to a startup adjustment, the Department would make the adjustment unless there was clear and convincing evidence that factors other than startup” were responsible for low production. Commerce explained: “[Ajccording to the [Statement of Administrative Action], the burden of proof undoubtedly rests with the party seeking a startup adjustment. Therefore, it is incumbent upon that party to (1) prove that the startup conditions [specified in the statute] existed during the period of ... review, and (2) as with any antidumping adjustment, document that fact to the Department’s satisfaction.” See Antidumping Duties; Countervailing Duties: Proposed Rule,-61 Fed.Reg. at 7340 (Preamble) (citing Statement of Administrative Action).

Moreover, as the Domestic Producers correctly note, this case certainly is no outlier. Commerce has denied requests for startup adjustments in other cases where there was insufficient record evidence to prove that limited production was the result of technical factors unique to startup. See Domestic Producers Response Brief at 25-26 (citing Notice of Final Determination of Sales at Less Than Fair Value: Certain Preserved Mushrooms from Chile, 63 Fed.Reg. 56,613, 56,618 (Oct. 22, 1998) (noting that respondent company failed to establish that its production levels were limited by technical factors associated with the initial phase of production); Certain Corrosion-Resistant Carbon Steel Flat Products and Certain Cut-to-Length Carbon Steel Plate From Canada: Preliminary Results of Antidumping Duty Administrative Reviews and Intent To Revoke in-Part, 63 Fed.Reg. 37,320, 37,324-25 (July 10, 1998) (unchanged in -Final Results) .(finding that respondent company’s production levels were limited not by technical factors unique to startup, but rather by "chronic production problems”)); see also, e.g., Notice of Preliminary Results of Antidumping Duty Administrative Review: Stainless Steel Bar From India, 68 Fed.Reg. 11,058, 11,062 (March 7, 2003) (unchanged in Final Results) (finding that respondent company’s low production levels were caused by "chronic production problems rather than technical factors associated with startup”).

30 . Commerce's supplemental Section D Questionnaire requested additional information on production start dates, capitalization of costs, and specifically how the technical factors ICDAS described limited the production levels that could be achieved. See Letter from Commerce to ICDAS (Dec. 21, 2004) (Pub.Doc. No. 97) at 3. In response, ICDAS provided the date that production began at Biga, as well as a breakdown of how costs were capitalized. But ICDAS gave no explanation whatsoever in response to Commerce’s request for information and documentation establishing exactly how the technical factors that ICDAS cited limited production levels. See Letter from ICDAS to Commerce (Jan. 25, 2005) (Pub. Doc. No. 116; Conf. Doc. No. 24) at 18.

31 . ICDAS asserts that its response to Commerce's Section D Questionnaire constituted adequate evidence that the assertedly limited production at the Biga facility in December 2003 was attributable to technical factors unique to startup. See ICDAS Brief at 30. ICDAS there stated:

Production levels were limited by technical factors associated with the initial phase of commercial production because the company had to 1) develop the production parameters of the new operations; 2) install, adjust, calibrate and test the new equipment; and 3) train new employees to operate the new equipment. Operations typically incur such technical problems because of the newness of the facility.

Section D Questionnaire Response of ICDAS Celik Enerji Tersane ve Ulasim Sanayi, A.S. (Pub.Doc. No. 67) at D-40.

ICDAS argues that the factors that it listed in its’ Section D Questionnaire Response closely parallel the factors cited by Commerce as a basis for the startup adjustment granted in another case, SRAMs from Taiwan. See ICDAS Brief at 30 (citing Notice of Final Determination of Sales at Less Than Fair Value: Static Random Access Memory Semiconductors From Taiwan, 63 Fed.Reg. 8909, 8930 (Feb. 23, 1998) ("SRAMs from Taiwan”) (finding that "the development of process pa rameters, cleaning of the ... facility, and installation, adjustment, calibration, and testing of new equipment” were technical factors unique to startup operations)). However, all ICDAS points to is the Federal Register notice in the SRAMs proceeding. ICDAS provided no information from the underlying administrative record to indicate the nature or quantum of evidence submitted to Commerce by the producer there to substantiate the causal link between the listed factors and the limited production it experienced in its initial phase of operations — in other words, the evidence to substantiate its claim that technical factors unique to startup were to blame for limited production in its initial phase of operations. There is thus no basis to conclude that Commerce granted the startup adjustment in SRAMs from Taiwan on the strength of an evidentiary record as thin as the record here. Moreover, it would seem to be a near-universal truth that new facilities everywhere must "develop the production parameters of ... new operations; 2) install, adjust, calibrate and test ... new equipment; and 3) train new employees to operate the new equipment” (quoting ICDAS' Section D Questionnaire Response). Indeed, ICDAS itself observed that "[ojperations typically incur such technical problems because of the newness of the facility.” Id. It is difficult to imagine that Congress could have intended that such bald, generalized statements of near-universal truth would suffice to satisfy the requirements that a producer seeking a startup adjustment "demonstrate that, for the period under investigation or review, production levels were limited by technical factors associated with the initial phase of commercial production and not by factors unrelated to startup, such as marketing difficulties or chronic production problems,” and, further. "explain their production situation and identify those technical difficulties associated with startup that resulted in the underutilization of the facilities.” See Statement of Administrative Action, H.R. Doc. No. 103-316, at 838, reprinted in 1994 U.S.C.C.A.N. at 4174. Indeed, the recitation of the requirements in the Statement of Administrative Action is longer than the sentence that ICDAS relies on as evidence to satisfy those requirements.

32 . In a back-door attempt to demonstrate that the asserted limitations on production at the Biga facility in December 2003 were due to technical factors unique to startup, ICDAS emphasizes that "[tjhere has been no allegation, nor does the record reflect, that production [at Biga in December 2003] was limited by 'factors unrelated to startup, such as marketing difficulties or chronic production problems.’ ” See ICDAS Brief at 30 n. 19 (quoting Statement of Administrative Action, H.R. Doc. No. 103-316, at 838, reprinted in 1994 U.S.C.C.A.N. at 4174).

As discussed above, however, it was ICDAS that bore the burden of affirmatively establishing that technical factors unique to startup were the cause of the assertedly limited production at the Biga facility in December 2003; thus, it was ICDAS that bore the burden of (at least implicitly) eliminating other potential causes of limited production. See n. 29, supra. Contrary to ICDAS' implication, neither Commerce nor the Domestic Producers was under any obligation to prove that any asserted limitations on production in the startup phase were attributable to "factors unrelated to startup, such as marketing difficulties or chronic production problems.” The fact that the record is devoid of evidence of any such "factors unrelated to startup” is therefore of no moment.

33 . As the Government notes, when Commerce receives insufficient information from an interested party to make a determination, the statute and regulations authorize the agency to fill in the gaps in the facts with "facts otherwise available.” See Def. Response Brief at 18 (citing 19 U.S.C. § 1677e(a)); see abo 19 C.F.R. § 351.308(a). If Commerce finds that the information is not available because the party "has failed to cooperate by not acting to the best of its ability to comply with a request for information,” the statute and regulations further provide that Commerce "may use an inference that is adverse to the interests of that party in selecting from among the facts otherwise available.” See Def. Response Brief at 18 (citing 19 U.S.C. § 1677e(b)); see also 19 C.F.R. § 351.308(a); Nippon Steel Corp., 337 F.3d at 1380-81 (summarizing operation of statutory and regulatory provisions governing use of "facts otherwise available” and adverse inference). As explained below, however, the concepts of "facts otherwise available” and adverse inference have no application here.

