In the latest chapter in this long-running battle over the United States Department of Commerce's assessment of antidumping duties against Mitsubishi Heavy Industries ("MHI") and Tokyo Kikai Seisakusho ("TKS") for their United States sales of large newspaper printing presses ("LNPPs"), TKS appeals from the final judgment of the Court of International Trade affirming the dumping determination. On appeal, TKS contests the Department of Commerce's determination that Japanese market LNPPs are a foreign like product under 19 U.S.C. 1677b(e)(2)(A). Because we conclude that the Department of Commerce's determination was supported by substantial evidence, and because TKS's allegations regarding the agency's statutory construction are not properly before us, we affirm.

* BACKGROUND

This case involves large newspaper printing presses exported to the United States from Japan. Although all LNPPs have similar design and function, individual LNPPs are custom-made per the customer's specification. The companies provide their customers with a menu of various components that can be built into the machine, and the customer decides what components to order. As a result, individual orders for LNPPs can vary to a greater or lesser extent, depending on what components the customer chooses. Because Japanese and United States newspapers have somewhat different characteristics in terms of size, use of color, etc., the LNPPs used to produce them also have somewhat different components. Thus, every contract for sale of an LNPP contains different terms-including price terms-because the LNPPs themselves have different components from contract to contract.

Upon a petition by Rockwell Graphics Systems, Inc., a U.S. competitor now known as Goss Graphics Systems, Inc. ("Goss"), the Department of Commerce ("Commerce") launched an antidumping investigation of two manufacturers, MHI and TKS. In due course, Commerce issued its final antidumping determination finding sales at less than fair value and announcing a dumping margin of 56.28 percent for TKS, the appellant here. Large Newspaper Printing Presses and Components Thereof, Whether Assembled or Unassembled, From Japan, 61 Fed. Reg. 38,139 (Dep't Commerce, July 23, 1996) ("Japan Final"), amended by 61 Fed. Reg. 46,621 (Dep't Commerce, Sept. 4, 1996) (antidumping duty order and amendment to final determination). In Japan Final, Commerce used constructed value ("CV") to calculate the dumping margin, see Japan Final, 61 Fed. Reg. at 38,140, and it used home market (i.e., Japanese) LNPPs as the foreign like product in its determination of profit, which is one component of CV, see 19 U.S.C. 1677b(e)(2) (1994), despite having earlier found that direct price-to-price comparisons with home market LNPPs were impracticable as a basis for normal value-a finding that led to its original decision to use CV as a basis for normal value.1 See Japan Final, 61 Fed. Reg. at 38,146.

TKS and MHI appealed numerous aspects of Commerce's determination in Japan Final, including its foreign like product determination. See Mitsubishi Heavy Indus., Ltd. v. United States, 15 F. Supp. 2d 807, 810, 828 (Ct. Int'l Trade 1988) (Mitsubishi I). TKS, in particular, argued that Commerce's reliance upon 19 U.S.C. 1677b(e)(2)(A) to calculate profit was inappropriate because "the findings that led Commerce to rely on CV rather than home-market sales in calculating normal value constitute[d] evidence that no foreign like product exist[ed] in the home market." Mitsubishi I, 15 F. Supp. 2d at 828-29. The profit calculation under 1677b(e)(2)(A) relies upon sales of "a foreign like product." 19 U.S.C. 1677b(e)(2)(A) (1994). Because Commerce did not describe adequately its profit calculation so as to permit judicial review, the Court of International Trade remanded the case to Commerce to explain upon which of the three statutory definitions of foreign like product it relied to make its profit calculation. Mitsubishi I, 15 F. Supp. 2d at 829. In its remand determination, Commerce explained that it had relied upon the definition of foreign like product in 19 U.S.C. 1677(16)(C), which requires, inter alia, that the foreign like product be merchandise that "the administering authority determines may reasonably be compared with" the exported merchandise subject to the investigation. 19 U.S.C. 1677(16)(C)(iii) (1994).

