Memorandum Opinion and Order
Goldberg, Judge:This matter is before the court for review on plaintiffs appeal from the final antidumping duty determination of the United States Department of Commerce, International Trade Administration (“ITA”), regarding circular welded non-alloy steel pipe from the Republic of Korea. Circular Welded Non-Alloy Steel Pipe From the Republic of Korea, 57 Fed. Reg. 42,942 (Sept. 17, 1992) (“Final Determination”). Defendant-intervenors1 also argue that certain aspects of the ITA’s Final Determination are unsupported by substantial evidence and not in accordance with law. Together, plaintiff and Defendant-inter-venors raise five primary issues for the court’s review. The court exercises its jurisdiction pursuant to 28 U.S.C. § 1581(c) (1988).
Background
On October 21,1991, in response to apetition filed by several U.S. producers including Laclede Steel Company (“Laclede”), the ITA published a notice of initiation of an antidumping investigation of circular welded non-alloy steel pipe from Korea. Circular Welded Non-Alloy Steel Pipe From Brazil, the Republic of Korea, Mexico, Romania, Taiwan, and Venezuela, 56 Fed. Reg. 52,528 (Oct. 21, 1991). The period investigated was from April 1 through September 30, 1991. 57 Fed. Reg. at 42,943. The ITA conducted verification of Defendant-intervenors’ sales responses in May and June 1992, and of Defendant-intervenors’ cost responses in *966June and July 1992. Id. at 42,942-43. The ITA published its Final Determination on September 17,1992. Id. at 42,942. The ITA subsequently amended its final determination to correct a ministerial error. Certain Circular Welded Non-Alloy Steel Pipe from Brazil, the Republic of Korea (Korea), Mexico, and Venezuela, 57 Fed. Reg. 49, 453 (Nov. 2, 1992).
Discussion
In this review of the ITA’s amended final determination, the court is charged to hold unlawful any determination, finding, or conclusion that is unsupported by substantial evidence on the record, or otherwise not in accordance with law. 19U.S.C. § 1516a(b)(l)(B) (1988). “Substantial evidence is something more than a ‘mere scintilla,’ and must be enough reasonably to support a conclusion.” Ceramica Regiomontana, S.A. v. United States, 10 CIT 399, 405, 636 F. Supp. 961, 966 (1986) (citations omitted) aff’d, 5 Fed. Cir. (T) 77, 810 F.2d 1137 (1987). Furthermore, the “ITA’s interpretation of the statute it administers must be reasonable and must not conflict with Congressional intent.” Floral Trade Council v. United States, 15 CIT 497, 498, 775 F. Supp. 1492, 1495 (1991) (citations omitted).
The following five issues are presented in this appeal: (1) Whether the ITA’s determination that home market sales of alleged “overruns” were actually within the ordinary course of trade (and thus should be included within the margin calculations), is supported by substantial evidence in the administrative record and in accordance with law; (2) Whether the ITA’s reliance on Defendant-Intervenors’ reported weight data for purposes of calculating cost of production (“COP”) is supported by substantial record evidence and in accordance with law; (3) Whether the ITA’s determination that Defendant-Intervenors were entitled to duty drawback adjustments is supported by substantial record evidence and in accordance with law; (4) Whether the ITA’s decision to use revalued depreciation expense data for Hyundai, in accordance with Korean generally-accepted accounting principles (“GAAP”), is supported by substantial evidence and in accordance with law; (5) Whether the ITA’s decision to proceed with a separate levels-of-trade analysis is supported by substantial evidence and in accordance with law. For the following reasons, the court sustains the ITA’s amended final determination in part, and remands in part.
A. Alleged “Overrun” Sales:
1. Ordinary Course of Trade:
The first issue presented is whether the ITA’s determination that certain alleged overrun Production of pipe was sold in the ordinary course of trade is supported by substantial evidence in the administrative record and in accordance with law. The ITA determined that Pusan and Hyundai failed to prove that their alleged overruns were sold outside the ordinary course of trade.
The ITA compares the prices of subject merchandise sold in the United States (“USP”) with the foreign market value (“FMV”) of such *967or similar merchandise. The pertinent statutory provision defines FMV as:
[T]he price * * * at which such or similar merchandise is sold, or, in the absence of sales, offered for sale in the principal markets of the country from which exported, in the usual commercial quantities and in the ordinary course of trade for home consumption * * *.
19 U.S.C. § 1677b(a)(l)(A) (1988) (emphasis added). The term “ordinary course of trade” is defined as:
[T]he conditions and practices which, for a reasonable time prior to the exportation of the merchandise which is the subject of an investigation, have been normal in the trade under consideration with respect to merchandise of the same class or kind.
