Slip Op. 23-
UNITED STATES COURT OF INTERNATIONAL TRADE
RISEN ENERGY CO., LTD.,
Plaintiff,
JINGAO SOLAR CO., LTD., ET AL.,
Consolidated Plaintiffs, Before: Jane A. Restani, Judge SHANGHAI BYD CO., LTD., TRINA SOLAR CO., LTD., ET AL., Consol. Court No. 20-03912 Intervenor Plaintiffs,
v.
UNITED STATES,
Defendant.
OPINION AND ORDER
[Commerce’s Remand Results in the Sixth Administrative Review of Commerce’s Countervailing Duty order on crystalline silicon photovoltaic cells from the People’s Republic of China are remanded for reconsideration consistent with this opinion.]
Dated: April 11, 2023
Gregory S. Menegaz and Alexandra H. Salzman, deKieffer & Horgan, PLLC, of Washington, DC, argued for Plaintiffs. With them on the brief was James K. Horgan.
Sarah M. Wyss, Mowry & Grimson, PLLC, of Washington, DC, argued for Consolidated Plaintiffs. With her on brief were Jeffrey S. Grimson, Bryant P. Cenko, Jill A. Cramer, Kristin H. Mowry, Yixin (Cleo) Li.
Craig A. Lewis and Jonathan T. Stoel, Hogan Lovells US LLP, of Washington, DC, for Intervenor Plaintiffs Shanghai BYD Co., Ltd.
Jonathan M. Freed, Kenneth N. Hammer, MacKensie R. Sugama, and Robert G. Gosselink, Trade Pacific PLLC, of Washington, DC, for Intervenor Plaintiffs Trina Solar Co., Ltd. Consol. Court No. 20-03912 Page 2
Joshua E. Kurland, Commercial Litigation Branch, U.S. Department of Justice, of Washington, DC, argued for Defendant. With him on the brief were Brian M. Boynton, Patricia M. McCarthy, and Reginald T. Blades, Jr. Of counsel on the brief was Spencer Neff, Office of Chief Counsel for Trade Enforcement & Compliance, U.S. Department of Commerce, of Washington, DC.
Restani, Judge: Before the court are the remand results of the U.S. Department of
Commerce (“Commerce”) pursuant to the court’s order in Risen Energy Co. v. United States, 46
CIT __, __, 570 F. Supp. 3d 1369, 1382 (2022), in the Sixth Administrative Review of the
countervailing duty order on crystalline silicon photovoltaic cells, whether or not assembled into
modules (“solar cells”), from the People’s Republic of China, covering the period from January 1,
2017, to December 31, 2017. See Redetermination Pursuant to Court Remand Order, ECF Nos.
93–94 (“Remand Results”). Plaintiff Risen Energy Co., Ltd. (“Risen Energy”) and Consolidated
Plaintiff Jingao Solar Co, Ltd. (“JA Solar”) (collectively, “Plaintiffs”)1 challenge the Remand
Results as unsupported by substantial evidence or otherwise not in accordance with law.
BACKGROUND
While the court presumes familiarity with the facts as set out in Risen, the court briefly
summarizes the relevant record evidence for ease of reference. In March 2019, Commerce began
the Sixth Administrative Review of the countervailing duty order on solar cells from the People’s
Republic of China. Initiation of Antidumping and Countervailing Duty Administrative Reviews,
84 Fed. Reg. 9,297, 9303–04 (Dep’t Commerce Mar. 14, 2019). On November 5, 2019, the U.S.
International Trade Administration selected JA Solar and Risen Energy as mandatory respondents
(“Mandatory Respondents”) in this review. See Dep’t Commerce, Resp’t Selection Mem. at 1–2,
1 Intervenor Plaintiffs Shanghai BYD Co., Ltd. (“Shanghai BYD”) and Trina Solar Co., Ltd. (“Trina”) are non-examined parties who seek the benefits of whatever relief the court grants. See Remand Results at 4; Trina Comments on Remand Results, ECF No. 99 (Nov. 7, 2022); Shanghai BYD Comments on Remand Results, ECF No. 100 (Nov. 7, 2022) (“Shanghai BYD Brief”). Consol. Court No. 20-03912 Page 3
P.R. 98 (Nov. 5, 2019).
