Slip Op. 25-
UNITED STATES
COURT OF INTERNATIONAL TRADE
Court No. 22-00152
COALITION OF AMERICAN MANUFACTURERS OF MOBILE ACCESS EQUIPMENT, Plaintiff,
v.
UNITED STATES,
Defendant,
and
ZHEJIANG DINGLI MACHINERY CO., LTD., Defendant-Intervenor.
Before: M. Miller Baker, Judge
OPINION
[Sustaining Commerce’s remand redetermination.]
Dated: December 11, 2025
Ned H. Marshak, Dharmendra N. Choudhary, and Jordan C. Kahn, Grunfeld, Desiderio, Lebowitz, Silverman & Klestadt LLP, New York, NY, on the comments for Defendant-Intervenor.
Tara K. Hogan, Assistant Director, and Kristin E. Olson , Trial Attorney, Commercial Litigation Branch, Civil Division, U.S. Department of Justice, Washing-
ton, DC, on the comments for Defendant. Of counsel for Defendant was Ruslan Klafehn, Attorney, Office of the Chief Counsel for Trade Enforcement & Compliance , U.S. Department of Commerce, Washington, DC.
Timothy C. Brightbill, Laura El-Sabaawi, and Theodore P. Brackemyre, Wiley Rein LLP, Washington, DC, on the comments for Plaintiff.
Baker, Judge: This dumping case involving self-propelled lifting machines from China returns after a remand on two issues. See Slip Op. 24-66, 2024 WL 2796654 (CIT 2024).
The first involved the surrogate value for ocean freight charges. As relevant here, the court instructed the Department of Commerce to explain why it was permissible for Zhejiang Dingli Machinery Co. (Dingli) to designate its shipping routes as business proprietary information. Slip Op. 24-66, at 7, 2024 WL 2796654, at **2–3. If the agency found it impermissible , it was to allow the company to redesignate the information as public and then give the Coalition of American Manufacturers of Mobile Access Equipment the corresponding opportunity to resubmit its proposed surrogate value data as public. Id. at 7–8, 2024 WL 2796654, at *3. Relatedly, the court instructed the Department to reconsider its finding that the shipping routes reflected in the Descartes, Freightos, and Drewry information are more specific to Dingli’s routes than the Maersk data and to “reconcile” its finding that the latter set of pricing information is not publicly available with contrary past pronouncements. Id. at 10–13, 2024 WL 2796654, at **4–5.
3
For the second issue, Commerce was to reconsider internal inconsistencies in its valuation of certain of Dingli’s inputs, “minor fabricated steel components.” Id. at 16–20, 2024 WL 2796654, at **6–7.
I
A
On remand, the agency found that Dingli’s designation of its shipping routes as business proprietary information “may have been an impermissible characterization .” Appx22052 (citing 19 C.F.R. § 351.105(c)(3), (5), and (6)). The company redesignated the information as public. Appx22052–22053. The Coalition did the same with the Maersk data. Appx22053.
The Department found that the Maersk figures were publicly available ocean freight prices and, based on past decisions, a reliable source of information. Id. It therefore compared them to Descartes, Drewry, and Freightos to determine the “best available information ” for valuing ocean shipping.
Commerce noted that Dingli reported using two routes to ship its goods during the period of investigation —Shanghai to New York and Shanghai to Oakland , with the latter employed for most of the company ’s relevant sales. Appx22053–22054. Maersk included pricing for Shanghai–Oakland. But Descartes was limited to Shanghai–New York; Drewry included that route along with Shanghai–Los Angeles and Shanghai–Houston; and Freightos included Shanghai –Long Beach and Ningbo–New York. Appx22054.
4
Stating that its practice is to use ocean freight surrogate values that are “specific to the port combinations a respondent uses for its shipments of merchandise under consideration,” the agency found Maersk superior because it was the only source with pricing for the same route Dingli used for most of its shipments. Appx22054–22055.
Commerce also found that Maersk included “all incidental ocean freight charges” and that the figures covered “export service fees, terminal handling service fees (origin), documentation fees, bunker adjustment factor fees, low sulfur surcharge, and terminal handling service fees (destination)” as part of the rates. Appx22055. The other three datasets, in contrast, “offer [ed] no accompanying rate details (or accompanying information explaining the methodology for price data collection and reporting),” such that it could not “be determined whether the ocean freight rates in the Descartes , Freightos, and Drewry information include all, or any, ocean freight incidental charges.” Id.
