Full opinion text
OPINION RIDGWAY, Judge. In this action, Plaintiffs Zhengzhou Harmoni Spice Co., Ltd., Jinan Yipin Corporation, Ltd., Jining Trans-High Trading Co., Ltd., Jinxiang Shanyang Freezing Storage Co., Ltd., Linshu Dading Private Agricultural Products Co., Ltd., Shanghai LJ International Trading Co., Ltd., and Sunny Import & Export Co., Ltd. — Chinese producers and exporters of fresh garlic — contest the final results of the U.S. Department of Commerce’s tenth administrative review of the antidumping duty order covering fresh garlic from the People’s Republic of China (“PRC”). See Fresh Garlic from the People’s Republic of China: Final Results and Partial Rescission of Anti-dumping Duty Administrative Review and Final Results of New Shipper Review, 71 Fed.Reg. 26,329 (May 4, 2006) (“Final Results”); Issues and Decision Memorandum for the Administrative Review and New Shipper Reviews of the Antidumping Duty Order on Fresh Garlic from the People’s Republic of China (April 26, 2006) (Pub. Doc. No. 462) (“Decision Memorandum”). Pending before the Court is the Motion for Judgment on the Agency Record filed on behalf of four of the plaintiffs in this matter — Zhengzhou Harmoni Spice Co., Ltd., Jinan Yipin Corporation, Ltd., Linshu Dading Private Agricultural Products Co., Ltd., and Sunny Import & Export Co., Ltd. (collectively “the Chinese Producers”). In their motion, the Chinese Producers challenge the methodology used in calculating “normal value,” as well as various other aspects of Commerce’s anti-dumping determination, and request that this matter be remanded to the agency for reconsideration. See generally Brief in Support of Plaintiffs’ Rule 56.2 Motion for Judgment on the Agency Record (“Pis.’ Brief’); Reply Brief in Support of Plaintiffs’ Rule 56.2 Motion for Judgment on the Agency Record (“Pis.’ Reply Brief’). The Government opposes the Chinese Producers’ motion. The Government maintains that Commerce’s determination is supported by substantial evidence and is otherwise in accordance with law, and that it should be sustained in all respects. See generally Defendant’s Memorandum in Opposition to Plaintiffs’ Rule 56.2 Motion for Judgment Upon the Agency Record (“Def.’s Brief’). The Defendanb-Intervenors, represenb ing the interests of domestic producers of fresh garlic, oppose the Chinese Producers’ motion as to two of the seven issues raised — ie., Commerce’s use of the agency’s intermediate input methodology and the valuation of garlic bulb — and, like the Government, similarly urge that Commerce’s determination should be sustained. See generally Defendant-Intervenors’ Brief in Response to Plaintiffs’ Motion for Judgment on the Administrative Record (“Def.-Ints.’ Brief’). Jurisdiction lies under 28 U.S.C. § 1581(c) (2000). For the reasons set forth below, the Chinese Producers’ Motion for Judgment on the Agency Record is granted in part. I. Background The underlying antidumping order here at issue, covering imports of fresh garlic from the PRC, dates back to 1994. See Antidumping Duty Order: Fresh Garlic from the People’s Republic of China, 59 Fed.Reg. 59,209 (Nov. 16, 1994) (“Anti-dumping Order”). In December 2004, Commerce initiated its tenth administrafive review of producers and exporters of fresh garlic from the PRC, including the Chinese Producers who are the plaintiffs in this action. See Initiation of Antidumping and Countervailing Duty Administrative Reviews and Request for Revocation in Part, 69 Fed.Reg. 77,181 (Dec. 27, 2004); see also Fresh Garlic from the People’s Republic of China: Preliminary Results and Partial Rescission of Antidumping Duty Administrative Review and Preliminary Results of New Shipper Reviews, 70 Fed.Reg. 69,942, 69,942-43 (Nov. 18, 2005) (“Preliminary Results”). In the course of conducting the administrative review, Commerce issued multiple questionnaires to the respondents (ie., various Chinese garlic producers, including Plaintiffs), requesting information concerning their organization, sales, and production costs, in order to determine the normal value of the subject merchandise. See Def.’s Brief at 4. In addition, Commerce issued supplemental questionnaires to address certain questions that had been raised in previous administrative reviews concerning the respondents’ reported growing and harvesting-related “factors of production.” See Decision Memorandum at 2-3; see also Issues and Decision Memorandum for the Administrative Review of the Antidumping Duty Order on Fresh Garlic from the Peoples’ Republic of China (Ninth Administrative Review), 2005 WL 2290660 (June 13, 2005) (“Ninth Garlic Review Memorandum”), at comment 1 (considering, and ultimately declining, use of intermediate input methodology). In their responses to Commerce’s questionnaires, the respondents provided the agency with suggested values for their factors of production. See Respondents’ Second Surrogate Value Submission (Pub.Doc. No. 418). The Domestic Producers supplied surrogate value information for the respondents’ factors of production, and requested that Commerce use the agency’s intermediate input valuation methodology — as the Domestic Producers had urged in prior reviews — due to asserted “anomalies and inconsistencies in the ... data submitted by all of the respondents.” Def. Ints.’ Brief at 3; see also Domestic Producers’ Surrogate Value Submission (Pub. Doc. No. 82); Domestic Producers’ Second Surrogate Value Submission (Pub.Doc. No. 143) (submitting information from previous reviews comparing respondents’ ranged factors of production data); Ninth Garlic Review Memorandum, 2005 WL 2290660, at comment l. Given the concerns expressed in prior reviews as to the reliability of respondents’ records, Commerce conducted onsite “harvest verifications” of six respondents in May and June 2005, to assist the agency in determining whether to value intermediate inputs rather than factors of production. See Preliminary Results, 70 Fed.Reg. at 69,943; see also Harvest Verification Reports (Pub.Doc.Nos.386, 392, 393). Unlike typical verifications, where Commerce focuses on respondents’ books and records (■ie., general ledgers, subledgers, etc.), these harvest verifications involved onsite visits to allow agency personnel to directly observe respondents’ actual cultivation and harvesting procedures. See Decision Memorandum at 2-3; see also Def.’s Brief at 5. In its Preliminary Results, Commerce concluded that “the books and records maintained by the [Chinese garlic producers] do not report or account for all of the relevant information and do not allow the respondents to identify all of the factors of production necessary to grow and harvest garlic.” See Preliminary Results, 70 Fed. Reg. at 69,949. Commerce therefore used its intermediate input method of valuation for the respondents’ growing and harvesting factors of production, and valued the intermediate input, raw garlic bulb (in lieu of the upstream factors of production used to produce that input), in calculating respondents’ dumping margins. Commerce’s Final Results similarly reflected the agency’s use of its intermediate input methodology to value raw garlic bulb, after again finding respondents’ data inadequate. See generally Decision Memorandum at 11-22. While Commerce found the harvesting factors of product data insufficient, Commerce found respondents’ reported data on their post-harvesting factors of production (ie., processing, packaging, shipping) to