Slip Op. 26-
UNITED STATES COURT OF INTERNATIONAL TRADE
TOYO KOHAN CO., LTD.,
Plaintiff, Before: Jane A. Restani, Judge v. Court No. 24-00261 UNITED STATES, Public Version Defendant,
THOMAS STEEL STRIP CORPORATION,
Defendant-Intervenor.
OPINION AND ORDER
Dated: May 2, 2026
[Sustaining the United States Department of Commerce’s final remand redetermination in its antidumping duty order review of diffusion-annealed, nickel-plated flat-rolled steel products from Japan.]
Daniel Lewis Porter, Pillsbury Winthrop Shaw Pittman LLP, of Washington, DC, for the plaintiff, Toyo Kohan Co., Ltd. Also on the brief were Gina Marie Colarusso, James Philip Durling, John Taishu Pitt, and William Charles Sjoberg.
Tara Kathleen Hogan, Lead Attorney, Commercial Litigation Branch, Civil Division, U.S. Department of Justice, of Washington, DC, for the defendant, the United States. Of counsel on the brief were Charlie Chung and William Mitchell Purdy, Office of Chief Counsel for Trade Enforcement and Compliance, U.S. Department of Commerce, of Washington, DC.
James R. Cannon, Jr., Cassidy Levy Kent (USA) LLP, of Washington, DC, for the defendant- intervenor, Thomas Steel Strip Corporation. Also on the brief were Nicole Brunda and Ulrika Kristin Skitarelic Swanson.
Restani, Judge: Before the court is the United States Department of Commerce’s
(“Commerce”) final remand redetermination pursuant to the court’s remand order in Toyo Kohan
Co. v. United States, 2025 WL 2986385 (CIT Oct. 23, 2025) (“Toyo Kohan I”). See Final Results Court No. 24-00261 Page 2 Public Version
of Redetermination Pursuant to Court Remand, ECF No. 65-1 (Feb. 18, 2026) (“Remand Results”).
In Toyo Kohan I, the court considered Commerce’s final results of its administrative review of
diffusion-annealed, nickel-plated flat-rolled steel products from Japan, covering period of review
(“POR”) May 1, 2022, through April 30, 2023. See Diffusion-Annealed, Nickel-Plated Flat-
Rolled Steel Products From Japan: Final Results of Antidumping Duty Administrative Review,
2022-2023, 89 Fed. Reg. 95,735 (Dep’t Commerce Dec. 3, 2024) (“Final Results”). In its Final
Results, Commerce used the earlier of the commercial invoice date or the shipment date—which,
in this case, was the shipment date—as the date of sale for Toyo Kohan Co., Ltd.’s (“Toyo Kohan”)
U.S. sales.1 See Final Results; Issues and Decision Memorandum for the Final Results of the
2022-2023 Administrative Review of the Antidumping Duty Order on Diffusion-Annealed,
Nickel-Plated Flat-Rolled Steel Products From Japan at 4, P.R. 108 (Nov. 25, 2024) (“IDM”);
Remand Results at 1 n.3. Commerce also applied its differential pricing analysis to Toyo Kohan’s
U.S. sales, the first step of which was the Cohen’s d test.2 Decision Memorandum for the
Preliminary Results of the 2022-2023 Administrative Review of the Antidumping Duty Order on
Diffusion-Annealed, Nickel-Plated Flat-Rolled Steel Products from Japan at 5–6, P.R. 85 (May
16, 2024) (“PDM”).
1 Commerce’s practice is to rely on the earlier of the shipment or invoice date as the date of sale because “where shipment date precedes invoice date, the shipment date better reflects the date when the material terms of sale are established.” Remand Results at 5. Commerce uses its variable “SALEDATU” to refer to the earlier of the shipment or invoice dates. Id. at 1 n.3. 2 The Cohen’s d test is a statistical measure of the extent of the difference between the mean of a test group and the mean of a comparison group. Decision Memorandum for the Preliminary Results of the 2022-2023 Administrative Review of the Antidumping Duty Order on Diffusion- Annealed, Nickel-Plated Flat-Rolled Steel Products from Japan at 5, P.R. 85 (Dep’t Commerce May 16, 2024) (“PDM”). Court No. 24-00261 Page 3 Public Version
The court remanded the final determination to Commerce to reevaluate its date of sale
determination and to reperform its differential pricing analysis. See Toyo Kohan I at *7. For the
following reasons, the court sustains Commerce’s Remand Results.
BACKGROUND
The court presumes familiarity with the facts of this case as set out in its previous opinion
ordering remand to Commerce, see Toyo Kohan I, and recounts only those facts relevant to the
court’s review of the Remand Results. On July 12, 2023, Commerce initiated its administrative
review of the antidumping (“AD”) duty order covering nickel-plated steel products from Japan for
POR from May 1, 2022, through April 30, 2023. See Initiation of Antidumping and Countervailing
Duty Administrative Reviews, 88 Fed. Reg. 44,262 (Dep’t Commerce July 12, 2023). Commerce
selected Toyo Kohan as the sole mandatory respondent in this review. See Respondent Selection
Memorandum at 1–2, P.R. 16 (Aug. 18, 2023); IDM at 1. On December 3, 2024, Commerce issued
its Final Results. See Final Results. Commerce calculated a final weighted-average dumping
margin of 4.44 percent. Id. at 95,736.
