Solarworld Americas, Inc. v. United States

Case: 19-1591 Document: 104 Page: 1 Filed: 06/24/2020

United States Court of Appeals for the Federal Circuit ______________________

SOLARWORLD AMERICAS, INC., Plaintiff-Cross-Appellant v.

UNITED STATES, Defendant-Appellee

DEPARTMENT OF COMMERCE, Defendant

v.

TRINA SOLAR (U.S.) INC., Defendant

YINGLI GREEN ENERGY HOLDING COMPANY LIMITED, YINGLI GREEN ENERGY AMERICAS, INC., YINGLI ENERGY (CHINA) CO., LTD., BAODING TIANWEI YINGLI NEW ENERGY RESOURCES CO., LTD., BEIJING TIANNENG YINGLI NEW ENERGY RESOURCES CO., LTD., TIANJIN YINGLI NEW ENERGY RESOURCES CO., LTD., HENGSHUI YINGLI NEW ENERGY RESOURCES CO., LTD., LIXIAN YINGLI NEW ENERGY RESOURCES CO., LTD., BAODING JIASHENG PHOTOVOLTAIC TECHNOLOGY CO., LTD., HAINAN YINGLI NEW ENERGY RESOURCES CO., LTD., SHENZHEN YINGLI NEW ENERGY RESOURCES CO., LTD., CANADIAN SOLAR, INC., CANADIAN SOLAR (USA), INC., CANADIAN SOLAR MANUFACTURING (CHANGSHU), INC., CANADIAN SOLAR MANUFACTURING

Case: 19-1591 Document: 104 Page: 2 Filed: 06/24/2020 2 SOLARWORLD AMERICAS, INC. v. UNITED STATES

(LUOYANG), INC., CANADIAN SOLAR INTERNATIONAL LIMITED, BYD (SHANGLUO) INDUSTRIAL CO., LTD., SHANGHAI BYD CO., LTD., Defendants-Appellees

CHANGZHOU TRINA SOLAR ENERGY CO., LTD., TRINA SOLAR (CHANGZHOU) SCIENCE & TECHNOLOGY CO., LTD., YANCHENG TRINA SOLAR ENERGY TECHNOLOGY CO., LTD., CHANGZHOU TRINA SOLAR YABANG ENERGY CO., LTD., TURPAN TRINA SOLAR ENERGY CO., LTD., HUBEI TRINA SOLAR ENERGY CO., LTD., Defendants-Appellants ______________________

2019-1591, 2019-1593 ______________________

Appeals from the United States Court of International Trade in Nos. 1:16-cv-00132-CRK, 1:16-cv-00134-CRK, 1:16-cv-00135-CRK, Judge Claire R. Kelly. ______________________

Decided: June 24, 2020 ______________________

TIMOTHY C. BRIGHTBILL, Wiley Rein, LLP, Washington, DC, argued for plaintiff-cross-appellant. Also represented by TESSA V. CAPELOTO, DOUGLAS C. DREIER, LAURA EL- SABAAWI, USHA NEELAKANTAN, STEPHEN JOSEPH OBERMEIER, JOHN ALLEN RIGGINS, ADAM MILAN TESLIK, MAUREEN E. THORSON, ENBAR TOLEDANO.

NEIL R. ELLIS, Sidley Austin LLP, Washington, DC, argued for defendants-appellees Yingli Green Energy Holding Company Limited, Yingli Green Energy Americas, Inc., Yingli Energy (China) Co., Ltd., Baoding Tianwei

Case: 19-1591 Document: 104 Page: 3 Filed: 06/24/2020 SOLARWORLD AMERICAS, INC. v. UNITED STATES 3

Yingli New Energy Resources Co., Ltd., Beijing Tianneng Yingli New Energy Resources Co., Ltd., Tianjin Yingli New Energy Resources Co., Ltd., Hengshui Yingli New Energy Resources Co., Ltd., Lixian Yingli New Energy Resources Co., Ltd., Baoding Jiasheng Photovoltaic Technology Co., Ltd., Hainan Yingli New Energy Resources Co., Ltd., Shenzhen Yingli New Energy Resources Co., Ltd., Canadian Solar, Inc., Canadian Solar (USA), Inc., Canadian Solar Manufacturing (Changshu), Inc., Canadian Solar Manufacturing (Luoyang), Inc., Canadian Solar International Limited. Also represented by SHAWN MICHAEL HIGGINS.

