Slip Op. 20-

               UNITED STATES COURT OF INTERNATIONAL TRADE

  CANADIAN SOLAR INC., ET AL.

                        Plaintiffs,

  SUMEC HARDWARE & TOOLS CO., LTD.,

                        Consolidated Plaintiff, and   Before: Jane A. Restani, Judge

  CHANGZHOU TRINA SOLAR ENERGY
  CO., LTD. ET AL.,                                   Consol. Court No. 18-00184

                        Consolidated Plaintiffs,
                 v.                                   PUBLIC VERSION

  UNITED STATES,

                        Defendant,

  SOLARWORLD AMERICAS, INC.,

                        Defendant-Intervenors.



                                      OPINION AND ORDER

                                                                 Dated: February 25, 2020

[Commerce’s Final Results in the Fourth Administrative Review of Commerce’s Countervailing
Duty Order on crystalline silicon photovoltaic cells from the People’s Republic of China are
partially sustained and partially remanded for reconsideration consistent with this opinion.]

        Jeffrey S. Grimson, Bryan P. Cenko, James C. Beaty, Jill A. Cramer, Kristin H. Mowry,
and Sarah M. Wyss, Mowry & Grimson, PLLC, of Washington D.C. for Plaintiffs, Canadian Solar
Inc., Canadian Solar International, Ltd., Canadian Solar Manufacturing (Luoyang) Inc., Canadian
Solar Manufacturing (Changshu) Inc., CSI Cells Co., Ltd., CSI Solar Power (China) Inc., CSI
Solartronics (Changshu) Co., Ltd., CSI Solar Technologies Inc., CSI Solar Manufacture Inc., CSI
New Energy Holding Co., Ltd., CSI-GCL Solar Manufacturing (YanCheng) Co., Ltd., Changshu
Tegu New Materials Technology Co., Ltd., Changshu Tlian Co., Ltd., Suzhou Sanysolar Materials
Technology Co., Ltd. and Canadian Solar (USA) Inc.

       Mark B. Lehnardt and Lindita V. Ciko Torza, Baker & Hostetler, LLP, of Washington D.C.
Consol. Court No. 18-00184                                                                 Page 2
Public Version
for Consolidated Plaintiff Sumec Hardware & Tools Co., Ltd.

       Robert A. Gosselink, Jarrod M. Goldfeder, and Jonathan M. Freed, and Kenneth N.
Hammer, Trade Pacific, PLLC, of Washington, D.C., for Consolidated Plaintiffs Changzhou Trina
Solar Energy Co., Ltd., Trina Solar (Changzhou) Science & Technology Co., Ltd., Changzhou
Trina Solar Yabang Energy Co., Ltd., Yancheng Trina Solar Energy Technology Co., Ltd., Turpan
Trina Solar Energy Co., Ltd., Hubei Trina Solar Energy Co., Ltd., and Changzhou Trina PV
Ribbon Materials Co., Ltd.

       Joseph H. Hunt, Jeanne E. Davidson, Tara K. Hogan, and Justin R. Miller, International
Trade Field Office, U.S. Department of Justice, of New York, NY for the Defendant. Of counsel
on the brief was Paul K. Keith, Office of Chief Counsel for Trade Enforcement and Compliance,
U.S. Department of Commerce, of Washington, D.C.

       Timothy C. Brightbill, Adam M. Teslik, Cynthia C. Galvez, Douglas C. Dreier, Enbar
Toledano, John A. Riggins, Laura El-Sabaawi, Maureen E. Thorson, Stephanie M. Bell, and
Stephen J. Obermeier, Wiley Rein, LLP, of Washington, D.C., for Defendant-Intervenor
SolarWorld Americas, Inc.

       Restani, Judge: This action is a challenge to the final determination made by the United

States Department of Commerce (“Commerce”) in the Fourth 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 (“PRC”) covering the period from January 1, 2015

to December 31, 2015.

       Plaintiffs and Consolidated Plaintiffs Changzhou Trina Solar Energy Co., Ltd., Trina Solar

(Changzhou) Science & Technology Co., Ltd., Changzhou Trina Solar Yabang Energy Co., Ltd.,

Yancheng Trina Solar Energy Technology Co., Ltd., Turpan Trina Solar Energy Co., Ltd., Hubei

Trina Solar Energy Co., Ltd., and Changzhou Trina PV Ribbon Materials Co., Ltd. (collectively,

“Trina”); Sumec Hardware & Tools Co., Ltd. (“Sumec”); 1 and Canadian Solar Inc., Canadian




1
  Sumec does not present its own arguments before the court, but rather adopts the arguments made
by Trina and Canadian Solar. See Rule 56.2 Mot. of Consol. Pl., Sumecht Hardware & Tools Co.,
LTD, for J. on the Agency R., ECF No. 39 (Feb. 8, 2019); Reply Br. of Consol. Pl. Sumecht
Hardware & Tools Co., LTDS, in Supp. of Rule 56.2 Mot. for J. on the Agency R., ECF No. 63
(July 30, 2019).
Consol. Court No. 18-00184                                                                 Page 3
Public Version
Solar International, Ltd., Canadian Solar Manufacturing (Luoyang) Inc., Canadian Solar

Manufacturing (Changshu) Inc., CSI Cells Co., Ltd., CSI Solar Power (China) Inc., CSI

Solartronics (Changshu) Co., Ltd., CSI Solar Technologies Inc., CSI Solar Manufacture Inc., CSI

New Energy Holding Co., Ltd., CSI-GCL Solar Manufacturing (Yancheng) Co., Ltd., Changshu

Tegu New Materials Technology Co., Ltd., Changshu Tlian Co., Ltd., Suzhou Sanysolar Materials

Technology Co., Ltd., and Canadian Solar (USA) Inc. (collectively, “Canadian Solar”) request that

the court hold aspects of Commerce’s final determination unsupported by substantial evidence or

otherwise not in accordance with law.

