WALLACH, Circuit Judge.

In the decision now on appeal, the United States Court of International Trade affirmed the Department of Commerce's ("Commerce") use of zeroing to determine antidumping duties in administrative reviews, even though Commerce no longer uses zeroing in investigations establishing antidumping orders. This court has twice considered whether such divergent practices constitute a reasonable construction of Commerce's governing statute, both times remanding for Commerce to provide an explanation. In the case now on appeal, Commerce has provided such an explanation. Union Steel, LG Hausys, Ltd., LG Hausys America, Inc., and Dongbu Steel Co., Ltd. (collectively, "Appellants") appeal from the Court of International Trade's decision that, in light of this explanation, Commerce's zeroing practices are a reasonable interpretation of statute. Union Steel v. United States, 823 F.Supp.2d 1346, 1360 (Ct.Int'l Trade 2012) ("Union Steel"). Because the Court of International Trade properly found that Commerce's interpretation of its governing statute is in accordance with law, we affirm.

BACKGROUND

Dumping occurs when imported merchandise is sold for a lower price in the United States than it is sold in its home market. This practice can harm domestic producers who are selling the same goods at market value. See Sioux Honey Ass'n. v. Hartford Fire Ins. Co., 672 F.3d 1041, 1046 (Fed.Cir.2012). The antidumping duty statute provides for the imposition of remedial duties to imported merchandise sold, or likely to be sold, in the United States "at less than fair value" when the relevant domestic industry is harmed. 19 U.S.C. § 1673. "Sales at less than fair value are those sales for 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.'" U.S. Steel Corp. v. United States, 621 F.3d 1351, 1353 (Fed.Cir.2010) (quoting 19 U.S.C. § 1677(35)(A)).

Commerce calculates a "dumping margin," which is "the amount by which the normal value exceeds the export price or constructed export price."1 19 U.S.C. § 1677(35)(A). Commerce relies upon three comparison methods to calculate dumping margins:

See Statement of Administrative Action accompanying the Uruguay Round Agreements Act, H.R. Doc. No. 103-316, vol. 1, at 842-43, reprinted in 1994 U.S.C.C.A.N. 3773 ("SAA").

Commerce calculates dumping margins both in investigations, which establish an antidumping order, and in subsequent administrative reviews of that order. Following an investigation, Commerce issues an antidumping order which imposes a duty based upon the dumping margin. See 19 U.S.C. §§ 1673a, 1673b(b), 1673b(d), 1673d(a), 1673d(c). Any exporter of the goods subject to the antidumping order may annually request an administrative review to determine the exact amount by which the foreign market value exceeds the U.S. price and assess the precise amount of duties owed for their exports.2 See 19 U.S.C. §§ 1675(a)(1), 1675(a)(2)(A).

As explained in the SAA accompanying the Uruguay Round Agreements Act ("URAA") Commerce had a practice of using average-to-transaction comparisons to calculate dumping margins in both investigations and administrative reviews. SAA at 842. After adoption of the URAA in 1995, Commerce switched to using average-to-average or transaction-to-transaction comparisons in antidumping duty investigations.3 Id.; 19 U.S.C. § 1677f-1(d)(1)(A). Commerce continued to use average-to-transaction comparisons as its general practice in administrative reviews.

In calculating the weighted average dumping margin, Commerce has historically used a methodology called "zeroing" where negative dumping margins (i.e., margins of sales of merchandise sold at nondumped prices) are given a value of zero and only positive dumping margins (i.e., margins for sales of merchandise sold at dumped prices) are aggregated. "That is, after [Commerce] computed an average dumping margin for each averaging group, if that averaging group ... product did not have a positive dumping margin, Commerce set the margin at zero rather [than] at a negative number that would offset a positive margin for another averaging group."4 Union Steel, 823 F.Supp.2d at 1350. The applicable statute, 19 U.S.C. § 1677(35)(A), does not mention zeroing. However, as authority for this method, Commerce has emphasized the part of the statute stating that the dumping margin "means the amount by which the normal value exceeds the export price or constructed export price of the subject merchandise." 19 U.S.C. § 1677(35)(A) (emphasis added).

