dissenting.
I disagree with the majority that 19 U.S.C. § 1671d(c)(5)(A)(i)-(ii) (“Section 1671d”) unambiguously requires that the voluntary respondent rates be included in the all-others rate calculation. Because I find the statute ambiguous and Commerce’s interpretation reasonable, I would affirm the Court of International Trade’s decision affirming a final all-others countervailing duty rate of 137.65% in this investigation. I therefore respectfully dissent.
BACKGROUND
A complete statement of the background of this case is essential to understanding the issues involved on appeal. The present appeal concerns a countervailing duty investigation on imports of aluminum extrusions from China initiated by Commerce on April 20, 2010. Aluminum Extrusions from the People’s Republic of China, 75 Fed.Reg. 22,114 (Dep’t of Commerce Apr. 27, 2010). The petition that triggered the investigation identified as *1247potential respondents 114 Chinese exporters and producers of aluminum extrusions. Due to limited agency resources, Commerce decided to limit the investigation to the three largest companies based on export volume (the “mandatory respondents”).
The investigation initially focused on 29 subsidy programs alleged in the petition. Commerce issued a questionnaire to the mandatory respondents regarding these programs, but none of the mandatory respondents submitted a response. Two other Chinese companies identified in the petition, however, requested “voluntary respondent” status and submitted questionnaire responses. The Government of China also submitted a questionnaire response.
Commerce published its preliminary results on September 7, 20.10. Aluminum Extrusions from the People’s Republic of China, 75 Fed.Reg. 54,302, at 54,321 (Dep’t of Commerce Sept. 7, 2010). Because the mandatory respondents failed to cooperate in the investigation by declining to respond to the questionnaire, Commerce calculated their preliminary countervailing duty rate using “adverse facts available” (AFA). Pursuant to 19 U.S.C. § 1677e(b), Commerce may use an adverse inference in selecting among facts derived from the petition, the record, or any previous investigation. Commerce explained that it typically relies on the rates calculated for cooperating respondents when calculating an AFA rate. Because in this case there were no cooperating mandatory respondents, Commerce used the highest non-de minimis rate calculated for the same or similar subsidy program in another China countervailing duty proceeding, in accordance with its established practice. Where no non-de minimis subsidy rates calculated for the same or similar program existed, Commerce applied “the highest calculated subsidy rate for any program otherwise listed that could conceivably be used by the non-cooperating companies.” Id. at 54,321. Following this procedure, Commerce arrived at a 137.65% preliminary countervailing duty rate for the mandatory respondents. Id. at 54,305. For the two voluntary respondents, Commerce calculated rates of 6.18% and 10.37% based on their separate, company-specific data. Id.
For the exporters and producers not designated as mandatory or voluntary respondents (the “all-others respondents”), Commerce calculated a preliminary “all-others” rate equal to the preliminary AFA rate for the mandatory respondents. As the basis for this methodology, Commerce relied on 19 C.F.R. § 351.204(d)(3), which provides that countervailing duty rates calculated for voluntary respondents should be excluded from the all-others rate calculation. Commerce acknowledged that, in prior investigations, it had calculated the all-others rate by averaging the AFA rates of the non-cooperating, mandatory respondents with the rate calculated for voluntary respondents. But Commerce explained that, upon further examination, it found that approach was not appropriate in this case because the “potential for voluntary respondents’ net subsidy rates to distort or manipulate the all-others rate is too great.” Id. at 54,321.
After receiving case briefs and rebuttal comments from interested parties, including Appellants, Commerce issued a final determination calculating a final AFA rate of 374.15% for the mandatory respondents. Aluminum Extrusions from the People’s Republic of China, 76 Fed.Reg. 18,521 (Dep’t of Commerce Apr. 4, 2011). Commerce included in the investigation and its calculation of the final AFA rate 25 additional subsidy programs identified by petitioners during the course of the inves*1248tigation. Commerce also increased the subsidy-specific rate for some programs. Commerce revised the rate for the two voluntary respondents to 8.02% and 9.94%.