34 . ICDAS correctly observes that — if the date of sale analysis conducted by Commerce in this case actually were the rule — the nature of the information provided to Commerce in questionnaire responses and the information confirmed by the agency through its verification process would be radically different. Rather than analyzing the nature of a respondent's sales process, Commerce would simply survey a respondent’s documentation to determine whether there had been any change in any material term of sale in any contract at issue. But the latter was not Congress’ intent; nor is it reflected in Commerce's own date of sale regulation. See Antidumping Duties; Countervailing Duties: Final Rule, 62 Fed.Reg. at 27,364 (Preamble) (indicating that Commerce will verify "a respondent's description of its selling processes” to determine appropriate date of sale); see also ICDAS Response Brief at 17 n. 13.

35 . The Government peppers its brief with repeated invocations of Commerce’s "discretion” (which, as noted above, it claims is "absolute”). See, e.g., Def. Supp. Response Brief at 6 (quoting Colakoglu Metalurji A.S. v. United States, 29 CIT 1238, 1240, 394 F.Supp.2d 1379, 1381 (2005), for proposition that, if material terms of sale were fixed at different time, Commerce "has the power to exercise discretion” by using different date of sale); id. at 7 (quoting Hornos Electricos de Venezuela, S.A. v. United States, 27 CIT 1522, 1536, 285 F.Supp.2d 1353, 1367 (2003) (Hevensa ) and its discussion of Commerce’s "discretion”); id. at 8 (citing Hevensa, and referring to Commerce’s "discretion,” and to agency’s use of a date of sale other than invoice date as a "discretionary act”); id. at 11 (asserting that, although Commerce "may exercise its discretion” to use date other than invoice date as date of sale, "it is not required to do so”); id. at 12 (indicating that, in case at bar, Commerce exercised "its discretion” in deciding to use invoice date as date of sale); id. at 12-13 (opining that, "even if Commerce did not possess discretion,” outcome of case would not differ); see also Remand Results at 18 (quoting Hevensa).

36 . The only "preference” expressed by Congress and Commerce is that the date of sale used in the agency’s analysis accurately reflect the date on which "the material terms of sale” were established. See discussion of Uruguay Round Agreements Act, supra (explaining that legislation incorporated into U.S. law the date of sale provision in relevant WTO agreement); 19 C.F.R. § 351.401(i); Antidumping Duties; Countervailing Duties: Final Rule, 62 Fed.Reg. at 27,349 (Preamble). And, in fact, that is not merely a "strong preference”; it is a requirement.

37 . Thai Pineapple simply stated that the rarity of changes in contractual terms, in and of itself, is not reason enough to use contract date as the date of sale. The court in that case therefore reversed Commerce's decision to use contract date, because the agency’s rationale was limited solely to the fact that changes in the terms of sale were infrequent. See also Final Results of Redetermination Pursuant to Court Remand, Habas Sinai ve Tibbi Gazlar Istihsal Endustrisi A.S. v. United States, No. 05-00613 (Ct. Int’l Trade March 3, 2008) (“Habas Remand Results”) at 41, 46 (rejecting, in companion case to this one, domestic producers’ claim that — under Thai Pineapple — deviation from use of invoice date as date of sale is warranted only in "unusual” circumstances).

At the same time, the Thai Pineapple court expressly acknowledged that there are a number of factors that have been recognized as supporting the use of contract date as the date of sale, none of which were present in that case:

Commerce does not cite industry practice or a lag between invoice and shipment, or any other unusual situation, indicating a date[] other than invoice date should be used. There appears to be no other case in which "rare instances” of changes after contract date ... was considered substantial reason to abandon the invoice date presumption.

Thai Pineapple, 24 CIT at 109.

Here — in contrast to Thai Pineapple — ICDAS has identified a range of factors supporting the use of contract date as the date of sale (in addition to the rarity of changes in terms of sale), including the lag time between invoice and shipment, and the practice of ICDAS (as well as other Turkish producers) to use formal contracts for sales of manufactured-to-order rebar to the U.S. market. See generally ICDAS Response Brief at 13 n. 9.

38 . As ICDAS correctly notes, contrary to the Domestic Producers' assertions, there is nothing at all "unusual” about Commerce using a date other than invoice date as the date of sale. In fact, in just the five-month period following Commerce’s issuance of its Decision Memo in this case, Commerce issued at least seven determinations in which the agency did not use invoice date as the date of sale. Compare Domestic Producers Reply Brief at 9 with ICDAS Response Brief at 13 n. 9 (citing Commerce determinations using "contract date,” "final contract date,” "date of final amendment to contract addendum,” "email confirmation date,” and "order acknowledgment date” as date of sale).

39 . Further, as ICDAS observes, neither the Remand Results nor the Government’s brief cited insufficient documentation as grounds for using invoice date as the date of sale. See ICDAS Supp. Reply Brief at 7 n. 4; see also id. (indicating that it is neither "legally required” nor "practically feasible” for a foreign producer "to submit complete sales traces for every transaction that may be included in Commerce's universe of sales”).

Indeed, in a companion case to this one, Commerce recently expressly rejected the exact same argument that the Domestic Producers raise here. See Habas Remand Results at 41-42, 46-47 (dismissing domestic producers’ argument that a respondent advocating use of contract date as the date of sale for its U.S. sales is required to supply all U.S. sales documentation).

40 . See also USEC Inc. v. United States, 31 CIT -, --, 498 F.Supp.2d 1337, 1343-44 (2007) (quoting SeAH Steel Corp. v. United States, 25 CIT 133, 135 (2001), to explain that agency may use a proposed date other than invoice date as date of sale either "if 'material terms’ are not subject to change between the proposed date and the invoice date” or if "the agency provides a rational explanation as to why the alternative date ‘better reflects’ the date when ‘material terms' are established”).

41 . Though Commerce summarily rejected it, ICDAS makes the fair point that — while Congress and the agency may have legitimate concerns about scenarios in which U.S. buyers pressure foreign producers to obtain the lowest possible price — the facts of this case are, in essence, exactly the opposite, and thus raise no such issues. Unlike the scenarios of concern, ICDAS here sought a price increase; it did not offer a price decrease. Nor was there any attempt to import a greater quantity of rebar than specified in the parties' contract. In other words, the change in this case is not the sort of change that concerned Congress as it considered the effect on the date of sale of a change to a material contract term. See Remand Results at 7-8, 21; Transcript of Oral Argument ("Tr.”) at 87.

42 . Commerce (in the Remand Results) and the Government (in its brief) seek to dismiss Pipes and Tubes from Thailand as irrelevant to this case, emphasizing that Pipes and Tubes from Thailand involved changes to contractually-specified quantities (rather than price), and that the changes there were within contractual tolerances. See Remand Results at 22; Def. Supp. Response Brief at 8-9. However, neither Commerce nor the Government disputes the accuracy of the quotation (above) from Commerce’s determination in that matter. Nor can they dispute the broader principle for which ICDAS cites the case— Commerce's own recognition of the need for "flexibility” in its date of sale analyses. See ICDAS Response Brief at 16 n. 11.

43 .See also, e.g., Pipe from Korea, 63 Fed. Reg. at 32,835-36; Issues and Decision Memorandum for the Final Determination of the Antidumping Duty Investigation of Polyethylene Retail Carrier Bags from Thailand, 2004 WL 3524397 (June 18, 2004) ("Polyethylene Retail Carrier Bags from Thailand”), at comment 2; Uranium from France, 2005 WL 2305751, at comment 11; Rebar from Latvia, 2006 WL 3702620, at comment 2; Issues and Decision Memorandum for the Administrative Review of Certain Cut-to-Length Carbon Steel Plate from Romania: Final Results of Antidumping Duty Administrative Review and Final Partial Rescission, 2007 WL 527754 (Feb. 12, 2007) (“Steel Plate from Romania"), at comment 1.