On October 23, 2025, the court remanded the Final Results to Commerce. Toyo Kohan I
at *7. First, the court held that Commerce’s use of the shipment date as the date of sale was
unsupported by substantial evidence because the record evidence suggested that Toyo Kohan’s
billing documentation at the time of shipping is “virtually meaningless, as it does not necessarily
reflect the quantity or price in the purchase order or the final invoice.” Id. at *5. Accordingly, the
court remanded for Commerce “to reconsider both the date of sale and the sales price, recognizing
that when there is an agreed-to pricing formula, there may be a disconnect in the record data for
the two matters.” Id. Court No. 24-00261 Page 4 Public Version
Second, the court remanded to Commerce to reperform its differential pricing analysis
consistent with the Federal Circuit’s holding in Marmen Inc. v. United States, 134 F.4th 1334 (Fed.
Cir. 2025) (“Marmen”) that Commerce may not apply Cohen’s d when the underlying data is not
normally distributed, equally variable, and equally and sufficiently numerous. Toyo Kohan I at
*7; see Marmen at 1348. The court noted that, although Toyo Kohan had failed to challenge
Commerce’s application of Cohen’s d in the underlying administrative proceedings, “any such
efforts would have been futile” because the Federal Circuit had not yet resolved the issue of how,
if at all, Cohen’s d could be applied, while Commerce had “adhered to the methodology faithfully
. . . after several opportunities to reconsider the matter.” Toyo Kohan I at *7. Accordingly, the
court remanded to Commerce to reperform its differential pricing analysis consistent with
Marmen. Id.
On February 18, 2026, Commerce filed its Remand Results. See Remand Results. First,
Commerce continued to find that Toyo Kohan’s reported shipment date is the appropriate date of
sale for its U.S. sales. Id. at 4. Second, Commerce reconsidered and discontinued its application
of the Cohen’s d test and, alternatively, applied the “price difference test.” Id. at 8. Toyo Kohan’s
estimated weighted-average dumping margin increased from 4.44 percent to 4.58 percent. Id. at
2. On March 20, 2026, Toyo Kohan filed its comments on the Remand Results. Pl.’s Conf.
Comments on Commerce’s Remand Results, ECF No. 68 (Mar. 20, 2026) (“Pl. Cmts.”). The
government and defendant-intervenor Thomas Steel Strip Corporation (“Thomas Steel”) filed their
reply comments on the Remand Results. Gov’t’s Reply to Comments on Remand Results, ECF
No. 72 (Apr. 20, 2026) (“Gov’t Reply”); Thomas Steel Strip Corp.’s Conf. Reply to Comments on
Remand Results at 5, ECF No. 74 (Apr. 21, 2026) (“Thomas Steel Reply”). Court No. 24-00261 Page 5 Public Version
JURISDICTION AND STANDARD OF REVIEW
The court has jurisdiction pursuant to 28 U.S.C. § 1581(c) and section 516A of the Tariff
Act of 1930, codified as amended, 19 U.S.C. § 1516a(2). The court sustains Commerce’s results
in an antidumping duty investigation 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); see Fujitsu Gen.
Ltd. v. United States, 88 F.3d 1034, 1038 (Fed. Cir. 1996). “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, 219 F. Supp. 3d 1300, 1307 (CIT 2017) (citation modified) (citation
omitted).
DISCUSSION
I. Commerce Reasonably Concluded That the Shipment Date Was the Date of Sale
Toyo Kohan argues that Commerce correctly continued to use the final invoice price as
Toyo Kohan’s U.S. sales price in its calculation, but that Commerce unreasonably determined that
the material terms of the sale were established on the shipment date. Pl. Cmts. at 7. Toyo Kohan
contends that Commerce unreasonably relied on its practice to use the earlier of the shipment or
invoice dates as the date of sale without carefully considering the record evidence. Id. at 8–9.
Toyo Kohan argues that the quantity and price of its sales are not established upon the shipment
date, meaning that the documentation issued at the time of shipment does not always reflect the
final quantity and price provided in the invoice. Id. at 15. Toyo Kohan explains that this is because
the pricing formula that it agrees with its customer “generates only a temporary or tentative price,
which is expressly subject to adjustment based on subsequent negotiations, in addition to the
exchange rates, and other commercial factors,” and that its U.S. database shows that the majority
of transactions experienced post-shipment price changes. Id. at 12–13. Toyo Kohan adds that the Court No. 24-00261 Page 6 Public Version
billing statement, generated at the time of shipment, reflects the price in Toyo Kohan’s system at
the time of shipment, which can change up to the time of invoicing to the U.S. customer. Id. at
10–11 (citing Toyo Kohan’s Section C Questionnaire Response at 24, P.R. 39, C.R. 48 (Oct. 10,
2023) (“Sec. C Resp.”); Toyo Kohan’s Section A Questionnaire Response at 36, P.R. 22–23, C.R.
3, 5–9 (Sep. 15, 2024) (“Sec. A Resp.”)). Because there is often a change in the price reflected on
the billing statement issued at the time of shipment and the final invoice, Toyo Kohan concludes
that the material terms of the sale are established on the invoice date rather than the shipment date
and that, accordingly, the invoice date is the date of sale. See id. at 15–16, 32.
