HIGHLAND INDUSTRIES INC., JIASHAN DALIN

Page 22024-1258, 2024-1259

Appeals from the United States Court of International Trade in Nos. 1:18-cv-00002-JCG, 1:18-cv-00011-JCG, 1:18-cv-00017-JCG, 1:18-cv-00018-JCG, 1:18-cv-00019JCG, Judge Jennifer Choe-Groves.

Page 3Decided: April 29, 2026

GREGORY S. MENEGAZ, The Inter-Global Trade Law Group PLLC, Washington, DC, argued for all plaintiffs-appellees. Plaintiffs-appellees Anhui Hoda Wood Co., Ltd., Celtic Co., Ltd., Far East American, Inc., Jiaxing Gsun Import & Export Co., Ltd., Jiaxing Hengtong Wood Co., Ltd., Linyi Chengen Import and Export Co., Ltd., Linyi Evergreen Wood Co., Ltd., Linyi Glary Plywood Co., Ltd., Linyi Hengsheng Wood Industry Co., Ltd., Linyi Huasheng Yongbin Wood Co., Ltd., Linyi Jiahe Wood Industry Co., Ltd., Linyi Linhai Wood Co., Ltd., Linyi Mingzhu Wood Co., Ltd., Linyi Sanfortune Wood Co., Ltd., Qingdao Good Faith Import and Export Co., Ltd., Shandong Qishan International Trading Co., Ltd., Shanghai Futuwood Trading Co., Ltd., Suining Pengxiang Wood Co., Ltd., Suqian Hopeway International Trade Co., Ltd., Suzhou Oriental Dragon Import and Export Co., Ltd., Xuzhou Andefu Wood Co., Ltd., Xuzhou Jiangyang Wood Industries Co., Ltd., Xuzhou Long-yuan Wood Industry Co., Ltd., Xuzhou Pinlin International Trade Co. Ltd., Xuzhou Shengping Import and Export Co., Ltd., Xuzhou Timber International Trade Co. Ltd. also represented by JAMES KEVIN HORGAN, ALEXANDRA H. SALZMAN, DeKieffer & Horgan, PLLC, Washington, DC.

JEFFREY S. NEELEY, Husch Blackwell LLP, Washington, DC, for plaintiff-appellee Zhejiang Dehua TB Import & Export Co., Ltd. Also represented by STEPHEN W. BROPHY.

JEFFREY S. GRIMSON, Mowry & Grimson, PLLC, Washington, DC, for plaintiffs-appellees Taraca Pacific, Inc., Richmond International Forest Products, LLC. Also represented by BRYAN PATRICK CENKO, JILL CRAMER, KRISTIN HEIM MOWRY, SARAH WYSS.

Page 4SOSUN BAE, Commercial Litigation Branch, Civil Division, United States Department of Justice, Washington, DC, argued for defendant-appellant United States. Also represented by BRIAN M. BOYNTON, TARA K. HOGAN, PATRICIA M. MCCARTHY; SAVANNAH MAXWELL, IAN ANDREW MCINERNEY, Office of the Chief Counsel for Trade Enforcement and Compliance, United States Department of Commerce, Washington, DC.

MAUREEN E. THORSON, Wiley Rein, LLP, Washington, DC, argued for defendant-appellant Coalition for Fair Trade of Hardwood Plywood. Also represented by STEPHANIE MANAKER BELL, TIMOTHY C. BRIGHTBILL, TESSA V. CAPELOTO.

Before LOURIE, STOLL, and STARK, Circuit Judges .

STARK, Circuit Judge .

The United States (“government”) and the Coalition for Fair Trade in Hardwood Plywood (“Coalition” and, with the government, “Appellants”) appeal from a final order of the Court of International Trade (“Trade Court”) setting antidumping duties for hardwood-plywood products produced in and imported from the People's Republic of China (“China”) by appellees Linyi Chengen Import & Export Co., Ltd. (“Chengen”), Richmond International Forest Products LLC ('Richmond International'), Taraca Pacific, Inc. (“Taraca Pacific”), and Zhejiang Dehua TB Import & Export Co., Ltd. (“Dehua TB”) (together, “Appellees”). Appellants raise three issues. First, they challenge the Trade Court's order requiring Commerce to accept into the administrative record certain documents, not discovered until the verification stage of Commerce's investigation, and treat the documents as complete and accurate. Second, they contend the Trade Court erred by rejecting Commerce's intermediate input method to measure one ofPage 5 Chengen’s factors of production. Lastly, the Coalition (but not the government) seeks reversal of Commerce’s exemption (affirmed by the Trade Court) of two non-mandatory respondents, Dehua TB and Jiangyang Wood Industries Co., Ltd. ('Jiangyang Wood'), from the all-others antidumping duty rate. We agree with Appellants on the first two issues and with the Trade Court on the third. Accordingly, we affirm-in-part, reverse-in-part, and remand to the Trade Court for further proceedings.

In 2016, in response to a petition filed by the Coalition, Commerce initiated an anti-dumping investigation into Chinese hardwood-plywood products. In such investigations, Commerce must “determine the individual weighted average dumping margin for each known exporter and producer of the subject merchandise.' 19 U.S.C. § 1677f1(c)(1). “A dumping margin reflects the amount by which the normal value (the price a producer charges in its home market) exceeds the export price (the price of the product in the United States) or constructed export price.” Nan Ya Plastics Corp. , 810 F.3d 1333, 1337 (Fed. Cir. 2016) (internal quotation marks omitted).