CRAIG A. LEWIS, Hogan Lovells US LLP, Washington, DC, for defendants-appellees BYD (Shangluo) Industrial Co., Ltd., Shanghai BYD Co., Ltd.

TARA K. HOGAN, Commercial Litigation Branch, Civil Division, United States Department of Justice, Washington, DC, argued for defendant-appellee United States. Also represented by JOSEPH H. HUNT, REGINALD THOMAS BLADES, JR., JEANNE DAVIDSON; BRENDAN SASLOW, MERCEDES MORNO, Office of the Chief Counsel for Trade Enforcement & Compliance, United States Department of Commerce, Washington, DC.

JONATHAN FREED, Trade Pacific PLLC, Washington, DC, argued for defendants-appellants. Also represented by ROBERT GOSSELINK. ______________________

Before PROST, Chief Judge, DYK and O’MALLEY, Circuit Judges. DYK, Circuit Judge. Defendants Changzhou Trina Solar Energy Co., Ltd. et al. (“Trina”) appeal decisions of the United States Court of International Trade (“CIT”) regarding the first

Case: 19-1591 Document: 104 Page: 4 Filed: 06/24/2020 4 SOLARWORLD AMERICAS, INC. v. UNITED STATES

administrative review of an antidumping duty order on crystalline silicon photovoltaic cells from the People’s Republic of China (“PRC” or “China”). Plaintiff SolarWorld Americas, Inc. (“SolarWorld”) cross-appeals. We affirm in part, vacate in part, and remand. BACKGROUND “Dumping occurs when a foreign firm sells a product in the United States at a price lower than the product’s normal value.” Home Prod. Int’l, Inc. v. United States, 633 F.3d 1369, 1372 (Fed. Cir. 2011). The Tariff Act of 1930, codified at 19 U.S.C. § 1202 et seq., authorizes the government to impose on dumped products “an antidumping duty . . . in an amount equal to the amount by which the normal value exceeds the export price” of the products. 19 U.S.C. § 1673. “For exporters based in market economy . . . countries, [the normal value] is generally the price at which the firm sells the product in its home market.” Home Prod., 633 F.3d at 1372 (citing 19 U.S.C. § 1677b(a)(1)(B)(i)). Where the exporter is located in a non-market economy, “the default rule is that [the normal value] is calculated based on a factors-of- production analysis whereby each input is valued based on data from a surrogate [market economy] country.” Id. (citing 19 U.S.C. § 1677b(a)(1)(B)(ii)). The government is required to separately determine a weighted average dumping margin for “each known exporter and producer,” unless “not practicable.” See 19 U.S.C. § 1677f-1(c). On December 7, 2012, the Department of Commerce (“Commerce”) issued an antidumping duty order on crystalline silicon photovoltaic cells from China. On February 4, 2015, Commerce initiated the first administrative review of this antidumping duty order, covering the period December 1, 2013, through November 30, 2014 (“Period of Review”). Included as mandatory respondents in this review were Trina, Yingli Green

Case: 19-1591 Document: 104 Page: 5 Filed: 06/24/2020 SOLARWORLD AMERICAS, INC. v. UNITED STATES 5