The government responds that, consistent with its normal practice, Commerce began its
analysis with the final invoice date and “determined whether the material terms of sale at the
invoice date remained the same since the shipping date.” Gov’t Reply at 8. The government
clarifies that “if Commerce finds that the material terms of sale were established before the invoice
date, and if the shipment date precedes the invoice date, Commerce selects the shipment date as
the date of sale.” Id. at 7. Thomas Steel notes that the court has previously affirmed this practice
because “when a party ships its product to a customer, it is reasonable to assume that the material
terms of the sale have been established with the customer.” Thomas Steel Reply at 5 (citing
Tension Steel Indus. Co. v. United States, 179 F. Supp. 3d 1185, 1194 (CIT 2016)) (citation
modified). The government argues that Toyo Kohan has failed to show that its pricing formula
only generates a temporary or tentative initial price, particularly when the sample sales provided
by Toyo Kohan reflect no price change between the shipment and invoice dates.3 Gov’t Reply at
3 Toyo Kohan repeatedly takes aim at Commerce’s reliance on a “single individual U.S. sales transaction” in its analysis, whereas Toyo Kohan provided evidence that “a substantial majority of transactions experienced post-shipment price changes, recorded in the VAL_ADJU field and confirmed by the billing statement and invoice documentation.” See Pl. Cmts. at 13–14. Yet, as Court No. 24-00261 Page 7 Public Version
9. The government adds that, while Toyo Kohan points to the billing documentation issued at the
time of shipment as evidence that the material terms of sale were not established at the shipment
date, Commerce could not rely on any purported difference between the terms of sale listed on the
billing documentation and the final invoice because the court held that the billing documentation
is “virtually meaningless.” Id. (citing Toyo Kohan I at *5); see Thomas Steel Reply at 6 (noting
that any alleged changes between the shipment and invoice dates “were simply internal changes
in Toyo Kohan’s accounting system, not actual price changes between Toyo Kohan and its
customers”).
In an antidumping review, Commerce conducts a “fair comparison” of the prices of a good
sold in the respondent’s home market (the “normal value”) with the prices for the same or similar
good sold in the U.S. market (the “export price”) to determine the dumping margin. 19 U.S.C.
§§ 1677b(a), 1677a, 1675(a)(2). The normal value must be “the price . . . at a time reasonably
corresponding to the time of the sale used to determine the export price.” Id. § 1677b(a)(1)(A).
Accordingly, Commerce determines the “date of sale,” which is the date “when the material terms
of sale are established.” Uruguay Round Agreements Act, Statement of Administrative Action,
H.R. Rep. No. 103-316 (1994) (“SAA”), reprinted in 1994 U.S.C.C.A.N. 4040, 4153.4 The
material terms generally include price, quantity, payment, and delivery terms. Eregli Demir ve
Celik Fabrikalari T.A.S v. United States, 308 F. Supp. 3d 1297, 1306 (CIT 2018) (citing Sahaviriya
Steel Indus. Pub. Co. v. United States, 34 CIT 709, 727, 714 F. Supp. 2d 1263, 1280 (2010), aff’d,
Thomas Steel points out, in its Section A Questionnaire, Commerce asked Toyo Kohan to provide sales trace documentation for a representative “sample” sale in the U.S. market, and Toyo Kohan provided this particular sale. See Thomas Steel Reply at 7–8; Sec. A Resp. at 23. Toyo Kohan has not shown that Commerce unreasonably relied on a sample sale that Toyo Kohan itself provided. 4 Congress designated the SAA as “an authoritative expression by the United States concerning the interpretation of the Uruguay Round Agreements and this Act.” 19 U.S.C. § 3512(d). Court No. 24-00261 Page 8 Public Version
649 F.3d 1371 (Fed. Cir. 2011)). Commerce will usually use the invoice date as the date of sale
but may use another date if that date “better reflects the date on which the exporter or producer
establishes the material terms of sale.” 19 C.F.R. § 351.401(i). Commerce has a “well-established
and long-standing practice of looking beyond the invoice date to the parties’ actual course of
conduct, as well as the parties’ expectations concerning the transaction, to determine whether an
earlier date . . . represents the point at which the parties reached a meeting of the minds on the
material terms of sale.” Nucor Corp. v. United States, 33 CIT 207, 259–60, 612 F. Supp. 2d 1264,
1308 (2009) (citation modified) (citation omitted).
Toyo Kohan’s regular sales process is to offer its customer a price based on a temporarily
agreed pricing formula. Sec. A Resp. at 16. The customer then sends a purchase order to Toyo
Kohan, and Toyo Kohan generates an electronic order confirmation containing the estimated price,
quantity, product specifications, shipment term, and a sales contract. Id. at 16–17. Toyo Kohan
produces the merchandise and ships it to the customer in the United States. Id. at 17. Toyo
Kohan’s system generates a tentative summary bill based on the shipment date and [[
]]. Id.; Sec. C Resp. at 16. Toyo Kohan later generates a payment
invoice reflecting the [[
]]. Sec. A Resp. at 17. Toyo Kohan’s U.S. sales database contains a
field marked “VAL_ADJU,” which reflects the difference between the price listed on the tentative
summary bill generated on the date of shipment and the final invoice. Sec. C Resp. at 25. Toyo
Kohan submitted example sales with its Section A Response. See Sec. A Resp., Exs. A-10A, A-
10B. The sale reflected in Exhibit A-10A includes a contract of purchase, dated [[
]], for [[ ]] metric tons at a unit price of [[ ]] per metric ton. Sec. A Resp.,
Ex. A-10A at 55. The sales contract, dated [[ ]], lists a slightly different quantity of Court No. 24-00261 Page 9 Public Version
[[ ]] metric tons at the same price of [[ ]] per metric ton. Id. at 58. The billing
statement issued at the time of shipment lists the same quantity as the contract of purchase,
[[ ]] metric tons, but a different price per metric ton than the price listed on the contract of
purchase and the final invoice. See id. at 59; Sec. C Resp. at 24; Sec. A Resp. at 36. The price
listed on the final invoice matches the price listed on the contract of purchase and sales contract.