To determine the appropriate margin, Commerce creates a list of importer-respondents and assesses whether individual investigation of each respondent is practicable. If it is not, Commerce may select a lesser number of “mandatory respondents,” each of whom is individually investigated and receives an individualized rate based on the facts discovered during the investigation. Importers who are not investigated by Commerce - termed 'non-mandatory respondents' - receive an 'all-others' rate. 19 U.S.C. § 1673d(c)(1)(B)(i)(II). Here, Commerce determined that individual investigation of all respondents would be impracticable and selected Chengen and Shandong DongfangPage 6 Bayley Wood Co., Ltd. (“Bayley”) as the only mandatory respondents.¹

Any entity not selected by Commerce for individual investigation (i.e., any non-mandatory respondent) may request voluntary respondent status. See 19 U.S.C. § 1677m(a)(1). If such a requesting party 'submits to [Commerce] the information requested from exporters or producers selected for examination,' in the manner and time prescribed, Commerce “shall establish . . . an individual weighted average dumping margin' for that party, treating it as a voluntary, individual respondent. Id. In this case, two non-mandatory respondents, Jiangyang Wood (an affiliate of Richmond International and Taraca Pacific) and Dehua TB sought to be treated as individual respondents not subject to the all-others rate.

When considering nonmarket economies (“NME”) like China, all respondents (mandatory and non-mandatory) are initially presumed to be controlled by the government and, hence, subject to a country-wide antidumping duty rate, also referred to as an “NME-entity rate.” 19 U.S.C. § 1677b(c); 19 C.F.R. § 351.108. Respondents may rebut this presumption by demonstrating independence from their government; if successful, they are assigned what is called a “separate rate.” See 19 C.F.R. § 351.108(b),(d). In the case before us, Commerce found that mandatory respondent Chengen and the non-mandatory respondents rebutted the presumption of government control, but mandatory respondent Bayley did not.² This left ChengenPage 7 as the only mandatory respondent subject to an individualized rate, instead of the NME-entity rate.

Commerce’s investigations of mandatory respondents typically begin with a questionnaire, in response to which respondents provide information and submit documentation regarding the nature of their operations and business. See 19 C.F.R. § 351.301. These questionnaires have the force of subpoenas. See 19 C.F.R § 207.8. Commerce reviews the responses and, thereafter, issues a preliminary determination of an appropriate antidumping duty. See 19 U.S.C. § 1673b; see also 19 C.F.R. § 351.205. Commerce then begins “verification,” a process that can involve visiting the facilities of a respondent and its relevant affiliates, and reviewing books and records, all in an effort to ensure the information that was provided in response to the questionnaires is complete and accurate. See 19 C.F.R. § 351.307. Verification is meant to be a “test” of the information already in the administrative record. Goodluck India Ltd. v. United States , 11 F.4th 1335, 1344 (Fed. Cir. 2021).

non-mandatory respondents) is equivalent to the separate

rate (which is assigned to successful separate rate appli-

cants not individually investigated). We use the terms in-

terchangeably. See Yangzhou Bestpak Gifts & Crafts Co.

, 716 F.3d 1370, 1374 (Fed. Cir. 2013) (not-

ing that in similar situations, separate rate is determined

using same method used to calculate all-others rate); Albe-

marle Corp. & Subsidiaries v. United States , 821 F.3d

1345, 1348 (Fed. Cir. 2016) (noting terms “separate rate”

and “all others rate” can be interchangeable). For these

same reasons, we also treat the terms “non-mandatory re-

spondent” and “separate rate respondent” as equivalent.

Page 8Completeness and accuracy are important in individual investigations because Commerce uses the record developed to determine the respondent’s factors of production, including raw material input volumes and energy and labor costs. See 19 U.S.C. §1677b(c)(3). The factors of production, in turn, allow determination of variables such as the normal value of the merchandise sold in the exporting country (here, China), allowing calculation of the dumping margin and antidumping duties. The applicable statute obligates Commerce to use the “best available information” to evaluate the factors of production, id. § 1677b(c)(1), but does not require Commerce to use information that is unverified, see id. § 1677m(e)(2).

Commerce’s initial questionnaire to Chengen requested 'all documents . . . generated/used/relied upon during production . . . .” J.A. 1445. A supplemental questionnaire asked Chengen to provide “a detailed explanation and supporting documentation for how [it] records the purchase and consumption of each input in the normal course of business.” J.A. 5079. Chengen’s responses to both the original and supplemental questionnaires stated that its volume determinations for its largest raw material input, poplar logs, were recorded on “material purchase invoices;” these responses further claimed that “accountants record purchases according to the warehouse in-tickets and invoices . . . in the normal course of accounting.” J.A. 314. Chengen never explained how the invoices or warehouse in-tickets were prepared.

Based on these responses, Commerce preliminarily determined that Chengen should receive a de minimis dumping margin. See 19 U.S.C. § 1673b (“[A] weighted average dumping margin is de minimis if the administering authority determines that it is less than 2 percent ad valorem or the equivalent specific rate for the subject merchandise.”). Because Commerce is required to disregard any de minimisPage 9 weighted average in calculating the all-others rate for nonmandatory respondents, see 19 U.S.C. § 1673b, Chengen’s rate did not factor into the preliminarily determined allothers rate. Certain Hardwood Plywood Prods. from the People’s Republic of China , 82 Fed. Reg. 28,629 (Dep’t Commerce June 23, 2017).