44 . See, e.g., Pipe from Korea, 63 Fed.Reg. at 32,835-36; Sulfanilic Acid from Portugal, 2002 WL 31493754, at comment 1; Uranium from France, 2005 WL 2305751, at comment 11; Rebar from Latvia, 2006 WL 3702620, at comment 2; Polyethylene Retail Carrier Bags from Thailand, 2004 WL 3524397, at comment 2.

45 . The Government attempts to distinguish several of the cases on which ICDAS relies to support this principle. See Def. Supp. Response Brief at 9 (seeking to distinguish Pipe from Korea); id. at 9-10 (seeking to distinguish Sulfanilic Acid from Portugal); id. at 10 (seeking to distinguish Polyethylene Retail Carrier Bags from Thailand).

.As ICDAS observes, however, ''[t]he Government's efforts to distinguish these determinations are wasted,” because "ICDAS never contended that [the cited cases] were factually identical to the instant case. Instead, ICDAS cited the[] determinations to illustrate the general propositions 'that evidence of change, while relevant, does not end the date of sale analysis' and that ‘when confronted with evidence of isolated change, Commerce must conduct a factual analysis of the contracting parties’ expectations and conduct to understand when they had a true meeting of the minds.' ” ICDAS Supp. Reply Brief at 6; see also ICDAS Response Brief at 19 n. 14 (discussing Polyethylene Retail Carrier Bags from Thailand). The Government's arguments do nothing to diminish in any way ICDAS’ point that Commerce’s date of sale analyses in other cases involving changes to material contract terms did not begin and end with the fact of the change.

46 . As ICDAS points out, both this case and Pipe from Korea are very different from Hevensa. See ICDAS Response Brief at 19 n. 15 (discussing Hevensa, 27 CIT 1522, 285 F.Supp.2d 1353). Unlike the respondent in Hevensa, both ICDAS and the Korean respondent have distinctly different home market and U.S. sales processes. Although the respondent in Hevensa reported contract date as the date of sale for both its U.S. and home market sales, the respondent failed to identify any aspect of either sales process to justify use of contract date as the date of sale. See Hevensa, 27 CIT at 1535-37, 285 F.Supp.2d at 1366-67. In contrast, ICDAS’ U.S. sales process differs markedly from its home market sales process. In contrast to typical home market sales, ICDAS’ U.S. sales involve formally-negotiated contracts to establish the material terms of sale on the date of the contract, because rebar sold for the U.S. market is produced, marked, packaged, and shipped differently than rebar sold in the home market. Accordingly, for ICDAS — as for the respondent in Pipe from Korea — " ‘invoice’ dates in both [the U.S. and the home] markets, while the same in name, are materially quite different” for purposes of Commerce’s date of sale analysis. See Pipe from Korea, 63 Fed.Reg. at 32,836.

47 . Thus, for example, it appears from the record that ICDAS’ sales prices (like other critical sales terms) are proprietary information. Under the circumstances, it is not clear how — as a practical matter' — the expectations of other parties to other contracts could have been affected in any way by a price change of which they were unaware. Other aspects of Commerce’s reasoning on this point are equally unclear.

48 .In the Remand Results, Commerce attempts to distinguish Pipe from Korea from the instant case on the grounds that "the percentage of ICDAS’s U.S. entries that had changes in material terms of sale [was] significant.” See Remand Results at 22. But — as a matter of pure logic — Commerce’s statement, without more, is insufficient to justify treating the two cases differently. It is entirely unclear how (particularly given the facts of this case) the portion or “percentage of ICDAS’s U.S. entries” affected by the price change is relevant to parties' expectations concerning the binding nature of their contracts. See also ICDAS Response Brief at 19 (characterizing Commerce’s attempt to distinguish Pipe from Korea from the case at bar as "without merit,” emphasizing that "the focus in Pipe from Korea is on the nature of the sales process and the frequency of change, not on the volume of entries that are affected by a single change”).

Commerce, the Government, and the Domestic Producers take pains to underscore the percentage of ICDAS’ U.S. sales affected by the price increase. See Remand Results at 4, 17, 19-20; Def. Supp. Response Brief at 2, 4, 11-12; Domestic Producers Reply Brief at 11-13. Indeed, Commerce attached tremendous (seemingly determinative) weight to that fact in the Remand Results:

We agree with ICDAS that a single change in price does not automatically disqualify contract date from selection as the proper date of sale. In this case, however, we find that the price change, while limited to a single contract, related to a significant percentage of ICDAS’s U.S. entries during the [Period of Review]. Under these circumstances, we determine that the contract date does not represent the date on which the parties had a real "meeting of the minds”....

Remand Results at 20. As discussed above, however, neither Commerce, nor the Government, nor the Domestic Producers has proffered a rational explanation of the significance of the portion or percentage of ICDAS' sales affected by the price increase vis-a-vis parties' expectations concerning the binding nature of their contracts (or any other factor central to Commerce’s date of sale determina- • tion).

Finally, to the extent (if any) that the percentage of ICDAS’ U.S. sales affected by the price change may be relevant to determining date of sale, it is worth noting that the figure is highly dependent on the universe of sales used in Commerce’s analysis — another issue which is hotly contested in this action. See ICDAS Response Brief at 12 n. 8; section III.F, infra (discussing issue of universe of sales).

49 . Somewhat perversely, Commerce and the Government seek to use the fact of the contract amendments in Sulfanilic Acid from Portugal' as a basis for' distinguishing that case from this one. Quoting Commerce’s Remand Results, the Government argues that "ICDAS did not amend the contract at issue, ‘but rather merely issued sales invoices reflecting different material terms of sale (i.e., price).’ ” See Def. Supp. Response Brief at 9-10 (quoting Remand Results at 23). However, "[m]ore than enumeration of factual differences between cases is required; [Commerce and the Government] must explain their relevance.” Burinskas, 357 F.2d at 827 n. 5. Neither Commerce nor the Government has explained the significance of the fact that the price change here was reflected in an invoice, rather than a formal contract amendment. Nor is it obvious why that difference would distinguish the expectations of the parties in the two cases. In both this case and Sulfanilic Acid from Portugal, the contract term which was changed was price; but — • while there was only one price increase here — in Sulfanilic Acid, the price was changed twice.

50 . See also, e.g., Uranium from France, 2005 WL 2305751, at comment 11 (using contract date "except for one instance where the terms of sale were revised, making the date of the amended contract the appropriate date of sale" for that one instance); Rebar from Latvia, 2006 WL 3702620, at comment 2 (using date of final contract amendment as date of sale).

51 . The Remand Results misconstrue one specific aspect of ICDAS' remand comments. In the Remand Results, Commerce states that ICDAS contends that there is a "three-factor test” governing date of sale analyses. In fact, ICDAS never made any such claim. Compare Remand Results at 23 and ICDAS Remand Comments (Remand Pub. Doc. No. 4) at 9-11. In its remand comments, ICDAS simply pointed out that Commerce’s past determinations identify a number of indicia that the agency has cited to support its use of contract date as the appropriate date of sale for purposes of antidumping analyses. Id. In its briefs filed with the court, ICDAS continues to assert that three such indicia — use of formal contracting procedures, made-to-order merchandise, and lag time — are present in this case. See, e.g., ICDAS Response Brief at 21-23, 23-25 (discussing use of formal contracting procedures), 25-27 (discussing lead time and made-to-order nature of merchandise); ICDAS Supp. Reply Brief at 7-9 (discussing use of formal contracting procedures), 9-12 (discussing lead time and made-to-order nature of merchandise).

52 . Commerce and the Domestic Producers have argued that the interests of administrative consistency counsel use of invoice date as the date of sale here. See Remand Results at 15 & n. 4, 20; Domestic Producers Reply Brief at 8, 14.

