Opinion
Restani, Judge:
In this action, CEMEX, S A. (“CEMEX”) and the Ad Hoc Committee of AZ-NM-TX-FL Producers of Gray Portland Cement and National Cement Company of California (collectively “petitioners”) contest the final results of the second administrative review in Gray Portland Cement and Clinker from Mexico, 58 Fed. Reg. 47,253 (Dep’t Comm. 1993) (final admin, review), by the International Trade Administration of the United States Department of Commerce (“Commerce”).
Facts
CEMEX is a Mexican manufacturer and exporter of gray portland cement, the primary component of concrete, and clinker, an intermediate material product of the cement manufacturing process used only to be ground into finished cement. On August 30,1990, Commerce issued an antidumping duty order against CEMEX in Gray Portland Cement and Clinker from Mexico, 55 Fed. Reg. 35,443 (Dep’t Comm. 1990) (anti-dumping duty order).1 The second administrative review of the anti-dumping order was commenced on September 28, 1992, for the period between August 1, 1991 and July 31, 1992. In its questionnaire [588]*588responses, CEMEX reported U.S. sales of Types II and V cement and home market sales of Type I, Type II and Type V cement.2
Following requests from petitioners, Commerce initiated a “fictitious markets” and cost of production (“COP”) investigation concerning CEMEX’s sales.3 In its preliminary determination, Commerce found that CEMEX’s Types II and V sales did not constitute a fictitious market. Commerce also determined that sales of Type II cement were made below cost, thus constructed value was used to determine FMV Gray Portland Cement and Clinker from Mexico, 58 Fed. Reg. 33,071, 33,072-73 (Dep’t Comm. 1993) (prelim, admin, results). From June 28 until July 2, 1993, Commerce conducted a verification in Mexico of CEMEX’s submissions regarding the fictitious markets and COP issues. Following verification, petitioners reiterated arguments as to the fictitious markets and COP issues, and further claimed that CEMEX’s Types II and V sales were outside the ordinary course of trade.
On September 8, 1993,. Commerce published the final results for the second administrative review. 58 Fed. Reg. 47,253. Commerce determined that CEMEX’s sales of Types II and V cement in Mexico were outside the ordinary course of trade.4 As a result, Commerce found it unnecessary to reach a conclusion as to the creation of fictitious markets or the COP investigations. Having rejected CEMEX’s home market sales of Types II and V cement, Commerce used constructed value to determine FMV
Petitioners and CEMEX initiated actions in this court challenging Commerce’s final results. The actions were consolidated by the court on November 23,1993. CEMEX raises three categories of claims: (1) Commerce erred in determining that CEMEX’s home market sales of Types II and V cement were outside the ordinary course of trade, (2) Commerce erred in calculating CEMEX’s FMV based on constructed value, and (3) Commerce erred in calculating U.S. price (“USP”). Petitioners contend that (1) Commerce erred in basing FMV on constructed value, rather than on sales of Type I cement, (2) Commerce erred in the cal[589]*589culation. of constructed value, and (3) Commerce erred in the calculation of CEMEX’s dumping margin.
Standakd of Review
In reviewing final determinations in antidumping duty investigations, the court will hold unlawful those determinations by Commerce found to be 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 that which ‘“a reasonable mind might accept as adequate to support a conclusion.’” Matsushita Elec. Indus. Co. v. United States, 750 F.2d 927, 933 (Fed. Cir. 1984) (quoting Consolidated Edison Co. v. NLRB, 305 U.S. 197, 229 (1938)).
Discussion
I. CEMEX’s Contentions:
A. Ordinary Course of Trade:
The relevant statutory provision defines the term “ordinary course of trade” as
the 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) (current version at 19 U.S.C.A. § 1677(15) (West Supp. 1995)); see 19 C.ER. § 353.46(b) (1992). Determining whether home market sales are in the ordinary course of trade requires evaluating not just ‘“one factor taken in isolation but rather * * * all the circumstances particular to the sales in question.’” Murata Mfg. Co. v. United States, 820 F. Supp. 603, 607 (Ct. Int’l Trade 1993) (quoting Certain Welded Carbon Steel Standard Pipes and Tubes from India, 56 Fed. Reg. 64,753, 64,755 (Dep’t Comm. 1991) (final)). Furthermore, an analysis of these factors should be guided by the purpose of the ordinary course of trade provision, namely “to prevent dumping margins from being based on sales which are not representative” of the home market. See Monsanto Co. v. United States, 12 CIT 937, 940, 698 F Supp. 275, 278 (1988).
