Tri Union Frozen Products, Inc. v. United States

227 F. Supp. 3d 1387, 2017 CIT 71, 2017 Ct. Intl. Trade LEXIS 72, 2017 WL 2558263

• Court: United States Court of International Trade
• Decided: June 13, 2017
• Docket: Consol. 14-00249
• Status: Published

Opinion

OPINION AND ORDER

Kelly, Judge:

Before the court for review is the U.S. Department of Commerce’s (“Commerce” or “the Department”) remand determination filed pursuant to the court’s order in Tri Union Frozen Products, Inc. v. United States, 40 CIT -, 168 F.Supp.3d 1265 (2016) (“Tri Union I”). See Final Results of Redetermination Pursuant to Court Remand, Sept. 1, 2016, ECF No. 118-1 (“Remand Results”).

In Tri Union I, the court granted Defendant’s request to remand “for Commerce to reconsider [Ad Hoc' Shrimp Trade Action Committee]’s arguments concerning Commerce’s reliance on Bangladeshi labor wage rate data” from the Bangladesh Bureau of Statistics (“BBS data”), a government source, to value the labor factor of production in this review. Tri Union I, 40 CIT at -, 163 F.Supp.3d at 1312-13. On remand, Commerce continued to use BBS data to value the labor factor of production (“FOP”) in this review, providing further explanation of its decision to do so in light of Ad Hoc Shrimp Trade Action Committee’s (“Ad Hoc Shrimp”) arguments that the Bangladeshi wage rate data is aberrational and unreliable due to systemic labor abuses in the Bangladeshi shrimp industry. Remand. Results 5-42. For the reasons that follow, the court remands again to Commerce for further consideration of Ad Hoc Shrimp’s argument that record evidence of alleged labor abuses in the Bangladeshi shrimp industry renders the BBS data aberrational, unreliable, and not reflective of actual labor conditions in a market economy at comparable economic development to the Socialist Republic of Vietnam.

BACKGROUND

On March 29, 2013, Commerce initiated the eighth administrative review of the antidumping duty (“ADD”) order covering certain frozen warmwater shrimp from the Socialist Republic of Vietnam (“Vietnam”) for the period of February 1, 2012 through January 31, 2013. See Initiation of Antidumping and Countervailing Duty Administrative Reviews and Request for Revocation in Part, 78 Fed. Reg. 19,197 (Dep’t Commerce Mar. 29, 2013); see also Frozen Warmwater Shrimp from the Socialist Republic of Vietnam, 70 Fed. Reg. 5,152 (Dep’t Commerce Feb. 1, 2005) (notice of amended final determination of sales at less than fair value and ADD order).

Prior to publication of the preliminary results, petitioner Ad Hoc Shrimp submitted comments regarding the primary surrogate country selection. Comments on Surrogate Country Selection, PD 133-137, bar codes 3152484-01-04 (Aug. 30, 2013) (“Ad Hoc Shrimp Surrogate Country Comments”). 1 In this submission, Ad Hoc Shrimp argued that Commerce should not select Bangladesh as a primary surrogate country because, “as a consequence of the pervasive labor abuses in Bangladesh[,] the two countries are not economically comparable.” Id. at 2. Ad Hoc Shrimp placed evidence on the record documenting alleged “aberrational labor conditions— comprised of severe abuse of labor and disregard for workers’ rights—permeating the entire supply chain of the Bangladesh shrimp industry.” Id.

On March 24, 2014, Commerce published its preliminary results. See Certain Frozen Warmwater Shrimp From the Socialist Republic of Vietnam, 79 Fed. Reg. 15,941 (Dep’t Commerce Mar. 24, 2014) (preliminary results of ADD administrative review; 2012-2013) and accompanying Decision Memorandum for Preliminary Results of Antidumping Duty Administrative Review: Certain Frozen Warmwater Shrimp from the Socialist Republic of Vietnam; 2012-2013, A-552-802, PD 191, bar code 3188821-01 (Mar. 19, 2014) (“Prelim. Decision Memo”). Despite Ad Hoc Shrimp’s comments on the surrogate country selection, Commerce selected Bangladesh as the primary surrogate country for the purpose of valuing the mandatory respondents’ FOPs for this review. Prelim. Decision Memo at 15. Regarding the labor FOP, Commerce explained its practice of valuing the labor input using industry-specific labor wage rate data from the primary surrogate country, and accordingly used the BBS data to value the labor input here. Id. at 23; see Surrogate Values for the Preliminary Results at 5-6, PD 192-198, bar codes 3188847-01-07 (Mar. 18, 2014) (“Prelim. Surrogate Value Memo”). Commerce explained that, although it considers the International Labor Organization Yearbook of Labor Statistics Chapter 6A: Labor Cost in Manufacturing (“ILO Chapter 6A data”) to be the best source of data for industry-specific labor rates, because the ILO does not include labor data for Bangladesh, the agency would use the BBS data. Prelim. Surrogate Value Memo at 5-6; see Prelim. Decision Memo at 23. 2

Following publication of the preliminary results, Ad Hoc Shrimp placed on the record additional documentation of alleged labor abuses in the shrimp industry in Bangladesh and ILO data for five countries (Guyana, India, Indonesia, Nicaragua, and Philippines) 3 that Ad Hoc Shrimp contended are economically comparable to Vietnam. Post-Prelim Evidentiary Submission Regarding Surrogate Country and Value Selection at Attach. 8, PD 221-222, bar code 3198211-01 (Apr. 28, 2014) (“Ad Hoc Shrimp Post-Prelim. Comments”). Ad Hoc Shrimp subsequently submitted a case brief to Commerce continuing to challenge the use of the BBS data, again arguing that the BBS data is aberrational and unreliable and highlighting the usable ILO data already on the record for economically comparable countries. Case Br. on Behalf of the Ad Hoc Shrimp Trade Action Committee 8-25, PD 234, bar code 3204785-01 (May 28, 2014) (“Ad Hoc Shrimp Agency Case Br.”). Ad Hoc Shrimp argued to Commerce that, notwithstanding the primary surrogate country selection, the BBS data should not be used to value the labor FOP. 4 Id. at 2, 24.