Commerce published its preliminary results on February 11, 2020, see Crystalline Silicon
Photovoltaic Cells, Whether or Not Assembled Into Modules, From the People’s Republic of
China: Preliminary Results of Countervailing Duty Administrative Review and Rescission of
Review, in Part; 2017, 85 Fed. Reg. 7,727 (Dep’t Commerce Feb. 11, 2020), along with the
accompanying Preliminary Issues and Decision Memorandum, Decision Memorandum for the
Preliminary Results of the Countervailing Duty Order on Crystalline Silicon Photovoltaic Cells,
Whether or Not Assembled Into Modules, from the People’s Republic of China, C-570-980, POR
01/01/2017–12/31/2017 (Dep’t Commerce Jan. 31, 2020) (“PDM”).
Commerce published its final determination on December 9, 2020. See Crystalline Silicon
Photovoltaic Cells, Whether or Not Assembled Into Modules, From the People’s Republic of
China: Final Results of Countervailing Duty Administrative Review; 2017, 85 Fed. Reg. 79,163
(Dep’t Commerce Dec. 9, 2020); see also Issues and Decision Memorandum for Final Results of
the Administrative Review of the Countervailing Duty Order on Crystalline Silicon Photovoltaic
Cells, Whether or Not Assembled Into Modules, from the People’s Republic of China, C-570-980,
POR 01/01/2017-12/31/2017 (Dep’t Commerce Nov. 27, 2020) (“I&D Memo”).
In Risen, the court upheld Commerce’s determination that Plaintiffs received regionally
specific electricity subsidies subject to countervailing duties. See 570 F. Supp. 3d at 1382. The
court remanded to Commerce to reconsider (1) the benchmark for land prices in China and (2) the
benchmark for determining the cost of ocean freight for subsidy calculations involving provisions
of raw materials for less than adequate remuneration. Id. at 1376, 1379. Additionally, the court
granted the United States’ (“Government”) request for remand on the Government of China’s
(“GOC”) Export Buyer’s Credit Program (“EBCP”) but instructed Commerce to attempt to verify Consol. Court No. 20-03912 Page 4
or to explain the reason that the court “should not provide some form of equitable relief.” Id. at
1373.
Following remand, Commerce recalculated (1) the land benchmark by averaging the prior
dataset with new data placed on the remand record and (2) the ocean freight benchmark by
adjusting the previous average to correct the overreliance on data related to United States to China
routes. See Remand Results at 1–2. Finally, Commerce attempted to verify nonuse of the EBCP
by the Mandatory Respondents, finding that JA Solar did not use the program but concluding that
Risen could not show nonuse. Remand Results at 1.
The Remand Results do not adequately address all of the court’s concerns in Risen and
they are not supported by substantial evidence. Accordingly, the court once again remands with
further instructions.
JURISDICTION & STANDARD OF REVIEW
The court’s jurisdiction continues pursuant to 19 U.S.C. § 1516a(a)(2)(B)(i) and 28 U.S.C.
§ 1581(c). The court sustains Commerce’s final redetermination results unless they are
“unsupported by substantial evidence on the record, or otherwise not in accordance with law[.]”
19 U.S.C. § 1516a(a)(2)(B)(i). “The results of a redetermination pursuant to court remand are also
reviewed for compliance with the court’s remand order.” U.S. Steel Corp. v. United States, 41
CIT __. __, 219 F. Supp. 3d 1300, 1307 (2017) (internal quotations and citations omitted).
DISCUSSION
I. Export Buyer’s Credit Program
The GOC’s EBCP promotes exports by providing credit at preferential interest rates to
qualifying foreign purchasers of GOC goods. See Clearon Corp. v. United States, 43 CIT __, __,
359 F. Supp. 3d 1344, 1347 (2019). During the investigation, Risen reported that none of its Consol. Court No. 20-03912 Page 5
customers used the EBCP during the period of review (“POR”) and confirmed that it had never
been involved in assisting customers in obtaining loans under the program; it also provided
certifications of nonuse from its U.S. customers attesting to this fact. See Risen Energy Section
III Questionnaire Response, at 23–24, Ex. 19, P.R. 144–162, C.R. 109–276 (Dec. 30, 2019); Risen
Unaffiliated Supplier II, Section III Questionnaire Response at 23, Ex. 15, P.R. 164, C.R. 277 (Jan.
6, 2020). JA Solar similarly reported that none of its affiliated or unaffiliated customers received
assistance under the EBCP and that it did not assist any customers in using the program. See
Questionnaire Response of JA Solar and Affiliates, at Volume 1, III 38–40, P.R. 132–138, C.R.