       The United States (“government”) and Defendant-Intervenor SolarWorld Americas, Inc.

(“SolarWorld”) ask that the court sustain Commerce’s Amended Final Results of its fourth

administrative review.

                                        BACKGROUND

       Commerce published a countervailing duty order on solar cells from the People’s Republic

of China (“PRC”) on December 7, 2012. See Crystalline Silicon Photovoltaic Cells, Whether or

Not Assembled Into Modules, From the People’s Republic of China: Countervailing Duty Order,

77 Fed. Reg. 73,017 (Dep’t Commerce Dec. 7, 2012). Commerce initiated its fourth administrative

review of this countervailing duty order, covering the period from January 1, 2015 to December

31, 2015. Canadian Solar and Trina were selected as mandatory respondents (“Respondents”) and

issued questionnaires along with the Government of the PRC (“GOC”). See Decision

Memorandum for Final Results of Countervailing Duty Administrative Review: Crystalline

Silicon Photovoltaic Cells, Whether or Not Assembled Into Modules, from the People’s Republic

of China; 2015, C-570-980 (Dep’t Commerce July 12, 2018) (“I & D Memo”). On January 10,

2018, Commerce published its preliminary results of the administrative review. Crystalline Silicon
Consol. Court No. 18-00184                                                                      Page 4
Public Version
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; 2015, 83 Fed. Reg. 1,235 (Dep’t Commerce Jan. 10, 2018) and accompanying

issues and decision memorandum (“Prelim. I & D Memo”). Commerce issued its final results on

July 23, 2018. Crystalline Silicon Photovoltaic Cells, Whether or Not Assembled Into Modules,

From the People’s Republic of China: Final Results of Countervailing Duty Administrative

Review; 2015, 83 Fed. Reg. 34,828 (Dep’t Commerce July 23, 2018), as amended by Crystalline

Silicon Photovoltaic Cells, Whether or Not Assembled Into Modules, From the People’s Republic

of China: Amended Final Results of Countervailing Duty Administrative Review; 2015, 83 Fed.

Reg. 54,566 (Dep’t Commerce Oct. 30, 2018) (“Amended Final Results”). Plaintiff and

Consolidated Plaintiffs challenge several aspects of the Amended Final Results.


                        JURISDICTION AND STANDARD OF REVIEW

          The court has jurisdiction pursuant to 28 U.S.C. § 1581(c) and 19 U.S.C. § 1516a(a)(2)

(2012). 2 The court will uphold Commerce’s results in a countervailing duty proceeding unless they

are “unsupported by substantial evidence on the record, or otherwise not in accordance with law[.]”

19 U.S.C. § 1516a(b)(1)(B)(i).

                                            DISCUSSION

    I.    Commerce Requests a Remand on Four Issues

          Commerce requests a remand on four of the issues before the court that are substantially

similar to the same issues presented in the Third Administrative Review of the order on the

merchandise at issue: the application of adverse inferences to facts otherwise available to find that




2
    All further citations to the United States Code will be to the 2012 edition unless otherwise noted.
Consol. Court No. 18-00184                                                                    Page 5
Public Version
respondents used the Export Buyer’s Credit Program, the determination that China’s provision of

aluminum extrusions is a specific subsidy, the decision to average two datasets from IHS

technology and U.N. Comtrade in calculating the benchmark for aluminum extrusions, and the

determination that China’s provision of electricity is a specific subsidy. See Def.’s Resp. in Opp’n.

To Pls.’ Mots. for J. Upon the Agency R., ECF No. 52, at 9–12 (June 13, 2019) (“Gov. Br.”).

Although Trina asks that the court grant the government’s request for a remand, Reply Br. of

Consol. Pls. Trina, ECF No. 61, at 3–4 (July 30, 2019) (“Trina Reply”), Canadian Solar argues

that the court should grant a remand only on the aluminum benchmark issue and should find that

it has enough information on the remaining issues to find Commerce’s decisions to be

“unsupported by substantial evidence and contrary to law.” Pls.’ Canadian Solar Reply Br. in Supp.

of Rule 56.2 Mot. for J. on the Agency R., ECF No. 59, at 1–3 (July 30, 2019) (“Canadian Solar

Reply”).

        The court concludes that in view of its opinions covering the Third Administrative Review

of certain photovoltaic cells from the PRC, Changzhou Trina Solar Energy Co., Ltd. v. United

States, 

(CIT 2018) (“Changzhou Remand I”) and, the following opinion on

the remand redetermination, Changzhou Trina Solar Energy Co., Ltd. v. United States, Slip Op.