19 U.S.C. § 1677(15) (1988). Defendant-Intervenors bear the burden of proving whether sales used in the ITA’s calculations are outside the ordinary course of trade. Nachi-Fujikoshi Corp. v. United States, 16 CIT 606, 608, 798 F. Supp. 716, 718 (1992); see also Koyo Seiko Co. v. United States, 16 CIT 539, 543, 796 F. Supp. 1526, 1530 (1992). Home market sales of overruns which are not sold in the ordinary course of trade are excluded from the ITA’s calculation of FMV See Mantex, Inc. v. United States, 17 CIT 1385, 841 F. Supp. 1290, 1305-09 (1993). Overrun sales which are, however, made in the ordinary course of trade, are included in the pool of products eligible to be matched to products sold in the United States.
The court reviews the ITA’s ordinary course of trade determination on an individual basis, taking into account all of the relevant facts and circumstances particular to the sales in question. Mantex, 841 F. Supp. at 1306 (citation omitted). Among the circumstances which the ITA normally examines in order to determine whether sales of alleged overruns were made in the ordinary course of trade are: (1) Whether the sales in question did, in fact, consist of production overruns or seconds; (2) The comparative volume of sales and number of buyers for such overrun production in the home market; (3) The differences in product standards and uses between overruns and ordinary production; (4) The price and profit differentials between alleged overrun sales and ordinary sales in the home market. Mantex, 841 F. Supp. at 1295.
Defendant-Intervenors allege that certain sales of ASTM and KSD pipe to Korean home market customers were overrun sales resulting from overproduction of pipe needed to fill particular orders. Defendant-Intervenors explain that because they produce pipe to order, any resulting overruns are more difficult to market. Moreover, because only small quantities of overruns result from producing to order, and because overruns are sold at a single metric ton price regardless of their size, pipe overruns are sold at lower prices than merchandise produced to order. Administrative Record Public Document (“AR Pub. Doc. ”) 140, frame 399; AR Pub. Doc. 273, frames 1804-05. Defendant-Intervenors therefore argue that their sales of overrun pipe were not made in the ordinary *968course of trade and should not have been included in the ITA’s FMV calculations.
The ITA states that Hyundai failed to present “any evidence whatsoever,” and that Pusan presented “inadequate information,” to support their claims that the identified sales in question were, in fact, overruns. ITABrief at 15. The ITA argues that many of the sales in question were not overruns because the sales involved pipe built to Korean specification, which would almost always have a demand in the Korean home market. Id. at 16. Furthermore, as Defendant-Intervenors concede, there is a ready market for ASTM pipe in Korea. Memorandum of Points and Authorities in Support of Motion by Defendant Intervenors Hyundai Pipe Co., Ltd. et al., For Judgment Upon the Agency Record (“Defendant-Intervenors’ Brief”) at 47. The ITA also argues that the record shows that average sales quantities of allegedly overrun pipe were similar to average sales quantities of pipe sold in the ordinary course of trade. ITA Brief at 18. Indeed, in some instances, the volume of alleged overrun sales was larger than the volume of regular commercial sales of the same model. Id. Finally, with respect to price and profit differentials, the ITA rejected Hyundai’s claim that its ASTM overrun sales were consistently priced below the average price of regular commercial sales, finding it to be unsubstantiated. 57 Fed. Reg. at 42,948; ITA Brief at 21.
It does not appear that the ITA fully considered all relevant facts and circumstances particular to the sales in question in making its ordinary course of trade determination. Specifically, the court finds that the ITA’s failure to adequately address: the number of home market customers buying alleged overruns; the product standards and uses of alleged overruns; and, the prices and profits on alleged overrun sales, necessitates a remand for reconsideration of these factors. First, in examining the volume of alleged overrun sales, the ITA failed to address whether the alleged overruns were routinely sold to a regular buyer or group of buyers in order to test Defendant-Intervenors’ claim that alleged overruns were difficult to market. Second, the ITA failed to address whether, despite the alleged overruns’ conformity with Korean or ASTM standards, buyers were in fact utilizing the alleged overruns for different purposes than merchandise sold in the ordinary course of trade. Third, the court finds that the record does not reflect that the ITA examined all relevant information submitted by Defendant-Interve-nors concerning the pricing of alleged overruns versus products sold in the ordinary course of trade. In particular, the ITA failed to adequately address the evidence which supports Pusan’s claim that its alleged overruns were sold at significantly lower prices than sales of the same model sold in the ordinary course of trade.2 Additionally, in concluding that Hyundai’s claim (that alleged overruns were consistently priced below *969products sold in the ordinary course of trade) was “unsubstantiated,” the ITA failed to adequately address evidence to the contrary.3
The court therefore finds that the ITA’s determination that Defendant-Intervenors failed to prove that any of the alleged overruns were sold outside the ordinary course of trade is not supported by substantial record evidence. On remand, the court directs the ITA to review the data on alleged overrun sales and consider all relevant facts and circumstances contained in the administrative record. Once the ITA has determined which alleged overruns were sold in the ordinary course of trade, if any, it must fully articulate the basis for its determination. All sales found not to have been made in the ordinary course of trade must be excluded from the ITA’s calculations.