Relevant circumstances surrounding CEMEX’s sales of Types II and V cement may include, among other factors, home market demand, relative volume as compared to Type I sales, the duration of production, shipping arrangements, relative profit margin as compared to Type I sales, and purpose. See generally Mantex, Inc. v. United States, 841 F. Supp. 1290, 1295, 1306-07 (Ct. Int’l Trade 1993); Timken Co. v. United States, 852 F. Supp. 1122, 1127 (Ct. Int’l Trade 1994); Murata, 820 F. Supp. at 607. These factors will be addressed in turn.
1. Home market demand:
Ordinarily, home market demand is one factor in the ordinary course of trade analysis. See Mantex, 841 F. Supp. at 1307 (reasoning that [590]*590“[although marginal demand for a product does not by itself indicate sales are outside the ordinary course of trade * * * such a factor is probative of whether sales ‘have been normal in the trade’”) (quoting 19 U.S.C. § 1677(15)). CEMEX established, and Commerce agreed, that a legitimate home market demand for Type II and V cement existed in Mexico. 58 Fed. Reg. at 47,255. Thus, lack of home market demand is not a factor supporting a finding that Type II and V cement were outside the ordinary course of trade.
In the final determination, however, Commerce stated that this finding “is not relevant to the issue of whether these sales of cement are within CEMEX’s ordinary course of trade in the home market.” Id. CEMEX contends that Commerce erred in finding home market demand “not relevant.” This is a poor choice of words. The fact is that Commerce concluded that CEMEX was not meeting the home market demand through sales that were within its ordinary course of trade, as will be explained.5
2. Volume of home market sales:
Commerce considers volume of home market sales as another factor in determining whether such sales are in the ordinary course of trade. See East Chilliwack Fruit Growers Co-Operative v. United States, 11 CIT 104, 106-07, 655 F. Supp. 499, 502-03 (1987) (upholding Commerce’s inclusion of substantial sales of bulk-packed fruit within ordinary course of trade, despite Co-op’s primary focus on processing juice and concentrate). In the present case, Commerce properly considered as a relevant factor that CEMEX’s sales of Types II and V cement in Mexico were a small portion of its total sales of cement. Commerce has previously looked to low sales volume as one of various factors in determining ordinary course of trade. See Mantex, 841 F. Supp. at 1307-08 (noting that sales volume was low both in isolation and relative to overall sales, thus supporting Commerce’s finding that sales were outside ordinary course of trade). While CEMEX’s Types II and V sales are low only with respect to Type I sales, the sales volume is significant in absolute terms. Pl.’s Mem. Supp. Mot. J. Agency R. at 44 n.7. Therefore, CEMEX’s relatively low volume of Types II and V cement sales must be assessed in relation to other factors.
[591]*5913. Sales patterns:
CEMEX began to sell Types II and V cement in Mexico during the mid-1980s, when it began exporting to the United States. CEMEX claims Commerce should have considered the sales history of Tolteca, which CEMEX acquired in 1989 and which has been selling cement in Mexico since 1964. Pl.’s Mem. Supp. Mot. J. Agency R. at 45-46.6 Alternatively, CEMEX claims that cement sales for seven years prior to Commerce’s review constitutes a reasonable period of time.7 Id.
Despite CEMEX’s long history in the cement business, CEMEX’s sales of Types II and V cement in Mexico commenced only after production for export markets began. The court does not believe Commerce overweighed this factor, but as with the other factors, it formed the background against which Commerce considered the key element, shipping arrangements.