In response to Dingli’s objection that the Maersk data were unreliable because they were based on price quotes, rather than actual transactions, the Department pointed to instances when it gave preference to price quotes over other sources’ actual transactions when data-quality considerations weighed more heavily . Appx22061–22063. It “determined that route specificity should be accorded more weight . . . because only one data subset proffered on the record is actually specific to Dingli’s shipping routes.” Appx22063.
5
That conclusion segued directly into the agency’s answer to the company’s objection to route specificity. While Dingli did not dispute that the Maersk routes were more specific than the ones it proffered, it raised other objections. The company claimed that its alternative routes were of similar distances such that “freight charges should be nearly equal,” that Shanghai –Los Angeles and Shanghai–Long Beach are “nearly as route specific as the Maersk data,” and that any difference between charges for shipment to the East and West Coasts is “insignificant.” Id. (emphasis in original). Commerce rejected those arguments, emphasizing that its practice is to prefer ocean freight figures “that are specific to the port combinations a respondent uses for its shipments of merchandise under consideration.” Id. So it did not analyze Dingli’s assertions about whether other routes “should be nearly equal.” Id. (emphasis removed).
The agency also disagreed with Dingli’s contention that the Maersk data were distortive because they double-counted certain costs already accounted for on the record by the World Bank’s Doing Business 2020 (Brazil) report. It explained that this source “likely” did not include the charges reflected in the Maersk information . Appx22065.
Finally, Commerce disagreed with Dingli that container weight is vital. It found that the Maersk data “are for a specified quantity of goods as opposed to a theoretical ‘container load’ of merchandise” and that the 18,000-kilogram weight Maersk used was both “within range for a 20-foot container capacity” and “very close” to the lowest 40-foot container capacity,
6
based on figures in the record. Appx22067. The agency also observed that because the record showed that a 40-foot container might carry a payload as low as 20,000 kg, the 18,000 kg figure was “not so anomalous as to be unreliable.” Appx22068. And it noted that Dingli cited no evidence that Commerce had ever declined to use Maersk figures because of low container weights. Id.
B
Dingli now challenges Commerce’s decision to use the Maersk price quotes instead of the other sources’ transaction-based pricing. ECF 92, at 4–13. The company contends that “this distinction” was “a sufficient standalone reason to reject” the former data. Id. at 8.
That might well be, but the Department—not the court—is the entity authorized to make that discretionary determination. It explained that a source’s use of price quotes “is not dispositive as to [its] reliability.” Appx22061–22062. On this record, it found that route specificity was the more important consideration. Appx22063.
The agency’s conclusion was both lawful and supported by substantial evidence. While this court has remanded the use of price quotes when the agency failed to explain why it favored them, it has also held that using them can be reasonable. See Jiangsu Zhongji Lamination Materials Co. (HK) v. United States, Slip Op. 23-84, at 28–29, 2023 WL 3863201, at *10 (CIT 2023) (citing Changzhou Trina Solar Energy Co. v. United States, 492 F. Supp. 3d 1322, 1330
7
(CIT 2021), and Changzhou Trina Solar Energy Co. v. United States, 532 F. Supp. 3d 1333, 1336–37 (CIT 2021)).
Dingli contends that Jiangsu Zhongji is “distinguishable based on its facts” because there, the non- Maersk datasets included Chinese figures that could reflect subsidies. ECF 92, at 10–11. But that misses the point—weighing the evidence is a task for Commerce , not this court. See Slip Op. 23-84, at 29, 2023 WL 3863201, at *10. 1 To the extent there are competing agency preferences—one for actual transactions, the other for route specificity—the Department may choose between them as long as it explains its reasoning . See Catfish Farmers of Am. v. United States, Slip
1 In the Changzhou Trina cases, in turn, the court did not
remand because the use of price quotes was impermissible, but rather because the agency did not explain why it deemed it reasonable to use them. See 492 F. Supp. 3d at 1330 (remanding the use of Maersk on that basis); 532 F. Supp. 3d at 1337 (sustaining the remand results after Commerce found Xeneta “more specific to the shipping routes used by” the respondent). Similarly, while Dingli disparages the agency’s citation of Jinko Solar Import and Export Co. v. United States, 701 F. Supp. 3d 1367 (CIT 2024), and argues that because Commerce used both Descartes and Maersk data there it cannot reject Descartes here, see ECF 92, at 12–13, the Jinko court made two important observations. First, “[a]lthough the Maersk data [do] not reflect actual transactions, it is reasonably apparent that Commerce viewed the data as reliable.” 701 F. Supp. 3d at 1385. The same is true here. Second, the agency relied on both datasets for route-specificity reasons. Id. at 1385–86. If anything, its emphasis on route specificity in those cases suggests that its heavily weighting that factor here is consistent with past practice.