Commerce continued to apply the AFA rate (which had increased to 374.15%) to the all-others respondents. In a memorandum responding to the comments submitted by interested parties, Commerce noted that 19 C.F.R. § 351.204(d)(3) mandates exclusion of the voluntary respondent rates from the all-others rate calculation. Quoting comments accompanying the promulgation of this rule, Commerce explained that the Tariff Act was silent regarding whether voluntary respondent rates should be included, and that Commerce had therefore looked at the World Trade Organization (WTO) Anti-Dumping Agreement for guidance:
Article 9.4 of the [Anti-Dumping] Agreement provides that the duties applied to “exporters or producers not included in the examination” (ie., “all-others”) may not exceed the weighted-average margin for the “selected exporters or producers.” This implies that those exporters or producers not “selected” are not considered to be included in the “examination.” Therefore, the better interpretation of section 705(c)(5) or 735(e)(5) of the Act is that producers who are not “selected” by the Department (ie., voluntary respondents) are not considered to have been “examined” (ie., investigated), so that their margins should not contribute to the “all-others” rate. In effect, the Department conducts parallel proceedings for voluntary respondents.
Joint Appendix (J.A.) at 245 (quoting Anti-dumping Duties; Countervailing Duties, 62 Fed.Reg. 27,296, at 27,310 (Dep’t of Commerce May 19, 1997)). Commerce also explained that excluding voluntary respondent rates from the all-others rate “serves the obvious purpose of preventing distortion or outright manipulation of the all-others rate” because exporters or producers most likely to submit voluntary responses “are those with reason to believe that they will obtain a lower margin by volunteering than by being subject to the all-others rate.” J.A. at 246 (quoting 62 Fed.Reg. at 27,310).
Appellants and other importers and Chinese producers filed suit in the Court of International Trade alleging that Section 1671d unambiguously requires that the voluntary respondent rates be included in the all-others rate calculation. The Court of International Trade disagreed, finding that the statute was ambiguous and that Commerce’s interpretation excluding voluntary respondent rates was reasonable. Maclean-Fogg Co. v. United States, 836 F.Supp.2d 1367, 1373-74 (Ct. Int’l Trade 2012). The court nevertheless found that Commerce had not established that excluding the voluntary respondent rates from the all-others rate calculation was “a reasonable method” under § 1671d(c)(5)(A)(ii) because there was no “logical connection” between the AFA rate and the all-others respondents. Id. at 1375-76. The court remanded for Commerce to recalculate the all-others rate or explain why adopting the AFA rate for the all-others respondents constituted a reasonable method. Id.1
On remand, Commerce retained its methodology for calculating the all-others rate based on the mandatory respondents’ *1249AFA rate. Commerce argued that its approach was reasonable because the three mandatory respondents represented a significant portion of the relevant market and were therefore representative of the all-others respondents. The Court of International Trade disagreed, finding that Commerce had failed to explain why the AFA rate was remedial and not punitive when applied to the all-others respondents, who Commerce had not found to be non-cooperative. See MacLean-Fogg Co. v. United States, 853 F.Supp.2d 1336, 1343 (Ct. Int’l Trade 2012). First, the court noted that the AFA rate assumed that the mandatory respondents benefited from all 54 subsidy programs included in the investigation, but Commerce had provided no reason why the same assumption applied to the all-others respondents, who are smaller than the mandatory respondents. See id. at 1342. Second, the court observed that Commerce had failed to justify its assumption that the all-others respondents benefited from subsidies specific to certain geographic areas. See id. at 1343. The court thus remanded the case again for Commerce to reconsider the all-others rate.