31–103 (Dec. 30, 2019). JA Solar provided customer declarations certifying nonuse of the EBCP.
Id. at Ex. 25.
The GOC did not provide all of the initially requested information to Commerce, stating
that the questions were inapplicable because “the GOC believes that none of the respondents under
review applied for, used, or benefitted from the alleged program.” GOC Initial Questionnaire
Response at 126, P.R. 140–143, C.R. 104–108 (Dec. 30, 2019); see I&D Memo at 27. Consistent
with the certifications of the Mandatory Respondents, the GOC, by searching the China Ex-Im
Bank’s loan database, corroborated that Mandatory Respondents and their customers did not use
the EBCP. GOC Initial Questionnaire Response at 126–28.
Following remand, Commerce explained that it was changing its practice relating to EBCP,
and, going forward, would issue supplemental questionnaires if a respondent provided complete
customer declarations of EBCP nonuse. Remand Results at 18–19. Commerce explained that it
would require respondents to provide customer financial records constituting “complete gap-filling
information.” Remand Results at 19. If the respondents could not provide complete gap-filling
information, Commerce would be “left only with the GOC’s non-cooperation,” and thus would Consol. Court No. 20-03912 Page 6
apply an adverse inference as to the available facts (“AFA”) based on that non-cooperation.
Remand Results at 19. Applying the new practice on remand, Commerce sent questionnaires to
JA Solar and Risen requesting information about loans, financing, and record keeping by U.S.
customers to determine if customers received EBCP assistance. Remand Results at 6. JA Solar
provided complete information for its sole importer, JA Solar USA, Inc. for Commerce to review
and verify nonuse. Remand Results at 6–7. Commerce verified that JA Solar USA received no
loans or financing connected with the GOC. Remand Results at 7; see also JA Solar Verification
Report Rem. P.R. 27, Rem. C.R. 24 (Aug. 31, 2022). Commerce removed the previously applied
EBCP subsidy rate from JA Solar’s total rate. Remand Results at 7.
Commerce reached a different result with respect to Risen. Remand Results at 7–9. Risen
had 12 U.S. customers but provided Commerce’s requested information for only 6 of the
customers, which Risen estimated were responsible for roughly 95% of its POR sales. Remand
Results at 7. Risen’s other six customers either did not respond, had ceased operations since the
POR, or answered that they did not use the EBCP. Remand Results at 7–8. As a result, Commerce
found that Risen had failed to provide sufficient information to fill the evidentiary gap in the
record. Remand Results at 8–9. Commerce reasoned that verification based on only a portion of
customers would provide Risen with “an opportunity to evade Commerce’s scrutiny” by providing
responses only from customers who did not use the EBCP. Remand Results at 9. Ergo, Commerce
concluded that Risen failed to fill the gap created by GOC’s failure to cooperate, and thus used
AFA to determine that Risen benefited from the EBCP. Remand Results at 9. Although Risen
suggested that Commerce adjust the AFA rate to account for customers who did provide requested
information, Commerce declined because there was no “precedent for doing so” and it had “no
way of knowing the size of the subsidies received by customers.” Remand Results at 20. Consol. Court No. 20-03912 Page 7
Now, Risen objects to the Remand Results, arguing that the record contains sufficient
information to verify nonuse of the EBCP. Risen Comments on Remand Results, ECF Nos. 97–
98 (Nov. 7, 2022) (“Risen Br.”). Risen points to the fully cooperating customers, whose responses
showed that they did not benefit from the EBCP, and asserts that there is no record evidence to the
contrary. Risen Br. at 3. Risen contends that these 6 customers, representing roughly 95% of
Risen’s sales, are a sufficient sample to verify nonuse. Risen Br. at 4. Risen argues that, at worst,
Commerce should modify the rate to account for the fact that there is record evidence
demonstrating that roughly 95% of sales were not benefited by the EBCP. Risen Br. at 7–8.
If “necessary information is not available on the record” or if a responding party “withholds
information” requested by Commerce, Commerce shall “use the facts otherwise available in
reaching the applicable determination[.]” 19 U.S.C § 1677e(a). Commerce may use AFA only
when information is missing from the record because a party “has failed to cooperate by not acting
to the best of its ability to comply with a request for information” from Commerce. 19 U.S.C. §
1677e(b). The application of adverse facts that collaterally impact a cooperating party is
disfavored. Fine Furniture (Shanghai) Ltd. v. United States, 36 CIT 1206, 1212 n.10, 865 F. Supp.