19-137, 

(CIT Nov. 8, 2019) (“Changzhou Remand II”), remand here is

appropriate. The administrative records of the Third and Fourth Administrative reviews and the

government’s legal rationales are similar and thus the determination at hand suffers from

essentially the same deficiencies that the court has noted in these prior opinions. The court remands

with instructions for Commerce to consult these prior opinions and reevaluate its decisions on

these four issues accordingly.

 II.   Polysilicon Benchmark
Consol. Court No. 18-00184                                                                       Page 6
Public Version
       Canadian Solar argues that Commerce erred in refusing to accept its “‘arms-length

purchases’ of polysilicon as a ‘tier-one’ benchmark source.” Mem. of P. & A. In Supp. of Rule

56.2 Mot. for J. on the Agency R. of Pls. Canadian Solar, ECF No. 38, at 28 (Feb. 8, 2019)

(“Canadian Solar Br.”). It points to Commerce’s regulation, which evinces a preference for a

“market-determined price for the good or service resulting from actual transactions in the country

in question.” 

C.F.R. § 351.51l(a)(2)(i)). Commerce found, pursuant to facts

otherwise available, that “the GOC’s intervention in China’s solar grade polysilicon market leads

to significantly distorted prices for solar grade polysilicon in China,” and thus discounted Canadian

Solar’s proffered import data. I & D Memo at 35. Commerce did not, however, elaborate on how

the GOC’s participation in only [[     ]] of the general polysilicon market led to distorted import

prices of solar-grade polysilicon, such that resorting to “tier-two” world market price benchmarks

was appropriate. See Canadian Solar Br. at 29. The government argues that the GOC’s failure to

provide information specific to solar-grade polysilicon allowed Commerce to rely on facts

otherwise available––namely documents from the petition and an article from IHS Technology on

solar glass prices–– to determine that the solar grade polysilicon market is distorted and disregard

Canadian Solar’s import data. Gov. Br. at 25–29.

       The court concludes that the polysilicon issue in this case is substantially like the one in

the Third Administrative Review and for the same reasons requires remand. As noted in

Changzhou Remand II:

       Commerce “must explain the evidence which is available, and must offer a rational
       connection between the facts found and the choice made.” See Motor Vehicle Mfrs.
       Ass’n of U.S., Inc. v. State Farm Mut. Auto. Ins., 

, 42 (1983) (quotation
       and citation omitted). It has not done so here. As noted in the Preamble;
       Countervailing Duties; Final Rule, unless a “government provider constitutes a
       majority, or in certain circumstances, a substantial portion of the market,” the effect
       on the market will normally be minimal. 63 Fed. Reg. 65,348, 65,377 (Dep’t
       Commerce Nov. 25, 1998) (“Preamble”); see also, 

.
Consol. Court No. 18-00184                                                                    Page 7
Public Version


Changzhou Remand II, Slip. Op. 19-137 at 20, 

at *10. Likewise, here,

Commerce did not sufficiently explain how the GOC’s minimal participation in the general

polysilicon industry led to the price distortion of imported solar-grade polysilicon. It is unclear to

the court whether and how the documents cited by Commerce rationally support its finding.

Accordingly, the court remands this issue with instructions to consider the court’s decisions in

Changzhou Remand I and Changzhou Remand II. Commerce shall either use Canadian Solar’s

import data as a tier-one metric, or provide sufficient evidence demonstrating that the GOC’s

participation in the solar-grade polysilicon market renders this import data unreliable.

III.   Exclusion of Xeneta Data in the Calculation of International Freight

       After relying on an average of Xeneta and Maersk data3 in computing the ocean freight

benchmark in its preliminary determination, Commerce relied solely on the Maersk data in the

final determination. See I & D Memo at 41. Commerce explained that the Xeneta data

inconsistently included or excluded terminal handling charges, which Commerce requires to be

included under 19 C.F.R. § 351.511(a)(2)(iv) as charges reflective of “the price that a firm actually

paid or would pay if it imported the product.” I & D Memo at 41–42; see also 19 C.F.R. §

351.511(a)(2)(iv).

       Trina and Canadian Solar argue that although Xeneta’s default methodology does not

include both origin terminal handling charges (“OTHC”) and destination terminal handling

charges (“DTHC”) (collectively “THCs”), the Xeneta data submitted by Canadian Solar [[




3
 Xeneta collects freight rate data from “short to long term rate-agreements,” see Exhibits 7B &
7C to Letter from Mowry & Grimson PLLC to Sec’y of Commerce Pertaining to Canadian Solar
Benchmark Submission, C.R. 104-146 (Nov. 1, 2017) (“Canadian Solar Xeneta Data”), while
Maersk publishes freight rate quotes. See I & D Memo, at 40–41.
Consol. Court No. 18-00184                                                                   Page 8
Public Version
                        ]]. Mem. in Supp. of Mot. of Pls. Trina for J. Upon the Agency R., ECF.

