Guangdong Chemicals Import & Export Corp. v. United States

*1366 OPINION RESTANI, Chief Judge. Restani, Chief Judge: Plaintiff Guang-dong Chemicals Import and Export Corporation (“Guangdong”) challenged the results of an administrative review of an antidumping duty order on sebacic acid from the People’s Republic of China (“China”). Following oral argument, the court remanded for the Department of Commerce (“Commerce”) to reconsider the reliability of data used to calculate the surrogate value of sebacic acid, and also to explain its choice to deduct a by-product credit from normal value, rather than from manufacturing costs. On remand, Commerce reexamined its data, excluded aberrational values, and explained its decision to change its policy with respect to byproduct credits. Following remand, Gu-angdong asserts that Commerce’s exclusion of aberrational values does not justify its use of less product-specific data. Gu-angdong also argues that Commerce’s practice of deducting by-product credits from normal value is arbitrary, capricious and unsupported by substantial evidence. The court finds that Commerce’s choice of data set and its treatment of the by-product credit are reasonable and supported by substantial evidence. I. Background In 1994, Commerce issued an order imposing antidumping duties on sebacic acid from China. See Sebacic Acid from the People’s Republic of China, 59 Fed.Reg. 35,909 (Dep’t Commerce July 14, 1994) (notice of antidumping duty order). On December 16, 2004, Commerce completed an administrative review of that order for the period of review (“POR”) from July 1, 2002, to June 30, 2003. See Sebacic Acid from the People’s Republic of China, 69 Fed.Reg. 75,303 (Dep’t Commerce Dec. 16, 2004) (notice of final results of antidump-ing administrative review) (“Final Determination”). Two of Commerce’s actions taken during that review are at issue in this case. The first issue involves Commerce’s valuation of sebacic acid. Because Guang-dong’s supplier of sebacic acid, Hengshui Dongfeng Chemical Co., produces a co-product, capryl alcohol, Commerce must allocate the supplier’s costs of manufacturing between the two products based on their relative sales values. See Section C and D Response of Guangdong Chems. Imp. & Exp. Corp. (Nov. 4, 2003), P.R. Doc. 21, at D-4 (“Section C & D Response ”). Because India does not produce sebacic acid, Commerce relied on statistics describing the price of sebacic acid imported into India from other countries. Prelim. Valuation of Factors of Prod. (July 30, 2004), P.R. Doc. 47 at 1-2. Commerce chose to use import statistics maintained by the Indian government, based on a six-digit Harmonized Tariff Schedule (“HTS”) category (“Indian government data”). Id., P.R. Doc. 47 at 4, Attach. 4. That category lumped together imports of sebacic acid with imports of azelaic acid. Id., P.R. Doc. 47 at Attach. 4. Guangdong advocated the use of product-specific data maintained by the publication Chemical Weekly in its Chemicals Import and Export trade database index (“Chemical Weekly data” or “ChemlmpEx”). Submission of Publicly Available Data for Use as Surrogate Value (Sept. 8, 2004), P.R. Doc. 62 at 2 (“Surrogate Value Submission”). That data was taken from a selection of information from the Indian government, but included a classification specific to sebacic acid. Id., P.R. Doc. 62. Because Guangdong’s data included limited data points (in fact, only two imports, both from Germany, totaling 1,400 kilograms), Guangdong submitted additional corroborating data to bolster its limited data set. Id., P.R. Doc. 62 at 2, Attach. 1. Without considering the impact of the corroborating data on the veracity of either data set, Commerce re *1367 jected the Chemical Weekly data and adopted the Indian government data. See Issues & Decision Memorandum for the 2002-2003 Antidumping Administrative Review of Sebacic Acid from the People’s Republic of China, A-570-825, at 6-9 (Dec. 10, 2004) available at http://ia.ita.doc.gov/ frn/summary/prc/E4-3678-l.pdf (“Issues & Decision Mem,.”). Because Commerce failed to consider Guangdong’s corroborating data, the court remanded this issue for additional consideration. Guangdong Imp. & Exp. Co. v. United States, 30 CIT -, -, 414 F.Supp.2d 1300, 1313 (2006). The second issue involves a change in Commerce’s treatment of by-product credits. Because Hengshui produces fatty acid and glycerine as by-products of sebacic acid, Commerce gave Guangdong a credit reflecting the value of the by-products. See Final Redetermination Pursuant to Court Remand (May 3, 2006), Remand P.R. Doc. 4 at 7 (“Final Redetermination”). In its preliminary determination, Commerce applied this credit to the cost of manufacturing sebacic acid. See id. In its final determination, Commerce applied the credit against normal value, after calculating overhead costs, “special general and administrative” (“SG & A”) expenses, and profits. See id. Commerce