This court has repeatedly addressed zeroing and has held 19 U.S.C. § 1677(35)(A) ambiguous and deferred to Commerce's reasonable interpretation of that statute. Timken Co. v. United States, 354 F.3d 1334, 1342 (Fed.Cir.2004) (applying Chevron analysis to determine that Commerce's practice of using zeroing in administrative reviews was a reasonable interpretation of the statute); Corus Staal BV v. Dep't. of Commerce, 395 F.3d 1343, 1347 (Fed.Cir. 2005) ("Corus") (extending Timken to encompass Commerce's practice of zeroing in investigations). Zeroing is controversial because some parties claim it does not fully account for all of the comparable export transactions, and so is not a "fair comparison" between export price and normal value as required by Article 2.4 and 2.4.2 of the Anti-Dumping Agreement. See Panel Report, United States — Laws, Regulations and Methodology for Calculating Dumping Margins ("Zeroing"), ¶¶ 7.32, 7.33, WT/DS294/R (Oct. 31, 2005); aff'd Appellate Report, United States — Laws, Regulations and Methodology for Calculating Dumping Margins ("Zeroing"), ¶ 146, WT/DS294/AB/R (Apr. 18, 2006); see also Dongbu Steel Co. v. United States, 635 F.3d 1363, 1366 (Fed.Cir.2011) (characterizing zeroing as "controversial").

Commerce's use of zeroing with average-to-average comparisons in certain antidumping duty investigations was challenged by the European Communities before the World Trade Organization's ("WTO") Dispute Settlement Body. See Panel Report, ¶¶ 7.32, 7.33. The WTO found Commerce's practice inconsistent with the United States' international obligations, and Commerce determined that it would cease using zeroing methodology in new and pending investigations. See Antidumping Proceedings: Calculation of the Weighted-Average Dumping Margin During an Antidumping Investigation; Final Modification, 71 Fed. Reg. 77,722 (Dec. 27, 2006). Instead, Commerce started using a method of "offsetting" to account for sales made at less than fair value, such that some of the dumping margins used to calculate a weighted average dumping margin would be negative. U.S. Steel, 621 F.3d at 1355.

In 2011, this court considered a challenge to Commerce's continuing use of zeroing in administrative reviews in an earlier review of the same antidumping duty order at issue in this case. In Dongbu, appellant Union Steel argued "that it is unreasonable to construe a single statutory provision [19 U.S.C. § 1677(35)] that applies to both investigations and administrative reviews as having opposite meanings depending on the nature of the antidumping proceeding." Dongbu, 635 F.3d at 1370. The court determined that "[a]lthough 19 U.S.C. § 1677(35) is ambiguous with respect to zeroing and Commerce plays an important role in resolving this gap in the statute, Commerce's discretion is not absolute. Commerce must provide an explanation for why the statutory language supports its inconsistent interpretation." Id. at 1372. The decision of the Court of International Trade was vacated and remanded for further proceedings "to give Commerce the opportunity to explain its reasoning." Id. at 1373.

Shortly thereafter, but before Commerce had the opportunity to provide an explanation, the court again addressed Commerce's practice of zeroing in administrative reviews but not in investigations in JTEKT, stating that Commerce

JTEKT Corp. v. United States, 642 F.3d 1378, 1384-85 (Fed.Cir.2011). Accordingly, the Court of International Trade's decision was vacated and the case was remanded "in order for Commerce to provide its reasoning." Id. at 1385.