8
Op. 25-24, at 13, 2025 WL 748133, at *5 (CIT 2025) (“Such weighing of preferences is the agency’s prerogative .”). That is what it did here.
Dingli also objects to Commerce’s route-specificity finding. It argues that the other datasets “are nearly as route-specific as Maersk’s Shanghai–Oakland route.” ECF 92, at 14 (emphasis added). The company complains that Commerce “summarily dismisse[d] this argument” because Maersk included the actual route the company used. Id. at 14–15. But the agency did no such thing. It reasonably explained that its practice is to “use ocean freight surrogate value data that are specific to the port combinations a respondent uses for its shipments of merchandise under consideration .” Appx22054. The Department’s finding as to route specificity is thus supported by substantial evidence .
Turning to brokerage and handling expenses, the company essentially contends that in previous proceedings , Commerce excluded various parts of Maersk data to avoid double-counting amounts reported elsewhere . Dingli complains that the Department failed to “distinguish” the World Bank Doing Business (2017) Thailand report used in the 2018 Solar Cells from China proceeding, 83 Fed. Reg. 35,616, “from the corresponding data reported in the Doing Business 2020 (Brazil) report” on the record here. ECF 92, at 22.
But the agency addressed that issue, observing that while the Thailand report may have allowed it to determine that the brokerage and handling value included the incidental freight fees, the Brazil report on
9
the record lacked information needed for such a finding . Appx22066–22067. While the company insists that “both reports provide substantially the same cost elements,” ECF 92, at 22, Commerce explained why it disagreed, see Appx22066–22067. That factual finding is supported by substantial evidence, even if the court were to agree with Dingli’s reading of the record.
Finally, the company argues that Commerce overlooked “a critical deficiency underlying the Maersk price quotes” because they all assumed a load of 18,000 kg in 40-foot dry containers. It asserts, “[T]his payload is unrealistically low as compared to the prescribed industry standard payload of 28,000 kg for a 40-foot container , published by Maersk itself.” ECF 92, at 24 (citing Appx6621–6622). It also argues that in past proceedings , “for Maersk price data, Commerce used an even higher container weight—28,870 kg (for 40 foot)—to value ocean freight.” Id. at 25 (citing Appx18593, Appx18596). “These facts render the anemic 18,000 kg data printed on the Maersk ocean freight quotes as anomalous and unreliable.” Id.
The government responds that Dingli conflates payload with container capacity and assumes, without evidence, that it is possible to load every container to its maximum capacity. 2 It argues that the company’s only support for its claimed industry standard payload for a 40-foot container is “a citation to the Coalition’s
2 The government offers the useful example of one cubic
foot of feathers versus one cubic foot of lead. They both take up the same amount of space in the container, but the lead weighs considerably more. ECF 94, at 26 n.3.
10
administrative case brief which provided a survey of various container capacities.” ECF 94, at 27–28 (emphasis in original) (citing ECF 92, at 24). The Coalition agrees, noting that Commerce was clear that the Maersk data were for a specific quantity of goods, rather than some theoretical volume, and asserts that “it is not inconsistent or anomalous for actual container loads to have lower weights than the theoretical maximum of container freight data.” ECF 93, at 11–12. Tying both points together, the government emphasizes that the agency must base its decision on the record and not on “theoretical assumptions about the payload .” ECF 94, at 28.
As discussed above, the Department found that the 18,000 kg figure was used to establish “a specified quantity of goods,” rather than speculating on what “a theoretical ‘container load’ of merchandise might be,” and it noted that a 40-foot container might carry a payload as low as 20,000 kg. Appx22067–22068. It said that Dingli provided no evidence in the original proceeding to show that Maersk’s container weights were unreliable and cited nothing on remand to merit a different result. Appx22068 (citing Appx1048).
The agency’s explanation is reasonable. Despite the company’s characterization of the Maersk data as “unreliable , anomalous, potentially distortive, and qualitatively inferior,” ECF 92, at 27, the Department explained why it rejected each of Dingli’s arguments and selected that dataset as the best available information on the record. The court sustains the use of Maersk as a surrogate value for ocean freight costs.
11
II
A
The other issue is the valuation of minor fabricated components. The court remanded because the agency’s analysis was internally inconsistent. Commerce first said the Coalition offered no evidence to show that Dingli’s suppliers fabricated steel plate into various components. See Appx1072. But it then said that while sometimes the suppliers do minimal work, “in some cases, they may do more in-depth fabrication or processing .” Id. The court directed the Department to resolve the contradiction. Slip Op. 24-66, at 19–20; 2024 WL 2796654, at *7.