failed to provide an opportunity for interested parties to comment on this change in methodology before issuing its final determination. Id. Commerce therefore requested a remand in order to explain its application of the by-product credit. Id. Commerce issued its Final Redetermination on May 3, 2006. As described more fully below, the Final Redetermination continued to use the Indian government data to value sebacic acid, but adjusted the Indian government data to eliminate aberrational values. Id., Remand P.R. Doc. 4 at 3, 5. Commerce also explained the rationale behind its application of Guangdong’s by-product credit. Id., Remand P.R. Doc. 4 at 7. Guangdong argues that Commerce’s choice of data set remains unreasonable, and that Commerce’s application of the by-product credit is unreasonable in light of generally accepted accounting procedures. See Pl.’s Comments on Def.’s Final Determination Pursuant to Court Remand at 1-2 (“PL’s Comments”). The court addresses each issue in turn. II. Commerce’s Use of the Indian Government Data to Calculate the Normal Value of Sebacic Acid Because India does not produce sebacic acid, Commerce relied on import statistics to estimate the value of sebacic acid. As mentioned, Commerce used statistics from the Indian Department of Commerce’s Import/Export Data Bank, based on a six-digit basket category in the Indian HTS, 1 which includes both sebacic acid and azelaic acid. Issues & Decision Mem. at 3. During the review, Guangdong offered more product-specific data compiled in an import and export database maintained on the website of the Indian publication Chemical Weekly. Guangdong Chems. Imp. & Exp. Co. Case Br. (Sept. 20, 2004), P.R. Doc. 65, at 4-6. Guangdong proposed using the Chemical Weekly data, which was based on a portion of the Indian government’s information, but was further subdivided and included a specific subheading for sebacic acid. 2 Surrogate Value Submission, P.R. Doc. 62 at 2. Based on this data, Guangdong argued that the value of sebacic acid in India during the POR was $3,551.73. 3 Id., P.R. Doc. 62 at *1368 2. Guangdong corroborated its proposed value with data from U.S. import statistics for sebacic acid, benchmark price data from the publication Chemical Market Reporter, and prices for oxalic acid, a chemical asserted to be similar to sebacic acid. Id., P.R. Doc. 62 at 2-3. In response to Guangdong’s proposed data, Commerce conducted additional research to determine whether prices of aze-laic and sebacic acid were similar. See Comparison of U.S. Int’l Trade Comm’n Dataweb Values for Sebacic Acid & Azelaic Acid Imps. to the United States (Dec. 10, 2004), P.R. Doc. 79 at 1 (“Price Comparison Mem.”). It concluded that the two products were similarly priced, varying only by $.30 per kilogram over a twenty-three-month period during which the price for sebacic acid ranged between $2 and $3 per kilogram. Id., P.R. Doc. 79 at 1. Commerce therefore used the broader Indian government data to arrive at a surrogate value of $15,826.30 for sebacic acid. See Issues & Decision Mem. at 9 (electing to use Indian government data); see also Prelim. Valuation of Factors of Prod., P.R. Doc. 47 at 4 (using Indian government data to arrive at $15,826.30 per-metrie-ton value for sebacic acid). In rejecting the Chemical Weekly data, Commerce reasoned that it could not determine how the Chemical Weekly data were derived from the Indian government information, and that the Chemical Weekly data lacked “a sufficiently broad range of import values.” See Issues & Decision Mem. at 7. Guangdong filed suit in this Court to challenge the results of the administrative review. See Guangdong, 30 CIT at-, 414 F.Supp.2d at 1300. Guangdong argued, inter alia, that Commerce had not supported its decision to use the Indian government data instead of the Chemical Weekly data. Id. at-, 414 F.Supp.2d at 1303. Because Commerce did not explain why it rejected the Chemical Weekly data without consideration of the corroborating data submitted by Guangdong, nor explained why the Indian government data were not aberrational, the court remanded for Commerce to address these infirmities in its reasoning. Id. at •-•, 414 F.Supp.2d at 1312-13. In its Final Redetermination, Commerce retained use of the Indian government data, but “examined the U.S. import statistics, the European Union import statistics, and the Chemical Market Reporter data that Guangdong provided on the record for benchmarking purposes.” Final Redetermination, Remand P.R. Doc. 4 at 5. Commerce noted that the value of seba-cie acid in the Final Determination, $15,826.30, was significantly higher than the value of the “benchmark data,” which showed a price of $3,061.54 for sebacic acid imported into the United States (excluding China, India and Korea), $3,098.42 for the European Union, and $4,187.60 developed from price quotes in the Chemical Market Reporter. Final Redetermination, Remand P.R. Doc. 4 at 5. On the basis of this evidence, Commerce found that its data for the POR were aberrationally high when compared with Guangdong’s corroborating data. Id., Remand P.R. Doc. 4. Consequently, Commerce reexamined its import data for India and determined that sales from the United States had skewed its results. Id., Remand P.R. Doe. 4 at 5-6. After removing the aberrational data, Commerce found the Indian import price of sebacic acid to be $4,901.88. Id., Remand P.R. Doc. 4 at 6. Despite these changes, Guangdong continues to argue that Commerce could not reasonably use the Indian government data when a more product-specific data set was on the record. 