TKS and MHI appealed the remand determination, and the Court of International Trade remanded again, this time because Commerce failed to explain the factual basis for its determination that the LNPPs sold in Japan and the United States could "reasonably be compared" as required by 19 U.S.C. 1677(16)(C)(iii). Mitsubishi Heavy Indus., Ltd. v. United States, 54 F. Supp. 2d 1183, 1197 (Ct. Int'l Trade 1999) (Mitsubishi II). The Court of International Trade was troubled because in its first remand determination, Commerce made statements that made it appear that it had previously conducted a difmer analysis2 and concluded that the home market and export LNPPs could not reasonably be compared. See id. at 1197. In its second remand determination, Commerce clarified that it had not conducted a difmer analysis. Second Remand Determination at 2-3. In addition, Commerce explained the factual basis for its finding that the home-market LNPPs could "reasonably be compared" with their United States counterparts, which included the common use to which the products are put (i.e., printing newspapers) and TKS's and MHI's responses to detailed questionnaires showing that the Japanese and United States LNPPs share the same set of detailed press characteristics. Id. at 11-12.

Based on Commerce's explanation of the factual basis underlying its comparability determination, the Court of International Trade affirmed the dumping determination. Mitsubishi Heavy Indus., Ltd. v. United States, 97 F. Supp. 2d 1203, 1209 (Ct. Int'l Trade 2000) (Mitsubishi III). The court denied TKS's motion for reconsideration, Mitsubishi Heavy Indus., Ltd. v. United States, 112 F. Supp. 2d 1170, 1175 (Ct. Int'l Trade 2000) (Mitsubishi IV), and this appeal by TKS followed. We exercise jurisdiction over this appeal from a final decision of the United States Court of International Trade pursuant to 28 U.S.C. 1295(a)(5).

* We review a decision of the Court of International Trade evaluating an antidumping determination by Commerce by reapplying the statutory standard of review that the Court of International Trade applied in reviewing the administrative record. Micron Tech., Inc. v. United States, 117 F.3d 1386, 1393 (Fed. Cir. 1997). We will uphold Commerce's determination unless it is "unsupported by substantial evidence on the record, or otherwise not in accordance with law." 19 U.S.C. 1516a(b)(1)(B)(i) (1994); see Micron Tech., 117 F.3d at 1393.

On appeal, TKS primarily argues that Commerce's determination that home and United States market LNPPs may reasonably be compared is not supported by substantial evidence. We note that in pursuing this argument, TKS has chosen a course with a high barrier to reversal. The Supreme Court has defined substantial evidence as "such relevant evidence as a reasonable mind might accept as adequate to support a conclusion." Universal Camera Corp. v. NLRB, 340 U.S. 474, 477 (1951) (quoting Consol. Edison Co. v. NLRB, 305 U.S. 197, 229 (1938)). The conclusion reached by Commerce need not be the only one possible from the record, for "[e]ven if it is possible to draw two inconsistent conclusions from evidence in the record, such a possibility does not prevent Commerce's determination from being supported by substantial evidence." Am. Silicon Techs. v. United States, 261 F.3d 1371, 1376 (Fed. Cir. 2001); see also Consolo v. Fed. Maritime Comm'n, 383 U.S. 607, 620 (1966). After reviewing the record, we conclude that substantial evidence supports Commerce's determination that home-market LNPPs are a foreign like product.

In its second remand decision, Commerce clarified the evidence underlying its decision to use home-market LNPPs as the foreign like product, explaining that "TKS's home market LNPP may reasonably be compared to its sales of LNPP in the United States based on evidence that LNPP in both markets share detailed product characteristics." Second Remand Determination at 2. Commerce noted that its conclusion was further "supported by the common use-to produce newspapers-to which both home market and U.S. LNPP are employed." Id. at 11. During the investigation, both TKS and MHI responded to a questionnaire sent by Commerce asking them to identify both United States and home-market LNPPs using the same set of detailed press characteristics. Id. TKS's and MHI's responses to this questionnaire, which indicated that their United States and home-market LNPPs do in fact share a majority of the same-or highly similar-characteristics, provide the principal factual predicate for Commerce's finding. TKS argues that this evidence is "self-serving" because Commerce prepared the questionnaire itself, forcing TKS and MHI to describe their Japanese and United States products using the same characteristics. To the extent that TKS accuses Commerce of stacking the deck against them, its argument is not well taken. As the agency to which Congress delegated the authority to determine antidumping duties, Commerce is responsible for gathering information to make dumping determinations. Commerce uses the information it collects in order to reach its decision-in this case that the home-market and United States LNPPs are reasonably comparable. Although Commerce is an agent of the United States government, it nevertheless makes its dumping determination based on an impartial analysis of the evidence. Furthermore, administrative acts by Commerce enjoy a presumption of regularity that includes, in this case, impartiality in its decision-making process, and one seeking to rebut that presumption carries a heavy burden. See Skinner v. United States, 594 F.2d 824, 830 (Ct. Cl. 1979). There is no evidence to suggest that Commerce made up its mind in advance and cunningly planned its questionnaire to support its position.