In the Remand Results, for example, Commerce asserted that its "preference is to select a date of sale methodology and apply it consistently across all segments of a proceeding, unless there is evidence of changes in selling practices between segments.” Remand Results at 20. However, administrative consistency has been roundly rejected as a rationale sufficient to justify Commerce's continued adherence to any particular date of sale methodology. See SeAH Steel, 25 CIT at 137 (noting that such logic “would obviate the need for any date of sale analysis in all reviews beyond the first administrative review,” and holding, inter alia, that Commerce "under the guise of ‘administrative consistency,’ may not ... abdicate its statutory duty to ensure that normal value is calculated 'at a time reasonably corresponding to the time of the sale used to determine the export price or constructed export price’ ") (citation omitted); see also Pipes and Tubes from Thailand, 65 ITADOC 60910, at comment 1 (using a date of sale different than that used by Commerce in previous administrative review, notwithstanding lack of change in respondent’s sales practices). The "administrative consistency” argument advanced by Commerce and the Domestic Producers thus has no basis in the law.

The factual foundation for the “administrative consistency” argument is just as weak. In their brief, the Domestic Producers sought to make much of the fact that Commerce "ha[d] found that invoice date [was] the appropriate date of sale for every respondent in every [administrative] review of this [anti-dumping] order.” See Domestic Producers Reply Brief at 8; see also id. at 14 (emphasizing "the consistent use of invoice date in every other review of this order”). But Commerce has since reversed itself, and has revised its analysis for the administrative review for 2002-2003 to use "order date” (contract date) — rather than invoice date — as the date of sale for U.S. sales by another Turkish rebar producer, Colakoglu. See Colakoglu Metalurji A.S. v. United States, 30 CIT 281, 281-82 (2006). It is thus no longer true that Commerce has consistently used invoice date as the date of sale for every respondent in every administrative review of the antidumping duty order in question. Moreover, the date of sale issue is being litigated by a third Turkish rebar producer, Habas, in a companion case challenging Commerce's determination in the 2003-2004 administrative review — the administrative review at issue here. Although Commerce initially used invoice date as the date of sale in that case, the issue was remanded to the agency for further consideration. In the remand results, Commerce concluded that contract date was the appropriate date of sale for Habas' U.S. sales, and recalculated Habas' dumping margin accordingly. See Habas Sinai Ve Tibbi Gazlar Istihsal Endustrisi A.S. v. United States, 31 CIT -, -, 2007 WL 3378201 at **5-8 (2007); Habas Remand Results at 19-22, 45-49.

Commerce, the Government, and the Domestic Producers seek to make a second, somewhat related point that similarly misses the mark. Both in the Remand Results and in their briefs filed in this matter, Commerce, the Government, and the Domestic Producers highlight the fact that — during the verification process — ICDAS representatives did not press Commerce to use contract date as the date of sale for the company’s U.S. sales. Commerce, the Government, and the Domestic Producers seem to intimate that the position of the company representatives at verification constitutes potent evidence in support of Commerce’s determination in the Remand Results. See Remand Results at 17, 20; Def. Supp. Response Brief at 12; Domestic Producers Reply Brief at 8, 14-15.

As ICDAS notes however, the company was not represented by counsel at verification; and, moreover, the pivotal significance of the date of sale issue became clear only after Commerce changed its universe of sales methodology. See ICDAS Response Brief at 6 n. 3, 9 nn. 5 & 6. More to the point, however, Commerce generally accords relatively little weight to respondent companies' views in the agency's date of sale determination, focusing (properly) on what the evidence shows instead. See, e.g., Pipe from Korea, 63 Fed.Reg. at 32,834-36 (using contract date as date of sale, despite respondent's assertions that invoice date was appropriate); see generally ICDAS Response Brief at 9 n. 5, 22 n. 16. Indeed, in the Final Results in this case, Commerce had no hesitation in finding that contract date was the proper date of sale, notwithstanding the ICDAS representatives’ position at verification. If the company representatives' position gave Commerce no pause at the Final Results stage, there is no apparent reason why their position should carry any significant weight now.

53 . The Domestic Producers reject out of hand ICDAS’ arguments concerning its use of a formal contracting process, as well as. its arguments based on the fact that the merchandise it sells in the U.S. market is manufactured-to-order. According .to the Domestic Producers, those factors "can only justify a deviation from the presumption in favor of invoice date if [it has already been demonstrated] that the material terms of sale are in fact definitively established as of the contract date” (which, the Domestic Producers maintain, "is clearly not the case here”). See Domestic Producers Reply Brief at 14; see also Remand Results at 17-18.

However, the Domestic Producers cite no authority to support their summary dismissal of ICDAS’ evidence. Nor does the Domestic Producers' position have any sound foundation in reason. In other words, contrary to the Domestic Producers’ implication, the fact that merchandise must be manufactured-to-order and the fact that a buyer and a seller enter into a formal,.written contract logically would seem to constitute evidence bearing directly on the question of when the parties had a meeting of the minds on the material terms of sale (which is central to determining date of sale). Indeed, ICDAS' position finds strong support in the Preamble to Commerce's date of sale regulation, which expressly states that "in situations involving large custom-made merchandise in which the parties engage in formal negotiation and contracting procedures, the Department usually will use a date other than the date of invoice” as the date of sale. See Antidumping Duties; Countervailing Duties: Final Rule, 62 Fed. Reg. at 27,349 (Preamble) (emphases added).

54 . The Preamble to Commerce’s date of sale regulation recognizes that the type of formal negotiation and contracting process used in ICDAS' U.S. sales is not the norm:

The Department ... has found that in most industries, the negotiation of a sale can be a complex process in which the details often are not committed to writing. In such situations, the Department lacks a firm basis for determining when the material terms, were established. In fact, it is not uncommon for the buyer and seller themselves to disagree about the exact date on which the terms became final.

Antidumping Duties; Counterváiling Duties: Final Rule, 62 Fed.Reg. at 27,349 (Preamble) (emphases added). The evidence cited by ICDAS suggests that — unlike the typical case referenced in the provision quoted above— Commerce here has (in the words of the Preamble) a "firm basis” for determining when ICDAS and its U.S. customers had a meeting of the minds on the material terms of sale. Id. Each step of the negotiation process was documented in writing; and, as to each and every sale, the process culminated in a formal written contract detailing the specific terms of that sale. No party has pointed to either evidence or legal authority to suggest that ICDAS’ written contracts were not legally binding instruments but instead mere "preliminary agreements] on terms,” where "renegotiation [was] common.” See Anti-dumping Duties; Countervailing Duties: Final Rule, 62 Fed.Reg. at 27,349 (Preamble).

55 . The lead time required to produce large, custom-made merchandise often results — as it did here — in significant "lag time” between the contract date and invoice date. As ICDAS points out, Commerce has acknowledged lag time as a factor warranting departure from the use of invoice date as the date of sale in a number of other cases, including Pipe from Korea and Extruded Rubber Thread from Malaysia. See Remand Results at 13 (citing Pipe from Korea, 63 Fed.Reg. at 32,836 (using contract date — rather than invoice date — as date of sale, due to lag time); Extruded Rubber Thread from Malaysia; Final Results of Antidumping Duty Administrative Review, 64 Fed.Reg. 12,967, 12,968 (March 16, 1999) ("Extruded Rubber Thread from Malaysia”) (using date of bill of lading — rather than invoice date — as date of sale, due to lag time)); ICDAS Response Brief at 25-26 & n. 18 (same).

