Home Meridian International, Inc. v. United States

OPINION AND ORDER RESTANI, Judge: This matter is before the court following a remand to the Department of Commerce (“Commerce”) in Home Meridian Int’l, Inc. v. United States, 865 F.Supp.2d 1311 (CIT 2012). This case involves challenges to Commerce’s final results in the fifth antidumping duty (“AD”) review of certain wooden bedroom furniture (‘WBF”) from the People’s Republic of China (“PRC”). See Wooden Bedroom Furniture from the People’s Republic of China: Final Results and Final Rescission in Part, 76 Fed.Reg. 49,729, 49,729 (Dep’t Commerce Aug. 11, 2011). The court determines that for the reasons below, Commerce failed to comply with the court’s remand order with respect to the valuation of Huafeng Furniture Group Co., Ltd.’s (“Huafeng”) factors of production (“FOPs”) and the use of Insular Rattan and Native Products’ (“Insular Rattan”) financial statement. BACKGROUND . The court assumes familiarity with the facts of this case as set out in the previous opinion, although they are summarized briefly below. See generally Home Meridian, 865 F.Supp.2d at 1315-20, 1326-27. In its previous order, the court instructed Commerce to address six issues raised by Plaintiffs and Intervenor Defendants in their motions for judgment on the agency record. Specifically, the court ordered Commerce to: 1) reconsider whether surrogate values or market-economy (“ME”) purchases should be used in valuing Huafeng’s FOPs for wood inputs; 2) reclassify Huafeng’s poly foam input; 3) explain its reliance on 2008 gross-national income (“GNI”) data for labor wage rates; 4) support its finding that Insular Rattan’s financial statements are acceptable for financial ratio calculations; 5) investigate whether combination rates are proper; and 6) explain its differing use of zeroing in administrative reviews and investigations. See Home Meridian, 865 F.Supp.2d at 1332. On remand, Commerce: 1) continued to rely upon surrogate values to calculate normal value based on Huafeng’s FOPs; 2) reclassified poly foam input as cellular plastic; 3) continued to rely on 2008 GNI data in calculating labor wage rates; 4) found Insular Rattan’s financial statements to be reliable and acceptable; 5) determined that combination rates were not appropriate; and 6) explained its use of zeroing in reviews. See Final Results of Redetermination Pursuant to Court Order (Dep’t Commerce Feb. 25, 2013) (“Remand Results ”), Dkt. No. 97. Plaintiffs Home Meridian International, Inc. and Import Services, Inc., as well as Consolidated Plaintiffs Great Rich (HK) Enterprises Co., Ltd. and Dongguan Liaobushangdun Huada Furniture Factory (collectively “HMI”), continue to challenge Commerce’s decision to use certain surrogate values. Plaintiffs argue that Huafeng’s pre-period of review ME input purchases must be used to value Huafeng’s FOPs. See Cmts. of Home Meridian Int’l, Inc. d/b/a Samuel Lawrence Furniture Co. and Pulaski Furniture Co.; Import Servs., Inc.; Great Rich (HK) Enters. Co., Ltd.; & Dongguan Liaobushangdun Huada Furniture Factory on Dep’t of Commerce Feb. 25, 2013 Final Results of Redetermination Pursuant to Ct. Order (“HMI Cmts.”) 2-29. HMI also contends that Commerce' has perpetuated a ministerial error in its *1370 surrogate value calculations. Id. at 30. Intervenor Defendants American Furniture Manufacturers Committee for Legal Trade and Vaughan-Bassett Furniture Co., Inc. (collectively “AFMC”) argue that Commerce’s reliance upon Insular Rattan’s financial statement is unsupported by substantial evidence and contrary to- agency practice. See AFMC’s Cmts. Concerning Commerce’s Final Results of Redetermination Pursuant to Court Remand (“AFMC Cmts-.”) 2-7. Defendant United States responds that Commerce’s determinations on both issues are supported by substantial evidence and in accordance with law. See • Def.’s Resp. to Pis.’ Remand Cmts.-(“Def.’s Resp.”) 3-17. 1 JURISDICTION AND STANDARD OF REVIEW The court has jurisdiction pursuant to 28 U.S.C. § 1581(c). The court will not uphold a determination by Commerce if it is “unsupported by substantial evidence on the record, or otherwise not in accordance with law.” 19 U.S.C. § 1516a(b)(l)(B)(i). DISCUSSION I. Value of Huafeng’s Wood Inputs HMI contends that Commerce violated the applicable statutory and regulatory framework when it used surrogate values to calculate the normal value of Huafeng’s products. HMI Cmts. 3-29. HMI insists that Commerce was required to use Huafeng’s ME purchases, all of which were made prior • to the period of review (“POR”). Id. Additionally, HMI asserts that even if Commerce’s methodology were permitted by the applicable statute and regulation, substantial evidence fails to support Commerce’s selection of surrogate values as the best information available when compared with Huafeng’s ME purchase prices. Id. Defendant asserts that Commerce has a reasonable practice of not using pre-POR ME input purchases in calculating normal value and that the chosen surrogate values constituted the best available information on the record. Def.’s Resp. 3-12. The court concludes that HMI’s interpretations of the applicable statute and Commerce regulation are not mandated, because both the statute and regulation are ambiguous. HMI’s claim that Commerce lacked substantial evidence to support its decision, however, has merit. In non-market economy- (“NME”) 2 AD cases, 3 Commerce “shall determine the normal value of the subject merchandise on the basis of the value of the factors of production utilized in producing the merchandise.” 4 19 U.S.C. § 1677b(c)(l). *1371 Among other costs, the factors of production include “quantities of raw materials employed.” Id. § 1677b(c)(3). In calculating normal value, ■ “the valuation of the factors of production shall be based on the best available information regarding the values of such factors in a market economy country or countries considered to be appropriate by the administering authority.” Id. § 1677b(e)(l). Furthermore, Commerce “shall utilize, to the extent possible, the prices or costs of factors of production in one or more market economy countries that are — (A) at a level of economic development comparable to that of the nonmarket economy country, and (B) significant producers of comparable merchandise.” Id. § 1677b(c)(4). 