Thai Pineapple Public Co. v. United States

' Opinion* Restani, Judge: This matter is before the court on a motion for judgment on the agency record by plaintiffs, The Thai Pineapple Public Co., Ltd. (“TIPCO”), Malee Sampran Factory Public Co., Ltd. (“Malee”), and Siam Agro Industry Pineapple and Others Public Co., Ltd. (“SAICO”), (collectively the “Thai plaintiffs”), and Siam Food Products Public Co., Ltd. (“Siam Food”); by plaintiff-intervenors, Dole Food Company, Inc., Dole Packaged Foods Co., and Dole Thailand, Ltd., (collectively “Dole”); and by defendant-intervenor, Maui Pineapple Co., Ltd. (“Maui”) pursuant to USCIT Rule 56.2. The administrative determination under review is the final antidumping duty determination of the United States Department of Commerce (“Commerce”) entitled Canned Pineapple Fruit [“CPF”] From, Thailand, 60 Fed. Reg. 29,553 (Dep’t Comm. 1995) (final determ. of LTFV sales) [hereinafter “Final Det.”], as amended, 60 Fed. Reg. 36,775 (Dep’t Comm. 1995). The Thai plaintiffs contend that Commerce erred in using their normal accounting systems for allocation of raw material fruit qosts for purposes of calculating cost of production (“COP”) and constructed value (“CV”) and in rejecting their weight-based allocation methodologies. Dole also claims that Commerce unlawfully rejected Dole’s weight-based fruit cost allocation and, in addition, improperly substituted best information available (“BIA”), used arbitrary weighting factors in calculating the weight averaged dumping margins, improperly calculated Dole’s indirect expenses, and failed to fully adjust for all expense differences. Maui argues that Commerce’s decision to treat Dole’s sale to Germany as outside the ordinary course of trade is contrary to law and unsupported by substantial evidence. — 1313 — In response, Commerce argues that all of its contested determinations were reasonable, supported by substantial evidence, and otherwise in accordance with law. Commerce does, however, request that the court remand this matter to Commerce so that it may use the same time period for weighting the dumping margin for all Dole products and for Commerce to instruct the United States Customs Service (“Customs”) to assess antidumping duties only upon Dole CPF from Thailand that was entered, or withdrawn from warehouse for consumption on or after February 22,1995, the date of publication of the amended preliminary less than fair value (“LTFV”) determination relating to CPF from Thailand. Commerce’s latter request for a remand is not opposed and is granted. Background On June 8, 1994, Maui and others petitioned Commerce and the United States International Trade Commission (“ITC”), alleging that the CPF from Thailand was being sold at LTFV Based upon the petition, Commerce initiated an antidumping investigation. Canned Pineapple Fruit From Thailand, 59 Fed. Reg. 34,408 (Dep’t Comm. 1994) (initiation of investigation). On January 11, 1995, Commerce found preliminary dumping margins for Malee, TIPCO, and SAICO ranging from 1.12 to 9.55 percent. Canned Pineapple Fruit From Thailand, 60 Fed. Reg. 2734, 2738 (Dep’t Comm. 1995) (prelim, determ.) [hereinafter “Prelim. Det. ”]. Commerce found a de minimis dumping margin for Dole of .30 percent. Id. On February 22, 1995, Commerce revised the preliminary dumping margin for Dole from .30 to .78 percent because Commerce found a significant ministerial error in its preliminary calculation. Canned Pineapple Fruit From Thailand, 60 Fed. Reg. 9820 (Dep’t Comm. 1995) (amended prelim. determ.). In February and March 1995, Commerce conducted verification and in June 1995, Commerce issued its final determination finding dumping margins for Dole, TIPCO, Malee, and SAICO ranging from 2.36 to 55.77 percent. Final Det., 60 Fed. Reg. at 29,571. On July 10, 1995, the ITC notified Commerce of its final affirmative injury determination, and on July 18, 1995, Commerce published an amended final determination and the antidumping duty order. Canned Pineapple Fruit From Thailand, 60 Fed. Reg. 36,775 (Dep’t Comm. 1995) (amended final determ. and order) [hereinafter “Amended Final Det. ”]. The revised final dumping margins for Dole, TIPCO, Malee, and SAICO, after correction of ministerial errors in the final determination, range from 1.73 to 51.16 percent. Id. at 36,776. — 1314 — Standard of Review The court must sustain a final determination of Commerce unless it is “unsupported by substantial evidence on the record, or otherwise not in accordance with law.” 19 U.S.C. § 1516a(b)(1)(B) (1988).1 Discussion I. Dole’s German Sale: In its preliminary margin calculations for Dole, Commerce matched all United States sales of two products (i.e., CONNUMUs 1211 and 1231)2 to a single sale of a “similar” German product3 (i.e., CONNUMT 1142). Final Concurrence Memorandum at 11 [hereinafter “Final Concur. Mem. ”]; C.R. Doc. 154; Defs.’ Conf. App., Ex. 10. The United States sales of the two products together accounted for about one percent by volume of Dole’s total United States database during the POI; the single German sale of the one product accounted for less than 0.01 percent by volume of Dole’s total German database during the POI. Id. This accounted for over 90 percent of Dole’s dumping margin. Id. Dole objected and argued that the single German sale was outside the ordinary course of trade. Commerce agreed with Dole and excluded the sale from the calculation of FMV. Final Det., 60 Fed. Reg. at 29,562-63. Commerce calculates FMV by following the general method prescribed by 19 U.S.C. § 1677b(a)(1)(A) (1988): The foreign market value of imported merchandise shall be the price, at the time such merchandise is first sold within the United States by the person for whom (or for whose account) the merchandise is imported to any other person * * * (A) at which such or similar merchandise is sold, or, in the absence of sales, offered for sale in the principal markets of the country from which exported, in the usual commercial quantities and in the ordinary course of trade for home consumption Id. § 1677b(a)(1)(A) (emphasis added). Under the statute, [t]he term “ordinary course of trade” means the conditions and practices which, for a reasonable time prior to the exportation of the merchandise which is the subject of the investigation, have been normal in the trade under consideration with respect to merchandise of the same class or kind. Id. § 1677(15) (1988). In determining whether a sale is outside the ordinary course of trade, Commerce must consider “all the circumstances particular to the sales in question.” CEMEX, S.A. v. United States, Slip Op. 95-72, at 6 (Apr. 24, 1995) (quotations omitted); see also Final Det., — 1315 — 60 Fed. Reg. at 29,563. Furthermore, “Commerce, in its discretion, chooses how best to analyze the many factors involved in a determination of whether sales are made with in the ordinary course of trade.” Laclede Steel Co. v. United States, Slip Op. 95-144, at 6 (Aug. 11, 1995). Commerce’s “analysis of these factors [is] 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 [or third country] market.” CEMEX, S.A., Slip Op. 95-72, at 6 (quoting Monsanto Co. v. United States, 12 CIT 937, 940, 698 F. Supp. 275, 278 (1988)). Maui contends that Commerce’s analysis failed to satisfy the legal standard for treating Dole’s German sale as outside the ordinary course of trade. Maui argues that Commerce failed to consider the totality of the circumstances surrounding the sale, as required by the statute, and ignored overwhelming record evidence demonstrating that the sale was not unrepresentative when assessed against other sales in Dole’s third-country sales database, including those prior to the POI. Maui argues that Commerce’s decision was based upon only three factors: volume, price, and demand, which alone are insufficient grounds for finding the sale was outside the ordinary course of trade, and that Commerce erred in failing to consider other factors (e.g., merchandise standards; use of the merchandise; whether the sale was promotional or a trial sale; or whether the price of the sale was specially negotiated). Finally, Maui claims that Commerce wrongly considered only gross unit prices rather than net prices and relied on unverified data. The court finds that Commerce’s decision to exclude Dole’s German sale was reasonable and supported by substantial evidence. In making its final determination, Commerce evaluated eight factors: customers, terms of sale, volume of sales, frequency of sales, sales quantity, sales price, profitability, and market demand. Final Concur. Mem. at 13—14; see also Final Det., 60 Fed. Reg. at 29,563. Commerce found that of the eight factors, only the first two did not support a finding