Id. at 62–63. The sale reflected in Exhibit A-10B includes a contract of purchase, [[
]], for [[ ]] metric tons with no unit price listed. Sec. A Resp., Ex. A-10B at 24. The
order confirmation for the sale, dated shortly after the contract, lists the same quantity and a unit
price of [[ ]]. Id. at 26. The sales contract, dated [[ ]], lists the same
values as the order confirmation. Id. at 1. The billing statement lists a different quantity but the
same unit price. Id. at 2. The final invoice lists the same quantity as the billing statement and the
same unit price as the order confirmation, the sales contract, and the billing statement. Id. at 4–5.
In its original determination, Commerce concluded that the material terms of the
documented sample sales, such as price and quantity, between Toyo Kohan and its U.S. customer
were established at the time of the shipment date, meaning the shipment date was the date of sale.
IDM at 4. The court remanded to Commerce to reconsider its date of sale analysis. See Toyo
Kohan I at *5. The court noted that Toyo Kohan’s billing documentation at the time of shipping
is “virtually meaningless,” and that “[w]hile the price reflected in the record at the shipment date
may have no substance, the sales actually may or may not be considered to have been made on the
date.” Id. In its remand results, Commerce continued to find that the material terms of the sale
for Toyo Kohan’s U.S. sales did not change between the shipment date and the invoice date and
that, accordingly, Toyo Kohan’s shipment date is the appropriate date of sale. Remand Results at
4. Commerce noted that the sample sale documentation provided by Toyo Kohan shows that “the Court No. 24-00261 Page 10 Public Version
price indicated on the customer’s purchase order matched Toyo Kohan’s pricing formula, as well
as the price shown on the final invoice.” Id. at 5. Commerce clarified that it used the price reflected
on Toyo Kohan’s final invoice to calculate the net U.S. price in its margin calculations. Id. at 7.
Commerce reasonably selected the shipment date as the date of sale for Toyo Kohan’s U.S.
sales. The parties agree that the only two choices for the date of sale are the shipment date and the
invoice date.5 See Pl. Cmts. at 23 n.1; Thomas Steel Reply at 9; Gov’t Reply at 8. Toyo Kohan’s
argument that the material terms of sale were not established until the invoice date relies on the
purported difference between the price listed on the billing documents issued at the time of
shipment and the final invoice, denoted in Toyo Kohan’s system as the “VAL_ADJU” value. Pl.
Cmts. at 13. Yet, as the court noted, and Toyo Kohan and the government agree, the price and
quantity listed on the documentation issued at the time of shipment are “virtually meaningless, as
[they] do[] not necessarily reflect the quantity or price in the purchase order or the final invoice.”
Toyo Kohan I at *5; see Pl. Cmts. at 15; Gov’t Reply at 9–10. Accordingly, any purported
5 Toyo Kohan also contends that Commerce failed to comply with the court’s remand order to consider the purchase order date as a possible date of sale, although Toyo Kohan specifically notes that it does not actually contend that the purchase order is the appropriate date of sale because of the “large number of U.S. sales transactions for which the final invoice price or quantity are different from the price or quantity on the purchase order.” Pl. Cmts. at 2, 3 n.1. Toyo Kohan notes that, in its draft remand redetermination, Commerce considered and rejected the purchase order as the date of sale but then removed this analysis from the Remand Results. Id. at 9. Toyo Kohan argues that, accordingly, Commerce failed to adhere to the court’s Remand Order. Id. The government responds that Commerce did in fact consider the purchase order date but dismissed it because “Toyo Kohan stated that the purchase order prices did not remain the same.” Gov’t Reply at 8. Thomas Steel argues that the purchase order date is not the appropriate date of sale because, even though the price was set before the shipment date, the final quantity [[ ]]. Thomas Steel Reply at 9. Because no party has actually argued on remand that the purchase order date is the date of sale, the court does not consider whether Commerce sufficiently considered the issue. Court No. 24-00261 Page 11 Public Version
difference between the price listed at the time of shipment and the price listed on the invoice does
not bear on the question of whether the material terms were established at the shipment date.6
The record evidence reasonably supports the conclusion that the price and quantity were
actually established by the time of shipment, and that these terms did not change between the
shipment and invoice dates.7 The sales reflected in Exhibits A-10A and A-10B demonstrate that
the prices on both invoices match the prices reflected in documentation dated prior to the shipment
date, such as the contracts and order confirmations. See Sec. A. Resp., Ex. A-10A at 55, 59, 62–
63; Sec. A. Resp., Ex. A-10B at 24, 26, The only remaining variable, then, is the quantity, and
Toyo Kohan has not suggested that the quantity ever changes after shipment. Accordingly,
Commerce reasonably concluded that the material terms of the sale were established on the
shipment date, and thus that the shipment date is the date of sale.