5 “Nowhere does the statute speak directly to any methodology Commerce must employ to value the factors of production, indeed the very structure of the statute suggests Congress intended to vest discretion in Commerce by providing only a' framework within which to work.” Shakeproof Assembly Components Div. of Ill. Tool Works, Inc. v. United States, 59 F.Supp.2d 1354, 1357 (CIT 1999); see QVD Food Co. v. United States, 658 F.3d 1318, 1323 (Fed.Cir.2011) (recognizing that Commerce is entitled to deference in interpreting the undefined term “best available information”). Nonetheless, selection of the best available information must be in line with the overall purpose of the antidumping statute, which the Federal Circuit has explained to be “determining current margins as accurately as possible.” Rhone Poulenc, Inc. v. United States, 899 F.2d 1185, 1191 (Fed.Cir.1990); see also Lasko Metal Prods., Inc. v. United States, 43 F.3d 1442, 1443 (Fed.Cir.1994) (“Lasko II ”) (“[T]here is much in thé statuté that supports the notion that it is Commerce’s duty to determine margins as accurately as possible, and to use the best information available to it in doing so.”). In calculating normal value in the NME context, the particular aim of the statute is to determine the non-distorted cost of producing such goods. See Lasko Metal Prods., Inc. v. United States, 810 F.Supp. 314, 316-17 (CIT 1992) (“Lasko I”). A. Commerce’s Methodology In applying this framework, Commerce has confronted situations in which an NME producer sourced its inputs from an *1372 ME supplier, paying for the goods in a convertible, ME currency. The court previously has held that the statute is silent regarding the methodology that Commerce must use under these circumstances; it also has held, however, that an interpretation prohibiting Commerce from considering actual prices paid by the producer, while possible, would conflict with the statute’s purpose. Lasko I, 810 F.Supp. at 317-18; see also Lasko II, 43 F.3d at 1446 (“In this case, the best available information on what the supplies used by the Chinese manufacturers would cost in a market economy country was the price charged for those supplies on the international market.”). To account for this gap in the law, Commerce has promulgated a regulation: The Secretary normally will use publicly available. information to value factors. However, where a factor is purchased from a market economy supplier and paid for in a market economy currency, the Secretary normally will use the price paid to the market economy supplier. In those instances where a portion of the factor is purchased from a market economy supplier and the remainder from a nonmarket economy supplier, the Secretary normally will value the factor using the price paid to the market economy supplier. 19 C.F.R. § 351.408(c)(1) (2012). Additionally, Commerce has established a general hierarchy for selecting the best available information that it will use to value an entity’s FOPs: We generally seek to value factors using (in order of preference): (1) Prices paid by the NME manufacturer for items imported from a market economy; (2) prices in the primary surrogate country of domestically produced or imported materials; (3) prices in one or more secondary surrogate countries reported by the industry producing subject merchandise in the secondary country .or countries; and (4) prices in one or more secondary surrogate countries from sources other than the industry producing the subject merchandise. Final Determination of Sales at Less Than Fair Value: Sparklers From, the People’s Republic of China, 56 Fed.Reg. 20,588, 20,590 (Dep’t Commerce May 6, 1991). In fact, Commerce has gone even further in stating that “using surrogate values when market-based values are available would, in fact, be contrary to the intent of the law.” Final Determinations of Sales at Less Than Fair Value: Oscillating Fans and Ceiling Fans From the People’s Republic of China, 56 Fed.Reg. 55,271, 55,275 (Dep’t Commerce Oct. 25, 1991) (“Oscillating Fans ”). Commerce has further refined the meaning of when “[Commerce] normally will use” ME supplier data, as stated in 19 C.F.R. § 351.408(c)(1). Specifically, Commerce will “disregard[] the prices of inputs that could not possibly have been used in the production of subject merchandise during the period of investigation or review.” Antidumping Methodologies: Market Economy Inputs, Expected NomMarket Economy Wages, Duty Drawback; and Request for Comments, 71 Fed.Reg. 61,716, 61,716 (Dep’t Commerce Oct. 19, 2006) {“AD Methodologies ”). This focus on inputs used in production is in line with Commerce’s preference of using contemporaneous ME purchases because Commerce “presumes that a factor purchased and paid for from an ME supplier is used by the respondent during that period.” Issues and Decision Memorandum for the Antidumping Duty Investigation of Certain Frozen and Canned Warmwater Shrimp from the Socialist Republic of Vietnam, A-552-802, POI: 4/1/03-9/30/03 (Nov. 29, 2004), Cmt. 8 (emphasis added), available at http://ia.ita.doc.gov/frn/ summary/vietnam/04-26977-l.pdf (last visited June 25, 2013). Additionally, Commerce has previously rejected non-con *1373 temporaneous ME purchases in favor of contemporaneous surrogate values, at least when non-contemporaneous ME inputs did not comprise 100 percent of the inputs actually used in production during the POR. See Home Meridian, 865 F.Supp.2d at 1317 n. 5 (collecting Commerce determinations). Where, however, the NME producer sources at least 33 percent of its inputs from an ME supplier during the POR, Commerce will use the weighted average of those actual purchases to value the entire input, including purchases from NME suppliers. AD Methodologies, 71 Fed.Reg. at 61,717-18 (discussing extensively what percentage of input purchases are necessary to render the prices “meaningful” in valuing a company’s entire input purchase). 