Opinion & Order
This case returns to the Court following a remand to the United States Department of Commerce pursuant to the Court’s order in Husteel Co., Ltd. v. United States, 31 CIT ---, ---, 491 F.Supp.2d 1283, 1296 (2007) (remanding Oil Country Tubular Goods, Other Than Drill Pipe, from Korea, 71 Fed.Reg. 13,091 (Dep’t Commerce Mar., 14, 2006) and associated Issues & Decision Mem.) (“Husteel /”). Plaintiffs, Husteel Company, Ltd. and SeAH Steel Corporation, Ltd. (together, “Respondents”), challenge the results of Commerce’s remand, in which Commerce continues to exclude certain of Respondents’ sales from the calculation of normal value. See Results of Redetermination on Remand Pursuant to Husteel Co., Ltd. & SeAH Steel Corp., Ltd. v. United States (“Remand Results”). Because Commerce’s Remand Results suffer from the same shortcomings as the original results that they were intended to rectify, the Court remands the Remand Results for further consideration.
Background
Respondents, who are Korean producers of Oil Country Tubular Goods (“OCTG”), participated in the ninth administrative review of the antidumping order on OCTG from Korea, covering the 2003-2004 period of review. In the final results to the administrative ' review, Commerce excluded certain of Respondents’ sales from the calculation of their respective normal values. The sales that Commerce excluded were made by each Respondent to independent trading companies located in Korea, who in turn resold the merchandise to buyers located in the People’s Republic of China (“China” or “PRC”), a nonmarket economy. At the time Respondents negotiated these sales with the trading companies, Respondents knew that the trading companies intended to resell the merchandise to *1359buyers located in China. As a result, Respondents classified these sales as sales to China when they reported them to Commerce. Respondents’ characterization of these sales as Chinese was correct and is not at issue in this case.
What is at issue is Commerce’s exclusion of the sales from the calculation of Respondents’ normal values.1 In the final results of the administrative review, Commerce excluded the sales on the grounds that the prices for the sales may not be “representative,” a statutory requirement for inclusion. See 19 U.S.C. § 1677b(a)(l)(B)(ii)(I) (2000). In the Court’s first decision in this matter, the Court affirmed that Commerce’s interpretation of representative as meaning, “determined on the basis of market principles” was permissible. Husteel I, 31 CIT at ---, 491 F.Supp.2d at 1290. Yet, the Court held that Commerce failed to adequately explain and support with substantial evidence on the record its decision to exclude Respondents’ sales as being unrepresentative. Id. at 1293.
Commerce’s explanation for excluding these sales rested on two assumptions: (1) that domestic prices in an nonmarket economy are not determined on the basis of market principles; and (2) that foreign suppliers to nonmarket economies compete with domestically-set prices. Based on these assumptions, Commerce concluded that sales from a market-economy seller to a buyer located in a nonmarket economy “may very well not be at prices that reflect the fair value of the merchandise.” Issues & Decision Mem. for the Final Results of the Admin. Rev. on OCTG from Korea 8 (Dept’ Commerce Mar. 7, 2006).
The Court identified two problems with Commerce’s explanation. First, the Court questioned why Commerce treated Respondents’ sales as sales into a nonmarket economy. Husteel I, 31 CIT at ---, 491 F.Supp.2d at 1291. Respondents sold OCTG to independent trading companies located in Korea; the trading companies then resold the merchandise to buyers located in China. The price data Respondents wanted to use was from the sales between Respondents and the Korean trading companies. Though it was correct to refer to the sales as Chinese as a matter of characterization, Commerce failed to explain why it was applying presumptions about conditions of sale in nonmarket economies to an arm’s-length sale between two independent entities both operating in a market economy — facts which had been separately verified by Commerce.
The second problem the Court identified centered around Commerce’s assumption that foreign suppliers to nonmarket economies compete with domestically-set prices in the nonmarket economy, and therefore *1360sell merchandise at distorted, nonmarket prices. Commerce’s assumption in this case appeared to contradict the agency’s position in a related line of antidumping duty investigations, where the producer being investigated for dumping is itself located in a nonmarket economy (referred to here as “NME-Producer cases”). In NME-Producer cases, Commerce regularly accepts sales price data from a market-economy supplier to a nonmarket-economy buyer (the respondent in those cases) to calculate normal value, see 19 C.F.R. § 351.408(c)(1) (2007), and Commerce did not give a persuasive explanation for why Respondents here should be treated differently.
On remand, Commerce attempted to address the Court’s concerns. Regarding the question of why Commerce treated Respondents’ sales as though they were made to nonmarket buyers, Commerce explained that knowledge of the trading companies’ intent to resell the merchandise to buyers in China influenced the price at which Respondents sold the merchandise to the trading companies. “Because Plaintiffs had knowledge of the destination country, Plaintiffs ... priced the OCTG sold to the trading companies based on the conditions in the PRC.” (.Remand Results 17.)