In reaching its de minimis conclusion for Chengen, Commerce applied the standard factors of production analysis, by valuing (among other inputs) Chengen’s largest input, poplar logs. Commerce rejected the Coalition’s request that it instead apply the intermediate input method, which would have involved valuing wood veneer – thin sheets of wood cut from logs by machine – an intermediate created in the course of producing the subject hardwood-plywood products. The intermediate input methodology is a deviation from Commerce’s default factors of production analysis. See Shenzhen Xinboda Indus. Co. v. United States , 456 F. Supp. 3d 1272, 1280-81 (Ct. Int’l Trade 2020).

After completing its preliminary determination, Commerce moved into the verification phase. As part of that process, it sent Chengen an on-site verification agenda, explaining:

verification is not intended to be an opportunity for submission of new factual information. New factual information will be accepted at verification only when: (1) the need for that information was not evident previously; (2) the information makes minor corrections to information already on the record; or (3) the information corroborates, supports, or clarifies information already on the record.

J.A. 6940.

Verification here included sending employees to China to visit the facilities of a Chengen-affiliated company called Linyi Dongfangjuxin Wood Co., Ltd. ('Dongfangjuxin'), Page 10which produces the merchandise that Chengen exports.³ During the on-site visit, Commerce learned that, contrary to Chengen’s questionnaire responses, Chengen’s production affiliate actually determines poplar log input volume using a two-page conversion chart ('Conversion Chart'), which considers two measurements: (1) log length (which is uniform across suppliers) and (2) log diameter, as measured by warehouse staff at the thin end of each log. The corresponding volume from the chart is subsequently recorded on warehouse in-tickets, company-maintained journals, and Chinese tax forms, which Dongfangjuxin provides to suppliers as receipts and also submits to Chinese tax authorities. Commerce further discovered at verification that Chengen and its affiliates did not possess any supplier-provided invoices listing the volume of poplar logs, notwithstanding Chengen’s suggestion in its questionnaire responses that it did.

Commerce verifiers requested a copy of the Conversion Chart. Chengen responded by providing a fuller document that included the Conversion Chart as well as a cover page and ten additional pages (the “12-Page Document”). Commerce’s verifiers only reviewed and accepted the Conversion Chart but not the remainder of the 12-Page Document as they viewed the excess to be untimely new factual information.

After verification concluded, the parties submitted briefing. Once again, the Coalition urged Commerce to apply the intermediate input methodology, rather than the standard factors of production analysis, arguing that the Chengen-provided log volumes are inherently inaccurate and subject to manipulation. Chengen responded that its volume calculation methodology, which it now insisted relies on the Conversion Chart, is reliable and accurate andPage 11 also complies with the Chinese National Standard for log volume calculation, which it asserted could be discerned from even cursory review of the 12-Page Document.

In November 2017, Commerce issued its final determination. It deviated from the approach it had taken at the preliminary determination stage, now deciding to apply the intermediate input methodology, valuing the veneers used in Chengen’s production process rather than poplar log volumes. Commerce explained that multiple discoveries at verification made the intermediate input method appropriate. In particular, Commerce reasoned that its discovery of Chengen's use of the Conversion Chart, as well as its lack of supplier-provided invoices, cast doubt on Chengen’s self-reported log volumes. Commerce additionally found that Chengen’s method of calculating volume, which relied on diameter measurements at the thin end of the log, was inherently imprecise and predicated on the incorrect assumption that all logs are uniform in shape. Thus, Commerce could not accurately determine the volume of Chengen's largest input, poplar logs, and would instead turn to valuing its intermediate product, veneers. Applying this methodology, Commerce calculated a 183.36% dumping margin for Chengen. Commerce assigned this same rate to the non-mandatory respondents as the all-others rate.

Chengen contested Commerce’s use of the intermediate input methodology as being plagued by ministerial error. See 19 C.F.R. § 351.224(e). Chengen argued that: (1) the entire 12-Page Document it provided at verification should be included in the record, and (2) Commerce improperly viewed Chengen’s method of measuring log volume as inherently imprecise. Commerce found no ministerial error.

Chengen and other interested parties then filed a complaint in the Trade Court challenging Commerce’s decision. In an opinion issued in June 2019, the Trade Court Page 12concluded that Commerce did not sufficiently explain how the record supports its determinations that the Chengencalculated and recorded log volumes are unreliable. The court remanded for Commerce to provide a more fulsome explanation of its decisionmaking. It reserved ruling on whether Commerce had appropriately applied the intermediate input methodology.

On remand, Commerce provided additional information but adhered to its determination that the log-volume information Chengen had submitted could not be verified and, thus, the intermediate input method was more appropriate. Commerce also explained, including via a declaration from a Commerce analyst who attended the on-site verification, that agency verifiers first saw the twopage Conversion Chart on a desk while touring Dongfangjuxin’s facilities. The declaration added that the verifiers had requested the Conversion Chart in order to use it to verify information that was already in the record. By contrast, Commerce had never requested the full 12-Page Document, which Chengen offered at verification, because the verifiers understood the additional pages to be new (and untimely) factual information outside the record. Commerce further explained that the accuracy of Chengen's volume measurements was not evident even from the Conversion Chart; to the contrary, Chengen's method was inherently imprecise because it involved measuring the diameter of the small end of irregularly-shaped logs. Finally, Commerce stated that Chengen alone was responsible for recording and documenting log volumes, adding that the resulting lack of third-party verification undermined Commerce’s confidence that Chengen’s poplar log intake volumes were accurate. As a result of these determinations, Commerce again set Chengen's dumping rate at 183.36% and deemed this same rate as the all-others rate.