that the sale was outside the ordinary course of trade and agreed with petitioners that the customer and terms of sale were not unique. Final Concur. Mem. at 13. Commerce concluded that the other six factors, however, supported a finding that the sale was outside the ordinary course of trade. Id. at 13-14; see also Final Det., 60 Fed. Reg. at 29,563. In particular, Commerce found that: (1) Dole’s single third country sale of this product constituted an insignificant portion of its total German sale volume; (2) the sale was of a product that was sold only once during the POI; (3) the sales quantity was significantly lower than the average sales quantity for the POI; (4) the sales price was significantly higher than the average sales price charged on other CPF products sold in the same can size during the POI; (5) the profit margin realized by Dole on this particular sale was substantially higher than the weighted-av — 1316 — erage profit earned on other sales of CPF in this can size during the POI; and (6) there was only one customer for this product in the third country market during the POI. Final Det., 60 Fed. Reg. at 29,563; see also Final Concur. Mem. at 13-14. Commerce properly weighed the contrary factors against the six other factors that were in favor of an outside of the ordinary course finding. These six factors have been found previously to be significant in ordinary course of trade analyses. See Laclede Steel Co., Slip Op. 95-144, at 9-11 (court sustained Commerce’s ordinary course of trade decision based upon consideration of sales price, profitability, market demand, terms of sale, sales quantity, sales volume, industry specifications, and special marking of merchandise); see also CEMEX, S.A., Slip Op. 95-72, at 6-14 (court sustained Commerce decision based upon consideration of market demand, sales volume, sales patterns, shipping arrangements, profitability, and corporate image). Moreover, “ordinary course of trade is determined on a case-by-case basis by examining all of the relevant facts and circumstances.” CEMEX, S.A., Slip Op. 95-72, at 13. Maui presents no convincing argument as to why the data relied on by Commerce was insufficient to allow Commerce to assess whether the circumstances surrounding the sale was truly aberrant, other than stating the record contains no evidence that demonstrates that the sale was a “trial sale” or affirmatively establishes the sale as non-representative outside the confines of the POI. The court finds that Commerce sufficiently examined the totality of the circumstances surrounding the German sale by analyzing the six month POI. Commerce determinations, whether in an investigation or an administrative review, are based upon record evidence for that particular segment of the proceeding. Maui further argues that Commerce should have considered net unit price instead of gross unit price as it is agency practice to rely on net unit prices to make “apples to apples” comparisons between transactions regardless of individual sales terms. In terms of net unit prices, Maui claims that the net price of the sale is not markedly different from that of the next most similar merchandise. See Maui’s Mot. For Summary J. on the Agency R. at 21 & Ex. 1. If Commerce had relied on net prices, Maui contends that the sale in question stood less than 30 percent higher than the next most similar merchandise as compared to the more than 80 percent difference in gross prices relied upon by Commerce. Id. at 21. Maui also contends that Commerce’s reliance on gross unit prices distorted the profitability of the German sale. Commerce, however, was making a judgment about the data, not performing a full margin calculation analysis and thus, an examination of net unit price was unnecessary where Commerce considered comparison of gross unit prices sufficient. Commerce is was not required to perform a cost test analysis upon the particular sale in order to analyze the profit of this sale relative to the weighted-average profit earned upon other sales of CPF in the same can size. — 1317 — Finally, Maui argues that the sale at issue was not anomalous in terms of sales quantity because there were two other sales of relatively small quantity and although the sale was to only one customer, other products were also sold to only one customer. The court finds, however, that Commerce correctly analyzed the sale quantity of the German sale relative to the average sales quantity of the remaining sales. It was less than one percent of the normal quantities. Despite the fact that the other sales were made to only one customer, Commerce may still determine that this factor was significant in its overall analysis. “[P]laintiffs have the burden of demonstrating the sales Commerce excluded from its [FMV] calculation were not outside the ordinary course of trade.” Timken Co. v. United States, 18 CIT 486, 492, 852 F. Supp. 1122, 1128 (1994) (emphasis in original). Maui has not met this burden. Accordingly, the court finds that Commerce’s decision to exclude Dole’s German sale from the FMV calculation was supported by substantial evidence and in accordance with law. II. Fruit Cost Allocation Methodologies: Commerce must compute a CV in lieu of FMV if significant sales are below cost of production. 19 U.S.C. § 1677b. The statute provides a formula for calculating CV: [T]he constructed value of imported merchandise shall be the sum of- (A) the cost of materials (exclusive of any internal tax * * *) and of fabrication or other processing of any kind employed in producing such or similar merchandise * * * (B) an amount for general expenses and profit * * * (C) the cost of all containers and coverings of whatever nature * * *. Id. § 1677b(e)(1). Constructed value is thus, defined as, “the combined cost of materials, fabrication or other processing of any kind, general expenses and profit, and other incidental shipping expenses.” IPSCO, Inc. v.. United States, 965 F.2d 1056, 1059 (Fed. Cir. 1992) [hereinafter “IPSCO III”]. “The cost of materials encompasses the cost of raw components in the manufacturing process.” Id. Thus, cost accounting was relevant to both Commerce’s COP and CV decisions. In addition to producing CPF,4 Dole, TIPCO, SAICO, and Malee produce additional pineapple products, such as pineapple juice and juice concentrate, which are not covered by the scope of the investigation. These products are made from the same raw material, fresh pineapple fruit and thus, share jointly the cost of this material. No method for segregating parts of the product and costing them separately has been proposed.5 — 1318 — In the proceeding below, each of the Thai plaintiffs claimed that their normal methodology for allocating the cost of fresh pineapple fruit between CPF and non-CPF co-products resulted in distorted and inappropriate cost of production figures. As a result, Dole, TIPCO, SAICO, and Malee each reported raw material fruit costs using an alternative allocation methodology based upon the output derived weight of the fresh fruit used in the various pineapple-containing products. This methodology resulted in a significantly lower fruit cost allocation to CPF and hence, a lower COP and CV. See Final Concur. Mem. at 23. In its final determination, Commerce rejected the Thai plaintiffs’ alternative weight-based allocation methodology and instead, relied upon the pineapple fruit cost allocations as calculated pursuant to each company’s normal accounting system, with the exception of Dole, because Commerce found that each Thai company’s allocation methodology was consistent with Thai generally accepted accounting principles (“GAAP”) and reasonably reflected the costs associated with production of CPF. See Final Det., 60 Fed. Reg. at 29,559-62. For Dole, Commerce relied upon an average of the fruit cost allocation percentages used by TIPCO, SAICO, and Malee because, although Dole’s allocation methodology was consistent with Thai GAAR Commerce found that it did not reasonably reflect the costs associated with production of CPF. Id. at 29,560. Juice costs were admittedly understated. TIPCO, SAICO, and Malee claim that their normal formulas make no attempt to measure costs accurately and in fact, grossly overallocate costs to CPF. The Thai plaintiffs argue that Commerce ignored the record evidence to this effect, which makes it clear that the allocation bears no resemblance to either a weight or value-based allocation, and Commerce’s use of the Thai plaintiffs normal allocation formulas was not supported by substantial evidence. Furthermore, the Thai plaintiffs assert that by utilizing accounting methodologies that distort costs, Commerce’s decision violated the statutory requirement that COP and CV be