4. Shipping arrangements:
Commerce focused on CEMEX’s shipping arrangements and absorption of high freight costs. It is not unusual for Commerce to consider shipping arrangements in determining whether sales are outside the ordinary course of trade. See Porcelain-On-Steel Cooking Ware from Mexico, 51 Fed. Reg. 36,435, 36,437 (Dep’t Comm. 1986) (final). CEMEX argues that its shipping arrangements for Types I, II and V cement are similar because the company typically absorbs all freight costs. Pl.’s Mem. Supp. Mot. J. Agency R. at 50-51. The statute, however, focuses not on the company’s similarity of product treatment but on whether the treatment of the particular product at issue, here Types II and V cement, is “normal in the trade.” See 19 U.S.C. § 1677(15). Thus, with regard to Type II and V cement, Commerce determined that shipments over great distances were abnormal and that absorption of the resulting enormous shipping costs was also abnormal.
CEMEX contends that the shipping arrangements are mandated by the geographic location of the cement-producing facilities, and that it could not compete with Apasco, S.A. de C.V, in the Mexico City and Guadalajara area, if it did not absorb the costs. Pl.’s Mem. Supp. Mot. J. Agency R. at 48; Statements at Oral Argument. CEMEX, however, consolidated production of Types II and V cement to one plant only after the issuance of the antidumping order. Pet’rs’ Br. Resp. Pl.’s Mot. J. Agency R. at 38. Ostensibly, it intended to serve more efficiently the U.S. market for Type II cement. Nonetheless, it is apparent that CEMEX has reduced its Mexican Types II and V sales so as to diminish the impact of the absorption of the abnormally-high freight rates. Whatever the real strategy behind the consolidation in the North, the result was an abnormal shipping arrangement for Types II and V cement, which weighs heavily in favor of a finding of sales made outside the ordinary course of [592]*592trade. Shipping arrangements are closely related to the next factor, profitability.
5. Profitability:
Commerce properly compared the relative profitability of CEMEX’s Types II and V cement sales. “Profit level is clearly one of several significant ‘conditions’” in assessing whether home market sales are outside the ordinary course of trade. See Mantex, 841 F. Supp. at 1308 (quoting 19 U.S.C. § 1677(15)). Furthermore, “sales do not have to be below cost in order for a profit margin to diverge from ‘the conditions and practices which * * * have been normal in the trade under consideration’” Id. (same).8 As CEMEX has, on a comparative basis, very low profits for these types of cement, this factor also weighs against finding home market sales of Types II and V cement as within the ordinary course of trade.
6. Corporate image:
Commerce concluded that the promotional quality of Types II and V cement sales supported its finding that sales were outside the ordinary course of trade. Commerce has previously considered the promotional quality of sales in other cases involving an ordinary course of trade analysis. See Sulfur Dyes, Including Sulfur Vat Dyes, from the United Kingdom, 58 Fed. Reg. 3,253, 3,256 (Dep’t Comm. 1993) (final) (finding that sales under “special agreement” to promote product at issue noticeably different from respondent’s other sales, thus supporting determination of sales outside ordinary course of trade).-
Although CEMEX’s sales of Types II and V cement did not involve a “special agreement,” they were either designed to maintain an artificially low FMV or were intended legitimately to promote other sales. During Commerce’s fictitious market investigation, a CEMEX executive stated that CEMEX continued Types II and V cement sales despite their low volume in order to encourage large quantity purchases of Type I cement. See Def.’s Mem. Partial Opp’n Mots. J. Upon Admin. R., App. 4, at 6.
CEMEX contends that these sales of Types II and V cement are within the ordinary course of trade because they were not special sales, for example, sales of second quality or non-standard merchandise. Pl.’s Mem. Supp. Mot. J. Agency R. at 60. See Aramid Fiber Formed of Poly-Phenylene Terephthalamide from the Netherlands, 58 Fed. Reg. 65,699, 65,700 (Dep’t Comm. 1993) (prelim.) (sales of second quality merchandise outside ordinary course of trade); Certain Circular Welded Carbon Steel Pipes and Tubes from Thailand, 59 Fed. Reg. 60,128, 60,128 (Dep’t Comm. 1994) (prelim.) (sales of non-standard merchandise outside ordinary course of trade); see also Monsanto, 12 CIT at 940, 698 F. Supp. at [593]*593278 (sales of obsolete merchandise ordinarily “not appropriate for comparison purposes”). CEMEX’s sales, however, were at a lower volume and profit level than the majority of its other cement sales and if made for legitimate reasons, were made to induce their customers’ continued business in more profitable areas. Such sales may be considered “inappropriate for comparison purposes,” that is, as non-representative of CEMEX’s ordinary course of trade, in the circumstances described here.