6 In summary, Commerce will normally use ME purchases to value inputs provided that the purchases were made and the inputs were used during the POR. Where only surrogate values are being compared, rather than choosing between surrogate and ME values, Commerce has a practice of selecting the best available information based on several factors including contemporaneity, specificity, and tax-exclusivity. See Remand Results at 6. The court previously has upheld certain applications of Commerce’s so-called mix- and-match methodology, finding that “[t]he cost for raw materials from a market economy supplier, paid in convertible currencies, provides Commerce with the closest approximation of the cost of producing the goods in a market economy country.” Lasko I, 810 F.Supp. at 317 (rejecting a claim that Commerce was required by the statute to use only surrogate values); s,ee also Lasko II, 43 F.3d at 1445 (contrasting ME prices and surrogate values, the latter of which Commerce uses in its “factors of production methodology to estimate [normal value] for the merchandise in question” (emphasis in original)). In explaining, why the methodology was reasonable, the court noted that the “[c]osts of production in a surrogate country will never perfectly approximate what the costs in an NME would be in an open market situation. Any inaccuracies in the mix-and-match approach likely stem from surrogate values analysis, not Commerce’s attempt to improve upon it” by using purchase prices from ME suppliers. Lasko I, 810 F.Supp. at 317 n. 4; see also Shakeproof Assembly Components Div. of Ill. Tool Works, Inc. v. United States, 268 F.3d 1376, 1382-83 (Fed.Cir.2001) (recognizing that because surrogate values are at best estimates, the best available information for valuing domestic inputs was the actual purchase price of imported ME inputs). The court, however, has sustained Commerce’s practice of favoring surrogate values that are contemporaneous with the POR, all other things being equal. 7 See *1374 Shakeproof Assembly Components Div. of Ill. Tool Works, Inc. v. United States, 30 CIT 1173, 1177, 1179, 2006 WL 2457626 (2006) (“Shakeproof III ”). HMI argues that the text of 19 U.S.C. § 1677b(c)(l) unambiguously prohibits the application of Commerce’s methodology in this case. The statute requires Commerce to base normal value on “the factors of production utilized in producing the merchandise.” § 1677b(e)(l) (emphasis added). HMI contends that this compels Commerce to use actual price data to the exclusion of any surrogate values when pre-POR inputs are “utilized” in production. See HMI Cmts. 4. Although “utilized” could be interpreted to require Commerce to rely on actual costs of the specific inputs used in production by the particular producer, this portion of the statute is ambiguous. Furthermore, HMI’s reading seems inconsistent with the rest of the subsection. For example, § 1677b(c)(3) defines “the factors of production utilized in producing merchandise” in general terms such as labor, raw materials, energy costs, and capital costs. This supports the conclusion that the clause relied upon by HMI simply is directing Commerce to value the types of inputs actually used by the particular producer, without specifically addressing any temporal aspects with regard to cost. Even if this were not the case, Commerce’s interpretation of the ambiguous provision is reasonable, and therefore the court defers to that interpretation. Similarly, HMI’s argument based on Commerce’s regulation requiring it to “normally” use ME data is without merit. “Normally” is clearly an ambiguous term designed to allow Commerce discretion in uncontemplated abnormal situations. Commerce as the promulgating agency may reasonably interpret its meaning. See Decker v. Nw. Envtl. Def. Ctr., -U.S. -, 133 S.Ct. 1326, 1337, 185 L.Ed.2d 447 (2013). But see id. at 1338 (Roberts, C.J., concurring); id. at 1339 (Scalia, J., concurring in part and dissenting in part) (questioning the continued application of Auer deference). This is not to say, however, that the general thrust of HMI’s argument is inaccurate, as Commerce must still calculate dumping margins as accurately as possible and in line with the producer’s actual experience, employing the best available information on the record. 8 See Remand Results at 7 (acknowledging that the aim of using ME prices, where possible, in the NME context is “to capture the company’s actual POR experience with market prices”). B. Substantial Evidence Even though HMI unsuccessfully argues that the applicable statute and regulation unambiguously require Commerce to use ME purchase prices in all cases to value FOPs, HMI correctly asserts that Commerce must support the application of its methodology in this particular case with *1375 substantial evidence. . Commerce has failed to do so here. “[W]hile various methodologies are permitted by the statute, it is possible for the application of a particular methodology to be unreasonable in a given case. Form should be disregarded for substance and the emphasis should be on economic reality.” Yangzhou Bestpak Gifts & Crafts Co. v. United States, 716 F.3d 1370, 1378 (Fed.Cir.2013) (internal quotation marks, citations, and brackets omitted). For example, the court has rejected Commerce’s selection of surrogate values based solely on the fact that they are more contemporaneous and non-aberrational where Commerce has failed to consider whether the less contemporaneous data “[was] an accurate reflection of the price paid” for the input by the producer. See Yantai Oriental Juice Co. v. United States, 26 CIT 605, 617-18, 2002 WL 1347018 (2002) (remanding to Commerce for an explanation of why domestic data, adjusted for inflation or deflation, “would not more accurately approximate the experiences of’ producers during the POR); see also Olympia Indus. v. United States, 7 F.Supp.2d 997, 1001 (CIT 1998) (rejecting Commerce’s blanket policy as a justification for not considering NME trading company data instead of comparing it to alternative surrogate value data). In its previous opinion, the court explained that based on the record, “[i]f the only wood inputs into the subject merchandise were market economy inputs, contemporaneity would not outweigh all other factors,” especially in light of flaws in the selected surrogate values. Home Meridian, 865 F.Supp.2d at 1318-19. The court noted that “[u]sing the actual [purchase prices of the] inputs, if available and where they yield reliable values, would seem to promote accuracy more than does using flawed surrogate values.” Id. at 1319. The court also explained that “Commerce cannot create a blanket rule that prevents it from comparing the merits of contemporaneous and non-contemporaneous data, and thereby prevents Commerce from determining the best available information.” Id. Finally, the court required Commerce to consider the factual allegation by AFMC that Huafeng had not used the ME-sourced inputs that it claimed to have used, an allegation that Commerce did not previously address. Id. Because this factual question is a threshold issue, the court turns tb it first. 