based on actual costs and thus, was not in accordance with law. See IPSCO III, 965 F.2d 1056 (requiring weight-based allocation of costs among joint pipe products). The legislative history of the antidumping statute states that “in determining whether merchandise has been sold at less than cost, [Commerce] will employ accounting principles generally accepted in the home market of the country of exportation if [Commerce] is satisfied that such principles reasonably reflect the variable and fixed costs of producing the merchandise.” H.R. Rep. No. 93-571, at 71 (1973). Commerce asserts that its normal practice is to adhere to an individual firm’s recording of costs in accordance with the GAAP of its home country if Commerce is satisfied that such principles reasonably reflect the costs of producing the subject merchandise. Final Det., 60 Fed. Reg. at 29,559. Commerce maintains that the rationale behind this practice is that normal accounting practices should, in most circumstances, provide an objective standard by which to measure costs, while allowing the — 1319 — respondents a predictable basis on which to compute those costs. Id. In those instances in which Commerce determines that a company’s normal accounting practices result in an unreasonable allocation of production costs, however, it will make certain adjustments or use alternative methodologies that more accurately capture the costs incurred. Id. It is not disputed that TIPCO, SAICO, and Malee’s normal fruit cost allocation methodologies were consistent with Thai GAAP TIPCO, SAICO, and Malee argued, however, that their normal allocation methodologies do not reasonably reflect costs, because the methodologies were designed to achieve certain managerial goals as opposed to providing accurate cost information.6 Id. Commerce, however, found that while these reasons may have been factors in the companies’ selection of the methodologies, this does not make the methodologies, or their resulting allocations, unreasonable, particularly where a company’s accounting methodology had been approved by independent auditors. Id. at 29,560. To support this position, Commerce cites Hercules, Inc. v. United States, 11 CIT 710, 673 F. Supp. 454 (1987). In Hercules, the court upheld Commerce’s decision to rely upon COP information from respondent’s normal financial statements maintained in conformity with GAAP 11 CIT at 755, 673 F. Supp. at 490. The respondent, SNPE, had argued that the accelerated depreciation method employed in its financial statements and records was for tax purposes and did not accurately reflect SNPE’S actual costs. Id. Consequently, SNPE submitted recalculated depreciation expenses under a straight-line methodology. Commerce rejected this alternate allocation methodology, which was based upon unverifiable allegations that straight-line depreciation methodology would more accurately reflect actual costs, in favor of the information contained in SNPE’s verified normal records and audited financial statements. 11 CIT at 755-56, 673 F. Supp. at 490-91. Commerce claims that in the present case, as in Hercules, Commerce chose to rely on verified and independently audited normal allocation methodologies, rather than the respondents’ unverified alternative methodologies. Final Det., 60 Fed. Reg. 29,560. Here, however, plaintiffs have demonstrated that the allocation formulas are unrelated to actual cost. The allocation formulas used in the normal course of business do not vary regardless of changes in raw fruit prices, finished goods prices, input materials’ utilization efficiencies, or relative production output. See Hearing Transcript at 34. Only Malee allocates according to some notion of either cost or value. It ¿locates more than 80 percent of fruit costs to solid products and less than 20 percent to juice products. See Malee Sec. D Resp. at 28 (Dec. 23, 1994); C.R. Doc. 56; Pls.’ Conf. App., Ex. 12; see also Cost Verification Report for Malee (Apr. 6, 1995) at 6; C.R. Doc. 118; Pls.’ Conf. App., Ex. 13.7 The — 1320 — cost verification exhibits show that this formula is based on rough estimates, averages, and standard numbers, rather than actual cost or sales records, and like the TIPCO and SAICO approaches it takes no account of fluctuations in relevant costs or prices. See Malee Sec. D Resp. at 28. Commerce explicitly recognized the lack of a factual basis in SAICO’s case. Commerce stated that: In SAICO’s normal accounting system, pineapple fruit costs are allocated on a basis which is purely nominal and has no basis in fact. Cost Verification Report for SAICO at 17 (Apr. 4,1995); C.R. Doc. 111; Pls.’ Conf. App., Ex. 7. The Thai plaintiffs also argue that although they have utilized these allocation formulas for several years, this fact does not make the allocation formulas any less distortive of actual costs for the purposes of anti-dumping law. See Final Det., 60 Fed. Reg. at 29,559 (Commerce “found that each company had used its recorded fruit cost allocation methodology for at least a number of years”). The Thai plaintiffs argue further that Commerce acted inconsistently when it deviated from the Thai plaintiffs’ normal accounting practices on several occasions in calculating COP and CV in the underlying investigation without factually distinguishing why it was appropriate in those occasions but not for fruit costs. Commerce admits that in certain instances (fruit receiving and common processing costs, sugar costs, plantation overhead costs, and direct labor costs) do represent departures from the Thai plaintiffs’ normal accounting records, but states these adjustments were necessary because the companies’ normal records did not provide at all, or provided only minimally, for allocation of these shared costs between CPF and other pineapple products. The court finds that Commerce erred in its determination to rely on the Thai plaintiffs’ normal accounting practices to allocate fruit costs. Whether or not those methodologies were consistent with Thai GAAR the court finds that the methodologies do not reasonably reflect actual costs as required by the statute. The methodologies are arbitrary and reflect neither the cost of the inputs nor any credible measure of the value of the outputs, as Commerce appears to recognize. Although Commerce repeatedly noted the unreliability of the Thai plaintiffs’ allocation methodologies, it continued to employ them simply because they were consistent with Thai’s GAAP Furthermore, there is really no meaningful distinction between Dole’s rejected “normal” methodology and the others which were accepted. Even if Malee’s allocation is a very rough estimate of value it is nonetheless unreliable, and the others differ for arbitrary reasons, not because they are more accurate. Commerce has never stated that foreign GAAP will suffice when it is unreliable. To the contrary, GAAP consistent methodologies are rejected when they do not reflect actual costs. See, e.g., New Minivans from Japan, 57 Fed. Reg. 21,937, 21,946 (Dep’t Comm. 1992) (final determ, of LTFV sales). Maui’s argument that respondents cannot — 1321 — reject their own GAAP consistent records is incorrect. The only issue is whether the records reflect actual costs. GAAP is concerned with overall financial performance and not product specific cost allocation for anti-dumping purposes. Commerce should have followed its normal practice and rejected GAAP accounting records because they were unreliable and distortive of actual costs. Thus, Commerce’s decision to rely on “normal” methodologies presented to it was arbitrary, capricious, not based on substantial evidence and contrary to law. The next question is what methodology or methodologies may be used. Lurking in the background is not Commerce’s desire to use distorted GAAP records, but rather its apparent change of heart regarding proper co-product cost allocation for antidumping purposes. Commerce is not seeking some “canned fruit” exception to the principle of weight/volume based allocation, as it appears to argue here. See discussion infra. It is stepping back from that methodology even in the steel area, where it vigorously supported weight-based allocation, in the past. Cf. IPSCO III (weight-based allocation of joint pipe product costs) and Certain Carbon and Alloy Steel Wire Rod from Canada, 59 Fed. Reg. 18,791, 18,795 (Dep’t Comm. 1994) (final determ, of LTFV sales) (time instead of weight used to allocate steel fabrication costs; “use of tonnage * * * would result in the same cost per ton regardless of the grade of steel”). In this case Commerce has proffered a desire for a value-based allocation as between solids and juice products, although it accepted weight-based allocations among solid