Chengen once again sought review at the Trade Court. The Trade Court issued its second decision in FebruaryPage 13 2020, this time concluding that Commerce abused its discretion by “rejecting the 12-page complete document representing the entirety of the Chinese National Standard.' J.A. 123. According to the Trade Court, the complete document “should be construed instead as information corroborating, supporting, or clarifying information already on the record (regarding Linyi Chengen’s method of calculating log volumes) that should be accepted pursuant to Commerce’s past practices at verification.” Id. In the court's view, the additional pages beyond the two-page Conversion Chart were not “prohibited new factual information,” and Commerce’s focus on the lack of third-party corroboration of the information lacked a legal basis. Id. Therefore, the Trade Court again remanded, this time instructing Commerce 'to accept the previously-rejected documents' as “complete and accurate” and to “reconsider modifying Linyi Chengen’s margin and the rate assigned to the Separate Rate Plaintiffs [i.e., the non-mandatory respondents].' J.A. 27, 126.

In its subsequent second remand determination, Commerce (1) reopened the record to include, under protest, the ten additional pages offered by Chengen at verification; (2) abandoned the intermediate input methodology in favor of the standard factors of production methodology; (3) revised Chengen’s rate from 183.36% to 0%; (4) found that Bayley was a China-wide entity and issued it a revised rate of 114.72%; and (5) determined an all-others rate of 57.36%, which was the simple average of the two mandatory respondents’ rates: Chengen’s 0% and Bayley’s 114.72%.

In the appeal from these second remand determinations, the Trade Court affirmed the 0% rate as to Chengen but reversed Commerce’s all-others rate. The court noted that Commerce’s use of a simple average of Chengen’s and Bayley’s margins departed from Commerce’s typical method of calculating the all-others rate, which is to take the weighted average of the dumping margins for Page 14mandatory respondents while excluding any zero and de minimis margins (and also excluding any margins based entirely on facts available). 19 U.S.C. § 1673d(c)(5)(A). At the same time, the Trade Court recognized that if all dumping margins assigned to individually investigated entities are de minimis or zero based on facts available, which is how the Trade Court viewed this case, then Commerce “may invoke an exception to establish a separate rate for exporters and producers not individually investigated,” allowing Commerce to use “any reasonable method to establish the estimated all-others rate for exporters.” J.A. 13839; see also 19 U.S.C. § 1673d(c)(5)(B). The Trade Court remanded because Commerce had not adequately explained its departure from its typical method or why its alternative simple-average method was reasonable.

In the ensuing remand, Commerce explained its use of the simple average methodology to calculate the separate rate. It pointed to differences in the prices at which Chengen and the separate rate respondents sold merchandise, as well as differences in their respective cost structures. According to Commerce, these differences make the alternative method of simple averaging reasonable because neither the Chengen rate nor the Bayley rate (which were the only two individually investigated rates of record), alone, reasonably reflected the separate rate respondents’ potential for dumping. The Bayley rate of 114.72% failed to account for the fact that the separate respondents, unlike Bayley, had demonstrated independence from the Chinese government. As for Chengen, which had also proven independence, differences in cost structures indicated that making Chengen’s 0% rate the all-others rate would inappropriately benefit the separate rate respondents, who had a greater potential for dumping than Chengen. Commerce concluded, then, that a simple average of the Bayley and Chengen rates was a reasonable way to account for these differences.

Page 15Thereafter, the Trade Court issued its fourth decision, again remanding Commerce's calculation of the rate assigned to the non-mandatory respondents. The Trade Court determined that the agency had discretion to depart from its typical methodology, and apply a simple average, but also found the agency had not produced substantial evidence that the resulting rate, 57.36%, was reasonably reflective of the separate rate respondents' potential for dumping.

On remand, Commerce once again concluded the 57.36% rate for separate rate respondents was the most reasonable rate based on the record, given Chengen’s 0% rate and the 114.72% China-wide rate (based on mandatory respondent Bayley). In reviewing that decision, the Trade Court remanded yet again, reiterating its finding that the agency had not provided substantial evidence that the 57.36% separate rate was reasonably reflective of the separate rate respondents' potential for dumping. The Trade Court “advised” Commerce not to submit the 57.36% separate rate again.

In the fifth remand, Commerce complied with the Trade Court's directive, assigning the separate rate respondents the same rate as Chengen, 0%. Commerce also excluded two individual non-mandatory respondents, Jiangyang Wood and Dehua TB, from the all-others rate, as they had requested voluntary respondent status and had submitted all questionnaire responses in a timely manner.

When the Trade Court reviewed Commerce’s fifth remand determination, it affirmed. The court entered judgment in October 2023.

The Coalition and the government now appeal. We have jurisdiction under 28 U.S.C. § 1295(a)(5).

Page 16“We review Commerce’s decision using the same standard of review applied by the Court of International Trade.” Nucor Corp. v. United States , 927 F.3d 1243, 1248 (Fed. Cir. 2019). “Commerce’s determination will be sustained unless it is 'unsupported by substantial evidence on the record, or otherwise not in accordance with law.’” Yangzhou Bestpak Gifts & Crafts Co. , 716 F.3d 1370, 1377 (Fed. Cir. 2013) (quoting 19 U.S.C. § 1516a(b)(1)(B)(i)). Thus, we “must sustain ‘any determination, finding or conclusion’ found by Commerce unless it is ‘unsupported by substantial evidence on the record, or otherwise not in accordance with law.’” Fujitsu Gen. Ltd. , 88 F.3d 1034, 1038 (Fed. Cir. 1996) (quoting 19 U.S.C. § 1516a(b)(1)(B)). Substantial evidence is “such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.” Consolidated Edison Corp. v. Labor Board , 305 U.S. 197, 229 (1938).