In sum, ordinary course of trade is determined on a case-by-case basis by examining all of the relevant facts and circumstances. Of the several factors examined by Commerce, only one, home market demand, is in CEMEX’s favor. As CEMEX is obviously not tryingto meet this demand, except in very limited circumstances, this factor is not of great importance. Other factors such as unusual shipping arrangements minuscule, (in relative terms) volume of sales, and very low profit margins weigh against CEMEX. Commerce’s determination that CEMEX’s sales were outside the ordinary course of trade involved a weighing of data and is supported by substantial evidence.
B. Home Market Credit Expenses, Discounts and Rebates9:
CEMEX also alleges Commerce failed to adjust FMV to account for differences in circumstances of sales, as authorized by 19 U.S.C. § 1677b(a)(4)(B) (1988) (current version at 19 U.S.C.A. § 1677b(a) (6)(C)(iii) (West Supp. 1995), such as home market credit expenses, discounts and rebates. Claiming to have failed inadvertently to calculate constructed value without these deductions, Commerce requests that the issue be remanded for recalculation of constructed value. Def.’s Mem. Partial Opp’n Mots. J. Upon Admin. R. at 33. Petitioners argue that CEMEX is not entitled to the requested COS adjustments to FMV because it failed to request these adjustments in a timely manner. Pet’rs’ Br. Resp. Pl.’s Mot. J. Agency R. at 56-60.
Under the statute, Commerce may correct ministerial errors in final determinations.10 19 U.S.C.A. § 1675(h) (West Supp. 1995) (amending 19 U.S.C. § 1675(f) (1988)). Petitioners argue that Commerce’s oversight was not a ministerial error but a substantive mistake. Pet.’rs’ Br. Resp. Pl.’s Mot. J. Agency R. at 59. Commerce is given fairly broad discretion to determine which types of unintentional error to regard as ministerial. See 19 U.S.C.A. § 1675(h); see also NTN Bearing Corp. v. United States, 826 F. Supp. 1435, 1438 (Ct. Int’l Trade 1993). This error is within the range of what may reasonably be considered ministerial. Therefore, Commerce may choose to consider the failure to make COS adjustments a ministerial error.
[594]*594C. Recalculation of Profit:
Commerce must include an adjustment to constructed value to account for profit. 19U.S.C. § 1677b(e)(l)(B) (1988) (current version at 19 U.S.C.A. § 1677b(e)(2)(A) (West Supp. 1995)). The relevant statute requires that the amount of profit included in constructed value must not be less than eight percent of the COP together with general expenses. Id. § 1677b(e)(l)(B)(ii) (current version at 19 U.S.C.A. § 1677b(e)(2) (B)(iii) (West Supp. 1995)). CEMEX claims that since Commerce erred in calculating COR application of the profit percentage resulted in overstating the amount of profit figured into constructed value, and the resulting FMV Pl.’s Mem. Supp. Mot. J. Agency R. at 68-69.11
Petitioners contend that CEMEX’s claim is not raised in a timely manner and that it should have brought this issue to Commerce’s attention before the issuance of the final results. Pet.’rs’ Br. Resp. Pl.’s Mot. J. Agency R. at 62-63. The failure to exhaust administrative remedies is an affirmative defense that must be presented in the pleadings. Manifattura Emmepi S.p.A. v. United States, 16 CIT 619, 621, 799 F. Supp. 110, 112-13 (1992). The failure to plead an affirmative defense results in a waiver of the defense. United States v. Atkinson, 6 CIT 257, 259, 575 F. Supp. 791, 793 (1983). Therefore, petitioners have waived any claim regarding CEMEX’s failure to exhaust administrative remedies and defendant, on remand, may recalculate profit, as it requests.
D. Imputed Limestone Depletion Expense:
In establishing the FMV of merchandise based upon constructed value, Commerce looks to those costs “which would ordinarily permit the production of that particular merchandise in the ordinary course of business.” 19 U.S.C. § 1677b(e)(l)(A). Thus, Commerce, in calculating constructed value, must arrive at the ordinary or typical cost, in the country of export, of producing the article subject to investigation.