No. 41-2, at 32, 34 (Feb.11, 2019) (“Trina Br.”); Canadian Solar Br. at 30. Trina claims this

inclusion is evident 4 because both Trina and Canadian Solar submitted Xeneta data, but Trina’s

submission failed to include [[                          ]]. Trina Br. at 34–37. By comparing the

two sets, Trina claims that Commerce should have realized that the Canadian Solar data included

all necessary charges. Trina Br. at 37. Trina claims that because Commerce’s decision to exclude

the Xeneta data results from its misreading the data as not including the THCs, its decision is

unsupported by substantial evidence. Trina Br. at 38. The government argues that “there is no

record information explaining whether the terminal handling charge methodology described by

Xeneta is only a [[                 ]], and that there is no record information describing the [[

                                    ]].” Gov. Br. at 30. In the absence of information explaining

whether the Xeneta data could be sorted to include all relevant charges, the government argues

that Commerce’s decision to exclude the Xeneta data was reasonable and supported by substantial

evidence. Gov. Br. at 30–31. SolarWorld additionally argues that it is not clear that the Canadian

Solar submission of Xeneta includes all relevant charges. SolarWorld Resp. to Pl.s’ Mot. for J.

Upon the Agency R., ECF No. 50, at 34–35 (July 10, 2019) (“SolarWorld Br.”).

       The court finds it necessary to remand for Commerce to reconsider whether the Xeneta

data submitted by Canadian Solar properly include all terminal handling charges such that it should

be averaged with the Maersk data in accordance with 19 C.F.R. § 351.511(a)(2)(iv). The court’s

review of the record shows that although Trina’s submission of Xeneta data did not include all

terminal handling charges, Canadian Solar’s submission [[




4
 Trina explains that results as to whether both THCs are included is “merely a matter of how the
user queries the database.” Trina Br. at 36.
Consol. Court No. 18-00184                                                                 Page 9
Public Version
                          ]]. This is most readily apparent from the screenshots of the relevant

graphs submitted by the parties. Further, Trina notes that Canadian Solar’s monthly breakdown of

ocean freight calculations from Xeneta from various ports is consistently markedly higher than the

rates submitted by Trina. See Trina Br. at 37; Compare Canadian Solar Xeneta Data with Letter

from Trade Pacific PLLC to Secretary of Commerce Pertaining to Trina Benchmark Submission,

C.R. 147–164, Exhibit 9.1 (Nov. 1, 2017). Although not definitive, this appears to indicate that

Canadian Solar’s submission included costs not included by Trina.

       On remand, Commerce should evaluate whether its decision to exclude Canadian Solar’s

Xeneta submission was based on a misreading of the evidence. If Commerce remains uncertain or

desires further clarification, it may seek additional information from respondents.

IV.    Imputing an Electricity Benefit to Respondents after April 20, 2015

       Commerce found that the GOC failed to provide adequate information regarding the

provision of electricity in the PRC. I & D Memo at 14–15, 33. Although Commerce acknowledged

the GOC’s claims that following April 20, 2015, the National Development and Reform

Commission (“NDRC”) was no longer involved in provincial price adjustments, it found that other

record information contradicted the GOC’s assertions. 

       Canadian Solar argues that record evidence supports that the NDRC has delegated its

authority to provincial government agencies after April 20, 2015. Canadian Solar Br. at 35–37.

The government avers that although the GOC claims that the NDRC delegated its authority to the

provinces, Commerce found that other evidence, specifically the NDRC Notices 748 and 2909,
Consol. Court No. 18-00184                                                                 Page 10
Public Version
undermined this claim, and so continued to find the provision of electricity to be a countervailable

subsidy after that date. Gov. Br. at 22–23. 5

         Because Commerce requests a remand on the electricity issue, the court need not address

this issue at this juncture. With that said, if the NDRC Notices state what Commerce claims, its

decision to find electricity countervailable after April 20, 2015, may be based on substantial

evidence. Commerce should be sure to put all relevant documents, such as the NDRC Notices,

on the record following remand as these documents do not appear to be in the submitted

appendices.

    V.   Translation Error in the Calculation of Canadian Solar’s Electricity Benefit

         After Commerce issued the preliminary results, Canadian Solar asserted that one of the

electricity schedules it submitted contained a mistranslation of “sharp” as “peak.” I & D Memo at

33–34. Commerce did not alter its benchmarks based on the schedules as originally understood

stating that “it is the responsibility of the respondents, and not Commerce, to ensure the documents

submitted on the record are translated fully and accurately.” I & D Memo at 34.

         Canadian Solar argues that Commerce’s stance “contravenes Commerce’s duty to calculate

a benefit ‘as accurately as possible.’” Canadian Solar Br. at 38 (quoting Borusan Mannesmann

Boru Sanayi ve Ticaret A.S. v. United States, 

, 1337 (CIT 2015)). It claims

that as it alerted Commerce soon after the preliminary results and because the agency had

everything it needed on the record to correct the “plainly evident” error, Commerce should have

done so. 

6 The government responds that the error was not plainly evident and that it




5
  The court’s understanding is that Commerce has not based its subsidy decision on the Provinces
as the governmental authority providing the countervailable benefit.
6
  Canadian Solar does not describe the resulting practical consequence of this alleged mistaken
translation.
Consol. Court No. 18-00184                                                                 Page 11
Public Version
was Canadian Solar’s duty to develop the record. Gov. Br. at 24–25. Commerce also asserts that

the “record documents do not indicate which translation of the headings is correct, the nature of

the translation error, or whether multiple possible translations exist.” 