On remand, the agency conceded the inconsistency and framed the issue as “whether 16 inputs [it] valued using raw steel categories under [Harmonized Tariff Schedule (HTS)] chapters 7208, 7211, and 7214 were products purchased by Dingli as minor fabricated steel parts or raw steel, i.e., plate, sheet, bar, rod.” Appx22056–22057. It found that the record showed the suppliers “fabricated” the inputs to varying degrees , based on the company’s specifications, for later incorporation into finished machines. As a result, the components were no longer raw steel. Appx22057. The company’s submissions confirmed that, but it disagreed that the suppliers’ processing and fabrication work had any significance. Id. Commerce found it essential that “Dingli relied on evidence summarizing ‘the degree of fabrication or processing’ for all sourced material inputs of production.” Id. “Once a set of fabrication processes dedicated to a specific outcome is
12
performed on the purchased primary steel, [it] is no longer usable to customers other than Dingli and is only usable for incorporation into” the company’s lift machines. Id.
The Department therefore concluded that the use of HTS subheadings “that reflect base steel material,” as Dingli advocated, was improper because fabrication transforms the raw steel into something else, resulting in “further value added.” Appx22057–22058. The agency selected subheading 8341.20.90, “Parts suitable for use solely or principally with machinery of heading 8427, other than Fork lift trucks.” Id.
The company objected and argued that Commerce had followed a strict Customs classification analysis, rather than selecting a subheading that would yield a more accurate value. Appx22069. The Department again disagreed, emphasizing that because the inputs were no longer raw steel, it was inappropriate to value them using HTS chapters covering that material. Appx22069–22070. It rejected the company’s reliance on two Federal Circuit cases, 3 finding they simply hold that the agency may employ Customs classifications as part of its analysis but is not required to do so. Appx22071. Commerce found that any code that did not cover minor fabricated steel components would be inappropriate, including those covering raw steel (which the inputs no longer were). Appx22073. It also
3 Downhole Pipe & Equip., L.P. v. United States, 776 F.3d
1369, 1378–79 (Fed. Cir. 2015), and SolarWorld Americas, Inc. v. United States, 910 F.3d 1216, 1223–25 (Fed. Cir. 2018).
13
observed that, in the pre-remand administrative briefing , Dingli had cited 8431.20.90 as the appropriate classification for “parts (i.e., not base materials) used in the production of subject merchandise” because the fact that the parts were dedicated for a particular purpose was “more important” than their composition. Id. (quoting Appx6675).
Finally, Commerce rejected Dingli’s argument that 8431.20.90 “is somehow discredited” because the average unit value under that heading is higher than the averages for the same inputs in the final determination . Appx22074. It observed that the company placed market-economy prices for fabricated components used in its machinery on the record “to corroborate the recommended [HTS] numbers and the corresponding [average unit values] for various fabricated component inputs.” Appx22075. The Department found that the average unit value was corroborated with market- economy prices for similar fabricated components, which it said was another reason to use 8431.20.90. Id.
B
Dingli begins its argument with a chart showing the differences in valuation between the final determination and the remand results for the 16 inputs in question. All of them increased between 461–680 percent . ECF 92, at 28. The company objects that the Department ’s revised analysis is “unsupported by substantial record evidence and controlling judicial precedent .” Id. It argues that the agency elevated adherence to classification principles over “normal practice
14
that favors [an] HTS subheading yielding a more reliable , representative, and accurate value.” Id. at 30.
Dingli’s theme is that 8431.20.90 might be appropriate for inputs that underwent major fabrication, but it results in significant overvaluation for those subject to minor changes like cutting and bending. Id. at 33–34. The company asserts that “simple machining operations” do not result in the “outsize value addition ” assigned by 8431.20.90 and that headings for raw steel plate, sheet, bar, and rod more “reasonably approximate ” the inputs’ actual value. Id. at 34–35. It also argues, again, that Downhole Pipe and Solar- World Americas require the agency to follow some approach other than “conducting a rigid classification analysis.” Id. at 30–32.
The government responds that Dingli ignores that the components are no longer usable as raw steel. It emphasizes Commerce’s finding that the reason for the fabrication work is to render the inputs fit “for a specific purpose,” such that they “have ceased to become ‘primary steel.’” ECF 94, at 50 (citing Appx22057–22058). “Dingli points to no flaws with HTS 8431.20.90 other than its unsupported belief that the value is too high.” Id. at 49.