In calculating COR Commerce is to “use the firm’s expenses as recorded in its financial statements as long as those statements are prepared in accordance with the home country’s generally accepted accounting principles (“GAAP”) and do not significantly distort the firm’s financial position or actual costs.” NTN Bearing Corp., 826 F. Supp. at 1441 (quoting IPSCO, Inc. v. United States, 12 CIT 1128, 1130 n.3, 701 F. Supp. 236, 238 n.3 (1988)). Commerce, however, will not use a foreign GAAP that will not reflect accurately the company’s actual costs. Id. at 1441-42.
CEMEX contends that Mexican GAAP applies and that it properly reflects cost. Pl.’s Mem. Supp. Mot. J. Agency R. at 70-71; Statements at Oral Argument. Mexican GAAP does not require companies to account for mine depletion expenses. CEMEX contends further that the limestone depletion should not be reflected in constructed value, [595]*595because limestone is plentiful in Mexico and any depreciation would be insignificant.
Commerce, however, decided to utilize depletion costs calculated by CEMEX in accordance with U.S. GAAE12 See 58 Fed. Reg. at 47,257. Commerce believed that “following Mexican GAAP would result in an inaccurate measurement of cost since the mine depletion expenses would not be included as part of COE yet all costs incurred must be reflected in COP.” Id. Commerce, in other situations has rejected foreign GAAE13 The court has upheld Commerce’s policy of not using the GAAP of a foreign country when to do so would distort the company’s actual costs. See, e.g., NTN Bearing Corp., 826 F. Supp. at 1442; Camargo Correa, Slip. Op. 93-163, at 6. CEMEX would have incurred the mining depletion costs, whatever their amount, regardless of whether or not the company had to include the costs in Mexican GAAE Even if these costs were not recorded, it is “irrelevant to whether these costs are actually incurred and should be included in [CEMEX’s] cost of production.” NTN Bearing Corp., 826 F. Supp. at 1442. Thus, as a general proposition, Commerce properly considered limestone depletion expenses in calculating CEMEX’s costs.14
CEMEX further contends that Commerce incorrectly calculated the limestone depletion expense by (1) basing its calculation on “actual” reserves rather than on “probable” reserves, and (2) estimating the residual value of the land following mineral extraction at zero, based on an alleged assumption by CEMEX. Defendant and petitioners counter that Commerce used the depletion expense calculated by CEMEX from its company records.
Commerce may exercise discretion in evaluating the evidence presented by CEMEX. It is clearly within Commerce’s “discretion to make reasonable interpretations of the evidence and to determine the overall significance of any particular factor or piece of evidence.” Maine Potato Council v. United States, 9 CIT 293, 300, 613 F. Supp. 1237, 1244 (1985). Further, “[v]alue determinations made in antidumping cases ‘must be based upon proof of actual costs of prices — not estimates, approximations or averages.’” IPSCO, Inc. v. United States, 12 CIT 384, 391, 687 F. Supp. 633, 639 (1988) (quoting F.W. Myers & Co. v. United States, 376 F. Supp. 860, 873 (Cust. Ct. 1974)). Thus, Commerce’s decision to use evidence of proven reserves when calculating the depletion expense, despite CEMEX’s citation to potential reserves, was not unreasonable.
[596]*596As to CEMEX’s second argument, the record does not clearly indicate, as Commerce contends, that CEMEX expressly relied upon the assumption that the residual land value should.be zero following extraction of the limestone. In fact, during verification, Commerce noted that “CEMEX’s experience has been that the land increases in value once the company has leveled the land.” Pet’rs’ Apps. to Br. Resp. Pl.’s Mot. J. Agency R., App. 54, at 13 (Commerce’s verification report). Commerce, however, relied upon the calculation provided by CEMEX, the submission of which apparently was intended to demonstrate “that — even under most adverse and unrealistic assumptions — [CEMEX’s] natural resource depletion expense would be negligible.” CEMEX’s Reply Br., Tab 3, at 48.