1. Factual Dispute as to Inputs Used The parties centinue to dispute HMI’s claim that 100 percent of the wood inputs used by Huafeng during the POR were purchased from an ME supplier. See HMI Cmts. 9-15; Def.’s Resp. 8-10. Ón remand, Commerce purportedly weighed the record evidence and determined that Huafeng had purchased its wood inputs from both ME and NME suppliers in the thirteen months prior to the POR, with the exception of two inputs that were purchased entirely from ME suppliers. Remand Results at 8. For all inputs, Commerce found that sufficient inventories of unknown origin existed, such that NME-supplied inputs could have been used by Huafeng to make all of the subject merchandise during the POR. Id. HMI continues to argue that it and Huafeng always have claimed that all of Huafeng’s wood inputs that were used during the POR were supplied by ME producers. HMI Cmts. 10-11. HMI also submits that these assertions were supported by Huafeng’s original and supplemental Section D questionnaire responses. Id. (citing Section D questionnaire responses). HMI claims that Commerce’s determination, is, at best, speculation and, at worst, the application of an impermissible adverse inference. Id. at 10-15. Defendant contends that Commerce based its decision on substantial evidence because HMI has failed *1376 to present sufficient evidence to prove its allegation. Def.’s Resp. 8-10. Commerce asserted in response to HMI’s challenges that “[rjather than identifying evidence that definitively shows that Huafeng’s consumption of lumber consisted entirely of ME-purchased lumber, Home Meridian relies on circumstantial evidence to support its claims.” Remand Results at 41 (citing Huafeng’s Section D supplemental responses). Commerce reached this conclusion despite the fact that it verified that Huafeng separated its inputs based on country of origin at the manufacturing site and segregated its workshops based on shipping destination; Commerce also did not find anything that contradicted Huafeng’s assertions as to the price or quantity of its ME purchases. Verification Report (Jan. 31, 2011) Bus. Proprietary App. to Mem. of Points and Auth. in Supp. of R. 56.2 Mot. for J. on the Agency Record by Pis. HMI (Feb. 29, 2012) (“HMI App.”), Tab 16, at 4-5, 26. Instead, Commerce merely noted that the handwritten material withdrawal slips did not specify the origin of the input when materials were withdrawn from supply and entered into production. See Remand Results at 41. This appears to be the piece of direct evidence Commerce required. See id. (“[T]here is no evidence on the record that ties either ME purchases or NME purchases of lumber to consumption during the POR.”) Although Commerce certainly has the authority to discredit Huafeng and HMI’s claims that a certain percentage of the wood inputs used were sourced from an ME supplier, such a decision must rest upon substantial evidence. Here, Commerce was presented with evidence of ME purchases, which it had an opportunity to verify, along with sworn statements by Huafeng that 100 percent of the inputs it used were purchased from an ME supplier. Commerce opted not to conduct an in-depth verification of these ME purchases because it considered the evidence irrelevant, but its failure to do so does not discredit Huafeng’s claim. 9 Indeed, as HMI has argued, some of Commerce’s verification report corroborates Huafeng’s assertions, such as the separation of ME and NME inputs at the work site, the division of workshops based on product destination, and the use of separate control numbers (“CONNUMS”) for products bound for the United States. See Verification Report (Jan. 31, 2011) HMI App., Tab 16, at 4-5, 15, 26. When presented with conflicting evidence that provides substantial evidence to support opposite conclusions, the court will defer to Commerce’s reasoned choice between the two. Where, however, as here, Commerce is presented with non-definitive but still substantial evidence to support one factual conclusion and zero evidence to the contrary, beyond mere speculation, the only reasonable choice for Commerce is the one for which evidence exists. 10 Accordingly, the court finds that Commerce lacked substantial evidence to discredit Huafeng and HMI’s claims that 100 percent of the inputs used in produc *1377 tion during the POR were purchased from an ME supplier. 11 2. Best Available Information In light of this conclusion, the court turns to Commerce’s decision as to what data set constituted the best available information. Although somewhat unclear, Commerce decided on remand that even if HMI’s factual allegations regarding the source of its inputs were true, Huafeng’s pre-POR purchase data still would not constitute the best available information on the record for valuing those inputs because the purchases were made during the year prior to the POR. Remand Results at 5-7. Commerce supported this determination by relying on its policy that it “normally calculates normal value by valuing the NME producers’ FOPs using prices from an ME that is at a comparable level of economic development and that is also a significant producer of corn-parable merchandise.” Id. at 5. Commerce then contended that it has created a narrow exception to the rule by permitting actual market economy purchase prices to be used rather than surrogate values when they are contemporaneous with the POR and the inputs are used within the POR. 12 Id. at 5. As set out above and in Home Meridian, selection of the best available information is, except in the rarest of cases, a balancing of competing options, not a binary decision made with respect to one data source in isolation. The selection cannot be made without consideration of the specific options available to Commerce based upon the particular facts on the record. Here, Commerce was presented with two potential sources for valuing Huafeng’s inputs, Huafeng’s nearly contemporaneous 13 but pre-POR ME purchases and contemporaneous surrogate values based on import data under three Philippine tariff subheadings. 