MHI's and TKS's responses to Commerce's information-gathering request provide ample support for Commerce's finding.3 First, the questionnaire responses confirm "that the LNPP sold in Japan and the LNPP sold in the United States share the detailed press characteristics that [Commerce] set out in its questionnaire." Second Remand Determination at 11. And within each characteristic, the responses indicated that the individual specifications for each press characteristic were also similar. Obviously, because the LNPPs are custom-made, each individual LNPP may contain a different mix of these common characteristics. However, it is apparent that they all reflect a choice from among similar characteristics. Based on the long list of shared features, Commerce could reasonably conclude that Japanese and United States LNPPs could reasonably be compared for calculating CV profit.

TKS retorts that whatever the value of the questionnaire, Commerce did not consider the whole record when making its comparability determination, because the weight of evidence points the other way. First, TKS notes that United States LNPPs often contained significantly more individual components than did their Japanese counterparts. However, because profit is calculated as a percentage of the sale price, the fact that Japanese LNPPs may have fewer components (and thus, perhaps, a lower overall price) is immaterial. The individual differences between the United States and Japanese models that TKS cites are significant (for example, the United States units use "tower printing units" instead of the "satellite printing units" and "spot color units" more prevalent in Japan). However, such differences are unavoidable in customized equipment. That a United States buyer chooses a somewhat different mix of components than does a Japanese one may preclude price-matching the two contracts, but it does not mean that the machines themselves may not reasonably be compared.

TKS also takes umbrage at Commerce's reference to the English-language and Japanese-language Spectrum product brochures that TKS submitted in response to Commerce's demand to provide all brochures relating to the merchandise under investigation. Commerce cited the brochures as an example of an LNPP model-the Spectrum model-marketed in both the United States and Japan, and noted that the Japanese and English versions of the brochure were identical. TKS argues that this brochure does not show that all United States Spectrums have identical characteristics as their Japanese counterparts, and that Commerce erred in citing the brochures as evidence of comparability.4 But this is simply another way of saying that the Spectrums, like all LNPPs, are custom-made. The critical point is, given that individual differences exist from order to order, can the custom-made merchandise from Japan and the United States be reasonably compared? Commerce, looking at a brochure offering identical menus of features to Japanese and United States purchasers, could reasonably conclude that one Spectrum LNPP described in the brochure would be reasonably-not perfectly, not identically, but reasonably-comparable to any other Spectrum model.

In short, TKS does not provide any compelling evidence to suggest that Commerce neglected its duty to base its decision on the whole record. To the extent that TKS urges that the evidence before Commerce could be open to multiple interpretations, its argument does not require, or even allow, reversal. See Matsushita Elec. Indus. Co. v. United States, 750 F.2d 927, 933 (Fed. Cir. 1984) (noting that "the possibility of drawing two inconsistent conclusions from the evidence does not prevent an administrative agency's finding from being supported by substantial evidence" (quoting Consol. Edison Co. v. NLRB, 305 U.S. 197, 229 (1938))). Obviously, TKS draws a different conclusion from the evidence of the variations between individual product specifications than did Commerce, but that cannot-and does not-mean that Commerce's interpretation should be overturned. Accordingly, we hold that substantial evidence supported Commerce's decision to treat Japanese market LNPPs as the foreign like product for its determination in this case.

This brings us to TKS's second ground for reversal, that Commerce's foreign like product determination was not in accordance with law because it applied an incorrect interpretation of the relevant statutory provision defining foreign like product.