On September 19, 2014, Commerce issued its final determination in the eighth administrative review of the ADD order covering certain frozen warmwater shrimp from Vietnam for the period of February 1, 2012 through January 31, 2013. See Certain Frozen Warmwater Shrimp From the Socialist Republic of Vietnam, 79 Fed. Reg. 57,047 (Dep’t Commerce Sept. 24, 2014) (final results of ADD administrative review, 2012-2013), as amended, 79 Fed. Reg. 65,377 (Dep’t Commerce Nov. 4, 2014) (amended final results of ADD administrative review, 2012-2013), and accompanying Issues and Decision Memorandum for the Final Results, A-552-802, (Sept. 19, 2014), ECF No. 27-4 (“Final Decision Memo”). In the final determination, Commerce continued to find that the BBS data was the best available information on the record to value labor in this review, stating that its finding is in keeping with its practice to use “industry-specific labor rates from the primary surrogate country.” Final Decision Memo at 47. Commerce again explained that it was unable to use data from its preferred source, ILO Chapter 6A data, as the ILO does not contain data from Bangladesh; therefore, Commerce used data published by the BBS to value the labor FOP. Id.

Plaintiffs Tri Union Frozen Products, Inc., Mazzetta Company LLC, Ore-Cal Corporation, Consolidated Plaintiff Quoc Viet Seaproducts Processing Trading and Import-Export Co., Consolidated Plaintiffs Vietnam Association of Seafood Exporters and Producers (including certain of its individual member companies), and Consolidated Plaintiff Ad Hoc Shrimp respectively moved for judgment on the agency record challenging various aspects of Commerce’s final determination. See Mem. Supp. Mot. Tri Union Frozen Products, Inc. J. Agency R., Mar. 30, 2015, ECF No. 48; Mem. Supp. Mot. Quoc Viet Seaproducts Processing Trading and Import-Export Co., Ltd. J. Agency R., Mar. 30, 2015, ECF No. 46; Resp’t Pis. VASEP and Individual VA-SEP Members’ Br. Supp. Mot. J. Agency R., Mar. 30, 2015, ECF No. 50; Mot. Ad Hoc Shrimp Trade Action Committee for J. Agency R. Under USCIT Rule 56.2, Mar. 30, 2015, ECF No. 49-3 (“Ad Hoc Shrimp Br.”). 5 Ad Hoc Shrimp challenged as unsupported by substantial evidence Commerce’s use of the BBS data to value the labor factor of production in this review, arguing that the BBS data is aberrational and unreliable and renders the, final results of the review unsupported by substantial evidence. Ad Hoc Shrimp Br. 15-30. Additionally, Ad Hoc Shrimp argued that Commerce failed to explain why the BBS data was reliable and non-distortive. See id. at 23-24. In response, Defendant requested remand for Commerce to consider Ad Hoc Shrimp’s arguments that the BBS wage rate data is aberrational. See Def.’s Resp. in Opp’n to Pis.’ Mots. J. Agency R. 88-89, Sept. 10, 2015, ECF No. 73. In Tri Union I, the court sustained Commerce’s final determination in all respects other than its use of Bangladeshi labor wage rate data to value the labor factor of production, granting Defendant’s request for remand to Commerce on that issue. Tri Union I, 40 CIT at-, 163 F.Supp.3d at 1312-13.

On September 1, 2016, Commerce issued the Remand Results. See generally Remand Results. On remand, Commerce continued to rely on the BBS data to value the labor FOP in this review. Id at 5-42. Commerce continued to find that the BBS data provided the best available information for valuing the labor FOP as it re- fleets the agency’s “strong preference to use surrogate values from the primary surrogate country,” is specific to the shrimp industry, and, while not contemporaneous, is closer to the period of review than other data on the record. Id. at 8-10. Commerce also contended that Ad Hoc Shrimp did not demonstrate the data to be aberrational and unreliable because Ad Hoc Shrimp did not provide a “measurable means (La, a benchmark)” by which to assess the data as distortive. Id.,,at 29. Commerce further emphasized that its statutory directive does not require it to consider socio-political factors that may influence industry wage rates. See id. at 17.

Following remand, Ad Hoc Shrimp continues to challenge the BBS data as aberrational, unreliable, and therefore not the best available information with which to value the labor factor of production in this review. See Consolidated PI. Ad Hoc Shrimp Trade Action Committee’s Comments on Final Results of Redetermination to Court Remand 6-30, Dec. 2, 2016, ECF No. 126 (“Ad Hoc Shrimp Remand Comments”). Ad Hoe Shrimp also argues that Commerce has failed to explain why the data is reliable and non-aberrational in light of the record evidence. See id. Defendant argues that substantial evidence supports Commerce’s selection of the Bangladeshi data as the best available surrogate value data, contending that Ad Hoc Shrimp has not demonstrated the data to be, aberrational or unreliable. See Def.’s Resp. to Comments on Remand Results 7-25, Mar. 23, 2017, ECF No. 133 (“Def.’s Remand Comments”);