14 *1378 Instead of comparing the substance and merits of these two options, Commerce again undertook a cursory, binary consideration. Commerce has not pointed to substantial evidence on the record to demonstrate that Huafeng’s input purchases are unreliable or aberrant, and it has not argued as much. At most, Commerce noted that a single supplier from the 2008 review paid higher input prices to a different supplier. Remand Results at 9. Commerce also failed to respond to any of HMI’s explanations differentiating its purchases from those of the other producer. HMI Cmts. 20-23. Commerce merely acknowledged that different suppliers may have different prices at different times. See Remand Results at 8-9. Even if the prices are pre-POR, Commerce has established a relatively simple way to adjust for the lack of contemporaneity of valuation data. It is able to add or subtract from the pre-POR prices an inflation or deflation adjustment to reflect changes in relative prices at the macroeconomic level. 15 Because this simple adjustment exists, if even needed here, Huafeng’s ME purchase prices are specific and easily convertible to reflect POR prices. In contrast, the court noted in Home Mendian its concerns with the surrogate values chosen by Commerce, especially with the basket tariff line used to value some inputs, as it is a basket subheading that includes many types of woods not used by Huafeng in its products. See 865 F.Supp.2d at 1319. On remand, Commerce partially corroborated the surrogate prices, which are largely based on imports from North American and Europe, with reference to a single respondent in 2008 in order to address the court’s concern regarding the large year-over-year increase in prices. 16 Remand Results at 9; see Surrogate Value Submission (Nov. 15, 2010), HMI App., Tab 19, Ex. 2 (listing the source countries underlying the Philippine import data). A partial corroboration of the surrogate values does not, however, trump Commerce’s stated preference for ME values as furthering statutory intent. Further, such corroboration does not address the concerns about the specificity of the basket tariff heading, 4407.99. 17 *1379 Although Commerce continues to cite to its general preference for contemporaneity among equally useful values, its- continued reliance on the court’s affirmation of that preference in Shakeproof III fails to recognize that such a preference may not be absolute; the preference is permissible only when there is “no dispute about the representativeness of Commerce’s chosen surrogate value.” Shakeproof III, 30 CIT at 1180 n. 7 (distinguishing Yantai, 26 CIT at 617). At any rate, the limited approval of such a policy in Shakeproof III is distinguishable from the unique facts of the present case where 100 percent of the inputs used were from pre-POR purchases. See id. Under the unique facts of this case, Commerce’s conclusion that its fictional surrogate values better reflect the actual experience of the producer, especially in light of the imprecision of surrogate values generally, is 'unsustainable. The court is unable to square Commerce’s asserted blanket policy of ignoring actual ME data unless the inputs are both purchased and used within the POR with Commerce’s statutory obligation to use the best available information on the record, and its previous statements about what kind of value data are the best available data. As noted supra, although Commerce reasonably may prefer both conditions be met in the abstract, such reasoning breaks down under the circumstances of the present case where no input purchases at all were made during the POR. Commerce’s stated preference for contemporaneous purchases is in part based on its presumption that those purchases will be used during the POR. Commerce has failed to explain its departure from this practice of focusing on when the inputs were used in production, especially where here 100 percent of the inputs were purchased from an ME supplier in contrast to the 33 percent threshold Commerce uses when considering whether to use contemporaneous ME purchases to value even the other 67 percent. 18 Additionally, statements by Commerce such as “[t]he true experience of the company during the POR is that it chose not to make ME purchases during that time frame” are unhelpful to the court. 19 Remand Results at 7. HMI’s proposed methodology does not seek to pretend that Huafeng made ME input purchases during the POR. Rather, it simply advocates that Commerce give consideration to the prices Huafeng actually paid. This advances Commerce’s own stated policies of using data that advances fairness, predictability, and accuracy. See Oscilating Fans, 56 Fed.Reg. at 55,275 (“Where we can determine that a NME producer’s input prices are market determined, accuracy, fairness, and predictability are enhanced by using those prices. Therefore, using surrogate values when market-based values are available would, in fact, be contrary to the *1380 intent of the law”). Commerce fails to recognize that its proposed valuation methodology, applied to the facts of this case, is far less reflective of Huafeng’s actual experience because it pretends that Huafeng purchased all of its inputs during the POR at the estimated surrogate values. This fails to gauge accurately whether Huafeng sold its products at dumped prices into the U.S. market. 20 Given the clear disparity between these two potential valuation methods and Commerce’s failure to provide any reasonable explanation of its preference for surrogate values on these facts, unless Commerce adequately supports a request to reopen this record to farther examine Huafeng’s wood purchases, Commerce must use Huafeng’s ME purchase prices to value wood FOPs in this case. Commerce may adjust the values for inflation or deflation, if needed, based on the time of purchase. 