ICDAS contends that ignoring the lag time between contract date and invoice date in this case yields “significant distortions” in Commerce’s dumping margin calculations in the Remand Results. See ICDAS Response Brief at 26 n. 18. Highlighting the significance of the lag time issue, ICDAS notes that Commerce itself explained in its original questionnaire: "Because the Department attempts to compare sales made at the same time, establishing the date of sale is an important part of the dumping analysis.” Id. (quoting Commerce Questionnaire (Pub.Doc. No. 11) at I-5). Commerce must accurately determine the date on which sales are finalized, because factors such as inflation, changing input prices, and market fluctuations can cause significant distortions in the antidumping analysis. See, e.g., Colakoglu, 29 CIT at 1240, 394 F.Supp.2d at 1381. ICDAS asserts that Commerce's blanket use of invoice date in the Remand Results here prevents the requisite •accurate comparisons, because Commerce is "effectively ... comparing home market sales in any given month to U.S. sales whose material terms were set” in an entirely different month, under different market conditions. See ICDAS Response Brief at 26 n. 18 (quoting Pipe from Korea, 63 Fed.Reg. at 32,836); ICDAS Supp. Reply Brief at 10. According to ICDAS, ”[t]he significant distortion caused by this inaccurate comparison is reflected in Commerce’s recalculation of ICDAS’ dumping margin from de minimis to 1.63 percent.” ICDAS Response Brief at 26 n. 18; see also ICDAS Supp. Reply Brief at 10.

Commerce, the Government, and the Domestic Producers all have sought to downplay the issue of lag time here. The Domestic Producers’ brief is silent on the subject. But, on remand, the Domestic Producers tried to distinguish Pipe from Korea from the case at bar on the grounds that the lag time in Pipe from Korea was greater than it is here; and the Government makes the same argument in its brief. See Remand Results at 18; Def. Supp. Response Brief at 9. As discussed above, however, the lag time between contract date and invoice date in this case makes the difference between a de minimis dumping margin (with revocation of the antidumping order as to ICDAS) and a dumping margin of 1.63% (which precludes revocation). The lag time issue here thus cannot be so readily dismissed. Cf. 19 C.F.R. § 351.224(g) (recognizing that ministerial error is "significant” when correction of error would make "a difference between a weighted-average dumping margin ... of zero (or de minimis) and a weighted-average dumping margin ... of greater than de minimis ”). Moreover, Commerce has recognized lag time as a factor warranting departure from the use of invoice date as the date of sale in other cases where the lag time was well short of that in Pipe from Korea.

Commerce's efforts to distinguish Extruded Rubber Thread from Malaysia are no more successful. In the Remand Results, Commerce asserted that ICDAS' reliance on that case is "misplaced,” because the case "merely illustrates the Department's policy of using the date of shipment as the date of sale where shipment occurs prior to invoicing.” See Remand Results at 24. This is specious. As ICDAS correctly notes, in Extruded Rubber Thread from Malaysia, Commerce used the bill of lading date — rather than invoice date— as the date of sale where merchandise was shipped directly from Malaysia to U.S. customers. Commerce specifically cited the "long lag time between the date of shipment to the customer and the date of invoice” as the factor rebutting the regulatory presumption of invoice date. See ICDAS Response Brief at 26 n. 18; Extruded Rubber Thread from Malaysia, 64 Fed.Reg. at 12,968. Because contract date was not at issue in that case, and because the foreign producer relinquished control over the merchandise at the time of shipment, the bill of lading date (rather than invoice date) reflected the date when the material terms of sale were established. Cf. Mittal Steel Point Lisas Ltd. v. United States, 548 F.3d 1375, 1385 (Fed.Cir.2008) (noting that "once goods have been shipped from a foreign port, the material terms of sale have been set, as the seller may not then sell those goods to another customer”). Extruded Rubber Thread from Malaysia thus stands for the proposition for which ICDAS cites it: Commerce has recognized "lag time” as a factor which may rebut the regulatory presumption favoring invoice date as the date of sale.

56 . In this sense, the case at bar is comparable to other cases dealing with large, custom-made merchandise. See generally ICDAS Response Brief at 26 n. 19 (citing Final Determination of Sales at Less Than Fair Value: Mechanical Transfer Presses from Japan, 55 Fed.Reg. 335, 341 (Jan. 4, 1990); Notice of Final Determination of Sales at Less Than Fair Value: Large Newspaper Printing Presses and Components Thereof, Whether Assembled or Unassembled, from Japan, 61 Fed.Reg. 38,-139, 38,159 (July 23, 1996)). As those cases illustrate, consistent with the agency's explanation in the Preamble to the date of sale regulation, Commerce generally has found that — when dealing with physically large, expensive merchandise that is made-to-order — it is appropriate to use the contract date as the date of sale, because the parties in such cases typically "engage in formal negotiation and contracting procedures.” See Antidumping Duties; Countervailing Duties: Final Rule, 62 Fed.Reg. at 27,349 (Preamble). The great expense involved in the production of such merchandise, in addition to the relative absence of an open market for merchandise that is (at least to some degree) customized, requires that parties commit themselves to sales at an early date, in contrast to the common practice in sales from inventory. See, e.g., Issues and Decision Memorandum for the Antidumping Duty Administrative Reviews on Large Newspaper Printing Presses and Components Thereof, Whether Assembled or Unassembled, from Germany — September 1, 1998, Through August 31, 1999, 2001 WL 193869 (Feb. 26, 2001), at comment 1.

In the Remand Results, Commerce attempted to distinguish the cases on which ICDAS relies; but Commerce's efforts are largely futile. Contrary to Commerce’s assertions, it is of no moment here that, in the cases cited by ICDAS, there were no changes in quantity or price outside contractual tolerances. See Remand Results at 22. The cases nevertheless stand for the proposition for which ICDAS has cited them: Consonant with the rationale set forth in the Preamble to its date of sale regulation, "in situations involving large custom-made merchandise in which the parties engage in formal negotiation and contracting procedures,” Commerce generally uses a date other than invoice date as the date of sale. See Antidumping Duties; Countervailing Duties: Final Rule, 62 Fed.Reg. at 27,349 (Preamble). Commerce’s observation that the newspaper printing presses in the cases that ICDAS cites were "highly customized” and thus "could not' be resold to alternative purchasers” is similarly unavailing. See Remand Results at 23. Again, even assuming the truth of Commerce's observations, they do nothing to detract from the fact that the cases cited by ICDAS still stand for the proposition for which ICDAS has cited them. Moreover, as discussed elsewhere herein, Commerce greatly overstates ICDAS' ability to sell in the home market merchandise that has been custom-produced and rolled in' inches in conformance with U.S. standards. The Domestic Producers’ argument has no greater merit. Compare Remand Results at 21-22; Def. Supp. Response Brief at 13 with n. 53, supra.

ICDAS candidly acknowledges that rebar may not be as structurally complex as the industrial presses at issue in the cases that it cites on this point. See ICDAS Response Brief at 26 n. 19. But no party has explained the relevance of the intricacy or complexity of the merchandise to the issue at hand. In any event, the cases that ICDAS cites are analogous to ICDAS’ U.S. sales here, because — like the sales of the industrial presses at issue in those cases — ICDAS’ U.S. sales here required formal contracts that covered the sale of tens of thousands of tons of merchandise custom-manufactured to U.S. specifications, which was worth millions of dollars and which required substantial lead time to produce.

57 . In the Remand Results, Commerce found that rebar produced pursuant to a contract with a U.S. buyer could be sold in ICDAS' home market if the contract with the U.S. buyer fell through. See Remand Results at 21-22. To the contrary, as Commerce correctly found at verification, rebar sold in the U.S. market is made-to-order. See ICDAS . Sales Verification Report (Conf.Doc. No. 44) at 4. Moreover, although rebar sold in the U.S. market and that sold in the home market may have equivalent strength specifications, the dimensions vary. Specifically, rebar for the U.S. market is rolled in inches; in contrast, rebar for the home market is rolled in millimeters.