       It is unclear to the court whether Canadian Solar is correct regarding the translation of the

schedules. Regardless, the court holds that Commerce erred in not considering whether the

translation was accurate when timely raised by Canadian Solar following the preliminary results.

Typically, Commerce is not required to correct alleged errors first identified after the issuance of

the final results, see Alloy Piping Prods., Inc. v. Kanzen Tetsu Sdn. Bhd., 

, 1293

(Fed. Cir. 2003) (noting that where “(1) the error was made by the respondent; (2) no request to

correct the error was made before the final determination; and (3) there was no showing that the

error was apparent to Commerce (or should have been apparent) from the record or the final

determination itself, Commerce was not required to correct the alleged error[.]”). In those

situations, “the requirement of administrative finality necessarily outweighed [respondent’s]

belated concern for correctness.” Chengde Malleable Iron Gen. Factory v. United States, 505 F.

Supp. 2d 1367, 1374 (CIT 2007). But the same concerns do not exist when a party seeks correction

of errors following the preliminary results. See Timken U.S. Corp. v. United States, 

,

1353 (Fed. Cir. 2006) (holding that the court “however, has never discouraged the correction of

errors at the preliminary result stage; we have only balanced the desire for accuracy in antidumping

duty determinations with the need for finality at the final results stage.”).

       The court has previously addressed the issue of translation errors raised after final results

are issued and refused to remand on that ground. See Bridgestone Am., Inc. v. United States, 

, 1367 (CIT 2010). But here, Canadian Solar raised this issue with ample time for

Commerce to correct that alleged error before Commerce issued its final results. Although a
Consol. Court No. 18-00184                                                                  Page 12
Public Version
prudent respondent would ensure that documents were translated properly before submission,

Commerce cannot decide to ignore potential mistakes simply on this basis, at least not in a situation

where the issue is easily rectified. To do so results in Commerce failing to assess a countervailing

duty rate as accurately as reasonably possible.

       The court does not opine on whether Canadian Solar’s arguments regarding translation are

correct, but simply remands for Commerce to consider whether that is the case, and to recalculate

the resulting electricity benchmark if necessary.

VI.    Incorrect Denominator in the Calculation of the GSDP Benefit

       Canadian Solar claims that Commerce should have combined the sales value of all of

Canadian Solar’s affiliates, including its cross-owned input suppliers, 7 in computing the

denominator for the Golden Sun Demonstration Program (“GSDP”). 8 Canadian Solar Br. at 39–

40. Canadian Solar argues that in previous reviews Commerce has used total sales but has since

changed its methodology without explanation. 

       The government first argues that Canadian Solar does not present a case or controversy

because even if Commerce alters the denominator as requested, the resulting subsidy rate would

be rounded to the same .58 percent. Gov. Br. at 32–33. 9 Even if the court does not dismiss the

claim for a lack of jurisdiction, the government argues that “Commerce’s calculations reflect a



7
  Commerce combined only some of the sales from Canadian Solar’s affiliates resulting in a
denominator of [[             ]] when that number, according to Canadian Solar should have
been [[           ]]. Canadian Solar Br. at 39.
8
  The benefit received is the “numerator” while the sales value is the “denominator” in calculating
a subsidy rate. Canadian Solar does not challenge the composition of the numerator.
9
  The current subsidy rate is .5843 percent, which Commerce rounds to .58 percent, whereas if
Commerce adjusted the denominator it would be .5769, which the government claims it would still
round to .58 percent. See Gov. Br. at 32.
Consol. Court No. 18-00184                                                                     Page 13
Public Version
straightforward application of its regulations,” as Commerce found that the input producers at issue

were cross-owned by Canadian Solar. 

(citing 19 C.F.R. § 351.525(b)(6)). Canadian Solar

argues that it presents a justiciable case given the potential for change at the three-digit decimal

level of the subsidy rate. See Canadian Solar Reply at 21 n.4. It further argues that the GSDP, as

an untied subsidy, “benefited all production related to photovoltaic power generation,” so all of

the sales of the input suppliers should also be found to have benefited from the program. 

       Assuming the government is correct that Commerce would round the countervailing duty

rate for GSDP to .58 percent regardless of whether input suppliers are included in the denominator,

then the government has a tenable argument that Canadian Solar fails to present a case or

controversy as no action by this court would redress the claimed injury. See e.g., Lujan v. Defs. of

Wildlife, 

, 561–62 (1992). The government, however, cites to no authoritative source

showing that this rounding would definitively occur and as the three-digit decimal rate would

change, the court exercises jurisdiction.

       Although Canadian Solar claims that this issue was addressed previously, the court did not

meaningfully consider the issue. Cf. Changzhou Remand 

(noting that

“an untied subsidy will be attributed to the firm’s total sales”). The court does not find that decision

dispositive of the more specific issue of whether the denominator used in the subsidy calculation

for Canadian Solar must include all the sales of the cross-owned input suppliers. That question is

answered by the regulations.

       Canadian Solar does not argue that the input suppliers at issue were not cross-owned by

the producers of the subject merchandise or that the input product is not primarily dedicated to the

production of the downstream product. The applicable regulations clearly state:

       Input suppliers. If there is cross-ownership between an input supplier and a
       downstream producer, and production of the input product is primarily dedicated
Consol. Court No. 18-00184                                                                 Page 14
Public Version
       to production of the downstream product, the Secretary will attribute subsidies
       received by the input producer to the combined sales of the input and downstream
       products produced by both corporations (excluding the sales between the two
       corporations).