2d 1254, 1262 n.10 (2012), aff’d, 748 F.3d 1365 (Fed. Cir. 2014). “When Commerce has access
to information on the record to fill in the gaps created by the lack of cooperation by the
government, as opposed to the exporter/producer, however, it is expected to consider such
evidence.” GPX Int’l Tire Corp. v. United States, 37 CIT 19, 58–59, 893 F. Supp. 2d 1296, 1332
(2013), aff’d, 780 F.3d 1136 (Fed. Cir. 2015); see also Guizhou Tyre Co., Ltd. v. United States,
42 CIT __, __, 348 F. Supp. 3d 1261, 1270 (2018) (“To apply AFA in circumstances where
relevant information exists elsewhere on the record — that is, solely to deter non-cooperation or
‘simply to punish’ — . . . that is a fate this court should sidestep.”) (citation omitted). Consol. Court No. 20-03912 Page 8
Information submitted by parties is subject to verification by Commerce. 19 U.S.C. §
1677m(i)(1). Commerce need not consider unverifiable information, but Commerce must show
that such information is not reasonably verifiable before it applies AFA. See Changzhou Trina
Solar Energy Co., Ltd. v. United States, 42 CIT __, __, 352 F. Supp. 3d 1316, 1327 (2018) (citing
19 U.S.C. § 1677m(e)); Papierfabrik August Koehler SE v. United States, 843 F.3d 1373, 1382–
83 (Fed. Cir. 2016) (holding that if the requirements of § 1677m(e) are not met, Commerce need
not consider information submitted by an interested party).
As Commerce makes plain, it is using AFA based on a gap in the record about the EBCP
that occurred when the GOC declined to provide all the information Commerce sought regarding
the program. Remand Results at 9. Although the GOC did not provide all of the information it
should have, another party, such as a mandatory respondent, may be in a position to provide the
information sought or to render that information irrelevant, so that no adverse inference need be
drawn. See Jiangsu Zhongji Lamination Materials Co. v. United States, 43 CIT __, __, 405 F.
Supp. 3d 1317, 1333 (2019); GPX Int’l Tire, 37 CIT at 58–59, 893 F. Supp. 2d at 1332. Here,
Risen has supplied information so that there is no relevant missing information about the EBCP.
Not only has Risen provided sworn declarations from each of its customers stating that they did
not use financing from the EBCP, see Risen Unaffiliated Supplier II, Section III Questionnaire
Response, at 23, Ex. 15, P.R. 164, C.R. 277 (Jan. 6, 2020), but, after remand, Risen supplied
financial, loan, and record information regarding 6 of its 12 customers, representing roughly 95%
of sales during the POR. See Risen EBC Questionnaire Response at 1–2, Rem. P.R. 12, Rem. C.R.
2–4 (July 8, 2022) (“Risen EBC Questionnaire Response”); see also Remand Results at 7–8.
Commerce’s refusal to verify the customer data and continued application of other facts available Consol. Court No. 20-03912 Page 9
is not supported by substantial evidence on this record because the information necessary to the
determination, assuming it is verified, is not lacking.
This case involves a specific agency record that drives the court’s conclusion. During the
review, and before remand, as other respondents had done in other cases, Risen provided
declarations from all of its nonaffiliated customers, which stated that they did not use the EBCP.
See Risen Unaffiliated Supplier II, Section III Questionnaire Response at 23, Ex. 15, P.R. 164,
C.R. 277 (Jan. 6, 2020). At the time, Commerce concluded that these declarations could not be
verified to show nonuse. See Remand Results at 5. During the proceedings at the court, however,
Commerce shifted its EBCP policy, instead choosing to send supplemental questionnaires
requesting extensive customer financial documents whenever a respondent provided the
declarations of nonuse. See Remand Results at 6, 18–19. After remand, years after the POR and
the sales at issue were made, Commerce required Risen to coordinate sensitive and voluminous
financial records from all of its nonaffiliated customers to show that it did not receive a benefit
from the EBCP. See Commerce’s Letters, “Export Buyer’s Credit Supplemental Questionnaire,”
at 3–4, Volume 1, Rem. P.R. 1–2 (July 7, 2022). Risen had to do this to attempt to avoid the
collateral impact of the GOC’s failure to cooperate, despite Risen’s cooperation at every step along
the way. That brings the court to this decision.