2. Treatment of Identical and Similar Matches of Alleged Overruns:
A subsidiary issue presented is how the ITA should account in its dumping margin calculations for those alleged overruns which are found to have been sold in the ordinary course of trade. Although the ITA determined that alleged overruns should be included in the calculation of FMVJ the ITA inadvertently failed to include them in its calculations for the Final Determination. The ITA subsequently amended its calculations to include alleged overruns that were identical matches to products sold in the United States.4 57 Fed. Reg. at 49,453. The ITA determined that incorporating identical matches of alleged overruns into its calculations was a ministerial error capable of being corrected pursuant to 19 U.S.C. § 1673d(e) and 19 C.F.R. § 353.28(d). In its amended margin calculations, therefore, the ITA accounted for both regular and alleged overrun products sold in the Korean market that were identical to products sold in the United States. The ITA further concluded, however, that modification of its calculations to account for overruns which were similar matches to United States sales would require more than the correction of a ministerial error. The ITA therefore seeks a court order of remand permitting it to render a decision regarding how to account for alleged overrun sales that may have constituted the “most similar” match in the home market. ITA Brief at 23-27.
The court agrees that a remand is necessary in order to permit the ITA to address the omission of similar matches from its calculations. On remand, the court directs the ITA to account for alleged overrun sales that it finds were sold in the ordinary course of trade and which constitute “most similar” matches in the home market. The ITA shall clearly articulate the matching methodology it employs, and the results drawn therefrom.
*970B. Theoretical versus Actual Weight:
The second issue before the court is whether the ITA’s reliance on Pusan’s and KSP’s reported weight data is supported by substantial record evidence and in accordance with law. Laclede argues that the steel coil input weights and pipe output weights used to calculate COP and constructed value (“CV”) are inaccurate, and thus should have been adjusted or rejected in favor of best information available (“BIA”). See Plaintiff’s Memorandum of Law in Support of Its Motion For Judgment on the Agency Record (“Laclede Brief”) at 15-25. Laclede claims that the industry formula used to convert the standard actual weight of the input coil to the theoretical weight of the finished pipe fails to accurately account for pipe-wall buildup during the production process. Therefore, Laclede argues, the resulting costs (i.e. CV and COP) that are based on these weights are necessarily inaccurate.
The ITA determined that respondents’ reported weight data is reasonably accurate, and verified that the data did not significantly distort respondents’ actual costs. The ITA made numerous traces to source documents, verified calculations, and obtained extensive documentation. See ITA Brief at 34-36 (providing numerous citations to the record). The ITA did acknowledge that “use of the nominal thickness of the coil to calculate the weight of the pipe may under- or over-state the actual weight of the pipe. As such, this calculation may have an effect on cost calculations.” 57 Fed. Reg. at 42,945. Nevertheless, the ITA rejected Laclede’s argument that Defendant-Intervenors’ weights were inherently understated.
In Avesta Sheffield, Inc. v. United States, 17 CIT 1212, 838 F. Supp. 608 (1993), the court similarly faced the issue of whether the ITA may calculate fabrication costs based on a standard industry formula rather than the true actual weight of the finished pipe. The Avesta Sheffield court upheld the ITA’s use of respondents’ data, noting that: (1) the conversion formula used by respondents to calculate their reported fabrication costs was standard in the industry; (2) respondents’ reported data was consistent with the data in respondents’ financial records; and (3) petitioner was unable to demonstrate any specific error resulting from such a conversion methodology. Avesta Sheffield, 838 F. Supp. at 612-13.
In the present case, the court finds that the ITA’s use of Defendant-Intervenors’ data and the industry conversion formula is supported by substantial record evidence and in accordance with law. The ITA verified that Defendant-Intervenors’ cost information was based on information maintained in their normal accounting records, raw material files, inventory records, and production records. Defendant-Inter-venors provided detailed examples taken from actual sales records, and explained how they arrived at the conversion factors and how they calculated the theoretical weight. ITA Brief at 34. The ITA traced each example to source documentation and discovered no significant prob*971lems regarding the accuracy of the weight reported.5 The court is unpersuaded by Laclede’s arguments that Defendant-Intervenors’ weight data is inherently unreliable. Although Defendant-Intervenors’ costs may sometimes have been under- or over-reported, the court finds no record evidence supporting Laclede’s claim that these costs are always underreported. Indeed, the ITA verified that Defendant-Intervenors’ approach did not significantly distort Defendant-Intervenors’ actual costs. Laclede has failed to demonstrate any error in this regard. For the foregoing reasons, the court holds that the ITA’s decision to accept the weight data submitted by Pusan and KSP is supported by substantial record evidence and is otherwise in accordance with law, and, therefore, is sustained.