On appeal, both Appellants – that is, the government and the Coalition – ask that we reinstate the 183.36% margin Commerce had assigned to Chengen as well as the separate-rate respondents in Commerce's second remand determination. The Coalition, but not the government, also requests that, regardless of our resolution of the first issue, we vacate Commerce’s fifth remand exclusion of two non-mandatory respondents from the all-others rate.⁴ We address the two issues in turn.

We affirm Commerce's determination to apply a 183.36% dumping margin to Chengen and to use this same number as the all-others rate. This disposition rests on the following conclusions: (1) Commerce did not abuse its discretion by taking into the record the two-page Conversion Chart and not the full 12-Page Document; and (2) substantial evidence supports Commerce’s finding that the intermediate input methodology is appropriate here.

Commerce verifiers discovered the two-page Conversion Chart during their site visit as part of the verification process. Specifically, Commerce verifiers noticed the Conversion Chart on a desk in the veneer storage area, and asked for a copy of the table; Chengen then produced the full 12-Page Document, that included the requested twopage Conversion Chart. At that point, Commerce verifiers rejected the additional ten pages and added to the record only the two-page Conversion Chart they had requested. Substantial evidence, including the sworn declaration of an agency official who was present during verification, supports Commerce’s finding as to how the verifiers discovered the Conversion Chart and their reasons for accepting from Chengen only the materials they expressly requested.

It is undisputed that Chengen did not provide the Conversion Chart, or any portion of the 12-Page Document, to

wide) rate are all the same: 183.36%. Mathematically, then, the weighted average of these rates, just like a simple average of them, yields the same rate: 183.36%. This renders it unnecessary to decide whether the simple average methodology is proper here. See Oral Argument at 8:4110:06 (Counsel for government agreeing that issue of separate rates only arises if we affirm Trade Court’s rejection of Commerce’s intermediate input methodology).

Page 18Commerce in responses to the original or supplemental questionnaires. It was only at verification, for the first time, that Chengen revealed that it used the two-page Conversion Chart, which might actually be the Chinese National Standard, to determine its poplar log input volume. Nor did Chengen, at any point prior to verification, even suggest that its method for determining poplar log input volumes was based on this chart, or on the purported Chinese National Standard. Instead, prior to verification, Chengen purported to use a process for determining poplar log volume inputs that depended on third-party invoices.⁵ Chengen failed, however, to produce such third-party documents.

In these circumstances, it was not an abuse of discretion for Commerce to take into the record at verification only the two-page Conversion Chart its verifiers noticed during their site visit and specifically requested, and toPage 19 decline to add to the record the additional portions of the 12-Page Document Chengen belatedly produced. We find no abuse of discretion even if Chengen is correct that Commerce’s verifiers tore the cover off the 12-Page Document, and that the cover by itself provided sufficient information to prove that Chengen, by using the two-page Conversion Chart, was following the Chinese National Standard.⁶ See J.A. 27 (Trade Court concluding that '[the] remaining pages provide context for understanding whether the conversion table for log volume in the initial two pages are part of the Chinese National Standard contained in the complete 12-page document”). It was within Commerce’s discretion to value finality, and deterrence of respondent noncompliance with document requests, over any potential reduction in accuracy that might result from rejecting the new, untimely additional pages offered by Chengen.

Our conclusion is consistent with the governing statutes, regulations, and caselaw, which provide Commerce broad discretion to create the record on which it will rely. “Judicial review of antidumping duty administrative proceedings is normally limited to the record before the agency in the particular review proceeding at issue.” QVD Food Co. v. United States , 658 F.3d 1318, 1324-25 (Fed. Cir. 2011). Because Commerce must predicate its decisions solely on the administrative record, it must take care as to what materials it places in that record, as well as when and how it does so. “Information that is submitted on a timely basis” must be sent to the other parties in the proceedingPage 20 “within a reasonable time” in order to permit them to “comment.” 19 U.S.C. § 1677m(g). Regulations set out the applicable deadlines for submission of factual information and specify that Commerce will reject untimely filed factual information. See 19 C.F.R. §§ 351.301(c)(1), 351.302(d), 351.104(a)(2)(i); see also MS Int’l, Inc. , 32 F.4th 1145, 1153 (Fed. Cir. 2022) ('Commerce may set such deadlines where the statute is silent, and must be permitted to enforce them in order to administer the trade remedy laws.”) (internal citations omitted).

“[T]he untimely submission of corrective information at verification results in a tension between finality and correct result.' Goodluck India , 11 F.4th at 1342 (internal quotation marks omitted). “[I]t is within Commerce’s discretion to decide which interest,” finality or accuracy, “outweighs the other on a case-by-case basis.” Id. at 1343.

Here, then, Commerce had discretion to determine that the appropriate balance between the interests of finality and accuracy warranted accepting the Conversion Chart into the record at the stage of verification while declining to further expand the record to include the remainder of the 12-Page Document. See Essar Steel Ltd. , 678 F.3d 1268, 1278 (Fed. Cir. 2012) (explaining that Commerce’s obligation to “administer the statute entrusted to it” provides it with discretion to reject untimely submitted information, in order to deter noncompliance and promote finality).