ICDAS’ home market sales database indicates that, during the period of review, there was only a single sale of U.S.-sized rebar; and that sale was for a relatively modest quantity. That single, insignificant sale cannot constitute substantial evidence that ICDAS could have sold in its home market — in

anything remotely resembling the normal course of business — the thousands of tons of U.S.-sized rebar that it sold in the U.S. market during the period of review. See generally ICDAS Response Brief at 27 n. 20; ICDAS Supp. Reply Brief at 11.

The Government’s argument on this point amounts to little more than a bald, unsupported assertion that the fact that Commerce found that the rebar ICDAS sold in the United States was made:to-order “does not preclude its resale.” See Def. Supp. Response Brief at 13. The Domestic Producers steer clear of the topic entirely; and their silence speaks volumes. In short, the record evidence strongly supports ICDAS’ claim that— contrary to Commerce's conclusion in the Remand Results — “[t]here is simply no meaningful home market demand for U.S.-sized re-bar.” See generally ICDAS Response Brief at 27; see also ICDAS Supp. Reply Brief at 11-12.

58 . The record evidence indicates that the price increase occurred sometime between the issuance of the letter of credit and the issuance of the invoice. See ICDAS Sales Verification Report (Conf.Doc. No. 44), Exh. 14 at domestic inland freight documentation.

59 . In the Remand Results, Commerce characterized as sheer “speculation” ICDAS’ statement that it would have honored its contract even if it had not succeeded in negotiating the price increase. See Remand Results at 21. But, given the facts of this case (as documented by record evidence), it strains credulity to contend that ICDAS would have breached a legally-binding contract of the magnitude at issue, where it had already rendered substantial performance, for a comparatively modest increase in sales revenue. See ICDAS Sales Verification Report (Conf.Doc. No. 44), Exh. 14 at contract & credit notice. ' '

Because ICDAS was — as a practical matter — unable to inventory or resell in the home market large quantities of rebar which were rolled to non-metric dimensions, ICDAS would have been stuck at the Turkish port with (quite literally) thousands of tons of re-bar, if it had breached the contract. In fact, as the record reflects, there would be no demand for the non-metric rebar until ICDAS’ next U.S. sale, which occurred months later.

Neither Commerce, nor the Government, nor the Domestic Producers has pointed to any concrete evidence to refute the basic facts recited above, which ICDAS has supported with ample citations to the record. Indeed, as discussed above, the other parties have gone to great lengths to emphasize the volume of rebar which was subject to the price increase, in their attempts to distinguish this case from other cases where Commerce has used contract date as the date of sale notwithstanding changes to material contract terms. See n. 48, supra. As noted there, however, no party has yet explained how the volume of rebar subject to the price increase is relevant to parties’ expectations concerning the binding nature of their contracts. Id. Nevertheless, by emphasizing the volume of rebar subject to the price increase, Commerce, the Government, and the Domestic Producers effectively (albeit unintentionally) bolster ICDAS’ argument here. The record evidence suggests that, given the volume of U.S.-sized rebar at issue, there is simply no way — as a practical matter — that ICDAS could have resold that merchandise in ICDAS’ home market. As a practical matter, ICDAS thus had little choice but to fulfill its obligations under the contract, whether or not it was successful in its efforts to negotiate a price increase.

60 . More specifically, ICDAS explains that "scrap is the single primary input in billet, and billet is the single primary input for rebar production.” See ICDAS Brief at 7.

61 . In SRAMs from Taiwan, Commerce determined that the use of quarterly averages for prices and costs resulted in a more accurate comparison than annual averages, in light of the significant decrease in the price of SRAMs throughout the period of investigation. See SRAMs from Taiwan, 63 Fed.Reg. at 8926. Similarly, in DRAMs from Korea, Commerce found that declining production costs and declining prices during the period of investigation in both the U.S. and the surrogate country markets justified the use of monthly weighted averages. See DRAMs from Korea, 58 Fed.Reg. at 15,476.

See also Statement of Administrative Action, H.R. Doc. No. 103-316, at 842-43, reprinted in 1994 U.S.C.C.A.N. at 4177-78 ("Price Averaging”).

62 . ICDAS states that, although the Final Determination in Stainless Steel Coils from Korea was amended in accordance with a WTO panel decision, "the rationale of [Commerce's] original holding remains valid.” See generally ICDAS Brief at 14 n. 8 (citing United States — Anti-Dumping Measures on Stainless Steel Plate in Coils and Stainless Steel Sheet and Strip from Korea, WT/DS 179/R (adopted Feb. 1, 2001); Issues and Decision Memorandum for the Final Determination in the Antidumping Duty Investigation of Live Swine from Canada, 2005 WL 2290627 (March 4, 2005), at comment 5 (finding that the external macroeconomic event that justified the use of two averaging periods in Stainless Steel Coils from Korea — i.e., the precipitous decline in the value of the Korean won— was not present in Live Swine from Canada)); see also Decision Memo at 6 n. 3.

63 . ICDAS asserts that “[t]he rationale underlying the requirement of a single primary input product flows from the significant impact that the rise of the price of the primary input has on the total cost of manufacturing.” ICDAS Brief at 15. As an example, ICDAS points to Canned Pineapple from Thailand, where, on remand, Commerce replaced the POR average cost that it had originally used in its analysis with two separate weighted-average costs, in light of the "almost fifty percent” increase in the cost of fresh pineapple — the single primary input for canned pineapple. See ICDAS Brief at 15 (citing Thai Pineapple Canning Indus. Corp. v. United States, 273 F.3d 1077, 1081-82 & n. 1 (Fed.Cir.2001)).

The Domestic Producers argue that scrap is not ICDAS' single, primary input. See generally Domestic Producers Response Brief at 6-9, 15. The Domestic Producers assert that, in prior determinations, Commerce has found that a given input meets the "single, primary” requirement only where the input accounts for at least a certain relatively high percentage of the total cost of manufacturing. See Domestic Producers Response Brief at 8 (citing Notice of Preliminary Results of Anti-dumping Duty Administrative Review and Intent to Revoke Order: Brass Sheet and Strip From the Netherlands, 64 Fed.Reg. 48,760, 48,762 (Sept. 8, 1999) ("Brass Sheet and Strip From the Netherlands Prelim”)); see also Decision Memo at 8 n. 6 (stating that, in Brass Sheet and Strip from the Netherlands Prelim, Commerce "found that significant and consistent declines in the cost of metal inputs, constituting approximately 70 percent of the total [cost of manufacturing], justified a departure from its normal practice”).

As ICDAS correctly observes, however, Commerce did not rely on the Domestic Producers' argument as a basis for the agency’s decision not to use quarterly average costs in this case. See ICDAS Reply Brief at 3 n. 3. Commerce's determination therefore cannot be sustained on that basis. See Burlington Truck Lines, Inc. v. United States, 371 U.S. 156, 168-69, 83 S.Ct. 239, 9 L.Ed.2d 207 (1962) (ruling that an agency's decision may only "be upheld, if at all, on the same basis articulated ... by the agency itself”).

ICDAS further argues that "the strict percentage test newly minted by [the Domestic Producers] does not exist,” and, moreover, that the difference between the percentage which the Domestic Producers claim is required and the percentage actually proved in this case is "insignificant.” See ICDAS Reply Brief at 3 n. 3.

64 . As evidence that "there is no special requirement linked to currency changes or technological advances" as a precondition to Commerce’s use of monthly or quarterly average costs, ICDAS points to Certain Pasta from Italy and to Brass Sheet and Strip from the Netherlands Final — two cases where no such factors were present. See ICDAS Brief at 16; Issues and Decision Memorandum- for Certain Pasta from Italy, 2000 WL 1880666, at comment 18; Notice of Final Results of Anti-dumping Duty Administrative Review and Determination Not to Revoke the Antidumping Duty Order; Brass Sheet and Strip from the Netherlands, 65 Fed.Reg. 742, 747 (Jan. 6, 2000) ("Brass Sheet and Strip from the Netherlands Final”).