Regarding the question of why Commerce did not accept Respondents’ Chinese sales data to calculate normal value when it regularly does so in NME-Producer cases, Commerce offered a two-fold response. The primary answer rehashed an argument that the Court rejected in the first decision: that the two situations are not analogous because different standards apply to the admissibility of evidence in the two types of cases. The second answer was that even if the two situations are analogous, Commerce does not accept sales price data if those sales might be distorted (i.e., dumped or subsidized). Commerce stated that it has reason to believe that Respondents’ merchandise may have been subsidized, with the implication being that Commerce need not use the price data from their Chinese sales to calculate normal value. (Remand Results 7.)
Commerce also presented evidence on two subjects to support the determination that Respondents’ sales where not “representative”: (1) data on average prices for Chinese OCTG as compared to average world prices for OCTG; and (2) analysis of the Chinese oil and gas industry, the sector of the Chinese economy that uses OCTG.
In the results to the remand, Commerce continued to conclude that Respondents’ sales were not representative, and therefore did not recalculate Respondents’ normal values or dumping margins. Respondents argue that Commerce did not comply with the Court’s remand order and that Commerce’s Remand Results are not supported by substantial evidence; Commerce and Defendant-Intervenors contend that the agency did so comply and that the Remand Results are in accordance with the Court’s order.
Standard of Review
The Court reviews Commerce’s Remand Results under the substantial evidence test. 19 U.S.C. § 1516a(b)(l)(B)(i) (2000) (“The court shall hold unlawful any determination, finding, or conclusion found ... to be unsupported by substantial evidence on the record, or otherwise not in accordance with law.”).
Discussion
I. Commerce may not presume that sales made from market-economy sellers to buyers located in nonmark-et-economies are made at distorted, nonmarket prices.
The first question raised by this case is whether Commerce can assume *1361that when a buyer located in a nonmarket economy purchases merchandise from a seller located in a market economy, the sales are made at distorted, nonmarket prices. The answer is that Commerce may not.
A. Commerce’s Presumption is Inconsistent With Its Practice in NME-Producer Cases.
As discussed in the Court’s first opinion, in NME-Producer cases, Commerce “regularly calculates normal value using price data from sales between market-economy sellers and nonmarket-economy buyers.” Husteel I, 31 CIT at, --- 491 F.Supp.2d. at 1293. Commerce regulations state that “where a factor is purchased from a market economy supplier and paid for in a market economy currency, the Secretary normally will use the price paid to the market economy supplier” to calculate normal value for the nonmarket-economy buyer (the respondent). 19 C.F.R. § 351.408(c)(1) (2007). The Court of Appeals for the Federal Circuit affirmed this practice. Shakeproof Assembly Components, Div. of Ill. Tool Works, Inc. v. United States, 268 F.3d 1376, 1382 (Fed.Cir.2001) (“accuracy, fairness, and predietability are enhanced by using” the sales price) (citation omitted).
Here, however, Commerce excluded similar data submitted by Respondents.2 Commerce argues that there are valid reasons to treat the sales price data submitted by Respondents differently than that submitted in NME-Producer cases. First, Commerce argues that data must meet a higher standard to be used in this case than is required to be used in NME-Producer cases. (Remand Results 9.) Second, Commerce argues that Respondents’ sales price data would be excluded under rules applying to NME-Producer cases, because Commerce suspected that Respondents’ prices might be distorted due to broadly available export subsidies in Korea. (Remand Results 7 n. 9, 28-29.) Each point will be refuted below.
1. “Representative” Data v. “Best Available Information”
The first point Commerce raises is that Respondents should be treated differently because, for their sales price data to be used, they need to be “representative,” while data in NME-Producer cases need only be the “best available information.” *1362(Remand Results 9; Def.’s Resp. 9.) Commerce explains that it is reasonable to apply different standards to the data because in NME-Producer cases the data are used to value “a single input used in the calculation of normal value,” whereas here the data are “used as the entire basis for normal value.”3 (Remand Results 9.)
Commerce is correct that the statute governing use of sales price data in the two types of cases use different words, “representative” in this case and “best available information” in NME-Producer cases. Compare 19 U.S.C. § 1677b(a)(l)(B)(ii)(I), with 19 U.S.C. § 1677b(c)(l). Yet, the argument misses the crux of the problem identified by the Court in the first opinion, which was that Commerce’s inconsistent treatment rests on two incompatible views of how the world works.