products. Final Det., 60 Fed. Reg. at 29,560-61 (rejecting Thai plaintiffs’ weight-based allocations). As indicated, Dole’s “normal” allocation methodology was rejected and an average of Thai company allocation ratios was used. Id. But Commerce stated that for future reviews and assessments Dole’s fruit costs should be allocated between CPF and pineapple juice products on the basis of historical net realizable value (“NRV”). Id. Commerce stated that: A reasonable fruit cost allocation methodology would be one which reflects the significantly different quality of the fruit parts which are used in the production of CPF versus those which are used in the production of juice products. One approach to deriving such an allocation methodology would be to compare the net realizable value of the CPF versus juice products over a period of years. Net realizable value (NRV) is commonly defined as the predicted selling price in the ordinary course of business less reasonable predictable costs of completion and disposal. Id. (citing Charles T. Horngren & George Foster, Cost Accounting: A Managerial Emphasis (5th ed. 1982) at 534 [hereinafter “Cost Accounting”]); see also Final Concur. Mem. at 30-31 & n.11. Commerce declined to utilize the NRV methodology in this investigation based on Dole’s verified sales data from the POI because the resulting allocation percentage differed significantly from the allocation percentages used in the books of the other respondents. Final Det., 60 Fed. Reg. at 29,560. This is not a complete explanation, as will be explained. See also Final — 1322 — Concur. Mem. at 30-32. Commerce also did not obtain historical information in this investigation. The Thai plaintiffs argue that weight is the appropriate basis for allocation because it is on the basis of weight that fruit costs are actually incurred. They claim that they purchase raw pineapple on a Thai baht per kilogram basis and that cost is the same for a kilogram of higher quality raw material used to make choice grade slices as it is for a kilogram of lower quality raw material used to make juice. Commerce, however, stated that, “[b]ecause the parts of the pineapple are not interchangeable when it comes to CPF versus juice production, it would be unreasonable to value all parts equally using a weight-based allocation methodology.” Final Det., 60 Fed. Reg. at 29,561. Commerce likens the situation to that of ahog, where if the joint costs of a hog were assigned to its various products based on weight, center-cut pork chops would have the same unit cost as pigs’ feet, lard, bacon, ham, etc., resulting in large profits for some cuts and consistent losses shown for others. Id. at 29,560. The hog example is found in Donald E. Keller, Management Accountants’ Handbook (4th ed. 1992) at 11:2 [hereinafter “Keller”]. Commerce reasons that like the hog, the pineapple is comprised of various parts, i.e., the cylinder, core, shells, etc. with different uses and values. Final Det. at 29,560-61. Commerce also noted that authoritative accounting literature provides examples of cost allocations in the canning industry dependent on two factors, a quantitative factor and a qualitative factor. Id. at 29,561 (citing Keller at 11:13). Commerce stated that: The output of finished products can be captured in the quantitative measure, which is used to allocate the direct preparation labor costs and other costs directly related to the quantity of raw fruit processed. The difference in the relative quality of the fruit used in each product is reflected in a qualitative factor, which is used to allocate the purchase cost of raw materials among products. The various grades or parts of the fruit are assigned a factor reflective of the quality of the fruit used for each product. With all of this in mind, we believe it is inappropriate to allocate fresh pineapple fruit costs to the various pineapple products solely on the basis of weight. Id. 8 The accounting citation is incorrect. It merely discusses the advantages of value-based allocation for normal costing purposes without discussing how to allocate in the canning industry specifically.9 The Thai plaintiffs claim that, while pineapple parts are not interchangeable, this fact is irrelevant because although the value of different pineapple parts may be different, the cost is all the same. The Thai plaintiffs do admit — 1323 — that the pineapple meat from which juice primarily is made is of a lower quality than the material used to make many, though not all, solid products, and because solid products have a higher market value than juice products, it can be said that the juice raw material has a lower value than at least some canned pineapple raw material. They argue, however, that they pay the same cost when they buy pineapples regardless of the mix of higher- and lesser-valued products ultimately processed. Dole argues that its weight-based methodology allocates fruit cost based on the weight of the net product output, e.g., only the weight of the juice pressed from the pineapple meat, and not the pulp left behind, is included in the net product weight. Therefore, Dole claims that its allocation methodology does not value all parts going on the distorted weight basis that Commerce asserts. Thus, Commerce’s shells, cores, etc. argument, as well as its hog analogy are overstated. Waste such as shells and pressed cores are given by-product treatment. Final Det. 60 Fed. Reg. at 29,566. One continuous fruit process yields a range of products from the whole fruit input. See, e.g. Malee Sec. D Resp. at 11-15; Pls.’ Conf. App., Ex. 12. As indicated, among solid products including crushed pineapple (a low value solid product) cost allocation by weight was found acceptable. Thus, apparently it is only for the juice vs. solid product that Commerce proposes to use a raw fruit value-based allocation, if Thai GAAP consistent “normal” methodologies are rejected. Dole also points out that the same accounting authorities relied upon by Commerce recognize the inappropriateness of a revenue-based or sales-price method when the purpose of the cost calculation is to validate prices: All sales-price methods are subject to one compelling criticism. Where cost is determined by price, price cannot be determined by cost. Therefore, all sales-price methods are circular for pricing decisions and inferentially so for many other types of decisions that have sales price as a fundamental factor. Keller at 11.18; see also Charles T. Horngren & George Foster, Cost Accounting: A Managerial Approach at 532 (7th ed. 1991) (stating, “[pjhysical measures are sometimes preferred to sales value methods in rate-regulation settings when the objective is to set a fair selling price. Why? Because it is circular reasoning to use selling prices as a basis for setting a selling price, as would be the case under the sales value at split-off method”).10 If joint inputs could be as easily distinguished as Commerce claims, there would be no need for allocation. In every co-product case where output prices vary greatly among co-products, value based allocation is intuitively appealing for antidumping purposes, but it is not permitted. In IPSCO, Inc. v. United States, 13 CIT 402, 405-06, 714 F. Supp. 1211, 12-13-15 (1989) [hereinafter “IPSCO II”], this court addressed the — 1324 — issue of how to allocate joint costs between co-products — “prime” and “limited-service” steel pipe or oil country tubular goods (“OCTG”).11 In the underlying investigation, Commerce allocated material, labor, and overhead costs uniformly to prime and limited service pipe because there was no ascertainable distinction in the costs actually incurred in producing these two products. 