65 . Although Commerce used monthly weighted-average metal costs in Brass Sheet and Strip from the Netherlands Final, it expressly declined to use monthly costs for "fabrication costs.” See Brass Sheet and Strip from the Netherlands Final, 65 Fed.Reg. at 748.

66 . Although Commerce's analysis is somewhat opaque (and the logical underpinnings of that analysis are unclear), the Decision Memo states that Commerce evaluated "the significance of the change in the [cost of manufacturing]” by “first identif[ying] the five highest volume home market control numbers and examining] the impact of using annual average costs of manufacturing versus quarterly average costs of manufacturing.” See Decision Memo at 11-12. According to the Decision Memo, Commerce found that "the difference ranged from approximately five to 10 percent of the [cost of manufacturing].” See id. at 12. The Decision Memo concludes that, “[i]n the past, the Department has not considered one to ten percent increases significant.” See id. (citation omitted).

The Domestic Producers’ discussion of this issue adds little to Commerce’s analysis as set forth in the Decision Memo. See generally Domestic Producers Response Brief at 12. The Domestic Producers assert that, "[w]hile ICDAS argues that there was an absolute increase in scrap costs of 57 percent over the POR, it is clear that this absolute increase did not produce a correspondingly significant increase in the total cost of manufacturing.” See id. Based on Commerce’s explanation in the Decision Memo, the Domestic Producers conclude that "use of a single cost-averaging period does not significantly affect the total cost of manufacturing over the period of review.” See id.

67 . ICDAS further asserts that “the 'consistent' increase in cost is ... confirmed by review of [ICDAS’] quarterly costs. Quarter by quarter during the POR, ... the weighted-average total COM for rebar increased by approximately 11%, 6%, and 30%.” See generally ICDAS Brief at 18 (discussing total cost of manufacturing data, quarter-by-quarter).

68 . Although the Decision Memo reflects Commerce's determination on the "significance” of the change in the cost of scrap, it does not appear that the agency analyzed, much less reached a conclusion as to, the "consistency” of the change in the cost of scrap.

Moreover, it is not clear why Commerce evaluated "significance” here in the way that it did. The methodology used in this case appears to differ from that used in other cases. Indeed, as discussed in greater detail below, ICDAS criticizes Commerce’s methodology here as "statistical manipulation.” See generally ICDAS Brief at 19-20.

69 . ICDAS further asserts that "the 'consistent' increase in cost is ... confirmed by review of [ICDAS’] quarterly costs. Quarter by quarter during the POR, scrap cost increased by approximately 10%, 6%, and 35%.” See generally ICDAS Brief at 18 (discussing scrap cost data, quarter-by-quarter).

Because the greatest increase in scrap cost occurred in the fourth quarter of the POR, ICDAS proposed to Commerce — as an alternative to use of quarterly costs — that the agency use .two weighted-average periods (one for the first three quarters of the POR, and the other for the fourth quarter of the POR). See ICDAS Brief at 19 n. 11 (citing Stainless Steel Coils from Korea Amended, 66 Fed.Reg. at 45,280 (using two cost-averaging periods of differing lengths)); see also Stainless Steel Coils from Korea, 64 Fed.Reg.' at 30,675-76 (discussing same). Although Commerce acknowledged ICDAS' alternative proposal in the Final Results, the agency, never addressed it. See Decision Memo at 7-8 ’(summarizing terms of ICDAS' alternative proposal).

70 . ICDAS distinguishes the cases on which it relies from other cases where Commerce has held that changes were insignificant. See ICDAS Brief at 17 n. 10 (distinguishing, inter alia, Issues and Decision Memorandum for Certain Pasta from Italy, 2000 WL 1880666, at comment 18 (where “the changes in prices from the - beginning to the end of the POR were only 12 and 10 percent”); Fujitsu, 88 F.3d at 1039 n. 4 (“where the difference in COP between the first and last month of the review period was very small — on the order of 1% of the total cost of production”)).

71 . As ICDAS notes, its criticisms would appear to apply with equal force to Commerce’s analysis of the significance of the change in the total cost of manufacturing in this case (discussed earlier), as well as Commerce’s analysis of the significance of the change in the cost of scrap. See ICDAS Brief at 19-20.

72 . Like Commerce, the Domestic Producers also contend that the proponent of the use of multiple cost-averaging periods must demonstrate a direct link between the input raw material costs and directly related sales transactions within the same cost-averaging period. See generally Domestic Producers Response Brief at 6-7, 13-14. As ICDAS notes, however, the Domestic Producers completely fail to address the fact that Commerce has used shorter cost-averaging periods in other cases where no such demonstration was required. See ICDAS Reply Brief at 4. ICDAS further notes that Commerce cited only a single case to support its new requirement. See id.

73 . Elsewhere, ICDAS explains:

[W]hen ICDAS set the prices for sales that were made in the first quarter, ICDAS had no way of knowing that its costs would increase dramatically three quarters later and that such an increase would significantly affect the average cost for the entire year of the POR. Those cost changes could not have been taken into account when ICDAS set sale prices months earlier. On a contemporaneous basis, ICDAS did in fact adjust its prices to take into account the significantly increased costs.... ICDAS could not, however, go back and retroactively adjust prices on sales that had occurred months earlier.

ICDAS Brief at 12.

74 . As the Domestic Producers note, two of the respondents in the underlying administrative review did not contest Commerce’s use of POR average costs in the agency's calculations. See Domestic Producers Response Brief at 16 n. 5. However, like ICDAS, one other respondent — Habas—is challenging the agency on the issue. See Habas Remand Results at 19 (reaffirming, on remand, Commerce’s decision not to depart from agency's normal practice of using the POR average cost), which are being contested by Habas.

75 . To support its opposition to the Government’s request for a voluntary remand, IC-DAS relies on Pittsburgh Logistics, Corns Staal, and Atlantic Sugar. See ICDAS Reply Brief at 1-2 (discussing Corus Staal BV v. United States, 29 CIT 777, 781-83, 387 F.Supp.2d 1291, 1295-97 (2005); Atlantic Sugar, Ltd. v. United States, 1 CIT 211, 211-13, 511 F.Supp. 819, 820-21 (1981)); ICDAS Reply Brief at 14 (discussing Former Employees of Pittsburgh Logistics Systems, Inc. v. U.S. Sec’y of Labor, 27 CIT 1301, 1308-09 (2003)). But the cases that ICDAS cites are not squarely on point, and do not compel denial of the remand requested here.

In Pittsburgh Logistics, for example, the agency had already had five bites at the apple — including one remand. See Pittsburgh Logistics, 27 CIT at 1308-09. In Corns Staal, the court found the agency’s request for a remand to be "both unsupported and unexplained." See Corns Staal, 29 CIT at 781-83, 387 F.Supp.2d at 1295-97. And the Atlantic Sugar court actually did not even deny the request for remand; it simply deferred the request until briefing on the merits of all claims was complete and the case fully submitted for decision. See Atlantic Sugar, 1 CIT at 211-13, 511 F.Supp. at 820-21.