On the one hand, in the NME-Producer methodology, Commerce implicitly acknowledges that there are at least some instances where a buyer located in a nonmarket economy purchases merchandise from a market-economy seller at a market price. We know this to be true because Commerce may not use data that is distorted, and must verify that the purchases are “arm’s-length, bona fide sales.” (See Remand Results 9 (citing Shakeproof, 268 F.3d at 1382).) Accordingly, Commerce must acknowledge that in those instances, the nonmarket-economy buyer cannot, merely by virtue of being located in a nonmarket economy, dictate the price it will pay based on the domestic price of the merchandise inside the nonmarket economy. In fact, the domestic economic conditions are irrelevant to such a transaction. Instead, the nonmarket-economy buyer must pay the price the merchandise fetches on the world market.4
On the other hand, in this case Commerce presumed that the entire chain of transactions between Respondents, the trading companies, and the ultimate Chinese buyers was somehow tainted by the domestic economic conditions in China. As Commerce explained it, “sales of any product are directly impacted by conditions in the country into which it is sold. There is no viable domestic market for OCTG in Korea.” (Remand Results 2.) “As such, price conditions in the PRC impacted the prices that Plaintiffs ultimately negotiated for these sales.” (Id. at 5.)
Setting aside the problem of Commerce glossing over the fact that Respondents actually sold their merchandise to unrelated trading companies located in Korea-a market economy5 — Commerce never considers the possibility that this could be one of those situations where a nonmarket-economy buyer purchases the subject mer-*1363ehandise at the world, or market, price. Given this possibility, it does not make sense for Commerce to start from the presumption that the Chinese buyers dictated a distorted, nonmarket price to Respondents. As a result, the Court will not allow Commerce to exclude Respondents’ sales price data as unrepresentative without presenting evidence to support that conclusion.
In Section II of the Discussion, the Court evaluates the evidence cited by Commerce, and concludes that Commerce’s determination that Respondents’ sales are not representative is not supported by substantial evidence on the record. But first, the Court addresses a red herring argument made by Commerce regarding subsidies.
2. Commerce’s Suspicion that Korea Provides Export Subsidies is Not a Valid Reason to Exclude Respondents’ Sales.
A secondary reason that Commerce provides for not accepting Respondents’ sales price data is that they might be distorted due to the availability of subsidies in Korea. Sales will be excluded from the calculation of normal value in NME-Producer cases if Commerce has reason to believe the sales were subsidized. The agency states here that it has “found in other proceedings that South Korea maintains broadly available, non-industry-specifíc export subsidies and, therefore, it is reasonable to infer that all exports to all markets from South Korea may be subsidized.” (Remand Results 7 n. 9.) Although Commerce does not explicitly state as much, the implication is that it was justified in excluding the sales.
However, the Court has reason to doubt Commerce’s suspicion. If Commerce was truly concerned that the price of Respondents’ sales was distorted as a result of export subsidies, this concern would not be export-country specific. Based on Commerce’s description, South Korea’s export subsidy programs do not appear to be limited to certain export countries. (See id.) Under such a premise, Commerce would not be able to use price data for subsidized sales, regardless of export-country. See 19 U.S.C. § 1677b(b). Yet, Commerce used data from Respondent SeAH’s sales to Canada without noting a concern that those sales were subsidized. As a result, the Court wonders if subsidization is truly such a concern here.
Furthermore, as a factual matter, it is unclear to the Court that Respondents were eligible for export subsidies for the relevant sales because it was the unrelated trading companies who exported the OCTG to China, not Respondents. On the record as it stands, there is not sufficient evidence to affirm Commerce’s implied determination that Respondents’ sales were subsidized. The Court now turns to the evidence Commerce presented to support its determination that Respondents’ sales were not representative.
II. Commerce’s Evidence Does Not Support the Conclusion that Respondents’ Sales are Not Representative.
Commerce presents evidence on two subjects as support for its determination that Respondents’ sales are not representative: (1) average price data for imports of OCTG into China as compared to the rest of the world; and (2) analysis of the Chinese oil and gas industry, the sector of the economy that uses OCTG imports. Because of problems with the evidence presented, the Court holds that Commerce’s determination that Respondents’ sales are not representative is unsupported by substantial evidence on the record.
*1364A. The OCTG Price Data is Not Detailed Enough to Support Commerce’s Conclusion.
Commerce presents empirical evidence that the price of OCTG sold to China is less than the price of OCTG sold to the rest of the world. Specifically, Commerce presents two graphs, one comparing “average OCTG import prices into the PRC in 2003 and 2004 [the period of review spans those two years] and average OCTG import prices into the rest of the world for the same period.” (Remand Results 11.) In the second graph, Commerce compares “Korean OCTG export prices to the PRC and Korea’s export prices to the rest of the world” during the same period. (Id.) Commerce indicates that the prices in both comparisons are “significantly different” and concludes that “the PRC prices are not representative of the prices found in the rest of the world.” (Id. at 11, 33.)