This is not to say that Commerce’s discretion at verification is unbounded. As we observed in Mid-Continental Steel & Wire, Inc. , 941 F.3d 530, 544-45 (Fed. Cir. 2019), “it can be unreasonable for an agency to refuse to obtain readily available, highly relevant information.” For example, Commerce may abuse its discretion “by refusing to accept updated data when there is plenty of time for Commerce to verify or consider it,” or when it “departs from a consistent practice without reasonablePage 21 explanation.” Goodluck India , 11 F.4th at 1342 (internal citations and brackets omitted). None of this occurred here. Commerce reasonably determined that, given the status of its investigation when the site visit took place, it lacked “plenty of time” to “verify or consider” the additional pages.

And Commerce provided a reasonable explanation for any perceived departure from its consistent practice. “Commerce has developed a practice of accepting new factual information at verification when (1) the need for that information was not evident previously, (2) the information makes minor corrections to information already in the record, or (3) the information corroborates, supports, or clarifies information already on the record.' TMK IPSCO , 179 F. Supp. 3d 1328, 1354 n.34 (Ct. Int’l Trade 2016) (internal quotation marks omitted). Commerce explained that it accepted the two-page Conversion Chart, but not the additional ten pages, because the Conversion Chart, and only the Conversion Chart, “called into question the accuracy of Chengen’s log purchase and consumption records, and its ability to substantiate such records.” J.A. 8540. Commerce determined that the use of the Conversion Chart demonstrates “Chengen’s derivation of log quantity is inherently imprecise” because it “relies only on the diameter of the smaller end of the log and its length,” which “unavoidably introduces inaccuracies to the reported volume.” J.A. 8541. Commerce further explained that, despite Chengen’s argument concerning the Chinese National Standard, “there is no evidence on the record that supports Chengen’s claim that the conversion table and formula used by Chengen elicits the log’s actual volume.” Id. Thus, contrary to the Trade Court’s finding, there was no record evidence that the additional ten pages discovered at verification 'corroborate, support, or clarify' information already in the record.

Our holding is further supported by our decision in Essar Steel , 678 F.3d 1268. There, the Trade Court had ordered ComPage 22merce to reopen the record and accept information about a manufacturing facility that the investigated entity had not disclosed in questionnaire responses. See id. at 1272. The Trade Court had required Commerce to include this untimely-produced information in the record despite finding that “Essar did not act to the best of its ability when responding to Commerce’s questions.” Id. at 1274-75. On appeal, we reversed the Trade Court. In doing so, we explained that respondents in an antidumping investigation bear the burden of creating an accurate record, yet Essar had 'failed to cooperate and gave false statements.” Id. at 1277; see also QVD Food , 658 F.3d at 1324 (“Although Commerce has authority to place documents in the administrative record that it deems relevant, the burden of creating an adequate record lies with interested parties and not with Commerce.”) (internal quotation marks and brackets omitted). We observed that while the Trade Court can “require that Commerce reopen and supplement the record” where the respondent-created record is plainly inaccurate or the product of fraud, those circumstances were not present. Essar Steel , 678 F.3d at 1278; see also Home Prods. Int'l, Inc. v. United States , 633 F.3d 1369, 1378 (Fed. Cir. 2011) (recognizing Commerce may be required to reopen record “where a party brings to light clear and convincing new evidence sufficient to make a prima facie case that the agency proceedings under review were tainted by material fraud”); Borlem S.A.-Empreedimentos Industriais , 913 F.2d 933, 937 (Fed. Cir. 1990) (same where Commerce’s decision is based on “inaccurate data' that purchasing party 'indicates are incorrect”). Hence, Commerce’s rejection of untimely-produced information from the respondent in Essar Steel was not an abuse of discretion.

The situation we confront today is quite similar. The burden to create an accurate record during the investigation was on Chengen. See generally J.A. 3425-30 (Coalition contesting, continuously from early in investigation,Page 23 Chengen’s allegedly deficient efforts to meet burden to create record of its factors of production). Chengen’s efforts to meet that burden failed, as the materials it timely provided Commerce mistakenly indicated that Chengen determined poplar log input volumes using third-party invoices, and did not reveal that Chengen (it now says) uses the two-page Conversion Chart (that may be part of a Chinese National Standard). The Trade Court did not find that Chengen’s responses to the questionnaires (though deficient and evidently incorrect) were materially fraudulent or resulted in a plainly inaccurate record. More importantly, the Trade Court did not find that any inaccuracy or fraud present in Chengen’s questionnaire responses would be corrected by taking into the record the 12-Page Document and not the Conversion Chart alone. Commerce, thus, acted within its discretionary authority.

Chengen’s contention that the Trade Court, and this court, can and should take judicial notice of the “fact” that the Conversion Chart is part of the Chinese National Standard is meritless. Federal Rule of Evidence 201 provides that “[a] court may take judicial notice of a fact only when it is either ‘generally known’ or ‘accurately and readily [discernible] from sources whose accuracy cannot reasonably be questioned.'' Apple Inc. v. Qualcomm Inc. , 992 F.3d 1378, 1384 (Fed. Cir. 2021) (quoting Fed. R. Evid. 201(b)); see also B.V.D. Licensing Corp. v. Body Action Design, Inc. , 846 F.2d 727, 728 (Fed. Cir. 1988) (“Courts may take judicial notice of facts of universal notoriety, which need not be proved, and of whatever is generally known within their jurisdictions.”). We have no basis on the record before us to hold that the Conversion Chart being part of the Chinese National Standard is “generally known” and 'cannot reasonably be questioned' within the United States (or any other relevant jurisdiction). Moreover, the record does not show Chengen having requested that Commerce or the Trade Court take judicial notice, or that Appellants were provided notice and an opportunity to object Page 24to such a request. See Fed. R. Evid. 201(e) (providing that judicial notice is improper when non-requesting party has not been provided opportunity “to be heard on the propriety of taking judicial notice and the nature of the fact to be noticed').⁷