C. Duty Drawback/Import Duties:
The third issue before the court is whether the ITA’s adjustment of Defendant-Intervenors’ data to account for import duties on raw materials was proper. Defendant-Intervenors argue that they were entitled to an adjustment for duty drawback on pipe exported to the United States because the Korean government had refunded them import duties paid for raw materials used to produce the exported pipe. The ITA agreed that Defendant-Intervenors’ claims for duty drawback were valid. Normally, such a finding results in an upward adjustment to USE In the present case, however, the ITA chose to grant an upward adjustment except in those instances where USP was matched to constructed value. The ITA declined to adjust for duty drawback any United States sale matched to CV, as a BIA penalty for Defendant-Intervenors’ failure to report duty-exclusive CV data.
The court will first examine whether Defendant-Intervenors substantiated their claims for duty drawback adjustments. The court will then examine whether Defendant-Intervenors’ failure to separate import duties from their CV data justified the ITA’s refusal to adjust for duty drawback any United States sale which was compared to CV
1. Duty Drawback Adjustment:
An upward adjustment to United States price (“USP”) for duty drawback is provided for by statute in 19 U.S.C. § 1677a(d)(1)(B) (1988); see also 19 C.F.R. § 353.41(d)(1)(ii) (1992). The ITA utilizes a two prong test to determine whether a party is entitled to a duty drawback adjustment:
1. The import duty and rebate must be directly linked to, and dependent upon, one another.
2. The company claiming the adjustment must demonstrate that there were sufficient imports of imported raw materials to account for the duty drawback received on exports of the manufactured product.
*972See, e.g., Avesta Sheffield, 838 F. Supp. at 611. This court has consistently upheld the ITA’s two-prong test. Id.; see Far East Mach. Co. v. United States, 12 CIT 972, 979, 699 F. Supp. 309, 315 (1988) (“Far East Mach. II”); Far East Mach. Co. v. United States, 12 CIT 428, 431, 688 F. Supp. 610, 612 (1988) (“Far East Mach. I”); Carlisle Tire & Rubber Co. v. United States, 11 CIT 168, 171, 657 F. Supp. 1287, 1289-90 (1987); see also U.S. Dep’t of Commerce, Int’l Trade Admin., Study of Antidumping Adjustments Methodology and Recommendations for Statutory Change 26-27 (Nov. 1985). The first prong analyzes the foreign government’s export rebate program, while the second prong tests the application of the program to the particular respondent claiming the adjustment. Far East Mach. II, 12 CIT at 975 n.4, 699 F. Supp. at 312 n.4. A duty drawback adjustment to USP prevents “dumping margins from arising [simply] because the exporting country rebates import duties * * * for raw materials used in exported merchandise * * *.” Carlisle Tire & Rubber Co. v. United States, 10 CIT 301, 307, 634 F. Supp. 419, 424 (1986).
It is undisputed that Defendant-Intervenors satisfied the first prong of the test. See Final Determination, 57 Fed. Reg. at 42,946. Laclede, however, contests the ITA’s conclusion that Defendant-Intervenors satisfied the second prong of the duty drawback test. Laclede Brief at 46-51. Specifically, Laclede objects to the ITA’s decision to accept Defendant-Intervenors’ average duty drawback data, instead of requiring transaction-specific data. Id. Defendant-Intervenors reported average duty drawback in this case because they do not maintain such bookkeeping records on a sales-specific basis. ITA Brief at 55. Thus, to have collected and submitted transaction-specific duty drawback data, Defendant-Intervenors would have had to manually trace thousands of documents, some of which covered multiple invoices. See AR Pub. Doc. 221, frames 2200, 2201. In rendering its Final Determination, the ITA noted that the second prong of its test encompasses the principle of drawback substitution. Based upon extensive record evidence addressing duty drawback in this case, the ITA determined that there were sufficient imports of raw materials to account for the duty drawback receivedby Defendant-Intervenors. 57 Fed. Reg. at 42,946; see ITABrief at 53-55.