Commerce's explanation is now before the court. At the Court of International Trade, Appellants challenged Commerce's application of zeroing methodology to the final results of the sixteenth administrative review for imports of certain corrosion-resistant carbon steel flat products from the Republic of Korea, the same antidumping duty order at issue in Dongbu. Union Steel, 823 F.Supp.2d at 1347-48; see Dongbu, 635 F.3d at 1373. The United States sought a voluntary remand from the Court of International Trade in light of this court's decision in JTEKT, and the court granted the motion. J.A. 96. On remand, Commerce discussed the inconsistent use of zeroing in administrative reviews and investigations, and explained why it believed its interpretation is reasonable, stating in summary:

Results of Redetermination Pursuant to Remand at 7 (Oct. 14, 2011) ("Remand Results").5 The Court of International Trade sustained Commerce's explanation, concluding that Commerce "did not abuse its discretion in changing its investigation methodology, but not its review methodology... in response to WTO decisions." Union Steel, 823 F.Supp.2d at 1360. This appeal followed. We have jurisdiction pursuant to 28 U.S.C. § 1295(a)(5).

DISCUSSION

While we recognize the Court of International Trade has unique and specialized expertise in trade law,6 its decision is reviewed de novo, applying anew the same standard used by that court in its consideration of Commerce's determination. Dongbu, 635 F.3d at 1369. Accordingly, Commerce's antidumping determination will be upheld 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). We apply a two-part inquiry to determine whether to sustain Commerce's statutory interpretation. See Chevron, U.S.A., Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 842-43, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984). First, we determine "whether Congress has directly spoken to the precise question at issue." Id. at 842, 104 S.Ct. 2778. If so, we "must give effect to the unambiguously expressed intent of Congress." Id. at 842-43, 104 S.Ct. 2778. "[I]f the statute is silent or ambiguous with respect to the specific issue," however, "the question for the court is whether the agency's answer is based on a permissible construction of the statute." Id. at 843, 104 S.Ct. 2778.

The question here, as in Dongbu and JTEKT, is whether it is reasonable for Commerce to use zeroing in administrative reviews even though it no longer uses zeroing in investigations. See Dongbu, 635 F.3d at 1369; JTEKT, 642 F.3d at 1384-85. Our analysis of the issue, however, is now aided by Commerce's explanation why the ambiguous statute, 19 U.S.C. § 1677(35), supports use of zeroing in administrative reviews and not in investigations. Thus, our task is now to determine whether Commerce's explanation of its zeroing practice is a reasonable interpretation of the statute.

Commerce's decision to modify its zeroing practice has previously been sustained by this court. In U.S. Steel, the court sustained Commerce's decision to cease zeroing when making average-to-average comparisons in antidumping duty investigations while recognizing Commerce intended to continue zeroing in other circumstances. See U.S. Steel, 621 F.3d at 1355 n. 2, 1362-63. The court relied upon the differences among various types of comparison methodologies, recognizing that 19 U.S.C. § 1677f-1(d)(1) allows Commerce to use average-to-transaction comparisons in investigations where certain patterns of significant price differences exist. Id. at 1362. Additionally, the court sustained Commerce's decision to use zeroing methodology in an administrative review using average-to-transaction comparisons. SKF USA Inc. v. United States, 630 F.3d 1365, 1375 (Fed.Cir.2011). In SKF, the court stated that "[e]ven after Commerce changed its policy with respect to original investigations, we have held that Commerce's application of zeroing to administrative reviews is not inconsistent with the statute." Id.

Contrary to Appellants' assertions, Commerce's reasonable interpretation of the statute is not foreclosed by this court's prior decisions.7 In Corus, the court held that the statute was equally ambiguous for both administrative reviews and investigations and thus Commerce may use zeroing in both despite using different comparison methodologies in each. Corus, 395 F.3d at 1347. Although this court noted, in dicta, that "[i]t may be that Commerce cannot justify using opposite interpretations of 19 U.S.C. § 1677(35) in investigations and in administrative reviews," Dongbu, 635 F.3d at 1373, neither Dongbu nor JTEKT foreclosed the opportunity for Commerce to provide an explanation. Instead, both cases specifically requested that Commerce do so. Id.; JTEKT, 642 F.3d at 1384-85. Commerce's explanation now on review demonstrates that its varying interpretations are reasonable given the distinction between the comparison methodologies used in investigations and administrative reviews. Moreover, Commerce attributes the differing interpretations as necessary to comply with international obligations, while preserving a practice that serves recognized policy goals. Each of these analyses will be discussed in turn.