Energy Holding Company Limited et al. (“Yingli”), and BYD (Shangluo) Industrial Co., Ltd. et al. (“BYD”). Commerce published its final determination (“Final Results”) on June 13, 2016. SolarWorld, Trina, Yingli, and BYD brought suit against the government in the CIT, each challenging aspects of Commerce’s Final Results under 19 U.S.C § 1516a(a)(2). SolarWorld, a domestic producer, argued that the antidumping duty rates were too low. Trina, Yingli, and BYD, foreign producers, argued that their antidumping duty rate was too high. After remands on October 18, 2017, and May 18, 2018, the CIT sustained Commerce’s determinations on December 13, 2018. Commerce calculated a dumping margin of 6.55% for Trina, 0% for Yingli, and 8.52% for BYD. SolarWorld, Trina, and BYD appeal. We describe the particular challenges to the antidumping determinations and the CIT’s rulings below. We have jurisdiction under 28 U.S.C. § 1295(a)(5). DISCUSSION We review the CIT’s decision to sustain Commerce’s final results and its remand redeterminations de novo. See U.S. Steel Corp. v. United States, 621 F.3d 1351, 1357 (Fed. Cir. 2010). We will affirm Commerce unless its decision is “unsupported by substantial evidence on the record, or otherwise not in accordance with law.” 19 U.S.C. § 1516a(b)(1)(B)(i). I We first address Trina’s argument that Commerce overstated its dumping duty by using Thai import data to value Trina’s nitrogen input. Where an exporter is from a non-market economy (here, China), 19 U.S.C. § 1677b(c)(1) directs Commerce to

Case: 19-1591 Document: 104 Page: 6 Filed: 06/24/2020 6 SOLARWORLD AMERICAS, INC. v. UNITED STATES

determine the “normal price” of merchandise subject to an antidumping duty “on the basis of the value of the factors of production utilized in producing the merchandise.” “The evaluation of the factors of production shall be based on the best available information regarding the values of such factors in a market economy country or countries considered to be appropriate by [Commerce].” Id. (emphasis added). This statutory directive reflects “the intent of Congress . . . that Commerce should avoid the use of distorted surrogate prices.” Nation Ford Chem. Co. v. United States, 166 F.3d 1373, 1378 (Fed. Cir. 1999). Section 1677b is implemented by 19 C.F.R. § 351.408, which provides, in relevant part, that “[Commerce] normally will use publicly available information to value factors,” and that Commerce “normally will value all factors in a single surrogate country.” Commerce has a practice of “resort[ing] to a secondary surrogate country . . . if data from the primary surrogate country [is] unavailable or unreliable.” J.A. 746. “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 the antidumping margins as accurately as possible.’” Zhejiang DunAn Hetian Metal Co. v. United States, 652 F.3d 1333, 1341 (Fed. Cir. 2011) (quoting Shakeproof Assembly Components v. United States, 268 F.3d 1376, 1382 (Fed. Cir. 2001)). Thus, a “surrogate value must be as representative of the situation in the [non-market economy] country as is feasible.” SeAH Steel VINA Corp. v. United States, 950 F.3d 833, 845 (Fed. Cir. 2020) (quoting Nation Ford, 166 F.3d at 1377). “This 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, 652 F.3d at 1341 (quoting Goldlink Indus. Co. v.

Case: 19-1591 Document: 104 Page: 7 Filed: 06/24/2020 SOLARWORLD AMERICAS, INC. v. UNITED STATES 7

United States, 431 F. Supp. 2d 1323, 1327 (Ct. Int’l Trade 2006)). Commerce here selected Thailand as “the primary surrogate country” under 19 C.F.R. § 351.408 and calculated a surrogate value for Trina’s nitrogen input using Thai import data published by the Global Trade Atlas (“GTA”). J.A. 704. The GTA records the quantity and value of imports into countries by Harmonized Tariff Schedule (“HTS”) classification. Commerce found the Thai GTA nitrogen data to be “reliable.” J.A. 707. The GTA data indicated an overall average unit value (“AUV”) of $11.68 per kilogram for nitrogen during the period of review, which Commerce adopted as the surrogate value for Trina’s nitrogen input. The CIT sustained that determination, concluding that Trina had failed to show that the Thai GTA data were aberrational. Trina asserts that Commerce erred in using the Thai GTA data because it was not the best available information to value Trina’s nitrogen gas input. Trina asserts that the Thai GTA data was “exceedingly aberrational” when compared to alternative surrogate values. Appellant’s Br. 12. We agree with Trina that Commerce has not provided a persuasive reason for using the Thai GTA data in light of unrebutted evidence of its unreliability. A table showing GTA nitrogen data for each of the six “potential surrogate countries” identified by Commerce, J.A. 705, is reproduced below:

Case: 19-1591 Document: 104 Page: 8 Filed: 06/24/2020 8 SOLARWORLD AMERICAS, INC. v. UNITED STATES

Source Quantity Value AUV % of (kgs) (USD) (USD/kg) Total Quantity

Bulgaria 10,657,309 953,544 0.09 86.44% Romania 1,575,456 199,446 0.13 12.78% Thailand 47,618 556,305 11.68 0.39%

South 24,043 131,325 5.46 0.20% Africa

Ecuador 18,462 89,414 4.84 0.15% Ukraine 6,750 531,537 78.75 0.05% Total 12,329,638 2,461,571 0.20 100%

J.A. 813–29. The government contends that Commerce’s use of Thai GTA data was reasonable because it “was within the range of [average unit values] for the other potential surrogate countries [other than Thailand], including Bulgaria ($0.09/kg), Romania ($0.13[/kg]), Ecuador ($4.84[/kg]), South Africa ($5.46[/kg]), and Ukraine ($78.75/kg).” United States’ Br. 17. In essence, the government argues that Commerce reasonably relied on a “bookend” methodology to find the Thai data reliable because its average unit value was neither the highest nor the lowest of the potential surrogate countries. Id. But the use of a bookend methodology here is illogical because it fails to account for the fact that countries on one end of the bookend (Bulgaria and Romania) account for the vast majority (99.22%) of the recorded nitrogen imports, that these countries have a substantially lower average unit value than that of Thailand, and that the other countries

Case: 19-1591 Document: 104 Page: 9 Filed: 06/24/2020 SOLARWORLD AMERICAS, INC. v. UNITED STATES 9

to which Commerce compares the Thai average unit value (the other end of the bookend) together represent only a fraction of a percent of the quantity (about 0.40%) recorded in the GTA. In this respect, Commerce’s decision here to use the Thai GTA data also appears inconsistent with its usual practice. Commerce’s longstanding “administrative practice with respect to aberrational data is ‘to disregard small-quantity import data [from the primary surrogate country] when the per-unit value is substantially different from the per-unit values of the larger quantity imports of that product from other [potential surrogate] countries.’” Shakeproof Assembly Components Div. of Illinois Tool Works, Inc. v. United States, 59 F. Supp. 2d 1354 (Ct. Int’l Trade 1999) (quoting Heavy Forged Hand Tools, Finished or Unfinished, With or Without Handles, From the People’s Republic of China; Final Results of Antidumping Duty Administrative Reviews, 63 Fed. Reg. 16758, 16761 (April 6, 1998)). The Thai GTA data is “small-quantity” (0.39%), and the per-unit value of the Thai GTA data ($11.68/kg) is “substantially different” from the per-unit values GTA data for the two countries with “larger quantity imports”: Bulgaria ($0.09/kg) and Romania ($0.13/kg). Under these circumstances, the use of Thai data appears to be inconsistent with Commerce’s own approach in past cases. Commerce has also failed to explain how the Thai GTA data can be reconciled with data from the United States International Trade Commission’s (“ITC”) Dataweb website, which records exports from the United States to other countries. As relevant here, that data is nitrogen exports from the United States to Thailand during the Period of Review. As previously mentioned, the Thai GTA data records imports into Thailand from other countries (including the United States) during the Period of Review. Commerce admits that the ITC data and Thai GTA data

Case: 19-1591 Document: 104 Page: 10 Filed: 06/24/2020 Case: 19-1591 Document: 104 Page: 11 Filed: 06/24/2020 SOLARWORLD AMERICAS, INC. v. UNITED STATES 11