13 CIT at 405, 714 F. Supp. at 1215. This court rejected Commerce’s methodology, remanding the case with instructions to account for output value differences in allocating costs, as Commerce would do here, first by accepting Malee’s rough value allocations and second, by proposing an historic NRV methodology for Dole. The court stated that, “[b]y declining to account for differences in value and treating prime and limited service products identically in its calculations of foreign market value, [Commerce] made an unreasonable fair value comparison.” Id. Following IPSCO II on remand, Commerce allocated costs between prime and limited-service pipe based on relative sales value. IPSCO III, 965 F.2d at 1061. This court approved that result. The Federal Circuit reversed this court’s IPSCO II decision, found that a value-based allocation violated the antidumping statute, and also found that Commerce’s original weight-based methodology was the correct interpretation of the law. IPSCO III, 965 F.2d at 1060-61. The appellate court in IPSCO III noted that, “[t]he broad language of section 1677b(e) does not at any point expressly authorize adjustment of these production costs to account for products of a lower grade or less value.” Id. The court then found that the legislative history of Title 19 “confirms the statute’s unambiguous intent to provide cost of production as an independent yardstick for deciding whether home and export sales prices are suitable for fair value comparisons.” Id. The Federal Circuit found that calculating costs for both limited-service and prime products on the basis of their relative prices in the United States market was unreasonable and circular as “[t]he selling price of pipe became a basis for measuring the fairness of the selling price of pipe.” Id. at 1061. The court stated that, “[t]his circular reasoning contravened the express requirements of the statute which set forth the cost of production as an independent standard for fair value.” Id. Additionally, the Federal Circuit found that as Commerce’s original weight-based methodology was a consistent and reasonable interpretation of the statute, this court erred in substituting “its own construction of a statutory provision for a reasonable interpretation made by the administrator of an agency.” Id. (quoting Chevron U.S.A. v. Natural Res. Defense Council, 467 U.S. 837, 844 (1984)). Thus, the Thai plaintiffs and Dole claim that the holding in IPSCO III prohibits Commerce from shifting costs from one co-product to the other simply because there is a difference between them in market values, and despite the appellate — 1325 — court’s final statement on this issue, the entire opinion speaks of no statutory ambiguity. In its final determination, Commerce concluded that “[c]ontrary to respondents’ arguments, however, IPSCO is not controlling in this case.” Final Det., 60 Fed. Reg. at 29,561. Commerce claims that IPSCO did not hold that in every instance value-based allocations aré legally impermissible and attempts to distinguish IPSCO on its facts. See id. Commerce notes that in IPSCO, “[t]here were no physical differences between the two grades of pipe, only differences in quality and market value.” Id.; see IPSCO III, 965 F.2d at 1058. Commerce also states, “the same materials, labor, and overhead went into the manufacturing lot that yielded both grades of pipe.” IPSCO III, 965 F.2d at 1058. Commerce allocated production costs equally between the two grades of pipe because they were produced simultaneously and had identical production costs. Id. Commerce argues that IPSCO III suggests that the courts will defer to Commerce’s preference for reliance upon respondents’ normal allocation methodologies, particularly where there are significant differences in raw material, i.e., the primary use of the cylinder in the production of CPF and the additional use of the shells, cores, and ends, in the production of juice and concentrate, as well as differences in processing, labor, and overhead. See Final Det., 60 Fed. Reg. at 29,561. The Thai plaintiffs counter that Commerce’s attempt to distinguish IPSCO III fails because although in IPSCO III materials, labor, and overhead costs were shared, whereas in this case, there are clear differences in labor and overhead costs, this fact is immaterial because the issue of cost allocation in this case pertains only to shared raw material costs. They are correct. They also claim that all pineapple using co-products incur identical baht per kilogram fruit costs, just as prime and limited service pipe incurred identical material, labor, and overhead costs. The Thai plaintiffs argue that to the extent that co-products have shared costs that are incurred uniformly, they should be allocated uniformly. Further, the Thai plaintiffs assert that in IPSCO III, as in this case, it is impossible to separate out the different quality inputs before the costs are incurred. Instead costs are incurred uniformly and all at once, at the exact same rate for high quality inputs as for low quality inputs. The Thai plaintiffs contend that Federal Circuit in IPSCO III held that in these circumstances, costs must be allocated uniformly among the resulting high-value and low-value products. As indicated, the court finds that Commerce overstates distinctions with IPSCO III and to some extent it also misstates the facts of this case, as explained previously. See supra at 23. The court finds that Commerce may not rely on output price-based allocation methodologies given the holding in IPSCO III. IPSCO II allowed Commerce to choose between any number of value-based methodologies. IPSCO II, 13 CIT at 406 n.6, 714 F. Supp. at 1215 n.6. The Federal Circuit approved none of those methodologies. Ascertaining actual costs per product is just as impossible for pineapple co-products as it was for pipe co-products. The court — 1326 — made no exceptions for intentionally produced lower value outputs, as may be the case here, and it certainly did not say that weight-based methodologies are to be used only when they work to respondents’ disadvantage. The court declines to reject the words of the Federal Circuit and to limit IPSCO III to its facts. The Federal Circuit specifically found that Commerce’s weight-based cost (non-value) methodology was correct, not just permissible. It in no way suggested that Commerce had discretion to choose a value-based methodology, rather it found output price-based methodologies circular. The Federal Circuit knows how to allow Commerce discretion. See, e.g., Federal-Mogul Corp. v. United States, 63 F.3d 1572, 1581-82 (Fed. Cir. 1995) (Commerce may choose among any number of permissible VAT methodologies).12 Despite the final paragraph which repeats the general rule of deference to agency decisions, fully three-fourths of IPSCO III is in mandatory language. The statute is declared unambiguous. Thus, no true Chevron deference to an agency’s choice among several permissible interpretations is involved. 965 F.2d at 1060. The court must simply defer to the one correct statutory interpretation. It is a rare opinion that could not be more clearly written. The court, however, declines to dismiss the bulk of IPSCO III as simply loose language. The court must assume that IPSCO III means what it says. Joint cost allocations for co-products are an accounting dilemma. No method is totally satisfactory. The IPSCO III court apparently decided that Congress made a choice to avoid circularity, despite the predictable distortive effects on COP and CV for some co-products. While this court may not agree that Congress made such a choice and indeed believes that the circularity problems can be controlled for, and, thus, that output price-based methodologies which avoid the worst distortions potentially caused by POI U.S. price-based allocations are permissible, this court has no authority to approve them.13 The appellate court has spoken on this issue and only it or Congress may change the law. Accordingly, this issue is remanded. Commerce may choose to accept the weight-based allocation methodologies put forth by the Thai plaintiffs and Dole if they are otherwise acceptable, because these are cost, not price-based, methodologies, or it may rely on another non-output price-based cost allocation methodology. III. Weighting the Dumping Margin for Dole’s United States Sales: Commerce allowed Dole to report all of its United States sales, including those of Philippine origin, for each product category for which Dole had shipments from Thailand during 1993, because Dole was unable to distinguish between its pineapple grown and canned in Thailand and its — 1327 — pineapple grown and canned in the Philippines. In its preliminary and final determinations, Commerce calculated a dumping margin for Dole by weight-averaging the dumping margin for each product category according to the ratio of shipments of CPF from Thailand to total volume shipped from both Thailand and the Philippines during the last seven accounting periods of 1993, i.e., July 19 through December 31, 1993. Prelim. Det., 60 Fed. Reg. at 2736; Final Det., 60 Fed. Reg. at 29,562. Commerce used the July-December accounting period as the basis for establishing the ratio rather than the entire 1993 period because Dole reported that its average inventory turnover rate was six to seven months. Prelim. Det., 60 Fed. Reg. at 2736. After Commerce issued its final determination, petitioners pointed out a flaw in Commerce’s rationale: for certain products sold in the United States in the first half of 1994, the shipment data showed that (a) they had been shipped from Thailand to the United States only in the first half of 1993, and (b) they were not shipped at all from the Philippines to the U.S. in 1993. As a result, the sales of these products, which the petitioners emphasized showed positive dumping margins were given zero weight using Commerce’s Period 7-13,1993 shipment ratios. While describing this flaw as an error in Commerce’s methodology, petitioners nevertheless asserted that the result was “unintentional,” and for that reason qualified as a clerical or ministerial error that could be corrected after issuance of the final determination. Commerce agreed that the result was “unintentional” and reasoned that its methodology: was only intended to influence the weight assigned to the PUDD14 calculated for those UPCs where Dole had shipments from both Thailand and the Philippines. Any elimination of PUDD values calculated on UPCs sourced exclusively from Thailand was an oversight on [Commerce’s] part and was unintentional. Ministerial Errors Decision Mem. at 6 (June 28, 1995) [hereinafter “Ministerial Errors Mem. ”]; P.R. Doc. 342; Defs.’ App., Ex. 6. Pursuant to 19 C.F.R. § 353.28(c) (1994), Commerce recalculated Dole’s dumping margin in its amended final determination so as to include products exclusively of Thai-origin that had been excluded from the final dumping calculation because they were shipped prior to the seven month accounting period examined by Commerce. See Amended Final Det., 60 Fed. Reg. at 36,776; Ministerial Errors Mem. at 4-6. For these products, Commerce used a full year 1993 weighting period as the basis for establishing the ratio; for the remaining sales, Commerce continued to use the July-December accounting period as the basis for the ratio. See Amended Final Det., 60 Fed. Reg. at 36,776. Dole claims that Commerce’s inconsistent use of different weighting factors in the amended final determination is arbitrary and unreason — 1328 — able. Dole argues that it is inherently arbitrary to calculate weighting factors based on an average inventory turnover for all products, but then make exceptions for individual products with different turnover ratios. Such selective application, Dole asserts, destroys the presumed congruence between the set of products used to calculate the average and the set of products to which the results are applied. Without such congruence, the justification for using the average inventory period is lost. Dole contents that in citing examples of certain products sold in the POI but not shipped in the second half of 1993, petitioners have simply demonstrated that the July to December 1993 shipment data do not accurately represent the source of the products sold in the POI, and that a longer period should be used. Upon reconsideration of this issue, Commerce agrees with Dole that it should have applied a consistent weighting factor across the board. The court agrees. This issue is remanded so that Commerce may use the full-year 1993 shipment volumes, supplied by Dole in its Section C Questionnaire Response, to weight the dumping margin for all Dole products. IV. Dole’s G&A Expense Factor: In its Section D Questionnaire Response, Dole computed a G&A factor based upon 1993 audited financial data for Dole Food Company, Dole Packaged Foods, and Dole Thailand. Dole Sec. D Resp. at 16 & Ex. D-12 (Dec. 19, 1994) [hereinafter “Dole Sec. D Resp.”]; C.R. Doc. 52; Pl.-Intvr.’s Conf. App., Ex. 10. Dole subsequently provided a revised G&A factor based upon 1994 audited financial data for Dole Food Company and upon 1994 unaudited financial data for Dole Thailand. Dole COP/CV Verification Report at 15 (Apr. 5, 1995); C.R. Doc. 115; Pl.-Intvr.’s Conf. App., Ex. 15; and Dole Letter to Commerce (Apr. 27, 1995); P.R. Doc. 295; Pl.-Intvr.’s Conf. App., Ex. 18 (Dole Food Company, Inc.’s 1994 annual report containing audited financial statements). In its final determination, Commerce calculated G&A expenses using Dole’s 1993 audited financial information consistent with Commerce’s practice of calculating G&A expenses from audited financial statements that most closely correspond to the POI. Final Det., 60 Fed. Reg. at 29,565; see also Furfuryl Alcohol from Thailand, 60 Fed. Reg. 22,557, 22,561 (Dep’t Comm. 1995) (final determ. of LTFV sales) (stating “[u]nder ordinary circumstances, the most appropriate full-year G&A period is that represented by the latest fiscal year for which the respondent has complete and audited financial statements”). Dole objects to Commerce’s determination to rely on its 1993 financial statements for both Dole Food Company and Dole Thailand. Dole argues that Commerce should have based Dole’s G&A expense factor upon a combination of Dole’s 1993 audited financial data for Dole Thailand and Dole’s 1994 audited financial data for Dole Food Company. Commerce claims, however, that this approach would have been inconsistent with Commerce’s practice and policy. In calculating G&A, Commerce asserts that it is looking at a “snapshot” of expenses incurred over — 1329 — the specific period of operations reported in the respondents’ financial statements. Under Dole’s proposed reporting methodology, Commerce claims that companies within its consolidated group would be allowed to mix and match financial data from different periods, which would defeat the snapshot approach to Commerce’s G&A factor calculation. Dole counters that, unlike a snapshot, which records the situation at a given moment in time, a full-year record of G&A expenses captures all such expenses incurred over the full fiscal year, thereby capturing both recurring daily expenses and periodic annual expenses. Such expenses, Dole contends, will change from year to year for a company, particularly where one-time restructuring expenses are incurred. Dole also argues that the mere possibility that in some cases a respondent might attempt to manipulate its data by withholding release of audited annual reports does not support creation of an irrebuttable presumption that such manipulation occurs in every case. .The court agrees with Commerce that Dole’s proposed methodology would lead to inconsistent practices and, in fact, is contrary to Dole’s position that a consistent time period must be used in regard to margin weighting. Commerce’s determination to rely on the audited 1993 financial reports for Dole Food Company and Dole Thailand was reasonable and otherwise in accordance with the law. Accordingly, the court affirms Commerce’s determination in regard to Dole’s G&A expense factor. V. Dole’s Inventory Carrying Cost: Inventory carrying cost measures the imputed cost incurred by a company for storing merchandise in inventory. Commerce is unable to calculate the actual cost for inventory carrying costs because such costs are not found in a company’s books. Thus, Commerce imputes the opportunity cost of holding the merchandise in inventory prior to the sale to an unrelated customer. Commerce deducts inventory carrying costs, measured from the date of production is completed through the date of shipment to an unrelated customer, from exporter sales price and FMV. See Torrington Co. v. United States, 44 F.3d 1572, 1579-81 (Fed. Cir. 1995) (upholding this practice). With respect to the interest rate used to calculate inventory carrying cost, Commerce explains that its practice is to use the short-term rate in the currency in which the cost of the inventory is incurred by the entity that bears the costs of producing or acquiring such inventory. See, e.g., Tapered Roller Bearings, Four Inches or Less in Outside Diameter, and Components Thereof, from Japan, 57 Fed. Reg. 4975, 4984 (Dep’t Comm. 1992) (final results) [hereinafter “TRBs Final Results”]. If there is evidence, however, that an entity other than the one holding the merchandise in inventory bears the cost of carrying the merchandise for a portion of the time the merchandise is held in inventory, Commerce will use this entity’s short-term interest rate to calculate that portion of the inventory carrying cost. Id. For example, in TRBs Final Results, Com — 1330 — merce used the Japanese parent company’s short-term interest rate to calculate United States inventory carrying cost because the payment terms that the parent company extended to its United States subsidiary indicated that the parent company bore the cost of maintaining the inventory in the United States. Id.; see also Timken Co. v. United States, 18 CIT 942, 948-49, 865 F. Supp. 881, 886-87 (1994) (upholding this determination). Here, in reporting inventory carrying cost, Dole applied a United States dollar denominated short-term borrowing rate to the actual cost of manufacture of the inventory for the time that the merchandise was held in inventory in the United States, Thailand, and Germany. See Final Concur. Mem. at 4-5. In its final determination, Commerce recalculated