1.

Commerce justifies using zeroing in administrative reviews but not in investigations in part based on the different comparison methodologies used in each. Commerce explained in its Remand Results that average-to-average comparison methodology typically used in investigations is useful for examining an exporter's or manufacturer's overall pricing behavior. Remand Results at 13. Overall pricing behavior helps determine the appropriateness of imposing an antidumping duty order on a particular product. Id. With an antidumping duty order already in place, administrative reviews typically use average-to-transaction comparison methodology which permits greater specificity to determine pricing behavior for individual transactions. Id. The greater specificity afforded through that methodology furthers the transactional accuracy interests of administrative review. Union Steel, 823 F.Supp.2d at 1359. We agree with the Court of International Trade's explanation that this distinction is supported by statute: "Specificity is less important in investigations in that [product group (CONNUM)] averages in investigations are not even monthly averages, as they are in reviews. Rather, they are averages over a broad time period compared to all other broad averages." Id. (citing 19 U.S.C. § 1677f-1(d); 19 C.F.R. § 351.414(d)(3), (e)). However, "when it comes to setting the final rates to be used for actual assessment, i.e., the review rates, it is reasonable for the agency to look for more accuracy, which it achieves in some measure through monthly averaging, and also for the agency to look for the full measure of duties resulting therefrom, which it better achieves through zeroing." Id. The Court of International Trade noted that "parties who are marginally dumping or not dumping may be excluded from the order pursuant to the looser standards of the investigation." Id. That is, in an investigation, margins of less than two percent are treated as de minimis, resulting in a party's exclusion from the order, while in an administrative review, margins of 0.5 percent or less are treated as de minimis. Id.

Commerce also explained that the average-to-average methodology justifies the use of offsetting (i.e., not zeroing) for reasons inapplicable to average-to-transaction comparisons. When using average-to-average comparisons, transactions are divided into "averaging groups." Remand Results at 11. Transactions are divided into averaging groups on the basis of physical characteristics and level of trade for the purpose of price comparison. Id. When calculating the average export price or constructed export price, Commerce calculates a comparison result for each averaging group, and averages together high and low export prices within the group. Thus, those export prices above normal value offset those below normal value within the averaging group. Commerce then aggregates the results of the comparison for each averaging group to calculate a weighted average dumping margin. Id. at 11-12. Accordingly, this comparison methodology masks individual transaction prices below normal value with other above normal value prices within the same averaging group.

In contrast, when Commerce uses the average-to-transaction comparison method, as it did in this administrative review, Commerce compares the export price (or constructed export price) for a particular export transaction with an average normal value for the comparable sales of foreign like products within the averaging group. Id. at 12. For specific export transactions, Commerce calculates a comparison result which establishes the amount that transaction is priced at less than its normal value. Id. Using this methodology, Commerce does not average export transaction prices before comparing the export price (or constructed export price) to normal value. Instead, Commerce uses a single export transaction price and aggregates the transaction-specific comparison result. The average-to-transaction comparison methodology thus reveals individual dumping.

Commerce's decision to use or not use the zeroing methodology reasonably reflects unique goals in differing comparison methodologies. In average-to-average comparisons, as used in investigations, Commerce examines average export prices; zeroing is not necessary because high prices offset low prices within each averaging group. When examining individual export transactions, using the average-to-transaction comparison methodology, prices are not averaged and zeroing reveals masked dumping. This ensures the amount of antidumping duties assessed better reflect the results of each average-to-transaction comparison.8 Commerce's differing interpretation is reasonable because the comparison methodologies compute dumping margins in different ways and are used for different reasons.