19 U.S.C. § 1677b (2000) provides that valuation of factors of production “shall be *1369 based on the best available information,” but does not mandate that Commerce use any particular data source. Id. § 1677b(c)(l)(B). This gap in statutory authority leaves Commerce with considerable discretion in selecting a data source to calculate normal value. See Nation Ford Chem. Co. v. United States, 166 F.3d 1373, 1377 (Fed.Cir.1999). The court will uphold Commerce’s determination 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)(i). The fact that the evidence on the record may support two inconsistent outcomes does not mean that the agency’s selection of one alternative is unreasonable. Goldlink Indus. Co. v. United States, 30 CIT -, -, 431 F.Supp.2d 1323, 1326 (2006) (citing Consolo v. Fed. Mar. Comm’n, 383 U.S. 607, 620, 86 S.Ct. 1018, 16 L.Ed.2d 131 (1966)). Thus, “the court may not substitute its judgment for that of the [agency] when the choice is ‘between two fairly conflicting views, even though the court would justifiably have made a different choice had the matter been before it de novo.’ ” Id. at-, 431 F.Supp.2d at 1326 (quoting Am. Spring Wire Corp. v. United States, 8 CIT 20, 22, 590 F.Supp. 1273, 1276 (1984)). In this case, Commerce identified “several factors, including the quality, specificity, and contemporaneity of the source information.” 4 Final Redetermination, Remand P.R. Doc. 4 at 2. Commerce must “conduct a fair comparison of the data sets on the record” with regard to these factors. Allied Pac. Food (Dalian) Co. v. United States, 30 CIT -, -, 435 F.Supp.2d 1295, 1313-14 (2006) (emphasis added). That is, Commerce’s analysis must do more than simply identify flaws in the data sets it rejects. Commerce must also apply the same criteria to the data upon which it relies, and explain how the preferred data meet these criteria, or why a given criterion should not apply to the preferred data. The fact that a rejected data set is superior with respect to one criterion is not determinative so long as Commerce explains why the preferred data set is superior overall, and what steps were taken to ameliorate weaknesses in the preferred data. This case involves two proposed data sets, the Chemical Weekly data and the Indian government data. The court’s remand focused on Commerce’s treatment of two criteria with respect to these data sets. First, the court remanded for Commerce to consider the “quality” of the Indian government data, i.e., its reliability in light of the evidence on the record. Guangdong, 30 CIT at-, 414 F.Supp.2d at 1313 (“Having failed to consider whether the $15,826.30 figure derived from the basket category was aberrational despite evidence of its wide variation from the value of the same basket category in another year, Commerce failed to present substantial evidence supporting its surrogate value for sebacic acid.”). Second, the court remanded for Commerce to explain its choice not to use the more product-specific Chemical Weekly data set, in spite of Guang-dong’s corroborating data. Id. at -, 414 F.Supp.2d at 1312 (remanding for Commerce to reconsider its “departure] from its generally expressed preference for product-specific data” based on Guangdong’s submission of corroborating evidence). *1370 A. Commerce’s Comparison of the Data Sets on the Record Was Reasonable On remand, Commerce considered Guangdong’s corroborating evidence and its implications for the “quality” of the Indian government data, which it had previously ignored. See Final Redetermination, Remand P.R. Doc. 4 at 5 (“In our Final Results, we did not address these data points that Guangdong provided for benchmarking purposes.”). Commerce determined that Guangdong’s corroborating evidence raised questions regarding the quality of the Indian government data. Id., Remand P.R. Doc. 4 at 5 (“[W]e find that the period of review ... average sebacic acid surrogate value from the Indian six-digit HTS category ... is significantly higher than the average import value from the previous POR ... and higher than the data provided by Guangdong from the European Union import statistics, the U.S. import statistics, and the Chemical Market Reporter.”). Commerce chose to address this problem by excluding aberrational values. 