Commerce has stated that it cannot verify that Risen did not benefit from the EBCP because
6 of Risen’s 12 customers did not provide sufficient financial information. See Remand Results
at 7–9. Of the missing six customers, three did not respond to Risen’s requests, one responded
that its loans “had nothing to do with the EBC program or Risen,” another that “it did not use any
financing,” and a final company that had ceased operating during the POR. See Risen EBC
Questionnaire Response at 2. But the parties agree that Risen provided financial information for Consol. Court No. 20-03912 Page 10
its other 6 customers, who collectively represented roughly 95% of Risen’s sales during the POR.
See Remand Results at 7; Risen Br. at 1.
Commerce claims that it cannot verify nonuse based on these responses because Risen
could evade Commerce’s scrutiny by hiding EBCP usage in data from the noncooperating
companies. See Remand Results at 9. And, Commerce explains, because Risen did not provide
“complete gap-filling information,” Commerce declined to attempt to verify Risen’s supplied
information. See Remand Results at 18, 20. Commerce concluded that there is still a gap in
verifiable information caused by the GOC’s failure to supply full information about the EBCP.
See Remand Results at 19–20. Ultimately, the question before the court is whether Commerce’s
explanation is reasonable. Cooper (Kunshan) Tire Co., Ltd. v. United States, 45 CIT __, __, 539
F. Supp. 3d 1316, 1333 (2021); see also Shandong Huarong Gen. Corp. v. United States, 25 CIT
834, 837, 159 F. Supp. 2d 714, 718 (2001) (stating that the court should uphold the agency
determination as long as “its factual findings are reasonable and supported by the record as a
whole”) (citations omitted), aff’d sub nom. Shandong Huarong Gen. Grp. Corp. v. United States,
60 F. App’x 797 (Fed. Cir. 2003).
Commerce cannot reasonably conclude that Risen did not supply information that rendered
the missing data from the GOC irrelevant and essentially eliminated any gap in the record. Risen
substantially complied with Commerce’s investigation efforts and provided near complete data for
Commerce to review even after the long passage of time. See Risen EBC Questionnaire Response
at 1. The missing financial data from roughly 5% of sales involved smaller importers who likely
did not have enough of an interest to justify sharing the sensitive financial information with Consol. Court No. 20-03912 Page 11
Commerce years after the POR.2 Considering that the POR was five years ago, that Commerce
changed its policy, and that Risen complied to the best of its ability, the court concludes it is
unreasonable for Commerce to require perfection. All of the record evidence points to nonuse of
the program at issue. Commerce’s concern about potentially hiding the use of EBCP in the
nonresponding companies is not reasonable when considering the collateral impact of AFA on the
fully cooperating Risen, the age of this case, and the still-relevant initial complete set of nonuse
declarations, which has not been seriously undermined. Substantial evidence does not support
Commerce’s continued application of AFA to Risen’s detriment on this record. On remand,
Commerce should attempt to verify Risen’s submissions to the extent Commerce finds appropriate,
and if that is successful, it should either accept the pro rata adjustment sought by Risen or conclude
that the EBCP was not used at all.
II. Land Benchmark
When remanding, the court questioned Commerce’s continual reliance on data from 2010
Coldwell Banker Richard Ellis Asian Marketview Report (“2010 CBRE Report”) for Thailand
land prices to calculate the tier-three benchmark for the value of land-use rights. Risen, 570 F.
Supp. 3d at 1375–76. Commerce indexed the 2010 Thai prices to the POR when calculating the
benchmark. I&D Memo at 51. The court expressed concerned as to why Commerce gave
controlling weight to geographic proximity in evaluating land data and why Commerce rejected
more contemporaneous data from other countries, such as Mexico and Brazil, or the supplemental
2 Commerce based the subsidy rate for EBCP on another program in a past proceeding. See PDM at 38–39. Because Commerce has not obtained evidence of the use of the program at hand, it has established no actual rate for the EBCP. It seems consistent with the record evidence of nonuse that the possibly small benefit involved has not motivated customers in the United States to pursue financing through the GOC’s export import bank. Whatever the motivation or lack therof, as indicated, there is no evidence that this program was used by the customers involved in the sales at issue. Consol. Court No. 20-03912 Page 12
Nexus reports in favor of using index data from 2010. Risen, 570 F. Supp. 3d at 1375–76; Letter
on Behalf of JA Solar to Dep’t of Commerce re: Land Benchmark Submission, at Ex. 1, P.R. 192
(Feb. 18, 2020) (“Nexus Reports”).3
After remand, Commerce placed Malaysian land values from the Malaysian Investment
Development Authority Cost of Doing Business Report on the record as well as additional
information on the comparability of Malaysia and Thailand to China. See Dept. Mem. Re.