JURISDICTION AND STANDARD OF REVIEW

The court has jurisdiction pursuant to section 516A of the Tariff Act of 1930, as amended,-19 U.S.C. § 1516a(a)(2)(B)(iii) (2012) 6 and 28 U.S.C. § 1581(c) (2012), which grant the Court authority to review actions contesting the final determination in an administrative review of an anti-dumping duty order. The court will uphold Commerce’s determination unless it is “unsupported by substantial evidence on the record, or otherwise not in accordance with law.” 19 U.S.C. § 1516a(b)(1)(B)(i). “The results of a redetermination pursuant to court remand are also reviewed ‘for compliance with the court’s remand order.’ ” Xinjiamei Furniture (Zhangzhou) Co. v. United States, 38 CIT -, -, 968 F.Supp.2d 1255, 1259 (2014) (quoting Nakornthai Strip Mill Public Co. v. United States, 32 CIT 1272, 1274, 587 F.Supp.2d 1303, 1306 (2008)).

DISCUSSION

Ad Hoc Shrimp argues on remand, as it did before the agency and before this court prior to remand, that the BBS data is unreliable and aberrational, and that Commerce has failed to explain why the data is reliable and non-aberrational in light of the record evidence. See Ad Hoe Shrimp Remand Comments 6-30; see also Ad Hoc Shrimp Br. 13-31; Ad Hoc Shrimp Agency Case Br. at 8-25. Defendant contends that Commerce’s- selection of the BBS data is supported by substantial evidence as Commerce found that it is the best available information on the record to value the labor FOP. Def.’s Retaand Comments 10-25. Defendant ■ also contends that Commerce reasonably determined Ad Hoc Shrimp cannot undermine Commerce’s finding without demonstrating quantitatively that the BBS data are aberrational or unreliable. Id. at 12-25.

In antidumping cases, when the exporting country is a nonmarket economy country (“NME”), 7 Commerce calculates normal value for subject merchandise using FOPs 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].” 8 19 U.S.C. § 1677b(c)(1); 19 C.F.R. §§ 351.408(a)-(c) (2013). 9 For each FOP Commerce selects data from a market economy country that is both economically comparable to the NME country and a significant producer of the merchandise in question. Remand Results 2-3; 19 U.S.C. § 1677b(c)(4)(A)-(B); 19 C.F.R. § 351.408(b). Commerce has a regulatory preference to value all FOPs using data from a single surrogate country, 19 C.F.R. § 351.408(c)(2), and determines what data constitutes the best information using criteria developed through practice. 10 Qingdao Sea-Line Trading Co. v. United States, 766 F.3d 1378, 1386 (Fed. Cir. 2014) (considering data quality, specificity, and contemporaneity); see Prelim. Decision Memo at 23. The objective of using surrogate values is to construct a “hypothetical market value representative of the foreign producers under investigation,” Nation Ford v. United States, 166 F.3d 1373, 1378 (Fed. Cir. 1999), and the use of a “primary surrogate country” serves this objective by providing consistency in data and accurately representing such a hypothetical market value. See Antidumping Methodologies in Proceedings Involving Non Market Economies: Valuing the Factor of Production: Labor, 76 Fed. Reg. 36,092, 36,093 (Dep’t Commerce Jun. 21, 2011) (“Labor Methodologies”).

Commerce endeavors to use data that is non-aberrational and reliable. See Antidumping Duties; Countervailing Duties, 62 Fed. Reg. 27,296, 27,366 (Dep’t Commerce May 19, 1997); Remand Results 18-19. Commerce considers data to be aberrational when it is an “extreme outlier,” Remand Results 11; id. at 23-24, 31-32 (citing prior agency practice), is distorted or misrepresentative, or is “somehow incorrect.” Id. at 30. Commerce has not affirmatively set forth a definition of reliable here. It implies that aberrational data, data shown to deviate from the norm, would be unreliable. See Remand Results 23 n.90, citing Polyethylene Terephthalate Film, Sheet, and Strip From the People’s Republic of China, 80 Fed. Reg. 33,241 (Dep’t Commerce Jun. 11, 2015) (final results of ADD administrative review and final determination of no shipments; 2012-2013). Nonetheless, Commerce has explained the need for reliable surrogate values by emphasizing that it is precisely “the unreliability of NME prices that drives [the agency] to use the special NME methodology in the first place.” See Antidumping Duties; Countervailing Duties, 62 Fed. Reg. at 27,366. Surrogate values serve as substitutes for what Commerce considers to be unreliable NME data.

One of the FOPs that Commerce values is labor, he., the wage rate for the industry in which the subject merchandise is produced. See 19 U.S.C. § 1677b(c)(3). Commerce’s practice is to value labor using industry-specific data from the primary surrogate country, as published in Chapter 6A of the ILO Yearbook of Labor Statistics. See Labor Methodologies, 76 Fed. Reg. at 36,093; see Remand Results 3. In developing its current practice, 11 Commerce had initially used earnings or wages reported in the ILO Yearbook of Labor Statistics “Chapter 5B: Wages and Manufacturing.” See Antidumping Methodologies in Proceedings Involving Non-Market Economies: Valuing the Factor of Production: Labor, 76 Fed. Reg. 9,544, 9,544-45 (Dep’t Commerce Feb. 18, 2011). Ultimately, Commerce chose ILO Chapter 6A data over ILO -Chapter 5B data for its preferred practice due to “a concern with under-counting” using Chapter 5B data, and what the agency called a “rebuttable presumption that Chapter 6A accounts for all direct and indirect costs.” Labor Methodologies, 76 Fed. Reg. at 36,094.