The Tariff Act provides three definitions of foreign like product. See 19 U.S.C. 1677(16)(A)-(C) (1994 and Supp. V 1999). Commerce relied upon the third and broadest of the three possibilities. The Act provides in relevant part that a foreign like product is "[m]erchandise-(i) produced in the same country and by the same person and of the same general class or kind as the subject merchandise, (ii) like that merchandise in the purposes for which used, and (iii) which the administering authority determines may reasonably be compared with that merchandise." Id. 1677(16)(C) (emphasis added). Commerce concluded that TKS's and MHI's home market LNPPs meet this definition, including the requirement that they "may reasonably be compared" with the United States merchandise. TKS has consistently argued that its home market LNPPs may not be reasonably compared with its exports to the United States, and it continues to press that argument on appeal. As discussed above, we reject the first component of this argument, i.e., that Commerce's determination is not supported by substantial evidence, and we now proceed to the second component, which involves Commerce's interpretation of the reasonable comparability prong of the foreign like product definition. In its second remand determination, Commerce opined that the reasonable comparability prong "must be interpreted based on context of the statutory provision to which the phrase is being applied." Second Remand Determination at 5. On appeal, TKS contends that this interpretation improperly varies depending upon the subsection of the statute to which the foreign like product determination is applied, and that Commerce's incorrect construction of the statute requires reversal of its foreign like product determination.

To understand the basis for TKS's allegation, it is necessary to describe the context in which Commerce articulated its flexible construction of the statute. In Mitsubishi II, the Court of International Trade directed Commerce to explain the factual basis for its determination that home market LNPPs could "reasonably be compared" with the subject LNPPs. Mitsubishi II, 54 F. Supp. 2d at 1197. This order was necessary because Commerce made ambiguous statements in its background section that suggested, in the Court of International Trade's view, that Commerce made a final difmer determination of >20%, giving rise to a presumption that the home market was not reasonably comparable to the exported LNPPs. Id. This presumption, if Commerce made it, would be inconsistent with finding the home market to be reasonably comparable for purposes of making a foreign like product determination. The Court of International Trade realized that the >20% difmer policy is only a guideline, but stated that after finding a difmer of over 20 percent, Commerce must affirmatively demonstrate why the merchandise is nevertheless reasonably comparable. Id. at 1196. This Commerce failed to do, necessitating a remand.

On remand, Commerce responded to the Court of International Trade's order by explaining that it never conducted a difmer analysis, and thus never made a presumptive finding of noncomparability. Second Remand Determination at 4. The reference to the difmer analysis in its first remand determination, explained Commerce, was purely background information-an assertion supported both from its context and from its location in the "Background" section of the first remand determination. See First Remand Determination at 15. But after providing this explanation, Commerce engaged in a lengthy, and seemingly unnecessary, discussion of why it was appropriate to construe the statutory term "may reasonably be compared" differently depending on the context in which the definition of foreign like product will be applied. The discussion of the flexible construction was unnecessary because, as discussed below, Commerce never used differing meanings of the reasonable comparability prong in this investigation. Nevertheless, TKS's challenge to the legal underpinnings of Commerce's foreign like product determination springs from Commerce's articulation of this somewhat novel interpretation of the statute.

The Court of International Trade expressed concern with Commerce's statutory construction, but ultimately declined to reach the issue because it concluded that regardless of Commerce's proposed construction, "it was apparent that Commerce had not in fact applied the reasonable comparability prong inconsistently in its investigation of Japanese LNPPs. Therefore, the issue was not directly before us." Mitsubishi IV, 112 F. Supp. 2d at 1174. We agree with the Court of International Trade that the statutory construction issue is not ripe for decision, because TKS does not present, and our independent review does not reveal, any evidence of an inconsistent application of the statute in the case now on appeal. Instead, TKS argues that when Commerce initially decided to use CV instead of price-to-price comparisons, it necessarily found that home market LNPPs were not reasonably comparable to the exported LNPPs. Then, goes the argument, Commerce reversed itself and found them comparable for purposes of its foreign like product determination. TKS accuses Commerce of using its "flexible" statutory construction to arrive at inconsistent foreign like product determinations in its CV profit determination and its decision memorandum adopting CV over price-to-price comparisons. Careful examination of the allegedly inconsistent uses reveals, however, that TKS is chasing a phantom inconsistency in this case.

