Usinas Siderurgicas de Minas Gerais, S.A. v. United States

*744Opinion and Order

Aquilino, Judge:

At issue in this action, which consolidates a whole series of filed complaints, are the Final Determinations of Sales at Less Than Fair Value: Certain Hot-Rolled Carbon Steel Flat Products, Certain Cold-Rolled Carbon Steel Flat Products, Certain Corrosion-Resistant Carbon Steel Flat Products, and Certain Cut-to-Length Carbon Steel Plate From Brazil of the International Trade Administration, U.S. Department of Commerce (“ITA”), published at 58 Fed.Reg. 37,091 (July 9, 1993). The titular plaintiff which is referred to from time to time hereinafter by its chosen acronym USIMINAS, contests the determinations via a motion pursuant to CIT Rule 56.2 for judgment upon the agency’s record.1 The motion articulates multitudinous reasons for relief which can be summarized as follows: (i) the ITA conducted its investigation(s) with “undue bias and hostility”, to quote from the amended complaint; (ii) the agency’s rejection of USIMINAS responses and resort to best information otherwise available was unwarranted; (iii) the ITA opted for punitive dumping margins even though USIMINAS was a cooperative foreign respondent; and (iv) bevelled plate should not have been held within the scope of the agency’s investigation of cut-to-length steel plate.

On their part, the encaptioned intervenor-defendants have also interposed a Rule 56.2 motion to the effect that the final margins are too low, based upon erroneous ITA “simple” averaging of their petition rates.

The court’s jurisdiction is pursuant to 28 U.S.C. §§ 1581(c), 2631(c), and its standard for review of the contested agency determination(s)2 is whether they are unsupported by substantial evidence on the record or otherwise not in accordance with law. 19 U.S.C. §1516a(b)(l)(B); 28 U.S.C. §2640(b).

I

The petitions filed with the ITA implicated manufacturers of steel flat products around the world, including USIMINAS and Companhia Sid-erúrgica Paulista (“COSIPA”) and Companhia Siderúrgica Nacional (“CSN”) in Brazil. The ensuing agency investigation(s) entailed questionnaires to those firms and the others regarding their home markets and circumstances of production and pricing. In the case of Brazil, the ITA was concerned from the outset about hyperinflation, which has been defined as “extreme inflation, usually over 100 percent per year, *745and * * * characterized by a devaluation of the currency,” North Star Steel Ohio v. United States, 17 CIT 459, 467 n. 7,824 F.Supp. 1074, 1081 n. 7 (1993), citing Samuelson & Nordhaus, Economics 81,975 (13th ed. 1989). Indeed, the finding in the determination(s) at bar is that Brazil’s rate of inflation “was never less than 20 percent a month” and “in each of the past five years * * * has never been lower than 475 percent”. 58 Fed.Reg. at 37,093. In such a situation, the ITA attempts to decipher foreign-market value on a monthly or even shorter basis, as opposed to the whole period of investigation.3 The same shorter-period approach is sought to be taken for cost of production and for constructed value. See Plaintiffs Appendix, Exhibit 6 (Section D).

When the respondents’ initial submissions did not lend themselves to this tack, the ITA notified COSIPA (and the plaintiff) that they

may not adequately compensate for inflation and that the period used to define “contemporaneous” sales should be shortened to two weeks or even one week, wherever possible.

Id., Exhibit 4, p. 1. The agency also stated that, if

COSIPA decides to provided diffmers on abiweekly or weekly basis, * * * [it would] calculate biweekly or weekly weighted-average FMV’s. In order to adhere to the principle of contemporaneity, the Department will not use diffmers if the sale date and shipment date of both the U.S. sale and the home-market sale(s) being compared occurred in different monthly, biweekly, or weekly periods. Dif-fmers and OOP’s should therefore be calculated on the same time-period basis * * * and should not straddle two different time periods.4

USIMINAS disagreed with this approach5, but to little avail. For similar merchandise, all price-to-price comparisons were required by the ITA to be based on sales within the same month, with diffmers to reflect the difference between variable replacement costs on the date(s) of sale; and when no home-market sales of similar merchandise were available for comparison, or if neither product was manufactured during the month of sale, constructed value, based on a replacement cost of production, was required to be used. See id., Exhibit 5, pp. 1-2.

When the agency reported its preliminary determination of sales at less than fair value6, it relied on USIMINAS submissions to calculate cost of production and constructed value (“CV”) but invited the company to supplement the data. Several ministerial errors were detected and *746corrected, which led to amendment of the preliminary determination. See 58 Fed.Reg. 28,393 (May 13, 1993). Upon attempted verification, however, the ITA came to conclude that the USIMINAS submissions had “serious, irreparable flaws”7, whereupon it decided to resort to the best information otherwise available (“BIA”) in lieu of the company’s data. The final determination lists a weighted-average margin of 42.08 percent for USIMINAS, 109.00 for COSIPA, and 75.54 percent for all others then producing cut-to-length carbon-steel plate in Brazil. See 58 Fed.Reg. at 37,099.

A

As indicated above, in bringing suit for judicial review and filing its instant motion for relief, the plaintiff claims that the ITA “repeatedly changed its mind, without notice, about the information it required and the manner in which it had to be presented * * * [and] refused or was unable to comprehend the complexities that a hyperinflationary economy, suchas Brazil’s, added to anantidumpinginvestigation.” Plaintiffs Memorandum, p. 21. The following questions are posed, among others:

(a) Whether the final adverse determination was the result of undue bias and hostility toward USIMINAS by Commerce Department officials, and part of a general pattern of unfair and unwarranted adverse determinations against foreign producers of flat-rolled steel.
(b) Whether the Commerce Department’s conduct of both the price-to-price and cost investigations made it impossible for USI-MINAS to receive a fair and impartial decision on the merits.

Amended Complaint, para. 6. The motion at bar also avers that the agency

staff gave inconsistent and ambiguous instructions to USIMINAS and its counsel (in many cases departing from precedent and prior practice), declined to clarify those instructions in a timely manner, refused to accept USIMINAS’ own clarifyinginformation (while accepting similar clarifications from respondents from other countries being investigated), and ultimately rejected all of the information USIMINAS submitted.

Plaintiffs Memorandum, pp. 21-22.

The defendants respond that any difficulties were not caused by the instructions but by the company’s inadequate comprehension. They also point out that difficulties with the ITA instructions could not have caused the fundamental flaws discovered in the USIMINAS database during attempted agency verification.

(i)

According to the plaintiff the ITA was required to provide special instructions to account for Brazil’s hyperinflation. It argues that the *747agency’s failure to give adequate guidance is most evident in the methodology used to pinpoint the home-market sales for price comparisons and in the requirements for cost-of-production analysis. The plaintiff claims to have requested instructions at the beginning of the investigation(s), only to be informed they were unavailable but “would be prepared shortly”. Id. at 24. Once available, the instructions were

without opportunity to comment, without any reference to prior practice or explanation of the basis for the change in prior practice, and with unrealistic deadlines for submission of the requested information.

Id. at 4. Moreover, the above-quoted agency diffmer methodology forecloses respondents in hyperinflationary economies from selecting similar home-market products for comparison with products sold in the United States unless the sale and shipment dates are within the same month in both markets. See id. at 26. Because of a two-to-three-month USIMINAS production cycle, the plaintiff asserts that the ITA “virtually guaranteed that [the company] would not be able to utilize price-to-price comparisons for any U.S. sales that did not have an identical product match in the same month as the U.S. sale.” Id.

The defendants respond that same-month comparisons of sales did not constitute a change in practice, that earlier investigations8 entailed similar matching methodology, and that company counsel seemingly understood the superiority of contemporaneous comparisons. See Defendants’ Memorandum, p. 131. The defendants deny plaintiffs assertion of unfairness. With respect to the claim that the agency further confused matters by modifying the product-matching method, the defendants argue that, in response to USIMINAS objections, there was a “liberalization” of its practice to permit a greater number of matches.