Nevertheless, the court finds that CEMEX should have submitted the appropriate calculation, as per Commerce’s request, for a per-ton estimate of limestone depletion expenses. Because, generally, mining of natural resources decreases the value of the land, CEMEX needed to substantiate further its claim on this point and to submit a calculation that accurately reflected its expenses. CEMEX misinterprets prior court decisions regarding burdens of proof that properly “should belong to the party in possession of the necessary information.” Zenith Elecs. Corp. v. United States, 988 F.2d 1573, 1583 (Fed. Cir. 1993); see U.S. Steel Group v. United States, 873 F. Supp. 673, 691 (Ct. Int’l Trade 1994) (noting that this principle “applies in ITA cases to such matters as a respondent’s claims to adjustments to its price data, and other matters clearly within a respondent’s knowledge”), appeal docketed, No. 95-1245 (Fed. Cir. Mar. 21, 1995).
Commerce did not have to accept the vague, unsubstantiated statements of CEMEX officials that the depleted land would be more valuable after extraction. Although the court is mindful of the difficulty of providing information during verification, CEMEX did not claim it could not comply. In fact, it did submit a calculation. Permitting CEMEX now to provide “more accurate” data would allow it a second chance at verification, or would allow its new information to escape verification entirely. Accordingly, Commerce’s limestone depletion expense adjustment to COP is sustained.
E. Allocation of Excess Capacity:
As indicated, value determinations should be based upon proof of actual costs or prices, not estimates, approximations or averages. F.W. Myers, 376 F. Supp. at 873. Commerce based its excess capacity expense on a weighted-average of actual costs incurred at each of CEMEX’s plants. 58 Fed. Reg. at 47,256. Commerce refused to pool excess capacity costs incurred by specific plants and allocate these costs to all plants. Id.
Commerce reasoned that “calculating excess capacity at the company-wide level would not reflect the costs incurred to manufacture the products under review.” Id. According to Commerce, following CEMEX’s methodology would have underestimated and overestimated [597]*597excess capacity expenses for certain plants, consequently “shifting these costs to products which did not incur these costs.” Id. It is Commerce’s policy to use actual costs of production of the subject merchandise whenever possible. See, e.g., Fresh and Chilled Atlantic Salmon from Norway, 58 Fed. Reg. 37,912, 37,915 (Dep’t Comm. 1993) (final). Therefore, Commerce could choose to use a plant-specific excess capacity charge in order to link the excess capacity expense to CEMEX’s actual costs. Accordingly, Commerce’s calculation of excess capacity expense is sustained.
F. Sales to C.L. Pharris:
CEMEX contends that Commerce overstated the amount of profit used in the U.S. value-added calculation by failing to use transfer prices between its distributor, Pacific Coast Cement, and C.L. Pharris, a company CEMEX purchased in 1992. Commerce requests a remand to determine whether sales made to Pharris were at arm’s length. The remand request is granted.
G. Other Errors in Calculation of USP:
CEMEX contends that Commerce failed to correct two errors, concerning bad debt expenses and interest income on overdue customer accounts, in the calculation of USP Commerce, however, treated CEMEX’s data as new factual information submitted untimely, and consequently refused to consider it. Specifically, CEMEX, on June 4,1993, sought to correct its submission of accrued bad debt expenses for its Sunbelt Cement subsidiary, due to Commerce’s policy of requiring actual bad debt expense data. As to interest income, CEMEX requested that Commerce treat interest income as an upward adjustment to U.S. price. Such treatment of interest income was permitted in Gray Portland Cement and Clinker from Japan, 58 Fed. Reg. 21,442, 21,442 (Dep’t Comm. 1993) (prelim.). Thus, CEMEX sought to change its submission in this regard.
CEMEX admits that the information was submitted after the expiration of the 180-day time period for submitting factual information following initiation of this review on September 28, 1992 (the deadline being March 27,1993),see 19C.F.R. § 353.31(a)(l)(ii) (1993), however, it argues that the data was not new unsolicited information. Rather, the data was submitted to correct information that was already placed in the administrative record. As the submission was made one month before verification, Commerce should have, according to CEMEX, accepted the data.