Dole’s inventory carrying cost for the time that the merchandise was held in inventory in Thailand using a Thai baht-denominated, short-term borrowing rate. Final Det., 60 Fed. Reg. at 29,554, 29,556; see also Final Concur. Mem. at 5. For the time that the merchandise was on the water and in inventory in the United States or Germany, Commerce used the United States dollar-denominated, short-term borrowing rate supplied by Dole and confirmed by Commerce at verification. Final Det., 60 Fed. Reg. at 29,554, 29,556. Dole contends that Commerce improperly substituted an imputed baht interest rate for Dole’s actual dollar cost of borrowing in calculating Dole’s inventory carrying cost. Dole argues that Dole’s use of Dole Food Company’s actual dollar-denominated short-term cost of borrowing as the applicable rate of interest was consistent with Commerce’s then-prevailing practice under LMI-La Metalli Industriale, S.p.A. v. United States, 912 F.2d 455, 460-61 (Fed. Cir. 1990), of calculating both credit expenses and imputed interest expenses using the lowest borrowing rate available to the respondent. See also Class 150 Stainless Steel Threaded Pipe Fittings from Taiwan, 59 Fed. Reg. 38,432, 38,434 (Dep’t Comm. 1994) (Commerce “is required to use the lowest rate at which the respondent has borrowed or to which the respondent has access”). Dole asserts that the court in LMI-La Metalli recognized that it defies commercial reality to presume that a company would borrow at a high home currency interest rate when a lower dollar rate is available. See LMI-La Metalli, 912 F.2d at 460 (stating, “we agree with LMI that it is not reasonable to presume that a commercial enterprise would borrow at almost twice the available rate”). Dole claims that in its final determination, Commerce adopted a modified methodology for calculating imputed credit costs, and then extended that methodology to inventory carrying costs. Dole argues that Commerce failed to properly apply its own stated methodology, which applies an interest rate using the currency of the pending transaction. Dole contends that because all of Dole Thailand’s export sales are in U.S. dollars, it is the cost of holding inventory pending receipt of U.S. dollars that defines the relevant currency under that methodology. Dole also points out that as a Thai corporation, Dole Thailand keeps its — 1331 — day-to-day financial statements in baht in fulfillment of local statutory requirements, but as a subsidiary of Dole Food Company, Dole Thailand also prepares dollar-denominated financial statements, which are used to consolidate its financial position in the financial statements of Dole Food Company. Finally, Dole claims that the imputation of a higher baht-denominated interest rate is inconsistent with commercial reality and Dole Thailand’s own borrowing practices. Dole claims the current liabilities consist overwhelmingly of payables to affiliates (approximately 80% of total current liabilities), with only about 3% of total current liabilities in debt to banks, which might be local debt. Dole Sec. D. Resp., Ex. D-4. Accordingly, Dole argues that the imputed cost of holding finished goods inventory is borne by the consolidated Dole group, and is properly calculated based on the lower dollar-based cost of borrowing available to Dole Food Company. Commerce counters that there is no record evidence to corroborate Dole’s statement that the imputed cost of holding finished goods inventory in Thailand is borne by the consolidated Dole group. In contrast, Commerce noted that in Timken Co., specific payment terms, allowing for delayed payment for the imported merchandise, formed the eviden-tiary basis for Commerce’s conclusion that the American subsidiary could have benefitted from its Japanese parent’s ability to borrow money in Japan. 18 CIT at 948-49, 865 F. Supp. at 886-87. Dole’s contentions have not been adequately addressed. If Dole’s borrowings are largely in U.S. dollars because they are debt to U.S. affiliates, such specific evidence would seem to be similar to that accepted in Timken. Commerce must address these arguments and make a rational determination. Accordingly, the court finds that Commerce’s determination in this respect not supported by substantial evidence and it is remanded. VI. Dole’s Pre-Sale Movement and Import Duty Expenses: Dole reported pre-sale movement and import duty expenses for its third country sales. In its final determination, Commerce concluded that Dole’s third country movement and import duty expenses did not bear a direct relationship to the sales under consideration and, therefore, could not be deducted fully as an ordinary circumstance-of-sale (“COS”) adjustment pursuant to 19 U.S.C. § 1677b(a)(4)(B) and 19 C.F.R. § 353.56(a)(1) (1994). Final Det., 60 Fed. Reg. at 29,563-64. Commerce also concluded that to treat similar United States movement and import expenses as indirect expenses for purposes of the ESP offset cap would be contrary to 19 U.S.C. § 1677a(d)(2)(A). See Sharp Corp. v. United States, 63 F.3d 1092, 1097 (Fed. Cir. 1995).15 Thus, the advantage to be gained through the ESP offset to FMV for indirect expenses was limited. — 1332 — Dole argues that the pre-sale movement and import duty expenses are directly related to its third country sales and, as such, should have been deducted fully from FMV To determine whether pre-sale movement expenses are direct, Commerce examines the respondent’s pre-sale warehousing expenses because Commerce considers the pre-sale movement charges incurred in positioning the merchandise at the warehouse to be linked, for analytical purposes, to pre-sale warehousing expenses. Final Det., 60 Fed. Reg. at 29,563; see also The Ad Hoc Committee of AZ-NM-TX-FL Producers of Gray Portland Cement v. United States, Slip Op. 95-91 at 5-6 (May 15, 1995), aff'd, 95 F.3d 1164 (Fed. Cir. 1996) (unpublished table decision). Commerce explains that it typically “treats expenses associated with inventory that is held for purposes of production planning and being able to ship the merchandise quickly with a regular turnover as indirect selling expenses because this inventory is maintained by the company as a service to all customers.” Final Det., 60 Fed. Reg. at 29,563. Commerce acknowledges, however, that in limited circumstances, it does recognize certain pre-sale expenses as direct. Id. For freight and warehouse expenses, Commerce claims that those circumstances usually involve products channeled or customized for certain buyers. See, e.g., Stainless Steel Bar From Italy, 59 Fed. Reg. 66,921, 66,928-29 (Dep’t Comm. 1994) (final determ, of LTFV sales) (allowing COS adjustment where pre-sale warehousing expenses incurred for designated amount of subject merchandise with certain specifications for particular customers). In the present case, Dole reported two types of third country warehousing expenses: (1) those associated with movingthe merchandise “in and out” of the warehouse; and (2) warehouse storage charges. Dole Sec. B Resp. at 11-12 (Sept. 20, 1994); P.R. Doc. 90; Defs.’ App., Ex. 9. Commerce found that Dole had not provided evidence to substantiate its claim that either pre-sale warehousing expense was directly linked to the sales under investigation. Final Det., 60 Fed. Reg. at 29,564. Commerce listed its reasons as follows: (1) The amount of time that passes between the date the merchandise arrives at the European warehouse and the date it is shipped to the third country customer; (2) in most instances the third country sales were made from inventory * * *; (3) the merchandise held in the European warehouses is not pre-designated for sale to a specific customer; (4) the merchandise sold from inventory was not specialty merchandise * * *; (5) the merchandise that was held in inventory was sold to numerous third country customers during the POI; (6) Dole incurs the cost of pre-sale warehousing expenses, not the customer * * *; and (7) in its questionnaire response Dole did not claim the warehouse storage charges as direct selling expenses; rather, Dole characterized warehouse storage costs as indirect expenses. — 1333 — Id. Thus, Commerce concluded that because Dole’s third country pre-sale warehouse expenses were indirect, the expenses involved in moving the merchandise to the warehouse were also indirect. Id. Dole contends that unlike other types of selling expenses, such as general advertising or sales staff salaries, third country movement expenses are directly tied to subsequent individual sales of the products because they add value directly to the individual articles sold. Dole claims that it is indisputable that to the German customer, a shipment sold delivered, duty paid in Germany has greater