Of course, the ITA has discretion to alter methodology — within certain limitations. See, e.g., Koyo Seiko Co. v. United States, 20 F.3d 1160, 1162-63 (Fed.Cir. 1994). And those precedents cited by the plaintiff9 do not hold otherwise. Here, the agency informed USIMINAS of its approach, requested pertinent information, and gave the respondent(s) opportunities to comment. The record reflects interaction between the ITA and them, the nature of which this court cannot call “unfair”. See, e.g., Defendants’ Appendix, Exhibit 10 (USIMINAS offered opportunity to clarify information prior to the final determination); Exhibit 21 (the company requested clarification of agency correspondence); Exhibits 35 and 36 (USIMINAS offered (and accepted) the opportunity to comment *748on the amended preliminary determination); Exhibits 3,4,7,16,25,29, 30(granting the company extensions of time to respond).

(ii)

The plaintiff contends that the ITA “repeatedly failed to define its ‘replacement cost’ methodology in a manner which provided adequate notice to respondents of how to develop adjustments for similar merchandise and adequate response” to the COP and CV questionnaires. Plaintiffs Memorandum, p. 31. It also claims that the only guidance USIMINAS received merely directed it to use replacement costs in the calculation of cost of production and constructed value and to provide them on a monthly basis. See id. at 32. Cf. Plaintiffs Appendix, Exhibit 6, p. 1. Even with a precise definition, the plaintiff argues that any attempt to calculate replacement costs was unduly burdensome, “requiring literally tens of thousands of calculations not normally required.” Plaintiffs Memorandum, p. 30.

The defendants counter that the ITA “provided adequate instructions, abundant opportunities for comments on its policies, substantial feedback, reasonable deadlines, and a steady stream of deadline extensions and other accommodations to USIMINAS.” Defendants’ Memorandum, p. 19. They point out that, in previous investigations, the agency required replacement costs if the adjusted actual costs were inaccurate, citing Camargo Correa Metais, S.A. v. United States, 17 CIT at 899-900 (“for all costs to be recovered in the normal course of trade, home market prices must be sufficient to recover the cost of replacing the materials used to manufacture the product”). Regarding the instructions to USIMINAS, it is argued that the agency did “not prescribe a single, fill-in-the-blanks approach to calculating cost of production[] and cost of manufacture”. Defendants’ Memorandum, p. 134. Furthermore,

it would not have been appropriate for the Department to elaborate and defend a single replacement cost methodology * * *, not only because cost reporting must be appropriate to the facts of individual cases, but also because fundamental flaws and omissions in USIMINAS’ databases prevented the Department from ever utilizing the cost data USIMINAS had reported.

Id. at 140-41.

Be that as it may, the plaintiff indicates that USIMINAS could have computed replacement costs, although to have done so would have been labor intensive.10 And the record reflects that the company was free to consult with the agency regarding methodology. See Plaintiffs Appendix, Exhibit 6, p. 3. In addition, there is no showingthat the errors in the USIMINAS data were created by the ITA’s instructions. Similarly, the *749claim that the company was prejudiced by unclear instructions and strict deadlines has little merit. The record indicates that the company had chances to comment; that any comments were considered, and some resulted in revisions; that the company was granted extensions of time to comply; and that it failed to raise at the administrative level all of its instant arguments.

As for the claim that “the limitation on submitting additional clarifying information imposed on Brazilian respondents was more restrictive than that imposed on respondents from other countries”11, the agency has discretion in accepting or not accepting data. That the agency’s approach for the respondents in Brazil may have differed from its approach for other respondents in other circumstances is not in itself a violation of law.

(iii)

At the time of the ITA’s steel investigations worldwide, the Department of Commerce, Office of the Inspector General audited the agency’s performance and promulgated a report entitled “Import Administration’s Investigations of Steel Industry Petitions.” Upon commencement of this action, the plaintiff interposed a motion to supplement the agency record with that report. The motion was denied by the Chief Judge.

The plaintiff renews the motion now, alleging, among other things, that the report contains “many admissions”. Plaintiffs Memorandum, p. 43. Even if true, this would not be ground for adding to the administrative record matter not before the decisionmaker at the time of its determination. On the other hand, the plaintiff has made the report a part of the judicial record. And the court is unable to find that it supports plaintiffs specific concerns. Ergo, the motion for reconsideration should be, and it hereby is, denied.12

II

To justify its resort to best information otherwise available, the ITA reports “serious, irreparable flaws in the respondent’s submission” viz. date of sale; cost of production, constructed value and diffmer; product concordance; and quantity and value of sales. See 58 Fed.Reg. at 37,092-93. Whereupon the agency stated that,

[bjecause USIMINAS cooperated with the Department’s investigations, * * * [it] used as BIA for that respondent the average dumping margin alleged in the petitions for each product.

Id. at 37,094. Nonetheless, the plaintiff takes the position that its data were more accurate than the “arbitrary rate” selected for USIMINAS. Plaintiffs Memorandum, p. 44.

*750A

The agency’s final determination states that the company reported

incorrect dates of sale in 38 percent of the U.S. sales that the Department reviewed at verification. In 23 percent of the reviewed sales (60 percent of the sales with erroneous dates), the correct sale date was outside the originally reported month of sale; in certain cases, the correct date of sale was outside the [period of investigation] entirely. * * * Without knowing whether the correct universe of sales and month of sale have been reported, the Department is unable to carry out two essential components of its LTFV calculations. First, the Department is unable to determine to which home-market sale(s) any given U.S. sale should be compared. * * * Second, the Department is unable to calculate adjustments for differences in the physical characteristics of the merchandise. * * *

58 Fed.Reg. at 37,093. Moreover:

* * * At such high monthly rates of inflation, it is crucial that LTFV comparisons be made, and diffmers calculated, on a monthly basis rather than over the entire [period of investigation] because erroneous cross-month comparisons — even if the date of sale is off by a single month — would distort LTFV margins by over 20 percent, which is unacceptable.13

The plaintiff contends that it correctly used the “specification and size assortment advice”, also referred to as the “final specification” document during verification, for calculating date of sale “since it was the first document establishing the price and quantity to be purchased.”14 In this action, the plaintiff claims that (i) its methodology was not flawed, (ii) that the errors discovered were not “serious”, and (iii), whatever those errors, they had no significance for the agency’s analysis. Plaintiffs Memorandum, pp. 47-48.

(i)

The plaintiff contends that the ITA failed to show “a single flaw” in the date-of-sale methodology and that the company’s approach “was the only correct [one] that could possibly have been used”. Id. at 48. With a voluminous database, encompassing hundreds of thousands of transactions, it is argued that handpicking the best possible document for each transaction was impossible. Id. at 49. During such investigations, the *751plaintiff claims that the agency selects a single document for date-of-sale purposes, notwithstanding any subsequent changes in terms.

The court cannot concur with this last representation. It can agree that USIMINAS’s decision to use the specification- and-size-assortment advice provided a convenient starting point for responding to the ITA’s questionnaire, hut not necessarily the fin'al or only one. Cf. e.g., Final Determinations of Sales at Less Than Fair Value: Certain Hot-Rolled Carbon Steel Flat Products, Certain Cold-Rolled Carbon Steel Flat Products, and Certain Cut-to-Length Carbon Steel Plate From Belgium, 58 Fed.Reg. 37,083,37,090 (July 9,1993)(date of sale to be reported as “the date that price and quantity are first set”). Here, upon comparison of earlier with later sales documents, the court observes that some transactions do exhibit differences in both pricing and quantity amounts. See Plaintiffs Appendix, Exhibit 15, pp. 39-44. Budd Co. v. United States, 14 CIT 595, 604, 746 F.Supp. 1093, 1100 (1990), established that it is

appropriate for Commerce to choose to effectuate the primary statutory purpose in favor of fair determinations based on contemporaneous comparisons. * * * [W]ere Commerce to do otherwise, dumping margins would result from the application of exchange rates, a circumstance which was wholly beyond the control of the exporters, an unacceptable outcome.

Citations omitted. See also General Housewares Corp. v. United States, 16 CIT 24, 783 F.Supp. 1408 (1992). In short, plaintiffs proposed exclusive reliance on the specification-and-size-assortment advice for determining date of sale is misplaced.