In this review, Commerce did not choose labor data from ILO Chapter 6A for the primary surrogate country, Labor Methodologies, 76 Fed. Reg. at 36,093, because ILO Chapter 6A data was not available for Bangladesh. Remand Results 4-5. Instead, it chose BBS data because the data was specific, closer in time to the period of review than the other data, 12 and from the primary surrogate country. Id at 8.

On remand, Commerce fails to address record evidence of widespread labor abuses in the Bangladeshi shrimp industry that undermines Commerce’s implicit findings that the BBS data is non-aberrational, reliable, and thus the best information available. Commerce’s practice is to not use aberrational or unreliable data. See Antidumping Duties; Countervailing Duties, 62 Fed. Reg. at 27,366; Remand Results 34. Plaintiff submitted record evidence alleging forced labor, child labor, and systemic labor abuses specifically within the Bangladeshi shrimp industry. See Ad Hoc Shrimp Surrogate Country Comments at Exs. 1-5; Ad Hoc Shrimp Post-Prelim. Comments at Attachs. 1-7. Although Commerce acknowledges this evidence, it finds it is insufficient to demonstrate aberration or unreliability absent a quantitative analysis that (i) compares the BBS wage rate to benchmark data, or (ii) demonstrates how the alleged labor abuses impact the BBS rate. Remand Results 11-12, 16.

' [8,9] Commerce’s reasons for dismissing the record evidence put forth by the Plaintiff without further investigation and analysis are inadequate given the substantial evidence standard. Commerce’s ultimate determination must be supported by substantial evidence and “[t]he substantiality of evidence must take into account whatever in the record fairly detracts from its weight.” Universal Camera Corp. v. NLRB, 340 U.S. 474, 488, 71 S.Ct. 456, 95 L.Ed. 456 (1951); Tudor v. Dep’t of Treasury, 639 F.3d 1362, 1366 (Fed. Cir. 2011) (same). The substantial evidence standard of review essentially asks whether, given the evidence on the record as a whole, the agency’s conclusion was reasonable. Nippon Steel Corp. v. United States, 458 F.3d 1345, 1351 (Fed. Cir. 2006). Plaintiffs evidence leads to the reasonable inference that the data is aberrational and not reliable and therefore fairly detracts from Commerce’s finding that the BBS data is the best information available. Therefore, Commerce must explain why the BBS data is nonetheless a reasonable choice for the best available information.

First, Commerce explains that, to determine whether a surrogate value is aberrational, it conducts, “a quantitative, meas-ureable analysis,” Remand Results 12, and that, in “its well established practice” for determining aberration, the agency “compares a dataset to a meaningful benchmark to determine whether a particular dataset is an outlier.” 13 Id. at 31-32. Plaintiff placed on the record evidence of the ILO Chapter 6A labor data from Guyana, India, Nicaragua, and Philippines, and Chapter 5B labor data from Indonesia, with which to compare the BBS data. 14 See Ad Hoc Shrimp Post-Prelim Comments at Attach. 8. Still, Commerce claims it has no specific data to serve as a benchmark with which to compare BBS data. Although ILO Chapter 6A data may not be as specific as the BBS data, it is Commerce’s preferred data source. See Labor Methodologies, 76 Fed. Reg. at 36,093. It is unclear to the court why this data cannot serve as a benchmark. 15 Further, Commerce suggests that the labor input is not well suited to a cross-country comparison to determine if the data is aberrational, noting the “many socio-economic, political and institutional factors, such as labor laws and policies unrelated to the size or strength of an economy, that cause significant variances in wage levels between countries.” Remand Results 13 (also noting “the variability in labor rates that exists among otherwise economically comparable countries is a characteristic unique to the labor input”), quoting Issues and Decision Memorandum for the Final Determination in the Anti-dumping Duty Investigation of Crystalline Silicon Photovoltaic Cells, Whether or Not Assembled into Modules, from the People’s Republic of China, A-570-979, 23 (Oct. 9, 2012), available at http://ia.ita.doc.gov/frn/ summary/prc/2012-25580-l.pdf (last visited Jun. 8, 2017). Thus, it appears that Commerce requires parties to demonstrate aberration in labor data by comparing historic wage rates within the same country. It is unclear to the court how aberration can be assessed where there is a claim of systematic labor abuses within a country and parties may only use historical wage rate data from that country.

Second, Commerce suggests that Plaintiff must not only compare the BBS data to benchmarks but also indicate how the labor abuses affected the wage rates. 16 Remand Results 39. It is unclear to the court what exactly Commerce wishes to see or how a party should endeavor to supply this information. Without more direction from Commerce, such a broad demand is unreasonable given the record evidence in this case. 17 It is Commerce’s practice not to use unreliable data. Although here Commerce does not define “reliable” as it applies to labor data, the Oxford dictionary defines “reliable” as “consistently good in quality; able to be trusted.” Reliable, Oxford Online Dictionary, https://en.oxford dictionaries.com/definition/reliable (last visited June 8,2017). The evidence of widespread labor abuses, including forced labor and child labor, in the specific industry under consideration includes evidence that workers are either not paid, or are not fully paid, what is owed to them. See, e.g., Ad Hoc Shrimp Surrogate Country Comments at Attach. 3; Ad Hoc Shrimp Agency Case Br. at 10-13, citing Research on Indicators of Forced Labor in the Supply Chain of Shrimp in Bangladesh, Verité (2012), available at http://www.verite.org/ wp-eontent/uploads/2016/ll/Research-on-Indicators-of-Forced-Labor-in-the-Bangladesh-Shrimp-Sector_9.16.pdf (last visited Jun. 8, 2017) (“Research on Indicators of Forced Labor”). Therefore, it is a fair inference that the wage pates of workers subject to systematic labor abuses cannot be trusted. That inference detracts from Commerce’s determination and therefore Commerce must address it. See Universal Camera Corp., 340 U.S. at 488, 71 S.Ct. 456 (“The substantiality of evidence must take into account whatever in the record fairly detracts from. its weight.”). Commerce needs to explain why this inference is incorrect or nonetheless should be disregarded. The absence of a quantitative analysis demonstrating how labor abuses affected wage rates does nothing to undermine this inference given this record.