The Court does not find there to be sufficient evidence contained in the Remand Results to evaluate Commerce’s conclusion. The primary problem with the data is that Commerce has included only average world price as a comparison to Chinese price. With a world average, the distribution of prices across countries is hidden. As a result, the Court cannot evaluate whether Chinese OCTG prices are “significantly different” than those in the rest of the world. For example, the distribution of country-prices might cluster around the Chinese price, but a single, large outlier would result in the average world price appearing “significantly different.” 6
The second issue with the data is that the mere fact that the prices are different is not in and of itself legally significant. Commerce does not indicate that the agency generally suspects the validity of a sale merely because the price for it differs from average world price. The difference in prices matters here because Commerce argues that it shows that the Chinese government is setting the price of OCTG inside China, and the distorted domestic price in China led Respondents to sell OCTG for export to China at distorted, nonmarket prices. However, Respondents raised a persuasive argument that the average price data relied on by Commerce actually undermines the agency’s claim.
Respondents claimed that “the data shows significant variation in [average prices] for all source countries for OCTG into the PRC, and this should not happen if, as Commerce alleges, prices for OCTG are set by the government.” (Remand Results 32 (Commerce summarizing Respondents’ argument).) Commerce responded that it “expects this type of variation to be found for all of the countries that collect this type of data.” (Id.) But why? If the nonmarket economy government sets the price, why would it differ from export country to export country? To explain the variation, Commerce relies on the very market considerations (volume of sales, specific grade of merchandise, etc.) that it argues are not present in sales to a nonmarket-economy buyer.7 Corn-*1365merce has not adequately dealt with the objection to the average price data raised by Respondents.
On remand, Commerce should (a) compare Chinese price to individual country prices of OCTG to determine whether the prices are “significantly different,” and (b) meaningfully address Respondents’ objections to the average price data.
B. The Evidence Regarding Government Control Over the Oil and Gas Industry in China is Irrelevant.
Commerce also presents evidence that the sector that uses OCTG in China, the oil and gas industry, is owned and controlled by the Chinese government. (Remand Results 10-11, Attach. 1.) Commerce believes that the Chinese government’s control of the oil and gas industry necessarily means that the prices for Respondents’ sales are not based on market conditions. The Court disagrees. The Chinese government’s control of the oil and gas industry, within China, does not inform on whether Respondents sold OCTG at representative prices. As Respondents explain:
Even assuming that the PRC government has complete control over the oil and gas sector in China and sets the prices for the sale of oil and gas in China, this begs the question as to how this PRC government control of the oil and gas sector within China impacts the pricing decisions of market economy companies outside of China making steel products.
(Pl.Br.21.) The relevant question is not whether the Chinese government controls the oil and gas industry within China, but whether its control over the oil and gas industry means that Chinese buyers of OCTG can dictate distorted, nonmarket prices to Respondents. And it is not enough to conclude that this is so; this is the issue on which Commerce must present evidence.
Because the evidence presented by Commerce does not show what Commerce purports it does, the Court cannot affirm Commerce’s determination that Respondents’ sales are not representative. If Commerce has persuasive evidence that the sales are not representative, it should be presented on remand. If Commerce does not, it may not exclude Respondents’ sales on that basis.8
Conclusion
Because Commerce’s Remand Results are not supported by substantial evidence on the record, the Court remands the Remand Results to Commerce for further consideration. Upon consideration of the papers submitted by all parties, and upon due deliberation, it is hereby
ORDERED that Respondents’ motion for oral argument is denied; and it is further
ORDERED that this case is remanded to Commerce for the agency to present persuasive evidence, if there is any, that Respondents’ sales for export to the China are not “representative” within the meaning of 19 U.S.C. § 1677b(a)(l)(B)(ii)(I) (2000); and it is further
*1366ORDERED that if Commerce cannot present persuasive evidence that Respondents’ sales are not representative, Commerce will determine that the sales are representative; and it is further
ORDERED that if Commerce determines that the sales are representative, Commerce will determine pursuant to 19 C.F.R. § 351.404(e) whether China or Canada should be selected as SeAH Steel Corp., Ltd.’s third-country comparison market; and it is further
ORDERED that if Commerce determines that China should be used as the third-country comparison market, Commerce will recalculate Respondents’ dumping margins accordingly; and it is further
ORDERED that the remand results shall be filed no later than September 2, 2008; that Respondents may file papers with the Court indicating whether they are satisfied or dissatisfied with the remand results no later than October 15, 2008; that Defendant and Defendant-Interve-nors may respond to Respondents’ comments no later than November 19, 2008; and that Respondents may reply to the responses no later than December 10, 2008.