(ii)

At verification, the ITA found that some 38 percent of USIMINAS’s reported dates of sale were incorrect.15 Since several of the sales folders covered more than one transaction, the plaintiff posits that the agency “very likely examined all of the transactions included on the examined invoices”, thereby leading to, at most, a ten percent rate of error.16 The plaintiff also points out that four of those transactions accurately specified the dates of sale, despite agency claims to the contrary. See Plaintiffs Memorandum, pp. 57-60.

No one denies that errors in reporting dates of sale can and did occur. The court should not simply consider the percentage of such error, it must also examine the record relied upon by the agency in concluding that that error was of moment. Here, examination reveals a number of *752USIMINAS sales in disparate months, or even outside the period of investigation altogether. See 58 Fed.Reg. at 37,093; Plaintiff s Appendix, Exhibit 16. And no satisfactory explanation is offered for those variations in price or quantity on the final specification documents and sales listings, leaving this court unable to conclude that the agency’s finding with regard to dates of sale is unsupported by substantial evidence.

(iii)

The plaintiff disputes the ITA’s position that it could not verify (a) whether USIMINAS had reported the correct universe of sales for the period of investigation17 and (b) whether the company had correctly matched U.S. sales with home-market sales in the same month. See 58 Fed.Reg. at 37,093.

(a)

The plaintiff argues that, since the agency “found no discrepancies or errors” upon testing for completeness, then it, in fact, verified the entire universe of sales as accurate. That is, the ITA “was able to tie [the] universe of reported U.S. sales to its financial statements to the ton.” Plaintiffs Memorandum, p. 62 (emphasis in original). On the other hand, the total number of reported sales should have varied significantly from the amount shown on USIMINAS’s books had its sales information been seriously flawed. See Plaintiffs Reply Memorandum, p. 62.

The defendants counter that the plaintiff takes the significance of the completeness test out of context. As the verification report and the ITA’s submission make clear, the agency’s purpose is not to show that the correct shipments were investigated, rather that the details of particular shipments match what has been reported to the ITA. See Defendants’ Memorandum, p. 33. Moreover, according to the agency, the completeness-test data are not cross-referenced against the final-specification or other documents. Instead, verifiers merely cross-check data against a sales listing at the moment of shipment.

The record shows that the agency verification did uncover errors in reporting “the universe of U.S. sales”, including a grade of alloy plate and other cut-to-length plate from the USIMINAS sales listing. See Plaintiffs Appendix, Exhibit 15, p. 37. The reference to “no discrepancies or errors” is only to one aspect of a multipart test.18 Moreover, while the sales universe may match USIMINAS’s financial records, it is clear that the ITA only cross-checked shipment data against a listing of sales.

*753(b)

The plaintiff maintains that, as long as a contemporaneous exchange rate is used in converting home-market prices to dollars, a change in the date of U.S. sale does not cause margin distortion, given the agency’s method for determining foreign-market value. This seemingly assumes that the cruzeiro/dollar exchange rate parallels inflation so that the conversion to U.S. dollars will neutralize the effects of inflation on the agency’s margin determination. See Plaintiffs Memorandum, p. 66. The plaintiff also contends that distortions due to hyperinflation would only occur if the ITA converted foreign-market value for one period with an exchange rate from another. See id. at 64.

The defendants respond that home-market prices do not necessarily follow fluctuations in the rate of exchange. Those prices may increase at a greater rate than the number of cruzeiros needed to purchase dollars, thereby resulting in foreign-market value denominated in dollars increasing faster than United States price. See Defendants’ Memorandum, p. 35. They also observe that, although FMV and USP may reflect those hyperinflationary effects, they must still be matched to the right month for the dumping margin to be properly calculated. See id. at 36.

Clearly, fluctuations in the rate of inflation, if unaccounted for, can engender margins that are distorted. Cf., e.g., Budd Co. v. United States, 14 CIT at 602, 746 F.Supp. at 1099 (“dumping margins should not be based wholly on the misapplication of exchange rates”). Here, the record indicateS\that home-market prices were not in lockstep with inflation or the exchange rate and, in fact, varied. Cf. Domestic Steel Producers’ Opposition Memorandum, Appendix, Tab 6, Exhibit 6. And no one denies thVt date of sale is critical in a hyperinflationary case. See, e.g., transcript of oral argument (“Tr.”), pp. 39, 51-52.

B

The plaintiff takes the position that the record lacks substantial evidence in support of the premise that its cost information was deficient. The agency’s final determination refers to coal cost, labor cost, depreciation, general and administrative expenses, pension costs, cost of extras, and standard costs for new products. See 58 Fed.Reg. at 37,096-98 (Comments 6-13).

(i)

That determination states that, in its questionnaire response,

USIMINAS reported projected coal costs based on its anticipated expenses as of the date at which USIMINAS takes title to the merchandise. USIMINAS’ anticipated expenses included ocean freight, port charges, taxes, and all related movement charges required to transport the coal to [its] mill * * *. Since USIMINAS takes possession of the coal at the time the ship leaves the foreign port, movement expenses * * * generally were estimated during the month before they were actually incurred. * * * USIMINAS’questionnaire *754response did substitute actual costs for the coal itself, but continued to use estimates for movement charges * * *. As a result, USIMI-NAS failed, at verification, to * * * accurately persuade the Department that its reported expenses accurately represent its current monthly costs for coal.

58 Fed.Reg. at 37,096. In response, the plaintiff now argues that

(1) the error had a nominal effect on the accuracy of the costs presented (in the range of 0.2%-0.4%); (2) the error affected only certain elements of the coal costs; and (3) the Department had information to allow it to adjust the costs to take into account the effects of the error on the costs presented.

Plaintiff s Memorandum, pp. 69-70. It continues that the recording of these costs is “quite complicated”19 because “movement expenses are first recorded based on projections and then adjusted to actual * * * costs when * * * known. ” Id. at 69. The plaintiff claims that it could tie its cost and constructed-value information directly to its financial statement when the replacement values of those costs are reversed and the historical values are substituted. See Plaintiff s Appendix, Exhibit 18, p. 61 (citing Exhibits 46B and 46C of the Cost Verification Report in support).

The defendants do not dispute the extent of the coal-cost discrepancy and even acknowledge some errors made by the agency. See Defendants’ Memorandum, pp. 39, 41 n. 22. Instead, they claim that USIMINAS’s reported costs could not be tied to its financial statements. Id. at 41. For one coal purchase, for example, only the forecast cost, along with an additional cost representing an exchange-rate variation, were entered into the general ledger. Other costs were not tied to the general ledger or satisfactorily explained. See Plaintiff s Appendix, Tab 9 (Exhibit 48). On the other hand, USIMINAS’s reporting of coal costs creates but a minor distortion, representing perhaps 0.2 to 0.4 percent of the costs of production, hardly the threshold, standing alone, necessary to justify resort to best information otherwise available. See, e.g., Asociacion Colombiana de Exportadores de Flores v. United States, 13 CIT 13, 16, 704 F.Supp. 1114, 1117 (1989) (“Defects in data which are minimal do not result in failure of verification requiring use of best information available in place of all the respondent’s data”), aff'd, 901 F.2d 1089 (Fed.Cir.), cert. denied, 498 U.S. 848 (1990).

(ii)

The plaintiff objects to the claim that USIMINAS understated labor costs by “not including accruals for holiday pay, bonuses, and ‘thirteenth-month’ salary[20], which are paid in other months.” 58 Fed.Reg. at 37,096. According to the ITA,

company officials admitted that they had not reported accruals as part of the fringe benefit costs reported in their labor cost response *755for the selected month * * *. These costs were included in the annual audited financial statement. Accordingly, the COP/CV data did not reflect all costs incurred.

Id. The plaintiff contends that any understatement of labor costs was caused by the agency’s failure to fully comprehend USIMINAS’s record-keeping. As indicated above, the ITA required the company to present its data using replacement costs, even though USIMINAS did not keep the information that way. Instead, the company utilized one of two record-keeping methods, either the “constant currency” system21 or the “corporate legislation financial statement.”22

Relying upon data from its constant-currency system, the plaintiff claims that (1) costs submitted on a replacement-cost basis will never tie directly to a financial statement based on historical costs, (2) the agency should distinguish between accruals related to bonuses, holiday pay, and thirteenth-month salary attributable to production in one month with different accruals occurring in other months of production caused by restatements in constant currency, and (3) the restatement of prior-month accruals is not appropriate in a replacement-cost system. Plaintiffs Memorandum, pp. 77-78.