5 Id., Remand P.R. Doc. 4 at 5-6. The elimination of aberrational values has been held to be a reasonable means for compensating for flaws in a data set. See Hebei Metals & Minerals Imp. & Exp. Corp. v. United States, Slip-Op. No. 04-88, 2004 WL 1615597, at *12 (CIT July 19, 2004) (ordering exclusion of aberrational values from one country to avoid distortions in the overall value for a specific import category); Issues & Decision Memorandum for Final Determination in Steel Wire Rope from the People’s Republic of China, A-570-859 (Feb. 14, 2001), available at http://ia.ita.doc.gov/frn/summary/ prc/01-4895-l.txt (stating that Commerce “has excluded — where appropriate — aberrational data that appear to distort the overall value for a specific import category”). The court finds that Commerce’s elimination of aberrational values constituted a reasonable step to compensate for some weaknesses in the Indian government data based on the evidence in the record. Having adjusted the Indian government data, Commerce then performed a comparison of the Indian government data with the Chemical Weekly data offered by Guangdong. Commerce acknowledged that “it may appear that the eight-digit category developed by ChemlmpEx is more specific than the six-digit [HTS] category,” Final Redetermination, Remand P.R. Doc. 4 at 4, and also that the price of azelaic acid was, on average, about 18.75% higher than the price for sebacic acid during the POR. Id. at 6. Despite these relative strengths of the Chemical Weekly data, Commerce found that, on balance, the Indian government data were the best available information on the record. Commerce also found that Guangdong’s corroborating information “d[id] not remedy the deficiencies in quality or the limited number of data points in the [Chemical Weekly] data provided by Guangdong for sebacic acid.” Final Redetermination, Remand P.R. Doc. 4 at 5. Commerce rejected the Chemical Weekly data primarily for two reasons. First, Commerce found that it was unclear how the data reported in Chemical Weekly’s ChemlmpEx database were selected. Although the parties appear to agree that the Chemical Weekly data were developed using information obtained from the Indian government, the Chemical Weekly data did not use all of the available data. See Pl.’s R. 56.2 Mot. J. Agency Record 16 (stating that the Chemical Weekly data *1371 were “derived from ship manifest data collected by Indian Customs authorities”); Final Redetermination, Remand P.R. Doc. 4 at 4 (stating that the Chemical Weekly data were “derived from the Daily Lists published by the customs authorities in India,” and noting that “ChemlmpEx does not provide the methodology on how the data was selected or from where the data [were] derived”). Only about half of the sebacic and azelaic acid imported into India is represented in the statistics from Chemical Weekly’s database. Final Rede-termination, Remand P.R. Doc. 4 at 4. Commerce found no information on the record showing why certain imports were included, while other imports were not. Consequently, Commerce could not be sure that the data were “truly representative of the full data set from which [they were] derived.” Id., Remand P.R. Doc. 4 at 4. A lack of information regarding the selection of data in a data set raises concerns distinct from concerns raised by the size of that data set. Without information on how transactions were chosen for inclusion in the Chemical Weekly data, Commerce could not be certain that the method used to select imports in the Chemical Weekly data was not biased. Second, Commerce noted that the Chemical Weekly data included only two data points, consisting of two sales from the same company in Germany to India. Id., Remand P.R. Doc. 4 at 4. By contrast, the Indian government data, even after removing aberrational values, represented imports from five different countries. Id., Remand P.R. Doc. 4 at 6. The use of broader product categories is reasonable, despite the availability of product-specific data, if a greater variety of data provides greater reliability. See Writing Instrument Mfrs. Ass’n v. U.S. Dep’t Commerce, 21 CIT 1185, 1195-96, 984 F.Supp. 629, 639-40 (1997) (approving use of basket category of import statistics from Pakistan, rather than more product-specific data from India, where Commerce found substantial evidence on the record suggesting that the Indian data were “aberrational and unreliable”). The court therefore finds that Commerce’s decision to use the Indian government data is supported by substantial evidence. 6 B. Guangdong’s Additional Arguments Guangdong claims that Commerce’s decision merely to adjust the Indian government data cannot be reasonable in view of the evidence Guangdong submitted showing prices of $32,045.58 for azelaic acid and $3,551.73 for sebacic acid. Pl.’s Comments at 4; see also PL’s R. 56.2 Mot. J. Agency Record at Attach. 