21 II. Insular Rattan Financial Statement AFMC argues that Commerce again erred in using the 2009 financial statement of Insular Rattan in calculating surrogate financial ratios. AFMC Cmts. 2-7. It argues first that Commerce’s finding that the statement is reliable and unaffected by subsidies is unsupported by substantial evidence. Id. at 2-5. It further argues that Commerce acted contrary to its prior practice by continuing to include the “questionable” financial statement in its calculations, despite the existence of eleven other reliable financial statements on the record. Id. at 5-6. Defendant acknowledges that the court remanded this same issue back to Commerce in another case, but nonetheless makes the same previously rejected arguments. Def.’s Resp. 3 n. 1; see Dongguan Sunrise Furniture Co. v. United States, 904 F.Supp.2d 1359, 1367-68 (CIT 2013). Defendant contends that Commerce did not blindly rely upon the “unqualified” statement by Insular Rattan’s auditor, but instead relied on the completeness of all aspects of the statement, other than the missing tax line. Def.’s Resp. 14. Defendant also claims that because Philippine generally accepted accounting principles require disclosures of subsidies, the lack of any such disclosure provides evidence that no countervailable subsidy was received by Insular Rattan. 22 Id. at 13-14. Commerce selected financial statements in this review based on whether the companies were: “producers of merchandise identical to subject merchandise which received no countervailable subsidies, and earned a before tax profit in 2009 for which [Commerce has] financial information.” Wooden Bedroom Furniture From the People’s Republic of China: Preliminary Results of Antidumping Duty Administrative Review and Intent To Rescind Review in Part, 76 Fed.Reg. 7534, *1381 7544 (Dep’t Commerce Feb. 10, 2011). Commerce also requires complete financial statements to ensure reliability. 23 See Dongguan Sunrise Furniture Co. v. United States, 865 F.Supp.2d 1216, 1243 (CIT 2012). On remand, Commerce continued to speculate as to how Insular Rattan’s financial statement, which facially conflicts with the applicable accounting standard, remains complete and reliable for use in surrogate financial ratio calculations. See Remand Results at 17-19; Resolution No. 24-00 (July 31, 2000) App. to AFMC’s Cmts. Concerning Commerce’s Final Results of Redetermination Pursuant to Ct. Remand, Tab. 2, at 21 (requiring financial statements to contain a separate tax line). The court already has held that the financial statement in question is “dubious” without some further basis to support Commerce’s reliance. See Home Meridian, 865 F.Supp.2d at 1326-27. Merely stating that the rest of the financial statement facially complies with the applicable accounting standards and that the statement is “unqualified” is insufficient to provide substantial evidence to negate the obvious lack of a required tax line. For the same reason, Commerce’s assumption that Insular Rattan would have disclosed any subsidies under applicable accounting principles, whether they existed at the time or not, is unreasonable in light of the tax line omission. Additionally, although Defendant argues that the tax line is not a critical component of the financial statement because it is not directly used in surrogate financial ratio calculations, it is directly related to another important tax line, profit, which is used in calculating financial ratios. See Dongguan, 904 F.Supp.2d at 1366. Commerce acknowledges that without proper disclosures, it is unable to reconstruct financial statements to conform the statement to proper accounting standards, eliminating any possible remedy for this defect. See Remand Results at 18 (“The Department does not have the ability to look behind financial statements used as surrogates by questioning the surrogate company.”). The lack of a tax line also limits Commerce apd the parties’ ability to determine whether an undisclosed tax subsidy was received. 24 Finally, Commerce’s attempt to distinguish its decision to include Insular Rattan’s statement in this review, despite refusing to do so in the third administrative review, is unreasonable. Although an agency is permitted to depart from its prior practice, it must provide a reasonable explanation for doing so. See NMB Singapore Ltd. v. United States, 557 F.3d 1316, 1328 (Fed.Cir.2009). Commerce’s only explanation for its shift in practice is that the financial statement in the third review lacked auditor notes while the statement here lacked a tax line. Remand Results at 17-19. Commerce does not assert necessity, as other statements are available for *1382 use in financial ratio calculations. Although notes are certainly different from a tax line, Commerce has not provided a reasoned basis for concluding that the omission of notes renders the statement incomplete and unreliable while the omission of a -tax line does not, especially in light of the requirement under applicable accounting principles that both be disclosed. Accordingly, Commerce’s finding that the statement was reliable was not based on substantial evidence but rather speculation as to why the apparent deficiency is neither real nor important, and its selection was contrary to Commerce’s practice. The court remands this issue so that Commerce may remove Insular Rattan’s facially defective financial statement from the pool for calculation of financial ratios. CONCLUSION For the foregoing reasons, Commerce has until July 15, 2013, to seek to reopen the record. Otherwise, the court remands the matter to Commerce to: 1) use Huafeng’s actual ME wood input purchases to calculate normal value, adjusted as needed, and 2) omit Insular Rattan’s financial statement in its financial ratio calculations. Commerce shall file its remand determination by August 26, 2013. The parties shall have until September 23, 2013, to file objections, and Defendant shall have until October 8, 2013, to file its response. Footnotes [1] No other consolidated plaintiff filed comments on the Remand Results. Additionally, no party has challenged Commerce’s determinations regarding poly foam, labor wage rates, or zeroing. Accordingly, those determinations by Commerce are sustained. The parties also voluntarily dismissed the claim related to combination rates. See Stipulation of Dismissal of Count Five (Combination Duty Rates) of Compl., Dkt. No. 100. [2] Commerce considers the PRC to be an NME within the definition of the statute, and no party has challenged that designation in this case. See 19 U.S.C. § 1677(18); Remand Results at 5. [3] Dumping is defined as the sale of goods at less than fair value, calculated by a fair comparison between the export price or constructed export price