Equally problematic in Commerce’s response to Plaintiffs record evidence in this case is its failure to ai'ticulate a standard with respect to labor data. Despite stating its preference to avoid aberrational or unreliable data, see Antidumping Duties; Countervailing Duties, 62 Fed. Reg. at 27, 366, Commerce adequately defines neither “aberrational” nor “unreliable” as these terms relate to labor data. Throughout the Remand Results, Commerce alludes to aberrational data as data which is an extreme outlier. Remand Results 11 (“The record must contain specific evidence showing the value is aberrational, e.g., an extreme outlier”); see also, e.g., id. at 23, 24, 32, 34, 41. Commerce provides no other example of what would constitute aberrational labor data and it does not explain a standard for identifying when labor data is an “extreme outlier.” Although Plaintiff points to record data that shows the Bangladeshi data is the lowest value data, see Ad Hoc Shrimp Agency Case Br. at 7; Ad Hoc Shrimp Post-Prelim. Comments at Attach. 8, Commerce offers neither the court nor the parties any insight into how far afield labor values must be to qualify as aberrational. It may be that the BBS data is not aberrational; it is not for the court to say. It is for Commerce to say what it considers aberrational and how a party can demonstrate that labor data is aberrational. The court can then assess whether that standard is reasonable and whethér the record evidence supports that determination. Although Commerce refers to the need to perform a quantitative analysis, it is not clear what that entails in this case or when such analysis would lead Commerce to find labor data aberrational or unreliable.

Further, here, Commerce does not adequately explain what it means by “reliable” with respect to labor data, although it implies that aberrational data would be unreliable. See Remand Results 23 n.90. Even when Commerce recounts what it has found to be “not unreliable,” the agency offers no insight into how a standard might be applied in this case. See Remand Results 16 (discussing Steel Wire Garment Hangers From the People’s Republic of China, 80 Fed. Reg. 13,332 (Dep’t Commerce Mar. 13, 2015) (final results of ADD administrative review, 2012-2013), and recounting that “we determined that ‘USTR reports do not make Thai import data unreliable or inferior to Philippine data’ ”), 16-17 (discussing Certain Steel Threaded Rod From the People’s Republic of China, 79 Fed. Reg. 71,743 (Dep’t Commerce Dec. 3, 2014) (final results of ADD administrative review; 2012-2013), and noting that, in response to a claim of data unreliability based upon political upheaval in Thailand, the “interested party ‘provided no specific record evidence showing how this event had any specific distortive impact on the Thai import data in general’ ”). Given the record evidence of systemic labor abuses in this case, the court fails to see how BBS data can be reliable given the court’s understanding of “reliable.” Nonetheless the court must await Commerce’s explanation of what constitutes reliable labor data to assess whether its definition is reasonable and whether record evidence supports Commerce’s determination given that definition.

Commerce also does not explain why, given this record, it finds the data to be an appropriate proxy for Vietnam, reflective of a hypothetical Vietnamese market economy. Surrogate values are used as proxies with which to create a hypothetical market economy price. See Nation Ford v. United States, 166 F.3d at 1378. Evidence of systemic abuse, including forced and child labor, specific to the Bangladeshi shrimp industry detracts from the representativeness of the BBS data and therefore undermines the market-based approach that Congress sought by requiring that Commerce use surrogate values when calculating normal value for subject merchandise from NME countries. 18 See Antidumping Duties; Countervailing Duties, 62 Fed. Reg. at 27,366.

Commerce’s reticence to accept Ad Hoc Shrimp’s arguments seems to stem from its “strong preference” to select data from the primary surrogate country. 19 See Remand Results 10. Commerce states that the BBS data is the only available wage rate data from Bangladesh, Id. at 8-10 (noting that Labor Methodologies indicates that the best methodology is to use “industry specific data from the primary surrogate country,” and that using the BBS data “results in the use of a uniform basis for FOP valuation—the use of data from a primary surrogate country.”)• Understandably, the Department prefers to use surrogate value data from the primary surrogate country to minimize distortion. Id. at 24 (data from the primary surrogate country “provides] the best interplay with other factors of production that are measured based on data from the same country.”). Even assuming that Commerce’s preference is reasonable as a general rule, Commerce cannot, without further analysis, justify the use of unreliable data. 20 Indeed Commerce’s own practice suggests that Commerce will depart from its normal practice of valuing FOPs from a primary surrogate country where that data is aberrational. See Certain Cased Pencils from the People’s Republic of China, 59 Fed. Reg. 55,625, 55,633 (Dep’t Commerce Nov. 8, 1994) (notice of final determination of sales at less than fair value). 21 With record evidence suggesting that factors within the Bangladeshi shrimp industry affect the reliability of the data and, therefore, its ability to reflect a hypothetical Vietnamese market economy, Commerce cannot continue to rely on the explanation that the data comes from the primary surrogate country.