1 (submitting data for azelaic acid). Guangdong argues that the presence of azelaic acid in the Indian government data must have skewed the price for sebacic acid upwards because the price of azelaic acid was almost ten-times that of sebacic acid. PL’s Comments at 4. This argument assumes that the price of azelaic acid shown in the Chemical Weekly data is reliable. Commerce has explained that the prices in the Chemical Weekly data are unreliable because they do not account for all imports of sebacic acid or azelaic acid into India, and it is unclear how the selected data were chosen. Final Redetermination, Remand P.R. Doc. 4 at 4 (“Although the data [in Chemical Weekly’s Chemicals import/export database] was originally derived from the Daily Lists published by the customs authorities in India, using a classification system that has been developed by Chemical Weekly, [it] does not provide the methodology on *1372 how the data was selected or from where the data was derived.”). Moreover, the evidence on the record supporting Guang-dong’s asserted price for azelaic acid is weaker than the evidence supporting Gu-angdong’s asserted price for sebacic acid. Guangdong has not presented any evidence corroborating the Chemical Weekly data’s price for azelaic acid. Evidence submitted by Guangdong shows that the price of azelaic acid exported to India from the United States was between four and eight times greater than the price of other azelaic acid imported by India in the same period. See Pl.’s R. 56.2 Mot. J. Agency Record at Attach. 1 (showing U.S. export price of 2855 rupees per kilogram of azelaic acid, as compared to 316 rupees per kilogram from Malaysia and 636 rupees per kilogram from Japan). In fact, as Guangdong points out, other evidence in the record shows that prices for azelaic acid in the United States were substantially lower than the $32,045.58 found in the Chemical Weekly data. Pl.’s Comments at 6 (“Commerce’s own analysis shows that U.S. prices for both azelaic acid and seba-cic acid are priced BELOW $3,000 per metric ton.”). Finally, Commerce conducted its own analysis of the prices of azelaic acid and sebacic acid, and found a much smaller variation. See Final Redetermi-nation, Remand P.R. Doc. 4 at 6, Attach. 1 (finding that the U.S. prices of sebacic acid were on average only 18.75 percent lower than those of azelaic acid during the POR). Commerce was therefore justified in concluding that inclusion of azelaic acid in the Indian government data did not skew the surrogate value of sebacic acid as much as Guangdong claims. In a similar argument, Guangdong attacks Commerce’s comparison of U.S. prices of azelaic and sebacic acid to establish the average variance in price between the two products. PL’s Comments at 5-6. Guangdong argues that Commerce has not adequately explained why it did not rely on the Indian prices from Chemical Weekly for the purpose of comparing the respective prices of sebacic and azelaic acid instead. Id. The impact of this argument is blunted by the absence of evidence corroborating the price of azelaic acid found in the Chemical Weekly data. Given the absence of evidence corroborating Guang-dong’s price, and the aberrationally high price of azelaic acid exported from the United States, it was reasonable for Commerce to conclude that U.S. domestic price data were the more appropriate benchmark for comparison. See Timken Co. v. United States, 26 CIT 434, 446-47, 201 F.Supp.2d 1316, 1328 (2002) (stating that use of U.S. data as a benchmark to “determine the reliability of ... surrogate data is within ‘Commerce’s statutory authority and consistent with past practice.’ ” (quoting Peer Bearing Co. v. United States, 22 CIT 472, 481, 12 F.Supp.2d 445, 455 (1998))). Finally, Guangdong argues that Commerce’s surrogate value of sebacic acid cannot be reasonable because it exceeds the prices reflected in Guangdong’s corroborating data. PL’s Comments at 5. The mere fact that Commerce’s surrogate value is higher than one or all of Guangdong’s benchmarks does not mean that that value is unreasonable per se. Guangdong notes that Commerce’s surrogate value is 60 percent higher than U.S. imports of sebacic acid, 58 percent higher than European Union imports of sebacic acid, 17 percent higher than price in the Chemical Market Reporter data, and 38 percent higher than Guangdong’s proposed surrogate value. Id. Still, Guangdong’s own corroborating evidence exhibits similar levels of variation. For example, the value found in the Chemical Market Reporter data ($4,187.60) is 36.7 percent greater than the value established by the U.S. import prices, and 35. 