In short, as we have previously observed, “[v]erification represents a point of no return” and serves to “test information provided by a party for accuracy and completeness.” Goodluck India , 11 F.4th at 1343-44. Like the respondent in Essar Steel , Chengen had multiple opportunities to provide Commerce with what it belatedly insisted was accurate information about how it calculates its poplar log input volumes. It did not do so. Rather than provide, in a timely manner in response to Commerce’s questionnaires, even the two-page Conversion Chart, or any portion of the 12-Page Document, Chengen repeatedly stated that it relied on third-party “material purchase invoices,” which at verification it was unable to produce.⁸ Under these circumstances, Commerce did not abuse its discretion byPage 25 balancing the considerations of finality and deterrence, on the one hand, against the interests of accuracy and completeness, on the other, by adding only the Conversion Chart to the record and rejecting Chengen's additional pages.

Accordingly, we reverse the Trade Court’s order that Commerce reopen the administrative record to include the entirety of the 12-Page Document provided by Chengen for the first time at verification.⁹

Having established that Commerce did not abuse its discretion in creating a record that includes the two-page Conversion Chart but not the remainder of the 12-Page Document, we now determine that this record contains substantial evidence supporting Commerce's decision to use the intermediate input methodology to measure Chengen’s factors of production.

In NME countries like China, Commerce calculates normal value – that is, the price the subject merchandise is sold at in the country of production – based on the factors of production. See 19 U.S.C. § 1677b(c)(1). Typically, this necessitates a finding as to the “quantities of raw materials employed' by a respondent. Id. § 1677b(c)(3). However, in Page 26some circumstances, including where a respondent cannot accurately account for its raw material consumption, Commerce will instead use an intermediate input, rather than raw materials, to assess factors of production.

Commerce has used an alternative intermediate input methodology in three cases to which the parties direct our attention: Honey from the People's Republic of China , 71 Fed. Reg. 34,893 (Dep’t Commerce June 16, 2006) (respondents unable to accurately record and substantiate complete costs associated with production); Fresh Garlic from the People's Republic of China , 71 Fed. Reg. 26,329 (Dep’t Commerce May 4, 2006) (respondents unable to record and substantiate costs of growing garlic); and Certain Frozen Fish Fillets from the Socialist Republic of Vietnam , 68 Fed. Reg. 4,986 (Dep’t Commerce Jan. 31, 2003) (factors of production misreported or underreported by respondents). In each of those prior situations, Commerce found the volumes of a raw material input to be unreliable.

Substantial evidence supports Commerce's finding that the same is true here for Chengen’s use of poplar logs. Commerce found Chengen's method unreliable because (i) it does not accurately reflect the actual cubic meter volume of the logs used, and (ii) the documents Chengen uses to record volume were under the exclusive control of Chengen, creating an opportunity for manipulation and misuse.

Commerce explained at length how Chengen’s process for measuring poplar volume 'is inherently imprecise.' J.A. 8540-41. It determined that Chengen typically measured the diameter at the thin end of the log and sometimes, for particularly asymmetric logs, it would measure diameter at both ends and then average them. J.A. 8540. This methodology, Commerce reasoned, would leave any taper in most logs “completely unaccounted for,” thereby “unavoidably introduc[ing] inaccuracies to the reported volume.' J.A. 8541. The agency further pointed out thatPage 27 Chengen used supplier-provided invoices to measure log volume for two other types of wood, birch and eucalyptus, yet did not do so for poplar.

From all this, Commerce concluded that Chengen could not “accurately report and substantiate its consumption of logs,' which 'is critical because an inaccurately reported log consumption quantity would undermine Chengen’s entire reporting methodology for the primary raw material in this case.' J.A. 8541. These considerations drove Commerce to decide to apply the intermediate input methodology and assess the cost of veneers, rather than rely on Chengen's potentially inaccurate self-reported volume of poplar logs.

Even assuming Chengen's Conversion Chart-based log-volume measurement methodology conforms to the Chinese National Standard, a reasonable mind could still agree with Commerce that such a methodology is inherently inaccurate, given the undisputed reality that logs are not uniform in shape and size. Further, as the Coalition points out, the 12-Page Document does not explain what assumptions underlie the Conversion Chart, providing another basis to reasonably conclude its use leads to unreliable results.

Chengen insists that its volumes were corroborated by third-party suppliers and, ultimately, Chinese tax authorities, leaving Commerce no valid basis to doubt the accuracy of the Conversion Chart and the Chinese National Standard. Chengen further asserts that the intermediate input methodology is applicable only in cases “where there were major costs missing from the surrogate financial ratios,” which did not (it contends) happen here. Commerce considered these contentions and rejected them. Substantial evidence supports Commerce’s doubts about the accuracy of Chengen's measurements notwithstanding that there may be evidence, including from third parties, that could have potentially persuaded Commerce otherwise, yet did not. See genPage 28erally Nippon Steel Corp. , 458 F.3d 1345, 1352 (Fed. Cir. 2006) (substantial evidence standard of review dictates that 'we must affirm [Commerce’s] determination if it is reasonable and supported by the record as whole, even if some evidence detracts from [Commerce’s] conclusion”). And we have no basis to limit Commerce’s discretion to use an intermediate input methodology only in situations in which major costs are otherwise missing, even if this is a characteristic of most situations in which Commerce has relied on this alternative methodology in past cases.