value than a shipment sold FOB Bangkok. Dole argues that by moving goods from Thailand to Germany and paying 24 percent ad valorem import duties, Dole has added identifiable value to those goods and upon their subsequent sale, that value is transferred to the customer. Dole likens this situation to selling costs incurred by the producer that inure to the benefit of the purchaser, such as advertising expenses “assumed by the producer * * * on behalf of the purchaser.” See 19 C.F.R. § 353.56(a)(2). Dole distinguishes this situation from that in Ad Hoc Committee, Slip Op. 95-91, where the movement was of inventory from the factory to the warehouse in the home market, which Dole argues is neither designated to any specific market nor significantly advanced in value. This distinction is not meaningful. Presumably both direct and indirect expenses add value to the merchandise. Indirect expenses are not ignored, they are simply treated differently. Commerce did not err in finding that there is an insufficient tie of these expenses to particular sales to warrant direct expense treatment. Dole also points out that the Federal Circuit has recently confirmed that pre-sale as well as post-sale movement expenses may be deducted from FMV as a COS adjustment. Torrington Co. v. United States, 68 F.3d 1347, 1353-56 (Fed. Cir. 1995). As indicated, limited pre-sales expenses have been found to be direct. Torrington did not change this. Commerce is correct that Dole’s reliance on Torrington is misplaced as the court did not reach the issue of whether Commerce’s analytical link between pre-sale movement expenses and pre-sale warehousing expenses is in accordance with law, as it did in Ad Hoc Committee. The court merely approved Commerce’s use of the COS standards to determine whether home market pre-sale movement expenses should be deducted in the calculation of FMV. Id. at 1356. This does not change the fact that, pursuant to 19 C.F.R. § 353.56(a)(1), Commerce will make a full COS adjustment only if the expenses are determined to be directly related to the sales under investigation. In the light of the foregoing, the court finds that Commerce’s determination that Dole’s pre-sale third country movement expenses were not directly related to the sales under investigation and were therefore not fully deductible under the COS provision was reasonable and supported by substantial evidence. Alternatively, Dole argues that if its third country pre-sale movement expenses must be characterized as indirect, then Commerce erred in refusing to treat similar United States movement expenses as indirect — 1334 — for purposes of the ESP offset cap, as well. Commerce, however, while recognizing that there is an “admitted imbalance” in its treatment of Dole’s United States movement expenses and its treatment of third country movement expenses, concluded, after examining the then applicable statutory and regulatory provisions, that it was “without the authority” to correct the imbalance. Final Concur. Mem. at 18. In its final determination, Commerce stated, “[w]e do not have the option of treating comparable expenses on U.S. sales as indirect in nature because such sales are ESP sales, and [19 U.S.C. § 1677a(d)(2)(A)] clearly requires the deduction of such expenses in arriving at USP.” Final Det., 60 Fed. Reg. at 29,564. Section 1677a(d)(2)(A), Title 19, United States Code provides in relevant part: The purchase price and the exporter’s sales price shall be adjusted by being— ‡ ‡ ‡ ‡ ‡ ‡ (2) reduced by— (A) * * * the amount, if any, included in such price, attributable to any additional costs, charges, and expenses, and United States import duties, incident to bringing the merchandise from the place of shipment in the country of exportation to the place of delivery in the United States; * * *. 19 U.S.C. § 1677a(d)(2)(A). The Federal Circuit recently came to the same conclusion, finding that treatment under the agency created ESP offset cap is inappropriate for expenses already subject to specific statutory adjustment. Sharp Corp. v. United States, 63 F.3d at 1097. Dole claims Sharp is not applicable here because (1) it involved home market selling expenses not substitute third-country transport and duty expenses; (2) a court cannot uphold an agency decision on grounds not relied upon by the agency; and (3) Sharp did not consider the legislative history of Congress’ recent amendment of the antidumping law to expressly provide for the deduction of all movement expenses from FMV Finally, Dole claims that 19 U.S.C. § 1677a(d)(2)(A) says nothing about the categorization of U.S. pre-sale movement and import duty expenses. Of course, the new statute does not apply and the old statute gives a clear answer. Thus, Dole’s arguments are either irrelevant or incorrect on this issue. The court finds that, in light of the Federal Circuit decision in Sharp, Commerce’s refusal to include U.S. pre-sale movement expenses in setting the ESP offset cap pursuant to 19 U.S.C. § 1677a(d)(2)(A) was in accordance with law. Accordingly, the court affirms Commerce’s decision in this respect. Conclusion This matter is remanded for Commerce to use a consistent time period for calculating the weighted average dumping margin for all Dole products, to correct the effective date of the antidumping duty order — 1335 — with respect to Dole, to consider Dole’s evidence in support of a U.S. dollar inventory cost measure, and to use a non-output price-based methodology to allocate joint costs between solid and non-solid outputs. Remand results should be returned within 60 days. Any new objections may be filed within 11 days thereof. Any responses are due within 11 days thereafter. Footnotes [*] Footnote 6 contains bracketed confidential information which is to be deleted from publicly distributed copies. [1] The applicable statutory and regulatory provisions in this case are those provisions as they existed on December 31, 1994, i.e., prior to the amendments made under the Uruguay Round Agreements Act. [2] “CONNUM” is the computer field name assigned to each unique product sold. “CONNUMU” indicates a United States product and “CONNUMT” indicates a third country product. [3] In calculating the dumping margins, Commerce relied upon sales to Germany as the basis for foreign market value (“FMV") because it determined that none of the four respondents had a viable home market. Prelim, Det., 60 Fed. Reg. at 2737; Final Det., 60 Fed. Reg. at 29,554. [4] CPF is defined as “pineapple processed and/or prepared into various product forms, including rings, pieces, chunks, tidbits, and crushed pineapple, that is packed and cooked in metal cans with either pineapple juice or sugar syrup added.” Final Det., 60 Fed. Reg. at 29,553. [5] No party argued that juice is a byproduct, such as pineapple waste. Thus, all parties agree that a co-product methodology for allocation of joint costs is necessary. [6] [ ] [7] SAICO and TEPCO’s normal allocations were approximately midway between Malee’s and Dole’s. [8] Commerce does not propose and no party here employed a fact-based unitary value/weight ratio for fruit allocation purposes. [9] An article referred to in Keller at 11:14 does discuss the canning industry. It suggests aqualitative approach for cost allocation for various grades of fruit. Jankowski, “Cost and Sales Control in the Canning Industry,” 36 N.A.C.A. Bull, at 376 (Nov. 1954). This is the opposite of Commerce’s approach here. That is, cost allocation by weight among solid pineapple products was allowed. [10] Although rate regulation perse is not involved here, antidumping laws relate to price setting, and, thus, are analogous. [11] Prime OCTG meets American Petroleum Institute (“API") standards and sells at premium prices, whereas lira-ited-service OCTG does not meet API standards tolerances and therefore sells at a discount. See IPSCO III, 965 F.2d at 1057-58. [12] The court here uses value as did the court in the IPSCO cases to refer to output price or net realizable value of the outputs. It does not refer to value as cost-based value. [13] The decision to use historical value allocation for Dole in the future appears to be another attempt to sidestep IPSCO III. While in IPSCO III the remand decision rejected by the appellate court used a POI-based value allocation, IPSCO Ill’s rejection of value-based allocations is unlimited. [14] The “PUDD” (potentially uncollectible dumping duties) is the absolute amount of the dumping dutieB (as opposed to a percentage of volume) that would have been collected from the United States sales under investigation if an antidumping duty order had been in effect. [15] The Sharp decision sets forth the background of the ESP offset and ESP offset cap.