The ITA argues that plaintiffs reliance on its constant-currency ac-countingis misplaced since “booked actual monthly costs already reflect the impact of inflation and, thus, can be reconciled to the historical financial statement.” Defendants’ Memorandum, pp. 45-46. The agency objects to USIMINAS’s constant-currency adjustment for those accrued labor costs since they appear on the company’s cost-of-manufacturing statement and should, therefore, already show the replacement costs.

This court concludes that the ITA’s interpretation of the corporate legislation financial statement does not fully consider the complexity of USIMINAS’s dual accounting and over-simplifies the data submitted. The agency fails to adequately explain how adjustments for prior months, which are charged as costs for those months, are taken into account. They do not clearly fall within the definition of replacement costs since some accrue in one month but then result in an expenditure in a. subsequent month.23 As has been suggested by the plaintiff differences between the submitted costs and the amounts shown on USIMINAS’s financial statements could also represent an accountingfor inflationary *756effects. In any event, the court is unable to concur that the reported fringe-benefit accrual costs are deficient.

(iii)

The ITA concluded that USIMINAS altered its depreciation costs by using unrealistic figures in calculating the useful lives of assets. Specifically, it changed

the depreciable lives of * * * asset[s] in the middle of these investigations[, which] substantially distorted the submitted COP/CV and diffmer data. Products manufactured early in the [period of investigation] reflect significantly different depreciation costs than those produced in the latter part of the investigation. Furthermore, the revised remaining lives are much longer than the life commonly utilized for such assets in the steel industry worldwide * * *. This inconsistency is even greater when consideringthe number of years the revalued assets have already been in use.

58 Fed.Reg. at 37,097.

The plaintiff contends that there is no evidence that these changes understated depreciation, nor were they challenged by USIMINAS’s shareholders or auditors, or by Brazilian tax or securities authorities. Moreover, it claims that the changes did not commence until after the period of investigation. Thus, they had no effect on diffmer calculations or any “material effect on the results.” Plaintiffs Memorandum, p. 99.

NTN Bearing Corp. v. United States, 17 CIT 713, 720, 826 F.Supp. 1435, 1441 (1993), citing Ipsco, Inc. v. United States, 12 CIT 1128, 1130 n. 3,701 F.Supp. 236, 238 n. 3 (1988), upheld the use of a firm’s expenses as recorded in its 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.” See Cinsa, S.A. de C.V. v. United States, 21 CIT 341, 343, 966 F.Supp. 1230, 1234-35 (1997). On the other hand, the ITA may consider expenses not recorded on company books even when that enterprise is relying on the home-market’s generally-accepted accounting principles. See, e.g., NTN Bearing Corp. v. United States, 19 CIT 1221, 1235-36, 905 F.Supp. 1083, 1095-96 (1995).

Here, the changes to asset lives were substantial and had an impact on reported costs. As the record shows,

USIMINAS revised the * * * lives of its plants [sic] assets. This change in accounting estimate was based on a four month study performed by USIMINAS’s engineers. [] For each asset the expected useful [life] was extended thereby reducingthe depreciation associated with these assets. This change reduced the submitted CV for those products shipped from July through October * * *.
* * * In reviewingthe submitted depreciation cost for June, the Department noted that the submitted amount * * * was less than the amount reflected in the monthly financial statement * * *. It may be appropriate to use the amount reported in the monthly financial *757statement as this was the amount reflected in the audited annual report.

Plaintiffs Appendix, Exhibit 9, pp. 21-22. The useful lives assigned by USIMINAS were indeed longer than those assigned by comparable companies within the steel industry. See Defendants’ Appendix, Exhibit 49, Exhibit 50, App. 1. Even though the revised accounting for deprecation started after the period of investigation, to the extent subsequent, affected data were required for proper analysis24, the agency could not ignore them.

(iv)

The plaintiff disputes the ITA’s finding that USIMINAS’s general and administrative (“ G&A”) expenses were deficient in that they did not include a “social contribution tax”25 and certain non-operating expenses. As for the latter, the ITA, reports that they

represent costs which are ordinarily incurred in a company’s business. The Department includes these expenses in revenue in G&A except when such expenses/revenues are not specifically related to other business activities of the company. The cost verification exhibit relating to non-operating income and expenses shows that the great majority of these expenses relate to provisions for losses on loans to the electric company ELETROBRAS, not from sales of real estate. The Department verified that USIMINAS provided these loans to ELETROBRAS as part of its electricity payments. Since energy costs are related to the production of subject merchandise, these expenses should have been included in G&A.

58 Fed.Reg. at 37,097.

The plaintiff responds that those outlays should not have been included because they “related to investments in assets which were not ‘used in production’ consistent with Department practice.” Plaintiffs Memorandum, p. 85. There is no record evidence showing that the real estate investments result in non-operating losses or that the investments in ELETROBRAS involved assets which were used in production. Id.

Accordingto the defendants, “since no information has been provided on the nature of the real estate involved, there is no indication that these investments are not associated in some way with the operations of the company.” Defendants’ Memorandum, p. 52. They also note that the ELETROBRAS “investments” or “loans” are billed on USIMINAS’s monthly electric statements. Cf Plaintiffs Appendix, Exhibit 9, p. 20. Thus, the ITA concluded that they were related to the provision of elec*758tricity and operation of USIMINAS and should therefore have been included in G&A expenses, which are defined on the record as

those expenses which relate to the activities of the company as a whole rather than to the production process. These would include expenses which are not identified with a particular operation * * *.

Domestic Steel Producers’ Appendix, Exhibit 14, p. 14.

There is insufficient evidence whether the real estate investments were associated with the operations of the company or not. Since the plaintiff does not present an answer, the court can only note that the losses, if proven, would represent a small percentage of the total nonop-erating expenses. See Defendants’ Appendix, Exhibit 51. See also Magnesium Corp. of America v. United States, 20 CIT 1092, 1104 n. 63, 938 F.Supp. 885, 897 n. 63 (1996)(complainants, not Commerce, bear the burden of persuasion in this kind of action).

With respect to ELETROBRÁS, the amounts apparently are listed separately from the charges for electricity, are recorded as assets, and earn interest26, but they have not been repaid, and USIMINAS does not appear to expect any net return. See Defendants’ Appendix, Exhibit 51, p. 1 (ELETROBRÁS loans referred to as probable losses). The payments are compulsory27 and appear related to the procurement of electrical power.

(v)

The domestic petitioners claimed that USIMINAS understated its pension costs based on “an independent actuary’s report [that] these costs may not be sufficient to cover USIMINAS’ ultimate liability.” 58 Fed.Reg. at 37,097. In its final determination, the agency agreed with the report28, which

the Department examined at verification [and which showed] that USIMINAS’s pension-fund accruals fell short of the amounts required to meet its pension obligations. USIMINAS should have recorded this shortfall in its labor costs for 1992.

Id.

The company indicated that it reported its pension expenses in accordance with its accountingmethods and Brazilian GAAP The plaintiff argues that, had it stated pension costs in accordance with the actuary’s report, then it would have been exposed to allegations that its responses were deficient since its reported costs would not have matched those costs underlying its financial statement. In addition, the plaintiff points out that the shortfall between the fund’s reserves and payment expenses was “cumulative” and represented “underfunding since the inception of the fund”; it was not just attributable to 1992. See Plaintiff s *759Reply Memorandum, pp. 32, 33. Retroactive adjustments noted by the ITA, according to the plaintiff merely reflected subsequent events and did not provide a basis for resort to best information otherwise available.

In contrast, the agency is ofthe view that there simply was “amanagement decision * * * which lowered * * * booked costs for the year in question”. Defendants’ Memorandum, p. 54. Assuming retroactive adjustments were made, as indicated by USIMINAS’s half-year financial statement29, then the ITA contends that the company still bore the responsibility of reporting actual costs to the agency.