19 C.F.R. § 351.525(6)(iv). In other words, Commerce disregarded the sales between the cross-

owned input suppliers and the producers in arriving at the denominator. Accordingly, Commerce

followed its regulations in excluding from the denominator certain sales made by cross-owned

input suppliers. Canadian Solar has not demonstrated that the regulation is in conflict with the

statute. Canadian Solar’s arguments that would require Commerce to deviate from a valid and

applicable regulation are unavailing.

VII.   Commerce’s Refusal to “Offset”

       In its final determination, Commerce did not assign a value to countervailable input

purchases (aluminum, solar glass, polysilicon, loans, and electricity) made by Canadian Solar at

or above the benchmark world market price. See I & D Memo at 45–46. Commerce did so because

“a positive benefit from certain transactions cannot be masked by ‘negative benefits’ from other

transactions.’” 

Commerce claimed that offsetting purchases made at less than adequate

remuneration (“LTAR”) with those made above is neither contemplated by the statute nor a

department practice. 

claims that this practice of “zeroing” 10 is not authorized by statute or

regulation. Canadian Solar Br. at 40–42. It claims this results in an over-collection of



10
   Canadian Solar continually refers to Commerce’s practice as “zeroing,” which is not a term
frequently used in the countervailing duty context. It is, however, often referred to in antidumping
proceedings as the practice where in “calculating the weighted-average dumping margin,
Commerce treats transactions that generate ‘negative’ dumping margins (i.e., a dumping margin
with a value less than zero) as if they were zero.” See Timken Co. v. United States, 

,
1338 (Fed. Cir. 2004). The court does not equate Canadian Solar’s claim of “zeroing” with that
practice.
Consol. Court No. 18-00184                                                                    Page 15
Public Version
countervailing duties as “Commerce’s calculation failed to reflect the actual market conditions.”

The government and SolarWorld argue that a benefit is conferred so long as respondent

purchases an input at a rate lower than would be available without the subsidy. Gov. Br. at 35–37;

SolarWorld Br. at 42–44. They assert that the statute allows for offsets in a narrow range of

circumstances not at issue here. See Gov. Br. at 36 (citing 19 U.S.C. § 1677(6) 11 and Kajaria Iron

Castings Pvt. Ltd. v. United States, 

, 1174 (Fed. Cir. 1998) (holding that 19 U.S.C.

§ 1677(6) “provides the exclusive list of permissible offsets”)); see also SolarWorld Br. at 43

(same). Canadian Solar responds that the government misunderstands its claim, and that it is not

asking for an offset, but is rather asking Commerce to consider “prevailing market conditions,” in

the form of accounting for its purchases of inputs made at or above LTAR. Canadian Solar Reply

at 21–23 (quoting 19 U.S.C. § 1677(5)(E)(iv)).

          Commerce’s method assumes that countervailing duties are not assessed on purchases at

or above the world market rate. The government is correct that Canadian Solar functionally asks

for the offsetting of its LTAR purchases with its at or above LTAR purchases. The government’s




11
     The statute defines “net countervailable subsidy” as:

          For the purpose of determining the net countervailable subsidy, the administering
          authority may subtract from the gross countervailable subsidy the amount of—

          (A) any application fee, deposit, or similar payment paid in order to qualify for, or
              to receive, the benefit of the countervailable subsidy,

          (B) any loss in the value of the countervailable subsidy resulting from its deferred
              receipt, if the deferral is mandated by Government order, and

          (C) export taxes, duties, or other charges levied on the export of merchandise to the
          United States specifically intended to offset the countervailable subsidy received.

19 U.S.C. § 1677(6).
 Consol. Court No. 18-00184                                                                      Page 16
 Public Version
 practice of not calculating a “negative benefit” is in accordance with the statute as a respondent

 still receives a countervailable benefit in situations where only some of its inputs were provided

 for less than adequate remuneration. See 19 U.S.C § 1677(5)(e)(iv). By countervailing only those

 purchases made below LTAR, Commerce is simply effectuating its statutory mandate. At base,

 Canadian Solar has benefitted from receiving reduced-cost inputs and Commerce properly

 countervailed those benefits, regardless of whether Canadian Solar also purchased some inputs at

 or above LTAR. A countervailable subsidy remains countervailable regardless of the extent of use

 by a respondent.

VIII.    Entered Value Adjustment on Canadian Solar Entries

         Commerce denied Canadian Solar an entered value adjustment (“EVA”) 12 to Canadian

 Solar’s subsidy rate. 13 I & D Memo at 47–48. Commerce stated that its practice is to grant an

 EVA only when six criteria are satisfied:

         1) the price on which the alleged subsidy is based differs from the U.S.
         invoiced price; 2) the exporters and the party that invoices the customer are
         affiliated; 3) the U.S. invoice establishes the customs value to which the CVD
         duties are applied; 4) there is a one-to-one correlation between the invoice that
         reflects the price on which subsidies are received and the invoice with the mark-up
         that accompanies the shipment; 5) the merchandise is shipped directly to the United
         States; and 6) the invoices can be tracked as back-to-back invoices that are
         identical except for price.