On the second remand, Commerce adopted the 137.65% preliminary AFA rate as the final all-others rate. Commerce found it reasonable to assume that the all-others respondents might not have benefited from the 25 subsidy programs added during the investigation, which had been examined solely with respect to the voluntary respondents. Commerce thus concluded that the preliminary rate was reasonable because it only included the 29 programs that were investigated in connection with the mandatory respondents. Commerce also noted that, although certain programs were location-specific, it had to assume use of those programs because the record was incomplete regarding the location of the all-others respondents and it was not reasonable to extrapolate the locations of manufacturing facilities or company affiliations from the addresses included in the petition.
The Court of International Trade affirmed the 137.65% all-others rate, finding that the rate was reasonable given “the limitations of the administrative record.” MacLean-Fogg Co. v. United States, 885 F.Supp.2d 1337, 1342 (Ct. Int’l Trade 2012). The court found that adopting the preliminary AFA rate — based on a reduced number of subsidy programs — sufficiently addressed its concern that the rate should not be punitive. The court rejected Appellants’ attempts to show that the all-others rate was based on program-specific rates that are aberrant and unrepresentative when compared to historical data, noting that Commerce is only required to provide a reasonable rate, not necessarily “a perfect one.” Id. The court also determined that the methodology used to calculate the preliminary rate for the mandatory respondents was reasonable as applied because the mandatory respondents account for the vast majority of aluminum exports from China. See id.
This appeal followed.
DISCUSSION
I
Section 1671d is ambiguous with respect to the precise question at issue: whether Commerce is required to include voluntary respondent rates when calculating the all-others rate under the general rule. At issue in this case is not the nomenclature developed by Commerce to distinguish classes of respondents, but how the statute dictates that rates should be calculated depending on whether a company is selected for investigation or not. Section 1671d states that Commerce shall use the weighted average rates calculated for re*1250spondents “individually investigated,” excluding rates that are zero, de minimis, or determined entirely under facts available. 19 U.S.C. § 1671 d(c)(5)(A)(i). The statute does not define “individually investigated” as used in Section 1671d.
I disagree with the majority that the requirement to calculate an “individual” rate for voluntary respondents implies that voluntary respondents are “individually investigated” within the meaning of Section 1671d. The fact that voluntary respondents must receive an individual rate does not directly address the question of whether such rate must be included in the all-others rate calculation. “Voluntary” respondents are not selected for individual-specific investigation. Commerce may accept voluntary respondents at its discretion, and only if they submit the relevant information in a complete and timely manner. See id. § 1677m(a)(l)-(2). The optional nature of the process suggests that it provides a mechanism for respondents “not initially selected” to be excepted from the all-others rate, and was not necessarily intended to inject an additional variable into the mandatory process of calculating the all-others rate.
Indeed, the statute clearly distinguishes voluntary respondents from those that were “initially selected for individual investigation” by Commerce. See id. § 1677m(a). The parties do not dispute that the respondents “initially selected for individual investigation” are the mandatory respondents. And § 1677Í-1 expressly provides that the individual countervailable subsidy rates determined for the mandatory respondents “shall be used to determine the all-others rate under section 1671d(c)(5) of this title.” Id. § 1677f-l(e). Because this section neither forecloses nor mandates inclusion of the voluntary respondent rates in the all-others rate calculation, I would hold that the statute is ambiguous with respect to the precise question presented here.2
I also find that Commerce’s interpretation of Section 1671 d — embodied in 19 C.F.R. § 351.204(d)(3) — is based on a permissible construction of the statute. See Chevron, U.S.A., Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 843, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984). In promulgating, the regulation, Commerce looked at the WTO Anti-Dumping Agreement for guidance, which provides:
When authorities have limited their examination in accordance with the second sentence of paragraph 10 of Article 6, any anti-dumping duty applied to imports from exporters or producers not included in the examination shall not exceed:
(i) the weighted average margin of dumping established with respect to the selected exporters or producers or,
(ii) where the liability for payment of antidumping duties is calculated on the basis of a prospective normal value, the difference between the weighted average normal value of the selected exporters or producers and the export prices of exporters or producers not individually examined,
provided that the authorities shall disregard for the purpose of this paragraph any zero and de minimis margins and *1251margins established under the circumstances referred to in paragraph 8 of Article 6. The authorities shall apply individual duties or normal values to imports from any exporter or producer not included, in the examination who has provided the necessary information during the course of the investigation, as provided for in subparagraph 10.2 of Article 6.