1 percent greater than the value of the imports into the European Union. See id, Moreover, Commerce’s price is lower than *1373 the surrogate values for sebacic acid found in other administrative proceedings. See Issues & Decision Memorandum for Final Results of Changed Circumstances Review in Sebacic Acid from the People’s Republic of China, A-570-825, at 17 (Mar. 23, 2005), available at http://ia.ita.doc.gov/frn/ summary/prc/E5-1401-l.pdf (finding surrogate value of sebacic acid to be $5,459.72 during the POR of July 1, 2002, to June 30, 2003, and noting a surrogate value of $5,388.66 for the administrative review conducted between July 1, 2000 and June 30, 2001, after adjusting for inflation). Given the variation among the corroborating data, the court finds that Commerce’s surrogate value is not unreasonably high. The court therefore affirms as reasonable Commerce’s analysis of the reliability of the Indian government data in view of the corroborating evidence submitted by Guangdong. III. Commerce’s Application of the ByProduct Credit to Normal Value The remaining issue in this case arises from Commerce’s treatment of byproduct revenue from sales of fatty acid and glycerine made in the process of manufacturing sebacic acid. Congress has mandated that, in cases involving imports from non-market economies, Commerce must calculate the normal value of a respondent’s factors of production using a surrogate data source from a country of similar size and economic development. 19 U.S.C. § 1677b(c)(4). The law requires Commerce to calculate normal value based on a number of factors of production, including labor, raw materials, energy used, and the cost of capital. Id. § 1677b(c)(3). After determining the costs of these materials, Commerce must also add “an amount for general expenses and profit plus ... other expenses.” Id. § 1677b(c)(1). These general expenses are calculated using “financial ratios” based on a surrogate’s overhead costs, SG & A, and profits. See Goldlink Indus., 30 CIT at-, 431 F.Supp.2d at 1333. In this case, Commence derived these ratios using data from the Reserve Bank of India Bulletin. See Prelim. Valuation of Factors of Prod., P.R. Doc. 47 at 8. Commerce calculated the “overhead ratio” by dividing total factory overhead by the cost of “direct items” (including, inter alia, raw materials, power and labor). See id. at Attach. 8. For SG & A, Commerce divided total SG & A expenses by the sum of direct items and factory overhead. See id., P.R. Doc. 47 at 8. Finally, to generate the profit ratio, Commerce divided the amount of pre-tax profits by the sum of direct items, factory overhead and SG & A. See id., P.R. Doc. 47 at 8. These ratios were then applied to the respondent’s surrogate values to determine the amount of overhead, SG & A and profits. See generally Hebei Metals & Minerals Imp. & Exp. Corp. v. United States, 29 CIT --,-, 366 F.Supp.2d 1264, 1277 n. 7 (2005) (describing calculation of financial ratios). 19 U.S.C. § 1677b(c) does not mention the treatment of by-products, nonetheless, Commerce sometimes grants a respondent a “credit” for a “by-product ... generated in the manufacturing process [that is] either reintroduced into production or sold for revenue.” Final Redetermination, Remand P.R. Doc. 4 at 7. Both by-products in this case are sold for revenue. Id., Remand P.R. Doc. 4 at 12. In its preliminary results, Commerce deducted by-product revenues from manufacturing costs, before applying the financial ratios. Id., Remand P.R. Doc. 4 at 7. In its final results, however, Commerce determined that the byproduct credit should have been deducted from normal value, after calculation of overhead, SG & A and profit amounts based on the cost of manufacturing. Id., Remand P.R. Doc. 4 at 7. In the past, Commerce’s practice was to apply by-product credits against the manu- *1374 factoring costs of the respondent, prior to the calculation of overhead, SG & A and profits. See id., Remand P.R. Doc. 4 at 7; see also Union Camp Corp. v. United States, 22 CIT 267, 270, 8 F.Supp.2d 842, 846 (1998) (noting that in its 1996 anti-dumping administrative review of sebacic acid from China, Commerce subtracted byproduct sales revenues from the manufacturing cost of sebacic acid). Because overhead, SG & A and profits are calculated based on manufacturing costs, a reduction in manufacturing costs reduces overhead, SG & A and profit amounts as well. Commerce recently adopted a new policy with respect to by-product credits. Final Re-determination, Remand P.R. Doc. 4 at 7-8. Commerce now looks to the financial statement of the company (or companies) used to calculate surrogate value and applies the by-product credit in the same manner as the surrogate does. Id., Remand P.R. Doc. 4 at 7-8. In the event that the surrogate financial statement does not state how by-product revenue is applied, Commerce will “consider other information on the record, such as whether the byproduct was re-introduced into the production process or sold for revenue purposes.” Id., Remand P.R. Doc. 4 at 8. In this case, Commerce found that deducting the byproduct credit from normal value, after applying the financial ratios, was “appropriate ... because it is reflective of the respondent’s practice to sell the by-product as opposed to reintroducing it into the production process.” Id., Remand P.R. Doc. 4 at 8. This methodology does not reduce manufacturing costs prior to the calculation of overhead, SG & A and profit amounts, which in turn results in a higher normal value and dumping margin. 