and normal value. See 19 U.S.C. §§.'1677(34), 1677b(a). [4] In market economy cases, Commerce typically calculates normal value based on the price of the goods iii the domestic market of the investigated entity. See 19 U.S.C. § 1677b(a)(l). Where this is not possible, the - prices of the goods in a surrogate third-country market are used, or alternatively, Commerce constructs a normal value by calculating the total costs of production; including an allowance for general expenses and profits. See id. These two methodologies ultimately target two types of pricing behavior, interna *1371 tional price discrimination (using the home market or third-country price methodology) and below-cost sales (using the constructed value methodology). See Ad Hoc Shrimp Trade Action Comm. v. United States, 596 F.3d 1365, 1371 (Fed.Cir.2010) (recognizing that the NME methodology does not relate to price discrimination but rather costs); see also John H. Jackson et ah, Legal Problems of International Economic Relations 756-58 (5th ed. 2008). [5] As explained below, Commerce's current methodology is a blend of the use of surrogate values, authorized in subsection (c) of the statute, and the use of actual costs for the producer, similar to subsections (e) and (f). In the parallel context of constructed value in ME cases, the statute requires that "[c]osts shall normally be calculated based on the records of the exporter or producer of the merchandise, if such records are kept in accordance with the generally accepted accounting principles of the exporting country ... and reasonably reflect the costs associated with the production and sale of the merchandise.” 19 U.S.C. § 1677b(f)(l)(A). This provision applies to subsections (b) and (e) of the section, covering cost of production and constructed value. Id. § 1677b(f). The NME AD methodology is analogous to the constructed value methodology, and the subsection addressing the NME methodology previously included a specific reference to subsection (e), 19 U.S.C. § 1677b(c)(l) (1988), which was deleted by the Uruguay Rounds Agreement Act in 1994. See Uruguay Round Agreements Act, 103 Pub.L. 465, 108 Stat. 4809, 4882-83 (1994). The legislative history indicates that the modification to this subsection was not intended to be substantive. S.Rep. No. 1 OS-412, at 73 (1994). [6] The AD Methodologies policy statement is not directly applicable here because it dealt with establishing the percentage of ME input purchases during the POR needed to allow Commerce to use those prices to value the' NME-sourced inputs. See 71 Fed.Reg. at 61,-717. The primary issue addressed in that statement was whether a small percentage of ME-supplied inputs was reflective of the price of the total inputs purchased because it was not possible for the entire amount of inputs to be sourced at the small-order ME price. Here, Huafeng purchased 100 percent of its actual lumber and veneer inputs from its ME sources. During the POR, however, it purchased no lumber and veneer inputs whatsoever, rendering the application of this practice meaningless as no percentage can be calculated with a" denominator of zero (the total input purchases during the POR). [7] Under the market economy methodology, the statute requires that the "normal value of the subject merchandise shall be the price described in subparagraph (B), at a time reasonably corresponding to the time of the sale used to determine the export price or constructed export price.” 19 U.S.C. § 1677b(a)(l)(A). Although this clause is not applicable in NME cases, which are analyzed *1374 under subsection (c), Commerce of course must make some reasonable connection between the normal value and the export price. It typically does so by using values for both that are contemporaneous with the POR. See Remand Results at 6. Commerce departs from the strict application of its preference for POR prices in other circumstances, such as through application of the 90/60 day contemporaneity rule whereby in ME cases Commerce may consider home market sales from 90 days prior to and 60 days after the POR. See 19 C.F.R. § 351.414(f)(2)-(3). [8] The court notes some tension between Commerce's asserted blanket methodologies. - Commerce focuses on contemporaneous purchases based on a presumption that those inputs will be used in production and excludes inputs not actually used in production. Despite this policy’s focus on production, Commerce purports to employ a methodology in this case that disregards production concerns, focusing instead exclusively on the time of purchase. Commerce has not based this shift in focus on evidence of record here. [9] Commerce declined to conduct further verification of the ME purchases because it decided to apply a blanket policy of excluding prePOR ME purchase data, even at the verification stage before surrogate values were calculated. [10] Where there is a gap in the record, Commerce may employ facts otherwise available and draw an adverse inference against a non-cooperating respondent under some circumstances. See 19 U.S.C. § 1677e. Commerce did not state that it was using such a tool here, and its application of such an inference likely would not be permissible. See Remand Results at 42 (recognizing that few, if any, companies keep the type of direct evidence •Commerce appears to be demanding). [11] Of course, under Commerce's own practice, a small percentage of ME inputs is normally sufficient to value the entire set of inputs. If Commerce has good reason to reopen the record at this late stage to address the factual issue of Huafeng's use of ME wood inputs, it may move for permission to do so. Otherwise, the conclusion that Huafeng used 100 percent ME wood inputs for its POR production will control the proceedings. [12] In its Remand Results, Commerce also compared actual ME price data to surrogate value data under the criteria for deciding between competing surrogate values, although Commerce recognized that actual price data is not a surrogate value. Remand Results at 5-6. It does not appear that Commerce relied on this comparison in arriving at its decision, and at any rate, it seems to make little analytical sense. It is unsurprising that actual ME purchases by a producer may be from an economically different country, fail to represent broad market averages, and may be inclusive of