Camau Frozen Seafood Processing Import Export Corp. v. United States

OPINION AND ORDER POGUE, Chief Judge: This consolidated action returns to court following remand for a second redetermination of the final results of the fifth administrative review of an antidumping duty order covering certain frozen warmwater shrimp from the Socialist Republic of Vietnam (“Vietnam”). 2 At issue is the estima *1331 tion by the United States Department of Commerce (“Commerce”) of a surrogate fair market labor wage rate for the shrimping industry in Vietnam, which Commerce treats as a non-market economy (“NME”). In its 2d Remand Results, Commerce claims that the court’s second remand order compelled the agency to use data from more than one country when calculating surrogate labor values in this review, contrary to Commerce’s new labor rate policy. But Commerce’s claim is incorrect. Rather, the court’s prior decisions required that Commerce address, evaluate, and weigh the conflicting record evidence regarding the appropriateness of its surrogate data choices for valuing the relevant factors in this review, including labor. Commerce has yet to do so. Consequently, the 2d Remand Results must again be remanded for additional consideration, consistent with Camau I, Camau II, and this opinion. BACKGROUND Because Commerce treats Vietnam as an NME country, 3 Commerce determines the normal value of merchandise from Vietnam by using surrogate market economy data to calculate production costs and profit. See 19 U.S.C. § 1677b(c)(l) (2006). In doing so, Commerce’s valuation of the factors of production (“FOPs”) must 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 [agency].” Id. “[T]o the extent possible,” Commerce is required to use data from countries that are both economically comparable to the NME and significant producers of comparable merchandise. Id. at § 1677b(c)(4). In the past, Commerce generally valued the labor FOP for NME countries by using “regression-based wage rates reflective of the observed relationship between wages and national income in. market economy countries.” 19 C.F.R. § 351.408(c)(3) (2010). 4 Regression-based NME wage rates “estimate[d] the linear relationship between yearly per capita gross national income (‘[GNI]’) and hourly wage rate (‘wage’)” to arrive at the wage for an NME country by using the NME’s GNI. 5 *1332 During the prior (fourth) administrative review of this antidumping duty order, however, 19 C.F.R. § 351.408(c)(3) was invalidated as contrary to the statute because it did not rely exclusively on data from economically comparable countries that are significant producers of comparable merchandise. 6 Consequently, in that prior fourth review Commerce used a new method for calculating the surrogate wage rate when determining the normal value of subject merchandise from Vietnam. Explaining its new method, Commerce specifically rejected proposals to calculate the surrogate wage rate using data solely from Bangladesh — the chosen primary surrogate country. Commerce declared: While information from a single surrogate country can reliably be used to value other FOPs, wage data from a single surrogate country does not constitute the best available information for purposes of valuing the labor input due to the variability that exists between wages and GNI. While there is a strong world-wide relationship between wage rates and GNI, too much variation exists among the wage rates of comparable [market economies]. As a result, we find reliance on wage data from a single country to be unreliable and arbitrary. 7 That is, Commerce rejected proposals to base Vietnam’s surrogate wage rate on data from Bangladesh because, although Bangladesh is sufficiently economically comparable to Vietnam for the purpose of valuing the other FOPs, the observed strong linear relationship between wage rates and GNI suggests that data from Bangladesh, which has a GNI roughly half that of Vietnam, 8 are unlikely to be representative of a fair market wage rate in Vietnam. The surrogate wage rate ultimately calculated for Vietnam in the fourth review was $0.89. ARk I & D Mem. cmt. 9 at 31. Before the results of this (fifth) review were finalized, however, Commerce published its determination that, in light of the recent judicial decisions constraining the available dataset for calculating surrogate FOP values in NME cases, 9 Commerce was changing its policy from a preference for using data from multiple market economies when constructing surrogate labor *1333 rates to a policy of relying on data from a single market economy to calculate all surrogate FOPs, including labor. 