Cf. F.lli De Ceceo Di Filippo Fara S. Martino, S.p.A. v. U.S., 216 F.3d 1027, 1034-35 (Fed.Cir.2000) (rejecting claim that trial court "erred in not allowing Commerce a second chance, after remand, to corroborate the ... petition rate” based on pre-existing data; reasoning, inter alia, that “[t]o do so ... would create a perverse incentive for Commerce not to make adequate efforts to corroborate the rate it selected in the first instance, as the statute requires”); Corus Staal, 27 CIT at 391, 259 F.Supp.2d at 1257 (denying Government’s request for voluntary remand, notwithstanding SKF, where "Commerce's brief [did] not provide any reason, policy or otherwise, for requesting a remand,” except that agency wished to " 'reconsider its decision’ ”; noting that "remand must be on account of appropriate reasons,” that "the agency must state its reasons for requesting remand,” and that "concerns for finality do exist”).

76 . See, e.g., Tr. at 14 (arguing that "the Government must give due regard to [the] finality [of] its decision,” that an agency “cannot simply ask for a do over anytime it wishes,” and that "[h]ere, the Government simply wants a chance to rewrite Commerce's decision”), 23-24 (asserting that Commerce is not entitled to "basically have an opportunity to redo its decision anytime it feels that there are weaknesses in the decision”).

77 . See, e.g., Tr. at 14 (asserting that Commerce's request for remand is, in essence, "an effort to make an end run around the prohibition on post hoc rationalizations”), 16 (characterizing Commerce’s request for remand "an effort to now come up with reasons for a decision that's already been made”), 23 (asserting that reason for requested remand is not that Commerce "want[s] to go back and reconsider the issue,” but — rather—that the agency "want[s] to bolster [its] support” for its existing determination).

78 . In Habas, the companion case to this case, the plaintiff filed a notice of subsequent authority based on Commerce's treatment of the issue of the use of multiple cost-averaging periods in the course of the Ninth Administrative Review of the same underlying antidumping order at issue here, concerning rebar from Turkey. The parties to this case were accorded the opportunity to comment on that subsequent authority. See Letter from ICDAS to Court (Dec. 1, 2008); Defendant’s Response to Notice of Subsequent Authority (Dec. I, 2008); Letter from Domestic Producers to Court (Dec. 1, 2008); see also Letter from Domestic Producers to Court (Feb. 2, 2009) (transmitting United States v. Eurodif S.A., 555 U.S. -, 129 S.Ct. 878, 172 L.Ed.2d 679 (2009), and highlighting Supreme Court’s holding that Commerce's interpretation of the statute "governs in the absence of unambiguous statutory language to the contrary or unreasonable resolution of language that is ambiguous,” and that "[t]his is so even after a change in regulatory treatment”; arguing that Commerce’s recent asserted "refinement of its test for considering whether to use multiple cost averaging periods” is just such a change in regulatory treatment).

In its comments, the Government argued that any recent change in Commerce's methodology is “irrelevant.” See Defendant’s Response to Notice of Subsequent Authority at 4. Remand will give Commerce itself the opportunity to consider the implications for this case — if any — of any asserted recent change in agency methodology, and will allow the agency to consider the views of ICDAS and the Domestic Producers on the matter, if appropriate.

79 . The Domestic Producers bristle at any suggestion that using the date of entry to define the universe of sales constitutes a new methodology. See generally Domestic Producers Response Brief at 4, 34-40. Indeed, the Domestic Producers assert not only that the use of the date of entry is standard agency practice, but also that the statute requires it (at least where the date of entry is known). See Domestic Producers Response Brief at 4, 35-40.

Contrary to the Domestic Producers’ claims, however, Commerce maintains that — • as ICDAS argues — the agency has the discretion to define the universe of sales using "as appropriate, entries, exports, or sales.” See Decision Memo at 22 (discussing 19 U.S.C. § 1675(a)(2)(A) and 19 C.F.R. § 351.213(e)); ICDAS Reply Brief at 13 n. 10. As Commerce explained in promulgating its regulation on the issue, "neither the [statute] nor the AD Agreement specifies whether sales or entries are to be reviewed.” See Antidumping Duties; Countervailing Duties: Final Rule, 62 Fed.Reg. at 27,314 (Preamble).

ICDAS further argues that defining the universe of sales using the date of sale (rather than the date of entry) has been sustained by the courts. See ICDAS Reply Brief at 13 n. 10 (citing FAG Kugelfischer Georg Schafer KGaA v. United States, 19 CIT 1177, 1180-81, aff'd, 86 F.3d 1179, 1996 WL 267277 (Fed. Cir.1996)). In its Final Results, however, Commerce sought to distinguish FAG Kugelfischer, on which ICDAS relies. See Decision Memo at 24 (stating, inter alia, that "FAG Kugelfischer addressed the appropriateness of using sales versus entry data in the context of a sampling situation, unlike the use of actual sales and entry data here”); see also Domestic Producers Response Brief at 39. Moreover, the Final Results stated that — although Commerce retains the discretion to define the universe of sales using “entries, exports, or sales” — the agency does not believe that those three bases are "equally preferable,” and that Commerce’s usual practice is to use date of entry. See Decision Memo at 22; see also id. at 24 (explaining that "it remains within [Commerce’s] discretion, and is, in fact, [Commerce's] preference and practice to restrict the universe to entries when the facts permit”).

80 . ICDAS relies on Shikoku for the proposition that Commerce is not permitted to make "minor but disruptive changes in methodology where a respondent demonstrates its specific reliance on the old methodology used in multiple preceding reviews.” ICDAS Brief at 38 (<siting Shikoku, 16 CIT 382, 795 F.Supp. 417). Specifically, the Shikoku court found that Commerce "[had] abused its discretion in adopting a slightly improved allocation methodology in the face of years of acceptance of the prior approach.” Shikoku, 16 CIT at 387, 795 F.Supp. at 420-21. The court noted that Commerce had failed to identify any new or important facts that would justify the new methodology, and had failed to explain how the new methodology "would reveal significant and heretofore undiscovered dumping.” Shikoku, 16 CIT at 387, 795 F.Supp. at 421. Although the new methodology at issue there would have been more accurate, its use would have resulted in a dumping margin slightly above de minimis, and thus would have denied the respondent the opportunity to have the antidumping order against it revoked after years of having zero or de minimis dumping margins. The court therefore held that Commerce’s obligation "to administer the anti-dumping laws fairly” precluded the agency "from changing its methodology at this late stage.” Shikoku, 16 CIT at 388, 795 F.Supp. at 421. The court emphasized that "[a]t some point, Commerce must be bound by its prior actions so that parties have a chance to purge themselves of antidumping liabilities.” Shikoku, 16 CIT at 387, 795 F.Supp. at 421. See generally ICDAS Brief at 38-40 (discussing Shikoku); ICDAS Reply Brief at 15 (same). But see Domestic Producers Response Brief at 36-37, 39-40 (seeking to distinguish Shikoku from the case at bar).

81 . As a threshold matter, the Domestic Producers highlight the fact that ICDAS did not advance its reliance argument at the agency level. The Domestic Producers therefore contend that the doctrine of exhaustion of administrative remedies bars ICDAS from making the argument here. See generally Domestic Producers Response Brief at 34-36 & n. 13. ICDAS maintains that, because Commerce used the date of sale to define ICDAS' universe of sales in the Preliminary Results, "IC-DAS could only challenge the use of entry date to define the universe of sales after Commerce published the Final Results.” See IC-DAS Reply Brief at 13 n. 10. It is far from clear whether ICDAS reasonably could have been expected to make its reliance argument in the rebuttal brief that it filed with Commerce. But, in any event, the Domestic Producers did not press their exhaustion claim at oral argument. See Corus Staal, 502 F.3d at 1381 (and cases cited there) (noting that "applying exhaustion principles in trade cases is subject to the discretion of the judge of the Court of International Trade”).

82 . See, e.g., Tr. at 14 (arguing that Commerce “cannot simply ask for a do over anytime it wishes,” and surmising that agency's request for voluntary remand is "an effort to make an end run around the prohibition on post hoc rationalizations”).