Upcoming Draft Remand Results – The Provision of Land for Less Than Adequate Remuneration
– Malaysia: Costs of Doing Business, Rem. P.R. 16 (Aug. 8, 2022); see also Dept. Mem. Re.
Benchmark Analysis of the Government Provision of Land-Use Rights in China for Countervailing
Duty Purposes at 30, P.R. 203 (Jan. 31, 2020) (“Land Use Memo”); Dept. Mem. Re. “Upcoming
Draft Remand Results – Population Densities of Countries”, Rem. P.R. 17 (August 8, 2022)
(“Commerce Population Density Data”); Dept. Memo. re. Upcoming Draft Remand Results –
Level of Economic Development, Rem. P.R. 18 (August 8, 2022) (“Commerce Economic
Development Information”).
Commerce considered the Commerce Population Density Data because it indicated that
Thailand and Malaysia had similar population densities to China’s. Remand Results at 11. China
had a population density of 140 persons per square kilometer (K2) while Malaysia had 99/K2 and
Thailand had 130/K2. See Commerce Population Density Data at 1, 6–7. Further, Commerce
considered that China, Malaysia, and Thailand were classified as “upper middle income” countries
by the World Bank. Remand Results at 11; Commerce Economic Development Information at 5–
3 The parties are not pursuing arguments related to Brazil, Mexico, or the Nexus Reports. Instead, the Mandatory Respondent seek use of Malaysian data without averaging it with the Thai data. JA Solar Comments on Remand Results, ECF No. 96 (Nov. 7, 2022) (“JA Solar Br.”) at 7–9; Risen Br. at 8–9. Accordingly, the court will not address these matters further. Consol. Court No. 20-03912 Page 13
7. Based on these factors, Commerce concluded that data from Thailand and Malaysia would best
approximate a benchmark for land in China. Remand Results at 11–13. Acknowledging, however,
the court’s concern that the Thailand data was stale, Commerce calculated a simple average of the
Malaysian and Thai datasets to construct the benchmark. Remand Results at 11.4
JA Solar presented data that showed that the difference between Thailand’s and China’s
gross national income per capita (“GNI”) was almost twice that of the difference between
Malaysia’s and China’s GNI. See JA Solar Rebuttal Comments on Level of Economic
Development at Ex. 1, Rem. P.R. 21 (Aug. 16, 2022). Commerce declined to adjust the benchmark
because Commerce does not “prioritize the closeness of per capita GNI in selecting a benchmark.”
Remand Results at 22. Once again at the court, Plaintiffs challenge the use of the 2010 CBRE
Report for Thailand data, now arguing that Commerce should rely solely on the contemporary
Malaysia dataset. JA Solar Br. at 7–9; Risen Br. at 8–9.
Under 19 U.S.C. § 1677(5)(E)(iv), Commerce must set benchmarks that reflect “prevailing
market conditions.” The statute further defines prevailing market conditions as including “price,
quality, availability, marketability, transportation, and other conditions of purchase or sale.” 19
U.S.C. § 1677(5)(E)(iv). 19 C.F.R. § 351.511(a)(2) provides additional guidance on how
Commerce sets benchmarks, setting out three methodological tiers. 19 C.F.R. § 351.511(a)(2).
A tier-three benchmark “measure[s] the adequacy of remuneration by assessing whether
the government price is consistent with market principles.” Id. § 351.511(a)(2)(iii). “If Commerce
determines that the government price is not consistent with market principles it will look to
4 Commerce explained that although this was a tier-three benchmark, “its approach is consistent with the methodology it employs for tier two, under which Commerce will average together available prices when more than one are available.” Commerce’s Resp. to Comments, ECF No. 105 (Jan. 18, 2023); see 19 C.F.R. § 351.511(a)(2)(ii). Consol. Court No. 20-03912 Page 14
construct an external benchmark.” Risen Energy, 570 F. Supp. 3d at 1374. When Commerce has
multiple datasets available, it “will average such prices to the extent practicable, making due
allowance for factors affecting comparability.” 19 C.F.R. § 351.511(a)(2)(ii).