2.

Commerce also explained the methodology for investigations was changed in response to an adverse WTO decision through a section 123 proceeding.9 In Dongbu, the government raised this rationale at oral argument, and this court indicated that "the government's decision to implement an adverse WTO report standing alone does not provide sufficient justification for the inconsistent statutory interpretations." Dongbu, 635 F.3d at 1372. Nevertheless, it is within Commerce's discretion to adopt reasonable practices to meet international obligations. Union Steel, 823 F.Supp.2d at 1357-58.10 Certainly, this information is relevant when considered in conjunction with the other explanations offered by Commerce.

Section 123 establishes how an adverse WTO decision may be implemented in domestic law. See 19 U.S.C. §§ 3533; 3538. The WTO's decision was limited; it found that Commerce's use of zeroing methodology with respect to average-to-average comparisons in antidumping duty investigations was inconsistent with the United States's international obligations. Panel Report ¶ 7.106. The Executive Branch responded by discontinuing its zeroing practice in new and pending investigations using average-to-average comparison methodology. Antidumping Proceedings: Calculation of the Weighted-Average Dumping Margin During an Antidumping Duty Investigation; Final Modification, 71 Fed. Reg. 77,722 (Dec. 27, 2006); see U.S. Steel, 621 F.3d at 1354-55. Commerce, did not, however, alter its practice with respect to the use of zeroing methodology in anything other than investigations using average-to-average comparisons. "[T]here is no reason to assume that [Commerce's] only legal option is to expand the exception to apply in all contexts." Remand Results at 18. Commerce may reasonably decline to take any action beyond that which is necessary for it to come into compliance. See ThyssenKrupp Acciai Speciali Terni S.p.A. v. United States, 603 F.3d 928, 934 (Fed.Cir. 2010) (affirming Commerce's determination to only address that which was necessary to bring its determination into accordance with a WTO ruling). Commerce's modification was limited to changes that were necessary to comply with the WTO decision.

Citing Clark v. Martinez, 543 U.S. 371, 125 S.Ct. 716, 160 L.Ed.2d 734 (2005),11 Appellants argue that it is unreasonable to construe a single statutory provision that applies in both investigations and administrative reviews as having different meanings depending on the type of antidumping proceeding. In Clark, the Supreme Court relied upon the rule of lenity to support a limiting construction of a statutory provision concerning detention of aliens subject to removal from the United States. Clark, 543 U.S. at 380-81, 125 S.Ct. 716. The Supreme Court applied traditional rules of statutory construction, and never considered Chevron deference in reaching its determination. See Clark, 543 U.S. 371, 125 S.Ct. 716. Here, the court has repeatedly held that 19 U.S.C. § 1677(35) is ambiguous and Commerce's explanation of its interpretation must be reasonable. See Timken, 354 F.3d at 1342; Corus, 395 F.3d at 1347; U.S. Steel, 621 F.3d at 1361. Thus, Clark is distinguishable here, where there is a conflict between two permissible interpretations simultaneously maintained.

No rule of law precludes Commerce from interpreting 19 U.S.C. § 1677(35) differently in different circumstances as long as it provides an adequate explanation. "[I]f the agency adequately explains the reasons for a reversal of policy, `change is not invalidating, since the whole point of Chevron is to leave the discretion provided by the ambiguities of a statute with the implementing agency.'" Nat'l. Cable & Telecomms. Ass'n. v. Brand X Internet Servs., 545 U.S. 967, 981, 125 S.Ct. 2688, 162 L.Ed.2d 820 (2005) (quoting Smiley v. Citibank (South Dakota), N.A., 517 U.S. 735, 742, 116 S.Ct. 1730, 135 L.Ed.2d 25 (1996)).12

CONCLUSION

For these reasons, the Court of International Trade's decision is affirmed.

AFFIRMED