This court has consistently held that “there is no requirement that [a] specific input be traced from importation through exportation before allowing drawback on duties paid * * Far East Mach. II, 12 CIT at 975,699 F. Supp. at 312; see also Chang Tieh Indus. Co. v. United States, 17 CIT 1314, 840 F. Supp. 141, 147 (1993); Avesta Sheffield, 838 F. Supp. at 612. Indeed, this court has previously upheld the ITA’s use of average duty drawback rather than transaction-specific duty drawback as being in accordance with law. See Carlisle Tire, 11 CIT at 172, 657 F. Supp. at 1290. Pursuant to the second prong, the only limit on the allowance for duty drawback is that the adjustment to USP may not exceed the amount of import duty actually paid. Far East Mach. II, 12 CIT at 974-75, 699 F. Supp. at 311-12. The court acknowledges that adjusting *973USP on the basis of average duty drawback may result in some sales receiving more or less of an adjustment than was rebated, or even a rebate where none was received. See Carlisle Tire, 11 CIT at 172, 657 F. Supp. at 1290. However, the sales-specific tracing requirement proposed by Laclede would eviscerate established substitution principles utilized by the ITA in situations where it would be difficult, if not impossible, to determine whether the raw materials used in producing the exported merchandise actually came from imported rather than domestic sources. Avesta Sheffield 838 F. Supp. at 612 (quoting Far East Mach. I, 12 CIT at 431, 688 F. Supp. at 612). In this case, the ITA verified that Defendant-Intervenors imported sufficient raw materials to account for the duty drawback received on exports of pipe. Final Determination, 57 Fed. Reg. at 42,946. After examining the extensive record on this issue, the court concludes that the ITA’s decision to accept average duty drawback data is supported by substantial evidence.6 For the foregoing reasons, the ITA’s determination that respondents submitted sufficient evidence to allow a duty drawback adjustment to USP is supported by substantial record evidence and in accordance with law, and, therefore, is sustained.
2. Denial of Duty Drawback Adjustment for Failure to Report CV Exclusive of Import Duties:
The next issue presented is whether the ITA’s decision to resort to BIA and deny any duty drawback adjustment to USP when compared to CV is in accordance with law and supported by substantial evidence in the record. The ITA justified its resort to BIA by explaining that Defendant-Intervenors failed to comply with the ITA’s request that Defendant-Intervenors report cost of materials (“COM”) data exclusive of import duties. See Final Determination, 57 Fed. Reg. at 42,951-52. COM is one component of constructed value. The court finds that because the ITA’s stated justification is without statutory basis, and because it amounts to a departure from longstanding practice for which the ITA has failed to provide an adequate explanation, the ITA’s application of BIA cannot be sustained.
First, 19 U.S.C. § 1677b(e) provides that CV shall include the sum of the cost of materials used in manufacturing such or similar merchandise. This provision explicitly excludes only internal taxes from the cost of materials.7 This court has previously recognized that the term “internal tax” denotes taxes other than import duties. See Serampore Indus. Pvt. Ltd. v. United States Dep’t of Commerce. Int’l Trade Admin., 11 CIT 866, 869, 675 F. Supp. 1354, 1357 (1987). The ITA has failed to offer any *974statutory basis for equating import duties with internal taxes for purposes of requiring the exclusion of import duties from reported CV data.
Not only does statute not require the exclusion of import duties from CV the court also notes that the ITA’s longstanding practice has been to require respondents to include import duties in constructed value. See, e.g., Offshore Platform Jackets and Piles From the Republic of Korea, 51 Fed. Reg. 11,795,11,796 (Apr. 7,1986) (ITA’s usual practice is to include import duties which would have been waived or rebated upon exportation in CV because such duties are added to USP). Indeed, in this case the ITA initially issued a standard cost questionnaire instructing Defendant-Intervenors that COP and CV data should include import duties. AR Pub. Doc. 139, frame 353. Subsequently, however, the ITA issued a supplemental questionnaire in which it changed its reporting format and required Defendant-Intervenors to exclude import duties from the overall cost of production for exported products. AR Pub. Doc. 249, frame 476. The ITA has failed to provide an adequate rationale for this change in its reporting requirement. While the ITA is afforded great discretion to evaluate and address the facts and circumstances particular to each case in reaching its determinations, this court will not sustain a new reporting requirement absent a rational explanation for the ITA’s departure from an established prior practice. See Citrosuco Paulista, S.A. v. United States, 12 CIT 1196, 1209, 704 F. Supp. 1075, 1087-88 (1988) (citations omitted).
The ITA’s request for duty-exclusive COM data in this case appears to be an unfounded attempt to add a new hurdle to the drawback test that is not required by statute. Accord Chang Tieh, 840 F. Supp. at 147. The ITA’s determination that respondents fully satisfied the requirements for duty drawback adjustment is in accordance with law and supported by substantial record evidence, and is dispositive of this issue given the ITA’s failure to articulate a rationale explanation for its change in methodology. Because the ITA has failed to adequately justify its decision to link Defendant-Intervenors’ eligibility for duty drawback adjustments to the reporting of CV data, the court remands this issue and directs the ITA to render a determination granting adjustments for duty drawback on all U.S. sales, including those compared to constructed value.
D. Korean GAAP and Depreciation Expenses:
The fourth issue before the court is whether the ITA properly relied upon the revalued depreciation expenses reported in Hyundai’s financial statements. Hyundai argues that the ITA should have valued depreciation expenses on a historical basis rather than a revalued basis, because the revalued depreciation expenses are distortive. Defendant-Intervenors’ Brief at 71-72. Hyundai also argues that valuing its depreciation on a historical basis is consistent with United States and Korean GAAR and with the ITA’s past practice. Id. at 55-56,65-70. The ITA responds to Hyundai’s argument by explaining that calculation of Hyundai’s depreciation on a revalued basis is consistent with Hyundai’s *975financial records, Korean GAAI] and with the ITA’s practice of applying foreign GAAP principles except where those principles are distortive. ITA Brief at 58-70.