Commerce explained its decision to use CV rather than price-to-price comparisons in its November 9, 1995, decision memorandum. Commerce began by noting that "[t]he issue of the usability of the foreign like product in determining normal value (NV) in this case is two-fold: (1) whether or not price-to-price comparisons based on disaggregation of contract prices are feasible; and (2) whether or not price-to-price comparisons are technically feasible." Throughout the decision memorandum, Commerce uses the term "foreign like product" to refer to LNPPs sold in the Japanese market. Indeed, it uses "home market" and "foreign like product" interchangeably in the decision memorandum. Thus, contrary to TKS's allegation, Commerce used home market LNPPs as the foreign like product both for purposes of its decision on price-to-price comparisons and for its CV profit determination. Far from demonstrating an allegedly improper inconsistency of application, the decision memorandum highlights a consistent use of the statute, because Commerce used the same Japanese LNPP foreign like product both to determine whether TKS's home market was viable and to calculate CV profit.

In the decision memorandum, Commerce proceeds to describe the statutory guidelines for determining normal value, as set forth in 19 U.S.C. 1677b(a)(1). As noted by Commerce, this section "establishes .general rules for determining when the Department may base normal value on home market sales in the exporting country," a determination Commerce calls the "viability test." Commerce found, and TKS concedes, that the home market was viable. But Commerce noted that "[c]onsistent with the new statute ... and notwithstanding the results of the above-described 'viability test,' the Department may determine that home market sales are inappropriate as a basis for determining [normal value] if the particular market situation would not permit a proper comparison." Commerce devoted the rest of its decision memorandum to an analysis of the particular market situation with respect to LNPPs. It explained that many factors underlie the "particular market situation" decision, the most important of which are "(1) the unique demand pattern prevalent in each national market; (2) the unique technical specifications required for each customized product sold; and (3) the very low volume of individual LNPP sales in the normal business cycle." The problem with direct price-to-price comparisons was that because each LNPP is custom made, Commerce would have to deconstruct each contract into its component parts, conduct a difmer analysis for each part, and then perform the overall comparison. After an exhaustive analysis, Commerce concluded that even if deconstruction were feasible (which it doubted), the actual calculation "would become an analytical exercise equivalent to the use of constructed value." In other words, the number of individual difmer adjustments that would be required for actual price-to-price comparisons was great enough that it would be more efficient simply to use constructed value. Importantly, this does not mean that the home market cannot reasonably be compared to the exported goods-it simply means that any comparison should use CV rather than direct price-to-price comparisons of individual models.

TKS claims that in the decision memorandum Commerce concluded "that the LNPPs sold to the home market are not 'reasonably comparable' to the LNPPs sold to the United States for 'price comparison purposes.'" TKS's argument has no merit because it does not reflect what Commerce actually did. Commerce simply decided that the particular market conditions rendered price-to-price comparisons impracticable-nothing more. It certainly did not decide that the home market LNPPs, in general, could not be a foreign like product under the statute.

In light of what Commerce actually decided in its price-to-price comparison decision memorandum, it is apparent that its variable interpretation of the reasonably comparable prong is not squarely before us, because Commerce did not apply that prong inconsistently in this determination. In the decision memorandum on price-to-price comparisons, Commerce did not discuss specifically its application of the reasonably comparable prong, but it used home market LNPPs as the foreign like product for the viability test, just as it did for the CV profit calculations. Therefore, these two applications were consistent with one another. This view is borne out by Commerce's own rejection of TKS's "inconsistent application" argument in Japan Final. Commerce stated its position as follows:

We disagree with TKS that there were no sales of the foreign like product in the home market during the [period of investigation.] TKS is incorrect to suppose that because we did not find home market sales which provided practicable price-to-price matches, no foreign like product existed. The foreign like product ... (i.e. sales of LNPP in Japan) did exist, as revealed by our examination of LNPP equipment sold in the home market for purposes of the Department's home market viability test.... However, the degree of unique customization for customers made the difference-in-merchandise adjustment for product price matching potentially so complex that the use of CV provided a more reliable and administrable methodology for establishing NV.

Japan Final, 61 Fed. Reg. at 38,146. In other words, Commerce's own explanation in Japan Final reveals that it used the same foreign like product, Japanese LNPPs, both for CV profit and in its decision to use CV. Hence, no inconsistency has been presented to frame our review in this appeal.

Because Commerce did not apply its flexible statutory construction of the reasonable comparability prong in this determination, we decline to reach the issue of whether Commerce's interpretation is contrary to law. Furthermore, because Commerce did not apply inconsistent interpretations of the statute when determining TKS's antidumping duty, we reject TKS's statutory construction argument for reversal.

For the reasons stated above, the decision of the Court of International Trade affirming Commerce's assessment of antidumping duties is affirmed in all respects.

No costs.