Defendant contends that Commerce cannot determine whether the BBS data is aberrational or unreliable because the Plaintiff has failed in its burden to make an adequate record so that Commerce can perform the quantitative analysis it wishes to perform. Def.’s Remand Comments 16, citing QVD Food Co. v. United States, 658 F.3d 1318, 1324 (Fed. Cir. 2011); see also Nan Ya Plastics Corp. v. United States, 810 F.3d 1333, 1337-38 (Fed. Cir. 2016). Although the burden lies with the parties to populate an adequate record, see QVD Food Co., 658 F.3d at 1324, here Plaintiff has provided the record evidence that undermines the agency’s determination and that must therefore be explained. Universal Camera Corp., 340 U.S. at 488, 71 S.Ct. 456. Commerce’s argument that the Plaintiff has not populated the record with necessary data suggests that Commerce rejects the inference that systemic labor abuses (including forced and child labor) detract from the reliability of BBS data as representative of Bangladeshi labor costs and/or labor costs in a hypothetical market economy which is economically comparable to Vietnam. Commerce therefore presumes the BBS data is reliable despite the evidence of systemic labor abuse. Such a presumption is not reasonable. The systemic labor abuses cited in the record by the Plaintiff at the very least, involve workers either not being paid for all of their labor or being underpaid for their labor. See, e.g., Ad Hoc Shrimp Post-Prelim. Comments at Attachs. 1-7. If workers are, as a result of forced or child labor practices, not compensated or not fully compensated for their work, then one cannot presume that the reported wage data represents either the actual labor costs in Bangladesh or the labor costs in a hypothetical market economy which is economically comparable to Vietnam.

Commerce contends that is not required “to conduct an analysis of working conditions when determining whether a particular set of data, like the BBS, is aberrational.” Remand Results 12. According to Commerce, socio-political factors are inap-posite to its determination under 19 U.S.C. § 1677b(c)(l)(B) and beyond its realm of expertise, as its standard practice for determining aberration is to conduct a “quantitative, measurable analysis,” either of comparable data across economically comparable countries or of historic wage rate data within the country at issue (Bangladesh). Id. at 12-13. However, Commerce is not being asked to analyze the working conditions of Bangladesh; Commerce is being asked to follow its practice, which instructs the agency not to rely on aberrational or unreliable surrogate value data. See Antidumping Duties; Countervailing Duties, 62 Fed. Reg. at 27,366.

Commerce also contends that the record evidence submitted by Ad Hoc Shrimp is not measurable, “specific evidence” that “the BBS data itself, as a dataset, is distorted.” Remand Results 18, citing Camau Frozen Seafood Processing Import Export Corp. v. United States, 35 CIT -, -, 929 F.Supp.2d 1352, 1356 (2013) (finding no evidence of aberration where the record demonstrated only that the Bangladeshi data was “the lowest price in a range of prices.”). The evidence submitted by Ad Hoc Shrimp details workers in the Bangladeshi shrimp industry who are not paid or are underpaid for their work. See, e.g., Ad Hoc Shrimp Surrogate Country Comments at Attach. 3; Ad Hoc Shrimp Agency Case Br. at 10-13, citing Research on Indicators of Forced Labor. This evidence makes the wage rates unreliable, as they cannot be trusted to fully reflect compensation for the labor under market principles.

On second remand, Commerce must either reconsider its choice of the BBS data to value labor or it must address the record evidence of widespread labor abuses in the Bangladeshi shrimp industry that undermines Commerce’s implicit findings that the BBS data is non-aberrational, reliable, and thus the best information available. Further, Commerce must articulate a reasonable method by which a petitioner can demonstrate aberration or unreliability where, as here, there is a claim of widespread, systemic labor abuse.

CONCLUSION

In accordance with the foregoing, it is hereby

ORDERED that Commerce’s remand determination is remanded for further consideration consistent with this opinion. Specifically, upon remand, Commerce must:

(1) articulate a reasonable method by which a petitioner can demonstrate aberration or unreliability where, as here, there is a claim of widespread, systemic labor abuse; and

(2) address the record evidence of widespread labor abuses in the Bangladeshi shrimp industry that undermines Commerce’s, implicit findings that the BBS data is non-aberrational, reliable, and thus the best information available; and

(3) explain why the Bangladeshi wage rate data is reliable and not aberrational, in light of the record evidence of systemic labor abuses; or

(4) if the data is aberrational and unreliable, explain why it is nonetheless the best available information; or

(5) reconsider its determination that the Bangladeshi data.is the best available information; and it is fur-; ther

ORDERED that Commerce shall file its second remand determination with the court within 45 days of this date; and it is further

ORDERED that Plaintiff shall have 30 days thereafter to file comments on the second remand determination; and it is further

ORDERED that Defendant shall have 15 days thereafter to file a reply to comments on the second remand determination.

1 . On December 8, 2014, Defendant submitted indices to the public and confidential administrative records, which identify the documents that comprise the public and confidential administrative records to Commerce’s final determination. See Index to Administrative Record, Dec. 8, 2014, ECF No. 27-1. All further references to the documents from the administrative record are identified by the numbers assigned by Commerce in these administrative records,

2 .Commerce stated in the Prelim. Decision Memo that its current preferred practice is to use ILO Chapter 5B data to value labor wage rates, Prelim. Decision Memo at 23; however, Commerce's current practice is to use ILO Chapter 6A data to value labor wage rates. See Prelim. Surrogate Value Memo at 5-6; Antidumping Methodologies in Proceedings Involving Non Market Economies: Valuing the Factor of Production: Labor, 76 Fed. Reg. 36,092, 36,093 (Dep’t Commerce Jun. 21, 2011). Commerce stated this practice correctly in its Prelim. Surrogate Value Memo and the Final Decision Memo in this review. See Prelim. Surrogate Value Memo at 5-6; Issues and Decision Memorandum for the Final Resuits, A-552-802, 47, (Sept. 19, 2014), ECF No. 27-4.