7 Gu-angdong argues that “[t]here is no rational basis for Commerce’s departure from its longstanding administrative practice of applying the by-product offset as an adjustment to production costs.” Pl.’s Comments 11. Commerce claims that this methodology was approved as reasonable in Sinopec Sichuan Vinylon Works v. United States, 29 CIT -, -, 366 F.Supp.2d 1339, 1351 (2005). Sinopec involved a respondent, Sinopec Sichuan Vinylon Works (“SVW”) that produced polyvinyl alcohol (“PVA”) in China. Id. at -, 366 F.Supp.2d at 1340. Commerce chose a company in India, Jubilant, which produced a precursor to PVA, polyvinyl acetate (“PVAc”), to act as a surrogate. Id. at-, 366 F.Supp.2d at 1340-41. SVW produced a by-product, acetic acid, when it converted PVAc into PVA. Id. at-, 366 F.Supp.2d at 1341. SVW “recovered] and reuse[d]” this by-product in its production process. Id. at --, 366 F.Supp.2d at 1351. Because Jubilant did not produce PVA, it did not produce acetic acid as a byproduct of converting PVAc into PVA. 8 Id. Commerce determined that SVW should receive a by-product credit for producing acetic acid, but because Jubilant did not produce acetic acid as a by-product of converting PVAc to PVA, Commerce determined that it should not apply the byproduct credit before applying Jubilant’s *1375 financial ratios to the cost of manufacture. Id. at -, 366 F.Supp.2d at 1349-50. Commerce’s practice was not intended to account for additional costs associated with the production of acetic acid, however, but to ensure that deduction of the by-product credit did not artificially distort the overhead, SG & A and profit expenses associated with the production of PYAc. Sinopec, 29 CIT at -, 366 F.Supp.2d at 1348 (stating that Commerce’s determination to apply the by-product credit after the financial ratios was intended “to equate the base on which the ratios were calculated with the base to which they were applied”). Commerce’s reasoning in Sinopec was entirely different from its reasoning in the Final Redetermination. In this case, Commerce never suggested that the application of financial ratios to cost of manufacturing data before deduction by-product revenues would mischaracterize Heng-shui’s cost of manufacturing sebacic acid. Rather, Commerce applied the by-product credit after the financial ratios to reflect the fact that, where a by-product is sold, “the by-product necessarily incurs expenses for overhead, SG & A, and profit.” Final Redetermination, Remand P.R. Doc. 4 at 12. Thus, the reasoning behind Sino-pec does not support Commerce’s determination in this case. Turning to Commerce’s explanation in the Final Redetermination, the court finds that the remainder of Commerce’s analysis provides a reasonable basis for its decision to apply by-product credits after a surrogate’s financial ratios where the by-product is sold. Hengshui sells two by-products, revenue from which may be used to offset the cost of producing sebacic acid. Where a by-product is sold, Commerce assumes that the respondent would incur overhead, SG & A and profit expenses in selling the by-product. Final Redetermi-nation, Remand P.R. Doc. 4 at 12. The Reserve Bank of India statistics, from which Commerce derived its surrogate financial ratios, include “selling commission[s],” “bad debts,” and advertising as sales expenses. See Prelim. Valuation of Factors of Prod. Mem., P.R. Doc. 47 at Attach. 8. A respondent’s sales of a byproduct would appear to incur each of these costs over and above what a surrogate spends to sell its primary products. If the surrogate does not produce a similar by-product, it would be reasonable for Commerce to conclude that the surrogate would not incur these expenses. Therefore, it is reasonable for Commerce to adjust a by-product credit to reflect the additional sales expenses incurred by the respondent. Guangdong argues that Commerce could not reasonably have chosen to account for separable costs associated with the sale of a by-product by changing the point at which it applies the by-product credit. Guangdong notes that, as a matter of accounting procedure, by-products are commonly subtracted from the cost of manufacturing a main product. See Pl.’s Comments at 11 (“[Generally accepted accounting principles ... normally treat both by-product income and by-products consumed in the production process as offsets to manufacturing costs.”) (emphasis removed); see also Charles T. Horngren & George Foster, Cost Accounting: A Managerial Emphasis 490 (6th ed. 1987) (“The estimated net realizable values of [by-products and scrap] are best treated as deductions from the cost of the main products.”) (emphasis removed). Nevertheless, it appears that “Considerable variation exists in accounting for by-products.” Wayne J. Morse & Harold P. Roth, Cost Accounting 157 (3d ed.1986). Indeed, in some circumstances, by-product sales may be credited to miscellaneous income. Id. at 158. The court’s opinion in Magnesium Corp. of America v. United States, 20 CIT 1092, *1376 1107-08, 938 F.Supp. 885, 900 (1996), supports Commerce’s treatment of the byproduct credit. In that case, in its preliminary determination, Commerce subtracted a by-product credit from the respondent’s cost of materials, before calculating the cost of manufacturing. 