Indeed, the ITA is not limited to financial records kept pursuant to the home-country GAAE It may accept those records, but it also may reject them if their acceptance would distort true costs. NTN Bearing Corp. v. United States, 74 F.3d 1204, 1206 (Fed.Cir. 1995) (citations omitted). See also Camargo Correa Metais, S.A. v. United States, 17 CIT at 899 (“If the finding that Brazilian GAAP does not reasonably reflect the cost of production is supported by substantial evidence on record, then the ITA’s rejection of those procedures must be affirmed”). Quite clearly, uncon-tradicted30 actuarial results for the period of investigation indicate that USIMINAS’s pension costs were understated.

(vi)

The plaintiff objects to the agency’s finding that its “methodology for assigning costs to products did not differentiate between the cost of extras, as requested in the Department’s questionnaire.” 58 Fed.Reg. at 37,093. USIMINAS maintains costs for final products, not for individual extras, using a four-digit code system. The ITA verified that the company

itemized extras separately on customer invoices and that each extra increased the gross price by a certain amount. Allocating the total cost of all extras over many products without regard to whether or not they underwent additional processing is inappropriate. USIMI-NAS should have itemized the cost of these extras in its cost submissions to enable the Department to calculate diffmers and make “apples-to-apples” comparisons. * * *
Furthermore, the fact that USIMINAS is able to itemize the price of extras individually at the invoice level suggests that USIMINAS has the ability to allocate the cost of extras to those products incurring extras.

Id. at 37,097.

The company claimed it was “unable to report costs on a product-specific basis in any other way without reconstructing its cost accounting system from the ground up.” Id. Instead, it argues that an average, not an itemization of the cost of extras, is required since it “calculates only the costs for a given combination of extras not individual extras”. Plain*760tiffs Memorandum, p. 92. The failure to submit individual-extra costs was irrelevant since USIMINAS was able to identify each product compared, specify the extras, and provide the relevant costs. Finally, the plaintiff contends that the agency neither established that its methodology was “distortive in any respect nor that USIMINAS had the capability to provide the information in the format requested.” Id. at 94.

The ITA requests that data be submitted in a specified format. Here, the agency directed the company to provide the cost of extras as follows:

For each product sold in the U.S. and in the home market the costs of each extra(s) should be separately presented. The cost of the extra should include the materials, variable and fixed overhead and movement expenses.

Plaintiff s Appendix, Exhibit 6, p. 11. Although USIMINAS claims that it was impossible to itemize extras by their production costs, the company does not deny that extras are itemized separately on customer invoices. See Plaintiffs Memorandum, p. 93. While the agency may have been able to verify the cost of extras as proposed by the plaintiff31, it may conduct an investigation notwithstanding resulting burdens on a party in attempting to be forthcoming.

(vii)

The domestic steel producers argued that USIMINAS’s use of estimated, as opposed to actual, costs resulted in understatement of production costs for new products. The ITA found that

USIMINAS’ methodology in developing standard costs for new products, using normalized slab costs (which include [a weighted average of the] costs for both conventional and continuous slab processes) can cause an underallocation of costs to new products using only the conventional slab process[32]. USIMINAS’ methodology for calculating standards for new products caused distortion in the allocation of costs. The standard costs for new products were therefore significantly different from the actual cost incurred for those products.

58 Fed.Reg. at 37,097-98.

The plaintiff takes the position that the agency simply decided not to accept its accounting33 but failed to show that that system was inaccurate, unreliable, or that the company could have derived the costs any other way. That is, there is

no evidence that the number of transactions or new products affected was significant, that any of the transactions to the U.S. were affected, that any of the comparable merchandise sold in Brazil was affected, or that the manner in which USIMINAS’ cost accouting *761system addresses new products affects the integrity of the overall system.

Plaintiffs Reply Memorandum, p. 36.

There is evidence, however, that the underreporting of costs for new products manufactured by conventional process created distortion.34 Since slab costs for new products are based upon a weighted average, the agency indicates that the reported amount will not be as accurate as those standards used for older products. And, in any event, the ITA does not automatically accept a respondent’s financial information. Cf. Camargo Correa Metais, S.A. v. United States, 17 CIT at 899. Several large deviations were detected, which, in this court’s view, are attributable to USIMINAS’s accounting method for new products. See Plaintiffs Appendix, Exhibit 9, p. 9.

C

The ITA found fault with USIMINAS’s product concordance, citing several deficiencies and claiming that the company failed to conform to agency instructions. It found that

USIMINAS included alloy steel products as well as carbon steel in its sales listings and product concordances. USIMINAS also compared products with high-strength characteristics to products with low-strength characteristics. The overall effect of these errors was that not only was the universe of reported products in itself incorrect, but products sold in the U.S. market were susceptible to being improperly matched to products sold in the home market. Information available on the record was insufficient to allow the Department to determine the proper product matches.

58 Fed.Reg. at 37,093.

The plaintiff explains that problems arose, in part, because the agency circulated a “concordance table”, establishing those categories the ITA assumed identical, and gave USIMINAS only 48 hours to submit written comments. See Plaintiffs Memorandum, p. 108. The table, however, did not reflect the company’s product listings or own corporate records. Id. Given the short response period, USIMINAS claimed then that it was unable to “comment meaningfully” on the table and complains now that the agency adopted it “virtually without modification.” Id.

G)

The domestic steel producers contend that the company improperly matched alloy steels, which were outside the investigation’s scope, with certain carbon steels in their product concordance. See 58 Fed.Reg. at 37,098 (Comment 14). The ITA states that, while it

*762is able to identify and delete alloy products from U.S. sales, it is unable to do so on the home-market side because numerous grades of steel were sold in the home market, many of which were manufactured to proprietary specifications or foreign standards. Because insufficient information exists on the record with regard to these specifications and standards, the Department is unable to carry out the detailed analysis of the chemical composition of each product which would be required for us to separate alloy from carbon steel products. It is not the Department’s responsibility to correct significant substantive errors in a respondent’s submission; rather, it is the respondent’s responsibility to provide the Department with a complete and accurate sales listing and product concordance. * * *

Id.

The plaintiff explains that this issue arose because “a separate identification field” was not created. Plaintiffs Memorandum, p. 115. Instead, USIMINAS’s computerized system mixed together “a few products that had precisely the same grade, measurement and surface treatments for both alloy and carbon-steel products.” Id. And the inclusion of alloy-grade in the sales listings and product concordances had only a limited effect, and resulted in misidentification in only a few specific instances.

When the ITA discovered such inclusion, it requested grade specifications for products sold in the home market. A review thereof confirmed that alloy products had been matched with carbon steel35 in the concordance, thus distorting product matching. Although the plaintiff minimizes the severity of the distortion as involving a limited number of products, the defendants point out that each “grade” can encompass many “products”, and each “product” can have many sales. Defendants’ Memorandum, p. 80. See also Domestic Steel Producers’ Opposition Memorandum, p. 51 (“even if only four grades were involved, the errors could still affect * * * possibly thousands of individual sales transactions”) (underscoring in original).

(ii)

The petitioners pointed out that USIMINAS misreported and mismatched certain steel as “high strength” in the product concordance. That term was defined in the hot-rolled and cold-rolled questionnaires as having a minimum yield strength of 35,000 pounds per square inch (“psi”) to 50,000 or more. See Defendants’ Appendix, Exhibit 1, App. V, pp. 5,7. The ITA found that USIMINAS was unable to provide the study results in support of its submitted classification. The agency indicates that it analyzed other data, apart from merely measuring carbon content, to conclude that the steel was not high-strength. It reports the company’s position that

SAE norms do not specify a yield strength and that USIMINAS, therefore, d[id] not carry out yield-strength tests on a regular basis. *763In order, therefore, for USIMINAS to include a yield-strength variable in the product description, as required by the Department, USIMINAS used the best information available to it at the time. In response to the Department’s statement that SAE-grade material “does not have enough carbon to meet [the] minimum yield strength” * * *, USIMINAS claims that the classification of such materials was correct, based on a study carried out by three major Brazilian integrated steel producers and on the classifcation of such products under the Harmonized Tariff Schedule (HTS) * * *. USI-MINAS also asserts that, should the Department persist in its view that certain SAE-grade products are low-strength, the Department could easily re-code and re-match those products itself.