10 For its final results of this review, therefore, Commerce employed the New Labor Rate Policy to arrive at the surrogate wage rate used to construct normal value. I & D Mem. cmt. 21. Using data solely from the primary surrogate country, Bangladesh, Commerce calculated a surrogate wage rate for Vietnam’s shrimping industry of $0.21. 11 Responding to the Ad Hoc Shrimp Trade Action Committee 12 (“AHS-TAC”)’s challenge to the application of Commerce’s New Labor Rate Policy in this review, this Court held that although the New Labor Rate Policy is reasonable on its face, Commerce’s conclusion that Bangladesh provided the best available data from which to value all FOPs in this review, including labor, could not be sustained without further evaluation and explanation. Camau I, — CIT at-, 880 F.Supp.2d at 1358-61. As the court explained, though Commerce may use a single surrogate country for all FOPs (which it is statutorily neither required to do nor prohibited from doing), the reasonableness of using that country’s data must be explained where the evidence and factual findings on record may fairly detract from the weight of Commerce’s determination. Id. 13 Specifically, the court remanded the Final Results because the record evidence included Commerce’s prior findings that 1) GNI is linearly correlated to wage rates 14 ; 2) Commerce’s economic comparability analysis allows for a fairly wide range of GNI values to satisfy the economic comparability criterion for surrogate market economy countries 15 ; and 3) Bangladesh’s GNI, equaling roughly half of Vietnam’s, is sufficiently disparate from that of Vietnam that, given 1) above, using solely the wage rate data from Bangladesh would likely understate the estimate for a fair *1334 market wage rate in Vietnam. 16 These are findings that fairly detract from the reasonableness of Commerce’s conclusion that the $0.21 wage rate derived from Bangladeshi data provides the best information available regarding the market wage rate that would be Vietnam’s if Vietnam were a market economy. Because Commerce did not address these findings and explain the continued reasonableness of its decision notwithstanding these factual circumstances, the court remanded Commerce’s determination for additional consideration and/or more explanation. Camau I, — CIT at-, 880 F.Supp.2d at 1358-61. Responding to the court’s first remand order, Commerce continued to insist that, given the court’s recent decisions directing the agency to keep within the bounds of its statutory authority, and given the relevant statutory constraints, the agency is now justified in using data, without further evaluation, that it had previously rejected as arbitrary. See 1st Remand Results. 17 The court remanded again, again holding that Commerce must explain its conclusion to account for the evidence that using Bangladeshi wage data would likely significantly undervalue the surrogate wage rate due to the roughly 50 percent GNI disparity between Bangladesh and Vietnam. See Camau II, —- CIT at-, 929 F.Supp.2d at 1354-58. Now, in its second redetermination, Commerce has thrown up its proverbial hands, maintaining “under respectful protest” that the court’s decisions have dictated to the agency to average data from multiple countries when determining surrogate labor FOP values, contrary to the New Labor Rate Policy. See 2d Remand Results at 7-8. STANDARD OF REVIEW This Court will uphold Commerce’s determinations on remand if they are in accordance with law, consistent with the court’s remand order, and supported by substantial evidence. See 19 U.S.C. § 1516a(b)(l)(B)(i); Trust Chem Co. v. United States, — CIT -, 819 F.Supp.2d 1373, 1378 (2012). Substantial evidence is “such relevant evidence as a *1335 reasonable mind might accept as adequate to support a conclusion,” Universal Camera, 340 U.S. at 477, 71 S.Ct. 456 (internal quotation marks and citation omitted), and the “substantiality of evidence must take into account whatever in the record fairly detracts from its weight.” Id. at 488, 71 S.Ct. 456. Although “a court is not to substitute its judgment for that.of the agency,” the court must ensure that the agency “examine[d] the relevant data and articulate[d] a satisfactory explanation for its action including a ‘rational connection between the facts found and the choice made.’ ” State Farm, 463 U.S. at 43, 103 S.Ct. 2856 (quoting Burlington Truck Lines, Inc. v. United States, 371 U.S. 156, 168, 83 S.Ct. 239, 9 L.Ed.2d 207 (1962)). In providing the required explanation for its action, the agency “must cogently explain why it has exercised its discretion in a given manner,” id. at 48, 103 S.Ct. 2856 (citations omitted), and “supply a reasoned analysis” that comports with its factual findings and the evidentiary record. Id. at 57, 103 S.Ct. 2856 (internal quotation marks and citation omitted). “The grounds upon which an [agency action] must be judged are those upon which the record discloses that [the] action was based.” Changzhou Wujin Fine Chem. Factory Co. v. United States, 701 F.3d 1367, 1377 (Fed.Cir.2012) (quoting SEC v. Chenery Corp., 318 U.S. 80, 87, 63 S.Ct. 454, 87 L.Ed. 626 (1943)). “Review of an administrative decision must be made on the grounds relied on by the agency. If those grounds are inadequate or improper, the court is powerless to affirm the administrative action by substituting what it considers to be a more adequate or proper basis.” Id. at 1379 (internal quotation marks and citations omitted). DISCUSSION Here, Commerce’s second remand redetermination was the result of a methodology applied “under respectful protest.” 2d Remand Results at 7-8. That is, the agency acted not from its own analysis and conclusions, but rather based upon a belief that it was compelled to act as it did by this Court’s decisions in Camau I and Camau II. See id. 18 But Commerce is incorrect that either of those decisions compelled the agency to act as it did in the 2d Remand Results. As discussed above, 19 the court did not order Commerce to do anything more than what is required of it pursuant to established principles of administrative law— namely, to provide a reasoned and reasonable explanation connecting its conclusion to the record evidence, including the evidence suggesting that Bangladesh’s wage data is likely to significantly understate the estimated fair market wage rate in Vietnam. 20 The requirement to rethink and/or further explain the agency’s ap *1336 proach in order to reach results that are supported by a reasonable reading of the entire record is not synonymous (as Commerce implies) with a compulsion to employ any particular approach. Because Commerce’s 2d Remand Results were reached based on an erroneous belief that the specific determinations contained therein were compelled by the decisions of this Court, these results cannot be affirmed on the basis provided. 21 While the court has ordered the agency to support its determination with sufficient explanation of the entire evidentiary record, it is for Commerce to weigh and analyze the conflicting evidence and provide a reasoned explanation for the outcome of such weighing. 22 Thus the 2d Remand Results — which abdicate the agency’s responsibility by announcing that the determinations contained therein are not the result of the agency’s own analysis but rather the apparent implementation of an erroneous reading of this Court’s decisions 23 — cannot be affirmed. This issue must therefore again be remanded for reconsideration, consistent with Camau I, Camau II, and this opinion. Importantly, it is simply not the case that the only alternative to Commerce’s Final Results and 1st Remand Results is to deviate from the New Labor Rate Policy and average the Bangladeshi wage data with other data. As the court held in Camau I, Commerce reasonably determined that, in general, the administrative costs of engaging in a complex and lengthy analysis of additional surrogate data for the labor FOP may outweigh the accuracy-enhancing benefits of doing so. See Camau I, — CIT -, 880 F.Supp.2d at 1358. But as the court also held in Camau I, the particular evidentiary record of this review includes Commerce’s prior finding that Bangladeshi wage data are likely to significantly understate the estimate for a fair market labor rate in Vietnam. See id. at 1360-61. All of the factual premises on which Commerce based its determination not to use the Bangladeshi wage data in the previous review remain in effect — Bangladesh’s GNI remains roughly half that of Vietnam’s and Commerce’s findings regarding the positive linear correlation between GNI and wage rates remain uncontroverted. 24 But in the Final Results and 1st Remand Results, Commerce did not address the relative weight of this prior finding when determining that data from Bangladesh provide the best available information from which to value all of the surrogate FOPs in this review. 25 One option that continues to be available to Commerce on remand, therefore, is to explicitly weigh the evidence that Bangladeshi wage data are likely to understate *1337 the surrogate fair market labor rate for the shrimping industry in Vietnam against the remaining evidence (if any) that Bangladeshi surrogate FOP data as a whole are nevertheless the best available data on record from which to value all of the surrogate FOPs in this review. 