Commerce, however, fails to provide sufficient reasons to continue relying on the stale
data from Thailand when it has the contemporary data from Malaysia. As Commerce explained,
it selected the Malaysian land values because Malaysia also had a comparable level of economic
development to China and was similarly grouped as an upper-middle income country with
Thailand. See Commerce’s Resp. to Objections, ECF No. 105 (Jan. 18, 2023) at 15–16; Remand
Results at 12; see also Commerce Economic Development Information. Although Commerce
decided to average the datasets together for the benchmark, Commerce offered no rationale for the
decision to utilize both data and conceded that “nowhere on remand did [it] pursue a comparison
between the two sources.” Commerce’s Response to Objections at 17. As long as both countries
satisfy basic comparability standards, Commerce does not have to decide whether Malaysia is
more, less or equally comparable to China as Thailand is. It has to decide if it has such defective
data that it should be rejected in favor of better data. The Malaysian data is contemporaneous
while the Thai data is outdated by seven years and requires inflationary adjustments. Commerce
fails to address the court’s concern regarding the staleness of the 2010 CBRE Report regarding
Thailand. Thus, Commerce did not comply with the court’s remand instructions. Because the
Malaysian data is useable and the Thai appears to be defective, at this late stage, Commerce must
provide a compelling reason for its continued use of the stale 2010 CBRE report or otherwise use
the Malaysian data only. Consol. Court No. 20-03912 Page 15
III. Ocean Freight
In its original determination, Commerce calculated a tier-two benchmark for the price of
ocean freight for bringing solar glass, polysilicon, and aluminum extrusions to China from various
cities as a world market price, pursuant to 19 C.F.R. § 351.511(a)(2)(ii). I&D Memo at 55–56. In
order to calculate the benchmark, Commerce used a simple average of two submitted data sets,
one from Descartes and the other from Xeneta. I&D Memo at 55. The Xeneta data reflected prices
from “monthly ocean freight data for shipping a 20-foot standard container to Shanghai” from
various ports across the world including Barcelona, Busan, Singapore, Jakarta, Los Angeles,
Rotterdam, and Mumbai. JA Solar, Benchmark Submission at Ex. 7C, C.R. 284–96, P.R. 166–68
(Jan. 13, 2020). At the same time, the Descartes data consisted of freight rates from various
American cities, including Los Angeles, Portland, San Francisco, Seattle, Chicago, Murrieta, and
Atlanta, to Shanghai. Petitioner, Submission of Benchmark Information at Exs. 5–7, P.R. 170–75
(Jan. 13, 2020). Many of the shipments had the same “Tariff Code” and freight forwarder code,
and some of the data also stated that the container size was less than a container load. Id.
In Risen, the court remanded the benchmark calculation to reconsider flaws in the
Descartes data raised by Plaintiffs. 570 F. Supp. 3d at 1379. The court was concerned whether
the potentially flawed Descartes data was necessary to arrive at a “world market price” when the
Xeneta data provided global routes compared to the Descartes data’s U.S.-focused routes. Id. The
court also considered that the Descartes data might be flawed because it appeared to be sourced
from limited samples based on the codes, was for less than a container load, and some of the routes
were from inland American cities, which could incur additional fees. Id. As a result, the court
remanded to Commerce to reconsider these flaws before determining if any use of the Descartes
data was appropriate. Id. Further, the court instructed Commerce that, if it used the Descartes Consol. Court No. 20-03912 Page 16
data, it should average it only with the Xeneta’s United States to China routes data instead of using
it in a simple average with world-wide data to prevent data limited to U.S. routes from improperly
ballooning the benchmark. Id.
On remand, Commerce excluded the double counting of inland freight by omitting that
data and averaged the Descartes route values with the Xeneta values for shipments from the United
States. Remand Results at 14. Regarding the other alleged flaws in Descartes data, Commerce
did not agree that they warranted excluding the dataset. Remand Results at 14–15. Commerce
stated that it did not believe that “a hypothetical importer would as a rule not buy less than a
container.” Remand Results at 15. Commerce also concluded that the record did not show that a
hypothetical importer would not use these shipments based on the tariff codes and freight
forwarder codes. Remand Results at 15. Commerce therefore continued to rely upon the Descartes
data in determining the benchmark, only with the modifications requested by the court. Remand
Results at 15.