II. Commerce Reasonably Conducted its Differential Pricing Analysis
Toyo Kohan also challenges the price difference test8 that Commerce has adopted in the
wake of the Federal Circuit’s decision in Marmen invalidating Commerce’s application of the
Cohen’s d test when the underlying data is not normally distributed, equally variable, and equally
6 Toyo Kohan agrees that Commerce reasonably used the final invoice price as Toyo Kohan’s U.S. sales price in its calculation. See Pl. Cmts. at 7. This, however, does not in turn mean that the date of sale must also be the date of invoice. As the court noted in Toyo Kohan I, “when there is an agreed-to pricing formula, there may be a disconnect in the record data for [the price and the date of sale].” Toyo Kohan I at *5. 7 While Toyo Kohan argues that its pricing formula leads to price changes between the shipment and invoice dates, Pl. Cmts. at 12–13, the only record evidence it has presented to demonstrate this proposition is the “VAL_ADJU” documentation, see id. at 13, which, as the court has already noted, is based on meaningless documentation issued at the time of shipment. Toyo Kohan I at *5. 8 As the court describes in more detail below, the “price difference test” is the first step in Commerce’s new differential pricing analysis to determine whether the subject merchandise is being, or is likely to be, sold at less than fair value. See Remand Results at 10–11; 19 U.S.C. § 1677b(a). Court No. 24-00261 Page 12 Public Version
and sufficiently numerous. See Pl. Cmts. at 16–32; Marmen at 1348. Toyo Kohan argues that 19
U.S.C. § 1677f-1(d)(1)(B) permits Commerce to depart from the statutory default average-to-
average comparison methodology only if Commerce finds a “pattern of export prices . . . for
comparable merchandise that differ significantly among purchasers, regions, or periods of time,”
and that Commerce’s price difference test does not actually test for a “pattern” as required by the
statute. Pl. Cmts. at 16–17 (citing 19 U.S.C. § 1677f-1(d)(1)(B)). Toyo Kohan contends that
Congress intended “pattern” in section 1677f-1 to encompass structured relationships between
prices and purchasers, regions, and period of time, whereas Commerce’s price difference test
establishes only whether an arbitrary number of transactions differ among purchasers, regions, or
time periods. Id. at 18–22. Toyo Kohan adds that the legislative history of the statute evidences
that Congress authorized the use of alternative comparison methodologies to address “deliberate
pricing practices that mask dumping, not ordinary or random variation.”9 Id. at 22 (citing SAA at
4177 (“In part, the reluctance to use an average-to-average methodology has been based on a
concern that such a methodology could conceal ‘targeted dumping.’”)). Toyo Kohan also argues
that Commerce’s continued use of the ratio test is unlawful because it only tests for a limited
number of transactions instead of the statutorily mandated “pattern.” Id. at 26. Toyo Kohan adds
9 Toyo Kohan argues that in 2008, Commerce even explicitly rejected the adoption of the two percent price difference test for purposes of implementing section 1677f-1(d)(1)(B). Pl. Cmts. at 24 (citing Certain Steel Nails from the People’s Republic of China, 73 Fed. Reg. 33,977 (Dep’t Commerce June 16, 2008)). The government and Thomas Steel assert that Toyo Kohan confuses the two percent price difference test with the “P/2 test,” which Commerce used in the determinations cited by Toyo Kohan but has not resurrected. Gov’t Reply at 17–18; Thomas Steel Reply at 13. The government adds that, specifically, the P/2 test is different from the two percent price difference test because it covered more steps in the analysis than the price difference test now before the court. Gov’t Reply at 18. Whether Commerce previously used and rejected the two percent price difference test does not influence the court’s reasoning on the reasonableness of the test. See Qingdao Sea-Line Trading Co. v. United States, 766 F.3d 1378, 1387 (Fed. Cir. 2014) (Each administrative review is a “separate exercise of Commerce’s authority.” (citation omitted)). Court No. 24-00261 Page 13 Public Version
that Commerce’s Remand Results “completely forego any statistical or analytical rigor, contrary
to the Federal Circuit’s intent in Marmen in interpreting [section 1677f-1(d)(1)(B)(i)].” Id. at 28.
The government responds that Commerce had only employed the Cohen’s d test in the first
step in Commerce’s differential pricing analysis, and that the Federal Circuit has already sustained
the next two steps in the analysis: the ratio and meaningful difference tests. Gov’t Reply at 13
(citing Stupp Corp. v. United States, 5 F.4th 1341, 1354–56 (Fed. Cir. 2021) (“Stupp I”)). The
government and Thomas Steel argue that, contrary to Toyo Kohan’s position, the price difference
test alone does not aim to identify a pattern. Id. at 16; Thomas Steel Reply at 11–12. Rather,
Commerce’s overall differential pricing methodology—which replaces Cohen’s d with the price
difference test and maintains the previously upheld ratio and meaningful difference tests—as a
whole identifies a “pattern” of prices within the meaning of section 1677f-1(d)(1)(B) because the
“price difference test determines whether prices differ significantly, and the ratio test determines
whether there is a pattern.” Gov’t Reply at 16 (citing Remand Results at 21); Thomas Steel Reply
at 11–12. The government adds that the two percent cutoff in the price difference test is reasonable
because Commerce uses a two percent threshold in other provisions of the statute that consider
price differences—19 U.S.C. §§ 1673b(b)(3), 1673d(a)(4)—and in Commerce’s practice for
determining whether prices to affiliates are made in the ordinary course of trade under 19 C.F.R.
§ 351.403(c).10 Gov’t Reply at 17; see Thomas Steel Reply at 14–15. Thomas Steel adds that in
Marmen, the Federal Circuit did not question the validity of Commerce applying its ratio test to
determine whether a pattern existed. Thomas Steel Reply at 12. Finally, the government argues
10 The government also argues that in the administrative proceedings, Toyo Kohan only challenged the price difference test and failed to challenge Commerce’s use of the ratio test and therefore failed to exhaust its administrative remedies. Gov’t Reply at 19–20. Because the court holds that Commerce’s overall methodology is reasonable, the government’s argument is moot. Court No. 24-00261 Page 14 Public Version
that the statute does not specify how Commerce may determine whether a pattern exists and
whether prices differ significantly. Gov. Reply at 22.