26 The agency may want to consider: — Is the data from Bangladesh with regard to the other FOPs so superi- or in quality to that from any other potential surrogate that the accuracy-enhancing benefits of using such data outweigh the accuracy-loss resulting from the wage rate undervaluation? — Is accuracy/data quality with regard to the other FOPs more important than accuracy with regard to labor? — How great is the effect of an undervalued wage rate on the accuracy of the resulting dumping margin? — How great is the administrative effort involved in analyzing data for the purpose of adjusting the Bangladeshi wage data to increase accuracy? — Does this effort outweigh its accuracy-enhancing benefits? — Is there anything about the interrelationship between the Bangladeshi data for the respective FOPs that makes the use of such data relatively more accuracy-enhancing than using FOP data from another surrogate country with a GNI closer to Vietnam’s? — Conversely, does Commerce’s inability to explain and account for the labor undervaluation suggest that perhaps another surrogate country choice may be more reasonable? — Are there additional data from Bangladesh that Commerce could use to adjust the wage data to correct, or at least diminish or ameliorate, the likely undervaluation? It may be that, upon weighing the evidence, Commerce decides that the reasons supporting the use of Bangladesh as the primary surrogate country outweigh the trade-off of losing some accuracy with regard to the labor FOP value. As the court has held, Commerce is not required to deviate from its New Labor Rate Policy and use data from more than one country when calculating the labor FOP in this case. But Commerce must address its prior finding that Bangladeshi wage data are likely to understate the fair market rate in Vietnam and weigh the impact of this finding on the accuracy of the resulting dumping analysis against the explicit benefit(s) — if any — that nevertheless support the continued use of Bangladeshi data to construct a normal value in this case. Should Commerce choose to engage in such evidence-weighing, however, the agency must explicitly lay out the value-choices and data preferences it is making, so that the path of its analysis may reasonably be discerned as based on some set of predictable standards, as well as to provide a basis for judicial review. For while it is Commerce’s job to weigh the evidence, the court’s role on review is to ensure that such weighing is done explicitly and reasonably. 27 The court cannot do so if Commerce, rather than laying out the reasonable value choices it makes in giving more or less weight to some aspects of the evidentiary record than to others, fails to *1338 acknowledge that the evidence is conflicting. 28 In sum, Commerce’s valuation of the labor FOP used to construct a normal value for the subject merchandise in this review remains without an adequate reasoned explanation linking it to the record evidence. In the original determination and in the 1st Remand Results, the decision to use solely the identified data from Bangladesh to value the market labor rate for Vietnam’s shrimping industry was not reasonably explained in light of Commerce’s outstanding and unaccounted-for prior finding that these data are likely to understate the estimate due to the GNI disparity between Vietnam and Bangladesh. And the 2d Remand Results do not provide the requisite support for affirmance because they are not grounded in an analysis of the factual record but are the apparent result of a mistaken belief in a compulsion to reach such results. Accordingly, this matter is again remanded for further consideration. CONCLUSION For all of the foregoing reasons, Commerce’s 2d Remand Results are remanded for further consideration, consistent with this opinion and the decisions in Camau I and Camau II. Commerce shall have until May 6, 2014, to complete and file its remand results. Plaintiffs shall have until May 20, 2014, to file comments. The parties shall have until June 3, 2014, to file any reply. It is SO ORDERED. Footnotes [2] See Certain Frozen Warmwater Shrimp from the Socialist Republic of Vietnam, 76 Fed.Reg. 56,158 (Dep't Commerce Sept. 12, 2011) (final results and final partial rescission of anti-dumping duty administrative review) ("Final Results ") and accompanying Issues & Decision Mem., A-552-802, ARP 09-10 (Aug. 31, 2011) (“I & D Mem.") cmt. 21; Camau Frozen Seafood Processing Imp. Exp. Corp. v. United States, - CIT -, 880 F.Supp.2d 1348 (2012) ("Camau