 12
   This is sometimes referred to as an “export value adjustment.” See I & D Memo at 47 n.258.
 The court gave a detailed explanation of EVA and Commerce’s practice of granting one in Jiangsu.
 See Jiangsu Zhongji Lamination Materials Co., Ltd. v. United States, 

, 1326–
 1331 (CIT 2019). At base, Commerce grants an EVA, which results in an adjusted subsidy rate,
 when a party can show that the entered value of the merchandise at issue is higher than the
 previously assessed sales value on which the subsidy rate was based. 

Without an
 EVA in this situation, Commerce potentially would collect excess countervailing duties. 

 13
    Commerce normally computes a subsidy rate by dividing the subsidy benefit (the numerator) by
 the sales value of a product on an f.o.b. (port) basis if the product is at issue is at the export stage.
 (the denominator). See 19 C.F.R. § 351.525(a).
Consol. Court No. 18-00184                                                                    Page 17
Public Version

(citations omitted). Here, after finding that respondent failed to demonstrate that

all sales made via its affiliate were for a higher customs value than the sales value initially

used to calculate the subsidy rate, Commerce found that Canadian Solar failed to meet the

fourth criterion. See 

–48; Canadian Solar Final Calculation Mem., C-570-980 at 3

(Dep’t Commerce July 12, 2018) (“Calc. Mem.”).

       Canadian Solar claims to have satisfied these criteria. Canadian Solar Br. at 42–45.

Canadian Solar argues that it does not need to show that all of its sales to made through its affiliate

[[ ]] were for a higher customs value to merit an EVA, rather, it needed only show that “the total

sales represented a higher overall customs value.” 

It argues that Commerce’s decision

results in an overcollection of countervailing duties. 

       The government argues that Commerce’s decision not to make an EVA is correct. Gov. Br.

at 38–40. Both the government and SolarWorld argue that Canadian Solar ignores that Commerce

will grant an EVA only if “there is a one-to-one correlation” between the invoice reflecting the

received subsidies and the invoice with the mark-up, which requires the respondent demonstrate

that all sales had a higher customs value. Gov. Br. at 40; SolarWorld Br. at 44. SolarWorld also

notes that “Canadian Solar submitted documentation for only a single sales transaction to support

its claim for an EVA.” SolarWorld Br. at 45. Additionally, SolarWorld claims that this single

sample transaction evinces a manipulation of the sales value in an attempt to lower the subsidy

margin. 

briefing, Canadian Solar argues that “there is no reason for Commerce to

require that all sales have a markup in order to qualify for an EVA,” Suppl. Br. Per Ct. Order of

Pls. Canadian Solar Related to Entered Value Adjustment, ECF No. 80, at 2 (Dec. 9, 2019)

(“Canadian Solar Suppl. Br.”), and that Commerce’s failure to grant an EVA in this instance results
Consol. Court No. 18-00184                                                                      Page 18
Public Version
in an overcollection of countervailing duties. 

The government states that it requires the

adjustment to be applied to all sales to serve the “evidentiary purpose of demonstrating that the

overall entered value is, with certainty, higher than the overall sales value,” as the record may not

always contain the entered value information for all sales. Def.’s Resps. to the Ct.’s Questions

Issued Nov. 13, 2019, ECF No. 81, at 2–3 (Dec. 9, 2019) (“Def. Suppl. Br.”). Further, the

government argues that the practice of allowing an EVA only when all sales enter with a mark-up

helps avoid potential manipulation in which a respondent could “selectively request an entered

value adjustment for marked-up sales while Commerce would be unaware that these markups

should have been offset by other marked-down sales.” 

SolarWorld argues that an EVA is

based on self-reported “untested, third-party transfer prices” and so is “largely beyond

corroboration.” Solarworld Supp. Br. Regarding Export Value Adjustment, ECF No. 82, at 2 (Dec.

9, 2019) (“SolarWorld Supp. Br.”). SolarWorld further avers that Canadian Solar failed to offer

sufficient evidence proving its entitlement to an EVA. 

Finally, SolarWorld posits that

given the prospective nature of CVD rates, respondents could strategically mark up sales in one

year “for the purpose of engineering a lower CVD rate and cash deposit rate” in the following year.

SolarWorld Suppl. Br. at 6. In rebuttal, Canadian Solar dismisses the potential manipulation

scenarios noted by SolarWorld and the government as either farfetched or failing to have a

distortive effect. Rebuttal Cmts. Related to Suppl. Br. on Entered Value Adjustment, ECF No. 84,

at 6–7 (Dec. 23, 2019) (“Canadian Solar Reb.”). The government responds that Canadian Solar

has failed to demonstrate that the higher aggregate value of its sales to its affiliate [[   ]] is due to

a higher aggregate value of U.S. sales. Def.’s Resp. to Canadian Solar’s Answers to the Ct.’s

Questions Issued Nov. 13, 2019, ECF No. 86, at 2–3 (Dec. 23, 2019) (“Def. Reb.”). Accordingly,

it is possible that “the overall entered value of Canadian Solar’s U.S. sales could have been marked
Consol. Court No. 18-00184                                                                  Page 19
Public Version
down, but when combined with marked up sales to other markets the aggregate sales value could

still show a markup.” 