II We next turn to Trina’s contention that Commerce erred by using records for which a zero quantity of imports was recorded to calculate the average unit value for dozens of its inputs. The GTA records the import value and the import quantity for import transactions into a country. To determine the surrogate value for an input, Commerce calculates the average unit value of all imports of the input into the surrogate country during the period of review. The average unit value for an input is calculated as the total sum of import values (i.e., the total dollar value of all inputs) divided by the total sum of import quantities. Here, the Thai GTA import data relied upon by Commerce to calculate the surrogate value for several of Trina’s inputs included transactions for which a non-zero value was recorded but a quantity of zero was recorded. 1 Trina asserts that the use of these transactions was “mathematically incorrect” because “[t]he result of including these data was that the numerators in the surrogate value calculations increased while the denominators remained the same.” Trina’s Appellant’s Br. 22. We do not reach Trina’s contention that Commerce’s use of zero-quantity data is incorrect because we find that Trina has not satisfied its burden to show that it suffered harm as a result of the purported error. “[T]he party that ‘seeks to have a judgment set aside because of an erroneous ruling carries the burden of showing that prejudice

1 Commerce asserts that these zero-quantity transactions were the result of the rounding down to zero of transactions with an actual quantity of less than 0.05 of the quantity unit.

Case: 19-1591 Document: 104 Page: 12 Filed: 06/24/2020 Case: 19-1591 Document: 104 Page: 13 Filed: 06/24/2020 SOLARWORLD AMERICAS, INC. v. UNITED STATES 13

the total value of imports of the HTS category. Meanwhile, the removed 647 zero-quantity imports account for 6.9 percent of the data points on the record. Removing the zero-quantity imports would have almost no impact on Trina’s margin, despite the zero-quantity entries accounting for 6.9 percent of the import data . . . . J.A. 794 (emphasis added). This analysis by Commerce suggests that the effect of the alleged error was so small as to be negligible. Trina provided no rebuttal to Commerce’s analysis, admitting at oral argument that it could not provide “the exact impact here,” and that “the impact on a percent basis is small.” Oral Argument, 12:27–35, 14:50–15:03. Trina’s position is that if Commerce made a mathematical error here, it would be per se prejudicial regardless of the practical effect of that error. There is no such per se rule, and Trina’s position is directly contrary to the harmless-error rule. We affirm the CIT’s decision to sustain Commerce’s use of zero-quantity data because Trina has failed to demonstrate that the alleged error, if there was an error, was harmful. III We next turn to cross-appellant SolarWorld’s contention that the dumping margin was understated because Commerce undervalued Trina’s and Yingli’s solar- module backsheets. Yingli’s backsheets are made primarily of polyethylene terephthalate (“PET”). Some of Trina’s backsheets are made primarily from PET, and others primarily from ethylene-vinyl acetate (“EVA”).

Case: 19-1591 Document: 104 Page: 14 Filed: 06/24/2020 14 SOLARWORLD AMERICAS, INC. v. UNITED STATES

To value Yingli’s backsheets and Trina’s PET backsheets, Commerce used Thai HTS subheading 3920.62, which covers PET plates and sheets. 3 To value Trina’s EVA backsheets, Commerce used Thai HTS subheading 3920.10, which covers plates and sheets of polymers of ethylene (such as EVA). 4 The CIT sustained that determination. SolarWorld contends that Commerce’s use of these HTS numbers undervalued Yingli’s and Trina’s backsheets because the numbers fail to take into account the technical complexity of the backsheets. As support for this contention, SolarWorld relies on evidence that the actual market-economy prices paid by Yingli and Trina were substantially higher than the average unit values calculated from the HTS numbers used by Commerce. To SolarWorld, Commerce should instead have used Thai HTS subheading 3920.99 (covering plates and sheets of “other” plastics, i.e., those not in any other subheading of 3920, such as 3920.10, 3920.62) for all of Yingli’s and Trina’s backsheets. SolarWorld has not met its burden to show error. As Commerce found and SolarWorld does not dispute, there is no HTS number specific to solar-grade backsheets. Commerce explained that it chose HTS subheading 3920.62 for Yingli’s backsheets and Trina’s PET backsheets because this classification takes into account that the backsheets are comprised primarily of PET. Commerce similarly explained that it chose HTS subheading 3920.10 for Trina’s EVA backsheets because this classification is specific to ethylene products (such as

3 More specifically, Commerce used Thai HTS No. 3920.62.00090. 4 More specifically, Commerce used Thai HTS No. 3920.10.00090.