3 . This submission included ILO Chapter 6A data for India, Guyana, Philippines, and Nicaragua, and ILO Chapter 5B data for Indonesia. See Ad Hoc Shrimp Post-Prelim. Comments at Attach. 8.

4 . In its agency case brief Ad Hoc Shrimp argued that due to the evidence it submitted to Commerce documenting labor abuses, Commerce should value the labor FOP in this review “either by using wage data from a secondary surrogate country or through averaging wage data from multiple countries.” Ad Hoc Shrimp Agency Case Br, at 2, 24.

5 . The challenges raised by Plaintiffs Tri Union Frozen Products, Inc., Mazzetta Company LLC, Ore-Cal Corporation, Consolidated Plaintiff Quoc Viet Seaproducts Processing Trading and Import-Export Co., and Consolidated Plaintiffs Vietnam Association of Seafood Exporters and Producers (including certain of its individual member companies) to Commerce’s final determination were rejected in Tri Union I; the court sustained Commerce’s final determination with regard to those issues, See Tri Union I, 40 CIT at -, 163 F.Supp.3d at 1256, 1267-1312, 1313.

6 . Further citations to the Tariff Act of 1930, as amended, are to the relevant provisions of Title 19 of the U.S. Code, 2012 edition.

7 . A NME country is "any foreign country that .,. does not operate on market principles of cost or pricing structures, so that sales of merchandise in such country do not reflect the fair value of the merchandise.” 19 U.S.C. § 1677(18)(A). Thus, Commerce must “determine the normal value of the subject merchandise on the basis of the value of the factors of production utilized in producing the merchandise ... [together with other costs and expenses].” 19 U.S.C. § 1677b(c)(1).

8 . For all FOPs, Commerce seeks the best available information due to its statutory directive pursuant to 19 U.S.C. § 1677b(c)(1) and as part of its mandate to determine dumping margins as accurately as possible. Rhone Poulenc, Inc. v. United States, 899 F.2d 1185, 1191 (Fed. Cir. 1990); Shakeproof Assembly Components, Div. of Illinois Tool Works, Inc. v. United States, 268 F.3d 1376, 1382 (Fed. Cir. 2001) (noting that "the purpose of [19 U.S.C. § 1677b(c)] is to determine antidumping margins as accurately as possible,”); see also Zhejiang DunAn Hetian Metal Co., Ltd. v. United States, 652 F.3d 1333, 1341 (Fed. Cir. 2011)). The Court of Appeals has emphasized that the Court's “duty is 'not to evaluate whether the information Commerce used was the best available, but rather whether a reasonable mind could conclude that Commerce chose the best available information.’ ” Zhejiang DunAn Hetian Metal Co., Ltd. v. United States, 652 F.3d 1333, 1341 (Fed. Cir. 2011) (quoting Goldlink Indus. Co. v. United States, 30 CIT 616, 619, 431 F.Supp.2d 1323, 1327 (2006)).

9 . Further citations to Title 19 of the Code of Federal Regulations are to the 2013 edition.

10 . To determine what constitutes the best available information, Commerce evaluates the quality of data sources from the countries offered to value respondents’ FOPs favoring data that is; (1) specific to the input in question; (2) representative of a broad market average of prices; (3) net of taxes and import duties; (4) contemporaneous with the period of review; and (5) publicly available. See generally Import Admin., U.S. Dep't Commerce, Non-Market Economy Surrogate Country Selection Process, Policy Bulletin 04.1 (2004), available at http://ia.ita.doc.gov/policy/bull 041.html (last visited June 8, 2017); see also Qingdao Sea-Line Trading Co. v. United States, 766 F.3d 1378, 1386 (Fed. Cir. 2014).

11 . Commerce’s approach to valuing labor as an FOP has recently changed. Prior to 2011, Commerce based labor surrogate value on a regression methodology, averaging wage rate data collected from multiple countries, in order to minimize the variability that exists in wages across countries, including even economically comparable countries. See Labor Methodologies, 76 Fed. Reg. at 36,093. However, the Court of Appeals held that the wage rate regression methodology was inconsistent with 19 U.S.C. § 1677b(c)(4), and invalidated 19 C.F.R. § 351.408(c)(3), Commerce’s regulation codifying the regression-based methodology. Dorbest Ltd. v. United States, 604 F.3d 1363, 1372-73 (Fed. Cir. 2010). Commerce subsequently determined that the countries whose data it could average in any given case going forward would be so limited that "there would be little, if any, benefit to relying on an average of wages from multiple countries.” Labor Methodologies, 76 Fed. Reg. at 36,094. Accordingly, while continuing to acknowledge the variability in wage rates across economically comparable countries, id. at 36,093, the agency determined that it would thereafter use ILO Chapter 6A wage rate data from the primary surrogate country to value labor, which would have the benefit of "a uniform basis for FOP valuation—a single surrogate,” Id.

12 . As Commerce explains, the BBS data is not contemporaneous; however, “the BBS wage data is from 2010, which, while not contemporaneous with the [period of review], is closer to the [period of review] than the labor surrogate values proffered by Petitioner, which range from 2004 through 2008.” Remand Results 8-10.

13 . More specifically, the agency compares "the prices for an input from all countries found to be at a level of economic development comparable to the NME whose products are under review from the [period of review] and prior years,” or "examines data from the same HTS number for the surrogate country whose data are allegedly aberrational over multiple years to determine if the current data appear aberrational compared to historical values.” Remand Results 12 n.55.