The legislative history of the COP statute states that “in determining whether merchandise has been sold at less than cost, [the ITA] will employ accounting principles generally accepted in the home market of the country of exportation if [the ITA] is satisfied that such principles reasonably reflect the variable and fixed costs of producing the merchandise. H.R. Rep. No. 571,93d Cong., 1st Sess. 71 (1973) (emphasis added). This court has upheld the ITA’s use of a firm’s expenses as they are recorded in the firm’s financial statements, as long as those statements are prepared in accordance with the home country’s GAAP and do not significantly distort the firm’s financial position or actual costs. Asociacion Colombiana de Exportadores de Flores v. United States, 13 CIT 526, 533 n.12, 717 F. Supp. 834, 841 n.12 (1989); see Hercules, Inc. v. United States, 11 CIT 710, 754-56, 673. F. Supp. 454, 490-91 (1987).
Korean GAAP normally requires Korean companies to calculate depreciation expenses based on historical costs. During the POI, however, Korean GAAP permitted a company to revalue its assets pursuant to Article 56-2 of the Tax Exemption and Reduction Control Act.8 The Korean government permitted companies to revalue their assets in order to encourage companies like Hyundai to offer their shares publicly. Defendant-Intervenors’ Brief at 56; Defendant-Intervenors’ Reply Brief at 76. In 1989, in anticipation of its first public stock offering, Hyundai elected to revalue its fixed assets to current market value and record a “revaluation surplus” above and beyond its actual historical cost. Defendant-Intervenors’ Reply Briefat 76-77; see AR Pub. Doc. 268, frame 1729. Existing fixed assets were marked up from their then-depreciated value to current market value; subsequent depreciation on fixed assets would thenceforth be on a revalued basis. See ITA Brief at 64-65. Thus, calculation of Hyundai’s depreciation expenses on either a revalued basis, i.e. the ITA’s methodology, or on a historical basis, i.e. Hyundai’s preferred methodology, appears consistent with Korean GAAP
The court finds that respondents have failed to demonstrate that the ITA’s decision to use Hyundai’s revalued depreciation expenses is dis-tortive. The verified revalued depreciation expenses were consistent with Korean GAAP and were based on information obtained directly from Hyundai’s financial statements. See AR Pub. Doc. 283, frame 2328; AR Non-Pub. Doc. 129, frame 1252. In addition, use of Hyundai’s reported depreciation expenses at historical value would be distortive because such a methodology would overlook the significant impact that revaluing assets has had on Hyundai. The ripple effects caused by reval*976uation of Hyundai’s assets include, inter alia, a decrease in tax liabilities due to increased amounts of depreciation; an increase in equity reflected on the company’s balance sheets; a potentially enhanced stock value resulting from more available equity; and, an improved ability to acquire debt resulting from an increase in equity. See ITA Brief at 65-66. Thus, while Hyundai’s depreciation expenses increased as a result of the asset revaluation, other costs, notably financing costs, decreased. Use of depreciation expenses reported on a historical basis rather than on a revalued basis would thus result in a skewed portrayal of cost of production; by isolating one of many variables affecting a company’s entire financial picture over a period of time, i.e. depreciation expenses in this case, and subjecting that variable to accounting principles different from those applied to other variables such as financing costs, Hyundai seeks to reap the benefits of revaluation with respect to additional available liquidity, a lower tax liability, etc., and yet turn back the clock to take advantage of diminished depreciation expenses solely for purposes of this antidumping investigation.
Hyundai erroneously argues that the ITA departs from basing depreciation expenses on historical costs only when confronted with companies in hyper-inflationary economies. See Defendant-Interve-nors’ Brief at 68-69. The ITA’s cost calculations turn not on the type of economy in which the company is situated, but on whether reported costs are reasonably representative of the costs faced by the company. In hyper-inflationary economies, the rapid escalation of prices means that assets valued at historical costs are necessarily understated, often to a substantial degree. In those cases, the ITA does not base depreciation expenses on distortive historical costs, but on a methodology which provides a more accurate measurement of the costs facing that company during the POI. Similarly, in this investigation, rather than base Hyundai’s depreciation expenses on distortive historical costs, the ITA elected to use revalued depreciation expense data because such data more accurately measured Hyundai’s costs during the POI. Furthermore, in this case Hyundai voluntarily chose to revalue its depreciation expenses in accordance with Korean GAAP; thus, the ITA used depreciation expense data which was drawn directly from Hyundai’s financial records. For the foregoing reasons, the court finds that the ITA’s decision to employ revalued depreciation expense data for Hyundai is supported by substantial record evidence and in accordance with law. The ITA’s determination is therefore sustained in this regard.