be reliable, both in the Preliminary Results and in the Final Results. The agency therefore added the surrogate values for those factors of production to the surrogate value of the raw garlic bulb inputs. See Decision Memorandum at 14; see also generally Final Results, 71 Fed.Reg. 26,-329. II. Standard of Review In reviewing Commerce’s final determination in an antidumping case, the agency’s determination must be upheld, except to the extent that it is found to be “unsupported by substantial evidence on the record, or otherwise not in accordance with law.” 19 U.S.C. § 1516a(b)(1)(B)(i); see also NMB Singapore Ltd. v. United States, 557 F.3d 1316, 1319 (Fed.Cir.2009). Substantial evidence is “more than a mere scintilla”; rather, it is “such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.” Universal Camera Corp. v. Nat’l Labor Relations Bd., 340 U.S. 474, 477, 71 S.Ct. 456, 95 L.Ed. 456 (1951) (quoting Consol. Edison Co. v. Nat’l Labor Relations Bd., 305 U.S. 197, 229, 59 S.Ct. 206, 83 L.Ed. 126 (1938)); see also Mittal Steel Point Lisas Ltd. v. United States, 548 F.3d 1375, 1380 (Fed.Cir.2008) (same). Moreover, any evaluation of the substantiality of evidence “must take into account whatever in the record fairly detracts from its weight,” including “contradictory evidence or evidence from which conflicting inferences could be drawn.” Suramerica de Aleaciones Laminadas, C.A. v. United States, 44 F.3d 978, 985 (Fed.Cir.1994) (quoting Universal Camera, 340 U.S. at 487-88, 71 S.Ct. 456); see also Mittal Steel, 548 F.3d at 1380-81 (same). That said, the mere fact that it may be possible to draw two inconsistent conclusions from the record does not prevent Commerce’s determination from being supported by substantial evidence. Am. Silicon Techs, v. United, States, 261 F.3d 1371, 1376 (Fed.Cir.2001); see also Consolo v. Federal Maritime Commission, 383 U.S. 607, 620, 86 S.Ct. 1018, 16 L.Ed.2d 131 (1966). Finally, while Commerce must explain the bases for its decisions, “its explanations do not have to be perfect.” NMB Singapore, 557 F.3d at 1319. However, “the path of Commerce’s decision must be reasonably discernable,” to support judicial review. Id. III. Analysis The Chinese Producers challenge multiple aspects of Commerce’s Final Results in the tenth administrative review of fresh garlic from the PRC. Specifically, the Chinese Producers contest: (1) Commerce’s decision to utilize an intermediate input methodology in valuing fresh garlic bulb; (1) Commerce’s valuation of fresh garlic bulb using a specific subset of domestic Indian pricing data; (3) Commerce’s wage rate calculation; (4) Commerce’s valuation of ocean freight; (5) Commerce’s valuation of packing cartons; (6) Commerce’s valuation of plastic jars; and (7) Commerce’s inclusion of certain labor expenses as part of manufacturing overhead. Pis.’ Brief at 2-4. As discussed in greater detail below, there is no merit to the Chinese Producers’ challenge to Commerce’s use of the intermediate input methodology, or their challenge to Commerce’s inclusion of certain labor expenses as part of factory overhead. See sections III.A & III.G, infra. In contrast, Commerce’s determinations concerning the valuation of garlic bulb, ocean freight, packing cartons, and plastic jars, and its calculation of the applicable wage rate, cannot be sustained on the existing record. See sections III.B-III.F, infra. The Chinese Producers’ Motion for Judgment on the Agency Record therefore must be granted in part, and this matter remanded to Commerce for further action. A. Intermediate Input Methodology The Chinese Producers first challenge the use of Commerce’s intermediate input methodology to determine the normal value of raw garlic bulb, asserting that the methodology is contrary to the plain language of the statute and not in accordance with past practice, and that its use did not produce the most accurate normal value. See generally Pis.’ Brief at 8-14. The Chinese Producers contend that, when Commerce finds respondents’ factors of production information inadequate, Commerce is statutorily required “to base normal value in its entirety on the price at which comparable merchandise produced in a comparable market economy is sold in other countries.” Pis.’ Brief at 8 (citing 19 U.S.C. § 1677b(c)(2)). The Chinese Producers further maintain that, even if application of the intermediate input methodology were appropriate in this case, the intermediate input valued must be a “component of the subject merchandise,” not the subject merchandise itself. See id. at 9. The Chinese Producers’ arguments, however, are without merit. Generally, Commerce calculates normal value in antidumping matters pursuant to 19 U.S.C. § 1677b(a). Where (as here) the subject merchandise was exported from a nonmarket economy (“NME”) country, and if Commerce finds that — in light of concerns about the sufficiency or reliability of the data — the available information “does not permit the normal value of the subject merchandise to be determined under [19 U.S.C. § 1677b(a) ],” the statute instructs Commerce to use surrogate values “based on the best available information regarding the values of such factors [of production] in a market economy country or countries considered to be appropriate” by the agency. See 19 U.S.C. § 1677b(c)(l). However, if Commerce finds the available information inadequate for purposes of determining the normal value of the subject merchandise pursuant to the factors of production method (described above), the statute provides that Commerce shall base normal value on the export price of comparable merchandise produced in one or more market economy countries that are economically comparable to the nonmarket economy country at issue. 19 U.S.C. § 1677b(c)(2). Within this general framework, the statute “accords Commerce wide discretion in the valuation of factors of production in the application of those guidelines”; indeed, the Court of Appeals has “specifically held that Commerce may depart from surrogate values when there are other methods of determining the ‘best available information’ regarding the values of the factors of production.” Shakeproof Assembly Components v. United States, 268 F.3d 1376, 1381 (Fed.Cir.2001) (citing Lasko Metal Prods., Inc. v. United States, 43 F.3d 1442, 1446 (Fed.Cir.1994)). In short, when “determining the valuation of the factors of production, the critical question is whether the methodology used by Commerce is based on the best available information and establishes antidumping margins as accurately as possible.” Shakeproof, 268 F.3d at 1382. In prior administrative reviews of the Antidumping Order at issue, Commerce used the factors of production data reported by the respondents (including both harvesting and post-harvesting factors of production), and based its calculation of the normal value of fresh garlic upon those data. See, e.g., Ninth Garlic Review Memorandum, 2005 WL 2290660, at comment 1. In the immediately preceding review, however, Commerce acknowledged various flaws in the respondents’ factors of production data, and considered whether use of the agency’s intermediate input methodology would be appropriate. See id. Although the agency ultimately declined to use its intermediate input methodology in the ninth review due to lack of sufficient information, Commerce there indicated that it planned to “fully examine all of these issues in the next administrative