Agreement on Implementation of Article VI of the General Agreement on Tariffs and Trade 1991, art. 9.4, Apr. 15, 1994, 1868 U.N.T.S. 201 (emphasis added).
The WTO Anti-Dumping Agreement was approved by Congress as part of the Uruguay Round Agreements Act. Pub.L. 103-465, § 101, 108 Stat. 4809, 4814-15 (1994) (codified at 19 U.S.C. § 3511). Absent express language to the contrary, a statute should not be interpreted to conflict with international obligations. See Fed.-Mogul Corp. v. United States, 63 F.3d 1572, 1581 (Fed.Cir.1995) (“GATT agreements are international obligations, and absent express Congressional language to the contrary, statutes should not be interpreted to conflict with international obligations.”). By making an implicit distinction between “selected” respondents and those “not included in the examination [that have] provided the necessary information during the course of the investigation^]” Article 9.4 supports Commerce’s conclusion that voluntary respondents are examined in “parallel” to those companies “included in the investigation.” See 62 Fed.Reg. at 27,310. Hence, Commerce’s conclusion that voluntary respondents are not “individually investigated” within the meaning of Section 1671d is reasonable.
Commerce’s interpretation does not create a “statutory conflict” by mandating imposition on the voluntary respondents of two different rates — the individual rate and the all-others rate. See Maj. Op. at 1242-43. When properly understood as examined in “parallel,” the voluntary respondents are not included in the group of respondents “not individually investigated” who must receive an all-others rate pursuant to § 1671d(c)(l)(B)(i). Likewise, this understanding that voluntary respondents are examined in parallel explains why Congress had no need to include voluntary respondent rates in the list of exclusions to the general rule of Section 1671d. See Maj. Op. at 1242-43.
The majority’s reliance on the “history of the countervailing duty statute” is also misplaced. See id. at 1245. It is true that, prior to the 1994 amendments to the Tariff Act, the calculation of the countrywide rate had to include any “individual rates” calculated for whom the majority views as the “forerunners” of the voluntary respondents. See id. at 1245-46. But the reason for such rule was that zero and de minimis rates also had to be included. See Kajaria Iron Castings Pvt. Ltd. v. United States, 156. F.3d 1163, 1178 (Fed.Cir.1998) (explaining that it would be inconsistent to exclude individual rates when zero and de minimis rates must be included). In contrast, the current statute explicitly excludes zero and de minimis rates from the all-others rate calculation. Accordingly, “the pre-1994 regime” not only does not support the majority’s conclusion, but actually supports Commerce’s determination that, under the current regime, it is reasonable to exclude voluntary respondent rates from the all-others rate calculation.
Finally, I agree with Commerce that excluding voluntary respondent rates from the all-others rate calculation serves the purpose of preventing distortion or manipulation of data. For example, exporters or producers “not initially selected” whose own data suggests an incentive to participate on a voluntary basis could decide not *1252to participate (or refuse to cooperate) should they determine that other respondents with lower countervailable subsidy rates could drive the all-others rate lower than their own voluntary respondent rate. I would hold that Commerce’s regulation excluding voluntary respondent rates from the calculation of the all-others rate is reasonable and entitled to Chevron deference. I therefore respectfully dissent.