I") (remanding Final Results ); Final Results of Redetermination Pursuant to [Camau /] (Nov. 15, 2012), ECF No. 90 ("1st Remand Results "); Camau Frozen *1331 Seafood Processing Imp. Exp. Corp. v. United States, - CIT -, 929 F.Supp.2d 1352 (2013) ("Camau II") (remanding 1st Remand Results); Final Results of Redetermination Pursuant to [Camau 17] (July 31, 2013), ECF No. 107-1 ("2d. Remand Results"). Familiarity with the facts and procedural posture of this case is presumed. Facts most relevant to the legal issues presented are briefly summarized in the 'Background' section of this opinion. [3] See Final Results, 76 Fed.Reg. at 56,160. [4] See Antidumping Duties; Countervailing Duties, 61 Fed.Reg. 7308, 7345 (Dep’t Commerce Feb. 27, 1996) ("[W]hile per capita [gross domestic product] and wages are positively correlated, there is great variation in the wage rates of the market economy countries that [Commerce] typically treats as being economically comparable. As a practical matter, this means that the result of an NME case can vary widely depending on which of the economically comparable countries is selected as the surrogate.... [U]se of [regression-based] wage rate[s] will contribute to both the fairness and the predictability of NME proceedings. By avoiding the variability in results depending on which economically comparable country happens to be selected as the surrogate, the results are much fairer to all parties.”). [5] Zhejiang DunAn Herian Metal Co. v. United States, - CIT -, 707 F.Supp.2d 1355, 1366 (2010) (footnote omitted), vacated on other grounds, 652 F.3d 1333 (Fed.Cir.2011); see also Dorbest Ltd. v. United States, 604 F.3d 1363, 1371 (Fed.Cir.2010) ("Commerce determines a linear trend that best fits the data, providing a way to predict the labor rate for a country with any given gross national income.”). [6] See Dorbest, 604 F.3d at 1372 (invalidating 19 C.F.R. § 351.408(c)(3) as contrary to 19 U.S.C. § 1677b(c)(4) because the regulation "improperly require[d] using data from both economically comparable and economically dissimilar countries, and it improperly use[d] data from both countries that produce comparable merchandise and countries that do not”). [7] Certain Frozen Warmwater Shrimp from the Socialist Republic of Vietnam, Issues & Decision Mem., A-552-802, ARP 08-09 (July 30, 2010) (adopted in 75 Fed.Reg. 47,771 (Dep’t Commerce Aug. 9, 2010) (final results and partial rescission of antidumping duty administrative review)) ("AR4 I & D Mem.”) cmt. 9 at 27. See also supra note 4. [8] See Camau I, - CIT at -, 880 F.Supp.2d at 1359-60 & n. 12 (discussing the GNI data on record). [9] See Dorbest, 604 F.3d at 1371-72 (holding that because the statute requires Commerce to use data from economically comparable countries "to the extent possible,” Commerce may not employ a methodology that requires using data from both economically comparable and economically dissimilar countries, in the absence of a showing "that using the data Congress has directed Commerce to use is impossible”); Shandong Rongxin Imp. & Exp. Co. v. United States, - CIT -, 774 F.Supp.2d 1307, 1316 (2011) (holding that because the statute requires Commerce to use, "to the extent possible,” data from countries that are "significant” producers of comparable merchandise, Commerce may not employ a methodology that requires using data from "countries which almost certainly have no domestic production — at least not any meaningful production, capable of having influence or effect”). [10] Antidumping Methodologies in Proceedings Involving Non-Market Economies: Valuing the Factor of Production: Labor, 76 Fed.Reg. 36,092 (Dep't Commerce, June 21, 2011) {“New Labor Rate Policy ”). [11] See Camau I, - CIT at -, 880 F.Supp.2d at 1359-60 & n. 12 (explaining how the $0.21 rate was derived from the evidence in this review). [12] AHSTAC is an association of manufacturers, producers, and wholesalers of a domestic like product in the United States that participated in this review. CompL, Court No. 11-00383, ECF No. 8, at ¶ 9. [13] Established principles of administrative law, while permitting the agency to change course and adopt a new policy within the scope of its statutory authority (as this new policy is), require the agency to explain how applying the new policy is consistent with the evidence and prior factual findings on record. See Motor Vehicle Mfrs. Ass’n of the U.S., Inc. v. State Farm Mutual Auto. Ins. Co., 463 U.S. 29. 