review.” Id. In the tenth administrative review at issue here, Commerce made good on its promise, and- — -indeed—decided to use its intermediate input methodology when determining the value of raw garlic bulb. See Decision Memorandum at 11-15; Preliminary Results, 70 Fed.Reg. at 69,947-50. As set forth above, in reaching its decision, Commerce sent multiple supplemental questionnaires to the respondent producers of Chinese garlic, and conducted onsite harvest verifications of six of those producers, all in an attempt to verify their harvesting factors of production. See Preliminary Results, 70 Fed.Reg. at 69,943; Harvest Verification Reports. Despite the agency’s efforts, Commerce found that “respondents are unable to accurately report and substantiate the complete costs of growing garlic.” See Intermediate Input Methodology Memorandum (Pub.Doc. No. 388), at 3. Specifically, Commerce stated: Evidence on the record of these reviews regarding the recording and accounting standards of the garlic industry in the PRC supports a finding that we cannot accurately quantify the consumption rates of all relevant [factors of production] used to grow, harvest and process the subject merchandise. Further, the respondents’ ability to measure and report accurate [factors of production] to the Department is greatly diminished by the fact that they lease the land on which the garlic is grown, and therefore cannot obtain information on other crops that are grown on this land in the off-season, which ultimately affect the [factors of production] for raw garlic. Finally, respondents do not keep the types of books and records that would allow the Department to establish the appropriateness or accuracy of the reported [factors of production]. Id. at 10-11. After determining the surrogate value for the intermediate input at issue- — -raw garlic bulb — Commerce added the post-harvesting factors of production, which it had deemed reliable, to the intermediate value and ultimately arrived at the overall normal value of the subject merchandise. See Decision Memorandum at 14. Commerce here used its intermediate input methodology within the statute’s traditional be st-available-factor-of-production-information valuation methodology, and not pursuant to the alternative methodology exception established by statute. See Decision Memorandum at 14. As Commerce explained in its Intermediate Input Methodology Memorandum, Commerce merely “modified]” its standard upstream factor of production methodology in applying its intermediate input methodology. See Intermediate Input Methodology Memorandum at 1. The only difference is that, in cases where the intermediate input methodology is employed, Commerce believes that, due to inaccurate or flawed data, or some other anomaly, the best way to value the factors of production used to produce an intermediate product (here, raw garlic bulb) is through the direct valuation of that intermediate input. See Decision Memorandum at 14. Commerce explained that: We [Commerce] also disagree with the contentions raised by ... respondents that we relied on [§ 1677b(c)(2)’s alternative methodology exception] to calculate normal value. Rather, using the intermediate input methodology ... is consistent with [§ 1677b(c)(1)(B) ] because we valued the respondents’ reported [factors of production]. The intermediate input methodology merely allows [Commerce] to value the intermediate product (in this case the raw garlic bulb) in lieu of valuing the upstream inputs used to produce that intermediate product. Valuing the intermediate input in this way constitutes the “best available information,” in accordance with [§ 1677b(c)(1)(B) ]. Decision Memorandum at 14 (discussing application of 19 U.S.C. § 1677b(c)). None of the Chinese Producers’ arguments undermine the soundness of Commerce’s reasoning to any significant degree. The Chinese Producers contend that the statute requires that, if Commerce finds the available factors of production data inadequate, the agency must abandon application of the traditional factor of production method, and instead apply an alternative valuation methodology, basing “normal value in its entirety on the price at which comparable merchandise produced in a comparable market economy is sold in other countries.” Pis.’ Brief at 8 (citing 19 U.S.C. § 1677b(c)(2)). But the Chinese Producers are mistaken. Because the statutory exception that the Chinese Producers invoke “is unclear as to the circumstances in which ‘the available information is adequate for purposes of determining the normal value,’ the question is whether Commerce’s interpretation is based on a permissible construction of the statute.” Def.’s Brief at 22 (citing Chevron U.S.A., Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 843, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984)). Here, Commerce properly read the statute to require use of an “alternative valuation method” only when “all respondents’ information is ‘inadequate’ or unuseable to determine normal value under [§ ] 1677b(c)(l).” Id. at 21. As the Domestic Producers explain, the statute “only directs Commerce to determine the ‘quantities of raw materials employed;’ [sic] it contains no detailed prescription for how physical [factors of production] are to be identified or a requirement that material [factors of production] be identified with the same level of detail in every case.” Def.-Ints.’ Brief at 9 (quoting 19 U.S.C. § 1677b(c)(3)(B)). Congress accorded Commerce broad discretion to develop one or more methodologies for valuing factors of production in NME cases, subject only to the fundamental requirement that the agency’s determinations be “based on the best available information” in every case. See 19 U.S.C. § 1677b(c)(1); see also Def.-Ints.’ Brief at 9-10. Commerce’s intermediate input valuation methodology was developed pursuant to the agency’s broad statutory authority, and in accordance with its statutory mandate to base its determinations “on the best available information” concerning the values of the factors of production at issue. See 19 U.S.C. § 1677b(c)(1); see also Def.Ints.’ Brief at 9-11. The Chinese Producers further claim that “the statute leaves no discretion for the agency to disregard [certain factors of production] and still attempt to calculate normal value pursuant to 19 U.S.C. § 1677b(e)(1).” See Pis.’ Reply Brief at 3. But Commerce is not forced to abandon § 1677b(c)(1) factors of production valuation merely because a portion of respondents’ factors of production data is inadequate. On its face, the § 1677b(c)(2) exception limits Commerce’s ability to determine normal value using that alternative valuation method — i.e., to determine normal value based on the price at which comparable merchandise is “produced in one or more market economy countries that are at a level of economic development comparable to that of the nonmarket economy country,” and “sold in other countries, including the United States” — to situations where the agency “finds that the available information is inadequate for purposes of determining the normal value of subject merchandise under [§ 1677b(c)(1) ].” 