Plaintiffs challenge the Remand Results, arguing that Commerce did not adequately
address the flaws in the Descartes data. See JA Solar Br. at 9–10; Risen Br. at 9–10; Shanghai
BYD Br. at 7–8. Plaintiffs assert that Commerce failed to consider the tariff codes and freight
forwarder codes. JA Solar Br. at 9; Risen Br. at 9. Plaintiffs suggest that the codes indicate that
the Descartes data is derived from limited route samples because each shared the same codes. JA
Solar Br. at 9; Risen Br. at 9. JA Solar contends that Commerce failed to explain what the
Descartes data added that the Xeneta data did not already include in the benchmark calculation.
JA Solar Br. at 10. Plaintiffs agree, however, that Commerce complied with the remand
instructions when omitting the inland routes from the Descartes dataset. See JA Solar Br. at 9–10;
Risen Br. at 9–10; Shanghai BYD Br. at 7–8. Consol. Court No. 20-03912 Page 17
For a tier-two benchmark, Commerce compares “the government price to a world market
price where it is reasonable to conclude that such price would be available to purchasers in the
country in question.” 19 C.F.R. § 351.511(a)(2)(ii). When there is more than one dataset
representing the world market price, then Commerce “will average such prices to the extent
practicable, making due allowance for factors affecting comparability.” Id. “This means that
Commerce must at least consider the factors in the course of evaluating potential benchmark
sources.” RZBC Grp. Shareholding Co. Ltd. V. United States, 40 CIT __, __, 2016 WL 3880773
at *9 (2016) (quotation marks omitted). Further, Commerce will “adjust the comparison price to
reflect the price that a firm actually paid or would pay if it imported the product.” Id. §
351.511(a)(2)(iv). This requires that Commerce calculate the benchmark based on a hypothetical
importer of the given product. See Changzhou Trina Solar Energy Co., Ltd. v. United States, 42
CIT __, __, 352 F. Supp. 3d 1316, 1338 (2018) (“As the court has indicated, however, Commerce
has determined that benchmark calculations are assessed based on a hypothetical importer making
a market-price purchase . . . .” (citation omitted)). But the court has been hesitant to endorse
benchmarks that use simple averages that give “undue weight” to small samples. See RZBC Grp.
Shareholding Co., Ltd. v. United States, 39 CIT __, __, 100 F. Supp. 3d 1288, 1308–11 (2015)
(holding that a simple average between a quantity unknown dataset and a high-cost, low-quantity
dataset may have distorted the benchmark by giving undue weight to small shipments).
Commerce did not sufficiently comply with the court’s remand order. In the remand order,
the court expressed a concern that the Descartes data “appear[ed] to be sourced from limited
samples because many of the shipments use the same tariff codes and freight forwarder codes.”
Risen, 570 F. Supp. 3d at 1379. On remand, however, Commerce only responded that it could not
conclude that these codes “indicate a shipment method that an importer of the materials at issue Consol. Court No. 20-03912 Page 18
would be unlikely to use.” Remand Results at 15. The court’s concern was not about whether this
was a shipping method a Chinese company would use, but instead about whether the Descartes
data was a high-cost, low-quantity dataset that improperly ballooned the benchmark when
averaging. See Risen, 570 F. Supp. 3d at 1379; see also RZBC Grp. Shareholding, 100 F. Supp.
3d at 1308–11. Commerce did not fully consider the potential impact that a small sample size
could have when affecting the comparability of the Descartes and Xeneta datasets. See 19 C.F.R.
§ 351.511(a)(2)(ii). Although Commerce may calculate the benchmark based on a hypothetical
importer of a product, the source must still be appropriate, and until Commerce addresses whether
the tariff codes and freight forwarder codes indicate a small sample size, the court cannot sustain
Commerce’s benchmark calculation. Thus, Commerce did not comply with the court’s remand
instructions, and the determination is not supported by substantial evidence. The Xeneta data
seems to provide a broadly based average. If Commerce cannot supply a convincing reason as to
why the Descartes data improves accuracy, Commerce should not use it.
CONCLUSION
As to EBCP, the Remand Results set standards not appropriately applied to this particular
factual record. As to land value and ocean freight, Commerce did not properly consider the quality
of data sets it averaged. Defects in data are not cured by averaging with better data. For the
foregoing reasons, the court remands to Commerce for a determination consistent with this opinion
on all three issues. The remand shall be issued within 60 days hereof. Comments may be filed 30
days thereafter and any response 15 days thereafter.
/s/ Jane A. Restani Jane A. Restani. Judge
Dated: April 11, 202 New York, New York