Case: 19-1591 Document: 104 Page: 15 Filed: 06/24/2020 SOLARWORLD AMERICAS, INC. v. UNITED STATES 15

EVA). Commerce rejected classification under HTS subheading 3920.99 because the subheading is not specific to the material. Commerce’s reasonable choice of the more specific HTS categories is supported by substantial evidence. See Lifestyle Enter., Inc. v. United States, 751 F.3d 1371, 1379 (Fed. Cir. 2014) (affirming Commerce’s use of one of two “imperfect” datasets after it “acknowledged and evaluated potential problems in using either” dataset). Commerce also did not err in discounting SolarWorld’s evidence that the backsheets purchased by Yingli and Trina from market economies were higher-priced than the average unit values of HTS categories 3920.10 and 3920.62. We find Commerce reasonably followed its practice of finding such information unpersuasive because “a respondent’s market economy purchase prices are proprietary information and are not necessarily representative of industry-wide prices available to other producers.” J.A. 727–28. We affirm the CIT. IV We finally turn to SolarWorld’s contention that Commerce’s use of Thai GTA data to value Yingli’s tempered glass input was appropriate and that the CIT erred in holding the data could not be used. The CIT concluded that Commerce failed to persuasively explain why it used Thai GTA data in light of distortions by aberrational imports from Hong Kong. The CIT remanded for Commerce to justify its use of the Thai data or adopt an alternative surrogate value. On remand, Commerce provided additional explanation for its use of the Thai data. The CIT found Commerce’s additional justification unpersuasive and again remanded. On this second remand, Commerce used under protest Bulgarian GTA data to value Yingli’s tempered glass input. The government does not on appeal argue that the CIT erred in remanding to Commerce, but SolarWorld does.

Case: 19-1591 Document: 104 Page: 16 Filed: 06/24/2020 16 SOLARWORLD AMERICAS, INC. v. UNITED STATES

We agree with the CIT that Commerce did not adequately justify its use of the Thai GTA data. The imports from Hong Kong have a unit value 191 times higher than the average unit value for imports into Thailand from other countries. Thus, even though the Hong Kong imports constitute only 1.6% of Thailand’s import quantity, their inclusion causes the Thai average unit value to be quadrupled. “There is no reason to incorporate the distortions in the surrogate market into a hypothetical respondent market.” SeAH, 950 F.3d at 845 (quoting Nation Ford, 166 F.3d at 1378) (alterations omitted). SolarWorld asserts that Commerce’s original decision to use the Thai data is consistent with its earlier decision in Wood Flooring. See Issues and Decision Mem. for the Final Results of the 2012–2013 Admin. Review of Multilayered Wood Flooring from the PRC, No. A-570-970 (July 8, 2015) (“Wood Flooring”). To SolarWorld, because the overall Thai average unit value for tempered glass ($4.14/kg) is in line with that of other potential surrogate countries (Ecuador at $2.75/kg, and Ukraine at $5.89/kg), Commerce reasonably concluded that the Thai data was not aberrational. As the CIT concluded, SolarWorld’s reliance on Wood Flooring is misplaced. In Wood Flooring, Commerce rejected an argument that imports to Thailand from Taiwan and the United States should be excluded from an average-unit-value calculation because they were aberrantly low in price. The “imports from Taiwan and the United States represent[ed] the vast majority of imports into Thailand (77.1%) and, therefore, [were] a true representation of market-driven prices.” Wood Flooring, at 43. Here, by contrast, imports from Hong Kong constitute only 1.6% of imports into Thailand, so cannot be said to be “a true representation of market-driven prices.” See id.