In general, the court has upheld Commerce’s rejection of untimely submitted factual information pursuant to 19 C.F.R. § 353.31(a). See, e.g., NSK Ltd. v. United States, 16 CIT 745, 749, 798 F. Supp. 721, 725 (1992). Commerce “is not required to correct a respondent’s errors when a respondent reported erroneous data but failed to timely correct it.” NSK Ltd. v. United States, Slip Op. 93-110, at 6 (June 17, 1993). Although the language of Commerce’s regulation is mandatory, Com[598]*598merce routinely accepts data after deadlines depending on the circumstances of each case. Thus, the court must determine whether Commerce “abused the wide discretion afforded to it in enforcing the timing provisions of its regulations.” Böwe-Passat v. United States, Slip Op. No. 93-68, at 5 (May 7, 1993).
As previously noted, the basis for rejecting CEMEX’s corrected information was that it was untimely “new factual information.” 58 Fed. Reg. at 47,258. The court agrees that the bad debt data was “new” and need not be accepted.15 As to the second item of information, CEMEX alleges that it was alerted to a change in Commerce’s policy with regard to interest income on April 6, 1993, after reading a Commerce memorandum. This policy was later articulated in a preliminary determination issued April 21, 1993. CEMEX also asserts that in reliance upon the old policy, it did not request the adjustment.
If Commerce has articulated a new policy with regard to the adjustment to be made for interest income, Commerce may have abused its discretion in not permitting CEMEX to claim such an adjustment, where CEMEX submitted the relevant information one month before verification and less than two months after Commerce revealed its alleged new policy. On remand, Commerce should examine whether it changed policies at a crucial time juncture. If so, CEMEX’s interest income data should be accepted.
II. Petitioners’ Contentions:
A. Use of Such or Similar Merchandise for Price Comparisons:
Commerce was required to calculate FMV pursuant to 19 U.S.C. § 1677b(a)(l)(A) (1988) (current version at 19 U.S.C.A. § 1677b(a)(l) (West Supp. 1995)), which provided that:
(1) In general
The foreign market value of imported merchandise shall be the price * * *
(A) at which such or similar merchandise16 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 * * *.
[599]*599Id. In the absence of identical merchandise, Commerce must ‘“choose the most similar merchandise’ for comparison.” Hussey Copper, Ltd. v. United States, 834 F. Supp. 413, 417 (Ct. Int’l Trade 1993) (quoting Timken Co. v. United States, 10 CIT 86, 96, 630 F. Supp. 1327, 1337 (1986)). Commerce may use constructed value “if [it] determines that the [FMV] of imported merchandise cannot be determined under paragraph (1)(A),” that is, on the basis of home market sales of such or similar merchandise. See 19 U.S.C. § 1677b(a)(2).
In the final determination, Commerce found sales of Type II and Type V cement to be outside the ordinary course of trade. According to Commerce, the next most similar merchandise for determining FMV was Type I cement. Commerce, however, did not request, and CEMEX did not provide, “difference in merchandise” (“difmer”) information that “would permit an accurate comparison of home market sales of Type I cement to U.S. sales of Type II and V cement.” 58 Fed. Reg. at 47,255. Thus, Commerce based FMV on the constructed value of Types II and V cement. Id.
There is a statutory preference for the use of similar merchandise in determining FMV. NTN Bearing Corp. v. United States, 14 CIT 623, 633, 747 F. Supp. 726, 735 (1990). Constructed value should only be used where Commerce has made a determination that “the exporter’s home market prices are inadequate or unavailable for the purpose of calculating FMV.” See H.R. Rep. No. 317, 96th Cong., 1st Sess. 76 (1979). Here, Commerce admits that Type I cement was similar merchandise upon which to base FMV and, that the only basis for excluding Type I cement sales was that difmer information was neither requested by Commerce nor provided by CEMEX. Although Commerce found that an “accurate comparison” could not be made without the difmer information, no determination was made as to whether the difmer information would have corrected the deficiency.