Chengen also argues there is no inherent “difficulty” in reporting log volumes, that Commerce has in other cases accepted log volume data like what Chengen offered here – including in Multilayered Wood Flooring from the People’s Republic of China , 76 Fed. Reg. 30,656 (Dep't Commerce May 26, 2011) (“ MLWF from China” ) – and Commerce has previously rejected application of an intermediate input method like the one applied here. Even accepting these contentions as true, they do not detract from our conclusion that substantial evidence, as described above, supports Commerce’s finding that Chengen’s poplar log data is unreliable. Commerce carefully assessed MLWF from China and persuasively distinguished it, for reasons including that the mandatory respondents there relied on supplierside invoices for log volume, items Commerce found were missing here. J.A. 8543-44. Moreover, while Chengen points to inaccuracies in Commerce’s method of measuring veneers, and contends they are worse than any inaccuracies that may plague Chengen's raw volume measurements, Commerce reasonably concluded otherwise. Commerce’s determination that Chengen’s “nominal measurements” for veneers “are based on the settings of a mechanical process . . . designed to produce consistent and reliable measurement results” is supported by substantial evidence, including that the veneers, unlike irregularlyPage 29shaped logs, are designed to be “rectangular and uniformly shaped.” J.A. 8544.

Therefore, Commerce’s use of the intermediate input methodology is supported by substantial evidence. We reverse the Trade Court’s determination to the contrary. Accordingly, we remand to the Trade Court for it to direct Commerce to apply a dumping rate of 183.36% to Chengen and the non-mandatory respondents.

The Trade Court approved Commerce's exclusion of two non-mandatory respondents, Jiangyang Wood and Dehua TB, from the all-others rate based on those entities achieving voluntary respondent status. So do we.¹⁰

Non-mandatory respondents may voluntarily request individual investigation and exclusion from the all-others rate. The governing statute requires that Commerce “shall establish an individual . . . weighted average dumping margin for any exporter or producer not initially selected for individual examination” if: (1) the individual exporter or producer timely submits 'the information requested from exporters or producers selected for examination [i.e., mandatory respondents];” and (2) the number of such companies is not so large so as to burden or delay the investigation of the individually examined companies. 19 U.S.C. § 1677m(a)(1). This provides Commerce discretion to “decline to fully investigate the voluntary respondent” where either of the two statutory conditions have not been met, or where the voluntary respondent has “submitted such information [that] is so large that individual examination of such exporters or producers would be unduly burdensome Page 30and inhibit the timely completion of the investigation.' Yangzhou Bestpak , 716 F.3d at 1373 (citing 19 U.S.C. § 1677m(a)(2)).

Commerce granted the individual rate requests of Jiangyang Wood and Dehua TB. The Trade Court affirmed its decision to do so, stating, “Commerce provided a reasonable explanation to justify its exclusion from the Order the voluntary applicants who submitted timely requests for voluntary respondent treatment with hundreds of pages of questionnaire responses and supporting documentation, providing the same information that mandatory respondents submitted by the same deadline in accordance with 19 U.S.C. § 1677m(a).” J.A. 82.

On appeal, the Coalition argues it was error to exclude Jiangyang Wood and Dehua TB from the all-others rate because the “exclusion was inadequately explained and supported.” Coalition Br. at 37. The Coalition relies largely on Passenger Vehicle & Light Truck Tires from the Socialist Republic of Vietnam , 86 Fed. Reg. 28,559 (Dep’t Commerce May 27, 2021) (' Tires '), in which Commerce did not exclude voluntary respondents even when they submitted timely materials. Dehua TB responds that Tires did not result in a final antidumping order and is not binding on Commerce.

Contrary to the Coalition’s argument, nothing in Tires , a preliminary determination by Commerce that binds neither Commerce nor us, suggests the Trade Court committed reversible error. Instead, as Commerce recognized, the investigation in Tires “was terminated by the International Trade Commission and did not result in an antidumping order;” the preliminary exclusion determination 'was never implemented or subjected to judicial review.” J.A. 569. There is nothing inconsistent, then, between the Tires non-final action and what Commerce did here. Additionally, “[w]e have rejected the notion that Commerce is forever bound by its past practices. Instead, eachPage 31 administrative review is a separate exercise of Commerce’s authority that allows for different conclusions based on different facts in the record.” Hyundai Elec. & Energy Sys. Co. , 15 F.4th 1078, 1089 (Fed. Cir. 2021) (internal quotation marks and citations omitted).

Commerce adequately explained its reasoning for excluding Jiangyang Wood and Dehua TB from the all-others rate. It explained that these two voluntary respondents timely provided all of the same information that the two mandatory respondents, Chengen and Bayley, submitted, and therefore satisfied the statutory requirements of § 1677m(a). Commerce reasoned that Jiangyang Wood and Dehua TB had submitted “hundreds of pages of questionnaire responses and supporting documentation, as well as sales and factor of production data bases,” which provided Commerce all the information it needed to “potentially select them as voluntary respondents and complete the investigation without undue delay,' warranting exempting them from the all-others rate. J.A. 568-69.

Accordingly, we affirm the Trade Court’s decision upholding Commerce’s exclusion of Jiangyang and Dehua TB from the all-others rate.

We have considered the parties’ remaining arguments and find they lack merit. Accordingly, the judgment of the Trade Court is affirmed-in-part, reversed-in-part, and remanded for further proceedings consistent with this opinion.

Costs to Appellants.