58 Fed.Reg. at 37,098 (Comment 15).

As indicated, the plaintiff claims that its classification was correct based on the Brazilian steel producers’ study which found that SAE-grade steel had an average yield strength of 40,000 psi. See Plaintiffs Memorandum, p. 110. Although USIMINAS did not submit that report for review, handwritten notes were presented which conclude that most of the SAE-grade materials had a yield strength of less than 275 mp, which, accordingto the defendants, equates to a minimum yield point of less than 40,000 psi. See Defendants’ Memorandum, pp. 70-72. Be that as it may, the court observes in passing that those handwritten notes are inconclusive of plaintiffs position. Cf. Plaintiff s Appendix, Exhibits 27, 28.

(iii)

The petitioners were of the view that USIMINAS incorrectly characterized some sales as hot-rolled sheet, when they should have been categorized as cut-to-length plate. The record is clear that, while the company defined any cut-to-length plate less than 5 mm in thickness as hot-rolled sheet, the agency considered that maximum to be 4.75 mm. See Plaintiffs Appendix, Exhibit 15, p. 37; Defendants’ Appendix, Exhibit 1, App. V

The plaintiff argues that this difference is irrelevant since none of its home-market sales of hot-rolled sheet thicker than 4.75 mm were used by the ITA in a product comparison for the hot-rolled concordance table. See 58 Fed.Reg. at 37,099 (Comment 16). Only a small percentage of sales, 1.755 or a total of 55236, were improperly classified, and none of the covered products were used in the comparison and therefore had any impact upon the hot-rolled-product concordance table. The plaintiff also takes the position that the agency could have reclassified those sales without resorting to best information otherwise available.

The ITA reports in its final determination that

this error had a significant effect on FMy UVJ and diffmers. Because “thick sheets” should have been included in plate thickness band “A” (i.e., less than 0.3125 in.), * * * the absence of these products in *764the plate database distorted weighted-average FMV calculations. This error also caused incorrect CV’s to be calculated, since CV’s are weight-averaged for each control number. Finally, this error caused incorrect diffmers to be calculated for plate falling within thickness band “A,” since USIMINAS did not include its “thick sheets” in the appropriate control numbers within the plate concordance. A significant number of USIMINAS’ U.S. plate sales fell within this thickness band and could have been matched to the missing “thick sheets”!.]

Id. And it now argues that “it was of paramount importance that all sales of steel products be reported in the context of the correct investigation.” Defendants’ Memorandum, p. 72.

It is uncontested that thinner plate requires more processing and is more expensive to produce than thicker ones. See, e.g., id. at 76; Plaintiffs Reply Memorandum, p. 79. Hence, the absence of thinner plate from the cut-to-length investigation was likely to skew the pricing data and lead to inaccurate determination of foreign-market value, constructed value and even margin of dumping. Despite the limited number of affected sales, the court is persuaded that the degree of error37 is of moment, since the capture of all U.S. sales at their actual prices is always at the heart of an ITA investigation. E.g., Florex v. United States, 13 CIT 28, 33, 705 F.Supp. 582, 588 (1989). Accordingly, respondents must report the volume and value of sales as precisely and completely as possible.

D

Notwithstanding the foregoing concerns of the ITA, the plaintiff claims to have been the victim of a “gotcha policy”38, whereby the agency

withholds scrutiny until verification, disagrees with the respondent’s position for the first time at verification, and then refrains from requesting the respondent to revise its information in accordance with the Department’s position.

Plaintiffs Reply Memorandum, p. 70. Cf. Böwe-Passat v. United States, 17 CIT 335, 343 (1993) (agency held arbitrary and capricious when it “sent out a general questionnaire and a brief deficiency letter, then effectively retreated into its bureaucratic shell, poised to penalize Bowe for deficiencies * * * that the ITA would only disclose after it was too late”).

Such was not and is not the case here. The ITA has discretion in determining when to turn to best information otherwise available. See, e.g., Geneva Steel v. United States, 20 CIT 7, 36-37, 914 F.Supp. 563, 590 *765(1996). While it may not do so on whim or inconsequential deficiency39, when errors are “cumulative and widespread, Commerce may reject the submitted information in toto.” Yamaha Motor Co. v. United States, 19 CIT 1349, 1360, 910 F.Supp. 679, 688 (1995), citing Rhone Poulenc v. United States, 13 CIT 218, 710 F.Supp. 341 (1989), aff'd, 899 F.2d 1185 (Fed.Cir. 1990). As discussed hereinabove, the record does reflect errors — in USIMINAS’s date-of-sale methodology, in its calculation of cost of production and constructed value, in developing its product concordance, for example — which satisfy the long-standing judicial definition of substantial evidence40 and which make the ITA’s determination to resort to the best information otherwise available in accordance with law.

Ill

In this action, the agency relied on total BIA, which means that the dumping margin is derived from sources other than any data submitted by the respondent to which it applies. See, e.g., Nat'l Steel Corp. v. United States, 18 CIT 1126, 1131, 870 F.Supp. 1130, 1135 (1994). Here, the rates alleged in each petition, including those based upon home-market prices and constructed values for all implicated Brazilian manufacturers, were averaged, resulting in final margins for USIMI-NAS of 42.08 percent for cut-to-length plate, 35.76 percent for cold-rolled sheet and 40.44 percent for hot-rolled sheet. See 58 Fed.Reg. at 37,099; Plaintiffs Appendix, Exhibit 30, p. 2.

The plaintiff takes the position that, since its errors discovered by the agency

were either reparable or insignificant, the Department’s determination to use total BIA was inappropriate. Even assuming that the alleged errors could not be corrected, the Department could have easily used partial BIA for these areas.

Plaintiffs Reply Memorandum, p. 85. It argues that, while allegedly “cooperative”, the rates are in reality punitive41 since they are higher than any other rate on record for USIMINAS. The plaintiff also claims that the U.S. and home-market prices which the petitions attribute to it42 are more accurate than the average of all petition allegations, which are predicated upon prices and values constructed by the domestic, U.S. industry for all Brazilian producers.

*766Those that constructed them, the intervenor-defendants herein, complain about the ITA’s factoring of their petitions’ home-market prices into the final margins. That is, the

only facet of the Department’s Final Determination at issue here is the inclusion of margins based on home market prices in the calculation of a BIA dumping margin for USIMINAS.

Domestic Steel Producers’ Memorandum in Support of Judgment Under Rule 56.2, p. 15. They contend their

petitions unequivocally demonstrated * * * that these home market prices were significantly below the corresponding costs of production and, hence, such prices could not be used as the basis of FMV

Id. at 18. See 19 U.S.C. §1677b(b) (1992).

The ITA has latitude in determining what constitutes best information otherwise available. It may reject inadequate or suspect submissions in toto, and it is not required to rely on BIA only for missing answers. See, e.g., Ad Hoc Comm. of AZ-NM-TX-FL Producers of Gray Portland Cement v. United States, 18 CIT 906, 914-15 and n. 21, 865 F.Supp. 857, 865 and n. 21 (1994). Moreover, the role of this court is “not to determine whether the information chosen by Commerce is the ‘best’ actually available, but whether the choice is supported by substantial evidence and is in accordance with law.” Novachem, Inc. v. United States, 16 CIT 782, 786, 797 F.Supp. 1033, 1037 (1992). The agency’s authority, however, is “subject to a rational relationship between data chosen and the matter to which they are to apply.” Manifattura Emmepi S.p.A. v. United States, 16 CIT 619, 624, 799 F.Supp. 110, 115 (1992). See also D & L Supply Co. v. United States, 113 F.3d 1220 (Fed.Cir. 1997) (“irrational” to uphold a rate when its foundation has been invalidated). But this “does not preclude Commerce from using a different methodology if it determines in its discretion that the circumstances so warrant. ” Hussey Copper v. United States, 21 CIT 217, 222, 960 F.Supp. 315, 319 (1997). The ITA may use

“all information that is accessible or may be obtained, whatever its source.”* * * This encompasses “the information submitted in support of the petition.” * * * Furthermore, best information available “is not necessarily accurate information, it is information whichbe-comes usable because a respondent has failed to provide accurate information.”