II
The majority does not reach the question of whether the methodology used by Commerce to determine the all-others rate in this case was a “reasonable method” pursuant to § 1671d(c)(5)(A)(ii). See Maj. Op. at 1240. Appellants argue that it is not, although they apparently concede that Commerce may use AFA rates “in some reasonable fashion” in calculating the all-others rate. See Reply Br. 12; see also Yangzhou Bestpak Gifts & Crafts Co. v. United States, 716 F.3d 1370, 1378 (Fed.Cir.2013) (calculating the all-others rate by taking the simple average of a de minimis rate and a total AFA rate assigned to two mandatory respondents is a “reasonable method”). Specifically, Appellants contend that Commerce is not permitted to apply to the all-others respondents, which were never found to be non-cooperative, a rate entirely based on AFA rates. Additionally, Appellants argue that the 137.65% rate is aberrant and punitive because it assumes “a worst-case scenario” that each of the all-others respondents benefited from all of the 29 original programs at the highest benefit rate ever calculated in a prior China countervailing duty investigation.
I disagree that Commerce may not adopt an all-others rate based entirely on AFA rates calculated for mandatory respondents. Section 1671d(c)(5)(A)(ii) expressly permits “averaging the weighted average countervailable subsidy rates” calculated for the mandatory respondents when they are all zero, de minimis, or determined entirely on the basis of facts available. There is nothing in the statute that forecloses the adoption of a weighted average of the mandatory respondents’ rates when they are all AFA rates.
Likewise, the Statement of Administrative Action, recognized by Congress as an “authoritative expression” of the interpretation and application of the Tariff Act, see 19 U.S.C. § 3512(d), explicitly contemplates that a “reasonable method” under the exception to the general rule of Section 1671d may include AFA rates:
In such situation, Commerce may use any reasonable method to calculate the all others rate. The expected method in such cases will be to weight average the zero and de minimis margins and margins determined pursuant to facts available, provided the volume data is available.
Statement of Administrative Action Accompanying the Uruguay Round Agreements Act, H.R. No. 103-316, vol. 1, at 873 (1994), reprinted in 1994 U.S.C.C.A.N. 4040, 4201. Indeed, adopting a weighted average of the mandatory respondents’ rate is the “expected” method unless it results in an average “that would not be reasonably reflective” of potential countervailing subsidy margins for the all-others respondents. See id. I thus would reject Appellants’ argument that an all-others rate may not be based entirely on AFA rates, regardless of whether a finding of non-cooperation has been made with respect to the all-others respondents.
I also find that adopting an all-others rate equal to the preliminary AFA rate is reasonable under the facts of this case. Commerce reasonably concluded that the all-others respondents might not have benefited from the additional 25 programs examined solely with respect to the voluntary *1253respondents, and appropriately adopted an all-others rate based only on the 29 subsidy programs included in the initial petition. Given “the limitations of the administrative record,” MacLean-Fogg, 885 F.Supp.2d at 1342, it was not unreasonable for Commerce to assume that the all-others respondents, like the mandatory respondents, received benefits under each of these programs at the highest rate on record for identical or similar programs in prior China countervailing duty investigations. I see no error in Commerce declining to take into consideration information pertaining to individual circumstances that Commerce had no obligation to consider. Had a majority of the panel upheld Commerce’s regulation, I would also hold that substantial evidence supports Commerce’s determination to adopt the preliminary AFA rate for the all-others respondents.
. The court declined to review the preliminary all-others rate. Maclean-Fogg, 836 F.Supp.2d at 1375 n. 11. In a subsequent opinion, the court clarified that the preliminary all-others rate was subject to the same “reasonable method’’ standard as the final rate but deferred review until receiving Commerce's remand determination. MacLean-Fogg Co. v. United States, 853 F.Supp.2d 1253 (Ct. Int'l Trade 2012).
. The majority states that "[t]he fact that some rates determined for mandatory respondents shall be used in calculating the all-others rate does not suggest that the rates determined for individually investigated voluntary respondents must be excluded from the calculation of the all-others rate.” Maj. Op. at 1243 (emphasis added). I agree, and thereby point out that the majority’s position on this issue highlights the ambiguity in the statute.