As already discussed, in an antidumping duty review, Commerce conducts a “fair
comparison” of the normal value and the export price to determine whether the subject
merchandise is being, or is likely to be, sold at less than fair value. 19 U.S.C. § 1677b(a).
Commerce may use an average-to-average (comparing the weighted-average normal values to
weighted-average export prices), transaction-to-transaction (comparing the normal values of
individual transactions with the export prices of individual transactions), or average-to-transaction
method (comparing the weighted-average normal values with the export prices of individual
transactions). 19 C.F.R. § 351.414(b). Commerce will use the average-to-average method unless
it determines that another method is appropriate in a particular case. Id. § 351.414(c). Commerce
may use the average-to-transaction method if “there is a pattern of export prices . . . for comparable
merchandise that differ significantly among purchasers, regions, or periods of time,” and
Commerce “explains why such differences cannot be taken into account using [the average-to-
average or transaction-to-transaction methods].” 19 U.S.C. § 1677f-1(d)(1)(B). The rationale
informing the average-to-transaction method “is that targeted dumping is more likely to be
occurring when export prices fit a pricing model that differs significantly among different periods
of time, different purchasers, or different regions of the United States.” Stupp I at 1345 (citing
Apex Frozen Foods Private Limited v. United States, 862 F.3d 1337, 1347 (Fed. Cir. 2017)).
Congress has not delineated exactly how Commerce shall assess whether there is a “pattern
of export prices . . . differ[ing] significantly among purchasers, regions, or periods of time.” Id.
(citation omitted); see 19 U.S.C. § 1677f-1(d)(1)(B). In 2013, Commerce began applying its
“differential pricing analysis.” See Differential Pricing Analysis; Request for Comments, 79 Fed. Court No. 24-00261 Page 15 Public Version
Reg. 26,720, 26,722 (Dep’t Commerce May 9, 2014); Xanthan Gum from the People’s Republic
of China, 78 Fed. Reg. 33,351 (Dep’t Commerce June 4, 2013). Commerce conducted the
differential pricing analysis in three steps: First, Commerce applied the Cohen’s d test to determine
whether sales prices in the test group are significantly different from the sales prices in the
comparison group. Stupp I at 1346–7 (citation omitted). Second, Commerce applied the ratio test
to sales that “passed” the Cohen’s d test (i.e., sales with significantly different sales prices from
those in the comparison group). Id. at 1347. If the total percentage of transactions that passed the
Cohen’s d test was 33% or less, Commerce used the average-to-average method to calculate the
weighted average dumping margin. Id. If the total percentage was 66% or more, Commerce
tentatively used the average-to-transaction method. Id. If the total percentage was between 33%
and 66%, Commerce tentatively applied the average-to-transaction method to those transactions
passing the Cohen’s d test and the average-to-average method to the remainder of the transactions.
Id. (citation omitted). Third, Commerce conducted the “meaningful difference test” in which it
assessed whether using the average-to-average or the other tentative calculation method led to
meaningfully different results.11 Id. (citation omitted). If the comparison led Commerce to
conclude that there was no meaningful difference, Commerce applied the average-to-average
11 The first step of the meaningful difference test was to calculate the weighted average dumping margin using the average-to-average method. Stupp I at 1347. The second step was to calculate the weighted average dumping margin with the tentatively selected method. Id. The third step was to compare the results. Id. If the margin for the average-to-average method was below the de minimis threshold—which is 2 percent in less-than-fair-value investigations, see 19 U.S.C. § 1673d(a)(4)—and the margin for the tentatively selected method was above that threshold, or if both were above the threshold and the margin for the tentatively selected method was 25 percent greater than the average-to-average margin, then Commerce considered there to be a meaningful difference, and selected the alternative approach. See id. If Commerce concluded there was no meaningful difference, Commerce applied the average-to-average method across the board. Id. Court No. 24-00261 Page 16 Public Version
method to all sales. Id. Otherwise, Commerce selected the alternative comparison method. Id.
(citation omitted).
The Federal Circuit held that Commerce’s ratio test, step two of the analysis described
above, “reasonably implements the statutory requirement that Commerce determine whether there
is a ‘pattern of export prices’ ‘differ[ing] significantly among purchasers, regions, or periods of
time’ before selecting the average-to-transaction method.”12 Id. at 1355 (quoting 19 U.S.C.