9 Id. at 1106, 938 F.Supp. at 899. In its final determination, Commerce changed its practice and applied the by-product credit after calculating the cost of manufacturing, thus increasing the amount of factory overhead. Id. Commerce did this to reflect “the byproduct processing costs, thereby eliminating the need for valuing any additional processing-related elements.” Id. (quotations omitted). Plaintiff argued that separable by-product processing costs should have been deducted from by-product revenues, and that the by-product revenue should have been deducted before calculating manufacturing cost. Id. The court disagreed, finding that Commerce’s decision to change the timing of the application of the by-product credit was a reasonable means of “accounting] for ... costs related to by-product processing” while avoiding “costly accounting procedures” not warranted for by-products. Id. at 1107, 938 F.Supp. at 900. Similarly, Guang-dong’s argument implies that Commerce should have calculated a separate overhead, SG & A and profit amount for Heng-shui’s by-products, deducted that amount from the by-product credit, and then deducted the remaining by-product credit from manufacturing costs. This would require Commerce to engage in just the “costly accounting procedures” that the court in Magnesium Corp. found to be unnecessary. As in Magnesium Corp., Commerce’s decision to change when it applies the by-product credit is a reasonable alternative means of accounting for additional overhead, SG & A and profit expenses associated with Hengshui’s sale of by-products. Even if Guangdong’s alternative approach to implementation of the statute were reasonable, the court could not substitute its own view of the statute for Commerce’s reasonable interpretation or implementation. Id. (citing Chevron, U.S.A. v. Nat. Res. Def. Council, Inc., 467 U.S. 837, 844, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984)). Therefore, the court upholds Commerce’s decision to account for separable costs associated with by-product sales by applying a by-product credit after application of financial ratios to manufacturing costs. IV. Conclusion The results of the remand determination are sustained in their entirety. Footnotes [1] The six-digit Indian HTS heading is 291713. [2] The eight-digit heading for sebacic acid is 291713.02. [3] All prices are in U.S. dollars per metric ton unless otherwise stated. [4] In its preliminary determination, Commerce stated that it will select, "where possible, the publicly available value which was (l)[a]n average non-export value; (2) representative of a range of prices within the POR or most contemporaneous with the POR; (3) product-specific; and (4) tax exclusive.” Sebacic Acid from the People’s Republic of China, 69 Fed. Reg. 47,409, 47,411 (Dep’t Commerce Aug. 5, 2004) (preliminary results of antidumping duty administrative review and notice of partial rescission). [5] Specifically, Commerce excluded imports from the United States, the price of which was ten-times greater than the price of imports from other countries during the POR. Final Redetermination at 5. [6] Because the court finds that these reasons are sufficient to justify Commerce's choice of data sets, the court does not address Commerce’s arguments concerning imports from Malaysia and the purity of sebacic acid imports in the Chemical Weekly data. [7] For example, assume that a respondent has manufacturing costs of $1000, financial ratios of 20%, and receives a by-product credit of $100. Using Commerce’s old methodology, that company would have a normal value of $1080. ($1000 — $100 = $900; $900 + ($900 x .2) = $1080). Using Commerce's new methodology, and assuming the by-product was sold, that company would have a normal value of $1100. ($1000 + ($1000 x .2) = $1200; $1200-$ 100 = $1100). The dispute in this case does not involve the size of the byproduct credit itself, but whether a respondent should receive the added benefit ($20 in the example above) associated with reduced overhead, SG & A and profit amounts. [8] Jubilant produced acetic acid as a by-product, but not at the relevant stage of production. Id. at -, 366 F.Supp.2d at 1351. [9] Cost of manufacturing is composed of the cost of materials plus factory overhead, which is calculated by multiplying cost of materials by a factory overhead ratio. See Magnesium Corp., 20 CIT at 1106, 938 F.Supp. at 899.