taxes. See id. This does not make such purchases somehow less accurate as a data source for valuing the inputs, and in fact, the court and Commerce have noted repeatedly that actual purchase data are generally preferable to surrogate estimates, all other considerations equal. See Lasko I, 810 F.Supp. at 317-18; Oscillating Fans, 56 Fed. Reg. at 55,275 ("[UJsing surrogate values when market-based values are available would, in fact, be contrary to the intent of the law.”). [13] [[Confidential Data Deleted ]] of Huafeng's ME purchases occurred during the [[Confidential Data Deleted]] of 2008, just prior to the POR. For example, [[Confidential Data Deleted ]] of its pine lumber purchases, [[Confidential Data Deleted ]] of its poplar lumber purchases, and [[Confidential Data Deleted ]] of its oak lumber purchases were made between [[Confidential Data Deleted ]]. Market Economy Inputs During 2008 or Before (Oct. 12, 2010) HMI App., Tab 15, Ex. S-39, at 3. [14] Philippine harmonized tariff schedule subheading 4407.10 covers “wood sawn or chipped lengthwise, sliced or peeled, whether or not planed, sanded or end-jointed, of a thickness exceeding 6mm, coniferous;” 4407.99 covers "wood sawn or chipped lengthwise, sliced or peeled, whether or not *1378 planed, sanded or end-jointed, of a thickness exceeding 6mm, other;” and 4408.90 covers “sheets for veneering (including those obtained by slicing laminated wood), for plywood or for similar laminated wood and other wood, sawn lengthwise, sliced or peeled, whether or not planed, sanded, spliced or end-jointed, of a thickness not exceeding 6 mm, other.” Letter from King & Spalding to Commerce re: Submission of Publicly Available Information to Value Factors of Production (Nov. 15, 2010), P.R. 435, Ex. 3, at 234-37. [15] Commerce contends that inflating ME prices is not appropriate because they are not based on broad market averages and inflation does not account for product-specific changes in price over time. See Remand Results at 7. Commerce, however, consistently inflates prePOR surrogate and actual values when needed, and it does not seem that an adjustment to non-contemporaneous surrogate values via an inflator accounts for price fluctuations any better than one applied to actual prices. [16] It remains particularly alarming that most of the surrogate values are [[Confidential Data Deleted ]] Huafeng's ME purchase prices, even if a single producer paid prices to [[Confidential Data Deleted ]] that were somewhat similar to those of the non-specific wood imports into the Philippines utilized by Commerce. See HMI Cmts. 27-29. [17] Commerce notes that no party challenged the particular tariff classifications used to calculate surrogate values, including this basket subheading. Remand Results at 9. The lack of objection, however, merely indicates that neither party argued for a more specific tariff subheading, possibly because one might not exist. It does not mean that the parties acknowledged that a general basket subheading is more specific than the actual ME purchase data submitted by Huafeng. [18] The court has recently held that in applying the 33 percent rule, Commerce may not blindly reject the use of actual ME purchase data where it amounts to 32.9 percent of total input purchases. See Gold. East Paper (Jiangsu) Co., Ltd. v. United States, 918 F.Supp.2d 1317, 1320-24, Ct. No. 10-00371, at 6 (CIT June 17, 2013) (finding application of this rigid policy to be "unreasonable, arbitrary and capricious under the circumstances” where Commerce failed to explain why the actual purchase data were not the best available information, notwithstanding Commerce’s preferred input threshold). [19] Commerce also explains that the use of pre-POR ME prices "would suggest that prePOR ME prices are preferable to POR surrogate values because they reflect the respondent’s POR purchasing' experience — which is not necessarily the case.” Remand Results at 9. It is unclear in what circumstances actual purchases, under market conditions, would not accurately reflect a producer's purchasing experience. These prices reflect the reality of what the producer actually paid as compared to the fiction Commerce proposes. In any case, such unusual circumstances do not exist here. [20] NME dumping methodology is essentially an adjusted below-cost methodology. Certainly if this were an analogous ME AD case based on the cost of production methodology, Commerce would be hard pressed to argue that a producer making well-timed purchases of inputs has sold its goods below cost simply because input prices rose between the time of purchase and the time of sale. [21] Because the court remands to Commerce to revalue Huafeng's wood inputs, it need not address HMI's ministerial error claim. See HMI Cmts. 30. [22] Defendant also contends that arguments involving the effective date of the relevant accounting principle, IAS 20, were not made before the agency and therefore have not been exhausted by HMI. Def.'s Resp. 15-16. Because the court remands Commerce's determination on other grounds, it need not consider the exhaustion question. The court notes, however, that HMI generally challenged Commerce’s reliance on the non-disclosure of subsidies to support its finding, although HMI did not directly refer to IAS 20. See Petitioners’ Cmts. Concerning Draft Results of Redetermination (Jan. 10, 2013) Dkt. No. 113-2, Attach. 2, at 8. [23] Commerce previously rejected Insular Rattan’s financial statement due to different defects than the ones at issue here. See Issues and Decision Memorandum for the Final Results of the 2007 Antidumping Duty Administrative and New Shipper Reviews, A-5 70-890, POR: 1/1/07-12/31/07, at 35 (Aug. 10, 2009), available at http://ia.ita.doc.gov/frn/summary/ PRC/E9-19666-1 .pdf (last visited June 25, 2013), aff'd Lifestyle Enter., Inc. v. United States, 768 F.Supp.2d 1286, 1310-11 (CIT 2011). [24] The court notes that Insular Rattan’s overhead expenses; selling, general and administrative ("SG & A”) expenses; and profits were significantly lower than those of the other companies used to calculate financial ratios, possibly indicating other issues with Insular Rattan’s financial statement. See Final Results Surrogate Value Memorandum (Aug. 5, 2011) App. to AFMC's Resp. Br. in Opp’n to Pis.’ Mots, for J. on the Agency Record, Tab 15, at 3.