E. Commercial Levels of Trade:
The final issue before the court is whether the ITA’s decision to compare products according to level of trade is supported by substantial evidence and in accordance with law. The applicable regulation states that calculations of FMV and USP will normally be based on “sales at *977the same commercial level of trade.” 19 C.F.R.§ 353.58.9 The ITA determines whether different levels of trade exist by examining the type and function (i.e. economic function) of the first unrelated buyers in the chain of commerce. Ad Hoc Committee of AZ-NM-TX-FL Producers v. United States, 16 CIT 1008, 1010, 808 F. Supp. 841, 843 (1992) (“Ad Hoc 11”), rev ’don other grounds, 12 Fed. Cir. (T)_, 13 F.3d 398 (1994). The identification of different levels of trade establishes a rebuttable economic presumption that levels of trade have an impact upon price and, ultimately, upon FMV Ad Hoc II, 16 CIT at 1010-11, 808 F. Supp. at 844. Once discernable levels of trade have been identified, challenging parties bear the burden of rebuttingthe presumption that prices and selling expenses are correlated to levels of trade.10 If a party provides sufficient evidence to rebut this presumption, the ITA then applies a correlation test to determine whether it should proceed with an analysis based upon the previously identified functional levels of trade.
As an initial matter, the court finds that the ITA’s identification of two levels of trade, i.e. end-users and distributors in the Korean market, is supported by substantial evidence in the record. Defendant-Intervenors argue that there is no discernable correlation between prices and the identified levels of trade. In support, Defendant-Intervenors point to record evidence which demonstrates that if the Korean market is analyzed in the aggregate with respect to sales of merchandise in the relevant class or kind, there are no discernible differences between prices and selling expenses on the one hand, and levels of trade on the other. The government acknowledges this record evidence, but argues that because Defendant-Intervenors rely upon an aggregate analysis (i.e. one which includes both subject and non-subject merchandise), they have failed to rebut the presumption that there is a correlation between price and levels of trade. ITA Brief at 74-75; see Defendant-Intervenors’ Brief at 79-80. The court disagrees.
The court finds that the evidence introduced by Defendant-Interve-nors in this case was sufficient to rebut the presumption of a correlation between prices and levels of trade. Once the presumption of a correlation is rebutted by sufficient evidence, the burden shifts from the challenging party to the government; it is then incumbent upon the ITA to *978conduct its own correlation test, and to base its final determination upon the results of that test. Ad Hoc II, 16 CIT at 10-11, 808 F. Supp. at 844. The court therefore remands this issue to the ITA; upon remand, the ITA is instructed to conduct its own correlation test, utilizing only the price factor in this case, to determine whether there is in fact a correlation between prices and levels of trade for the subject merchandise. Based upon the results of that analysis, the ITA shall amend its determination accordingly.
Conclusion
In conclusion, the court sustains the ITA’s amended final determination with respect to two issues, i.e. the ITA’s reliance upon Defendant-Intervenors’ weight data; and, the ITA’s reliance upon the revalued depreciation expenses reported in Hyundai’s financial statements. With respect to the three remaining issues, however, the court directs the ITA upon remand to reconsider its determination in accordance with the court’s opinion.
First, with regard to the alleged overruns, the ITA shall review the alleged overrun sales to determine whether they were made in the ordinary course of trade, taking into account all of the relevant facts and circumstances particular to the sales in question. The ITA is specifically directed to address: the number of home market customers buying alleged overruns; the product standards and uses of alleged overruns; and, price and profit differentials between alleged overruns and sales made in the ordinary course of trade. The ITA shall fully articulate the bases for the conclusions it reaches. Only those identically matched alleged overruns which are found to háve been sold in the ordinary course of trade are to be included in the ITA’s calculations. The ITA is further directed to account for alleged overrun sales that it finds were sold in the ordinary course of trade and which constitute “most similar” matches in the home market. The ITA shall provide a full explanation of the matching methodology it employs and the results drawn therefrom.
Second, with regard to duty drawback, the court directs the ITA to render a determination granting adjustments for duty drawback on all U.S. sales, including those compared to constructed value.