N.A.R., S.p.A. v. United States, 14 CIT 409, 416, 741 F.Supp. 936, 942 (1990) (citations omitted). Finally, in relying upon best information otherwise available, there is no requirement that the agency select a margin restricted to the firm to which the rate is assigned. Cf. Industria de Fundicao Tupy v. United States, 20 CIT 870, 874, 936 F.Supp. 1009, 1018 (1996).

On the record presented herein, it is not clear which, if any, of the petition margins are more probative than others. The plaintiff indicates *767that margins based on constructed value and pricing for all steel producers in Brazil favor manifestly less accurate over more accurate information. See Plaintiffs Memorandum, p. 127. But those “more accurate” data come from the petitions herein, not from USIMINAS itself. This court is also unpersuaded that margins derived from constructed value are invalid simply because they may be based on estimates rather than empirical evidence. Here, the record supports the ITA’s determination that it could not and therefore would not rely on USIMINAS’s own data. Hence, the intervenor-defendants’ contention that the ITA “implicitly concluded”43 that margins had to be based on constructed value because of USIMINAS home-market sales below cost is also off the mark since the agency did not make any findingto this effect and, in fact, rejected as deficient data submitted in regard thereto. In sum, the court finds that the ITA’s simple average of petition rates is supported by the record and otherwise in accordance with law.

IV

The plaintiff claims that the agency improperly expanded the scope of its cut-to-length steel investigation to include “non-rectangular cross-section” or “bevelled plate”44. While acknowledging that the agency may clarify the scope of an investigation, the plaintiff contends that the ITA abused its authority, given the exact wording of the petition in question, which refers only to “[rjolled products of solid rectangular (other than square) cross section”. Domestic Steel Producers’ Appendix, Exhibit 2, p. 3 n. 3. Cf. Plaintiffs Appendix, Exhibit 33. Since the petition did not explicitly include plate with a non-rectangular cross-section, the plaintiff also argues that it thereby unambiguously excluded bevelled plate. Moreover, the ITA’s belated addition of such plate adversely affected the preliminary determination of the ITC, which allegedly had not collected any relevant data on this type of product. See Plaintiffs Memorandum, pp. 152-60. Hence, the “expansion” was invalid since “later stages of antidumping proceedings must maintain the scope set forth in the earlier proceedings”. Id. at 160.

The defendants claim two ambiguities in the petition, to wit, “nonrec-tangular shapes are specifically excluded * * * while nonrectangular cross-section products are not”45, and there is an inconsistency between the tariff general notes, which only cover rectangular-cross-section products, and the explanatory notes, which classify nonrectangular and rectangular-cross-section products together. The defendants also point out that the iTC’s preliminary determination

*768reflected an ambiguity as to whether bevelled plate was included, comparable to the ambiguity contained [in] the petition, which was clarified by the Department. The ITC clarified this ambiguity in its final determination, which defined the products in terms of Commerce’s product descriptions.

Defendants’ Memorandum, p. 121.

Although it is clear that products of nonrectangular shape were explicitly excluded from the petition, the same is not true of nonrectangu-lar cross-sections. The record shows that “from the outset [the petitioners] intended that such products be within the scope of these investigations and did not believe language in the initiation notices regarding cross-sectional shape to exclude them.” Plaintiffs Appendix, Exhibit 33, p. 10. Furthermore, it is established that

the ITA has a certain amount of discretion to expand the language of a petition to encompass the literal intent of the petition, not to the exclusion of other factors, but certainly, with the purpose in mind of preventing the intentional evasion or circumvention of the antidumping duty law.

Mitsubishi Elec. Corp. v. United States, 12 CIT 1025, 1046, 700 F.Supp. 538, 555 (1988), aff’d, 898 F.2d 1577 (Fed.Cir. 1990). See also Wirth Limited v. United States, 22 CIT 285, Slip Op. 98-40 (April 3, 1998), appeal docketed, No. 98-1391 (Fed.Cir._June 30, 1998).

According to the record, bevelled plate is created through a relatively minor processing step, involving the burning off of “material at specified inclines.” Plaintiff’s Appendix, Exhibit 35, p. 5. The agency credibly demonstrates that including such product within its proceeding was not “a major step” since “ [p]roducts of nonrectangular cross-section may be within the same class or kind as the products currently in the scope. ” Id., Exhibit 40, p. 9. In fact, the explanatory notes to HTSUS heading 720846 include bevelled plate within the definition for flat-rolled products. Finally, the record reflects that information for such plate may have been obtained prior to the ITC’s preliminary determination. See, e.g., id., Exhibit 44. And the ITC’s Office of Investigations has confirmed that its

questionnaire did not explicitly exclude all nonrectangular cross-section products and that there is no evidence demonstrating that the questionnaire responses did not include data on products of nonrectangular cross-section.

Id.

Based on the foregoing, the court concludes that the ITA’s determination to include bevelled plate within the ambit of its proceeding is supported by substantial evidence on the record and otherwise in accordance with law.

*769V

To the extent that matters raised by the parties have not been addressed herein by the court, suffice it to state that they do not warrant specific reference. In the light of the discussion in parts I through IV of this opinion, plaintiffs and intervenor-defendants’ motions for judgment on the agency record should each be denied and this consolidated action dismissed. Judgment will enter accordingly.

Subsequent to the filing of this motion, counsel for the encaptioned intervenor-plaintiff interposed a motion for voluntary dismissal pursuant to CIT Rule 41(a)(2), which motion can be, and it therefore hereby is, granted.

The court notes that in Certain Flat-Rolled Carbon Steel Products From Argentina, Australia, Austria, Belgium, Brazil, Canada, Finland, France, Germany, Italy, Japan, Korea, Mexico, the Netherlands, New Zealand, Poland, Romania, Spain, Sweden, and the United Kingdom, USITC Pub. 2664 (Aug. 1993), the International Trade Commission (“ITC”) determined that an industry in the United States was materially injured by reason of imports of plate from Brazil, ergo the antidumping-duty order published at 58 Fed.Reg. 44, 164 (Aug. 19,1993), but also that there was no such injury, or threat thereof by reason of hot-rolled, cold-rolled or certain corrosion-resistant products from that country. To the extent there has been judicial review of those final negative determinations, they have been affirmed. See United States Steel Group v. United States, 18 CIT 1190, 873 F.Supp. 673 (1994), aff'd, 96 F.3d 1352 (Fed.Cir. 1996); Kern-Liebers USA, Inc. v. United States, 19 CIT 87 (1995); Nippon Steel Corp. v. United States, 19 CIT 450 and 827 (1995).

See, e.g., Camargo Correa Metais, S.A. v. United States, 17 CIT 897 (1993). The subsequent history of that ease in court is reported at 18 CIT 330 (1994), vacated and remanded, 52 F.3d 1040 (Fed.Cir. 1995), on remand, 21 CIT_, Slip Op. 97-159 (Nov. 25,1997).

Plaintiff s Appendix, Exhibit 4,p. 2. The acronyms “FMV” and “COP” refer to foreign-market value and cost of production, respectively The term “diffmer” refers to allowances for the differences in merchandise which are calculated when similar, rather than identical, goods are used in determining foreign-market value and when it is established to the agency’s satisfaction that pricing for the U.S. and the home market is due to such differences. Cf. 19 U.S.C. §1677b(a)(4) (1992). The ITA will normally consider costs of production in making those allowances. See, e.g., 19 C.F.R. §353.57(b) (1992).

See Plaintiffs Appendix, Exhibit 3.

See 58 Fed.Reg. 7,080 (Feb. 4, 1993).

58 Fed.Reg. at 37,092.

E.g., Final Determination of Sales at Less Than Fair Value: Industrial Nitrocellulose from Brazil, 55 Fed.Reg. 23,120, 23,122 (June 6, 1990).

In Creswell Trading Co. v. United States, 15 F.3d 1054 (Fed.Cir. 1994), an ITA ruling was held “inherently unfair” since the agency first indicated that certain information was unnecessary and then later declared “the missing information so relevant that it could not render a determination in its absence”. In Sigma Corp. v. United States, 17 CIT 1288, 841 F.Supp. 1255 (1993), the agency changed its position without giving notice to a respondent by issuing a final determination which was “the complete opposite” of its earlier position. In Mitsui & Co. v. United States, 18 CIT 185 (1994), the ITA’s use of best information otherwise available was held unfair since the agency had neither provided clear instructions on the method of compiling information nor given the respondent an opportunity to revise its original submission.