§ 1677f-1(d)(1)(B)(i)). The Federal Circuit reasoned that the ratio test “is [] a conventional method
for quantifying comparisons across discrete groups: counting the number of divergent sales prices,
as identified by an effect-size test, and calculating the population percentage of those divergent
sales prices.” Id. at 1354. The Federal Circuit further held that Commerce’s selection of the 33%
and the 66% cutoffs in the ratio test is reasonable. Id. at 1354–55. The Federal Circuit also held
12 Toyo Kohan also argues that, while the Federal Circuit accepted the ratio test as part of the differential pricing test in Stupp I at 1353, the Federal Circuit’s reasoning rested on Commerce’s discretion to determine a reasonable methodology. Pl. Cmts. at 29. Toyo Kohan concludes that the decision is therefore no longer applicable in the wake of Loper Bright Enters. v. Raimondo, 603 U.S. 369 (2024) (“Loper Bright”). Id. at 29. In Loper Bright, the Supreme Court overruled Chevron, U.S.A., Inc. v. Nat. Res. Def. Council, Inc., 467 U.S. 837 (1984), holding that courts must exercise their independent judgment in deciding whether an agency has acted within its statutory authority. Loper Bright at 412. Courts exercise their independent judgment in deciding statutory meaning, id. at 394, and may conclude that Congress explicitly delegated authority to the agency to give meaning to the terms so long as it is consistent with the Constitution. Id. at 413. Congress afforded flexibility to Commerce in section 1677f-1(d)(1)(B) by including the phrase “differ significantly,” which is an open-ended qualifier. Garg Tube Export LLP v. United States, 740 F. Supp. 3d 1355, 1366–67 (CIT 2024) (“Congress’ mandate to Commerce to assess not merely whether prices differ, but whether they differ ‘significantly’ necessarily affords Commerce flexibility to assess the degree of difference depending on the particular context.” (citing Loper Bright at 395)); see 19 U.S.C. § 1677f-1(d)(1)(B). The SAA confirms Congress’ delegation: “[T]he Administration intends that in determining whether a pattern of significant price differences exist[,] Commerce will proceed on a case-by-case-basis . . . .” SAA at 4178. Because the text and legislative history of this subsection of the statute confirm that Congress intended that Commerce have discretion in making the relevant determination, the Federal Circuit’s holding in Stupp I is undisturbed by the Supreme Court’s holding in Loper Bright. Court No. 24-00261 Page 17 Public Version
that Commerce’s use of the “meaningful difference test,” step three in the analysis described
above, is reasonable. Id. at 1356 (citing Apex Frozen Foods, 862 F.3d at 1348–49).
In Marmen, the Federal Circuit held that Commerce unreasonably used the Cohen’s d test,
the first step of its differential pricing analysis, “to determine whether prices differ significantly
when the underlying data is not normally distributed, equally variable, and equally and sufficiently
numerous.” Marmen at 1348. The Federal Circuit remanded, directing that “Commerce may re-
perform a differential pricing analysis, and that analysis may not rely on the Cohen’s d test for data
sets like those here.” Id. The Federal Circuit “[did] not preclude Commerce from fashioning and
justifying a statistical analysis that uses some of the ideas underlying [the Cohen’s d] analysis of
group differences as long as the resulting analysis is itself justified as sound for gauging differences
in the data sets at issue.” Id. Following the Federal Circuit’s decision, Commerce sought
information and public comment on alternatives to the Cohen’s d test. See Alternatives to the Use
of the Cohen’s d; Request for Comment, 90 Fed. Reg. 21,277 (Dep’t Commerce May 19, 2025).
In its remand redetermination, Commerce discontinued its application of the Cohen’s d test
to Toyo Kohan’s U.S. sales data and instead applied the “price difference test” as the first step in
its differential pricing analysis. Remand Results at 8. Commerce outlined that “[f]or comparable
merchandise, the price difference test examines whether the weighted-average net price to a given
purchaser, region, or time period is within two percent of the weighted average net price to all
other purchasers, regions or time periods.” Id. at 10 (citation modified). If the weighted-average
net price to the given purchaser, region, or time period falls outside of the two percent band, then
those prices “pass” the price difference test. Id. at 10–11. Commerce then proceeds to steps two Court No. 24-00261 Page 18 Public Version
and three: the ratio13 and meaningful difference14 tests. Id. at 11–12. Using its new differential
pricing analysis, Commerce preliminarily found that 100 percent of the value of Toyo Kohan’s
U.S. sales passed the price difference test, and “confirm[ed] [with the ratio test] the existence of a
pattern of prices that differ significantly among purchasers, regions, or time periods.” Id. at 12.
Commerce also preliminarily applied the meaningful difference test and accordingly applied the
average-to-transaction method to calculate the weighted-average dumping margin for Toyo
Kohan. Id.
When reviewing Commerce’s choice of methodology, the court considers whether the
methodology “reasonably implements a given statutory directive,” Stupp I at 1351 (quoting Stupp
Corp. v. United States, 365 F. Supp. 3d 1373, 1378 (CIT 2019)). Commerce’s two percent price
difference test reasonably implements section 1677f-1(d)(1)(B)’s directive. As discussed above,
Congress has not specified how Commerce shall assess whether there is a “pattern of export prices
. . . differ[ing] significantly among purchasers, regions, or periods of time.” 19 U.S.C. § 1677f-
13 Commerce’s ratio test now differs from that described above because Commerce has discontinued the use of the “mixed method” in administrative proceedings. Remand Results at 8. Now, “[i]f 33 percent or less of the total value of sales passes the price difference test, then the results of the price difference and ratio tests do not support consideration of the [average-to- transaction] method. If more than 33 percent of the total value of U.S. sales passes the price difference test, then Commerce will find that a pattern of prices existed during the POR.” Id. at 11. Toyo Kohan, however, does not argue that this difference in the ratio test renders it unreasonable now. See Pl. Cmts. at 27–29. Rather, Toyo Kohan argues that the ratio test does not identify a pattern of sales. Id. As the court will discuss, this argument fails. 14 Commerce’s meaningful difference test described in the Remand Results matches the pre- Marmen meaningful difference test described in note 11, supra; see Stupp I at 1347. Specifically, Commerce considers whether using the average-to-transaction method yields a meaningful difference in the weighted-average dumping margin as compared to that resulting from the average-to-average method. Remand Results at 11. Commerce considers a difference to be meaningful if “(1) there is a 25 percent relative change in the weighted-average dumping margins between the [average-to-average] method and the [average-to-transaction] method where both rates are above the de minimis threshold; or (2) the resulting weighted-average dumping margins between the [average-to-average] method and the [average-to-transaction] method move across the de minimis threshold.” Id. at 11–12. Court No. 24-00261 Page 19 Public Version
1(d)(1)(B); see Stupp I at 1346. Toyo Kohan shoots arrows at Commerce’s price difference test
but does not explain how the test may be improved or propose a meaningful alterative. See Pl.