19 U.S.C. § 1677b(c)(2). Using a short-cut to determine factors of production value in certain, limited situations where the factors of production data submitted are inadequate certainly falls within the broad discretion accorded Commerce under the statute. Commerce here did not conclude that all of respondents’ factors of production data was inadequate for determining the normal value of subject merchandise; only a portion was found to be inadequate. As a result, in the present review, there was no need for the agency to resort to the alternative methodology exception in § 1677b(c)(2). See Def.’s Brief at 21-23; Def.-Ints.’ Brief at 10. Contrary to the Chinese Producers’ “all-or-nothing” approach, because the only inaccurate factors of production data concerned the factors of production used to produce raw garlic bulb, Commerce reasonably decided to value that intermediate input (raw garlic bulb) pursuant to the intermediate input methodology (and within the traditional valuation methodology prescribed by § 1677b(c)(l)), while also using the post-harvesting factors of production data that the agency deemed accurate. See Def.’s Brief at 22-23. Thus, as the Government and the Domestic Producers maintain, both Commerce’s interpretation of the statute and the agency’s decision to use its intermediate input methodology in this case are reasonable. See id. at 14; Def.Ints.’ Brief at 10. Turning to Commerce’s application of its intermediate input methodology, the Chinese Producers correctly note that, in past cases, Commerce has used that methodology to value a component of the subject merchandise. Here, according to the Chinese Producers, raw garlic bulb — the intermediate input at issue — is “the actual product subject to the dumping order.” Pis.’ Brief at 9. The Chinese Producers’ argument is unavailing. The Government explains that, “because the raw garlic bulb is harvested from the ground, and is ‘not immediately shipped to the United States,’ [and] requires ‘at least a minimum amount of processing and packing,’ the raw garlic bulb cannot be the subject merchandise.” See Def.’s Brief at 19 (quoting Decision Memorandum at 13). The Government notes that “Commerce’s experience in several administrative and new shipper reviews of fresh garlic confirmed this conclusion, showing ... ‘that the garlic harvested from the ground is, at minimum, cleaned to remove the outer skins in order to give the garlic bulb its characteristic white, fresh appearance,’ ” and “ ‘the whole bulb garlic is then typically packed in mesh bags and cartons for shipment.’ ” See id. at 19 (quoting Decision Memorandum at 13). Contrary to the Chinese Producers’ assertions, the fact that raw garlic bulb is an advanced intermediate input does not set this case apart from prior cases where Commerce has used its intermediate input methodology. The Domestic Producers’ analysis differs slightly from that of the Government, but reaches the same end. The Domestic Producers concede that raw garlic bulb may be covered by the Antidumping Order. See Def.-Ints.’ Brief at 12. However, the Domestic Producers go on to note that “the propriety of applying the intermediate input methodology has nothing to do with whether the intermediate inputs in question are within the scope of the anti-dumping order or not; it depends on whether this methodology achieves a more accurate and legally-defensible result in the particular case.” Id. at 12. In any event, the Chinese Producers’ emphasis on the scope of the Antidumping Order is misplaced, because (as both the Government and the Domestic Producers note) the subject merchandise in the present case is not the raw bulb, but — rather— fresh garlic, which, as described above, obviously requires at least some degree of processing in order to be ready for shipment. See Decision Memorandum at 13. Finally, the Chinese Producers assert that Commerce failed to adequately explain the need to depart from the agency’s practice in prior reviews, and its basis for concluding that use of the agency’s intermediate input methodology would result in a more accurate margin. See Pis.’ Brief at 12. As the Chinese Producers acknowledge, “Commerce is permitted to deviate from ... past practice, at least where it explains the reason for its departure.” Allegheny Ludlum Corp. v. United States, 346 F.3d 1368, 1373 (Fed.Cir.2003) (citing Atchison, Topeka & Santa Fe Ry. Co. v. Wichita Bd. of Trade, 412 U.S. 800, 808, 93 S.Ct. 2367, 37 L.Ed.2d 350 (1973)). The Chinese Producers seek to make much of the fact that “respondents to this review maintained their books and records in accordance with the methodology employed in previous reviews of this order.” Pis.’ Brief at 12. But that is of little moment here. As the Government and the Domestic Producers underscore, in the administrative review immediately preceding the review at issue here, Commerce unequivocally expressed its concerns about the accuracy and reliability of the factors of production data submitted by the respondents and stated the agency’s intentions to examine and address those concerns in future reviews. See Ninth Garlic Review Memorandum, 2005 WL 2290660, at comment 1; see also Def.’s Brief at 17-18; Def.-Ints.’ Brief at 14-15. The Domestic Producers explain: Although Commerce was not persuaded to depart from use of the traditional upstream [factors of production] valuation method in the eighth and ninth reviews themselves, it put all parties on notice in the final results of the ninth review that it intended to scrutinize the respondents’ reported upstream [factors of production] very carefully in this tenth review with a view toward moving to use of the intermediate input valuation method if its review of the upstream data warranted it. Def.-Ints.’ Brief at 14. Commerce adequately explained its intermediate input methodology and the agency’s reasons for using it in this case (i.e., to obtain the most accurate results) in the Preliminary Results, in the Intermediate Input Methodology Memorandum, and in the Decision Memorandum. See generally Preliminary Results, 70 Fed.Reg. at 69,949-50; Intermediate Input Methodology Memorandum; Decision Memorandum at 11-15. Moreover, the agency’s reasoning in this case is reinforced by the agency’s handling of similar issues in other cases in the past. The law does not require more. See generally NMB Singapore, 557 F.3d at 1319 (although agency must explain the bases for its decisions, “its explanations do not have to be perfect,” so long as “the path of Commerce’s decision ... [is] reasonably discernable”). In sum, Commerce acted within its discretion in deciding to use the agency’s intermediate valuation methodology. Commerce thoroughly explained its reasons for deviating from its practice in prior administrative reviews of the Antidumping Order at issue; and the agency adequately supported its use of the intermediate input methodology within factors of production valuation. The Chinese Producers’ challenges to Commerce’s use of the intermediate input methodology in this instance must therefore be rejected. B. Garlic Bulb Valuation The Chinese Producers fare better on their challenge to Commerce’s surrogate valuation of the raw garlic bulb input in its calculation of normal value for fresh garlic. According to the Chinese Producers, Commerce erred in selecting the Indian Agricultural Marketing Information Network (“Agmarknet”) data for the “China” variety of garlic as the basis for the surrogate value of respondents’ garlic bulb input. The Chinese Producers further assert that Commerce improperly rejected more representative data that the respondents submitted. See generally Pis.’ Brief at 16-17, 20-21. As section III.A above explains, in a nonmarket economy case such as this, Commerce must base its surrogate values on “the best available information” from an appropriate market economy country or countries. See 19 U.S.C. § 1677b(c)(l). Because the statute does not define “best available information,” Commerce has broad discretion to determine the best available information “in a reasonable manner on a case-by-case basis.” See Rhodia, Inc. v. United States, 25 CIT 1278, 1286, 185 F.Supp.2d 1343, 1351 (2001) (quotation marks and citation omitted). That discretion, however, is “curtailed by the purpose of the statute, i.e., to construct the product’s normal value as it would have been if the NME country were a market economy country.” Rhodia, 25 CIT at 1286, 185 F.Supp.2d at 1351 (citing Nation Ford Chem. Co. v. United States, 166 F.3d 1373, 1375 (Fed.Cir.1999)). In the present case, following its decision to apply the intermediate input methodology to determine normal value, Commerce was required to choose a surrogate, value for the intermediate input in question, raw garlic bulb. Initially, Commerce found that Indian import statistics derived from the World Trade Atlas constituted the best publicly available data for this purpose, and dismissed the use of Agmarknet data. See Preliminary Results, 70 Fed.Reg. at 69,949-50. In the Preliminary Results, Commerce explained its dissatisfaction with the Agmarknet data, noting that the agency could not “ascertain the quality or nature of the garlic products (i.e., bulbs, loose cloves, etc.).” Preliminary Results, 70 Fed.Reg. at 69,950. According to Commerce, this shortcoming was important, because respondents’ garlic is a large, high yield, high-quality type of garlic that is distinct from the overwhelming majority of garlic grown in India. See Decision Memorandum at 40-41. Thus, while the respondents submitted data from the broader Agmarknet database during the course of the review, Commerce initially rejected the use of Agmarknet data based on the agency’s belief that the data reflected only small-bulb Indian garlic. In addition, Commerce was concerned that it could not determine whether the Agmarknet figures were tax-exclusive. See id. at 40; see also Def.’s Brief at 25. In the Final Results, Commerce reversed course and concluded that the Agmarknet data reflecting values for the Indian domestic garlic identified as “China” variety was the best overall source of Indian price information. See Decision Memorandum at 40-41. Explaining its decision to rely on the Agmarknet data, Commerce stated: [T]he [Agmarknet] database represents daily garlic bulb prices from wholesale markets in 21 out of 28 Indian states plus the National Capital Territory of Delhi. Therefore, we find that Agmarknet data are broadly representative of garlic bulb prices throughout India as noted by several respondents. Second, the database represents market transactions covering the period November 1, 2003 through October 31, 2004, and therefore, we find the data to be contemporaneous with the review. Id. at 42 (emphasis added). Commerce also noted that the Agmarknet database was publicly available; but the agency had to concede that it still had no firm indication as to whether the Agmarknet figures were inclusive of taxes. See id. Finally, and most importantly, Commerce noted that Agmarknet lists prices for six separate varieties of garlic, one of which is the “China” category. Conspicuously absent from the Agmarknet data, however, is any description of those categories. See id. at 40, 42. Obviously conscious that its task was “to value large-sized garlic bulb because this is the primary characteristic that distinguishes the type of garlic exported by the PRC respondents from the majority of garlic sold in India,” Commerce found the “China” variety of garlic to be most similar to the respondents’ garlic — even though no description of “China” variety garlic was provided, and seemingly based on nothing more than perhaps the name of the variety, and the fact that it had a higher weighted-average price. See Decision Memorandum at 42-44. To be sure, Commerce enjoys broad discretion in determining what constitutes the best information available. Nevertheless, Commerce may not act arbitrarily. “In determining the valuation of the factors of production, the critical question is whether the methodology used by Commerce is based on the best available information and establishes antidumping margins as accurately as possible.” Shakeproof 268 F.3d at 1382. The statutory objective of calculating dumping margins as accurately as possible can be achieved only when Commerce’s choice as to what constitutes the best available information evidences a rational and reasonable relationship to the factor of production that it represents. And, as the Chinese Producers assert, Commerce’s speculation here — that higher-price-equals-bigger-bulb — cannot suffice to establish the requisite rational and reasonable relationship between respondents’ garlic bulb input and the Agmarknet “China” variety of garlic. See Pis.’ Brief at 23. In short, absent evidence on the nature and characteristics of Agmarknet’s “China” variety of garlic bulb, Commerce’s decision to use the “China” variety prices was impermissibly speculative. See id. at 17. Commerce’s reliance on the Domestic Producers’ Market Research Report (submitted in the course of the administrative review) does little, if anything, to buttress the agency’s decision. As the Market Research Report explains it, Chinese garlic exported to the United States has an average bulb diameter of greater than 40 millimeters, while garlic typically grown and sold in the Indian market has an average bulb diameter of only 20 to 40 millimeters. Decision Memorandum at 42-43; see also generally Market Research Report (Pub. Doc. No. 41). The Market Research Report stated that this typical Indian garlic is cultivated in “short-day” zones, rather than “long-day” zones (which have longer periods of sunlight, and which are where larger, high-yield garlic is cultivated). Decision Memorandum at 41 (citing Market Research Report at ll). The Market Research Report specifically identified Agrifound Parvati (a large Indian garlic variety) as one of the three varieties of garlic with characteristics similar to the Chinese garlic grown by respondents. Decision Memorandum at 41; id. n. 110 (citing Ninth Garlic Review Memorandum, 2005 WL 2290660, at comment 2; Market Research Report at 18). According to Commerce, “Agrifound Parvati is a clonal garlic believed to be of Chinese genetic origin that was developed by NHRDF and is the closest variety (in terms of genetic origin, specifications, etc.) to Chinese garlic.” Id. at 41 (citing Market Research Report at 18). Commerce further indicated that “use of such clonal varieties, developed mainly by NHRDF, are increasing in certain areas in India due to the efforts of NHRDF and other institutions, and that this development is in sharp contrast to the rest of the county where local varieties dominate.” Id. (citing Market Research Report at 13-17,18). The Chinese Producers contend that— from Commerce’s recitation of certain information from the Domestic Producers’ Market Research Report (outlined above) — the agency then “made a substantial leap ... and concluded without any evidentiary support that the minute group of sales designated as ‘China’ variety in the Agmarknet data must represent sales of this larger garlic being cultivated in the ‘long-day’ regions.” See Pis.’ Brief at 17; see also Decision Memorandum at 41. The Chinese Producers’ assessment is not far off the mark. The language of Commerce’s summation of the Market Research Report, and the agency’s phrasing of its ultimate conclusion on this point, are telling. The stated rationale for the agency’s selection of Agmarknet’s “China” variety data as the surrogate value for raw garlic bulb in this case was largely speculative and conclusory, and lacks adequate support in the evidentiary record. Even the Domestic Producers are “not prepared to defend Commerce’s choice of the price for garlic designated as ‘China’ in the Agmarknet data as the surrogate value for raw garlic bulb.” Def.-Ints.’ Brief at 17. All that the Domestic Producers are prepared to defend is “the principle underlying Commerce’s choice — that is, the notion that the surrogate value for raw garlic bulb should, as much as possible, be representative of the large-bulb garlic exported by respondents to the United States.” Id. Such a “notion,” however, lends Commerce no support in its decision to select the “China” variety of Agmarknet garlic, in particular, as a basis for surrogate value here. The Government struggles to mount a defense of Commerce’s decision, gamely arguing that the price differential between “China” variety garlic and other Indian garlic “would be expected if the ‘China’ variety garlic reflected the larger Agrifound Parvati bulb.” See Def.’s Brief at 28 (emphases added). However, the conjecture and contingency reflected in “woulds” and “ifs” belie the sound evidentiary basis required to support the agency’s conclusion “that the Agmarknet ‘China’ variety garlic was the most appropriate surrogate for respondents’ garlic.” See id. at 28. The Government emphasizes that “the information regarding garlic bulb derived from the other reported garlic varieties in the Agmarknet database ... did not reveal any physical characteristics demonstrating that they were similar to respondents’ large bulbed garlic.” Defl’s Brief at 28. But the very same can be said of the information on the “China” variety. See Decision Memorandum at 42 (“there are no descriptions provided by Agmarknet which define these variety categories”). The Agmarknet database provides no physical description of any of the listed varieties of garlic bulb, including the “China” variety. Distilled to its essence, the Government’s argument in support of Commerce’s selection of the “China” variety data as the best available information is basically the same as the basis set forth by the agency itself in its Decision Memorandum — that is, that a higher price may be indicative of a larger bulb, and that such a larger bulb may be the Agrifound Parvati bulb, which, in turn, may be similar to the respondents’ garlic bulb. See Def.’s Brief at 26-28; Decision Memorandum at 41-48. The tenuous nature of that logic is self-evident. The Chinese Producers point to other problems with Commerce’s selection of the Agmarknet “China” variety data as well. See generally Pis.’ Brief at 17, 21. For example, the Chinese Producers note that “China” variety garlic is apparently grown only in three Indian states (Punjab, Gujarat, and Haryana). But only one of those states (Punjab) is within the long-day growing region where Agrifound Parvati— the Indian variety which reportedly may be the most similar to the respondents’ Chinese garlic — is found. See id. at 20-21. In other words, in an effort to base surrogate value on the “China” variety of garlic bulb, Commerce selected a dataset made up primarily of wm-long-day-zone regions, and rejected the Chinese Producers’ effort to select a dataset more representative of long-day zones (which the Domestic Producers contend, and Commerce accepted, is most likely where garlic comparable to respondents’ garlic is grown). See id. Such inherently illogical reasoning undermines Commerce’s selection and use of the “China” variety as the basis for surrogate value here. And there are other seeming contradictions in Commerce’s position that are problematic. For example, the Chinese Producers emphasize that Commerce assertedly predicated its decision to use Agmarknet data on the fact that the Agmarknet data were broad-based and country-wide. Yet Commerce ultimately based the surrogate value on data from a subset of Indian garlic (the “China” variety) grown only in three Indian states, which essentially contradicted the agency’s rationale for using the Agmarknet data in the first place. See Pis.’ Brief at 20-21; see also Decision Memorandum at 42 (emphasizing that “Agmarknet data are broadly representative of garlic bulb prices throughout India”). Underscoring the apparent irony of Commerce’s decision, the Chinese Producers note that Commerce rejected their proposal to use sales prices contained in the Agmarknet database for states where the Market Research Report indicates that high-yield, high-quality garlic predominates (ie., the states of Himachal Pradesh, Uttaranchal, and Jammu and Kashmir), based on the agency’s reasoning that: it is [Commerce’s] practice to use country-wide data instead of regional data when the former is available. Moreover, we attempt to find the most representative, least distortive ... market-based value. The more broad-based the value, the greater the likelihood that the value [is] representative. Pis.’ Brief at 20-21 (quoting Decision Memorandum at 45). Thus, not only is Commerce’s selection of “China” variety garlic as the “best available information” unsupported by the evidence, but it also seems inherently inconsistent with Commerce’s stated rationale for rejecting the respondents’ proposed data (ie., that the data were not country-wide). Finally, the Chinese Producers maintain that, because the Agmarknet data that Commerce used constituted price information representative of the final product (fresh garlic sold at market), the data intrinsically already reflected post-harvest factors of production. See Pis.’ Brief at 11-12. Thus, the Chinese Producers contend, Commerce impermissibly inflated the surrogate value of fresh garlic by adding additional post-harvest factors of production (e.g., sales, packing, and transportation costs) to a figure that already reflected such costs. See id. at 11-12; see also Pis.’ Reply Brief at 5. While there is a certain logic to the Chinese Producers’ argument, they have not demonstrated that use of the intermediate input methodology in fact results in double counting. On remand, however, Commerce should consider this possible inconsistency and the potential for double counting that may result when using data from the Agmarknet database, which presumably contains information regarding Indian market transactions and is representative of the final garlic product rather than an intermediate garlic product (ie., garlic bulb). See Pis.’ Brief at 12-14. As the Chinese Producers emphasized at oral argument, Commerce’s use of Agmarknet data for the intermediate input appears to conflict with its prior claim that raw garlic — the intermediate input — is not the subject merchandise because it requires further processing to become market-ready fresh garlic (the final product). See Recording of Oral Argument at 1:50:00. As Commerce noted, the post-harvest factors of production in India may differ from those of respondents. But, on remand, Commerce should be mindful that, when valuing an intermediate product in an NME country case, it must find a surrogate representative of that intermediate product. See Decision Memorandum at 52-53; see also Def.’s Brief at 29-30. Here, it is unclear whether Agmarknet’s Indian garlic is adequately representative of the respondents’ “intermediate” garlic bulb. Commerce has broad discretion in this arena; but its decisions nonetheless cannot be arbitrary, and must have a solid foundation in the evidentiary record. The Government here contends that Commerce selected the “China” variety of garlic bulb as listed in the Agmarknet database because that garlic comes from three different states in different regions (unlike the regional data proposed by the Chinese Producers, which — the Government claims' — ■ may be subject to regional price distortions). See Def.’s Brief at 31. The Chinese Producers’ proposed data set may or may not be the “best available information”; but, contrary to the Government’s assertions, Commerce has not established that the long-day-region data proffered by the Chinese Producers suffers from any distortions. See id. In light of the numerous flaws and discrepancies in Commerce’s reasoning (as summarized above), Commerce has failed to adequately support its selection of Agmarknet’s “China” variety garlic bulb as the basis for its surrogate valuation of respondent’s garlic bulb input. This issue therefore must be remanded to the agency, so that it may reconsider its valuation of respondents’ garlic bulb input, taking into consideration the evidence and the arguments that the Chinese Producers have advanced here. C. Wage Rate Calculation Taking issue with Commerce’s wage rate calculation, the Chinese Producers contend that the PRC wage rate used in this review — calculated pursuant to the agency’s “long-standing regression-based methodology” — does not reflect the wage rates in market economy countries comparable to the PRC; that Commerce failed to explain why the results produced by the regression methodology are more accurate than the alternatives proposed by the Chinese Producers; and that Commerce’s calculation contravenes established precedent. See generally Pis.’ Brief at 24-31; Pis.’ Reply Brief at 8-9; see also Def.’s Brief at 55 (referring to Commerce’s “long-standing regression-based methodology”); Expected Non-Market Economy Wages: Request for Comment on Calculation Methodology, 70 Fed.Reg. 37,761 (June 30, 2005) (same). Because Commerce has failed to adequately explain, justify, and support its methodology (including its application in this case), this issue must be remanded to Commerce for further consideration. When constructing the normal value of a product from an NME country, Commerce must determine the “hours of labor required” as a factor of production. See 19 U.S.C. § 1677b(c)(3). Like other factors of production, Commerce is directed to value labor “utilizing], 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.” 19 U.S.C. § 1677b(c)(4). In doing so, Commerce essentially creates a “hypothetical” market value to approximate the production experience in the NME country. See Nation Ford, 166 F.3d at 1377-78. The valuation of an NME-country producer’s cost of labor, however, is treated differently from the valuation of other factors of production. See Dorbest v. United States, 30 CIT 1671, 1703, 462 F.Supp.2d 1262, 1291 (2006), appeal docketed, No.2009-1257, -1266 (Fed.Cir. Mar. 20, 2009) (“Dorbest /”) (citing 19 C.F.R. § 351.408(c)(3) (2003)); see also Decision Memorandum at 50. Commerce is permitted to depart from typical surrogate valuation and to value factors of production according to source data outside of the data from the chosen surrogate country — provided that the “methodology used by Commerce is based on the best available information and establishes antidumping margins as accurately as possible.” Shakeproof 268 F.3d at 1381-82 (“we have specifically held that Commerce may depart from surrogate values when there are other methods of determining the ‘best available information’ regarding the values of the factors of production”). Commerce has determined that “in calculating wage rates, an analysis different in some aspects from valuing other [factors of production is] warranted in light of [the agency’s] concerns about wide variances in wage rates between comparable economies.” See Decision Memorandum at 50. Accordingly, in valuing the cost of labor in NME cases, Commerce employs “regression-based wage rates reflective of the observed relationship between wages and national income” in a variety of market economy countries. 19 C.F.R. § 351.408(c)(3) ; see also Dorbest I, 30 CIT at 1703, 462 F.Supp.2d at 1291. In other words, unlike its valuation of other factors of production in an NME case, Commerce bases its surrogate wage rate on data from a broad “basket” of countries, and does not limit itself to market economy countries at a level of economic development comparable to the NME country in question. See Dorbest I, 30 CIT at 1706, 462 F.Supp.2d at 1293. In the case at bar, Commerce employed a labor rate of $0.97 per hour — • Commerce’s calculated wage rate for 2003 for the PRC, derived through the agency’s regression-based analysis of a wage rate data set from some 50-plus market economy countries, including “the most industrialized and advanced countries in the world,” using data from the Yearbook of Labour Statistics published by the International Labour Organization (“ILO”). See Decision Memorandum at 48-51; see also Preliminary Results, 70 Fed.Reg. at 69,950. The Chinese Producers contested the Preliminary Results, raising three main objections. First, the Chinese Producers argued that the statute requires Commerce to calculate a wage rate based on wage rate data from the principal surrogate country, India — a market economy country “at a level of economic development comparable to that of the nonmarket economy country” which is a “significant producer!] of comparable merchandise.” See Decision Memorandum at 47; Pis.’ Brief at 24-25; 19 U.S.C. § 1677b(c)(4). The Chinese Producers further challenged the inclusion in Commerce’s regression analysis of non-comparable countries (such as Switzerland, the United Kingdom, Norway, and Germany), in conflict with the asserted statutory directive to use surrogate values derived from economically comparable countries. See Decision Memorandum at 47-48; Pis.’ Brief at 24-25;19 U.S.C. § 1677b(c)(4)(A). Finally, the Chinese Producers argued that Commerce’s labor rate calculation “inexplicably ignored” the available 2003 data for 14 countries, and should include all available data. See Decision Memorandum at 48; Pis.’ Brief at 25. In its Final Results, Commerce rejected the Chinese Producers’ arguments, and reaffirmed that, for the administrative review at issue here, “the appropriate surrogate value for the wage rate for the PRC respondents continues to be the wage rate of $0.97/hour.” See Decision Memorandum at 49-51. The Chinese Producers argue that the Final Results on the surrogate value for the wage rate in this case comprise “arguments, conclusions and findings that are directly contradicted by record evidence,” such that Commerce’s determination is not supported by substantial evidence. See Pis.’ Brief at 26-27. The Chinese Producers target, in particular, Commerce’s conclusion that its regression model “reflects ‘market economy wage rates at a comparable level of economic development.’ ” Id. at 26 (quoting Decision Memorandum at 50). The Chinese Producers contend that Commerce’s conclusion “is directly contradicted by record evidence ... [which] show[s] that the regression model does not reflect, and in fact drastically overstates, the market wage rates in countries at a comparable level of economic development.” Id. The Chinese Producers emphasize that Commerce’s calculated wage rate of $0.97/ hour is “400% higher than India’s actual wage rate of $0.23/hr. and much higher than the wage rates of other countries found to be economically comparable to China.” Id. The Chinese Producers further note that Commerce’s calculated wage rate for China “is much higher than the actual rates of Pakistan ($0.38), Sri Lanka ($0.34), and the Philippines ($0.80), countries found by Commerce to be economically comparable to China.” Id. The Government does not specifically dispute any of the facts as stated by the Chinese Producers. Nor does the Government respond directly to the Chinese Producers’ charge that Commerce’s conclusion that the agency’s regression model reflects “market economy wage rates at a comparable level of economic development” is not supported by substantial evidence (except that the Government baldly asserts at one point in its brief that “Commerce’s wage rate calculation is supported by substantial evidence”). See Def.’s Brief at 54 (emphasis added); see generally id. at 54-58. Instead, the Government argues that Commerce properly decided “not to rely on the sole wage rate from the selected surrogate country because, while per capita [Gross National Product] rates and wages are positively correlated, there is great variation in the wage rates of the market economy countries that it treats as economically comparable.” See Def.’s Brief at 55 (quoting Decision Memorandum at 49) (internal quotation marks omitted). The Government explains that “[u]n-like other surrogate values for valuing merchandise, such as water, ocean freight, etc., which do not vary significantly from country to country, labor is unique because ‘immigration, welfare and general wage support programs’ can greatly influence wage rates in each country and ‘two economically comparable economies’ can have very different wage rates as a result.” Id. 0quoting Decision Memorandum at 49). Accordingly, the Government states, “since 2000, the regression methodology has been ‘based on data from a wide range of [56] market economy countries’ to enhance ‘the accuracy, predictability and stability of the wage rate.’ ” Id. (quoting Decision Memorandum at 49). The Government similarly argues that Commerce properly declined to expand the “basket” of labor data used in its analysis to include additional countries. See id. at 56-58. However, mere arguments and explanations do not constitute “evidence” — much less “substantial evidence.” As the Chinese Producers point out, Dorbest I took note of the same types of seeming distortions (vis-a-vis Commerce’s labor rate calculation for 2002), and remanded the matter to Commerce. See Pis.’ Brief at 27 (citing Dorbest I, 30 CIT at 1710-13, 462 F.Supp.2d at 1296-98). Dorbest I similarly questioned the bases for Commerce’s assertion that its regression methodology in that case enhanced “the accuracy, predictability and stability” of the wage rate used there' — the same justification that Commerce invokes here. See Dorbest I, 30 CIT at 1710-12, 462 F.Supp.2d at 1297-98; Decision Memorandum at 49 (stating that Commerce “finds that its regression methodology, based on data from a wide range of market economy countries, enhances the accuracy, predictability and stability of the wage rate”). Just as Commerce’s determination in Dorbest I was found to be unsupported by substantial evidence, so too Commerce’s determination here is lacking. The matter therefore must be remanded to the agency for further consideration. The Chinese Producers also criticize the Final Results on the grounds that Commerce “failed to give valid reasons for the selection of its methodology and [its] rejection of the alternatives” that the Chinese Producers proposed — specifically, “calculating the labor rate based on 1) the wage rate in India; or 2) all available data from market economy countries.” See Pis.’ Brief at 28-29. The Government first asserts generally that “Commerce has no affirmative obligation to prove that its regulation [providing for the use of regression-based wage rates] is superior” to alternatives, citing Lasko Metal for the proposition that “policies adopted by Commerce are reasonable when they enhance accuracy, fairness, and predictability.” See Def.’s Brief at 56 (citing Lasko Metal, 43 F.3d at 1446). As noted above, however, it is Commerce’s unsupported assertion that the agency’s methodology enhanced “the accuracy, predictability and stability of the wage rate” that lies at the very heart of this dispute. See Decision Memorandum at 49; see generally Dorbest I, 30 CIT at 1710-12, 462 F.Supp.2d at 1297-98. Moreover, the Government’s argument does not speak to the Chinese Producers’ challenge to the size and composition of the “basket” of countries selected for Commerce’s regression analysis. The Government’s argument thus fails to meet the thrust of the Chinese Producers’ point. As to the Chinese Producers’ first proposed alternative — that Commerce value wage rates using only India wage data'— the Government argues simply that the wage rate methodology set forth in its regulations “expressly rejects the India-only methodology posited by [the Chinese Producers], and is consistent with 19 U.S.C. § 1677b(c) and Commerce’s overriding obligation to construct the most accurate normal value.” See Def.’s Brief at 55-56. But, as the Chinese Producers note, Commerce’s stated underlying rationale is that “using data from a single market-economy country comparable to China (i.e., India, Indonesia, Pakistan, Sri Lanka, and the Philippines) ‘would not provide [Commerce] with a sufficiently large data set to conduct a reliable regression analysis.’ ” Pis.’ Brief at 28 (quoting Decision Memorandum at 50). The Chinese Producers further note that Dorbest I faulted Commerce for its failure to explain its selection of the regression methodology over other possible alternatives, stating: “While Commerce’s model may be the best information available on the record, ... Commerce has failed to give a viable explanation for its choice in light of [the] possible distortion of its predicted wage rates for countries such as the PRC.” See id. (quoting Dorbest I, 30 CIT at 1711, 462 F.Supp.2d at 1297). Similarly, Commerce in this case has failed to adequately explain its decision “to use a regression based methodology that [at least arguably] results in distorted and unrepresentative wage rates versus a single wage rate from India, the country [Commerce] has found most comparable to China.” See id. The Government also seeks to brush aside the Chinese Producers’ second proposed alternative (that Commerce expand its “basket” of labor data to include data from 19 additional countries), asserting that there is no “record evidence demonstrating that such a change would result in a more accurate wage rate.” Def.’s Brief at 56; see also id. at 56-58. As the Chinese Producers observe, however, the Government is attempting to gloss over key findings and determinations in the Final Results, such as Commerce’s finding that its analysis in this case was based “on a basket of countries ... that is sufficient