The plaintiff asserts that replacement costs encompass inventory value, which “is the raw material acquisition cost in the month for which cost is provided.” Plaintiffs Memorandum, p. 30. Calculation of such costs “involves extracting from the normal cost accounting system information on yields, rolling time, hourly costs, recirculation rates for costs from indirect cost centers, etc. and applying these to the ‘replacement’ value extracted from the normal system.” Id.

Plaintiffs Memorandum, p. 41 n. 45.

In addition, defendants’ motion to strike all references to the report contained in plaintiffs reply memorandum should be, and it hereby is, denied.

58 Fed.Reg. at 37,093. Date of sale was defined by the agency as

the purchase order date, the contract date, or, where written confirmation, is given, the order confirmation date (i.e., the point in the transaction where the basic terms of the contract, particularly price and quantity, are agreed to by the parties involved) * * *.

Defendants’ Appendix, Exhibit 1, App. II, p. 1.

Plaintiffs Memorandum, p. 51 (emphasis in original). USIMINAS indicated that its sales process begins with purchases through unrelated foreign trading companies. Thereafter, a

trading company’s order is finalized in the form of a purchase order. The purchase order is acknowledged and agreed to by USIMINAS with a sales confirmation. * * *
Approximately two weeks after the sales confirmation is issued, the customer defines the precise specifications of the material ordered and confirmed by the sales confirmation. The document is referred to as the "specification and size assortment advice.”

Plaintiffs Appendix, Exhibit 13, p. 10.

The agency examined 39 U.S. sales and found mistakes in 15 transactions. It does appear that two of the 15 with incorrect dates of sale, however, were not transactions initially targeted.

See Plaintiff’s Reply Memorandum, p. 54. For verification, the ITA randomly selected 40 “target” transactions. Since each folder examined encompassed other transactions along with the target, the plaintiff claims that the agency actually considered 145 rather than 40, two of which involved the same invoice.

The ITA denies having done so. And the defendants have served a motion for leave to file their Clarification of Certain Statements Contained in Defendants’ Memorandum in Opposition to Plaintiffs’ Motion for Judgment upon the Agency Record which is hereby granted. Disregardingadditional errors discovered in connection with nontarget transactions, the defendants continue to maintain that 13 of the 40 transactions selected revealed mistakes in the date of sale, resulting in the error rate of approximately one-third. See Defendants’ Memorandum, p. 29.

The agency relies on a “completeness test”, which determines “whether all sales which should have been reported were reported and properly classified, and that only sales which should have been reported were included in the sales listings.” Defendants’ Memorandum, p. 33 n. 16.

The ITA found “no discrepancies” after examining five randomly-selected sales from each of the product listings. As part of this “reverse check” process, certain random dates are selected, and then shipments made on that date are checked against the relevant sales listings. See Plaintiffs Appendix, Exhibit 15, p. 37. In another portion of the agency’s general discussion, however, it points to “various errors in the universe of U.S. sales subject to this investigation.” Id.

Plaintiff’s Reply Memorandum, p. 38.

«phis term refers to the accrual each month of a set amount, payable after twelve months.

This entails restatement of costs monthly, and cumulatively at the end of the fiscal year, to provide a clearer picture of a company’s operations. See Plaintiffs Memorandum, pp. 76-77.

This approach reflects Brazilian generally-accepted accounting principles (“GAAP”) by recording a temporary monetary correction as a capital reserve until the capitalization receives shareholders approval at their annual meeting. See Plaintiffs Appendix, Exhibit 9, p. 7 (Cost Verification Report).

With regard to the thirteenth-month payment, for example, the initial accrual in the first month would be insufficient to cover the company’s liability for the 13th month salary in December As a result, an additional accrual for the month one provision of 1 is requiredin order for the aggregate of the provision at the end of month 2 to be sufficient in terms of the value of the currency at the end of month two to be equivalent to two-twelfths of the 13th month payment * * *. It is not * * * a replacement cost of production in that month.

Plaintiffs Reply Memorandum, pp. 19-20 (emphasis in original).

As the intervener-defendants point out, since

normal lead time between the date of sale and the date of shipment was two to four months for USIMINAS, constructed values for March through October were matched with U.S. sales made in January through June.

Domestic Steel Producers’ Opposition Memorandum, p. 70 (footnote omitted).

Apparently, this is assessed only when a company has generated income. See Plaintiff’s Memorandum, p. 84.

Since there is no reference to this tax now by the defendants or intervenor-defendants, the court concludes that it was not an element of the ITA’s attempted computation of COP/CV and thus need not be considered herein. Cf. Tr. at 36.

See Plaintiff’s Appendix, Exhibit 7, p. 84.

See id., Exhibit 9, p. 20 (indicating that any companies owned or formerly owned by the state “pay for the development, maintenance and upgrade of Brazil’s hydro-electric power”).

See Defendants’ Appendix, Exhibit 52.

See Plaintiffs Appendix, Exhibit 24, n. 15.

Id., Exhibit 7, p. 88.

See, e.g., Defendants1 Memorandum, p. 55.

The conventional method is apparently more expensive than the continuous process.

USIMINAS uses an “INDOOR” or standard variation to measure variance in asingle month or semester between a single product and the average of all products in a 4-digit product specification (which represent actual costs). The INDCORis thenusedto adjust actual costs to arrive at a 27-digit product specification. See Plaintiffs Appendix, Exhibit 9, p. 8.

See, e.g., Defendants’ Appendix, Exhibit 54. In most of the INDCOR transactions presented, the increase from semester to semester is relatively small. Accordingly, “INDCOR changes of less than 25% are unlikely to be due to changes in the use of conventional versus normalized slab costs.” Plaintiff's Appendix, Exhibit 7, p. 91. A few transactions, however, do show differences well in excess of 25 percent.

gee Plaintiffs Appendix, Exhibit 15, p. 18. See also Defendants’ Appendix, Exhibits 60,61,62.

See Plaintiffs Memorandum, p. 120.

Albeit only some 3.5 percent. See Plaintiffs Reply Memorandum, p. 76.

Tr. at 8.

See, e.g., Asociacion Colombiana de Exportadores de Flores v. United States, 13 CIT 13, 16, 704 F.Supp. 1114, 1117 (1989).

See, e.g., Universal Camera Corp. v. NLRB, 340 U.S. 474, 477 (1951); Matsushita Electric Indus. Co. v. United States, 750 F.2d 927, 933 (Fed.Cir. 1984); Timken Co. v. United States, 12 CIT 955, 962, 699 F.Supp. 300, 306 (1988), aff'd, 894 F.2d 385 (Fed.Cir. 1990).

Punitive rates are the result of rejection of low-margin information in favor of high-margin information that is demonstrably less probative of current conditions. Rhone Poulenc, Inc. v. United States, 899 F.2d 1185, 1190 (Fed.Cir. 1990).

See Plaintiffs Memorandum, p. 127. According to USIMINAS, had dumping margins been calculated on a price-to-price basis, they would have been 0.00% for cut-to-length plate, 0.31% for cold-rolled steel and 13.98% for hot-rolled products. See Plaintiffs Appendix, Exhibit 30, p. 2.

Domestic Steel Producers’ Memorandum in Support of Motion for Judgment Under Rule 56.2, p. 18.

The intervenor-defendants describe these characteristics as follows:

[P]roducts of “non-rectangular cross-section” are rectangular carbon steel flat products whose edges have been cut or bevelled to meet customer specifications. Bevelledplate refers to abroad spectrum of rectangular plate products that are processed to meet specific customer requirements. * * * Petitioners refer to products of non-rectangular cross-section as “bevelledplate.”

Domestic Steel Producers’ Opposition Memorandum, pp. 127-28 n. 316.

Defendants’ Memorandum, p. 109 (emphasis in original).

See 3 Harmonised Commodity Description & Coding System § X\£ 11-72.08 (1993).