Third, the court instructs the ITA to conduct a correlation test, utilizing only the price factor, to determine whether there is a correlation between price and levels of trade for the subject merchandise. The ITA shall fully articulate its methodology and, based upon the results of the test, shall amend its determination accordingly. For the foregoing reasons, it is hereby:
Ordered that Plaintiffs Motion For Judgment On The Agency Record is granted in part and denied in part; it is further
Ordered that Defendant-Intervenors’ Motion For Judgment Upon The Agency Record is granted in part and denied in part; it is further
Ordered that the U.S. Department of Commerce, International Trade Administration, shall, within sixty (60) days of the date of this *979Memorandum Opinion and Order, issue a remand determination in accordance with the instructions provided herein; it is further
Ordered that the parties may, within twenty-five (25) days of the date on which the ITA issues its remand determination, submit memoranda addressing the ITA’s remand determination, not to exceed thirty (30) pages in length; and it is further
Ordered that the parties may, within fifteen (15) days of the date on which memoranda addressing the ITA’s remand determination are filed, submit response briefs, not to exceed twenty (20) pages in length.
Three of the four foreign producers who participated in the investigation have challenged the Final Determination, i.e. Pusan Steel Pipe Company, Ltd. ("Pusan”); Hyundai Pipe Company, Ltd. (“Hyundai”); and Korean Steel Pipe Company, Ltd. (“KSP”). The court will refer to these foreign producers collectively as “Defendant-intervenors”. A fourth company, Masan Steel Tube Works Company, Ltd., does not challenge the ITA’s Final Determination.
See Verification Exhibit 21, frames 556-58; AR Non-Pub, Doc. 119, frame 860; AR Pub. Doc. 140, frame 399; AR Pub. Doc. 80 at 8-9.
See AR Pub. Doc. 81 at 8-9; AR Pub. Doc. 288, frame 2353; see also AR Non-Pub. Doc. 132, frame 1269(ITAinternal memorandum regarding overrun sales fails to address price and profit differentials between overruns and sales made in ordinary course of trade). The court does not base its decision on an exhibit which Hyundai claims the ITA apparently forgot to take at verification. See ITA Brief at 21 n.9.
Because respondents failed to report sales information for alleged overruns, the ITA programmed its computer to resort automatically to constructed value information in the home market database. The ITA amended its Final Determination to reflect this correction. 57 Fed. Reg. at 49,453.
ITA Brief at 34-35 (providing numerous citations to the record); see AR Pub. Doc. 268, frames 1730-31,1733-34; AR Pub. Doc. 269, frames 1749-51; AR Pub. Doc. 272, frames 1788-92; AR Pub. Doc. 284, frame 2335; AR Pub. Doc. 288.
See, e.g,, AR Pub. Docs. 80-82; AR Pub. Doc. 117, frames 33-37; AR Pub. Doc. 140, frames 401-05; AR Pub. Doc. 221; AR Pub. Doc. 247; AR Pub. Doc. 264, frames 920-21; AR Pub. Doc. 270, frames 1771-73.
19 U.S.C. § 1677b(e) defines the constructed value of Imported merchandise as the sum of, inter alia:
[T]he cost of materials (exclusive of any internal tax applicable in the country of exportation directly to such materials or their disposition, but remitted or refunded upon the exportation of the article in the production of which such materials are used) and of fabrication or other processing of any kind employed in producing such or similar merchandise * * *. (Emphasis added).
This Act, which was enacted on November 28,1987, was subsequently abolished, effective December 31,1990. Defendant-Intervenors’Memorandum of Points and Authorities In Reply To Plaintiff and Defendant’s Memoranda of Law In Opposition To Defendant-Intervenors’ Motion For Judgment On the Agency Record ("Defendant-Intervenors ’ Reply Brief”) at 76 n.151.
19 C.F.R. § 353.68 states in full:
The Secretary normally will calculate foreign market value and United States price based on sales at the same commercial level of trade. If sales at the same commercial level of trade are insufficient in number to permit an adequate comparison, the Secretary will calculate foreign market value based on sales of such or similar merchandise at the most comparable commercial level of trade as sales of the merchandise and make appropriate adjustments for differences affecting price comparability.
The ITA’s traditional correlation test for Ievel-of-trade analysis has entailed a comparison of prices and functional levels of trade. See, e.g., Titanium Sponge from Japan, 57 Fed. Reg. 567, 657 (Jan. 7, 1992); Porcelain-on-Steel Cooking Ware From Mexico, 55 Fed. Reg. 21,061, 21,065 (May 22, 1990). Recently, however, the ITA has exercised its discretion by including selling expenses as an additional factor relevant to the correlation test. U.S. Dep’t of Commerce, Int’l Trade Administration, Import Administration Policy Bulletin 92/1 at 2 (July 29, 1992) ("Policy Bulletin”). In this case, Defendant-Intervenors only presented evidence that prices did not vary by level of trade; they failed to present evidence that selling expenses were not correlated to levels of trade. The court notes, however, that the Policy Bulletin was issued three months after the preliminary determination in this case. Moreover, the Policy Bulletin specifically states that “[t]his policy will be implemented in all future cases and outstanding administrative reviews where the necessary information can be gathered and used in the preliminary results without delayingthe completion of the review.” Id. at 3 (emphasis added). Thus, the “selling expenses” factor is not relevant to the present case.