Cmts. at 16–32. Toyo Kohan dedicates much of its brief to its argument that Commerce’s “new
approach continues to ignore the statutory test to find a ‘pattern’ of prices.” Id. at 16. Yet, as
Commerce pointed out in its Remand Results, the price difference test—which has replaced the
Cohen’s d test and is just the first step of Commerce’s differential pricing analysis—aims to
determine whether prices differ significantly, not whether a pattern exists. See Remand Results at
17–18. Instead, Commerce uses the ratio test, the second step in the differential pricing analysis,
to determine whether there is a pattern to the price differences identified by the two percent price
difference test. Id. at 18. While Toyo Kohan also takes aim at the ratio test, see Pl. Cmts. at 25–
27, as discussed, the Federal Circuit has already affirmed Commerce’s use of the ratio test as a
reasonable approach to detect a pattern of sales in the light of the broad language in section 1677f-
1(d)(1)(B).15 See Stupp I at 1354.
15 Toyo Kohan also challenges the ratio test because Commerce “appl[ies] the 33 percent test to the total, and not by the different purchasers, regions, or time periods, and not by the different categories of comparable merchandise.” Pl. Cmts. at 26. The government responds that the Federal Circuit has already held that the aggregation of sales across purchases, regions, or time periods categories in comparing sales “is not inconsistent with the statute.” Gov. Reply at 23 (citing Dillinger France S.A. v. United States, 981 F.3d 1318, 1325 (Fed. Cir. 2020)). Thomas Steel responds that the ratio test “necessarily considers such distinctions [between purchasers, regions, and time periods] because the output from the first step of the differential pricing analysis—which measures whether there are significant price differences between purchasers, regions, or time periods—is used as an input in the second step of the differential pricing analysis, the ratio test.” Thomas Steel Reply at 13 (citing Marmen at 1346). In Stupp I, the Federal Circuit explained that before Commerce applied the Cohen’s d test as the first step of its differential pricing analysis, Commerce segmented the respondent’s exports sales into product groups and then segmented each product group into subsets based on the purchasers, regions, and time periods. Stupp I at 1346. Once Commerce conducted the first step, which at the time of Stupp I was the Cohen’s d test, Commerce applied the ratio test to the observations in each product group that passed the Cohen’s d test. Id. at 1347. The Federal Circuit held that this use of the ratio test “reasonably implements the statutory requirement that Commerce determine whether there is ‘a Court No. 24-00261 Page 20 Public Version
Accordingly, the court considers only whether the two percent price difference test
reasonably determines whether prices differ significantly as directed by section 1677f-
1(d)(1)(B)(i). The language of the statute affords Commerce discretion to fashion its methodology.
See Garg Tube Export, 740 F. Supp. 3d at 1367 (“Congress’ mandate to Commerce to assess not
merely whether prices differ, but whether they differ ‘significantly’ necessarily affords Commerce
flexibility to assess the degree of difference depending on the particular context.”). The Federal
Circuit’s holding in Marmen also leaves Commerce discretion to amend its differential pricing
analysis: The court “[did] not preclude Commerce from fashioning and justifying a statistical
analysis that uses some of the ideas underlying [the Cohen’s d] analysis of group differences as
long as the resulting analysis is itself justified as sound for gauging differences in the data sets at
issue.” Marmen at 1348. The court left Commerce the option to “re-perform a differential pricing
analysis” that does not rely on Cohen’s d for data sets that were not normally distributed, equally
variable, and equally and sufficiently numerous. Id. Commerce took the latter route: Rather than
apply the ideas underlying Cohen’s d, Commerce constructed a new price difference test. As
already discussed, the language of the statute affords Commerce discretion to fashion its
methodology. See Garg Tube Export, 740 F. Supp. 3d at 1367 (“Congress’ mandate to Commerce
to assess not merely whether prices differ, but whether they differ ‘significantly’ necessarily
affords Commerce flexibility to assess the degree of difference depending on the particular
context.”); see supra note 12. Accordingly, Toyo Kohan’s argument that the price difference test
pattern of export prices’ ‘differ[ing] significantly among purchasers, regions, or periods of time’ before selecting the average-to-transaction method.” Id. at 1355 (citing 19 U.S.C. § 1677f- 1(d)(1)(B)(i)). Just as before, Commerce’s new differential pricing analysis evaluates all U.S. sales by purchaser, region, and time period. See Remand Results at 10. Commerce’s application of the ratio test to the total value of sales by the respondent in the United States during the period of review, therefore, does not render the test unreasonable. See Remand Results at 11; Stupp I at 1355. Court No. 24-00261 Page 21 Public Version
foregoes any statistical or analytical rigor, see Pl. Cmts. at 28, does not succeed. The task Congress
has set in section 1677f-1(d)(1)(B) has no obvious solution. Toyo Kohan has not suggested what
route Commerce should take and has not shown that Commerce has unreasonably implemented
the statutory directive of section 1677f-1(d)(1)(B). The court therefore sustains Commerce’s
differential pricing analysis.
CONCLUSION
For the foregoing reasons, Commerce’s remand redetermination, ECF No. 65-1, is
sustained. Judgment will be entered accordingly.
/s/ Jane A. Restani Jane A. Restani, Judge
Dated: May 2, 2026 New York, New York