The Coalition notes the Department originally used HTS numbers for primary or raw steel because it found the group had not supported its claim that Dingli was purchasing anything other than raw steel—the finding the court remanded as inconsistent. ECF 93, at 15 (citing Appx22056). And it points out that, on remand, the agency cited the company’s own
15
acknowledgment that the inputs underwent fabrication or processing. Id. at 16–17 (citing Appx22057). The Coalition emphasizes that Dingli has not challenged Commerce’s finding, observing that if there were “any evidence on the record that these components were actually raw steel—and thus should be valued as such—then, presumably, [the company] would have made that argument.” Id. at 18.
The points made by the government and the Coalition are persuasive. Commerce reasonably explained that fabrication of a component adds value, so it is no longer a raw input and should not properly be valued as one. Appx22058. Dingli does not really dispute that point, instead quibbling with the amount of value added. Thus, it has tacitly admitted that the raw steel headings result in undervaluation. 4 But rather than suggesting an alternative, it simply says the undervaluation was the “most accurate” surrogate value. ECF 92, at 35. Essentially, Dingli asks the court to substitute its judgment for the agency’s. The court declines.
Downhole Pipe and SolarWorld Americas do not dictate otherwise. The former emphasizes the agency’s “discretion to determine what constitutes the best available information, as this term is not defined by statute.” 776 F.3d at 1375. While Dingli is correct that
4 Dingli’s emphasis on the average unit values tied to the
candidate headings fails for a similar reason. As the Coalition observes, the company’s argument ignores the Department ’s finding “that a fabricated component is necessarily going to have a higher value than its underlying raw material .” ECF 93, at 20 (citing Appx22074).
16
Downhole Pipe states that the Department need not follow the classification principles Customs and Border Protection uses, see id. at 1379, the flip side is also true. Nothing in that opinion prohibits Commerce from using such principles if it concludes they would yield the most accurate result. The point is that the agency must explain its rationale such that “a reasonable mind could conclude that [it] chose the best available information.” Id. (quoting Zhejiang DunAn Hetian Metal Co. v. United States, 652 F.3d 1333, 1341 (Fed. Cir. 2011)). And SolarWorld Americas confirms that a subheading may be the best available even if it is not a perfect match. See 910 F.3d at 1223 (citing Home Meridian Int’l Inc. v. United States, 772 F.3d 1289, 1296 (Fed. Cir. 2014) (“The data on which Commerce relies to value inputs must be the ‘best available information,’ but there is no requirement that the data be perfect.”)).
Finally, the Department observed that Dingli’s administrative briefing argued that 8431.20.90 is appropriate for valuing factors of production “that are parts (i.e., not base materials) used in the production of subject merchandise.” Appx22073 (quoting Appx6675). The company responds that its argument was based on valuing “major customized fabricated subassemblies .” ECF 92, at 34 (emphasis added).
But its argument actually swept more broadly. 5 It recognized that 8431.20 “trumps a basket heading
5 It is true that the cited discussion appeared in a section
of the company’s case brief titled “Fabricated Steel (footnote continues on next page)
17
based on component material, since ‘the fact that the parts were dedicated for use in the subject merchandise is more important than the fact [they] are composed of certain materials.’” Appx6675 (ellipses and brackets omitted) (quoting Vertical Shaft Engines from China I&D Memo at 39 (Mar. 5, 2021), accompanying 86 Fed. Reg. 14,077). 6 The Department, in turn, followed that principle in finding that “once a set of fabrication processes . . . dedicated to a specific outcome is performed on the purchased primary steel, [it] is no longer usable to customers other than Dingli and is only usable for incorporation into” the company’s lift machines. Appx22074. Dingli has not explained why that rule should apply to “major” components but not “minor” ones.
In sum, while the company contends—as did the Downhole Pipe appellants, see 776 F.3d at 1380—that the value Commerce selected is too high, its solution is to insist that the Department use subheadings the agency found inappropriate. But because Commerce gave a reasonable explanation of why those
Components,” see Appx6668, and that a separate section addressed “Minor Fabricated Steel Components,” see Appx6686. The latter advocated for treating the minor components as raw steel. Appx6687. 6 Vertical Shaft Engines made the point even more strongly
than Dingli suggested. It said that “in cases similarly involving machinery made from numerous parts and materials , Commerce has determined that the fact that the parts were dedicated for use in the subject merchandise was more important than the fact that the parts were composed of certain materials.” Vertical Shaft Engines I&D Memo at 38–39.
18
subheadings were unsatisfactory, the court must sustain its decision as supported by substantial evidence. The agency reasonably found it important not to value components as something they are not (raw steel). That decision was its to make.
* * *
The court sustains Commerce’s redetermination. A separate judgment will issue. See USCIT R. 58(a).
Dated: December 11, 2025 /s/ M. Miller Baker New York, NY Judge