46-48, 103 S.Ct. 2856, 77 L.Ed.2d 443 (1983) (holding that an agency may not change course without addressing the continued relevance of factual findings on which the agency’s prior policy was based); FCC v. Fox Television Stations, Inc., 556 U.S. 502, 537, 129 S.Ct. 1800, 173 L.Ed.2d 738 (2009) (J. Kennedy, concurring in part and concurring in judgment) (explaining that State Farm followed the principle that an agency "cannot simply disregard contrary or inconvenient factual determinations that it made in the past, any more than it can ignore inconvenient facts when it writes on a blank slate”). For agency action to be based on substantial evidence, the agency must explain why evidence that fairly detracts from the reasonableness of its determination does not outweigh that which supports it. See Universal Camera Corp. v. NLRB, 340 U.S. 474, 488, 71 S.Ct. 456, 95 L.Ed. 456 (1951). [14] See supra notes 4 and 7. [15] Id. [16] See AR41 & D Mem. cmt. 9 at 27-29. [17] Commerce emphasized recent court decisions that have reminded the agency of the statutory constraints on its construction of normal value for NME-originating merchandise, restraining the agency from overstepping its statutory authority by using a wider dataset than is statutorily permitted. See supra note 9 (discussing Dorbest and Shandong ). What the courts held in those cases is that the statute requires Commerce to use, to the extent possible, data from countries that are economically comparable and significant producers of comparable merchandise. As demonstrated in the challenges underlying these judicial decisions, rather than evaluating the extent to which it was possible to base its calculations on such statutorily prescribed data, Commerce was instead formulaically relying on data from countries that did not satisfy one or both of these statutory requirements. Accordingly, the agency’s approach to normal value construction in NME cases required reconsideration. In Camau I, the court suggested that one option for Commerce, on this particular record, may be to use data from the Philippines (for which Commerce also undertook a full potential surrogate analysis, consistent with 19 U.S.C. § 1677b(c)) to analyze (and perhaps correct for) the magnitude of potential undervaluation involved in relying on the Bangladeshi data alone. See Camau I, - CIT-, 880 F.Supp.2d at 1360-61. In response, Commerce unreasonably claimed that the Bangladeshi and Philippine data were wholly incomparable, despite the agency’s routine use of data from both of the sources at issue to value surrogate labor rates (suggesting sufficient comparability in Commerce’s view for these datasets to be at least theoretically interchangeable for the purpose of valuing labor). See Camau II,-CIT-, 929 F.Supp.2d at 1357. [18] See also GPX Int’l Tire Corp. v. United States, - CIT -, 942 F.Supp.2d 1343, 1348 n. 2 (2013) (“The only legitimate purpose of registering a protest in a remand determination is to preserve a particular issue for appeal where the agency has been compelled to take a particular step that results in an outcome not of its choosing.”) (emphasis added). [19] See the ‘Background’ section of this opinion. [20] Contrary to Commerce’s stance in the 2d Remand Results, the court’s insistence on the provision of this required reasoning does not force the agency to abandon its new policy in favor of multi-country averaging because, as discussed below, averaging multiple countries’ wage data is not the only method by which Commerce can reach a result that is consistent with a reasoned and reasonable reading of the evidentiary record. [21] See Changzhou, 701 F.3d at 1379 ("Review of an administrative decision must be made on the grounds relied on by the agency.”) (internal quotation marks and citation omitted). [22] See, e.g., Legacy Classic Furniture, Inc. v. United States, - CIT -, 867 F.Supp.2d 1321, 1328-29 (2012) (requiring Commerce to provide a reasoned analysis or explanation for how it weighed conflicting record evidence). [23] See 2d Remand Results at 7-8. See also supra note 19. [24] As noted above, Commerce has found that, as a matter of economic fact, labor wage rates in market economies tend to be linearly correlated with GNI. One logical implication of this finding is that surrogate market economy countries with a GNI that, though treated as "economically comparable,” is fairly divergent from that of the NME at issue, will provide labor wage data that similarly under- or over-states the estimation of fair market labor rates in the NME. [25] See I & D Mem. cmt. 21 at 24; 1st Remand Results at 7-8. [26] Nor is the agency prohibited from opening the record to obtain additional evidence to adjust or otherwise reconsider the Bangladeshi data. [27] See supra note 23. [28] See supra notes 7, 25-26.