       Commerce makes an adjustment for the calculated subsidy rate when the “sales value used

to calculate that subsidy rate does not match the entered value,” and where a respondent satisfies

the six criteria noted above. Countervailing Duty Investigation of Certain Uncoated Paper from

Indonesia: Issues and Decision Memorandum for the Final Affirmative Determination at 12, C-

560-829, POI 1/1/2014–12/31/2014 (Dep’t Commerce Jan. 8, 2016). An appropriately granted

EVA ensures that the countervailing duties imposed at entry accurately correspond to the

calculated subsidy for the product at issue. As described in Ball Bearings from Thailand:

       [W]e have divided the F.O.B. value of the exports of the subject merchandise before
       the inter-company transaction (before mark-up) by the value of the same
       merchandise as it entered the United States (after mark-up). This results in a ratio,
       which reflects the difference in the export and import values. We then multiplied
       this ratio by the subsidy rate to obtain an adjusted rate of 8.51 percent ad valorem.

Ball Bearings and Parts Thereof from Thailand; Final Results of Countervailing Duty

Administrative Review, 57 Fed. Reg. 26,646, 26,647 (Dep’t Commerce June 15, 1992). What

appears to matter in this calculation is the total value of the exports before the mark-up and the

total value after. If the total value of sales after mark-up is greater than the total value of

merchandise before mark-up, then a failure to grant an EVA could lead to an overcollection of

countervailing duties. See 

. (noting that over-collection of duties

occurs when Customs imposes a duty based on a subsidy rate that did not account for a mark-up

reflected on the invoice used by Customs to calculate the duty owed).

       In its response to Canadian Solar’s supplemental brief, however, the government dismisses

overcollection concerns in this case because Canadian Solar has failed to demonstrate that the

higher aggregate value of sales by its affiliate is caused by a higher aggregate value of U.S. sales.
Consol. Court No. 18-00184                                                                   Page 20
Public Version
Def. Reb. at 2–3. Accordingly, it is possible that the entered value of merchandise to the U.S. was

not higher than the total value of that merchandise assessed prior to any mark-up. Although

Canadian Solar asks for an EVA on all of its sales through its affiliate [[ ]], it has not proved that

in the aggregate all sales entered into the United States include a mark-up. See Calc Mem. at

Attach. II. 14 Commerce is understandably focused on limiting an EVA to the mark-up experienced

on U.S. entries as an EVA is granted to avoid overcollection, which has the potential to occur only

when Customs assesses duties on merchandise that enters the United States. Without record

evidence demonstrating that U.S. entries experienced a mark-up that was not accounted for, the

court can give no credence to Canadian Solar’s claims of overcollection.

           Further, Canadian Solar submitted only a single invoice and corresponding customs form,

which demonstrates a lower percentage mark-up than the percentage based on the raw aggregate

amounts. Calc. Mem. at 2. Even if Canadian Solar meets the remaining criteria, this cannot satisfy

Commerce’s requirement that there be “a one-to-one correlation between the invoice that reflects

the price on which subsidies are received and the invoice with the mark-up that accompanies the

shipment.” See I & D Memo at 47–48. Canadian Solar may well be correct that a sample invoice

could suffice to demonstrate this “one-to-one” requirement, but here Canadian Solar seeks a mark-

up that is not demonstrated to be applied consistently. Thus, logically a single sample does not

suffice.

           The court leaves for another day the validity of Commerce’s practice to grant an EVA only

where every U.S. sale is marked up, rather than where U.S. sales are marked up in the aggregate.

Based on what has been presented in this case, the court acknowledges that granting an EVA



14
   Although Canadian Solar claims that all sales to the United States were made via [[ ]] the
converse, that all sales made via [[ ]] went to the United States, does not appear to be true. See

II.
Consol. Court No. 18-00184                                                                    Page 21
Public Version
appears feasible only when a mark-up is added consistently and as a matter of course, given

verification concerns. See Def. Suppl. Br. at 2–3; see also Ball Bearings from Thailand, 57 Fed.

Reg. at 26,647 (granting an EVA, in part, because respondent in that case has an accounting system

“set up to track the mark-up for each individual shipment of bearings via back-to-back invoices

that are identical except for price.”). The court, however, does not discount the possibility that

respondents could find a way to demonstrate the appropriateness of an EVA in circumstances

where not every U.S. sale has a mark-up, consistent or otherwise, but where the respondent can

show, nonetheless, that the entered value of sales to the U.S. is in the aggregate higher to a specific

degree than the value on which Customs assessed dutiable value. But that is not the case here. The

court concludes that Commerce’s decision to deny an EVA to Canadian Solar is supported by

substantial evidence and otherwise in accordance with law.

                                          CONCLUSION

       The court sustains Commerce’s decision regarding the calculation of Canadian Solar’s

GSDP benefit, the refusal to offset certain purchases made by Canadian Solar at or above the

benchmark world market price, and the denial of Canadian Solar’s request for an EVA. For the

foregoing reasons, the court remands to Commerce for a determination consistent with this opinion

on the remaining issues. Remand results should be filed by April 27, 2020. Objections are due

May 27, 2020 and Responses to Objections are due June 11, 2020.



                                                               _/s/Jane A. Restani______
                                                               Jane A. Restani, Judge

Dated: February 25, 2020
       New York, New York