Under the statute, if similar, rather than identical, merchandise is used in determining FMV Commerce is required to make allowance for differences in merchandise to the extent that Commerce is satisfied that the amount of any price differential is wholly or partly due to such differences. See 19 U.S.C. § 1677b(a)(4)(C); 19 C.F.R. § 353.57(a) (1993); Hussey Copper, 834 F. Supp. at 427. Commerce’s failure to request such information was unreasonable and inconsistent with the statute’s purpose of making an “apples to apples” comparison.17 See NTN Bearing Corp., 14 CIT at 639, 747 F. Supp. at 740 (“Section 1677b(a)(2) sanctions the use of constructed value only when the ITA determines that the foreign market value of the imported merchandise cannot be determined [600]*600by sales of ‘such or similar’ merchandise in the home market.”).18 According to the view of Commerce expressed here, Commerce would never ask for difmer information or use prices of similar merchandise, because it would only ask for difmer information after it decided to reject the identical category of comparison merchandise, at which time it would be too late to seek difmer information. This approach effectively rewrites the statute. Thus, as the court has sustained Commerce’s finding that sales of Types II and V cement are outside the ordinary course of trade, Commerce shall analyze19 difmer information to determine whether sales of Type I cement are suitable for comparison to U.S. sales of Types II and V cement.
B. Fixed Factory Overhead:
In calculating constructed value, Commerce requested quarterly cost data covering July 1, 1991 through June 30, 1992, the period most closely corresponding to the period of review, August 1, 1991 through July 31,1992. Thus, CEMEX reported quarterly cost data for 1991 and three quarters of 1992. 58 Fed. Reg. at 47,255-56. Petitioners contend that Commerce understated constructed value by failing to annualize CEMEX’s fixed factory overhead costs for 1992. By excluding CEMEX’s fourth quarter cost data for 1992, or failing to adjust the data for the first two quarters in accordance with such new data, CEMEX’s COP data for the period of review was, according to petitioners, grossly understated. Petitioners also argue that Commerce’s methodology here is inconsistent with the approach taken in the first administrative review. See Gray Portland Cement and Clinker from Mexico, 58 Fed. Reg. at 25,807 (“CEMEX contends that [Commerce] should calculate CEMEX’s cost of production on an annual basis rather than a monthly basis * * *. [Commerce] agreefs] * * *. We have calculated and used annual 1990 and 1991 cost of production figures in our final results.”).
Defendant and CEMEX counter that the cost data was appropriately based on a contemporaneous time period to the actual period of review. See 19 U.S.C. § 1677b(a)(l) (directing calculation of FMV “at the time such merchandise is first sold within the United States”). They also contend that the calculation of COP was not inconsistent with the first administrative review, where two COP figures were used (one for the review months in 1990 and another for review months in 1991). This was apparently followed in the administrative review at issue. See Gray Portland Cement and Clinker from Mexico, 58 Fed. Reg. at 33,072 (prelim.). Finally, although defendant recognized that it had annualized cost [601]*601data in previous determinations, such action generally has been limited to investigations in which the period of review was for less than 12 months and annualizing costs was necessary to prevent distortion due to seasonal business cycles. See Nihon Cement Co. v. United States, Slip Op. 93-80, at 25-27 (May 25, 1993) (sustaining Commerce’s use of annualized fixed costs given seasonal fluctuations).
Here, the review period covered 12 months, thus seasonal concerns were not present, but the court is concerned that the data used was inaccurate for other reasons. If data for the entire year of 1992 indicates that there was a serious understatement of costs for the first two quarters, this error is to be corrected.20
C. Calculation of the Dumping Margin:
Petitioners contend that due to a computer error, Commerce inadvertently based CEMEX’s dumping margin solely on the margin calculated for cement sales, without including the separate margin on ready-mixed concrete sales. Defendant and CEMEX agree that remand is appropriate to correct this error.
Conclusion
The court sustains Commerce’s determination as to the ordinary course of trade issue, but remands this matter for collection of difmer data and consideration of Type I cement sales prices for calculation of FMV In addition, Commerce’s limestone depreciation adjustment to COP is sustained. The court also sustains Commerce’s excess capacity adjustment. The court conditionally remands the issues of profit calculation and 1992 fixed factory overhead and ministerial errors in the COS adjustment. As to the calculation of USI] the court sustains Commerce’s rejection of bad debt expense information as untimely, but remands for reconsideration rejection of CEMEX’s interest income data and the Pharris transfer pricing issue. The court also orders remand of the dumping margin calculation to rectify inadvertent errors by Commerce.
Remand results are due within 60 days hereof. Any objections are to be filed within 20 days thereafter and responses are due 10 days thereafter.