14 . Commerce claims in its remand results the BBS data is the best information available because it is more specific than other available data. Remand Results 32. Yet at the same time it also claims that it has no specific data to serve as a benchmark with which to compare the BBS data. IcL at 10, 32. These seemingly conflicting statements stem from the fact that, under category 15 of ILO Chapter 6A data, rates would be determined using a category covering “Manufacture of Food Products and Beverages,” while the BBS data covers only the shrimp industry. However, Commerce’s own practice prefers ILO Chapter 6A data. Labor Methodologies, 76 Fed. Reg. at 36,093. The BBS data pertains to the shrimp industry more specifically. Commerce implies that, had ILO Chapter 6A data for Bangladesh been available, it would have used that data despite the fact that ILO data covers industries beyond the shrimp industry. Remand Results 4, quoting Final Decision Memo at 47 ("Bangladesh does not report labor data to the ILO ... [films, we are unable to use ILO's Chapter 6A data or wage data reported under ILO’s Chapter 5B, as is the preference.”). Commerce rejects Guyana and Indonesia as a suitable primary surrogate country as neither country was identified as a significant producer of comparable merchandise. Id. at 7. Commerce rejects India, Nicaragua, and Philippines as a suitable primary surrogate country either because the record lacked data for whole shrimp (India) or lacked financial statements (Nicaragua and Philippines). Yet, it does not explain why the data from these countries cannot be used as a benchmark for assessing the reliability of the BBS data. Id That those countries may not have been suitable surrogate countries does not explain why the ILO Chapter 6A data on the record for those countries cannot be used as benchmarks to evaluate the claim that the BBS data is aberrational.

15 . Commerce does explain that Guyana and Indonesia are not significant producers of comparable merchandise and therefore ILO data for those countries was not useable. Remand Results 7. However, Commerce still had ILO Chapter 6A data from India, Nicaragua, and Philippines. See Ad Hoc Shrimp Post-Prelim. Comments at Attach. 8.

16 . Commerce states that "Petitioner has provided no measurable correlation between the evidence it placed on the record and the BBS wage data itself, but instead has merely speculated a cause and effect with no support for its allegations.” Remand Results 39.

17 . There may be reasons to use a quantitative analysis in other cases. However, in the present case, it is not reasonable to require "measurable” evidence that the dataset is unreliable. The evidence of systemic labor abuses, including forced and child labor, indicates that the data cannot be trusted because it suggests that, at the very least, workers are not paid or are not fully paid for their labor.

18 . Commerce contends that considering labor practices in determining aberration is problematic because "labor practices may potentially influence” the wage rate in each country which does not "necessarily” make the wage rate aberrational. Remand Results 13. Commerce’s statement mischaracterizes Petitioner’s argument, which is not that labor conditions necessarily make wage rates aberrational, but rather that the labor conditions specific to Bangladesh are unique and influence the Bangladeshi wage rates and, thus, render the data unreflective of a hypothetical Vietnamese market economy. See id.; see generally Ad Hoc Shrimp Remand Comments.

19 . Commerce reiterated the reasons the agency did not select the other countries for which Ad Hoc Shrimp provided surrogate value data as the primary surrogate country. Remand Results 7-8. Ad Hoc Shrimp placed ILO Chapter 6A data on the record for India, Guyana, Philippines, Nicaragua, and Indonesia on April 24, 2014, following publication of the preliminary determination in March 2014. See Ad Hoc Shrimp Post-Prelim. Comments at Attach. 8. Commerce did not select Guyana or Indonesia as the primary surrogate country because it found that those countries were neither significant producers of comparable merchandise nor on the surrogate country list. Remand Results 7. Commerce did not select Nicaragua, Pakistan, or Philippines as the primary surrogate country due to lack of available data, stating that no interested party placed data on the record that would require it to reconsider its disqualification of these countries in the preliminary stages from the primary surrogate country selection. Id. Commerce "did not select India as the primary surrogate country because the record did not contain Indian source for whole shrimp surrogate values." Id. at 7-8.

20 . Although there is a regulatory preference for valuing all FOPs in a single surrogate country, tire statute specifically allows for valuing FOPs with surrogate values from more than one market economy country. 19 U.S.C. § 1677b(c)(l) (instructing that surrogate values are to be taken from "a market economy country or countries.”). In noting that the statute allows, but does not require, Commerce to use the same source for all factors of production, the Court of Appeals emphasized that the statute "merely requires the use of the 'best available information’ with respect to the valuation of a given factor of production.” Nation Ford v. United States, 166 F.3d at 1378.

21 . In the Remand Results, Commerce responded to Plaintiff's citation to Certain Cased Pencils from the People’s Republic of China, 59 Fed. Reg. 55,625, by stating that it was an example of "another case where the Department [selected an alternative surrogate value from a country other than the primary surrogate country to avoid using aberrant surrogate values] based on a quantitative comparison of the surrogate value data from the alternative surrogate country.” Remand Results 36. However, the cited Certain Cased Pencils determination does not detail the "quantitative comparison of the surrogate value data from the alternative surrogate country” conducted by Commerce, stating only that Commerce "analyzed the Indian factor values for erasers, ferrules, paint, animal glue, and plastic foil. We compared these factor values with Pakistani and U.S. values based on U.S. costs taken from the petition and found the Indian factor value for erasers, ferrules and paint to be aberrational." Certain Cased Pencils from the People's Republic of China, 59 Fed. Reg. at 55,633.