Case: 19-1591 Document: 104 Page: 17 Filed: 06/24/2020 SOLARWORLD AMERICAS, INC. v. UNITED STATES 17

SolarWorld’s contention that the Hong Kong imports were not in fact aberrationally high is similarly unpersuasive. SolarWorld notes that the average unit prices of imports from the United States and the Netherlands into Thailand were greater than the average unit price of imports from Hong Kong. But imports from those countries constitute an infinitesimal fraction (0.0236% and 0.0003%, respectively) of total imports into Thailand, and do not indicate that the Hong Kong price was representative of the market. We affirm the CIT’s remand to Commerce to further explain or reconsider its decision to use the Thai GTA data to value Yingli’s tempered glass input. CONCLUSION We vacate the CIT’s judgment sustaining Commerce’s decision to use Thai GTA data to value Trina’s nitrogen input. We affirm the CIT’s judgment sustaining Commerce’s decision to use zero-quantity data. We affirm the CIT’s judgment declining to set aside Commerce’s use of Thai HTS subheadings 3920.62 and 3920.10 to value Trina’s and Yingli’s backsheets. We affirm the CIT’s judgment remanding to Commerce to further justify or reconsider its use of Thai GTA data to value Yingli’s tempered glass input. We remand the case for further proceedings consistent with this opinion. AFFIRMED IN PART, VACATED IN PART, AND REMANDED COSTS No costs.

Footnotes

[10] SOLARWORLD AMERICAS, INC. v. UNITED STATES for imports from the United States thus relate to the same real-world transactions. The ITC data indicated that about 136 times more nitrogen was exported from the United States into Thailand (roughly 586,305 kg) than the GTA data indicated was imported into Thailand from the United States (4,298 kg). And the ITC data indicated an average unit value for nitrogen of $0.16/kg, as opposed to the GTA value of $11.68/kg. The ITC data and the Thai GTA data cannot both be correct, as Commerce appears to admit. See Oral Argument, 35:55-36:04 (when asked whether the Thai GTA data and the ITC data can both be correct, the government stated “probably not”). Commerce has not explained why the Thai GTA data is a more accurate record of these transactions than the ITC data, admitting that it “just do[es]n’t know” which is accurate. Oral Argument, 36:40-49. The government asserts that the ITC data is irrelevant because “Commerce’s preference [is] to use import data for surrogate values[] [to enable] comparison between similar datasets,” and that “the value of U.S. exports to Thailand, which reflects only one data point, does not represent a broad-market average.” United States’ Br. 19. But Commerce’s preference for GTA data does not excuse its failure to reconcile the admitted inconsistency. In sum, Commerce has not provided sufficient justification for its conclusion that the Thai GTA data was the “best available information” from which to value Trina’s nitrogen input. See SEC v. Chenery Corp., 318 U.S. 80, 95 (1943). We remand to the CIT for Commerce to either adequately explain why the Thai GTA data is not aberrational or to adopt an alternative surrogate value for Trina’s nitrogen input.

[12] SOLARWORLD AMERICAS, INC. v. UNITED STATES resulted.’” Shinseki v. Sanders, 556 U.S. 396, 409 (2009) (quoting Palmer v. Hoffman, 318 U.S. 109 (1943)). “Consequently, the burden of showing that an error is harmful normally falls upon the party attacking the agency’s determination.” Id. In the antidumping context, a party challenging a purported error by Commerce must show that it was harmed as a result of the error. See Suntec Indus. Co. v. United States, 857 F.3d 1363, 1367 (Fed. Cir. 2017) (affirming judgment against foreign exporter because exporter failed to show prejudice caused by Commerce’s purported error). 2 The record indicates that Commerce’s decision to use zero-quantity data—whether or not correct—had essentially no impact on Trina’s antidumping duty rate. Commerce explained: Importantly, removal of the zero-quantity entries has almost no impact on almost any of the AUVs. For 70 of the 76 different HTS categories of GTA import data on the record, the difference between the AUV of the data including zero-quantity imports and the AUV of the data without zero- quantity imports rounds to zero percent; for four other HTS categories, the difference is one percent. The differences for the other two HTS categories are two and five percent. The average difference of the 76 different HTS categories is 0.16 percent of 2 The review in Suntec was under 28 U.S.C. § 2640(e). Section 2640(e) expressly incorporates section

[706] of the Administrative Procedure Act (“APA”), which includes a harmless-error review provision. The review here is under 28 U.S.C. § 2640(b), which does not expressly refer to section 706. Even so, section 706 review applies since no law provides otherwise. Dickinson v. Zurko, 527 U.S. 150, 154 (1999); 5 U.S.C. § 706.